EPA-600/2-76-249b
September 1976
Environmental Protection Technology Series
                  CHARGED  DROPLET  SCRUBBER FOR
                             FINE  PARTICLE CONTROL:
                                PILOT  DEMONSTRATION
                                 Industrial Environmental Research Laboratory
                                      Office of Research and Development
                                      U.S. Environmental Protection Agency
                                Research Triangle Park, North Carolina 27711

-------
               RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection  Agency,  have  been grouped  into five  series. These five broad
categories  were established to facilitate further development and application of
environmental technology.  Elimination of traditional grouping was consciously
planned to  foster technology transfer and a maximum interface in related fields.
The five series are:

     1.    Environmental Health Effects Research
     2.    Environmental Protection Technology
     3.    Ecological Research
     4.    Environmental Monitoring
     5.    Socioeconomic Environmental Studies

This report has been  assigned  to the  ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and  non-point sources of pollution. This
work provides the new  or improved technology required for the control  and
treatment of pollution sources to meet environmental quality standards.
                    EPA REVIEW NOTICE

This report has been reviewed by  the U.S. Environmental
Protection Agency, and  approved for publication.   Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield. Virginia 22161.

-------
                                   EPA-600/2-76-249b

                                   September 1976
  CHARGED DROPLET SCRUBBER

  FOR FINE PARTICLE CONTROL:

       PILOT DEMONSTRATION
                     by

         W. F.  Krieve and J. M. Bell

    TRW Defense and Space Systems Group
               One Space Park
      Redondo  Beach, California 90278
           Contract No. 68-02-1345
            ROAPNo.  21ADL-043
         Program Element No.  1AB013
    EPA Project Officer;  Dale L.  Harmon

 Industrial Environmental Research Laboratory
   Office of Energy, Minerals, and Industry
      Research Triangle Park, NC  27711


                Prepared for

U.S. ENVIRONMENTAL PROTECTION AGENCY
      Office of Research and Development
           Washington, DC 20460

-------
                                  ABSTRACT
This report presents the results of a successful  Charged Droplet Scrubber
(CDS) pilot demonstration of coke oven emissions  control.   It includes a
description of the design, installation and checkout of the demonstration
system.
The CDS is a device that uses electrically sprayed water droplets, accelerated
through an electric field to remove particulate material from a gas stream.
The pilot demonstration was a continuation of the laboratory and bench scale
studies for application of the CDS to fine particle control.  The pilot demon-
stration system included, in addition to the CDS, the ducting, flow transitions
and blower necessary to circulate process gas through the  CDS.
The test was performed at the Kaiser Steel Company coke oven facility, Fontana,
California.  A large fraction of the coke oven emissions were submicron and
composed of carbon particles and hydrocarbon aerosol.  After the system check-
out was completed during which the CDS operating parameters were established,
the demonstration test series was performed.  Results of the demonstration
test indicates that the CDS is an effective pollution control device for con-
trolling coke oven stack emissions.
This report was submitted in fulfillment of Contract Number 68-02-1345, by
TRW Defense and Space Systems Group, Preliminary Design and Fluid Systems
Department, under sponsorship of the Environmental Protection Agency.   Work
was completed as of March 1976.

-------
                              ACKNOWLEDGMENTS
The support of Kaiser Steel  Company, Fontana, California  personnel  is
acknowledged with sincere appreciation.   Special  thanks is extended  to
Messrs. C. Kingsbury and S.  Vitt of the  KSC  Engineering Department  for
their assistance.
                                      IV

-------
                                CONTENTS


                                                                 Page No.

  I.   CONCLUSIONS 	      1

 II.   INTRODUCTION  	      2

III.   DEMONSTRATION TEST DESIGN AND ASSEMBLY  	      5

 IV.   DEMONSTRATION TEST CHECKOUT AND STARTUP 	     16

        Design Verification Tests 	     16
        Startup	     18

  V.   DEMONSTRATION FIELD TEST	     20

        Sampling Procedures	  .  '   20
        Screening Tests 	     24
        Matrix Tests  	     25
        Development Tests 	     28
        Long Duration Tests	     29

 VI.   DISCUSSION	     32

        Methods of Analysis	     32
        Process Characterization  	     32
        Results of Analysis 	     34

      APPENDIX A

      APPENDIX B

      APPENDIX C

-------
                             LIST OF FIGURES


Figure No.                                                         Page  No.

    1        Kaiser-Fontana Coke Oven Battery A Stack  	       5

    2        CDS Structural Configuration 	       8

    3        CDS Housing with Header Electrode Support
            Insulator Column Exposed 	       9

    4        CDS Inlet Turning Section  	      10

    5        Pipe Nest Component Layout and  Electric
            Circuit Schematic  	      12

    6        Plan View of CDS Pilot Demonstration  System   ....      14

    7        Flow Distribution at 2.80 m/s Average Velocity  ...      17

    8        Modified E.P.A.  Sampling Train  with  In-Stack
            Cascade Impactor 	      22

   10        Gas Flow Velocity Distribution  After  Completion
            of Test Series   	      30

   11        Gas Maldistribution Sensitivity to Specific
            Collecting Area-RMS Velocity Profile  Vs
            Average Velocity   	      31

   12        Coke Oven Flue Gas  Process Variability, the
            Relationship of Inlet Dust Load to Particle Size  .  .      33

   13        Inlet and Outlet Size Distributions at
            Low Inlet Load   	      35

   14        Inlet and Outlet Size Distributions at
            Low Inlet Load   	      36

   15        Inlet and Outlet Size Distributions at
            Low Inlet Load   	      37

   16        Inlet and Outlet Size Distributions at
            High Inlet Load	      38

   17        Inlet and Outlet Size Distributions at
            High Inlet Load	      39 .

   18        Efficiency Correlation,  Total Unedited
            Data Set With  Continuous Wall Wash	      40
                                    vv

-------
                       LIST OF FIGURES (CONTINUED)
Figure No.                                                        Page No.

   19        Efficiency Correlation, Total Unedited
             Data Set Without Wall Wash	 .  .  .     41

   20        Effect of Gas Flow Rate on Overall
             Collecting Efficiency . .  .	     44

   21        Insensitivity of Efficiency to Second
             Stage Electrode Voltage 	     46

   22        Fractional Efficiency Data for the Higher
             Inlet Loadings	     49
                                    vii

-------
                            LIST OF TABLES

Table No.                                                       Page No.
    1        Coke Oven Exhaust Stack Gas Composition 	    5
    2        Kaiser Steel Water Sample Analysis  .	    6
    3        CDS Design Summary	  .    7
    4        Summary of Sampling Procedures Employed 	    21
    5        Test Matrix Parameters	    26
    6        Suspect Kaiser CDS Data	    43
    7        Summary of Regression Analyses  	    47
    8        Kaiser Observed CDS Stack Opacities 	    51
                                  vi i i

-------
                            I.   CONCLUSIONS

     The Charged Droplet Scrubber (CDS) demonstrated effective control
of the emissions from a coke oven battery over widely fluctuating process
conditions.  Particle removal efficiencies up to 95% were measured and
were an increasing function of the time averaged particle loading.
Improvements in the gas distribution internal to the equipment should
result in additional improvements in collecting efficiency.
     The average inlet particulate load varied between 0.05 and 0.33 gr/
dscf, associated aerodynamic mean diameters varied between 0.4y (hydro-
carbon aerosol) and 1.5y (carbon black).  The most sensitive design
variable affecting efficiency was the gas volume flow rate through the
equipment.  Low total energy and water consumptions, 0,8-1,2 watts/acfm
and 0.8-1.0 gal/1000 acfm, respectively were demonstrated over most of
the test conditions.  Operation of the CDS with intermittent (8 hour cycle)
collector plate over sprays was adequate for deposit control.  The CDS
gas discharge was not saturated during nominal operating conditions.
                                   -1-

-------
                            II.  INTRODUCTION

     This report describes the results of a pilot demonstration of the
Charged Droplet Scrubber (CDS) to control coke oven flue gas particulate
emissions.  Heretofore, it was the general concensus of the industry that
there was no suitable control technology for this process because of the
wide process fluctuations.  Emissions consisted of varying relative concen-
trations of submicron sticky hydrocarbon and micron sized high conductivity
carbon black.  The CDS applies electrohydrodynamically sprayed water drop-
lets to remove particulate material from a gas stream.  The droplets have
a size in the range of 60 to 250 ym in diameter and have a surface charge
density near that allowed by surface tension forces.  The charged droplets
are accelerated through the gas stream by an applied electrostatic field.
     The objectives of the pilot demonstration were to verify the appli-
cability for removing fine particles from an industrial effluent stack,
to determine the influence of CDS operating variables on performance and
behavior of the CDS under long term operation.  The demonstration unit
was to be of sufficient size, at least 17,000 m3/hr (10,000 acfm), to
adequately describe the behavior of a full size unit.  The emission source
was to be characteristic of those requiring control with a relatively large
fraction of particulate material in the submicron range.
     The pilot demonstration phase is a continuation of the laboratory and
bench scale studies for the application of the CDS to fine particle control.
These studies included an analysis of the particle removal interactions between
particulate material and charged droplets.  The laboratory scale studies included
the determination of charged droplet characteristics under system operating
conditions.  The results of these studies were used to verify some of the models
used in the fine particle removal analysis.  The particle removal efficiencies
of a small size CDS operating under simulated process conditions were measured
during the bench scale studies.  The results of these tests indicated that the
CDS should be effective for fine particle control and was sufficiently devel-
oped for a pilot demonstration test.  The laboratory and bench scale studies
are reported in Reference 1.
                                    -2-

-------
A pilot demonstration phase was started at the conclusion of the first phase.
The tasks of the pilot demonstration phase are:
     t  Site Selection and Hardware Preparation
     •  Pre-Delivery Design Verification Tests
     •  Pilot Demonstration Field Test
     •  Summary and Conclusions

 The  site selection and  hardware  preparation  task covered  all of the back-
 ground work necessary for conduction  of the  demonstration program.  The
 selected site  was  to provide  an  environment  consistent with  the objectives
 of the program.   It must be adaptable for installation of a  CDS and provide
 the  required utilities  within a  reasonable time span.  An acceptable  working
 agreement with the organization  providing the site was necessary to allow
 successful  completion of the  demonstration test.  A  test  matrix was developed
 as part of this  task to delineate  the CDS operating  parameters  that would be
 varied during  the  test  phase.
 The  hardware preparation portion of the first task included  the design and
 fabrication of the CDS, duct  work  and supporting structure for  the demon-
 stration unit.   All  interfacing  requirements  between the  site and demon-
 stration unit  were established.  Any  special  auxilary equipment was identified
 and  either purchased or fabricated.
 Prior to delivery  to the test site,  the CDS  was subjected to design and oper-
 ational verification tests.   These tests  of  Task 2 were  designed so that the
 CDS  would be operational when delivered to the test  site  with a minimum of
 rework or retrofitting.   The  checkout of  the CDS included high  voltage
 component electrical breakdown tests, electrode water flow distribution
 and water flow rates, and pressure and power requirements.
 Task 3, Pilot  Demonstration Field  Tests,  covers the  on site hardware  assembly,
 checkout  and  operational  testing  of  the  CDS.  Hardware  checkout and  parameter
 checks such as gas flow distribution  proceeded simultaneously with the assembly.
 This would allow any measurements  or  modifications to be  made with a  minimum
 of reassembly.   Final checkout of  the unit was made  after complete assembly
                                    -3-

-------
and connection to the process stream.  The testing includes screening tests
to determine the operating parameter levels of the CDS for the test matrix.
Following the performance demonstration tests is a long duration test, 500
hours, to identify maintenance requirements and potential failure modes.
The final task, of which this report is a part, includes documentation of
the design, fabrication, assembly and checkout of the demonstration unit and
a summary of the test results.
                                   -4-

-------
              III.  DEMONSTRATION TEST DESIGN AND ASSEMBLY
The site selected for the CDS pilot demonstration test was  a coking oven
battery exhaust stack at Kaiser Steel  Co.  in Fontana,  California.   A photo-
graph of the stack site is shown in Figure 1.   The emissions as  originally
specified consist of hydrocarbon aerosol  and carbon particles of which  42
percent by weight are less than one pm in  diameter and 96 percent  by weight
are less than ten pm.  This particle size  range was consistent with the
requirements for the demonstration test.   The nominal  effluent temperature
is 400°F, and the gas composition as measured in May of 1974 is  shown in
Table 1.
           Figure 1.  Kaiser - Fontana Coke Oven Battery A Stack

             Table  1.  Coke Oven Exhaust Stack Gas Composition
                 Constituent               Volume Percent
                   CO,
                   H20
                   CO
                   S00
                   NOX as
  4.2
 12.6
 15.0
 95 ppm
300 ppm
102 ppm
                                      -5-

-------
An existing port used during previous experiments allowed access through the
stack wall.  The necessary utilities were available either at the stack lo-
cation or could be supplied through existing piping and conduit.  There were
two water sources available to use in the experiment.  One was recycled
industrial water and the other domestic or fresh water.  Average values of
the water properties taken during the months of June and August 1974 are
shown in Table 2.
                Table 2.  Kaiser Steel Water Sample Analysis
    Source                Conductivity           Hardness         pH
    	                ymho/cm                ppm	      	
    Industrial
      6/74                   572                   137           7.51
      8/74                   635                   152           8.33
    Domestic
      6/74                   214                    58           8.45
      8/74                   169                    57           9.02

Either of these water sources, based on the tabulated properties, could be
used without conditioning in the scrubber.
The particulate material properties and loading varied with time during the
coke oven operating cycle.   A typical cycle included loading an individual
oven with coal, discharging coke, and coking with all ovens in the battery
closed.  The time period during which coal was charged and coke discharged
constitute a high stack emission condition.  A nominal operating period
occurs when all ovens are closed.  The previously measured particle loading
was 290 mg/m3  (0.1 gr/ft3) and the stack Ringlemann 5 during a high emission
condition.  It was found during the test phase of this program that under
these conditions particle loading could be as much as a factor of 3 higher.
Under normal operating conditions, the particle loading is of the order
of 1.37 mg/m3  (0.006 gr/ft3) and the stack Ringlemann 0 to 1.

These sufficiently higher inlet loads resulted in a requirement to derate
the equipment flow capacity to achieve higher overall collection effi-
ciencies.  As will be discussed later, this resulted in a substantial
gas maldistribution problem and some loss in efficiency.
                                   -6-

-------
The Charged Droplet Scrubber to be used on the program was purchased
from the Development & Applications Division of TRW Inc.  The unit was
a cost-effective modified version of a Model 300 with a capacity of
51,000 m3/hr(28,500 acfm) at 1.83 m/s (6 ft/s) gas flow rate.  An iso-
metric sketch of the CDS is shown in Figure 2 and a photograph of the
delivered unit in Figure 3.  A design summary is shown in Table 3.
The scrubber contained three electrostatic spraying stages arranged in
series with parallel collecting plates on 0.127 m (5 inch) centers. It
had 19 collecting modules.  The scrubber structural members and collector
plates were fabricated from mild steel.  Material thickness of the scrubber
housing and collector plates was 20.3 mm (0.080 inches).  Although the
compatibility problem of mild steel in the stack gas environment was recog-
nized, it was felt that the material would maintain its integrity during
the test period.  The electrodes which distribute high voltage and water
were fabricated from type 316 stainless steel tubing.  Each electrode stage
was supported from parallel main headers mounted exterior to the gas pass-
ages and supported on corner insulator posts.  A series of 12 doors, 6
on each side, was provided to allow access to the electrode headers and
electrodes. Inspection and alignment of all electrodes on the three stages
and the collector plates could be made through these doors.
                    Table  3.   CDS  Design Summary
            •   Three high  voltage  scrubbing  stages with  0.127 m  (5  in.)
               collector  plate  spacing.
            •   Nineteen  collecting modules,  3.05  m  (10  ft)  long.
            •   Flow  cross  sectional  area,  7.36 m2  (79.2  ft2).
            •   High  voltage  electrode,  type  316 stainless  steel  tubing
               190 mm (0.75  in.) diameter,  flattened  to  0.127 mm (0.5 in.)

            •   High  voltage  electrodes  contained  67 spray  tubes  each on
               44.5  mm (1.75 in.)  centers.
            •   Spray tubes,  titanium with  a  12.7  mm  (0.050 in) O.D.  by
               1.52  mm (0.006 in)  wall  and  protruding  25.4 mm  (1.0  in)
               from  the  electrode.
            •   Collector  plates  3.05 m  (10  ft.)  long  by  1.83 m  (6  ft.)
               high  by 20.3  mm (0.080 in)  thick mild  steel.
            •   Wall  wash  system  covering each  collecting surface.
                                    -7-

-------
                                  COLLECTOR PLATES
I
CO
FLANGED
GAS EXIT
                     ELECTRICAL
                     UPPER SECTION
           HIGH TENSION
           SUPPORT
           HOUSING
          GAS
          DISTRIBUTION
          LOWER
          SECTION
                            FLANGED
                            GAS INLET
                                                   HIGH TENSION
                                                   WATER HEADERS
                                                                                SLURRY
                                                                                DISCHARGE
                                                                                                        HIGH TENSION
                                                                                                        CONNECTOR
                                                                                                        PANEL
                           OVERFLOW
                                                                                                                      MAINTENANCE
                                                                                                                      PLATFORM
                                               Figure  2.   CDS Structural  Configuration

-------
                                                       GAS EXIT
HEADER ELECTRODE
COMPARTMENT
ACCESS DOORS
    INSULATOR COMPARTMENTS
HEADER ELECTRODE
SUPPORT INSULATOR
            Figure 3.   CDS  Housing with Header Electrode Support Insulator Column  Exposed

-------
A turning section,  shown  in  Figure 4, was also part of the  scrubber unit.
This section was designed to accept a horizontal flow duct,  deflect the flow
into a vertical direction and distribute it uniformly over  the  scrubber cross
section.  The  turning  section contained right angle vanes to deflect the gas
flow and moveable baffle  elements to distribute the flow.
                                   „ MATING FLANGE TO
                                     CLEANING SECTION
SUPPORT FOR ADJUSTABLE
ANGLE IRON FLOW
                 Figure 4.  CDS Inlet Turning  Section
                                    -10-

-------
Electrode feed water is brought up to high voltage through a long  length
of non-conducting pipe. The piping is sized to restrict the leakage current
through the water to an acceptable level  when high voltage is applied at  the
electrode end and the water supply end grounded and to have a tolerable
pressure drop.  The pipe lengths are supported or nested as close  as possible
without developing electrical breakdown gradients between the lengths to
minimize the containment volume.  The pipe nest was enclosed in a  sealed
housing separate from the scrubber housing.  Copper tubing, extending
through penetration insulators in both the scrubber shell and pipe nest
housing, was used to connect both the water and high voltage to the scrubber
electrodes.  This pipe nest was formed from thirteen 1.83 m (6 ft.) lengths
of Schedule 80 polypropylene pipe interconnected with pairs of elbows.

Isolation resistance between the stages was provided through three nominally
0.38 m (15 in) lengths of 6.35 mm (.25 in) I.D.  by 6.35 mm (.25 in)  wall
Tygon tubing.  The tubing was used to complete the water flow path from the
end of the pipe nest to the copper tube extension of the individual  stages.
The high voltage input was connected upstream of the three flow distribution
tubes.  Therefore, the current path to each stage was through the  water in
the connecting tubes.   The water resistance provided the electrical  isolation
between stages.  A schematic of the pipe  nest and electrical circuit is shown
in Figure 5.
A purge fan system was installed on the scrubber to provide an ambient air
flow over the stage electrode support and penetration insulators.   This air
flow reduced direct contact of the insulators with the process stream gas
eliminated fouling.  The air purge fan had a design capacity of 1020 m3/hr
(600 CFM) at a 25.4 mm (1.0 in) water pressure differential.  The  purge fan
and duct was located external to the scrubber housing with the inlets to  the
scrubber on the top end of each support insulator housing.
The high voltage source was a transformer/rectifier set with a maximum output
of 75 kV at 400 ma and a single phase 480 v Input.  The output voltage from the
transformer/rectifier set could be controlled either manually with a potenti-
ometer or by an automatic control circuit using an arc rate sensor for feed
                                  -11-

-------
WATER LEAKAGE
CURRENT MONITOR
A CARBON STEEL ELL
x AND STREET ELL
* Fll TF» TAPArlTriB
INLET J_ ' • . y$ • \

V f
HIGH VOLTA

\ I ARC EXIINCU|SHINC CAPACITOSS (3 PL/
> POLYPROPYLENE PIPE
/ *?' ?
/
^ 1 1
^ 	 : 	 o 	 c ' H
' ; ^=a 	 ^L'r^
TYGON TUBING ^ C |
• (3 PLACES) 1
POTENTIAL DIVIDERS 	 '
. (3 PLACES) ^^^n— f ||-
VOLTAGE 	 	 1 r^ r
MONITORING ^ i (
LEADS -i 	 N }-"•
(3 PLACES) JL I |
PENETRATION ^^ |]
INSULATORS •"
; /3 pi ACES1 WATER A
PIPE NEST COMPONENTS ' ' LEADS TC
HIGH VOLTAGE INPUT
f
-j,
li
*tr-
m
i T
MO POWER
STAGES
GE LEAD
ICES)
	 TYGON TUBING
(3 PLACES)
FILTERING T
CAPACITOR -ir
' ' ' ^- ARC EXTINGUISHING
^^"^ CAPACITORS POWER TO ELECTRODES
^^-^ (3 PLACES) ^^^ (3 PLACES)
MONITOR J=*» — CAPACITOR DROPPING i^~ 	 POTENTIAL DIVIDER
II' WATER SUPPLY ISOLATING RESISTANCE "i /^RESISTANCE % (3 PLACES)
1 \ / <3PLACES) ' t 	 ' — X°^!/"ONIT0'
vVAWWAWAW ^ AAWA-J^ 1 • f
LEAKAGE CURRENT ^ "T J -Jr
SHUNTING RESISTANCE ^* -±- 4
• . . STAGE ISOLATING ^ - I ,
• RESISTANCE • I :
n n f,r^\ ,rt^nnnnn^ • ™/- ... *. • >

PIPE NEST ELECTRICAL CIRCUIT
     Figure 5.   Pipe Nest Component Layout and Electrical
                 Gfrcuit Schematic
                           -12-

-------
back control.  The voltage when in the automatic mode would be either the
maximum of the transformer/rectifier set or that required to maintain a pre-
set arc rate.  An rf coil located within the scrubber housing was  used to
sense the arcs.  The operating voltage on the electrodes was the output of the
transformer/rectifier set minus the voltage drop across the stage  isolating
resistors.  Voltage on each stage was monitored through a potential  divider
located on the input to the stage.  The total scrubber current was monitored
directly on the ground return leg of the transformer/rectifier.  Leakage
current through the pipe nest was monitored with a current shunt near the
pipe nest ground end.  These monitoring circuits are shown in Figure 5.
The structural support and ducting design along with the final working drawings
for the CDS  installation was made by Trade West of Corona, California.  Trade
West was the manufacturer of the  CDS.  A top view of the scrubber experimental
layout is shown in  Figure 6.  Kaiser Steel Manufacturing Division performed
the structural and  ducting fabrication and installed all components of the
demonstration  test  unit.  A portion of the ducting used was salvaged from the
previous experiment at the stack  location.  These items included a 1.52 m
(5 ft) diameter flow control damper and a 50 horsepower blower used for draw-
ing gas from the stack and exhausting it through the scrubber.  The blower
had a capacity of approximately 68,000 m3yhr(40,000 ACFM) at a draft of 100 mm
(4 in) of water.  The nominal vacuum in the stack at the position gas was
withdrawn was  38 mm (1.5 in) of water.
The ducting  from the stack to the blower turning and transition section was
1.52 m (5 ft)  in diameter.  The straignt duct section between the elbow at
the stack and  the inlet to the blower transition was approximately 3.66 m
(12 ft) long.  It was angled approximately 20° relative to the horizontal to
accommodate  the difference in elevation between the ground level blower inlet
and the stack  port.  Gas stream pre-cooling water spray nozzles were located
0.3 m (1 ft) up stream of the blower transition.  The pre-cooling system con-
tained 10 spray nozzles located on the upper 300° section on the duct periphery.
CDS inlet gas  stream sampling ports were located in the duct straight section
2.9 m (9.5 ft) downstream of the elbow from the main stack.  The two sampling
ports were located  on the top and side of the straight duct section.  Each
port was equipped with an extension structure to support the sampling trains.
The position of the pre-cooling system and inlet sampling ports are shown in
Figure 6.

-------
                                 TRANSFORMER
      LADDER
                                                                                440V-150 AW
                                                                                ELECTRICAL PANEL
Figure  6.  Plan View  of CDS  Pilot  Demonstration  System

-------
The flow control damper was bolted directly to the stack outlet flange.  It
was not designed for complete flow cut-off; therefore, during scrubber non-
operating periods, there was a back flow of ambient air through the scrubber
into the stack.  The damper was pre-set manually to control  the flow rate
through the CDS during test operation.  Because of the fan capacity, the
damper was only partially opened during CDS operation.

A flow  scoop  extended  into  the stack  to assist  in diverting gas flow into  the
scrubber  ducting  system.  The scoop was a  half  section of a 1.52 m  (5  ft)
diameter  pipe.   It was  angled at  15°  downward from the horizontal into the
stack and extended inward 3.05 m  (10  ft).  Although the scoop may not  have
been necessary to insure adequate  flow through  the CDS because of the  fan
capacity,  it  would help provide a  more representative stack gas sample.
The  transition section  between the blower  and scrubber turning section con-
tained  four equally  spaced  straightening vanes  to assist in distributing
the  gas flow  across  the scrubber  cross section.  Gas sampling ports were
located in each of the  five flow  passages  formed by the vanes and transition
walls.  These sampling  ports were  used by  APT,  Inc. of San Diego, California
who  performed the designated EPA  sampling  tests.  The scrubber inlet gas
temperature monitoring  thermocouple was located in this blower transition
section.
The  outlet transition  section of  the  CDS was a  quadrihedron, truncated by
0.91 m  (3 ft) diameter  vertical exhaust stack.  The exhaust stack was  3.05 m
(10  ft) long.  Its exhaust  plane was  12.2  m (40 ft) from ground level.  Three
sampling  ports, located in  a horizontal plane, were in the exhaust  stack,
three feet below  its exit plane.   Two of the ports were at right angles and
the  third was in  their  enclosed quadrant.  This port arrangement allowed
simultaneous  outlet  sampling by the APT group during a test.
A single  overhead catwalk was incorporated into the scrubber support structure
for  access to the outlet sampling  ports.   Temporary scaffolding was erected
for  access to the CDS header electrode compartment doors and the vertical
inlet sampling port.  The temporary scaffolding was also used during scrubbing
installation  for  access to  various areas for alignment and checkout.
                                    -15-

-------
               IV.  DEMONSTRATION TEST CHECKOUT AND STARTUP
 The  fabricated equipment was first assembled and subjected to factory
 functional electrical and hydraulic tests.  After installation additional
 dynamic tests were performed and modification work identified as described
 below.
 Design Verification Tests
 The  CDS unit was subjected to design verification tests prior to
 delivery to the installation site.  These included electrical breakdown
 tests of the high voltage components, both at the scrubber support and
 penetration insulators and within the pipe nest.  Electrode and wall wash
 water flow distribution were also visually observed.   The high voltage
 isolation tests indicated that the electrical isolation design was capable
 of withstanding the peak voltage stresses expected during operation. The
 pipe nest provided an acceptable isolation from ground with leakage currents
 in the range of 10 to 12 percent of the total scrubber current.   The arc
 rate between the spray tubes and collector plates appeared to be excessive
 at the CDS operating voltage range.  A part of the arcing problem was due
 to collector plate misalignment.  There was adequate  collector plate ad-
 justment to allow proper alignment; however, realignment of the  plates was
 deferred until after installation at the test site.
 The water spray pattern of both the electrodes and wall wash were adequate
 for proper scrubber operation.   The wall wash gave no indication of the
 impending problems that were to be encountered later  in the program. The
 nozzles maintained a steady spray pattern on the collector plates with
 no splashing.   The wall wash was operated on domestic water without
 filtering.   After completion of these tests, the CDS  was shipped to
 the test site for installation.

 Typical profiles of the gas flow, taken after the distribution baffles
were adjusted, are shown in Figure 7.   The profiles at an average velocity
 of 2.70 m/s (8 fps), are marginally outside the design requirements of
 the scrubber.   As the anticipated operating average gas velocity was
 between 1.5 and 1.7 m/s, no further gas velocity adjustments were attempted
at this time.   It is necessary  to maintain the best possible uniformity
 in gas flow distribution across the CDS cross sectional area to  achieve
                                   -16-

-------
                                                                GAS FLOW VELOCITY (M/S)
                         — • tO NJ CO CO CO
CQ
c
 VI
 rh
 O
 3

 0)
 ft)
 -S
 O)
(Q
 fD
 ft)
 O
 O
 I  5
 C  m
 —  5

 ?  I
 s  S
 O  m
 5  5
 —<  ^
 c  •<
 z
1  + NJ
NJ*. N
Oi — O
393? S
    8
                                    s 8
                                        —i NJ Ni CO  CO CO

                                        8 8 3 5  k 8
                                                                      —• tO IO CO CO CO
Cn


o
                     *
                 z


                 00
                 pn
                     CO
in
                     00
                                                                                                                   — • NO NJ CO CO CO
s
s   s
                                                                                                                                      o
                                                                                                                 o

                                                                                                                 —i
                                                                                                                 »


                                                                                                                 o
                                                                                                                 CO
                                                                                                                 m
                                                                                                                 O

                                                                                                                 5
                                                                                                                 z
                                                    CO
                                                DISTANCE FROM INLET END OF TURNING SECTION (M)

-------
high particle removal efficiencies.  Particle removal efficiency has a
non-linear inverse dependence on gas flow rate; therefore, large variations
in flow velocity in the scrubber cross section will result in a reduced
efficiency.

The high voltage was applied to the electrodes for one brief period prior
to completion of the installation to check the integrity of the high
voltage isolation.  No sustained high voltage operation could be made until
after completion of the installation when all high voltage elements were
completely enclosed.  After installation was complete, specific high voltage
tests and adjustments were made.  The tests included locating potential
breakdown paths in the isolation, determining maximum operating voltages
on the electrodes and monitoring corona and leakage currents.  The adjust-
ments included setting the arc rate controls for automatic mode operation
and establishing the water tube lengths which served as the dropping resis-
tors to the stages.

Several other site specific, minor operating problems occurred due to the
inadvertent hookup of the equipment to the wrong water source. Feed water
filter plugging due to abnormally high suspended solids concentrations
and inadequate supply pressures were corrected by hookup to the domestic
water supply.

Also, at this time the OEM tygon tubing in the pipe nest was replaced with
polypropylene pipe as recent commercial experience in Japan indicated this
to be a more stable material in a high electrical  field gradient and
temperature environment.
Startup
The CDS was then put on process and marked change in high voltage electrical
stability noted as compared to operation with ambient air.   Two effects were
noted which potentially could have deteriorated collecting efficiency:
         •  Abnormally high spark rate in the first electrostatic
            spraying stage.
         t  Sustained arcing without sufficient quenching,  primarily
            in the first stage.
                                    -18-

-------
The high electrical sparking resulted from the presence of a low mobility
particle space charge, primarily within the first or inlet spraying stage.
The current from an electrode will decrease with increasing low mobility
space charge resulting in the first stage having the smallest voltage drop
across its isolating resistor.  Under these conditions the first stage
controls the operating voltage on the other stages.   The dropping resistors
to each stage were resized so that each successive downstream stage had
a higher nominal operating voltage.

The CDS would operate in a self-extinguishing arcing mode, except during
periods of heavy particle loading.  When the particle loading was heavy
some arcs appeared to become sustained burning arcs.  These sustained arcs,
actually a series of arcs, resulted from reignition of the plasma left from
the previous arc.  The sustained arc mode was eliminated by changing the
recharge rate, i.e., adding capacitance to each stage.

The gas temperature conditioning spray system was close coupled to the
blower in order to position the sampling ports as far as possible from the
stack outlet elbow.  Unvaporized water droplets were ingested into the fan,
but were demisted by the fan blades.  Although this did not create an
operating problem, the water mass balance was more difficult to estimate.
The unvaporized water was drained from a port in the blower inlet trans-
ition and the net water flow vaporized was determined measuring both the
preceding and overflow water flow rates.  The precooler was capable of
reducing the gas stream temperature up to 80 C (175°F).
                                   -19-

-------
                    V.  DEMONSTRATION FIELD TEST
The objective of the field test was to determine the operating characteristics
and particle collecting efficiency of a CDS under actual process conditions
as a function of operating parameter levels.  The parameters included:
         •  Electrode voltage
         t  Gas stream velocity
         •  Electrode water flow rate
         •  Electrode polarity

Sampling Procedures
Particle removal efficiency was determined by isokinetically and simultaneously
measuring the inlet and outlet particulate material loadings during a test
run.
Three different sampling procedures were used to measure particle collection
efficiency.  During the statistically designed tests both the EPA Method 5
procedure (Reference 2) for overall efficiency and a modified procedure
using a Washington State impactor for fractional efficiency were employed.
Additional testing was performed with a shortened tes" period using EPA
Method 5.  The three procedures are described in more detail in Table 4
and the modified APT sampling train shown in Figure 8.  Greased aluminum
substrates were used in the impactors to prevent particle bounce and
minimize wall losses.  The gas stream temperature and kinetic pressure
head were measured at each sampling point with the thermocouples and S-pitot
tubes on the probes. These values were integrated over the duct areas to
determine mean temperature and volumetric flow rates.  The temperature
monitored at the outlet sampling port corresponded to the outlet temperature
of the CDS.  The temperature measured at the inlet sampling port was near
that of the gas stream in the coke oven battery main stack.  The inlet gas
temperature to the scrubber was measured with a thermocouple located down-
stream of the gas precooler and blower.   At the CDS inlet, eddy mixing and
condensation in the pi tot tube was evident indicating negative velocity heads
and recirculating gas flow.   The best one point sampling location was used
during the impactor runs.   The eddy mixing would indicate that the inlet
gas velocity average and the gas flow rate (volume/time) are questionable

                                   -20-

-------
                                            TABLE 4

                                       SUMMARY OF SAMPLING

                                       PROCEDURES EMPLOYED
i
ro
          Method
    1.   EPA #5
2.   EPA #5

3.   Modified
    EPA #5
                   Equipment
                  Joy Emission
                   Analyzers
                      Washington,
                      Mark III
                      Impactor
 Sample Location
 Inlet    Outlet

Circular Circular
Duct,Up-  Stack
Stream of
Fan
Rectang- Circular
ular Duct, Stack
CDS Inlet
  Number of
Sample Points
Inlet  Outlet
 32'
32"
                                                       16-32^  16-32'
Sample
Period
 Min.

90-120
Organ-
ization

  TRW
                   30-90      TRW

                   30-45      APT
    Notes
     'Performed along two radial traverses orthogonal to each other.
    2,
      Performed along one or two radial traverses,  sometimes on a coarser  grid.

-------
i
ro
ro
i
                         CASCADE
                         IMPACTOR
                                                        IMPINGER TRAIN
                                 STACK
                                 WALL
  I
  I	;	I-CJEJB Ajj L.
THERf--10.MET.ERS        T
                                                                 	J


o
>
                                                                    ROTOMHTP.R
                                                                     VACUUM
                                                                     GAUGE
                       ORIFICE METER    DRY GAS METER  VACUUM
                                                        PUMP
             Figure 8.  Modified E.P.A.  Sampling Train with In-stack Cascade Impactor

-------
based on the inlet traverse.  This also was the cause of poor gas distri-
bution internal to the CDS.   The outlet port was located three duct
diameters downstream of the nearest disturbance and one duct diameter
upstream of the stack outlet.  Velocity traverses of the outlet revealed
fully developed flow profiles.

In addition to the CDS operating parameters, other observations which
affect CDS performance were recorded from the power supply meters during
a run.  These included:

     t  Electrode arc rate
     t  Nominal electrode voltages
     •  Nominal electrode and leakage currents
     •  Voltage and current of the. electrode power supply
The quantities associated with these observations varied during a course
of a run.  Ranges of values when appropriate are included with the tabulated
data and are helpful in interpreting the test results.
The CDS performance testing was divided into four series.  These included:
     t  Screening Tests
     •  Variable Parameter Tests
     •  Development
     t  Long Duration Tests

The screening tests were used to establish the parerneter levels for the
parameter tost matrix.  The test matrix runs constituted the najor portion
of the testing sequence and the results were used to establish the CDS
operating performance.  The development tests were performed by the TRW
division (DAD) supplying the CDS and the results are presented in this
report.  These tests were performed to obtain CDS performance data under
specific process conditions and to determine means of improving the device's
performance.  The DAD tests were performed during and after the test matrix
sequence.  A long duration test of 500 hours was scheduled to start after
completion of the test matrix runs.  This test was included in the program
to identify CDS maintenance schedule requirements and potential failure
modes.
                                    -23-

-------
The approved test matrix was drafted during the installation phase of the
program.  The matrix was based on a fractional factorial, two-level design
for five independent variables.
Screening Tests
A screening test was first performed to identify several of the
parameters and their levels.  The following conclusions were drawn from
the screening test (see Screening Test Summary, Appendix A, Table 1-A).
         t  Positive polarity resulted in unstable, high voltage
            operation and was eliminated as a test variable.
         •  Insufficient residence time for precooler droplet
            evaporation resulted in carry-over into the fan.
         •  Considerable reentrainment of the collector plate
            auxiliary wash system water spray resulted in atypi-
            cally high particle outlet loads.
         •  Extremely variable gas inlet conditions were observed
            and measured which sometimes exceeded the original
            design criteria (240 mg/Nnr*) on a time averaged basis
            and were significantly higher during part of a test
            period.
         t  Electrode operating voltages (31 kv) were significantly
            lower than the tests with ambient air (36-38 kv).
Based upon the above a fractional factorial design blocked with respect
to with and without wall wash was adopted.   All tests were at negative
power supply polarity with manual voltage control.   A nomenclature was
derived during the screening tests to describe the "effluent from the
coke oven battery stack.  The main coke oven stack description which evolved
was:
                      Normal
                      Atypical
                         Black
                         Gray
                         White
The condition could be identified also by observing the scrubber arc rate
and electrode voltages.   During normal  operation,  the stack effluent was

                                 •  -24-

-------
clear.   The CDS operated at its  pre-set  voltage with a low, less than 100
arcs/min arc rate.   During an  abnormal condition, the stack effluent
appeared black, various shades of  gray and white and high electrode arc
rates were measured, up to 1000  arcs/min.  For periods of up to 10 minutes
during black stack  conditions, the spray pattern on the first stage and
intermittently on the second stage would collapse.  The scrubber operated
normally during gray stack conditions with arc rates up to 350 arcs/min.
During white stack  operation,  the  scrubber arc rate was very sporadic,
less than 250 to in excess of  500  arcs/min. Derating  the gas velocity
20% substantially  improved  the electrical operation stability.

MATRIX TESTS
The matrix tests are designed  by a letter and number series of the form
TM-XXX-XXN where the X's are digits.  The last two digits are used to
identify the test relative to  the  run number.  The first three digits
correspond to the sequence in  which the  test was run during the matrix
test series.
The operating parameter levels established during the screening tests are
shown in Table 5.  Levels for  each parameter were the same within both
the wall wash and no wall wash blocks except voltage.
High level voltages were selected  to conform to conditions of about
equivalent spark rate, low level 4-5 kv  lower.

A fixed damper opening was used  to establish the volumetric flow rate for
the test runs.  The nominal  flow rate at different degrees of damper
opening was calibrated with pi tot  traverses at the sampling ports.  The
actual  gas flow rate changed during and  between test runs because of
variations in the main stack draft which coincided with the coke oven
operating cycle. The damper positions determined during the screening
tests were 5th notch for the low level and the 6th for high level. After
start of the matrix tests, abnormal conditions were encountered that could
not be handled by the scrubber at  6th notch operation.   This level was
subsequently changed to Notch  4.
                                   -25-

-------
           Table 5.   Test Matrix Parameters
Electrode
 Voltage A
Gas Flow Rate B
                   Wall  Wash
                           AQ = 31 kV
                           A] = 35 kV
                                        B  =
                                B,  =
    No Wall  Wash
       33 kV
       38 kV

4th notch (Damper Setting)3
5th notch
Electrode Water
 Flow Rate C
Pre-Cool ing Water
 Flow Rateb D
                                        CQ = 12 gpm
                                        C, = 16 gpm

                                        D  =3.75 gpm for B
                                         o                 o
                                           = 4.50 gpm for B,
                                        D.| =-5.00 gpm for B
                                           = 6.00 gpm for B,

         a.  The damper setting referenced to clamping notches  on handle
             track.  First notch is closed position.
         b.  The actual pre-cooling water absorbed by the gas  is  the dif-
             ference between the set flow rate, D, and the quantity  measured
             in the overflow.
The first matrix test series runs included the block  with wall  wash  and
are numbered 9 through 16 in the test matrix.   The results of these  tests
are summarized in Appendix A, Table 2-A.   The scrubber stack appearance
can be characterized as containing a large quantity of entrained  water
droplets.  The presence of these water droplets in the effluent and  the
location of the stack relative to the structures  complicated Ringlemenn
determinations.  Average, high inlet load collecting  efficiencies were
about 87%.   A screen mesh demister was installed  in the stack to  remove
the entrained droplets, but plugged rapidly because of intermittent
equipment operation which resulted  in screen  dryout.
At the conclusion of the first series of matrix tests,  a  development test
series was  performed to investigate intermittent  wall  wash  operation and
automatic voltage control.  Intermittent wall  wash with adequate  plete
                          -26-

-------
deposit control would allow higher voltage operation and    eliminate
the entrained droplets  in the stack gas stream.  It was determined during
these tests that the scrubber could operate up to eight hours without wall
wash and when used, required approximately 5 minutes of operation to clean
the system.  The wall wash cycle could coincide with periods of clean
main stack without jeopardizing performance.  There were no entrained
droplets in the exhaust stack when operated without wash.

Therefore, the second test block of the matrix, test numbers 1 through 8
were performed without wall wash.  The voltage levels used for the tests
were increased from those of the first block.  The results of these tests
are summarized in Appendix A, Table 3-A.  with the exception of test
number TM-016-03N, the particle removal efficiencies of this series exceeds
the first.  High inlet  load average efficiencies were about 94%.  The
cause of the one low run may be attributable to pick up of rust particles
in the sampling probe. This was the last test of the matrix run and at
this point in time rust flakes from the ducting were noted in the scrubber
stack.  The outlet probe water wash contained 108 mg of material at com-
pletion of the run and accounted for 61.7% of all material collected. The
average weight of material collected in the water wash of the previous
runs was 12.2 mg and accounted for 26% of the material collected.  If
this run were corrected, based on the sample distribution from previous
tests, the particle removal efficiency would be approximately 80%.

The main reason for the improved collection efficiency of the last test
series over the first is attributed to the elimination of dirty, entrained
water droplets. A more detailed analysis of the tests results and conclu-
sions is presented in Section VI.

A measure of the S02 and $03 removal was made during test number TM-015-05N.
These measurements  were made in accordance  with  EPA Method 8.   The first
impingers of the inlet and outlet sampling  trains  contained isopropyl
alcohol  to remove S03 from the sample  gas  and the  second  and third impingers
of each  train contained hydrogen  peroxide  for S02  removal.   The inlet S02
concentration was determined to be 253.6 ppm and the outlet,  160.7 ppm.  These
values  correspond to a removal  efficiency of 36.6%.   The  inlet  and outlet
                                   -27-

-------
S03 concentrations were 3.00 ppm and 3.45 ppm respectively.   The higher outlet
concentration probably resulted from conversion of SO^ to SO- by ozone gen-
erated in the droplet formation corona.
DEVELOPMENT TESTS
These tests, performed by DAD, were designed with specific objectives  of
determining means to improve the CDS performance, of further characterizing
the process stream and of complementing the contractual  test program.  The
sampling procedure and equipment used during these tests were the same as
those used in the program tests with the exception that the sampling times
at each point were shorter and in some cases only one traverse was used per
run.  The abbreviated test times were necessary so that a complete test
would coincide with a normal or atypical stack emission period.

The  tests  were  conducted  during  a period between  the  two blocks  of  the
matrix  tests  and after  the matrix tests were completed.  A  summary  of these
tests may  be  found in Appendix  A, Table  3-A. Included with  these  tests are
the  Ringlemann  range estimates of both the  CDS stack  and the main stack.
The  opacity values are  more  significant for these tests because  of  their
shorter  duration and of their specific intent of  being  run  during a  limited
range of main stack conditions.

The  tests  were  run in automatic  voltage mode. During  several of  the  tests,
only two stages  were operated.   The object  of the two stage operation
tests was  to  determine  the actual influence of staging on scrubber  perform-
ance.

The  main stack  and CDS  inlet particle concentration were measured simul-
taneously  during atypical and normal coke plant operations.  The main stack
had  a particulate material loading of 764 mg/m3 (0,334 gr/scf) while the
inlet to the CDS was 812 mg/m3 (0,355 gr/scf) at high emission load,  A
second run made  during  nominal operation indicated a  main stack  loading
of 68.0  mg/m3 (0.0297 gr/scf) and a scrubber inlet loading  of 97.7 mg/m3
(0.0427  gr/scf).  As only six sampling points along one traverse axis were
used in  the 16 foot diameter main stack, the sample may not be representative.
There is a possibility  that particulate material is stratified along one
side of  stack.   If any  stratification in the main stream occurred, then
                                   -28-

-------
the region from which the CDS flow was withdrawn could have a higher
particle concentration.

An access port was cut into the outlet transition section of the CDS
after completion of the matrix tests to allow a gas stream velocity
profile measurement across the CDS outlet.  The results of these tests
are shown in Figures 10 and 11.  These velocity profiles have considerably
more variation than those taken during the system checkout period and
shown in Figure 7 and become progressively worse as the average velocity
is decreased.  At original design flow rates through the scrubber, the
pressure drop of the baffles (25% open) appears to be sufficient.  Post
test calculations indicate that a 12% open area was required at the actual
test flow rates.
LONG DURATION TESTS
The long duration test was started at the conclusion of the matrix tests.
It was performed concurrently with additional development tests.  The CDS
was then operated for an additional eight hour day. Operating parameters
of the CDS were varied during the sampling periods.  However, during the
extended daily operating period the parameters were adjusted to correspond
to run number 2 of the matrix.

The sixteen hour day scrubber operation was terminated during the second
week because of operation problems that had developed which were attributed
to the excessive collector plate corrosion.  The arc rate was increasing with
time and it was more difficult to maintain a pre-set operating voltage
because roughened surfaces were initiating arcs.  The total operating time
during the long duration test was 100 hours.  An additional 260 hours of
operation were accumulated during the previous testing giving a total of
360 hours.
                                  -29-

-------
    MODULE NO. 3
1.40
1.20
1.00
0.80
0.60
0.40
0.20
   0
S^
                                           AVERAGE VELOCITY 0.88 M/S
                                                               + 59%
                                           MAXIMUM DEVIATION  -
    MODULE NO. ^7
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0

O— >








v
H,
^v
•x.
>>-








O (





S


^^




iX>^_
^x^>
^








^^
'^h^
"^



0.5 1.0 1.5 2,0 2.5 3.
                  DISTANCE FROM GAS FLOW INLET END (M)
    Figure 10.  Gas Flow Velocity Distribution After Completion
                of Test Series
                                -30-

-------
I
CO
o*

oe
O

uj
£

u
O
UJ

ISI
                       60

                       50

                       40


                       30

                       25

                       20


                       15
                 FINAL MODIFICATION   _.-•
                          8.0
  I
INITIAL DESIGN POINT
  ®
  I       AVERAGE GAS VELOCITY, fpi
 5.0
                      4.0
3.0
2.5
2.0
                        0.08  .10  .12  .14  .16   .18  .20   .22  .24  .26   .28  .30   .32  .34  .36

                                          SPECIFIC COLLECTING AREA, fiP/cfm
                           iFigurell.   Gas  Maldistribution Sensitivity to
                                        Specific Collecting Area-RMS Velocity
                                        Profile  Vs.  Average Velocity

-------
                           VI.   DISCUSSION

The test data developed from the statistically designed  experiment
previously discussed was used to estimate equipment performance over the
range of the test parameters.  However, efficiency changes due to process
variability dominated over the efficiency changes affected by parameter
level settings, thus, substantially weakening the statistical power of the
experiment.  Therefore, resort was made to data pooling,  editing and
regression analysis to develop the sensitivities of efficiency to the
parameter levels.
Methods of Analysis
Three data sets were pooled, the APT and TRW original  matrix test data, and
the development test data.   Preliminary data plotting  indicated that the
efficiency data with and without continuous wall wash  were different,
therefore, the data were initially classified into these  two populations.
The APT particle size distribution data was analyzed and  it was determined
that the efficiency variability was related to changes in particle aero-
dynamic mean size.   Low inlet loads corresponded to submicron high resis-
tivity hydrocarbon particulate.  Observations of the stack plume color
under these conditions substantiated these condlusions.   High dust loads
corresponded to micron sized carbon particulate.
Therefore, a regression analysis was first performed of efficiency vs.
inlet dust load for the two segregated populations.  This regression
analysis was then used to identify statistical outliers and test anomalies.
Judicious editing of this data finally evolved an analysis wherein the
dominant parameter variables were identified and quantified.
Process Characterization
     The APT emission and size distribution data was first investigated
to determine if the process conditions could explain the  variability in
collecting efficiency.  Figure 12 is a plot of mean aerodynamic diameter
vs. inlet particle load.  It can be seen that there is a  correlation,
lower inlet loads producing the smallest mean particle sizes.   The
mean particle sizes vary from about .4 microns to about 1.5 microns over
the dust load range.  Somewhere in the range of 200-250 mg/Nm3 the aero-

                                  -32-

-------
             SL
TJ
 K
oe.
LLI
LLJ


Q


1

y
S


1
O
               2.0
               1.5
               1.0
               0.9
               0.8
               0.7
               0.6

               0.5
               0.4
                 100    150  200     300

                       INLET PARTICULATE LOAD, MG/dNm
                                 500   700   1000
                                          3
Figure  12.   Coke Oven Flue  Gas Process Variability, the Relationship
             of Inlet Dust Load to Particle  Size
                                 -33-

-------
dynamic mean diameters become more constant.   Below this range mean size
fluctuates widely.  Examination of the particle size distributions plotted
on log-probability paper, Figures 13-15, suggest that below about 220
mg/Nm3 the inlet particle size distribution is highly bimodal.  At the
higher inlet loadings and for most all outlets distributions, Figures
16 and 17, the distribution approachs a log normal distribution.  The bimodal
distribution is composed of a hydrocarbon aerosol having a mean size of
perhaps 1 or 1 1/2 microns.  This suggests a statistical approach to the
data wherein the efficiency is first correlated to the inlet dust loads.

Results of Analysis
Regression analyses were first performed on the two data populations
classified as with continuous wall wash and without. All of the test
points from all three data sets were analyzed within each population.
High correlation coefficients were obtained.   The results of the regression
analysis are shown on Figures 18 and 19.  The data analyzed with the
continuous wall wash was shown statistically to be significantly lower
than the data without wall wash at  a  high confidence level.  In both
sets of data the efficiencies tend to assymtote and become nearly constant
                                             3
at inlet dust loadings of about 220-260 mg/Nm.   This corresponds to the
inlet load at which mean aerodynamic diameter becomes constant.   It would
appear that the most statistical  power would lie in the region of the
                          q
curves exceeding 220 mg/Nm .

Efficiencies were about 4 1/2% higher for the population of data wherein
the wall  wash was not operated.   As was previously discussed this was
attributed to the elimination of the wash water spray droplet reentrain-
ment problem which gave an atypically high outlet dust loading.   Also
the standard error of the residuals was substantial lower for the case of
no wall wash, again reflecting the high variability of the outlet emission
data when the wall wash was operated.   The three data sets scattered
randomly about the regression line, indicating they were all part of the
same population.

In order to determine the parameter effects an analysis of the residual
deviations was made (i.e., deviation about the efficiency-inlet load

                                   -34-'

-------
     4.0
 E
 a.
  8.
ffi
O
y

5
Q

O
     2.0
      1.0
     0.5
     0.2
             SYM. POSITION



                O INLET

                A OUTLET
LOADING

 mg/dNm3

  182

   22
                         I    I    I    I   I    I
                  10    20   30  40  50  60  70  80


                          MASS PERCENT  UNDERSIZE
                         90   95
98
Figure  13.   Inlet and. Outlet Size Distributions at Low Inlet  Load
                               -35-

-------
 E
 a.
  o
  Q.
 ULJ

 fc



 <

 Q

 y
Q

O
      6.0


      5.0
      2.0
1.0
     0.5
     0.2
            SYM. POSITION




               O INLET


               A OUTLET
                      LOADING
                             o
                       mg/dNm


                        201


                        23
                          I
                        J	I
I	I
I
                  1 10    20   30  40  50 60  70  -80


                           MASS  PERCENT UNDERSIZE
                                                 90    95
                         98
Figure  14.   Inlet and  Outlet Size Distributions  at Low Inlet  Load
                              -36-

-------
    '6.0
    5.0
    2.0
  a
  a.
     1.0
o
O
0.5
    0.2
         SYM. POSITION

           O INLET
           A OUTLET
LOADING
       3
mg/dNm
  217
  255
                  10     20  30  40 50  60 70   80
                         MASS PERCENT UNDERSIZE
                                                 90    95
                                   98
 Figure 15.   Inlet and Outlet Size Distributions  at Low Inlet Load
                              -37-

-------
 <
  O
  a.
 <

 Z
 >-
 Q

 u



 <

 Z


 O

 O
 0£
 LU
6.0



5.0



4.0





3.0







2.0




1.5







1.0



0.8




0.6


0.5



0.4




0.3






0.2
             SYM POSITION   LOADING
                                   O

                            mg/dNm
                O IN LET

                AOUTLET
                              I    I    I   I    I
                   10    20   30  40  50  60  70   80



                          MASS PERCENT UNDERSIZE
                                                 90
95
98
Figure  16.   Inlet and  Outlet Size  Distributions  at High Inlet Load




                               -38-

-------
  s.
 85
 te

 1
 a
 y

 <

 Q
 §
      6.0

      5.0
      2.0
1.0
0.5
      0.2
SYM POSITION

O INLET
A OUTLET
LOAD
mg/dNm3
508
23.1
         I	I
                          I    I    I   I    I   I     I
                   10    20  30  40 50  60  70   90

                          MASS PERCENT UNDERSIZE
I	I
                                                80   95
          98
Figure 17.   Inlet and Outlet  Size Distributions  at High Inlet Load


                               -39-

-------
  65.
165.
265.
365.
465.
565.
665.  •
765.
865
    PflRT IN
Figure 18. Efficiency Correlation, Total Unedited Data Set With Continuous Wall Wash

-------
65.
165.
265
365.
465.
565.
665.
   PflRT IN
   Figure 19.   Efficiency Correlation, Total Unedited Data Set Without Wall Wash

-------
regression line).  The parameters considered were inlet gas temperature,
gas flow rate and voltage and interactions thereof.  Analyses were per-
formed after eliminating obvious statistical outliers and reviewing the
data for test anomalies, Table 6.  Generally, the outliers could be
explained in terms of test anomalies particularly with respect to unstable
electrode voltage operation or measurement procedure.

For the case of the continuous wall wash the only statistically significant
parameter identified was the inlet temperature.  Increasing temperature
slightly decreases efficiency.  The data with the continuous wall wash
system is atypical of the real potential of the CDS and is a situation that
can be corrected by an improved design.  Therefore, the remaining analyses
were directed at analyzing the data without wall wash in an attempt to
develop the efficiency sensitivities to the parameter levels.

A detailed summary of the regression analyses may be found in Appendix C.
The first analysis of the residual deviations for the data population without
wall wash and incorporating edited data indicated an interaction between
gas flow rate and inlet temperature.  However, this data incorporated the
low inlet dust loading tests with the highest degree of variability.   The
more steady process conditions on the assymtotic portion of Figure 18
suggests that more statistical power may be developed with this part of
the data.  Therefore, efficiency data greater than 90% or inlet loads
                      •j
greater than 227 mg/Nm  were incorporated into a final data pool.  A
regression analysis was performed with first and second stage electrode
voltages, inlet gas temperature and actual  gas flow rate.   The statistical
analysis showed that gas flow rate was a highly significant variable having
                                               -4   •?
a negative coefficient with a value of 2.9 x 10  %/nrVm, Figure 20.   High
gas velocities result in lower efficiencies.   Over a 50% decrease in gas
flow rate the efficiency only increased about 2%.  This sensitivity is
significantly less than that experienced in other pilot programs where
the regression coefficients have been a factor of 2 higher.   The relatively
low sensitivity is attributed to the fact that the gas distribution grew
progressively worse as gas flow rate was reduced.  Post test velocity pro-
files of three of 19 CDS modules were measured over a range of gas flow rates.
A typical distribution is shown in Figure 10 and the rms error over the test
velocity range shown in Figure 11.   At a velocity of about 3 fps it is
                                  -42-

-------
                                TABL2 6

                        SUSPECT KAISER CDS DATA
Data
No.
8
13
10
4
S-2
S-3
3N
12N
006
Point
Source
APT
APT
APT
APT
' Screening
Test
Screening
Test
EPA
EPA
DAD
COMMENTS
Statistical
Outlier
Q
p
Outlier
dP
Outlier
vs load
Outlier
vs load
Outlier
vs load
Outlier
vs load
Outlier
vs load
Outlier
vs load
Outlier
vs load
on load vs
on load vs
on efficiency
on efficiency
on efficiency
on efficiency
1
on efficiency
on efficiency
on efficiency
Equipment Condition
Voltage Wallwash
Low voltage
operation ( ~30KV)
Low voltage
operation ( ~30KV)
.Upset 1st stage
voltage
Upset 1st stage
voltage
Positive polarity
Arc control on
1st stage only
Satisfactory
Arc control on
1st and 2nd
stages only
1st stage voltage
collapse
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
No
 Outlet sampling error.

                         Measurement Anomally    Position

    1       APT     5 & 6th Stages-50% Plugged    Inlet

4,10,18     APT     Not 100% Simultaneous With    Outlet,
                    Inlet,  Process Change         Inlet

    9       APT     6 & 7th Stages Wet            Outlet

   15       APT     3rd & 5th Stages, 20-30%      Inlet
                    Plugged
Efficiency
 Effect

Fractional

Fractional and
Overall

Fractional

Fractional
                                    -43-

-------
V)
c
o
£
  -2




  -3




  -4
   10
           12
14
16
18
        20
                         Measured Flow Rate sm3/m(xl03)
22
24
                              Figure 20

  Effect  of Gas Flow Rate  on Overall Collecting  Efficiency

-------
 substantially out of specification.   The  rms  velocity  error  increased
 from 15% to 50%.    The non-linear influence on  velocity would  therefore
 result in efficiency losses  at low gas  velocities  (2.5 fps)  estimated  at
 between 2 1/2 and 3%.  The increasingly  poor gas distribution is  attri-
 buted to the loss of pressure drop at the baffle system.

 Several regression analyses  were attempted to bring  out the  effects  of
 electrode voltage.   The regression of the residual deviations  was  first
 attempted using the second stage voltage  which  is  somewhat typical of
 the average voltage operation across  all  stages.   As no correlation  could
 be developed an additional analysis was conducted  by introducing both  the
 first and second stage voltages.   It  was  felt that this would  more typically
 reflect the first stage space charge  effect on  cleaning efficiency.  Again,
 no influence was uncovered,  Figure21.  As can  be  seen the data  fairly well
 scatters on both sides of the zero deviation  line  over a  range of  33-41
 kilovolts.   When the voltage is  pushed  much beyond 40  kv  greater statistical
 variability occurs which may be  indicative of the  high spark rate  conditions.
 The insensitivity of voltage is  surprising and  it  is attributed  to the low
 gas velocity operation.   It  is hypothesized that the high space  charge due
 to conducting carbonblack and water droplets  and their correspondingly low
 mobility is affecting  the ability of  the  high voltage  to  establish a suf-
 ficiently higher voltage gradient outside of  the corona field.   This effect
 would not be anticipated at  higher gas  velocities  where the  space  charge
 is more evenly distributed through the  equipment volume.

 A summary of the regression  analyses  discussed  above may  be  found  in Table 7.
 Further analyses did not result  in any  additional  correlations.

 Based upon  the above the particle size  distribution  data  obtained  in the  APT
 test was analyzed to develop fractional efficiencies as a function of  the
 aerodynamic diameter.   Only  those tests with  no test or statistical  anomalies
 and without wall  wash  were analyzed.  The APT data were segregated into
 both high and low inlet loading  conditions as previously  discussed.  The
 efficiencies at low inlet conditions  (typically 180  mg/Nm^ to  220),  Figure
22, are very similar for all  three tests,  Nos. 14,  16 and  19.   Efficiency
 improves with the smaller aerodynamic diameters  and  reaches  about 90%  at
 0.4 microns.   This  is  atypical of other types of electrostatic and wet

                                   -45-

-------
in
c
o
  0




  -1



  -2




  -3




  -4
           33
34       3T       36       37

     Electrode Voltage (Stage 2) KV
38
                                                              39
                                                                      40
                          41
                                       Figure  21


          Insensitivity of  Efficiency to Second Stage Electrode Voltage

-------
INDEPENDENT VARIABLES
Inlet load,
Residual Errors plus
Test Parameters, T.°C
Inlet load,
Residual Errors plus
Test Parameters, Q, —
                                       TABLE 7

                            SUMMARY OF REGRESSION ANALYSES
  DATA SET
Pooled Data
                                        WALL  WASH
                                           Yes
Edited Data
 STD. ERROR
OF RESIDUALS
     8.9
Yes
No
No
6.4
3.6
1.57
   SIGNIFICANT
  PARAMETER AND
    REG. COEFF.
89.1 (1-e "


40.6 - .318T

93.5 (l-e-014Ci)

5.6-2.9 x 10~4Q

-------
scrubbing equipment, but has been experienced in other CDS pilot programs.
It should be noted that the particulate under these conditions has a bi-
modal distribution of high resistivity hydrocarbon and carbon black.  Some
hydrocarbon might have condensed on the carbon black increasing its surface
electrical resistivity.  Therefore, lower efficiency might result as com-
pared with the higher inlet loadings  (420-510 mg), Figure 22, at the same
cut diameter. It is interesting to note that the low inlet load fractional
efficiencies were almost independent  of both electrode voltage and gas flow
rate.  Values of these parameters varied between 33 and 38 kilovolts and 12
and 15 x 103 m3/hr.

The fractional efficiency data for the higher inlet loadings, Figure 22 ,
are more typical of the results normally expected from precipitators and
high energy wet scrubbers.   Efficiencies in the 1 to 2 micron range approach
99%.   For this size particulate removal mechanisms consist of both high
energy impaction of charged water droplets and electrostatic charging and
precipitation.  Below 1 micron the collection mechanism is probably by
electrostatic precipitation either through charge transfer upon inter-
action of a charged water droplets or through direct corona charging of
the plasma.  The fractional efficiencies at high inlet load and submicron
particle sizes approach those of the  low inlet load case.   This is again
attributed to the fact that most of the particulate in this range is
hydrocarbon having the same electrical and physical properties as that
of the lower inlet load case.

Some additional  equipment optimization probably could have been affected
by running unbalanced electrode waterflow rates between the stages.  This
would result in optimally sizing the water droplet for high energy impaction
and removal of the carbon black in the first stage and the electrostatic
precipitation of the hydrocarbon in the second stage.

Limited testing with a two-stage unit, (i.e., the third stage removed from
the process) showed that about a 2 or 3% loss in efficiency results.
Conversely had a fourth stage been added it is believed that approximately
a 1.5 to 2.0% efficiency improvement could have been affected.   The fourth
stage would also protect against infrequent high dust loads which would
normally cause a collapse of the first stage. In this event even with a

                                  -48-

-------
LU
a,
 *•.


O


LU

(J

LL.
u_
LU
   99
98
95
   90
8  80
u

g:  70
   50
                                                           INLET LOAD,   SYM
                                                           mg/dM Nm3   RUN
                                                                         NO.
                                                     X
                                                       o
                                                    X  419-508
                                                            182 - 217
017

015
                                                                      O16

                                                                      D19

                                                                      A 14
                  0.3   0.4     0.6       1.0     1.5   2.0

                          AERODYNAMIC DIAMETER, MICRONS
   Figure 22.  Fractional  Efficiency  Data for the Higher Inlet Loadings
                                     -49-

-------
 partial  loss  of  the  first  stage.the  equipment would  basically perform as
 an  optimized  three stage unit.

 The above  performance was  achieved in  a  unit  consuming  very  little  power
 and at a low  liquid  to  gas  ratio. Under  typical  gas  flow rate conditions  of
 about 15,000  Nm3/n (13,000  acfm), the  total power  supply consumption  was
 of  the order  0.7-1 watt/hr/Nm3  (.85-1.2  watts/acfm)  at  a nominal  35 kv
 electrode  voltage.   Pressure  drops between the CDS inlet and the  stack
 discharge  were extremely low, about  1.3  cm/we.   System  pressure drops,
 coke oven  breeching  to  CDS  stack, were of the order  of  10 cm/we.  This  is
 probably an atypically  high pressure drop in  that  there  were numerous
 expansions and elbows of ductwork which  would increase  the pressure drop.
 The liquid to gas ratio at  the  nominal conditions  were  of the  order of
         <5
 0.18 £/Nrrr (0.9  gal/1000 acf) gas treated.  Energy consumption is typical
 of  energy  requirements  for wet  electrostatic  precipitators.  Liquid  to gas
 ratios are substantially lower  than high energy  Venturis  and  most other
 types of high performance wet systems.

 Limited CDS stack opacity observations were made to correlate time  ave-
 raged outlet loads capable of meeting a  20% opacity criteria, Table 8.
 The  stack opacities varied between 5 and 25%.   The 15% opacity was  not
 exceeded during Test No. 002 for an average discharge load of 33.4  mg/
    o                                      3
 dNm  (.0146 gr/dscf).  Values of 23 mg/dNm  consistently  resulted in a
 stack opacity of 0-5, Tests Nos. 003, 010, Oil, 012 and 015.   During Test
 No.   .006, performed during a major upset, the 20%  opacity was exceeded
 during two 5-minute periods.  The actual discharge load during this
                                                                        3
 period most certainly exceeded  the measured average value of  37.8 mg/dNm
 as  the first stage electrical voltage collapsed.    It would appear that a
                                            o
design point  emission discharge  of 23 mg/dNrrr  would satisfy the requirement.

 Using 95%  as the maximum efficiency for  high  voltage on  three  stages,
 the  maximum permissable inlet load corresponds to  700 mg/dNm3  (.3 gr/scf).
 For inlet  loads  less than this  progressively  larger margins  of safety
 result.  It should be noted that most  data from  coking emissions  is in the
 range of 230 to  500 mg/dNm3.  The Kaiser Coking  Furnace  conditions  and
 production rates are such that  abnormally high smoking conditions are
 currently  being  experienced.  It is felt that this results in  a sufficiently
 conservatively CDS design and operating point.
                                  -50-

-------
Test
No.
002
003
010
Oil
012
006
015
Outlet
mg/N
33.4
16.9
26. 8
23. 8
23. 8
37.8
22.9
                                TABLE 8

                  KAISER OBSERVED CDS STACK OPACITIES
        Outlet ,  Dust Load   Observed Opacities, %         Average
        	 / TWT	*•'  IV •*• I ft f> f         T*	 O _	.	
3


   .0146

   .008

   .0117

   .0104

   .0104

   .0165

   . 0100
                              15,10,10,15,15,10,10
                                  5,  5,  10,15,©,©,(2<

                              (5,  5,  5,  5,  5,©  52)
12

 5

 5

 5

 5

ll»

 7
 '  1st stage voltage severely deteriorated
2,
   Every other observation recorded
3' Circled nos.  Stack A B0% opacity
                                 -51-

-------
                                REFERENCES
1.   "Charged Droplet Scrubber for Fine Particle Control  - Laboratory Study",
    Report prepared for the Environmental  Protection Agency under Contract
    Number 68-02-1345 by TRW Systems Group,  February 13,  1976.

2.   Federal Register, Vol 36, No. 247, pp 24887-24890, December 23, 1971.

3.   Private Communication from Di-Gerald Shaughnessy of the University of
    Dayton, dated January 31, 1975.

4.   Charged Droplet Scrubber Pilot Demonstration for Fine Particle Control,
    Test Matrix; Contract Number 68-02-1345, prepared for the Environmental
    Protection Agency, July 28, 1975.
                                  -52-

-------
          LIST OF TABLES AND FIGURES - APPENDICES A  - B  -  C
Table No.                                                         Page No.
A-l
A- 2
A- 3
A- 4
A- 5
B-l
B-2
B-3
B-4
B-5
B-6
B-7
B-8
B-9
B-10
B-ll
B-12
B-13
C-l
C-2
Figure No.
C-l
C-2
Screening Test Summary 	 .
Block One Matrix Test Summary - With Wall Wash ....
Development Test Summary 	
Block Two Matrix Test Summary - No Wall Wash 	
Inlet and Outlet Size Distribution Data Summary . . .
Inlet and Outlet Sample Particle Data for Run #1 . . .
Inlet and Outlet Sample Particle Data for Run #2 ...
Inlet and Outlet Sample Particle Data for Run #6 ...
Inlet and Outlet Sample Particle Data for Run #8 ...
Inlet and Outlet Sample Particle Data for Run #9 .
Inlet and Outlet Sample Particle Data for Run #10 . .
Inlet and Outlet Sample Particle Data for Run #13 . .
Inlet and Outlet Sample Particle Data for Run #14 . .
Inlet and Outlet Sample Particle Data for Run #15 . .
Inlet and Outlet Sample Particle Data for Run #16 . .
Inlet and Outlet Sample Particle Data for Run #17 . .
Inlet and Outlet Sample Particle Data for Run #18 . .
Inlet and Outlet Sample Particle Data for Run #19 . .
Analysis of Data - With Wall Wash 	
Analysis of Data - Without Wall Wash 	

Particle Loading Effect on Removal Efficiency With
Wall Wash.
Particle Loading Effect on Removal Efficiency Without
Wall Wash.
A-l
A- 2
A- 3
A- 4
A- 5
B-l
B-l
B-2
B-2
B-3
B-3
B-4
B-4
B-5
B-5
B-6
B-6
B-7
C-l
C-2

C-4
C-5

-------
        APPENDIX A





SUMMARY OF PERFORMANCE DATA

-------
                                                                                                                                                                                                                                                        Table A-l.  Screening Test Summary
TEST NO.
AND DATE
1
9/23/75
2
9/24/75
3
10/7/75
4
10/8/75
5
10/9/75
6
10/10/75
VOLTAGE a(kV)
STAGE TR
1 2 3 SET
30 34 32 73
29 32.5 32.5 65
32 32 30 65.5
29 31 30 62
31 31.5 31.5 7T
29.5 30 29.5 64.5
WATER FLOW RATE (1/min)
WALL PRE-
ELECTRODE WASH COOLING
62 106 22.7
59 114 • 22.7
61 106 170
5.19e
61 106 17. Oe
6.706
45 106 17.0
6.516
61 106 17.0
6.32e
GAS FLOW
RATE (sm3/hr)
INLET OUTLET
20 835 13 528
18 750 12 537
18 443 13 344
19 770. 17 417
17 930 14 554
21 470 18 130
GAS TEMPERATURE (°C)
STACKb INLETC OUTLET
199 107 52
205 120 46
191 138 52
201 162 53
187 128 54
191 156 58
PERCENT
MOISTURE
INLET OUTLET
8.7 14.4
6.6 73.2
4.55 13.0
10.4 15.2
8.27 13.2
8.97 14.8
PARTICLE LOADING f
(mg/sm3)
INLET OUTLET
211.9 42.79
73.2 37.41
132.0 62.70
343.9 88.10
215.8 55.84
151.3 59.50
CLEANING
EFFICIENCY
PERCENT
79.8
48.9
52.5
74.4
74.1
60.7
DAMPER
SETTING
(NOTCH)
5
5
5
6
5
6
OUTPUT
CURRENT (mA)
TOTAL LEAKAGE
195 45.5
195 43
170 25
175 25.2
170 23
160 22
POWER
INPUT (AC)
VOLTS AMPS

. 330 36
368 38
388 38

372 37.5
ARC RATE
(ARCS/MIN.)d
NOMINAL MAXIMUM
50 150
125 300
70 >500
150 >500
50 150
150 230
COMMENTS
Negative polarity. One upset during
test. Main stack light gray, except
black during major load.
Positive polarity. Main stack medium
gray to clear. Sustained arcing during
test. Power supply off periodically
to quench arcs.
Negative polarity. Main stack clear to
medium gray. One major load. Arc
quenching capacitor on first stage only.
Negative polarity. Main stack light gray
to black. Power supply off periodically
to quench sustained arcs. Arc quenching
capacitor in first stage only.
Negative polarity. Main stack light gray.
Major load started toward end of run. Arc
quenching capacitor in first stage only.
Negative polarity. Main stack light gray.
Arc quenching capacitor in first stage
only. Several collecting troughs were
overflowing.
a.   Nominal  set  voltage.   Mean  voltage  during periods of high arc rate was considerably lower.
b.   Stack  gas  temperature  measured  at same position as inlet particle sampling port.
c.   Inlet  gas  temperature  to  CDS turning section.  Measured after pre-cooling.
d.  Maximum arc rate readable - 500 arcs/min.
e.   Net flow rate adsorbed by gas  stream.
f.   gr/dscf =  2288 mg/dNm3
                                                                                                                                                                                                                                                           A-l

-------
 PAGE NOT
AVAILABLE
DIGITALLY

-------
Table A-3.   Development Test3 Summary
TEST NO.
AND DATE
D-002
10/29/75
D-003
10/29/75
D-004
10/30/75
0-006
10/31/75
D-007
11/3/75
D-008
11/20/75
D-009
11/20/75
D-010
11/21/75
D-011
11/21/75
D-012
11/24/75
D-013
11/25/75
D-014
12/1/75
D-015
12/1/75
D-016
12/4/75
D-017b
12/5/75
D-018b
12/9/75
VOLTAGE (kV)
STAGE
1 2 3
36
36
38
36
36
35
39
40
35
38
40
40
40
40


32
37
36
35
41
40
40.5
40
40
42
40
41
41
40


37
36
39
39
40
39
39.5
39
0
0
40
38
38
40


OUTLET GAS
FLOW RATE
(sm3/hr)
11 693

8 160
10 079
9 689
11 433
9 219
10 245
10 058
10 639
11 200
14 338
10 804
8 629
66 110C
64 528C
GAS TEMPERATURE
INLET OUTLET
121
135

120
123
135
132
144
146
137
144
141
140
139
21 Od
226d
51
68
51
69
74
69
69
64
71
77
73
77
72
66
176e
202e
PERCENT
MOISTURE
OUTLET
12.2
6.8
12.3
11.6
12.9
12.0
22.2
12.0
12.8
11.7
11.8
12.8
11.2


ELECTRODE
CURRENT
(mA)
285
220
250
358
233
248
258
244
222
227
277
260
260
261


ARC RATE
LOW/HIGH
(ARC/MIN.)
250/575
100/375
50/100
400/500
150/270
>500
400/450
250/350
?500
>500
>500
?500
250/500
300/>500


SAMPLING
NO. OF NO. OF
TRAVERSES POINTS
1
2
1
2
2
2
2
2
2
2
2
1
1
2


16
32
16
32
32
32
16
16
16
16
16
8
16
32


SAMPLE
TIME
(min)
3
3
3
3
3
3
4
4
4
4
4
3
4
3


DAMPER
SETTING
(NOTCHES)
4
4
2
4
4
4
4
4
4
4
4
5
4
4


PARTICLE LOADING
(mg/sm )
INLET OUTLET
294.7
244.6
364.8
505.5
230.0
642.1
342.1
225.9
253.6
204.8
248.5
551.5
153.3
233.4
68. Of
764. 3f
33.41
16.93
16.93
37.76
12.81
31.58
20.37
26.77
23.80
23.80
37.76
54.23
22.88
21.51
97. 79
CLEANING
EFFICIENCY
(PERCENT)
88.7
92.9
95.4
92.5
94.5
95.0
94.0
88.2
90.6
88.4
84.8
90.2
85.1
90.8

812. 49
! STACK OPACITY
(PERCENT)
MAIN CDS
35-95
15


15


10
10
82


80



10-15
10-15





<5
<5
<5


15



COMMENTS
With wall wash.
Pi tot probe data in error.
Main stack medium gray
Main stack medium gray.
Main stack clear to light gray.
Main stack black to dark gray.
Main stack black.
Main stack white.
Two stage operation.
Two stage operation.
Main stack black to light gray.
Main stack black to dark gray.
Main stack light to dark gray.
Main stack black to light gray.
Nominal Stack Condition
Load
I
a. All DAD tests were run with an electrode flow rate of 53 i/min. and in automatic voltage control- mode.
b. Tests D-017 and D-018 were run to obtain a comparison between particle loading in main stack gas stream and stream to CDS.
c. Volumetric flow rate of main stack.
d. Main stack gas stream temperature.
e. Cut off stream to CDS temperature.
f- Particle loading in main stack gas stream.
          A-3

-------
 PAGE NOT
AVAILABLE
DIGITALLY

-------
                 Table A-5.  Inlet and Outlet Size Distribution Data Summary
                             Air Pollution Technology, Inc.
Run
No.
APT
1
4
6
8
9
10
13
14
15
16
17
18
19
10P(1)
13P(1)
TRW
001
003
004
005
005
006
008
009
010
010
on
on
012


Inlet
Load ,
mg/NmJ
232
163
188
247
132
166
953
201
419
182
508
168
217
244
1789
Mean
Size
dpa,MmA
1.50
0.89
0.96
0.59
0.41
1.60
1.02
1.00
1.35
0.67
1.55
1.55
1.00
0.78
1.92
Std.Dev.
ag
2.7
2.0
2.0
2.1
2.6
4.3
1.9
4.2
2.1
2.5
2.2
8.0
2.8
1.7
1.8
Outlet
Load
mg/Nm3
35.8
61.0
11.0
29.2
41.5
84.5
56.6
23.1
35.5
22.2
23.1
21.9
25.5
166
953
Mean
Size
dpa'MmA
0.57
0.69
0.96
0.41
0.61
0.95
1.13
1.00
1.40
0.83
1.06
1.25
0.98
2.10
1.02
Std.Dev.
"g
2.6
1.9
2.0
2.7
1.8
2.1
2.0
2.2
2.5
2.3
2.2
2.7
2.5
4.3
1.9
Efficiency
85.7
62.9
94.2
88.4
69.5
49.6
94.2
88.9
91.6
88.0
95.6
87.3
88.6


(1)     Runs 10P and 13P  were  samples  taken before the water quench sprays;
       therefore,  the  inlets  for  Run  10 and  13 are the outlets for Runs 10P and 13P
       respectively.
                                            A-5

-------
              APPENDIX B






SUMMARY OF FRACTIONAL EFFICIENCY DATA

-------
Table B-l.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #1
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter .
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
232
232
232
232
230
180
99.8
50.8
42.9
d
pc
(pmA)

27.3
12.0
5.7
2.3
1.3
0.74
0.45
	
0.134

OUTLET
"cum
(mg/DNm3)
35.8
35.8
35.0
34.7
34.0
33.2
27.0
20.2
15.9
V
(ymA)
--.-
23.8
10.4
4.0
2.0
1.2
0.64
0.37
	
0.585

Table B-2.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #4
IMPACTOR
STAGE
NUMBER

Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
163
159
159
159
158
157
' 120
65.8
" 28.9
d
PC
(ymA)
	
27.4
12.0
4.6
2.3
1.3
0.74
0.45
	

0.087

OUTLET
M
cum
, (mg/DNn3)
61.0
61.0
60.8
60.6
59.8
58.9
45.4
22.9
17.2
d
V
(umA)
	
. 23.6
10.3
4.0
2.0
1.2
0.64
0.37
	

0.475

                           B-l

-------
Table B-3.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #6
IMPACTOR
STAGE
NUMBER

Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)

188
187
186
186
186
177
132
67,0
26,4
d
pc
. (ymA)


25.6
11.2
5.3
2.2
1.3
0.69
0.42


0.099

OUTLET
M
cum
(mg/DNm3)

11.0
10.8
10.6
10.6
10.6
10. .0
7.2
2.8
1.0
d
VL
pc
(ymA)


24.3
10.6
4.1
2.1
1.2
0.66
•0.3S


0.602

Table B-4.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #8
IMPACTOR
STAGE
NUMBER

Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
247
247
246
245
245
238
- 194
134
- 59.0
V
(ymA)
	
21.6
9.5
4.5
1.8
1.1
0.58
0.35
	

0.134

OUTLET
"cum
(mg/DNm3)
29.2
27.5
26.0
25.3
25.3
25.3
24.9
21.0
14.2
• V
(ymA)
	
26.0
11.4
4.4
2.2
1.3
0.70
0.41


0.466

                           B-2

-------
Table B-5.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #9
IMPACTOR
STAGE
NUMBER

Precutter
5 Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
132

119
113
113
113
113
113
95.4
65.21
d
V
(ymA)


24.5
10.7
4.2
2.1
1.2
0.66
0.40
	

0.106

OUTLET
M
cum
(mg/DNm3)
4.1.5

41.3
40.6
40.6
40.6
40.6
39.5
23.0
13.9
d
PC
(ymA)


26.6
11.6
4.5
2.3
1.3
0.72
0.41
	

0.438

Table 6-6.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #10
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET PRIME
M
cum
(mg/DNn3)
244
244
244
244
244
244
224
137
5 7. -6
V
(ymA)
	
32.1
14.0
5.4
2.7
1.6
0.87
0.52
	
0.045
INLET
Mcum
(mg/DNm3)
166
107
105
103
100
100
54.7
29.6
27.4
dpc
(ymA)
	
24.7
10.8
5.1
2.1
1-2
0.67
0.40
	
0.044
OUTLET
Mcum
(mg/DNm3)
84.5
84.5
83.6
82.8
81.6
76.5
48.8
22.4
16.2
V
(ymA)
	
26.0
11.4
4.4
2.2
1.3
0.70
0.40
. 	
0.355
                           B-3

-------
Table B-7.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #13
IMPACTOR
STAGE
NUMBER

Precutter
5 Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET PRIME
M
cum
(mg/DNm3)
1789
1678
1651
1636
1569
1247
477
108
27.7
d
pc
(ymA)
	
26.9
11.8
4.6
2.3
1.3
0.73
0.44
_•__

0.115

INLET
M
cum
(mg/DNm3)
953
945
938
935
923
857
644
233
113
d
pc
(ymA)
	
26.6
11.7
4.5
2.3
1.3
0.72
0.42
	

0.088

OUTLET
M
cum
(ng/DNm3)
56.6
54.8
53.7
52.0
50.2
44.1
29.0
12.5
2.5
d
pc
(ymA)
	
23.2
10.2
3.9
2.0
1.1
0.63
0.36
	

0.279

Table B-8.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #14
IMPACTOR
STAGE
NUMBER

Precutter
§ Nozzle
1
2
3
4
5
6
7 .
Filter
Sample
Volume
(DNm3)
INLET
M
cuin
(mg/DNm3)
201
168
156
. 146
138
- 126
' 90.8
41.4
5.9
d
pc
(ymA)

11.1
4.6
2.6
1.4
0.88
0.48
0.28
	
0.051

OUTLET
M
cum
(mg/DNm3)
23.1
23.1
22.1
21.0
19.5
16.2
- 10.5
4.4
1.0
d
pc
(ymA)
	
17.3
7.6
2.9
1.5
0.85
0.47
0.27
---
0.771

                           B-4

-------
Table B-9.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #15
IMPACTOR
STAGE
NUMBER

Precutter
5 Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
419
410
399
382
360
195
84.5
36.8
7.4
d
pc
(ymA)
	
8.4
4.2
2.4
1.3
0.81
0.45
0.26
	

0.054

OUTLET
M
cum
(mg/DNm3)
35.5
35.2
33.5
31.4
27.2
21.6
12.3
4.9
1.8
d
pc
(ymA)
	
18.8
8.2
3.2
1.6
0.92
0.51
0.29
	

0.554

Table B-10
 INLET AND OUTLET SAMPLE PARTICLE DATA
 FOR RUN #16
IMPACTOR
STAGE
NUMBER

Precutter
5 Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DN'm3)
INLET
M
cum
(mg/DNm3)
182

177
174
168
163
149
- 108
58.0
' . 5.5
d
V
CymA)


10.1
4.2
2.4
1.3
0.80
0.44
0.25
	

0.036

OUTLET
M
cum
(mg/DNm3)
22.2

22.2
21.8
21.2
20.0
17.6
12.9
7.4
1.4
d
PC
(ymA)


18.0
7.9
3.1
1.5
0.89
0.49
0.28
	

0.933

                           B-5

-------
Table B-ll
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #17
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
. (DNm3)
INLET
M
cum
(mg/DNm3)
508
476
467
456
433
214
119
50.2
9.3
d
pc
(vimA)
	
9.4
4.8
2.7
1.5
0.90
0.50
0.29
	
0.054
OUTLET
M
cum
(mg/DNm3)
23.1
23.0
22.2
21.1
19.1
15. -9
10.3
4.1
1.4
d
pc
(ymA)
	
18.6
8.1
3.2
1.6
0.91
0.50
0.29
	
0.876
Table B-12.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #18
IMPACTOR
STAGE
NUMBER
Precutter
§ Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
168
129
126
121
116
111
103
78.4
- 24.5
dpc
(MmA)
	
11.3
4.7
2.7
1.5
0.89
0.49
0.28
	
0.061
OUTLET
Mcum
(mg/DNm3)
21.8
21.8
21.0
19.8
17.5
14.3
9.8
5.2
1.1
V
(pmA)
	
19.2
8.4
3.2
1.6
0.94
0.52
0.30
	
0.943
                          B-6

-------
Table B-13
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #19
IMPACTOR
STAGE
NUMBER

Precutter
5 Nozzle
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
INLET
M
cum
(mg/DNm3)
217
170
166
160
151
135
77,5
50.3
15.0
d
pc
(pmA)
	
8.7
4.4
2.5
1.4
0.83
0.46
0.27
	

0.074

OUTLET
M
cum
(mg/DNm3)
25.5
25.5
24.6
23.7
21.9
19.0
13.3
6.3
2.1
d
pc
(ymA)
	
1.9.5
8.5
3.3
1.7
0.96
0.53
0.30
	

0.663

                          B-7

-------
         APPENDIX C





REGRESSION ANALYSIS SUMMARY

-------
                                          Table  C-l.   Analysis  of Data  - With Wall Wash
                          OBJECTIVE
                                                 DATA SAMPLED
              RESULTS AND COMMENTS
o
i
            1.  Fit curve to eliminate effect of  inlet
                particle loading on efficiency, n.
                Must meet initial  conditions that at
                X = 0   n =  0, and at X = », n  <100.
2.   Determine additional  reduction in
    residuals after correcting for inlet
    loading.   Parameters  considered were
    inlet gas temperature, outlet gas
    flow rate, and  interaction.***
    The outlet gas  flow rate was corrected
    to standard dry condition.
                                           TRW DATA:9N, ION, UN, 12N,
                                           13N, 14N, 15N, 16N, D002,
                                           SI * S2, S3, S4, S5, S6.
                                           ATP DATA: ION, UN, 12N, 13N,
                                           15N.
                                                       TRW DATA ONLY.**
                                                       (Same as No. 1 above)
n = KQ(1 - e"KlX), KQ = 89.1,  K-,  =  .008746,
s(K0) = 4.7, s(K1) = .0013, s(R)  =  8.9.

(See Figure 1)
This curve is significantly different from
the curve obtained without wall wash.
Indicated addition correction  for inlet  gas
temperature, T.  The correction is
          40.6 - .318T.
The standard error of residuals,  s(R) = 8.0,
was reduced to 6.4.

(Completely opposite of results without
 wall wash)
            **
            ***
       denotes  screening  tests.
       ATP data wasn't  utilized because it was assumed that all input data was not available.
       Electrode Voltage  was  eliminated via visual examination of the plots of the parameters versus the residuals.

-------
                                        Table  C-2.   Analysis  of  Data  - Without Wall Wash
                                                                                                    RESULTS AND COMMENTS
           OBJECTIVE
      DATA SAMPLED
o
i
ro
             1.
             3.
Fit curve to eliminate  effect of  inlet
particle loading,  X,  on efficiency, n.
Must meet initial  conditions that at
X = 0,  n = 0,  and  at  X  = », n <100.
Determine addition  reduction  in
residuals after correcting  for
inlet loading.   Parameters  con-
sidered* were inlet gas  temperature,
outlet gas flow rate and interaction.
The outlet gas  flow rate was  corrected
to standard dry condition.

Determine additional reduction  in
residuals after correcting  for  inlet
loading. Censured data.
(n >90%, X >227 mg/m3).
Parameters considered were  electrode
voltage (stage 1 and stage  2),  inlet
gas temp and the actual  measured flow
rate  at the prevailing ambient
condition.
TRW DATA:  IN,  2N,  4N,  5N,
6N, 7N, 8N,  D003,  D004,
D006, D007,  D008,  D009,
D010, DOll,  D012,  D014,
D015, D016,  D019,  D020,
D021, D022,  D023,  D024,
D025.
ATP DATA:  2N,  4N,  7N.7N,
8N, 8N.

TRW DATA; **
IN, 2N, 4N,  5N, 6N, 8N,
D006, D007,  D008,  D009,
D010, D011,  D012,  D014,
DO!5, DO!6.
TRW DATA:***
IN, 2N, 4N, 5N, 6N, 8N,
D006, D007, D008, 0009,
DOll, D014, D016
ATP DATA
15, 17.
N = K0(l - e"KlX), KQ = 93.5,  K]  =  .014468,
                 ) = .0009669, s(R)  =  3.6.
                                                                                    s(K0)  =  .92,
                                                                                    (See  Figure  2)
                                                                                    This curve significantly different from
                                                                                    the curve obtained with wall wash.
Indicated additional correction for gas  flow
rate and interaction.  The correction  is
      4.1 + .00057R - .000007RT
The standard error of residuals, s(R)  =  1.7
was reduced to 1.5.
(Completely opposite of results with wall  wash)

Indicated additional correction for gas  flow
rate.  The correction is
      5.6 - .00029R.
The standard error of residuals, s(R)  =  1.76
was reduced to 1.57.

-------
                                 Table C-2.   Analysis of Data * Without Wall Wash  (Continued)
                              OBJECTIVE
      DATA SAMPLED
                 RESULTS AND COMMENTS
o
CO
            4.  Same as No. 3.  except gas flow rate
                corrected to standard dry condition.
            5.  Same as No. 3.  except used actual
                efficiencies.  It is assumed  the
                censored data is in a region  where
                inlet loading has a minimal effect.

            6.  Same as No. 5.  except used gas flow
                rate corrected to standard dry
                conditions.
Same as No.  3.
Same as No.  3.
No additional  correction  indicated.
No additional  correction  indicated.
Same as No.  3.
No additional  correction  indicated.
            *     Electrode Voltage was eliminated  via visual examination of the plot of the parameters versus the residuals.

            **    ATP data wasn't utilized because  it was assumed that all input data was not available.

            ***   All data obtained after 12/31/75  was deleted. Test equipment considered to be degraded.

-------
o
             95.
             85.
1 75.


u
z
LLJ
y

£ 65.
             45.
               65.       165.        265.        365.       465.        565.       665.       765.


                                                  INLET LOADING (MG/M3)
                                                                                         865.
                       Figure  C-1.  Particle Loading  Effect on  Removal Efficiency With  Wall Wash

-------
o
on
                                                                 DATA SOURCE
                                                               • APT
                                                               • DAD
                                                               A MATRIX
                                            265.        365.        465.
                                              INLET LOADING(MG/M3)
                  Figure C-2,  Particle  Loading Effect on Removal  Efficiency Without Wall Wash

-------
                               TECHNICAL REPORT DATA
                         (Please read Instructions on the reverie before completing)
1. REPORT NO.
 EPA- 600/2 -76-249b
                           2.
                                                      3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
 CHARGED DROPLET SCRUBBER FOR FINE
 PARTICLE CONTROL: PILOT DEMONSTRATION
                                 5. REPORT DATE
                                  September 1976
                                 6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

W. F. Krieve and J. M. Bell
                                                     8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 TRW Defense and Space Systems Group
 One Space Park
 Redondo Beach, California 90278
                                  10. PROGRAM ELEMENT NO.
                                  1AB013; ROAP 21ADL-043
                                  11. CONTRACT/GRANT NO.
                                  68-02-1345
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC  27711
                                  13. TYPE OF REPORT AND PERIOD COVERED
                                  Phase: 7/74-6/76	
                                  14. SPONSORING AGENCY CODE
                                   EPA-ORD
15. SUPPLEMENTARY NOTES JERL-RTP project officer for this report is D.L. Harmon, Mail
Drop 61, 919/549-8411,  Ext 2925.
16. ABSTRACT
          The report gives results of a successful Charged Droplet Scrubber (CDS)
pilot demonstration of coke oven emissions control,  It also describes the design,
installation, and checkout of the demonstration system.  The CDS uses electrically
sprayed water droplets, accelerated through an electric field, to remove particulate
material from a gas stream.  The pilot demonstration was a continuation of laboratory
and bench scale studies for application of the CDS to fine particle control.  The pilot
demonstration included, in addition  to the CDS, the ducting,  flow transitions, and
blower necessary to circulate process gas through the CDS.  The test was performed
at the  Kaiser Steel Company coke oven facility, Fontana, California.  A large fraction
of the  coke  oven emissions  were submicron and composed of carbon particles and
hydrocarbon aerosol.  After the system checkout was completed,  during which CDS
operating parameters were established,  the demonstration test series was performed.
Results of the demonstration test indicate that  the  CDS is an effective pollution control
device for controlling coke  oven stack emissions.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                      b.lDENTIFIERS/OPEN ENDED TERMS
                        c. COSATI Field/Group
Air Pollution
Scrubbers
Dust
Electrostatics
Coking
Carbon
Hydrocarbons
Aerosols
A ir Pollution Control
Stationary Sources
Particulate
Charged Droplets
Charged Droplet Scrub-
  ber
13B
07A
11G
20C
13H
07B
07C
07D
18. DISTRIBUTION STATEMENT
 Unlimited
                                          19. SECURITY CLASS (This Report)
                                          Unclassified 	  	
                                              21. NO. OF PAGES

                                                  82	
                      20. SECURITY CLASS (This page)
                      Unclassified
                                              22. PRICE
EPA Form 2220-1 (9-73)
                                      82

-------