EPA-650/2-75-057-C
July 1975
Environmental Protection Technology Series
SURVEY OF FLUE GAS
DESULFURIZATION SYSTEMS
PHILLIPS POWER STATION, DUQUESNE LIGHT CO.
Ul
CD
-------
EPA-650/2-75-057-C
SURVEY OF FLUE GAS
DESULFURIZATION SYSTEMS
PHILLIPS POWER STATION, DUQUESNE LIGHT CO,
by
Gerald A. Isaacs
PEDCo-Environmental Specialists, Inc.
Suite 13
Atkinson Square
Cincinnati, Ohio 45246
Contract No. 68-02-1321, Task 6c
ROAP No. 21ACX-130
Program Element No. 1AB013
EPA Project Officer: Norman Kaplan
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
WASHINGTON,D.C. 20460
July 1975
-------
EPA REVIEW NOTICE
Tins report has been reviewed by tin.- National Environmental Research
Center Research Triangle. Park, Office of Research and.Development,
EPA. and approved"'for publication. Approval does not signi iy^that. the
con lent;-; necessarily-reflect the views and policies of the Environmental
Pre'eclion Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Rest arch reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
* ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This reporl has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY scries. This series describes research performed to
develop and demonstrate instrumentation, equipment and methodology
to repair or prevent environmental degradation from point and non-
point sources of pollution. This work'provides the new or improved
technology required for the control and treatment of pollution sources
;u meet environmental quality standards.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
Publication No. EPA~650/2-75-057-c
u
-------
ACKNOWLEDGMENT
This report was prepared under the direction of Mr.
Timothy W. Devitt. The principal author was Dr. Gerald A.
Isaacs.
Mr. Wade H. Ponder, former EPA Project Officer, had
primary responsibility within EPA for this project report.
Information and data on plant operation were provided by Mr,
S. L. Pernick, Duquesne Light Company, and Mr. Paul Chopra,
Chemico, Inc., during and subsequent to the survey visit.
The author appreciates the efforts and cooperation of
everyone who participated in the preparation of this report.
111
-------
TABLE OF CONTENTS
Page
ACKNOWLEDGMENT ii
LIST OF FIGURES iv
LIST OF TABLES iv
SUMMARY v
1.0 INTRODUCTION 1-1
2.0 FACILITY DESCRIPTION 2-1
2.1 Plant Location 2-1
2.2 Boiler Data 2-1
2.3 Pollution Controls 2-2
3.0 FLUE GAS DESULFURIZATION SYSTEM 3-1
3.1 Process Description 3-1
3.2 Installation Schedule 3-7
3.3 Cost Data 3-9
4.0 FGD SYSTEM PERFORMANCE ANALYSIS 4-1
4.1 Start-up Problems and Solutions 4-1
4.2 Performance Test Run 4-3
4.3 Performance Parameters 4-4
4.4 Process Modification for Future 4-5
Installations
APPENDIX A PLANT SURVEY FORM A-l
APPENDIX B PLANT PHOTOGRAPHS B-l
APPENDIX C OPERATIONAL REPORTS 'C-l
-------
LIST OF FIGURES
Figure
3.1
General Flow Diagram of the FGD System
at the Phillips Power Station
Page
3-3
LIST OF TABLES
Table
2.1
3.1
3.2
Data on Plant Design, Operation and
Emissions Duquesne Light Company -
Phillips Power Station
Summary of Scrubber Data - Phillips
Power Station
Summary of Tank Data - Phillips Power
Station
2-4
3-8
3-8
VI
-------
SUMMARY
Four parallel Chemico venturi scrubber trains have been
installed at the 400 MW Phillips Power Station of Duquesne
Light Company.
Each of the four trains has a first stage venturi for
S02 and particulate removal; one of the four trains includes
a second stage venturi absorber for the additional control
of S02 emissions. The system uses slaked lime as the
absorbing medium and has been operational since July 1973.
An overall SO2 removal efficiency for the plant of 50-60
percent has been attained. The two stage train has achieved
approximately 90 percent S02 removal efficiency. The use of
magnesium oxide-modified lime is being investigated in an
effort to increase the S02 removal efficiency of the existing
system.
Numerous start-up and operational problems have been
encountered and corrected. A ten-month evaluation of the
system was initiated in December 1974 to determine whether
additional SO2 removal equipment will be installed.
Pertinent data on the facility and FGD system operation
are summarized in the following table.
vn
-------
SUMMARY OF FGD DATA - PHILLIPS
Station Rating
Fuel Characteristics
Prime A/E
FGD Vendor
Process
New or Retrofit
Start-up Date
FGD Modules
Efficiency, overall
Particulate
so2
Water Make-up
Sludge Disposal
Unit Cost
373 MW (net) after scrubber derating
Coal: 11,350 BTU/lb, 18.2% ash,
2.15% sulfur
Gibbs and Hill, Inc.
Chemico
Wet lime
Retrofit
August 1973 (SC<2 absorber)
Four first stage; one second stage
99 percent
50-60 percent
1.7 gpm/MW (net)
Stabilization and haul to landfill
$38 million ($102/net KW) capitala
$15 million/yr (5.9 mills/net KWH)
operating including indirect charges
on capital
An estimated $15 million that may be required for sludge
disposal facilities are not included in this cost.
Vlll
-------
1.0 INTRODUCTION
The Control Systems Laboratory of the U.S. Environmental
Protection Agency has initiated a study to evaluate the
performance characteristics and degree of reliability of
flue gas desulfurization (FGD) systems on coal-fired utility
boilers in the United States. This report on the Phillips
Power Station of Duquesne Light Company is one of a series
of reports on such systems. It presents values of key
process design and operating parameters, describes the major
start-up and operational problems encountered at the facility
and the measures taken to alleviate such problems, and
identifies the total installed and annualized operating
costs.
This report is based upon information obtained during a
plant inspection on August 22, 1974, and on subsequent data
provided by Duquesne Light Company and Chemico personnel.
Section 2.0 presents pertinent data on facility design
and operation including actual and allowable particulate and
S02 emission rates. Section 3.0 describes the flue gas
desulfurization system and Section 4.0 analyzes FGD system
performance.
1-1
-------
2.0 FACILITY DESCRIPTION
2.1 PLANT LOCATION
The Phillips Power Station of the Duquesne Light Company
is located on the Ohio River in Allegheny County, Pennsylvania,
20 miles northwest of Pittsburgh, Pennsylvania. The area is
heavily industrialized, and the plant supplies electricity
to residential, commercial, and industrial users in Allegheny
and Beaver Counties.
2.2 BOILER DATA
Six generating units at the station constitute a total
nominal generating capacity of 400 MW. The net station
capacity will be 373 MW with four scrubber modules and four
absorber modules installed. At the present time only one
absorber module has been installed. All the generators are
cycling base load units. Unit No. 6 is the largest generator,
having a net capacity of 143 MW. All the boilers were
manufactured by Foster-Wheeler and are dry-bottom, pulverized-
coal-fired units. The first unit was installed in 1942, and
the sixth one in 1956. Coal at this station has an average,
as received, gross heating value of 11,350 BTU/lb. Ash
content, on a dry basis, is 18.2 percent, and sulfur content
is 2.15 percent.
2-1
-------
2.3 POLLUTION CONTROLS
Primary particulate emission control at this plant is
provided by Research-Cottrell mechanical collectors followed
by a Research-Cottrell electrostatic precipitator (ESP) on
each boiler. The collection efficiency for this system was
originally estimated to be 90 percent. However, disruption
of the gas flow distribution through the ESP resulted from
the design of the ducts which tie in to the FGD system so
that apparently the combined efficiency through the mechanical
and electrostatic system is only about 80 percent.
Final particulate control is achieved by four parallel
Chemico venturi scrubbers with a design efficiency of 99
percent. Outlet grain loadings from the scrubbers have been
measured to be less than 0.04 grains per standard cubic
foot. These scrubbers also remove about 50 percent of the
S0_ entering them when lime is added (for pH control).
SO- emissions from one scrubber are further controlled
by a Chemico second-stage venturi absorber recirculating a
lime solution. S02 removal efficiency for this two-stage
train has averaged about 90 percent based on performance
guarantee tests. Overall SO- removal efficiency is there-
fore about 50-60 percent when pH can be maintained between
6-7.
At the present time the FGD absorber facilities consist
of a two stage absorber module processing about 125 MW
equivalent of flue gas and three single stage absorber modules
2-2
-------
processing gas from the balance of the station's nominal 400
MW capacity. Present plans provide for the installation of
up to three more FGD absorbers if they are found to be
necessary.
The scrubber and absorber facilities derate the station
capacity by 14 MW. This effects an increase in station heat
rate from a previous level of 11,500 BTU/KWH (net) to about
11,900 BTU/KWH (net).
Table 2.1 presents pertinent data on plant design,
operation, and atmospheric emissions.
2-3
-------
Table 2.1 DATA ON PLANT DESIGN,
OPERATION AND EMISSIONS
DUQUESNE LIGHT COMPANY - PHILLIPS POWER PLANT
Station Data
Item
Maximum theoretical continuous generating
capacity, MW (net) with 4 scrubbers and
1 absorber operating
Average capacity factor (1974), %
Boiler manufacturer
Year placed in service
Station heat rate, BTU/KWH (net)
Maximum heat input MM BTU/hr
Stack height above grade, ft
Maximum flue gas rate, acfm @ 360°F
Emissions Controls:
Particulates
S02
Particulate Emission Rate:
Allowable, Ib/MM BTU
Actual, Ib/MM BTU
S02 Emissions Rate:
Allowable, Ib/MM BTU
Actual, Ib/MM BTU
373
70
Foster-Wheeler
1942-1957
11,900
4463
340
1,850,000
Multicyclones ESP
Venturi Scrubber
Venturi Absorber
0.08
0.08 est
0.6
1.9 est
2-4
-------
3.0 FLUE GAS DESULFURIZATION SYSTEM
3.1 PROCESS DESCRIPTION3
Lime to be introduced to the FGD system at Phillips is
initially activated by slaking it to form calcium hydroxide.
This slaked lime, along with the recirculation calcium
sulfite-rich slurry constitute the scrubbing liquor. Reaction
with SO2 in the flue gas takes place in the liquid phase.
Following are the principal chemical reactions in the lime-
base flue gas desulfurization processes:
Lime slaking: CaO + H20 *Ca(OH)2 (1)
Sulfite formation: Ca(OH)2 + S02 *• CaSO- + H2O (2)
Bisulfite formation: CaS03 + S02 + H20-»-Ca (HS03) 2 (3)
Sulfate formation: 2CaS03 + 02 >• 2CaS04 (4)
Formation of sulfate (reaction 4) in the absorber module is
undesirable and should be suppressed.
The Phillips FGD system may ultimately consists of four
identical parallel trains, each train consisting of a particulate
scrubber module and a S02 absorber module in series. At the
present time, all six boilers have been tied in to the
Adapted from "Duquesne Light Company, Phillips Power Station
Lime Scrubbing Facility," Steve L. Pernick, Jr. and R. Gordon
Knight, Presented at EPA FGD Symposium, Atlanta, November
4-7, 1974, and supplemented with information obtained during
plant visit. See Appendices A and C.
3-1
-------
scrubbing system; only one of the four scrubbers has a
second-stage absorber installed in conjunction with it for
the removal of S02. The exhaust streams of all six boilers
at the station are manifolded together so that the components
of the emission control system are not directly associated
with specific boilers. This system allows maximum flexibility
for emission control when various components of the FGD
system must be shut down for maintenance or repair. There
is also provision to bypass the whole FGD system under
emergency conditions. However, in order to accomplish the
bypass it is necessary to shut down the boilers and to
remove blank-off plates from the exhaust gas stream. Sixteen-
foot diameter ducts lead to each of the four first-stage
scrubbers.
Figure 3.1 is a schematic flow diagram for this installa-
tion. Hot flue gas (about 500,000 acfm at 340°F) containing
fly ash and sulfur dioxide (about 1400 ppm SO-) enters the
first-stage scrubber and impinges upon the upper cone. One-
half (8250 gpm) of the total scrubber recycle liquor for
each module is introduced into the vessel through the bull
nozzle where it is sprayed over the surface of the upper
cone, to initiate the scrubber action. The rest of the
scrubber recycle liquor enters through tangential nozzles
above the adjustable throat damper. The flue gas and scrubbing
liquor are vigorously contacted in the throat section of the
scrubber where the particulates and some S02 are removed.
3-2
-------
GJ
I
U)
HOT FLUE GAS 400.000cfm 340°Ft
FROM BOILERS
16,500 gpm.
120°F
1st-STAGE •
^SCRUBBER
COLD FLUE GAS
DEMISTER SPRAY
2nd-STAGE
ABSORBER
FUEL
OIL COMBUSTION
I lAIR
MAKE-UP FROM
OVERFLOW PUMPS»
BLEED TO
, THICKENER
AND CLARIFIER
TO THICKENER TROUGH
V
sV
IBRA
LIME
STORAGE
SILO
_BY_P_AS_S
DEGRITTER
j r
WEIGH
FEEDER
y^LAKER
/TRANSFER
** TANK
LIME SLAKER
LIME SLURRY
PUMP (2)
Figure 3.1. General flow diagram of the FGD system
at the Phillips Power Station.
-------
The gas and liquor continue downward to the separator section
where, after being separated from the scrubbing liquor, the
flue gas enters a spray-washed demister where entrained
liquor is removed. The gas leaves the scrubber and enters a
booster ID fan which is equipped with fresh water sprays to
remove any accumulation of solids resulting from scrubber
carryover.
The ID fan housings are lined with 1/4-inch thick
natural rubber. Wheel material is Carpenter 20 Cb 3, a
stainless steel containing niobium and tantalum. The fan
shaft is 316L stainless steel. Each fan is driven by a 5500
horsepower, 1200 rpm, 4160 volt electric, motor. A closed
system supplies cooling water to the fan bearings.
Outlet gas temperature from the first stage scrubber is
monitored and is normally in the 110-120°F range. At 175°F
a control valve automatically opens to admit emergency
cooling water to the upper cone. Additional temperature
rise automatically shuts down the fan and closes the isola-
tion dampers.
The gas leaving the ID fan enters one of two vessels
depending on the train. In the case of the three single-
stage scrubber trains, the gas enters an entrainment sepa-
rator in which entrained scrubbing liquor that has been
diluted by the fan sprays is separated and collected. These
separators will be replaced by second-stage absorber vessels
where the test program indicates that additional absorbers
are feasible. In the case of the S02 prototype train, gas
3-4
-------
leaving the ID fan enters a second stage absorber vessel,
which is identical in size and design to the first stage
scrubber. This second vessel is equipped with lime slurry
injection nozzles.
Lime is fed from a storage silo at a controlled rate to
a lime slaker where it is mixed with fresh make-up water.
The slaked lime overflows to a slaker transfer tank where
make-up water is added to provide a constant flow of lime
slurry with a 15-percent solids concentration. Lime composi-
tion is 95 percent pure CaO, balance inert. Design lime
requirement is 130 percent of the stoichiometric requirement.
At a station heat rate of 11,900 BTU/KWH, a coal gross heat
content of 11,350 BTU/lb and a sulfur content of 2.1 percent,
design lime consumption for the single train is calculated
to be about 4650 Ib/hr.
The scrubbing liquor is circulated by two recycle pumps
(Carpenter 20 Cb 3). Some of the recycle liquor is bled to
the first-stage scrubber and part of the first-stage recycle
stream is bled to the thickeners.
Total recycle flow is about 16,500 gpm per stage with a
615 gpm bleed to the thickeners. All single-stage vessels
have been fitted with reagent nozzles so that pH control can
be maintained and so that an approximate SO2 removal efficiency
of 50 percent can be accomplished.
The exit gases leaving either the second stage absorber
or the entrainment separator enter a common wet duct lined
with Flakeline 103 (a product of the Ceilcote Company consisting
3-5
-------
of a glass-flake-filled modified polyester resin, subsequently
referred to as Ceilcote) which leads to a new 340 ft, acid-
resistant, brick-lined, concrete stack. A 316L stainless
steel section of the duct, preceding the stack is equipped
with a direct oil-fired reheater unit that can raise stack
gas temperatures as much as 30°F. Normal reheat is about
20°F.
Bleed streams from the absorber and the scrubbers enter
a trough which feeds two 75-foot diameter thickeners. The
overflow from these thickeners runs to an overflow tank
where it is pumped by two, 2000 gpm 316L stainless steel
pumps to the make-up line for return to the scrubber system.
Underflow from each thickener, 105 gpm, at 40 to 45
percent solids concentration, is pumped by a 15 horsepower
pump to one of three clay-lined sludge holding ponds. Each
pond has a capacity of about 6500 cubic yards. Just prior
to discharge to the holding ponds, the sludge enters a
mixing tank where a stabilizing agent is added at a predetermined
rate based on the density of the sludge as it leaves the
thickener.
Settling and curing take place in the pond. Supernatant
liquid reenters the scrubber system via the thickeners.
Overflow from the thickeners is pumped back to the make-up
line for return to the scrubber system. The sludge curing
ponds provide interim storage while stabilization is taking
place. One pond receives thickener underflow while a second
is curing and the third is being excavated for final off-
site disposal.
3-6
-------
A service (river) water system independent from the
power station service water system is provided. It includes
a pair of 900 gpm, 100 horsepower pumps taking suction from
the existing condenser discharge tunnel and a 10-inch distri-
bution header for the scrubber facility. It provides fan
spray water, pump seal water, demister spray water, instrumenta-
tion flush water, reagent mixing tank water, and emergency
water for the scrubbers.
With present loads on the FGD system an excess of water
is entering the system so that closed-loop operation is not
possible. Excess water is discharged to the bottom ash
settling ponds and from these into the river. An early
attempt to operate in closed-loop mode resulted in a buildup
in chloride concentration to 4000 ppm, causing significant
corrosion damage to Carpenter 20 steel fan parts, scrubber
internals, and dampers. Operating parameters are summarized
in Tables 3.1 and 3.2.
3.2 INSTALLATION SCHEDULE
In December 1969 Gibbs and Hill, Inc. was retained to
conduct a comprehensive study of the most feasible means of
complying with current and anticipated emission regulations
applicable to the Phillips and Elrama Power Stations of
Duquesne Light Company. In September 1970 Gibbs and Hill
submitted the results of that study indicating that stack
gas desulfurization was the most practical means of achieving
compliance at both stations. The consultant also concluded
that a dual-stage venturi scrubbing system should be pilot
3-7
-------
Table 3.1 SUMMARY OF SCRUBBER DATA
PHILLIPS POWER STATION
Particulate
Scrubber
Modules
FGD
Scrubber
Modules
Superficial gas velocity, ft/sec
Module size
Equipment internals
Material of Construction
Shell
Internals
40
40'diam x 50'h
Venturi
Mild steel,
Ceilcoted
Some 316L
stainless,
Ceilcoted
40
45'diam x 50'h
Venturi
Mild steel,
Ceilcoted
Some 316L
stainless,
Ceilcoted
Table 3.2 SUMMARY OF TANK DATA
PHILLIPS POWER STATION
Item
No. of units
Unit size
Retention time
at full load
Temperature °F
PH
Solids Cone. %
Materials of
Construction
Scrubber
Recirculating
Tank
20'diam x 30'h
1.2 minutes
120°F
5
mild steel
Thickener
70' diam x 40'
ambient
4-6
40-45
(Underflow)
mild steel,
Celicoted
3-8
-------
tested. Pilot scrubber specifications were prepared and
issued to five vendors in September 1970. The successful
vendor, Chemico, Inc., installed a 1500 cfm pilot unit in
February 1971 and completed tests in May of that year. In
July 1971 Chemico was awarded a contract to build a full
scale system consisting of four particulate scrubbers and a
second-stage SO- absorber connected to one of the four
scrubbers along with lime injection, recycle pump and reheater
equipment. Gibbs and Hill, Inc. the Architect/Engineer firm
for the project was responsible for the remaining design of
the system. This construction was completed in July 1973 to
a point where start-up was possible. The single SO_ absorber
is to serve as a basis for decisions concerning the installation
of additional S02 facilities at the Phillips and Elrama
Stations. Five particulate scrubbers have been installed at
the Elrama Station.
3.3 COST DATA
As of April 1, 1975, the system at Phillips represents
a capital investment of approximately $33 million. It is
expected that an additional $7 million may be required for
additional equipment necessary to comply with the S02
emission limitations. This cost includes all process
equipment, engineering costs, contractors' fees, and interest
on capital during construction. This will represent a total
investment of $40 million or approximately $103 per net
kilowatt, exclusive of the costs of additional sludge disposal
facilities that may be required.
3-9
-------
A recent estimate indicates that the annual operating
and maintenance expenses at Phillips, excluding additional
sludge disposal requirements, but including fixed charges
estimated at 14.5 percent of capital costs, with full SOj
scrubbing will be about $15 million per year, or 5.9 mills
per kilowatt hour.
In an effort to estimate the total capital and operating
costs that would utimately be incurred for the scrubber
system at the Phillips Station, including provision of
adequate sludge disposal facilities for full SO- removal, a
preliminary evaluation was made of the additional property
required for development into a disposal area and the
additional expenses involved in disposing of the sludge.
This estimate indicates total capital expenditures approxi-
mately $147 per KW.
Present estimates of sludge disposal costs are $5
to $7 per wet ton of sludge (40 percent solids).
3-10
-------
4.0 FGD SYSTEM PERFORMANCE ANALYSIS
4.1 START-UP PROBLEMS AND SOLUTIONS
The initial schedule for this installation called' for
start-up of the FGD system on January 1, 1973, to be fol7
lowed by a two-month shakedown period and a ten-month test
period to evaluate the feasibility of the system, so that by
January 1974 decisions could be made regarding additional
equipment installations at the Phillips and Elrama Power
Station.
Actual start-up lagged the target date by about six
months, and the ten-month test program initiation was de-
layed nearly two years. A detailed description of the
causes for these delays has been published and is included
in this report as Appendix C. Problems and solutions are
outlined briefly in the following paragraphs.
The six month start-up delay was attributed to several
factors. Some engineering delays occurred as a result of
efforts to incorporate ongoing experience gained at other
Chemico FGD installations. Equipment delivery delays and
fabrication errors also occurred. Some defective equipment,
such as fans, had to be field repaired. Failures of elec-
trical switchgear and transformers occurred during field
tests preliminary to start-up.
4-1
-------
The scrubber system was put in operation on July 9,
1973, although the lime additive system was not ready at
that time. Operation without lime resulted in excessive
scrubber liquid acidity that quickly corroded some com-
ponents of the FGD system, including pumps and piping. The
system operated for two weeks before it was shut down for a
two-week repair period due to a leak in the stack seal. On
August 5, 1973, the system was restarted with the lime feed
system operational, and ran until October 8, 1973 when it
was shut down until March 17, 1974.
Scale deposits developed in the scrubbers, and the
scale had to be removed periodically. Corroded fan dampers
and scrubber cones were Ceilcoted. Exhaust gas expansion
joints were replaced, fan welds were reground, and stack
mortar joints were repaired. Many parts were found to be
constructed of off-spec 304 stainless and had to be replaced
with originally specified 316L material.
An early attempt to operate the system in a closed-loop
mode resulted in a buildup of chloride ion concentration in
the scrubber recirculating system to about 4000 ppm, in-
tensifying the existing corrosion problems and causing
additional ones.
Installation of the ductwork to the FGD system caused
a flow disruption through the existing ESP so that the
particulate efficiency of the unit was reduced considerably.
This resulted in excessive fly ash carryover into the scrubber
and the consequent generation of excessive quantities of
4-2
-------
sludge. Additional sludge was also created by treating the
single stage scrubber trains with lime. As a result the
spare thickener capacity had to be utilized constantly, and
the sludge holding pond capacity was reduced to about a 10-
day fill period as compared to a 28 day design period. This
situation plus the high outage rate for the stabilizing
agent injection equipment often made it necessary to transport
sludge by truck to the disposal area in a fluid state.
4.2 PERFORMANCE TEST RUN
A ten-month test program was initiated in December,
1974. At that time five of the six boilers at the Phillips
Plant were connected to the scrubber system. This program
will evaluate the efficiency, reliability, and practica-
bility of the system. A decision should then be made late
in 1975, concerning the installation of additional SO-
removal equipment for compliance with the SO- emission
limitations. The number of additional second stage ab-
sorbers required for compliance with the regulations will be
determined during the ten-month test program, and this will
establish the completion date for installation of all addi-
tional FGD equipment.
In the hope of obtaining higher efficiencies of SO-
removal with the use of lime, Duquesne Light is conducting
tests using a lime containing 4 to 6 percent magnesium
oxide. Various tests on pilot systems indicate that the
higher reactivity of such a lime results in SO2 single-stage
removal efficiencies above 80 percent. In preparation for
4-3
-------
these tests, one of the single-stage scrubbers was isolated
along with one of the thickener tanks. This isolation was
necessary to establish and maintain the necessary conditions
in the test scrubber and one thickener. A test run with the
high MgO was made in September, 1974. However, it has been
difficult to achieve a sufficiently high rate of utilization
of the magnesium oxide. A few cursory tests under these
conditions indicate SO2 removal efficiencies somewhat better
than those with high calcium lime, but also significantly
less than anticipated. The tests were suspended in Septem-
ber to further evaluate means of increasing the reactivity
and utilization of the magnesium lime. Additional tests
with lime in early November showed that better pH control
could be obtained if the lime slurry was heated. It is
believed that this reagent holds promise, but the tests were
not conclusive because of inconsistencies in the quality of
the lime supplies and the difficulty in blending and obtain-
ing sufficient supplies.
4.3 PERFORMANCE PARAMETERS
The ultimate S0_ removal efficiency needed to comply
with emission regulations is about 83 percent. However, the
present equipment installed using only one two-stage system
and three single-stage scrubber trains with lime is pro-
ducing an estimated SO_ removal efficiency of 50-60 percent.
Availability, was reported to be 100 percent in January
1975, but this figure is misleading since there is con-
siderable scrubber redundancy in the system at the present
4-4
-------
time. (Percent availability is defined as 100 times the
ratio of system operating hours to boiler operating hours).
4.4 PROCESS MODIFICATIONS FOR FUTURE INSTALLATIONS
The design of additional installations will rely
heavily on operating experience gained from this system.
The discussion in Section 4.1 points out many areas where
system improvements can be made, especially regarding
material selection, quality control, and equipment sizing.
Ongoing operational experience during the ten-month test
program will help to optimize process design parameters for
future control systems, if such systems are indicated.
4-5
-------
APPENDIX A
PLANT SURVEY FORM
A-l
-------
PLANT SURVEY FORM
NON-REGENERABLE FGD PROCESSES
A. COMPANY AND PLANT INFORMATION
1. COMPANY NAME Duquesno Light Co.
2. MAIN OFFICE 4^5 fiixt-h Avpniie, Pittsburgh, Pa.
3. PLANT MANAGER T.S. Ahhoht
4. PLANT NAME Phillips Power Station
5. PLANT LOCATION Cresant Township. Allegheny County. Pa.
6. PERSON TO CONTACT FOR FURTHER INFORMATION S.L. Pernick. Jr.
7. POSITION Manager Environ. Affairs
8. TELEPHONE NUMBER (412) 471-4300
9. DATE INFORMATION GATHERED
10. PARTICIPANTS IN MEETING AFFILIATION
Wade H. Ponder EPA
John Busik EPA
Tim W. Devitt PEDCo
Fouad K. Zada PEDCo
Robert D. O'Hara D. L. Co,
I. S. Abbott D. L. Co.
R. G. Knight D. L. Co.
A-2
5/17/74
-------
B. PLANT DATA. (APPLIES TO ALL BOILERS AT THE PLANT).
C.
CAPACITY, MW (Gross)
SERVICE (BASE, PEAK)
FGD SYSTEM USED
BOILER NO.
1 2
35 35
P-B P-B
Wet Gc
3
65
P-B
is Ventui
4
65
P-B
i Scrubb
5
65
P-B
2r
6
148
P-b
BOILER DATA. COMPLETE SECTIONS (C) THROUGH (R) FOR EACH
BOILER HAVING AN FGD SYSTEM.
1. BOILER IDENTIFICATION NO.
2. MAXIMUM CONTINUOUS HEAT INPUT
1-6
4463
MM BTU/HR
3. MAXIMUM CONTINUOUS GENERATING CAPACITY 387 MW (net)
4. MAXIMUM CONTINUOUS FLUE GAS RATE. 2.300.000 ACFM @ 380 °F
5. BOILER MANUFACTURER Foster-Wheeler
6. YEAR BOILER PLACED IN SERVICE 1942-1956
7. BOILER SERVICE (BASE LOAD, PEAK, ETC.)
8. STACK HEIGHT 340' Scrubber Stack
9. BOILER OPERATION HOURS/YEAR (197 )
10. STATION CAPACITY FACTOR* 70%
11. RATIO OF FLY ASH/BOTTOM ASH 85
* DEFINED AS: Kw" GENERATED TN YEAR
MAX. CONT. GENERATED CAPACITY IN KW x 8760 HR/YR
A-3
5/17/74
-------
D. FUEL DATA
1. COAL ANALYSIS (as received)
GHV (BTU/LB.)
S %
ASH %
MAX.
12600
2.8
23
MIN.
10700
1.0
9.5
AVG.
11350
2.15
18.2
2. FUEL OIL ANALYSIS (exclude start-up fuel)
GRADE
S % .
ASH %
N/A
E. ATMOSPHERIC EMISSIONS
1. APPLICABLE EMISSION REGULATIONS
a) CURRENT REQUIREMENTS
AQCR PRIORITY CLASSIFICATION
REGULATION & SECTION NO.
MAX. ALLOWABLE EMISSIONS
LBS/MM BTU
b) FUTURE REQUIREMENTS,
COMPLIANCE DATE
REGULATION & SECTION NO.
MAXIMUM ALLOWABLE EMISSIONS
LBS/MM BTU
PARTI CULA'iES
0.081b/10 BTU
Same as abov
SO-
0.6 lb/106BTU
Article 18 Allegheny County
County Health Department
Same as above
Same as above
Same as abovs
PLANT PROGRAM FOR PARTICULATES COMPLIANCE
Connect all six boilers to common header to four particulate
scrubbers
3. PLANT PROGRAM FOR S02 COMPLIANCE Operate 10 months for SO2
prototype module including single stage scrubbers with lime.
Based on outcome of tests determine how many additional
second stage scrubbers required to come into compliance.
Roughly assume 28 months for second scrubber, 19 months for
third scrubber, 18 months for fourth scrubber.
A~4 5/17/74
-------
F. PARTICIPATE REMOVAL
1. TYPE
MANUFACTURER
EFFICIENCY: DESIGN/ACTUAL
MAX. EMISSION RATE* LB/IIR
GR/SCF
LB/MMBTU
MECH.
p-r
E.S.P.
R-C
FGD
Chemico
99/N.A.
0.08
DESIGN BASIS, SULFUR CONTENT
2.3%
G. DESULFURIZATION SYSTEM DATA
1. PROCESS NAME
2. LICENSOR/DESIGNER NAME:
ADDRESS:
PERSON TO CONTACT:
TELEPHONE NO.:
Lime scrubbing
Chemico Co.
1 Penn Plaza NY, NY 10001
Paul Chopra
(212) 239-5100
3. ARCHITECTURAL/ENGINEERS, NAME: Gibbs and Hill, Inc.
ADDRESS: 393 Seventh Avenue NY, NY 10001
PERSON TO CONTACT:
TELEPHONE NO.:
Mr. Joseph V. Armao
PROJECT CONSTRUCTION SCHEDULE:
DATE
a) DATE OF PREPARATION OF BIDS SPECS. Pilot scrubber, 9/70
b) DATE OF REQUEST FOR BIDS
c) DATE OF CONTRACT AWARD 12/70 G&H 7/71 Chemico
d) DATE ON SITE CONSTRUCTION BEGAN 7/71 Phase I
10/70 pilot, 3/71 full
scale
c) DATE ON SITE CONSTRUCTION COMPLETED 7/73
f) DATE OF INITIAL STARTUP July 1973
g) DATE OF COMPLETION OF SHAKEDOWN
*At Max. Continuous Capacity
A-5
In Progress
5/37/74
Phase I
-------
6
7
8
LIST MAJOR DELAYS IN CONSTRUCTION SCHEDULE AND CAUSES:
Equipment delivery, Electrical, Duct Work, Fans
I.D. Fan defects in welds
Lack of space to install equipment restricted workers
Boiler house ductwork erection, method of handling at
site difficult
Selection and development of sludge disposal techniques,
NUMBER OF S02 SCRUBBER TRAINS USED 1
DESIGN THROUGHPUT PER TRAIN, ACFM @ 340 °F 547,000
DRAWINGS: 1) PROCESS FLOW DIAGRAM AND MATERIAL BALANCE
See G & Hill or Chemico
2) EQUIPMENT LAYOUT
H.
S02 SCRUBBING AGENT
1
2
3
4
TYPE
SOURCES OF SUPPLY
Lime
See Dravo Corp. Mr. Win.
Lord Neville Island Pitts,
Pa.
CHEMICAL COMPOSITION (for each source)
SILICATES
SILICA
CALCIUM OXIDE
MAGNESIUM OXIDE
EXCESS SCRUBBING AGENT USED ABOVE
STOICHIOMETRIC REQUIREMENTS
5. MAKE-UP WATER POINT OF ADDITION
6. MAKE-UP ALKALI POINT OF ADDITION
30%
Scrubber Recycle
Scrubber Vessel Basins
A-6
5/17/74
-------
J . SC RBERT RAI N SPEC 1FICATIONS
1. SCRUBBER NO. 1
TYPE (TOWER/VENTURI) Veniuir i_. .
L1QUJD/GAS RATIO, G/MCF @ 120 °F ____
GAS VELOCITY THROUGH SCRUBBER, FT/SEC '
MATERIAL OF CONSTRUCTION Carbon Steel
TYPE OF L1NJNG Ceilcote
INTERNALS:
TYPE (FLOATING BED, MARBLE BED, ETC.) Variable Throat Vonturi
NUMUKR OF STAGES 1
TYPK AND SIZE OF PACKING MATERIAL
PACKING THICKNESS PKR STAGE*b)
MATERIAL OF CONSTRUCTION, PACKING:
SU P PORTS:
2. SCRUBBER NO. 2 ^
TYl'K (TOWER/VENTURI) Same as 1
LIQUID/GAS RATIO. G/MCF @ °F
GAS VELOCITY THROUGH SCRUBBER, FT/SEC
MATERIAL OF CONSTRUCTION
TYPE OF LINING
INTERNALS:
TYPE (FLOATING BED, MARBLE BED, ETC.)
NUMBER OF STAGES
TYI'i: AND SIXK OF PACKING MATERIAL
a) Scrubber No. 1 is the scrubber that the flue qascs first
enter. Scrubber 2 (if applicable) follows Scrubber No. 1.
b) For floating bed, packing thickness at rest.
A-7
VI-7/7/1
-------
PACKING THICKNESS PER STAGE ^
MATERIAL OF CONSTRUCTION, PACKING:
SUPPORTS:__.
3. CLEAR WATER TRAY (AT TOP OF SCRUBBER)
TYPE N_/A_
L/C RATIO _
SOURCE OF WATER
UEMJSTER
TYPE (CHEVRON, ETC.) Chevron
NUMBER OF PASSES (STAGES)
SPACE BETWEEN VANES
ANGLE OF VANES
TOTAL DEPTH OF DEMISTER
DIAMETER OF DEMISTER
DISTANCE BETWEEN TOP OF PACKING
AND BOTTOM OF DEMISTER
POSITION (HORIZONTAL, VERTICAL)
MATERLAL OF CONSTRUCTION
METHOD OF CLEANING Internal Auto Sprays
SOURCE OF WATER AND PRESSURE
FLOW RATE DURING CLEANINGS, GPM
FREQUENCY AND DURATION OF CLEANING
REMARKS
5. REHIIATER
TYPE (DIRECT, INDJKECT) Direct OJ1 Fired
b) For floating bed, packing thickness at rest.
A-8 5/17/74
-------
DUTY, MMBTU/HR
HEAT TRANSFER SURFACE AREA SQ.FT
TEMPERATURE OF GAS: IN OUT
HEATING MEDIUM SOURCE No. 2 Oil
TEMPERATURE & PRESSURE
FLOW RATE 7.3qal/min LB/ilR
REIIEATER TUBES, TYPE AND
MATERIAL OF CONSTRUCTION
REHEATER LOCATION WITH RESPECT TO DEMISTER Downstream
of Demister prior to scrubber stack entrance
METHOD OF CLEANING
FREQUENCY AND DURATION OF CLEANING
FLOW RATE OF CLEANING MEDIUM LB/HR
REMARKS Problems with oil pumps, burners and temperature
control so far very little operation
6. SCRUBBER TRAIN PRESSURE DROP DATA INCHES OF WATER
PARTICULATE SCRUBBER
S02 SCRUBBER
CLEAR WATER TRAY
DEMISTER 2"
REHEATER
DUCTWORK
TOTAL FGD SYSTEM
5/17/74
-------
7.
FRESH WATER MAKE UP FLOW RATES AND POINTS OF ADDITION
TO: DEMISTER Service Water fClarifier)
QUENCH CHAMBER Service Water (Clarifier)
ALKALI SLURRYING Service Water (Clarifier)
PUMP SEALS Service Water (Clarifier)
OTHER
Service Water (Clarifier)
TOTAL Make-up service water to thickener overflow
tank from river.
FRESH WATER ADDED PER MOLE OF SULFUR REMOVED 180
8. BYPASS SYSTEM filank Qff ^
CAN FLUIC GAS BE BYPASSED AROUND FGD SYSTEMS But not automatic
GAS LEAKAGE THROUGH BYPASS VALVE, ACFM No bypass valve
K. SLURRY DATA
LIME/LIMESTONE SLURRY MAKEUP TANK
I'ARTICULATE SCRUBBER EFFLUENT
HOLD TANK (a)
S02 SCRUBBER EFFLUENT HOLD
TANK (a)
PH
%
Solids
Capacity
(gal)
Hold up
time
L. LIMESTONE MILLING AND CALCINING FACILITIES: INDICATE BOILERS
SERVED BY THIS SYSTEM.
TYPE OF MILL (WET CYCLONE, ETC.) Dravo ,Gibbs & Hill
NUMBER OF MILLS
CAPACITY PER MILL
RAW MATERIAL MESH SIZE
PRODUCT MESH SIZE
T/HR
A-10
5/17/74
-------
SLURRY CONCENTRATION IN MILL
CALCINING AND/OR SLAKING FACILITIES
SOURCE OF WATER FOR SLURRY MAKE UP OR
SLAKING TANK
M. DISPOSAL OF SPENT LIQUOR
1. SCHEMATICS OF SLUDGE & FLY ASH DISPOSAL METHOD
(IDENTIFY QUANTITIES OR SCHEMATIC)
2. CLAUIFIEHS (THICKENERS)
NUMBER __2
DIMENSIONS 75' diameter
CONCENTRATION OF SOLIDS IN UNDERFLOW 30-40%
ROTARY VACUUM FILTER
NUMBER OF FILTERS
CLOTH AREA/FILTER
CAPACITY TON/111? 'WET CAKE)
CONCENTRATION OF SOLIDS IN CAKE
PRECOAT (TYPE, QUANTITY, THICKNESS)
REMARKS
4. SLUDGE FIXATION
Mixing tank prior to
POINT OF ADDITIVES INJECTION sludge holding pond
FIXATION MATERIAL COMPOSITION
FIXATION PROCESS (NAME) Calcilox (Dravo)
FJXATLON MATERIAL REQUIREMENT/TONS OF DRY SOLIDS OF SLUDGE
Nominal 10%
A-ll 5/17/74
-------
ESTIMATED POND LIFE, YRS.
CONCENTRATION OF SOLIDS IN FIXED SLUDGE
METHOD OF DISPOSAL OF FIXED SLUDGE Truck haul to land fill
INITIAL SOLIDIFICATION TIME OF FIXED SLUDGE
SLUDGE QUANTITY DATA
POND/LANDFILL SIZE REQUIREMENTS, ACRE-FT/YR
IS POND/LANDFILL ON OR OFFSITE Off-site
TYPE OF LINER Hypalon on demonstration
pond only
IF OFFSITE, DISTANCE AND COST OF TRANSPORT N/A
POND/LANDFILL DIMENSIONS AREA IN ACRES N/A
DEPTH IN FEET N/A
DISPOSAL PLANS; SHORT AND LONG TERM
N. COST DATA
$33,000,000 Phase I,
1. TOTAL INSTALLED CAPITAL COST Phase II -$7.000.000
without fixed charges Phase I is
2. ANNUALIZED OPERATING COST S5fQQQfQQQ Phase II 7.273.000
Including fixed charges Phase I
10,000,000 Phase II 13,661,000
A-12 5/17/74
-------
3.
COST BREAKDOWN
A.
COST ELEMENTS
CAPITAL COSTS
S02 SCRUBBER TRAINS
LIMESTONE MILLING
FACILITIES
SLUDGE TREATMENT &
DISPOSAL POND
SITE IMPROVEMENTS
LAND, ROADS, TRACKS,
SUBSTATION
ENGINEERING COSTS
CONTRACTORS FEE
INTEREST ON CAPITAL
DURING CONSTRUCTION
ANNUALIZED OPERATING COST
FIXED COSTS
INTEREST ON CAPITAL
DEPRECIATION
INSURANCE & TAXES
LABOR COST
INCLUDING OVERHEAD
0 & M COSTS
LABOR
RAW MATERIAL
UTILITIES
MAINTENANCE
INCLUDED IN
ABOVE COST
ESTIMATE
YES
a
X
a
X
a
x
a
X
a
X
E
NO
a
a
a
a
a
a
n
a
a
a
s
a
a
a
ESTIMATED AMOUNT
OR 1 OF TOTAL
INSTALLED CAPITAL
COST
G & H
7.5%
20.5%
19.0%
15.5%
DISPOSAL
37.5%
Based on full S02 removal to meet regulations of
0.6% Ib S0-/million BTU emissions.
A-13
5/17/74
-------
4. COGT [-'ACTORS
a. ELECTRICITY
b. WAT UK
C. STEAM (OR FUEL FOR REHEATING)
d. FIXATION COST N/A
X
C. RAW MATERIAL PURCHASING COST
f. LABOR: SUPERVISOR
OPERATOR
OPERATOR HELPER
MAINTENANCE
S/TON OF DRY SLUDGE
$/TON OF DRY SLUDGE
_ HOURS/WEEK WAGE
U. MAJOR PROBLEM AREAS: (CORROSION, PLUGGING, ETC.)
1. S02 SCRUBBER, CIRCULATION TANK AND PUMPS.
a. PROBLEM/SOLUTION
.Deposits on skirts of scrubber, Gusset Plate
Throat parnper_ __
Under Upper Cone
2.
3.
Bleed Valves
Recycle Pumps_
DEMISTER
PROBLEM/SOLUTION,
Lime Deposits
REIIEATER
PROBLI-:M/SOLUTION_ _ .
J?.unip_s_, __Blo.wer.s.,_..Tenip_. .Co_ntr_ol
A-14
5/17/74
-------
4. VENTURI SCRUBBER, CIRCULATION TANKS AND PUMPS
PROBLEM/SOLUTION
Recycle pumps_ 33% Impeller and casing erosion in 3000 hrs
Better volute design, possibly rubber line pumps
5. I.D. BOOSTER FAN AND DUCT WORK
PROBLEM/SOLUTION
Dampers Corrosion - Ceilcote same
Fan Yiflrf s-t-rp»nnhh . r^nrrnsinn — rnhhpr 1 i np shaft:, Tnf-prp
_ ^ f— ' Wash
Exp. Joint .Asbestos-Butyl compound material for joints
better than s.s.
Fan Strengh^LWeld mat overlay install doubler plates to
reduce operating yield strength
6. LIMESTONE MILLING SYSTEM OR LIME SLAKING
PROBLEM/SOLUTION
Degritting - No redundancy 10% down time
better if indoor operation, spare capacity
7. SLUDGE TREATMENT AND DISPOSAL
PROBLEM/SOLUTION
.Over_ 10OQ_ton/day sludge disposal requir_ed_t Purchase
and Development
Leachate Still not known if leachate is acceptable
more testing required.
A-15 5/17/74
-------
8. MISCELLANEOUS AREA INCLUDING BYPASS SYSTEM
PROBLEM/SOLUTION
Instrumental:!On need good pH cells, constant attention
. stack opacity monitor useless
Boiler Instrument
Stack
Noise Fan noise and associated equipment
InstrumentativG Sump pumps high erosion wear
DESCRIBE FACTORS WHICH MAY NOT MAKE THIS A REPRESENTATIVE
INS TALL AT I ON What is representative? No such thing
Possibly Land limitations
Q. DESCRIBE METHODS OF SCRUBBER CONTROL UNDER FLUCTUATING
LOAD. IDENTIFY PROBLEMS WITH THIS METHOD AND SOLUTIONS.
IDENTIFY METHOD OF pH CONTROL AND LOCATION OF pH PROBES.
Boiler gas _flow._ fed into damper control and venturi throat ___
damper control. System has worked ok. but because plant hag only
jr.ecejitJ.Y-r.un.. with all, boilers rnnnect'p'ri t-O g
had in_s.ufficie_nt_time_ to. observe. _ Have lost trains occasionally
pH is controlled by maintaining scrubber .re_cy_cle .solution., at a
given level. _ _ _ _ _
A-16 5/17/74
-------
APPENDIX B
PLANT PHOTOGRAPHS
B-l
-------
Photo No. 1 General view of the Phillips Power Station from
the sludge disposal area. The scrubber stack appears at the
center of the picture.
Photo No. 2 View of the Phillips lime storage tank and
slaking facility. The storage tank has a -day capacity
of pounds of lime.
B-2
-------
Photo No. 3 View of the Phillips FGD scrubber module is
almost completely obscured by structural steel. Gas inlet
duct appears at the right side slightly below the center of
the picture.
Photo No. 4 View of ductwork between scurbber and stack.
The direct-fired reheater is located immediately to the right
of the scrubber module.
B-3
-------
Photo No. 5 View of Phillips FGD scrubber recirculating pumps
?t ***
Pf ft ' Jt* *Ji-*'*;
Photo No. 6 View of the Phillips scrubber control panel
B-4
-------
Photo No. 7 View of edge of empty Phillips sludge disposal
pond showing plastic barrier liner which surrounds the pond.
Leachate collection pipes are installed in the bottom of the
pond for sampling and analysis.
Photo No. 8 Stabilized sludge in Phillips test disposal
pond. Barrier liner at perimeter can be seen in foreground
and background.
B-5
-------
APPENDIX C
OPERATIONAL REPORT
C-l
-------
DUQUESNE LIGHT COMPANY
PHILLIPS POWER STATION
LIME SCRUBBING FACILITY
By
Steve L. Pernick, Jr
Manager-Environmental Affairs
Duquesne Light Company
Pittsburgh, Pennsylvania
and
R. Gordon Knight
Superintendent-Technical Services
Duquesne Light Company
Pittsburgh, Pennsylvania
Prepared for presentation
at
EPA Flue Gas Oesulfurization Symposium
Atlanta, Georgia
November 4-7, 1974
c-2
-------
DUQUESNE LIGHT COMPANY
Phillips Power Station
Lime Scrubbing Facility
CONTENTS
Page
Introducti on V-4
Pilot Plant Operation C-5
Installation Schedule C-6
Description of System .C-8
Initial Operation C-ll
Extended Shutdown for Repairs C-13
Second Start-up Operation C-16
Sludge Treatment and Disposal C-18
Future Operating Plans C-21
Availability and Reliability C-23
Capital and Operating Expenses C-23
Future Objectives C-24
c-3
-------
DUQUESNE LIGHT COMPANY
Phillips Power Station
Lime Scrubbing Facility
INTRODUCTION
Duquesne Light Company is an investor-owned electric utility
serving approximately 1/2 million customers in southwestern Pennsyl-
vania including Pittsburgh. The Company has a net generating capa-
bility of approximately 2700 MW and presently operates three coal
burning power stations, two of which, Phillips and Elrama, are dis-
cussed in this paper.
In 1969, Gibbs & Hill, Inc. was engaged to conduct a com-
prehensive study of the most feasible means of complying with: the
emission regulations being promulgated at the time by Allegheny County,
which has control over the Phillips Station; the regulations of the
Commonwealth of Pennsylvania, which has jurisdiction over the Elrama
Station, and the most stringent regulations that could reasonably be
conceived. The study was completed in September, 1970. It repre-
sented a comprehensive evaluation of all test data concerning emis-
sions from the two power stations and an intensive investigation of
all possible means of compliance with the applicable and anticipated
regulations. The study concluded that: (1) the use of low sulfur
western coal was not a feasible means of compliance since it would
not achieve compliance with the particulate emission limitations of
the regulations with the existing dust collection equipment; (2) that
conversion from coal to low sulfur oil was not feasible due to the
uncertainties of obtaining a sufficient supply of the necessary
large quantities of such oil; and (3) that the most practical means
of sulfur dioxide emission control was desulfurization of stack
gases. The consultants then conducted a thorough study of existing
methods of sulfur dioxide removal from stack gases. At the time of
their investigation in 1969 to 1970, only two full scale stack gas
scrubbing systems were in operation in the United States. One of
these has since been abandoned and the second has had a poor record
of operating reliability. Furthermore, the injection of limestone
directly into the boiler, which was the basis of these SO? removal
processes, was not at all suitable for consideration at the Elrama
and Phillips Power Stations since it would result in additional
boiler tube erosion to which the boilers were already particularly
susceptible. Furthermore, available data indicated that S02 removal
efficiency might not be sufficiently high to comply with the regu-
lations which were among the most stringent in the country.
Because of the characteristics of our boilers and our re-
quirements for a system that would remove both particulars and sul-
fur dioxide, the report concluded that a tail-end scrubbing system
was the best available means of control and recommended that we in-
stall a dual stage venturi scrubbing system. Some of the technical
C-4
-------
reasons for choosing the system are as follows:
1. It was thought that a dual stage scrubbing system would enable
separating the sludges resulting from scrubbing for fly. ash
in the first stage and scrubbing for S02 in the second stage,
in the event the sludge from the S02 removal stage became a
problem.
2. A dual stage scrubber had the potential of being converted
to a magnesium oxide regenerative system in the event it
became desirable or practical to do so.
3. Venturi scrubbing for particulate removal was fairly well
proven in the steel industry. Furthermore, if only one
of the scrubber trains was equipped with a dual stage S02
prototype, and this was proved to be impractical or un-
workable, we would still have a workable single stage ven-
turi scrubber system for particulate removal. We could
then investigate an alternate method of S02 removal with
minimal loss of investment.
In addition to the above reasons was the fact that preliminary
time schedules indicated that these scrubbers could be delivered with the
least delay. Duquesne accepted the recommendations of Gibbs & Hill, and
in view of the unproven nature of the wet scrubbers for sulfur dioxide
removal, decided that the most prudent approach to the installation was
to begin by equipping only a portion of the scrubber system at the Phillips
Station with sulfur dioxide removal equipment. Subsequently, Gibbs & Hill
was instructed in September, 1970, to proceed with the designing and in-
stallation of the Phillips scrubber system. The scrubber installation
at this plant was to include four scrubber trains. Three of these were
single stage scrubbers for dust removal, and one was a dual stage scrub-
ber (two single stages in series) which is the prototype for sulfur di-
oxide removal. Each scrubber was to be capable of approximately 540,000
CFM, which is the equivalent of approximately 125 megawatts. With the
station rated at 387 megawatts, this meant that the station would be
equipped with essentially a spare scrubber train. This was a criterion
specified by Duquesne Light Company to assure sufficiently high avail-
ability of the system.
It was anticipated that after a trial period, the operation of
the prototype would serve as a basis for decision as to the installation
of additional similar S02 removal equipment at this station and other
coal burning facilities to comply with the applicable regulations.
PILOT PLANT OPERATION
In February, 1971, a pilot plant size Chemico venturi type,
dual stage scrubber system with a capacity of 1500 CFM was installed
at the Phillips Power Station to study the performance and operating
characteristics of the dual stage venturi scrubber system. The pilot
C-5
-------
plant was connected to the duct entering the mechanical dust collector
of No. 5 boiler. The scrubber was operated with various types of lime-
stone and lime. In addition, other conditions such as the stoichiometry,
liquid to gas flow ratio and pressure drop were varied in order to deter-
mine the conditions which would yield optimum collection efficiency and
operating conditions. After determination of these parameters in approxi-
mately 3 months of operation, the pilot scrubber was dismantled on May 8,
1971.
In the meantime, in March, 1971, invitations to bid were issued
by Duquesne Light Company to Chemico and Combustion Engineering for a
scrubber system capable of removing sufficient S02 and particulates to
comply with existing and anticipated future regulations. The bid was
subsequently awarded to the Chemico Corporation in July, 1971. The
battery limits for Chemico was limited to the scrubbers and associated
pumps and controls between the inlet hot gas duct manifold to the exit
wet gas header, including the reheater but excluding the new induced
draft fans.
In obtaining approval of our plan, compliance programs were
negotiated with Allegheny County for the Phillips Station and with the
State for the Elrama Power Station. The program included the instal-
lation of five single stage venturi scrubbing trains for particulate
removal only at the Elrama Station, to be installed concurrently with
the Phillips scrubber system as described above. The plan was to use the
prototype sulfur dioxide removal installation at Phillips as a basis for
decisions concerning additional SC>2 removal equipment at both stations
in order to comply with SC>2 emission limitations.
INSTALLATION SCHEDULE
Although in Diiquesne's best judgment, the scrubber system could
not be made operational, even for particulate removal, prior to July 1,
1973, which was 34 months from the date of our decision to install a
scrubber system, Duquesne Light eventually acceded to pressure from the
State of Pennsylvania and established a target date of January 1, 1973.
Following startup, the schedule called for a tv/o-month "debugging" period
followed by a ten-month test program on the prototype SO^ scrubber at
Phillips. This would determine if stack gas scrubbing with alkaline in-
jection on a plant scale basis was a feasible and reliable means of re-
moving sulfur dioxide from the gases. The plan was to study operating
conditions during this period to establish optimum SC"2 removal efficiency
using various types of lime, the operating mode most conducive to reliable
trouble-free operation, and other essential information. The schedule
called for a decision to be made by January 1, 1974, as to whether or not
the operation of the prototype S02 scrubber was satisfactory or if an
alternate system should be investigated. If the operation of the prototype
sulfur dioxide scrubber proved satisfactory, the schedule called for design,
procurement and installation of additional SO? removal equipment by January 1,
1975, for both stations. At that time, compliance with County and State
regulations would be achieved. (For a list of significant dates see Figure 5)
C-6
-------
Despite our best efforts, no part of the Phillips scrubber
installation was placed in operation until July 9, 1973, and then for
only two boilers. At that time, two of the scrubber trains were
placed in operation, and tv/o of the six boilers were connected to the
scrubber system. This was consistent with our startup schedule which
called for connecting one boiler at a time to the scrubber system so
as not to jeopardize the availability of a significant portion of the
plant's generating capability in the event of failure of the scrubber
system. The tv/o boilers connected have a combined steam generating
capability equivalent to approximately 80 megawatts.
The July 9, 1973 startup date represents a slippage of approxi-
mately 6 months from the original schedule. This slippage was attribut-
able to the following factors:
1. The venturi scrubber system was not a completely engineered
system at the time it was purchased, and a considerable amount
of developmental engineering was required to adapt the scrubbers
to a system that would meet our requirements. Chemico was al-
so charged to incorporate, to the limits of their responsibility the
latest available operating and maintenance experiences at the Four
Corners Plant in New Mexico, the Dave Johnston Plant in Wyoming,
Mitsui Plant in Japan, and any other Chemico installation having
experiences adaptable to our proposed installations.
2. Equipment suppliers and fabricators did not meet promised de-
livery dates due to the volume of orders, decreased produc-
tivity, and inadc 'ate quality control. Considerable money
was spent to expccite and inspect equipment and materials for
our project. In spite of this effort and expense, many late
deliveries and mistakes in fabrication occurred. Some examples
are as follows:
a. The induced draft fans were scheduled for delivery
in Phillips by July 15, 1972. Delivery was not
completed until December 12, 1972, a delay of five
months. Not only were the fans late, but numerous
cracks and defects in the welds on the fan rotors
were found after delivery. It was necessary to
grind out all of the defective welds and make the
necessary repairs in the field.
b. An approximate five-month delay was experienced in
delivery of the ductwork. In addition, because of
the huge size and weight of the ductwork, many
structural support schemes had to be studied before
an acceptable method of reinforcing the existing
station building to accept the unusual weight load
of the ductwork could be determined. After a rein-
forcement plan was devised, errors were discovered
in the implementation of the plan, requiring cor-
rective design work.
C-7
-------
c. Delays of up to six months were experienced in the
delivery of electrical equipment such as switch-
gear and transformers. In addition to the delay
of equipment, errors were found in the wiring dia-
grams for this equipment. Once the equipment was
received, design errors were also discovered, such
as improperly designed bus connections on the switch-
gear and improper alignment of bus bars on power
centers. In addition, failures of switchgear and
transformers were experienced during field testing.
These are only a few examples of the numerous delays and errors
by equipment suppliers which affected the final completion date
of the first phase of the Phillips project.
3. Major delays were also incurred in the field, such as the follow-
ing:
a. Because of engineering delays and late delivery of
equipment, construction schedules had to be con-
tinually revised which seriously affected pro-
ductivity.
b. Space limitations existing at the job site, more
serious than anticipated, affected productivity
because of the lack of accessability and maneuver-
ability. Special equipment and construction methods
for the installation of the equipment were required
beyond those originally anticipated, thus contribu-
ting to additional delays.
c. The availability of boiler outages for tie-in of the
new ducts, presented additional problems. Outages
had to be scheduled months in advance, and any re-
vision in schedule was dependent upon availability
of purchased power and the outage schedules of
other generating equipment, both on our system and
on interconnections.
DESCRIPTION OF SYSTEM
Design of both the Phillips and Elrama systems was a joint
effort by Gibbs & Hill and Chemico, each having certain battery limits.
Construction was the responsibility of Gibbs & Hill. Only the Phillips
system will be described in detail.
The Phillips system comprises four parallel trains, Figure 1,
and was designed to handle a total gas volume of 2,190,000 ACFM with
all trains in service. With one train out of service, the three
trains have a capacity of 1,600,000 ACFM, which is sufficient for
400 MW, assuming normal excess air.
C-8
-------
Three scrubber trains were provided with first stage scrub-
bing for fly ash removal, Figure 2. The fourth train, known as the
"prototype" SOp train, was fitted with a first stage particulate re-
moval scrubber and a second stage SOo absorption scrubber as shown
on Figure 3.
The ductwork on each boiler was modified to enable operation
with the gases going either to the scrubbers or to the original gas
path after passing through the existing dust collection equipment.
A mechanical and electrical dust collector are located in series
ahead of each individual boiler stack. The existing ducts leading
to the old ID fans and stacks were blanked off with removable steel
plates to enable diversion of the flue gas to a new common duct.
The new ductwork is routed to the scrubber building where 16 foot
diameter takeoffs lead to each of the four first stage scrubbers.
The hot flue gas (about 340°F) containing fly ash and sul-
fur dioxide (about 1400 ppm S02) enters the first stage scrubber and
impinges upon the upper cone. One half of the total scrubber liquor
(8250 gpm) is introduced, Figure 4, into the vessel through the bull
nozzle where it is sprayed over the surface of the upper cone, to
initiate the scrubber action. A second stream of 8250 gpm scrubber
recycle liquor enters through the tangential nozzles at a point
above the adjustable throat damper. The flue gas and scrubbing
liquor are brought into intimate contact in the throat section of
the scrubber where ths particulates and some 862 are removed. The
gas and liquor continue downward to the separator section where, after
being separated from the scrubbing liquor, the flue gas enters a de-
mister where entrained liquor is removed. The gas leaves the-scrubber
and enters a new wet ID fan which is provided with water sprays to re-
move any accumulation of solids resulting from carryover from the
scrubber. Fresh water is used for spraying.
The ID fan housings are lined with 1/4 inch thick natural
rubber. Wheel material is Carpenter 20 Cb 3. The shaft is 316 L
stainless steel. Each fan is driven by a 5500 HP, 1200 RPM, 4160V
electric motor. Bearings are of the water-cooled type, served from a
closed cooling water system.
Outlet gas temperature, normally 110-120°F, from the first
stage scrubber is monitored. At 175°F a control valve is automatically
opened to admit emergency cooling water to the upper cone. Additional
temperature rise would automatically shut down the fan and close the
isolation dampers.
The gas leaving the ID fan enters one of two vessels de-
pending on the train. In the case of the three single stage scrubber
trains, the gas enters an entrainment separator in which entrained
water is separated and collected. The separators or mist eliminators
are separate vessels which will be replaced by second stage scrubber
vessels if the test program indicates this to be necessary, in the
C-9
-------
case of the S02 prototype train, gas leaving the ID fan enters a
second stage scrubber vessel, which is identical in size and design
to the first stage. This second stage vessel is equipped with a
reagent (lime) injection nozzle in the bottom cone. The scrubbing
liquor is then picked up by the two recycle pumps (Carpenter 20
Cb 3). Some of the liquor is bled off to the first stage scrubber, where
some is removed from the cycle to thickeners, and the balance is
recycled.
Total recycle flow is about 16,500 gpm with a 615 gpm
bleed to the thickeners. All single stage vessels have also been
fitted with reagent nozzles and feeds to enable control of pH and
incidental S02 removal.
The exit gases leaving either the second stage scrubber
vessel orthemist eliminator vessel enter a common wet duct lined
with Flakeline 103 (a product of the Ceilcote Company consisting
of a glass flake filled modified polyester resin, subsequently
referred to as Ceilcote) which leads to a new concrete acid re-
sistant brick lined chimney. Prior to entering the 340 foot
chimney, a section of the wet gas duct, constructed of 316L SS is
equipped with a reheater, employing a direct oil-fired burner system.
The two-burner system is capable of reheating 30°F, although the in-
tent is to normally reheat 20°F.
Bleed off from all scrubbers is directed to a trough feeding
two 75 foot diameter thickeners. The overflow from these thickeners
is directed to a collecting tank where it is pumped by two-2000 gpm
SS pumps to the make-up line for return to the scrubber system.
Thickener underflow, at 35 to 40% solids concentration, is
pumped by one of two 105 gpm, 15 HP pumps to one of three clay lined
sludge holding ponds. Each pond has a capacity of about 6500 cubic
yards. Just prior to discharge to the holding ponds, the sludge
enters a mixing tank where a stabilizing agent can be added at a
predetermined rate based on the density of the sludge as it leaves
the thickener. After the additive is intimately mixed with the
sludge, the mixture discharges by gravity to one of the ponds.
In the pond, both settling and curing take place. Super-
natant liquid is withdrawn and recirculated to the scrubber system
via the thickener. The sludge curing ponds provide interim storage
while stabilization is taking place. One pond received thickener
underflow, while a second is curing and the third is being excavated
for final off-site disposal.
Lime is fed from a storage silo at a controlled rate to the
lime slaker where it is slaked with fresh make-up water. The slaked
lime overflows into a slaker transfer tank, where make-up water is
added to provide a constant flow of lime slurry with a concentration
of about 15%.
C-10
-------
A service (river) water system independent from the power
station service water system is provided. It includes a pair of 900
gpm, 100 HP pumps taking suction from Lhs existing condenser dis-
charge tunnel and a 10 inch distribution header for the scrubber
facility. It provides fan spray water, pump seal water, mist elim-
inator and scrubber demister spray water, instrumentation flush
water, chemical mixing tank, and emergency water for scrubbers.
INITIAL OPERATION
Although the lime addition system was not yet completed,
partial start-up of the scrubber system began on July 9, 1973. The
decision to proceed with startup, without the lime addition system,.
was made in light of the fact that wo thought the scrubber system
was suitable for operation at a low pH. Almost all of the major
components and piping were either Ceilcoted, rubber lined or fabri-
cated of a corrosion resistant material to withstand a low pH.
During the first five days of operation, a pH of approximately 1.3
to 2.0 developed in the scrubbing liquor c/hile burning 1.8-2.0% sul-
fur coal. Because of pressure to prour_>d with the debugging operation,
and because of our reluctance to incur additional delays, the decision
was made to continue operation with the low pH rather than shut down
the scrubbers until the lime system became operational. Unfortunately,
as a result, the scrubber system was forced out of service due to the
corrosion failure after five days of the two thickener overflow re-
turn pumps, which were intended for operation at a pH of 5 to 6. They were
promptly replaced with two 316 L SS purrps. Use of the lime system
feed to the thickeners was obtained after the first week of operation.
During the first two weeks of operation, additional problems
developed and were corrected in a two-week outage following the first two
weeks of operation:
1. A piece of rubber linrr was found in the suction of one of
the recycle pumps. The individual scrubber was shut down
in an effort to locate the source; however, it could not
be determined until the corrosive nature of the scrubber
media caused the failure of a 20"/16' Tee within 10 days.
2. Leaks occurred in the lead floor lining of the stack which
were a potential threat tu the stack (oncrete structure.
The lead sheets had been solder-r! rather than "burned."
The lining was replaced.
3. Leaks developed in four 405 *5S expcinsion joints in the
gas ducts leading to the stack. The appearance of the
failed metal was variously escribed cis that of "lace
curtain" or "swiss cheese." Fhi., resulted from partial-
load type of operation of the scrubber system requiring us
to partially open t^c vacuum breakers and allow ambient air
c-11
-------
to supplement the gas flow through the scrubbers. Conden-
sation resulted. The joints were replaced with an asbestos/
butyl material.
4. The reheater, which utilizes direct firing of No. 2 oil in
the cold gas duct as it enters the stack, is not yet oper-
ational. We have been unable to obtain satisfactory oper-
ation of the reheater due to burner problems and oil pump
problems. We have placed no priority on solving these prob-
lems due to our reluctance to operate the reheater consid-
ering the present energy crisis and oil shortage. Whenever
the reheater becomes operational, we plan to operate it
during periods of adverse meteorological conditions to ob-
tain a higher effective stack height.
5. During the outage, evidence of corrosion was found on the
ID fan shaft shrouds, inlet dampers and stiffener bars - all
316 L stainless steel. TWO corrective measures were taken -
redesign of the fan sprays and application of 1/4-inch
rubber coating to the affected parts. After two weeks of
operation the chloride concentration in the liquor (closed
cycle)had reached 800-900 pprn. This, no doubt, contributed
to the corrosion noted above. It was eventually to reach
3500-4500 ppm.
On August 5, 1973, the scrubber system was returned to service
with the lime feed system for pH control continuing in operation. In
addition to No. 1 and No. 2 boilers, the No. 3 boiler was also connected
to the scrubber system, which made a total of approximately 145 mega-
watts.
During the following weeks of operation, we were informed by
Chemico of a major change in the operating requirements of the scrubbers;
namely, the necessity to bleed more fluid to the thickener tanks so
that the concentration of solids in the recycling fluid could be kept
to a minimum in order to protect the recycle pumps, and to minimize the
settling of solids in the bleed lines. This caused a change in the de-
sign concept by essentially doubling the bleed rate. This meant that
the two 900 gpm thickener tank overflow return pumps, which had been
provided with the system, were inadequate since they no longer had
sufficient capacity to supply the needs of the scrubber system under
one pump operation (the second being a spare) with all boilers con-
nected to the scrubber system. Two 2000 gpm pumps were obtained for
later installation in parallel with the two 900 gpm pumps.
During the two months of operation following the August 5,
1973 start-up date, operation of the scrubbers was alternated between
trains. In other words, as problems arose on one train, it was shut
down and another train was put in service to take its place. No effort
was made to operate the dual stage prototype S02 scrubber for full SOg
removal because it was felt that debugging operations could more effect-
ively proceed with the most basic operation possible, in other words,
C-12
-------
with the single stage scrubbers.
EXTENDED SHUTDOWN FOR REPAIRS
During the operation of the scrubbers in August and Septem-
ber, 1973, additional problems came to light which caused an extended
shutdown of the scrubber system, and a further delay in the operating
schedule, namely:
1. A routine inspection of the new stack revealed additional
moisture leaks from condensation (pH 2.0) seeping through
the mortar joints and running down the outside of the acid
brick lining. Seepage was occurring in more than 100 areas
of the brick lining. The condensate was collecting at the
bottom of the stack in the annul us between the lining and
the shell. The acid condensate was also seeping through
construction joints and down the outside of the concrete
shell permitting it to attack the concrete, which could
possibly impair the structural stability of the stack. The
contractor advised us that a four week outage would be
necessary to permit inspection of all mortar joints and
to repoint as required. This outage for other reasons to
be described, extended for about five months.
2. Significant scaling and buildup had occurred in various
portions of one or more of the scrubbers and in the asso-
ciated piping in varyimj degrees as described below:
a. Buildup was found on many of the upper cone sup-
port struts. These football-size accumulations
appeared to be creating a disturbance in the water
flow pattern over the cone resulting in accumu-
lations of deposit en the cone surface.
b. Deposits were found in the throat area of two of
the scrubbers, sufficiently thick in some areas
to prevent the maintenance of the six inch de-
sign pressure drop across the throat.
c. A thin buildup of hard scale was observed on the
face of the throat dampers of on? of the scrubbers.
d. Deposits of fly ash were found in and around the
operating mechanisms of the adjustable throat
dampers. These were sufficient to prevent full
adjustment of the dampers.
3. Corrosion/erosion was found in numerous areas:
a. Severe pitting and corrosion appeared on all of the 316 L
clad ID fan damper blades and frames. The cladding was
C-13
-------
removed and the carbon steel was Ceilooted.
b. In all scrubbers, a moist layer of fly ash was observed
on the carbon steel skirt that directs the flue gas
down the cone. Extensive rusting was found under the
fly ash and on the underside of the vessel tap as well.
The latter may have resulted from frequent outages for
inspection, during which alternate wetting and drying
occurred.
c. The upper cones of the scrubber vessels were fabricated
with a 316 L SS apex. The balance was carbon steel with
a Ceilcote coating. The uncoated apex showed erosion/
corrosion in all vessels. Similar attack was shown on
the inner surface of the 316 L bull nozzle directly
above the cones. Both affected areas were coated during
the outage.
d. Baffle plates in some tangential nozzles (12 per vessel)
were found badly corroded. The specified material of
construction was 316 L SS, but it was found that the
corroded baffles had been fabricated of 304 SS. Six-
teen baffles out of a total of 60 required replacement
with 316 L.
e. Similarly, the throat damper arms (12 per vessel) were
supposed to be 316 L SS. Thirty-two out of a total of
60 were badly corroded and were found to be 304 SS. They
were replaced with 316 L.
f. The top and bottom throat damper scraper blades were
practically destroyed. The also were found to have
been 304 SS instead of 316 L. They were replaced with
316 L.
4. An initial inspection and analysis of the ID fans by Franklin In-
stitute Research Laboratory (FIRL) indicated the existence of
chloride stress corrosion cracking of the parent metal and
structural welds, as well as general corrosion on all exposed
surfaces. At that time, they advised us that the problems re-
sulted from the hostile environment in which the fans were
operating. In addition, they made a specific recommendation
that the fans not be operated in the condition found at that
time, due to possible catastrophic failure.
As a result of their search for protective coatings, FIRL
recommended the application of a Ceilcote material, Coroline
505AB. This is a hard, impervious, acid resistant, epoxy base
material which is applied at a nominal thickness of approxi-
mately 1/4". The application of the Coroline is a tedious,
time consuming hand application by trowel and requires close
C-14
-------
quality control. The application of the material began in
November, 1973. Although Coroline had not been extensively
used on fan blades, or proven to any degree, we were given
reasonable assurance that it would be successful. As a
result, we began application of this material to as many
fans as manpower permitted.
On January 22, 1974, the first fan was completely coated
and was started up for balancing. After a balance move
was attempted, the fan was forced out of service due to
excessive vibration resulting from separation of a portion
of the coating from one of the blades. The Ceilcote Com-
pany informed us that the separation was localized and
probably caused by a faulty batch of material which did
not properly adhere to the metal. The Coroline was re-
applied and the fan was started again on February 5, 1974
and again was forced to shut down due to separation of the
material in a second area. At that time, the Ceilcote Com-
pany was unable to adequately explain the failure of the
coating to adhere to the fan surfaces. As a result, the
decision was made that the application of Coroline as
protection against the adverse effects of the flue qases
was not the answer to the problem. It is believed that the
flexing of the blades during operation may have contributed to
the lack of adhesion between the Ceilcote and the blades.
By this time, FIRL had been able to conduct more exhaustive
investigations and analyses on the fan blade material. As a
result, they revised their initial conclusion and reported
that the fans were suffering from a much lesser degree of
stress corrosion cracking of the weld metal and a degree of
pitting attack on all fan blades. The report also indicated
that fen failure was no longer considered imminent.
In review of the ID fan problems, it appears that the corrosion
problems which we had experienced might be partially due to the operation
of the scrubber system at a lower pH and a higher chloride concentration
(3500-4500 ppm) than had been foreseen by the scrubber designers. Several
steps were taken to try to alleviate the situation:
1. Operation of the dual stage scrubber with full lime addition,
which we felt would be adequate for maintaining optimum pH
control in that train;
2. Addition of supplemental lime feed to the other operating
scrubbers, and
3. The installation of a redesigned spray system, which had proven
effective at other installations.
C-15
-------
These steps in addition to frequent scheduled fan inspections
were felt to be adequate safeguards for placing the system
back in operation.
Our ability to start up the scrubber system and initiate full
S02 removal in the dual stage prototype was further improved by the
completion of the sludge stabilizer addition system. We had been re-
luctant to add large quantities of lime to the single stage scrubbers,
and equally reluctant to begin operation of the dual stage scrubber,
both of which result in the production of considerable amounts of
sulfate sludge, without the capability of adding a stabilizer.
SECOND START-UP OPERATION
Removal of the Coroline material from the dual stage scrubber
fan No. 1 was completed, and startup of the scrubber system once again
began March 17, 1974. At that time, the dual stage Scrubber Train No. 1,
and Scrubber Train No. 4 were placed in operation with boilers No. 3
and 4 connected to the scrubber system. Since then there have been
these additional developments:
1. Excessive Fly Ash
The sludge handling system was designed so that after 30
days of normal operation with all boilers tied in and
with the dual stage S02 prototype and two of the three
single stage scrubbers, the first curing pond should be
approximately full. However, after the first two weeks,
one pond was completely filled with only two boilers con-
nected to the scrubber system during the first week and
three boilers connected to the system during the second
week. This means that, with an average of 120 megawatts
connected to the scrubber system, we were producing approxi-
mately 7,000 tons of sludge in a two week period. Or,
stated another way, if the entire plant were connected to
the scrubber system, we would fill a pond within a week.
The major factor contributing to this change in capacity
proved to be an increased fly ash loading of the scrubber
system. The sludge disposal system was based on a 90%
combined collection efficiency of the mechanical dust
collectors and electrostatic precipitators. The transition
of the new scrubber ductwork to the old ductwork is such
that the gas flow pattern of the gases through the pre-
cipitators proved to be greatly disrupted. Velocity
traverse data indicated that laminar gas flow no longer
exists. Rough estimates indicated that the amount of
particulates leaving the precipitators and entering the
scrubber system may have trippled. Additional details on
the sludge handling will be covered later.
c-16
-------
2. ID Fan Stress
As part of the continuing investigation to determine the
conditions of the ID fans, Structural Dynamics Research
Corporation (SDRC) was engaged to conduct a series of
strain gage tests to determine the structural stability
of the fans. Their tests indicated that the yield
strength was being exceeded in several portions of the
fan blades, and that a degree of metal deformation was
taking place. After additional strain gage testing,
SDRC recommended the installation of doubler (reinforcing)
plates on each of the blades to reduce the stresses to
acceptable levels. After actual tests with stressed welded
specimens in the fan atmosphere, it was also recommended
that the doubler plates be welded with an Inconel 112 rod,
rather than the Carpenter 20, 4 NIA or 8 N12 rods pre-
viously used. Since the installation is on a fan-by-fan
basis, so that a spare scrubber train would always be
available for emergency backup, it is expected that all
four fans will not be completed until the end of November,
1974.
3. ID Fan Attack
Frequent shut down of the ID fans for inspection revealed that
significant attack was still taking place. Despite the addi-
tion of lime for S02 removal and pH control, a low pH was being
encountered in the ID fan drains. This results from additional
S02 scrubbing occuring with the fan spray system. In a trial
installation, caustic was added to the fan spray water. How-
ever, an inordinate quantity is required to obtain even a pH
of 4, and the trial will soon be ended. As another attempt
to alleviate the attack, a new type of fog nozzle was installed
on one of the ID fans in an attempt to increase the effective-
ness of spraying. No evaluation is yet possible.
4. ID Fan Deposits
In addition to the cleaning being accomplished by both the
original and the new fan spray systems, it was necessary to
periodically remove the ID fans from service to remove de-
posits by manual methods. In recent, weeks, this phenomenon
has not required the need to clean the blades as frequently.
5. Recycle Pumps
An unsatisfactory degree of wear (about 30% in three months)
is being experienced on the recycle pumps to the point where
pump parts are being diverted from the Elrama scrubber system
C-17
-------
to Phillips at such a rapid rate that the stock of installed pumps
at Elrama will soon be depicted. Experimental pump impellers and
wear rings of three different materials and design will be installed.
6. Thickener Capacity
Modifications were made to the thickener tank overflows to increase
their hydraulic capacity and thus compensate for the increased dust
loading and provide spare capacity.
7. Throat Dampers
Significant unacceptable accumulations are still being found in and
around the scrubber throat damper operating mechanisms. These have
restricted the damper operation and resulted in operation at pressure
drops of up to 10 inches instead of the normal 6 inch drop. Cleaning
requires removal of the train from service for manual removal of the
mechanism and hand cleaning. About 288 manhours per scrubber are
required.
8. Sump Pumps
The ID fan spray sump pumps have been wearing out much faster than
anticipated. This is due in part to the low pH and to the higher
solids concentration in the fan drains. Once again, we have had to
draw on the Elrama installation for replacements.
9. Closed Loop Operation
A closed-loop system may be possible at maximum load with all boilers
cut-in. However, the Phillips Power Station is not a base load station,
and as a result, the load fluctuates between 30% and full load. At the
lower loads, less water is evaporated from the system, while many of the
points of water addition (purge, seal and spray water) continue. This
results in an excess of water in the system. Temporary permission was
obtained from the Pennsylvania Department of Environmental Resources
(PDER) to discharge the excess to our existing bottom-ash settling ponds.
This permits the settling out of particulate matter and dilution of the
blowdown prior to discharge to the river.
10. SOg Guarantee Performance Tests
In July, 1974, several tests were made on the dual stage prototype scrubber
for S02 removal, with inlet S02 concentrations ranging from 1263 ppm vol.
to 1,603 ppm. at stoichiometric ratios from 1.01 to 1.15 the S02 removal
ranged from 86% to 93%. The guaranteed S02 removal is 80%.
SLUDGE TREATMENT & DISPOSAL
At the time the present period of operation started, the sludge
additive system had been completed to the point where it was
considered operable. However, it soon developed that the pumping
system for the additive was operating below design capacity. The
c-18
-------
necessitated pumping a slurry with higher solids content,
and the 750 ft. long transfer lines were frequently out of
service due to plugging.
This stabilization additive chosen for use is Calcilox,
a proprietary product of the Dravo Corporation. This
material was developed in the approximate period of
October through December, 1972, when we re-installed the
Chemico pilot scrubber for the sole purpose of making
sludge. Industry knowledge of sulfate-containing scrubber
sludge and means of stabilization appeared almost totally
lacking at that time. The scrubber was connected so it
could be supplied from either upstream or downstream of
the existing dust collectors of the No. 4 boiler to enable
experiments with both clean and dirty gas. The material
developed, Calcilox, is of the consistency of dry cement and,
on the basis of laboratory tests, has the property of con-
solidating the sludge so that it can be disposed of as land
fill without an objectionable leachate. The degree and time
of consolidation depends on the amount added (3-15% by dry
weight of sludge), the pH of the sludge (10.5 appears suit-
able for our curing time), and temperature (lower temperatures
slow the process). Since actual field leaching tests had not
been made, we installed two membrane-lined ponds in an inactive
ash disposal area to enable collection and retention of leachate
from Calcilox-treated sludge. Bottom drain systems were pro-
vided with valves to enable leachate sampling at intervals.
One of the ponds has 22,860 yd3 capacity, the other 5720 yd3.
These capacities may be increased by 502 if the material proves
suitable for successive layering and compacting.
During the previously noted difficulties with the additive
system, Calcilox was either added sporadically or not added
at all. As a result, the sludge after residence time in the
ponds was of a "soupy" nature, which made it difficult to
excavate with a clam shell and to transport in open trucks.
Prior to placing the material in the trucks, fly ash from the
ash silo was deposited in the rear of the truck to seal the
tailgate. Because of the difficulties in handling the sludge,
additional boilers could not be connected to the scrubber
system since it would tax our ability to remove the sludge
from the ponds as rapidly as it was being produced. Since
the sludge did not have the amount of additive which would
normally be added prior to discharging it to the ponds, no
appreciable degree of stabilization took place in the ponds.
We did not feel that the "soupy" material was representative
of that which would result fron the normal sludge disposal
system, and were reluctant to use one of the lined test
ponds for this material. Therefore, we requested interim
permission from the State's Solid Waste Disposal Division
of the PDER to dispose of the material on the ash disposal
area where we normally dispose of the fly ash and bottom
C-19
-------
from the generating plant. The ^tote consented to our
request, and v/e started disposal r,f the unstabilized sludge
on our ash disposal site. It was spread in thin layers,
mixed with additional fly ash and then compacted.
Handling of the sludge was extremely difficult. The trucks
could only be loaded to approximately 1 '2 cepacity because
of the danger of the sludge flowing over the sides if the
truck stopped or started too abruptly or made turns at
excessive speed. In adJition, because- it was impossible to
prevent sludge from dripping onto the si deb of the truck
during the truck loading operations, it was necessary that
the trucks he washed prior to leaving the plant premises.
Even with such precautions, occasions arose when sludge was
deposited on the roadways. After several warnings from the
local police and the Pennsylvania Pepartmont of Transportation,
it was necessary to rent a street cleaner and a water sprinkler
truck to clean the streets during the hours of sludge hauling.
In approximately May, 1974, modifications were completed on
the additive system, and CalcT.ox was then added on a continu-
ous basis to the sludge with dround-the-clcck technical super-
vision. Filling time per ponri was about- 10-14 days with the
increased quantity of fly ash, instead of the 28-30 days which
had been planned. Similarly, the curing time was reduced from
30-60 days to 14. However, v/ith about 107, Calcilox addition
the sludge did consolidate to the extent that drag-line ex-
cavation was possible, truck tailgates did not require sealing,
and it was possible to fully load the trucks. Observations
of the sludge during excavation showed that although the top
layer resembled hard cloy in consistency, it became earth-like
and then almost fluid near the Lot ton.. It was necessary to
mix the lower-most layer with dryer matfri.il from the upper
layers to obtain a satisfactory hfindlir.1. quality.
The consolidated material was trucked to and deposited in
one of the lined areas for leachate Monitoring. However, the
thixotropic nature of the material was such as to prevent
either level inn or successive layering of truckloads. It
was necessary to tailgate the sludge and then let it stand
(cure?) for about six more weeks before it could bp leveled,
and dozers and trucks did not get ''huny-up." This, of course,
slows the disposal process consi.lorahly.
Our present procedure is to ado consolidated sludge to the
lined area whenever possible. A* other times, it is mixed
and compacted with dry fly asr- or. the normal disposal area.
We have now been monitoring tl.e hcttom drains from the lined
area at weekly intervals for six weeks. In that time, the
total dissolved solids (TDS) have varied from an original
C-20
-------
2203 mg/1 to the present (October 2) 3380 mg/1. A sample analyzed
by the PDER on September 22 verified the results on our own
sample of that same date. Mo record of total volume of leachate
has been maintained, and it is realized that this quantity can
affect the measured IDS appreciably. Similarly, it is realized
that the method of disposal by tailgating and our inability to
level the top of the sludge to a possibly impervious surface
has also had an effect.
In addition to the above tests, leachate studies were con-
ducted on a more closely controlled basis. About June 1,1974,
we used a concrete mixture to prepare four sludge mixtures
with various additives. The additives were fly ash, 2%
hydrated lime, 5% hydrated lime and 11% Calcilox. Each mix
was then added to a plastic wading pool which had previously
been fitted with a valved underdrain system for leachate
sampling. Since the tests started, we have added water to
each pool once each week so as to cover the sludge after
first sampling and draining all water from the previous week.
After fourteen weeks, the most meaningful tests are those with
the 5% lime and the Calcilox. They show that the total dissolved
solids (TDS) on the lime test started at 1450 mg/1 and are now
460 mg/1. The TDS on the Calcilox started at 3370 and are now
2000. Penetrometer readings in that time have increased to
2.7 tons/ft^ on the lime and 4.5+ on the Calcilox. In summary,
while more salts are leached from the Calcilox mix, the degree
of consolidation is higher than that of the lime.
FUTURE OPERATING PLANS
Due to improvements in the method of sludge treatment and
handling, and due to close control of the thickener tank underflow
operation, it was possible to cut-in a fourth boiler to the scrubber
system on August 18, 1974. This increased the capacity connected to the
scrubber system to approximately 180 megawatts. Our plan is to operate
the scrubber system with the four boilers for a period of time to deter-
mine the extent of any aggravation of existing problems and the occurrence
of any new problems. If the problems become intolerable, it may be neces-
sary to return the scrubber system to a three boiler operation. However,
if the problems are not severe, we anticipate the possibility of adding
a 5th boiler to the scrubber system. We are extremely cautious in our
optimism because of the number of unknowns in the sludge disposal system
such as the following:
1. Since the lime feed system is essentially operating at
maximum capacity, additional boilers will reduce the pH
in the scrubber system. We are told that a reduced pH would
adversely effect the stabilization of the sludge. However,
we do not know to what degree it will be affected. Also,
if additional boilers are connected to the scrubber system,
c-21
-------
the rate of sludge production may not permit optimum stabili-
zation period which now appear to be greatly in excess of 14
days.
2. With the oncoming cold winter months, it is not known to what
extent low temperatures will hinder stabilization.
3. If additional boilers are connected to the scrubber system,
and it becomes impossible to optimize stabilization of the
sludge prior to removing it from the pondc, it may be necessary
to again seal the tailgates with fly ash prior to loading the
trucks with sludge. This would then be conducive to sludge
material dropping on to the road, freezing and causing com-
plaints from the local citizens.
If we can connect five out of the six boilers to the scrubber
system by November JO, 1974, we will begin our ten-month test program
at that time to evaluate the efficiency, reliability, and practicability
of the scrubber system. A decision would then be made on September 30,
1975, concerning the installation of additional S02 removal equipment
for compliance with the S02 emission limitations. The number of
additional second stage scrubbers required for compliance with the
regulations will be determined during the ten-month test program, and
this will establish the completion date for installation of all additional
scrubbers.
In the hope of obtaining higher efficiencies of SOp removal with
the use of lime, Duquesne Light is conducting tests using Thiosorbic lime
(a Drayo development) which is a lime having a concentration of magnesium
oxide in the order of 4 to 6%. Various tests on pilot scrubber systems
indicate that the higher reactivity of such a lime results in S02 single-
stage removal efficiencies in the 90's. In preparation for these tests,
one of the single stage scrubbers was isolated along with one of the
thickener tanks. This isolation was necessary to establish and main-
tain the desired chemistry, since lime slaking capacity is insufficient
to maintain the necessary conditions in all scrubbers and both thickeners.
A test run with Thiosorbic lime was made in September, 1974. However,
we encountered difficulties in achieving a sufficiently high rate of
utilization of the magnesium oxide. A few cursory tests under these
conditions indicated S02 removal efficiencies somewhat better than those
with high calcium lime, but also significantly less than anticipated.
The tests were suspended later in September to further evaluate means
of increasing the reactivity and utilization of the Thiosorbic lime.
Resumption is planned for about October 21.
As of August, 1974, construction on the Elrama scrubber system,
which is essentia^y identical to the Phillips scrubber system, had pro-
gressed to a point where no additional meaningful work could be accom-
plished until the problems with the Phillips scrubbers are resolved.
With the modifications and test programs now being conducted at Phillips,
it is anticipated that solutions to the present problems may progress
to the point where startup of the Elrama scrubber system may be possible
in May, 1975.
C-22
-------
AVAILABILITY AND RELIABILITY
The availability of the Phillips scrubber system is meaning-
less at this point since we have essentially 100% backup. In other
words, there are two spare scrubber trains which can be placed in
service whenever a problem arises on an operating train. The service
hours for each train for the period from March 17, 1974 through
September 30, 1974 are 3430 hours for No. 1 train (dual stage),
1678 hours for No. 2 train, 1931 hours for No. 3 train, and 2803
hours for No. 4 train.
Reliability was built into the design of the system which
called for a spare scrubber train enabling us to carry almost full
station load on three scrubber trains. In addition, a degree of
reliability is built into each scrubber train through the instal-
lation of spare pumps, thickener capacity and backup electrical con-
trols.
We do not view bypassing as the solution to maintaining
acceptable reliability. However, we have incorporated a manual type
bypass into our scrubbing system as the best approach to avoid the
serious consequences of scrubber failure. The system uses blanking
plates to redirect the gas flow to the original stacks, and the
boilers have to be shut down during the plate change. Under normal
circumstances, this process would consume several days depending on
the availability of manpower. There really was no alternative in
designing a bypass system for our plants. Mechanical bypass systems
utilizing a damper arrangement require tight dampers, and sufficiently
tight dampers have yet to be developed. We also had serious space
limitations on installation of a mechanical damper system.
CAPITAL AMD OPERATING EXPENSES
As of this date, the scrubber systems at Phillips and Elrama
represent a capital investment of approximately $61 million. It is
expected that an additional $19 million will be required for equip-
ment necessary to comply with the S02 emission limitations. This will
represent a total investment of $80 million or approximately $91 per
kilowatt, exclusive of the cost of any additional sludge disposal facili-
ties required.
A recent estimate indicates that the annual operating and
maintenance expenses, including fixed charges, at Phillips and Elrama
with full S02 scrubbing will be in excess of $30 million per year, or
5.5 mills per kilowatt hour, exclusive of additional sludge disposal
facilities which will be required for full S02 removal operation. These
expenses would represent an increase of approximately $11 per ton of
coal consumed. Expressed in different terms, these expenses, without
fixed charges, would represent approximately a 50% increase over the
actual operating and maintenance expenses for 1973.
C-23
-------
In an effort, to estimate tl',3 t-'U'i capital and operating
costs that would ultimately be incurred for the scrubber systems at
the Phillips and Elrama Stations, including provision of adequate
sludge disposal facilities for full S02 removal, a preliminary
evaluation was made of the additional properfy required for develop-
ment into a disposal area and the additional expenses involved in
disposing of the sludge. This estimate ii\'itares possible capital
expenditures of $110 million for scrubber insiallation and property
acquisition and development, of which 551 mi Hi or* has thus far been
spent. This represents an investment of approximately $124 per kw.
Present estimates of the sludge disposal costs are $7 to
$10 per wet ton, or $15 to $20 per dry ton of sludge.
Future Objectives
Although we feel vie ray be able to overcome the present oper-
ating and equipment problem' with ad'J'tion'il time, effort and expense,
there are five levels of performance which must be satisfactorily re-
solved if this flue gas desulfurization syrtem is to be operationally
feasible and economically acceptable:
1. Reliability - meaning me scrubber ^arility snould meet
the degree of availabi'ity nonr-.l:y e/pected of power
generating equipment.
2. Turndown - the capability of the flue ga? desulfurization
equipment to follow the normal cycling operation of a
power generation facility without series disruption to
both the scrubber system and the qoneivfing plant.
3. Closed Loop Operation - the ability :,i ~,rvr,ite without
the discharge of objectionable li^ui:; Affluents as
dictated by the applicable water quality requirements.
4. Sludge - the technique for disposing of sludge from the
flue gas desulfurization system >«/itliout any adverse
ecological effects.
5. Cost consideration - capital and o>-- Bating expenses should
not be of such a magnitude as \e unreasonable
financial burden on the operator/ owner and the consumer.
Until all these areas of concern are resolved, our system
cannot be considered a successful operation as reports by others have
indicated.
c-24
-------
CALCILOX CALUL.OX
BLPG. SILO
ELECTBiCAl.
SWlTCHGtzAR
BLDG.
AIR POLLUTION CONTROL EQUIPMENT LAYOUT
PHILLIPS POWER STATION
-------
HOT FLUE 64S 71? /If
/£/" STAGS NO 3 I f£CYCLE
>ymt,(-No4 \ *efo
__ --VJtw) SCRUBUL-R. HO 1
_ «U/M
HO 3
,/Vd 3MIS.T ELIM
,NO 4 Miyr ELIM
fl ItAIMKTOI?
!
FUEL COMB
OIL. AIR F
\ \
IA/t^f CrA
-------
DUAL STAG* SCRUBBER TRAIN-PHILLIPS POWER STATION
HOT FI nc r;/\s
COLD FLUE GAS
FUEL COMRUSTION
OIL. A IK
i r
-------
8ULL MOZZLE;
SPRA)
TANGENTIAL NOZZLES
V
ADJUSTABLE THR.OAT
PAMPERS
LOWER CONE
RECYCLE PUMP
SUCT/ON
in
VZNTUZI SCRUBBER SECTION
FIGURE 4 C-28
-------
SIGNIFICANT DATES
December, 1969
September, 1970
October, 1970
February-May, 1971 -
May, 1971
July, 1971
December, 1971
July, 1973
July 9, 1973
October 8, 1973
March 17, 1974
August 26, 1974
October, 1974
Gibbs & Hill, Inc. requested to study methods
of compliance.
Gibbs & Hill study recommends scrubber and
Duquesne Light Company gave authority to
Gibbs & Hill to proceed with projects at
Elrama and Phillips Power Station.
Invitations were extended to bid on scrubber
pilot plant.
Chemico pilot scrubber tests at Phillips under-
way.
Des-ign of full scale system started.
Contract for scrubber process including scrubber
recycle pumps, reheater and associated controls
and instrumentation awarded to Chemico.
Construction started at both Phillips and Elrama
plant sites.
Construction of Phase I at Phillips essentially
complete. Phase I consists of one dual stage
S02 scrubber train and 3 single stage particuldte
scrubber train with associated lime and thickener
equipment.
Operation began at Phillips with two and eventually
three of station's six boilers.
Scrubber system shutdown for extensive repair.
Scrubber system began with three of six boilers.
Fourth boiler cut into scrubber system.
Fifth boiler cut into scrubber system.
Figure 5
C-29
-------
TECHNICAL RCPORT DATA
r\ ,7i7 litiinir!i
------- |