EPA-660/2-75-020
JUNE 1975
Environmental Protection Technology Series
Refinery Effluent Water Treatment
Plant Using Activated Carbon
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development,
U.S. 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 STUDIES 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 REVIEW NOTICE
This report has been reviewed by the Office of Research and
Development, EPA, 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.
-------
V'
EPA-660/2-75-020
JUNE 1975
REFINERY EFFLUENT WATER
TREATMENT PLANT USING
ACTIVATED CARBON
By
Gary C. Loop
Grant No. 12050 GTR
Program Element 1BB036
ROAP 21AZP/Task 027
Project Officer
Leon H. Myers
Robert S. Kerr Environmental Research Laboratory
National Environmental Research Center
P. 0. Box 1198
Ada, Oklahoma 74820
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
For sale by the Superintendent of Documents, U.S. Government
Printing Office, Washington, D.C. 20402
-------
ABSTRACT
Reduction of Chemical Oxygen Demand (COD) in petroleum refinery effluent
wastewater by adsorption onto activated carbon was demonstrated on
a commercial level during a two-year project at Carson, California.
The plant contained over 750,000 pounds of carbon, regenerated
1,644,000 pounds of carbon, processed 172 million gallons of water,
and removed 408,000 pounds of COD.
The carbon was exhausted at the rate of 9.5 pounds per 1000 gallons
of water processed. At an average feed COD concentration of 250 ppm
and an average effluent COD concentration of 50 ppm, the carbon was
loaded to an average of 0.26 pounds of COD per pound of carbon.
Following solution of initial startup problems, the unit was operated
at a cost of 40 cents per 1000 gallons of water treated, or 18 cents
per pound of COD removed.
This report was submitted in fulfillment of EPA Grant No. 12050 GTR,
under the partial sponsorship of the Environmental Protection Agency.
Work was completed by Atlantic Richfield Company, Carson, California,
January 1974.
ii
-------
CONTENTS
Section Page
I Conclusions 1
II Recommendations 2
III Introduction 3
IV Process Description 6
V Description of Activated Carbon Plant 8
VI Design Basis 20
VII Operation and Evaluation of Water Treatment
Facilities 22
VIII Operation and Evaluation of Regeneration
Facilities 57
IX Test Methods and Their Evaluation 63
X Quantities and Costs Based on Conditions
During the Project 70
XI Appendices 73
iii
-------
FIGURES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
CARBON ADSORPTION PLANT FLOW DIAGRAM
OVERALL VIEW OF CARBON PLANT
IMPOUNDING RESERVOIR
WATER TREATMENT UNIT
VIEW INTO AN ADSORBER CELL
CARBON STORAGE TANKS
CARBON REGENERATION FURNACE AND GAS SCRUBBER
CARBON REGENERATION FURNACE
CAUSTIC TREATING TEST RUN
RELATIVE EFFICIENCY PROFILE OF CELL NO. 3 AFTER
208 HOURS <§ 250 GPM AND 650 TO 600 PPM COD IN FEED
CELL NO. 3 COD VS. TIME FIRST RAINS
FEED AND EFFLUENT COD DATA, FIRST RAINS
COD LOADING, FIRST RAINY PERIOD, CELL NO. 1
COD LOADING, FIRST RAINY PERIOD, CELL NO. 2
COD LOADING, FIRST RAINY PERIOD, CELL NO. 3
COD LOADING, FIRST RAINY PERIOD, CELL NO. 4
COD LOADING, FIRST RAINY PERIOD, CELL NOS. 5
AND 7 THROUGH 12
COD LOADING, FIRST RAINY PERIOD, CELL NO. 6
TWO-STAGE SYSTEM
9
10
11
11
13
15
16
18
24
27
28
30
33
34
35
36
37
38
39
iv
-------
FIGURES (CONT'D.)
Page
20 COD CONCENTRATIONS, TWO-STAGE TEST RUN 41
21 CARBON LOADING, TWO-STAGE TEST RUN 42
22 FIVE-CELL STAGGERED OPERATION TEST RUN 47
23 SECOND RAINS RECYCLE OPERATION 48
24 ADSORPTION DATA, SECOND RAINS, CELL 1 50
25 ADSORPTION DATA, SECOND RAINS, CELL 3 51
26 ADSORPTION DATA, SECOND RAINS, CELL 5 52
27 ADSORPTION DATA, SECOND RAINS, CELL 7 53
28 ADSORPTION DATA, SECOND RAINS, CELL 9 54
29 ADSORPTION DATA, SECOND RAINS, CELLS 2, 4, 6, 8 55
10, 12, 11
30 FURNACE OPERATION 61
-------
TABLES
Page
1 COD DATA BEFORE AND AFTER VIGOROUS BACKWASHING 25
AT 5000 GPM
2 LABORATORY DATA ON CARBON PLANT FEED AND EFFLUENT 31
3 ESTIMATE OF MAXIMUM WATER THROUGHPUT 46
4 TIME REQUIRED FOR REGENERATION 58
5 DATA FROM CARBON REGENERATION 60
6 ACCURACY OF COD TESTING PROCEDURES 64
7 COMPARISON OF ABD AND RE TESTS 67
8 COMPARISON OF RELATIVE EFFICIENCY TESTS FOR 69
REGENERATED CARBON
9 ADSORPTION DATA 71
10 COST DATA 72
vi
-------
ACKNOWLEDGMENTS
Personnel and facilities at Carson, California were furnished by
Atlantic Richfield Company.
Mr. C. R. Adams, Engineering Manager, and Mr. M. A. Prosche,
Refinery Technology Manager, served as Project Directors, and
provided administrative and technical assistance.
Mr. Gary G. Loop, Associate Process Engineer, provided technical
direction and prepared the final report.
Mr. E. F. Dumas, Refinery Technology Supervisor, and Mr. P. L.
Mehta, Process Design Engineering Supervisor, provided required
technical assistance.
Mr. R. P. Strand, Waste Water Disposal Supervisor, provided
operational guidance of the plant.
Mr. W. P. Ellertson, Analytical Services Supervisor, Mr. G. R.
Meador, Chemical and Lubricants Laboratory Supervisor, and their
staff performed the laboratory tests and provided interpretation
and comparisons of the test results.
The Environmental Protection Agency provided partial financial
support, and Mr. Leon H. Myers, Project Officer, provided neces-
sary guidance.
vii
-------
SECTION I
CONCLUSIONS
1. Reduction of Chemical Oxygen Demand (COD) content in refinery
waste water effluent has been demonstrated to be feasible by
using activated carbon as an adsorbent.
2. The system performed well in that it demonstrated an ability
to start up and shut down without delay or difficulty. This
gives the process a distinct advantage over biological units
for use in handling intermittent rainfall.
3. During the two-year project the unit was operated at an over-
all average cost of 49 cents per 1000 gallons of water treated
or 24 cents per pound of COD removed from the effluent waste
water. The first year's operational costs were 62 cents per
1000 gallons of water treated, or 34 cents per pound of COD
removed. After improvements over the first year's operation,
the costs during the second year were reduced to 40 cents per
1000 gallons of water treated, or 18 cents per pound of COD
removed.
4. The plant demonstrated excellent reliability.
5. The carbon adsorption plant has demonstrated that when the
feed COD is controlled to an average of 233 ppm, the plant can
treat refinery water using 8.5 pounds of carbon per 1000 gal-
lons treated and reduce the effluent COD to an average of
48 ppm with a high level of 95 ppm.
-------
SECTION II
RECOMMENDATIONS
This project demonstrated that activated carbon can be used on a
commercial scale to reduce the Chemical Oxygen Demand (COD) con-
centration of petroleum refinery effluent waste waters. However,
several areas need further investigation. These are summarized
below.
1. Further determination of the quantities and types of COD
materials that do and do not adsorb onto activated carbon.
2. Determine feasibility of pretreatment to reduce load on
carbon.
3. Determine optimum number of stages in adsorption process
with controlled feed COD concentrations.
Variables to consider when constructing commercial plant:
1. Be certain that desired effluent COD concentration is attain-
able for the particular circumstances involved. This should
be done by use of a pilot plant representative of the com-
mercial plant.
2. The feed COD concentration range should be determined and
controlled to eliminate surges. This could be done by re-
cycling effluent water, or possibly by some pretreatment
facility.
-------
SECTION III
INTRODUCTION
Petroleum refineries are faced with the problem of disposing of
hugh volumes of waste water. These large amounts of water come
from a wide variety of sources. These sources include process
water used for heat transfer, wash water, and rain water runoff
which collects oils and chemicals from within the refinery.
In the past, separation of the visible, floatable oils were con-
sidered satisfactory to allow discharge into adjacent waterways.
However, new concern over the ecological balance of our environ-
ment has called this practice into question. In an effort to
protect our waterways from these harmful discharges, new and
improved technology is needed.
One major pollutant existing in refinery discharge waters is
oxygen demanding material. Oxygen demand, which refers to the
demand for oxygen by chemicals and oils, lowers the water's
available dissolved oxygen content, a vital need for marine life.
In 1968, the Los Angeles Regional Water Quality Control Board made
a study of the Dominguez Channel in Los Angeles County. They de-
termined that petroleum and chemical plant discharges were causing
a problem due to their oxygen demand. The Control Board, in ac-
cordance with these findings, issued a resolution in February 1968,
which limited the total chemical oxygen demand (COD) from all
industrial discharges into the Channel. These discharges included
rain water runoff. The resolution was to be complied with by
February 1971.
3
-------
Atlantic Richfield Company's Watson Refinery is one of about 16
industrial facilities discharging water into the Dominguez Channel.
As defined by the resolution, the Watson Refinery was limited to
1330 pounds per day of COD in its discharge water to the Channel.
Meeting this requirement meant reducing the COD in its discharge
waters by 95% if the water was discharged to the Channel.
The Watson Refinery, as a taxpayer of Los Angeles County, made an
agreement with the Los Angeles County Sewer District to have its
process waste water handled in the County's primary treatment unit.
However, due to limitations in the County unit, the County was un-
able to handle rain water runoff. This presented a problem for the
Watson Refinery due to the fact that the rain water collection
facilities were interconnected with the process waste water collec-
tion system. Therefore, during periods of rainfall; the process
waste water and rain water mixture could not be sent to the sewer
district facilities due to the presence of rain water, nor could
it be sent to the Dominguez Channel due to high COD content of the
process waste water.
To solve this problem a system was needed which would treat all
the process water plus rain water as it was produced, or an im-
pounding plus processing system which would allow large volume im-
pounding during the rain followed by low volume processing. In
either case, a system was needed which could be started up easily
when rain fell and then shut down when no longer required. A
biological unit requires continuous feed and thus the conventional
technology of today was not satisfactory. Therefore, it was
decided to use impounding followed by activated carbon treatment
to adsorb the COD material from the impounded rain diluted process
water.
Construction of the first commercial sized carbon adsorption plant
for treatment of petroleum refinery waste water was completed in
-------
1971 and contained over one-half million pounds of activated carbon.
This report describes the first two year's operation of the unit
and is submitted in fullfillment of Project Number 12050 GTR under
the sponsorship of the Water Quality Research Division of Applied
Science and Technology of the Environmental Protection Agency.
The specific objectives of the project were:
1. Determine feasibility of activated carbon as a treatment
system for storm water runoff and refinery process waters.
2. Evaluate performance of the system.
3. Determine operating costs.
4. Assess reliability of the system.
-------
SECTION IV
PROCESS DESCRIPTION
ADSORPTION
Adsorption is defined as physical attraction of molecules onto a
surface, such as the pore struction of activated carbon. Acti-
vated carbon is carbon that has been processed to obtain surface
areas in the order of 1,000 square meters per gram by penetrating
the particules with molecular size pores. The total surface area
is essentially unchanged even by grinding the material to a fine
powder, as the surface area of the outside of the particule is
small compared to the pore surface area.
Strong adsorption takes place by capillary effect as the pore size
nears the size of the molecules adsorbed. Dissolved organics are
generally more strongly adsorbed than inorganic compounds. Higher
molecular weight compounds generally displace the lower molecular
weight compounds in the pores. Non-polar compounds are usually
more strongly adsorbed than polar compounds, and they will also
displace the polar materials. These and other factors, influence
the net adsorption effect under flow condition in a granular acti-
vated carbon bed.
Activated carbon was used during the two-year project to adsorb
organic chemical oxygen demand materials from the impounded rain
water and process water mixture. Filtrasorb 300, the adsorbent
used for the Watson Carbon adsorption Plant, is an activated car-
bon with granules having an approximate size range of 8 to 30 mesh
made from bituminous coal. It is made to high hardness standards
6
-------
to minimize attrition loss in handling, regeneration, and hydraulic
transport. It has a broad spectrum of pore sizes to meet adsorp-
tion requirements for a braod range of organic molecule sizes.
THERMAL REGENERATION
Filtrasorb 300 is regenerated by selective oxidation of the organic
impurities in the pores. This is done at high temperatures
(1600°F-1750°F) and with a controlled low oxygen atmosphere in a
multiple hearth furnace. As the carbon is heated, the more
volatile organic compounds are vaporized. With further heating
additional organics are pyrolysed. The remaining organics are
then oxidized selectively by addition of air. The carbon is then
quenched in water. Time, temperature, and atmosphere are the
controllable parameters for regeneration. Free oxygen must be
carefully controlled in the lower hearths of the furnace to avoid
burning up the regenerated carbon.
CHLORINATION
As there is an inorganic COD background level in the total COD,
provision was made for chlorination after adsorption. Chlorine
can be added to the effluent stream to permit operating the carbon
beds with some breakthrough of organics as chlorine will reduce
organic COD level as well as inorganic. This provides greater
flexibility in loading the carbon or in handling an unusual COD
load.
-------
SECTION V
DESCRIPTION OF ACTIVATED CARBON PLANT
The water treatment facility is made up of four main systems plus
an impounding reservoir. Figure 1 describes this system. An
overall view of the adsorption plant is given in Figure 2.
RESERVOIR
The reservoir is a 1.2 million barrel holding basin which impounds
all refinery waste water and rain water runoff when the Los Angeles
County Sewer District will not accept it. The reservoir is shown
in Figure 3.
WATER TREATMENT CARBON ADSORPTION UNIT
The water treatment unit is shown in Figure 4. This unit reduces
the organic COD of water impounded in the refinery prior to dis-
charge to the channel. The water treatment unit consists of
twelve identical adsorber cells, (V-l through V-12), each 12' x
12' x 23' deep. Each cell originally contained a 13' deep bed
of carbon having a dry weight of approximately 48,700 pounds.
Supporting the carbon bed is a one-foot layer of gravel on top of
a Leopold underdrain system. The depth of the carbon was altered
in 1972 for reasons discussed later in this report.
Impounded water is delivered to the influent water distribution
trough through a 14" line from the impounding reservoir. The
water is distributed to any or all of the twelve adsorber cells
(V-l through V-12) by slide gates. Flow to each cell is regulated
by a handwheel operated slide gate.
8
-------
1
fff/atat
*r/*-/,-*r.
^sS*^S
ff.'&fSff
tj 'J
Saffftp
V-fS
^V/93J>
J*orf
|
l"-7*
|
!
\.
t ^-i
Figure 1. Carbon Adsorption Plant
Flow Diagram
-------
Figure 2. Overall view of Carbon Plant
10
-------
Figure 3. Impounding Reservoir
Figure 4. Water Treatment Unit
11
«
-------
The waste water first hits a splash plate designed to prevent
packing or pocketing of the bed. A view into an adsorber cell is
shown in Figure 5. The water level will be above the surface of
the carbon bed, and the level varies depending on flow rate and
pressure drop from solids accumulation on the bed. The water
passes down through the carbon bed where it collects in the under-
drain system.
The treated water flows through 6" lines from each cell to a 24"
collection header leading to the effluent retention sump (V-15).
Each 6" cell discharge line has a sample point and an air-operated
pinch valve which can be shut during backwashing.
A chlorine-water solution may be injected into the incoming
treated water stream at the inlet to the effluent retention sump
(V-15). Approximately 15 minutes retention time is allowed for
chlorine contact in this sump. The chlorinator (X-l) injection
rate is manually adjusted to further reduce the COD content. The
treated water collected in sump V-15 overflows through a 24"
underground drain through an outfall box to the channel.
BACKWASH
Each carbon bed must be backwashed whenever it will not pass its
share of water flow due to buildup of solids on top of the carbon.
Frequency of backwashing depends on the rate of build-up of solids
on the carbon bed. A bed is backwashed whenever it will not pass
its share of water flow. When this occurs, flow from the bed is
stopped by closing the pinch valve on the 6" discharge line.
Treated effluent is pumped (P-2) from the backwash sump (V-20) up
through the bed to expand the bed and flush out accumulated solids.
Surface washers, which spray water from nozzels on a pipe rotating
just above the carbon bed, will enhance this action. Bed expansion
is affected by flow rate, water temperature, density, organic
12
-------
Figure 5. View Into An Adsorber Cell
13
-------
loading, and surface wash action. The turbid water overflows into
the backwash troughs to the backwash effluent sump (V-16). From
there it is pumped (P-3) back to the reservoir for settling and
retreating.
CARBON HANDLING SYSTEM
The carbon handling system includes the storage tanks for spent
carbon (V-18), and regenerated carbon (V-17); plus the pump,
eductor, piping, and controls to move spent carbon from any cell
to tank V-18, and regenerated carbon from V-17 back to any cell.
The storage tanks are shown in Figure 6.
When the carbon in a bed is exhausted, the flow of water is stop-
ped, the spent carbon removed, and the bed refilled .with regen-
erated carbon. After a backwashing to remove fines and to
stratify and level the bed, the flow of water is restarted.
The spent carbon is removed from the bed as a water slurry through
a valve in the side of the adsorber cell. The carbon slurry flows
by gravity in a concrete trench to the suction of the spent carbon
transfer pump (P-l). The carbon slurry is transferred by pump
(P-l) to the spent carbon tank (V-18) where it is stored prior to
regeneration.
Regenerated carbon slurry is transferred by gravity from its
storage tank (V-17) via hose to the proper cell. Utility water is
added to assist the transfer.
REGENERATION FURNIACE
The regeneration furnace and gas scrubber are shown in Figure 7.
The slurry of spent carbon is delivered from the bottom of its
storage tank (V-18) to the dewatering screw (M-l) at a controlled
rate. Motive water is provided from the utility system to eductor
J-l. To maintain a constant carbon delivery rate, the water level
in the spent carbon tank (V-18) is kept constant. The carbon
14
-------
Figure 6. Carbon Storage Tanks
15
-------
Figure 7. Carbon Regeneration Furnace and Gas Scrubber
16
-------
settles out from the slurry in the feed end of the dewatering screw
(M-l) located above the furnace. The dewatering screw is set at
an angle so that the water drains from the unit counter-current to
the flow of carbon. The drained carbon discharging into the fur-
nace contains approximately 50% water (wet basis). Excess water
overflows"from the dewatering screw and is returned to the reser-
voir. The dewatered carbon flows by gravity from the dewatering
screw to the top hearth of the regeneration furnace.
The regeneration furnace is a 56" I.D., six hearth multiple hearth
unit. A diagram of the furnace is shown in Figure 8. It is gas-
fired on two hearths. Supplemental air and steam are added on two
hearths. A center shaft rotates arms with teeth which move the
carbon across the hearths and downward through the furnace. The
burners on the furnace automatically control furnace temperatures
at the desired levels via thermocouple element and controllers.
Steam and air addition rates are manually set.
The afterburner section is separately gas-fired to raise off-gas
temperatures to about 1450°F. This is required to combust organic
vapors in the furnace exit gas.
The hot gas from the after burner goes through a quencher where it
is cooled by water injection, is pulled through induced draft fan
(K-4) again with water injection for scrubbing, and exits through
an entrainment separator where entrained water and particulate
matter scrubbed out of the off-gases are removed. The clean flue
gases are exhausted via stack to the atmosphere. Hot air from the
center shaft is added to the stack to reduce the humidity and
minimize the vapor plume.
Regenerated carbon discharges down a chute from the furnace in
periodic slugs as the rabble arms pass over the drop hole on the
bottom hearth. This chute has two legs. Normal carbon flow is
17
-------
Figure 8. Carbon Regeneration Furnace
18
-------
vertically into the quench tank (V-19). Water level in the quench
tank is kept above the "bottom of the chute to prevent air from
being drawn into the furnace. A trash screen is provided in the
quench tank to protect the eductor.
The 45° leg on the furnace discharge chute is used to bypass the
quench tank in case transfer problems are encountered. Opening
the dump gate permits hot carbon to drop directly into drums and
allows continued furnace operation.
Water sprays are provided to quench the hot carbon to eliminate
sudden evolution of steam which would upset furnace pressures.
Quenched carbon is educted to the regenerated carbon tank (V-17)
where it is stored until needed for refilling an adsorber cell.
19
-------
SECTION VI
DESIGN BASIS
The volume of carbon in the adsorbers and Its exhaustion rate are
set by the volume of waste water to be treated, the concentration
and types of adsorbable material in the influent, and the per-
missible COD concentration in the effluent water. Design of the
full scale unit was based on pilot tests performed on diluted
refinery waste water.
The carbon exhaustion rate was difficult to establish from the
pilot plant tests because the influent COD concentration varied
considerably, both above and below the limits expected under
actual operating conditions. The actual design of the unit was
based on the following criteria:
30 days per year of rain (maximum)
300,000 barrels water per rainy day (maximum)
9,000,000 barrels of water per year (maximum)
250 ppm COD average influent concentration
37 ppm COD average effluent concentration
1 pound of carbon exhausted per 1000 gallons
water treated
Based on these criteria, the unit was designed to handle 100,000
barrels of water per day. Thus, the unit would run 90 days if the
maximum rains were received without having to replace or regenerate
any carbon. Based on this operation premise, regeneration of all
carbon would be done during the summer, non-rainy season.
20
-------
The regenerator furnace was designed to regenerate 8,500 pounds of
carbon per day. This is equivalent to 11.3 GPM of slurry from
V-18. The approximate utility consumption, based on design, was
as follows:
Steam - 1.0 Ib. of steam per Ib. of carbon = 354 Ibs./hr
Refinery fuel gas (including afterburner = 3,000 SCFH
21
-------
SECTION VII
OPERATION AND EVALUATION OF WATER TREATMENT FACILITIES
This section of the report covers the operation and evaluation of
the adsorber cells and backwashing facilities.
FIRST PERIOD OF OPERATION
The first period of operation covers the first year from May 1971,
through June 1972.
Initial Operation
The unit was first started in May 1971. The purpose was to test
operation of the unit prior to the rainy season. Test water was
synthesized by mixing high COD process water with service water
and impounding. Impounded water was fed to cell 3 at the design
rate of 250 GPM. With a feed COD of 650 ppm, the effluent remained
in the 44-80 ppm range for the 208 hour run. In addition to the
COD's being above the design level of 37 ppm, the effluent water
was cloudy and had a septic odor. This condition of the effluent
was believed to be caused by anaerobic bacteria growing on the
carbon.
Caustic Treatment of Carbon Beds
It was decided to caustic wash four beds in order to kill the
bacteria and improve COD removal efficiency. The beds were back-
washed with 2% caustic until a pH of 11 broke through the bed
surface. A white precipitate formed, a sour smelling gas was
evolved, and the wash solution had a yellow color. The white
22
-------
precipitate was easily backwashed out, and is believed to have been
a carbonate compound. The sour gas is believed to have been H S,
and the yellow color is believed to have been hydrocarbons releas-
ed from the carbon and dissolved in the caustic solution.
A test with the four treated cells in parallel with four untreated
cells was run for 21 hours to determine the effect of the caustic
wash. The results are given in Figure 9. Although the effluent
COD's were almost identical for the treated and untreated cells,
the treated cells had very little problem with odor and clarity
in the effluent water.
Because of the improved odor and clarity of the water from the
caustic treated cells, all 12 cells were treated. With continued
operation it was soon apparent that the adsorption plant was not
adsorbing as much as expected,
Backwashing
Since the carbon was not adsorbing the COD in the amounts it had
in pilot plant work, it was felt that trapped air might be de-
creasing the available surface area on the carbon. In an attempt
to decrease the trapped air, a more vigorous backwash was used for
three cells. The backwash rate was increased from 3,300 GPM to
5,000 GPM, which resulted in an increased bed expansion from 20%
to 50%.
Although the effluent COD's were temporarily reduced to below
30 ppm, the improvement lasted only one day. The results from
cell three are shown in Table 1. Since the increased backwash rate
showed no permanent improvement, it was discontinued.
The frequency of backwashing was studied during the test runs
described below. In run 4, backwashing was infrequent, while in
run 5 the cells were backwashed daily. Despite a lower feed rate
and lower feed COD in run 5, both run 5 and run 4 had equal effluent
23
-------
£30$-
n
500
— 400
300
NON-CAUSTIC
TREATED CELLS
CAUSTIC TREATED CELLS
8
10 12 14
ELAPSED TIME, HOURS
9 . Caxiatijc. Ti:e.a.tijn.e. Test
16
18
20
22
-------
Table 1. COD DATA BEFORE AND AFTER VIGOROUS BACKWASHING
AT 5000 GPM
CELL NO. 3
FEED RATE = 250 GPM
J3ATE
9-15-71
9-24-71
J-25-71
>-26-71
TIME
11:00 AM
1:00 PM
3:00 PM
11:45 AM &
For 10 and
3:00 PM
6:00 PM
7:30 PM
11:00 PM
7:00 AM
3:00 PM
11:00 PM
7:00 AM
11:00 PM
FEED ppm EFFLUENT ppm
194
189
264
83
89
114
12:45 PM backwashed @ 5000 GPM
11 minutes respectively
75
—
167
119
119
113
116
181
118
<30
42
53
72
76
74
77
33
76
25
-------
COD's. Increased frequency of backwashlng is not felt to be an
Improvement for COD removal.
COD Penetration Into Carbon Bed
Cell number three was stopped after 208 hours of operation to study
the problem of high effluent COD. Besides the testing of caustic
washes and backwashing, data was collected to determine the rela-
tive efficiency at various depths in the carbon bed. The data is
shown in Figure 10. The relative efficiency test, which is de-
scribed in Appendix B, is an indication of the carbon's ability to
adsorb COD compared with virgin carbon. The reliability of the
test is discussed in the "Test Methods and Their Evaluation" sec-
tion of this report. The data in Figure 10 indicates that the
top two feet of carbon were completely exhausted while the last
five feet were still at 65% of virgin adsorption ability.
Exhaustion Run
Cell three was run for 247 more hours during which time the plant
was stopped once on September 17, 1971, and restarted again on
September 24, 1971. The results for the last 247 hours are shown
in Figure 11. During the seven days the plant was down, the feed
COD decreased about 180 ppm. It appears that natural oxidation
reduced the COD in the impounding reservoir.
Cell three was regenerated when feed and effluent COD's converged
in the 431st hour. The exhaustion rate could not be calculated be-
cause by design criteria the bed is exhausted when the effluent
COD exceeds 37 ppm. The effluent was rarely below 40 ppm. Based
on the entire run, the carbon was used at the rate of seven pounds
of carbon per 1,000 gallons of water treated.
Varying Flow Rates
As rainfall was extremely light, there was very little water pro-
cessed by the carbon plant in the winter of 1971 and 1972. In
addition to the light rainfall, the sanitation district accepted
26
-------
z
LJJ
•—i
CJ
u_
u_
LLJ
LJ.I
ISJ
0
45678
DEPTH DOWN INTO CARBON BED, FEET
10
Figure 10. Relative Efficiency Profile of Cell No. 3 After 208 Hours at 250 GPM
650 to 600 PPM COD In Feed
-------
320
280
240
o.
Q.
Q
O
120
80
40
I I
200 220 240 260 280 300
320 340
TIME, HOUR
360 380 400 420 440 460
Figure 11. Cell Number Three COD Vs Time First Rains
-------
process waters with COD's above 450 ppm 24 hours after rainfall
stopped. Therefore, water was only processed when there was water
in the impounding reservoir with a COD below 450 ppm. Throughout
the winter, six runs were made with varying flow rates. Data
from the six runs is given in Figure 12.
Run 1
The carbon plant was started at 4 p.m. on December 25, 1971, with
a design feed rate of 3,000 GPM. All 12 cells were in service.
The run lasted 44 hours with an average feed COD of 326 ppm and
an average effluent COD of 43 ppm. The run was stopped when the
feed COD jumped above 450 ppm. During this run, samples of feed
and effluent were tested for COD, turbidity, color, odor, and
suspended solids. The results of these tests are listed in Table 2.
Run_2_
Run 2 was started when the feed COD dropped to 415 ppm. This run
which lasted 47.5 hours had an average feed COD of 360 ppm and an
average effluent COD of 48 ppm. Halfway through the run, the rate
was reduced to 2,000 GPM to comply with channel discharge limita-
tions.
•SyjL_3
Again 2,000 GPM was continued as effluent COD's average 80 ppm.
The feed averaged 374 ppm and the run lasted 38.5 hours.
J*un_4_
The fourth run lasted only 18 hours with a feed rate of 2,000 GPM.
The average feed COD was 310 ppm and the effluent averaged 67 ppm.
|un_5_
The effluent was still well above 37 ppm COD with an average COD
°f 66 ppm. The feed rate was cut to 1,000 GPM to assure com-
pliance with the COD discharge limitation. The feed averaged
29
-------
900 —
800 —
1-RUN 6-|
20
40
60
80
100 120 140
TIME , HOURS
160
180
200 220 240 260 280
"Figure 12. "Feed and Effluent COD Data "First Rains
-------
Table 2. LABORATORY DATA ON CARBON PLANT FEED AND EFFLUENT
Date
•*••*• !•! 1
12-27-71
12-28-71
12-29-71
Effluent
12-27-71
12-28-71
12-29-71
COD
690
415
370
43
15
40
Turbidity
J.T.U.
56
41
36
14
14
12
Color
65
72
60
11
15
13
Odor
8
8
12
1
1
1
Total Suspended
Solids, ppm
122
64
132
11
20
34
237 ppm COD if the extremely high peak in Figure 12 of 900 ppm COD
is discredited. This was the longest of the six runs and lasted
95 hours.
Run six lasted 22 hours with an average feed COD of 147 ppm. Al-
though the feed COD was low, it had a green algae which was not
adsorbed or filtered by the carbon bed. The effluent which
averaged 93 ppm COD was also green in color. The problem with
algae occurred again during later testing and is discussed under
the "Two-State Test Run" heading of this section.
^valuation of First Rain's Operation
several attempts to improve operation, the carbon adsorp-
tion plant was unable to process water at the design rate of
100,000 barrels per day, and still meet county regulations. Fort-
unately, rains were light, and the sanitation district accepted
high COD water.
31
-------
The COD loadings of each cell vs. gallons of water treated are
given in Figures 13 through 18. The loadings varied from 0.2 pounds
of COD per pound of carbon to about 0.3, at the end of the first
rainy season. Since the effluent COD's remained about 37 ppm, a
precise determination of COD loadings at breakthrough is not
possible. Therefore, an estimate was made in September, 1972 as
follows:
Feed Loading Carbon Exhaustion Rate
COD, ppm Lbs. COD/Lb. Carbon Lbs. of Carbon/1,000 Gals. Treated
150 0.35 3
250 0.30 6
This estimate is based on the assumption that the beds are extreme-
ly sensitive to surges of COD. The surges may penetrate deep into
the bed and prevent the necessary COD removal in the low range.
Since rainfall was light, dilution of the COD with rain water was
greatly decreased. The result was extremely erratic feed COD con-
centrations and poor carbon plant operation. It should be empha-
sized, however, that the estimates are based on assumptions which
have not been validated and the estimated loadings have never
been achieved prior to breakthrough.
TWO-STAGE TEST RUN
The data in Figure 10 shows that the carbon near the top of the
bed adsorbs much more COD than the lower portion of the bed. There-
fore, if the bed could be cut in half, only the more heavily
loaded top portion would need to be regenerated. This would allow
a much higher COD loading. Since an individual cell cannot be
split in half, a system was devised to test two cells in series.
Description of System and Operation
Cells number two, six, and seven were used in a two-stage test run
from July 12, 1972, through August 30, 1972. Both phases of the
downflow upflow system are shown in Figure 19. In order to
32
-------
CO
u>
a
o
CD
o
o
o
CO
a:
0
FIGURE 13
COD LOADING
FIRST RAINY PERIOD
CELL NO, 1
1000 2000 3000
VOLUME OF WATER TREATED, THOUSANDS OF GALLONS
4000
Figure 13. COD Loading First Rainy Period Cell No. 1
-------
o
m
a:
<
o
Co
O
<
O
O
CQ
QC
o
LU
LU
o;
0
1000
2000
3000
4000
VOLUME OF WATER TREATED, THOUSANDS OF GALLONS
14. COB l^oaditig ¥lrst Rainy Period Cell No. 2
_J
5000
-------
.30
z
o
CQ
OC
<
O
.20
LO
Q
O
CJ
.10
O
PQ
o:
<
u
UJ
o:
UJ
ill
CJ3
UJ
Of-
2000
3000
4000
6000
10000
VOLUME OF WATER TREATED, THOUSANDS OF GALLONS
Figure 15. COD Loading First Rainy Period Cell No. 3
-------
1000 2000 3000
VOLUME OF WATER TREATED, THOUSANDS OF GALLONS
Figure 16. COD Loadings First Rainy Period Cell No. 4
4000
-------
COD LOADING
O
PQ
O
FIRST RAINY PERIOD
CELL NOS, 5 AND NO, 7 THROUGH 12
0
1000
2000
3000
4000
VOLUME OF WATER TREATED, THOUSANDS OF GALLONS
Figure 17. COD Loading First Rainy Period Cell Nos. 5
and No. 7 through 12
-------
CD
CQ
o
3
<
o
CQ
OL
0.20
oq
O
u
0.10
0
1
1
1000
2000 3000 4000
VOLUME OF WATER TREATED THOUSANDS OF GALLONS
Figure 18. COD L.oadin.g, ¥irst Rainy Period Cell 15o. 6
'3000
-------
PHASE 1 OF TWO STAGE SYSTEM
•MTER FROM
RESERVOIR
1
STAGE
1
CELL
2
TO CHANNEL
I
STAGE
2
CELL
6
CELL
7
- ^
* t
PHASE II OF TWO STAGE SYSTEM
'/ATER FROM RESERVOIR
TO CHANNEL
CELL
2
J
STAGE
1
CELL
6
!
STAOE
2
CELL
7
».
4 t
Figure 19. Two Stage System
39
-------
increase the adsorptive capacity of the unit, 15,000 pounds of
carbon were added to the existing 48,700 pounds in each of the
cells used. This increased the bed depths from 13 feet to 17 feet
to give a total of 63,700 pounds in each cell. The first phase
used cell number two as the first stage with downflow, and cell six
as the second stage with upflow. At the beginning of the second
phase, cell number two was taken out of service, cell number six
became stage one -with downflow, and cell number seven was placed
in service as the second stage with upflow. Since no rain fell
during the summer, the feedwater was made up in the impounding
reservoir every two or three days using several water sources in
an attempt to maintain the COD concentration at 250 ppm.
The feed rate to each cell during the first phase of the test run
was held constant at 500 GPM. This is equivalent to the design
rate of 3,000 GPM for the entire plant. The feed rate was reduced
to 250 GPM for most of the second phase to see if a lower effluent
COD could be reached.
Results
The results of the two-stage run are shown in Figures 20 and 21.
Figure 20 shows the COD concentrations of the feed, first stage
effluent and second stage effluent as a function of gallons of
water processed in the unit. Figure 21 shows the COD loading of
each cell vs. gallons of water processed. At the end of Phase 1,
the first stage had a loading of 0.39 pounds of COD per pound of
carbon and the second stage had a loading of 0.11. The effluent
COD at the end of Phase 1 was well above the desired 37 ppm.
There was also a problem with algae growth, which, as discussed
later in this section, was adversely affecting the test run re-
sults. Since lower COD effluents were not detected during Phase 2
and because the unit had to be readied for the next rainy season,
the run was terminated. As can be seen in Figure 20, the effluent
remained below 100 ppm virtually the entire run.
40
-------
400
2000 4000
6000 8000- 10000 12000 140QO 16000 18000 20000 22000
VOLUME OF WATER TREATED, THOUSANDS OF GALLONS
2*000 26000 28000
Figure 20. COD Concentrations Two-Stage Test Run
-------
S.30[
CL.
<
O
9
o
o
J.20
a
<
o
o
OQ
. 10
FIGURE 21
CARBON LOADING TWO-STAGE TEST RUN
CELL-2
STAGE 1
2000
4000
6000
8000 10000 12000 14000 16000 18000 20000 22000
VOLUME OF WATER TREATED, THOUSANDS OF GALLONS
26000 28000
Figure 21. Carbon Loading Two-Stage Test Run
-------
Validity of Results
By the definition used for the Watson Carbon Plant, the carbon is
considered loaded when it contains so much COD that its adsorption
capability is reduced to the point where it cannot adsorb enough
COD to keep the effluent below 37 ppm. Ideally, this loading could
be determined by running the carbon plant under the expected rainy
season conditions until the effluent becomes greater than 37 ppm,
Or> in other words, until "breakthrough" occurs. Unfortunately,
the system cannot be operated at a steady feed concentration of
250 ppm COD until there is enough rainfall to dilute the COD to
250 ppm. The COD during test runs could not be kept at the desired
concentration on a day to day basis because it takes several hours
to determine the COD, and thus, any concentration surge would be
detected too late. Therefore, an estimate of operation under ideal
conditions free of surges and algae was made for the two-stage
system as was previously done for the single stage system. This
estimate is that the two-stage system can load to 0.39 pounds of
COD per pound of carbon. This is the same as attained by the
first stage during Phase 1 of the test run.
Growth
growth in the water impounding reservoir first appeared at
the end of the first rainy season. The algae was easily identified
as it gave the water a green coloration. The carbon adsorption
Plant was adversely affected by the algae as indicated in run six
Curing the first rains. During this test run, the carbon plant
had one of the lowest average feed COD concentrations (147 ppm)
tested, and yet, due to the presence of algae, the effluent averaged
93 ppm COD.
*he algae problem appeared again in July 18 in the water fed dur-
the two-stage test run. A study of reservoir algae control
43
-------
revealed that algae growth can be kept low by use of small concen-
trations of copper sulfate. On July 28, solid copper sulfate was
spread over the surface of the reservoir to give a content of about
0.0002 weight percent. Within 72 hours the algae was greatly re-
duced. Since the effect of the CuSO, is not permanent and no more
addition of CuSO was made, the algae became so concentrated that
the carbon plant had to be shut down on August 18. Copper sulfate
was again spread over the reservoir on August 22 to reduce the
algae. Three days later it was added again which kept the algae
at low levels until the end of the two stage test run.
Since the algae problem was never experienced again, no further
investigation was made into its control.
Evaluation of Two-Stage System
The one-stage system can load to an estimated 0.30 while the two-
stage system can load to an estimated 0.39. This gives the two-
stage system a distinct advantage over the previous single stage
operation. The two-stage system also has the advantage that it
ran longer than any test run and had one of the lowest average
effluent COD concentrations (47 ppm). It should be re-emphasized,
however, that the predicted loadings of the carbon used for com-
parison have never actually been achieved prior to breakthrough.
SECOND RAINY PERIOD
Single Stage System
Although plans called for converting the carbon plant to a two-
stage system if it proved superior, a change in Los Angeles
County sanitation regulations allowed rain diluted waste water to
be pumped to the sewer. Previously, only limited quantities of
impounded water were permitted, and then only if COD was above
450 ppm. The plant was used as a single stage system to process
impounded water from October 1972, through July 1973, since
pumping limitations prevented discharging all impounded water to
the sewer.
44
-------
The 12 carbon cells were divided into two groups: (1) those for
testing purposes; and (2) those for handling the bulk rain water.
To allow the accumulation of additional data under controlled
conditions relative to alternative modes of operation, five of the
cells were designated for use only for testing purpose. The re-
maining seven were designated for use in processing the bulk rain
Water as it was impounded.
.Staggered Mode Operation Test Run
The choice of five cells for test purposes is based on the calcu-
lations shown in Table 3. These calculations were used to estimate
the maximum number of cells which could be operated continulusly
based on predicted plant limitations. The calculation is simply a
determination of how much water can be processed to load carbon at
the same rate it can be regenerated. A staggered operation was
devised to allow three cells to operate at all times without delay
for regeneration. In order to start the test run, it was neces-
sary to start one cell each week. At the end of three weeks, a
fourth cell was started while the first cell was moved to the
spent carbon tank and regenerated. At the end of the fourth week,
the first cell had been regenerated and was ready for use. However,
a fifth cell was used and the first was put on stand-by to be used
at the end of the fifth week. The extra fifth cell was used as a
buffer against temporary mechanical holdups. The mode of operation
is shown in Figure 22.
Since the impounding reservoir had a high COD content, effluent
Water was recycled back to dilute the feed as is shown in Figure 23.
The total feed rate to each cell, which includes the recycle was
maintained at 250 GPM while the unit was in operation. Air was
dispersed into the effluent sump to kill anerobic bacteria that
might be recycled to the beds.
45
-------
Table 3. ESTIMATE OF MAXIMUM WATER THROUGHPUT
Estimated Feed COD 300 ppm
Estimated Loading .30 Lb. COD/Lb. Carbon
Estimated Carbon Regeneration Ability 9,000 Lb. Carbon/Day
Estimated Effluent COD 30 ppm
8.34 x 10 6 Lb. COD . (300 ppm - 30 ppm) =
(Gal. H20)(ppm COD)
2.252 x 10~10 Lb. COD
(Gal. H20)
9,000 Lb. Carbon/Day (.30 Lb. COD/Lb. Carbon) = 2,700 Lb. COD/Day
2,700 Lb. COD/Day
2.252 x 10-3 Lb. COD/(Gal. H20) 24 Hrs./D 60 M/Hrs. = 833 GPM
With a feed rate to each cell of 250 gpm, 3 cells were used for a
total of 750 gpm.
46
-------
TWO
THREE
FOUR
FIVE
SIX
FIRST
CELL
LN
USE
IM
USE
I • v
USE
REGEN.
STAND
3Y
IN
USE
SECOND
CELL
in
USE
IN
USE
IN
USE
REG EN.
STA'!:')
3Y
THIR:)
CELL
IN
USE
IN
USE
IN
USE
REGEr!.
FOURTH
CELL
IN
USE
IN
USE
IN
USE
FIFTH
CELL
IN
USE
IN
USE
Figure 22. Five-Cell Staggered Operation Test Run
47
-------
FRESH FEED
1 "-^ " ••
i
-t*
00
k
1 1
BULK
TEST PROC.
CELLS CELLS
"
EFFLUENT
SUMP
CI2 INJECTION | |
j
RECYCLE
TO CHANNEL
Figure 23. Second Rains Recycle Operation
-------
The COD concentrations and loadings of each, of the five cells are
plotted in Figures 24 through 28.
After the cells had finished one cycle, it was apparent that the
carbon could not keep the effluent COD below the desired 37 ppm
f°r a full week, despite the dilution of the feed. Therefore, it
vas decided to chlorinate the effluent in the hope that COD mole-
cules which had not been adsorbed on their first pass through the
b^d might form a halogenated molecule which would be more likely
to be adsorbed. The chlorine was added at 100 pounds per day for
the remainder of the test. Since it was difficult to determine
the effects of the chlorine addition, the cells were not switched
as scheduled. They were run an extra two weeks before it was
decided that there were no noticeable improvements with the
chlorine.
Bulk Rain Water
n of the 12 cells were used for processing bulk rain water dur
the second rainy season. Bulk rain water refers to water in
reservoir when the level in the reservoir is high enough to
limit the ability to impound water beyond reasonable expectations.
About 40.6 million gallons of water were processed through the
8even cells. The influent and effluent COD's along with the
carbon loading are plotted vs. gallons of total water treated in
^igure 29. The chlorine addition start time is also indicated in
pigure 29. When the rain season ended, the cells were loaded to
°-3l pounds of COD per pound of carbon.
•Valuation of Second Rainv Period
g the second rainy period 102,000,000 gallons of water were
Processed. The average diluted feed COD concentration was 233 ppm
vhile the effluent averaged 48 ppm with a high of 95 ppm. The
carbon loaded to an average of 0.26 pounds of COD per pound of
.
49
-------
E
o.
a
o
O
3OO
aoo
100
o
S
^ 0.40
o
o
o
S 0.30
5 O.EOH
a
1
te.
0.10
0
FEED
EFFLUENT
o
ID
UJ
UJ
o
Ul
a:
1
I
2000 4000 6000 8000 0 2000 4000 6000 8000 10000 12000
VOLUME OF WATER TREATED, thousands of gallons
Pigure 24.
Adsorption Data Second Rains
Cell 1
50
-------
a
Q.
Q
8
300
200
100
§ 0.40
-------
300
200
o
8 100
o
o
^ 0.40
o
o
o
„ 0.30
o
5 O.20
o
§
0.10
z
o
OQ
t£.
<
O
FEED
— EFFLUENT
o
UJ
o
Ul
I
I
I
0 2000 4000 6000 8000 0 2000 4000
VOLUME OF WATER TREATED,
thousands of gallons
Figure 26.
Adsorption Data Second Rains
Cell 5
52
-------
E
a
a
a
O
o
1
a
u
I
2 0.20 —
5
§
CD
a:
2
300
200
100
0.40
030
0.10 —
FEED
EFFLUENT
o
uj
(C
IU
z
UJ
o
UJ
or
I
I
I
I
0 2000 4000 6000 8000 10000 12000 14000
VOLUME OF WATER TREATED, thousands of gallons
Figure 27.
Adsorption Data Second Rains
Cell 7
53
-------
E
a.
Q.
o
o
o
300
200
100
0
e
o
O
u
Q
O
O
0.40
0.30
2
o
z
5
<
o
o
CD
r 0.20
0.10
FEED
EFFLUENT
z
UJ
o
UJ
cc
I
I
I
1
1
0 2000 4000 6000 8000 10000 12000 14000
VOLUME OF WATER TREATED, thousands of gallons
Figure 28.
Adsorption Data Second Rains
Cell 9
54
-------
FEED
E
a.
o
o
u
300
200
100
EFFLUENT
c
o
o
u
o
O
O
0.40
0.30
o
I
Uf
o
Ul
DC
0.20-
o
<
3
o
o
oc
0.10-
I
I
I
I
2000 4000 6000 8000 10000
VOLUME OF WATER TREATED,
thousand! of gallons
Figure 29.
Adsorption Data Second Rains
Cell Nos. 2,4,6,8,10,12,11
55
-------
Although the carbon plant was operated without violating the limit
of 1330 pounds per day of COD, a large portion of the impounded
water was not processed in the plant. The change in county regula-
tions relieved the need for meeting design operation. The con-
trolled COD level in the feed allowed for the highest average carbon
loading during the two-year project.
The five-cell test run showed that control of the feed COD concen-
tration Improves carbon loading.
56
-------
SECTION VIII
OPERATION AND EVALUATION OF REGENERATION FACILITIES
This section of the report covers the operation of the carbon re-
generation furnace and carbon transfer equipment.
FIRST PERIOD OF OPERATION
Twelve cells were regenerated during the first period of operation.
The time required for regeneration for the first seven cells is
tabulated in Table 4. The time includes all downtime for repairs.
The regenerations vary in time from 4.0 days to 14 days with an
average of 9.8 days. The unit was designed to regenerate a bed in
5.7 days. The causes of the extended regeneration time are dis-
cussed below.
.Temperature Excursions
During the first regenerations the temperature on hearths four and
six would fluctuate from 1700°F to 1850°F in 10 to 20 minute cycles.
Several attempts were made to adjust the controllers but there was
no improvement. The thermowells were changed from 3/4" standard
wall to 1" heavy wall stainless steel and the thermocouples were
changed from iron constintan to chrome alumal. The temperature
controller on hearth six was found in need of repairs and was
fixed. The temperature excursions were reduced to only 50°F in
future runs.
jj-ductor Weajr
The original design used 1" cast iron eductors and 110 pound per
square inch motive water to transport carbon to the furnace screw
57
-------
Table 4. TIME REQUIRED FOR REGENERATION
TIME REQUIRED
CELL NO.
3 9 AM,
2 2 PM,
6 9 AM,
8 4 PM,
10 3 PM,
12 2 PM,
11 1 AM,
START
Nov. 2,
Feb. 17,
April 19
May 11,
May 26,
June 12,
June 20,
1971
1972
, 1972
1972
1972
1972
1972
10 PM
4 AM,
6 PM,
10 PM
3 PM,
12 N,
3 PM,
FINISH
, Nov. 10
March 1,
April 27
, May 20,
June 9,
June 17,
June 30,
, 1971
1972
, 1972
1972
1972
1972
1972
Total
Average
hours
205
302
201
222
336
118
254
1638
234
days
8.5
12.
8.
9.
14.
4.
10.
68.
9.
6
4
3
0
9
6
HM
3
8
feeder and away from the furnace regenerated carbon quench tank.
The capacity of these eductors when in new condition was about 6.5
pounds of carbon per minute. Due to erosion of the throat, both
of the eductors lost efficiency after three regenerations. This
loss of efficiency limited the feed to the furnace. In addition to
the eductors wearing out, they had many problems with plugging, but
this was alleviated with screens to stop large chunks from entering
the eductors.
High Gas Velocities
A problem with fines blowing out the stack was encountered, but
was alleviated by lowering the steam rate.
58
-------
High Shaft Rates
Due to mechanical problems, the shaft which moves the rabble arms
over the carbon was rotating at three RPM instead of the design
rate of 1 RPM. The rate was reduced to two RPM's after adjustments
were made. During the seventh regeneration, the bolt which locks
the motor drive shaft to the drive gear broke. This may have been
due to running the rabble arms above design rate.
.Evaluation of Furnace Problems During the First Period of Operation
Most of the problems encountered with the furnace operation were
mechanical in nature. Therefore, they could be solved as more
experience was obtained.
TWO-STAGE TEST PERIOD
Operation
The three cells used in the two-stage test run were regenerated
without stoppage for mechanical problems. The major reason for
the success was larger eductors and reduced motive water pressure.
The 1" cast iron eductors were replaced with 1 1/2" stainless
steel eductors. The changes allowed the carbon to be moved without
Wear or plugging. The uninterrupted feed of carbon allowed for a
steady operation and the temperature excursion problem ceased to
exist. The new eductors allowed the flow of carbon to go as high
as 10 pounds per minute. The data from the regeneration is given
in Table 5. Figure 30 shows a plot of regenerated carbon relative
efficiency vs. carbon feed rate to the regenerator.
^valuation of Improved Regeneration
The larger eductor is believed to have solved most of the regener-
ator problems.
pigure 30 is based on limited data, and thus, the accuracy is
questionable. However, the general trend of decreased quality of
*egenerated carbon with increased feed rate is believed correct.
59
-------
Table 5, DATA FROM CARBON REGENERATION
DATE
September 18
19
20
21
22
23
24
25
26
27
CARBON FEED RATE APPARENT
pounds /min. SPENT
.59
.58
6-6.5 .58
7-7.5
8-8.5
.52
10
.55
.55
DENSITY
REGEN.
.495
.52
.515
.49
.48
.48
.475
.465
.48
.46
RELATIVE EFFICIENCY
SPENT REGEN .
95.8%
94.3%
88.0%
88.0%
64.6% 79.7%
86.5%
27.6% 86.5%
67.7% 85.9%
SECOND RAINY PERIOD
During the second rainy period, regenerations continued with no
problems. A major portion of the time was spent collecting data
for improving testing techniques which are discussed in detail in
the "Test Methods and Their Evaluation" section of this report.
On April 23, 1973, the APCD tested the regenerator furnace stack
for CO and found the CO content over the allowable concentration of
0.2% by volume. This forced a shutdown of the furnace. It was
felt that the insufficient combustion efficiency was caused by the
combustion air damper set nearly closed and a low afterburner
temperature.
60
-------
96 L
7 8 9 10
CARBON FE€D RATE TO REGENERATOR FURNACE, LB,/MINUTE
11
Figure 30. Furnace Operation
-------
Upon obtaining a test permit the furnace was restarted in October
1973, with the combustion air damper wide open and the temperature
controller increased from 1400°F to 1550°F. With these changes
the CO is well below pollution concentration.
62
-------
SECTION IX
TEST METHODS AND THEIR EVALUATION
GOD DETERMINATION
There are several means of determining the chemical oxygen demand
(COD) of a solution. Three of these methods, which are described
in Appendix A are EPA HIGH LEVEL method (STORET No. 00340), EPA
LOW LEVEL method (STORET No. 00335), and the AquaRator method.
The criteria considered in deciding which method to use for COD
determination were time necessary to obtain results, accuracy, and
acceptability of results. Both EPA methods are well established
hut are time consuming. The AquaRator, on the other hand is
extremely new and yields results in minutes. A series of tests
Were run using two types of standard solutions to determine the
accuracy of each of these methods. The results are given in-
Table 6.
The high level method was selected because it is widely accepted
and because it covers the required range of COD's. It has a
""ajor drawback, however, because it takes about four hours to
obtain results. The low level method was not used because of its
limited range, and because it was not felt to have any superior
a^curacy when compared to the other methods in the 30 to 50 ppm
COD range.
The usefulness of the AquaRator was the subject of continuous
discussion. Results could be obtained in about 20 minutes. At
tines the device would yield reproducible and fairly accurate
63
-------
Table 6. ACCURACY OF COD TESTING PROCEDURES
mg/1
STANDARD SOLUTION
RUN NUMBER
EPA HIGH LEVEL
COD METHOD
STORET NO. 00340
SUCROSE
SODIUM ACETATE
EPA LOW LEVEL
COD METHOD
STORET NO. 00335
SUCROSE
SODIUM ACETATE
AQUARATOR
COD METHOD
SUCROSE*
ON
•P-
COD
STANDARDS
mg/liter
10
30
50
120
250
500
15
32
52
124
266
551
14
28
48
125
267
549
14
32
54
123
261
551
10
26
44
106
240
484
12
26
44
107
230
478
12
21
52
108
227
476
18 12 14 11 9 8 13
35 36 32 31 31 32 29
55 57 57 46 46 46 45
116
221
423
12
30
44
117
215
428
12
31
46
118
223
423
* For AquaRator, sodium acetate solutions were used as standards.
-------
results, while at other times it yielded erratic results. Due to
these problems, it was not felt that the AquaRator could be used
for determining effluent COD's despite the speed of obtaining re-
sults. The AquaRator was used during the second rainy period to
aid in determining recycle rate. Because of the need to keep net
feed level at about 250 pptn COD, up to date results were required.
The EPA method was too time consuming to be used as a control aid.
Another subject of concern was the identification of COD materials,
COD is referred to in this report as if it were a known, identi-
fiable substance, but actually it is a wide variety of materials
which can consume dissolved oxygen. These materials can vary in
size from several atoms to large complicated molecules and can be
organic or inorganic. When these various materials are adsorbed
onto activated carbon they will all affect the loading of the car-
bon in different ways. For this reason, it would be of value to
know what materials make up the COD in the feed and effluent
streams. With a knowledge of what is adsorbed and what is not
adsorbed, improvements in operations might be made.
An attempt was made to identify the COD materials, but due to the
small amounts and difficulty of analyzing hydrocarbons and water
mixtures, very limited results were obtained. These results were
obtained by the United States Environmental Protection Agency's
Petroleum-Organic Chemicals Wastes Section, Treatment and Control
Technology Branch and are given in Appendix C. These findings
indicated the organic compounds in the effluent have a higher per-
centage of higher molecular weight compounds than the feed.
More work is needed in the area before a true understanding COD
adsorption onto activated carbon can be attained.
65
-------
REGENERATED CARBON TEST METHODS
A major effort was put forth during the two-year project to find a
reliable means to determine the quality of regenerated carbon. The
major reason for the study was to give operating personnel a target
for furnace operation.
Two tests which were run on a regular basis during regenerations
are the apparent bulk density (ABD) test and the relative efficiency
(RE) test. These test methods are described in detail in Appendix
B. The apparent bulk density test takes about two minutes to run
and is simply a quick determination of unpacked density of the car-
bon. The relative efficiency test is a time consuming comparison
of the adsorptive capacity of carbon sample with virgin carbon.
The criteria considered in selecting which test method to use as
an operational tool for the regenerator furnace were time required,
accuracy, and reliability of results. During regenerations, operat-
ing personnel have relied on the ABD test simply because it can be
run on site in a couple of minutes. If the delta ABD between
spent and regenerated carbon changes significantly from 0.06 gr/cc,
the operating personnel alter the burning rate. The RE tests take
several hours to run and are, therefore, not used as an operational
tool. The RE tests do, however, allow a more direct comparison of
adsorptive ability compared to virgin carbon. The mode of opera-
tion used has resulted in carbon of essentially constant quality
based on RE tests within the limits of the determinations Repre-
sentative ABD, Delta ABD, and RE data are given in Table 7.
Since, as described previously in this section, there are many
types of COD material with different adsorption characteristics,
it was decided to use process water instead of sucrose as a stan-
dard for the RE test. This was done in the hope that the process
water would be more likely to be adsorbed in a similar matter as
66
-------
Table 7. COMPARISON OF ABD AND RE TESTS
First Rains
Date
11/3/71
11-5-71
11-6-71
11-7-71
11-9-71
11-10-71
Second Rains
3-3-73
3-9-73
3-10-73
3-13-73
3-15-73
3-17-73
3-24-73
Regenerations
Regenerated Delta
ABD ABD
0.531
0.504 0.07
0.587
0.499
0.499
0.493
Regenerations*
^*
0.50 0.05
0.49 0.06
0.50 0.06
0.49 0.07
0.50 0.07
0.47 0.07
0.48 0.05
Relative
Efficiency
66%
59%
69%
72%
83%
82%
102
102
102
101
98
98
98
*Further data is given in Table 8 for the second rains regeneration.
67
-------
the carbon plant feed. The results of the RE tests using the two
standards are compared in Table 8. The RE tests using processs
water give erratic results. This is believed to be due to the
fact that the process water used for the tests is obtained on
different days and thus is not a good standard.
Therefore, even though the sucrose RE test is not necessarily
representative of how COD material will adsorb, it appears to be
a better test for comparing regenerated carbons.
68
-------
TABLE 3
COMPARISON OF RELATIVE EFFICIENCY TESTS
OR REGENERATED CARBON
DATE
1-19-73
1-20-73
1-21-73
1-22-73
1-23-73
1-24-73
1-26-73
1-27-73
1-28-73
1-29-73
1-30-73
1-31-73
2-1-73
2-2-73
2-3-73
2-4-73
2-5-73
2-21-73
2-22-73
2-23-73
2-24-73
2-25-73"
2-28-73
3-1-73
3-3-73
3-4-73
3-5-73
3-6-73
3-9-73
3-10-73
3-11-73
3-12-73
3-13-73
3-14-73
3-15-73
3-16-73
3-17-73
3-18-73
3-19-73
3-20-73
3-21-73
3-23-73
3-24-73
3-25-73
ABD _
0.44
0.41
0.44
0.44
0.43
0.43
0.47
0.48
0.47
0.45
0.43
0.45
0.47
0.48
0.48
0.46
0.45
0.46
0.50
0.50
0.47
0.44
0.48
0.52
0.50
0.52
0.46
0.45
0.49
0.50
0.47
0.48
0.49
0.48
0.50
0.48
0.47
0.46
0.47
0.50
0.48
0.50
0.48
0.46
RE
USING
SUCROSE
98
99
98
100
100
100
99
100
98
97
99
96
99
99
99
99
99
102
102
102
102
102
102
99
102
101
101
100
102
102
102
102
101
97
98
98
98
98
99
97
98
99
98
99
RE
USING
PROCESS
WATER
101
103
114
101
101
98
99
100
99
96
98
101
113
111
114
114
111
116
116
120
116
118
114
113
111
111
111
110
115
116
118
118
113
105
113
113
109
109
108
108
108
108
110
111
69
-------
SECTION X
QUANTITIES AND COSTS BASED ON CONDITIONS DURING THE PROJECT
During the two-year project, 172,040,000 gallons of water was pro-
cessed to load 1,643,700 pounds of carbon with 407,890 pounds of
COD. This resulted in an average carbon loading of 0.25 pounds of
COD per pound of carbon. The carbon was used as the rate of 9.5
pounds per 1000 gallons of water treated. The average feed COD
was 249 ppm and the average effluent was 50 ppm. Averaged data
for each of the three periods of operation is given in Table 9.
The data in Table 9 shows that the second rains single stage operation
had a slightly higher loading than the first rains single stage
operation, and a lower effluent COD.
The overall average cost to operate the plant was 49C per 1000 gal-
lons of water treated, or 24C per pound of COD removed from the
water. The cost summaries are given for both years' operation in
Table 10.
As can be seen in Figure 10, the cost was greatly reduced the sec-
ond year of operation due to improvements in the system. The
operating labor and carbon costs remained close to the same while
the repair labor and utilities costs accounted for the major de-
crease.
With the improved operation of the second year, the plant demon-
strated the ability to operate at only 40? per 1000 gallons of water
treated, or 18
-------
Table 9. ADSORPTION DATA
FIRST RAINS - SINGLE STAGE
TWO STAGE TEST RUN
SECOND RAINS - SINGLE STAGE
TOTAL
FEED COD, ppm
AVERAGE HIGH
377 490*
235 395
233 345
249 490*
EFFLUENT
AVERAGE
67
47
48
50
COD, ppm
HIGH
158
105
95
158
CARBON LOADING
Ib. COD/lb. carbon
AVERAGE
0.23
0.22
0.26
0.25
HIGH
0.30
0.39
0.39
0.39
*The high value of 490 ppm COD is based on the assumption
that the peak of 900 ppm COD during run 5 of the first
rains is invalid.
-------
COST
First Year
Ctnts Per Cans ?«r
Thousand Gallons !.b, of COD
S Of Patar 'freatad
TABLE 10
COST DATA
Second Year
Cat-.ts Per Cents Per
Thovsmcl Calicos lb. of CCS
S Of V'ifrpr Treated Ueraovad
Total
Cents ?er Cents Per
Thousand Gallons Lh. of COD
S Of tfetar Trs-aMd
UTILITIES
12209
17
8858
21067
12
RE?AIU
7553
11
2699
1C 251
OPJSMIHG
847Q
11
23-i/O
-J
KS
CAJfflOH
10247
15
11532
11
21779
13
5113
2ns
7249
TOTAt
40197
40
83789
*X£SCELLABZCGS lUCL
1, Maiotenanc« overhead
2. Transportation
3. Ktlatenance coats other than Labor
-------
SECTION XI
APPENDICES
Appendix
A. Test Methods Used for Analysis of Water
High Level COD 74
Low Level COD 76
AquaRator COD 77
Suspended Solids and Volatile Solids 80
Total Oil 81
Turbidity 81
Color 81
Odor 81
Miscellaneous Test Results 81
Table 1. Lab Water for Adsorption Plant 82,83
During Second Rains
B. Test Methods Used for Analysis of Carbon 84
Relative Efficiency 84
Apparent Bulk Density 85
Figure 1. Apparent Density Vibrator Feed 86
Figure 2. Conditions - Glass or Metal 87
Figure 3. Metal Vibrator 26 Gauge Galvantized 88
Sheet Metal
C. Determination of COD Materials 89
EPA Internal Report 89
73
-------
APPENDIX A
TEST METHODS USED FOR ANALYSIS OF WATER
HIGH LEVEL COD
The method given below is a brief summary of EPA STORET No. 00340
taken from Methods for Chemical Analysis of Water and Wastes 1971,
Environmental Protection Agency Water Quality Office, Analytical
Quality Control Laboratory, Cincinnati, Ohio. p.17.
HIGH LEVEL WET COD
(50 mg/1 & higher)
APPARATUS
500 ml flat bottom boiling flasks 24/40
Condensers - 300 mm Liebig or equivalent
Hot plates - 9 watts/in2
REAGENTS
0.25 N K2Cr2C7 (Containing 120 mg Sulfamic acid/liter
to eliminate nitrite N interference up
to 6 mg/1 N in a 20 ml sample)
Cone. H2S04 + 22 mg AgS04/9# bottle
0.10 N Fe(NH4)2(S04)2'6H20 (40 gm FAS + 20 ml cone H2S04
diluted to 1 liter)
Ferroin Indicator
PROCEDURE
1. Place approximately 0.4 gm HgS04 and a few glass
74
-------
beads in a 500 ml boiling flask.
2. Add 20 ml of sample of aliquot + H20.
3. Pipette 10 ml 0,25 N K2Cr20y into the boiling
flask and connect it to the water cooled condenser,
4. Slowly pour 30 mis 1^04 + AG2S04 down the top of
the condenser while swirling the boiling flask.
(CAUTION wear safety glasses and rubber gloves)
Reflux for 2 hours (or less if sufficient).
5. Cool and wash condenser down with distilled water
(90 mis).
6. Remove flask from condenser and cool to ambient
temperature. Add 2-3 drops Ferroin indicator.
7. Titrate to Organge-Brown color with 0.10 N FAS
which has been recently standardized.
8. Run blank with 20 mis distilled water and all
reagents.
COD mg/1 -' (A-B)C x 8000
mis sample
Where A = mis 0.10 N FAS for blank
B = mis 0.10 N FAS for sample
C - Normality of FAS
STANDARDIZATION OF FAS
1. Pipette 10 mis K2Cr207 into a 500 ml boiling flask
and dilute to 100 mis.
2. Add 30 mis cone. H2S04 without Ag2S04. Cool and
titrate with FAS.
Normality FAS = 2.5
mis FAS
75
-------
LOW LEVEL COD
The method given below is a brief summary of EPA STORE! No. 00335,
taken from Methods for Chemical Analysis of Water and Wastes 1971
Environmental Protection Agency Water Quality Office, Analytical
Quality Control Laboratory, Cincinnati, Ohio, p. 19.
LOW LEVEL WET COD
(5 mg/1 to 50 mg/1)
APPARATUS
500 ml flat bottom boiling flasks 24/40
Condensers - 300 mm Liebig or equivalent
Hot plates - 9 watts/in2
REAGENTS
0.025 N K2Cr20y
Cone. H2S04 + 23.5 gm AgS04/9# bottle
HgS04
Ferroin Indicator
PROCEDURE
1. Place 1 gm HgS04 + 5.0 ml cone,
H2S04 ancj a few glass beads in a
500 ml boiling flask.
2. Place in ice bath and add 25 ml of
0.025 N K2Cr207 and 70 ml
Cone. H2S04 (with AgS04).
3. Add 50 ml of sample or aliquot + H20.
4, Apply heat to flask and reflux for two hours.
76
-------
5. Cool and wash condenser down with distilled
water (25 ml).
6. Remove flask from condenser and cool to ambient
temperature. Add 8 to 10 drops of Ferroin indicator.
7. Titrate to reddish hue with 0.025 Fe(NH4)2(S04)2
6H20
8. Run blank with 50 ml of distilled water and all
reagents.
COD mg/1 = (A-B) C x 8000
S
Where A = mis 0.025 N Fe(NH4)2(S04)2 for blank
B = mis 0.025 N Fe(NH4)2(S04)2 for sample
C = Normality of Fe(NH4)2(S04')2
S = mis of sample
AQUARATOR COD
The theory of the AquaRator is given below. The operation of the
instrument is not given in this report, but is given in "LIRA
INFRARED Analysed Model 300, Theory - Operation and Service Manual"
Mine Safety Appliances Company, 201 North Braddock Avenue, Pitts-
burg, Pennsylvania.
.Theory of Operation
Quoted from "Lira Infrared Analyser Model 300, Theory -
Operation and Service Manual" Mine Safety Appliance
Company, 201 North Braddock Avenue, Fittsburg, Pennsyl-
vania, pp. 10, 11.
The LIRA is based on the principle of infrared absorp-
tion. It is a common physical fact that all molecules,
with the exception of elemental gases, exhibit character-
istic absorption spectra that is related to the number
77
-------
configuration, and type of atmos in this molecule.
The more simple the molecular structure, the simpler
the absorption spectrum and conversely, heavy com-
plicated molecules exhibit quite complex spectra
By examination of the infrared spectra of the com-
ponents in a process stream, it is normally possible
to locate an infrared absorption band unique to the
"component of interest." The LIRA detector gas,
filter gas, interrupter and window materials are
selected so that the optics are sensitized ("tuned")
only to this unique absorption band of the "component
of interest." The cell length is determined by the
intensity of the absorption band and the calibration
range of the instrument. LIRA optics and sensitiza-
tions are in good agreement with the "Lambert-Beer"
law of light absorption.
In general, the LIRA sources direct two identical
infrared beams through two parallel, gold plated,
polished stainless steel gas cells housed in a solid
aluminum block. One cell contains a known comparison
gas, the other the sample (unknown) gas. After the
radiation beams pass through the gas cells they are
directed into a single detector unit that contains
a sealed-in gas. As the gas in the detector absorbs
infrared radiation, there is a temperature and result-
ing pressure increase of the detector gas. The in-
creased pressure moves a sensitive membrane in the
detector unit. The movement results in a capaci-
tance change in the detector and this capacitance
change is converted to an output signal by the
electronic amplifier.
78
-------
Between the infrared sources and the gas cells, a
half-circle interrupter element or beanj-^chopper
rotates at two cycles per second. As it rotates,
It alternately blocks the infrared radiation beam
from each source, permitting only one beam at a time
to pass through the gas cells and enter the detector
unit. The detector thus alternately responds to the
infrared adsorption of the gases in each cell. As
long as the energy at the detector is equal in both
beams, a properly aligned instrument will read zero.
When the gas to be analyzed (the "component of interest")
is introduced into the sample cell, it absorbs some of
the detectable infrared energy and thus reduced the beam
radiation that reaches the detector unit from the
sample cell. As a result, the two beams become unequal
and the radiation entering the detector flickers as
the beams are alternated. (The brighter comparison
gas beam, then the dimmer sample gas beam, etc.)
The detector gas expands and contracts with this
flicker and directly indicates the energy difference
between the two beams. This variation generates an
electrical signal which is proportional to the infra-
red energy difference between the two beams.
The electronic circuit is tuned so that only variations
between the intensity of the beams entering the detector
unit produce an output signal. When the beams are un-
equal the instrument produces an output signal that can
be measured. The output signal is quickly read (90%
of final reading within 5 seconds) on the meter or tran-
scribed on an auxiliary recorder.
79
-------
Selectivity to the "component of interest" is obtained
by (a) the detector gas, (b) window and interrupter
materials, and (c) filter gases. These sensitizing
components and sample cell lengths vary with the
application. Most LIRA Model 300 Analyzers will
operate on gains 2, 3, or 4, but some special high
sensitivity applications will perform adequately
on gain position 5.
SUSPENDED SOLIDS AND VOLATILE SOLIDS
The method given below is a brief summary of the Suspended Solids
and Volatile Solids test given in Standard Methods for the
Examination of Water and Wastewater. 13th Edition, P 538, Method
224D (1971). The equipment and furnace temperature are altered
slightly from the cited method.
Equipment: Milipore Filter Apparatus
Crucibles
Glass Fiber Filters - Reeve Angel 4.25 cm
Grade 934 AH
Procedure: 1.
Dry crucible and glass filter in 103°C
oven and place in dessicator to cool.
Obtain tare weight (A).
2. Place glass filter on Millipore filter
apparatus and vacuum 20-50 mis of sample
through filter.
3. Rinse any adhering material from sides
of Millipore filter apparatus with dis-
tilled water.
4.
Dry glass filter in a 103°C oven for
2 hours.
5. Cool in a dessicator and weigh (B).
SUSPENDED SOLIDS ppm - (B-A) x 10
6.
mis sample
Place crucible and filter in 600°C
(1112°F) Furnace for 10 minutes.
80
-------
7. Cool in a dessicator and reweigh (C).
CB-C
mis sample
VOLATILE SUSPENDED SOLIDS - CB-C x IQ6)
TOTAL OIL
The total oil was determined by APRA Standards, Part 137:
Petroleum Ether Extraction, 13th Ed., pp 254-257.
TURBIDITY
The turbidity was determined using a Each Turbidimeter using
Jackson turbidity units.
COLOR
The color was measured by ASTM D-1209 using cobalt-platinum color
standard.
ODOR
The odor was measured by APHA, Part 136: Odor.
MISCELLANEOUS TEST RESULTS
The total oil and Suspended Solids test were run on a regular
basis along with the pH. The laboratory data from January 1973
is given in Table 1. This data is typical of that obtained
throughout the project.
81
-------
APPENDIX A
Table 1. LAB DATA FOR ADSORPTION PLANT DURING SECOND RAINS
DATE
1-18-73
1-19-73
1-20-73
1-21-73
1-22-73
1-23-73
1-24-73
1-25-73
1-26-73
1-27-73
1-28-73
1-29-73
STREAM
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
FEED
EFFLUENT
PH
7.2
6.9
7.2
7.1
7.0
6.8
7.0
7.2
7.1
7.1
6.9
6.6
6.8
6.7
7.4
7.3
6.8
6.8
7.1
6.9
6.8
6.8
6.8
6.8
COD
320
95
245
25
265
25
335
20
335
20
345
50
295
40
270
50
260
45
250
70
240
55
220
55
TOTAL
OIL
34
1
27
9
29
13
18
11
23
6
20
17
26
6
29
7
32
16
22
8
19
7
SUSPENDED
OIL
29
12
16
10
24
12
35
17
17
14
27
17
14
9
16
8
16
11
51
15
15
2
14
3
82
-------
Table 1 (cont'd). LAB DATA FOR ADSORPTION PLANT DURING SECOND RAINS
TOTAL SUSPENDED
DATE STREAM PH COD OIL OIL
1-30-73 FEED 7.1 205 32 20
EFFLUENT 7.1 50 16 6
1-31-73 FEED 7.0 175 53 23
EFFLUENT 7.1 40 41 4
83
-------
APPENDIX B
TEST METHOD USED FOR ANALYSIS OF CARBON
RELATIVE EFFICIENCY
The method given below is the relative efficiency test developed
by Atlantic Richfield Company.
RELATIVE EFFICIENCY TEST
FOR ACTIVATED CARBON
1. About five grams of each carbon sample to be
analyzed is oven-dried for three hours at 110°C.
2. Upon cooling to room temperature, each sample
is individually ground to yield a few grams of
+325 mesh powder.
3. One gram (± 0.01 gm) portions of each sample is
mixed with a 50 ml aliquot of standard solution
(COD concentration = A ppm) in 125 ml Erlenmeyer
flasks. The flasks are well stoppered and shaken
for one hour on a laboratory shaker.
4. After shaking, the COD of the aqueous phase is
measured (B ppm).
5. The 100% efficiency reference is provided by treat-
ing a virgin sample of activated carbon according
to steps 1 through 4.
6. The relative efficiency is calculated as
Relative Efficiency, % (A-B)
m A Sample
TOT"
A Reference
84
-------
APPARENT BULK DENSITY
The method given below is the Apparent Bulk Density test used
during the project.
APPARENT DENSITY TEST
EQUIPMENT REQUIRED
1. Vibrator Feeder - See Figures 1 through 3.
2. Cylinder, graduated, capacity 100 ml.
3. Balance having a sensitivity of 0.1 g.
PROCEDURE
1. 100 ml. of the carbon is dried to constant
weight at 150 ± 5°C or taken dry from furnace
discharge chute.
2. Sample is placed into the reservoir funnel so
that the material does not prematurely flow into
the graduated cylinder.
3. yhe sample is added to the cylinder from the
vibrator feeder through the feed funnel.
4. Fill the cylinder at a uniform rate of 0.75 to
1.0 ml. per second, up to the 100 ml mark. The
rate can be adjusted by changing the slope of the
metal vibrator and/or by raising or lowering the
reservoir funnel.
5. Transfer the contents from the cylinder to a
balance pan and weigh to the nearest tenth of a
gram (O.lg).
CALCULATIONS
Calculate the apparent density as follows:
Apparent Density, g/ml «= weight of activated carbon-pn
100
85
-------
RESERVOIR FUNNEL
CLAMPED TO RING STAND
RING STAND
METAL VIBRATOR
DOOR BELL "BUZZER"
(FOR 10 VOLT SERVICE)
60 CYCLE
—
—
^.
/
7
FEED FUNNE
TO RING SI
inn-mi. A5
CYLINDER
/ rUTTPH
(SPST BA1
-rfr /
h1 //
PI //
TRANSFORMER
(PRI. VOLTS
VA
SCC. VOLTS 6-
CYCLES 50-ou
,
'
Appendix B, Figure 1. Apparent Density Vibrator Feeder
86
-------
00
•y i /o" .. .
O I / C. " -•••••—'- —™..™-— «.——
p-15/16 ~
Appendix B, Figure 2. Conditions: - Glass or Metal
-------
00
00
Appendix Bf Figure 3. Metal Vibrator - 26 Gauge Galvanized Sheet Metal
-------
APPENDIX C
DETERMINATION OF COD MATERIALS
On April 17, 1973, the United States Environmental Protection
Agency in a report to Mr. Leon H. Myers, Chief, Petroleum-
Organic Chemicals Wastes Section, Treatment and Control Technology
Branch; from Mr. Billy L. DePrater, Supervisory Research Chemist,
Petroleum-Organic Chemicals Wastes Section, Treatment and Control
Technology Branch, gave the results of numerous tests on various
Watson Carbon Plant streams. These tests included an attempt at
a determination of the types of COD involved in the feed and
effluent. The report is given below.
The numerous gas chromatographs referred to are not given in this
report, and interested readers are referred to the original report.
EPA INTERNAL REPORT
ORGANICS IN ARCO SAMPLES
The drganics were extracted from the Arco samples for
analyses by gas chromatography to obtain a "finger-
print" of the organics present and also to measure the
amount of organics as determined by the area under the
chromatograms.
The procedure for extraction of the water samples was
to measure 500 ml of sample (750 ml in the case of
"Eff New Cell") into a two liter separatory funnel and
add 20 ml of redistilled chloroform. After shaking
about two minutes the phases were allowed to separate,
and the chloroform phase was drained into a small
column of sodium sulfate which had been washed with
89
-------
chloroform. The sodium sulfate column removed water
and broke any emulsion that was present. Two more
extractions were made on the sample using 15 ml of chloro-
form each time, so that the total volume of chloroform
used was 50 ml.
After extracting the water samples, the extracts were
concentrated by evaporating the chloroform with a
gentle stream of air until the solutions were con-
centrated enough to run gas chromatographic analyses.
The two effluent extracts contained a very small amount
of organics, so the chloroform was evaporated until the
volume of each was 0.1 ml. The other extracts were
sufficiently concentrated after evaporation of the
chloroform to a remaining volume of 1 ml.
Samples of spent carbon and regenerated carbon were
extracted in a Soxhlet extraction apparatus. The
spent carbon first was placed in a shallow pan on a
steam bath to evaporate the water. After two hours
the spent carbon appeared to be dry so samples of about
100 ml of each carbon was weighed and put into the
extractors with glass wool at the bottom to prevent
carbon particles from getting into the siphon tubes.
Each flask contained 300 ml of chloroform and steam
was used to heat the flasks and maintain a cycling
rate of four times per hour for 24 hours. The extracts
were then distilled to reduce the volume from 300 down
to 50 ml before analysis by gas chromatography.
Gas chromatographic analyses were run with a 6* x 1/8"
stainless steel, Dexsil 300 column. The temperature
was programmed from 50 to 350°C at eight degrees per
minute, and an electronic integrator was used to
90
-------
measure the total area under the chromatogram for quan-^
titation of the sample. Solvent oil mixtures were used
for calibration by running mixtures containing 0.06,
0.12, and 0.24 mg of oil. The area measurements obtained
were plotted versus the milligrams of sample. A mixture
of n-paraffins was run under identical conditions to
indicate the boiling range of the samples by the chromato-
grams. Data from the area of the chromatogram, volume
of water extracted (or grams of carbon extracted), volume
of concentrated extract, and volume charged to the gas
chromatograph gave the following information:
Eff. New Cell 0.63 mg oil/liter
Carbon Plant Total Eff. 1.73 " " "
Carbon Plant Total Feed 34.8 " " "
Carbon Plant Total Feed 34.3 " " "
Res 505 111 " " "
Feed to Res 505 68.8 " " "
Spent Carbon 19 mg oil/gm.
Regenerated Carbon 0 " " "
Copies of the chromatograms are attached. The total
area between four minutes: and 42 minutes was measured
using the baseline set at the beginning of the run.
Corrections were made for area due to the solvent tail
and drift due to stationary phase bleed that occurs
toward the end of the run.
The following observations are made by comparison of
Eff. New Cell and Carbon Plant Total Eff. chromatograms.
It is obvious that the carbon plant total effluent
contains a greater amount of organics since the chromato-
graph is higher above the baseline throughout most of the
run. The prominent peak at 28 minutes retention time
is very likely a phthalate compound used as a plasti-
cizer in the liner of the screw.cap on the container.
91
-------
The caps, were supposed to have Teflon liners, but
appeared to be only a Teflon coating which apparently
was not impervious to the chloroform solvent. The
carbon was less efficient for removing the organic
compounds represented by peaks at 12 and 12.4 minutes
retention time.
The composition of the organics remaining after carbon
treatment consists of a higher percentage of high boiling
compounds when compared to the organics in the feed.
This is illustrated by plotting corrected area percent-
ages of the total area between four and 42 minutes
retention time on the chromatograms for carbon plant
total effluent and carbon plant total feed. The
plots show that 70 percent of the organics in the feed
have a retention time of less than 22 minutes or boil
below 657°F. The 70 percent point of the effluent is
at about 28.8 minutes retention time or about 789°F.
If a visual comparison of peaks and areas is made, it
must be remembered that the effluent sample-solvent
volume was 0.1 ml while the feed sample-solvent volume
was 1 ml. There is a factor of two difference in
attenuation on the feed and effluent chromatograms
or a total magnification of 20 on the effluent chromato-
gram. Several normal paraffin retention times obtained
from the temperature calibration runs are shown on the
feed and effluent chromatograms, and it is noted that
they correspond closely to some of the sample peaks.
All of the chromatograms may be compared on the basis
of peak retention times and approximate boiling points
and boiling ranges; but as noted before, visual area
comparisons and peak height comparisons must take into
consideration the attenuation and concentration of the
sample.
92
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-660/2-75-020
3, RECIPIENT'S ACCESSIOC+NO.
4. TITLE AND SUBTITLE
Refinery Effluent Water Treatment Plant Using
Activated Carbon
5. REPORT DATE
" 1975
May
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Gary C. Loop
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Atlantic Richfield Company
Carson, California
10. PROGRAM ELEMENT NO.
1BB036
11. CONTRACT/GRANT NO.
12050 GTR
12, SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
U.S. Environmental Protection Agency
Robert S. Kerr Environmental Research Laboratoify*
P.O. Box 1198
Ada, Oklahoma 74820
Demo 1971^1974
. SPONSORING AGENCY CODE
IB. SUPPLEMENTARY NOTES
Prepared in cooperation with the Petroleum/Organic Chemicals Wastes
Section, Robert S. Kerr Environmental Research Laboratory, Ada, Okla,748
16. ABSTRACT
Reduction of Chemical Oxygen Demand (COD) in petroleum refinery effluent
wastewater by adsorption onto activated carbon was demonstrated on a
commercial level during a two-year project at Carson, California. The
plant contained over 750,000 pounds of carbon, regenerated 1,644,000
pounds of carbon,'processed 172 million gallons of water, and removed
408,000 pounds of COD.
The carbon was exhausted at the rate of 9.5 pounds per 1000 gallons of
water processed. At an average feed COD concentration of 250 ppm and an
average effluent COD concentration of 50 ppm, the carbon was loaded to
an average of 0.26 pounds of COD per pound of carbon. Following solutioi
of initial startup problems, the unit was operated at a cost of 40 cents
per 1000 gallons of water treated, or 18 cents per pound of COD removed.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Activated carbon, Adsorption, Organi
loading, Oxygen demand, Pollution
abatement, Settling basins, Waste*
water treatment, Carbon regeneration
Chemical oxygen demand
Intermittent carboi
treatment plant, Ca.'.
gon process, Rainfa]
runoff
8. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)
21. NO. OF PAGES
100
20. SECURITY CLASS (Thispage)
22. PRICE
Form 2220-1 (9-73)
* U. S. GOVERNMENT PRINTING OFFICE: 1975-698.909 /4 REGION 10
------- |