EPA-R2-73-146
JANUARY 1973 Environmental Protection Technology Series
Pilot Plant for
Tertiary Treatment of
Wastewater with Ozone
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
<4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-R2-73-146
January 1973
PILOT PLANT FOR TERTIARY TREATMENT
OF WASTEWATER WITH OZONE
by
Dr. Clayton S. Wynn
Dr. Bradley S. Kirk
Dr. Ralph McNabney
Contract #14-12-597
Project #17020 DYC
Project Officer
Francis L. Evans III
U.S. Environmental Protection Agency
National Environmental Research Center
Cincinnati, Ohio 45268
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402
Price $2.flO domestic postpaid or $2,25 QPO Bookstore
-------
EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents
necessarily reflect the views and policies of the
Environmental Protection Agency, nor does mention
of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
Tertiary treatment of wastewater with ozone in a nominal
50,000 gal./day pilot plant at Blue Plains, Washington, D.C.,
is described. Plant feeds (about 10 to 100 ppm COD) were
effluents from other pilot processes involving nine different
biological and physical treatments of the Blue Plains waste-
water. Major COD reductions were realized, and product water
was practically free of bacteria and oxygen saturated.
The pilot plant used three major process steps: (1) genera-
tion of ozone gas from oxygen, including: preconditioning of
the gas feed and means of recirculating the gas; (2) dissolu-
tion of ozone from the oxygen carrier gas into the water; and
(3) retention of the ozonated water for a period sufficient
for the organic contaminants to be oxidized.
Plant performance for each feed is described in terms of COD
reduction characteristics and the effects of pH, ozone concen-
tration, feed pretreatment and initial COD on reaction rate.
Data are given for ozone solubility and half-life in pure
water and various wastewaters. Bacteria kills are reported.
Estimates of capital and operating costs are presented for
large plants to treat wastewater with ozone and a procedure
is given for optimization of costs for large plants.
This report was submitted by Airco, Inc., Murray Hill, New
Jersey, in fulfillment of Project Number 17020-DYC, Contract
14-12-597, under the sponsorship of the Office of Research
and Monitoring of the Environmental Protection Agency.
111
-------
CONTENTS
Section Page
I CONCLUSIONS 1
II RECOMMENDATIONS 5
III INTRODUCTION 7
IV PILOT PLANT AND EQUIPMENT DESCRIPTION 13
V PILOT PLANT OPERATION 21
VI SOURCES OF FEEDWATER 33
VII TECHNICAL RESULTS 43
VIII CHARACTERISTICS OF PILOT PLANT EQUIPMENT 103
IX ECONOMIC ANALYSIS OF LARGE PLANTS FOR 127
OZONE TREATMENT OF WASTEWATER
X ACKNOWLEDGMENTS 171
XI REFERENCES 175
XII PUBLICATIONS 177
XIII GLOSSARY 179
XIV APPENDICES 185
V
-------
FIGURES
Page
1 FLOW DIAGRAM OF PILOT PLANT 14
2 EQUIPMENT LAYOUT OF PILOT PLANT 15
3 POST-OZONATION APPARATUS I9
4 RELATION BETWEEN COD AND BOD IN OZONE 29
TREATED WATER
5 RELATION BETWEEN TOG AND COD IN OZONE 30
TREATED WATER
6 DISTRIBUTION CURVES FOR HIGH COD FEEDS 38
7 DISTRIBUTION CURVES FOR LOW COD FEEDS 39
8 DISTRIBUTION CURVES FOR 3B + L + F FEED 40
9 TYPICAL COD REDUCTION CURVES FOR MC AND UN 45
PRETREATED FEEDS: LOG COD vs REACTION TIME
10 EXTENDED COD REDUCTION CURVES FOR UN 47
PRETREATED FEEDS INCLUDING END-POINTS
FROM POST-OZONATION APPARATUS
11 COD REDUCTION vs LOG OF REACTOR NUMBER 48
PLUS ONE
12 CHANGES OF pH DURING OZONE TREATMENT OF 50
3B + L + F AND UN + L + F PRETREATED FEED
13 CHANGE OF pH DURING OZONE TREATMENT OF 51
MC + CL + C PRETREATED FEED
14 CHANGE OF pH DURING OZONE TREATMENT OF 52
MC PRETREATED FEED
15 CHANGE OF pH DURING OZONE TREATMENT OF 53
3B + F AND DC2 + L PRETREATED FEED
16 AVERAGE OF PILOT PLANT COD REDUCTION CURVES: 56
3B + L + F AND UN + L + F
17 AVERAGE OF PILOT PLANT COD REDUCTION CURVES: 57
MC + CL + C AND UN
VI
-------
FIGURES (Continued)
Page
18 AVERAGE OF PILOT PLANT COD REDUCTION 58
CURVES: MC
19 AVERAGE OF PILOT PLANT COD REDUCTION 59
CURVES: MC -I- CL AND 3B + F
20 AVERAGE OF PILOT PLANT COD REDUCTION 60
CURVES: DC2 AND DC2 + L
21 PERCENT COD REMAINING vs OZONE DOSAGE 68
22 EFFECTS OF pH AND REACTION TIME ON 70
CALCULATED SLOPE (b) OF DOSAGE CURVE:
3B + F
23 CORRELATION OF COD REMOVAL AND OZONE 73
DOSAGE FOR ALL FEEDS
24 RELATION BETWEEN PERCENT COD REDUCTION 75
AND LB OZONE DISSOLVED/LB COD FEED AND
LB OZONE DISSOLVED/LB COD REMOVED
25 COD IN MG/L vs LB OZONE DISSOLVED PER 76
1000 GAL WATER TREATED
26 PLOT OF LOG DISSOLVED OZONE (Z*) vs 80
LOG DOSAGE, D*
27 RELATION BETWEEN SMOOTHED VALUES OF 81
DOSAGE (D*) AND REACTION TIME (t*) FOR
Z* = 1 MG/L AND INDICATED pH
28 GRAPH OF REACTION RATE CORRELATION FOR 84
CONSTANT Z* = 1 MG/L
29 GRAPH OF REACTION RATE CORRELATION FOR 85
CONSTANT pH 7
30 OXIDATION OF NITROGENOUS COMPOUNDS BY 88
OZONE: UN
31 OXIDATION OF NITROGENOUS COMPOUNDS BY 89
OZONE: MC
32 OXIDATION OF NITROGENOUS COMPOUNDS BY 90
OZONE: 3B+F
VI1
-------
FIGURES (.Continued)
Page
33 OXIDATION OF NITROGENOUS COMPOUNDS BY 91
OZONE: DC2
34 TURBIDITY REDUCTION FOR MC AND UN 93
PRETREATED FEEDS
35 TURBIDITY REDUCTION DURING OZONE TREAT- 94
MENT OF DC2
36 AVERAGE EFFICIENCY OF PILOT PLANT OZONE 106
GENERATOR
37 SCHEMATIC OF PILOT PLANT MIXER ASSEMBLY 108
38 SOLUBILITIES OF OZONE IN PURE WATER 113
39 RELATION BETWEEN OZONE HALF-LIFE AND COD 114
REMAINING IN WATER
40 OZONE SOLUBILITY IN WATER vs RECIPROCAL 116
HALF-LIFE
41 PILOT PLANT MIXER TRANSFERENCE (T) vs 120
MECHANICAL EFFICIENCY (T/KW) FOR VARIOUS
WATER FLOW RATES
42 PILOT PLANT MIXER TRANSFERENCE (T) vs 122
GAS TO LIQUID VOLUME RATIO (G/L) FOR TWO
WATER FLOW RATES
43 CORRELATION OF PILOT PLANT MIXER PERFORM- 124
ANCE: T vs (T/KW) FOR VARIOUS GAS TO
LIQUID VOLUME RATIOS AND WATER FLOW RATES
44 FLOW DIAGRAM FOR 1 MGD WASTEWATER TREAT- 131
MENT PLANT
45 FLOW DIAGRAM OF 10 MGD WASTEWATER TREAT- 132
MENT PLANT
46 MULTILEVEL PLAN VIEW OF 10 MGD WASTEWATER 134
TREATMENT PLANT
47 CENTRIFUGAL PUMP CAPITAL COSTS 135
48 OZONE GENERATOR CAPITAL COSTS 136
VI11
-------
FIGURES (Continued)
Page
49 INJECTOR-MIXER CAPITAL COSTS 137
50 OXYGEN COMPRESSOR CAPITAL COSTS 138
51 GAS DRYER CAPITAL COSTS 139
52 CAPITAL COSTS OF INTEGRATED OXYGEN 140
RECYCLE SYSTEMS
53 TYPICAL COSTS OF MERCHANT LIQUID 146
OXYGEN AS A FUNCTION OF USAGE RATE
54 TOTAL OPERATING COSTS FOR OZONE 149
TREATMENT OF WASTEWATER AT DIFFERENT
OZONE FEED RATIOS
55 BREAKDOWN OF UNIT COST FOR TREATING 150
WASTEWATER WITH OZONE
56 ESTIMATED TOTAL PRODUCTION COST OF 154
OZONE IN A TYPICAL WASTEWATER
TREATMENT PLANT
57 EXPANDED GRAPH OF ESTIMATED TOTAL 155
PRODUCTION COST OF OZONE
58 PERCENT OF OZONE FEED DISSOLVED IN 159
TYPICAL MIXER
59 INJECTOR TESTING APPARATUS 186
IX
-------
TABLES
No.
Page
1 Range of Variables Examined in Ozone 11
Pilot Plant
2 Pretreatment of Wastewater Feed to Pilot 36
Plant • Descriptive Summary
3 Index of Feed Waters Treated in 36
Pilot Plant
4 Average COD Results from Post-Ozonation 55
Apparatus
5 Average Cumulative Dissolved Ozone Concen- 62
tration (Z*, mg/1) at the Outlet of
Reactors
6 Fit of 2-Component Model to Averages of 64
Raw Data
7 Analysis of COD Reduction vs D* (Ib Ozone 72
Dissolved/lb COD Feed)
8 Bacteriological Counts on Wastewater Samples 96
from Pilot Plant
9 Response to Change in Feed COD at Constant 101
Hydraulic Load of a Plant Capable of Dis-
solving 0.45 Ib Ozone/1000 gal Feedwater
10 Response to Changes in Hydraulic Load at 101
Constant Lf = 40 mg/1 of a Plant Capable
of Dissolving 0.45 Ib of Ozone per 1000 gal
Feedwater
11 Design Parameters for Initial Cost Estimates 130
for Large Ozone Wastewater Treatment Plants
12 Capital Cost Estimates for Plants to Treat 142
Wastewater with Ozone
13 Estimated Annual Operating Costs of Ozone 147
Wastewater Treatment Plants in Thousand
Dollars
14 Percentage Breakdown of Estimated Annual 148
Operating Costs of Ozone Wastewater Treat-
ment Plants
x
-------
TABLES (Continued)
No. Page
15 Percent of Ozone Feed Dissolved in Typical 160
Mixer
16 Values of Adjustable Costs, £/1000 gal, 166
for Example Ozone Wastewater Treatment
Plant
17 Optimum Parameters and Costs for Various 168
Plant Specifications
18 Summary of Pressure vs Flow Characteristics 190
of Injectors
19 Summary of Mass Transfer Tests 191
20 Plant Profile Data Sets 212
XI
-------
SECTION I
CONCLUSIONS
1. The Blue Plains wastewaters subjected to nine different
combinations of pretreatment were successfully treated in
the ozone tertiary treatment pilot plant. Average influent
COD values ranged from 13 to 99 mg/1.
2. For domestic wastewaters similar to that at Blue Plains,
the relations between feedwater COD, product COD, feedwater
pH, reaction time, dissolved ozone concentration, and the
quantity of ozone dissolved can be calculated from correla-
tions developed here.
3. Estimated capital and unit operating costs (including
maintenance and amortization) for large plants to reduce
COD from 40 to 20 mg/1 with ozone are:
Plant capacity, mgd 1 10 50
Fixed capital cost,
1000 dollars 500 2,400 9,300
Unit cost, cents/
1000 gal. 24 12 10
4. Ozone production costs, including gas recycling, vary
with plant scale and the concentration of ozone generated.
In a plant generating 7500 Ib ozone/day, estimates based
on pilot plant data (see Figure 56) show the combined
equipment amortization and power costs for producing ozone
to be 8 to 9£/lb in the range of 1 to 2-1/2 wt percent
ozone in oxygen. A shallow cost minimum is exhibited at
about 1.7 wt percent. Above about 2-1/2 wt percent, costs
increase rapidly with increasing ozone concentration.
1
-------
5. Relative ozone consumption, expressed as Ib ozone dis-
solved per Ib COD feed, increases exponentially with percent
COD removal.
6. For a specified percent COD removal, the relative ozone
consumption is practically independent of treatment time,
feedwater pH, and the exact nature of feedwater pretreatment.
It is simply related to the feedwater COD, as shown by
Equation 21 and Figure 31.
7. The rate of COD reduction is increased by increased
feedwater pH and by increased dissolved ozone concentration.
This relation is described by Equations 23 and 24.
8. There appear to be no technical limits on the COD
removal. However, reaction rates become very slow and
relative ozone consumptions very high after 50 to 70%
removal.
9- Typical ozone tertiary treatment involves handling gas
volumes of the same magnitude as the volume of water treated.
10. The gas-liquid mixers used in the pilot plant are most
effective at water flow rates greater than nominal capacity
and at gas to liquid volumetric ratios of about one. Their
performance is summarized by Figure 11.
11. Cocurrent gas-liquid mixers, as used in the pilot plant,
cannot dissolve most of the ozone from the gas stream and
simultaneously maintain high dissolved ozone concentrations.
-------
12. Improved mixer designs allowing 90% of the ozone to be
stripped from the gas, e.g., multistage mixers operated
stage-wise in gas-liquid counterflow, could reduce ozone
treatment costs perhaps 20%.
13. The solubility of ozone in wastewater is 10% to 30% less
than in pure water.
14. The half-life of dissolved ozone in wastewater shows a
dependence on the COD concentration (see Figure 39). For
wastewater subjected to extended ozone treatment, the half-
life is roughly 45 minutes.
15. Water subjected to ozone tertiary treatment is prac-
tically free of bacteria.
16. Whether acidic or basic, the pH of the wastewater
changes toward neutrality during ozone tertiary treatment.
17. Nitrogenous matter was not removed by ozone treatment.
At high pH, much of it was oxidized to nitrate.
18. Turbidity in the feedwater was reduced by ozone
treatment.
19. The pilot plant results confirmed the observations con-
cerning the benefits of ozone treatment of wastewater which
are outlined in the Introduction.
-------
SECTION II
RECOMMENDATIONS
1. Construct and operate a full-scale demonstration plant
for ozone tertiary treatment of wastewater using the
technology described in this report.
2. Concurrently, develop improved counter-flow gas-liquid
mixer systems which can simultaneously maintain high
dissolved ozone concentrations and dissolve a large
fraction of the ozone feed.
-------
SECTION III
INTRODUCTION
Increasing populations and improving standards of living are
placing increasing burdens on water resources. The main-
tenance or improvement in quality of our finite supplies of
fresh water and, in the not distant future, the necessity for
direct recycling of water in some parts of the country will
demand improved technology for the removal of contaminants
from wastewater.
The contaminants in wastewater are many and continually vary-
ing, and they are not well characterized according to chemical
species. Commonly, the level of contamination is described by
either a biological oxygen demand (BOD) or a chemical oxygen
demand (COD). BOD- is a measurement of the oxygen consumption
by aerobic microbes feeding on the contaminants during a
5-day incubation. COD is the oxygen equivalent of the oxida-
tion of the contaminants in hot dichromic acid. In general,
the COD analysis is quicker and more reproducible than the
BOD,, analysis. For these reasons, COD will be used in this
report to characterize the contaminants in wastewater.
Wastewater treatment is usually divided into three stages:
primary, the removal of settleable solids; secondary, the
removal of readily biodegradable contaminants; and tertiary
treatment. Tertiary treatment is generally the further treat-
ment of wastewater after prior treatment has reduced the COD
to less than about 60 mg/1 and the BOD- to less than about
20 mg/1. It may also include the removal of nitrogenous com-
pounds and phosphates.
Tertiary treatment processes include lime (or other chemical)
clarification, filtration, activated carbon adsorption, and
-------
ozone treatment. Ozone, 0_, is a very powerful oxidant which
can oxidize most of the oxidizable contaminants in wastewater.
It is, however, an unstable material which must be generated
at the point of use and in fairly low concentrations.
Ozone has been used for disinfecting drinking water in other
countries for many years. And, it has been used to treat
some special industrial wastes, notably for removing cyanides
and phenols. However, ozone has not been utilized for general
wastewater treatment on any large scale.
Several years ago, Airco ran experiments on ozone tertiary
treatment of secondary-treated wastewaters. The favorable
results led to a pre-pilot, bench-scale experimental program
sponsored by the Federal Water Pollution Control Adminis-
tration. (1)
On the basis of these laboratory investigations, it was
established that ozone tertiary treatment potentially offers
the following advantages:
(1) There is a major reduction in both BOD and COD.
The reduction is accomplished by destroying, not
simply separating the organic material; hence,
there are no problems of solids handling or con-
centrated wastes disposal.
(2) There is pronounced reduction of odor,, color,
turbidity and surfactants.
(3) All bacteria are killed.
(4) Except for oxygen, no residual chemical additives
remain in the product water.
(5) The product water has a high dissolved oxygen con-
centration (DO). Especially, when using pure
oxygen to prepare ozone, the DO is generally greater
-------
than the residual BOD or COD. The effect of such
effluent on receiving streams would be that of a
negative BOD.
Because of these potential advantages of the process, Airco,
in contract with EPA, designed, built and operated a 50,000
gal/day pilot plant for ozone tertiary treatment of waste-
water at Blue Plains, Washington, D.C.
The objectives of the pilot plant program were to:
1. Establish the effect of scale-up on:
a) The rate of ozone transfer from the gas phase
to the aqueous phase;
b) The efficiency of ozone in reducing the chemical
oxygen demand (COD);
c) The removal of organic pollutants from a variety
of pretreated waste streams.
2. Determine the effect on the ozone process caused by:
a) Hydraulic and quality changes in the incoming
waste stream;
b) Various clarification procedures used in pre-
treatment;
c) Variation in treatment time.
3. Obtain process data for the design of full-scale
plants.
4. Establish the economics of the process. Optimize
the important process variables, such as ozone feed
ratio to each stage of reaction, the number of
stages, process residence time, and effluent pre-
treatment.
This report contains a description of the pilot plant and
its operation, the experimental programs, technical results
-------
and their correlation, and an economic assessment of
large plants for the ozonation of wastewater. Feed waters
for the pilot plant were the products of nine different bio-
logical and physical treatments of Blue Plains wastewaters.
The tertiary treatment of wastewater with ozone involves three
major steps: (1) generation of ozone gas, including precon-
ditioning of the gas feed and means of recirculating the gas ;
(2) dissolution of ozone from the oxygen carrier gas into the
water; and (3) retention of the ozonated water for a period
sufficient for the organic contaminants to be oxidized.
The pilot plant contained six 440-gallon reactors arranged in
hydraulic series. At the top of each reactor there was a
special, injector-type mixer for dissolving ozone from an
ozone-oxygen gas stream. Spent gas from the reactors was
dried by adsorption on a desiccant under pressure and re-
cycled to the ozone generator. To maintain the desired dis-
solved ozone concentration in the reactors, the ozone feed
to the mixers was adjusted by varying the gas flow and ozone
generator power. The range of variables examined in the pilot
plant operation is given in Table 1.
Plant performance for each of the nine pretreated feeds is des-
cribed in terms of COD reduction characteristics. The effects
of pH, ozone concentration, feed pretreatment, and initial COD
on reaction rate are presented. Data are also included on
ammonia nitrogen reactions, turbidity reduction, ozone genera-
tor power consumption, and the solubility and half-life of
ozone in pure water and various wastewaters. Bacteriological
test results from ozonation of the nine different feeds are
tabulated. Estimates of capital and operating costs are pre-
sented for ozone wastewater treatment plants with design
capacities of 1, 10 and 50 mgd. A procedure for optimization
of costs for large plants is described and illustrated by an
example.
10
-------
Table !•
RANGE OF VARIABLES EXAMINED
IN OZONE PILOT PLANT
Variable Minim-urn Maximum
Feedwater:
Flow, gpm 14 56
COD, mg/1 7.0 149
pH 4.9 10.9
Temperature, °F 61 78
Total treatment time, min. 40 188
Lb ozone dissolved/lb COD feed 0.4 5.4
Ozone concentrations
Gas from generator, wt % 0.85 5.25
Outlet water from reactor, nil 9.1
mg/1
Number mixers operated 2 6
Mixer flow/feedwater flow 0.54 2.75
11
-------
SECTION IV
PILOT PLANT AND EQUIPMENT DESCRIPTION
Plant Location
The pilot plant is located at the Blue Plains wastewater
treatment plant in Washington, D.C., a large plant serving
the Washington metropolitan area. The wastewater to this
plant is nearly all domestic waste and has a fairly low
level of COD and BOD. An important advantage of this loca-
tion is the presence of a complex of pilot scale, advanced
wastewater treatment processes operated cooperatively by the
Environmental Protection Agency and the District of Columbia.
These facilities made possible a variety of different waste-
water feeds to the ozone treatment pilot plant.
Pilot Plant
A schematic flow diagram of the pilot plant is shown in
Figure 1, and a layout of the major equipment is shown in
Figure 2. Nominal design capacity of the plant is 35 gpm
(50,000 gal/day) of a wastewater feed having a COD level of
40-60 mg/liter. The nominal treatment time of wastewater
with ozone is about 1-1/3 hours. Pumps and control equip-
ment are sized to permit plant operations at rates up to
7 0 gpm.
Wastewater enters the plant through a neutralization tank
which has provisions for automatic adjustment of pH by a
carbon dioxide sparge system. Dissolved gases, mostly nitro-
gen, are removed in a spray type, vacuum deaerator operating
at about 28 inch Hg vacuum. The water then flows through a
series of six 440 gal reactors, each 20 ft high x 2 ft in
diameter, in which the actual ozone treatment takes place.
Finally, the treated water is detained in a 2000 gal holding
tank to permit the residual dissolved ozone to decompose
13
-------
DEAERATOR
/ OZONE
""^GENERATOR
• MAIN WATER STREAM
• INJECTOR WATER STREAM
=GAS STREAM
LEGEND
E OZONE INJECTOR DISPERSER
P PUMP
R REACTOR
FCR FLOW CONTROLLER RECORDER
Fl FLOW INDICATOR
LC LEVEL CONTROLLER
PCI PRESSURE CONTROLLER INDICATOR
PH-C PH-CONTROLLER
Figure 1. FLOW DIAGRAM OF PILOT PLANT
-------
i ' ' ' ' ' ' ' ,,,,,•- r-^T ,,,,,,::,,,,,, ^_^ ,,.,..
.'-"" ~^x T"
/ \ r ---- - - - --J
/ \ LONGITUDINAL CLARIFER
\9i« n ARIFFP ' OPERATED BY
1 ' by EPA ' ENVIRONOMENTAL PROTECTION AGENCY
/ ___,
^_^^ !
ir ""
11 I"""! AIR
LEGEND INSTRUMENT " " "^"^
PUMPS (P-i TO P-7) AIR COMPRESSOR
REACTORS (R-l TO R-6)
FLOOR AREA 40' * 40'
1 DRAINAGE TRENCH —
/ \ O f R-€jP-6( R-5)p-s( R-4jP-4( R-3)p-3( R-2J
/HOLDiNG\^7 ^/ovCyovIyovIyovIy
I TANK J
OZONE OXYGEN
GENERATOR COMPRESSOR
" — ""* (1 [ D^)
^ v _ ,^^
gui
LABORATORY ^ _J , ,
s
\
\
^ DRYER
>,
\
J m \-—-^ —h — "
T
R
E
N
C
H
>-J
O
I
•
1 1
1 1
NEUTRALIZATION
/^VTANK
\g5rXwCUUMn
© PUMP \j
( l' ^
/^^ ^-^
R-i VACUUM
Vll/ DEAERATOR
MOTOR
CONTROL
Figure 2. EQUIPMENT LAYOUT OF PILOT PLANT
15
-------
before it is. discharged from the plant. Part of the treated
wastewater from the holding tank was used as cooling water
for the ozone generator and as seal water for the recycle
gas compressor and deaerator vacuum pump.
Ozone is produced from oxygen by electrical discharge in a
commercial ozone generator having a nominal capacity of 60
pounds per day.
Pure oxygen, rather than air, is used as the feed gas to the
ozonator in order to realize greater power efficiency in the
generation. In addition, ozone generators require a very dry
gas for maximum efficiency and low maintenance. The use of
pure oxygen, rather than air, greatly reduces the volume of
gas that must be dried.
Although pure oxygen is not very expensive in large quanti-
ties, its cost, together with the limited concentrations of
ozone (l%-4% by wt) that can be produced efficiently by exist
ing types of ozone generators, makes recycling of the oxygen
an economic necessity. The deaeration of the wastewater
feed to the plant prevents the build-up of inert gases in
the gas cycle. In the pilot plant, the gas recycle stream
usually was 90-95 vol % oxygen, the remainder being nitrogen
and carbon dioxide roughly in the ratio of 2:1.
The gas cycle through the plant is as follows: Gas from the
ozone generator, consisting of an ozone in oxygen mixture, is
fed in parallel to the dissolving equipment of the six
reactors. The spent gas is collected from the reactors in
parallel, compressed to about 50 psig in a Nash water ring
compressor, and dried to a dew point of -60°F in an adsorp-
tion dryer which is automatically regenerated by low pres-
sure purge. The spent gas from the reactors contained some
16
-------
residual ozone, but most of it was lost through decomposi-
tion in the dryer and by dissolution in the seal water of
the compressor. Make-up oxygen from liquid storage tanks
was automatically fed to the suction side of the compressor
as required to maintain a constant suction pressure. The
consumption of make-up oxygen was measured by a positive
displacement gas meter.
Emergency gas vent lines and vents from the gas sampling
system passed through an electrically heated furnace operat-
ing at 320°C. This assured the decomposition of any residual
ozone before discharge of the gas to the atmosphere.
The operating characteristics of the major pieces of pilot
plant equipment are described below, and the equipment speci-
fications are outlined in Appendix 2.
Materials of Construction
The neutralizing tank, deaerator, and reactors are made of
mild steel. The inside surfaces of the reactors have a
hot-applied, coal-tar bitumastic lining, and the inside of
the vacuum deaerator tank is coated with a baked epoxy coat-
ing. Pipes carrying dry gas are aluminum (Spec. 6061-T-6),
and those carrying the wet, spent gas are type 316 stainless
steel. Water piping and mixers are of polyvinyl chloride
(PVC), Type I. All wetted parts of pumps and the compressor
are made of stainless steel (type 316). Pipe flanges are
sealed with Teflon enveloped asbestos gaskets, and fluori-
nated elastomers are used for shaft seals. Throughout the
test period, no corrosion of these components was evident.
Originally, the mild steel holding tank was lined with an
epoxy-bonded enamel. This coating was attacked very severely
by the ozone and, after two months use, the residual coating
17
-------
was removed by sandblasting and a solvent coal-tar bitumastic
coating was applied. This coating showed some evidence of
blistering in the gas space after six months of service.
Appendix 3 contains additional corrosion test results for
several common metals of potential usefulness in equipment
fabrication.
Post-Ozonation Apparatus
The apparatus shown in Figure 3 was used to gather additional
process data. Product water from the pilot plant is placed
in a 3-liter flask and subjected to an ozone sparge for an
additional 4-8 hours in order to carry the ozone treatment
nearly to completion. The gas sparge is then stopped; con-
temporaneously, the liquid and gas are analyzed for ozone
concentration, and a sample for COD analysis is taken. Sub-
sequently, several analyses of the dissolved ozone (DO.,) are
made over a 40 to 60 minute period.
The COD value affords a point on the reaction curve near the
reaction end point. From the contemporaneous gas and liquid
analyses, ozone solubility in the wastewater can be calcu-
lated. By plotting the log of dissolved ozone concentration
against time, as is shown in the bottom of Figure 3, the rate
of ozone decomposition in the wastewater (assuming first
order kinetics) can be obtained.
18
-------
OZONE
VENT
VENT
WET TEST METER
DO 3 SAMPLE
o>
ro
O
Q
GAS 1.66 VOL%03
I
I
0
10 20 30
TIME, MIN,
40 50
Figure 3. POST^OZONATIQN APPARATUS
19
-------
SECTION V
PILOT PLANT OPERATION
General
After a several week start-up period, routine pilot plant
operations on wastewater were started on February 1, 1971 and
continued to September 30, 1971 with only minor interruptions
(caused mostly by equipment failure and process upsets in the
upstream pretreatment processes). The ozone pilot plant was
normally operated twenty-four hours per day, five days per
week. It was staffed by three operators, one for each shift,
a chemical analyst, and a senior research engineer.
During most of the operations, wastewater was piped into the
neutralizing tank from the upstream pretreatment processes.
The desired plant feed flow was pumped from there into the
vacuum deaerator, and the excess wastewater overflowed from
the neutralizing tank to drain. The lime clarified secondary
effluent from the main Blue Plains plant was pumped into the
deaerator directly from the longitudinal clarifier system
location in the ozone pilot plant building (see Figure 2).
The available flows for several of the feeds examined were
less than the 35 gpm nominal capacity of the pilot plant. The
plant was operated both in a constant flow mode and in a vari-
able flow mode in which the flow approximated the diurnal flow
to the Blue Plains plant.
Adjustment of pH in the neutralizing tank was accomplished by
using carbon dioxide. A small quantity of air was continu-
ously sparged into the tank to keep it stirred. Carbon dioxide
was automatically introduced into the air stream as required to
maintain a constant pH.
The pilot plant incorporated only a few automatic controls;
21
-------
these were; feedwater flow rate, feedwater pH, water level in
the vacuum deaerator, make-up oxygen flow, and compressor out-
let pressure. All other plant parameters were controlled
manually. Water flow and gas flow to each mixer were indi-
cated by glass rotameters and were controlled by hand-operated
valves. Pressures of gas and liquid streams were measured
with bourdon gauges. The ozone feed was adjusted by manually
varying both the gas flow rate and electric power input to the
ozone generator. Coarse adjustment of ozone feed was made in
proportion to feedwater flow rate, and fine adjustment was
made to maintain the desired dissolved ozone concentration in
the overflow from each reactor. For the most part, manual
analyses of dissolved ozone were used as the control basis,
but as is mentioned in the discussion of analytical methods,
below, a "bread-board" continuous dissolved ozone analyzer
was used for this purpose for a period of several weeks.
The pilot plant responded very rapidly to changes in operating
parameters. It could be started up in less than 20 minutes
and shut down in less than 10 minutes. Because of the rapid
response to changes, composite sampling of the water and gas
streams was not necessary.
A number of parameteric studies were made on the various
pilot plant feeds. These include studies of the effects of
feedwater pH, treatment time, ozone dosage, number and
sequence of mixers operated, and ozone distribution to the
mixers. The nature of these studies is evident from the data
listings in Appendix 6. In many of these studies, advantage
was taken of the rapid plant response by "interleaving" the
parameteric changes. In examining the effects of several
values of pH, for example, instead of making several runs at
one pH, then several at the next pH, and so on, the pH was
changed every shift until four to six sets of data were ob-
22
-------
tained for each pH value. Thereby, the average data for
each of the several pH values similarly reflected any pos-
sible long term drifts in the feedwater characteristics.
Safety
Ozone treatment of wastewater involves the handling of two
potentially dangerous substances: ozone and pure oxygen.
Ozone is toxic, and oxygen can increase fire hazards. How-
ever, with moderate precautions, both materials can be han-
dled with complete safety. A copy of the safety rules and
procedures employed in the pilot plant is shown in Appendix
4. With these procedures, no accidents or hazardous situa-
tions associated with oxygen or ozone were experienced during
the more than nine months of pilot plant operation.
An oft quoted measure of the toxicity of gases is the Thresh-
old Limit Values (TLV) set by the American Conference of
Governmental Industrial Hygienists which "represents conditions
under which it is believed that nearly all workers may be
(2)
repeatedly exposed day after day without adverse effect".
The TLV of ozone in air is 0.1 parts per million by volume
(ppmv). However, ozone has a very strong, pungent odor.
The minimum concentration detectable by odor is only about
0.02 ppmv, and at the TLV concentration, the odor is very
strong and irritating. This makes it unlikely that anyone
would voluntarily remain in dangerously high ozone concen-
trations .
The degree of possible hazards from ozone can be placed in
perspective by comparison with chlorine, a material used
(3)
routinely in wastewater treatment. Chlorine is con-
sidered to be less toxic than ozone, the tentative TLV being
1 ppmv. However, the minimum concentration of chlorine
detectable by odor is about 3 ppmv. Thus, the TLV of ozone
23
-------
is much greater than the concentration detectable by odor,
while the converse is true for chlorine. Moreover, no large
quantities of ozone would ever be accumulated in a wastewater
treatment plant; being unstable, it is generated only as
needed, and then only in low concentrations. And in the event
of system leaks, ozone generators can be shut down instantane-
ously. Chlorine, on the other hand, is commonly transported
and stored in large quantities as liquid under considerable
pressure; any catastrophic leak in such a chlorine system can
release large quantities of concentrated chlorine, thereby
creating a major hazard. Thus, it appears that even though
ozone is more toxic than chlorine, in the practical applica-
tions to wastewater treatment ozone should be judged less
hazardous than chlorine.
It may be mentioned that high concentrations of ozone are very
unstable and can decompose with explosive violence. But
under no known circumstances can such high ozone concentra-
tions be produced in conventional electric ozone generators.
Pure oxygen is not toxic; twenty percent of normal air is
oxygen. But high concentrations of oxygen grossly enhance
the ease of ignition and flame temperature of any combustible
substance. High concentrations of oxygen cannot be detected
by the physical senses—odor, taste, or color—and for this
reason, possibly hazardous situations are not always immediate-
ly obvious. Thus, oxygen may present a more insidious hazard
than ozone. However, very large quantities of pure oxygen
are handled daily for industrial and medical purposes with
virtually no accidents. Proper safety procedures are simple
and well established; a minimum of combustible materials and
possible ignition sources are allowed in the vicinity of pure
oxygen systems, and the areas must be well ventilated.
24
-------
The ozone in oxygen treatment of wastewater does introduce
two particular, if somewhat remote, hazards. First, combus-
tible materials, e.g., oils and greases, must be removed
from the wastewater before ozone treatment. If there is any
practical possibility that such materials could pass through
the upstream treatment processes, reliable means for detec-
tion of combustibles must be incorporated at the inlet of
the ozone treatment plant with provisions for automatic
plant shut-down. The second possible hazard is more remote,
involving the high dissolved oxygen concentrations in the
product water. The partial pressure of oxygen in the product
water is about five times that in air. Oxygen will tend to
desorb from the water into the air. In the rare combination
of (a) a large body of undiluted, oxygen saturated product
water, (b) no wind, and (c) a strong atmospheric temperature
inversion, oxygen enriched air could accumulate and create a
flammability hazard. This rather unlikely hazard can be
completely removed by arranging the plant outfall so that the
product water is rapidly mixed with the receiving waters.
Data Collection
Wastewater sampling taps are located at the outlets of the
neutralizing tank and vacuum deaerator, in the water overflow
and recycle lines of each of the six reactors, and at the
outlet of the product holding tank. Samples were routinely
collected from these taps, and if the sample had been exposed
to ozone, any residual dissolved ozone was removed by an
inert gas sparge.
Gas sampling taps are located at the outlets of the ozone
generator, the compressor and the dryer, and in the gas
return lines from each reactor. These sampling taps were
connected by metal tubing to a valve manifold in the laboratory
where the various gas samples were analyzed for ozone concen-
tration.
25
-------
Plant operating parameters were monitored continously,
once each eight-hour shift a completed set of plant data was
obtained. These data were collected according to a schedule
based on the feedwater flow rate so that the data collection
followed the flow of a single slug of water through the treat-
ment process. Data collected included:
1. All gas and liquid flows and pressures.
2. The COD as the water slug leaves each reactor.
3. The ozone concentration in the gas from the ozone
generator and in the spent gas from each reactor.
4. The pH, temperature and turbidity of the feed and
product water.
5. The dissolved ozone concentration in the water in
each reactor.
6. The ozone generator power parameters.
7. Make-up oxygen meter readings.
The operating parameters were recorded on log sheets; the
analytical results were recorded in notebooks; and the per-
tinent information from these basic sources was then recorded
on "Reactor Profile Data Sheets" for later computer reduc-
tion of the data. A copy of the data form is shown in
Appendix 6, together with a summary listing of the 296 com-
plete sets of plant data collected during operations.
In addition to the complete plant data sets, other waste-
water samples were collected from the pilot plant at less
frequent intervals. These were subjected to additional
analyses which included bacteriological assay by an inde-
pendent commercial laboratory and analyses for BOD TOC
and total Kjeldahl, ammonia, nitrite, and nitrate nitrogen
concentrations by the EPA analytical group at Blue Plains
26
-------
Analytical Procedures
General. The analytical procedures used in the pilot plant
work are described in detail in Appendix 5. Most of the
analytical methods employed were either standard proce-
(4 5)
dures ' or slight modifications thereof.
COD Analyses. For reasons mentioned in the introduction and
in Appendix 5, COD was selected as the method of character-
izing the contaminants in wastewater. Since the investiga-
tion of COD reduction during ozone treatment was a primary
objective of this work, and because many of the COD levels
encountered were very low, much effort was devoted to obtain-
ing accurate and precise COD data.
As finally evolved, the low level COD analytical method
(4)
described by EPA was used with one modification. Instead
of adding mercuric sulfate to each reflux flask as the powder,
it was dissolved into the acidified standard potassium
dichromate solution beforehand. This was necessary because
traces of solid organic contaminants were found in many
batches of reagent grade mercuric sulfate. If added to each
analytical flask as the solid powder, each flask would receive
varying quantities of the heterogeneously distributed con-
taminant, a situation which could not be corrected from the
blank analysis run with each batch of COD samples. It was
also found that the quality of the distilled water was criti-
cal at low COD values. The use of bottled, pyrogen-free
distilled water enhanced the precision of the analyses.
The collection and handling of the COD samples required
great care. Only a tiny speck of organic dirt accidentally
introduced into the sample produces large errors in the very
low range COD analyses. All the samples were collected in
glass bottles which had been cleaned with chromic acid
27
-------
solution and well rinsed with, good distilled water. The
samples were stabilized with reagent grade sulfuric acid
during collection and were then immediately sparged with
helium gas to remove any residual dissolved ozone.
From the results of replicate analyses, and from the consist-
ency of the COD data obtained from the various reactors
(described below), the low range COD determinations (about 5
to 30 mg/1) are believed to be precise within about 1 mg/1.
BOD and TOC. During the latter part of the pilot plant
studies, a number of additional water samples were collected,
and these were submitted to the EPA analytical group at Blue
Plains for BOD , total organic carbon (TOC), and nitrogen
analyses. (The samples for BOD analysis were not acid stabil-
ized but were refrigerated, and the analyses were started on
the same day as collected.)
A comparison of the BOD5 analyses with the plant COD analyses
is shown in Figure 4. (See Table 2 for explanation of efflu-
ent abbreviations.) A similar comparison with the TOC analy-
ses is shown in Figure 5. The correlations are not very
good, a matter further discussed in Appendix 5.
Dissolved Ozone. The acidic KI method of determining dis-
solved ozone concentrations used in the prepilot work was
used unchanged during the first part of the pilot plant work.
But in late June, no feed to the plant was available for
several days, and a series of runs on tap water were begun
to gather additional data on mixer performance. With the
ozone consumption by chemical reaction no longer a masking
factor, it was found that an ozone mass balance around indi-
vidual reactors could not be completed. This eventually led
to the conclusion that the acidic KI procedure leads to
erratic and usually high results.
28
-------
ro
vn
50
40
o
UJ
i
UJ
30
I
u.
o
o
o
00
20
10
o
• •
0
O
0 o g
o
O
O
0
O
O
o —
0 10 20 30 40 50 60 70
COD IN WASTE WATER FEED, MG/L
PRETREATMENT-~3B+F MC UN DC2
STAGE (DIN OIN A IN O IN
OF
-------
OJ
o
50
40
30
0
oh-
o
8
°
0
O w A
O
o
o
.A A
® ©
o
A o
A AO A
• UNTREATED BIOLOGICAL EFFLUENTS
O OZONATED BIOLOGICAL EFFLUENTS"
AUNTREATED MC EFFLUENTS
AOZONATED MC EFFLUENTS
20 40
60 80 100
COD, MG/L
120
140 160
Figure 5. RELATION BETWEEN TOG AND COD IN OZONE
TREATED WATER
-------
The alkaline KI method described in Appendix 5 was adopted
on June 24, 1971 and used for all subsequent analyses. This
method was found to be in generally good agreement with the
more cumbersome sparging procedure described in Standard
Methods.(5)
The errors in the earlier dissolved ozone analyses do not
enter into calculations for ozone consumption since such
calculations are based on an ozone gas balance. But the
earlier erratic analyses were judged to be too suspect for
use in mass transfer and reaction rate correlations and,
therefore, (as described later) were not used for these
purposes.
Ozone in Gas. The analysis for ozone concentrations in
various gas streams was based on the standard potassium
iodide method without modification. A description of this
procedure is given in Appendix 5.
Continuous Dissolved Ozone Analyses. A simple dissolved
ozone analyzer was devised and used to a limited extent for
plant control. A sample wastewater stream containing dis-
solved ozone was passed through a special metering valve
which diverted a small fixed flow from the stream into a
sparge chamber. The diverted portion was sparged with a
large, measured flow of clean air. Ozone in the air stream
was determined by a commercial ozone detector. The meter
reading was proportional to the dissolved ozone concentra-
tion in the water sample.
This relatively simple device operated satisfactorily on well-
filtered plant feeds and demonstrated the feasibility of
continuously controlling ozone dosage on the basis of dis-
solved ozone levels. However, even small quantities of
31
-------
suspended matter in the water were found to repeatedly
plug the metering valve. We believe that a refined device of
this type, or probably any of the several available dissolved
chlorine analyzers using the iodometric methodology, could be
used in a large plant to continuously measure dissolved ozone
levels for control purposes.
32
-------
SECTION VI
SOURCES OF FEEDWATER
Available Pretreatments
The Blue Plains Water Pollution Control Plant, the location
of the ozone pilot plant, is a very large (225 Mgd) facility
serving the Washington, B.C. metropolitan area. Treatment
consists of primary clarification, activated sludge secondary
treatment, and chlorine disinfection. The wastewater is
nearly all domestic waste and, therefore, has a fairly low
BOD5 (about 130 mg/1) and COD (about 300 mg/1).
In addition to providing a source of fairly typical domestic
wastewater, a great virtue of the location is the presence of
a complex of pilot scale (50,000 to 200,000 gpd) advanced
wastewater treatment processes operated cooperatively by the
Environmental Protection Agency and the District of Columbia.
These facilities afforded a large variety of secondary and
tertiary pretreatments of the feedwater to the ozone pilot
plant. The pretreatments available are described below,
together with the abbreviations assigned to each.
Each of the EPA/DC pilot processes has been described else-
(7 8 9)
where. ' ' Since several of them can be operated in
various modes, only a brief description of the processes as
they were used to feed the ozone pilot plant will be given
here.
.3J3 - The three-stage biological process used three biologi-
cal processes for removing both carbonaceous and nitrogenous
materials. It consists of (1) a more or less conventional,
high rate, activated sludge plus clarification step for removal
of carbonaceous materials, followed by (2) an aerobic sludge
33
-------
plus clarification step for converting nitrogenous material
to nitrate, followed by (3) an anaerobic sludge plus clarifi-
cation step for reducing nitrates to nitrogen gas. In order
to supply carbonaceous nutrients for the denitrifying organisms,
carefully controlled quantities of methanol are added during
the third step.
UN - The Unox^ process is a Union Carbide multi-stage acti-
vated sludge process in which dissolved oxygen is supplied
from pure oxygen gas instead of air. High BOD reduction rates
are realized by operating with high concentrations of mixed
liquor biological solids. The pilot unit consisted of four
stages in series, each equipped with submerged turbine type
oxygen dissolution equipment.
MC - In the two-step mineral clarification process , raw
wastewater was first gravity clarified with comparatively
high concentrations of lime, and then it was further gravity
clarified with ferric salts in a second unit.
L_ - Lime clarification is also useful as a tertiary treatment
for removing both phosphates and suspended solids. In the
pilot unit, effluent from one of the secondary biological
processes was gravity clarified with comparatively small con-
centrations of lime. The clarified overflow was carbonated
with C02 to a specified pH (8 to 10.5), then reclarified.
F - The pilot filtration unit removed suspended solids by
gravity filtration through a dual media bed of granular
anthracite atop sand.
C - The carbon adsorption pilot unit consisted of five acti-
vated carbon columns operating in series, in which the carbon
was regularly back-washed and periodically replaced. Both
34
-------
residual suspended solids and dissolved organic matter were
removed.
CL - The break-point chlorination pilot process oxidized
dissolved ammonia in secondary effluents to nitrogen gas by
the carefully controlled addition of chlorine. The "break-
point" is reached when the ammonia concentration is reduced
to zero, free available chlorine is detected, and the total
residual chlorine is minimized. Good mixing, pH control,
and chlorine dosage control are required to minimize undesir-
able byproducts.
DC2 - In addition to the wastewaters pretreated in the pilot
processes, secondary effluent from the main plant was avail-
able for feed to the ozone pilot plant. A stream was pumped
from the secondary clarifiers before chlorine disinfection.
Feedwater Treated
Wastewaters subjected to nine combinations of pretreatment
were examined in the ozone pilot plant. A summary of these
nine pretreatments, together with the symbols used to
designate each one throughout this report, are listed in
Table 2.
The periods during which each feed was examined are listed
in Table 3. Also shown are the arithmetic average COD's of
each feed during the period. Feed COD's range from those
of a marginally treated secondary effluent to those so well
pretreated that ozone treatment is only a final polishing
and disinfection step. The period that each feed was
examined was determined by availability from upstream proc-
esses and the extent of parametric studies planned.
35
-------
Table 2.
PRETREATMENT OF WASTEWATER PEED TO PILOT PLANT
DESCRIPTIVE' SUMMARY
I. EPA/District of Columbia Pilot Plants
Process
Abbreviation
3B + F
3B + L + F
UN or UNOX
UN + L + F
MC
MC
+ CL
MC + CL + C
Processing Scheme
A 3-stage biological treatment consisting
of high rate activated sludge, aerobic
nitrification, and anaerobic denitrifica-
tion, plus dual media filtration.
3-stage biological treatment plus lime
clarification plus dual media filtration.
Oxygen activated sludge.
Oxygen activated sludge plus lime clarifi-
cation plus dual media filtration.
Lime and mineral clarified raw wastewater
Lime and mineral clarified waste water
plus break-point chlorination.
MC + CL plus activated carbon adsorption.
II. Blue Plains Plant Effluent
DC 2
DC2 + L
Secondary effluent from District of Colum-
bia Water Pollution Control Plant.
DC2 effluent plus lime clarification.
Table 3.
INDEX OF FEEDWATERS TREATED IN PILOT PLANT
Pretreatment
3B + F
3B + L + F
UN
UN + L + F
MC
MC + CL
MC + CL + C
DC2
DC2 + L
Average
COD
13.5
25.9
26.1
16.9
53.7
46.6
14.7
98.9
72.4
Date
8/27-9/12/71
2/1-5/7/71
8/2-8/18/71
5/10-5/20/71
7/12-7/30/71
8/20-8/27/71
5/21-7/9/71
9/13-9/23/71
9/24-9/30/71
36
-------
Feedwater Characteristics
COD frequency distribution charts are shown in Figures 6, 7
and 8 for the nine feeds to the ozone pilot plant.
The points for the moderate and high COD feeds shown in
Figure 6 (log COD vs probit) generally fall along a straight
line showing that the percentage variation in COD about the
mean is normally (Gaussean) distributed. For the lower COD
feeds shown in Figure 7 (COD vs probit), the straight line
fall of the data indicates that the COD values (rather than
percentage deviation) are normally distributed about the
mean. On can easily suspect, however, that this reflects
the normal error in COD analyses at the low COD levels, and
that the percentage variation is normally distributed for all
of the feeds.
The 3B + L + F feed is the exceptional case, as is shown in
Figure 8. About half the data show a normal distribution at
a low COD, typical of the other feeds subjected to 3B pre-
treatment. The other half, however, show a non-normal distri-
bution at a much higher COD level. This was due to excess
methanol resulting from poor control of the methanol addition
to the denitrification stage of the 3B process during the
early period of its operation. This also accounts for the
fact that the average COD shown for this feed in Table 3 is
higher than that for the less thoroughly pretreated 3B + F
feed. The 3B + L + F feed exhibited one other anomaly. On
at least two occasions, some substance appeared which liber-
ated iodine from acidic KI solution. The exact identity of
this substance is not certainly known, but a nitrosamine is
suspected. In any event, it appeared to be destroyed in the
first stage of ozone treatment.
The suspended solids content (as reflected by turbidity) of
37
-------
200
8
12 5 10 20 30 50 70 80 90 95 98 99
PERCENT OF RUNS WITH FEED COD'S
LESS THAN OR EQUAL TO ORDiNATE
Figure 6, DISTRIBUTION CURVES FOR HIGH COD FEEDS
38
-------
! 2 5 10 20 30 50 70 80 90 95 98 99
PERCENT OF RUNS WITH FEED COD'S
LESS THAN OR EQUAL TO ORDINATE
Figure 7. DISTRIBUTION CURVES FOR LOW COD FEEDS
39
-------
100
80
60
40
20
o
8
10
8
6
4
12 5 10 20 30 50 7080 90 95 9899
PERCENT OF RUNS WITH FEED COD'S
LESS THAN OR EQUAL TO ORDINATE
Figure 8. DISTRIBUTION CURVES FOR 3B + L + F FEED
40
-------
the feeds varied according to the degree of pretreatment.
Feed turbidities are discussed below in sections dealing with
turbidity reduction.
The MC, MC + CL, DC2, and DC2 + L feeds contained a quantity
of surfactants sufficient to cause minor frothing problems
in the first stage of the ozone pilot plant. However, the
ozone dose in the first stage alone was enough to eliminate
the frothing problem in subsequent stages.
The feeds pretreated by the 3B and CL processes (3B + F,
3B + L + F, MC + CL, MC + CL + C) contained very little
ammonia nitrogen. The other feeds contained ammonia in
varying degrees. The latter are described in the discus-
sion of nitrification by ozone treatment, below.
The only modification of the feeds to the ozone pilot plant
was pH adjustment, generally using only carbon dioxide to
lower the pH. For the MC + CL + C feed, however, it was
desired to operate at pH values higher than that of the feed.
In this one case, the pH was raised to a preset value by a
constant rate addition of sodium hydroxide solution combined
with automatically controlled carbon dioxide sparging.
A major objective in treating the MC + CL pretreated feed
was to determine the removal of residual chloramines and the
effects of available chlorine on the ozone process. The data
obtained were not as reliable as those for the other feeds.
First, the upstream CL pilot equipment began malfunctioning
just about as soon as the feed was introduced into the ozone
plant. Second, attempts to analyze for chloramines and avail-
able chlorine in the presence of dissolved ozone produced
erratic results. Analyses were made with KI on samples
before and after sparging with helium to strip out the ozone.
41
-------
Often more iodine was liberated from KI after the sparge
than before. After about five days, treatment of this
feed was abandoned.
42
-------
SECTION VII
TECHNICAL RESULTS
Data Handling
During the period of February through September, 296 complete
sets of plant data were obtained for nine feeds divided into
a total of 23 groups according to feed and pH. Each complete
set of plant data contains 83 recorded quantities (some null).
Aside from a thousand or so other data such as BOD, TOC, and
nitrogen analyses, the raw technical results from the plant
data sets alone are comprised of more than 21,000 individual
data.
The analysis and correlation of such a huge quantity of data
is feasible only with the aid of an automatic digital com-
puter. An IBM 1130 computer with 8K core memory and magnetic
disk file, located at Airco's Murray Hill laboratories,
was used.
As described earlier, and in Appendix 6, each complete set of
plant data was entered onto a special data sheet by the plant
operators. At the end of each week, the data sheets were
sent to Murray Hill where the data were punched onto cards.
Each data set was then processed by the computer program to
yield three or four pages of reduced data for each data set.
These weekly analyses of data served to guide further opera-
tions of the pilot plant. Also, they were the basis of a
preliminary report of this work published elsewhere.
Upon completion of the plant operations in September 1971,
all the data sheets and punched cards were carefully checked
for accuracy. The results of all COD, ozone concentration,
pH, and turbidity analyses had also been recorded in
laboratory notebooks which served as a check for the data
43
-------
sheets. And all plant operating parameters could be checked
against the independent record in the plant operating log.
Occasionally, some data would be missing from a data sheet.
For example, the facilities for COD analyses were limited to
t
two batches of 10 samples per day, including blanks. For
this reason, on two of the three data sets normally collected
each day, COD analyses were not made on samples from all six
reactors. However, the data sheets were encoded to indicate
which reactors were not sampled, and the computer program
interpolated values for these missing COD's using a procedure
described in the next section.
If ozone concentrations in the vent gas or outlet water from
any reactor were not measured due to mistake or malfunction,
values were estimated from similar and contemporaneous runs.
That such values were estimated and not measured was encoded
into the data set so that lower statistical weight could be
assigned during correlation.
After being carefully checked and the necessary special
encoding performed, the complete group of data sets were
stored on a magnetic disk file for easy and rapid access
by the computer.
All data reduction and correlation programs (about twenty
were used) were written in Fortran compiler language. The
results of each program were verified by at least one manual
computation or by comparison with results of previously veri-
fied programs.
COD Reduction Characteristics
Typical COD reduction curves for two plant feeds are shown
in Figure 9 in which log COD is plotted vs reaction time
44
-------
60
40
MC
O>
£
Figure 9. TYPICAL COD REDUCTION CURVES FOR MC
AND UN PRETREATED FEEDS: LOG COD vs
REACTION TIME
45
-------
with ozone. Two characteristics are evident. The initial
reaction is quite rapid, but as the reaction continues, it
slows down so drastically that one might suspect that it has
neared completion.
That such is not the case is illustrated in Figure 10 in
which log COD (normalized to feed COD) is plotted against
log reaction time. These data are for the UN pretreated feed
and include end-points from the post-ozonation apparatus-,
described above. Data for the other feeds yield similar
curves. The convex shape of these curves indicates^ that
practically all of the COD can eventually be oxidized by
ozone. Thus, in the technical sense, there appears to be no
.-.
fraction of the COD that is refractory to ozone treatment.
However, the reaction becomes so very slow after 50%-70:%! COD
reduction that the economic practicality of further reduc-
tion becomes questionable.
There is another interesting, purely empirical way of plot-
ting the pilot plant COD data that is useful because it
linearizes the data. In Figure 11 is shown a plot of COD at
the outlet of each reactor against the log of the reactor
number plus 1.0. These typical data are arbitrarily selected
runs on several feeds which cover a wide COD range. The gen-
eral linear relation is particularly useful for interpolating
COD values for reactors which were not sampled for COD in
collecting a plant data set. Such a procedure has been used
in the data reduction programs.
It must be pointed out, however, that the linearizing scheme
in Figure 11 does not, in general, apply when the feed COD
points are included in the plot. Thus, the COD from reacto^
1 cannot be accurately interpolated from the feed COD and
that from reactor 2.
46
-------
100
80
60
40
o
UJ
cc
o
o
o
o
LJ
UJ
U_
I-
I 2<>
UJ
Q.
UN
10
5 10 50 100 500 1000
REACTION TIME, MIN
Figure 10,
EXTENDED COD REDUCTION CURVES FOR UN
PRETREATED FEEDS INCLUDING END-POINTS
FROM POST-OZONATION APPARATUS
47
-------
30
20
o
o
o
10
o
2
3
STAGE
456
Figure 11. COD REDUCTION vs LOG OF REACTOR
NUMBER PLUS ONE
48
-------
The typical data shown in Figure 11 also qualitatively
illustrate the precision and internal consistency of the
entire procedure of collecting, handling, and analyzing COD
samples. On this sort of linear plot, the raw COD data
generally fall within about one mg/1 of the average line.
It is also evident from Figure 11 that even with the compara-
tively high precision in the raw COD data, any attempt to
determine COD reduction rates by graphical or numerical dif-
ferentiation of the raw data would produce badly scattered
results. This fact must be considered in selecting methods
of correlation.
In the initial computer reduction of each data set, rate
data were desired. On the basis of plots similar to Figure 11,
the raw COD data were analytically smoothed by a least-square
fit to Eq (1) in which a7 may be zero.
LN+1 = ao + aiLn(N+1> + a2/Ln(N+l) (1)
where L = COD, mg/1
N = stage number
The nonlinear term is included to permit proper fit to the
feed COD [1^ in Eq (1) ] .
pH Changes
The pH of the wastewater is an obvious process parameter, and
part of the experimental program was devoted to examining its
effects.
It was discovered, however, that the pH consistently changes
toward neutrality during ozone treatment. In Figures 12, 13,
14, and 15 are shown the pH changes from feedwater to product
during many runs on several different plant feeds. In these
figures , the circular symbols representing the feedwater pH
49
-------
01
o
PH
10
9
8
7
c
—
3B+L+F
__
0
10 ^9
Il'll
<•
i i
A
<•
0
r
1
1
i
o FEED
A? PRODUCT -
p
t
i
i
D
1
UN+L + F
J"K.
c^
1
9
i
i
' 1
i
^v ••»•
Q A
t I
1 >
o t
f A 4
a T
At i
r
Figure 12„ CHANGES OF pH DURING OZONE TREATMENT OF
3B + L + F AND UN + L + F PRETREATED FEED
-------
10
Ul
I-1
8
x
CL
o FEED-
i i i i i i
MC+CL+C
1 1 1 I 1
A? PRODUCT
c
— —
- 1
6
i ,
-ll
)
C
-------
Ul
to
10
PH
8
• FEED MC
A7 PRODUCT
I
F
1
iUt
;t
7 7
,1"
DAYS
Figure 14. CHANGE OF pH DURING OZONE TREATMENT OF
MC PRETREATED FEED
-------
ui
00
8
1
_
T
A o
— •
A
A
(
O
1
1 1 1
38-hF
A r> A
w \j a
? 0 ^
O A ^
o
<
>
1 1 1 1
1 1 1 I 1 1 i 1
0 FEED 9
AV PRODUCT o T
e.
:
'U
(
1
I?
1 T7
o| 1
V V 1
^8 A9 X
v 0 X V AO
o v go
DC2+L
1 1 1 1 1 1 1 1
DAYS
Figure 15. CHANGE OF pH DURING OZONE TREATJYIENT OF
3B + F AND DC2 + L PRETREATED FEED
-------
are connected by vertical lines to the triangular symbols
representing the product water pH. Whether the feedwater is
acidic or basic, it can be seen that the product water is almost
always nearer neutrality.
While no firm theoretical explanation for the shift of pH
toward neutrality has been demonstrated, the phenomenon was
consistently observed in the ozone pilot plant, and it has
also been noted by others.(11) Limited tests showed that
most of the pH change takes place in the initial stages of
ozone treatment. And qualitatively, the greater pH changes
seem to be associated with the higher COD feeds.
The ozone pilot plant had no provisions for controlling pH
throughout the reaction. Therefore, in examining the effects
of pH on ozone treatment of wastewater, only the feedwater pH
could be used as an independent variable. Throughout this
report, pH of a water treated in ozone pilot plant will mean
the pH of the water as it entered the first reactor unless
explicitly stated otherwise.
Observed COD Reductions
The feed and product COD's observed in all of the complete
plant data sets are tabulated in Appendix 6. Feedwater flow
rates and pH are also included.
COD data from the post-ozonation apparatus are summarized in
Table 4. The total treatment time entries include the ozone
treatment time both in the pilot plant and in the post-ozona-
tiori apparatus.
Raw Plant Data. It is impractical to show graphs of all the
raw COD data and the pH of each set. In order to summarize
the data, the data sets for each feed were grouped according
54
-------
Table
4.
AVERAGE COD RESULTS FROM
POST-OZONATION APPARATUS
Feed
UN
MC+CL
3B+F
DC2
DC2+L
No.
Runs
12
4
7
10
3
*Antilog of
Avg
COD
mg/1
4.8
8.3
2.6
12.7
13.3
average
Avg* Avg Total
% Treatment
Reduction Time, hr
82.4
83.0
80.8
88.8
83.3
of log
9.17
8.43
8.50
7 = 65
6.93
percent reduction
Avg
End
PH
6.3
6.4
6.5
7.0
7.2
to pH. These groups were initially selected as integral pH
values plus or minus 0.5 pH. The pH values for all data sets
within the group were then arithmetically averaged to obtain
the average pH for the group.
For each pH group, the COD's corresponding to several selected
even values of ozone treatment time were interpolated from the
raw COD vs time data using the linearizing scheme described in
the discussion of Figure 11. For each of the even times, the
interpolated COD's were normalized to feed COD, and all of
the resulting values were averaged.
i
The averages of the COD data interpolated to a common time
basis are shown by pH groups and feed pretreatment in
Figure 16 through Figure 20. The averaged data fall smoothly,
and they all have the same general shape as do the typical
individual data sets in Figure 9- In general, the higher pH
feeds exhibit more rapid COD reductions. However, one must
be careful in drawing quantitative conclusions from these
data.
The average representations of the raw data in Figure 16
55
-------
IUU
PERCENT COD REMAINING PERCENT COD REMAINING
4> CT> 00 O •* OJ 00
O 0 OO-O O 0
\0^
\ *
X,
\
D^^
v
^
<
3 20 4
REAC
K
\\
\
\
v
—
^
\
A
\
:
^•^•^
V '
>v
<
i ~°~~
i
!^^c-
>^ ^^
"^5-
1
3B+ L+ F
D
o"^
-. 0 ^
o--.
pH -
-D^9
-^.8.1
^73"
•
i
UN + L+F
20 40 60 80
REACTION TIME, MINUTES
100
120
Figure 16.
AVERAGE OF PILOT PLANT COD REDUCTION
CURVES: 3B + L + F AND UN + L + F
56
-------
100
100
o
^ 80
UJ
I60
LJ
O
cr
UJ
Q_
40
20 40 60 80 100
REACTION TIME, MINUTES
120
UNOX
PH
6.0
20 40 60 80 100
REACTION TIME, MINUTES
120
Figure 17: AVERAGE OP PILOT PLANT COD REDUCTION CURVES
MC + CL + C AND UN
57
-------
PERCENT COD REMAINING _ PERCENT COD REMAINING
J> 0> 00 O J> ^ S £
o oooo o o c
V
\
\N^
*
x^X
\
N
A\^
< -v.
X
k. P
7
>v
>>7
MC
H
.1
'.2
D 20 40 60 80 100 120
REACTION TIME, MINUTES
\
\
\N
X
^
MC
Xc^ foHo
^Xp Q
0
20 40 60 80
REACTION TIME, MINUTES
100
120
Figure 18.
AVERAGE OF PILOT PLANT COD REDUCTION
CURVES: MC
58
-------
100
CD
z 80
z
UJ
QL
Q
O
O
60
UJ
o
cc
UJ
a.
40
0
MC+CL
.7.2
20
40 60 80
100
120
PERCENT COD REMAINING
J> 0> CO C
O O O C
V
\\
\\
<
\
X
* \
>. \ ^
v^
^Q
^^
XA
"o, ^
)
\
6T9
3B+F
PH
^ 5.2
'^^t^[ r
V
0 20 40 60 80
REACTION TIME, MINUTES
100^ S20
Figure 19,
AVERAGE OF PILOT PLANT COD REDUCTION
CURVES: MC + CL AND 3B + F
59
-------
IUU
*r\ Q^\
O 8O
2
2
PERCENT COD REMAINING PERCENT COD REMA
* 8 § 8 „ S § ,
X
N
X.
X
) 20 4
REAC
V
V
\
X
*
>ax
X
^^
DC 2
\
^
PH
7.0
K) 60 80 100 12
mON TIME, MINUTES
^
'v ^
\,
X,
>\
"^
DC2+L
V
PH
iO
^«
1
0 20 40 60 80 100 120
REACTION TIME, MINUTES
Figure 20. AVERAGE OF PILOT PLANT COD REDUCTION CURVES
DC2 and DC2 + L
60
-------
through Figure 20 must be taken as nothing more than that.
The effect of another very important variable is reflected
in these averages, namely, the ozone dosage. Ozone dosage
varied from run to run, and from feed to feed. Such varia-
tion does not affect the general shape of the curves as is
illustrated by the smoothness of the average data. But it
does affect the slope of the curves.
One measure of the ozone dosage is the time weighted average
of the dissolved ozone concentration at the outlet of the
reactors, Z*. This variable will be used in the final corre-
lations and is defined by Eq (2).
N N
Z * = (£Z t, )/(Et. ) (2)
IN 1 1 1
where Z* is evaluated for the outlet
of the N-th stage
Z. is the measured dissolved
ozone concentration at the
outlet of the i-th stage
t. is the treatment time in the
i-th stage.
For the case of equivolume reactors as in the pilot plant,
Eq (2) reduces to Eq (3).
N
Z * = (ZZ.)/N (3)
1
Values of Z* linearly interpolated to a common time base are
listed by pH group and feed in Table 5. Even with purely
manual control in the pilot plant, it can be seen that the Z*
values were held reasonably constant for each pH group. How-
ever the values of Z* vary significantly between pH groups
and feeds. It is this group to group variation in Z* that is
reflected in the averaged raw COD data.
61
-------
Feed
3B+L+F
UN+L+F
MC+CL+C
UN
MC
MC+CL
3B+F
DG2
DC2+L
Table 5.
AVERAGE CUMULATIVE DISSOLVED
OZONE CONCENTRATION (.Z*, mg/1]
AT THE OUTLET OF REACTORS
Avg
pH
5.
7.
8.
9.
6.
8.
6.
7.
8.
9.
6.
6.
7.
7.
8.
10-
7.
5.
6.
6.
7-
7.
8.
5
3
1
1
7
9
1
0
2
0
0
1
2
9
9
0
2
2
2
9
0
2
6
CUMULATIVE TREATMENT TIME, MIN.
20
2.
1.
1.
0.
2.
2.
3.
3.
-
—
1.
1.
1 .
0.
0.
0.
1.
2.
4.
2.
0.
0.
0.
26
68
41
94
89
18
85
65
70
12
01
97
91
48
61
93
50
59
47
92
36
_4
3.
2.
2.
0.
3.
2.
3.
3 .
2.
2.
1.
1.
1.
1 „
1.
0.
1.
3.
4.
3.
0.
1.
0.
0
03
19
06
89
01
71
20
14
99
34
63
28
21
14
05
75
61
31
79
05
58
36
80
3
2
2
0
2
2
2
2
2
2
1
1
1
1
1
0
1
3
4
3
0
1
1
60
.37
.34
.28
.99
.94
.89
.80
.98
.77
.33
.91
.45
.43
.44
.30
.99
.66
.38
.92
.25
.67
.75
.10
90
3.
2.
2.
1.
_
-
2.
2.
2.
2.
-
_
_
_
_
-
-
„
5.
-
0.
2.
1.
45
46
45
22
73
90
58
37
30
85
05
35
120
3.49
2.26
2.46
1.16
_
-
2.46
2.68
2.35
2*34
-
_
_
_
-
-
_
_
-
-
2.08
1.52
62
-------
"Two Component" Model. The plant data furnish COD reduc-
tion values for the first one to two hours of ozone treatment.
The POA data furnish an additional point for 7-9 hours treat-
ment. In order to compare these data separated by more than
a four to one time ratio, some analytic interpolating func-
tion is needed.
An inspection of the data in Figures 16 through 20 suggested
that they might be fit by a "two-component" model. The
contaminants in the wastewater were assumed to consist of
only two components, each exhibiting a different first order
reaction rate during ozone treatment. Component 1 was
assumed to comprise a fraction x, of the contaminants and to
have a first order reaction rate constant k-, . Similarly, the
(1-x,) fraction of component 2 had a rate constant k~• For a
first order reaction Eq (4) applies.
d Ln(L/Lf)/dt = -k (4)
For the "two-component" model, Eq (4) can be integrated as
follows.
L/Lf = x1e~klt + (l-x1)e~k2t (5)
Implicit here is the assumption that the multistage pilot
plant can be approximated as a continuous plug flow reactor.
By a somewhat involved iterative least-square procedure, the
values of x,, k, and k- were adjusted to fit the average raw
COD data for each feed and each pH group. The values so
obtained are listed in Table 6. The solid curves shown in
Figures 16 through 20 are the least-squares fit of the "two-
component" model to the data. From the figures and from the
standard deviations, s, listed in Table 6, it is seen that
this simple model fits the average data (L/L_) within about a
percent. (See Appendix 7 for a discussion of s.)
63
-------
Table 6-.
FIT OF 2-COMPONENT MODEL
TO AVERAGES OF RAW DATA
= x-'V + (l-x)e~k2t
Feed
3B+L+F
II
II
UN+L+F
II
MC+CL+C
ii
ii
ii
MC
II
II
II
II
UN
H
MC+CL
II
3B+F
II
II
II
DC 2
ii
DC2+L
H
II
pH
5.5
7.3
8.1
9.1
6.7
8.9
6.1
7.0
8.2
9.0
6.1
7.2
7.9
8.9
10.0
6.0
(6.0)
7.2
(7.2)
5.2
(5.2)
6.2
6.9
7.0
(7.0)
7.2
(7.2)
8.6
Xl
.2631
.1626
.3095
.4466
.1595
.3147
.2730
.2904
.3500
.4458
.3314
.4408
.3157
.3492
.3438
.2862
(.3408)
.2539
(.2213)
.4767
(.3722)
.2892
.4304
.2272
(.3415)
.2892
(.3207)
.3617
kl
2.010
4.764
2.760
1.452
4.601
6.422
3.268
4.224
2.936
3.705
3.449
3.231
6.454
4.594
5.054
7.224
(5.447)
3.385
(3.846)
2.354
(2.891)
5.609
4.342
1.704
1.354
2.362
(2.287)
2.380
(i-x^
.7368
.8373
.6904
.5533
.8404
.6852
.7269
.7095
.6499
.5541
.6685
.5591
.6842
.6507
.6561
.7137
(.6591)
.7460
(.7786)
.5232
(.6277)
.7107
.5695
.7727
(.6584)
.7107
(.6792)
.6382
k2
.00698
.3243
.1623
.1754
.3197
.3776
.1100
.1237
.09832
.07592
.2632
.1785
.4560
.3646
.3304
.2290
(.1448)
.1403
(.1800)
.001465
(.1387)
.4062
.2701
.3032
(.2314)
.2277
(.2025)
.2680
s
.018
.012
.010
.008
.010
.019
.004
.004
.007
.015
.012
.010
-Oil
.012
.003
.007
(.012)
.003
(.006)
.023
(.023)
.016
.008
.009
(.007)
.007
(.006)
.013
n
2
16
25
17
6
10
27
27
8
9
7
6
7
15
4
34
(12)
12
( 4)
4
( 7)
8
17
21
(10)
8
( 3)
5
The values shown in parentheses are for least
squares fits including POA data. The values
of y and sample deviation, s, are expressed in
fraction of feed COD. Times, t, are in hours,
thus, the units of k]_ and k2 are per hour. The
values of n not parenthesized are the number of
plant data sets averaged. Those in parentheses
are the numbers of POA data averaged.
64
-------
The post-ozonation apparatus was available for only the
last five of the nine feeds treated in the ozone pilot
plant. For these feeds, the "two-component" model parameters
were fitted to the average plant COD data combined with the
average data from the post-ozonation apparatus (POA) shown
in Table 4. The values so obtained are shown as the
parenthesized entries in Table 6 and the results are shown
as the broken curves in Figures 17, 19 and 20. The curves
fitted to plant and POA data passed within 0.1% of the
average L/Lf value from POA, and as can be seen in the
figures and from the values of s in Table 6, fitting to both
plant and POA data certainly does no violence to the plant
data.
In view of the excellent fit of the "two-component" model to
the average raw COD data over a range of treatment time from
10 minutes to 8 hours, and COD reduction from 10% to 90%,
it is very tempting to draw broad theoretical conclusions.
However, because of the uncompensated variations of ozone
dosage in the raw data, this probably is not warranted.
Instead, the "two-component" model should, for the present,
be regarded simply as an excellent empirical relation for
fitting and, if necessary, extrapolating COD reduction
rate data. For example, if the curves fitted to the plant
data alone are extrapolated four to eight-fold in time to
predict the POA data, the average predicted percent reduction
is within 6% of the experimental POA averages.
Correlations for COD Reduction
General. From detailed inspection of the raw data, one
expects the following variables to be important in developing
overall correlations for COD reduction by ozone treatment:
the nature of the pretreatments given to the wastewater, pH,
65
-------
treatment time, ozone dosage, and dissolved ozone concentra-
tions. There are other, but probably less important, vari-
ables such as surface tension, suspended solids, and dissolved
inorganic salts. Water temperature is probably an important
variable, but one over which there was no control in the pilot
plant. Although there was a significant variation in feedwater
temperature over the entire period of plant operation, the
variation during the treatment of any one feed was insignif-
icant. Therefore, the effects of temperature on reaction
rate cannot be included in the overall correlation of the
plant data.
Dosage Curves. Ozone dosage is probably the most important
variable in determining overall COD reduction. Unfortunately,
the term dosage is somewhat ambiguous in general usage. It
has been used to mean anything from the quantity of ozone
leaving the generator to the net quantity of dissolved ozone
consumed in removing COD, and it has been expressed in many
different units. To avoid any possibly ambiguity and for
the lack of a more precise term, in this report a variable D*
is defined as the total pounds of ozone dissolved in the
wastewater per pound of COD in the feedwater at the start of
the ozone treatment. This is also numerically the same as
ppm ozone dissolved/ppm COD feed.
From the plant data, the value of D., the pounds of ozone
dissolved in the i-th reactor per pound COD feed, is
calculated from a gas balance around the reactor.
DJL = 0.0828 G± (Cg - cv)/100 Mf (6)
where 0.0828 is the density of oxygen, Ib/scf
Gi is the gas flow to the i-th stage, SCFH
Cg is the wt % ozone from generator
GV is the wt % ozone vented from the i-th
stage
Mf is the Ib/hr COD feed to the plant
66
-------
The value of DN* at the outlet of the N-th stage is simply
the cumulative sum of D. for the case of equivolume stages.
D* = ?D. (7)
In order to examine the effects of D* on COD reduction,
values of D* and L/Lf were interpolated to a .common set of
selected even values of treatment time. Several sets of 'the
resulting data for individual pH and feed groups were
plotted in order to discover a correlating function. A
typical such plot is shown in Figure 21. Although the data
exhibit scatter as most wastewater treatment data do, the
following equation is a good fit to the data at any selected
reaction time.
Ln (L/Lf) = -bD* (8)
The next step was to make a least-squares fit of Eq (8) to
the data interpolated for the several selected values of
reaction time for each pH and feed group. An example of
the values of b so obtained are illustrated for the 3B+F
feed in Figure 22. In general, this group of graphs showed
small, but irregular, variations of b with treatment time
and a somewhat larger, but still irregular, variation with
pH.
In order to find if the variations of b with time and pH
were significant, a simple statistical analysis of the data
(12)
was made. ' The values of L/Lf, total reaction time (t*),
and D* were calculated for the outlet of each reactor. Only
measured COD values were considered; interpolated values
were not used.
One adjustment was made of some' of the raw data for the
DC2+L feed only. After changing the plant inlet plumbing to
accept this last feed tested, the feedwater sample tap
67
-------
100
0 I 2 3 45
LB OZONE DISSOLVED/LB COD FEED
Figure 21. PERCENT COD REMAINING VS OZONE DOSAGE
68
-------
actually yielded a sample of the water flowing to the first
stage mixer, a fact not noticed until the runs were nearly
completed. When no recycle was being drawn from the bottom
of the first reactor, i.e., when the mixer flow did not ex-
ceed the feedwater flow, the tap yielded a good feedwater
sample, and no correction was necessary. But when the
mixer flow exceeded the feed flow, the tap yielded a mix-
ture of feedwater and recycle water in proportions known
from flow meter readings. In this latter case (9 data sets),
the feedwater COD was calculated by mass balance from the
measured mixture COD assuming that the COD of the recycle
water had been reduced by 30% of the measured COD reduction
in the entire first stage reactor.
These raw plant data were fit by least squares to the
following two equations and to Eq (8) for each pH and feed
group separately and then for all points in each feed
group.
Ln(L/Lf) = (g + ht)D* (9)
Ln(L/Lf) = (u + j LnLf)D* (10)
The resulting parameters for Eq (9) and Eq (8) are listed in
Table 7. In this table, s, represents the standard deviation
of the coefficient h in Eq (9), s, that of coefficient b in
Eq (8), and s the standard sample deviation. (See Appendix
7.)
The coefficient h in Eq (9) represents the possible variation
of b in Eq (8) with reaction time. In general, the standard
deviation for h is as large as the value of h. Therefore, it
can be concluded that the variation of b with time is
insignificant within the 3/4 to 3 hr range of treatment
69
-------
0.6
0.5
0.4
0.3
O.I
38 4- F
1
I
20 40 60 80
REACTION TIME, MIN.
100
Figure 22.
EFFECTS OF pH AND REACTION TIME ON CALCULATED
SLOPE (b) OF DOSAGE CURVE: 3B +' F
70
-------
times examined in the pilot plant.
In examining the values of b and s, in Table 7 it can be
seen that, for almost all cases, the difference between
b's determined by pH group and the b fitted to all of the
pH groups within a feed is less than the combined uncertainty
in the b values. Therefore, in the statistical sense, the
value of b is not dependent on the pH of the feedwater.
The coefficients for Eq (10) are not shown in Table 7.
However, for each feed the values of the coefficients u and
j were similar, suggesting that b might depend upon L_.
Therefore, Eq(10) was fitted to the 1031 experimental points
for all feeds except 3B+L+F. Those data were not used primar-
ily because at least half the feed COD's contained excess
methanol, a substance atypical of domestic wastewater. A good
fit was obtained as shown by the coefficient values in Table
7. In figure 23 are shown the b values obtained for each
feed plotted against log of the average feed COD. The
vertical bars represent the one s, uncertainty in each
value. The point for the 3B+L+F feed is shown by a dotted
symbol. And, the solid line represents the fit of Eq(10)
to all of the data except 3B+L+F.
The relation between COD reduction during ozone treatment
and D* is as follows:
Ln(L/Lf) = (0.1290 - 0.1746 Ln Lf) D* (11)
Eq (11) is a broad and general relation. According to the
analysis of the pilot plant data, the relation between COD
reduction and D* is:
a. strongly dependent on the feedwater COD,
b. independent of treatment time within the 0.7 to
3 hour periods examined,
71
-------
7.
ANALYSIS OF COD REDUCTION vs.
D* (LB OZONE DISSOLVEP/LB COD FEED)
Ln(L/Lf) «
(g + ht)D* Ln(L/Lf) - -bD*
Feed
3B+L+F
UN+L+F
MC+CL+C
MC
UN (All)
MC+CL(A11)
3B+F
DC2 (All)
DC2+L
(9 data
sets ad-
justed)
pH
5.5
7.3
8.1
9.1
All
6.7
8.9
All
6.1
7.0
8.2
9.0
All
6.1
7.2
7.9
8.9
10.0
All
6.0
7.2
5.2
6.2
6.9
All
7.0
7.2
8.6
All
h
0.101
-0.071
-0.028
0.049
0.040
0.090
0.115
0.102
0.016
0.020
0.012
0.006
0.014
-0.003
-0.053
0.006
-0.076
0.009
-0.041
0.175
-0.139
0.083
-0.054
-0.030
-0.010
0.143
-0.011
-0.075
-0.038
Sh
0.135
0.161
0.101
0.038
0.039^
0.080
0.053
0.047
0.038
0.044
0.047
0.078
0.026
0.027
0.058
0.077
0.048
0.028
0.021
0.049
0.061
0.098
0.125
0.090
0.064
0.102
0.060
0.122
0.070
b
-0.168
-0.434
-0.333
-0.261
-0.310
-0.304
-0.315
-0.313
-0.307
-0.322
-0.307
-0.256
-0.295
-0.573
-0.580
0.642
-0.570
-0.586
-0.582
-0.433
-0.468
-0.305
-0.471
-0.419
-0.417
-0.634
-0.505
-0.650
-0.568
Sb
0.084
0.074
0.052
0.022
0.023
0.047
0.037
0.028
0.026
0.029
0.044
0.057
0.015
0.046
0.053
0.067
0.055
0.056
0.028
0.052
0.080
0.045
0.071
0.043
0.032
0.058
0.049
0.099
0.058
Se
0.161
0.474
0.472
0.328
0.443
0 .115
0.219
0.188
0.173
0.214
0.150
0.351
0.220
0 .089
0.110
0.141
0.226
0 .080
0.166
0.263
0.205
0.190
0.275
0.341
0.317
0.206
0.124
0.233
0.198
No
Points
17
89
145
95
346
30
53
83
123
116
33
39
311
35
29
27
71
19
181
140
49
19
30
75
124
94
27
22
49
All Feeds Ln(L/Lf) = (u + j LnL )D*
u = 0.1290 + 0.0956
j = 0.1746 + 0.0327
s = 0.231
72
-------
co
0.7
0.6
*o 0.5
04
0.3
0.2
O.I
3B+F<
MC+CL+Ci
UN+L+F
MC
DC2+LO
ln(L/Lf)=-bP'
10
15
20 30 40 50 60 70 80 100
Lf,FEED COD, MG/L
Figure 23. CORRELATION OF COD REMOVAL AND OZONE DOSAGE
FOR ALL FEEDS
-------
c. practically independent of the feedwater pH in the
range 5 to 10, and
d. practically independent of the exact nature of
pretreatment.
To the extent that the Blue Plains wastewater is a typical
domestic wastewater, there is no evident reason to expect that
Eq (11) should not apply to other wastewaters from mostly
domestic sources.
To illustrate some of the implications of Eq (11) , the
percent COD reduction and the Ib ozone dissolved/lb COD
removed are shown in Figure 24 as a function of D* for feed
COD's of 20, 40, and 80 mg/1. This illustrates that the
higher COD feeds are the more susceptible to ozone treat-
ment; a given D* producing a greater percent COD reduction
for the higher COD feeds. Similarly, the ozone consumption
per unit COD removed is less for the higher COD feeds. This
should be expected, especially for oxidative pretreat-
ment, since lower COD feeds reflect more extensive pretreat-
ment and the removal of the more easily oxidizable contamin-
ants .
It must be remembered, though, that D* is normalized to feed
COD. To emphasize this, in Figure 25 is shown COD reduction
in mg/1 as a function of the Ib ozone dissolved per 1000 gal
water treated. As the feed COD increases, the ozone
requirement per 1000 gallons must increase very rapidly to
maintain the same percent COD reduction, and even more
rapidly to maintain a constant COD product.
COD Reduction Rates
General. Inspection of the averaged raw data and the
fact that the dependence of COD removal on D* is time
independent suggests that the rate of COD removal at any
74
-------
Q3AOIN3H 000 ai/Q3A"IOSSia 3NOZO 91
^J- CJ
Figure 24. RELATION BETWEEN PERCENT COD REDUCTION
AND LB OZONE DISSOLVED/LB COD FEED AND
LB OZONE DISSOLVED/LB COD REMOVED
75
-------
I 2 3
LB OZONE DISSOLVED /1000 GALLONS
Figure 25.
COD IN MG/L vs LB OZONE DISSOLVED PER,
1000 GAL WATER TREATED
76
-------
point during ozone treatment is strongly dependent on the
degree of ozone treatment up to that point. Thus, the
correlation of COD reduction rates must be made on an
integral, rather than a differential, basis.
Seeing that COD reduction as a function of D* exhibits no
significant pH dependence, but that the averaged raw data
do depend on pH, it is evident that pH must have an
important effect on the COD reduction rate.
'.LI
IT*
5 .•• ~
From simple chemical intuition, the reaction rate should also
depend on the dissolved ozone concentration. The dissolved
ozone concentration varied throughout each reactor, that at
the bottom of the reactor being higher than that at the top,
and one should expect that in the mixer and dissolver tube
the concentrations were higher still. Presumably, there is
some more-or-less constant relationship between dissolved
ozone concentrations at various locations within the reactor.
Because only dissolved ozone values at the reactor outlets
could be obtained for all plant data sets, these values were
selected for use in correlating the data. As was pointed out
above, dissolved ozone concentrations measured by the
acidic KI procedure prior to 6/24/71 are considerably less
reliable than those measured by the alkaline KI procedure
after that date. For this reason, only the plant data sets
collected after 6/24/71 were used for the reaction rate
correlation. It should be noted that the latter period
includes the complete range of feed COD's from the lowest,
3B+F, to the highest, DC2.
Having recognized, though, that the reaction rate correlation
must be on an integral basis in order to take into account
the previous history of ozone treatment up to any point, one
must use some integral function of the dissolved ozone
77
-------
concentration at the outlet of each reactor, Z* (the time
weighted average of dissolved ozone concentration) as defined
by Eq (3) is just such a function.
A reaction rate correlation implies the use of COD removal
as the dependent variable. But because COD removal is a
simple function of D* and Lf, one independent variable, L^,
can be eliminated from the reaction rate correlation by using
D* rather than L/Lf as the dependent variable. Moreover, the.
use of D* has certain advantages in the curve-fitting
mathematics. Any correlated value of D* can be simply
related to COD reduction using Eq (11).
For the reasons outlined above, the general "reaction rate"
correlation will be aimed at describing D* as a function of
cumulative reaction time, t*; pH of the entering feedwater;
and Z*- (It should be noted that time, t, in all equations
is in hours, whereas the time axes are in minutes for all
data or correlation plots which follow.)
Average Raw Data. By pH and feed groups, the raw plant
data for the period after 6/24/71 were used to obtain inter-
polated values of Z* and D* at selected even values of
treatment time. (The data for the MC+CL feed were not used
because the residual chlorine destroyed the accuracy of the
dissolved ozone measurements.) These interpolated data were
plotted on various coordinates and it was found that on log
Z* vs log D* plots the data fell linearly. A typical plot
is shown in Figure 26 for all data in the pH range 6.5 to
7.5 interpolated for 60 min treatment time. Although there
is some scatter, the data for all five feeds clearly fall
alqng a straight line.
The next step was to least-square a straight line through
78
-------
the interpolated, log Z* and log- D* data, for all points in
each pH group and for each 'common value of reaction time.
From these smoothing functions, the values of D* could be
interpolated for a number of selected even values of Z*,
t*, and pH. In Figure 27 is shown a plot of the smoothed
and interpolated values of D* vs t* for Z* - 1 mg/1 and for
three pH groups. (Recognize that each group contains data
for several different feeds.) Plots for other values of Z*
yielded curves of similar shape but displaced upward with
increases'in "Z*. Thus, in Figure 27 one can see an
approximation of the shape that the "reaction rate" correlation
curves must assume. Log D* vs log t* is a straight line for
pH 7. Other pH curves initially diverge from the pH 7 curve,
but then begin approaching the pH 7 values as treatment time
is increased. This characteristic is entirely consistent
with the observed change of pH toward neutrality during
ozone treatment.
Final Correlation. The final correlation of D* as a
function of t*, pH, and Z* was obtained using a more-or~less
standard multiple-regression procedure modified to
discard "wild points" (2.5 s criterion) from the initial fit
before making the final fit. Only raw experimental data
were used. In the few cases that either Z* or D* incorporated
estimated data, the point was assigned 1/2 normal statistical
weight. If both incorporated estimated data, the statistical
weight 1/4 was assigned.
Ln D* = aQ + a^ Ln t* + a2 LnZ* + a3 (pH-7)/t*
+a4[ (pH«7)/t*]Ln t* (12)
where aQ - 0.0072 + 0.0137
a-L - 0.4461 + 0.0156
a2 = 0.5044 + 0.0147
79
-------
10.0
6.0
4.0
3.0
2.0
_ O MC+CL+C
o MC
* 3B + F
h X DC2 + L
a DC2
*N
0.8
0.6
0.4
0.3
0.2
PH*7
t*= 60 MIN.
I
I
0.3 0.4 0.6 0.8 1.0 2.0 3.0 4.0 6.0
D*LB OZONE DISSOLVED/LB COD FEED
Figure 26.
PLOT OF LOG DISSOLVED OZONE (Z*)
vs LOG DOSAGE, D*
80
-------
I
I
o oo
try
^tCt oj —•' O O O'O
Q33d 000 ai/a3A~IOSSIO 3NOZO
O
O
ro
(O
o
0<
ro LJ
a:
CvJ
OJ
Ci
Figure 27.
RELATION BETWEEN SMOOTHED VALUES OF DOSAGE
(D*) AND REACTION TIME (t*) FOR Z* ~ 1 MG/L
AND INDICATED pH
81
-------
a3 - 0,1359 t 0.0094
a4 =q 0,0575 + 0,0067
The coefficients were obtained from a final fit to 774
experimental data points after discarding 23 points as "wild".
The standard sample deviation on In D* is 0.267, thus the D*
data scatter about 30% around this function. The uncertainty
shown for the coefficients are the one s& values. Note that
for practical purposes, aQ is zero.
In Figure 28, Eq (12) is graphed for the value Z* = 1 mg/1.
Figure 29 is a graph of Eq (12) for pH 7. From Eq (12) it
may be noted that the displacement of the curves due to pH is
independent of Z*, thus by the use of conventional graphical
techniques, the values for any set of D*, t*, and Z* can be
determined from Figures 28 and 29.
Eq (11) can be combined with Eq (12) to yield an actual
reaction rate function.
Ln(L/Lf) = (0.1290 - 0.1746 LnLf)t* (0'4461 +0.0575(PH-7)/t*)
Z*0.5044e0.1359 (pH-7)/t* (13)
For pH = 7, Eq (13) is simplified to Eq (14)
Ln(L/Lf) = (0.1290 - 0.1746 LnLf)Z*°'5044t°'4461 (14)
Again, to the extent that the Blue Plains wastewater is
typical of domestic wastewater, there is no evident reason
to expect that Eq (14) should not hold for other domestic
wastewaters. But because wastewater streams vary in their
pH buffering capacity, Eq (13) is probably not so generally
applicable for pH values much removed from neutrality.
Applying the Correlations. In order to illustrate the
use of the correlated data, a specific example will be used.
Consider an ozone treatment plant receiving a pretreated
82
-------
waste from primarily domestic sources with a COD of 40 mg/1
and a pH of 7. The required product COD is 20 mg/1. Assume
that the plant will consist of three equivolume reactors
each with a detention time of 15 minutes.
Total Ozone Dissolved. From Eq (11) or Figure 23, the
value of -b for domestic wastewater with Lf = 40 is 0.513.
From Eq (10) or (11) or Figure 24, the required value of D*
for L/Lf = 0.5 is 1.35 Ib ozone dissolved/lb COD feed
(or 54 mg/1). From Eq (12) or Figure 29, attaining D* of
1.35 in 45 min (3x15 min) at pH 7 requires a Z* of about
2.3 mg/1.
Ozone Dissolved per Stage. The only technical require-
ment for the values of Z. in each of the three stages is
that Eq (3) be satisfied, i.e., Z1 + Z2 + Z3 = 3 x Z*=6.9 mg/1
Without economic optimization, there are no further require-
ments .
For this example, the following Z. values will be used:
Z, = 1.4, Z2 = 2.5, Z3 = 3.0 mg/1. Again using Eq (14) or
Figure 29, D* at t* = 0.25 hr and Z* = 1.4 mg/1 is
0.63. D * at t * = 0.50 hr and Z * = (1.4 + 2.5)/2=1.95 mg/1
2. £ & | "
is 1.02. And D^* at tg* - 0.75 hr and Z* = 2.3 mg/1 is, of
course, 1.35.
Recognizing from Eq (3) that D. = D.* - D*_,, the Ib ozone
dissolved/lb feed COD in each of the three reactors is 0.63,
0.39, and 0.33, respectively.
To proceed, this example will be placed on a 1000 gal treated
basis. The COD feed is 40 Ib COD/106 Ib water x 1000 gal x
8.33 Ib/gal = 0.333 Ib COD feed/1000 gal. The quantity of
ozone to be dissolved is 1.35 x 0.33 = 0.45 Ib ozone/1000 gal
83
-------
20 30 40 6O 80 100 200 300
CUMULATIVE REACTION TIME, MIN.
Figure 28, GRAPH OF REACTION RATE CORRELATION FOR
CONSTANT Z* = 1 MG/L
-------
6.0
10 20 30 40 60 80 100 200 300
CUMULATIVE REACTION TIME,
Figure 29. GRAPH OF REACTION RATE CORRELATION FOR
CONSTANT pH 7
85
-------
treated, 0.21 in the first stage, 0.13 in the second stage,
and 0.11 in the third stage.
Oxygen Consumption
There are four sources of oxygen consumption in the ozone
pilot plant.
1. Chemical reaction of ozone to reduce COD: one
Ib of oxygen is required to remove one Ib COD.
2. Dissolution of oxygen in the product water: after
deaeration, the feedwater contains practically no
- > ;'
dissolved oxygen, and the product water is
practically saturated with one atma of pure oxygen.
3. Dissolution of oxygen and residual ozone gas in
the compressor seal water.
4. System leakage, including gas sampling.
Make-up oxygen to the pilot plant was measured by an integra-
ting, positive displacement meter. The meter readings
(usually about 25 cu ft/hr) were generally in agreement with
the oxygen consumption calculated for items 1, 2, and 3,
above, which indicates that system leakage was small. Oxygen
losses in the compressor seal water, which accounted for
roughly 40% of the total oxygen consumption, could be elemin-
ated either by recycling the seal water to one of the latter
treatment stages or by using a different type of compressor.
Nitrogen Reactions
The reactions of nitrogenous substances in wastewater
during ozone treatment were investigated to a limited
extent for the last five feedwaters tested in the pilot
plant: UN, MC, 3B + F, DC2, and DC2 + L. Usually, one set
of wastewater samples per day were collected and submitted
to the EPA analytical group for nitrogen analyses. The
samples were analyzed for total Kjeldahl nitrogen (TKN) ,
86
-------
ammonia nitrogen (NH-j-N) , nitrite nitrogen (N02-N) , and
nitrate nitrogen (N03-N).
In general, ozone treatment did not reduce the overall
nitrogen content of the wastewater. At high pH values,
however, some of the ammonia and total Kjeldahl nitrogen
was oxidized to nitrates. No nitrite formation was ever
observed. Interestingly, there was no detectable ammonia
removal in the vacuum deaerator, even at high pH values.
The nitrogen analytical data for five feedwater pretreat-
ments are summarized in bar chart form in Fig 30 through
33. Each bar in these figures represents the average
results for a number of pilot plant runs, and each bar is
subdivided into TKN, NH--N, and NO -N. Bars are shown for
the feedwater, water from the outlet of the first reactor
(#1), and water from the outlet of the last reactor (#6).
UN. Average results for the Unox pretreated wastewater
are summarized in Fig 30. The feedwater was fairly well
nitrified by the pretreatment, and ozone treatment at pH
5.9 to 6.6 produced no significant additional nitrification.
MC. Average results for the feedwater pretreated by mineral
clarification is shown in Fig 31 for three different pH
values. Pretreatment had oxidized very little of the
nitrogenous materials. These data illustrate the strong
dependence of nitrification by ozone treatment on feedwater
pH. Nitrification is significant only at pH 10 and in the
final stages of the treatment process. At pH 6, only about
5% of the TKN was oxidized to nitrate by ozone; only about
8% was oxidized at pH 9; but more than 35% of the TKN was
nitrified by ozone at pH 10.
87
-------
20i
UNOX
15
e>
00
CO
10
o
or
T
SETTS
P • 1»-
FEED # I
Figure 30. OXIDATION OF NITROGENOUS COMPOUNDS BY OZONE: UN
-------
20
MC
'CO
15
,-10
UJ
e>
o
o:
FEED
pH6
FEED #\
pH 9
1
i
ro
O
FEED #\ #6
pH 10
Figure 31. OXIDATION OF NITROGENOUS COMPOUNDS BY OZONE: MC
-------
10
UJ
0 K
o 5
IT
h-
3B-I-F
\\VsT
._,..
^N03-N
0r
FEED
Figure 32. OXIDATION OF NITROGENOUS COMPOUNDS BY OZONE: 3B+F
-------
20-
o
UJ
10
DC2
DC2+L
N03-N
.
xl-
il
FEED # I
FEED
Figure 33. OXIDATION OF NITROGENOUS COMPOUNDS BY OZONE: DC2
-------
3B + F. The three-stage biological process is specifically
designed for nitrogen removal. The average data for the 3B
+ F pretreated feed, illustrated in Fig 32, shows that there
was, indeed, very little nitrogen left in this feedwater.
Ozone treatment in the pH range 5.4 to 6.9 produced virtually
no additional nitrification of the small, residual TKN..
DC2 and DC2 + L. Average results obtained during ozone
treatment of the secondary effluent from the main Blue Plains
plant, with and without additional lime clarification-, are,
summarized in Fig 33. Both feedwaters had relatively high
TKN concentrations and very low nitrate concentrations.
Little nitrification by ozone treatment was observed for
either feedwater in the pH range 6.9 to 8.1.
Turbidity Reduction
Wastewater turbidity was consistently reduced by ozone
treatment. For unfiltered pilot plant feeds, turbidities
expressed in Jackson Turbidity Units (JTU) were reduced by
about two-thirds during ozone tertiary treatment. In
treating filtered feedwaters which had low initial
turbidities, product water turbidities of less than one JTU
were obtained routinely.
Typical pilot plant data for two unfiltered feedwaters,
MC and UN, are illustrated in Fig 34. Feedwater turbidity
(circular points) and product water turbidity (triangular
points) are shown for a number of runs. The data shown for
the MC pretreated feed encompass a pH range 6.1 to 10.9,
and those for the UN pretreated feed, a pH range 5.8 to 6.6.
In comparison with the biologically pretreated feed (UN) ,
one might expect a greater portion of the suspended solids
in the physicochemical pretreated feed (MC) to be ozone
92
-------
i—rn
FEED FROM MC
DAYS
I I I I I I
FEED FROM UNOX
PRODUCT
I I I I
DAYS
Figure 34. TURBIDITY REDUCTION FOR MC AND UN
PRETREATED FEEDS
93
-------
1000
_ DC2
DC2+ L
ENTERING COD
100
o
Q
o
o
^
(-;
•o
g
CD
cc
10
I
Figure 35.
TURBIDITY REDUCTION DURING OZONE
TREATMENT OF DC2
94
-------
resistant inorganic materials, e.g., calcium carbonate.
But from Fig 34, it can be seen that the turbidity removal
from the MC pretreated feed is, if anything, greater than
that from the UN pretreated feed. A possible explanation
is an increase in the solubility of suspended inorganic
solids due to the pH shift which accompanies ozone
tertiary treatment.
A similar graph of turbidity data for the DC2 and DC2 + L
feeds is shown in Fig 35 which also includes data for the
feedwater COD (square points). It can be seen that the
turbidity reductions are, again, quite substantial and that
there is little correlation between feedwater COD variations
and the turbidity reduction.
Bacteria Kills
During the course of ozone treatment of each type of
effluent, samples were taken ^rom four points in the pilot
plant for determination of bacteria kill. The points
chosen were the neutralization tank and the discharge pipes
from the first, third, and sixth reactor stages. The
sample systems were thoroughly flushed with the treated
water prior to sample collection, and the samples were then
drawn into sterile bottles, taking care to avoid contamina-
tion during sampling. The samples were taken immediately to
the C.W. England Laboratories in Beltsville, Md., so
preliminary procedures could be started in about an hour
from the time of sampling. Standard plate count, Coliform,
and E. Coli were determined on these samples.
Table 8 summarizes the results of these analyses. The No. 1
samples are untreated water and have a substantial total
plate count. The wastewater feeds which have been clarified
and filtered show lower standard plate counts in the beginning
95
-------
Table 8.
BACTERIOLOGICAL COUNTS ON
WASTEWATER SAMPLES FROM PILOT PLANT
Date of
Sample
Type of
Collection jWastewater
1971
4-28
5-20
7-9
7-20
8-18
Feed
3B+L+F
UN+L+F
MC+ CL+ C
MC
UNa
Sample
Point No.
&
(Treatment
Stage)
1 (0)
2 (1)
3 (3)
4 (6)
1 (0)
2 (1)
3 (3)
4 (6)
1 (0)
2 (1)
3 (3)
4 (6)
1 (0)
2 (1)
3 (3)
4 (6)
1 (0)
2 (l)b
3 (3)
4 (6)
Standard
Plate
Count
per ml
2,300
0
0
0
12,000
0
0
0
110,000
0
0
0
150,000
7
0
0
16,000
12,000
130
110
Coliform Bacteria
Most Probable No.
per 100 ml
46
<3
<3
<3
23
<3
<3
<3
3.6
<3
<3
<3
>1,100
<3
<3
<3
>1,100
>1,100
<3 !
<3
E. Coli Bacteria
Most Probable No.
per 100 ml
2.3
<3
<3
<3
<3
<3
<3
<3
<3
<3
<3
<3
<3
<3
<3
<3
23
23
<3
<3
CTl
-------
Table 8 (Contd)
Date of
Sample
Collection
1971
8-26
9-9
9-17
9-30
Type of
Wastewater
Feed
MC+CL
3B+F
DC2
DC2+L
Sample
Point No.
&
(Treatment
Stage)
1 (0)
2 (1)
3 (3)
4 (6)
1 (0)
2 (1)
3 (3)
4 (6)
1 (0)
2 (1)
3 (3)
4 (6)
1 (in)c
2 (0) d
3 (1)
4 (3)
5 (6)
Notes: a - Approx. 36 hrs elapsec
Standard
Plate
Count
per ml
12
8
14
74
7,600
0
0
0
54,000
5
1
0
63,000
54,000
17
1
0
Coliform Bacteria,
Most Probable No.
per 100 ml
<3
<3
<3
<3
>1,100
<3
<3
<3
>1,100
<3
<3
<3
>1 , 100
>1 ,100
3.6
<3
<3
E. Coli Bacteria
Most Probable No.
per 100 ml
<3
<3
<3
<3
240
<3
<3
<3
9.1
<3
<3
<3
>1,100
>1,100
<3
<3
<3
between sample collection and start of
analysis .
b - No ozone was fed to first stage.
c - Sample of District of Columbia secondary wastewater.
d - Sample of District of Columbia secondary wastewater after lime
clarification.
-J
-------
and these are reduced to zero in the first ozonation stage.
The mineral clarified, chlorinated and carbon treated
(MC+CL+C) effluent had a high plate count indicating that
bacteria not only survived chlorination but also multiplied
in the carbon column. The bacteria are quickly killed by
ozone, however, because the bacteria are not protected by
suspended matter.
Other effluents, which had not been filtered, contained
more suspended matter as indicated by higher turbidities
(4-20 JTU). The standard plate count in these cases did not
reduce to zero after the first stage, but usually was
reduced to a low value.
To summarize, these tests demonstrate that ozone is a very
effective disinfectant for bacteria even in the presence of
suspended matter. However, the suspended particles require
a longer exposure to obtain a zero standard plate count.
Response to Plant Load Variations
The response of an ozone wastewater treatment plant to varia-
tions in load depends entirely upon the details of the plant
design. For maximum flexibility, a number of the pilot plant
components were intentionally oversized. Therefore, the
pilot plant response to load variations is, in itself, of
little value for predicting the behavior of a large plant.
However, one important point is evident from the pilot plant
data: an ozone wastewater treatment plant will respond very
rapidly to any load changes. Thus, there is limited surge
capacity inherent in the process, but on the other hand,
there is no lingering after-effect from large load variations
A fair approximation of the response of a large ozone waste-
98
-------
water treatment plant to load variations can be calculated
from the correlation of dosage data, Eq (11). The
necessary assumption is that the plant has a fixed capacity
(Ib/hr) for dissolving ozone. As is shown in the discussion
of economic optimization, below, the dissolution efficiency
of the gas-liquid mixers increases slightly with increasing
plant load due to the decrease in dissolved ozone concentra-
tion (Z). Therefore, the assumption is a conservative one. •
As an illustration, again consider the example used in
applying the COD correlations for a plant required to reduce
COD from 40 mg/1 to 20 mg/1. The quantity of ozone to be
dissolved was shown to-be 0.45 Ib ozone/1000 gal feedwater
at nominal hydraulic load. Assuming this quantity to be
invariant with changes of load, the calculated effects of
variations in feed COD (L,-) at constant hydraulic load are
shown in Table 9. The calculated effects of hydraulic load
variations at constant 40 mg/1 COD feed are shown in Table 10.
From Table 9, it can be seen that as the feed COD increases,
the product COD (L ) increases, but in less than direct
proportion to the load. At twice the nominal feed COD, a,
35% COD reduction is still realized. Recalling that the
product water contains about 40 mg/1 of dissolved oxygen, the
product COD would not exceed the dissolved oxygen content
until the COD load exceeded about 165% of nominal load. Up
to that point, the product water could exert no net oxygen
demand on the receiving stream.
From Table 10, one sees that the percent COD reduction
decreases with, increasing hydraulic load, but in this example,
the product COD would not exceed the dissolved oxygen con-
centration. Also tabulated is the concentration of ozone
dissolved at various hydraulic loadings. The work of others
99
-------
has shown that ozone doses of 7 to 15 mg/1 are adequate to
meet regulatory disinfection requirements. Thus, even at
hydraulic loads of 250% of nominal capacity, this example
plant would still have reserve disinfecting capacity.
A practical ozone wastewater treatment plant naturally would
incorporate provisions for meeting normal load variations.
But even in this restricted example of a completely inflex-
ible plant, relatively large overloads are seen to produce
neither health nor ecological hazards.
Summary
The anticipated technical virtues of ozone wastewater
treatment were all demonstrated in the pilot plant. COD,
turbidity, and surfactants were significantly reduced, and
the product water was nearer neutral pH, oxygen saturated,
and practically free of bacteria. No extraneous chemical
additives remained in the water, and there were no residual
materials to be disposed of.
For domestic wastewater similar to that at Blue Plains, the
relations between Lf, L^, pH, t*, Z* and D* can be estimated
from the correlations shown in this section.
100
-------
Table 9.
RESPONSE TO CHANGE IN FEED COD AT CONSTANT
HYDRAULIC LOAD OF A PLANT CAPABLE OF DISSOLVING
0.45 LB OZONE/1000 GAL FEEDWATER
Load Lf Lp % COD
Factor mg/1 mg/1 Reduction
0.50 20 6.9 65.4
0.75 30 13.0 56.5
1.00 40 20.0 50.0
1.25 50 27.6 44.9
1.50 60 35.4 41.0
2.00 80 52.1 34.9
2.50 100 69.5 30.5
Table 10.
RESPONSE TO CHANGES IN HYDRAULIC LOAD
AT CONSTANT Lf = 40 mg/1 OF A PLANT
CAPABLE OF DISSOLVING 0.45 Ib OF
OZONE PER 1000 GAL FEEDWATER
Hydraulic Lp % COD mg/1 Ozone
Load Factor mg/1 Reduction Dissolved
0-50 10.0 75.0 108.0
0.75 15.9 60.3 72.0
1.00 20.0 50.0 54.0
1.25 23.0 42.5 43.2
1.50 25.0 37.0 36.0
2.00 28.3 29.3 27.0
2.50 30.3 24.2 21.6
101
-------
SECTION VIII
CHARACTERISTICS OF PILOT PLANT EQUIPMENT
Having at this point described and discussed the technical
results obtained from the ozone wastewater treatment
process, the next major consideration is the projected
costs for large plants. Since these cost projections
will be based largely on scaling-up the pilot plant process,
it is necessary to first describe and correlate the perfor-
mance of several pieces of the pilot plant equipment.
Ozone Generator
The ozone generator used in the pilot plant was a commer-
cial, 34 tube unit which is further described in Appendix 2.
In the first few months of operation, the individual fuses
to four of the thirty-four tubes opened. This is said to
be expected on initial start-up, and that upon replacement
of the weak glass dielectric tubes, very few additional
failures should be expected. However, the four tubes were
not replaced because the machine still slightly exceeded
the manufacturer's performance specifications, and the unit
operated reliability and without further incident for the
remainder of the nine-month pilot plant operation.
The electric power consumption of the ozone generator is a
major part of the total ozone water treatment cost. For
power measurements, the generator was supplied with an
ammeter, voltmeter, and electrodynamometer kilowatt meter
connected between the variable transformers and the high
voltage transformer. Data from these three instruments were
recorded for all plant runs.
103
-------
The treatment of the MC pretreated feed, which had a moderate-
ly high COD, offered the first opportunity for operating the
ozone generator at high power inputs. It then became apparent
that the calculated generator efficiencies were uncommonly
high and the power factors suspiciously low. Recognizing
that the highly distorted a.c. wave forms produced by the
ozone generator could easily introduce errors into the power
measurements, a check on the accuracy of the electrodynamo-
meter kilowatt meter was made.
The physically accurate measurement of very distorted a.c.i
power is difficult. But from the engineering viewpoint,
physical accuracy is not so necessary as is economic accuracy:
the practically important engineering parameter is the cost
of power for operating the ozone generator. It was, therefore,
decided to check the indicating kilowatt meter against the
type of meter commonly used for utility billing purposes,
namely, an induction watt-hour meter.
A commercial watt-hour meter with the necessary matching
transformer was loaned for this purpose by the Potomac
Electric Power Co. (PEPCO). This instrument was connected to
the ozone generator, and a series of calibrations were made
covering the full range of electrical power inputs and gas
flows. It was found that the electrodynamometer kilowatt
meter was, indeed, reading low, especially at higher powers.
(At the same time, tests were made to confirm that the losses
in the variable transformers were negligibly small.)
On the basis of this calibration, it was found that at lower
power inputs, the relation between the two meters was as
follows:
Kw = 1.537 (Kw ) - 0.0607 (Kw )2 + 0.0025 (Kw )3-661. (15)
W
104
-------
where Kw is the reading from the induction watt-hour
meter, and
Kww is the reading from the electrodynamometer
kilowatt meter.
However, above Kw readings of about 5 Kw, the electrodynamo-
meter instrument responded little to further increases of
power.
From the voltmeter and the ammeter readings, the Kva could
be compared with the induction watt-hour meter. It was
found that over the range 1 to 10 Kw input, the following
relation existed.
Kw = 0.748 (Kva) -0.00619 (Kva)2-153 (16)
For both Eq (15) and Eq (16), no significant effects of gas
flow rate were observed. Many comparisons of the several
instruments during subsequent plant operations verified the
original calibrations which are estimated to be precise
within about ten percent.
Since only data from the ammeter, voltmeter, and electrodyn-
amometer kilowatt meter had been collected during the
earlier plant operations, this procedure was continued, and
all of the power data was then converted to the induction
watthour meter basis according to Eq (15) and Eq (16). Both
the ammeter and voltmeter had nonlinear scales which
precluded precise readings at low powers. Therefore, Eq (15)
was used for Kw values up to 4.3 Kw, and Eq (16) was used
for the higher power inputs.
On the basis of the corrected power inputs, the efficiency
of the ozone generator, Kwh/lb O , was computed for all of
the plant runs from February through September. The
average efficiencies of the generator during that period are
shown in Figure 36. These curves offer a realistic description
105
-------
10
15
20
SCFM
01234
WEIGHT PERCENT OZONE
Figure 36.
AVERAGE EFFICIENCY OF PILOT PLANT
OZONE GENERATOR
106
-------
of the performance that can be expected from this type of
ozo.ne generator in practical plant operations, including
as it does, the effects of less than perfect maintenange,.
Gas-Liquid Mixers
General. In the ozone treatment of wastewater, the ozone
must be dissolved from the gas into the wastewater before? any
reaction with contaminants can occur. Since ozone can b,p
safely and economically produced only in relatively low
concentrations, and since the cost of ozone production is
a major part of the total cost of ozone treatment, highly
effective gas-liquid mixing equipment is necessary for
dissolving the ozone.
In the tertiary treatment of wastewater, the required ozone
dosage is comparatively high. In many practical cases, the
volume of ozone gas mixtures handled is roughly the same
as the volume of water treated. Mixing devices suited to
roughly equal volumetric flows of gas and liquid are not
common. The prepilot development of the mixing devices used
in the ozone pilot plant is described in Appendix 1.
Each of the six reactors in the ozone pilot plant incorpor-
ated identical mixing devices. A typical mixer is shown
schematically in Figure 37. Water is forced through the
nozzle of an eductor atop each reactor where it is mixed
with the ozone bearing gas. The dispersed gas in liquid
stream immediately enters a high-shear zone where it is
impinged by jets of additional water. The turbulent
dispersion is then forced down a 4-inch ID dissolver tub,e
to near the bottom of the reactor. There, a cap distributes
the flow evenly in the reactor cross-section, and the gas
bubbles are allowed to ascend through the water in the
reactor.
107
-------
I-1
o
CD
2" PVC PIPE
4 PVC PIPE TEE
3/4*O.D. STAINLESS TUBE
MIXING CHAMBER
4° DISSOLVER PIPE
PRIMARY WATER INLET
GAS IN
SECONDARY WATER
DISCHARGE SLOT
SWAGELOK
FITTING
SECONDARY
WATER
INLET
SECTION-A A
SECTION-BB
Figure 37, SCHEMATIC OF PILOT PLANT MIXER ASSEMBLY
-------
The water and gas feeds to each mixer were manually adjusted
and were individually metered. Water feed is supplied by a
separate pump for each mixer. Each pump, except the first,
draws from a line connecting the outlet of the preceding
reactor to the bottom of the reactor it feeds. Thereby,
the water flow through a mixer does not have to be matched
to the feedwater flow. If mixer water flow is greater than
the feedwater flow, part of the water to the mixer is recycled
from the bottom of the reactor. Conversely, if the mixer
flow is less than the feedwater flow, part of the overflow
from the preceding reactor does not pass through the mixer,
but is bypassed directly into the bottom of the reactor.
Definition of Mixer Performance: Transference (T). The
processes occuring in the pilot plant mixers are complicated.
There are four separate zones of mixing; first, in the
eductor, then in the high turbulence zone immediately below
the eductor, then in the dissolver tube and, finally, in the
reactor itself. And this complexity is compounded by
concurrent chemical reaction. A detailed analysis of all of
these simultaneous and interrelated processes would
obviously be hopelessly complex. The alternative is to make
bold, but well defined, simplifying assumptions. Such has
been done.
The combined mixer and reactor are considered to be a
"black box" cocurrent mixer. Gas and water flows at the
inlet and outlet and the ozone concentrations in each
stream are experimentally known, but nothing is known or
assumed about ozone concentratipns anywhere within the
mixer-reactor unit. To this "black box" model a basic form
of rate equation is then applied: mass flux equals a
conductance factor times an average driving force. For
cocurrent processes, it is known that the log mean average
109
-------
is the generally applicable averaging procedure for the driv-
ing force. The mass flux, Q, is expressed as Ib ozone
dissolved/hour, and the driving force as atmospheres ozone
partial pressure. Lacking a more specific term, the
conductance factor will here be called transference (Ib ozone
dissolved/hr/atm average ozone partial pressure driving
force), and it will be assigned the symbol T.
T = Q/Apm (17)
where Ap is the log mean average of the difference in ozone
partial pressures in the gas and in the liquid, at the mixer
inlet and the reactor outlet, in atmospheres.
At the mixer inlet of the i-th reactor, the partial pressure
driving force, Ap., , is as follows:
AP-L= P-jVg - Zi_1/H (18)
where P, is absolute gas pressure, atm
v is the volume fraction ozone in the gas from the
generator
Z._, is the dissolved ozone concentration from the
outlet of the preceding reactor
H is Henry's Law solubility constant for ozone
mg/l/atm
The relation between volume fraction and wt percent ozone
in oxygen is v = 2c/(300-c), or to a good approximation for
small values of c, v = 0.00667c.
Similarly, the partial pressure driving force,Ap2, at the
outlet of the reactor, is as follows:
Ap2 = P2vv ~ Zi/H <19>
where P2is tne absolute pressure in the ullage of the reactor
(suction pressure of the recycle gas compressor), atm
vv is the volume fraction ozone left in the spent gas
Z^ is the dissolved ozone concentration at the outlet
of the i-th reactor
110
-------
The log mean average is defined as follows:
Apm = (Ap-L - Ap2)/Ln(APl/Ap2) (20)
In writing the expression for Apm in terms of experimentally
measured variables, one small simplifying assumption can be
made immediately, namely, Z^ = Z^ This assumption has no
effect on Ap , and because the ozone partial pressure in the
gas is very much larger than that in the liquid at the mixer
inlet, the effect on Ap, is negligibly small. It will also
be convenient to express P, = P (1+d) . Also for conven-
ience, the relation between v and v can be expressed in
terms of the fraction of the ozone feed dissolved, f.
f = % - V/Vg (21)
Combining Eq (18) , (19) , (20) , and (21) , the log mean ozone
partial pressure driving force for the i-th mixer-reactor
combination is as follows :
P v (d+f)
Ap =
m Ln[ (HP2vg(l+d)-Zi)/(HP2vg(l-f)-Zi) ] (22)
For calculating T from Eq (17), Q can be easily calculated
from the experimental gas flow and concentration data. In
Eq (22), the value of d corresponds to the hydrostatic head
of about half the submerged depth of the dissolver tube,
eight feet. Since P? is about 2 psig, d is about 0.21. The
only variable in the Eq (22) for calculating Ap that is not
available from the plant data is the solubility constant, H.
Solubility and Half-Life of Ozone in Wastewater. Among the
objectives of the post-ozonation experiment was the direct
determination of ozone solubility in wastewater and the rate
that dissolved ozone disappears by the combined processes
of reaction with COD and autodecomposition. There appear to
be no published data for these properties. In fact, the
published data for the solubility of ozone in pure water are
111
-------
somewhat discordant.
Two sets of published data for the solubility of ozone in pure
water, designated G and M are shown in Fig 38.
Also shown as solid circles are points measured for distilled
water at 25°C in the post-ozonation apparatus which demonstrate
that the solubility measurements are reasonably reproducible
for pure water. For comparison purposes, the short curve
segment, A, has been drawn through the average of the four
experimental points and approximately parallel to the
published curves.
The rate at which dissolved ozone disappears from water can
be approximated by a first-order reaction.
Ln(Z/Z )= -kt (23)
O £i
For these kinetics, the reaction can be characterized by a
dissolved ozone half-life which is often easier to concept-
ually relate to actual processes than is the value of k .
Zi
At any point in the reaction, half the remaining dissolved
ozone will disappear in one half-life, three-quarters in two
half-lives, seven-eights in three half-lives, etc.
Half-life values for dissolved ozone obtained from the post-
ozonation apparatus for several different feeds (which are
defined in Table 2) and for distilled water are shown in
Figure 39 as a function of the COD remaining at the end of
the post-ozonation experiment. The data scatter rather
badly. This may possibly be caused by variation in dissolved
salt concentration, residual suspended solids, pH, etc.
Nevertheless, there is a clear trend for the half-life to
increase with decreasing residual COD. These data offer
some guide as to how long it will take for residual dissolved
ozone to disappear from the product water after ozone
treatment.
112
-------
OZONE SOLUBILITY, MG/L1TER/ATM.
ro -P> o> oo O r>
§O o o o c
o o o o c
v\
\
\
\^ G
N^>
•s^
V
^
3 10 20 30 4
TEMPERATURE, °C
^
0 5
i I i I i 1 i 1 i 1 i i i 1 i l i
40 60 80 100 120
TEMPERATURE, °F
Figure 38. SOLUBILITIES OF OZONE IN PURE WATER
113
-------
o 3B + F
A UN
O DC2 + L
D DC2
X DISTILLED WATER
40 60 80 100 120 140
HALF-LIFE OF DISSOLVED OZONE, MIN.
Figure 39. RELATION BETWEEN OZONE HALF-LIFE AND COD REMAINING TCN WATER
-------
The data for ozone solubility in well treated wastewater
obtained from the post-ozonation apparatus are also
scattered. But taken at their face value, they suggest
that ozone solubility decreases with increasing residual COD.
This is perhaps real, but much more likely, it indicates
mass transfer limitations in these experiments; i.e., the
dissolved ozone was being consumed at a rate comparable
to the dissolution rate from the gas. (However, before
starting routine operation of the post-ozonation apparatus,;'
tests were made on relatively low COD product water /(UN)
using several rates of ozone gas sparging. No correlation
between the measured solubility and the rate of ozone'
Sparging was observed, indicating that there was-no mass
transfer limitation.)
Since the half-life of dissolved ozone determined in each
experiment is a measure of the actual rate of ozone consumption,
the ozd-ne solubility in the absence of any ozone consumption,
and hence in the absence of any mass transfer limitation,
can be estimated by extrapolating the experimental data to
the limit of infinite dissolved ozone half-life. Such
extrapolations are commonly accomplished using reciprocal
-plots: extrapolating the reciprocal of a variable to zero
is equivalent to extrapolating the variable to infinity.
Ozone solubility in wastewater, expressed as percent of the
measured solubility in distilled water at the same
temperature, is plotted against the reciprocal dissolved
ozone half-life in Fig 40. Each point represents one run in
the post-ozonation apparatus, and points are included for
wastewaters subjected to several different pretreatments
(see Table 2).
The scatter of these data is such that several interpretations
115
-------
>96
O 3B-I-F
A UN
ODC2 + L
D DC2
O. 20
CO
200 10075 50 40 30 25 20
HALF-LIFE OF DISSOLVED OZONE, MIN.
Figure 40. OZONE SOLUBILITY IN WATER vs RECIPROCAL HALF-LIFE
-------
are possible. A straight line bracketed by dashed lines
approximating the scatter limits has been drawn through
the data in Fig 40. Extrapolation to zero reciprocal
half-life indicates a solubility in the range 80 to 93
percent of that in distilled water. However, one can
equally well interpret these scattered data as showing no
significant dependence of solubility on half-life for
half-lives greater than about 30 minutes.
Assuming that there are no gross sampling or analytical
errors reflected in these data, it appears that the solubility
of ozone in well-treated wastewater must depend on other
variables besides dissolved ozone half-life and temperature.
Other possible variables are dissolved salt concentrations,
suspended solids, and small differences in pH. But from the
fall of the experimental data, the average solubility appears
to be generally within the range 70% to 90% of the solubility
in pure water.
From these data and from some results of the mixer performance
correlation described below, the solubility of ozone in
these wastewaters was assumed to be 85% of the distilled
water solubility, i.e., 541 mg/l/atm at 70° F.
Correlation of Mixer Performance. Having selected a value
for solubility, sufficient data were available for calculating
transference of the mixers from plant data. Values of T
were computed using Eq (17) , (18) , (19) and (20) with one
additional refinement. When the water flow through a mixer
was greater than the feedwater flow, it was a combination of
the outlet from the preceding reactor and water recycled
from the bottom of the reactor. Since in these cases Z data
were available for both streams, the value of Z._^ in Eq (18)
was adjusted accordingly; the correction was very small.
117
-------
Most of the time, the mixers in the pilot plant were operated
at one of three water flows, about 20, 35, and 55 gpm.
Dividing the applicable part of the plant mixer data, intp
these three flow groups, the values of T and the mechanical,
work consumed by each mixer were computed. Because of the
suspect accuracy of the dissolved ozone analyses made prior
to 6/24/71, only plant data collected after that date were
used. Because of the masking effect of available chloripe
on the dissolved ozone analyses, the data for the MC+CL
pretreated feedwater were not used. No estimated values for
dissolved ozone concentrations or spent gas compositions
were used. Mechanical work was calculated from stream flows
and pressure drops assuming pump and adiabatic compression
efficiencies of 70 percent. Compressor work was taken as
that necessary to compress the gas from the pressure in the
reactor ullage to that at the mixer inlet.
Examination of the data showed that, on log-log coordinates,
values of T and mixer efficiency, transference/KW
mechanical work, fell along a straight line. Lines fitted
by least squares to each of the three flow groups, together
with the experimental points for the 35 gpm group, are shown
in Figure 41., Although not separately symbolized, the datav
points include data for different feeds, pH, and mixers.
Clearly all data fall .on a line, and they do= not scatter .too
badly. The three fitted lines suggest that at constant
efficiency, the transference is proportional to the flow
rate. And, it is evident that mixer efficiency increases
with increasing transference.
As a matter of additional interest, a zero gpm line is .
shown which was determined from a few runs made on UN-feed•
in whxch gas was simply blown down the dissolver tube and
allowed to ascend through the reactor as large, irregularly
118
-------
shaped bubbles. While comparatively little mechanical work
was consumed, the transferences were very low.
The least-square fitting of the data for the three flow
groups was conducted using a number of assumed solubility
Values. The overall best fits were obtained using values
of H in the 520 to 560 mg/l/atm range. This was one of the
reasons for selecting the 85% of distilled water value for
H. Fits were even tried with the assumption that H varied
with COD; however, the improvement in fit was negligibly
small.
On the basis of the data for the three flow groups, all of
the T vs T/KW data were fit to the following expression :
Using about the same modified multiple regression procedure
described above for obtaining Eq (12).
Ln T = aQ + a^LnfT/KW) + a2Wm (24)
where W is the water flow to the mixer, gpm
a = -3.558 + 0.062
o —
a± = 1.232 + 0.013
a2 = 0.03899 + 0.00054
The coefficient values represent a fit to 636 data points
after discarding 12 "wild" points, and the indicated uncer-
tainties represent one standard deviation. The standard
sample deviation in Ln T is 0.153, indicating that the data
Scatter 16% about the fitted value of T.
The effects of gas flow rates on mixer performance must
also be correlated. Rather than using gas flow directly,
the gas to liquid volumetric ratio, G/L (SCF gas/CF water)
Was used as the correlating parameter.
119
-------
10
20 30 40 60 80 100
TRANSFERENCE/ KW
200 300
Figure 41. PILOT PLANT MIXER TRANSFERENCE (T) vs
MECHANICAL EFFICIENCY (T/KW) FOR VARIOUS
WATER FLOW RATES
120
-------
Again grouping the data into three groups according to water
flow,.values for T and G/L were computed for each mixer from
raw plant data. These values are plotted in Figure 42 on
log-log coordinates for the 20 and 55 gpm flow groups. These
T vs G/L data exhibit considerable scattering. Part is due,
no doubt, to experimental error, but one might suspect that
variation in water properties, notably surface tension
variations caused by surfactants, could also be responsible
for part of the scattering.
In any event, the data for 20 gpm fall pretty much along a
straight line. The 35 gpm data (not shown) do likewise!.
But the 55 gpm data show a very pronounced curvature. No
other plotting coordinates were found which would make all
of the flow groups linear and also satisfy the physical
requirement that T approach zero as G/L approaches zero.
Thus, any correlating function fitting all of the data must
contain a term strongly nonlinear in W .
A number of correlating functions were examined using the
modified multiple regression procedure. The most reasonable
fit was obtained with the following equation.
Ln T = aQ + a^ + a2Ln(G/L) + a^3 (G/L) (25)
where aQ = 2.478 + 0.050
a-jL = 0.0122 + 0.0022
a2 = 0.4888 + 0.0270
a3 = 0.0000526 + 0.0000066
The coefficient values represent a fit to 638 data points
after discarding 10 "wild" points, and the indicated uncer-
tainties represent one standard deviation. The standard
sample deviation in Ln T is 0.306, indicating that the
data scatter 36% about the fitted value of T. The curves for
121
-------
A 17-25 GPM
• 52-58 GPM
0.2 0.3 0.4 0.6 0.8 1.0 2.0
G/L, SCFM GAS/CFM WATER
3.0
Figure 42.
PILOT PLANT MIXER TRANSFERENCE (T) vs
GAS TO LIQUID VOLUME RATIO (G/L) FOR
TWO WATER FLOW RATES
122
-------
20 and 55 gpm in Figure 42 represent Eq (25).
Equations (24) and (25) constitute a complete correlation
of the mixer performance based on the pilot plant data.
These two equations are shown graphically in Figure 43.
From Figure 43, it is evident that maximum transference and
maximum mixer efficiency are attained at the highest G/L
values. But, in the prepilot testing of the mixers, it was
found that when G/L was increased much above one, the
mixers tended to become hydraulically unstable. At a G/L
value of about one, the transference increases with increas-
ing water flow, at least up to the approximately 55 gpm
maximum examined in the pilot plant. And at G/L=1, the
mixer efficiency is nearly independent of flow rate;
transference is directly proportional to the mechanical
power expended.
Practically, it can be concluded that optimum mixer perform-
ance occurs at G/L=1, and the maximum transference verified
by extensive plant data is obtained at water flow of 55 gpm.
Under these conditions, a transference of about 80 can be
expected with a T/KW of about 110.
Vacuum Deaerator
The vacuum deaerator consisted of a 120 gallon cylindrical
vacuum chamber 30 in. in diameter by 36 in. high. The water
entered through a spray nozzle in the middle of the head.
The vacuum connection was concentric with the spray nozzle.
Vacuum was produced by a 1 h.p., single stage, liquid-ring
vacuum pump rated for 6.5 cfm at 27 in. mercury vacuum.
The maximum amount of gas coming out of the water at 27 in. Hg
vacuum is 2.3 cfm at 60 gpm feed rate.
Water level was maintained near the bottom of the tank by a
123
-------
100
20 30 40 60 80 100
TRANSFERENCE /KW
200 30»
Figure 43. CORRELATION OF PILOT PLANT MIXER PERFORMANCE:
T vs (T/KW) FOR VARIOUS GAS TO LIQUID VOLUME
RATIOS AND WATER FLOW RATES
124
-------
displacement-type level control which actuated the throttling
valve on the discharge of the transfer pump at the bottom of
the ^barometric leg.
Dryer
Theirecycle dryer used in the pilot plant consisted of two
activated alumina beds through which the gas flow alternated
about every 10 minutes. While one bed was on stream, the
other was being regenerated by the pressure-swing technique.
The water saturated recycle gas was first compressed to
about 55 psig and cooled to ambient temperature whereupon
about three-quarters of the moisture condensed and was
physically separated. The compressed gas then flowed
through the adsorbent bed for further drying to a dew point
of less than -60°F. A portion, about one-quarter, of the
dry compressed gas was expanded to low pressure and used
to purge the off-stream adsorbent bed for regeneration. The
purge gas was ultimately returned to the compressor inlet.
The remainder of the dry compressed gas was expanded through
a pressure regulating valve and fed to the ozone generator.
Pressure-swing adsorbent dryers are compact, inexpensive,
and simple and safe to operate. Thus, they are very well
suited for pilot scale operations. However, because of the
high gas pressures required, their operating costs are
comparatively high.
For large installations, it has generally been found more
economical to use a refrigerated predryer to reduce the gas
dew point to about 35°F by condensation followed by a heat
regenerated adsorption dryer. Such units are available in
many sizes from several different manufacturers.
125
-------
SECTION IX
ECONOMIC ANALYSIS OF LARGE PLANTS FOR
OZONE TREATMENT OF WASTEWATER
Introduction
Economic analysis is necessarily an iterative process.
Equipment sizes and process parameters must be specified
before cost estimates can be made, and the cost estimates
must be available before equipment sizes and process
parameters can be optimumly specified. The approach taken
here is to make initial cost estimates on the basis of
reasonable estimates of equipment sizes and process
parameters and later to linearly perturb these estimates
to arrive at more nearly optimum sizes and parameters.
Because the ozone dosage required for a constant quality
product water increases rapidly with increasing feedwater
COD (see Fig 25), ozone wastewater treatment appears to be
best suited for final, tertiary treatment. In the regime
of tertiary treatment, the capital and operating costs of
large ozone treatment plants will depend primarily on the
COD of the plant feedwater and the hydraulic load.
The feedwater for a large ozone treatment plant may come
directly from a secondary treatment process, or it may
have received varying degrees of tertiary pretreatment. Thus,
the feedwater COD might range from about 20 to 60 mg/1.
The required ozone dosage will vary with the specified feed
and product COD values. For initial cost estimating
purposes, however, the use of ozone dosage as an independent
127
-------
design parameter is somewhat inconvenient in that it requires
assumptions about mixer efficiencies. A more convenient
parameter is the ozone feed ratio, F*, which is defined as
the total milligrams of ozone generated per liter of water
treated (numerically equal to pounds ozone generated per
million pounds of water treated).
In order to examine the economic effects of varying ozone
requirements, plant cost estimates were prepared for three
different ozone feed ratios: 40, 75, and 100 mg/1. And to
examine the effects of plant scale, estimates for plants of
1, 10, and 50 million gallons per day (Mgd) nominal capacity
were prepared for each ozone feed ratio. (Siijice the overall
oxygen consumption partly depends on the actual COD
reduction, for this purpose, COD reductions of 10, 20, and
25 mg/1 were assumed for the ozone feed ratios of 40, 75,
and 100 mg/1, respectively.)
The cost estimates for large plants are based,on a direct
scale-up of the pilot plant equipment with only two minor
changes. First, a combination of a refrigerated predryer
and a heat-regenerated adsorption dryer was selected instead
of the pressure-swing adsorption dryer. And second, three
instead of six reactors are used. Each of the reactors has
a 15 minute detention time for a total treatment time of
45 minutes, and the product holding tank has a detention
time of 15 minutes.
The cost estimates for all of the projected plants are
based on the following assumptions.
1. The plant will operate continuously at designed
load.
2. At least two parallel trains of equipment are
required for process reliability.
128
-------
3. All mechanical and electrical equipment is to be
inside buildings.
4. A 20% excess ozone generating capacity will be
installed.
5. Utility connections are available at site.
6. No feedwater pH adjustment is required.
7. Land costs are not included.
8. No service or administrative buildings are
included.
The basic design parameters used for initial estimating
purposes are summarized in Table 11. These cost estimates
are generally comparable with published estimates for other
tertiary treatment processes. However, these estimates are
intended to be conservative. For example, the capital
costs of major equipment were, where possible, taken as the
average of two or more quoted costs instead of the lowest
quoted cost.
One Mgd Plant
A schematic diagram of a one million gallon per day plant
is illustrated in Fig 44. Two trains (only one is shown)
of four, above-ground, bitumastic lined, cylindrical steel
tanks are employed. The first three tanks in each train are
reactors, and the fourth is the product holding tank. Each
tank is 8 ft diameter. The first tank in each train is 25 ft
high, and the remainder are 20 ft high. The additional
height of the first reactor is for handling possible
surfactant foam.
The process arrangement is essentially that used in the
pilot plant. In order to accommodate variations of
feedwater flow rate, water recycle around the first reactor
is automatically controlled by a float valve in the sump
129
-------
TABLE 11
DESIGN PARAMETERS FOR INITIAL COST ESTIMATES
FOR LARGE OZONE WASTEWATER TREATMENT PLANTS
Feedwater flow, 106 gal/day 1 10 50
Parallel trains of three,
15 min reactors 2 44
Ozone concentration, wt % 1.7 1.7 1.7
Ozone installed capacity, Ib/day
F* = 40 mg/1 400 4,000 20,000
F* = 75 mg/1 750 7,500 37,500
F* = 100 mg/1 1,000 10,000 50,000
Oxygen consumption, Ib/day
F* = 40 mg/1 450 4,500 22,500
F* = 75 mg/1 520 5,200 26,000
F* =100 mg/1 580 5,800 28,800
Electric power, Kw
F* = 40 mg/1 150 1>500 7,600
F* = 75 mg/1 220 2,200 10,80b
F* =100 mg/1 250 2,500 12,700
Gas flow, std ft /min
F* = 40 mg/1 160 1,600 8,000
F* = 75 mg/1 310 3,100 15,500
F* =100 mg/1 410 4,100 20,500
130
-------
u>
RECYCLE GAS
OZONE MIXTURE,
±
II
II
II
It
ll
II
11
-------
VACUUM
PUMP -
ROOF
H-
LO
VACUUM DEAERATOR
PRV COMPRESSOR
REFRIGERATED DRYER
OXYGEN
STATION
STAGE|
O
fc
I
1
i
1
INJECTOR-MIXERS
•4
F<
SVOX
f
*
TI-MZ
I J
[t 4^ 1
INTERCONNECTION
WITH OTHER
UNITS
ADSORPTION
DRYER
OZONE
GENERATOR
E INTERSTAGE
PARTITIONS
ELEVATION OF ONE OF FOUR PARALLEL
REACTOR UNITS
TREATED WATER
TO OUTFALL
LIST OF ABBREVIATIONS
I-M-INJECTOR-MIXER
L.C. -LEVEL CONTROL
P- PUMP
PRV. - PRESSURE REGULATING VALVE
S. V. - STOP VALVE
Figure 45. FLOW DIAGRAM OF 10 MGD WASTEWATER TREATMENT PLANT
-------
feeding the first stage pump. The water recycle-bypass
arrangement on the other reactors is the same used in the
pilot plant which allows the mixer water flows to be varied
independently of the feedwater flow.
10 and 50 Mgd Plants
Elevation and plan views of a projected 10 million gallon
per day plant are shown in Pig 45 and 46. The process flow
pattern and tank depths are the same as for the 1 Mgd
plant. Four parallel trains of four, rectangular, reinforced
concrete tanks are used: for larger plants, concrete tanks
are less expensive than the bitumastic lined steel tanks.
The concrete tanks form two walls of a building which houses
the mechanical and electrical equipment on two floors. The
lower floor contains the pumps, compressors and dryers, and
the upper floor contains the ozone generators. The vacuum
deaerators are mounted on the roof of the structure, and
the vacuum pumps and ventilating fans are in an adjoining
penthouse.
The layout of a 50 Mgd plant would be a direct scale-up
of the 10 Mgd plant.
Costs of Major Equipment
The capital costs for the major items of process equipment
are shown in Fig 47 through 52. These costs are based on
vendors' quotations FOB factory during the fourth quarter
of 1971. In several of the figures, the range of prices
quoted by several manufacturers and the average of these
prices are shown.
Costs of centrifugal pumps for driving the mixers are shown
as a function of pump capacity in Fig 47. Ozone generator
prices are shown in Fig 48 as a function of capacity when
133
-------
WASTEWATER
IN
OJ
i*!
\ V
DEAERATION TANKS
(ON ROOF)
ROOF
VACUUM PUMP
a
VENTILATION
PENTHOUSE
I i
I i
i i
LEGEND
COMPRESSOR
COOLING WATER
LEVEL CONTROL
OZONE GENERATOR
PUMP
VACUUM DEAERATOR
TREATED
WATER TO
OUTFALL
Figure 46. MULTILEVEL PLAN VIEW OF 10 MGD WASTEWATER TREATMENT PLANT
-------
COST OF PUMP,
1,000 DOLLARS PER UNIT
00
TJ
C
•o
o
•O
O
O
O
O
o
-0
en
•X
m
>
o
p
b
fc
\V
Ul
B
Ul
O
U
H
P4
<
U
PM
t4
H
W
O
01
M
3
tn
•K
135
-------
OZONE GENERATOR PRICE, 1000 DOLLARS
— re
b c
- o o c
??0 0 0 C
RAr
Av/r
AVL
JGE OF GENERATOR PRICES
:RAGE PRICE
/
/
x
/
/
,
/
/
X
/
/
x '
/
r
x
X
/
-
X
;
Xx
X
X
/
/
/x
X
/x
x '
y
/
/
X
, X
/
x
/
X
-
/
X
/'
x
/
/
/
/
x
>o 1,000 topoo
GENERATOR CAPACITY, LB. OZONE / DAY
Figure 48. OZONE GENERATOR CAPITAL COSTS
136
-------
IV
\-
COST OF INJECTOR-MIXERS, 1000 DOLLARS PER UNI
0 .-
0- 0
i
— —
- — R
— A
1
ANGE OF COSTS
VERAGE COST
'
;
^
^
f
^
^
^
^
^
^
>
^
**
s'
•^
^
^
**
-^
. —
>^
^
^
^
^
-------
100
ir
<
8
o
o
o
10
a:
o
w
to
LU
(T
a.
2
o
O
cn
o
u
O.I
_
- RAr
A\/C
«V L
^-
S
C(
'
^x
x^
X
X
DNDIT
S
DI9
JGE OF COSTS
'.RAGE COST
*•*
x^
X
'
ION
UCT
CHA
"
^
r- LIQUID
\
X
s:
ION P
RGE P
^
X
*-
x
-F
X
*C
>T
\
RING COMPRESSOR
ARY LOB
WITH AFT
E COI^/
ERCO
RESSURE: o PSIG
RESSURE1I5 PSIG
1PRE
OLEI
:ss
3
OR
1.0
COMPRESSOR CAPACITY, 1000 SCFM
10.0
Figure 50. OXYGEN COMPRESSOR CAPITAL COSTS
.138
-------
z
Z)
£T
UJ
CL
CO
o:
<
_J
_J
o
Q
O
•z.
<
CO
o
X
£
o
CJ
CE
Ul
i
10
1
0
ADS
(10 C
^
ORPTI
)EWPG
ON
!NT=
, — — -
^x-~
DR
F'™" t
^
X
^
YE
Q*
X
R
Fl
^
x
~1
-
I
X
^^
'RIGERATiEl^ Rf$|r-~ \
yl^O 41 4C;)1* p )' ' ' '.'
i
's*Tf R
| | |<
GAS FLOW. THOUSAND STIR. JSU- PiT, PER NIN-
.Figure 51, GAS DRYER CAPITAL COSTS
139
-------
1000
o
o
o
o
o
Ui
h-
yj
CE
UJ
t
i
8
fe
CO
O
O
100
10
F
\
*ANGE OF COMPLETE RECYCLE SYSTEMS COSTS
WERAGE COST
X
x*
/
x
X
X
xl
/
x^
X
X
/
,/
s
/
/
/
/
s
/ /
/ / \
/ / ,'
/ ,'
r X"
/
f
/
•'/,
x
X"
s
/
V
!/
/
/
/
/
/
/
/
/
>
x
jf
s
/
/
/
/
-
100
1,000
OXYGEN FLOW, SCFM
10,000
.Figure 52.
CAPITAL COSTS OF INTEGRATED OXYGEN
RECYCLE SYSTEMS
140
-------
generating 1,7 wt percent ozone in oxygen. The costs of
injector mixers are shown as a function of nominal water
flow rate in Fig 49. Prices for two types of oxygen
compressors, liquid ring compressors as used in the pilot
plant and rotary lobe compressors with aftercooler, are
shown as a function of capacity in Fig 50. Capital costs
for refrigerated predryers and heat-regenerated adsorption
dryers are shown in Fig 51. Capital costs for integrated
gas recycle systems - - compressors and dryers - - are shown
in Fig 52.
Capital Cost Estimates
The capital cost estimates for nine combinations of plant
sizes and ozone feed ratios are summarized in Table 12.
Costs for the major items of process equipment were taken
from the average curves in Fig 47 through 51, and allowance
was made for shipping and erecting. Costs of foundations,
buildings, piping, instruments, and electrical work were
estimated from approximate plant layouts and pipe drawings
using either vendors' quotations or published estimating
tables (16~19'. Engineering, contractor's overhead, and
fees were taken as 30% of direct materials and labor.
The estimated costs reflect average construction costs for
the fourth quarter of 1971. These estimates are intended to
be conservative, but costs at specific locations may easily
vary + 35% from such general estimates.
Operating Costs
Estimates of the operating costs for the nine combinations
of plant sizes and ozone feed ratios are summarized in Table
13 and are shown graphically in Fig 54. A percentage break-
down of estimated operating costs is shown in Table 14. A
graphical break-down of the operating costs of large plants
141
-------
Table 12
CAPITAL COST ESTIMATES FOR PLANTS TO
TREAT WASTEWATER WITH OZONE
Plant Capacity, mgd
Ozone Feed Ratio, Ib 03/106 Ib I
Item
Deaerators
Process pumps
Injector mixers
Reactors & Holding tanks
Oxygen compressors
Dryers
Ozone generators
Sumps & Drains
Ozone decomposer
Piping
Electrical
Instrumentation
Painting
Site work
Equipment supports & building
Total Materials & Labor
Engineering, Contractor's
overhead and fee
1
120 40
Capital
8
20
7
70
8
11
53
10
1
34
20
21
5
10
50
328
98
10
40
Costs, 1000
51
67
31
200
48
64
440 1
35
6
110
120
44
10
20
140
1,386 5
416 1
50
40
Dollars
160
390
140
450
230
310
,980
85
15
480
530
80
40
60
800
,750
,725
Total Capital Cost 426 1,802 7,475
142
-------
Table 12 continued
CAPITAL COST ESTIMATES FOR PLANTS TO
TREAT WASTEWATER WITH OZONE
Plant Capacity, mgd
Ozone Feed Ratio, Ib Q.,/10 Ib H,
j i
Item
Deaerators
Process pumps
Injector mixers
Reactors & Holding tanks
Oxygen compressors
Dryers
Ozone generators
Sumps & Drains
Ozone decomposer
Piping
Electrical
Instrumentation
Painting
Site work
Equipment supports & building
Total Materials & Labor
Engineering, Contractor's
overhead and fee
Total Capital Cost
1
,0 75
Capital
8
20
7
70
10
13
94
10
1
34
24
21
6
10
50
378
113
491
10
75
Costs, 1000
51
67
31
200
76
104
770
35
6
136
140
44
10
20
150
1,840
552
2,392
50
75
Dollars
160
390
140
450
280
390
3,465
85
15
520
550
80
40
60
800
7,425
2,227
9,652
143
-------
Table 12 continued
CAPITAL COST ESTIMATES FOR PLANTS TO TREAT
WASTEWATER WITH OZONE
Plant Capacity, mgd
Ozone Feed Ratio, Ib O3/106 Ib H,
Item
Deaerators
Process pumps
Injector mixers
Reactors & Holding Tanks
Oxygen compressors
Dryers
Ozone generators
Sumps & Drains
Ozone decomposer
Piping
Electrical
Instrumentation
Painting
Site work
Equipment supports & building
Total Materials & Labor
Engineering, Contractor's
Overhead and fee
Total Capital Cost
1
?o 100
Capital
8
20
7
70
12
15
122
10
1
34
26
21
6
10
50
412
124
536
10
100
Costs ,
51
67
31
200
90
122
980
35
6
136
153
44
10
20
150
2,095
628
2,723
50
100
1000 Dollars
160
390
140
450
380
490
4,410
85
15
520
600
80
40
60^
800
8,620
2,586
11,206
144
-------
with an ozone feed ratio of 100 mg/1 is shown in fig 55.
The operating cost estimates are based on continuous opera-
tion at designed capacity. Electrical power costs were
assumed to be 1.4, 1.1, and 0.9 £/Kwh for the 1, 10, and 50
Mgd plants, respectively. Oxygen costs are based on
typical merchant liquid oxygen prices (including customer
station rental) which are shown as a function of usage
rate in Fig 53.
Plant amortization is based on 25 years life and 5% annual
interest which corresponds to an annual amortization charge
of 7.1 percent of capital cost. Maintenance costs were taken
as one percent of capital costs per year. Incremental
operating labor was assumed to be 4 man-hours/day for the
1 Mgd plant and 8 man-hours/day for the 10 and 50 Mgd plants
at $5.00 per man-hour plus 30% for overhead and supervision.
Cost Optimization
All of the preceding cost estimates are based on an initial
set of reasonable, but assumed plant parameters. Now, in
order to arrive at more nearly optimal plant parameters, a
procedure for linearly perturbing the initial estimates to
minimize the projected operating costs will be described and
illustrated.
Once the requirements for a specific ozone wastewater
treatment plant are established, a number of plant parameters
should be concurrently adjusted to yield a minimum total
treatment cost. Many such parameters can be optimized only
for specific situations and specific plant locations.
However, three plant parameters can be optimized for more
general cases: the concencentration of ozone generated (c ),
the mixer sizes and flow rates, and the distribution of
145
-------
LIQUID OXYGEN PRICE, DOLLARS/TON_
p
ho
1
O
m
o
m
ro
o
o
o
09 O
O O
Z
7
o
c
H
EH
O
CQ
W
O
>H
X
O
p
H
p)
a
H
o
z
x.
o
o
CO
EH
en
o
o
U
H
O
in
QJ
5-1
3
tn
•H
En
146
-------
TABLE 13
ESTIMATED ANNUAL OPERATING COSTS
OF OZONE WASTEWATER TREATMENT PLANTS
IN THOUSAND DOLLARS
1 MILLION GAL/DAY PLANT
F*, mg/1 =•
Electric power
Amortization
Oxygen
Operation & Maintenance
Total
0/1000 gal
40
19
30
9
15
73
20.0
75
26
35
11
16
88
24,1
100
31
38
12
16
97
26.6
10 MILLION GAL/DAY PLANT
F*, mg/1 =
Electric power
Amortization
Oxygen
Operation & Maintenance
Total
0/1000 gal
40
147
128
37
41
353
9.7
75
208
170
44
47
469
12.8
100
245
193
48
50
536
14.7
50 MILLION GAL/DAY PLANT
F*, mg/1
Electric power
Amortization
Oxygen
Operation & Maintenance
Total
0/1000 gal
40
600
531
135
99
1365
7.5
75
855
685
159
120
1819
10.0
100
1002
796
174
136
2108
11.6
147
-------
TABLE 14
PERCENTAGE BREAKDOWN OF
ESTIMATED ANNUAL OPERATING COSTS
OF OZONE WASTEWATER TREATMENT PLANTS
1 MILLION GAL/DAY PLANT
F*, mg/1 =
Electric power
Amortization
Oxygen
Operation & Maintenance
40
26.0
41.2
12.3
20.5
75
29.5
39.8
12.5
18.2
100
31.9%
39.2
12.4
16.5
10 MILLION GAL/DAY PLANT
*F, mg/1 =
Electric power
Amortization
Oxygen
Operation & Maintenance
40
41.6
36.3
10.5
11.6
75
44.4
36.2
9.4
10.0
100
45.7%
36.0
9.0
9.3
50 MILLION GAL/DAY PLANT
F*, mg/1 =
Electric power
Amortization
Oxygen
Operation & Maintenance
40
44.0
38.9
9.9
7.2
75
47.0
37.7
8.7
6.6
100
47.5%
37.8
8.2
6.5
148
-------
100
e>
o
o
o
\
en
z
UJ
o
I-
co
8
10
UJ
UJ
CE
h-
cc
UJ
I
UJ
fc
i
NOTES:
I. COSTS INCLUDE AMORTIZATION, OPERATION AND MAINTENANCE
2. AMORTIZATION-5%, 25 YEARS
3. THREE OZONE FEED RATIOS ARE INDICATED BY THE CFRCLED
NUMBERS. THE FEED RATIO IS DEFINED AS LB. OZONE FED
PER MILLION LB. WATER.
J L
10
WASTE WATER TREATED, MOD,
50
100
Figure 54.
TOTAL OPERATING COSTS FOR OZONE TREATMENT
OF WASTEWATER AT DIFFERENT OZONE FEED
RATIOS
149
-------
. 100
o
o
o
o
i
UJ
o
UJ
UJ
a:
\-
(T
UJ
UJ
h-
cn
u_
o
en
o
o
10
1.0
0.
TOTAL COST
ELECTRIC POWER COST
OPERATIONS
AND MAINTENANCE
OZONE FEED RATIO : 100 LB/IO6 LB WATER
1.0
10
100
WASTE WATER TREATED, MOD
Figure 55. BREAKDOWN OF UNIT COST FOR TREATING
WASTEWATER WITH OZONE
150
-------
dissolved ozone concentrations (Z.) along the train of reactors
As was illustrated in the discussion of the COD correlations,
once the COD and flow requirements of an ozone wastewater
treatment plant are specified, the quantity of ozone to be
dissolved can be calculated. With the additional specifi-
cation of treatment time and feedwater pH, the time weighted
average dissolved ozone concentration (Z*) can be calculated.
For cost optimization, one further requires estimates for
ozone production costs as a function of c , correlations
for the performance of the mixers, and some simple method
for characterizing the Z. distribution. (Note that estab-
lishing the value of Z* does not uniquely define the Z.
distribution.)
Ozone Production Costs. Ozone production costs depend on
the concentration generated, scale of production, power
rates, and the specific type of generators used. The costs
associated with producing a pound of gaseous ozone are
taken to be all those power and capital amortization costs
required to deliver a pound of gaseous ozone in oxygen to the
gas-liquid mixers at 11 psig, assuming recycle oxygen at
2 psig and a 4 psi pressure drop in the gas system. Specif-
ically included are the amortization and power requirements
for the gas compressors, gas dryers, and ozone generators.
The scale of production will be assumed to be that of the 10
mgd plant with an ozone feed ratio of 75 Ib/million Ib water,
described in the initial cost estimates, which has an installed
ozone generating capacity of 7500 Ib/day. Linear scaling
rules will be used for adjusting amortization and power
costs from this base. Power rate is taken as 1.1^/Kwh and
amortization charge as 7.1% of capital investment/year.
151
-------
Ozone generation costs are taken from two sources: equip-
ment price quotations for units generating 1.7 wt % ozone
in oxygen, and the operating characteristics of the pilot
plant ozone generator. Power requirements were taken from
Figure 36 (the 20 scfm curve), and equipment capacity as a ;
function of c was calculated from the same data assuming a ,
maximum 10 Kw input to the pilot plant generator. Gas
compression and drying costs were taken from price quota-
tions and manufacturers1 equipment specifications.
The resulting estimated costs for producing a pound of gaseous
ozone are shown as a function of c in Figure 56. For
mathematical purposes of optimization, precise and smooth
production costs are required. Therefore, the total produc-
tion cost curve is shown on expanded scale in Figure 57.
As the ozone concentration increases, the costs for
generator power and amortization increase. But because
higher concentrations decrease the volume of gas to be
handled, the costs of compressing and drying decrease with
increasing concentration. These opposing trends combine to
yield a shallow minimum in the total production cost at
about 1.7 wt % ozone.
The magnitude of the various components of ozone production
costs is noteworthy. The major part is associated with the
ozone generators. The largest fraction is ozone generator
power costs, and the next largest is generator amortization.
Costs for compressing and drying the recycle gas vary from
20% of the total cost at 2 wt % ozone to 6% at 4 wt % ozone.
The statement is often made that producing ozone from pure
oxygen is less expensive, even including the cost of oxygen,
than is production from air. The total production costs
152
-------
shown here offer a basis for quantifying that statement,
Generallyf an ozone generator operating at constant power
will produce about half the concentration of ozone operating
on air that it will produce operating on oxygen. Thus, the
production of a pound of ozone from air will consume twice
the power and require twice as large generators as would be
required for production from oxygen. The volume of gas to be
compressed and dried would also be.doubled using air. Thus
the cost of producing a pound of ozone at 1% concentration in
air would be about 16C compared to 8£/lb at 2% ozone using
oxygen. Comparison of production costs at the same ozone
concentrations are even more favorable for oxygen.
The substantial savings in ozone production costs are partly
offset by the additional cost of oxygen. In ozone wastewater
treatment, oxygen is consumed both by COD reduction, which is
proportional to ozone dosage, and by the increase of dissolved
oxygen (DO) concentration, which is proportional only to the
quantity of water treated. For the 10 Mgd plant with F* of
75 mg/1, the oxygen costs are about 1.6^/lb ozone produced,
about 1/3 being attributable to COD reduction and 2/3 to
increasing the DO concentration. Thus in this case, the
additional oxygen cost is small compared with the savings in
production costs.
For very small ozone doses, however, most of the oxygen
consumption would be from increasing the DO concentration, and
the oxygen cost per pound of ozone produced might be greater
than the savings in ozone production costs. It must be
emphasized that the additional oxygen costs would not be
without benefit; for the high DO content in the product water
is generally very desirable. And of course, it is technically
feasible- to recover a major part of the DO by vacuum deaera-
tion after ozone treatment.
153
-------
20
6.0
4.0
LU
Z
o
82.0
m
LJ
u
_Q6
X
o
0.4
0.2
O.I
0.08
2 3
WEIGHT PERCENT OZONE
Figure 56. ESTIMATED TOTAL PRODUCTION COST OF
OZONE IN A TYPICAL WASTEWATER
TREATMENT PLANT
154
-------
14
CD
13
o
o
II
o
010
o
or
o_
o
g
2 3
WT. PERCENT OZONE
Figure 57.
EXPANDED GRAPH OF ESTIMATED TOTAL
PRODUCTION COST OF OZONE
155
-------
Optimizing Mixer Sizes
Optimizing the size and operating parameters of the gas-
liquid mixers-is based on scaling the mixers used in the
pilot plant. In this section, primed (') symbols will be
used for the full scale plant variables and not-primed ;
symbols will signify a pilot plant variable. Since the
pilot plant mixers were found to operate most effectively
at 55 gpm water flow, a plant scaling factor X' will be
P
defined on that flow basis.
X' = (designed plant feed, gpm)/(55 gpm) (26)
In order to allow adjustment of (G/L) for each mixer, a
mixer scaling factor X' must be defined.
X^ = (flow through mixer)/(plant feed flow) (27)
X^ is also the recycle ratio for a mixer. If X1 is greater
than one, part of the mixer flow is recycle from the bottom
of the reactor. If X' is less than one, part of the feed is
by-passed to the bottom of the reactor. In the pilot plant,
X^ values from 0.54 to 2.75 at mixer flows of 20 and 35 gpm
were used. There was no evident effect of X' on transference,
m
and there would presumably be none at 55 gpm.
The basic mixer scaling rule is as follows.
T' = X^T (28)
For the i-th stage, the value of D.^ (Ib 0_ dissolved in the
i-th stage/lb COD feed) can be calculated as shown in the
discussion of the COD correlations. The actual pounds of
ozone to be dissolved per hour, Q!, is as follows.
Q' = 0.02749 DiLfXI (29)
From the definition of T' = Q'/Ap , the required average
partial pressure driving force, Ap , is as follows.
Apm = 0.02749 D-L/T (30)
156
-------
Using Eq (30) and the base mixer flow, the gas to liquid
volume ratio is as follows:
(G/L) = 0.07526 D^/X^c f (31)
where c is the wt % ozone in gas
f is the fraction of feed ozone dissolved.
Solving Eq (30) and (31) for X^/D^ and equating the
results, the following equation is obtained in which it is
noted that Apm is a function of c , f and Z. [See Eq (22)]
cgf/APni(cg,f,Z) = 2.7376 T/(G/L) (32)
Notice that all scaling factors have disappeared from Eq (32)
A simple function of c , f, and Z is proportional to the
ratio of transference to G/L for any size mixer in any size
plant.
A numerical inspection of c f /Apm shows it to increase when
f increases. Since maximum values for f are desired, it
follows that the maximum value of T/(G/L) is desired. An
inspection of the mixer performance correlation, Figure 43,
shows this maximum for the pilot plant mixers to be at 55 gpm
and G/L = 1 where T = 80. Substituting the latter two values
into Eq (32) the following is obtained.
cgf/Apm(cg,f,Z) = 219 (33)
Equation (33) offers a basis for optimizing the mixers. For
specific values of c and Z, Eq (33) can be solved for f, the
fraction of feed ozone that can be dissolved by mixers of the
type used in the pilot plant operating at optimum conditions.
Having f, the value of X1 can be calculated directly from
Eq (31) using (G/L) = 1.
Unfortunately, there is no analytic solution for Eq (33).
It is transcendental in f and c and therefore requires an
157
-------
iterative solution. For convenience, values of f for various
values of c and Z are given in Table 15 for a typical
situation in which ?2 = 2 psig, d = 0.21, and H = 541 mg/l/atm.
From this table, it is evident that the fraction ozone dis-
solved increases with increasing c and decreasing Z, exactly
as a cocurrent mixer should behave.
The appropriate mixer power is simply that for liquid
pumping; the compression work for the gas has been included
in the ozone production cost. At 55 gpm and G/L = 1, the
pilot plant mixers exhibited a 17-1/2 psi water pressure
i
drop (and a gas pressure drop of 9 psig). The mixer power ,
i
costs at l.lC/Kwh and a pump efficiency of 0.7 is as follows.
Pumping costs (C/1000 gal feed) = 0.2 X^ (34)
As several cases are examined, it will be seen that the
pumping costs are very small compared to the ozone production
costs.
Dissolved Ozone Distribution Ratio (S)
For a specified technical performance of an ozone wastewater
treatment plant, the COD reaction rate correlation, Eq (13),
yields only the time weighted average dissolved ozone
concentration, Z*, that is required. But even with Z*
specified, the values of Z^ for each of the N reactors may
be varied subject only to the constraint of Eq (2) or Eq (3),
Since the ozone dosage requirement, D., will generally be
different for each of the reactors, there should be, for
every combination of c and mixer sizes, one optimum Z.
distribution that yields a minimum total operating cost.
For the purposes of this discussion, only a linear distribu-
tion for equivolume stages will be considered.
zi = zi + fcr'W 05)
158
-------
2345
Z, mg/LITER
Figure 58. PERCENT OF-OZONE FEED DISSOLVED IN
TYPICAL MIXER
159
-------
TABLE 15•
PERCENT OF OZONE FEED DISSOLVED IN TYPICAL MIXER
OUTLET DISSOLVED OZONE CONCENTRATION* MG/L
WT PCT •
OZONE 0.30 0.50 0.70 1.00 1.50 2.00 2.50 3.00 3.50 4.00 5.00 6.00 8.00
1.00 78.91 74.94 ?0.97 65.01 55.05 45.0& 35.00 24.81 14.28 2.70 0.00 0.00 0.00
1.25 80.il 76.95 73.77 69.00 61.06 53.09 45.11 37.08 28.96 20.72 2.78 O.UO 0.00
1.50 80.94 78.30 75.65 71.69 65.08 58.45 51.82 45.15 38.46 31.74 17.98 2.86 0.00
1.75 81.53 79.26 77.01 73.61 67.94 62.27 56.59 50.90 45.20 39.47/27,91 16.02 0.00
/
(—•
° 2.00 81.98 80.01 78.03 75.06 70.10 65.15 60.18 55.21 50.23 45.,24 35.20 25.04 3.03
2.25 82.32 80.58 78.82 76.17 71.78 67.^9 62.98 58.57 54.14 49.72 40.84 31.89 13.40
2.50 82.62 81.05 79.47 77.08 73.13 69.17 65.22 61.25 57.28 53.31 45.33 37.32 21.02
2.75 fi2.87 81.42 79.99 77.83 74.25 7Q.65 67.05 63.46 59.83 56.24 49.01 41.74 27.07
3.00 83.06 81.76 80.44 78.47 75.18 71.88 68.59 65.28 61.98 58.67 52.Q7 45.42 32.Q5
3.25 83.25 82.03 80.81 79.00 75.96 72.91 69.89 66.83 63.80 60.75 54.64 48.54 36.23
3.50 83.41 82.28 81.15 79.46 76.65 73.83 71.00 68.18 65.35 62.52 56.86 51.20 39.81
3.75 83.55 82.48 81.43 79.85 77.22 74.60 71.98 69.34 66.71 64.07 58.79 53.51 42.90
4.00 83.65 82.68 81.70 80.22 77,74 75.28 72.83 70.37 67.90 65.42 60.49 55.53 45.61
-------
where Z^ is the outlet dissolved ozone from the
i-th stage in a N stage plant.
Each such linear distribution can be characterized by a
distribution ratio.
S = ZN/Z1 (36)
From Eq (35) and (36) and the rules for summing finite
series, the following relations can be written for the
i-th stage of a N stage plant.
iN N — 1
7, * = 7 * I" 1 4- "*" ~ I c _ "M/91/M 4- n x, {C. — 'Ml IT Q \
" -! " n; L J- ^ vf T~ l>3 — J.;/zJ/[± T U.D(.o — L) J \-JO^
1 N N — 1
Simultaneous Optimization of c , X', and S
On the basis of the estimated cost of ozone as a function
of c , the mixer scaling correlation, and the linear dis-
solved ozone distribution, a procedure for simultaneously
optimizing c , the mixer sizes, and the dissolved ozone
distribution can be described.
Since not all of the plant operating costs will be influenc-
ed by the optimization procedure, it is convenient to divide
the plant operating costs into three groups:
1. adjustable costs - - those costs strongly
influenced by adjusting c , X^, and S which are
the total ozone generation costs and mixer pump
.power,
2. oxygen costs, and
3. nonadjustable costs - - the remaining operating
costs which include labor, maintenance, overhead,
and amortization and power costs for all equipment
other than ozone generators, compressors, and
dryers. (Mixer pump amortization is considered a
nonadjustable cost, and .pumping power adjustable.)
161
-------
The optimization procedure described here is based on linear
cost adjustments from the initial plant estimates. For these
small adjustments of plant operating parameters, it is
assumed that both the ozone production cost (C/lb ozone) and
the nonadjustable costs remain constant.
Certain overall plant parameters must be specified initially.
These are feedwater flow, COD,, and pH; product water COD,
treatment time, and the number of reactors (N). From these
parameters, the required value of D* can be calculated from
Eq (11), the required Z* from Eq (12), and the value of X'
from Eq (26) .
From this point, the optimization procedure is doubly
iterative in c and S as follows.
g
a. For a trial value of S in the first iteration, Z. and
Zt values are obtained from Eq (37) and (38).
b. Applying Eq (12) or Fig 28 and 29 sequentially to the
N reactors, the values of D? are obtained for each
reactor, and noting from Eq (7) that D. = D* - D* , the
values of D, are calculated. The values of Q! can then
be calculated from Eq (29) which for convenience can be
placed on a 1000 gal basis as follows.
Q|(lb 03/1000 gal) = 0.00833 L^ (40)
c. In the second iteration for each assumed value of S, a
value of c is assumed. For each reactor, the value of
f. for the trial values of c and Z. is obtained from
i g i
Table 15.
d. Taking (G/L) = 1, the value of X^ for each stage is
obtained from Eq (31).
e. The ozone feed to each stage is Q!/f., and the cost is
calculated using Fig 57.
f. Pumping power costs for each stage are calculated from
162
-------
Eg, (34)
g. The adjustable costs fop: each stage are the sum of, items
e and f.
h. The adjustable costs for the plant for the trial values
of c and S are the sum of item g for the N stages in
the plant.
The optimum values of c and S are those yielding the
minimum costs in item h. The total operating cost is
obtained by adding the oxygen cost and nonadjustable cost
to the minimized adjustable costs. For a 10 Mgd plant, the
oxygen costs are about as follows.
Oxygen cost («/1000 gal water) = 0.0167(Lf - L +40) (41)
Based on the initial cost estimates for a 10 Mgd plant, the
nonadjustable costs are about 5^/1000 gal water.
The overall fraction of ozone dissolved, f*, is as follows.
N N
f* = D*L /F* = EO! / Z(QVf. ) (42)
Specific Example
Continuing the example started in the discussion of the COD
correlations, consider a 10 mgd plant receiving 40 mg/1 COD
at pH 7 with a desired reduction to 20 mg/1 COD in 45 min-
utes in a three-stage plant. The required value of D* is
1.35 Ib ozone dissolved/lb COD feed and Z* is 2.32 mg/1.
X1 is 126.3
Now to follow the procedure outlined above.
a. Assume S = 1.7
Z., = 1.72, Z9 = 2.32, Z- = 2.92 mg/1
_L ^- -^
Z* = 1.72, Z* = 2.02, Z* = 2.32 mg/1
163
-------
b. D* = 0,689, D* - 1.033, D* = 1.34 Ib 03/lb COD feed
D = 0.689, D2 - 0.344, DS « 0,305 Ib Og/lb COD feed
Qj_ = 0.230, Q£ = 0.114, Q3 = 0.102 Ib 03/1000 gal feed
c. Assume c = 2.0 wt % ozone in oxygen.
From Table 16 for cg = 2.0 and Z1 = 1.72,
the value f, = 0.679 can be interpolated.
Similarly, f2 = 0.619 and fg = 0.559.
d. From Eq (31) using G/L = 1, the values of X^ are 1.53,
0.83, and 0.82 for mixers 1, 2, and 3, respectively.
These values are well within the 0.54 to 2.75 range
tested in the pilot plant.
e. The ozone production costs are 2.79*, 1.52*, and 1.50*
per 1000 gal feed for stages 1, 2, and 3, respectively.
f. The water pumping costs are 0.31*, 0.17*, and 0.16* per 1000
gal feed for stages 1, 2, and 3 respectively.
g. The total adjustable costs for S = 1.7 and c = 2.0
wt % are 3.10*, 1.69*, and 1.66* per 1000 gal
feedwater for stages 1, 2, and 3 respectively.
h. For all three stages, the total adjustable costs are
6.46*/1000 gal feedwater.
The oxygen costs for this case are 1.00*/1000 gal feedwater,
one-third being due to COD reduction and two-thirds due to
adding 40 mg/1 dissolved oxygen to the wastewater. Adding
this 1.00* and the 5*/1000 gal for nonadjustable costs, the
total treatment cost at S = 1.7 and c = 2.0 is 12.5*/1000
gal treated. The value of f* is 0.632.
The mixer sizes for the actual plant must be expressed as a
scale factor applied to the pilot plant mixers. For stage 1,
for example, X^X' = 1.53 x 126.3 = 193.2. Thus the first
stage mixer should have 193.2 times the rated capacity of
164
-------
the pilot plant mixer. This implies linear dimensions in-
creased by the square root of 193.2 = 13.9. Naturally, a
number of smaller mixers could be used, It is important to
note, however, that the mixer size is rather critical; there
is little leeway in specifying its size.
Values for the total adjustable costs for all three stages
(item h) are shown in Table 16 for other assumed values of
S and c The minimum is at c = 2,5 wt % ozone and S =1.7.
At these optimum conditions, the total treatment costs are
about 6.38 + 1.00 + 5 = 12.4* per 1000 gal.
It is noteworthy that varying S at constant c produces an
extremely shallow minimum in this case. In examining a
number of other cases, the optimum value of S generally lies
in the range 1 to 2. Therefore, the optimization procedure
may generally be performed only for S = 1.7 with little
error in the resulting minimum.
It is also important to note that the minimum obtained by
varying c at constant S is not particularly sharp. There-
fore, the exact c value selected is not very critical to
plant economics.
Other Examples
In Table 17 are summarized the results of costs optimizations
for a number of different plant specifications. All cases
are for pH 7 feedwater. The total operating costs given are
the sum of the adjustable costs, oxygen cost, and 5C/1000
gal non-adjustable cost. The value f* is the fraction of
the ozone feed to all stages that is dissolved. It bears
repeating that these estimates are based on linear scaling
from 10 mgd plant with ozone generator capacity of 7500
Ib/day at 1.7 % ozone, and that in some of these examples,
165
-------
Table 16.
VALUES OF ADJUSTABLE COSTS, £/1000 GAL,,
FOR EXAMPLE OZONE WASTEWATER TREATMENT PLANT
s
1.0
1.3
1.7
2.0
2.5
3.0
4.0
6.0
7.5
10.0
1.7
7.06
6.92
6.89
6.92
7.00
7.11
7.34
7.71
7.92
8.19
2.0
6.60
6.50
6.46
6.47
6.52
6.59
6.73
6.97
7,10
7.27
V
2.5
6.49
6.41
6.38
6.38
6.41
6.45
6.54
6.69
6.77
6.88
Wt %
3.0
6.81
6.74
6.71
6.72
6.74
6.76
6.83
6.94
7.01
7.09
3.5
7.44
7.38
7.35
7.35
7.37
7.39
7.45
7.55
7.60
7.67
4.0
8.37
8.31
8.28
8.28
8.29
8.31
8.37
8.46
8.51
8.58
166
-------
the scaling is large.
Cases 1-5 show the effects of varying Lf at a required L
of 20 mg/1 in 45 min. The treatment costs increase rapidly
with increasing Lf, and part of this very rapid increase is
due to the smaller fraction of ozone dissolved.
Cases 6-8 show the effects of varying the number of stages
while holding the treatment time constant. The effects are
negligibly small.
Cases 2, 9, and 10 show the effects of varying the treat-
ment time while holding other plant specifications constant.
There is a reduction in treatment costs with increasing
treatment time, and this reduction is due entirely to
increased values of f*.
Comparisons of cases 10-14 with cases 1-5 show the effects
of the longer treatment times for higher values of Lf.
Here the effects are pronounced.
Cases 15-18 show the effects of decreasing the required
product COD to 10 mg/1 holding the treatment time at 45 min.
Cost for more than 50% COD reduction is high.
Cases 19-22 compared to 15-18 show, again, the lower costs
realized by extending the treatment time.
From all of the cases shown in Table 17 one point becomes
clear; mixers of the type used in the pilot plant cannot
economically handle high dissolved ozone requirements. This
is not from a lack of mixing efficiency, but is a fundamental
limitation of pure cocurrent mixing. For example, in a
reactor operating at Z = 5 mg/lp a perfect cocurrent mixer
167
-------
Table 17.
OPTIMUM PARAMETERS AND COSTS
FOR VARIOUS PLANT SPECIFICATIONS
Plant Specification,
pH = 7.0
Case
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Lf
mg/1
30
40
50
60
80
40
40
80
40
40
30
50
60
80
20
25
30
40
20
25
30
40
Lp
mg/1
20
20
20
20
20
20
20
20
20
20
20
20
20
20
10
10
10
10
10
10
10
10
t*
Min
45
45
45
45
45
45
45
45
90
180
180
180
180
180
45
45
45
45
180
180
180
180
N
Optimum Parameters
Stages S
3
3
3
3
3
4
6
6
3
3
6
6
6
6
3
3
3
3
6
6
6
6
2.0
1.7
1.7
1.7
1.3
2.0
2.0
1.7
2.0
2.0
2.5
2.5
2.5
2.0
1.7
1.3
1.3
1.3
2.5
2.0
2.0
2.0
eg
Wt %
2.0
2.5
2.5
3.0
3.0
2.5
2.5
3.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
3.0
3.5
4.0
2.0
2.0
2.5
2.5
Z*
mg/1
0.
2.
3.
4.
6.
2.
2.
6.
1.
0.
0.
1.
1.
1.
3.
5.
7.
9.
1.
1.
2.
2.
98
32
49
48
04
32
32
04
25
68
28
02
31
77
95
70
09
18
15
67
08
69
•F*
i
.765
.679
.588
.569
.466
.68
.68
.470
.734
.789
.823
.756
.729
.685
.553
.489
.465
.413
.715
.696
.696
.650
Total
Operating
Costs
<=/1000 Gal
8.
12.
17.
23.
39.
12.
12.
39.
11.
11.
8.
14.
18.
26.
10.
15.
21.
36.
9.
11.
14.
19.
44
37
47
78
68
36
35
42
64
33
31
69
37
48
95
47
19
08
52
84
38
80
168
-------
would leave 1.2 % ozone in the vent gas.
It appears, then that cocurrent mixers such as were used in
the pilot plant are best used at Z values of about 1 mg/1
which can be realized by increasing the treatment time.
Under such conditions, about 75% of the feed ozone would be
dissolved.
If higher dissolved ozone values are required to attain
higher reaction rates, then other types of mixers must be
developed. They must embody some variant of gas-liquid
counterflow, and the water entering the mixers must have a
low dissolved ozone content. This would require a plant
flow system completely different from the one used in the
pilot plant.
Selection of Pretreatment
Any detailed cost effectiveness analysis of the myriad
possible pretreatments that could be introduced before ozone
tertiary treatment is entirely beyond the scope of this
work. However, having found that the performance of the
ozone process is apparently not dependent on the exact
nature of the pretreatment, some guide in this area is
available from the data in Table 17.
Attention is directed specifically to cases 10-14 and 19-22.
These illustrate the approximate incremental cost reductions
for ozone treatment that might be realized from the further
removal of COD by a pretreatment process.
As an example, consider a plant requirement for 20 mg/1
product COD and an initial feed COD of 60 mg/1. Assume
that some pretreatment process under consideration could
reduce the feed COD to 40 mg/1. Comparing case 13 with case
169
-------
10, it can be seen that the savings in operating costs for
ozone treatment would be 18,37 - 11.33 = 7.04C/1000 gal.
Therefore, if the pretreatment process being considered
costs less than 7.04C/1000 gal, it would be economically
attractive.
It should be noted, though, that many of the possible pre-
treatment processes, in contrast to ozone treatment, simply
separate the contaminants from the wastewater, and the
costs of ultimate disposal of the concentrated wastes must be
included in such economic comparisons of processes.
Comparison of Current Estimated Cost with Earlier Estimates
The cost estimates which are presented herein are based on
the ozone treatment technology demonstrated in the pilot
plant. The estimated cost for operation with an ozone feed
to water ratio of 75 Ib of ozone per million Ib of wastewater
is 12.4C per 1000 gallons for a plant with design capacity of
10 mgd.
In a previous contract for a laboratory investigation of
wastewater treatment with ozone, the estimated cost for a
large scale treatment plant was 7.7C per 1000 gallons ^ .
The higher estimated cost which is now presented is due, in
the main, to an increase of over 30% in construction costs
as shown by the EN-R cost index , an increase in elec-
trical power cost from 0.8* to l.lC/Kwh, increased labor
costs, and from change in scope of the plant design. The
earlier estimate was based only on a projection of labora-
tory data.
170
-------
SECTION X
ACKNOWLEDGMENTS
The design, construction and operation of the Blue Plains
pilot plant, analytical work, data analysis and report
preparation was performed by a team from the Research &
Development Department of Airco, Inc. Personnel participating
in the project were: Dr. C. S. Wynn, Project Manager; Dr. B.S.
Kirk and Dr. R. McNabney, Research Chemical Engineers; Mr. J.M.
Wynne, Design Engineer; Mr. H. Bottjer, Computer Operator;
Miss Margaret Carson, Secretary; and Messrs. R. L. Allen, D. M.
Figert, Jr., P. Osman, W. J. Powers and R. Reed, Operator-
Technicians .
The support of the project by the Office of Research &
Monitoring, Evironmental Protection Agency and, in partic-
ular,' Mr. J. M. Cohen, Chief, Physical & Chemical Treatment
Research; and Mr. F. L. Evans, III, Project Officer, is
acknowledged with sincere thanks.
The cooperation of Mr. D. F. Bishop and staff and the
services of the analytical laboratory at Blue Plains is,
likewise, deeply appreciated.
The cooperation of the International Nickel Company, Inc.,
New York, N. Y. for analyzing the corosion test results is
acknowledged with thanks.
The cooperation of the firms mentioned below is gratefully
appreciated in supplying cost information for commercial
scale equipment.
It should be emphasized that mention of trade names or
commercial products does not constitute endorsement or recom-
mendation for use by the Environmental Protection Agency.
171
-------
Ozone Generators
*Crane Co., Cochrane Division
800 3rd Ave., P. 0. Box 191
King of Prussia, Pa., 19406
W. R. Grace Co., Davison Chemical Div.
Charles and Baltimore Streets
Baltimore, Md. 21203
Pollution Control Industries
507 Canal Street
Stamford, Conn. 06902
*The Welsbach Corp.
Ozone Systems Division
3340 Stokley Street
Philadelphia, Pa. 19129
*Also supplied cost data for integrated oxygen recycle systems,
Gas Dryers
The C. M. Kemp Manufacturing Co.
Glen Burnie, Md. 21061
Kahn and Company, Inc.
885 Wells Road
Wethersfield, Conn. 06109
Pall Trinity Micro Corp.
Cortland, New York 13045
Recycle Blowers
North American Rockwell
Pneumatics Division
P. 0. Box 2877 Commercial Station
Springfield, Mo. 65803
Injector-Mixers
Croll-Reynolds Co.
751 Central Avenue
Westfield, N. J. 07091
Penberthy Div., Houdaille Industries, Inc.
P. 0. Box 112
Prophetstown, 111. 61277
172
-------
Liquid-Ring Compressors; Vacuum Pumps
Dresser Industries, Machinery Group
3201 Wolf Road
Franklin Park, 111. 60131
Spray Nozzles
Spraying Systems, Inc.
3201 Randolph Street
Bellwood, 111. 60104
Centrifugal Pumps
Buffalo Forge Co.
Pump Division
Buffalo, N. Y. 14205
173
-------
SECTION XI
REFERENCES
1. Huibers, D. Th. A., McNabney, R., and Halfon, A., "Ozone
Treatment of Secondary Effluents from Waste Water
Treatment Plants", Report No. FWPC-4, Robert Taft
Sanitary Engineering Center, Cincinnati, 1969. U. S.
Clearinghouse of Federal Scientific & Technical
Information, PB Rep. 1969, PB 187758, 26 pp.
2. Sax, N. I., "Dangerous Properties of Industrial Materials",
3rd Ed., Reinhold, New York (1968).
3. Sconce, J. S., "Chlorine, Its Manufacture, Properties
and Uses", Reinhold, New York (1962).
4. "FWPCA Methods for Chemical Analysis of Water and
Wastes", FWPCA Analytical Water Quality Laboratory,
Cincinnati, Ohio, 1969.
5. APHA-AWWA-WPCF, "Standard Methods for the Examination of
Water and Wastewater", 13th Ed., American Public Health
Association, New York, 1971.
6. Ingols, R. S., Fetner, R. H., and Eberhardt, W. H.,
"Determining Ozone in Solution", "Ozone Chemistry and
Technology", p. 102, American Chemical Society, 1959.
7. Bishop, D. F. , O'Farrell, T. P., Stamberg, J./ and
Porter, J. W., "Advanced Waste Treatment Systems at the
Environment Protection Agency -- District of Columbia
Pilot Plant", in "Water - 1971" (L. K. Cecil, ed.)
AIChE, Symposium Series £8(124) 11-24 (1972).
8. Cassel, A. F., Pressley, T. A., Schuk, W. W., and Bishop,
D. F., "Physical-Chemical Nitrogen Removal from
Municipal Wastewaters" in "Water-1971" (L. K. Cecil, ed.)
AIChE Symposium Series 6_8_(124) 56-64 (1972).
9. Stamberg, J. B., Bishop, D. F., and Kumke, G.,
"Activated Sludge Treatment with Oxygen" in "Water-1971"
(L. K. Cecil, ed.) AIChE Symposium Series 68 (124)
25-34 (1972).
10. Kirk, B. S., McNabney, R., and Wynn, C. S., "Pilot
Plant Studies of Tertiary Wastewater Treatment with
Ozone" in "Ozone in Water and Wastewater Treatment"
(F. L. Evans, III, ed.) Ann Arbor Science Publishers,
Inc., Ann Arbor, Mich., pp. 61-82 (1972).
175
-------
11. Nebel, C. W., "Ozone Treatment of Secondary Effluent",
paper presented at Symposium on Ozone in Sewage Treat-
ment, University Extension, The University of Wisconsin,
Milwaukee, Wis., (Nov. 10, 1971).
12. Lark, P. 0., Craven, B. R., and Bosworth, R. C. L.,
"The Handling of Chemical Data", Pergamon Press, New
York (1968).
13. Carnahan, B., Luther, H. A., and Wilkes, J. 0.,
"Applied Numerical Methods", John Wiley, New York, 1969.
14. "Gmelins Handbuch der Anorganische Chemie", Vol. 3.4,
p. 1122. Verlag Chemie, Weinheim, Germany.
15. Manley, T. C. and Niegowski, S. J., "Ozone",
Encyclopedia of Chemical Technology, Vol. 14, p. 417,
Interscience-John Wiley, New York, 1967.
16. Guthrie, K. M., "Capital Cost Estimating". Chemical
Engineering, Vol. 76, No. 6, March 24, 1969, p. 114.
17- Mells, H. E., "Costs of Process Equipment". Chemical
Engineering, March 16, 1964, p. 133.
18. Guthrie, K. M., "Pump and Valve Costs", Chemical
Engineering, Vol. 78, No. 23, Oct. 11, 1971, p. 151.
19. Godfrey, Robert S., "Building Construction Cost Data
1971", Robert Snow Means Co., Inc., Duxbury, Mass.,
02332, 1971.
20. Staff Report "Construction Cost Index", Engineering
News - Record, p. 56, Mar. 23, 1972, McGraw-Hill,
New York.
176
-------
SECTION XII
PUBLICATIONS
1. Kirk, B. S., McNabney, R., and Wynn, C. S., "Pilot
Plant Studies of Tertiary Wastewater Treatment with
Ozone" in "Ozone in Water and Wastewater Treatment"
(F. L. Evans, III, ed.) Ann Arbor Science Publishers,
Inc., Ann Arbor, Mich., pp. 61-82 (1972).
2. Wynn, C. S., Kirk, B. S., and McNabney, R., "Pilot
Plant for Tertiary Treatment of Wastewater with Ozone",
presented 73rd National Meeting of AIChE, Minneapolis,
Minnesota, Aug. 1972. To be published in "Water - 1972"
AIChE Symposium Series.
177
-------
SECTION XIII
GLOSSARY
TABLE OF SYMBOLS
a Fitted coefficient, arbitrary subscripts
b Slope of dosage curve, see Eq (8)
c Weight percent ozone in oxygen
G Weight percent ozone in oxygen from generator
y
GV Weight percent ozone in oxygen vented from
reactor
D. Lb ozone dissolved in i-th reactor per Ib COD
in plant feedwater
D* Cumulative sum of D . from first through N-th
stage, See Eq (7)
DO., Dissolved ozone concentration in post-ozonation
apparatus, see Fig. 3
d (P, - P )/P2, about 0.21 in pilot plant
e Base of natural logarithms
F* Ozone feed ratio, Ib ozone generated/million
Ib water treated
f Fraction of ozone feed to a mixer-reactor that
is dissolved, see Eq (21)
f* Fraction of total ozone feed to all stages
that is dissolved, see Eq (42)
G. Gas flow rate to i-th mixer, SCF/hr
(G/L) Volume ratio of gas to water feed to a mixer,
SCF gas/cu ft water
g Fitted coefficient in Eq (9)
H Henry's Law solubility constant for ozone in
water, mg/l/atm
h Fitted coefficient in Eq (9)
j Fitted coefficient in Eq (10)
179
-------
Electrical power to ozone generator in Kilovolt-
amps calculated from voltmeter and ammeter read-
ings
Kw Electrical power to ozone generator, Kilowatts
P
measured with integrating watt-hour meter
Kw Electrical power to ozone generator, Kilowatts
measured with electrodynamometer power meter
KW Actual Kilowatts of mechanical power consumed
by a mixer, combined water pumping power and gas
compression power (appears only in combination
T/KW)
k-, , k? First order COD reduction rate constant in "two-
component" model, see Eq (5)
k First order dissolved ozone consumption rate
2
constant, see Eq (23)
L COD concentration, mg/1
Lf COD of feedwater entering ozone treatment plant,
mg/1
L COD of product water leaving ozone treatment
plant, mg/1
Ln Natural logarithm (to base e)
Log Common logarithm (to base 10)
Mf Lb/hr COD feed to plant
N Number of stages or reactors in ozone treatment
plant
n Number of experimental data points in a statis-
tical fit
nc Degrees of freedom removed by a statistical fit;
generally, the number of coefficients fitted
P2_ Absolute pressure at mixer inlet, atm
P2 Absolute pressure at reactor outlet, atm
A? Ozone partial pressure driving force between
gas and liquid, atm; subscript 1 for mixer
inlet, subscript 2 for reactor outlet, and
subscript m for log mean of 1 and 2, see Eq (20)
180
-------
Q Lb ozone dissolved/hr, see Eq (29); or Ib
ozone dissolved/1000 gal water treated see
Eq (40)
S Dissolved ozone distribution ratio, see Eq (36)
s/ SG Standard sample deviation, see Appendix 7
3a' sb' sh standard deviation on a fitted coefficient,
see Appendix 7
T Transference of a cocurrent gas-liquid mixer of
the type used in the pilot plant, see Eq (17),
Ib ozone dissolved/hr/atm mean partial pressure
driving force
t* Cummulative ozone treatment time, hr
t^ Treatment time in i-th reactor, hr
u Fitted coefficient in Eq (10)
v Volume fraction ozone in oxygen produced by
generator
v Volume fraction ozone in oxygen in spent gas
vented from reactor
W Water flow rate to mixer, gal/min
X' Plant scaling factor, see Eq (26)
X1 Mixer scaling factor, see Eq (27)
x Fraction of component 1 in "two component"
model
y Calculated value of a general dependent variable,
see Appendix 7
y Experimental value of general dependent variable,
see Appendix 7
Z Dissolved ozone concentration at the outlet (top)
i
of i-th reactor, mg/1
2* Time weighted average of Z. from first through
N x
N-th reactor
181
-------
SUPERSCRIPTS
Time weighted average or cumulative variable
Parameter for large plant as distinguished from
pilot plant parameter
SUBSCRIPTS
f Feedwater
g Ozone generator output
i i-th of N reactors numbered sequentially from
plant inlet
p Product water
v Spent gas from reactor
182
-------
ABBREVIATIONS
BOD Five-day- biological oxygen demand
C Feed pretreatment process, see Table 2
CL Feed Pretreatment process, see Table 2
COD Chemical oxygen demand
DC2 Feed pretreatment process, see Table 2
F Feed pretreatment process, see Table 2
gpm Gallons per minute
JTU Jackson turbidity units
L Feed pretreatment process, see Table 2
MC Feed pretreatment process» see Table 2
Mgd Million gallons per day
NH--N Ammonia nitrogen
NOp-N Nitrite nitrogen
NO..-1 Nitrate nitrogen
POA Post-ozonation apparatus, see Fig 3
SCF Standard cubic feet
SCFM Standard cubic feet per minute
TKN Total Kjeldahl nitrogen
TOG Total organic carbon
UN Pretreatment process, see Table 2
3B Pretreatment process, see Table 2
183
-------
SECTION XIV
APPENDICES
APPENDIX 1
DEVELOPMENT OF GAS-WATER MIXING SYSTEM
This section of the report summarizes experimental work
which was done to optimize water-jet injectors as the
primary gas-water mixing devices for the reactors of the
Blue Plains pilot plant, as it had been shown in previous
work that injectors had good potential for achieving
effective mixing ' '.
Based on the earlier work, it was concluded that six reaction
vessels with a total residence time of about 60 minutes
(at 35 gpm rate) should be used for the Blue Plains plant to
provide flexibility in testing a variety of effluents with
respect to both flow rate and level of contaminants.
Previous results indicated that about five seconds contact
time was needed in the vertical dissolver tube into which
the injector discharged to obtain a high rate of transfer of
ozone to the water ^ ' .
A test unit was built and operated at the Research Laborato-
ries of Airco, Inc., Murray Hill, N. J., for evaluating
injector devices for mixing gas and water, simulating
operating conditions which were expected for the Blue Plains
pilot plant. A diagramatic sketch of the apparatus used for
these evaluation tests is given in Figure 59. The water jet
injector discharged into the top of a vertical glass pipe
which is designed as the dissolver pipe. The end of this
dissolver pipe was inserted through the top closure of a
185
-------
TAP
WATER
GAS
CENTRIFUGAL
PUMP
— GAS
OUTLET
Fl ROTAMETER
PI PRESSURE GAUGE
DRAIN
Figure 59. INJECTOR TESTING APPARATUS
186
-------
55-gallon drum and, terminating about half-way inside the
drum. A 1/2-inch diameter line carried off gas which rose
to the cop of the drum, along with a small amount of water.
Most of the water was removed from the system through a
2-inch bottom outlet and into a section of pipe 20 feet high
for maintaining a static head on the drum, equivalent to
that expected for the pilot plant reactors. The gas bubbles
which formed in the drum were observed through clear plastic
windows.
Two sizes of dissolver pipes were used, namely 4-inch I.D.
and 2-inch I.D. glass pipes, to show the effect of residence
time in the dissolver tube. When the four-inch dissolver
pipe was used, a glass tee was placed between the discharge
of the injector and the top of the pipe into which an
auxiliary stream of water was fed to achieve an extension of
the turbulent mixing zone into the dissolver pipe. Various
mixing and baffle arrangements were tested. The first
injectors tested were standard PVC water-jet eductors used
for vacuum service.
Operating data, consisting of liquid and gas flows, pressures,
and stability of the liquid-gas mixtures in the dissolver
pipe, were obtained. With certain combinations of nozzle and
baffle arrangements and of fluid flow rates, the gas and
liquid phases separated when the gas flow exceeded a certain
rate. Various injector dispersion systems were tested over
the same range of liquid and gas rates that were anticipated
for the pilot plant.
An attempt was made to show the performance of the various
dispersion systems by taking simultaneous high speed
(1/50,000 sec.) photographs of the gas bubbles and of a
187
-------
transparent millimeter scale for estimating bubble size,
Although these pictures permitted some approximate estimates
of bubble surface, correlation with column performance was
not possible. The photographic method is useful, however,
in showing relatively large differences in dispersion systems.
Since the molecular weight of carbon dioxide is only 4 units
smaller than ozone, and because of similar solubility values
in water, we used CO,, in lieu of ozone for obtaining mass
transfer data for our dispersion systems. We assumed that
these data represented to a reasonable degree the perform-
ance, except for the chemical reactions, that would be
expected in the pilot plant using ozone. Carbon dioxide was
used because we did not have an ozone generator of sufficient
capacity.
For obtaining quantitative transfer data, CO- was titrated
with 0.02 N sodium hydroxide by a standard method on water
samples taken from the outlet of the absorber. Simultan-
eously saturation solubilities of the C02-air mixtures used
were checked in a saturation apparatus attached to the system.
These solubilities were found to check those given in the
literature, using Henry's law. Tests were made in both 4-
inch and 2-inch diameter columns to determine if column size
was significant for performance of the dispersion systems.
The operating characteristics of the injectors tested are
given in Table 18.
The results from the mixing tests have been calculated in
terms of mass transfer coefficient (Percentage of the
cocurrent equilibrium concentration of dissolved CO
present in the water at the bottom of the tank.) and trans-
fer rate per fluid horsepower. The mass transfer results
for the C02 - air-water system are summarized in Table 19.
188
-------
A special eductor which operated at a lower water-inlet
pressure (30 psig at 35 gpm) gave the highest energy
efficiency and just as high a mass transfer efficiency as
the eductors which operated at 60 psig.
At low C02 dosages (less than 20 mg C02 per liter of water),
a 2-inch dissolver column showed as high efficiency as a
4-inch column with a secondary shear zone. However, at
higher dosages, the 4-inch tube with secondary liquid jets
gives more mass transfer surface.
The nozzle of the 2-inch, high-pressure eductor was reamed
out to reduce the inlet water pressure to about 30 psig.
Energy efficiency was improved without impairing mass trans-
fer efficiency.
From our test results, and from cost considerations, it was
concluded that the eductors for the pilot plant should
operate at a water feed pressure of less than 28 psig when
the gas flow is about 8 scfm. A 2-inch dissolver tube could
be considered as an alternative, particularly in the later
stages where the ozone consumption is low.
The eductors ordered for the pilot plant were designed to
reproduce the characteristics of the experimental model with
polyvinyl chloride construction. The water feed pressure
was lowered some more in these units. One of these units was
set up on the injector testing apparatus with the proposed
secondary water nozzle system. The final configuration of
this system as installed in the pilot plant is shown in
Figure 37.
In its operation, half the water flow is mixed with the gas
stream in the eductor. This discharges through a short 2-inch
189
-------
Table 18
SUMMARY OF PRESSURE vs.
FLOW CHARACTERISTICS OF INJECTORS
!
Injector
A (1-1/2")
B (2") Stock Model
C (2") Modification
of B
D Special
Water
Flow
Gal./Min
35
35
25
25
35
25
'35
35
25
Water
Pressure
psig
55
66
37
34
36
22
30
28
14
Gas
Flow
scfm
10
8
11.4
8
8
8
11.4
8
8
Gas
Inlet
Pressure
psig;
3
1 . 3
5
5
2
2
5
4
4
190
-------
Table 19
SUMMARY OF MASS TRANSFER TESTS
TEST GROUP NO. 1 1
. j
i
1. Make of injector i S-
2. Diameter of dissolver pipe, in.
3. Water nozzle pressure of injec- ^
tor at 35 gpm flow - psig
i
1
4. Fluid power, input - horsepower j
per gpm flow at 35 gpm ! 0 . (
i
5, Average percent of entering CO,.,
transferred to water, experi- d
mental
6. Average percent of entering C0«
transferred to water, calcu- I 6
lated for cocurrent equilibrium j
7. Average mass transfer efficiency
(5) x 100 ! S
(6) |
[a) ', ,(a) _
i & «J
-1 i C-l C-l
i
414 2
1
iO 30 30
1
)18 0.021 0.019
50 60 45
55 65 51
!
1
1 91 89
4
S-l
2
60
0.033
39
47
83
j
5
s-z(c)
2
32
0.017
49
56
87
6(b)
C-2
4
20
0.012
56
61
92
(a)
(b)
(c)
Secondary liquid jets used. Water flow to injector reduced to 30 gpm or less
with total water flow at 55 gpm.
Configuration used at pilot plant.
Water nozzle enlarged to lower pressure.
-------
pipe to the head of the 4^-inch dissolver pipe, which carries
the gas-liquid mixture to the bottom of the reactor tank. The
other half of the water flow impinges upward and inward in two
opposed flat jets against the descending gas-liquid mixture
from the eductor discharge. This creates turbulence at the
head of the dissolver pipe, concentrating most of the
dispersion energy in this intimate mixture of gas and liquid
as it descends to the bottom of the tank. The secondary
liquid nozzles consist of stainless steel tubes passing
horizontally on either side of the eductor discharge stream.
These tubes are plugged at their inner ends. The water is
discharged through 3/8 inch by 1-1/2 inch longitudinal slots
which are faced inward and upward as shown. Flow passages
into the mixing zone are well streamlined, so that unproductive
friction and turbulence losses are minimized. The nozzle
areas were sized small for the lower anticipated flows, with
the intention of enlarging them later.
Column 6 of Table 19 shows the mass-transfer test results
using the pilot plant system. It uses less power and gives
mass transfer as good or better than any of the others.
Feed pressures can be reduced further, particularly for
injectors handling large recycle flows.
192
-------
APPENDIX 2
OUTLINE SPECIFICATIONS OF MAJOR PILOT PLANT EQUIPMENT
Neutralization Tank. 4'-0" dia x 12'-2". Tank is constructed
of mild steel. Inside of tank is protected with a cold
application of an epoxy^bitumastic paint. Water capacity is
1200 gal.
Vacuum Deaerator. 30" O.D. x 36" to tangent lines. Tank
capacity: 100 gal. Working pressure: 2 psia. Internal
test pressure: 30 psig. Material of construction: ASTM
A285, Grade C, flange steel. Inside coated with 6 mil
layer of baked epoxy enamel.
Spray nozzle for water feed to the deaerator is a full cone
center jet nozzle, No, 11-E, as supplied by Spray Engineering
Co., New York, N. Y. 10036.
Vacuum Pump. A Nash pump, No. MD-15, 439-49015, in all-
bronze construction, with a one HP motor was used.
Reactors (R-l through R-6). 2'-0" I.D. x 20'-0" between
bottom and top flanges. Shells are constructed of schedule
40 steel pipe. All interior surfaces of reactors are coated
with 100 mils of a hot-applied Koppers 70-B bitumastic
enamel.
Eductor Mixers. Details of the pilot plant eductor mixers
are shown in Fig 37. The eductor, itself, is a Croll-Reynolds,
2-inch Aqua-Ductor fabricated in unplasticized polyvinyl-
chloride to resemble No. 22 Aqua-Vactor.
Holding Tank. 6'-0" O.D. x 10'-0" high. This tank is
193
-------
constructed Of.mild,steel with flat bottojn and a-cone roof,
The interior walls are protected with a solvent type bitumastic
coating, applied in the field.
Tank capacity to operating level is 2000 gal.
Pumps (P-l through P-7). These pumps are Dean Brothers,
Model D1206 inline centrifugal pump of 316 stainless steel
construction for all wetted parts and fitted with John Crane
mechanical seals.
Pump P-l has a 3600 rpm, 5 h.p. electric motor. The rest of
the pumps are equipped with 3600 rpm, 3 h.p, motors.
Ozone Generator. This unit was supplied by the Welsbach
Corporation, Philadelphia, Pa,, 19129. It is their model
C1-34-D19L and has a rated capacity of 60 pounds per day of
ozone from oxygen at 70°F and with a dew point of -60 F.
The unit consists of the generator, transformer, variable
transformer and operating controls. Material of construction
is 304 stainless steel for all surfaces in contact with
ozone. Gaskets are made of "Hypalon" (a chlorosulfonated
polyethylene).
Absorption Dryer. The dryer is a package unit with filters,
blow-off valve, instruments and controls. It was supplied
by Gas Drying, Inc., Wharton, N. J., and is their Type V
"Dehyditrol" size 60 unit. The desiccant is activated
alumina.
Oxygen Compressor. Nash Engineering Co. Model AL-673, water
ring compressor. Capacity: 50 scfm of oxygen at outlet
pressure = 55 psig, saturated with water at 70°F. Inlet
pressure 0 to 5 psig. All wetted parts of the compressor
194
-------
are 316 stainless steel. Unit operates at 3500 rpm and
requires 21 brake horsepower.
Ozone Decomposer. This unit is Chromalox Model GCH-330
circulation electric heater, with AR type thermostat as
manufactured by Edwin L, Wiegand Co., Clifton, N, J. 07013
The heater unit has 3 kilowatts capacity.
195
-------
APPENDIX 3
CORROSION TESTS
Corrosion test specimens were placed in the product water
holding tank at the water-gas interface near the effluent
overflow. The following metals were tested: Stainless steel
types 302, 304, 316, 409, 430; Incoloy alloy 800, Incoloy
alloy 825, Carpenter stainless steel 20 Cb-3; aluminum 6061;
high-strength low alloy steel Cor-Ten, and carbon steel.
The temperature of the water during the test ranged from
65-75°F. The total time of the test was 105 days.
Corrosion rates, expressed as inches penetration per year,
were less than 0.00001 for the stainless alloys, except a
type 304 sensitized type which was 0,00003 in./'yr. The rate
of Incoloy 800 and 825, and Carpenter 20 Cb-3 were also less
than 0.0001 in./yr. Aluminum 606 had a corrosion rate of
0.004 in./yr. High-strength, low alloy Cor-Ten steel and
ordinary mild carbon steel had rates of 0.0125 and 0.0129
in./yr. , respectively. In general, the stainless alloys
showed up well in the test. Slight local corrosion was
observed with Carpenter 20 .Cb-3 and with sensitized type 304
stainless. This latter effect indicates that type 304
stainless, which is intended for welding or heat treatment,
should be the low carbon type 304L. Although aluminum 6061
had a low weight loss rate, it did show rather deep pitting.
Carbon steel and the high-strength, low alloy Cor-Ten
corroded generally at an undesirably high rate and with no
significant difference between them.
197
-------
APPENDIX 4
COPY OF SAFETY INSTRUCTIONS ISSUED
TO PILOT PLANT PERSONNEL
General Outline
Oxygen and ozone are used in the pilot plant for treatment
of secondary effluents from wastewater treatment plants.
Ozone is made from pure oxygen gas by its reaction in an
electrical discharge. A mixture of ozone in oxygen results -
the maximum concentration of ozone will be about 3%. Both
oxygen and ozone have properties which differ considerably
from air and, consequently, practices and work habits will
differ from those when working with air. The purpose of
this safety section is to discuss and itemize the safe
practices to be followed in the pilot plant.
Properties of Oxygen
Oxygen is a colorless, odorless and tasteless gas at
atmospheric temperatures and pressures. About one-fifth of
the atmosphere is oxygen. Oxygen will combine, either
directly or indirectly, with almost all other chemical
elements. This outstanding property enables it to sustain
life and to support combustion. Although oxygen itself does
not burn, materials which burn in air will burn much more
vigorously and at a higher temperature in oxygen. Grease
and oil become highly flammable in the presence of oxygen and
must not be allowed to contaminate oxygen piping or equipment,
Since oxygen is an invisible gas, it is impossible to see an
oxygen enriched atmosphere; therefore, good housekeeping is
essential. Storage of combustible materials is not permitted
in areas where oxygen enriched atmospheres could exist. For
the same reason, vessels and other equipment must be purged
199
-------
with air until a normal ojxygen-in-air concentration is
reached before repair work is started. In addition, general
ventilation of the building must be in operation to remove
any local concentrations of oxygen.
Retention of gas within a container at a pressure higher than
the design pressure of the container can be dangerous whether
it is oxygen or air. However, it is standard industrial
practice to prevent excessive pressure by use of relief
devices to maintain pressures at a safe level.
Properties of Ozone
Ozone, 03, is an unstable, blue gas with a pungent character-
istic odor. It is generally encountered in dilute form in a
mixture with air or oxygen. The color is not noticeable at
the concentrations used in wastewater treatment.
Ozone is a very powerful oxidizing agent in both inorganic
and organic systems. Ozone is much more reactive than oxygen.
This strong oxidizing property of ozone is utilized in the
wastewater treatment process,
Toxicity of Ozone
The maximum allowable concentration of ozone for continuous
exposure is 0.1 part per million. Since toxicity is a
result of ozone concentration and time of exposure, higher
concentrations can be tolerated for a short time. An
exposure to one part per million for 10 minutes is consid-
ered non-toxic by health authorities.
Oxygen, Ozone and the Oxidation Process
Pure oxygen is piped into the building from two liquid oxygen
supply tanks which are located on a pad along the north side
of the building. Oxygen is fed automatically under controlled
200
-------
pressure of 10 to 15 psig into the oxygen recycle system.
The oxygen feed pins the oxygen that is being recycled
pass into an electrical device for the generation of ozone.
A mixture of ozone (1 to 3%) in oxygen is produced. This
gas mixture is used to treat the wastewater in six reactor
columns.
Any gas that is released from the recycle system is passed
through an electric furnace at 600°F in which any residual
ozone is destroyed before it can be discharged into the
atmosphere.
Building Ventilation and Air .Monitoring
The two roof exhaust fans are operated when the plant is
operating. These fans have capacity to completely change
the air of the building every 6 minutes. An ozone analyzer
is provided for measuring ozone concentrations of the
building air. In addition to the analyzer, the human nose
is probably the most sensitive analyzer of all for detection
of ozone. Unpleasant sensations will most likely develop
before dangerous levels of ozone are reached.
Emergency Conditions
The pilot plant for ozone treatment of wastewater is well
designed and every attempt was made to use the best equipment
available for the job. Safety of personnel was a prime
consideration in the design. Nevertheless, in the event that
an unforeseen mishap occurs, shut down the ozone generator in
the following sequence:
1. Push the STOP buttons.
2. Turn off cooling water.
3. Allow dry air or oxygen to pass through the
ozonator for 15 to 20 minutes before the
generator outlet valve is closed.
201
-------
Correct any unsafe condition that may have developed before
starting the ozonator again and resuming normal operation.
Minimal Safety Rules and Practices for Operators of the Ozone
Treatment Process
1. D£ not Smoke in the main equipment area. Smoking is
permissible in the control room - but here, too, care
should be taken to keep away from oxygen streams that
are being analyzed.
2. Cleanliness and good housekeeping are essential. Do not
wear dirty or greasy work clothing or gloves. Do not
store combustibles in the pilot plant area.
3. Check all safety devices periodically; be sure that
cooling water flows to the ozonator, the oxygen
compressor, vacuum pump and all seals are at the proper
level.
4. Make certain the ventilating system of the building is
operative. In case of leakage or suspected oxygen gas
leakage, ventilate the building by opening all doors to
dilute the oxygen concentration. Inform the Supervisor
of the leak.
5. Do not perform any repair work on equipment without per-
mission of the Supervisor. Any repair work on the
Kryo-Flow oxygen system will be performed by the
supplier.
6. Use only the recommended pipe fitting compounds and
gaskets on the oxygen lines.
7- Never use oxygen as a substitute for compressed air.
8. Read the instruction manuals pertaining to each mechani-
cal piece of equipment for proper handling of these items,
9. Know where the emergency shut-off valve for the oxygen
supply is located.
202
-------
APPENDIX 5
ANALYTICAL PROCEDURES
The analytical methods -used were adapted from "Standard
Methods for the Examination of Water and Wastewater"^
and from "FWPCA Methods for Chemical Analysis of Water and
( A\
Wastes'a ' .
Chemical Oxygen Demand (COD) (Low Level)
In the previous prepilot work on ozone treatment, monitoring
of the pollutant removal was done by measuring COD (Chemical
Oxygen Demand) and TOC (Total Organic Carbon) ^ '. At first,
measurement of TOC was favored because it was a rapid
instrumental analysis and seemed to give satisfactory results.
Analysis for BOD,- was much too slow and cumbersome to supply
the large number of analyses, and also there was some
question as to the accuracy of BOD analyses when applied to
water in which the organic substrate had been radically
altered by ozone treatment.
When low pollutant levels were reached in the ozone treatment,
reading of the TOC analyzer recorder charts became difficult,
and trouble was encountered in the electronic systems. TOC
also tended to approach COD in numerical value in the
neighborhood of 8 mg/1, indicating that carbon dioxide (the
substance measured in the TOC analyzer) was being evolved
from inorganic bicarbonates and sesquicarbonates along with
organic sources of extra C02.
Therefore, in the Blue Plains pilot plant work, main
reliance was put on COD analyses, starting with the low-level
(4)
procedure given in the FWPCA methods manual
203
-------
The first effluents treated involved ver.y low levels of COD
(4 to 20 xng/1) , Extreme precautions were required to get
consistent, replicate results. It was found that distilled
water from the laboratory stills was not of consistently good
quality. A bottled, pyrogen-free distilled water was
substituted, which improved the reproducibility of the results
somewhat. Great care was taken in cleaning all glassware and
sample bottles with chromic acid cleaning solution, rinsing
with good distilled water, and providing protection from con-
tamination during sampling and analysis.
Another source of error with the standard procedure was
found as a result of receiving an unusually bad bottle of
mercuric sulfate. In the usual procedure, this is added to
the sample as a powder. Unbelievably erratic results were
obtained when use of this lot of mercuric sulfate was
started. The trouble was traced to the presence of non-
uniformly distributed organic dirt in the reagent. Since
dirt is always present to some extent in reagents, dry
powders should not be added to these samples. One small
particle of dirt weighing as little as 0.1 mg can cause a
gross error at the very low COD levels. The procedure was
modified to introduce all reagents in solutions. In this
way, any organic contaminant can be corrected by the blanks
which are run with each batch of samples.
The standard low level COD procedure calls for the dissolu-
tion of 1 gm of mercuric sulfate in 5 ml concentrated sul-
furic acid in each reflux flask followed by the addition of
25 ml of 0.025 N potassium dichromate. In the modified pro-
cedure, these reagents were combined into one solution as
follows. 80 gm of mercuric sulfate was dissolved in 100 ml
cone, sulfuric acid diluted with 900 ml water. To this
solution was added 300 ml cone, sulfuric acid and 500 ml of
204
-------
0,1 N potassium dichromate solution, and the results were
diluted to 2000,0 ml, After filtering and aging for about
a day, it was then cross^standardized against straight
0.025 N potassium dichromate using ferrous ammonium sulfate.
To each reflux flask, 25 ml of the modified solution was
added, and from that point, the analysis followed the
standard procedure.^ '
,0'zone in Gas
Ozone in oxygen was determined by the standard
iodometric method . For application to our system, the
gas sample was passed through a gas washing bottle with a
fritted glass sparger containing 300 ml of water and 100 ml
of 10% KI solution. The gas discharge was measured in a
wet-test meter. Usually the sample volume was 1/2 cubic
foot. A 25 ml aliquot was pipetted from 400 ml of KI
solution, 25 ml of 10% sulfuric acid and 50 ml of distilled
water were added, and the aliquot titrated with 0.1 N
thiosulfate solution which was standardized daily.
When there was difficulty in getting a 0.5 cu ft gas sample,
the sample volume was reduced to as little as 0.1 cu ft with
a corresponding increase in the aliquot taken.
The absorption was checked by running two absorbers in series
and virtually no iodine was ever liberated in the second
absorber.
Ozone Dissolved in Water
To determine dissolved ozone in water, the liquid sample was
taken in a 500 ml glass graduated cylinder. First, 80 ml of
10% KI solution and 10 drops of 5% KOH solution were put in
the cylinder- The sample line was then purged, and the
cylinder was filled as rapidly as possible to the 480 ml
205
-------
mark with samplef taking care tg keep the tip of the sample
tube near the surface of the liquid in the cylinder, The
contents of the cylinder- were shaken; then 20 ml of 10%
sulfuric acid was added. The entire contents of the cylinder
was titrated with the 0.1 N thiosulfate solution, using
starch indicator added near the end point.
As mentioned in the reference (6) the reaction between ozone
and KI to liberate iodine must take place at a pH of 9.5
or higher in order to obtain accurate results. The titration
with thiosulfate, however, must be done in acid solution,
hence the acid addition must be made just before the titration,
Oxygen Content of the Recycle Gas
Oxygen analyses were made on samples obtained from the gas
feed stream to the ozone generator.
The oxygen content of the gas stream was determined by
saturating distilled water in a flow cell with the gas. This
cell contained the probe of a Precision No. 68859 Galvanic
Cell Oxygen Analyzer. The oxygen concentrations indicated
for samples obtained from the gas sample outlet were compared
with results for streams of pure oxygen and of air. Readings
were taken from these three streams in succession at least
three times and were averaged for each stream. The ratio of
the average readings of the sample and of pure oxygen times
100 gives the percent oxygen in the recycle stream.
In addition, several analyses of the gas were made with a
conventional Orsat apparatus to measure oxygen, carbon
dioxide and inerts (assumed to be nitrogen).
Turbidity
Turbidities were determined on incoming and ozone treated
206
-------
water with a Hach Model 2100 turbidimeter. This is a
nephelometer which reads directly in Jackson Turbidity
Units (JTU).
pH
Values of pH were measured with a Hach Model 2075-00 glass
electrode pH meter, which was calibrated each shift with
standard buffer solution.
Other Analyses
For part of the operating period, special samples were col-
lected in the early morning and taken immediately to the
EPA analytical laboratory at Blue Plains for BOD, TOC, and
ammonia, nitrite, nitrate and Kjeldahl nitrogen analyses.
Samples of untreated feed were taken before and after
vacuum deaeration, in a few runs.
The samples were all run by the standard methods used in the
EPA laboratory.
The BOD values are subject to some question because of the
peculiar behavior of ozone-treated water in the BOD determina-
tion.
207
-------
APPENDIX 6
PLANT PROFILE DATA SETS
During normal pilot plant operationsf a complete plant pro-
file data set was collected once each eight-hour shift.
These data sets included the necessary data for evaluating
the COD reduction, ozone consumption, mechanical and electri-
cal work expended, etc. for each stage of the plant. In
addition to being recorded in plant logs and laboratory note-
books, these data were also entered onto a special data sheet
by the plant operators and the chemical analyst.
A copy of the actual data sheet used is shown on page
211. It is laid out to facilitate key punching of the data
onto five IBM computer cards for subsequent computer process-
ing. Each block on the data sheet corresponds to a column
on a standard IBM card. Most of the entries are self-
explanatory. The latter columns for card 2 are for data
from the post-ozonation experiment. The blocks "IDEN" are
for a numeric code designating the shift and date. The
"time" blank for each of the six reactors (not key punched)
is the time at which each sample for COD analysis was taken;
these were selected according to the feedwater flow rate to
approximately follow a slug of water through the train of
reactors. All entries are fixed format numeric except for
"Source of Plant Feed" which is encoded alphanumerically.
Following the copy of the data sheet is a tabulation of all
the complete plant profile data sets collected in the course
of this investigation. Only the independent plant para-
meters and the final COD are listed. Not included in these
tables are the numerous other data sets showing only plant
inlet and outlet conditions.
209
-------
The tabulation is arranged chronologically, an.d it is
divided according to feedwater pretreatmentg. The column
headings are, for the most part, self-explanatory. Times
are on the 2400 basis. Feedwater GPM (gal/min) values are
the net feeds to the plant. Under the "ozone generator"
heading, "PCT" refers to the weight percent ozone produced
by the ozone generator, and "KW" refers to the actual
kilowatts of electrical power drawn by the generator
corrected to the watt-hour meter basis. Entries in which
both water and gas flows to a mixer are shown as zero
indicate that the mixer was not being used.
210
-------
REACTOR PROFILE DATA SHEET
Airco Ozone Water Treatment Pilot Plant - Washington, D.C.
_SOUfiCi: JDF PLl MT FEE
Key punched
211
-------
K>
Table 20 - PLANT PROFILE DATA SETS
FEED WATER PRETREATMENT- 3-STAGE BIOLOGICAL + LIME + FILTER (3B+L+F)
DATE
2/ 1
2/ 2
2/ 3
2/ 4
2/ 5
2/ 8
2/ 9
2/10
2/11
2/12
2/16
2/17
2/18
2/19
2/22
2/23
2/25
3/ 1
3/ 2
3/ 3
3/ 4
3/ 5
3/ 9
3/11
TI^E
1000
1000
1600
1700
1700
2000
2000
2000
2000
2000
2000
2000
1000
0
2000
2000
2000
1600
1600
1600
1600
1600
1600
1600
FEEDWATER
GPM
20.0
14.0
22.0
24.0
22.5
21.5
le.o
19.0
19.0
18.0
20.0
22.0
12.0
17.5
17.5
22.0
24.0
28.0
27.5
30»0
30.0
30.0
16.0
21.0
PH
8*6
8.7
9.1
8.9
8.8
9.6
9.3
9.2
9.4
9.2
8.5
9.2
9.3
9.8
7.6
7.1
7.3
7.5
7.1
7-3
7.6
7.3
7.6
6.6
COD»
IN
48*4
34.6
52.8
52.8
69.0
67.6
45.6
22.9
16.5
19.2
16.2
15.0
15.0
13.5
13.3
12.8
16.0
9.7
18.1
13.3
54.0
65.0
89.2
68.6
MG/L
OUT
22.0
13.0
26.5
12.5
22.0
48«6
1">.7
5.3
7.0
9.3
6.1
3.5
4.2
6.2
8.2
5.3
8.2
1*6
10.3
6*8
22.4
39.8
27.4
19.6
OZONE
GENERATOR
PCT KW
1.94
2.12
2.18
2.04
2.02
1.70
2.86
2.79
2.79
2.03
1.86
2.24
1.95
1.20
1.08
1.11
1.04
1.73
1.56
1*54
2.53
2.46
2.81
2,13
3.2
3.1
3.1
2.9
3.0
2.8
4.4
4*4
4.4
3,1
3.3
3«2
2.8
2*9
2«9
2,8
2.7
2.7
2.8
2*8
6*7
6.9
8.4
8.6
NO 1
20/1*9
14/1*8
22/1,8
24/1.8
22/1.8
21/1.8
18/1.8
19/1*8
19/1.8
18/1.8
20/1,8
22/1.8
12/1.8
17/2.7
17/2.7
22/2.7
24/2.7
28/1.8
27/0.9
30/0*9
30/0.9
30/1.8
16/1.9
21/3.6
NO 2
55/1.8
55/1.8
55/1.8
55/1.8
55/1.8
55/1.8
55/1.8
55/1.8
55/1,8
55/1,8
55/1.8
55/1.8
55/1.8
55/2.7
55/2.7
55/297
55/2,7
55/1.8
55/2.8
55/2.8
55/3.2
55/3.2
56/3.2
55/4.5
MIXER FLOWS
4 WATER/SCFM G/
NO 3 NO 4
55/1.8
55/1.8
54/1.8
55/1.8
55/0.9
55/1.8
55/1,8
55/1.8
55/1,8
55/1.8
55/1.8
55/1.8
55/1.8
55/2.7
55/2.7
55/2.7
55/2,7
55/1.8
55/1.4
55/1.4
55/1.4
55/2.3
55/2.3
53/3.6
55/0*9
55/0.9
54/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0*9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/2.7
55/2.7
55/2.7
55/2.7
54/0.9
55/0.9
55/0*9
55/1.8
55/1.8
55/1.8
55/1.8
\5
NO 5
55/0*9
55/0.9
54/0.9
55/0,9
54/0.9
5i3/0.9
55/0.9
55/0.9
55/0,9
55/0.9
55/0.9
55/0.9
55/0.9
55/2*7
55/2.7
55/2,7
55/2.7
55/0.9
56/0.9
55/0*9
55/1.8
55/1,8
55/1.8
55/1.8
NO 6
54/0*9
55/0.9
54/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0*9
55/0.9
55/0,9
55/0.9
55/0.9
55/0.9
55/2.7
55/2.7
55/2,7
55/2.7
55/0.9
55/3.2
55/3.4
55/1.8
55/1.8
55/1.8
55/1.8
(CONTINUED)
-------
Table 20 (Continued) - PLANT PROFILE DATA SETS
FFFD WATFR PRETRFAT^ENT- 3-STAGE BIOLOGICAL + LiME + FILTER (3B+L+F)
(CONTINUED)
DATE TIME"
3/12 1600
3/15 2000
3/16 2000
3/17 2000
3/18 1700
3/22 ?°00
3/23 2000
3/24 2000
M 3/25 2000
w 3/26 2020
3/31 1920
4/ 1 1820
4/ 2 800
4/ 5 1920
4/ 6 1920
4/ S 1920
4/12 1920
4/13 1920
4/14 2000
4/15 1920
4/16 1900
4/19 1920
4/20 1920
4/21 1920
FEEDWATER
GPM
21.0
20.0
20.0
20*0
21.0
20.0
20,0
20.0
21.0
20.0
20.0
20.0
20.0
21.0
21.0
22.0
21.0
22.0
18.0
20.0
20.0
21.0
21.1
21.4
PH
7*4
7*4
7.1
7.4
7.8
7.5
8.2
8.2
7.3
7.4
7.6
a.o
8.0
5.6
5.6
5,4
8.5
8.3
8.3
8.3
8.3
8.2
8.0
8.1
COD»
IN
112.0
57.2
13.6
lo.o
12.1
11.4
12.2
13.0
13.0
4.3
10,6
10.4
12.2
13.3
12.1
10.4
11.1
10.9
19,0
20. 1
14.0
19.2
7.6
12.4
MG/L
OUT
49*6
6.0
4.8
4.3
8.6
0.4
7.3
6.2
10.8
3.2
0.1
7.3
6.8
9.9
8.3
9.0
4.8
r>,9
8.5
7.6
7.1
7.1
7.1
8.0
OZONE
GENERATOR
PCT KW
2.04
2.30
0.97
i.io
1.13
0.86
0.86
0.85
0.94
0.86
0.77
1.10
1.03
1.01
1.36
1.29
1.16
1.39
1.09
3.08
1.16
1.55
1.06
1.22
8*7
8*6
2.8
2.8
2.8
1.4
1.4
1.4
1.4
1.4
1.5
1.5
1,4
1*6
1.5
2.1
1.8
2.0
1.6
6.7
2.5
2.8
1.8
2.1
NO 1
21/3*6
20/3.6
20/3.2
20/3.2
21/3.2
20/2. 7
20/2.7
20/2.8
20/2.8
20/2.7
20/3.7
20/1.8
20/1.8
21/1.4
20/1.4
44/1.3
51/1.4
51/1.4
54/1.4
52/1.8
52/1.4
53/3.7
52/3.2
53/3.2
NO 2
55/4.5
55/4.5
55/2.7
55/2.7
55/2.7
55/4.6
55/4.6
55/4.6
55/4.6
55/4.6
55/4.6
55/2.8
55/2.8
55/2.8
55/2.8
55/2.7
55/2.8
55/2*8
55/2.8
55/2.8
55/2.8
55/2.7
55/2.7
55/2.7
MIXER FLOP'S
/I WATER/SCF'1 G>
NO 3 NO 4
55/3.6
55/3.6
55/2.3
55/2.3
55/2.3
56/2. 1
55/2.7
54/2.8
55/2.8
55/2.7
53/3.7
54/1.8
55/1.8
55/1.4
55/1.4
56/1.3
55/1.4
55/1.4
55/2.3
55/1.4
55/1.4
54/1.8
55/1.3
55/1.3
55/1.8
55/1.8
55/1.8
55/1.8
55/1.8
55/0.4
55/0.4
55/0.4
55/0.4
55/0.4
54/1.8
57/0.9
55/0.9
55/0*9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
•\ o
NO 5
55/1.8
55/1.8
55/1.8
55/1.8
55/1.8
55/0.0
55/0.0
56/0.0
55/0.0
55/0.0
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
56/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
NO 6
55/1.8
54/1.8
55/1.8
55/1.8
55/1.8
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
54/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
56/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0,9
55/0.9
(CONTINUED)
-------
Table 20 (Continued) - PLANT PROFILE DATA SETS
1NJ
FEED WATER PRETREATMENT- 3-STAGE BIOLOGICAL + LIME + FILTER ( 3B+L+F )
(CONTINUED)
DATE
4/22
4/23
4/26
4/27
4/27
4/28
4/28
4/29
5/ 3
5/ 4
5/ 5
5/ 6
TIME
1920
1920
1920
800
1920
800
2130
1920
1920
1920
1920
1920
FEEDWATER
GPM
21*5
21.5
21.5
22.0
21*5
22.0
21,5
21*5
21.5
22.3
2 1 , 5
22*3
PH
6«6
8*1
7.8
7.5
8« 0
8.0
8,1
8*8
8.5
8.6
8.8
8.8
COD*
IN
11*1
15.1
13.7
13.0
13.0
I4o2
17.0
16.5
21,2
23,8
41,2
51,2
MG/L
OUT
4.5
6«S
7.6
6.3
6.9
7.1
10.6
8*2
6.5
9.3
10.9
10.3
OZONE
GENERATOR
PCT KW
1.24
1.22
1*24
1.11
1.64
1.18
1.28
1.68
1.59
1.52
3.41
5.25
2*1
2.1
2.1
2.2
2.1
2.2
2.1
2*8
2.9
2.8
7.0
8.6
MO 1
53/3.2
53/2.9
55/3.2
54/3.2
55/3.2
54/3.2
55/3.2
55/3.2
52/3*2
52/3e2
52/2,8
55/3.7
NO 2
55/2.7
55/2.7
53/2.7
55/2.7
56/2.9
56/2.7
54/2,9
55/2*7
55/2.7
55/2.7
55/1.4
55/1,8
MIXER FLOWS
4 WATER/SCFM G;
NO 3 NO 4
55/1*3
55/1.3
56/1.3
55/1,3
55/1.3
55/1.3
54/1.3
55/1.3
55/1,3
55/1.3
55/0,9
55/1.8
55/0*9
56/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0,9
55/^*9
55/ti,9
5 5 / 0 » 9
55/0,9
55/0.9
NO 5
55/0«9
55/0.9
56/0.9
55/0.9
56/0.9
55/0.9
55/0.9
55/Oey
55/0,9
55/0.9
55/0.9
55/0,9
NO 6
62/0*9
62/0.9
62/0.9
63/0.9
63/0.9
63/0.9
61/0,9
65/^,9
63/0.9
62/0.9
62/0.9
62/0.9
5/ 7 1920 21*5 8.1 40.9 8.9
3.20 8.4 56/3.7
-COMPLETED-
55/1,8 55/0.9 55/0.9 55/0,9 62/0.9
-------
(_n
Table 20 (Continued) PLANT PROFILE DATA SETS
FEED WATER PRETREATMENT" UNOX * LIME + FILTER
DATE TIME
5/10 1920
5/11 1300
5/11 1900
5/12 1030
5/12 1900
5/13 1030
5/13 1900
5/14 1030
5/14 1900
5/17 1015
5/17 1900
5/18 1015
5/18 1900
5/19 800
5/19 1900
5/20 1900
FEEDWATER
GPM
22*3
37.0
37.0
36.0
37.0
36.0
36,5
36.0
37.0
37.0
37.0
37.0
37.0
36.5
37.0
38.0
PH
8-9
9.0
8.8
9.1
9.5
7.3
7.0
6.6
6.5
6,1
6.4
8.4
8.6
8.6
9.4
8.4
COD,
IN
17»8
22.1
20.0
19.8
18.8
18.6
14.8
16.5
13.0
13.8
13.8
17.0
15.0
20.0
15.5
14.6
MG/L
OUT
6-1
8.8
7.5
3.2
B.4
9.3
8.6
9.2
10.3
8.3
8.0
9.0
7.9
7.3
7.3
8.4
OZONE
GENERATOR
PCT KW
3*26
3.20
2.88
1.82
1.53
1.94
1.60
1.61
1.56
1.64
1.39
1.64
1.54
1.38
1.54
1.58
8*7
8.7
8.6
5.4
5.3
5.4
2.9
3.0
2.9
2.8
1.7
2.3
2.8
2.8
2.8
3.3
NO 1
50/3*7
37/3.7
37/3.7
36/5.6
37/5.9
36/5.9
36/3.2
36/3.2
37/3.2
37/3.2
37/1.8
37/1.8
37/5.1
36/5.0
37/5.1
38/4.9
NO 2
55/1.8
55/1.8
55/1.8
54/3.6
55/3.6
55/3.6
55/2.3
55/2.3
55/2.3
55/2.3
56/0.9
55/0.9
54/1.8
55/1.8
54/1.8
37/1.8
MIXER FLOWS
1 WATER/SCFM G>
NO 3 NO 4
55/1.8
55/1.8
55/1.8
54/2.7
55/2.7
56/2.7
55/1.4
55/1.4
55/1.4
55/1.3
56/0.9
55/0.9
55/1.3
55/1.3
54/1.3
37/1.3
55/0*9
55/0.9
5S/0.9
55/1.8
55/1.8
55/1.8
55/0.9
55/O.V
55/0.9
55/0.9
55/0,9
55/0.9
55/0.9
55/0.9
55/0.9
37/0.9
•sa
NO 5
55/0*9
56/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
55/0.9
54/0.9
37/0.9
NO 6
63/0*9
63/0.9
63/0.9
63/0.9
63/0.9
63/0.9
63/0.9
63/0.9
63/0.9
62/0.9
63/0.9
63/0.9
63/0.9
59/0.9
63/0.9
37/0.9
-COMPLETED-
-------
Table 20 (Continued) - PLANT PROFILE DATA SETS
FFED WATER PRETREAT^ENT- MINERAL CLARIFIED + CHLORINE + CARBON
DATE
5/21
5/24
5/25
5/25
5/26
5/26
5/27
5/27
5/28
5/28
6/ 1
6/ 1
6/ 2
6/ 2
6/ 3
6/ 3
6/ 4
6/ 4
6/ 4
6/ 7
6/ 7
6/ 7
6/ 8
6/ 8
TIME
1900
1920
930
1920
1300
1920
900
1920
900
1920
900
1920
200
900
830
1920
100
900
1520
400
908
1700
200
800
FEEDWATER
GPM
18.0
20,8
32.0
20.0
20.0
20.0
19,0
19.0
19.5
20.0
20.0
20.0
20*0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20,0
20.0
PH
6.6
7.0
6.4
6.4
5.8
6.0
5*6
5.6
6.0
6.0
6,5
6.9
7*0
6.6
6.1
7.0
7.0
6.9
6.8
6,9
6,2
6.9
6.9
7.0
COD.
IN
12.0
11.7
14.8
9.4
9.2
10.9
12.7
13.1
15.2
11.8
8.7
7.0
11*7
14,3
13.7
13.8
18.9
17.4
18.5
10.3
11.0
14.1
15.8
13.6
MG/L
OUT
5.6
4.7
7.1
4.3
4.2
6.0
7.6
7.8
9.8
7.4
4.0
4.3
7*4
9.5
B.5
7.8
11.7
9,3
9.1
5.3
6,3
6.5
6.4
9.5
OZONE
GENERATOR
PCT KW
1*72
1.95
1.90
1.55
1.33
2.05
2.21
2.23
2.49
2.27
2.48
2.46
2«26
2.05
2.06
1.77
1.54
2.15
1.84
2,02
2.51
2.44
2,07
2.15
3.1
2.7
2.8
1.7
1.7
1.6
1.8
1.7
1.8
1.6
1.7
1.7
1*6
1.7
1.7
1.7
1.7
1,7
1.7
1.7
1.7
1.8
1.7
1.7
NO 1
18/4.8
36/1.8
32/1.8
33/1.8
20/1.8
20/1.8
19/1.8
19/1.8
19/1.8
20/1.8
20/1.8
20/1,8
20/1*8
20/1,8
20/1,8
40/1.8
39/1,8
19/1.4
19/1,4
20/1,4
20/1,4
20/2.3
20/2.3
20/2,3
NO 2
36/1.8
37/0.9
37/0.9
37/0.9
37/0.9
37/0.9
36/0.9
37/0.9
36/0.9
37/0.9
37/0.9
37/0.9
37/0*9
36/0.9
37/0.9
37/0.9
37/0.9
25/0.9
25/0.9
25/0.9
25/0,9
0/0,0
0/0,0
0/0.0
MIXER FLOWS
1 WATER/SCFM G/
NO 3 NO 4
36/1,3
37/0,9
37/0.9
37/0.9
37/0,9
0/0,0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0*0
0/0.0
0/0.0
o/o.o
0/0,0
25/0,7
25/0.7
25/0.7
25/0.6
25/0.9
25/0.9
25/0,9
36/0*9
37/0.9
37/0.9
37/0.9
37/0.9
37/0.9
37/0.9
37/0.9
37/0.9
37/0.9
36/0.9
37/0.9
37/0.9
36/0.9
37/0,9
37/0.9
37/0.9
25/0.7
25/0.7
25/0.7
25/0.6
0/0.0
o/o.o
0/0.0
\s — — — — -
NO 5
36/0*9
37/0.9
37/0.9
37/0.9
37/0.9
0/0.0
0/0.4
0/0.4
o/u.o
0/0.0
o/o.o
0/0.0
0/0*0
o/o.o
0/0.0
0/0.0
0/0,0
0/0,0
0/0,0
0/0,0
0/0.0
25/0.4
25/0.4
25/0,4
NO 6
37/0.9
37/0.9
37/0.9
37/0.9
37/0.9
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
o/o.o
0/0.0
0/0.0
0/0.0
0/0,0
0/0,0
0/0,0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
(CONTINUED)
-------
Table 20 (Continued) - PLANT PROFILE DATA SETS
FEED WATER PRETREATMENT- MINERAL CLARIFIED + CHLORINE + CARBON (MC+CL+C)
(CONTINUED)
DATE
6/ 9
6/ 9
6/10
6/10
6/10
6/11
6/16
6/17
6/17
6/17
6/18
6/18
6/18
6/21
6/21
6/22
6/22
6/23
6/23
6/24
6/24
6/24
6/25
6/25
TIME
1108
1700
200
945
1700
200
1600
0
1020
2000
300
930
1800
930
2100
930
1ROO
0
2000
130
1730
2100
200
800
FEEDWATER
GPM
20.0
20.0
20.0
20.0
20.0
20.0
19,0
20.0
20.0
20.0
2C.O
20.0
20.0
20.0
20.0
2U.O
20.0
20.0
20.0
17.0
17.0
17.0
17.0
18.0
PH
7.0
6.2
5.7
6.0
5.8
6.3
6.0
6.6
6.5
6.3
6.9
7.1
6.9
6.9
6.6
6.4
6.3
6.5
7.1
8.5
9.2
8.9
8.2
9.0
COD»
IN
14.0
10.8
14.8
9.3
11.7
13.3
16.9
14.1
16.5
17.7
16.3
16.4
13.6
10.0
13.2
17.6
18.1
16.0
19.6
17.6
15.8
18,5
21,8
16,1
MG/L
OUT
9.7
6.8
7.1
3.5
6.0
6.4
11.4
10.2
12.8
10.5
10.1
12.6
4.0
3.9
6*0
13,2
11,8
10,9
9.6
10.8
11,1
12.8
10.4
8.1
OZONE
GENERATOR
PCT KW
1.89
1.75
1.97
2»22
2.57
2.59
1.79
1.90
1.67
2.54
2.09
1.88
2.84
3,09
1.73
2.61
1,85
1.97
4.64
4.21
4.31
1.99
2.38
2.07
1.7
1.6
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
2.1
1.7
1.7
1.7
1.7
6.7
6.7
6.7
2.8
2.8
2.8
NO 1
40/2.8
40/2.8
39/2.8
40/2.8
20/1.4
20/1.4
19/1.4
20/2,3
20/1.4
20/1,4
0/0.0
0/0,0
20/1,8
20/1.4
20/1.4
20/1.4
55/3.7
55/3,7
54/3.7
55/3.7
55/3.7
17/2,7
17/2.7
17/2.7
NO 2
0/0.0
o/o.o
0/0,0
o/o.o
0/0.0
0/0.0
30/0,4
0/0,0
30/0,4
0/0.0
40/2.8
40/2.8
37/0.9
0/0,0
30/0,4
o/o.o
0/0,0
0/0,0
0/0,0
0/0.0
0/0,0
25/0.9
25/0.9
25/0,9
MIXER FLOWS
1 WATER/SCFM G>
NO 3 NO 4
0/0,0
0/0,0
0/0,0
0/0.0
0/0,0
0/0.0
30/0,4
25/0,9
30/0,4
0/0,0
0/0.0
0/0.0
0/0,0
0/0,0
30/0.4
0/0.0
0/0,0
0/0.0
0/0.0
0/0,0
0/0,0
25/0.9
25/0.9
25/0.9
20/0.9
20/0.9
20/0,9
20/0.9
20/0.4
20/^.4
30/0,4
0/0,0
30/0,4
20/0.4
25/0.9
25/0.9
37/0.9
20/0.4
30/0.4
20/0.4
0/0,0
0/0.0
0/0.0
0/0.0
0/0.0
25/0,4
25/0.4
25/0,4
NO 5
0/0.0
0/0.0
0/0.0
o/o.o
0/0.0
0/0.0
30/0,4
24/0,4
30/0,4
o/o.o
o/o.o
o/o.o
0/0.0
0/0,0
30/0.4
0/0,0
0/0,0
0/0.0
0/0,0
0/0,0
0/0,0
25/0,4
25/0,4
25/0.4
NO 6
0/0.0
0/0,0
0/0,0
o/o.o
0/0.0
0/0.0
30/0,4
0/0,0
30/0,4
0/0,0
0/0.0
0/0.0
0/0,0
0/0.0
30/0,4
0/0,0
0/0-.0
0/0,0
0/0,0
0/0,0
0/0,0
25/0,4
25/0,4
25/0,4
(CONTINUED)
-------
Table 20 (Continued) -
FEED WATER PRETREATMENT- MINERAL CLARIFIED
(CONTINUED)
PLANT PROFILE DATA SETS
CHLORINE + CARBON
(MC+CL+C)
DATE
6/25
6/23
6/28
6/29
6/29
6/29
6/30
7/ 1
7/ 1
7/ 1
7/ 2
7/ 2
7/ 2
7/ 6
7/ 6
7/ 6
7/ 6
7/ 7
It 7
7/ 8
7/ 9
7/ 9
7/ 9
TIME
1500
1130
1800
230
1230
1730
1730
130
826
1630
100
826
1530
412
900
1700
2340
900
1900
1900
30
900
1600
FEEDWATER
GPM
18.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
PH
9.1
7.1
5.9
8.4
7.4
7.3
8.0
7.0
6.2
8.7
8.1
6.9
6.3
8.9
8.2
7.1
6.5
9.0
8.1
8.9
8.1
8.9
6.2
COD»
IN
14.3
21.6
15.6
21.2
13.8
15.1
19.2
19.6
19.0
17.3
19.8
18.7
14.3
12.7
13.3
13,3
11.1
13.1
16.5
15.6
12.7
15.7
15.3
MG/L
OUT
2.6
11.4
11.3
12.1
8*6
7.8
7.1
9.8
10.6
8.4
9.7
11.4
7.1
6.8
6.8
6.8
7.6
5.5
10.6
4.2
6.3
7.1
6.6
OZONE
GENERATOR
PCT KW
1.77
1.16
1.24
1.52
1.45
1.67
2.29
2.25
2.32
2.41
2.59
2.55
2.18
2.63
2.38
1.95
1.69
3.08
1.73
2.86
1.63
2.84
1.92
2.8
1.4
1.4
1.4
1.4
1.4
2.1
2.8
2.8
3.1
3.1
3.1
3.1
3.9
3.0
2.5
2.3
4.6
2.1
5.4
2.4
4.4
2.8
NO 1
18/2.7
17/2.8
17/2.7
17/2.7
17/2.7
0/0.0
0/0.0
0/0.0
0/0.0
0/0,0
0/0,0
0/0.0
0/0.0
0/0,0
0/0,0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
NO 2
25/0.9
25/0.9
25/0.9
25/0.9
25/0*9
38/2.8
40/2.8
40/2.8
40/2.8
40/2.7
40/2.8
40/2.7
40/2.7
40/2,8
40/2,8
40/2,8
40/2,8
40/2,8
40/2,8
50/3.7
50/3.7
50/3.7
50/3,7
MIXER FLOWS
A WATER/SCFM G/
NO 3 NO 4
25/0.9
25/0,9
25/0,9
25/0.9
25/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0,9
35/0.9
35/0.9
35/0,9
35/0,9
35/0.9
35/0.9
35/0.9
35/0.9
35/1.4
35/1.4
35/1.4
35/1.4
25/0.4
25/0.4
25/0.4
25/0.4
25/0.4
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0,9
35/0.9
35/0.9
35/0.9
35/0.9
40/0.9
40/0.9
40/0.9
40/0.9
\s
NO 5
25/0*4
25/0.4
25/0.4
25/0.4
25/0.4
35/0.4
35/0,4
35/0.4
35/0,4
35/0.4
35/0.4
35/0.4
35/0.4
35/0.4
35/0.4
35/0,4
35/0.4
35/0.4
35/0.4
50/0.4
50/0.4
Su/0.4
50/0.4
NO 6
25/0.4
25/0.4
25/0.4
25/0.4
25/0.4
35/0.4
35/0.4
35/0.4
35/0.4
35/0,4
35/U.4
35/0,4
35/0,4
35/0,4
35/0,4
35/0.4
35/0.4
35/0.4
35/0.4
43/0.4
43/0.4
43/U.4
43/0.4
-COMPLETED-
-------
Table 20 (Continued) - PL-ANT PROFILE DATA SETS
FEED WATER PRETREATMENT- MINERAL CLARIFIED (MC)
DATE
7/12
7/13
7/13
7/13
7/14
7/14
to 7/14
i-1 7/15
7/15
7/15
7/16
7/16
7/16
7/19
7/19
7/19
7/20
7/20
7/20
7/21
7/21
7/21
7/22
7/22
TIVE
2000
200
915
1730
100
1030
1630
200
1030
1630
200
1020
1700
300
930
1700
0
1300
2000
300
1330
2000
300
1200
FEEDwATER
C,pM
32.0
32.0
32,0
32.0
32.0
32.0
32.0
32.0
32.0
32.0
32.0
32.0
32.0
32.0
32,0
32.0
20.0
37.0
3780
20«0
35.0
35.0
20.0
35.0
PH
8«9
8.0
7.3
6.4
9.1
8.1
7.1
6.2
8.7
7.9
7,1
6«1
9.8
7.8
8.6
6.0
9.1
7.6
6.0
10*9
7.1
9.8
8.7
7.6
COD. MG/L
IN OUT
49«8 20*4
64.4 27.5
55.2 24.6
39.0 19.3
59.5 26.1
44.5 15.5
38.6 16.1
64.5 31.7
50el 21.5
38.9 15,5
75.7 33.8
49.9 21.4
46.0 20.7
40.1 11.4
34.1 11.1
53.9 27,5
55.5 21.0
43.0 17.3
66.0 37.6
61.5 23.1
39.1 15,3
65.3 28.6
67,7 22,2
68.1 27.7
OZONE
GENERATOR
PCT
NO 3 NO 4
55/3.6
55/3*6
55/3.6
55/3,6
55/3,6
55/3.6
55/3.7
55/3.7
55/3.6
55/3,8
55/3.8
55/3.8
55/3,8
55/3.8
55/3.8
55/3,8
55/3.8
55/3.8
55/3,8
55/3.8
55/1,9
55/1.9
55/1,9
55/1.9
55/2.7
55/2.7
55/2,7
55/2.7
55/2.7
55/2.7
55/2,7
55/2.7
55/2.7
55/2.8
55/2.8
55/2.8
55/2.8
55/2.8
55/2.8
55/2,8
55/2.8
55/2.8
55/2,8
55/2.8
55/1.9
55/1.9
55/1.9
55/0.9
^i
NO 5
55/2*7
55/2.7
55/2,7
55/2,7
55/2,7
55/2.7
55/2.7
55/2.7
55/2.7
55/2.8
55/2.8
55/2.8
55/2.8
55/2.8
55/2.8
55/2,8
55/2.8
55/2.8
55/2.8
55/2-8
55/1,9
55/1,9
55/1.9
55/0,9
NO 6
0/0*0
0/0,0
0/0,0
0/0.0
0/0,0
0/0,0
0/0,0
0/0.0
0/0,0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
o/o.o
0/0.0
0/0.0
o/o.o
0/0.0
o/o.o
0/0.0
0/0.0
CONTINUED)
-------
Table 20 (Continued) - PLANT PROFILE DATA SETS
FEED WATFR PRETREATMENT- MINERAL CLARIFIED (MO
(CONTINUED)
DATE
7/22
7/23
7/23
7/23
7/26
7/27
7/27
7/27
7/28
7/28
7/28
7/29
7/29
7/29
7/30
TIviE FEEDWATER
1900
300
1200
1900
1930
300
1200
1900
300
1200
1900
330
1530
2230
500
GPW
35.0
20.0
35.0
41.0
35.0
20.0
35.0
35.0
20.0
35.0
35.0
20.0
35.0
35.0
20.0
PH
7«1
6.1
8.7
7.1
6.2
9.0
9.2
8.8
9.2
8.9
9.6
9.2
8.9
8.4
9.0
COD.
IN
72.5
68.9
69,9
52.8
55.5
57,5
42.2
50.8
51.2
42.2
56.0
52.4
48.4
54.5
52.4
MG/L
OUT
41.1
27,7
36.9
19.1
26.4
18.8
17.0
16.8
15.6
17.0
21.9
18.3
19.2
19.8
18.0
OZONE
GENERATOR
PCT KW
3.26
3.08
3.22
3.22
2.31
2.36
2.41
2.40
2.42
2.56
2.36
2.48
2.77
2.94
2.12
10«9
11.0
11.0
11.0
7.7
7.6
7.7
7.7
7.7
7.7
7.8
7.9
11.0
11.1
6.7
NO 1
53/6*7
46/6.7
42/6.7
42/6.7
51/6.7
52/6.7
53/6.7
53/6.7
52/6.7
53/6.7
52/6.7
52/6.7
52/6.7
52/6.7
52/6.7
MIXER FLOWS
NO 2 NO 3 NO 4 NO 5
55/1.9
55/1.9
55/1.9
55/1.9
55/0.0
55/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
55/1*9 55/0*9 55/0*9
55/1.9 55/U.9 55/0.9
55/1.9 55/f>.9 55/0.9
55/1.9 55A'.9 55/U.9
55/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
50/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
55/3.8 55/1.9 55/1.9
NO 6
o/o.o
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
o/o.o
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
0/0.0
-COMPLETED-
-------
NJ
K)
Table 20 (Continued) - PLANT PROFILE DATA SETS
FEED WATEF: PRETREATMENT- LlNOX
-------
Table 20 (Continued) - PLANT PRO-FILE DATA SETS
K;
FEF_[X WATER PRETREATMENT- UNOX (UN)
(CONTINUED)
DATE
8/13
R/13
8/16
8/16
3/16
8/17
8/17
8/17
8/18
8/18
TIME
300
1200
230
1100
2000
200
1100
1900
436
1020
FEEDwATER
GPM
35.0
54.0
36.0
54.0
52.0
35.0
54.0
51.0
35.0
54.5
PH
5*9
5.8
6.0
6.0
5.9
5.9
6.2
6.2
6.2
6.2
COD,
IN
23.6
42*0
27.6
32.5
33.7
29.2
34.4
33.6
26.5
38.7
MG/L
OUT
10*8
17,6
14, 8
18.8
19*6
19,2
17.8
17*8
14.6
2^.4
OZONE
GENERATOR
PCT KW
3.02
2.58
2.16
1.97
1.65
1.86
2.32
1.54
2.30
2.04
4.4
4*4
2.8
2.8
2*8
2,8
3.1
2.8
2.8
2.9
NO 1
34/2.4
0/0.0
36/2,4
0/0,0
52/3.8
35/2.4
o/o.o
51/3.8
35/2,4
0/0.0
NO 2
35/0*9
55/3.8
35/0,9
55/3.8
50/1,4
35/0,9
55/2.7
50/1.4
35/0.9
55/3.8
MIXER FLOWS
4- WATER/SCFM Qf
NO 3 NO 4
35/0.4
55/0.9
35/0.4
55/0.9
50/1,4
35/0.4
55/0.4
50/1,4
35/0,4
55/0.9
35/0.4
55/0,9
35/0.4
55/0.9
50/0,9
35/0.4
55/0.9
5C/0.9
35/0,4
55/0.9
NO 5
35/0.4
55/0,9
35/0,4
55/0.9
50/0,9
35/0,4
55/0.9
50/0.9
35/0,4
55/0.9
NO 6
0/0,0
0/0.0
0/0.0
0/0.0
50/0.9
0/0,0
o/o.o
50/0,9
0/0,0
0/0,0
-COMPLETED-
-------
Table 20 (Continued) - PLANT PROFILE DATA SETS
FEED WATER PRETREATMENT" MINERAL CLARIFIED + CHLORINE (MC+CL)
DATE
8/20
8/20
8/23
8/24
M
to 8/24
w 8/25
8/25
8/25
8/26
8/26
yR/26
8/27
TIWE FEEDWATER
1238
400
1200
30
2030
330
1200
1930
330
1200
1800
130
GPM
20.0
20.0
20.0
20.0
35.0
23.0
35.0
35.0
20.0
35.0
35.0
20.0
PH
7.3
7.6
7.5
8.3
7.4
7.5
6.9
6.5
7.6
6.5
7.5
6.0
COD,
IN
46.4
48.8
38.7
43.3
55.5
63.0
58.3
^3.3
42.2
36.4
36.4
47.5
MG/L
OUT
31.2
28*6
9.7
15.5
34.3
42.2
43.7
32.1
27.2
20.4
25.3
17.8
OZONE
GENERATOR
PCT KW
2.22
2.35
2.52
1.73
2.12
1.95
2.47
2.41
1.76
2.34
2.58
1.65
3.9
3.9
8.2
3.9
6.4
3.9
6.3
6«3
4.1
6.3
6.3
3.6
NO 1
0/0.0
0/0.0
20/4.8
0/0.0
35/4.8
0/0.0
35/3.8
35/3.8
20/3.8
35/3.8
35/3.8
20/4.8
NO 2
0/0.0
0/0.0
0/0.0
0/0.0
35/2.8
0/0.0
35/2.4
35/2.4
35/2.4
35/2.4
35/2.4
35/2.4
MIXER FLOWS
1 WATER/SCFM Gt
NO 3 NO 4
30/3.8
35/3.4
35/3.8
35/3.8
35/1.9
35/3.8
35/1.4
35/1*4
35/1.4
35/1.4
35/1.4
35/1.4
35/1.9
35/1.9
35/2.8
35/1.9
35/1.9
35/1.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
^o
NO 5
35/1.9
35/1.4
35/0.9
35/1.9
35/0.9
35/1.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
NO 6
35/1.4
35/0.9
35/0.9
35/1.4
35/0.9
35/1.4
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
35/0.9
-COMPLETED-
-------
Table 20 (Continued) - PLANT PROFILE DATA SETS
FEED WATER PRETREATMENT- 3-STAGE BIOLOGICAL + FILTER 13B+F)
DATE
8/27
8/27
8/30
R/30
8/31
8/31
8/31
9/ 1
9/ 1
9/ 1
9/ 2
9/ 2
9/ 2
9/ 3
9/ 3
9/ 3
9/ 7
9/ 7
9/ 7
9/ 8
Q/ 8
9/ 8
9/ 9
9/ 9
TIME
1330
1900
1215
1930
300
745
2030
100
1208
2000
230
1300
2000
300
1230
1600
300
800
2000
300
' 750
2000
300
1230
FEEDWATER
GPM
35.0
35.0
35.0
35.0
20.0
20,0
35.0
20.0
20.0
35.0
20.0
35.0
35.0
20.0
35.0
35.0
20,0
20.0
35.0
20.0
20.0
35.0
20.0
35.0
PH
6.6
6*8
6.6
6.8
6.6
6.8
4.9
6.5
5.2
6.9
6.9
7.0
6.4
6.1
5.8
6*8
7.1
7.1
5.2
6.5
6.2
7,5
5.4
6.6
COD»
IN
23.0
17.2
14.1
12.5
14.0
15.2
14.0
12.4
13.2
13.2
13,2
12.8
13.2
14,0
16.2
14.2
14.8
11.1
14.0
11.1
11.3
13,3
9.3
13.2
y>3/L
OUT
9.7
8.2
5.1
5.1
5.1
2.8
8.8
6.4
6*0
8.9
4.0
7.1
4.8
6.0
9.3
8.9
5.3
4.9
9.5
4.1
3.5
4.4
4.0
6.0
OZONE
GENERATOR
PCT KW
1.72
1.76
1.47
0.86
2.47
3.56
0.93
2.09
2.46
2.52
3.36
1.20
2.46
3.24
1.06
0.99
2.37
3.29
1.10
1.63
1.45
1.33
2.59
1.47
4*4
4*4
3.4
1.7
3.0
6.7
1.7
2.8
6.7
6.7
6.7
1.7
6.7
6.7
1.7
1.7
2.8
6.7
1.7
1.7
2.8
2.8
2.8
2.8
NO 1
35/3.8
35/3.8
35/3.8
35/3.8
0/0,0
0/0,0
37/3,8
0/0,0
20/3.8
35/3,8
0/0,0
35/3.8
35/3.8
o/o.o
35/3.8
35/3,8
0/0.0
0/0.0
35/3.8
0/0.0
0/0.0
35/3.8
0/0.0
35/3.8
NO 2
35/2.9
35/2.9
35/2.9
37/1.9
25/2.9
25/2.9
37/1,9
25/2.9
37/1.9
37/1.9
25/2.9
37/1.9
37/1.9
25/2.9
37/1.9
37/1.9
25/2.9
25/2.8
37/1.9
25/2.9
25/2.9
37/1.9
25/2.9
38/1.9
MIXER FLOWS
1 WATER/SCFM G;
NO 3 NO 4
35/1*4
35/1.4
35/1.4
37/1.4
25/1,4
25/1.4
37/1.4
25/1.4
37/1.4
37/1.4
25/1.4
37/1,4
37/1,4
25/1.4
37/1.4
37/1.4
25/1.4
25/1.4
37/1.4
25/1.4
25/1.4
37/1.4
25/1.4
38/1.4
35/0.9
35/0.9
35/0.9
37/0,9
25/0.9
25/0.9
37/0.9
25/0.9
37/0.9
37/0.9
25/t\9
37/0,9
37/0,9
25/0.9
37/0.9
37/0.9
25/0,9
25/0.9
37/0.9
25/0,9
25/0.9
37/U.9
25/0.9
38/0.9
^b
NO 5
35/0.9
35/0.9
35/0.9
37/0.9
25/0.4
25/0.4
37/0.9
25/0.4
37/0.9
37/0.9
25/0.4
37/0.9
37/0.9
25/0.4
37/0,9
37/0.9
25/0.4
25/0.4
37/0,9
25/0.4
25/0,4
37/0.9
25/0.4
38/0.9
NO 6
35/0.9
35/0.9
35/0.9
37/0.4
0/0.0
0/0.0
37/0.4
0/0.0
37/0.4
37/0.4
0/0,0
37/0.4
37/0.4
0/0,0
37/0,4
37/0.4
0/0,0
0/0.0
37/0,4
0/0.0
^5/0.0
37/0.4
0/0.0
38/0,4
CONTINUED
-------
Table 20 (Continued)
PLANT PROFILE DATA SETS
PEED WATER PRETREATMENT- 3-STAGE BIOLOGICAL + FILTER
(CONTINUED)
(3B+F)
DATE
9/ 9
9/10
9/10
9/13
9/13
TIME
2000
300
730
400
1230
FEEDWATER
GPM
35.0
20.0
20.0
20.0
35.0
PH
7.0
6.3
6.0
7.4
6.8
COD.
IN
12.8
12.4
14.2
11.0
11.4
MG/L
OUT
2.8
2.4
2.4
0.8
5.1
OZONE
GENERATOR
PCT KW
3.13
3*65
3.51
1.59
1.13
9*4
8.4
8.4
1.7
1.7
NO 1
35/3.8
0/0.0
0/0.0
0/0.0
35/3.8
MIXER FLOWS
NO 2 NO 3 NO 4 NO 5
37/1.9
25/2.9
25/2.9
25/2.9
37/1.9
37/1.4
25/1.4
25/1.4
25/1.4
37/1.4
37/0.9
25/0.9
25/0.9
25/0.9
37/0.9
37/0.9
25/0.4
25/0.4
25/0.4
37/0.9
NO 6
37/0.4
0/0.0
0/0.0
0/0.0
37/0.4
-COMPLETED-
NJ
U1
-------
KJ
Table 20 (Continued) - PLANT PROFILE DATA SETS
FEED WATER PRElREATMENj- D.C. PLANT SECONDARY EFF^^T
HATE
9/13
9/14
9/14
9/14
9/15
9/15
9/15
9/16
9/16
9/17
9/17
9/20
9/20
9/20
9/21
9/21
9/21
9/21
9/22
9/22
9/23
T I ME
1900
400
1230
2000
300
1030
1900
500
1730
430
1730
300
1200
2000
0
1330
1900
2338
1250
1900
200
FEEDW
GPM
35.0
35*0
35*0
35.0
35,0
35.0
35.0
35*0
35.0
35.0
35.0
20.0
20.0
35.0
35.0
35.0
20,0
20.0
20,0
20.0
20.0
ATER
PH
6,8
7.1
668
6*6
6.9
6*6
6.6
6*6
7*0
7*0
7.3
6.8
7,0
7,1
7,0
6,8
7,1
7,1
7.1
7,1
7.1
COD»
IN
92,0
110,8
86,6
62.3
113.3
128*7
121,1
122.8
82®6
95.3
91.3
113,1
96,4
115.4
149,1
85,7
85.9
82.6
67.1
73.7
101.9
MG/L
OUT
48.6
67.6
34.8
41.0
67,2
56.4
62.2
74,0
33e 2
48.0
49.4
69.6
59.0
65,1
75,0
37.9
34,0
44,5
27.5
29.1
48,5
OZONE
GENERATOR
PCT KW
1*99
2.45
2.63
2.55
2.55
2-75
2.68
3.81
2.55
2.90
2.79
2.88
3.40
2.81
2.16
2.83
2.91
3.29
3,16
2.55
2.51
4.4
8.5
11,3
11.2
11.1
11,3
11.0
11*3
11.1
11.1
11,0
11.1
11.1
11.3
4»4
11.1
11.3
11.0
11.4
7.6
7.5
MIXER FLOWf
NO 1
35/4.7
35/4*7
35/2.8
35/4.7
35/4.7
35/4,7
35/4*7
0/0.0
35/4.4
0/0.0
35/3,3
20/3,7
20/3.6
34/3,7
35/0.9
35/2.9
20/1.9
20/1,9
20/1,9
20/1.9
20/1,9
NO 2 NO 3 NO 4 NO 5
35/4,7
35/4,7
35/1.8
0/0,0
0/0.0
57/1,8
57/1.9
0/0.0
55/1,9
35/1,7
50/1,8
0/0.0
0/0,0
0/0,0
0/0.0
0/0.0
45/3,8
45/1.9
50/1.9
45/1.9
45/1.9
35/2»8 35/1,8 35/0,9
35/2,8 35/1,8 35/0.9
35/4,7 35/3,2 35/2.3
35/4e7 35/3,8 35/2.8
35/4,7 35/3,8 35/2.8
35/4,7 35/2.3 35/2.8
35/4*7 35/2.3 35/2,8
35/2.3 35/2.0 35/4.2
35/4e7 35/3,8 35/2,6
35/3,3 35/4.7 35/3.0
35/4,7 35/3,7 35/2.8
20/4,7 20/4,7 20/4,7
20/292 20/2.2 20/1.8
35/4.6 35/3.7 35/3.7
35/4,7 35/4*2 35/4.2
35/4»8 35/3.8 35/4.3
35/3.8 35/3,8 35/2.9
35/1.9 35/1.9 35/1.9
35/1,9 35/1.9 35/1.9
35/1,9 35/1.9 35/1.9
35/1.9 35/1.9 35/1.9
NO 6
35/0.9
35/0.9
35/2.3
35/2,8
35/2,8
35/2,3
35/2.3
35/2.1
35/2.8
35/3.8
35/2.8
20/4.7
20/4.5
35/3.7
35/4.2
35/4.3
35/1.9
35/1.9
35/1.9
35/1.9
35/1.9
-COMPLETED-
-------
Table 20 (Continued) - PLANT PROFILE DATA SETS
FEED WATER PRETREATMENT- D.C. PLANT SECONDARY EFFLUENT + LIME (DC2+L)
DATE
9/24
9/24
9/27
9/27
9/27
9/28
9/28
9/28
9/29
9/29
9/29
9/30
T I ME
1230
1800
300
1030
?ooc
300
930
1900
300
900
1930
300
FFEDWATEP
GPM
20*0
20»0
20,0
20.0
20*0
2090
20»0
20*0
20,0
20aO
20.0
20,0
PH
7.6
8*3
8.1
7,1
8.8
7»0
7*3
7.1
808
763
9,0
6.9
COD,
IN
71.2
77.8
110.1
58.0
56,4
55,5
61.3
69.5
71.8
63,9
55.4
69.4
MG/L
OUT
34.0
26«5
39.0
24.0
22.7
20.9
30.2
31.1
24.5
27.0
15.6
26.6
OZONE
GENERATOR
PCT
-------
APPENDIX 7
SIGNIFICANCE OF STANDARD DEVIATION VALUES
The parameter s and sg is the unbiased estimate of the
sample standard deviation or the sample standard error of
estimate. It is a measure of the data scatter about a
fitted curve. About 68% of the data should fall within +s
of the fitted curve, 90% within +1.64s, and 95% within +1.96s.
Throughout this report, the values of s and s have been
calculated as follows:
s = [MYi - Y,)2/(n - n )]°'5
x i i c
where y. is the experimental value of the dependent variable
y. is the fitted value of the dependent variable
n is the number of data points fit
n is the number of coefficients adjusted
No corrections were made for the finiteness of n (e.g.
Student's t).
The values of s , s, , etc. are estimates of the standard
3. JD
deviation of a fitted coefficient. It is a measure of how
far the fitted value of the coefficient may be from the
"true" value. There is about 68% probability that the "true
value" of a lies within +s of the least-squares fitted
— — 3.
value of a. It should be noted that s and SQ are generally
independent of the number of points fitted, but s^ decreases
with an increase in the number of points fit, being about
-I/2
proportional to n
229
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
. •£,
w
4. T,tle PILOT PLANT FOR TERTIARY TREATMENT OF
WASTEWATER WITH OZONE
.,/ . . - -_„
March 1972
7. Aifbor(s)
Wynn, Clayton S. ; Kirk, Bradley S.; McNabney, Ralph
Airco, Inc., Murray Hill, New Jersey 07974
Research and Development Department
;' Report No. ,
tac>:i a.
17020-DYC
Organ?
WQO 14-12-59.7
ji', Type /Repai'and..
•• ' P-aod Covered Final ,
July $9S$rMalfClx:l972
Environmental Protection Agency report
number, EPA-R2-73-146, January 1973.
M. Abstract Tertiary treatment of wastewater with ozone in a nominal 50,000
gal./day pilot plant at Blue Plains, Washington, D.C., is described.
Plant feeds (10 to 100 ppm COD) were effluents from other pilot processes
involving nine different biological and physical treatments of the Blue
Plains wastewater. Major COD reductions were realized, and product water
was sterile and oxygen saturated.
The pilot plant used three major process steps: (1) generation
o£ ozone gas from oxygen, including preconditioning of the gas feed and
means of recirculating the gas; (2) dissolution of ozone from the oxygen
carrier gas into the water; and (3) retention of the ozonated water for
a period sufficient for the organic contaminants to be oxidized.
Plant performance for each feed is described in terms of COD
reduction characteristics and the effects of pH, ozone concentration,
feed pretreatment and initial COD on reaction rate. Data are given for
ozone solubility and half-life in pure water and various wastewaters.
Bacteria kills are reported. Estimates of capital and operating costs
are presented for large plants to treat wastewater with ozone and a
procedure is given for optimization of costs for large plants.
17a. Descriptors
* Ozone
* Tertiary Treatment
* Wastewater
* Pilot Plant
Optimization
Water Treatment
Oxygen
Economic Evaluation
Cost Minimization
17b. Identifiers
05D
' |>. Security Class.
'''
Set -
(Page)
21. .' ,fftf. .of ~
Pages
,i'l, Ptivi
Send To:
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U S. DEPARTMENTOFTHE INTERIOR
WASHINGTON. D. C. 2O24O
231
AU.S. GOVERNMENT PRINTING OFFICE: 1973 514-152/179 1-3
------- |