United States
Environmental Protection
Agency
Robert S Kerr Environmental Research
Laboratory
Ada OK 74820
EPA 600 2 79 06G
M.HI li 1979
Research and Development
&EFA
Treatment of
Refinery
Wastewater Using a
Filtration-Activated
Carbon System
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental 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.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-066
March 1979
TREATMENT OF REFINERY WASTEWATER USING A
FILTRATION-ACTIVATED CARBON SYSTEM
by
Bruce A. McCrodden
BP Oil Inc.
Marcus Hook, Pennyslvania 19061
Demonstration Grant
No. 12050GXF
Project Officer
Leon H. Myers
Source Management Branch
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr Environ-
mental Research Laboratory, U.S. Environmental Protection Agency,
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S.
Environmental Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or recommendation
for use.
11
-------
FOREWORD
The Environmental Protection Agency was established to coor-
dinate administration of the major Federal programs designed to
protect the quality of our environment.
An important part of the agency's effort involves the search
for information about environmental problems, management tech-
niques and new technologies through which optimum use of the na-
tion's land and water resources can be assured and the threat
pollution poses to the welfare of the American people can be min-
imized.
EPA's Office of Research and Development conducts this
search through a nationwide network of research facilities.
As one of these facilities, the Robert S. Kerr Environmental
Research Laboratory is responsible for the management of programs
to; (a) investigate the nature, transport, fate and management
of pollutants in groundwater; (b) develop and demonstrate methods
for treating wastewaters with soil and other natural systems;
(c) develop and demonstrate pollution control technologies for
irrigation return flows; (d) develop and demonstrate pollution
control technologies for animal production wastes; (e) develop
and demonstrate technologies to prevent, control or abate pol-
lution from the petroleum refining and petrochemical industries,
and (f) develop and demonstrate technologies to manage pollution
resulting from combinations of industrial wastewaters and indus-
trial/municipal wastewaters.
This report contributes to the knowledge essential if the
EPA is to meet the requirements of environmental laws that it
establish and enforce pollution control standards which are
reasonable, cost effective and provide adequate protection for
the American public.
C.
W.C. Galegar
Director
Robert S. Kerr Environmental
Research Laboratory
111
-------
ABSTRACT
The objective of this project was to demonstrate the appli-
cation for a dual media filtration-activated carbon adsorption
system for total treatment of refinery wastewaters.
BP Oil, Inc.'s Marcus Hook Refinery has operated a waste-
water treatment system consisting of dual media filtration for
removal of oil and suspended solids followed by granular activat-
ed carbon adsorption for,removal of dissolved organic material.
Associated equipment includes backwash holding tanks, sludge
thickners, two-stage centrifugation for oil-water-solids separa-
tion and a multiple hearth furnace for carbon regeneration.
The 2.2 MGD wastewater treatment plant has demonstrated
average removals by the dual media filters of 58, 67, and 22
percent reduction for oil, suspended solids, and first stage
Ultimate oxygen demand, respectively. Average removals by the
activated carbon absorbers have been 70, 27, 32, and 39 percent
reduction for oil, first-stage ultimate oxygen demand, chemical
oxygen demand, and phenol, respectively.
Constructed on a one-quarter acre plot, the capital cost of
the wastewater treatment plant was $ 1,812,000 with an annual
operating cost of $223,980.
This report was submitted in fulfillment of demonstration
grant number 1205GXF by BP Oil, Inc. under the partial sponsor-
ship of the U.S. Environmental Protection Agency. This report
covers a period from October 1, 1973, to December, 1975, and
work was completed as of June 1, 1978.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables vii
Acknowledgments viii
1. Introduction 1,2
, 2. Conclusions 3
3. Recommendations 4
4. Process Description 5-8
5. Pilot Plant Study 9-18
6. Construction and Start-up 19-21
7. Wastewater Treatment System Design . . . 22-38
8. Wastewater Treatment System Performance . 39-52
9. Economic Evaluation 53-58
Appendices
A. Wastewater Concentration Histograms . . . 59-78
B. Metric Conversion Table 79
v
-------
FIGURES
Number
1 Filtration Pilot Plant Schematic 10
2 Adsorption Pilot Plant Schematic 14
3 Adsorption Isotherm for Filtered API Separator
Effluent 15
4 Summary of Time Sequential Activities During Startup
of Filtration/Adsorption System 20
5 Wastewater Treatment Plant Schematic 23
6 Dual Media Filter Cross Section 24
7 Carbon Adsorber Cross Section . . 27
8 Effluent Septum Cross Section 29
VI
-------
TABLES
Number Page
1 Filtration Pilot Plant Data 11
2 Analysis of Pilot Plant Filter Backwash Water 13
3 Carbon Adsorption Pilot Plant Data 16
4 Dual Media Filter Design Data 26
5 Activated Carbon Adsorption Design Data 30
6 Thermal Regeneration Design Data 33
7 Wastewater Analysis Period 1 40
8 Wastewater Analysis Period II 41
9 Wastewater Analysis Period III 42
10 Wastewater Analysis Period IV 43
11 Wastewater Analysis Period V 44
12 Analysis of Filter Backwash Water 45
13 Analysis of Water Removed from Filters Prior to
Backwash 45
14 Analysis of Filter Draindown Water 47
15 Wastewater Treatment Plant Capital Cost 54
16 Dual Media Filters Operating Costs 54
17 Activated Carbon Adsorption Operating Costs 55
18 Activated Carbon Regeneration Operating Costs 56
19 Solids Dewatering Operating Costs 57
20 Wastewater Treatment Plant Annual Operating Costs 58
21 Wastewater Treatment Plant Project Cost Summary 58
Vll
-------
ACKNOWLEDGMENT
The cooperation of the Standard Oil Company's (Ohio)
Engineering and Research and Development groups is gratefully
acknowledged for their participation in the design of the pro-
ject and sampling program.
The aid of Mr. Leon H. Myers, U.S. Environmental Protection
Agency, Robert S. Kerr Environmental Research Laboratory was
particularly valuable during the project and reviewing the
demonstration grant report.
The support of the Calgon Corporation Water Management
Division is acknowledged for their evaluation of the activated
carbon adsorption system and recommended systems improvements
(Calgon report to BP Oil Corporation, Report No. C-850, dated
June 13, 1974).
Vlll
-------
SECTION 1
INTRODUCTION
A 2.2 MGD filtration-carbon adsorption Wastewater Treatment
Plant was placed in operation at the Marcus Hook Refinery of BP
Oil Inc., a subsidiary of the Standard Oil Company (Ohio), in
March, 1973.
The Marcus Hook Refinery is a 150,000 BPD Class B Refinery
located in Southeastern Pennsylvania. During the project period
the refinery was modernized to take full advantage of its design
capacity. Prior to December 1974, the refinery was operated at
105,000 BPD.
In 1969, a compliance schedule of 48 months was established
to meet discharge standards prescribed by the Delaware River
Basin Commission (DRBC). The initial effort toward achieving
compliance was the evaluation of the existing API oil-water-sol-
ids Separator, through which all process wastewater flow is
directed. Monitoring of API Separator influent and effluent
first stage ultimate oxygen demand determined an average 68
percent removal, far below the DRBC's required 89.25 percent for
process wastewater streams.
Accordingly, a project to determine the treatability of the
API Separator effluent and a project to reduce the API Separat-
or's hydraulic loading were undertaken. The latter project had
as its basis an in-plant water use survey which concluded that a
reduction in process wastewater flow to the API Separator could
be accomplished by installation of a brine cooler; replacement
of barometric condensers with surface condensers; segregation of
sanitary wastes from the process wastewater stream; and further
segregation of oily water and once through cooling water streams.
The results of this project are evidenced by a reduction in the
hydraulic loading from 3750 to 1700 GPM.
Treatability of the API Separator effluent was investigated
through the operation of a bench scale activated sludge unit and
an extended aeration pilot plant. With accumulated data as the
basis, a preliminary biological treatment system design was pre-
pared. The proposed full scale design required intermediate
facilities for oil removal, two 369,000 gallon aeration basins,
final clarifiers, an anaerobic digester, and both biological and
oily sludge dewatering facilities.
-------
Review of an estimated biological treatment capital cost of
$2,500,000, an estimated annual operating cost of $220,500,
biological treatment variability, land requirements, and excess
sludge generation led to investigation of a filtration-activated
carbon adsorption treatment system.
With filtration/adsorption pilot plant data as the basis, a
preliminary filtration-carbon adsorption design was prepared.
Comparison of an estimated capital cost of $2,000,000; an esti-
mated annual operation cost of $179,000; and the reduced land
area requirements, with the biological treatment preliminary
design, led to the decision to construct a filtration adsorption
wastewater treatment system.
This report covers the first two and one half years of
operation of this system and is submitted in fulfillment of
Demonstration Grant No. 12050GXF from the Research and Monitor-
ing Division of the U.S. Environmental Protection Agency.
The objectives of this project were to:
1. Assess the feasibility of a filtration/adsorption
treatment system for petroleum refinery process
wastewater.
2. Evaluate performance of the system.
3. Determine capital and operating costs.
4. Determine economic feasibility of carbon regeneration.
5. Assess the treated effluent for reuse possibilities.
*Metric Conversion Table - p. 79
-------
SECTION 2
CONCLUSIONS
1. The filtration/adsorption wastewater treatment system did not
produce an effluent equal to design expectations.
2. The factors contributing to the poor performance of the
treatment system were:
A. The waste load on the filters/adsorbers increased
over that experienced during the pilot plant operation.
B. A 407o decrease in the adsorptive capacity of the
regenerated carbon was observed following 18 months of
operation.
C. A change occurred in the wastewater characteristics
as evidenced by a decrease in the theoretical carbon
capacity.
3. the carbon adsorber design flow rate of 1500 GPM could not be
maintained due to plugging of the effluent septums by carbon
fines.
4. The static bed activated carbon pilot plant did not provide
an adequate basis for design of a full scale pulse bed system,
5. Carbon losses stabilized to 670 per regeneration cycle.
6. The production of sulfide across the carbon columns was not
only the result of bacterial action but was also a function
of influent characteristics.
7. The dual media filters demonstrated consistent removal of
suspended solids and oil during 18 months of operation. The
filters' removal efficiency decreased as media was gradually
lost from the filter vessels.
8. The solids dewatering system created a recycle of solids and
oil to the API Separator influent which in turn impacted on
all down stream units.
-------
SECTION 3
RECOMMENDATIONS
1. All wastewater streams to be treated in the full scale units
should be included in the pilot plant influent.
2. Activated carbon pilot plants should be dynamic systems that
model the full scale unit. Pilot scale regeneration systems
should be operated to obtain scale up parameters and assess
losses in adsorptive capacity.
3. Further investigation as to the production of sulfide during
treatment of refinery wastewater should be undertaken.
4. Further investigation as to the optimum regeneration furnace
temperature profiles should be undertaken.
5. The design of activated carbon systems should provide ade-
quate storage capacity for regenerated and spent carbon to
allow for regeneration furnace shutdowns.
6. The design of activated carbon systems should include facili-
ties for removing fines from the regenerated carbon.
7. Pulsed bed carbon adsorption columns should be designed as
pressure vessels to permit increased flow rates.
-------
SECTION 4
PROCESS DESCRIPTION
FILTRATION
Filtration is the process of "straining" suspended, insol-
uble matter from a liquid stream. Various mechanisms are respon-
sible for the translation of suspended particles from the main
stream of flow to the filter grains in a deep-bed system. There
are also various forces which will retain the particle once it is
brought into contact with the filter medium.
It appears that the random movement of suspended particles
in a flow path is mainly responsible for transporting the parti-
cles either directly to the grains or close enough to the grain
surfaces for other forces to become effective. Removal mecha-
nisms can be discussed as a function of particle diameter for
small particles, with diameters one micron and smaller, random
movement dominates in bringing the particles to, or close to, the
grain surfaces. Van der Waals forces will accomplish capture
once particles are within 0.05 to 0.1 micron range, outside of
which gravity forces will dominate. For particles of very small
diameter, Brownian diffusional forces become increasingly impor-
tant in the last stage of the contact. In most cases, retention
is due to Van der Waals forces, but for fine particles with
positive charges, electrokinetic forces will be responsible.
Chemisorption, ion exchange, and chemical bonding will operate
in the atomic and molecular ranges.
For large particles with diameters above 50 microns, direct
contact predominates, especially for particles to be captured on
multiple contact sites. Sedimentation is of secondary importance,
When a particle is captured on a multiple-contact site, it re-
duces the passageway and causes capture of progressively finer
particles. On the bottom side of the grains, Van der Waals
forces may contribut to the capture mechanism. On the bottom
side of the filter grains, Van der Waals forces are strongly
opposed by the earth's gravitational pull. On the top side of
the grains, these forces are additive and more sediment is expec-
ted to have much electrokinetic potential; electrokinetic forces,
therefore, are not important in retention. Friction forces and
fluid pressure may help retain particles captured in multiple-
contact sites.
-------
For medium size particles with diameters one to 50 microns
most of the mechanisms and forces applicable to the other two
classes will have relevance. The larger particles in this class
may still be subject to capture by direct interception on multi-
ple-contact sites in fine filter media (grain diameter =0.3 mil-
limeter or less). The majority, however, will be captured by a
combination of random movement and sedimentation, with Van der
Waals forces increasingly more important as the particle size
and density decrease.
Friction and fluid pressure play a part in connection with
the larger particles captured on multiple-contact sites. Elec-
trokinetic adhesion forces may operate in the small particle
range in the case of opposing charges, but, in the bulk of this
class, Van der Waals adsorption and gravity forces predominate.
On the upper surfaces these two forces are additive; on the
bottom surfaces, they oppose each other. Since the superiority
of Van der Waals forces over gravity decreases with the inverse
second power of the particle diameter, larger particles will have
a greater preference for the upper surfaces than the small part-
icles. It is also to be expected that the particle adhesion
forces will diminish with each layer of particle deposition,
since these layers are usually not as dense as the filter medium.
This fact again tends to increase the relative importance of
gravity over Van der Waals forces. In other words, it results
in a thicker sediment layer on the top than on the bottom sur-
faces with increasing particle size.
The substances removed during filtration are distributed
irregularly over the grain surfaces and are not dislodged by the
passing fluid under normal operating conditions. Interstices,
however, are narrowed down by accumulating deposits and some may
be completely closed. Particles entering pores still open are
then transported deeper into the bed, until they reach grain
sites still able to accept them. Only when particles fail to
find such sites do they pass into the effluent.
ACTIVATED CARBON ADSORPTION
Activated carbon removes organic contaminants from water by
the process of adsorption or the attraction and accumulation of
one substance on the surface of another. In general, high sur-
face area and pore structure are the prime considerations in
adsorption of organics from water; whereas, the chemical nature
of the carbon surface is of relatively minor significance.
Granular activated carbons typically have surface areas of
500-1,400 square meters per gram. Activated carbon has a prefer-
ence for organic compounds and, because of this selectivity, is
particularly effective in removing organic compounds from aqueous
solution.
-------
Much of the surface area available for adsorption in granu-
lar carbon particles is found in the pores within the granular
carbon particles created during the activation process. The
major contribution to surface area is located in pores of molecu-
lar dimensions. A molecule will not readily penetrate a
pore smaller than a certain critical diameter and will be ex-
cluded from pores smaller than the designated critical diameter.
Activated carbon is manufactured by a process consisting of
raw material dehydration and carbonization followed by activa-
tion. The starting material is dehydrated and carbonized by
slowly heating in the absence of air. The activated carbon used
in this project was made from bituminous coal.
Adsorption by activated carbon involves the accumulation or
concentration of substances at a surface or interface. Adsorp-
tion is a process in which matter is extracted from one phase and
concentrated at the surface of another, and is therefore termed
a surface phenomenon. Adsorption from wastewater onto activated
carbon can occur as a result of two separate properties of the
wastewater-activated carbon system, or some combination of the
two: (1) the low solubility of a particular solute in the
wastewater and (2) a high affinity of a particular solute in the
wastewater for the activated carbon. According to the most
generally accepted concepts of adsorption, this latter surface
phenomenon may be predominantly one of electrical attraction of
the solute to the carbon, of Van der Waals attraction, or of a
chemical nature.
There are essentially three consecutive steps in the ad-
sorption of dissolved materials in wastewater by granular acti-
vated carbon. The first step is the transport of the solute
through a surface film to the exterior of the carbon. The
second step is the diffusion of the solute within the pores of
the activated carbon. The third and final step is adsorption of
the solute on the interior surfaces bounding the pore and
capillary spaces of the activated carbon.
There are several factors which can influence adsorption by
activated carbon, including: (1) the nature of the carbon
itself; (2) the,nature of the material to be adsorbed, including
its molecular size and polarity; (3) the nature of the solution,
including its pH; and (4) the contacting system and its mode of
operation.
THERMAL REGENERATION
Thermal regeneration of granular carbon consists of three
steps: (1) drying; (2) pyrolysis of adsorbates; and (3) activa-
ting by oxidation of the carbon residues from decomposed adsor-
bates. Drying is accomplished at 212ฐF, baking between 212 and
1500ฐF, and activating at carbon temperatures above 1500 F. All
-------
of these steps are carried out in a direct-fired hearth furnace.
Time, temperature and atmosphere are the controllable variables
for regeneration. Free oxygen must be controlled by the Addition
of steam in the lower hearths of the furnace to avoid burning of
the granular carbon itself.
CENTRIFUGATION
inr;irr-r--|i-|iปT~--- I TIT I __. , ^ j
Centrifugation may be defined as sedimentation under the
influence of forces greater than gravity. A centrifuge can
clarify, classify,.or separate components of a given stream as a
function of the difference in the component's specific gravities.
The disc and scroll type centrifuges are discussed in this report.
The fundamental difference in these types is the method by which
components are collected in, and discharged from the bowl. The
method of discharge determines the size and nature of the parti-
cles which are suitably collected and handled in each of the
centrifuges.
Within a centrifuge, centrifugal force acts on a suspended
particle, causing it to settle through the liquid component. By
rotating the settling vessel at high speed, the settling forces
acting on a particle can be increased by several orders of
magnitude.
The degree of removal within the centrifuge is both a func-
tion of the average retention time and the effective centrifugal
force acting on the component to bfe separated. Separation of the
components is effected when the settling velocity imparted to
them by the centrifugal force exceeds the overflow velocity of
the suspending liquid. The rate of separation is dependent on
the differential density between the individual components.
Liquid viscosity relationships apply to centrifugal separa-
tion. As the temperature of the liquid increases, the viscosity
decreases, and thus the rate of subsidence of the particle to be
separated increases.
Particle distribution or concentration is interrelated with
particle size and shape, and more generally affects the concen-
tration of the scroll centrifuge cake. The nature and compres-
sibility of the solids also affect the concentration of the cake.
-------
SECTION 5
PILOT PLANT STUDY
DUAL MEDIA FILTRATION
A pilot plant filtration/adsorption treatment of API Separa-
tor effluent was investigated over a period of six weeks in
August and September, 1970. The equipment used for this pilot
plant study was a dual media filter with a cross-sectional area
of one square foot.. The unit was square, 17 feet in overall
height and had one'face of transparent plexiglass. The filtering
media consisted of 5 ft. of sand and 2.5 ft. of anthrafilt sup-
ported by 16 inches of gravel. A schematic of''the filtration
pilot plant is shown in Figure 1.
API Separator effluent was fed to the filtration pilot plant.
The filter effluent was discharged to a 55-gallon drum for use as
feed to the activated carbon columns. The filter was backwashed
with influent water when the pressure drop reached 13.5 psi. The
backwash procedure consisted of an initial air scour followed by
a water rinse.
The influent to and effluent from the sand filter were com-
posited during each run and were analyzed for oil, phenol,
suspended solids, 600$, and TOC. A total of 37 runs were con-
ducted using flow rates of 12 to 18 gal./min./sq. ft. with most
runs at the lower end of the range.
The first five runs were conducted using only 5.5 ft. of
sand. The results indicated that performance was good but that
the filter runs were quite short, approximately 5 hours. There-
fore, to improve filter runs, 2.5 feet of anthrafilt were added.
Filter runs were conducted at 12 gpm./sq. ft. filtration rates,
16 gpm/sq. ft. and 18 gpm/sq. ft. In the sand anthrafilt runs,
performance of the filter did not change significantly when the
filtration rate was varied from 12 to 18 gpm/sq. ft.
Use of coagulants at dosages of 1 to 5 mg/L did not improve
the performance of the sand filter.
The performance data for the filtration of API Separator
effluent is presented in Table 1. Average removals of suspended
solids, BOD5, TOC, and oil were 77, 42, 48, and 79 percent
-------
Backwash Out
Filter In
1
T ฎ~
24"
4- ฎ-
12"
+ <ง>-
12"
12" r
12"
12"
12"
21"
__[_
r-r
^
^ ^
't&'$t,
lf ~ **
t
n
**"
^Vrji
nnnf
101
|
2'-6'
i
v^
5'-6
ซ
Anthrofilt
M
1 Sand
l'- 4 "Grovel
, _L
Backwash In
| - Filter Out
Figure 1. UHR test filter Functional diagram
10
-------
TABLE 1
Filtration Pilot' Plant Data
Suspended Solids BOPr TOO Oil Phenol
IHR/1
Date Run No.
8/13/70
6/13/70
8/14/70
8/17/70
8/18/70
8/19/70
8/20/70
8/21/70
8/23/70
8/24/70
8/25/70
8/26/70
8/29/70
8/31/70
9/01/70
9/02/70
9/03/70
Average
7
'8
9
14
16
18
20
22
26
28
29
30
31
36
37
38
39
Inf.
20
70
90
35
26
20
10
40
55
40
25
15
25
26
25
28
20
35
tff.
21
5
14
7
3
4
5
5
20
15
5
10
5
6
5
12
6
8
Percent
Removal
93
84
80
89
80
50
88
64
63
80
33
80
77
80
57
70
77
BJt/1 '
Inf.
54
87
73
84
30
66
45
83
68
45
50
48
53
60
55
60
Eff.
35
24
29
54
20
39
33
49
29
34
35
30
35
53
27
35
Percent
Removal
35
72
60
36
33
41
27
41
57
25
30
38
34
12
50
42
rag/1
Inf.
116
137
68
137
81
79
77
39
77
46
56
45
35
50
70
74
Eff.
45
45
38
50
61
48
41
33
39
41
26
34
25
36
35
--
40
Percent
Removal
61
67
44
64
25
40
47
2
49
11
53
20
29
28
50
48
rag/1 Percent
Inf.
178
36
48
35
51
58
44
56
50
44
51
50
60
71
60
Eff. Ejnoval
17
7.5
13.6
9.5
11
10
__
11
6.9
17.2
11
8.2
11
14.2
11.2
14.2
90
79
72
73
80
83
75
88
66
75
85
78
77
85
79
mR/1 Percent
Inf.
0.19
0.80
11.6
0.83
4.7
__
7.5
2.0
4.4
4.1
3.9
3.6
5.7
0.8
3.85
Eff. Re
0.35
0.76
9.8
0.79
3.7
6.8
2.3
4.8
3.0.
4.2
3.5
5.3
0.7
3.54
novaj
5
10
4
21
10
27
7
12
8
-------
respectively.
The filter runs varied from 12 to more than 24 hours, de-
pending on the incoming TOC concentration, i.e., the higher the
TOC concentration, the shorter the filter run.
The release of contaminants from the filter occurred during
backwashing at 30 gal./min./sq. ft. An analysis of the oil and
suspended solids content of the pilot plant filter backwash is
presented in Table 2. The results indicate that a 6-minute
backwashing was sufficient to clean the filter.
The average pilot plant influent and effluent values are
summarized below:
Filtration Pilot Plant Average Performance
Oil Suspended Solids BODc
mg/1 Percent mg/1Percent mg/1 Percent
Inf Eff Removal Inf Eff Removal Inf Eff Removal
60 11 79 35 8 77 60 35 42
ACTIVATED CARBON ADSORPTION
The carbon columns used in the study were 5 inch diameter
plexiglass columns. A schematic of the adsorption pilot plant is
shown in Figure 2. Four columns were used in series; the first
column was filled with 3 feet of carbon; the remaining columns
contained 5 feet, giving a total carbon depth of 18 feet. The
carbon used during the study was 8 x 30 mesh Filtrasorb 300. The
sand filter effluent was fed to the carbon column at a rate of
0.5 gal./min. giving a rate of 3.6 gal./min./sq. ft. and a con-
tact time of 36 minutes. A total of approximately 11,000 gallons
of wastewater was passed through the carbon. The effluent from
the carbon columns was composited to correspond to one complete
sand filter run and analyzed for oil, phenol, suspended solids,
BOD^ and TOC. In addition, grab samples of effluent from carbon
columns one and four were periodically analyzed for TOC.
The adsorption isotherm for filtered API Separator effluent
at ambient temperature is shown in Figure 3. The results indi-
cate that an effluent TOC of 3 mg/1 could be obtained by carbon
adsorption. The intersection of isotherm at the initial TOC
concentration of 36 mg/1 gives the theoretical capacity of the
carbon when it is in equilibrium with the influent concentration.
For this particular wastewater, the theoretical capacity was 0.3
Ibs. of TOC adsorbed per pound of carbon. This is equivalent to
1 pound of carbon exhausted per 1,000 gallons of wastewater treat-
ed.
Approximately 11,000 gallons of wastewater were treated
12
-------
TABLE 2
Summary of Analyses of Sand Filter Backwash Water
Time After
Initiating
Backwash, mg/1
Run No. Mins. Oil Suspended Solids
17 0.5 14,500 7,550
2.0 2,175 640
3.0 830 596
4.0 430 524
6.0 175 102
28 0.5 23,000 12,050
1.5 8,500 4,550
2.5 1,155 1,140
4.0 412 220
7.0 72 40
Backwash procedure consists of air and water scrub for 5
minutes followed by water rinse at 30/gal/min/ft.
13
-------
Filtrasorb 400 Columns
Raw
Feed
Tank
ri'ซ
ซ*
Sand
Filter
Pump
Carbon
Feed
Tank
I
I
1
I
^^^
>
_T
Pump
8-35 Mesh Granular
Darco Columns
Product
Water
Tank
Product
Water
Tank
Figure 2. Laboratory adsorption system
-------
X
CO
(0
o
10 20 50 100
TOG Remaining MG/L
Figure 3. Carbon adsorption isotherm of filtered waste water
15
-------
TABLE 3
Summary of Performance of Carbon Column 4
Suspended Solids
Pate
8/13/70
8/17/70
8/18/70
8/19/70
8/2O/70
8/21/70
8/23/70
8/24/70
8/25/70
8/26/70
8/27/70
8/31/70
9/01/70
9/O2/70
9/OJ/70
Run
No.
8
14
16
18
20
22
26
28
29
3O
31
36
37
38
39
ปK.
Int.
5
7
3
4
5
5
20
15
5
IO
5
6
5
12
6
M
Err.
i
3
t
2
1
5
10
15
I
1
1
2
2
2
1
Percent
Kcmuva 1
SO
57
67
50
8O
O
50
O
80
90
80
67
60
83
84
BOD..
ra
Inf
29
54
2O
39
33
49
29
34
35
30
35
53
27
A
.Eff.
4
6
4
4
6
5
4
3
4
7
7
10
5
Percent
Renova 1
86
89
80
90
81
90
86
91
89
77
80
81
82
KSUOD
ret/ 1
Tnr. Err.
50 5
46
84
46
71
50
42
6
15
17
8
5
8
Percent
Removal
90
87
81
63
89
90
81
TOC
Oil
Phenol
"R/l
Inf.
61
41
33
41
26
34
36
35
I
Ert.
12
17
14
13
10
13
1 1
16
10
Percent
Removal
80
59
sa
68
62
62
SJ
56
71
ป
~
Inr.
17
7.5
9.5
11
17.2
11.0
14.2
11.2
Etr. ;
2.3
1.3
0.9
2
__
0.8
2.5
2.1
2.3
Percent
Removal
87
83
91
82
_
96
77
85
80
: mft
. InfT
0.35
0.76
O.86
0.79
6.8
4.8
4.2
ซ.
5.3
0.7
/I 1
EFF. I
O
0
O.I
0
0
.ซ.
0
O.O2
_
o.oa
O.Ol
'ercent
leooval
10O
100
89
10O
100
ซ.ซ.
100
100
.
99
100
Average
62
36
85
57
83
37 13
65
12.3 1.8
2.7 0.02
99
-------
through the carbon columns. These results indicated that the
carbon column was not exhausted during the study, although it was
approaching exhaustion. Projection of these data to breakthrough
indicates that treatment of 14,000 gallons would exhaust carbon
column one. This projected throughput corresponds to an exhaus-
tion rate of 0.86 pounds of carbon/1,000 gallons of throughput.
The exhaustion rate obtained through column testing was quite
close to that obtained from the adsorption isotherm discussed
above.
The results of the pilot plant indicated that carbon column
four produces an effluent with a TOC concentration of approximate-
ly 10 mg/1 for influent TOC concentrations varying from 25-61
mg/1. The performance of carbon column four in reducing organics,
oil, and phenol is shown in Table 3. The average removal with
respect to suspended solids, BODc, TOC, oil, and phenol were 62,
83, 65, 85, and 99 percent, respectively.
The pilot plant carbon columns were backwashed to maintain
an acceptable pressure drop across the columns. However, samples
for wastewater analysis were not taken during these backwashes.
Samples of exhausted carbon from the lead column were oven
dried for three hours at 150 C prior to bench regeneration test-
ing. The analytical results are presented below.
Iodine Numbers
Virgin Carbon Exhausted Carbon Regenerated
Carbon
Top to Middle 900, min. 453 950
Middle to Bottom 900, min. 518 973
Based on these results, it was projected that spent carbon could
be regenerated to its original adsorptive qualities based on
iodine numbers using standard additions of steam and air/gas at
1750ฐF.
The average pilot plant influent and effluent values are
summarized below.
Activated Carbon Adsorption Pilot Plant Average Performance
Oil Suspended Solids BOD5
mg/1 Percent mg/1Percent mg/1 Percent
Inf Eff Removal Inf Eff Removal Inf Eff Removal
12.3 1.8 85 8 3 62 57 9 83
17
-------
Activated Carbon Adsorption Pilot Plant Average Performance
(cont'd)
TOC Phenol
mg/1 Percent mg/1 Percent
Inf Eff Removal Inf Eff Removal
37 13 65 2.7 0-02 99
SOLIDS DEWATERING STUDIES
Thickening of filter backwash and of a composite of filter
backwash, API Separator bottoms, and emulsion treater bottoms
were investigated. Laboratory tests indicated that the filter
backwash could be thickened to 2 percent solids at a solids
loading of 23 Ib/sq. ft./day, and that the composite sludge
could be thickened to a solids concentration of 1.7 to 3.0 per-
cent at solids loadings of 23 to 62 Ib/sq. ft./day.
Bench scale centrifugation of thickened sludge was investi-
gated. The bench scale evaluation of centrifuge operation con-
sisted of heating the sludge to 180-200ฐF. and centrifuging in a
solid bowl centrifuge. The effluent from the solid bowl centri-
fuge was treated in a disc centrifuge to separate oil from the
water layer.
18
-------
SECTION 6
CONSTRUCTION AND START-UP
The following schedule was maintained for this project:
Engineering Complete March 01, 1972
Bids Received April 10, 1972
Contract Issued May 01, 1972
Start of Construction May 04, 1972
Completion of Construction February 25, 1973
Start-up Commenced February 26, 1973
Start-up of the filters, adsorbers, and solids dewatering
facilities followed a predetermined sequence. A summary of the
start-up period is reported below. This summary of time sequen-
tial activities during start-up is shown in Figure 4.
Start-up of the dual media filters commenced on day one with
a total flow of 400 gpm. Problems were encountered with the
automatic butterfly valves, and the filters were shut down on
day two for repairs. The filters were placed on line again on
day three. Control adjustments were made on days four and five,
and the flow was increased to 900 gpm on day six.
The adsorbers and the carbon storage tanks were topped out
with carbon, using the carbon blow pot on days 23-26. At this
time some irregularities were observed on the skirts of the
adsorber vessels. During the next 14 days, 4x4 angles were
welded to the skirts for additional structural support. Waste-
water was reintroduced to the adsorber on day 48 with a total
flow of 600 gpm. The influent rate was increased to 900 gpm on
day 50.
On day 53, the flow to the filters and the adsorbers was
increased to 1500 gpm and the units were switched over to auto-
matic level control. As the level in the surge basin was lower-
ed, the suspended solids which had settled out in the surge basin
were scoured to the filter. The level in the surge basin had to
be raised to reduce the suspended solids loading to the filters.
The level in the surge basin was then slowly lowered at the rate
of one inch per day.
The carbon regeneration furnace was initially started on
19
-------
M NO P Q
FFFFCHHHHHHHHHHHHHHHHH I J
0 10 26 3D"" 4O'1 50 BO ro"' ซw ซ
I. Dual Media Filter Startup
A. Filter startup commenced with flow of 400 gpm.
B. Filters shutdown for repair of automatic butterfly valves.
C. Filters placed on line again.
D. Filter control adjustments made.
E. Filter flow increased to 900 gpm.
K. Influent increased to design rate of 1500 gpm and units switched to automatic level control.
II. Carbon Adsorber Startup
F. Adsorbers and Carbon Storage Tanks topped out with carbon.
G. Irregularities observed in skirt of Adsorber vessels.
H. Additional structural support added to Adsorbed skirts.
I. Influent reintroduced to Adsorbers at 600 gpm.
t^> J. Influent rate increased to 900 gpm.
0 K. Influent increased to design rate of 1500 gpm and units switched to automatic level control.
III. Carbon Regeneration Furnace Startup
L. Carbon Regeneration at design rate of 120 Pounds/Hour.
IV. Solids Dewatering System Startup^
M. Filter backwash water introduced into Dewatering System.
N. Solid bowl centrifuge started.
0. Disc centrifuge started.
P. API Separator sludge introduced into Dewatering System.
Q. API Separator sludge discontinued due to plugging problems.
R. Comminuter installed and sludge transfer from API Separator resumed.
Figure 4. Summary of time sequential activities
during startup of Filtration/Adsorption System.
-------
day 26. Continuous regeneration of carbon at the design rate of
120#/hr. began on day 58. The solids dewatering system was
started up on day 24 with the backwash water from the filters.
The solid bowl centrifuge was started on day 29 and the disc
centrifuge on day 30- On day 37 API Separator sludge was pumped
to the sludge blending tank at the head of the solids dewatering
system where it was passed through a one-quarter inch screen
before entering the tank. This screen became clogged after one
minute. Also, small sticks and other material which passed
through the screen or were in the filter backwash water became
clogged in the one-quarter inch pump impellers down stream of
the sludge blending tank. It became necessary to discontinue the
pumping of the API Separator sludge and the emulsion treater
bottoms to the solids dewatering system, due to plugging of the
screen and pump impellers.
A comminuter was installed on the solids collection tank
outlet by day 85 and sludge transfer from the Separator was
resumed.
21
-------
SECTION 7
WASTEWATER TREATMENT SYSTEM DESIGN
A schematic flow diagram of the filtration/adsorption/regen-
eration/centrifugation wastewater treatment system is presented
in Figure 5.
A discussion of the design of each individual system follows;
DUAL MEDIA FILTRATION DESIGN
Three parallel dual media filters were designed to remove
oil and suspended solids from the API Separator effluent. Design
removals were those achieved during pilot operation. An inter-
mediate basin was included in the design to control flow surges
and equalize influent overloads. Figure 6 is a cross section
view of one of the dual media filters.
Each filter is a carbon steel vessel 10 feet in diameter by
eighteen feet six inches overall height and is epoxy lined. Flow
enters the bottom of the vessel and rises vertically through a
10 inch pipe in the center of the filter. A rated flow of 1000
gpm per filter corresponds to a superficial hydraulic loading of
12.8 gpm/ft. .
Flow to the filter system is controlled by a level control-
ler which maintains a constant level in the intermediate surge
basin. The effluent flow from each filter is sensed by indivi-
dual flow indicators. The flows are summed and equally divided
among the three filters by throttling each filter's effluent
control valve.
Under normal filtering conditions the vessel is full of
water to the vent connection on top. The water flow is down
through the filter media of 2.5 feet of anthracite and 4.5 feet
of sand; through the support gravel; and through the nozzles
which are inserted in the supporting tube sheet. The water be-
neath the tube sheet flows out through the outlet connection to
a 30,000 gallon filtered water holding tank.
Removal of suspended solids and oil trapped by the filters
is accomplished by backwashing with water stored in the filtered
water holding tank.
22
-------
API SEPARATOR AIR - ^AL MEDIA FILTERS
SLUDGE
BLENDING
TANK
INFLUENT
~n i h
SLUDGE BASIN
i
Lg
*
JL
t
I
i
i
-ป
A
t
ซi
J
i
i
i
i
BACKWASH'""" API
WATER ,t f|~~SLUDGE
FILTERED FILTERED
EFFLUENT WATER
fit .TANK
-ป* '
EFFLUENT-4-
INFLUENT TO FILTORS " "BACKWi
INFLUENT TO ADSORBERS PUM
ACTIVATED CARBON ADSORBER
UNBROKEN EMULSION
r-i CARBON
COLLECTION
I TANK
I THICKENER
p-^VDEWATERING SCREW SCROLL CENTRIFUGJE
HOLDING
TANK
,JU
REGENERATION
FURNACE SLUDGE
DISPOSAL
SUMP
CENTRATE
-STEAM OIL TO RECOVERY
.MAKEUP
CARBON WATฃR TO API
SEPARATOR
WATER
WATER
DISC
CENTRIFUGE
Figure 5. Wastewater Treatment Plant
schematic flow diagram.
23
-------
Figure 6. Dual media filter cross section
24
-------
The initial step in backwash is to remove the water remain-
ing in the filter by applying air pressure to the top of the
vessel. This allows an up-flow air and water scour to follow
and effectively remove adhering suspended solids and oil from the
sand and anthracite particles. Scour rates are 7.1 gpm/ft.2 and
7.1 SCFM/ft. . As the scour water reaches the top of the vessel.
the air is shut off and the water rate increases to 25.1 gpm/ft.2,
thereby flushing the filter of trapped suspended solids and oil.
The backwash water overflows into the center standpipe and is
directed to a 30,000 gallon sludge blending tank.
The backwash cycle is automatically operated by a Programmed
Timer which can be initiated by an interval timer, high differen-
tial pressure, or manually by pushbutton. The three filter sys-
tem is designed to allow only one filter to backwash at any one
time. The filters will automatically backwash in numerical se-
quence. Although the filters were designed to operate with one
off-line, the mode of operation is to have an individual filter
off-line only during its backwash cycle. If the level in the
filtered water holding tank is low or the level in the sludge
blending tank is high, the backwash cycle cannot proceed and an
alarm is sounded.
Table 4 summarizes the dual media filter design data.
ACTIVATED CARBON ADSORPTION DESIGN >
Three parallel activated carbon adsorbers were designed to
remove soluble organic matter from the filter effluent at a
maximum flow rate of 2000 gpm. Design removals were those ob-
tained during pilot)operation. Figure 7 is a cross section view
of one of the carbon adsorbers.
Each adsorber is a carbon steel vessel 10 feet in diameter
by 65 feet overall height and is lined with 12-15 mils of Plas-
tite. The adsorbers each contain 92,000 pounds of granular ac-
tivated carbon in a bed depth of 45 feet. An additional 8000
pounds of carbon occupies the upper and lower end cone areas.
The upper and lower cone angles are 90 and 46 degrees respective-
ly, based on the angle of repose of granular activated carbon
immersed in water.
Flow to the three adsorbers is controlled by the level in
the filtered water holding tank, which acts as a feed surge basin.
The influent to each adsorber is distributed through a circum-
ferential manifold located just above the lower cone section.
The flow is directed downward under an internal cone, then upward
through a 3-foot diameter opening in the internal cone. A design
flow to each adsorber of 667 gpm corresponds to an empty bed con-
tact time of 40 minutes.
25
-------
TABLE 4
Dual Media Filter Design Data
Filter Media
Anthracite
Depth
Volume
Particle Diameter
Sand
Depth
Volume
Particle Diameter
Gravel Support
Depth
Rated Flow (Each of Three Filters)
Filter Diameter
Center Standpipe
Filter Area
Hydraulic Loading
Liquid Capacity With Media Installed
Maximum Allowable Pressure Drop
Thru Media
Design Pressure
Backwash Interval
2.5 FT
195 FT3
0.25 IN
4.5 FT,
350 FT-
1 mm
1.25 FT
1000 GPM
10 FT
10 IN
78 FT2
12.8 GPM/FT"
4800 GAL
6.5 PSI
47.5 PSI
12 HOURS
Backwash Water Flow As Percent of Filtrate 1.3%
Low Rate
High Rate
Backwash Air Flow
550 GPM 7.1 GPM/FT*
1960 GPM 25.1 GPM/FT2
550 SCFM 7.1 SCFM/FTZ
26
-------
5'-7/8"RE
Figure 7., Carbon Adsorber cross section.
27
-------
The upward flow through the packed bed at a superficial
hydraulic loading of 8.5 GPM/ft.2, is discharged through eight
internal septums which extend vertically from the upper cone.
The septums are stainless steel well screens which retain the
1.5 mm diameter activated carbon particles in the adsorber.
Filtered service water is provided at each septum for backflush-
ing, should plugging due to carbon fines occur. Figure 8 is a
cross section view of one of the effluent septums.
Continuous adsorption is dependent upon the removal of ex-
hausted carbon from the adsorbers and the addition of regenerated
carbon. One thousand pounds per day of spent carbon is pulsed
from each of the three adsorbers. This equates to 1.1 percent of
the total bed of an individual adsorber. During the pulse period,
which occurs for each vessel every 24 hours, the adsorber is tak-
en out-of-service. The hydrostatic pressure available at the
lower cone apex is used to transport the carbon slurry to a
flooded collection tank. A pulse period of 1.4 seconds allows
the desired 1000 pounds of carbon to be transferred under velo-
cities of 5 feet per second. Transfer lines are 4-inch schedule
40 carbon steel with schedule 80 long radius sweeps. Ball valves
are used in carbon slurry service. During this pulse period,
regenerated carbon is added to the top of the adsorber from a
carbon storage tank located above each vessel.
As the ball valve at the adsorber apex closes to stop spent
carbon flow, filtered service water is introduced to flush the
line, thereby preventing carbon bridging and corrosion. Freezing
problems are avoided by draining the transfer line following com-
pletion of the water flush.
A cone bottom carbon collection tank receives the spent
carbon and acts as the regeneration furnace feed tank. A ball
valve at the apex of the collection tank pulses carbon for 8
seconds into a dewatering screw at two minute intervals. Filter-
ed water is added at the apex to prevent carbon bridging, and is
added to the dewatering screw to further wash the carbon of free
oil which was "filtered out" in the adsorber. Overflows from the
collection tank and dewatering screw are directed to a carbon
settler from which the carbon is ejected into the dewatering
screw, and the water overflows to be reprocessed.
Table 5 summarizes the activated carbon adsorption design
data.
THERMAL REGENERATION DESIGN
A five foot diameter multiple hearth furnace was designed to
thermally regenerate the spent carbon. The dewatered carbon
enters the six hearth furnace through an 8-inch inlet for regen-
eration at a design rate of 125 pounds per hour. The regenera-
28
-------
FLANGED W.S.
PIPE WELD TO
TANK
CARBON
COLUMN
TANK
PIPE FLANGE ON
PIPE SECTION
PIPE SECTION INSIDE
SCREEN 4" 0 S.S. WITH
0 HOLES
REMOVABLE WELL
SCREEN, 304 SS.
W 0.015" OPENINGS.
CLOSED BAIL
BOTTOM, 304 SS.
Figure 8. Carbon Adsorber effluent
septum cross section.
29
-------
TABLE 5
Activated Carbon Adsorption Design Data
Rated Flow (Each of Three Adsorbers) 667 GPM
Adsorber Diameter 10 FT
Adsorber Bed Depth 45 FT
Contact Time (Empty Bed) 40 MIN 9
Hydraulic Loading 8.5 GPM/FT^
Design Inlet Pressure 60 PSI
Pressure Drop Thru Carbon 35 PSI
Carbon Inventory
Carbon Bed 92,000 LB
Adsorber Total 100,000 LB
Theoretical Carbon Capacity 0.3 LB TOC/LB
Carbon
Carbon Dosage 0.86 LB Carbon/1000 GAL
Throughput
Activated Carbon Properties
Filtrasorb 300
Total Surface Area 950-1050 M2/g
(N2 BET Method) ~
Bulk Density 26 LB/FT
Particle Density Wetted in Water 1.3-1.4 g/cc
Mean Particle Diameter 1.5-1.7 rnm
Iodine Number, minimum 950
Ash Max 8%
Moisture Max 2%
30
-------
tion furnace is capable of handling up to 250 pounds per hour.
The carbon is moved downward through the fire brick lined hearths
by cast iron rabble arms. In the first hearth, which is unfired
but maintains a temperature of 1100ฐF, any remaining moisture is
vaporized. Hearths four and six, numbered from the top, are
tangentially fired by two burners using refinery fuel gas at
rates of 188 and 68 CFH respectively, to maintain respective
temperatures of 1725 F and 1750ฐF.
In an atmosphere controlled by addition of steam at a design
rate of 125 pounds per hour, the adsorbed organics are volatiliz-
ed and oxidized. To assure complete oxidation, all flue gases
pass through an integral afterburner fired by refinery fuel gas
and maintained at a temperature of 1350ฐF. Recirculation of shaft
cooling air provides sufficient oxygen for combustion. Prior
to emission to the atmosphere, the flue gases pass through a two-
foot diameter, four plate, wet scrubber using filtered service
water for gas cooling and particulate removal to 0.04 grains per
standard cubic foot (dry).
Temperature indicator controllers maintain the desired temp-
erature in the fired hearths. Furnace safety features include
ultra-violet flame scanners and alarms which annunciate should
the combustion air blower, induced draft fan, or the shaft cool-
ing air fan fail. Abnormally high or low fuel gas pressure will
cause 'the main gas safety valve to close, resulting in a flame-
out at all burners.
Regenerated carbon is discharged from the furnace into a 12
cubic foot cone bottom quench tank flooded with filtered service
water. Temperature reduction, the addition of make-up carbon,
and the formation of a carbon slurry occur in the quench tank.
As the carbon level in the quench tank increases, a rotating
bindicator is stopped and a timed sequence is initiated to trans-
fer the regenerated carbon to one of three 96 cubic foot carbon
storage tanks located above each adsorber.
During the time controlled sequence, the carbon slurry flows
by gravity into a 5 cubic foot blow case. Filtered service
water is then introduced into the blow case to pressure the car-
bon at velocities of 5 feet per second through 2-inch transfer
lines of schedule 40 carbon steel with schedule 80 long radius
sweeps. The slurry transfer is followed by a water flush and an
air drain to clear the line. In the event a high level is indi-
cated by a storage tank bindicator, the carbon is automatically
transferred to the next storage tank.
Carbon addition to the adsorbers from the storage tanks,
which occurs during the pulsing of spent carbon from the bottom
cone, is judged complete by a bindicator located in the upper
cone. Should the bindicator indicate a low level, the adsorber
may not be brought back into service.
31
-------
An additional safety feature is an atmospheric vent from the
top of the adsorber to its carbon collection tank. In the event
a number of septums plug simultaneously, excess flow will be
vented, and overflow the collection tank to the carbon settler.
A pressure gage is located on the vent line to indicate such an
occurrence.
i
Table 6 summarizes the regeneration furnace design date.
SOLIDS DEWATERING SYSTEM DESIGN
A solids handling system was designed to separate the sludge
removed at the Filters, API Separator, and Emulsion Treater into
an oil, water, and solid phase* Upon separation, the oil is
recovered, the water is returned for reprocessing,,and the solids
are disposed at an offsite licensed sanitary landfill.
The three intermitent sludge streams noted above are mixed
in a 30,000 gallon sludge blending tank and transferred to a
26,000 gallon circular thickener at a rate of 60 gpm. The design
loading of 30 pounds per square foot per day results in an under-
flow concentration of 2.5 percent solids. The thickener under-
flow of 20 gallons per minute and overflow of 40 gallons per min-
ute are directed to the sludge holding tank and API Separator
respectively. Should an emulsion layer .accumulate on the thick-
ener, it is skimmed directly to the sludge holding tank. The
sludge holding tank acts as a feed surge basin for the scroll
centrifuge.
Feed to the scroll centrifuge, flowing at 20 gpm, passes
through a double pipe heat exchanger which maintains an outlet
temperature of 150ฐF. Operating at 2600 RPM, the scroll centri-
fuge discharges a stream of 50 percent solids, and an oil-water
stream. The solids are carried by conveyor belt to a holding
container to await disposal. The liquid centrate is directed
to the disc centrifuge feed sump.
The disc centrifuge feed, at 20 gpm. passes through a double
pipe heat exchanger which maintains an outlet temperature of
180ฐF. An additional 25 gpm of filtered service water also enters
the disc machine at 180ฐF to establish a nozzle seal. Operating
at 6350 RPM, the disc centrifuge discharges an oil stream for
recovery, a water stream for reprocessing, and a solids-water
stream, also for reprocessing through the API Separator.
THERMAL REGENERATION PERFORMANCE
Spent carbon regeneration has been achieved using a six-
hearth furnace fired by refinery fuel gas. The maximum regenera-
tion rate has been 250 pounds per day. The maximum steam addi-
tion rate for control of the furnace atmosphere has been 250
pounds per hour.
32
-------
TABLE 6
Thermal Regeneration Design Data
Furnace 60" x 6 Hearth with
Integral Afterburner
Regeneration Rate 125 LB/HR
Steam Addition Rate 125 LB/HR
Fuel Refinery Fuel Gas
Fuel Rate
Hearth 4 188 CFH
Hearth 6 68 CFH
Afterburner 310 CFH
Combustion Air Rate
Hearth 4 5000 CFH
Hearth 6 1800 CFH
Afterburner 8120 CFH
Design Temperatures
Hearth 4 1725ฐF.
Hearth 6 1750ฐF.
Afterburner 1250ฐF.
33
-------
A 40 percent decrease in the adsorptive capacity of the re-
generated carbon was observed following eighteen months of opera-
tion. Regenerated carbon iodine numbers in the range of 560-680
have been determined and show a decrease from virgin carbon
iodine numbers, which are in the range of 950-1000. Regenerated
carbon molasses numbers of 280 show an increase over the virgin
carbon molasses number of 230. A decrease in micro pores and an
increase in macro pores in the activated carbon are indicated by
the above results.
Adsorption isotherms were prepared using both regenerated and
virgin carbon and resulted in the following loadings at current
influent concentrations.
Virgin Carbon Loading Regenerated Carbon
0.17 Pounds TOC/Pound Carbon 0.096
0.73 Pounds COD/Pound Carbon 0.35
0.04 Pounds Phenol/Pound Carbon 0.03
In order to achieve greater regenerated carbon adsorptive
capacity, a revised regeneration furnace profile is currently
under evaluation. Rather than the design gradual temperature
increase through the furnace hearths, the revised profile main-
tains a temperature of 1200ฐF in number four hearth and increases
to 1750ฐF in number five hearth. The purpose of this rapid tran-
sition is to pass through the coking range, thereby preventing
plugging of the carbon micro pores.
Carbon regeneration, as measured by the carbon's apparent
density and comparing the value with virgin carbon, has been de-
termined to be a function of regeneration rate. With the carbon
out interval set at 4.5 hours, i.e., a 3000 pounds per day re-
generation rate, the regenerated carbon's density averaged
51.8g/100 cc. With the carbon-out interval set at 8 hours, i.e.,
a 1500 pounds per day regeneration rate, the regenerated carbon's
density averaged 50.0g/100 cc. The spent carbon averaged 59.4
and 59.5 during these periods respectively.
A tar-like substance consisting of carbon fines and water has
caused plugging of the .flue gas transfer line and wet scrubber.
This plugging results in a backpressure on the furnace which
activates the automatic furnace shutdown mechanism. The scrubber
trays and piping must then be removed for cleaning.
Two one-quarter inch lines were installed in the quench tank
to provide filtered water to flush the regenerated carbon away
from the furnace drop chute. Prior to this installation, re-
generated carbon had backed up on the chute and into No. 6 hearth
of the furnace, causing excessive metal temperatures.
34
-------
Carbon losses have recently been six percent per regenera-
tion cycle. Previous carbon losses were in the eight to ten
percent range. These initial high loss values may be attributed
to mechanical problems, carbon loss in the effluent, and carbon
losses which occur when an adsorber was overpressured arid vented
to the atmosphere. The losses attributable to each of thesfe
factors has not been determined.
The percent of time that the regeneration furnace has been
off line for maintenance has increased during each year of oper-
ation* During the first year of operation, the furnace was off
line 6 percent of the time; The percent offline during the
second year was 16. During the third year the furnace was off
line 22 percent of the time*
At various times .the furnace was shutdown in order to: (1)
clean the scrubber and flue gas line of built-up tar; (2) clean
the combustion air lines and gas lines and controls of corrosion
products; (3) repair cracks in the flue gas line; (4) replace the
furnace sand seal; (5) realign the ultraviolet flame sensors: or
(6) other routine maintenance.
One particular furnace shutdown was caused by the inability
to transfer regeneraterated carbon due to the number of leaks
which developed in the two-inch carbon transfer lines. The
leaks were the result of corrosion along the bottom section of
the transfer line in the horizontal run between the blow case
and vertical pipe section* The determination of pipe wall thick-
ness along the remainder of the carbon steel pipe and bends
revealed no other significant reduction in wall thickness. It
is suspected that carbon lying in the horizontal line resulted
in corrosion of the carbon steel wall.
A number of shutdowns were caused by failure of the flue
gas line from stress corrosion cracking probably the result of
chloride attack.
The corrosion in the combustion air lines and controls was
determined to be the result of carbon fines entering the
suction of the combustion air blower. The source of the fines
was the addition of makeup carbon to the quench tank which was
located adjacent to the combustion air blower. A filter and
protective cover were added to the blower to precent the entrance
of carbon fines and excess water.
SOLIDS DEWATERING SYSTEM PERFORMANCE
The scroll centrifuge has demonstrated average removals of
30 percent for suspended solids. The disc centrifuge has demon-
strated average removals of 91 percent for oil. Performance data
for the scroll and disc centrifuges follows:
35
-------
Centrifuge performance
Percent Percent
Solids Oil
Scroll Centrifuge Feed 11.6 2.7
Scroll Centrifuge Centrate 0.7 5.9
Scroll Centrifuge Sludge 33.6 5.5
Disc Centrifuge Water Discharge 0.1 0.3
Disc Centrifuge Solids Discharge 0.4 0.3
Scroll Centrifuge 13 gpm
Disc Centrifuge 43 gpm (including 30 gpm utility water)
A major problem associated with the solids dewatering system
is the recycle of solids and oil to the API Separator influent.
This recycle has increased the solids and oil loading and emul-
sion volume on all downstream treatment units. This increase has
in turn overloaded the solids dewatering units and 'has resulted
in the "back-up" of sludge and emulsion in the API.Separator.
The average individual contributions to the solids system
based on two one-week surveys are reported below:
ft/Day
Source Gal/Day OilSolids
Emulsion Treater 21,300 46,150 5,160
Filter Backwash 45,000 480 520
API Separator 2,000 630 720
Total 68,300 47,260 6,400
During this period 27,500 pounds of oil and 5,700 pounds per
day of solids were recycled to the separator influent. The
sources of this recycle are the thickener overflow and disc
centrifuge.
Sludges from the API Separator and Emulsion Treater caused
severe plugging problems in the solids dewatering system. The
inlet screens at the sludge blending tank, two inch transfer
lines, pump impellers, and the double pipe heat exchangers ex-
perienced plugging due to the debris contained in these sludges.
The problem has been corrected by the installation of a comminut-
or.
In order to improve operation of the sludge thickener, and
reduce recycle to the separator, unbroken emulsion from the
Emulsion Treater currently bypasses the sludge thickener and is
transferred directly to the sludge holding tank. Rather than
using the design method of continuously transferring to the
36
-------
sludge holding tank, thickener bottoms are pumped directly to the
scroll centrifuge for 30 minutes during each shift.
The wear plates on the sludge pumps had to be renewed after
two years of operation. The abrasive carbon fines are suspected
to be the cause of the excess wearing.
In order to maintain a minimum sludge velocity of five feet
per second, the scroll centrifuge feed piping has been reduced
from two inch to one and a quarter inch diameter. Following the
installation, an improvement in both the scroll and disc centri-
fuge operation has been observed. The solids concentration of
the scroll cake has increased and the disc centrifuge has main-
tained a longer run time between shutdowns.
An internal inspection of the scroll centrifuge, following
nine months of operation, revealed that the hard surface coating
on the internal flights had worn away exposing the stainless
steel. One-half inch thick, layered, deposits of carbon fines
and grit were found inside the scroll. The scroll was returned
to the manufacturer for resurfacing, and a spare scroll was ob-
tained.
The internal flights required a second resurfacing after
being in operation for 24 months. The increase in on line time
may be attributed to a decrease in the recycle of carbon fines to
the solids handling system.
A strainer with continuous backwash was installed upstream
of the disc centrifuge to remove solids which escaped the scroll
centrifuge. Carbon fines not removed by the scroll centrifuge
have on occasion eroded the screen of the strainer.
The disc centrifuge sealing water has been flow controlled
at 25 gpm to prevent the flushing of bearing grease which occur-
red previously, causing a bearing burnout and resulting in an
extended shutdown period. The maximum period that the disc
centrifuge has been on stream without plugging has been 34 days.
The downtime required for removal of the disc stack and cleaning
can extend to three days. During this downtime, centrate from the
scroll centrifuge is recycled to the API Separator. Visual in-
spection of the disc stack has revealed a coating of carbon fines.
The factors discussed above have contributed to increasing
downtime for the centrifuge system. During the first year of
operation the centrifuge system was off-line 29 percent of the
time. During the second year the off-line time increased to 37
percent and increased to 42 percent during the third year.
The rate of oil recovered from the disc centrifuge is a
function of the operation of the API Separator, the Emulsion
37
-------
Treater, and the sludge handling facilities. Oil recovery has
approached 20 gpm on occasion.
Construction of inlet sludge pumping stations at the API
Separator has permitted removal of sludge from the separator
inlets. This has resulted in improved separator operation, and
a reduction in the volume of sludge directed to the solids han-
dling facilities.
38
-------
SECTION 8
WASTEWATER TREATMENT SYSTEM PERFORMANCE
DUAL MEDIA FILTRATION PERFORMANCE
The dual media filters demonstrated average removals of 68
percent for suspended solids and 75 percent for oil during the
first year of operation. These averages are taken from the data
reported in Tables 7-10. Although individual operating period
characteristics have varied, performance of the filters remained
uniform during the first two years of operation. The filter's
removal efficiency decreased as media was gradually lost from the
filter vessels. This loss in efficiency is evidenced by the data
reported in Table 11 for operations during the third year after
start-up. For the period January, 1975-December, 1975 the mean
removal for suspended solids was 59 percent and 36 percent for
oil. Table 11 also reports the filter influent and effluent mean
and maximum concentrations for suspended solids, oil, TOC, COD,
phenol, and sulfide for the period January-December, 1975.
During this operating period the mean flow to the filters was
1939 gpm with a maximum of 2290 gpm.
Additional data on the filter operations is reported in the
appendix. Histograms are plotted for the frequency of occurrance
of suspended solids, oil, TOC, COD, phenol and sulfide concentra-
tions for the year January-December, 1975. Also reported is the
mean, standard deviation, maximum and minimum for each parameter.
Table 12 presents an analysis of filter backwash water.
Based on this data, the backwash high rate flush duration was
increased to 7 minutes and the flow rate was increased to greater
than 2000 gpm. The total backwash duration averages 20 minutes
and is dependent upon the time required for the pressurized re-
moval of water remaining in the filter. The time required is a
function of the differential pressure across the filter when a
backwash is initiated. The backwash interval is set at 4 hours,
i.e., each individual filter is backwashed every 12 hours.
Although the design included a differential pressure over-
ride to initiate backwash, this option has not been used since
the maximum differential pressure reached during the above back-
wash interval has been 3 psi.
39
-------
TABLE 7
Wastewater Analyses
Period 1
April 1973
- Initial Operation
- Virgin Carbon
- Foul Condensate Not Included
Parameter
suspended
Solids
BOD5
COD
Oil
Phenol
Sulfide
^fHiCm JJ fr^-
Flow
Concentration, ppm
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Filter
Influent
70
95
76
386
478
21
34
21
21
-
-
1485
1520
Adsorber
Influent
16
29
46
216
248
10
19
20
20
0.19
0.35
15.2
16.2
1375
1710
Adsorber
Effluent
11
14
40
63
77
0.34
0.72
0.023
0.028
13.8
18
19
20
Percent
Removal
Filter
77
39
44
52
5
-
-
Adsorber
31
13
71
97
,99-9
Increase
Increase
40
-------
TABLE 8
Wastewater Analyses
Period II
July 1973
- Early Regeneration
- Carbon Bed Not Turned Over
- Foul Condensate Included
Parameter
Suspended
Solids
BDD5
COD
Oil
Phenol
Flow
Concentration, ppm
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Filter
Influent
68
115
75
388
522
74
143
14
19
1850
1900
Adsorber
Influent
34
52
78
296
361
25
66
12
22
1405
1505
Adsorber
Effluent
21
67
72
133
184
7'7
28
4.5
8.4
Percent
Removal
Filter
61
-
24
66
14
A&sorber
38
8
55
69
63
41
-------
TABLE 9
Wastewater Analyses
Period III
October 1973
- Carbon Bed Turned Over
Maximum Regeneration Rate
- Foul Condeneate Included
Parameter
Suspended
eolids
COD
Oil
rhcnol
euin.dc
Affoonia
now
Avg
Max
Avg
Kax
Avg
Max
Avg
Max
Avg
Mas
Avg
Kax
Avg
Max
Concentration, ppm
Filter
Influent
67
126
415 _
660
67
lift
32.5 ,
34.5
-
"
1855
I960
Adoorber
Influent
20
34
322
400
10.4
16.3
32
33.5
7'5 16.6
87
103
1420
1600
Adsorber
Effluent
16
43
242
300
2.1
3.2
12.9
20.0
37
11.2
93
110
Percent
Removal
Filter
70
22
64
-
-
-
Adsorber
20
25
60
60
Increase
Increase
42
-------
TABLE 10
Wastewater Analyses
Period IV
February 1974
- Effluent Septums Bent
- Foul Condensate Not Included
- One Year of Operation
Parameter
Suspended
Solids
BODg
COD
OH
Phenol
Sulfide
ABB&onia
flow
Concentration, ppm
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Avg
Max
Filter
Influent
64
74
88
408
500
78
125
2.0
3.8
-
-
1770
1970
Adsorber
Influent
16
29
55
301
390
13
22
1.9
3.7
0.6
0.9
12.3
13.5
1030
1175
Adsorber
Effluent
55
99.0
65
252
330
8
16
ฐ-7
0.8
13
25
13.0
15.0
Percent
Removal
Filter
75
36
25
83
-
-
-
ป
Adsorber
Increase
Increase
12
3B
63
Increase
Increase
43
-------
TABLE 11
Wastewater Analyses
Period V
January 1, 1975-December 31, 1975
Parameter
Suspended
Solids
TOC
COD
.OIL
Phenol
Sulfide
Flow, CPN
Concentration, ppm
Mean
flax
Mean
Max
Mean
Flax
Mean
flax
Mean
flax
Mean
Wax
Tie an
flax
Filter
Influent
116
507
162
479
475
940
47
16B
14.2
B5
-
1939
2290
Adsorber
Influent
48
396
122
362
331
680
30
110
13.5
BO
2.3
32
731
1252
Adsorber
Effluent
45
999
71
3B4
195
1012
14
99
15.1
76
14.3
71
Percent
Removal
Filter
59
25
31
36
5
Adsorber f
6
42
40
53
Increase
Increase
44
-------
TABLE 12
Analyses of Sand Filter Backwash Water
Time After
Initiating
High Rate
(1960 GPM)
Backwash Suspended Solids
Mins .
0 87
1 19,370
2 9,010
3 4,160
4 3,010
5 850
6 300
TABLE 13
Analyses of Water Removed from Sand Filters by
Pressure Prior to Backwash
Time After
Air Pressurization Suspended Solids
Minutes _ _ ppm _
0 13
1 42
2 296
3 196
4 156
5 77
6 91
7 2,010
8 5,140
9 4,900
45
-------
The factor limiting the backwash interval has been the
capacity of the sludge blending tank which receives backwash
water.
An operating problem encountered was a decrease in effluent
quality due to a backwash cycle. The two filters remaining on
line experience a "shock" as the individual flow rates increase
to include that portion of flow previously handled by the third
filter. The decrease in effluent quality was due to the expel-
ling of suspended solids and oil trapped in the filter, and the
decrease in removal at the increased hydraulic loading. An
increase in effluent quality was obtained by setting the controls
to maintain the established flow rate to each on-line filter
during a backwash cycle.
A decrease in effluent quality was also observed during the
removal of water remaining in the filter by applying air pressure
prior to backwash. Table 13 presents an analysis of this water
and indicates that suspended solids are removed from the filter
and contaminate the effluent water stored in the filtered water
holding tank.
A draindown system to return water pressurized from the
filters during the first stage of backwash to the API Separator
was installed to increase the quality of filtered water. The
suspended solids removal efficiency increased from a range of
27-68 percent to a range of 45-76 percent following activation of
the draindown return to the API Separator. The oil removal ef-
ficiency increased from a range of 47-52 percent to a range of
68-86 percent. Evaluation of selective similar influent grab
sample values for oil, occuring on days with approximately equiv-
alent flow rates led to the above reported percent reductions.
The reported suspended solids values were derived from a 24 hour
composite collected on the day of the reported oil values. The
data base for the above evaluation is presented in Table 14.
Over a period of 18 months of operation, 13 inches of
anthrafilt was lost from the filter vessels. During the subse-
quent 18 months of operation, an additional 6 inches was lost
from the vessels. Also, over this 36-month operating period, 12
inches of sand was lost from the vessels.
ACTIVATED CARBON ADSORBER PERFORMANCE
The adsorbers demonstrated varying performance over the one
year operating period from March, 1973, through February, 1974,
as a function of individual operating period characteristics.
Table 7 presents influent and effluent concentrations observ-
ed during the Period 1 when the adsorbers were in initial opera-
tion with minimum carbon bed pulsing and virgin carbon. The foul
46
-------
TABLE 14
Dual Media Filters Draindown Data
OIL
Draindown to Filtered Water Tank
Flow Dual Media Filter
Influent Effleunt 7ซReduction
Draindown to API Separator
Flow Dual Media Filter
GPM
1502
1757
1405
1530
ppm
132
111.7
71.8
53.1
Influent Effluent ^Reduction
ppm
69.7
57.7
40
25.2
47%
487.
447.
527.
GPM
1470
1684
1467
1539
ppm
135.5
123.1
75.9
57.4
ppm
25.3
39.6
13.4
8.2
817.
687.
827.
867.
SUSPENDED SOLIDS
Draindown to Filtered Water Tank
Flow Dual Media Filter
Influent Effluent 7.Reduction
GPM
1502
1757
1405
1530
ppm
91
97
41
38
ppm
66
67
23
12
277.
317.
447.
687.
Draindown to API Separator
Flow Dual Media Filter
Influent Effluent 7ซReduction
GPM
1470
1684
1467
1539
ppm
98
73
55
53
ppm
54
28
13
13
457.
627.
767.
757.
-------
condensate stream from the Fluid Catalytic Cracker was not
included with the process wastewater. With minimum carbon bed
pulsing, an essentially static bed resulted in removal of oil to
0.34 ppm as measured by Freon extraction and infra-red absorption.
This removal was not observed following full scale regeneration
and normal bed movement. Phenol removal approached 100 percent
during this period due to the low influent loading and a low
spent carbon wave front. The increase in sulfide concentration
from an influent of 0.19 ppm to an effluent of 13.8 ppm, is
sulfide production occurring during initial operation.
The influent concentration from organic sulfur compounds was
determined to be 0.01 ppm. The sulfide present in the effluent
created an effluent odor problem and further investigation indi-
cated the presence of butyl mercaptan, thiophene, and dimethyl
sulfide.
The production of sulfide was not observed during a period
when the refinery was shut down for maintenance and modernization.
Total organic carbon results on the adsorber influent and efflu-
ent indicated that organic material was being removed during the
shutdown period. The production of sulfide was also not observed
during operation of a trial pH adjustment system at the API
Separator. With the API Separator inlet pH maintained at 6.5-7.0,
an increase in sulfide concentration did not occur across the
carbon adsorbers. It was concluded that it is not just the
presence of bacterial action which results in sulfide production,
but rather it is a function of the influent content. This
conclusion was not pursued further.
The difference between the reported filter and adsorber flow
rates is due to the utilization of filtered water as unit service
water and as backwash water.
The influent and effluent concentrations observed during
Period 2 are presented in Table 8. The foul condensate stream
from the Fluid Catalytic Cracker was included with the process
wastewater during this period.
The increase in effluent phenol concentration observed
during this period may be attributed to the introduction of
100 gpm of stripped foul condensate containing an average 300 ppm
phenol. The adsorptive capacity of the carbon for phenol was
0.03 pounds of phenol per pound of carbon, based on an adsorption
isotherm. The introduction of foul condensate therefore resulted
in an influent phenol overload.
During this period the spent carbon wave front moved upward
through the carbon bed. In order to achieve increased removals,
the carbon regeneration rate was increased to 250 pounds per hour,
thereby providing additional adsorptive capacity. It was observ-
ed that during the period in which the spent carbon wave front
48
-------
had moved upward in the adsorber, an increase in effluent phenol
concentrations occurred at low adsorber influent concentrations.
This was a result of the adsorbed phenol achieving equilibrium
with the phenol in solution in the wastewater. Again, lowering
the spent carbon wave front will provide additional adsorptive
capacity and eliminate the occurrence of this phenomenon.
Table 9 presents influent and effluent concentrations during
Period 3 when the carbon bed had turned over, the regeneration
rate was at the maximum 250 pounds per hour, and the foul con-
densate stream was included with the process wastewater.
The increase in effluent COD concentrations observed during
Periods 3 and 4 may be attributed to a change in influent waste-
water characteristics. During the design stages, the adsorption
isotherm prepared using virgin carbons indicated a theoretical
loading of 0.3 pounds TOC per pound carbon. However, current
isotherms using virgin carbon resulted in an average loading of
0.17 pounds TOC per pound carbon, thus indicating a change in
influent characteristics. This loading corresponds to an ex-
haustion rate in the range of 2.9-6.3 pounds of carbon per 1000
gallons of throughput, as compared to the exhaustion rate of one
pound of carbon per 1000 gallons of throughput predicted by the
initial isotherm.
The influent and effluent concentrations observed during
Period 4 are presented in Table 10.
The Wastewater Treatment Plant had been in operation for a
one-year period. The adsorbers operated during this period with
flow rates ranging 630 to 1175 gpm. It is suspected that exces-
sive carbon fines in the carbon beds and fines plugging the
effluent septums were causing the flow restriction.
Table 11 presents adsorber influent and effluent concentra-
tions during Period 5, January-December, 1975. The adsorbers
operated during this period with a mean flow rate of 731 gpm.
Carbon fines were continuing to plug the carbon beds and effluent
septums. During this operating period, the mean removal of COD
was 40 percent, for TOC the removal was 42 percent and for oil
the mean removal was 53 percent.
Additional data on the adsorber operations is reported in
the appendix. Histograms are plotted for the frequency of
occurrance of suspended solids, oil, TOC, COD, phenol and
sulfide concentrations for the year January-December, 1975. Also
reported is the mean, standard deviation, maximum and minimum for
each parameter. ';
The excessive suspended solids discharge reported in Tables
10 and 11 was due to holes in the effluent septums. The effluent
49
-------
septums were bent inward creating openings in the screens. In
order to correct the septum bending problem, new septums were in-
stalled. The new septum design included an internal four inch
diameter stainless steel sleeve with one-half inch bored holes
to provide structural strength to the external 0.015 inch slotted
stainless steel screen. A deformation of the upper adsorber
cone has been observed at the location of various septums. It
is suspected that this deformation is the result of force trans-
mitted to the three-eighth inch thick carbon steel cone when a
septum was bent by the rising carbon bed.
Carbon fines had been removed from the system by flushing
the carbon adsorbers. The flush is accomplished by closing the
adsorber effluent valves and introducing 150 gpm into the base
of the vessel and allowing the bed to "fluff" into carbon stor-
age tank above the adsorber. The flush water containing the
carbon fines overflows the carbon storage tank to the carbon
settler, from which the water is directed to the API Separator
and the settled carbon fines are transferred to the regeneration
furnace where they are combusted.
The adsorbers were flushed after every "carbon out" cycle.
This flush, together with a reduced regeneration rate, resulted
in a decrease in carbon fines discharged with the effluent water
when full flow is gradually returned to the adsorbers. The
flushing also resulted in a significant decrease in adsorber ef-
fluent oil concentrations. However, flushing the adsorbers to
remove carbon fines were discontinued. It was suspected that
this flushing was generating additional fines, was contributing
to the septum bending problem, and was causing gaps in the car-
bon bed within the adsorber.
Eight new septums were installed in each of the three carbon
adsorbers. Prior to installation of the new septums, the total
adsorber flow rate ranged from 500-1015 gpm, while the effluent
suspended solids concentration ranged from 2-415 ppm. Following
installation of the new septums, the total adsorber flow rate
increased in range from 1035-1535 gpm, while the effluent sus-
pended solids concentration decreased in range from 9-32 ppm.
Attempts were made during a "carbon out" cycle to backflush
the effluent septums while the carbon was at low level, and the
adsorber influent flow was shut off. This attempt did not result
in an increase of flow through the adsorber.
An activated carbon pilot plant was operated in parallel
with the full scale adsorbers in an effort to determine the
cause of the flow restrictions observed in the adsorbers.
The pilot plant was monitored by taking flow and pressure
readings through four 4 inch diameter by 4 foot length glass
columns. The pressure gauges were located to permit the deter-
50
-------
mination of pressure drop across the carbon bed or across the
inlet or outlet septum. Regenerated carbon was used to fill the
columns.
Initially, 50 psi inlet pressure was required to achieve the
design flow of 2.9 gpm. The pressure drop was occurring at the
outlet septum of each column. No readable pressure drop was
occurring across the columns per se, only across the septum.
After one hour of operation, the design flow could not be
maintained. Visual observation indicated that carbon fines were
concentrating at the top of each column around the outlet septum.
The columns were hydraulically defined and the pilot plant
operation was continued. This time only 5 psi feed pressure was
required to achieve the design flow of 2.9 gpm.
The pressure drops were about equally divided over the four
towers and outlet septum. The pilot plant was operated contin-
uously for a two week period. The feed pressure had to be in-
creased to 20 psig to maintain the design flow rate. During the
first week, obvious visual evidence of biological growth was
present in the carbon beds. Hov?ever, the pressure drop through
the four beds amounted to a total of less than 0.5 psi. The
remainder of the pressure drop occurred at the outlet septum.
The pilot plant was operated for another two weeks and the same
effect was observed, except that a feed pressure of 40 psig was
required to maintain the target flow rate.
At the end of the one month period, the towers were dis-
assembled. The outlet septums were caked with a gritty black
powder which x-ray identified as carbon fines. The carbon beds
were coated with a slimy brown-black mass which from visual
appearance and unique odor was bacterial. A bacterial plate ^
count of the material around the outlet septum showed 15 x 10
organisms/gm, while the count through the bed was greater than
10ฐ organisms /gpm.
In conclusion, bacterial growth was not the cause of the
flow reduction observed in the pilot plant. Fines pluggage of
the outlet septum appears to be the main cause of the flow
reduction observed in the carbon columns.
The septum backwash transfer piping has been increased to
three inch, with a two inch manifold at each adsorber to provide
adequate flow to remove carbon fines that plug the septums and
restrict flow. The design provides for a backflush of 90 gpm,
which is equivalent to the effluent flow from each septum.
An operating problem encountered was the decrease in efflu-
ent quality due to the pulsing of spent carbon from an off-line
adsorber. The two adsorbers remaining on-line experience a
51
-------
"shock" as the individual established flow rates increased to in-
clude that portion of flow previously handled by the third adsor-
ber. The decrease in effluent quality was due to the expelling
of oil and carbon fines from the adsorber when flow was restored.
In order to reduce the "shock" to the carbon bed, the times re-
quired to open and close were set to maintain the established
flow rate to each on-stream adsorber during the carbon pulse cycle,
The time period for a "carbon out" cycle is a function of
the time required to backwash the septums and the time required
to return full flow to the adsorber while preventing the discharge
of carbon particles with effluent. The maximum time an adsorber
was off line during a "carbon out" cycle was two hours.
Excessive carbon losses have occurred when an adsorber is
overpressured and vents to the atmosphere. In order to retain
This carbon, the adsorbers are now operated as pressure vessels.
Rupture discs and vacuum relief valves were installed in the
adsorber vent line. A pressure sensor was installed in the vent.
line and signals for an automatic shutdown in the event an adsor-
ber, is overpressured.
, After one year of operation, thickness measurements were
made on the adsorber vessels using a sonoray instrument. Concen-
trated measurements were made on the lower cone of each adsorber.
All measurements indicated no appreciable loss in wall thickness.
Subsequent to two years of operation, leaks developed in the
lower cone of each adsorber. External repairs were made by se-
curing a rubber gasket and a piece of rolled steel over the
leaking area.
52
-------
SECTION 9
ECONOMIC EVALUATION
Table 15 presents the Wastewater Treatment Plant capital
cost which totaled $1,812,000. This figure is reported in 1973
dollars.
Tables 16-19 present detailed operating costs for the fil-
ters, adsorbers, carbon regenerator, and solids dewatering
system. The total annual operating cost for the period October,
19:73-September, 1974, was $233,980 (Table 20). The carbon regen-
eration system accounted for 46 percent of this operating cost.
The unit cost of regeneration including carbon makeup was 8.87ฃ/#
carbon regenerated. The unit cost without including carbon make-
up was 5.84ฃ/# carbon regenerated. The filters accounted for 10
percent of the annual operating cost, the adsorbers 17 percent
and the solids dewatering system accounted for 27 percent.
Table 21 summarizes project capital and operating costs. In
all systems, operating labor and utilities accounted for the
majority of the operating costs. Operating labor accounted for
44 percent of the annual operating costs while utilities account-
ed for 25 percent. Makeup carbon for the regeneration system
accounted for 16 percent of the annual operating costs. Mainten-
ance accounted for approximately 10 percent.
Major needs contributing to maintenance costs were resurfac-
ing of the scroll centrifuge internal flights, modifications to
the solids dewatering feed system to eliminate clogging problems,
addition of structural support to the carbon adsorber skirts and
replacement of the carbon adsorber effluent septums to correct
the septum bending problem.
53
-------
TABLE 15
Wastewater Treatment Plant
Capital Cost
Item
Engineering
Surge Basin
Sand Filters
Carbon System
Carbon Charge
Solids Dewatering
Tanks
Pumps
Building
Piping
Electrical
Instrumentation
Structural
Foundations
Concrete
Cost
125
33
187
348
111
152
93
13
149
179
199
137
22
45
12
,000
,900
,100
,200
,200
,300
,800
,400
,200
,600
,900
,500
,100
,900
,800
Total $1,812,000
Operating Labor
Power
Instrument Air
Steam
Maintenance
Supplies
TABLE 16
Dual Media Filters
Operating Costs
10/1/73 - 9/30/74
Basis Cost
Refinery Records
(12% of total WTP) $12,340
79 HP ($78.68 HP/YR) 6,220
21 SCFM ง0.03/1000 SCF 330
(35% of total WTP)
Steam Tracing 350
Refinery Records 560
Refinery Records 400
$20,200
54
-------
TABLE 17
Activated Carbon Adsorption
Operating Costs
10/1/73 - 9/30/74
Basis Cost
Operating Labor Refinery Records $24,690
(25% of total WTP)
Power 115 HP ($78,68/HP/YR) 9,020
Instrument Air 6 SCFM @ $0.03/1000 SCF 100
(10% of total WTP)
Steam Steam Tracing 400
Maintenance Refinery Records 3,180
Supplies Refinery Records 400
$37,790
55
-------
TABLE 18
Operating Labor
Carbon Makeup
Power
Instrument Air
Steam
Fuel
Maintenance
Supplies
Activated Carbon Regeneration
Operating Costs
1,195,920# Regenerated
10/1/73 - 9/30/74
Basis Cost
Refinery Records $37,030
(38% of total WTP)
86,020# 36,130
42^/lb. carbon cost
(equates to 7.270 makeup)
31.5 HP ($78.68/HP/YR) 2,470
21 SCFM @ $0.03/1000 SCF 330
(35% of total WTP)
1# steam/#carbon regen. 3,850
$2.75/1000* steam
+ $560 Steam Tracing
Refinery Records 16,590
Based on Regen. Rate
$0.80/MM BTU 20,740 MM BTU
Refinery Records 9,260
Refinery Records 400
Unit Costs Including
Makeup
Unit Costs Not Including
Makeup
$106,060
8.87<ฃ/# carbon
5.840/1 carbon
56
-------
Operating Labor
Power
Instrument Air
Steam
Maintenance
Supplies
Sludge Disposal
TABLE 19
Solids Dewatering
Operating Costs
10/1/73 - 9/30/74
Basis Cost
Refinery Records $24,690
(25% of total WTP)
149 HP ($78.68/HP/YR) 11,700
12 SCFM @ $0.03/1000 SCF 200
(20% of total WTP)
Steam Tracing and Heat 3,850
Exchangers
Refinery Records 8,580
Refinery Records 400
Refinery Records 10,510
$59,930
57
-------
TABLE 20
Wastewater Treatment Plant
Annual Operating Costs
Annual Operating Cost % Of Total
Dual Media Filtration
Activated Carbon Adsorption
Carbon Regeneration
Solids Dewatering
TOTAL
$ 20,200
37,790
106,060
59,930
223,980
TABLE 21
Wastewater Treatment Plant
Project Cost Summary
10
17
46
27
100
Item
Capital
Operating
Labor
Utilities
Carbon Makeup
Sludge Disposal
Maintenance
Supplies
TOTAL
Cost
1,812,000
98,750
55,410
36,130
10,510
21,580
1,600
223,980
% of Total
44
25
16
5
10
<1
100
58
-------
APPENDIX A
WASTEWATER CONCENTRATION HISTOGRAMS
CARBON ADSORBER INFLUENT
FREQUENCY 3 13 2 6 7 23 21 36 6 7
36
30
20
10
I I I I I
= ,n
n,
L
1.000 251.200 501.400 751.600 1001.800 1252.000
126.100 376300 626.500 876.700 1126.900
FLOW, GPM
Figure A-l. Flow Histogram
59
-------
FILTER INFLUENT
FREQUENCY
6
16 36 24 23
36
34
32
30
28
26
24
22
20
16
16
14
12
10
8
6
4
2
. 1
i
1
1 I 1 1 1 1
i i i Hi Mi
|
M
| 1 II
|
i
mm
\
-.
|
^
|
1137.000 1367.600 1598.200 1828.800 2859.400 2290.000
1252.300 1482.900 1713.500 1944.100 2174.700
FLOW, GPM
Figure A-2. Flow Histogram, filter influent
60
-------
CARBON ADSORBER EFFLUENT
FREQUENCY
I 5 15 19 34 19 20
34
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
: i i r
ป
-
-
ป
*
>
, n, m
mm
]
TI r
**i
i ,
i
i
m
1 1 1 1 |
|
|
mm
f
-
1 1 1
5.500 6.380 7.260 8.140 9.020 9.900
5.940 6.80 7.700 8.580 9.460
pH
Figure A-3. pH Histogram, carbon adsorber effluent
61
-------
CARBON ADSORBER EFFLUENT
FREQUENCY 61 20 15 6 4 5 2 6 4 2
60
50
40
30
20
10
n
i
-
-
-
-
-
i-
. . . .fl.ll.n. .(In.
.220 14.376 28.532 42.688 56B44 71.000
7.298 21.454 35.610 49.766 63.922
SULFIDE, ppm
Figure A-4. Sulfide Histogram
62
-------
CARBON ADSORBER INFLUENT
FREQUENCY
110
/
10
8
6
4
2
n
1
<
II
145100102
i i i i i i i i i i
nr
ii
, , n, , , n, , 1 li
.1 6.1 12.0 20.0 26O 32.0
3.3 9.7 14.0 22.0 28.0
SULFIDE, ppm
Figure A-5. Sulfide Histogram
63
-------
FREQUENCY
CARBON ADSORBER EFFLUENT
54 26 20 9
54
50
40
30
20
10
l
I I I I
I
I. n. PI. n. II.
.500 15.600 30.700 45.800 60.900 76.000
6.050 23.150 38.250 53.350 68.450
PHENOL, ppm
Figure A-6. Phenol Histogram
64
-------
CARBON ADSORBER INFLUENT
FREQUENCY 6725 15 5 4 22311
inป I ' ' ' i I i I I
60
50
40
30
20
10
n.n.n.n.n.
n.
.750 16.60032.4504830064.15080.000
8.675 24525 40.375 56.225 72.075
PHENOL, ppm
Figure A-7. Phenol Histogram, filter effluent
65
-------
FILTER INFLUENT
FREQUENCY 69 27 II 7 123032
69
60
50
40
30
20
10
n
1
i
1
^m
1 1 1 1 1 1 1 1 1
a
p
, , ,.n,ri,n, . ,n,n,
.260 17.208 34.156 51.104 68.052 85.000
8.734 25.682 42.630 59.578 76.526
PHENOL, ppm
Figure A-8. Phenol Histogram
66
-------
CARSON ADSORBER EFFLUENT
FREQUENCY 68 27 18 3 1 3 I 2 1 1
68
60
50
40
30
20
10
ซ
'
j-
ป
M^^B
MB
|
1 1 1 I 1 1 1 1
MM
p
i 1 1 1, ni 1 h DI Hi ni Hi
1.000 20.600 40.200 59800 79.400 99.000
10.800 30.400 50.000 69.6OO 89.200
OIL, ppm
Figure A-9. Oil Histogram
67
-------
CARBON ADSORBER INFLUENT
FREQUENCY 42 24 19 8 10 4 7 35 3
42
40
30
20
10
r r-nr
I
I
I
JL Jl
1.000 22.800 44.600 66.400 88.200 110.000
11.900 33.700 55.500 77.300 99.100
OIL, ppm
Figure A-10. Oil Histogram, filter effluent
68
-------
FILTER INFLUENT
FREQUENCY 29 33 27 6 18 12 I 3 2 I
.33
30
20
10
n
: i r
m
m
ซ
i
|
i
1 1
^i
-
i i i T r r
I
MB
|
M
.n.fl.n.n,
1.000
17. TOO 51.100 84.500 117.900 151.300
OIL, ppm
Figure A-11. Oil Histogram
69
-------
CARBON ADSORBER EFFLUENT
FREQUENCY 42 46 20 8 2 I 3 I 0 2
46
40
30
20
10
n
I 1
*
-
-
- 1
MMHM
|
HH
1 | 1 1 1 1 1 1 1
|
m^m
*,
i i Hi ni 1 L ni i Hi
15.000 214.400 413.800 613.200 812.600 1012.000
114.700 314.100 513.500 712.900 912.300
COD, ppm
Figure A-12. COD Histogram
70
-------
CARBON ADSORBER INFLUENT
FREQUENCY 12 23 17 28 15 12 5 6 4 3
28
20
10
^1iir
i
15.000 188.000 361.000 534.000 707.000 830.000
101.500 274.500 447.500 620.500 793.50
COD, ppm
Figure A-13. COD Histogram, filter effluent
71
-------
FILTER INFLUENT
FREQUENCY 9 13 16 15 25 18 10 7 9 3
25
20
10
n
I 1 1 1 1
" 1
1
mtm
mmm
IM
I 1 1 1 1 1
ซ
"i
n
100.000 268.000 436.000 604000 772.000 940.000
184.000 352.000 520.000 688.000 856.000
COO, ppm
Figure A-14. COD Histogram
72
-------
CARBON ADSORBER EFFLUENT
FREQUENCY 39 14 7 4 I 2 I 0 0 I
3TFnpri r~iiiiiii
30
10
i rii I li rii i
16.000 89.600 163.200 236.800 310.400 384.000
52.800 126.400 200.00 273.600 347.200
TOC.ppm
Figure A-15. TOG Histogram
73
-------
FREQUENCY
FILTER EFFLUENT
CARBON ADSORBER INFLUENT
22 15 8 6
I I
22
20
10
n
I
ป
w
m
1
^B
1 1
|
M
|
I
| | 1 1 1 1 1 1
|
|
ซ
, ,n, , n, n,
8.000 78.000 149.600 220.400 291.200 362.000
43.400 114.200 185.000 255BOO 326.600
TOC, ppm
Figure A-16. TOC Histogram, filter effluent
74
-------
FILTER INFLUENT
FREQUENCY 4 17 18 17 5 3
0 0 I
18
10
n
-
-
|
^^^
i
1
|
a
| I 1 i i i I
, 1 1 1, n, , , n,
83.900 171.700 259.500 347.300 435.100
TOC, ppm
Figure A-17. TOC Histogram
75
-------
CARBON ADSORBER EFFLUENT
FREQUENCY 115 32300000 2
H5H"r
10
n.n.n.
1.0 20.0 40.0 60.0 80.0 100.0
10.0 30.0 60.0 70.0 90.0
SUSPENDED SOLIDS, ppm
Figure A-18. Suspended Solids Histogram
76
-------
CARBON ADSORBER INFLUENT
FREQUENCY 67 40 13 3 I 00001
' i i I i i i i i i
60
50
40
30
20
10
n
.Hi
i i
i ifli
1.000 83000 B9.000 236000 317030 396X80
40.500119500196.500 277.500 366500
SUSPENDED SOUDS, ppm
Figure A-19. Suspended Solids Histogram, filter effluent
77
-------
FILTER INFLUENT
FREQUENCY 20 49 34 10 2 4 3 20 I
49
40
30
20
10
n
i
-
-
m
mm
1
I^BM
1
|
H
1
|
Hi
I 1 I 1 1 1 ' 1
f
, , n. . II, n, , n,
2.000 103.000 204X>00 305.000 406.000 507.000
52.500 153500 254.500 355.500 456.500
SUSPENDED SOLIDS, ppm
Figure A-20. Suspended Solids Histogram
78
-------
APPENDIX B
ENGLISH-TO-METRIC UNIT CONVERSIONS
Multiply
This
Ibs
short tons
short tons
inches
feet
statute miles
gallons
barrels
Btu
SCF
Btu/lb
Btu/CF
Btu/SCF
109 Btu/day
106 Btu/day
MM Btu/hr
SCFD
MM SCFD
SCF/MM Btu
Ibs/MM Btu
Ibs/CF
psi
gpm
acre-ft/year
horsepower
nautical miles
knot
By
This
0.4536
0.9072
907.2
2.54
0.3048
1.609
3.785
0.1590
0.252
0.02679
0.5556
8.899
9.406
252
252
252
0.02679
0.02679
0.1063
1.8
16.02
0.07031
0.227
0.1408
745.7
1.852
1.852 :
To Obtain
This
kg
metric tons
kg
cm
m
km
1
m3
kcal
run
kcal/kg
kcal/m3
kcal/nm
Gcal/day
Meal/day
Meal A^
nm /day
106 nm3/day
(Mnm3/day)
ran /Gcal
kg/Gcal
kg/m3
kg/cm
m3/hr
m3/hr
W
km
kn/hr
kilograms
metric tons (1000 kg)
kilograms
centimetres
metres
kilometres
litres (1000 litres = 1 m3)
cubic metres
kilocalories
normal cubic metres
kilocalories/kilogram
kilocalories/cubic metre
kilocalories/normal cubic metre
gigacalories/day
megacalories/day
megacalories/hour
normal cubic metres/day
million normal cubic metres/day
(mega normal cubic metres/day)
normal cubic metres/gigacalorie
kilograms/gigacalorie
kilograms/cubic metre
kilograms/square centimetre
cubic metres/hour
cubic metres/hour
watts
kilometres
ki lometres/hour
(a) h SCF of gas is measured at 60ฐF and atmospheric pressure, and a
run3 of gas is measured at 0ฐC and atmospheric pressure.
(b) Exponential
,3
jnqlish
SI Metric
10-
10;
10
10
12
M or thousand
MM or million
billion (U.S.)
billion (U.K.)
k or kilo
M or mega
G or giga
T or tera
79
-------
TECHNICAL REPORT DATA
(/'lease read ImUructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-066
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Treatment of Refinery Wastewater Using a Filtration-
Activated Carbon System
5. REPORT DATE
March 1979 (issuing date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Bruce A. McCrodden
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
1BB610C
BP Oil Inc.
Marcus Hook, PA
19061
11. CONTRACT/GRANT NO.
12050 GXF
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Ada, Oklahoma 74820
13. TYPE OF REPORT AND PERIOD COVERED
Final 10/73 - 6/78
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The objective of this report was to demonstrate the applicability
of a dual media filtration-activated carbon adsorption system for the
treatment of petroleum refinery wastewater. Constructed on a one-quarter
acre plot, the capital cost of the wastewater treatment plant was $1,812,000
with an annual operating cost of $223,980.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Petroleum Refining
Waste Treatment
Industrial Wastes
Activated Carbon
Regeneration (engineering)
Wastewater Management
68 D
91 A
3. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
88
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
80
#U.S. GOVERNMENT PRINTING OFFICE: 1979-657-060/1629 Region No. SHI
------- |