vvEPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-450/3-89-27C
August 1989
Air
Municipal Waste
Combustors-
Background
Information for
Proposed Standards:
Post-Combustion
Technology
Performance
This document is printed on recycled paper.
-------
MUNICIPAL WASTE COMBUSTORS --
BACKGROUND INFORMATION FOR
PROPOSED STANDARDS: POST-COMBUSTION TECHNOLOGY PERFORMANCE
FINAL REPORT
Prepared for:
Michael G. Johnston
U.S. Environmental Protection Agency
Industrial Studies Branch (MD-13)
Research Triangle Park, North Carolina 27711
Prepared by:
Radian Corporation
3200 E. Chapel Hill Rd./Nelson Hwy.
Post Office Box 13000
September 22, 1989
-------
DISCLAIMER
This report has been reviewed by the Emission Standards Division
of the Office of Air Quality Planning and Standards, EPA, and
approved for publication. Mention of trade names or commercial
products is not intended to constitute endorsement or
recommendation for use. Copies of this report are available
through the Library Services Office (MD—35), U.S. Environmental
Protection Agency, Research Triangle Park NC 27711, or from
National Technical Information Services, 5285 Port Royal Road,
Springfield VA 22161.
11
-------
TABLE OF CONTENTS
Page
1.0 INTRODUCTION 1-1
1.1 Overview of Report 1-1
1.2 Overview of Uncontrolled Emissions 1-1
1.2.1 Particulate Matter 1-1
1.2.2 Metals 1-2
1.2.3 CDD/CDF 1-5
1.2.4 Acid Gas 1—6
1.3 References 1-8
2.0 ELECTROSTATIC PRECIPITATORS 2-1
2.1 Process Description 2-1
2.2 Summary of Test Data 2-5
2.2.1 Mass Burn MWC’s 2-5
2.2.1.1 Alexandria 2-5
2.2.1.2 Baltimore RESCO 2-7
2.2.1.3 Bay County 2-9
2.2.1.4 Dayton 2-9
2.2.1.5 McKay Bay 2-19
2.2.1.6 North Andover 2-22
2.2.1.7 Peekskill 2-25
2.2.1.8 Pinellas County 2-33
2.2.1.9 Quebec City 2-36
2.2.1.10 Tulsa 2-45
2.2.2 RDF MWC’s 2-49
2.2.2.1 Lawrence 2-49
2.2.2.2 Niagara Falls 2-49
2.2.2.3 Red Wing (Northern States Power) 2-55
2.2.3 Modular MWC’s 2-60
2.2.3.1 Barron County 2-60
2.2.3.2 Oneida County 2-60
2.2.3.3 Oswego County 2-63
2.2.3.4 Pigeon Point 2-67
2.2.3.5 Pope/Douglas 2-78
111
-------
TABLE OF CONTENTS (Continued)
Page
2.3 Sumary of Performance . 2-81
2.3.1 Particulate Matter 2-81
2.3.2 Metals 2-88
2.3.3 CDD/CDF 2-91
2.4 References 2-95
3.0 FURNACE SORBENT INJECTION 3-1
3.1 Process Description 3-1
3.2 Suninary of Test Data 3-2
3.2.1 Alexandria 3-2
3.2.2 Dayton 3-6
3.3 Sunm ary of Performance 3-17
3.3.1 Acid Gas 3—17
3.3.2 PartIculate Matter 3-18
3.3.3 Metals 3-18
3.3.4 CDD/CDF 3-19
3.4 References 3-20
4.0 DUCT SORBENT INJECTION FOLLOWED BY AN ELECTROSTATIC
PRECIPITATOR 4-1
4.1 Process Description 4-1
4.2 Suninary of Test Data 4-1
4.2.1 Dayton 4-2
4.3 Suiimiary of Performance 4-12
4.3.1 Acid Gas 4-12
4.3.2 Particulate Matter 4-13
4.3.3 Metals 4-13
4.3.4 CDD/CDF 4-13
4.4 References 4-14
i v
-------
TABLE OF CONTENTS (Continued)
Page
5.0 DUCT SORBENT INJECTION FOLLOWED BY A FABRIC FILTER 5-1
5.1 Process Description 5-1
5.2 Summary of Test Data 5-2
5.2.1 Claremont 5-2
5.2.2 Dutchess County 5-5
5.2.3 Quebec City 5-16
5.2.4 Springfield 5-23
5.2.5 St. Croix 5-26
5.2.6 Wurzburg 5-33
5.3 Summary of Performance 5-37
5.3.1 Acid Gas 5-37
5.3.2 Particulate Matter 5-37
5.3.3 Metals 5-37
5.3.4 CDD/CDF 5-40
5.4 References 5-42
6.0 SPRAY DRYING FOLLOWED BY AN ELECTROSTATIC PRECIPITATOR 6-1
6.1 Process Description 6-1
6.2 Summary of Test Data 6-3
6.2.1 Millbury 6-3
6.2.2 Munich 6-11
6.2.3 Portland 6-18
6.3 Sumary of Performance 6-20
6.3.1 Acid Gas 6-20
6.3.2 Particulate Matter 6-24
6.3.3 Metals 6-24
6.3.4 CDD/CDF 6-24
6.4 References 6-26
7.0 SPRAY DRYING FOLLOWED BY A FABRIC FILTER 7-1
7.1 Process Description 7-1
7.2 Summary of Test Data 7-1
V
-------
TABLE OF CONTENTS (Continued)
Page
7.2.2
7.2.3
7.2.4
7.2.5
7.2.6
7.2.7
7.2.8
7-59
7.3.1 7-59
7.3.2 7-65
7.3.3 7-65
7.3.4 7-66
7-69
7.2.1 Biddeford
Commerce
Long Beach
Mid-Connecticut
Marion County
Penobscot
Quebec City
Stan i si aus County
7.3 Summary of Performance
7-2
7-9
7-17
7-22
7-28
7-40
7-42
7-52
Acid Gas
Particulate Matter
Metals
CDD/CDF
7.4 References
APPENDIX A. Hydrogen Chloride and Sulfur
Dioxide Emissions Data
A-i
vi
-------
LIST OF TABLES
______ Page
1-1 1-2
1-2 1-3
1-3 1-5
1-4
1-7
2-1 Particulate Data 2-6
2-2 Particulate Data 2-8
2-3 Metals Emissions 2-10
2-4 Particulate Data 2-11
2-5 Particulate Data 2-13
2-6 Metals Data for 2-16
2-7 CDD/CDF Data for 2-17
2-8 Particulate and 2-21
2-9 Particulate Data 2-23
2-10 CDD/CDF Data for 2-24
2-11 Particulate Data 2-28
2-12 Metals Emissions 2-31
2-13 CDD/CDF Data for 2-32
2-14 Particulate Data 2-35
2-15 Metals Emissions 2-37
2-16 CDD/CDF Data for 2-38
2-17 Particulate Data 2-42
2-18 Metals Emissions 2-43
2-19 CDD/CDF Data for 2-44
2-20 Particulate Data 2-46
TABLE
Summary of Uncontrolled Particulate Matter Concentrations...
Summary of Uncontrolled Metals Concentrations
Summary of Uncontrolled CDD/CDF Concentrations
Summary of Uncontrolled Acid Gas (SO 2 and HC1)
Concentrations
For Alexandria Without Lime Injection
for Baltimore RESCO
Data for Baltimore RESCO
for Bay County
for Dayton Without Sorbent Injection
Dayton Without Sorbent Injection
Dayton Without Sorbent Injection
Metals Emissions Data for McKay Bay
for North Andover
North Andover
for Peekskill
Data for Peekskill
Peekskill
for Pinellas County
Data for Pinellas County
Pinellas County
for Quebec City
Data for Quebec City
Quebec City
for Tulsa
vii
-------
LIST OF TABLES (Continued)
TABLE
2-21
2-22
2-23
2-24
2-25
2-26
2-27
2-28
2-29
2-30
2-31
2-32
2-33
2-34
2-35
2-36
2-37
2-38
2-39
2-40
2-41
3-1
3-2
Metals Emissions
CDD/CDF Data for
Particulate Data
CDD/CDF Data for
Particulate Data
CDD/CDF Data for
Particulate Data
Metals Data for
CDD/CDF Data for
Particulate Data
Metals Emissions
Particulate Data
Metals Emissions
CDD/CDF Data for
Particulate Data
CDD/CDF Data for
Particulate Data
Metals Emissions
CDD/CDF Data for
Particulate Data
Metals Emissions
Page
2-47
2-48
2-50
2-51
2-53
2-54
2-56
2-58
2-59
2-61
2-62
2-64
2-65
2-66
2-68
2-71
2-75
2-76
2-77
2-79
2-80
2-82
3-4
3-5
Data for Tulsa
Tulsa
for Lawrence
Lawrence
for Niagara Falls
Niagara Falls
for NSP Red Wing
NSP Red Wing
NSP Red Wing
forBarronCounty
Data for Barron County
for Oneida County
Data for Oneida County
Oneida County
for Oswego County
Oswego County
for Pigeon Point
Data for Pigeon Point
Pigeon Point
for Pope/Douglas
Data for Pope/Douglas
2-42 CDD/CDF Data for Pope/Douglas
Acid Gas Data for Alexandria with Lime Injection
Particulate Data for Alexandria with Lime Injection... ......
v i i i
-------
LIST OF TABLES (Continued)
TABLE Page
3-3 CDD/CDF Data for Alexandria with Lime Injection 3-5
3-4 Acid Gas Data for Dayton with Furnace Sorbent Injection 3-7
3-5 Particulate Data for Dayton with Furnace Sorbent Injection.. 3-13
3-6 Metals Data for Dayton with Furnace Sorbent Injection 3-14
3-7 CDD/CDF Data for Dayton with Furnace Sorbent Injection 3-16
4-1 Acid Gas Data for Dayton with DSI 4-3
4-2 Particulate Data for Dayton with OS! 4-9
4-3 Metals Data for Dayton with DSI 4-10
4-4 CDD/CDF Data for Dayton with DSI 4-11
5-1 Acid Gas Data for Claremont 5-4
5-2 Particulate Data for Claremont 5-6
5-3 CDD/CDF Data for Claremont 5-7
5-4 HC1 Data for Dutchess County 5-9
5-5 SO 2 CEM Data for Dutchess County 5-10
5-6 Particulate Data for Outchess County 5-14
5-7 Metals Data for Dutchess County 5-15
5-8 CDD/CDF Data for Dutchess County 5-17
5-9 Acid Gas Data for Quebec City Pilot DSI/FF 5-19
5-10 Metals Data for Quebec City Pilot DSI/FF 5-24
5-11 CDD/CDF Data for Quebec City Pilot DSI/FF 5-25
5-12 Acid Gas Data for St. Croix 5-28
5-13 Fabric Filter Inlet Acid Gas Data for St. Croix 5-30
5-14 Particulate Data for St. Croix 5-32
5-15 Metals Data for St. Croix 5-32
ix
-------
LIST OF TABLES (Continued)
TABLE Pane
5-16 COD/COF Data for St. Croix 5-34
5-17 Acid Gas Data for Wurzburg 5-34
5-18 Particulate Data for Wurzburg 5-35
5-19 Metals Data for Wurzburg 5-35
5-20 CDD/CDF Data for Wurzburg 5-36
6-1 Acid Gas Data for Milibury 6-5
6-2 Particulate Data for Millbury 6-8
6-3 Metals Emissions Data for Millbury 6-9
6-4 CDD/CDF Data for Millbury 6-10
6-5 Acid Gas Data for Munich 6-13
6-6 Particulate Data for Munich 6-16
6-7 Metals Emissions Data for Munich 6-17
6-8 AcId Gas Data from Portland 6-19
6-9 CDD/CDF Data for Portland 6-22
7-1 Acid Gas Data for Biddeford 7-4
7-2 Particulate Data for Biddeford 7-7
7-3 Metals Data for Biddeford 7-8
7-4 CDD/CDF Data for Biddeford 7-10
7-5 AcId Gas Data for Conm*rce 7-12
7-6 Particulate Data for Coniuerce 7-14
7-1 Metals Emissions Data for Con’mierce 7-15
7-8 CDD/CDF Data for Comerce 7-16
7-9 Acid Gas Data for Long Beach 7-19
7-10 Particulate Data for Long Beach 7-20
x
-------
LIST OF TABLES (Continued)
TABLE Pag
7-11 Metals Data for Long Beach 7-21
7-12 CDD/CDF Data for Long Beach 7-23
7-13 Particulate Data for Mid-Connecticut 7-24
7-14 Metals Emissions Data for Mid-Connecticut 7-26
7-15 CDD/CDF Data for Mid-Connecticut 7-27
7-16 Acid Gas Data for Marion County 7-30
7-17 Particulate Data for Marion County 7-38
7-18 Metals Data for Marion County 7-39
7-19 CDD/CDF Data for Marion County 7-41
7-20 Acid Gas Data for Penobscot 7-43
7-21 Particulate Data for Penobscot 7-43
7-22 Metals Data for Penobscot 744
7-23 CDD/CDF Data for Penobscot 7-45
7-24 Acid Gas Data for Quebec City Pilot SD/FF 7-47
7-25 Particulate Data from Quebec City Pilot SD/FF 7-50
7-26 Metals Emissions Data for Quebec City Pilot SD/FF 7-51
7-27 CDD/CDF Data for Quebec City Pilot SD/FF 7-53
7-28 Acid Gas Data for Stanislaus County 7-55
7-29 Particulate Data for Stanislaus County 7-57
7-30 Metals Data for Stanislaus County 7-58
7-31 COD/COF Data for Stanislaus County 7-60
xi
-------
LIST OF FIGURES
Figures Page
2-1 Electrical Resistivity of Municipal Incinerator Dust.... 2-3
2-2 Outlet PM Concentration as a Function of Inlet
PM Concentration at the Dayton MWC 2-15
2-3 Outlet CDD/CDF Concentration as a Function
of ESP Inlet Temperature at Dayton 2-18
2-4 Outlet COD/COF Concentration as a Function of Inlet
CDD/CDF Concentration at Dayton 2-20
2-5 CDD/CDF Removal Efficiency as a Function of Inlet
CDD/CDF Concentration at the North Andover MWC 2-26
2-6 PM Removal Efficiency as a Function of Inlet PM
Concentration at the Peekskill MWC 2-30
2-7 CDD/CDF Removal Efficiency as a Function of Inlet
CDD/CDF Concentration at the Peekskill MWC 2-34
2-8 Outlet CDD/CDF Concentration as a Function of Inlet
CDD/CDF Concentration at the Pinellas County MWC 2-39
2-9 Relative Increase of CDD/CDF Across ESP as a Function
of Inlet CDD/CDF Concentration at the Pinellas County
MWC 2-40
2-10 PM Removal Efficiency as a Function of Inlet PM
Concentration at the Oswego County MWC 2-69
2-11 Outlet PM Concentration as a Function of Inlet PM
Concentration at the Oswego County MWC 2-70
2-12 CDD/CDF Removal Efficiency as a Function of ESP
Inlet Temperature at the Oswego County MWC 2-72
2-13 CDD/CDF Removal Efficiency as a Function of Inlet
CDD/CDF Concentration at the Oswego County MWC 2-73
2-14 PM Removal Efficiency and Outlet Concentration as a
Function of the Number of ESP Fields 2-83
2-15 PM Removal Efficiency as a Function of Actual SCA 2-85
2-16 Outlet PM Concentration as a Function of Design SCA 2-86
2-17 Relationship Between ESP Inlet and Outlet
PM Concentrations 2-87
xii
-------
LIST OF FIGURES (Continued)
Figures Page
2-18 Metals and PM Removal Efficiency for ESP’s with PM
Removal Efficiency Above 98 Percent 2-89
2-19 Metals and PM Removal Efficiency for ESP’s with
Removal Efficiency Below 98 Percent 2-90
2-20 CDD/CDF Removal Efficiency as a Function of ESP Inlet
Temperature at MWC’s with ESP’s 2-92
2-21 Outlet CDD/CDF Concentration as a Function of ESP Inlet
Temperature at Dayton 2-94
3-1 Inlet SO Concentration as a Function of ESP Inlet
Temper ture and Sorbent Feed Rate for the FSI
System at Dayton 3-10
3-2 Inlet HC1 Concentration as a Function of ESP Inlet
Temperature and Sorbent Feed Rate for the FSI System
at Dayton 3-11
4-1 SO , Removal Efficiency as a Function of Stoichiometric
Ratio at the Dayton DSI/ESP System 4-6
4-2 HC1 Removal Efficiency as aFunction of Stoichiometric
Ratio at the Dayton DSI/ESP System 4-7
5-1 Outlet SO Concentration as a Function of Inlet SO 2
Concent ation at Dutchess County ‘5-12
5-2 SO 9 Removal Efficiency as a Function of FF Inlet
Temperature at the Quebec City DSI/FF System 5-21
5-3 HC1 Removal Efficiency as a Function of FF Inlet
Temperature at the Quebec City DSJ/FF System 5-22
5-4 SO 9 Removal Efficiency as a Function of Stoichiometric
Ratio at the St. Croix DSI/FF System 5-31
5-5 SO, Removal Efficiency as a Function of FF Inlet
Temperature for DSI/FF Systems 5-38
5-6 HC1 Removal Efficiency as a Function of FF Inlet
Temperature for DSI/FF System 5-39
5-7 Outlet CDD/CDF Concentration as a Function of FF
Inlet Temperature for DSI/FF System 5-41
xiii
-------
LIST OF FIGURES (Continued)
Figures Pane
6-1 Outlet SO and HC1 Concentrations as a Function of
Inlet S 2 Concentration at Milibury 6-7
6-2 SO , and HC1 Removal Efficiency as a Function of
Stoichiometric Ratio at Munich 6-14
6-3 SO , and HC1 Removal Efficiency as a Function of
SP Inlet Temperature at Munich 6-15
6-4 SO, Removal Efficiency as a Function of ESP Inlet
1 Temperature at Portland 6-21
7-1 SO, Removal Efficiency as a Function of Stoichiometric
Ratio at Biddeford 7-5
7-2 HC1 Removal Efficiency as a Function of Stoichiometric
Ratio at Biddeford 7-6
7-3 SO, Removal Efficiency as a Function of Stoichiometric
Ratio at Marion County 7-32
7-4 HC1 Removal Efficiency as a Function of Stoichiometric
Ratio at Marion County... 7—33
7-5 SO Removal Efficiency as a Function of Inlet SO 2
oncentration at Marion County 7-35
7-6 HC1 Removal Efficiency as a Function of Inlet SO 2
Concentration at Marion County 7-36
7-7 HC1 Removal Efficiency as a Function of SO 2
Removal Efficiency at Marion County 7-37
7-8 SO, Removal Efficiency as a Function of Stoichiometric
Ratio at Quebec City 7-49
7-9 SO, Removal Efficiency as a Function of Lime Slurry
Feed Rate at Stanislaus County 7-56
7-10 SO, Removal Efficiency as a Function of St 8 ichiometric
Ratio for SD/FF Systems at Less than 300 F 7-61
7-11 HC1 Removal Efficiency as a Function of St 8 ichiometric
Ratio for SD/FF Systems at Less than 300 F 7-62
7-12 SO Removal Efficiency as a Function of FF Inlet
T mperature for SD/FF Systems 7-63
xiv
-------
LIST OF FIGURES (Continued)
Figures Page
7-13 HC1 Removal Efficiency as a Function of FF Inlet
Temperature for SD/FF Systems 7-64
7-14 Effect of Inlet PM on Mercury Emissions for Systems
with Spray Dryer/Fabric Filter Systems 7-67
7-15 Effect of Inlet CDD/CDF on Mercury Emission for MWC5
with Spray Dryer/Fabric Filter Systems 7-68
xv
-------
1.0 INTRODUCTION
1.1 OVERVIEW OF REPORT
This document evaluates the performance of various air pollution control
devices applied to new and existing municipal waste combustors (MWC’s). The
control devices analyzed include electrostatic precipitators (ESP’s), furnace
sorbent injection systems with ESP’s, duct sorbent injection systems with
ESP’s and fabric filters (FE’s), and spray dryers with ESP’s and FF’s. Each
control device is discussed in a separate section of this document.
Performance capabilities of each control device are evaluated for the
following pollutants: particulate matter (PM), metals (arsenic, cadmium,
chromium, lead, mercury, and nickel), chlorinated dibenzo-p-dioxins and
dibenzofurans (COD/CDF), and acid gases (sulfur dioxide [ SO 2 ] and hydrogen
chloride [ HC1]).
These evaluations include assessments of the key parameters affecting
control device performance for the various pollutants and the development of
correlations between parameters and performance to establish operating
conditions that yield high removal efficiencies and low emission rates for
each pollutant.
1.2 OVERVIEW OF UNCONTROLLED EMISSIONS
This section presents information on uncontrolled air emissions from
MWC’s. This information is presented to provide a basis for comparison with
controlled emissions at facilities where data on uncontrolled emission levels
are not available. Information is presented for PM, metals, CDD/CDF, and
acid gases (HC1 and SO 2 ) for each major combustor type.
1.2.1 Particulate Matter
Particulate emissions data collected at the inlet to the pollution
control device are summarized in “Municipal Waste Combustion Assessment:
Combustion Control at Existing Facilities” and a.re presented in Table 1-1.
From mass burn refractory wall combustors, typical uncontrolled PM
concentrations are 3 grains per dry standard cubic foot (gr/dscf) at
7 percent oxygen (02). Typical uncontrolled concentrations for mass burn
waterwall, mass burn rotary waterwall, and modular excess air combustors are
1—1
-------
TABLE 1-1. SUMMARY OF UNCONTROLLED PARTICULATE MATTER CONCENTRATIONS 1
Concentration
Combustor Type (gr/dscf at 7% 02)
Mass burn refractory wall 3
Mass burn waterwall;
Mass burn rotary waterwall; 2
Modular excess air
Refuse-derived fuel fired 4
Modular starved air 0.15
2 gr/dscf. At refuse derived fuel (ROF) combustors, typical uncontrolled PM
concentrations are 4 gr/dscf and at modular starved air combustors, typical
uncontrolled PM concentrations are 0.15 gr/dscf. Particulate data presented
in this document are generally normalized to 12 percent carbon dioxide (C0 2 ),
which results in values approximately six percent lower than when normalized
to 7 percent 02.
1.2.2 Metals
Emissions of metals in flue gas from MWC’s potentially have adverse
health effects. Certain metals considered to have the greatest effect--
arsenic (As), cadmium (Cd), chromium (Cr), lead (Pb), mercury (Hg), and
nickel (Ni) -- have been measured in flue gas emissions from a number of
MWC’s. Uncontrolled metal concentrations measured at individual MWC’s are
suninarized in Table 1-2 along with average concentrations for each combustor
type. All concentrations are presented in micrograms per dry, standard cubic
meter (ug/dscm) normalized to 7 percent 02.
Based on the available data, the highest uncontrolled arsenic levels are
from RDF combustors which average 615 ug/dscm and range from 203 to 1,060
ug/dscm. The data on mass-burn combustors average 216 ug/dscm and range from
1-2
-------
TABLE 1-2. SUMMARY OF UNCONTROLLED METALS CONCENTRATIONS
Concentration
Site As Cd Cr
(u ldscm at 7% 02)
Pb Hg Ni Reference
Mass Burn
Baltimore RESCO 226 NMa 2,960 NM NM NM 2
Commerce 1987 b 220 2,700 730 50,000 450 680 3
Commerce (1988) 74 1,600 3,450 17,200 450 4,000 4
Dayton 234 1,550 185 38,400 1,030 94 5
Gallatin 422 3,130 1,040 36,300 248 NM 6
Marion County NM 1,120 422 20,500 NM 12 7
Quebec City (pilot) 128 1,220 1,870 34,700 373 1,300 8
Average 216 1,890 1,520 32,800 510 1,220
Modular
Cattaraugus County 34 1,090 1,210 20,400 1,130 1,260 9
Prince Edward Island 14 1,120 71 18,400 921 553 10
Tuscaloosa 99 NM 34 NM NM NM 11
Average 49 1,110 436 19,400 1,020 905
RDF
Biddeford 583 1,280 3,170 31,300 440 NM 12
Mid-Connecticut 1,060 1,070 927 37,400 1,010 541 13
NSP Red Wing 203 805 381 24,300 140 344 14
Average 615 1,050 1,490 31,000 530 443
aNM Not measured.
bTests conducted firing a mixture of residential and commercial refuse. Tests
firing only comercial refuse yielded similar emissions, but are not reported
here.
1-3
-------
24 to 422 ug/dscm. Modular combustors average 49 ug/dscm for uncontrolled
arsenic, with a range of 14 to 99 ug/dscm.
Uncontrolled cadmium levels are similar for all combustor types. At
mass burn combustors, uncontrolled cadmium concentrations average 1,890
ug/dscm. From RDF combustors, uncontrolled cadmium emission average 1,050
ug/dscm. Uncontrolled cadmium emissions from modular combustors average
1,110 ug/dscm.
Chromium emissions from the combustor vary widely from facility to
facility. At mass burn combustors, uncontrolled emissions range from 185 to
3,450 ug/dscm and average 1,520 ug/dscm. Similarly, uncontrolled chromium
emissions from RDF combustors vary between 381 and 3,170 ug/dscm and average
1,490 ug/dscm. At modular units, uncontrolled chromium levels range between
34 and 1,210 ug/dscm and average 436 ug/dscm.
Uncontrolled lead emissions are the highest of any of the metals
presented in Table 1-1 and are relatively consistent within a combustor type.
Uncontrolled lead emissions from mass burn combustors average 32,800 ug/dscm.
From RDF combustors, the average uncontrolled lead emissions are
31,300 ug/dscm. Uncontrolled lead emissions from modular MWC’s average
19,400 ug/dscm.
Uncontrolled mercury emissions vary widely from facility to facility.
From mass burn combustors, uncontrolled mercury emissions range from 250 to
1,030 ug/dscm and average 510 ug/dscm. Uncontrolled values for ROE
combustors are between 140 and 1,010 ug/dscm, averaging 530 ug/dscm. The
available uncontrolled mercury emissions data from modular units are 920 and
1,130 ug/dscm, for an average of 1,020 ug/dscm.
Uncontrolled nickel emissions also vary significantly from each
facility. For mass burn combustors, uncontrolled nickel emissions range from
12 to 4,000 ug/dscm and average 1,220 ug/dscm. The uncontrolled nickel
emissions data available for RDF combustors are 344 and 540 ug/dscm for an
average of 443 ug/dscm. Uncontrolled nickel emissions for two modular
combustors are 550 and 1,260 ug/dscm, for an average of 905 ug/dscm.
The average uncontrolled metals concentrations are used in this document
to estimate removal efficiencies at individual MWC’s where uncontrolled
metals emissions are not available. Arsenic, cadmium, chromium, lead, and
1-4
-------
nickel are associated with particulate and are consistently removed to
relatively low levels across a particulate control device. Thus, the
estimated removal efficiency is probably a reasonable approximation of the
actual removal efficiency. For mercury, however, with the relatively wide
variation in uncontrolled mercury concentrations and the general
understanding that mercury is not associated with particulate, a realistic
mercury removal efficiency cannot be estimated with the available data unless
the outlet mercury concentration is substantially lower than the lowest
reported uncontrolled mercury concentration. Thus, outlet mercury emissions
data in this document are only compared to the range in uncontrolled mercury
concentrations in Table 1-2 to indicate whether mercury removal may have
occurred.
1.2.3 CDD/CDF
Uncontrolled CDD/CDF emission levels are summarized in “Muncipal Waste
Combustion Assessment: Combustion Control at Existing Facilities” and are
presented in Table 1-3. All concentrations reflect the sum of the tetra-
through octa-chiorinated COO/CDF homologues and are presented in nanograms
per dry, standard cubic meter (ng/dscm) normalized to 7 percent 02. From
mass burn refractory combustors, typical uncontrolled CDD/CDF concentrations
TABLE 1-3. SUMMARY OF UNCONTROLLED CDD/CDF CONCENTRATIONS’ 5
Concentrati on
Combustor Type (ng/dscm at 7% 02)
Mass burn refractory 4,000
Mass burn waterwall - large 500
Mass burn waterwall - midsize; 200
Modular excess air
Mass burn waterwall - small;
Refuse-derived fuel fired; 2,000
Mass burn rotary waterwall
Modular starved air 400
1-5
-------
are 4,000 ng/dscm. At large mass burn waterwall combustors, typical
concentrations are 500 ng/dscm. Midsize mass burn waterwall and modular
excess air combustors have typical uncontrolled COD/COF concentrations of
200 ng/dscm. Three combustor types: small mass burn waterwall, RDF, and
mass burn rotary waterwall combustors have typical uncontrolled CDD/CDF
concentrations of 2,000 ng/dscm. At modular starved air combustors, typical
concentrations are 400 ng/dscm.
1.2.4 Acid Gas
Uncontrolled acid gas (SO 2 and HC1) emissions are summarized in
Table 1-4. All acid gas concentrations are on a dry basis and are presented
in parts per million by volume (ppm) normalized to 7 percent 02. For mass
burn combustors, uncontrolled SO 2 concentrations range from 59 to 330 ppm and
average 180 ppm; HC1 concentrations range from 450 to 900 ppm and average
650 ppm. Data from the one RDF combustor indicate uncontrolled SO 2 emissions
of 100 ppm; HC1 emissions are 580 ppm. For modular combustors, uncontrolled
SO 2 concentrations range from 66 to 150 ppm and average 110 ppm; HC1
concentrations range from 190 to 570 ppm and average 420 ppm. Average
uncontrolled SO 2 and HC1 concentrations of 200 and 500 ppm are used in this
document to estimate removal efficiencies at MWC’s for which uncontrolled
acid gas data are not available.
1-6
-------
TABLE 1-4. SUMMARY OF UNCONTROLLED ACID GAS (SO 2 AND HC1) CONCENTRATIONS
(ppm, dry at 7% 0 .1
Site HC References
Mass Burn
Claremont NMa 450 16
Commerce (1987) 270 900 17
Commerce (1988) 110 650 18
Long Beach 140 NM 19
Marion County (1986) 180 570 20
Marion County (1987) 330 680 21
Milibury (Unit 1) 210 770 22
Milibury (Unit 2) 300 730 23
Munich 92 630 24
Portland 300 NM 25
Quebec City 130 450 26
Stanislans County (Unit 1) 67 NM 27
Stanislans County (Unit 2) 59 NM 28
Average 180 650
Modul ar
Cattaraugus 150 190 29
Prince Edward Island 66 490 30
St. Croix 120 570 31
Average 110 420
RDF
Biddeford 100 580 32
aNM = Not measured.
1-7
-------
1.3 REFERENCES
1. US. Environmental Protection Agency. Municipal Waste Combustion
Assessment: Combustion Control at Existing Facilities.
EPA-600/8-89-058. August 1989.
2. PEI Associates, Inc. Method Development and Testing for Chromium, No. 2
Refuse-to-Energy Incinerator, Baltimore RESCO. (Prepared for U. S.
Environmental Protection Agency. Research Triangle Park, North Carolina
EMB Report 85-CHM8. EPA Contract No. 68-02-3849. August 1986.
3. McDannel, M. D., L. A. Green, and B. L. McDonald (Energy Systems
Associates). Air Emissions Tests at Commerce Refuse-to-Energy Facility,
May 26 - June 5, 1987. Prepared for County Sanitation Districts of Los
Angeles County. Whittier, California, July 1987.
4. McDannel, M. D., 1. A. Green, and A. C. Bell (Energy Systems
Associates). Results of Air Emission Test During the Waste-to-Energy
Facility. Prepared for County Sanitation Districts of Los Angeles
County, Whittier, California. December 1988.
5. RadIan Corporation. Preliminary Data from October - November 1988
Testing at the Montgomery County South Plant, Dayton, Ohio.
6. Cooper Engineers, Inc. Air Emissions Tests of Solid Waste Combustion in
a Rotary Combustion/Boiler System at Gallatin, Tennessee. Prepared for
West County Agency of Contra Costa County, California. July 1984.
7. VancIl, M. A. and C. L. Anderson (Radian Corporation). Summary Report,
CDD/CDF, Metals, HC1, SO ,, NO , CO and Particulate Testing, Marion
County Solid Waste-to-Energy facility, Inc., Ogden Martin Systems of
Marion, Brooks, Oregon. Prepared for U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina. EMB Report No.
86-MIN-03A. September 1988.
8. The National Incinerator Testing and Evaluation Program: Air Pollution
Control Technology. EPS 3/UP/2, Environment Canada, Ottowa, September
1986.
9. New York State Department of Environmental Conservation. Emission
Source Test Report -- Preliminary Test Report on Cattaraugus County ERF.
August 5, 1986.
10. Environment Canada. The National Incinerator Testing and Evaluation
Program: Two Stage Combustion. Report EPS 3/up/i. September 1985.
ii. PEI Associates, Inc. Method Development and Testing for Chromium,
Municipal Refuse Incinerator, Tuscaloosa Energy Recovery, Tuscaloosa,
Alabama. Prepared for U.S. Environmental Protection Agency. Research
Triangle Park, North Carolina. EMB Report 85-CHM-9. EPA Contract No.
68-02-3849. January 1986.
1-8
-------
12. Klan ii, S., G. Scheil, M. Whitacre, J. Surman (Midwest Research
Institute), and W. Kelly (Radian Corporation). Emission Testing at an
RDF Municipal Waste Combustor. Prepared for U. S. Environmental
Protection Agency, North Carolina. EPA Contract No. 68-02-4453, May 6,
1988.
13. Anderson, C. 1. (Radian Corporation). CDD/CDF, Metals, and Particulate
Emissions Summary Report, Mid-Connecticut Resource Recovery Facility,
Hartford, Connecticut. Prepared for U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina. EMB Report
No. 8a-MIN-o9A. January 1989.
14. Interpoll Laboratories. Results of the March 21 - 26, 1988, Air
Emission Compliance Test on the No. 2 Boiler at the Red Wind Station,
Test IV (High Load). Prepared for Northern States Power Company,
Minneapolis, Minnesota. Report No. 8-2526. May 10, 1988.
15. U.S. Environmental Protection Agency. Municipal Waste Combustion
Assessment: Combustion Control at Existing Facilities.
EPA-600/8-89-058. August 1989.
16. Almega Corporation. SES Claremont, Claremont, NH, NH/VT Solid Waste
Facility. Unit 1 and Unit 2, EPA Stack Emission Compliance Tests, May
26, 27, and 29, 1987. Prepared for Clark-Kenith, Inc. Atlanta,
GeorgIa. July 1987.
17. McDannel, M. 0., 1. A. Green, and A. C. Bell (Energy Systems
Associates). Results of Air Emission Test During the Waste-to-Energy
Facility. Prepared for County Sanitation Districts of Los Angeles
County, Whittier, California. December 1988.
18. Radian Corporation. Preliminary Data from October - November 1988
Testing at the Montgomery County South Plant, Dayton, Ohio.
19. Ethier, 0. 0., 1. N. Hottenstein, and E. A. Pearson (TRC Environmental
Consultants). Air Emission Test Results at the Southeast Resource
Recover6y Facility Unit 1, October - December, 1988. Prepared for Dravo
Corporation, Long Beach, California. February 28, 1989.
20. Vancil, M. A. and C. L. Anderson (Radian Corporation). Summary Report,
CDD/CDF, Metals, HC1, SO ,, NO , CO and Particulate Testing, Marion
County Solid Waste-to-En rgy aci1ity, Inc., Ogden Martin Systems of
Marion, Brooks, Oregon. Prepared for U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina. EMB Report No.
86-MIN-03A. September 1988.
21. Anderson, C. L., et. al. (Radian Corporation). Characterization Test
Report, Marion County Solid Waste-to-Energy Facility, Inc., Ogden Martin
Systems of Marion, Brooks, Oregon. Prepared for U. S. Environmental
Protection Agency, Research Triangle Park, North Carolina. EMB Report
No. 86-MIN-04. September 1988.
1-9
-------
22. Entropy Environmentalists, Inc. Emissions Testing Report, Wheelabrator
Milibury, Inc. Resource Recovery Facility, Millbury, Massachusetts,
Unit Nos. 1 and 2, February 8 through 12, 1988. Prepared for Rust
International Corporation. Reference No. 5605-B. August 5, 1988.
23. Entropy Environmentalists, Inc. Emissions Testing Report, Wheelabrator
Milibury, Inc. Resource Recovery Facility, Millbury, Massachusetts,
Unit Nos. 1 and 2, February 8 through 12, 1988. Prepared for Rust
International Corporation. Reference No. 5605-B. August 5, 1988.
24. Hahn, J. L., et. al., (Cooper Engineers) and J. A. Finney, Jr. and B.
Bahor (Belco Pollution Control Corp.). Air Emissions Tests of a
Deutsche Babcock Anlagen Dry Scrubber System at the Munich North
Refuse-Fired Power Plant. Presented at: 78th Annual Meeting of the Air
Pollution Control Association. Detroit, Michigan. June 1985.
25. Engineering Science, Inc. A Report on Air Emission Compliance Testing
at the Regional Waste Systems, Inc. Greater Portland Resource Recovery
Project. Prepared for Dravo Energy Resources, Inc. Pittsburg,
Pennsylvania. March 1989.
26. The National Incinerator Testing and Evaluation Program: Air Pollution
Control Technology. EPS 3/UP/2, Environment Canada, Ottowa, September
1986.
27. Hahn, J. 1. (Ogden Projects, Inc.)
for Stanislaus Waste Energy Company
Report No. 177R. April 7, 1989.
28. Hahn, J. 1. (Ogden Projects, Inc.)
for Stanislaus Waste Energy Company
Report No. 177R. April 7, 1989.
30. Environment Canada.
Program: Two Stage
31. Interpoll Laboratories, Inc.
Performance Test at the St.
in New Richmond, Wisconsin.
Englewood, Colorado. Report
Environmental Test Report.
Crows Landing, California.
Environmental Test Report.
Crows Landing, California.
Prepared
OPI
Prepared
OP I
29. New York State Department of Environmental Conservation. Emission
Source Test Report -- Preliminary Test Report on Cattaraugus County ERE.
August 5, 1986.
The National Incinerator Testing and Evaluation
Combustion. Report EPS 3/up/i. September 1985.
Results of the June 6, 1988, Scrubber
Croix Waste to Energy Incineration Facility
Prepared for Interel Corporation.
No. 8-25601. September 20, 1988.
32. Kiam, S., G. Scheil, M. Whitacre, J. Surman (Midwest Research
Institute), and W. Kelly (Radian Corporation). Emission Testing at an
ROF Municipal Waste Combustor. Prepared for U. S. Environmental
Protection Agency, North Carolina. EPA Contract No. 68-02-4453, May 6,
1988.
1-10
-------
2.0 ELECTROSTATIC PRECIPITATORS
Section 2.0 describes the performance of electrostatic precipi-
tators. Electrostatic precipitators have been used extensively to reduce
particulate matter emissions from MWC’s. Section 2.1 describes ESP
design and operating characteristics. Section 2.2 describes the
available emissions test data from ESP-equipped MWC’s. In Section 2.3,
ESP performance is evaluated for PM, metals, and CDD/CDF. Because ESP’s
are not designed to reduce emissions of acid gases, data on SO 2 , HC1, and
nitrogen oxides (NOr) are not presented in this section.
2.1 PROCESS DESCRIPTION
Electrostatic precipitators consist of a series of high voltage (20
to 100 kV) discharge electrodes and grounded metal plates through which
PM-laden flue gas flows. Negatively charged ions formed by this high
voltage field (known as a “corona”) attach to PM in the flue gas, causing
the charged particles to migrate toward and be collected on the grounded
plates. As a general rule, the greater the amount of collection plate
area, the greater the ESP’s PM collection efficiency. The most common
ESP types used by MWC’s are: (1) plate-wire units in which the discharge
electrode is a bottom-weighted or rigid wire and (2) flat plate units
which use flat plates rather than wires as the discharge electrode.
Plate-wire ESP’s are generally better suited for use with fly ashes with
large amounts of small particulate and with large flue gas flow rates
(>200,000 actual cubic feet per minute [ acfm]). Flat plate units are
less sensitive to back corona problems and are thus well suited for use
with high resistivity PM. 1 Both of these ESP types have been widely used
on MWC’s in the U. S., Europe, and Japan.
The layer of particles collected on the plates is removed by
rapping, washing, or other methods. When this dust layer is removed,
some of the collected PM is reentrained in the flue gas. To assure good
PM collection efficiency during plate cleaning and electrical upsets,
ESP’s have multiple fields located in series along the direction of flue
2-1
-------
gas flow that can be energized and cleaned independently. Particles
reentrained when the collected PM is removed from one field can be
recollected in a downstream field. 2 Because of this phenomena,
increasing the number of fields generally improves particulate removal
efficiency.
In general, fly ashes with resistivities between 1 x i0 8 and
5 x 10 ° ohm-cm and with a minimum of very fine particles (<1 micron) are
most efficiently collected. If the resistivity of the collected PM
exceeds roughly 2 x 1010 ohm-cm, the collected PM layer may have
sufficient electrical charge to create a “back corona” phenomenon that
interferes with the migration of charged fly ash particles to the
collecting electrode and significantly reduces collection efficiency. At
resistivities below io8 ohm-cm, the electrical charge on individual
particles may be so low that reentrainment of collected dust during
electrode cleaning or simply as a result of contact with moving flue gas
can become severe. 3 Resistivity is greatly affected by temperature, as
shown in a graph of resistivity versus temperature for three municipal
solid waste (MSW) fly ashes in Figure 2-1. Most ESP’s on MWC’s have
traditionally operated at 440 to 550°F (225 to 290°C) to avoid potential
problems with ash resistivity. 4 However, individual ESP’s with
temperatures as low as 380°F (195°C) and as high as 600°F (315°C) are
currently operating in the U. S. A concern with ESP operation at lower
temperatures is the potential for acid corrosion of cool material
surfaces due to HC1 condensation.
Small particles generally migrate toward the collection plates more
slowly than large particles, and are therefore more difficult to collect.
This factor is especially important to MWC’s because of the amount of
total fly ash less than 1 micron. For MWC’s, 20 to 70 percent of the fly
ash at the ESP inlet is less than 1 micron. 5 In comparison, for
pulverized coal-fired combustors, only 1 to 3 percent of the fly ash is
generally less than 1 micron. Effective collection of MWC PM will
require greater collection areas and lower flue gas velocities than many
other types of PM.
2-2
-------
TEMPERATURE, °F
100 200 300 400 500 600
TEMPERATURE, °C
TRADITIONAL
ESP
OPERATING
TEMP.
A AREA OF BACK
CORONA
- DEVELOPMENT
RANGE OF
RESISTIVITY
BEST SUITED
FOR ESP
OPERATION
1 Samples taken at furnace outlet on a 250 ton/day municipal
using a dry separation chamber for particulate control.
incinerator
2 Samples taken at furnace outlet and exhaust stack inlet on a 250 ton/day
municipal incinerator using a wet baffle cooling chamber for particulate
control.
3 Samples taken at furnace outlet and exhaust stack outlet on a 120 ton/day
municipal incinerator using a vertical wetted baffle particulate
collection device.
Source: Walker, A.B. and Schmitz, Characteristics of Furnace Emissions
from Large chanically-Stoked Municipal Incinerators ,
Re search-Cottre 11
Figure 2-1. Electrical Resistivity of Municipal Incinerator Dust
3
E
U
0
I-,
( -S
Cd)
C l)
‘U
U
‘U
‘U
1013
1012
1011
1010
10
10$
iO
50 100 150 200 250 300 350
2-3
-------
The collection efficiency of an ESP can be estimated using the
Deutsch- Anderson equation:
Collection Efficiency (%) = (1 - exp(-Aw/V))100
where exp is the natural log (2.718...), A is the surface area of the
collecting electrodes (ft 2 ), w is the effective migration velocity of
individual PM particles toward the collecting electrode (ft/sec), and V
is the actual flue gas flow rate (acfm). However, because of variations
in the size and resistivity of individual particles in the flue gas, the
effective migration velocity of bulk fly ash is not easily defined.
To account for these variations in PM characteristics, the modified
Deutsch-Anderson equation is used:
Collection Efficiency (%) = (1 - exp( Aw/V)k)1O0
where k is a constant (generally around 0.5, but can vary between 0.4 and
0.8) that depends on the electrical resistivity and size of the fly ash
particles and w is now an empirically derived migration velocity.
As an approximate indicator of collection efficiency, the specific
collection area (SCA) of an ESP is frequently used. The SCA is
calculated by dividing the collecting electrode plate area by the actual
flue gas flow rate (A/V in the Deutsch-Anderson equation) and is
expressed as square feet of collecting area per 1,000 acfm of flue gas.
In general, the higher the SCA, the higher the collection efficiency.
One problem encountered with existing ESP’s is the amount of
particulate bypass around the ESP ionizing fields and collection plates.
This bypass limits the level to which outlet PM emissions can be reduced.
New ESP’s can be designed to significantly reduce the amount of bypass.
A recently identified concern with the operation of MWC ESP’s is the
potential for formation of CDD/CDF across the ESP. The mechanism and
extent of formation is poorly understood, but is believed to be promoted
by copper in MWC fly ash, carbon in the PM, and surface area for reaction
at temperatures between roughly 450 and 650°F. 6
2-4
-------
2.2 SUMMARY OF TEST DATA
Section 2.2 provides emissions data from various ESP-equipped MWC
facilities. The facilities are limited to those which were constructed
after 1984 and were designed for good particulate control. Data for
ESP’s built before 1984 show a substantially lower level of performance.
Section 2.2.1 covers mass burn combustors, Section 2.2.2 covers modular
(excess-air and starved-air) combustors, and Section 2.2.3 covers RDF
combustors. A summary of the emissions data is provided for each
facility.
2.2.1 Mass Burn MWC’s
2.2.1.1 Alexandria. 7 The Alexandria/Arlington Resource Recovery
Facility in Alexandria, Virginia, consists of three identical 325 tonI
day (tpd) Martin GrnbH waterwall combustors. Emissions from each
coinbustor are controlled with a 3-field ESP. The flue gas temperature at
the ESP outlet is generally about 340°F with a gas flow of about
67,000 acfm (40,000 dry standard cubic feet per minute [ dscfm]). The
ESP’s are designed to meet a PM emission limit of 0.03 gr/dscf at
12 percent CO 2 . Dry hydrated lime can be injected into the combustor
with the overfire air for acid gas control.
In December 1987, tests were performed to demonstrate compliance
with operating permits. The tests were conducted under normal operating
conditions with dry hydrated lime injection on Unit 1 and without lime
injection on Units 2 and 3. The test results without lime injection are
reported here. The test results with lime addition are reported in
Section 3.2.1. Flue gas was sampled at the ESP outlet for PM on the
three test runs conducted without dry lime injection.
The PM data from the tests without lime injection are summarized in
Table 2-1. Flue gas flow rate and temperature were relatively constant
among each of the runs. Measured PM emissions from Unit 2 ranged from
0.029 to 0.031 gr/dscf at 12 percent CO 2 and averaged 0.030 gr/dscf.
2-5
-------
TABLE 2-1. PARTICULATE DATA FOR ALEXANDRIA WITHOUT LIME INJECTION
Test Condition
Run
a
Number
ESP Inlet
Tempsrabure
( F)
Flue Gas
Flow
(acfm)
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Combustor = Normal
ESP = Normal
2-1
2-2
2-3
360
366
366
64,500
66,200
64,900
0.029
0.031
0.029
Average (Unit 2)
364
65,200
0.030
3-1
3-2
3-3
367
363
362
62,800
66,200
66,500
0.013
0.012
0.051
Average (Unit 3)
364
65,200
0.025
aRUfl Number contains the unit number followed by the run number on that unit.
bEstimated from measured 0 temperature at ESP outlet and an assumed temperature
drop across the ESP (20 F).
2-6
-------
Particulate emissions from Unit 3 ranged from 0.012 to 0.051 gr/dscf and
averaged 0.025 gr/dscf. Thus, both units are capable of meeting the permit
PM emission limit of 0.03 gr/dscf.
2.2.1.2 Baltimore RESCO. 8 ’ 9 The Baltimore RESCO facility in Baltimore,
Maryland, consists of three identical 750 ton/day, Von Roll, waterwall
combustors. Particulate emissions from each combustor are controlled by a
4-field, wire/plate ESP designed by Wheelabrator Frye. Each ESP has a design
SCA of 577 ft 2 /1,000 acfm. The flue gas flow at the ESP inlet is typically
about 240,000 acfm (123,000 dscfm) at a temperature of 460°F, although the
ESP has operated at lower temperatures (380°F). The ESP exhaust streams are
separately ducted and routed through an induced-draft (ID) fan into a common
stack.
In January 1985, tests were conducted to demonstrate compliance with
permit conditions for all three combustor trains. At the ESP outlet, three
runs were conducted on each unit in which PM, SO 2 , fluorides, solid and
gaseous chlorides, and NO were measured. In May 1985, testing was done on
Unit 2 at the ESP inlet and outlet as part of a method development effort for
chromium. These tests were conducted to measure chromium, arsenic, and PM.
Although these tests were designed to provide emissions data under normal
operating conditions, the steam load was only 85 percent of normal during
these tests.
Particulate data from both test programs are presented in Table 2-2.
For all runs conducted, outlet PM emissions did not exceed 0.007 gr/dscf at
12 percent CO 2 . The average outlet PM concentrations were 0.0020, 0.0044,
and 0.0010 gr/dscf at Units 1,2, and 3, respectively, for the compliance
tests. At Unit 2 at 85 percent steam load, the outlet PM emissions averaged
0.0027 gr/dscf, with an average PM removal efficiency of greater than 99.9
percent. Although the flue gas flow rate was approximately 60 percent higher
during the 85 percent steam load tests on Unit 2, yielding lower SCA values
(570 versus 910 ft 2 /1,000 acfm), outlet PM emissions were similar. This
suggests that increasing the SCA above the design value by decreasing the
flue gas flow does not necessarily improve PM performance of the ESP.
2-7
-------
TABLE 2-2. PARTICULATE DATA FOR BALTIMORE RESCO
aRUfl Number contains unit number followed by run number on that unit.
bEt.td from temperature measured at ESP outlet and an assumed temperature
based on measured vaLues for other tests.
CNM = Not measured.
N.)
Test Condition
Run
a
Number
ESP InLet
Temperature
C F)
Flue
Gas Flow
(acfm)
Inlet PM
Concentration
(gr/dscf
at 12% C0 2 )
OutLet PM
Concentration
(gr/dscf
at 12% C0 2 )
PM
RemovaL
Efficiency
(%)
Combustor z NormaL
ESP NormaL
1-1
1-2
1-3
385
385 b
377
176,100
171,600
161,400
NMC
MM
MM
0.0004
0.0017
0.0038
-
-
Average (Unit 1)
382
169,700
NM
0.0020
-
2-1
2-2
2-3
365 b
366 b
369
153,700
150,800
149,100
NM
NM
NM
0.0066
0.0032
0.0035
-
-
-
Average (Unit 2)
367
151,200
NM
0.0044
-
3 .
3-2
33
375 b
371
370
147,400
134,400
136,700
NM
NM
NM
0.0014
0.0004
0.0013
-
-
-
Average (Unit 3)
372
139,500
MM
0.0010
Combustor 85% Load
ESP = NormaL
2-1
2-2
2-3
465
463
462
239,000
251,400
236,800
2.27
2.03
1.86
0.0030
0.0030
0.0020
99.9
99.9
99.9
Average (85% Load)
453
242,400
2.05
0.0027
99.9
drop across the ESP (11°F)
-------
Chromium and arsenic data from Baltimore are presented in Table 2-3.
Removal efficiencies for both metals were high, 97 percent for arsenic and
99 percent for chromium.
2.2.1.3 Bay County. 10 The Bay County Resource Management Facility in
Panama City, Florida, is designed to conibust MSW or MSW with wood chips in
two identical 255 tons/day Westinghouse-O’Connor rotary waterwall combustors.
Emissions from each combustor are controlled by a 3-field wire/plate ESP
manufactured by Environmental Elements Corporation. At the ESP inlet, the
design flue gas temperature is 400 0 F and the flow is 56,000 acfm. Each ESP
has an SCA of 350 ft 2 /1,000 acfm and is designed to achieve 99 percent PM
removal. The permitted outlet PM concentration is 0.03 gr/dscf at 12 percent
CO 2 . The flue gas exits through individual flues in a single 125-ft high
stack.
Compliance tests were conducted in June 1987. Particulate emissions
were measured at the ESP outlet for three tests at each unit. The results
from the tests are presented in Table 2-4. Only MSW was combusted during the
tests. The PM emissions from Unit 1 ranged from 0.014 to 0.024 gr/dscf at 12
percent CO 2 and averaged 0.019 gr/dscf. At Unit 2, PM emissions ranged from
0.019 to 0.029 gr/dscf and averaged 0.024 gr/dscf. Inlet PM was not
measured. Both units were able to demonstrate emissions less than the permit
level of 0.03 gr/dscf.
2.2.1.4 Dayton. 11 The Montgomery County South Incinerator plant in
Dayton, Ohio, includes three nearly identical Volund refractory-lined
combustors. Two 300-tpd units were built in 1970 and a third 300-tpd unit
was built in 1988. Limestone can be injected through a single injection port
into the furnace ignition chamber of each unit to reduce SO 2 emissions.
Typically, the unit operates with 250 lb/hr limestone injection. Water
sprays are used to cool the flue gas exiting the combustor. Emissions are
controlled by United-McGill 3-field plate-to-plate ESP’s made of Cor-ten
steel. Each ESP has a design SCA of 326 ft 2 /1,000 acfm. The design flue gas
flow rate at the ESP inlet is 100,000 acfm at a temperature of 450 to 600°F,
2-9
-------
TABLE 2-3. METALS EMISSIONS DATA FOR BALTIMORE RESCO
Test Condition
Run
Numbera
ESP Inlet
Temp 8 rature
C F)
Outlet PM
Concentration
(gr/dscf
at 12% CD 2 )
Inlet
Concentration
7%
Outlet
Concentration
(ug/dscm at 7%
Removal
Efficiency
(ua/dscm at
As
021
Cr
As
021
Cr
As Cr
Combustor 85%
load
2-1
465
0.0030
NMb
3,600
NM
MM
36.5
16.2
-- 99.0
-. 99.4
ESP Normal
2-2
2-3
463
462
0.0030
0.0020
MM
226
2,560
2,710
5.8
34.8
97.4 98.7
Average
463
0.0027
226
2,960
5.8
29.2
97.4 99.0
5 RUfl Number contains unit number followed by run number on that unit.
— b
a NM Not measured.
-------
TABLE 2-4. PARTICULATE DATA FOR BAY COUNTY
Test Condition
Run
a
Number
ESP Inlet
Temp 8 ra ure
( F)
Flue Gas
Flow
(acfm)
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Combustor Normal
ESP = Normal
1-1
1-2
1-3
450
454
452
52,400
55,100
52,800
0.014
0.024
0.020
Average (Unit 1)
452
53,400
0.019
2-1
2-2
2-3
454
474
476
52,600
58,100
59,000
0.025
0.019
0.029
Average (Unit 2)
468
56,600
0.024
aRUfl Number contains the unit number followed by the run number on that unit.
bEstimated from measured 0 temperature at ESP outlet and an assumed temperature
drop across the ESP (25 F).
2-11
-------
with a typical value of 575°F. Flue gas is exhausted through a common stack
for Units 1 and 2. Unit 3 exhausts through a dedicated stack.
In November and December 1988, testing was conducted by EPA on Unit 3.
Tests were conducted with furnace sorbent injection, duct sorbent injection,
and without sorbent injection. The purpose of these tests was to evaluate
ESP performance for the removal of PM, metals, and CDD/CDF at reduced flue
gas temperatures at the ESP inlet and to evaluate the effect of sorbent
injection. The results from the tests without sorbent injection are reported
here. Results of the tests with furnace sorbent injection are reported in
Section 3.2.2. Results from the duct sorbent injection tests are reported in
Section 4.2.
Testing was performed in two phases: screening and parametric. The
screening tests were performed to help select operating conditions for the
parametric tests. During the screening tests, measurements of SO 2 and HC1
were taken at the ESP inlet during all tests and at the ESP outlet for some
of the sorbent injection tests. During the parametric testing, flue gas was
sampled simultaneously at the ESP inlet and outlet and analyzed for PM, SO 2 ,
CDD/CDF, metals (16 metals including arsenic, cadmium, chromium, lead,
mercury, nickel, and others), and volatile organics. Hydrogen chloride was
measured at the ESP outlet only except for the tests with duct injection
during which HC1 was measured at both the ESP inlet and outlet. CDD/CDF was
also measured at the mixing chamber during several of the tests. Nine test
runs were conducted during parametric testing without sorbent injection.
The particulate data from testing without sorbent injection are
presented in Table 2-5. The outlet particulate emissions ranged from 0.003
to 0.023 gr/dscf at 12 percent CO 2 and averaged 0.011 gr/dscf for seven of
the nine test runs. The ESP malfunctioned during Runs 6 and 7, yielding
excessive outlet PM levels. During Runs I through 5, PM emissions averaged
0.0065 gr/dscf. Particulate removal efficiency during the seven acceptable
runs averaged 98.5 percent. No trends were observed in PM removal versus SCA
or inlet PM concentration. Outlet PM concentrations tended to be higher with
2-12
-------
TABLE 2-5. PARTICULATE DATA FOR DAYTON WITHOUT SORBENT INJECTION
Test Condition Run
Number
ESP Inlet
Temp rature
( F)
Flue
Gas Flow
(acfm)
InLet PM
Concentration
(gr/dscf
at 12% C0 2 )
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
PM
Removal
Efficiency
(%)
Combustor = Norma ,
ESP = Normal (575 F)
1
2
3
567
548
564
71,800
78,000
80,300
0.376
0.568
0.983
0.0030
0.011
0.0057
99.2
98.1
99.4
Average
560
76,700
0.642
0.0066
98.9
Combustor = NormaL
ESP = Low (600°F)
4
5
6
402
400
400
79,200
72,500
66,200
0.590
0.527
0.578
0.0038
O.OO8
0.105
99.4
98.4
818 a
Average
401
72,600
0.565
0 • 0063 b
989 b
Combustor = Low mix ng
chamber temp. (15 0 F)
ESP = NormaL (525 F)
7
8
9
559
559
545
73,400
68,900
82,000
0.378
0.725
0.978
0.12?
0.022
0.023
675 a
97.0
97.7
Average
554
74,800
0.684
0023 b
974 b
aErroneous data points because of unstable ESP voltage. Values not included in average.
bA excLudes erroneous data points for runs 6 and 7.
-------
higher inlet PM concentrations, but outlet concentrations below 0.01 gr/dscf
were achieved even at inlet PM concentrations approaching 1.0 gr/dscf, as
shown in Figure 2-2. Thus, the 3-field ESP at Dayton with an SCA of about
320 ft 2 /1,000 acfm operating at 575°F or less, achieved outlet PM levels of
0.01 gr/dscf at 12 percent CO 2 or less.
The metals data from Dayton are presented in Table 2-6. During Runs 1
to 5, removal efficiencies for arsenic, cadmium, and lead were about 98 to
99 percent and were similar to the PM removal efficiencies. Chromium removal
efficiencies were somewhat lower, around 95 percent. The metals removal
efficiencies were somewhat lower for Runs 8 and 9, as were the PM removal
efficiencies. Mercury concentrations at the inlet were not significantly
different from the outlet values, indicating that no mercury removal occurred
across the ESP.
In Table 2-7, CDD/CDF data are presented. At an ESP inlet set point
temperature of 575°F, outlet CDD/CDF concentrations ranged from 7,790 to
22,500 ng/dscm at 7 percent 02 and averaged 17,100 ng/dscm. The inlet
CDD/CDF concentrations were between 112 and 391 ng/dscm and averaged 252
ng/dscm. At an ESP inlet set point temperature of 525°F and a lower mixing
chamber temperature, outlet CDD/CDF emissions of 13,800 to 15,700 ng/dscm at
7 percent 02 were measured, with an average of 14,500 ng/dscm. The
corresponding ESP inlet concentrations were between 150 and 247 ng/dscm and
averaged 214 ng/dscm. At the mixing chamber, CDD/CDF concentrations of 552
and 2,450 ng/dscm were measured. At an ESP inlet set point temperature of
400°F, outlet CDD/CDF concentrations ranged from 317 to 1,580 ng/dscm at 7
percent 02 and averaged 866 ng/dscm. The corresponding ESP inlet CDD/CDF
concentration ranged from 10.2 to 51 ng/dscm and averaged 32.8 ng/dscm. The
mixing chamber CDD/CDF concentrations ranged from 798 to 3,120 ng/dscm and
averaged 1,870 ug/dscm for the three runs conducted.
The flue gas temperature at the ESP inlet appears to significantly
affect CDD/CDF emissions at the ESP outlet, as shown in Figure 2-3. At
400°F, CDD/CDF emissions are roughly an order of magnitude lower than at 525
to 575°F. Except for Run 1, outlet CDD/CDF emissions at 525°F are 25 to
40 percent lower than emissions at 575°F.
2-14
-------
0.023 - _____ ____________________________________
0.022 - —- ---—--—— 0
• Condition 1(575 deg F)
0.021 —
f Condition 2 (400 dsg F)
0.02 Condition 6 (525 d.g F)
o 0.019-
° 0.018 -
(‘1
0.017 -
o 0.016 -
(n
0.015 -
I . .
2’ 0.014 -
C
o 0.013 —
0.012
C
a) 0.011 -
N)
— 0.01
C.)
0.009 -
+
0.008
a>
0.007 -
0
0.006 --
U
0.005 -
0.004 +
0.003-——— U
0.3 0.5 0.7 0.9
inlet PM Concentration (gr/dscf 12% C0 2 )
Figure 2-2. Outlet PM concentration as a function of inlet PM
concentration at the Dayton MWC.
-------
TABLE 2-6. METALS DATA FOR DAYTON WITHOUT SORSENT INJECTION
Outlet PM
ESP Inlet Concentration
Run Temp r.tur. (grldscf at Inlet CQric.ntr.tlon (uald$Cm at 7% °,L Outlet Concentration (ua/decm at 7% 0,).., Removal Efficiency 1% )
Test Condition Number ( F) 12% C0 2 ) As Cd Cr Pb Hg Ni As Cd Cr Pb Hg NI As Cd Cr Pb Hg NI
Combustor Norma 1 567 0.0030 189 489 114 27,700 845 98.6 2.17 10.5 5.91 322 625 3.80 98.9 97.9 94.8 98.8 26.0 96.2
tSP Normal (575 F) 2 548 0.011 308 2,460 170 67,300 890 68.4 3.52 53.9 4.91 973 1,034 2.86 98.9 97.8 97.1 98.6 -16.2 95.8
3 564 0.0057 168 2,905 84 13,160 1,150 94.7 1.85 24.4 3.52 309 1,390 1.58 98.9 99.2 95.8 97.7 -20.9 98.3
Average 560 0.0066 222 1.950 123 36,050 962 87.2 2.51 29.6 4.78 535 1,016 2.74 98.9 98.3 95.9 98.4 3.7 96.8
Combustor Normal 4 402 0.0038 165 1,455 150 15,635 920’ 78.4 1.53 16.2 4.44 602 631 2.12 99.1 98.9 97.0 96.2 31.4 97.3
ESP Low temp (400 0 F) 5 400 O.008 223 930 154 21,057 750 59.3 4.3k 2? 4 9.7 515 769 9.14 98.1 97.6 93.7 98.1 - 2.5
6 400 0.105 247 1,527 261 56,415 1.495 104 42.9 283 46.3 22,640’ 2.053 21.5 526 S 815 a 82.7’ 59.9’ -37.3 79.3
Average 401 0.0063 212 1,304 190 33,056 1,055 80.6 2.94 19.3 7.11 559 1,150 5.63 98.6 98.3 95.4 96.7 - 2.8 90.9
Combustor Low mix ,ng 7 559 O.l23 124 669 156 44,048 2,255 74.9 55.6’ 702 a 58.7 18,577 1,816 29.6’ 55.2’ 4•95 62.4’ 57.8 19.5 60 • 5 a
chamber temp (ISgO F) 8 559 0.022 230 1,994 242 50,598 975 148 10.3 69.0 11.1 1,836 1,076 9.25 95.5 96.5 95.4 96.6 -10.2 93.8
ESP Normal (525 F) 9 545 0.023 518 1,626 309 53,781 949 123 11.4 89.5 10.2 3,202 1,128 4.52 97.8 94.5 96.7 94.1 -18.9 96.3
Average 554 0.023 291 1,430 236 49,476 1,393 115 10.9 79.3 10.7 2,519 1,229 6.89 96.7 95.5 96.1 95.3 -3.2 95.0
Overall Average 505 0.011 242 1,561 183 39,527 1,137 94.4 5.03 40.9 7.14 1,109 1,148 4.75 98.2 97.5 95.8 97.0 -3.2 94.6
5 Not included in average because the first field of the tsp was unstable during the test, leading to unusually poor performance.
-------
TABLE 2-7. CDD/CDF DATA FOR DAYTON WITHOUT SORBENT [ NJEC1 iON
—I
Test Condition
Run
Number
ESP InLet
Temperature
( F)
Mixing Chamber
CDD/CDF
Concentration
(rig/dscm at 7% 02)
InLet CDD/CDF
Concentration
(ng/dscm
at 7% 02)
OutLet CDD/CDF
Concentration
(ng/dscm
at °2
CDD/CDF
RemovaL
Efficiency
Combustor = Norrna
ESP NormaL (575 F)
1
2
3
573
568
573
NMa
MM
NM
254
391
112
7,790
22,500
21,100
-2,970
-5,650
-18,800
Average
571
NM
252
17,100
-9,140
Combustor = NormaL
ESP = Low temp. (400°F)
4
5
6
396
407
386
552
2,450
MM
10.2
37 ,1
51.0
317
704
1,580
-3,010
-1,790
-3,000
Average
396
1,500
32.8
866
-2,600
Combustor = Low mixing
chamber temp. (15 0 F)
ESP NormaL (525 F)
7
8
9
548
525
528
798
3,120
1,680
150
247
247
14,000
15,700
13,800
-9,230
-6,260
-5,490
Average
534
1,870
214
14,500
-6,990
aNN not measured.
-------
24000 -
22000 - U
U
20000 -
18000 -
16000 *
C
14000- U
. 12000 -
C
0
U
10000 -
C.)
8000
( )
— N
co
o 6000-
C .)
4000-
2000- U
N
0— I I I
380 400 420 440 460 480 500 520 540 560 580
ESP Inlet Temperature (°F)
Figure 2-3. Outlet CDD/CDF concentration as a function of
ESP Inlet temperature at Dayton.
-------
The test runs at Dayton at ESP inlet temperature setpoints of 400 and
525°F showed an 81 to 98 percent decrease in CDD/CDF concentrations from the
mixing chamber to the ESP inlet. CDD/CDF was not measured at the mixing
chamber during the 575°F runs. At 400°F, the CDD/CDF concentrations
decreased by up to 71 percent from the mixing chamber to the ESP outlet. At
525°F, the ESP outlet CDD/CDF concentrations were greater than at the mixing
chamber. This suggests that the [ SF’ provides surface area and residence time
for CDD/CDF formation, with the amount of formation at least partially
dependent on the temperature.
Inlet CDD/CDF may affect outlet CDD/CDF concentration, as shown in
Figure 2-4. At ESP inlet concentrations of 10 to 50 ng/dscm, emissions of
320 to 1,990 ng/dscm resulted. At [ SF’ inlet concentrations of 150 to
390 ng/dscm, CDD/CDF emissions of 13,800 to 22,500 ng/dscm resulted with the
exception of Run 1. Run 1 had an outlet CDD/CDF concentration of 7,790
ng/dscm and an inlet concentration of 254 ng/dscm. However, as discussed
previously, this observation may actually be a result of inlet temperature.
At a given temperature, neither the outlet CDD/CDF concentrations nor the
degree of COD/CUE formation change with the inlet CDD/CDF concentration.
2.2.1.5 McKay Bay.U 3 The McKay Bay Refuse-to-Energy Facility in
Tampa, Florida, consists of four identical process lines each designed to
combust 250 tons/day of MSW. The combustors are Volurid refractory units
which use drying and ignition grates followed by a rotary kiln. Emissions
from each combustor are controlled by a F. L. Smidth 2-field ESP, each with
an SCA of 445 ft 2 /1,000 acfm and design PM removal efficiency of
99.45 percent. At the ESP inlet, the flue gas flow is typically about
83,000 acfm (38,000 dscfm) at a temperature of about 565°F. Treated flue gas
is released through two stacks (two process lines per stack).
A compliance test was conducted at the facility in September 1985. The
facility was operating at normal combustor and ESP conditions. During the
test program, three runs were conducted at each unit, with PM. SO 2 , HF, THC.
N0 , mercury, lead, and beryllium emissions measured at the ESP outlet.
Particulate data from the test are reported in Table 2-8. The average
outlet PM concentrations for each of the units were 0.013, 0.012, 0.0042, and
0.0079 gr/dscf. No inlet PM data were collected.
2-19
-------
24000 -
U
22000
U
20000 -
F ,-
E 18000
U
to
16000
C
14000- U
. 12000 -
C
10000
C)
I I . -
C)
N
C)
. 4000-
0
2000 - U
U
- -- If!
0 100 200 300 400
inlet CDD/CDF Concentration (ng/dscm @ 7% 02)
Figure 2-4. Outlet CDD/CDF concentration as a function of inlet
CDD/CDF concentration at Dayton.
-------
TABLE 2-8. PARTICULATE AND METALS EMISSIONS DATA FOR McKAY BAY
N)
aRUfl Number contains unit number folLowed by run number on that unit.
bEsp inlet temperature estimated from measur d temperature at ESP outLet and a estimated
drop based on previously measured vaLues (6 F for Unit 1, 30°F for Unit 2, 12 F for Unit
Unit 4).
temp r8tUre
3, 8 F for
Test Condition
Run
Number
a
ESP Inlet b
Tempe atUre
C F)
Flue
Gas Flow
(acfm)
OutLet PM
Concentration
(gr/dscf
at 12% CO 2 )
Outlet Concentration
(ug dscm at 7% 021
Pb Hg
Combustor = NormaL
ESP = Normal
1-7
i-a
1-9
536
553
564
82,000
86,700
89,400
0.015
0.013
0.010
1,630
763
808
573
452
1,000
675
Average (Unit 1)
551
86,000
0.013
1,070
1,280
2-7
2-8
2-9
552
570
570
82,400
82,700
82,500
0.012
0.012
0.011
1,240
1,140
1,040
522
913
905
Average (Unit 2)
564
82,500
0.012
1,140
3-7
3-8
3-9
558
557
564
77,300
77,300
77,300
0.0040
0.0049
0.0036
885
1,020
1,020
676
972
Average (Unit 3)
560
77,300
0.0042
4-7
4-8
4-9
554
545
536
91,200
90,100
84,600
0.012
0.0024
0.0094
1,190
1,050
1,450
672
1,024
1,135
Average (Unit 4)
545
88,600
0.0079
1,230
-------
Lead and mercury emissions at McKay Bay, presented in Table 2-8,
averaged 1,090 and 922 ug/dscm at 7 percent 02, respectively. Based on
typical uncontrolled metals concentrations (Section 1.2), lead was removed at
97 percent efficiency. Mercury was apparently not removed by the ESP since
the outlet concentration was within the range of typical uncontrolled mercury
emission concentrations (248 to 1,030 ug/dscm).
2.2.1,6 North Andover. 15 The North Andover, Massachusetts, facility
consists of two identical mass burn, waterwall combustors each designed to
combust 750 tons/day of MSW on Martin reciprocating grates. The air
pollution control system consists of two identical Wheelabrator ESP’s with
3 fields, each designed to reduce PM to 0.05 gr/dscf at 12 percent C0 2 , which
corresponds to a design collection efficiency of 98 percent. Specific
collecting area and other design data for the ESP’s are considered
confidential by the ESP manufacturer and are therefore not available. At the
ESP inlet, the flue gas flow is about 210,000 acfm at a temperature of 590 to
600°F.
In July 1986, compliance testing was conducted at the North Andover MWC.
Concurrent with the compliance testing, a test program to measure pollutant
levels at the ESP inlet was undertaken to quantify performance of the ESP.
At the ESP inlet and outlet, PM, metals, and CDD/CDF were sampled. Inlet PM
was sampled only during Runs 8 and 9. Inlet CDD/CDF samples for Runs I and 2
were not analyzed because of sampling problems. Because the metals analyses
were performed using an uncertified method, the data are suspect and are not
included in this report.
Particulate data from the testing at North Andover are presented in
Table 2-9. Outlet PM concentrations for six runs were between 0.0018 and
0.0054 gr/dscf at 12 percent CO 2 and averaged 0.0036 gr/dscf. Greater than
99 percent removal of PM occurred during the two runs in which it was
measured. Because the ESP collection area is unknown, the SCA of the ESP
cannot be calculated.
Test data for CDD/CDF emission measurements are presented in Table 2-10.
Simultaneous inlet and outlet measurements were conducted on three runs and
outlet measurements only on the additional runs. During each of the
2-22
-------
TABLE 2-9. PARTICULATE DATA FOR NORTH ANDOVER
Inlet PM
Outlet PM
ESP Inlet
Flue
Concentration
Concentration
PM Removal
Test Condition
Run
Number
Temp 8 rature
( F)
Gas Flow
(acfm)
(gr/dscf
at 12% C0 2 )
(gr/dscf
at 12% CO 2 )
Efficiency
(%)
Combustor = Normal
2
615
218,500
NMa
0.0050
-
ESP = Normal
3
4
5
8
9
580
584
591
600
609
195,300
201,000
207,800
221,500
219,300
NM
NM
NM
0.737
0.922
0.0013
0.0032
0.0023
0.0044
0.0054
-
-
-
99.5
99.4
Average
597
210,600
0.830
0.0036
99.5
, 3
I ’ , )
(A)
aNM Not measured.
-------
TABLE 2-10. CDD/CDF DATA FOR NORTH ANDOVER
I ’ )
Inlet CDD/CDF
Outlet CDD/CDF
CDD/CDF
ESP Inlet
Concentration
Concentration
Removal
Test Condition
Run
Number
Temp 8 rature
( F)
(ng/dscm
at 7%
(ng/dscm
at 7%
Efficiency
(‘ )
Combustor = Normal
1
607
NMa
396
-
ESP = Normal
2
3
4
5
615
580
584
591
NM
397
139
200
384
524
204
302
-
-31.8
-46.4
-51.0
Average
595
245
362
-43.1
aNM = Not measured.
-------
of the three runs for which both inlet and outlet CDD/CDF concentrations were
measured, CDD/CDF concentrations were higher at the ESP outlet than at the
inlet. The ESP inlet concentrations ranged from 139 to 397 ng/dscm and the
corresponding outlet concentrations ranged from 204 to 524 ng/dcsm. The
inlet concentration averaged 245 ng/dscm while the outlet concentrations
averaged 362 ng/dscm. The average CDD/CDF removal efficiency was -43
percent. The ESP inlet temperature during these tests averaged nearly 600°F,
which is well within the suggested range for CDD/CDF formation.
The data suggest that there may be an effect of inlet CDD/CDF
concentration on CDD/CDF removal efficiency. At the highest inlet
concentration, 397 ng/dscm, the smallest relative increase in CDD/CDF
concentration occurred. Lower inlet CDD/CDF concentrations yielded higher
relative increases in CDD/CDF concentrations across the ESP.
Plotted in Figure 2-5 is inlet CDD/CDF concentration versus outlet
COD/COF concentration. Although limited to only three data points, these
three points suggest that there exists a relatively strong relationship
between inlet CDD/CDF concentration and CDD/CDF formation at temperatures
approaching 600°F.
2.2.1.7 Peekskill. 16 ’ 17 The Westchester RESCO facility in Peekskill,
New York, consists of three identical waterwall combustors with Von Roll
reciprocating grates. Each unit has a design capacity of 750 tons of refuse
per day. Emissions from each combustor are controlled by a 3-field,
wire and plate ESP built by Wheelabrator Air Pollution Control in 1984. Each
ESP has a design SCA of 428 ft 2 /1,000 acfm (flue gas flow of 175,900 acfm)
with a superficial velocity of 3.55 ft/sec. The design flue gas temperature
at the ESP inlet was 425°F, but the system typically operates at about 455°F.
The ESP outlet temperature is typically about 435°F. Each ESP is designed to
achieve a PM removal efficiency of 99 percent.
Two separate test campaigns have been conducted at the Peekskill
facility. In April 1985, sampling was conducted as part of the New York
State Department of Environmental Conservation’s program to assess the health
effects of municipal waste combustion. At the ESP outlet, flue gas was
2-25
-------
540 - -——-_____________ ___—________
520- U
500-
480-
I s
460-
E
440-
420-
C
400-
0
I-
360-
340-
8 280-
260-
o 240-
220 —
200 - --— i I - 1 —
120 160 200 240 280 320 360 400
Inlet CDD/CDF Concentration (ng/dscm 7% 02)
Figure 2-5. Outlet CDD/CDF concentration as a function of inlet CDD/CDF
concentration at the North Andover MWC.
-------
sampled for CDD/CDF, PM, HC1, SO 2 , NON, metals (arsenic, beryllium, mercury,
cadmium, chromium, lead, manganese, nickel, vanadium, and zinc), and other
organics. The PM and some of the metals results from this test campaign have
been invalidated because of a problem with systematic contamination of the
collected samples. Specifically, extremely high PM levels were measured.
This suggests that the samples may have been contaminated by material that
was dislodged from the inside of the test ports. Except for arsenic,
beryllium, and mercury, which were collected separate from the PM sample, the
results for other metals are also considered suspect due to the contamination
problem and are thus not reported here.
In November 1985, extensive sampling was conducted under various
parametric combustor conditions and normal ESP conditions. Tests were
conducted under five different operating conditions: (1) normal load, end of
campaign (dirty heat transfer surfaces); (2) start-up (transient conditions);
(3) normal load, start of campaign (clean heat transfer surfaces); (4) high
load (115 percent of design); (5) low load (85 percent of design).
Triplicate sample runs were performed for each test condition except for
start-up, where only two sample runs were performed. The first start-up test
was performed following a 7-day maintenance outage after the end of campaign
tests. The second start-up test was performed several months later. The
pollutants measured include CDD/CDF, PM, SO 2 , HC1, and NOR, and other
organics. Flue gas was simultaneously sampled at the boiler inlet, ESP
inlet, and ESP outlet. The CDD/CDF samples were not collected simultaneously
with the PM or HC1 samples, but were taken under similar combustor operating
conditions. Measurements of SO 2 and NO were collected by CEM. Generally
these data were taken simultaneously with both the CDD/CDF and PM samples.
During the start-up runs, only CDD/CDF samples were collected.
Particulate data from the November test program are presented in
Table 2-11. Twelve simultaneous inlet and outlet sample runs were conducted
under varying combustor conditions but normal ESP conditions. Inlet PM
concentrations ranged from 0.28 gr/dscf (Run 6) to 2.7 gr/dscf (Run 8). The
outlet PM concentrations ranged from 0.0083 to 0.036 gr/dscf (Run 3) at
12 percent GO 2 , and averaged 0.017 gr/dscf. Excluding Run 6, removal
efficiencies for PM ranged from 97.1 to 99.5 percent and averaged 98.8
percent.
2-27
-------
TABLE 2•11. PARTICULATE DATA FOR PEEKSKILL
InLet PM
OutLet PM
ESP InLet
FLue
Concentration
Concentration
PM Removal
Test Condition
Run
Number
Temperature
C F)
Gas FLow
(acfm)
(gr/dscf
at 12% C0 2 )
(gr/dscf
at 12% CD 2 )
Efficiency
(%)
Combustor Norma Load,
end of campaIgn
1
2
461
472
181,200
200,300
1.68
1.53
0.0083
0.016
99.5
98.9
ESP $ Normal
3
479
198,000
1.68
0.036
97.9
Average
471
193,200
1.63
0.020
98.8
Combustor NormaL Aoad,
start of campaign
5
6
457
459
202,400
204,000
2.11
0.28
0.016
0.018
99.2
93.2
ESP NormaL
7
466
180,700
2.12
0.011
99.5
Average
454
195,700
2 • 12 b
0.015
994 b
Combustor High Load 5
8
462
210,700
2.70
0.015
99.4
ESP NormaL
9
10
464
462
201,300
207,000
1.89
1.72
0.011
0.019
99.4
98.9
Average
463
206,300
2.10
0.015
99.2
Combustor = Low Load 5
11
436
172,400
0.91
0.019
98.0
ESP NormaL
12
13
436
418
165,600
150,400
0.66
1.33
0.019
0.0087
97.1
99.3
Average
430
162,800
0.97
0.016
981 b
cx
November 1985 testing.
bAverage excLudes Run 6 due to low measured
inLet PM concentration.
-------
Analysis of the effect of SCA on PM removal efficiency showed no
significant trend. This is likely due to the relatively narrow variation
around the design SCA encountered during the test program.
Analysis of the effect of inlet PM on outlet PM emissions at Peekskill
showed that outlet PM concentrations were relatively insensitive to inlet PM
concentration. Except for one run, outlet PM concentrations ranged narrowly
from about 0.01 to 0.02 gr/dscf at 12 percent CO 2 over a wide range of inlet
PM concentration (0.3 to 2.7 gr/dscf). In no case did the average outlet PM
emissions from each test condition exceed 0.02 gr/dscf. As shown in
Figure 2-6, PM removal efficiency tended to increase with inlet PM
concentration at Peekskill.
Results for mercury and arsenic from the testing in April are presented
in Table 2-12. Emissions of mercury and arsenic averaged 1,790 and
2.17 ug/dscm at 7 percent 02, respectively. The mercury emissions compared
to typical uncontrolled mercury concentrations (see Section 1.2.) suggest
that no mercury removal occurred. In contrast, the arsenic emissions suggest
that a removal efficiency of 99 percent was achieved.
The CDD/CDF data from the April and November, 1985 tests are presented
in Table 2-13. Concentrations of CDD/CDF during normal and high load
combustion conditions (Runs 1-3 and 5-10) ranged from 333 to 861 ng/dscm at 7
percent °2 at the ESP inlet and 82 to 297 ng/dscm at 7 percent 02 at the ESP
outlet. Removal efficiencies ranged from 37 to 80 percent and averaged
63 percent. During the low load tests (Runs 11-13), CDD/CDF concentrations
were 132 to 281 ng/dscm at the ESP inlet and 96 to 250 ng/dscm at the ESP
outlet. Removal efficiencies during these tests ranged from 11 to 64 percent
and averaged 34 percent. During start-up tests (Runs 4 and 14), CDD/CDF
concentrations were near 10,000 ng/dscm at both the ESP inlet and outlet.
Removal efficiencies during the two start-up tests were 11 and 20 percent.
Of the ESP’s in the database for which both inlet and outlet CDD/CDF
measurements are available, only the ESP at Peekskill achieved a positive
CDD/CDF removal efficiency. Data for the other ESP’s in the database show
higher outlet CDD/CDF concentrations than inlet concentrations.
2-29
-------
100
+
+
99- 0
98-
• U
V.,
I-
0
0.
1 A
U End of Campaign
94 + Start of Campaign
High Load
L Low Load
I I I I I I I I I
0.2 0.6 1 1.4 1.8 2.2 2.6
Inlet PM Concentration (gr/dscf 12% C0 2 )
Figure 2-6. PM removal efficiency as a function of inlet PM concentration
at the PeeksklH MWC.
-------
TABLE 2-12. METALS EMISSIONS DATA FOR PEEKSKILL
ESP Inlet
Outlet PM
Concentration
Outlet Concentrationa
Test Condition
Run
Number
Temp 8 rature
( F)
(gr/dscf
at 12% CO 2 )
(u /dscm
at
7% 021
As
Hg
Combustor = Normal
b
1
NMC
--
2.33
1,320
ESP = Normal
2
3
NM
NM
--
--
2.05
2.14
1,510
2,540
Average
458 d
0017 e
2.17
1,790
aparticulate samples were analyzed for Cd, Cr, Pb, and Ni, but the results are suspect
and are not reported here.
bMetals results from April 1985 testing.
CNM = Not measured.
diemperature assumed same as for November 1985 testing.
eAverage PM results from November 1985 testing. Particulate samples from April were
invalidated because of systematic contamination.
r; .)
(A)
I- ..
-------
TABLE 2-13. CDD/CDF DATA FOR PEEKSK LL
8 ApriL 1985 testing.
bEsp inLet temperature not measured during ApriL 1985 test. Assumed to
November 1985 normaL toad tests (start and end of campaign).
CNb 1985 testing.
dinctudes mono- through octa- CDD/CDF.
CA ,
N,)
Test Condition
Run
Number
ESP InLet
TempSrature
( F)
InLet CDD/CDF
Concentration
(ng/dscm
at 7% 02)
OutLet CDD/CDF
Concentration
(ng/dscm
•t 7% °2
CDDICDF
RemovaL
Efficiency
Combustor z NormaL 6
ESP NormaL
1
2
3
458 b
458 b
458 b
•
•
.
85.8
98.2
138
-
Average
458 b
.
107
Combustor Nor aL Load,
end of campaign
ESP NormaL
1
2
3
472
471
471
584
411
861
222
143
173
62.0
65.1
79.9
Average
471
617
179
69.0
Combustor = Norma Load,
start of campaign
ESP = NormaL
5
6
7
444
445
447
473
525
436
297
224
267
37.3
57.4
38.8
Average
445
478
263
44.5
Combustor High LO.dC
ESP = NormaL
8
9
10
461
460
441
566
416
333
169
127
81.9
70.1
69.5
75.4
Average
454
438
126
71.7
Combustor = Low LoadC
ESP = NormaL
11
12
13
437
437
436
271
281
132
97.7
250
95.6
63.9
11.0
27.4
Average
437
228
148
34.1
Combustor = start.upC
ESP = NormaL
4
14
383
455
13,782
9082 d
11 ’ 080 d
8,060
19.6
11.2
Average
419
11,432
9,570
15.4
equaL average from
-------
Flue gas temperatures at the ESP inlet during all of these tests
(excluding Run 4) were between 436 and 472°F. These ESP inlet temperatures
are at the lower end of the range where CDD/CDF formation across an ESP has
been suggested to occur. However, because of the narrow range in ESP
operating temperature during these tests, it is not possible to evaluate a
correlation between temperature and ESP performance, even though there are 12
simultaneous measurements of CDD/CDF at the ESP inlet and outlet.
In Figure 2-7, CDD/CDF removal efficiency (excluding start-up tests) is
plotted as a function of inlet CDD/CDF concentration. As the plot indicates,
there does not appear to be a strong relationship between removal efficiency
and inlet CDD/CDF concentration.
2.2.1.8 Pinellas County.’ 8 The Pinellas County Resource Recovery
Facility in St. Petersburg, Florida, includes three identical mass burn
waterwall combustors. The combustors have Martin reciprocating grates and
each is designed to combust 1,000 tons/day of MSW. Emissions are controlled
by 3-field rigidwire and plate ESP’s built by Wheelabrator Air Pollution
Control. The SCA of the ESP’s is considered confidential by the ESP
manufacturer and is not available. At the ESP inlet, the flue gas flow is
typically 270,000 acfm at a temperature of 540°F. The controlled emissions
are discharged through individual stacks.
Compliance tests were conducted at Unit 3 of the facility in
February 1987. The combustor and ESP were under normal operating conditions
during the tests. Simultaneous ESP inlet and outlet measurements were made
for PM and CDD/CDF. Additional measurements were collected at the ESP outlet
for N0 , SO 2 , HC1, HF, mercury, beryllium, arsenic, hexavalent chromium, and
organics including benzene, chlorobenzene, chiorophenol, PAH’s, PCB’s, and
PCP’s.
Particulate data from the three sample runs conducted are presented in
Table 2-14. Outlet concentrations ranged from 0.0018 to 0.0026 gr/dscf at
12 percent CO 2 and averaged 0.0023 gr/dscf. The removal efficiency exceeded
99.7 percent for all three runs.
2-33
-------
100 -
90
80-
C
0 0
U
• 70-
>-. A
o I
C 60
0
0
Ui 50
(5
>
++
20 — U End of Campaign
+ Start of Campaign
O High Load
10 — A LowLoad
0- I I I
100 300 500 700 900
Inlet CDD/CDF Concentration (ng/dscm @ 7% 02)
Figure 2-7. CDD/CDF removal efficiency as a function of inlet CDD/CDF
concentration at the Peekskill MWC.
-------
TABLE 2-14.
PARTICULATE DATA FOR PINELLAS COUNTY
Test Condition
Run
Number
ESP Inlet
Tem 8 erature
( F)
Flue
Gas Flow
(acfm)
Inlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
PM Removal
Efficiency
(%)
Combustor = Normal
ESP = Normal
1
2
3
514
527
537
259,000
271,000
271,400
1.16
0.838
0.881
0.0024
0.0018
0.0026
99.8
99.8
99.7
Average
526
267,100
0.960
0.0023
99.8
-------
Metals data from three runs conducted at the ESP outlet are summarized
in Table 2-15. A comparison of the outlet data with typical uncontrolled
concentrations (see Section 1.2) for arsenic, cadmium, chromium, lead, and
nickel suggest that the ESP is achieving a relatively high level of removal
(greater than 98 percent) for these metals. For mercury, however, the outlet
concentration suggests that little or no removal is achieved.
CDD/CDF results from the six test runs conducted at the ESP inlet and
outlet at from Pinellas County are presented in Table 2-16. Inlet CDD/CDF
concentrations mesured during the six runs ranged from 31 to 103 ng/dscm and
averaged 54 ng/dscm. These concentrations are unusually low for any
combustor type and suggest that good combustion generates little CDD/CDF.
Outlet CDD/CDF concentrations ranged from 50 to 163 ng/dscm and averaged 100
ng/dscm. For each sample run, the outlet CDD/CDF concentration was higher
than that measured at the inlet. The ESP was operating at 524 to 552°F
during the test, which is within the temperature range suggested for CDD/CDF
formation.
Figure 2-8 is a plot of outlet CDD/CDF concentration as a function of
inlet CDD/CDF concentration. Increasing inlet concentration appears to
increase outlet CDD/CDF concentrations. A plot of relative CDD/COF formation
(negative removal efficiency) as a function of inlet CDD/CDF concentration is
shown in Figure 2-9. This plot shows no causal relationship between the two
variables. Because PM and CDD/COF were not measured simultaneously, the
effect of inlet PM on CDD/CDF formation or outlet emissions cannot be
evaluated.
2.2.1.9 Quebec City.’ 9 The Quebec City, Canada, MSW facility contains
four separate waterwall combustors. These were originally built in 1974 and
1975 with Von Roll reciprocating grates. Waterwall arches were added to the
combustor chambers in 1979. One unit was further modified in 1985/1986 to
include a “bull nose” in the combustion chamber and a modified secondary air
flow. Each unit is designed to combust 250 tons/day of MSW. Emissions are
controlled by 2-field ESP’s which then exit through a common stack for all
four lines. The SCA of the ESP is unavailable, but the ESP is designed to
achieve 98.5 percent PM removal. The design flue gas flow at the ESP inlet
is 100,000 acfm at a temperature of 540°F.
2-36
-------
TABLE 2-15.
METALS EMISSIONS DATA FOR PINELLAS COUNTY
r
( )
Test Condition
Run
Number
ESP Inlet
Temp 8 rature
( F)
Outlet PM
Concentration
(gr/dscf
at 12% GO 2 )
Outlet
As
Concentration
Cd Cr
(u Idscm
Pb
at 7%
Hg
02)
Ni
Combustor = Normal
1
539
0.0024
3.61
9.03 1.08
123
1,102
1.08
ESP = Normal
2
3
541
549
0.0018
0.0026
4.57
2.33
7.27 10.6
6.90 0.86
230
105
892
547
6.O
ND
Average
543
0.0023
3.50
7.73 4.18
153
847
2.38
aND = Not detected. Considered as zero in evaluating averages.
-------
TABLE 2-16.
CDD/CDF DATA FOR PINELLAS COUNTY
Inlet CDD/’COF
Outlet COD/COF
CDD/CDF
ESP Inlet
Concentration
Concentration
Removal
Test Condition
Run
Number
Tem 8 erature
( F)
(ng/dscm
at 7% 02)
(ng/dscm
at 7% 02)
Efficiency
(%)
Combustor = Normal
1
552
103
163
- 58
ESP = Normal
2
3
4
5
6
524
536
546
523
539
35
43
46
31
65
79
127
82
50
97
-127
-194
-80
- 62
-50
Average
537
54
100
- 95
0,
-------
170 -
160
150-
140-
E
130-
120-
C
0
110-
100-
U
0
o 90-
U.
C’) 0 80-
o 70-
0
4 -
0
50 -
40
30 50 70 90 110
Inlet CDD/CDF Concentration (ng/dscm Q )
Figure 2-8. Outlet CDD/CDF concentration as a function of inlet CDD/CDF
concentration at the Pinellas County MWC.
-------
200
I
190
180
C
I
120
0
U
C
100 -
0 90
40-
30 50 70 90 110
Inlet CDD/CDF Concentration (ng/dscm 7% q)
Figure 2-9. RelatIve Increase of CDD/CDF across ESP as a function
of Inlet CDD/CDF concentration at the Pinellas County MWC.
-------
From May through June 1986, testing was conducted on the modified
combustor. The test program was designed to collect process and emissions
data over a range of different combustor operating conditions in order to
relate combustor operating conditions with emissions of metals and organics,
ash quality, and boiler efficiency. The process parameters were varied to
yield five distinct operating conditions: (1) low feed rate (70 percent of
design), good combustion conditions; (2) design feed rate, good combustion
conditions; (3) high feed rate (115 percent of design) good combustion
conditions; (4) design feed rate, low combustor temperature; and (5) design
feed rate, poor air distribution. The ESP was kept at normal operating
conditions throughout the test program. Flue gas was sampled at the ESP
outlet and analyzed for CDD, CDF, PCB’s, PAH, chlorinated benzene,
chlorinated phenol, PM, SO 2 , HC1, THC, N0 , and metals (including arsenic,
cadmium, chromium, lead, mercury, and nickel). Additional testing at this
combustor using pilot-scale spray drying and dry sorbent injection systems
with a fabric filter was conducted in March 1985. Results of this test
program are presented in Sections 5.2 and 7.2, respectively.
Particulate data from Quebec City are presented in Table 2-17. Outlet
particulate concentrations for 14 sample runs ranged from 0.009 to
0.032 gr/dscf and averaged 0.018 gr/dscf at 12 percent CO 2 .
Metals data collected at Quebec City are presented in Table 2-18. Test
results are not available for the individual runs, but are summarized by test
condition. Uncontrolled metals concentrations were not measured during this
test. However, when compared with typical uncontrolled values presented in
Section 1.2, removal efficiencies for chromium and nickel could be as high as
99 percent; cadmium and arsenic slightly less, around 97 percent; and lead
about 95 percent. The outlet mercury concentration would suggest that little
or no mercury control was achieved.
CDD/CDF data from Quebec City are presented in Table 2-19. Outlet data
are available for 13 sample runs conducted under the combustor conditions
described above and normal ESP operation. No inlet data were collected.
Outlet CDD/CDF concentrations ranged from 46 to 690 ng/dscm and averaged
370 ng/dscm at 7 percent 02.
2-41
-------
TABLE 2-17. PARTICULATE DATA FOR QUEBEC CITY
Outlet PM
ESP Inlet
Flue Gas
Concentration
Run
Test Condition Number
Temp 8 ra ure
( F)
Flow
(acfm)
(gr/dscf
at 12% C0 2 )
Combustor = Low feed 2 410 57,400 0.011
rate, good combustion; 10 434 60,400 0.013
ESP = Normal 11 434 58,000 0.012
Average 426 58,600 0.012
Combustor = Design feed 5 423 56,800 0.007
rate, good combustion; 6 434 59,000 0.009
ESP = Normal 12 457 57,000 0.014
Average 438 57,600 0.010
Combustor — High feed 7 448 65,100 0.016
rate, good combustion; 9 471 71,600 0.016
ESP = Normal 13 484 75,500 0.024
Average 468 70,700 0.019
Combustor = Design 3 464 79,600 0.030
feed rate, low comb. 4 470 81,000 0.019
temp.; ESP = Normal
Average 467 80,300 0.025
Combustor = Design 14 470 73,400 0.032
feed rate, poor air 15 462 73,500 0.023
dist.; ESP = Normal
Average 466 73,500 0.028
aEstimated from measured 0 temperature at ESP outlet and an assumed temperature
drop across the ESP (25 F).
2-42
-------
TABLE 2-18. METALS EMISSIONS DATA FOR QUEBEC CITY
Test Condition
Run Number
Combustor Low Feed Rate, 2,10,118
Good Combustion; ESP = Normal
Combustor = Design Feed Rate, 5,6,128
Good Combustion; ESP = Normal
Corubustor = High Feed Rate, 7,98
Good Combustion; ESP = Normal
Combustor = Design Feed Rate, 3 , 4 a
Low Comb. Temp; ESP = NormaL
Combustor Design Feed Rate, 14,158
Poor Air Dist.; ESP Normal
Average
aResuLts are averages for the runs Listed.
ESP Inlet
Temp rature
C F)
406
417
439
448
446
OutLet PM
Concentration
(gr/dscf
at 12% C0 2 )
As
OutLet Concentration
Cd Cr
(j4gJdscm
Pb
at T X
Hg
O)
Ni
0.012
1.8
31.0
12.4
1,147
918
10.4
0.0099
2.7
23.4
7.2
655
685
5.1
0.016
4.8
43.1
15.7
1,699
927
8.5
0.024
6.8
93.3
21.4
2,123
843
8.3
0.028
6.7
79.7
15.2
2,627
655
7.0
431 0.018
4.6
54.1
14.4
1,650
806
7.9
-------
TABLE 2-19. CDD/CDF DATA FOR QUEBEC CITY
ESP
Inlet
Outlet
CDD/CDF
Temp 8 ra ure
Test Condition Run Number ( F)
Concent
(ng/dscm
ration
at 7%
°2
Combustor = Low feed 2 410 333
rate, good combustion; 10 434 129
ESP = Normal 11 434 112
Average 426 191
Combustor = Design feed 5 423 66
rate, good combustion; 6 434 46
ESP Normal 12 457 81
Average 438 64
Combustor = High feed 7 448 155
rate, good combustion; 9 471 193
ESP = Normal 13 484 300
Average 468 216
Combustor = Design 3 464 690
feed rate, low comb. 4 470 630
temp.; ESP = Normal
Average 467 660
Combustor = Design 14 470 563
feed rate, poor air 15 462 537
dist.; ESP = Normal
Average 464 550
aEstimated from temperat 8 re measured at ESP outlet and assumed temperature
drop across the ESP (20 F).
2-44
-------
At the design feed rate and good combustion conditions (Runs 5, 6, and
12), outlet CDD/CDF concentrations ranged from 46 to 81 ng/dscm and averaged
64 ng/dscm. Under the off-specification combustor conditions (low and high
feed rate, low combustion temperature, and poor air distribution), the
measured outlet CDD/CDF concentrations were substantially higher, ranging
from 112 to 690 ng/dscm. Although it cannot be ascertained because no inlet
CDD/CDF data were collected, these higher outlet CDD/CDF results are likely a
result of combustor (causing higher ESP inlet values) and ESP performance
both.
2.2.1.10 Tulsa. 2 ° The Walter B. Hall Resource Recovery Facility in
Tulsa, Oklahoma, consists of two identical 375-ton/day Martin GmbH mass burn,
waterwall combustors. Emissions are controlled by 3-field ESP’s with SCA’s
of 325 ft 2 /1,000 acfm and design PM removal efficiencies of 98.5 percent. At
the ESP inlet, the design flue gas flow is 89,500 acfm at a temperature of
515°F. The ESP normally operates between 375 and 505°F. The ESP exhaust
streams exit through a common stack.
In June 1986, compliance testing was performed with the combustor and
ESP under normal operating conditions. At each ESP outlet, three separate
runs were conducted to measure PM, NON, SO 2 , and HC1. At the common stack,
three measurements were taken of CDD/CDF, volatile organics, and metals
(lead, beryllium, and mercury).
Particulate data from the outlet of both ESP’s are presented in Table
2-20. Outlet PM concentrations from Unit 1 ranged from 0.0069 to
0.012 gr/dscf at 12 percent CO 2 and averaged 0.0094 gr/dscf. The outlet PM
concentrations from Unit 2 ranged from 0.0036 to 0.0056 gr/dscf and averaged
0.0049 gr/dscf.
Table 2-21 presents lead and mercury emissions data from the common
stack. Lead and mercury averaged 412 and 418 ug/dscm, respectively. Again,
when compared to typical uncontrolled levels measured at other facilities
(see Section 1.2), it appears that the ESP’s achieved a relatively high level
of lead control (99 percent removal efficiency), but little or no mercury
control.
The results of three CDD/CDF sample runs conducted at the common stack
are presented in Table 2-22. CDD/CDF concentrations ranged from 34 to
2-45
-------
TABLE 2-20. PARTICULATE DATA FOR TULSA
Test Condition
Run
a
Number
ESP Inlet
Temp rature
( F)
Flue Gas
Flowb
(acfm)
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Combustor = Normal
ESP = Normal
1-1
1-2
1-3
NM
NM
NM
69,500
77,800
81,000
0.012
0.0093
0.0069
Average (Unit 1)
375 C
76,000
0.0094
2-1
2-2
2-3
NM
NM
NM
77,600
80,800
81,400
0.0054
0.0056
0.0036
Average (Unit 2)
375 c
79,900
0.0049
alhe unit number is given first followed by the run number.
bFlue gas flow rate (acfm) calculated based on reported dscfm, average
moisture value of 17.6 percent, and estimated temperature at the ESP inlet
of 375 F. Flue gas flow in acfm is equal to the flow in dscfm times 1.92..
cTemperature not measured, but estimated based on a measured value at the
stack dunn 8 the same test program and an assumed temperature drop across
the ESP (20 F).
2-46
-------
TABLE 2-21. METALS EMISSIONS DATA FOR TULSA
Test Condition
Run
Number
ESP Inlet
Temp 8 ra ure
( F)
Outlet PM
Concentration
(gr/dscf
at 12% CO 2 )
Outlet Concentration
(u Idscm
Pb
at
7%
0 j.
H
Corubustor = Normal
1
362
- -
420
412
ESP = Normal
2
3
383
379
- -
--
490
326
454
389
Average
375
0 0023 b
412
418
estimated from a measur 8 d value at the ESP outlet and an assumed
drop across the ESP (20 F).
samples not collected simultaneously with metal samples.
results given.
N)
-.4
aTemperature
temperature
bparticui ate
particul ate
Average
-------
TABLE 2-22. CDD/CDF DATA FOR TULSA
Test Condition
Run
Number
ESP
Temp
(
Inlet
ra ure
F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Combustor = Normal
1
375
35.8
ESP = Normal
2
3
375
375
33.7
38.5
Average
375
36.0
aTemperature estimated based on a measured value at the stack d 8 ring the
metals runs and an assumed temperature drop across the ESP (20 F).
2-48
-------
39 ng/dscm at 7 percent 02 and averaged 36 ng/dscm. No inlet CDD/CDF data
were collected. The estimated ESP inlet temperature of 375°F is
substantially below the temperature range suspected for CDD/CDF formation.
This low temperature and correspondingly low CDD/CDF emission level suggest
that a well designed and operated ESP coupled with a well designed and
operated combustor can achieve relatively low CDD/CDF emission levels.
2.2.2 RDF MWC’s
2.2.2.1 Lawrence. 21 ’ 22 The Lawrence, Massachusetts, Thermal Conversion
Facility consists of a single unit designed to combust 1,000 tons/day of ROE.
Emissions are controlled using a Belco wire and plate hot-side ESP with 3
fields. The SCA of the ESP is unavailable. At the ESP inlet, the flue gas
flow is typically 190,000 acfm at about 540°F. Following the ESP is an air
preheater, an ID fan, and the stack.
Test data are available from September 1986 and September 1987. In
September 1986, flue gas at the ESP outlet (following the air preheater) was
sampled to determine the CDD/CDF emissions from the facility. In
September 1987, testing was performed to demonstrate compliance with State
permit requirements. At the ESP outlet (downstream of the air preheater),
CDD/CDF, N0 , and PM were measured. All tests were conducted under normal
combustor and ESP operating conditions.
Particulate emissions data from the September 1987 tests are presented
in Table 2-23. The measured particulate emissions from the three test runs
ranged from 0.0054 to 0.014 gr/dscf at 12 percent CO 2 and averaged
0.010 gr/dscf. No inlet PM or SCA data are available.
In Table 2-24, CDD/CDF data are presented from both the 1986 and 1987
tests. The 1986 tests (3,300 ng/dscm) show approximately 30 times greater
CDD/CDF concentration as the 1987 tests (111 ng/dscm). Although modifica-
tions were made to the combustor and ESP between the two test periods, the
details of the modifications are unavailable. Neither inlet PM nor inlet
CDO/CDF were measured.
2.2.2.2 Niagara Falls. 23 ’ 24 ’ 25 The RDF facility located in Niagara
Falls, New York, is operated by the Occidental Chemical Corporation. Each of
the two combustors are designed to combust 1,200 tons/day of RDF. Emissions
are controlled by Belco 4-field, wire-plate ESP’s, each with a design SCA of
2-49
-------
TABLE 2-23. PARTICULATE DATA FOR LAWRENCE
Test Condition
Run
Number
ESP Inlet
Temp 8 rature
( F)
Flue Gas
Flow
(acfm)
Outlet PM
Concentration
(gr/dscf
at 12% CU 2 )
Combustor Normal
ESP = Normal
1
2
3
NMa
NM
NM
166,000
166,600
160,300
0.014
0.011
0.0054
Average
NM
164,300
0.010
aNM Not measured. Assuming the same temperature drop across the ESP and
air preheater as measured in 1986, the temperature at the ESP inlet would be
507°F. However, because of the modifications made to the system this cannot
be verified.
2-50
-------
TABLE 2-24. CDD/CDF DATA FOR LAWRENCE
Test Condition
Run
Number
ESP
Temp
(
Inlet
rature
F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7%
Combustor = Normal
1
538
2,320
ESP = Normal
2
545
2,780
(1986)
3
544
4,810
Average (1986)
542
3,300
Combustor = Normal
1
NMa
115
ESP = Normal
2
NM
159
(1987)
3
4
NM
NM
78
90
Average (1987)
NM
111
aNM = Not measured. Assuming the same temperature drop across the ESP and
air preheater as measured in 1986, the temperature at the ESP inlet would be
508°F. However, because of the modifications made to the system, this
cannot be verified.
2-51
-------
520 ft 2 /1,000 acfm. At the ESP inlet, the design flue gas flow is
280,000 acfm at 550°F. The facility was built in 1981; modifications to the
ESP’s were made in 1986. These modifications included reduction of air
inleakage, improving gas distribution, and improving the rapping system.
In May and June 1985, testing was conducted as part of the New York
Department of Environmental Conservation’s assessment of risk from municipal
waste combustion. The sampling was conducted at Unit 1 of the facility under
normal operating conditions. Flue gas at the ESP outlet was sampled and
analyzed for CDD/CDF, PM, HC1, SO 2 , NOR, metals (lead, manganese, mercury,
zinc, beryllium, chromium, cadmium, nickel, and vanadium), and other
organics. However, because these results reflect operation of the ESP prior
to modifications, the PM and CDD/CDF data are presented only for comparison
to later results. Because metals data are available from later testing, the
metals results from these tests are not included.
Throughout 1986, PM testing was performed in conjunction with
modification of the ESP’s in order to assess improvements in performance.
The combustor steam load was varied during these tests to evaluate the range
of inlet conditions typically encountered. In February, prior to modifica-
tion of the ESP’s, PM was measured at the ESP inlet. Following completion of
modifications, PM testing at the ESP outlet was conducted in May and August.
In April 1987, testing at the Unit 1 ESP outlet for CDD/CDF only was
conducted to assess CDD/CDF emissions. The combustor and ESP were operating
normally during this testing.
Particulate data from the 1985 and 1986 tests are presented in
Table 2-25. The modifications to the ESP produced significant reductions in
PM emissions. Outlet measurements conducted on Unit 1 in 1985 prior to
modifications averaged nearly 0.1 gr/dscf at 12 percent CO 2 . In 1986, outlet
emissions ranged from 0.010 to 0.025 gr/dscf and averaged 0.016 gr/dscf for
Unit 1 and ranged from 0.023 to 0.028 gr/dscf and averaged 0.025 gr/dscf for
Unit 2.
Table 2-26 presents COD/COF emissions data. CDD/CDF emission
measurements were conducted at the ESP outlet of Unit 1 both before (1985)
and after (1987) modifications made to the ESP. Emissions measured during
test periods were relatively high (>2000 ng/dscm) and are similar suggesting
that the ESP modifications had little effect on CDD/CDF emissions.
2-52
-------
TABLE 2-25. PARTICULATE DATA FOR NIAGARA FALLS
ainlet and outlet PM measurements not taken simultaneously. Cannot calcuLate removal efficiency per run.
bAll inlet data collected in February 1986.
CNM = Not measured.
dAugust 1986 tests. Other data taken after ESP modifications were coLlected in May 1986.
eAverage of four tests at target steam flow of 270,000 Lb/hr (90 percent of design).
Average of two tests at target steam flow of 250,000 Lb/hr steam (83 percent of design).
of two tests at target steam fLow of 230,000 lb/hr stream (77 percent of design).
hRenovat efficiency calculated from average inlet and outlet vaLues for post-modifications tests.
r
01
( ,J
Test Condition
Run
Number
ESP Inlet
Tempgrature
( F)
Flue
Gas Flow
(acfm)
at
InLet PM
Concentration
(gr/dscf b
12% CO 2 )a ,
at
Outlet PM
Concentration
(gr/dscf
12% CO 2 ) 8
PM Removal
Effici ncy
(%)
Before Modifications (1985)
1-1
NMC
NM
NM
0.118
-
Combustor = Normal
1-2
NM
NM
NM
0.073
-
ESP = Normal
1-3
NM
NM
NM
0.096
-
Average before modifications
-
-
-
0.096
-
After Modifications (1986)
Combustor = Normal
ESP Normal
1-1
1-2
1-3
1-4
1-5
1-6
1-7
1-8
1-9
1-10
1-11
1-12
627
627
628
642
600 e
600 e
6 00 e
600 e
6O3
603’
62o
62 o
273,000
285,000
285,000
291,000
302000 e
302000 e
302000 e
302000 e
282,000
282,000’
286 , 000 g
286 ,000
3.87
3.42
3.18
2.51
2.48
4.03
3.91
4.24
3.91
NM
NM
NM
O.O18
°° 15 d
°•° 14 d
0.018
0.025
0.020
0.019
0.019
0.011
0.015
0.011
0.010
-
-
-
-
-
-
-
-
-
-
Average (Unit 1)
614
290,000
3.51
0.016
99 5 h
2-1
2-2
2-3
2-4
602
618
606
606
281,000
287,000
280,000
275.000
2.90
2.65
NM
NM
0023 d
0024 d
° 023 d
0.028
-
Average (Unit 2)
610
281,000
2.78
0.025
99 • 1 h
-------
TABLE 2-26. CDD/CDF DATA FOR NIAGARA FALLS
Test Condition
Run
Number
ESP
Temp
(
Inlet
8 rature
F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7%
Combustor = Normal
1
NRa
3,140
ESP = Normal, prior
2
NR
2,380
to modification (1985)
3
NR
2,200
Average, prior to
NR
2,560
modification
Combustor = Normal
3
NR
3,770
ESP = Normal, after
4
NR
5,390
modification (1987)
5
NR
3,700
Average, after
613 b
4,290
modification
aNR = not reported.
biemperature assumed same as measured for PM tests following ESP
modifications.
2-54
-------
2.2.2.3 Red Wing (Northern States Power). 26 ’ 27 The Northern States
Power (NSP) Company Red Wing, Minnesota facility consists of two Foster
Wheeler RDF spreader stoker combustors with Detroit Stoker grates capable of
firing 360 tons/day of coarse RDF each. In 1987, the two combustors were
converted by Babcock & Wilcox from firing coal to firing RDF. Emissions from
each combustor are controlled by a cyclone and Belco 4-field wire and plate
ESP that were installed in 1981. Each ESP is equipped with 47,250 ft 2 of
plate surface area and is operated at an SCA of about 570 ft 2 /1,000 acfm,
based on a measured flue gas flow rate of 83,000 acfm at the ESP inlet. The
temperature at the ESP inlet is designed to be between 260 and 450°F.
Compliance tests were conducted at the facility on two separate
occasions. The first compliance test was conducted in March 1988 on Unit 2.
Measurements were made at the ESP inlet and outlet under high load combustor
conditions (106,000 to 114,600 lb/hr steam) for PM, metals, CDD/CDF, and
other organics. Acid gases and NOx were also measured at the ESP outlet.
However, the CDD/CDF samples obtained at the ESP inlet had analytical
problems and have been invalidated.
The second compliance test was conducted in May 1988 on Unit 1 under
high load combustor conditions (105,700 to 113,700 lb/hr steam) and on Unit 2
under low load combustor conditions (60,400 to 62,400 lb/hr steam). At
Unit 1, PM measurements were made at the ESP outlet. At Unit 2, metals and
CDD/CDF measurements were made at the ESP outlet. No inlet samples were
collected on either unit.
Particulate data from both test periods are presented in Table 2-27.
Outlet PM emissions from three runs ar Unit 2 under high load conditions
ranged from 0.015 to 0.046 gr/dscf at 12 percent CO 2 and averaged 0.024
gr/dscf. The corresponding removal efficiencies ranged from 98.3 to 99.3
percent and averaged 98.9 percent. Outlet PM emissions from four runs at
Unit 1, under similar high load conditions, ranged from 0.031 to
0.055 gr/dscf at 12 percent CO 2 and averaged 0.041 gr/dscf. There is
insufficient information available to determine why the emissions from Unit 1
were generally higher than from Unit 2. The highest outlet PM concentrations
from both units resulted when sootblows occurred. This generated a 40 to 80
2-55
-------
TABLE 2-27. PARTICULATE DATA FOR NSP RED WING
0 ,
o.
Test Condition
Run
Numbera
ESP Inlet
Temperature
( F)
Flue
Gas Flog
(acfm)
Inlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Outlet PM
Concentration
(gr/dscf
at 12% CD 2 )
PM
Removal
Efficiency
(%)
Combustor = High
ESP = Normal
Load
2 -1
2-2
416
417
83,900
82,700
1.50
271 C
0.011
0 • 046 C
99.3
98.3
Average
2-3
407
413
82,300
83,000
1.95
2.05
0.015
0.024
99.2
98.9
Combustor = High
ESP = Normal
Load
1 1
1-2
1-3
1-4
454 d
457
464
458u
90,600
90,900
89,600
100,800
NMe
NM
NM
NM
018 f,g
0055 C,9
0036 g
00
--
--
Average
458
92,900
NM
0.041
--
ag Number contains the unit number followed by the run number for each unit.
bFlue gas flow rate measurement believed to be high due to turbulence at sampling location.
Cicludes sootbiow.
diemperature estima ed from measured value at the stack and a previously measured temperature drop
across the ESP (50 F).
eNM not measured.
Considered erroneous by testing company. Not included in average.
condensible fraction that cannot be separately quantified based on test report data.
This fraction was previously measured at Unit 2 to be 5.4 to 16.8 percent of the total particulate
sample.
-------
percent higher inlet PM loading, as shown by Unit 2, which in turn increased
the outlet PM concentration. The PM removal efficiency also decreased during
the sootblow.
Metals data from three runs each at low and high load conditions on Unit
2 are presented in Table 2-28. During high load operation, cadmium and lead
were consistently removed at greater than 99 percent efficiency, similar to
the PM removal efficiency. Removal efficiencies for arsenic, chromium,
mercury, and nickel were lower. Arsenic removal efficiencies ranged from
78.0 to 99.3 percent and averaged 90.9 percent. Mercury removal efficiency
ranged from 74.8 to 86.0 percent and averaged 81.6 percent. This is highly
unusual since mercury removal has not been observed across any other ESP. In
the test report, it is suggested that mercury is removed because it is in the
form of salts. However, other ESP’s operating at similar conditions have not
demonstrated mercury reductions. Chromium removal efficiencies ranged from
67.3 to 97.5 percent and averaged 80.7 percent. Nickel removal efficiencies
were relatively low, ranging from 41.9 to 82.3 percent and averaging 64.5
percent.
Outlet metals emissions were lower at low load than at high load for all
metals except mercury. There are no data on ESP operation to suggest a
reason for the generally better performance at low load. The lower metals
emissions may be due to lower PM emissions because the combustor generates
less PM at low load conditions or due to a lower flue gas flow resulting in a
higher SCA, which provides better PM removal.
CDD/CDF data are presented in Table 2-29 for six outlet sample runs on
Unit 2. Three runs were conducted under high load conditions and three runs
were conducted under low load conditions. At high load, outlet CDD/CDF
concentrations ranged from 26.5 to 34.1 ng/dscm at 7 percent 02 and averaged
29.3 ng/dscm. Concentrations obtained during low load were similar, ranging
from 22.0 to 48.7 ng/dscm at 7 percent 02 and averaging 32.7 ng/dscm.
Although the £SP inlet temperature during low load was about 45°F less than
at high load, the CDD/CDF emissions were not significantly different.
2-57
-------
TAILE 2-28. METALS DATA FOP NSP RED WING
OutLet PM
Teat Condition
ESP Inlet Conc.ntration Inlet Conc.ntration
Outlet
Concentration
Run Ts . ature (sr/dscf at Cualdac. at 7% 0
Nu m b e r C F) 12% C0 2 ) As Cd Cr Pb
)_
p
Ni
As
(uaidscp
Cd Cr
at 7%
Pb
Hg
Ni
As
RemovaL Efficiency C X)
Cd Cr Pb Hg
Ni
Combustor • Nigh Load
ESP • Normal
2-i NMb NM 109 777 459 Z8 9I0
2-2 MM NM 326 935 286 17 3S0
2-3 NM NM 173 702 399 26,580
131
114
176
408
296
327
24
15
1.2
5.6
2.7
0.1
150
65
10
77
39
27
33
16
28
237
91
58
78.0
95.4
99.3
99.3
99.7
99.9
67.3
77.3
97.5
99.7
99.8
99.9
74.6
86.0
64.1
41.9
69.3
82.3
Average
Combuetor Low Load
ESP NoruaL
413 c 24 d 203 805 381 24,280
2-1 NM NM NM NM NM NM
2-2 NM MN NM NM MM NM
2-3 NM NM NM NM NM NM
140
NM
NM
NM
344
NM
NM
NM
13
5.2
1.4
3.4
2.8
1.0
1.1
0.9
75
24
21
16
48
5.3
3.8
13
26
61
15
111
129
40
26
37
90.9
-
-
99.1
-
60.7
-
-
99.8
-
-
61.6
-
-
64.5
-
-
-
Average
390 C NM NM NM NM NM
NM
MM
3.3
10
20
7.4
62
34
-
-
-
-
-
5 RUn Number contains
- not me.sur.d.
the unit number followed by the run number for each unit.
CAverage temperature
dpM not measured simu
for other tests at the same condition.
Itaneousty. Average for Unit 2 reported.
-------
TABLE 2-29. CDD/CDF DATA FOR NSP RED WING
Test Condition
Run
Number
a
ESP
Temp
Inlet
e ature
( F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7%
02)
Combustor = High
Load
2-1
421
26.5
ESP Normal
2-2
2-3
424
431
27.4
34.1
Average
425
29.3
Combustor = Low
ESP = Normal
Load
2-1
2-2
2-3
378
388 b
389
22.0
27.5
48.7
Average
380
32.7
aRUfl number contains the unit number followed by the run number for that
bunit.
Temperature estimated from measured value at 0 the stack and a previously
measured temperature drop across the ESP (50 F).
2-59
-------
2.2.3 Modular MWC’s
2.2.3.1 Barron County. 28 The Barron County Resource Recovery Facility
in Almena, Wisconsin, consists of two Consumat Model Number CS-1600 starved-
air combustors. Each combustor has a rated capacity of 50 tons/day.
Emissions are controlled by wire-plate 2-field ESP’s, manufactured by
Precipitair Pollution Control. Each ESP has an SCA of 230 ft 2 /1,000 acfm.
The design flue gas flow at the ESP inlet is 12,500 acfm at 525°F.
In November 1986, compliance tests were conducted at the facility under
normal combustor and ESP operating conditions. Flue gas was sampled during
three runs at the ESP outlet for PM, HC1, NOR, and trace metals (lead,
arsenic, cadmium, chromium, and nickel).
Particulate data from the testing at Barron County are presented in
Table 2-30. The outlet PM concentration ranged from 0.010 to 0.011 and
averaged 0.010 gr/dscf at 12 percent CO 2 . No inlet PM data were collected.
The metals emissions data from Barron County are presented in
Table 2-31. Lead, cadmium, arsenic and chromium concentrations at the ESP
outlet averaged 270, 22, 21, and 2.9 ug/dscm at 7 percent 02, respectively.
Nickel was not detected. Compared to typical uncontrolled metals
concentrations from starved-air modular MWC’s (Section 1.2), removal
efficiencies for cadmium, chromium, nickel, and lead probably exceeded
98 percent. Arsenic removal efficiency appears to be somewhat lower, perhaps
60 percent.
2.2.3.2 Oneida County. 29 ’ 3 ° The Oneida County Energy Recovery Facility
in Rome, New York, combusts MSW in four Clear Air Model Number CA 4000A
starved-air modular combustors. Each is designed to combust 50 tons/day MSW.
United-McGill ESP’s, each with 2 fields and a design SCA of
232 ft 2 /1,000 acfm, control emissions from each combustor. At the ESP inlet,
the design flue gas flow is 18,000 acfm at 400°F. The design PM removal
efficiency is 90 percent. Controlled emissions are released through
independent stacks for each unit.
Testing was conducted at the facility in August 1985 as part of the New
York State Department of Environmental Conservation’s MWC test program.
Emission tests were conducted at the outlet of Unit 1 for PM, CDD/CDF, other
organics (chrysene, PCB’s, benzo(a)pyrene, formaldehyde), HC1, HF, various
2-60
-------
TABLE 2-30. PARTICULATE DATA FOR BARRON COUNTY
Outlet PM
ESP Inlet
Flue Gas
Concentration
Test Condition
Run
Number
Temp 8 ra ure
( F)
Flow
(acfm)
(gr/dscf
at 12% CU 2 )
Combustor = Normal
1
437
23,900
0.011
ESP = Normal
2
3
444
448
24,600
25,200
0.010
0.010
Average
443
24,600
0.010
aTemperature estimated from measured 0 value at ESP outlet and an assumed
temperature drop across the ESP (40 F).
2-61
-------
TABLE 2-31. METALS EMISSIONS DATA FOR BARROM COUNTY
Test Condition
Run
Number
ESP intet
Temp 8 ra ure
( F)
Outtet PM
Concentration
(gr/dscf
at 12% CD 2 )
Outtet Concentration
As Cd Cr
(ua/dscm at 7% 0 2 L...
Pb Ni
Combustor : NormaL
1
437
0.011
19
19
NDb
235
NO
ESP NormaL
2
3
444
448
0.010
0.010
22
22
24
24
4.3
4.4
287
287
ND
ND
Average
443
0.010
21
22
2.9
270
ND
aTempera ure estimated from a measured value at the ESP outlet and en assumed temperature drop across the
ESP (40 F).
bND not detected. Considered zero for evaluating averages.
-------
metals (arsenic, beryllium, mercury, cadmium, chromium, lead, manganese,
nickel, vanadium, and zinc), $02, and N0 . Both the combustor and ESP were
operated normally during testing. However, temperature data were not
reported for any location.
Particulate data from the three sample runs conducted are presented in
Table 2-32 The outlet PM concentration ranged from 0.013 to 0.033 gr/dscf
and averaged 0.026 gr/dscf at 12 percent CO 2 . No inlet PM data are
available.
The metals data from Oneida County are presented in Table 2-33. Outlet
metals concentrations averaged 5.0, 92, 150, 430, 2,100, and 130 ug/dscm for
arsenic, cadmium, chromium, lead, mercury, and nickel, respectively. Based
on typical uncontrolled metals concentration (Section 1.2), removal of
arsenic, cadmium, and nickel was moderate, about 90 percent. Lead removal
was somewhat higher (about 98 percent), while chromium removal was soernwhat
lower (about 70 percent). The outlet mercury levels suggested that no
removal of mercury was achieved.
Table 2-34 presents CDD/CDF data for the two outlet sample runs
conducted. The CDD/CDF concentrations measured were 327 and 597 ng/dscm at 7
percent 02, for an average of 462 ng/dscm.
2.2.3.3 Oswego County. 3 ’ The Oswego County Energy Recovery Facility in
Fulton, New York, includes four identical 50 tons/day starved-air combustors
manufactured by Consumat. Each combustor is equipped with a 1-field ESP
manufactured by PPC Incorporated. Each ESP is designed to treat 15,250 acfm
of flue gas at 450°F The design SCA is 294 ft 2 /1,000 acfm and the design PM
removal efficiency is 80 percent. The actual SCA is between 220 and
250 ft 2 /1,000 acfm. Cleaned flue gas from each incinerator is exhausted
through a separate stack.
In August 1986, a test program was conducted at the facility to
determine operating characteristics which minimize formation of CDD/CDF and
assess performance of the air pollution control device. Specifically,
relationships between combustion gas variables and CDD/CDF were analyzed.
Twelve test runs were conducted during the test program, with triplicate runs
at four test conditions: (1) start of campaign (normal operation, cleaned
heat transfer surfaces); (2) mid-range secondary chamber temperature
2-63
-------
TABLE 2-32. PARTICULATE DATA FOR ONEIDA COUNTY
Test Condition
Run
Number
ESP Inlet
Ternp rature
( F)
Flue Gas
Flow
(acfm)
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Combustor = Normal
ESP = Normal
1
2
3
NMa
NM
NM
NAb
NA
NA
0.013
0.033
0.033
Average
NM
NA
0.026
aNN = Not measured.
bNA = Not available. Flue gas flow not present in test report.
2-64
-------
TABLE 2-33. METALS EMISSIONS DATA FOR ONEIDA COUNTY
r ’)
U,
Test Condition
Run
Number
ESP InLet
Tempgrature
( F)
OutLet PM
Concentration
(gr/dscf
at 12% C0 2 )
As
OutLet Concentration
Cd Cr
(ug/dscm
Pb
at 7%
Hg
°2
Ni
Combustor = NormaL
1
NMa
0.013
5.09
89.2 16
636
3,200
44
ESP = NormaL
2
3
NM
NM
0.033
0.033
6.09
3.91
78.6 274
107 159
308
352
1,690
1,290
222
110
Average
NM
0 .026
5.03
91.6 150
432
2,060
125
aWN = Not measured.
-------
TABLE 2-34. CDD/CDF DATA FOR ONEIDA COUNTY
Test Condition
Run
Number
ESP
Temp
(
Inlet
8 rature
F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Combustor = Normal
1
NMa
597
ESP = Normal
2
NM
327
Average
NM
462
aNM = Not measured.
2-66
-------
(1,750°F); (3) end of campaign (normal operation, dirty heat transfer
surfaces); and (4) low secondary chamber temperature (1,650°F). Flue gas was
sampled simultaneously at the secondary chamber exit, ESP inlet, and ESP
outlet for CDD/CDF, other organics, and HC1. PM was also measured
simultaneously at the ESP inlet and outlet. Combustion gases (including CO,
C0 2 , 0 , N0 , and SO 2 ) were measured at the secondary chamber exit and ESP
outlet.
Results of the particulate tests conducted at Oswego County are
presented in Table 2-35. The outlet PM concentration ranged from 0.011 to
0.042 and averaged 0.020 gr/dscf at 7 percent 02 (approximately equal to
0.019 gr/dscf at 12 percent C0 2 ). The PM removal efficiency averaged
92 percent, which is significantly higher than the design value. Variations
in flue gas flow rate, and thus, in SCA were too narrow to examine any effect
of SCA on PM removal efficiency. Figure 2-10 presents a plot of PM removal
efficiency versus inlet PM concentration. PM removal efficiency increased
with inlet PM concentration for each test condition except during the tests
with low secondary combustion chamber outlet temperature. Figure 2-11
presents a plot of outlet PM concentration as a function of inlet PM
concentration. The data from the individual test conditions do not show any
causal relationship.
Table 2-36 presents the results of the COD/COF test runs conducted at
Oswego County. An increase in CDD/CDF concentration across the ESP was
observed during all runs except one. The ESP inlet temperature, which ranged
from 460 to 500°F, was within the temperature region suggested for CDD/CDF
formation (Section 2.1). In Figure 2-12, CDD/CDF removal efficiency is shown
as a function of ESP inlet temperature. Formation of CDO/CDF was lower
(removal efficiency less negative) at temperatures below 475°F. Figure 2-13
presents a plot of CDD/CDF remo:al efficiency as a function of inlet CDD/CDF
concentration. As shown, CDD/CDF formation was also lower for runs with
higher inlet CDD/COF concentrations.
2.2.3.4 Pigeon Point. 32 The Energy Generating Facility in Pigeon Point
(Wilmington), Delaware, consists of five Vicon excess-air modular combustors.
Each combustor has a rated capacity of 120 tons per day of RDF, but
2-67
-------
TABLE 2-35. PARTICULATE DATA FOR OSUEGO COUNTY
Test Condition
Run
Number
ESP InLet
Temp 8 reture
C F)
Flue
Gas Flow
(acfm)
InLet PM a
Concentration
(gr/dscf
at 7% 02)
OutLet PM a
Concentration
(gr/dscf
at 7% 02)
PM RemovaL
Efficiency
Combustor Normal,
start of campaign
ESP NormaL
1
2
3
496
493
502
19,700
20,400
20,100
0.495
0.331
0.203
0.026
0.020
0.028
94.8
94.0
86.2
Average
496
20100
0.343
0.025
91.7
Combuator = Mid-ra 9 ge
temperature (1,750 F)
ESP = Normal
4
5
6
486
492
487
19,300
19,600
18,300
0.182
0.224
0.155
0.013
0.012
0.023
92.9
94.6
85.2
Average
488
19,100
0.187
0.016
90.9
Combustor Normal,
end of campaIgn
ESP = Normal
7
8
9
487
497
506
19,700
19,900
19,400
0.301
0.151
0.183
0.013
0.011
0.011
95.7
92.7
94.0
Average
497
19,700
0.212
0.012
94.1
Combus or Low temperature
(1,650 F)
ESP = NormaL
10
11
12
460
472
477
17,700
19,100
19,000
0.264
0.398
0.372
0.019
0.042
0.024
92.8
89.5
93.6
Average
670
18,600
0.344
0.028
92.0
aPM data in gr/dscf at 7 percent 0 because CO . VaLues are approximately 6 percent Larger than if the
concentration were normalized to 2 percent C 2 , based on the theoretical relationship between 02 and CO 2 .
-------
96 -
0
95 -
+
94
+ A
> 92-
C)
C
a)
::i
N.) 89-
‘.0
88 -- U Start of Campaign
+ Mid-rang. S.condary Temperatur.
87 -
O End of Campaign
- A Low S.condary Temperature
85 + ______________________________________
0.15 0.25 0.35 0.45
Inlet PM Concentration (gr/dscf @ 7% 02)
Figure 2-10. PM removal efficiency as a function of inlet PM
concentration at the Oswego County MWC.
-------
0.042 -
0.04 —
0.038 -
c5
0.036 -
0.034 -
U
4
V -
0.03-
C
0.028 -
(0
0.026- R
0.024 -
-f
0.
— 0.02-
A U Start oi Campaign
+ Mid-rang. S.condary T.mp.ratur.
0.016 -
O End of Campaign
0.014 -
+ o A Low S.condary T.mp.ratur.
0.012 - +
0
0.01- I
0.15 0.25 0.35 0.45
Inlet PM ConcentraUon (gr/dscf O )
Figure 2-11. Outlet PM concentration a f jnction of inlet PM
concentration at the Oswego County MWC.
-------
TABLE 2-36. CDD/CDF DATA FOR OSWEGO COUNTY
I ’)
—.4
I- .
Test Condition
Run
Number
ESP Inlet
Tempgrature
C F)
Inlet CDD/CDF
Concentration
(ng/dscm
at 7% 02)
Outlet CDD/CDF
Concentration
(ngfdscm
at 7% 02)
CDD/CDF
RemovaL
Efficiency
Combustor = Normal.,
start of campaign
ESP NormaL
1
2
3
497
489
496
205
208
112
366
349
343
- 78.8
- 67.7
-207
Average
494
175
353
-118
Combustor Mid-range
temperature (1 ,750°F)
ESP Normal.
4
5
6
483
485
481
222
188
175
357
277
268
- 60.4
- 47.3
- 52.5
Average
483
195
301
- 53.4
Combustor = NormaL,
end of campaign
ESP = Normal
7
8
9
484
490
500
299
505
274
377
280
579
- 26.2
44.5
-111
Average
491
359
412
- 30.9
Combustor Low
temperature (1,650°F)
ESP = Normal
10
11
12
465
461
474
876
669
650
965
818
674
- 10.2
- 22.2
- 3.6
Average
467
732
819
- 12.0
-------
60-
40-
20 -
0-
>1 0
C)
-40-
S ::: + + +
-100-
N) .
-120- ________________
N)
—140 - • Start of Campaign
C)
— 1 60 - + Mid-rang. S.condary T.mp.ratur.
—180 - 0 End of Campaign
t Low S.condary T.mp.raturi
-200 - ______________________
-220-
460 470 480 490 500
ESP Inlet temperature (°F)
Figure 2-12. CDD/CDF removal ef tic. ncy as a function of ESP inlet
temperature at the Oswego County MWC.
-------
>‘
C)
U
a-
U i
>
0
E
r.’) 0
( ) U.
0
0
N
0
0
C)
60
40
20
0
-20
-40
-60
-80
-100
-120
-140
-160
-180
-200
-220
K>
A
A
A
++
-
-
+
.
U
.
K>
U
Start of Campaign
+
Mid-rang. S.condary T.mp.ratur.
K>
End ot Campaign
.
A
Low S.condary T.mp.ratur.
100
300 500 700
Inlet CDD/CDF Concentration (ng/dscm @ 7% Q 2 )
Figure 2-13. CDD/CDF removal efficiency as a function of inlet
CDD/CDF concentration at the Oswego County MWC.
900
-------
unprocessed MSW can also be fired. The RDF ranges in size from 3 to
18 inches. Emissions are controlled by four 3-field ESP’s manufactured by
Precipitair Pollution Control. Two ESP’s control emissions from three
combustors and two ESP’s control emissions from two combustors. The ESP’s
have design SCA’s of 307 ft 2 /1,000 acfm at 90,000 acfm flue gas flow and
400°F. The design PM removal efficiency of the ESP’s is 97 percent.
In December 1987 and January 1988, compliance tests were conducted.
Three simultaneous inlet/outlet measurements were made for PM across all four
ESP’s. In addition, each ESP outlet was sampled for SO 2 , SO 3 , HC1, and NO
during the same tests. At the outlet of the number 2 ESP flue gas was also
tested for COD/COF and metals (arsenic, beryllium, chromitn, lead, mercury,
and nickel).
Results of the particulate tests from all four ESP’s at Pigeon Point are
presented in Table 2-37. Outlet PM emissions were very low, ranging from
0.0012 to 0.0080 gr/dscf at 12 percent CO 2 . The average outlet PM
concentrations were 0.0029, 0.0015, 0.0019, and 0.0053 gr/dscf at flues 1, 2,
3, and 4, respectively. The average PM removal efficiencies for all four
ESP’s ranged from 98.7 percent at flue 4 to 99.8 percent for flues 1 and 3.
There was no apparent effect of inlet PM concentration on outlet PM
concentration or PM removal efficiency.
Metals data from three runs at the outlet of flue 2 at Pigeon Point are
presented in Table 2-38. Outlet concentrations for arsenic, chromium, lead,
mercury, and nickel averaged 0.83, 24, 150, 360, and 44 ug/dscm at 7 percent
02. Based on typical values for uncontrolled metals emissions for RDF MWC’s,
(Section 1.2), the ESP appears to have achieved high removal (greater than 9
percent) of lead, arsenic, and chromium. Nickel control appears to be
somewhat less, at approximately 90 percent. Mercury concentrations at the
ESP outlet were roughly 30 percent lower than typical uncontrolled levels.
However, because of the substantial variation in uncontrolled mercury
concentrations from RDF-fired MWC’s, it is uncertain whether mercury
reductions occurred.
CDD/CDF data from three runs at the outlet of flue 2 at Pigeon Point ar
presented in Table 2-39. Outlet CDD/CDF concentrations ranged from 71 to 14
ng/dscm at 7 percent 02 and averaged 105 ng/dscm. The ESP inlet temperature
2-74
-------
TABLE 2-37. PARTICULATE DATA FOR PIGEON POINT
Test Condition
Run Number
a
ESP InLet
Temp 8 rature
( F)
Flue
Gas FLow
(acfm)
InLet PM
Concentration
(gr/dscf
at 12% C0 2 )
OutLet PM
Concentration
(gr/dscf
at 12% C0 2 )
PM Removal
Efficiency
(%)
Combustor r NormaL
ESP = Normal
1-1
1-2
1-3
434
430
435
67,300
74,600
70.300
0.695
0.940
1.45
0.0017
0.0019
0.0052
99.8
99.9
99.6
Average (Flue 1)
433
70,700
1.03
0.0029
99.8
2-1
2-2
2-3
434
422
435
84,100
80,100
88,900
0.481
2.07
0.562
0.0013
0.0012
0.0021
99.7
99.9
99.6
Average (Flue 2)
430
84,400
1.04
0.0015
99.7
3-1
3-2
3-3
404
393
381
57,000
61,600
52,200
0.410
1.77
0.513
00154 b
0.0014
0.0023
962 b
99.9
99.6
Average (Flue 3)
393
56,900
0.898
0.0019
99.8
4-1
4-2
4-3
418
421
402
64,900
68,500
61,600
0.406
0.461
0.429
0.0080
0.0021
0.0058
98.0
99.5
98.7
Average (FLue 4)
414
65,000
0.432
0.0053
98.7
aRun Number contains the fLue number followed by the run number for each flue. FLue numbers 1 and 2 are for
three incinerators at 120 tpd each. Flue numbers 3 and 4 are for two incinerators at 120 tpd each.
bNozzle on sample train at outlet may have scraped port wall, causing contamination. Outlet concentration
and removal efficiency not included in average.
I )
u - I
-------
TABLE 2-38. METALS EMISSIONS DATA FOR PIGEON POINT
Test Condition
Run
a
Number
ESP Inlet
Temp 8 ra ure
C F)
OutLet PM
Concentration
(gr/dscf
at 12% C0 2 )
ation
(u /dscm at
7%
Outlet
As
Cr Pb
Hg
O 2 L_..
Ni
Combustor z Normal
ESP Normal
2-1
2-2
403
413
--
•-
0.925
0.878
45.8
8.8
190
127
370
385
47.1
7.4
2-3
415
--
0.695
16.4
146
333
77.2
Average
610
O.OO 3 O
0.833
23.7
154
363
43.9
5 RUfl Number contains the flue number foLLowed by the run number for each flue.
bTemperature estimated based on measured value at ESP outlet and previously measured temperature drop across
the ESP (38°F).
cPM samples not collected simultaneousLy. Average result reported.
-------
TABLE 2-39. CDD/CDF DATA FOR PIGEON POINT
Test Condition
a
Run Number
ESP
Temp
(
Inlet
8 ra ure
F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Combustor = Normal
2-1
411
102
ESP = Normal
2-2
2-3
414
411
70.8
142
Average
412
105
aRUfl Number contains the unit number followed by the run number on that unit.
biemperature estimated based on measured valu 8 at ESP outlet and previously
measured temperature drop across the ESP (38 F).
2-77
-------
remained at about 410°F during all three test runs. Although these CDD/CDF
values are relatively low, whether this is due to combustor or ESP
performance or both cannot be ascertained because no inlet CDD/CDF
measurements were made.
2.2.3.5 Pope/Douglas. 33 ’ 34 The Pope/Douglas Waste to Energy Facility
in Alexandria, Minnesota, includes two 38 tons/day excess-air modular
combustors manufactured by Cadoux, Inc. Emissions are controlled by 2-field
plate/plate ESP’s made by United McGill. The SCA of each ESP is
466 ft 2 /i,000 acfm with a design PM removal efficiency of 93 percent. The
design flue gas flow at the ESP inlet is 16,900 acfm at 400°F. The
controlled emissions exit through a common stack.
A compliance test was conducted in July 1987. All tests were performed
under normal combustor and ESP operating conditions. Particulate
measurements were made at the outlet of both units. Additional measurements
for CDD/CDF, lead, beryllium, arsenic, HC1, SO 2 , NOR , and other organics were
made at the outlet of Unit 2.
Results of the three test runs at each of the two units at Pope/Douglas
are presented in Table 2-40. The outlet particulate emissions from Unit 1
ranged from 0.022 to 0.029 gr/dscf at 12 percent CO 2 and averaged 0.025
gr/dscf. Emissions from Unit 2 ranged from 0.034 to 0.042 gr/dscf at
12 percent CO 2 and averaged 0.037 gr/dscf. The higher PM emissions measured
for Unit 2 may have been due to the rapping system in one of the ESP fields
having been out of service prior to testing. Although the rapping system was
repaired, PM buildup on the collection plate may not have had time to fully
clear before the testing.
Metals data from the outlet of Unit 2 are presented in Table 2-41.
Outlet concentrations from three runs for arsenic and lead averaged 1.2 and
980 ug/dscm, respectively. The one sample collected for mercury had a
concentration of 130 ug/dscm. Based on typical uncontrolled metals
concentrations, lead and arsenic were removed relatively effectively (greater
than 97 percent). Although mercury emissions are about 90 percent lower than
typical uncontrolled values, the wide variability in incontrolled mercury
levels and lack of inlet mercury data at Pope/Douglas prevent drawing any
conclusions.
2-78
-------
TABLE 2-40. PARTICULATE DATA FOR POPE/DOUGLAS
Test Condition
a
Run Number
ESP Inlet
Temp 8 ra ure
( F)
Flue Gas
Flow
(acfm)
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Combustor = Normal
ESP = Normal
1-1
1-2
1-3
480
483
477
13,200
13,800
13,800
0.024
0.029
0.022
Average (Unit 1)
484
13,600
0.025
2-1
2-2
2-3
480
490
482
13,700
13,500
13,400
0.042
0.034
0.034
Average (Unit 2)
484
13,500
0.037
aRun Number contains the unit number followed by the run number on that unit.
bTemperature estimated from measured temperature at the ESP outlet and an
assumed temperature drop across the ESP (65 F).
2-79
-------
TABLE 2-41. METALS EMISSIONS DATA FOR POPE/DOUGLAS
aRUfl Number contains the unit number followed by the run number for each unit.
bNM = not measured.
Clemperature assumed to be at average value of PM tests.
N)
Test Condition
Run
Numbera
ESP Inlet
Temperature
(°F)
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Outlet Concentration
(uci/dscm
As
at 7%
Pb
02)
Hg
Combustor = Normal
2-1
NMb
1.35
773
133
ESP = Normal
2-2
2-3
NM
NM
--
--
0.95
1.15
1,492
685
NM
NM
Average
482 C
0 • 031 d
1.15
983
133
dpM not measured simultaneously with metals. Average PM concentration reported.
-------
Table 2-42 presents CDD/CDF emissions data from three runs at the outlet
of Unit 2. CDD/CDF concentrations ranged from 358 to 529 ng/dscm at 7
percent 02 and averaged 416 ng/dscm. No inlet PM or CDD/CDF data are
available. The inlet temperature was consistently about 500°F during all
three runs.
2.3 SUMMARY OF PERFORMANCE
As discussed in the preceding review of available data from individual
units, the measured values of key operating variables that affect the
performance of an individual ESP generally cover too narrow a range to allow
meaningful analysis of their effect on ESP performance. This section
examines the combined data from all of the facilities tested to evaluate
relationships between key operating parameters and performance. In general,
more data are available to characterize performance in terms of outlet
emissions rather than removal efficiency due to the absence of simultaneous
measurements at ESP inlets at most facilities.
2.3.1 Particulate Matter
As discussed in Section 2.1, several measurable parameters can affect PM
performance. These include the number of fields, temperature, resistivity,
SCA, and inlet PM concentration. Figure 2-14 shows the the available data on
PM removal efficiencies and outlet PM concentrations as a function of the
number of ESP fields. Most of these data are from compliance test reports
and represent averages of multiple runs, typically three. Because data are
available for only one 1-field ESP, the plotted data for 1-field ESP’s
represent 12 runs at Oswego. As is shown, performance generally improves as
the number of fields increases, although there are no distinct boundaries
between the ranges. All but one of the ESP’s tested achieved an average PM
concentration of 0.03 gr/dscf or less. Several of the ESP’s demonstrated
outlet PM emissions of 0.01 gr/dscf or less. Five of nine 2-field ESP’s, ten
of fifteen 3-field ESP’s, and three of seven 4-field ESP’s tested achieved
average PM concentrations of 0.01 gr/dscf or less. These facilities include
three units at Baltimore RESCO, designed to achieve outlet PM concentrations
of 0.017 gr/dscf, and Unit 3 at Dayton, designed to achieve 0.015 gr/dscf.
Thus, the ESP’s achieving outlet PM concentrations of 0.01 gr/dscf are
2-81
-------
TABLE 2-42. CDD/CDF DATA FOR POPE/DOUGLAS
Test Condition
a
Run Number
ESP
Temp
(
Inlet
8 ra ure
F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Combustor = Normal
2-1
490
453
ESP = Normal
2-2
2-3
503
495
529
358
Average
496
446
aRUfl Number contains the unit number followed by the run number on that unit.
biemperature estimated from measured temperature at the ESP outlet and an
assumed temperature drop across the ESP (65 F).
2-82
-------
8
N
• 0
N
N
N
N
.
0
0 PM R•moval Eff ci.ncy
N OutI.t PM Conc.ntrat on
4
100
99
98
97
96
Number of ESP Fields
Figure 2-14. PM removal efficiency and outlet concentration as a
N
0
< ‘I
0
N. .. - -
K>.-
Oc
C)
N
U
C,)
C,)
0
(I)
C ’)
E
co
0
0.045 -
0.04 -
0.035 -
0.03 -
0.025 -
0.02 -
0.015 -
0.01
0.005
0
N
N
N
N
N
N
U
N
•
-u
- 95
94
-93
-922.
0
— 01 (•)
- 90
- 89
- 88
- 87
- 86
- 85
0
0;
•
I
I
I N
U
1 2 3
function of the number of ESP fields.
-------
generally state-of-the-art units designed for that level of outlet PM. All
of the 3-field ESP’s that did not achieve 0.01 gr/dscf were designed and
operated to meet a PM emission limit of 0.03 gr/dscf. However, the SCA’s of
these ESP’s were not generally lower than those of the 3-field ESP’s
achieving 0.01 gr/dscf. The 4-field ESP’s were installed on RDF-fired MWC’s
and were permitted to achieve a 0.03 gr/dscf emission limit (Niagara Falls)
or a limit of 0.1 gr/dscf (NSP Red Wing).
Nine of the ten 3- and 4-field ESP’s for which PM removal efficiency was
measured achieved a PM emission reduction efficiency of 98.7 percent or
greater. In addition, an increase in PM removal efficiency with increases in
the number of fields is observed.
The Deutsch-Anderson equation (discussed in Section 2.1) states that PM
collection efficiency increases with SCA. As discussed for individual ESP’s
in Section 2.2, however, changes in flue gas flow rate and SCA at the
individual facilities tested are generally too small to observe this
relationship. Figure 2-15 is a graph of removal efficiency as a function of
actual SCA for all of the facilities tested and discussed in Section 2.2 for
which removal efficiency data are available. With the exception of Oswego
County, which has a 1-field ESP, all of the ESP’s shown on the figure have 3-
or 4-fields. These data show that high PM removal efficiencies can be
obtained with SCA’s of 200 to greater than 600 ft 2 /1,000 acfm.
Figure 2-16 shows outlet PM concentration as a function of design SCA.
There is no consistent trend observed with the data. At SCA’s from 200 to
greater than 600 ft 2 /1,000 acfm, PM concentrations of less than 0.005 to
about 0.020 gr/dscf can be obtained.
Figure 2-17 shows the relationship between inlet and outlet PM
concentrations. This plot is based on the data from individual test runs
presented in Section 2.2. Note that the outlet PM concentrations from
individual ESP’s do not increase as a function of inlet PM concentration.
For example, outlet PM concentrations at Peekskill were between 0.009 and
0.019 gr/dscf at ESP inlet concentrations ranging from 0.3 to 2.7 gr/dscf.
Outlet PM concentrations of less than 0.01 gr/dscf were achieved by several
ESP’s with inlet PM concentrations from 0.4 to over 2.0 gr/dscf.
2-84
-------
100 -
C 0 +
99 --- U
S K>
98
—S 97
>
0
96 Mass Burn
+ Baitimor.
<) P..kakill
A Dayton
RDF R.d Wing
0 Modular
N) It
X Osw.go
93 jJJ Pig.on Point 1
0 Pig.on Point 2
92 -- x 0 Pig.on Point 3
• Pig.on Point 4
91
90 - - - I I I I I I I
150 250 350 450 550 650
Actual SCA (sq. ft./1000 acfm)
Figure 2-15. PM removal efficiency as a function of actual SCA.
-------
0.045 - ____________________________________
Sit. (U UP Plaid.)
0 5W •Osw.go(1) DIN .Dayton(3) RW
__ 0.04 ON • On.ida (2) pp • Pig•on Point (3)
- Barren County (2) PK P..kskiii (3)
0 M I Mckay Buy (2) ILl • Ialtim.rs ( ) P/D
NP • Niagara PaIls (4)
N 0.035 - P/D .Pop./Dougla .(2)
RW - R•d Wing (4)
__ BAY • Bay County (3)
ILS - Tulsa (3)
o 0.03-
‘ S
ON °
0.025 - NF
BAY RW
‘S
I-
0.02 - OSW
0 BAY
U
C PK
o NF
0 0.015 -
N)
—
0
0.01 BC
TLS
o MB
DTN
0.005 TLS
MB BLT
BET
0 - I I I
200 300 400 500 600
*
Modular atarv.d-air unit. with low lnI .t PM conc.ntrations.
Design SCA (eq. ft./1000 acfm)
Figure 2-16. Outlet PM concentration as a function of design SCA.
-------
0.045 -
U _______________
X Mas . Burn
__ 0.04 - 0 North Andovsr
61 P,.kskili
C,) t Dayton
0.035 — Q Pin•llas County
V Baitimor. RESC(
I-
RDF
• NSPR.d Wing
0.03 — Modular
0) + Pig.on Point
V x
N X Oswego
I - x ________
.! 0.025 -
x
x
0
0.02- X
0 0
0
0 0
0.015- +
N) o
xx
x
- 1 0.01- > < N 0
o 0
0.005- + D +
V V
++++o R o
+
0— I I I I I I
0 1 2 3 4
inlet PM Concentration (gr/dscf @ 12% C0 2 )
Figure 2-17. Relationship between ESP inlet and outlet PM
concentrations.
-------
These data demonstrate that PM emissions of 0.03 gr/dscf at 12 percent
CO 2 are achievable by existing ESP’s over a range of the number of fields,
SCA, and inlet PM concentrations. It is not possible, however, to predict
ESP performance based solely on the number of fields and SCA. The key
factors affecting ESP performance not accounted for by the above parameters
are the uniformity of flue gas distribution across the ESP and the design of
the ESP’s plate/electrode geometry that limits bypass of PM around the
ionizing fields and collection surfaces. Because these factors cannot be
changed during normal rebuild and repair of an ESP, the ability to increase
the performance of existing ESP’s to 0.01 gr/dscf may be limited.
2.3.2 Metals
Only limited metals removal efficiency data with simultaneous
measurements at the ESP inlet and outlet are available (Baltimore RESCO and
Dayton). However, it is possible to estimate a removal efficiency based on
typical uncontrolled metals concentrations presented in Section 1.2. Because
uncontrolled metals concentrations are fairly similar at different sites, the
estimated removal efficiency is believed to be representative. Figures 2-18
and 2-19 show the metal and PM collection efficiency for individual ESP’s
discussed in Section 2.2. Figure 2-18 covers ESP’s with estimated or
measured PM collection efficiencies of 98 percent or greater. Figure 2-19
covers ESP’s with PM collection efficiencies of less than 98 percent.
Mercury is not included on either figure due to the absence of mercury
removal occurring at any of these facilities.
As shown in Figure 2-18, for high efficiency ESP’s with collection
efficiencies of 98 percent or greater, removal efficiencies for arsenic,
cadmium, chromium, lead, and nickel were generally from 0 to 3 percent less
than the PM removal efficiency at the same facility. This suggests a high
association of these metals with PM, with some enrichment of metals on the
particulate that is less efficiently captured by high efficiency ESP’s.
For ESP’s with PM collection efficiencies of less than 98 percent,
metals removals show somewhat more scatter, with collection efficiencies for
individual metals being both higher and lower than for PM. As with high
2-88
-------
100— ——- 0
p
o +
R
x
99- X 0 0 0 I I i + 0
V +
± X A
98
0 +
97- - >K
V C) V
96- X
x
95 V
94 -
93 -
• PM
92- X Cr
+ As
V Pb
0 Be
91 A Cd o NI 0
90— I I I 1 I I
o > =
U
U)
I— >. C C
0 C 0 0
0
• 0 • a
• U • a.
a a
a.
a. 0
C a.
• &
Figure 2-18. Metals and PM removal efficiency for ESP’s
with PM removal efficiency above 98 percent.
-------
100
H V
p
U
90 -
U
()
>1
o 80-
C
w
.
> 70-
0
E
0
x
‘a
• PM
60- X Cr
+ As
V Pb
O Be +
o N
A Cd
50- I
- ‘ N >. >.
=
N C C
C I L.
0 0
o 0 0
N C 15
15 o 0
o N
.
15 C
0
Figure 2-19. Metals and PM removal efficiency for ESP’s with
removal efficiency below 98 percent.
-------
efficiency ESP’s, however, there is a strong association between PM and
metals collection efficiencies. In most of these instances, metals
collection efficiencies from these units exceeded 90 percent.
For metals other than mercury, the control of PM emissions will also
achieve reductions in metal emissions. If the PM removal efficiency is
98 percent or greater, the removal efficiency for arsenic, beryllium,
cadmium, chromium, lead, or nickel will generally be at least 95 percent.
Mercury is not removed by an ESP.
2.3.3 CDD/CDF
Outlet CDD/CDF concentrations have been measured at 13 MWC facilities
with ESP’s. Average ESP outlet concentrations ranged from 33 to 17,000
ng/dscm at 7 percent 02. Emissions averaging less than 500 ng/dscm were
measured at 10 of the 13 facilities. Inlet and outlet CDD/CDF concentrations
were measured at five MWC’s with ESP’s. Three (Pinellas County, North
Andover, and Peekskill) are mass burn waterwall units with heat recovery
systems upstream of the ESP; one (Oswego County) is a modular starved-air
unit with heat recovery; and one (Dayton) is a mass burn refractory wall unit
with a spray quench chamber (no heat recovery) upstream of the ESP for flue
gas cooling. Outlet emissions were higher than inlet emissions at four of
these facilities, indicating formation across the ESP. The fifth facility
(Peekskill) had lower outlet than inlet emissions, and thus had positive
removal efficiency. An analysis of all five data sets indicates that ESP
operating temperature is the principal parameter that affects performance on
CDD/CDF emissions.
Figure 2-20 presents a plot of CDD/CDF removal efficiency as a function
of ESP inlet temperature for the data from the four facilities with heat
recovery systems upstream of the ESP. The data suggest increased CDD/CDF
removal (less formation) for decreasing temperature. The temperatures
evaluated range from 430 to 590°F, which include temperatures below which
CDD/CDF formation is hypothesized to occur (450 to 650°F). Excluding
Peekskill, higher outlet CDD/CDF concentrations than inlet (negative removal
efficiency) are observed for 20 of 21 data points between 460 and 590°F. The
Peekskill data, collected at ESP inlet temperatures of 435 to 479°F, are the
only data set to have consistently positive CDD/CDF removal efficiency.
2-91
-------
a)
C.)
I -
a)
>
C.)
a)
C.)
1 -
LU
>
0
E
a)
N)
N ,)
a
a
0
ESP Inlet Temperature (°F)
Figure 2-20.
CDD/CDF removal efficiency as a function of ESP
inlet temperature at MWC’s with ESP’s.
A
80
60
40
20
0
-20
-40
-60
-80
-100
-120
-140
-160
-180
-200
-220
A
L
A —___
A 0 North Andov.r
4. D Pin.II s County
— A A A P..k.klII
A + Osw.go County
A
+
+
+
+
0
++ 0 0
- + 0 0
+
- + 0
+
o
0
+
I
I
I
I
430 450
470
490
510
I
I
I
I
530
550
570
590
-------
Thus, although there is no clear distinction, the data suggest that around
470°F or less, positive removal can be achieved.
The Dayton results are not included in Figure 2-20 because the MWC is a
refractory unit and uses a spray quench chamber rather than heat recovery for
flue gas cooling. Results for Dayton are shown in Figure 2-21. The quench
chamber both cools the flue gas and removes most of the larger PM particles.
Some removal of CDD/CDF and HC1 may also occur in the quench chamber. The
rapid cooling of flue gas in the quench chamber may limit reaction time at
temperatures required for CDD/CDF formation prior to the ESP. As a result,
the ESP inlet CDD/CDF concentrations are relatively low, and are 80 to 90
percent lower than measured upstream at the mixing chamber. However, in the
ESP, sufficient residence time, surface area, and temperature are apparently
available to promote formation of CDD/CDF, as evidenced by increases in
CDD/CDF concentration of 300 to 20,000 ng/dscm across the ESP. Because of
the differences between Dayton and MWC’s with heat recovery systems, the
CDD/CDF data from Dayton may not be comparable with the data from the other
MWC’s.
As with MWC’s with heat recovery, however, reductions in ESP operating
temperature at Dayton reduced outlet CDD/CDF concentrations as shown in
Figure 2-21. By decreasing the ESP inlet temperature from 575 to 400°F,
outlet CDD/CDF concentrations decrease from 17,000 ng/dscm to 1,000 ng/dscm.
This is a similar trend to that observed in Figure 2-20, suggesting that a
decrease in temperature at the ESP inlet will improve ESP performance for
CDD/CDF.
The inlet CDD/CDF concentration may have a secondary effect on
performance, but the data are inconclusive. There is no apparent effect of
inlet PM concentrations on performance.
Based on analysis of the available data, ESP performance for CDD/CDF can
be improved by lowering the ESP inlet temperature. For MWC’s with heat
recovery, limiting ESP inlet temperature to 450°F or less appears to
eliminate formation of CDD/CDF across the ESP.
2-93
-------
24000 - — ---______ ____ ___
22000 -
20000
18000 —
16000-
a
10000 -
C.)
t I
8 8000-
a
o 6000-
C.)
0,
— 4000-
0
2000 - a
a a
0 - 1 I I I I
380 400 420 440 460 480 500 520 540 560 580
ESP Inlet Temperature (°F)
Figure 2-21. Outlet CDD/CDF concentration as a function of
ESP inlet temperature at Dayton.
-------
2.4 REFERENCES
1. Turner, J. H., P. A. Lawless, T. Yamamoto, D. W. Coy, G. P. Greiner,
J. D. McKenna, and W. M. Vatavuk. Sizing and Costing of Electrostatic
Precipitators (Part I. Sizing Considerations). Journal of Air Pollution
Control and Waste Management (JAPCA). April 1988. pp. 459-460.
2. Reference 1, PP. 458-459.
3. Sedman, C. B., and T. G. Brna. Municipal Waste Combustion Study: Flue
Gas Cleaning Technology. U. S. Environmental Protection Agency,
Research Triangle Park, NC. EPA Publication No. EPA/530-SW87-021d.
June 1987. pp. 2-3 to 2-4.
4. California Air Resources Board. Air Pollution Control at Resource
Recovery Facilities. Sacramento, CA, May 24, 1984. pp. 153-156.
5. Reference 4, pp. 147-151.
5. Vogg, H., and L. Stieglitz. Chemosphere. 1986. Vol. 15, p. 1373.
‘. Zurlinden, R. A., et. al., (Ogden Projects, Inc.). Environmental Test
Report, Alexandria/Arlington Resource Recovery Facility, Units 1, 2,
and 3. (Prepared for Ogden Martin Systems of Alexandria/Arlington,
Inc.) Alexandria, VA. Report No. 144 A (Revised). January 1988.
Entropy Environmentalists, Inc. Baltimore RESCO Company, L. P.,
Southwest Resource Recovery Facility. Particulate, Sulfur Dioxide,
Nitrogen Oxides, Chlorides, Fluorides, and Carbon Monoxide Compliance
Testing, Units 1, 2, and 3. (Prepared for RUST International, Inc.)
January 1985.
PEI Associates, Inc. Method Development and Testing for Chromium, No. 2
Refuse-to-Energy Incinerator, Baltimore RESCO. (Prepared for
U. S. Environmental Protection Agency. Research Triangle Park, NC. EMB
Report 85-CHM8. EPA Contract No. 68-02-3849. August 1986.
Beachier, 0. 5., et. al. (Westinghouse Electric Corporation). Bay
County, Florida, Waste-to-Energy Facility Air Emission Tests. Presented
at Municipal Waste Incineration Workshop, Montreal , Canada.
October 1987.
Radian Corporation. Preliminary Test Data from October-November 1988
Testing at the Montgomery County South Plant, Dayton, Ohio.
Clean Air Engineering, Inc. Report on Compliance Testing for Waste
Management, Inc. at the McKay Bay Refuse-to-Energy Project located in
Tampa, FL. October 1985.
2-95
-------
13. Campbell. J., Chief, Air Engineering Section, Hilisborough County
Environmental Protection Commission, to Martinez, E. L., Source Analysi
Section/AMTB, U. S. Environmental Protection Agency, May 1, 1986.
14. Telecon. Weigle, G., McKay Bay Refuse-to-Energy Project, with
Vancil, M. A., Radian Corporation. April 7, 1988. Specific collecting
area of ESP.
15. Anderson, C. L., et. al. (Radian Corporation). Summary Report, CDD/CDF,
Metals and Particulate, Uncontrolled and Controlled Emissions, Signal
Environmental Systems, Inc., North Andover RESCO, North Andover, MA.
(Prepared for U. S. Environmental Protection Agency. Research Triangle
Park, NC. EMB Report No. 86-MINO2A. EPA Contract No. 68-02-4338.
March 1988.
16. Fossa, A. J., et. al. Phase I Resource Recovery Facility Emission
Characterization Study, Overview Report. New York State Department of
Environmental Conservation. Albany, NY. May 1987.
17. Radian Corporation. Results from the Analysis of MSW Incinerator
Testing at Peekskill, New York (DRAFT). (Prepared for the New York
State Energy Research and Development Authority). Albany, NY.
March 1988.
18. Entropy Environmentalists, Inc. Stationary Source Sampling Report,
Signal RESCO, Pinellas County Resource Recovery Facility, St.
Petersburg, Florida, CARB/DER Emission Testing, Unit 3 Precipitator
Inlets and Stack. February and March 1987.
19. Lavalin, Inc. National Incinerator Testing and Evaluation Program: The
Combustion Characterization of Mass Burning Incinerator Technology;
Quebec City (DRAFT). (Prepared for Environmental Protection Service,
Environmental Canada). Ottowa, Canada. September 1987.
20. Seelinger, R., et. al. (Ogden Products, Inc.) Environmental Test
Report, Walter B. Hall Resource Recovery Facility, Units 1 and 2.
(prepared for Ogden Martin Systems of Tulsa, Inc.). Tulsa, OK.
September 1986.
21. Knisley, D. R., et. al. (Radian Corporation). Emissions Test Report,
Dioxin/Furan Emission Testing, Refuse Fuels Associates, Lawrence MA.
(Prepared for Refuse Fuels Association). Haverhill, MA. June 1987.
22. Entropy Environmentalists, Inc. Stationary Source Sampling Report,
Ogden Martin Systems of Haverhill, Inc., Lawrence, Massachusetts Thermal
Conversion Facility. Particulate, Dioxins/Furans and Nitrogen Oxides
Emission Compliance Testing. September 1987.
2-96
-------
23. Preliminary Report on Occidental Chemical Corporation EFW. New York
State Department of Environmental Conservation. Albany, NY.
January 1986.
24. Trip Report, Occidental Chemical Corporation EFW in Niagara Falls, NY.
Epner, E., Radian Corporation. December 1987. Attachments.
25. H. J. Hall, Associates. Summary Analysis on Precipitator Tests and
Performance Factors, May 13-15, 1986 at Incinerator Units 1, 2 -
Occidental Chemical Company. Prepared for Occidental Chemical Company
EFW. Niagara Falls, NY. June 25, 1986.
26. Interpoll Laboratories. Results of the March 21 - 26, 1988, Air
Emission Compliance Test on the No. 2 Boiler at the Red Wing Station,
Test IV (High Load). Prepared for Northern States Power Company,
Minneapolis, Minnesota. Report No. 8-2526. May 10, 1988.
28. Memorandum. Perez, Joseph, AM/3, State of Wisconsin,
of Stack Test Performed at Barron County Incinerator.
1987.
29. Reference 16.
31. Radian Corporation. Data Analysis Results for Testing
Modular MSW Combustor: Oswego County ERF, Fulton, NY.
York State’s Energy Research and Development Authority.
November 1988.
33. Interpoll Laboratories. Results of the July 1987 Emission Performance
Tests of the Pope/Douglas Waste-to-Energy Facility MSW Incinerators in
Alexandria, MN. (Prepared for HDR Techserv, Inc.). Minneapolis, MN.
October 1987.
34. Telecon. Chamberlain, L., Minnesota Pollution Control Agency, with
Vancil, M. A., Radian Corporation. March 30, 1988. Specific collection
area of ESP’s at Pope/Douglas.
27. Interpoll Laboratories.
Compliance Test on Unit
NSP Red Wing Station.
Minneapolis, Minnesota.
Results of the May 24 - 27, 1988 High Load
1 and Low Load Compliance Test on Unit 2 at the
Prepared for Northern States Power Company.
Report No. 8-2559. July 21, 1988.
30. Telecon.
April 4,
to Files.
February
Review
24,
DeVan, S. Oneida ERF, with Vancil, M. A., Radian Corporation.
1988. Specific collecting area of ESP’s.
32. Letter with attachments from Gehring, Philip,
Point Energy Generating Facility), to Farmer,
OAQPS, U. S. Environmental Protection Agency.
at a Two-Stage
Prepared for New
Albany, NY.
Plant Manager (Pigeon
Jack R., Director, ESO,
June 30, 1988.
2-97
-------
3.0 FURNACE SORBENT INJECTION
Section 3 describes the technology and performance of furnace sorbent
injection (FSI) systems with an ESP for PM control as applied to MWC
facilities. No data are available for FSI systems followed by a FF for PM
control as applied to MWC facilities. In Section 3.1, operation and design
of FSI systems are described. Section 3.2 describes MWC facilities equipped
with FSI/ESP systems for which emissions data are available and summarizes
the available emissions test data. In Section 3.3, the performance of
FSI/ESP systems relative to the control of acid gas, PM, metals, and CDD/CDF
emissions is described.
3.1 PROCESS DESCRIPTION
Furnace sorbent injection is a technique for controlling acid gas
emissions which has been applied to conventional and fluidized bed MWC’s.
Lime (CaO) or limestone (CaCO 3 ) is injected into the furnace section of the
MWC. If limestone is used, sufficient temperature must be available for
calcination of CaCO 3 to CaO and CO 2 . Limestone is rapidly calcined to lime
in the furnace at temperatures of 1,400 to 2,000°F. The lime, in turn,
reacts with SO 2 and HC1 to form calcium sulfate (CaSO 4 ) and calcium chloride
(CaC1 2 ), which can be removed by the PM control device. In conventional
MWC’s, sorbent can be injected into the furnace using existing overfire air
jets or through separate ports located above the fuel bed. In fluidized bed
combustors, the sorbent is injected either into or above the fuel bed.
Furnace sorbent injection provides for effective SO 2 removal because
lime and SO 2 readily react at typical furnace temperatures of 1,600 to
2,200°F. Furnace sorbent injection also provides extended contact time
between lime and acid gases, as they are in contact in the flue gas starting
in the combustor, through the heat recovery sections, and ending in the
particulate control device. Furnace sorbent injection can also potentially
reduce CDD/CDF formation by removing chlorine prior to chlorination of
CDD/CDF. However, HC1 and lime reportedly do not react at temperatures
above 1,400°F.’ Potential disadvantages of FSI include fouling and erosion
of convective heat transfer surfaces by the injected sorbent.
Use of FSI will increase the PM loading to the particulate control
device. Because of this, a particulate control device needs to be sized
3-1
-------
larger than if FSI were not used. However, the characteristics of the PM
are not changed sufficiently by injection of sorbent to affect the ability
of the control device to remove particulate matter. Operation and design of
ESP’s is described in Section 2.1. Fabric filter design and operation is
discussed in Section 4.1.
3.2 SUMMARY OF TEST DATA
Section 3.2 presents the available emissions data for MWC facilities
equipped with furnace sorbent injection systems followed by an ESP. A
description of the facility and a summary and analysis of the emissions data
are provided for each facility. The data presented in this section are
based on short-term testing (less than 3 hours). Although either an ESP or
fabric filter can be used as the PM control device, data are available only
from systems with an ESP.
The effects of stoichiometric ratio and ESP inlet temperature on acid
gas removal are discussed in each section. The effect of SCA on PM removal
and the effect of ESP inlet temperature on CDD/CDF removal are discussed as
well.
Because simultaneous measurements of uncontrolled and controlled acid
gases cannot be made at MWC’s using furnace sorbent injection, removal
efficiencies for acid gases cannot be directly calculated. However, acid
gas removal efficiencies for a given MWC are estimated based on typical acid
gas concentrations at that MWC operating without FSI if those data are
available or based on typical uncontrolled acid gas emission levels (see
Section 1.2).
3.2.1 Alexandria. 2 ’ 3
The Alexandria/Arlington Resource Recovery Facility in Alexandria,
Virginia, consists of three identical 325 ton/day Martin GmbH waterwall
combustors. Emissions from each combustor are controlled with a 3-field
ESP. The flue gas temperature at the ESP outlet is generally about 340°F
with a gas flow of about 67,000 acfm (40,000 dscfm). The ESP’s are designed
to meet a PM emission limit of 0.03 gr/dscf at 12 percent CO 2 . Dry hydrated
lime can be injected into the combustor with the overfire air for acid gas
control and has been used on these combustors since startup in January 1987.
3-2
-------
In December 1987, tests were performed to demonstrate compliance with
iperating permits. The tests were conducted under normal operating
conditions with dry hydrated lime injection on Unit 1 and without lime
injection on Units 2 and 3. The tests with lime injection are reported
here. The tests without lime addition are reported in Section 2.2.1.1. The
hydrated lime feed rate during testing was 150 lb/hr. Flue gas was sampled
at the ESP outlet for SO 2 , HC1, PM, NOR, and CDD/CDF.
Acid gas data are presented in Table 3-1. Acid gases were measured at
the ESP outlet of Unit 1 only. Outlet SO 2 concentrations ranged from 32 to
45 ppm at 7 percent 02 over three runs and averaged 37 ppm. Assuming a
typical uncontrolled SO 2 concentration of 200 ppm (see Section 1.2),
approximately 80 percent SO 2 removal occurred. Outlet HC1 concentrations
ranged from 148 to 200 ppm at 7 percent 02 and averaged 166 ppm. Based on a
typical uncontrolled HC1 concentration of 500 ppm (see Section 1.2), 60 -
70 percent HCT removal occurred. Because uncontrolled SO 2 and HC1
concentrations cited above were not measured, the actual stoichiometric
ratio could not be evaluated. However, based on typical uncontrolled SO 2
and HC1 concentrations cited above, the hydrated lime feed rate of 150 lb/hr
corresponds to a stoichiornetric ratio of about 0.9. The temperature at the
ESP inlet varied over a limited range (366 to 373°F), preventing analysis of
the effect of this parameter on FSI/ESP performance.
In Table 3-2, particulate data from testing with furnace sorbent
injection are presented. Flue gas flow rate and temperature were relatively
constant for all three runs. Outlet PM emissions ranged from 0.016 to
0.035 gr/dscf and averaged 0.024 gr/dscf. Particulate data for Alexandria
without furnace sorbent injection are presented in Section 2.2.1. Without
FSI, outlet PM emissions averaged 0.027 gr/dscf. Therefore, the use of FSI
did not detrimentally affect ESP performance during the tests.
In Table 3-3, CDD/CDF data are presented. Outlet CDD/CDF emissions
ranged from 52 to 59 ng/dscm and averaged 55 ng/dscm. Uncontrolled CDD/CDF
emissions were not measured. However, other combustors with Martin GmbH
grates have demonstrated low uncontrolled COD/CDF emissions (e.g., Marion
County with 40 ng/dscm [ Section 7.2.5] and Pinellas County with 54 ng/dscm
3-3
-------
TABLE 3 -1. ACID GAS DATA FOR ALEXANDRIA WITH LIME INJECTION
Outlet
Acid
Gas
Test
Condition
Run
Numbers
ESP Inlet
Tempegaturea
( F)
Stoichiometric
Ratio
Concentrations
IDPm, dry at 7%
1
SO 2
02
HC1
Combustor
= Normal
1
366
NMb
45
200
FSI/ESP =
Normal
2
3
364
373
NM
NM
34
32
150
148
Average
368
0 • 9 C
37
166
alemperatue estimated fr 8 m a measured value at the stack and an assumed temperature
drop across the ESP (20 F).
bNM Not measured.
CEstimated from known lime feed rate of 150 lb/hr and assumed inlet SO 2 and HC1
concentrations of 200 and 500 ppm (see Section 1.2), respectively.
-------
TABLE 3-2. PARTICULATE DATA FOR ALEXANDRIA WITH LIME INJECTION
Outlet PM
ESP Inleta Flue Gas Concentration
Test Run Temperature Flow (gr/dscf at
Condition Number ( F) (acfm) 12% C0 2 )
Combustor = Normal 1 366 70,196
FSI/ESP = Normal 2 364 67,605
3 373 73,376
0.016
0.035
0.022
Average 368 70,392
0.024
aTemperature estimated from measured 0 value at the stack
temperature drop across the ESP (20 F).
and an assumed
TABLE 3-3. CDD/CDF DATA FOR ALEXANDRIA WITH
LIME INJECTION
ESP Inleta
Test Run Tempe ature
Condition Number ( F)
Outlet CDD,’CDF
Concentration
(ng/dscm at 7% 02)
Combustor = Normal 1 358
FSI/ESP = Normal 2 364
3 364
51.8
53.8
59.0
Average 362
54.9
aTemperature estimated from measured 0 value at the stack and an assumed
temperature drop across the ESP (20 F).
3-5
-------
Section 2.2.1.8fl, suggesting that injection of hydrated lime into the
furnace combined with ESP operation at 360°F, resulted in little or no net
effect on the control of CDD/CDF emissions.
3.2.2 Dayton 4
The Montgomery County South incinerator plant in Dayton, Ohio includes
three nearly identical Volund refractory-lined combustors. Emissions are
controlled with ESP’s. A complete description of the facility is provided
in Section 2.2.1.4.
In November and December 1988, testing was conducted by EPA on Unit 3.
Tests were conducted with furnace sorbent injection, duct sorbent injection,
and without sorbent injection. A complete description of the test program
is provided in Section 2.2.1.4. The results from the tests with furnace
sorbent injection are reported here. The results from tests without sorbent
injection are reported in Section 2.2.1.4. The results of tests with duct
sorbent injection are reported in Section 4.2.1.
Acid gas data from the screening and parametric tests are presented in
Table 3-4. A total of 30 test runs were conducted under a variety of
sorbent feed rates and ESP inlet temperatures. It should be noted in
reviewing the data that the water spray in the quench chamber provided some
removal of HC1 from the flue gas. Also, high stoichiometric ratios were
used, which are not typical of normal operation. Thus, the performance
demonstrated at Dayton may not be indicative of a system under commercial
operation.
In addition, because limestone is injected directly into the furnace,
it is not possible to measure uncontrolled acid gas concentrations
and calculate removal efficiency. Therefore, typical uncontrolled
concentrations of 500 ppm for HC1 and 200 ppm for SO 2 are assumed. (See
Section 1.2.)
Finally, because the true stoichiometric ratio could not be determined
for each test, the effect of this parameter is not evaluated. Instead, the
effect of limestone feed rate on system performance is evaluated.
Variations in temperature while maintaining in limestone feed rate allow
evaluation of the effect of temperature on system performance.
3-6
-------
TABLE 3-4. ACID GAS DATA FOR DAYTON WITH FURNACE SORBENT INJECTION
Test
Condition
Run
Number
ESP Inlet
Temperature
(°F)
Stoichiom tric
Ratio
Acid Gas Concentration
(ppm, dry at 7%
ESP
Inlet E P Out’ et
HO !. SO 2 HCI
SO 2
Combustor Normal P10 397 3.3 97.0 90.0 218
FSI= 500 sb/hr P11 391 3.3 19.8 NM 34.0 80.3
ESP = 400 F inLet P12 395 3.3 28.5 NM 48.1 62.8
Average 394 3.3 48.4 MM 57.4 120
Combustor = Normal P13 297 3.3 15.9 NM 21.6 11.5
FSZ = 500 tb/hr P13A 297 3.3 11.6 NM 21.1 NM
ESP = 300°F inlet P14 301 3.3 62.3 NM 31.1 79.5
P14A 301 3.3 18.5 NM 17.7 NM
P15 296 3.3 4.5 NM 19.1 13.6
P15A 296 3.3 9.3 NM 14.5 MM
Average 298 3.3 20.3 NM 20.8 34.9
Combustor = Low mixing SBA 550 1.6 192 152 MM NM
chamber temp. (1550 F) SBB 550 1.6 201 168 NM NM
FS I = 250 Lb/hr SBC 550 1.6 103 180 NM NM
ESP 550°F inlet
Average 550 1.6 165 167’ NM NM
Combustor = Low mixing S5A 550 1.6 72.5 141 46.8 MM
chamber temp. (1600 F) SSB 550 1.6 NM 197 52.9 NM
FSI = 250 Lb/hr S5C 550 1.6 NM 211 37.5 NM
ESP = 550°F inLet
Average 550 1.6 72.5 183 45.7 NM
Combustor = Normal 56 550 1.6 76.6 MM 58.0 MM
FSI = 250 tb/hr S7 550 1.6 101 NM 77.1 NM
ESP = 550°F inlet
Average 550 1.6 88.6 NM 67.6 NM
Combustor Normal S1A 400 1.6 122 171 NM NM
FSI = 250 Lb/hr SIB 400 1.6 113 193 NM NM
ESP = 400°F inLet SiC 400 1.6 128 173 NM NM
Average 400 1.6 124 179 MM NM
Combustor Normal SIZ 350 1.6 83.0 119 NM NM
FSI = 250 Lb/hr
ESP 350°F inlet
Combustor Normal S2A 400 4.9 77.4 26.3 MM NM
FS 1 = 750 Lb/hr S2B 400 4.9 75.4 73.1 NM NM
ESP = 400°F inLet S2C 400 4.9 45.5 35.4 NM NM
Average 400 4.9 66.1 44.9 MM NM
Combustor Normal S3A 350 4.9 24.8 53.9 NM NM
FSI = 750 lb/hr S3B 350 4.9 51.6 43.0 MM NM
ESP = 350°F inLet S3C 350 4.9 35.0 41.4 NM NM
Average 350 4.9 37.1 46.1 MM NM
Combustor NormaL S4A 300 4.9 7.2 45.3 NM NM
FSI = 750 tb/hr S4B 300 4.9 20.3 39.4 NM NM
ESP = 300°F inLet S4C 300 4.9 15.4 92.1 NM NM
Average 300 4.9 14.3 58.9 NM NM
aS Ihf . ratio caLculated using known Limestone feed rate and typica!. uncontrolled SO 2 and HCI.
concentrations (200 ppm at 7 percent 02 for SO 2 and 500 ppm at 7 percent 02 for HCI).
bNM = Not measured.
3-7
-------
The parametric tests involved FSI operation at a constant sorbent
injection rate of 500 lb/hr with three test runs at an ESP inlet temperature
of 400°F and six test runs at an ESP inlet temperature of 300°F. The ESP
inlet SO 2 concentrations ranged from 4.5 ppm to 97 ppm over the nine test
runs, averaging 48 ppm at an ESP inlet temperature of 400° and 20 ppm at an
ESP inlet temperature of 300°F.
Assuming a typical uncontrolled SO 2 concentration of 200 ppm (see
Section 1.2), the measured outlet SO 2 concentration during the parametric
tests suggest removal efficiencies of 55 to 93 percent. The ESP outlet HC1
concentrations ranged from 12 to 218 ppm over six test runs, averaging
120 ppm at an ESP inlet temperature of 400°F and 34.9 at an ESP inlet
temperature of 300°F. HC1 removal efficiencies of 56 to 98 percent are
estimated assuming typical uncontrolled HC1 concentrations of 500 ppm (see
Section 1.2).
The screening tests involved FSI operations at sorbent injection rates
of 250 and 750 lb/hr at four ESP inlet temperatures (550, 400, 350, and
300°F). A total of 12 test runs were conducted with 250 lb/hr of sorbent
injected. Eight of these tests were conducted at an ESP inlet temperature
of 550 0 F (three with 1550° mixing chamber temperature, three with 1600°F
mixing chamber temperature, and two with normal [ 1800°F] mixing chamber
temperature), three tests at an ESP inlet temperature of 400°F, and one test
with a temperature of 3500. These last four tests were at normal mixing
chamber temperatures. The ESP inlet SO 2 concentrations ranged from 72 to
200 ppm over all 12 runs and averaged 125 ppm at ESP inlet temperature of
550°F, 121 ppm at 400°F, and 83 ppm at 350°F. Based on the variability in
run to run SO 2 measurements, no clear relationship between SO 2 levels and ESP
temperature are apparent over this temperature range. At the ESP outlet, SO 2
concentrations averaged 54 ppm during the tests conducted at an ESP inlet
temperature of 550°F. Data on SO 2 concentrations at the ESP outlet were not
collected during the other tests. Based on a typical uncontrolled SO 2
concentrations of 200 ppm, these data suggest removal efficiencies of about
75 percent.
During these same test conditions, HC1 concentrations at the ESP inlet
ranged from 119 to 211 ppm and averaged 175 ppm at an ESP inlet temperature
3-8
-------
of 550°F, 179 nprn at an F irl t te p rat re of 4 J0L zrd flY ppm at an
ESP inlet t : :r “ Compared to a typical uncontrolleC N
conc ; on ot 500 ppm, these measurements suggest HC1 removal
et’ficic . . s of 50 to 80 percent, but no clear relationship with
temper?i u .
Nine test runs were conducted at a sorbent injection rate of 750 lb/hr
during the screening tests. The mixing chamber temperature was normal
during all of these three runs. Three runs were conducted at each of three
ESP inlet temperatures--400, 350, and 300°F. ESP inlet SO 2 concentrations
ranged from 7.2 to 77 ppm and averaged 66, 37, and 14 ppm at ESP inlet
temperatures of 400, 350, and 300°F, respectively. Compared to a typical
uncontrolled SO 2 concentration of 200 ppm. These measured concentrations
indicate removal efficiencies of 70 to 90 percent and suggest increased
removal of SO 2 as ESP inlet temperatures are decreased over this range of
temperatures.
Also with 750 lb/hr of sorbent, HC1 concentrations at the ESP inlet
ranged from 26 to 92 ppm and averaged 45, 46, and 59 ppm at 400, 350, and
300°F, respectively, at the ESP inlet. Compared to a typical uncontrolled
HC1 concentrations of 500 ppm, these HC1 measurements indicate removal
efficiencies of about 90 percent.
In Figure 3-1, SO 2 concentration at the ESP inlet is shown as functions
of limestone feed rate and ESP inlet temperature. Concentrations measured
when no sorbent was injected at Dayton are shown for comparison.
Figure 3-2, plots HC1 concentrations at the ESP inlet. Two effects are
noteworthy from these plots. First, increasing lime feed rate decreases
both SO 2 and HC1 emissions. At limestone feed rates of 750 lbs/hr, SO 2 and
HC1 emissions are roughly one-half to one-third the levels measured at 250
lbs/hr of limestone and equivalent temperatures. Second, at ESP inlet
temperatures of less than 400°F, decreasing ESP operating temperature
results in lower SO 2 emissions. No similar relationship was noted between
inlet HC1 concentrations and ESP temperature. However, data collected at
the ESP outlet during the parametric test, with a feed rate of 500 lb/hr of
sorbent, show an HC1 concentration of 120 ppm at 400°F and a concentration
of 35 ppm at 300°F.
3-9
-------
170
160
No 5o,b•flt
150 - + 2$OIb/hr
140 -i 0 5001b/hr
— A 750 Ib,hr
o 130— ________
p_ 120-
110-
100- 0
111.
200 300 400 500 600
ESP fist Temp.ratur. (°F)
170 -
A
160 -
150 - 300’F
+ 360F
140 - 0 400w
d 130- _____
,. 120- X
• 110-
100- A
90-
C +
o A
0 70- A
U) 60-
•
0
40 -
30 -
20
10 - __________
0 100 200 300 400 600 600 700 800
Sorbent Feed Rat. (lb/lw)
Figure 3-1. Inlet SO 2 concentration as a function of ESP inlet temperature
and sorbent feed rate for the FSI system at Dayton.
3-10
-------
280
260 To osor .nt
+ 250 lb/hr
240 o sooiai w 0
22 - A 750 1b/hr
0 ____
Ip 200-
E
. 180- +
a
I + +
10o
80 ’
60 A
40-
200 300 400 500 600
ESP Inlet Temperature (°F)
280 -
260 - _______
240 300P
4
0400
220- A5S0 P
‘ - 200-
180-
a A
160-
U
C
o 4-
+
100-
80
60
40-—
0 100 200 300 400 500 600 700 800
Sorbent Feed Rate (lb/hr)
FIgure 3-2. Inlet HCI concentration as a function of ESP Inlet temperature
and sorbent feed rate for the FSI system at Dayton.
3-11
-------
Simultaneous measurements of SO 2 at the ESP inlet and outlet during
parametric testing at 300 and 400°F indicate no additional SO 2 removal
occurred across the ESP with furnace sorbent injection. At the ESP inlet,
SO 2 concentrations averaged 48 and 20 ppm for the two parametric test
conditions. At the ESP outlet, the corresponding SO 2 concentrations were 57
and 21 ppm. During the screening tests conducted at an ESP inlet
temperature of 550°F (conditions 5, 6, and 7) averaged 54 ppm, or roughly 35
percent lower than the average SO 2 levels at the ESP inlet of 83 ppm.
In Table 3-5, particulate data from the parametric test with FSI at
Dayton are presented. Six test runs were conducted. Although the
temperatures and SCA’s differed significantly between the two test
conditions, no difference in performance is observed. At flue gas flows of
about 85,000 acfm and 70,000 acfm, SCA’s of 383 and 466 ft 2 /1,000 acfm,
respectively, PM removal efficiencies and outlet emissions were very
similar. Outlet PM concentrations ranged from 0.017 to 0.024 gr/dscf over
three runs and averaged 0.021 gr/dscf for Runs 10 to 12 conducted at an ESP
inlet of 400°F temperature. For Runs 13 to 15 conducted an ESP inlet
temperature of 300°F, outlet PM concentrations ranged from 0.0068
to 0.030 gr/dscf and averaged 0.022 gr/dscf. Under both sets of conditions,
average ESP removal efficiency was 98 percent. There was no apparent effect
of inlet PM on performance.
Particulate data from Dayton without furnace sorbent injection are
presented in Section 2.2.1. Outlet PM concentrations without sorbent
injection ranged from 0.0063 to 0.023 gr/dscf over the course of nine test
runs, averaging 0.011 gr/dscf. With average inlet PM concentrations
approximately twice as high during the tests with FSI, 1.13 gr/dscf with FSI
and 0.63 gr/dscf without sorbent, the outlet PM concentrations were also
approximately twice as high, averaging 0.022 gr/dscf. The PM removal
efficiency of the ESP during tests with and without FSI are roughly equal.
Metals data for six parametric test runs are presented in Table 3-6.
Removal efficiencies for arsenic, cadmium, lead, and chromium were
consistently at 96, 94, 93, and 84 percent, respectively, and did not depend
on the ESP inlet temperature. These removal efficiencies are 2 to
14 percent lower than the corresponding PM removal efficiency. Removal
3-12
-------
C .)
TABLE 3 -5. PARTICULATE DATA FOR DAYTON WITH FURNACE SORBENT INJECTION
InLet PM
ESP Inlet
Concentration
Concentration
Removal
Test
Condition
Run
Number
Temperature
C F)
FLue Gas
Flow (acfm)
(gr/dscf at
12% C0 2 )
(gr/dscf at
12% C0 2 )
Efficiency
(percent)
Combustor NormaL
10
397
86 .500
0.72
0.024
96.7
98.5
FSI 500 Lb/hr
ESP 600°F inLet
11
12
391
395
82 .900
86,400
1.37
1.49
0.017
98.9
Average
394
85,300
1.19
0.021
98.0
Combustor NormaL
13
297
72,700
1.06
0.030
97.2
97.4
FSI 500 Lb/hr
ESP 300 0 F inlet
14
15
301
296
70,600
66,900
1.07
1.25
0.028
0.0068
99.5
Average
298
70,100
1.13
0.022
98.0
-------
YULE 3-6. METALS DATA FOR DAYTON WITH FURNACE SORSENT INJECTION
Test Condition
Run
Number
ESP Inlet
Tuip 5 ratur.
C F)
Outlet PM
Concentration
(gr/d.cf at
12% C0 2 )
Inlet Concentration
(u Idac. at 7% 02)
Outlet Concentration
(uS/dacu at 7%
RYflOv
at Elf
Iciency
CX)
As
Cd
Cr Pb Hg
Hi
As
95.8
Cd Cr Pb
91.5 95.1 93.7
Hg
2.8
Ni
91.7
As
Cd
Cr
Pb Hg
NI
Combustor • Normal
FSI • 500 tb/hr
ESP - 400°F Inlet
10
I I
12
397
391
395
0.024
0.021
0.017
196
184
365
1,110
1,189
1,158
217
40
41
34,581 108
21,872 1,270
24,358 940
125
81
100
6.84
3.97
6.25
71.6
56.2
34.3
8.7
6.1
10.6
1,786 567
1,350 909
1,644 795
757
9.05
8.54
8.72
97.6
98.3
97.2
94.6
97.0
94.4
81.3
74.5
83.6
93.0
93.3
93.3
18.8
15.3
12.8
87.4
91.5
90.2
Average
394
0.021
248
1,152
99
26,939 973
102
5.69
56.0
5.6
1,593
602
9.67
95.1
89.8
68.8
86.7
27.7
81.6
Combustor • Normal
FSI • 500 (b/hr
ESP • 300°F inlet
13
14
15
297
301
296
0.030
0.028
0.0068
249
91
72
880
1,385
1,005
41
138
50
18,550 939
40.006 135
22,086 1,046
59
41
49
10.7
6.64
ND
19.2
76.1
9.5
13.1
10.3
5.0
2,180
2,111 521
230 1,005
709
8.69
13.1
10.5
92.7
100
96.0
94.5
99.1
94.5
92.5
90.3
83.9
94.7
99.0
93.5
29.1
6.9
21.3
78.8
74.0
78.1
Average
298
0.022
137
1,090
78
26,881 907
49
5.78
54.9
1,507
(A)
-------
efficiency for nickel averaged 90 percent for one test condition (400°F) and
78 percent for the other test condition (300°F). However, this difference
is probably due to the normal variation in ESP performance for nickel rather
than changes in temperature and SCA. The removal efficiencies for arsenic,
cadmium, lead, chromium, and nickel with FSI were 2, 4, 5, 10, and up to 16
percent lower, than the corresponding metals removal efficiencies measured
without sorbent injection. Thus, use of FSI may result in lower ESP
performance for arsenic, cadmium, lead, chromium, and nickel relative to
performance with no sorbent injection. Mercury removal efficiencies
averaged 12 percent for three runs at 400°F and 21 percent for three runs at
300°F. This difference may be due to temperature, but may also reflect
variability in sampling accuracy between runs. No mercury removal was
observed for the tests conducted at 400° and 550°F without sorbent
injection. Thus use of FSI may increase mercury removal compared to
operations without sorbent injection, but the removal of mercury will be
small.
CDD/CDF data from the parametric testing are presented in Table 3-7.
At an ESP inlet temperature of 400°F, outlet CDD/CDF concentrations were
between 1,018 and 1,990 ng/dscm at 7 percent 02 and averaged 1,480 ng/dscni
for the runs. The inlet CDD/CDF concentrations ranged from 22.3 to
49.1 ng/dscm and averaged 38.2 ng/dscm. At an ESP inlet temperature of
300°F, outlet CDD/CDF concentrations ranged from 352 to 1,030 ng/dscm and
averaged 659 ng/dscm for the runs. The inlet CDD/CDF concentrations ranged
from 7.4 to 23.6 ng/dscm and averaged 13.9 ng/dscm.
Both test conditions with limestone injected into the furnace showed an
increase in CDD/CDF concentrations across the ESP despite low ESP inlet
temperatures. At the ESP inlet, CDD/CDF concentrations were at most
46 ng/dscm. Due to significant amounts of PM assumed to be removed in the
quench spray chamber, these low concentrations may indicate that much of the
CDD/CDF associated with PM was removed in the quench spray chamber. The
CDD/CDF concentrations at the mixing chamber for the 300 and 400°F tests
were 1,070 and 2,450 ng/dscm at 7 percent 02. At the ESP outlet, all
CDD/CDF concentrations were at least 352 ng/dscm. These data indicate that
the level of CDD/CDF concentration at the ESP outlet depends on temperature.
3-15
-------
a..
aNM = not measured.
TABLE 3-7. CDD/CDF DATA FOR DAYTON WITH FURNACE SORBENT INJECTION
Test
Conditions
Run
Number
ESP InLet
Temp 8 rature
C F)
Mixing Chamber
CDD/CDF Concentration
(ng/dscm at 7% 02)
InLet CDD/CDF
Concentration
(ng/dscm at
7% 02)
OutLet COD/CDF
Concentration
(ng/dscm at
CDD/CDF
RemovaL
Efficiency
Combustor NormaL
FSI 500 tb/hr
ESP = 400°F inLet
10
11
12
397
391
395
NNa
MM
2,450
49.1
22.3
43.0
1,990
1,634
1,018
- 3,950
- 6,330
- 2,267
Average
394
2,450
38.2
1,480
- 4,180
Combustor = NormaL
FSI 500 Lb/hr
ESP = 300°F inLet
13
14
15
297
301
296
NM
1,240
910
8.4
23.6
10.6
1,030
596
352
-13,800
- 2,425
- 3,221
Average
298
1,070
13.9
659
- 6,490
-------
At 400°F an average of 1,550 ng/dscm of CDD/CDF was emitted from the ESP.
At 300°F, an average of 659 ng/dscm was emitted from the ESP.
In comparison, data collected at Dayton without sorbent injection at
400°F showed ESP outlet CDD/CDF emissions of about 865 ng/dscm (Section
2.2.1.4). This suggests that the use of a FSI/ESP system at Dayton did not
lower CDD/CDF emissions.
3.3 SUMMARY OF PERFORMANCE
Performance of individual units with furnace sorbent injection followed
by an ESP was discussed in Section 3.2. The data are evaluated as a whole
in this section. Section 3.3.1 evaluates acid gas performance,
Section 3.3.2 evaluates particulate performance, Section 3.3.3 evaluates
metals performance, and Section 3.3.4 evaluates CDD/CDF performance.
3.3.1 Acid Gas
As discussed in Section 3.1, SO 2 readily reacts with lime at typical
furnace temperatures (1400 to 1800°F). The removal of SO 2 depends on the
amount of lime calcined from limestone and the mixing of the lime and flue
gases in the furnace. HC1 is not reactive with lime at typical furnace
temperatures, but is removed at lower temperatures downstream of the
furnace. These differing reaction temperatures can lead to preferential
removal of SO 2 over HC1 in systems operating at below or only slightly above
stoichiometric sorbent feed rates.
Of the operating parameters that influence FSI acid gas control
performance, stoichiometric ratio appears to be the most significant. As
demonstrated at Dayton, increasing the limestone feed rate (stoichiometric
ratio) decreased SO 2 and HC1 emissions. A secondary factor affecting SO 2
emissions is the flue gas temperature at the ESP inlet. At flue gas
temperatures of less than 400°F with sufficient limestone injection, SO 2
emissions at Dayton decreased as the flue gas temperature was lowered.
There may be an effect of ESP inlet temperature on HC1 emissions, but the
data are inconclusive because of water scrubbing of HC1 during flue gas
cooling. Combustors using heat recovery to cool flue gas may not perform
similarly. Finally, based on the Dayton data, it appears that most of the
SO 2 and HC1 removal occurs prior to the ESP inlet and that little removal
occurs across the ESP.
3-17
-------
At 300°F at Dayton, SO 2 and HC1 emissions averaged 21 and 35 ppm at
7 percent 02, respectively. However, high limestone injection rates
(estimated stoichiometric ratios of 1.6-5) were used during these tests.
Testing conducted at the Alexandria MWC at an ESP inlet temperature of 360°F
and a lower lime feed rate yielded average outlet SO 2 and HC1 concentrations
of 37 and 166 ppm, respectively.
These data indicate use of furnace sorbent injection of lime or
limestone can reduce both SO 2 and HC1 emissions by 50 percent or more.
Higher levels of removal can be achieved by increasing sorbent stoichiometric
ratio and reducing flue gas temperatures to less than 400°F.
3.3.2 Particulate Matter
Both of the FSJ/ESP systems evaluated yielded outlet PM emissions
averaging about 0.025 gr/dscf. However, both systems were only designed to
achieve 0.03 gr/dscf or lower. As shown at Dayton, use of FSI can increase
both inlet and outlet PM emissions over operation without FSI at the same
facility. However, Alexandria showed no increase in outlet PM emissions
with FSI as compared to emissions without FSI.
Outlet PM emissions of less than 0.03 gr/dscf at 12 percent CO 2 can be
achieved by FSI/ESP systems applied to existing MWC’s. At new MWC’s, a
level of 0.01 gr/dscf at 12 percent is attainable, based on analyses with
SD/ESP (see Section 6.0) systems. As discussed in Section 2, however, more
ESP fields may be required to meet this level.
3.3.3 Metals
In general, metals removal by FSI/ESP systems, as demonstrated at
Dayton, appears to be consistent with PM removal. The measured PM removal
efficiency at Dayton with FSI was 98 percent. Removal of arsenic, cadmium,
and lead was consistently higher than 93 percent. Chromium removal was
84 percent. Nickel removal was erratic, with average removal efficiencies
of 78 and 90 percent at 300 and 400°F, respectively. Removal efficiencies
for all these metals would be expected to increase with more effective PM
removal. Although every run yielded lower outlet than inlet mercury
concentrations, mercury removal efficiencies were less than 30 percent for
all runs.
3-18
-------
With an FSI/ESP system able to achieve outlet PM emissions of
0.03 gr/dscf, arsenic, cadmium, and lead can be removed at greater than
90 percent efficiency. Roughly 80 percent removal efficiency of chromium
and nickel can be achieved. Limited reductions in mercury (up to
20 percent) may also be achievable.
3.3.4 CDD/CDF
The available data show wide variations in CDD/COF removal
performance. Outlet CDD/CDF emissions above 1,000 ng/dscm and below
60 ng/dscm have been measured at two separate MWC facilities with FSI/ESP
systems. At Dayton, a mass burn refractory combustor using a relatively
high limestone feed rate, average CDD/CDF emissions of 659 ng/dscm at an ESP
inlet temperature of 300°F and 1,550 ng/dscm at 400°F were measured at the
ESP outlet. Tests conducted at Dayton with and without FSI at an ESP inlet
temperature of 400°F showed similar emission levels of CDD/CDF. At
Alexandria, a mass burn waterwall cornbustor using a low lime feed rate and
an ESP inlet temperature of 360°F, low outlet CDD/CDF concentrations were
measured (<60 ng/dscm) which are similar to CDD/CDF emissions measured at
two other combustors of the same design. There are insufficient data to
determine if the use of lime rather than limestone affects performance.
Based on the similarity in outlet concentrations of CDD/CDF with and without
FSI at similar operating conditions, FSI does not significantly reduce
CDD/CDF emissions from an ESP.
3-19
-------
3.4 REFERENCES
1. Albertson, D.M. and M. L. Murphy. (Energy Products of Idaho). City of
Tacoma Steam Plant No. 2 Pilot Plant Testing and Ash Analysis Program.
Prepared for City of Tacoma, Department of Public Utilities. Tacoma,
Washington. December 1987.
2. Zurlinden, R.A., A. Winkler, and J.L. Hahn (Ogden Projects, Inc.
Environmental Test Report, Alexandria/Arlington Resources Recovery
Facility, Units 1, 2, and 3. Prepared for Ogden Martin Systems of
Alexamdria/Arlington, Inc. Alexandria, Virginia. Report No. 144B. March
9, 1988.
3. Zurlinden, R.A., H.P. Von
Inc.). Environmental Test
Recovery Facility, Units
of Al exandri a/Ar 1 i ngton,
(Revised). January 8, 1988.
Dem Fange, and J.L. Hahn (Ogden Projects,
Report, Alexandria/Arl ington Resource
1, 2, and 3. Prepared for Ogden Martin Systems
Inc. Alexandria, Virginia. Report No. 144A
4. Radian Corporation. Preliminary Data from October-November 1988 Testing
at the Montgomery County South Plant, Dayton. Ohio.
3-20
-------
4.0 DUCT SORBENT INJECTION FOLLOWED BY AN ELECTROSTATIC PRECIPITATOR
Section 4 describes the technology and performance of duct sorbent
injection (DSI) systems with an ESP for PM control. In Section 4.1, DSI/ESP
operation and design is described. Section 4.2 presents descriptions of
MWC’s equipped with DSI/ESP systems for which emissions data are available
and summarizes the available data. In Section 4.3, the performance of
DSI/ESP systems relative to the control of SO 2 , HC1, PM, metals, and CDD/CDF
emissions is discussed.
4.1 PROCESS DESCRIPTION
Duct sorbent injection is designed to control acid gas emissions.
Powdered sorbent, usually hydrated lime [ Ca(OH) 2 1, is injected into the flue
gas upstream of the particulate control device. The sorbent is generally
injected through a venturi or into a reactor vessel just upstream of the
ESP. The residence time prior to the ESP is generally one to two seconds.
The sorbent reacts rapidly with the acid gases to form salts which, in
addition to the flyash and unreacted sorbent, are removed by the ESP. Lower
system operating temperatures increase sorbent reactivity and increase
removal of acid gases as well as condensible metals and organics. At the
injection point, the flue gas temperature can range from less than 300°F to
roughly 600°F. Flue gas can be cooled upstream of the injection point by
heat exchange, such as by an economizer or air-to-air heat exchanger,
humidification, or addition of lower temperature air.
The operation and design of ESP’s following OSI is similar to the
operation of an ESP without sorbent injection. The increased particulate
loading and changed particulate characteristics due to the injected sorbent
are not expected to significantly decrease PM removal efficiency by an ESP.
However, larger ESP specific collection areas (SCA’s) may be needed to
achieve the same outlet PM emissions as systems without sorbent injection.
Design of ESP’s is described in Section 2.1.
4.2 SUMMARY OF TEST DATA
Section 4.2 presents the available emissions data for MWC facilities
with DSI/ESP systems. The only data available are for the Dayton MWC. A
4-1
-------
description of this facility and a summary and analysis of the emission data
are provided.
The effects of stoichiometric ratio, ESP inlet temperature, and inlet
acid gas concentration on acid gas removal are discussed in this section.
The effect of SCA on ESP performance for particulate removal is also
discussed. Finally, this section discusses the effects of inlet CDD/CDF
concentration and ESP inlet temperature on CDD/CDF removal.
4.2.1 Dayton 1
The Montgomery County South Incinerator plant in Dayton, Ohio, includes
three nearly identical Volund refractory-lined combustors. PM emissions are
controlled by ESP’s. A complete description of the facility is provided in
Section 2.2.1.4.
In November and December 1988, testing was conducted by EPA on Unit 3.
Tests were conducted with furnace sorbent injection, duct sorbent injection,
and without sorbent injection. A complete description of the test program
is presented in Section 2.2.1.4. Results from the duct sorbent injection
tests are reported here. Hydrated lime was injected into a vertical duct
prior to the ESP at a temperature of 350 to 400°F. The results from tests
without sorbent are presented in Section 2.2.1.4. The results from tests
with furnace sorbent injection are reported in Section 3.2.
The acid gas data are presented in Table 4-1. Sixteen test runs were
conducted at ESP inlet temperatures of 300, 350, and 400°F and sorbent
injection rates of 160, 320, and 480 lbs/hr. Runs 11 through 14 were
conducted in November as part of the test screening phase. Runs 16, 17, and
18 were conducted in December as part of the more detailed parametric
testing phase. Because the duct injector system was installed primarily to
protect the ESP from acid gas corrosion during the testing and was not
designed to achieve maximum sorbent utilization, the acid gas performance
data collected during the tests may not be indicative of performance levels
achievable by commercial Os! systems and should be used only to show trends
in performance. The relatively low acid gas concentrations at the ESP inlet
compared to other MWC’s may be due to acid gas scrubbing by the water quench
chamber located ahead of the ESP inlet sampling point.
4-2
-------
TABLE 4 -1. ACID GAS DATA FOR DAYTON WITH DSI
Test
Condition
Run
Number
ESP Inlet
Temperature
( 0 F)
Stoichiometric
Ratioa
Acid Gas Concentration
(ppmv. dry at 7% O, __.
Acid Gas Removal
Efficiency CX)
Inlet OutL t
so 2 b HcLb $02 HCL
SO 2 HCL
Combustor = Normal
DSI = 160 lb/hr
ESP = 400°F inlet
hA
11B
11C
400
400
400
NMC
4.9
4.7
NM 141 NM 29.7
106 204 52.9 37.9
123 217 56.7 35.6
-- 78.9
50.1 81.4
53.9 83.6
Average
400
4.8
114 187 54.8 34.4
52.0 81.3
Combustor = Normal
DSI 480 Lb/hr
ESP = 400 0 F inlet
12A
12B
12C
400
400
400
13
11
10
115 142 32.2 16.3
141 211 34.8 32.2
130 191 36.8 19.7
72.0 88.5
75.3 84.7
71.7 89.7
Average
400
11
129 181 34.6 22.8
73.0 87.6
Combustor = NormaL
DSI = 160 Lb/hr
ESP = 350°F inLet
13A
136
13C
350
350
350
4.5
4.6
6.6
108 159 65.6 50.8
117 177 56.9 34.9
138 263 54.6 34.9
39.3 68.1
51.4 80.3
60.4 86.7
Average
350
5.2
121 200 59.0 40.2
50.4 78.4
Combustor = Normal
DSI = 480 Lb/hr
ESP = 350°F inLet
13D
350
15
111 126 35.7 17.4
67.8 86.2
Combustor = Normal
OSI = 320 Lb/hr
ESP = 300°F inLet
16
17
18
306
305
306
9.0
9.6
7.0
93 119 30.0 12.0
122 128 46.8 6.8
139 86 40.0 8.0
67.9 89.9
61.5 94.7
71.2 90.7
Average
306
8.5
119 111 38.9 8.9
66.9 91.8
Combustor = NormaL
DSI 480 Lb/hr
ESP = 300°F inLet
14A
14B
14C
300
300
300
30
28
28
67 93 38.3 11.7
74 90 43.6 13.0
76 99 45.2 10.9
42.8 87.4
41.2 85.6
40.3 89.0
Average
300
29
72 94 42.3 11.9
41.5 87.3
acaLculated based on measured HC( and SO 2 concentrations at the ESP inlet.
bAcid gas concentrations at ESP inLet may reflect partial scrubbing of $02 and HC1 by water sprays in
quench chamber located between the combustor and ESP inLet.
CNN = Not measured.
dHdrated Lime feed rate = 250 Lb/hr.
-------
During screening test Runs hA, 11B, 11C, 13A, 13B, and 13C, the
sorbent feed rate was 160 lbs/hr. The ESP inlet temperature during these
runs was set at 400°F for Runs hA, B, and C and at 350°F for Runs 13A, B,
and C. SO 2 concentrations at the ESP outlet during the runs conducted at
400°F ranged from 53 to 57 ppm at 7 percent 02 and averaged 55 ppm while the
runs conducted at 350°F ranged from 55 to 66 ppm and averaged 59 ppm.
Average SO 2 removal efficiencies across the ESP at both temperatures were
relatively consistent at roughly 50 percent. Outlet HC1 concentrations
during the runs at 400°F ranged from 30 to 38 ppm and averaged 34 ppm while
the runs at 350°F ranged from 35 to 51 ppm and averaged 40 ppm. Across the
ESP, HG] removal efficiencies at both temperatures were also consistent at
roughly 80 percent.
During screening test Runs 12A, 12B, 12C, 13D, 14A, 14B, and 14C, the
sorbent feed rate was 480 lbs/hr. The ESP inlet temperature during these
runs was set at 400°F for Runs 12A, B, and C; 350°F for Run 13D; and 300°F
for Runs 14A, B, and C. SO 2 concentrations at the ESP outlet during the
runs conducted at 400°F ranged from 32 to 37 ppm and averaged 35 ppm.
Outlet SO 2 concentrations during the single run conducted at 350°F was
36 ppm and during the three runs conducted at 300°F ranged from 38 to 45 ppm
and averaged 42 ppm. SO 2 removal efficiencies across the ESP were
relatively consistent at both 400°F and 350°F, ranging from 68 to 75 percent
and averaging 70 percent. At 300°F, the SO 2 removal efficiency across the
ESP ranged from 40 to 43 percent and averaged 41 percent. In addition to
the somewhat higher outlet SO 2 concentration during the runs at 300°F, the
reduced SO 2 removal efficiency at 300 0 F also reflects a lower inlet SO 2
level of 67 to 76 ppm versus inlet concentrations of 111 to 141 ppm during
the 350 and 400°F runs. The reduced inlet SO 2 concentration and lower
removal efficiency at 300°F may have been partially due to problems with
clumping of hydrated lime around the sorbent injector nozzles and plugging
of the sorbent feed system at these lower temperatures. Outlet HC1
concentrations during the runs at 400°F ranged from 16 to 32 ppm and
averaged 23 ppm and was 17 ppm during the single run at 350°F. Outlet HC1
concentrations during the runs at 300 F ranged from 11 to 13 ppm and
averaged 12 ppm. HC1 removal efficiencies across the ESP during all of
4-4
-------
these runs were relatively consistent at all of the temperatures, ranging
from 85 to 90 percent and averaging 87 percent.
During parametric test Runs 16, 17, and 18, the ESP inlet temperature
was set at 300°F although the actual averages during each run were slightly
higher. The sorbent feed rate was 320 lbs/hr during Run 16, but was reduced
to 250 lbs/hr during Runs 17 and 18 due to the sorbent clumping and
injector plugging problems mentioned above, SO 2 concentrations at the ESP
outlet were 30 ppm during Run 16, and 47 and 40 ppm during Runs 17 and 18.
Average SO 2 removal efficiencies across the ESP during all three runs ranged
from 62 to 71 percent and did not show any clear differences as a function
of lime feed rates. Outlet HC1 concentrations were 12 ppm during Run 16,
7 ppm during Run 17, and 8 ppm during Run 18. Average HC1 removal
efficiencies across the ESP during all three runs ranged from 90 to 95
percent and did not show any clear differences as a function of lime feed
rates.
Taken in aggregate, the Dayton data indicate that increasing the
sorbent feed rate will reduce outlet acid gas concentrations. However, over
the range of temperatures examined, ESP operating temperature had relatively
little effect on outlet acid gas emissions or removal efficiency.
Overall SO 2 and HC1 removal efficiencies are shown in Figures 4-1 and
4-2, respectively, as a function of stoichiometric ratio. Both SO 2 and HC1
removal efficiency generally increase as the stoichiometric ratio increases.
The lower SO 2 removal efficiencies measured at high stoichiometric feed
rates during Runs 14A, B, and C (not shown on either figure) are a result of
unusually low inlet SO 2 concentrations and possible sorbent injector
problems encountered at low temperatures, high lime feed rates, and high
flue gas humidity. The outlet SO 2 concentrations measured during these runs
are similar to those measured during the other tests with stoichiometric
ratios above 7. These data suggest that increasing the stoichiometric ratio
above about 7 did not provide additional acid gas removal because of
constraints on lime and flue gas mixing associated with the injection system
used during the Dayton testing.
The removal efficiency and outlet emission data for SO 2 and HC1 do not
show any significant relationship with ESP inlet temperature. At the
highest sorbent feed rate, outlet SO 2 emissions remain consistently near 35
4-5
-------
80 -
o
0 0 0
70 -
C
U
0
60- +
0
U
‘I-
55-
0
0
E +
• 50- 0
U)
45 - [ 1] 400°F
4 350°F
40- + 0 300°F
35 — I I I I I i —————r———— ——i—
4 6 8 10 12 14 16
Stoichiometric Ratio
Figure 4-1. SO 2 removal efficiency as a function of stoichiometric
ratio at the Dayton DSI/ESP system.
-------
96 - -—----—--- ----- —-—---- — ---—— — --——-
K?
94 -
92
90
-S
4- . a
88 -
U
+
86- +
0.
U
o 84-
C U
0
I __
I + 350°F
300°F
72
70
68 - -- -+j — I I I
4 6 8 10 12 14 16
Stoichiometric Ratio
Figure 4-2. HCI removal efficiency as a function of stoichiometric
ratio at the Dayton DSI/ESP system.
-------
to 40 ppm for temperatures of 300 to 400°F. At these same feed rates, HC1
emissions were between 10 and 20 ppm for all but one run. There is no
apparent effect of inlet SO 2 and HC1 concentrations on performance.
Particulate data are presented in Table 4-2 for the three parametric
test runs. Outlet PM concentrations ranged from 0.0018 to 0.0039 gr/dscf at
12 percent CO 2 and averaged 0.0032 gr/dscf. The corresponding PM removal
efficiencies were between 99.2 and 99.7 percent and averaged 99.4 percent.
Although the flue gas flow rate was significantly higher for Run 16, giving
a lower SCA, PM removal efficiency was highest during that run. The outlet
PM emissions from the tests of DSJ are similar to the results from tests
with no sorbent injection, which demonstrated 0.0063 gr/dscf at 12 percent
CO 2 . The tests with furnace sorbent injection yielded much higher outlet PM
emissions of about 0.020 gr/dscf.
Metals data for the three runs conducted are presented in Table 4-3.
Average outlet metals concentrations were not detected for arsenic,
4 ug/dscm for chromium and nickel, 11 ug/dscm for cadmium, 360 ug/dscm for
lead, and 490 ug/dscm for mercury. Removal efficiencies averaged greater
than 98.9 percent for arsenic, cadmium, and lead. Chromium and nickel
removal efficiencies were 97.8 and 96.7 percent, respectively. The removal
efficiency for mercury was measured at 25 percent.
CDD/CDF data, which were collected simultaneously with the PM and
metals data, are presented in Table 4-4. Outlet COD/COF concentrations
ranged from 13.7 to 132 ng/dscm at 7 percent 02 and averaged 57.2 ng/dscm.
Increases in CDD/CDF concentrations were observed across the ESP, with the
outlet concentrations being 2.6 to 380 times higher than those measured at
the ESP inlet. Inlet CDD/CDF concentrations ranged from 0.31 to 8.4 ng/dscm
and averaged 5.3 ng/dscm. The low inlet COD/CDF concentrations may have
been caused by the water spray quench chamber removing a significant amount
of particulate as discussed in Section 3.2. At the mixing chamber, the
CDD/CDF concentration during Run 16 was 5,350 ng/dscm. There are
insufficient variations in ESP inlet temperature, inlet CDD/CDF
concentration, and inlet PM to allow analyses of the effects of these
parameters relative to CDD/CDF removal.
The outlet CDD/CDF emissions with DSI, 57 ng/dscm, are significantly
lower than measured during tests with no sorbent addition in
4-8
-------
TABLE 4-2. PARTICULATE DATA FOR DAYTON WITH OS!
Test
Condition
Run
Number
ESP Inlet
Temp 8 rature
( F)
Flue Gas
Flow
(acfm)
Inlet PM
Concentration
(gr/dscf at
12% CU 2 )
Outlet PM
Concentration
(gr/dscf at
12% C0 2 )
PM
Removal
Efficiency
(%)
Combustor = No 9 al
DSI = 320 lb/hr
ESP = 300°F inlet
16
17
306
305
83,700
66,200
0.604
0.596
0.0018
0.0039
99.7
99.4
18
306
63,200
0.481
0.0039
99.2
Average
306
71,000
0.560
0.0032
99.4
aHydrated lime feed rate changed to 250 lb/hr for Runs 17 and 18.
-------
a
Nydr.ted I.e feed r.t. for runs 17 and 1$ was 230 tb/hr.
bNO Not detected.
TAILE 4-3. NETALI DATA FOE DAYTON UITN DII
Test Condition
Nun
N,. sr
NIP Inlet
T. .r 1 tur.
( F)
Outlet PN
Concentration
(gr/d.cf at
12% C0 2 )
Inlet
Concentration
at 7%
Qutl.t
( IdsC.
Concentration
• °2
Cr Pb iig
NI
Re.ovat Efficiency
Cd Cr Pb NE
As
Cd Cr
02)
Pb
u
ii
as
Cd
As
Ni
Co uetor a Nor sI
DII • 320 lb/hr
NIP — 300°F inlet
16
17
18
306
305
306
0.0010
0.0039
0.0039
263
290
76
1 312
1 660
1 5$2
170
223
183
35 .6$
33 $70
39 .273
782
679
607
59
101
ITO
0 b
NO
ND
7.6
13.3
11.5
3.19
4.09
4.20
303 462
375 490
403 522
ND
5.3
6.7
100
100
100
100
99.4
99.1
99.2
99.2
97.9
98.0
97.5
97.8
99.0
98.8
98.9
98.9
36.4
22.8
17.7
25.0
100
94.4
95.7
96.7
Avera .
306
0.0032
210
1 S18
192
34863
716
110
ND
10.8
3.83
361 491
-------
TABLE 4-4. CDD/CDF DATA FOR DAYTON WITH DUCT SORBENT INJECTION
InLet
Mixing Chamber
InLet CDD/CDF
Concentration
OutLet CDD/CDF
Concentration
CDD/CDF
RemovaL
CDDICDF Concentration
(ng/dscm at
(ng/dscm at
Efficiency
Test
Conditions
Run
Number
Temperature
( F)
(ng/dscm at 7% 02)
% 02)
7% 02)
Combustor = Wor aL
DSI = 320 Lb/hr
ESP = 300°F inLet
16
17
18
306
305
306
5 ,3S
NM
NM
7.17
8.44
0.31
26.0
132
13.7
- 263
-14,600
- 4,219
-
Average
306
5,350
5.31
57.2
6,360
I aHydrated lime feed rate for Runs 17 and 18 was 250 Lb/hr.
— b
— NM Not measured.
-------
Section 2.2.1.4, 866 to 17,100 ng/dscm and furnace sorbent injection in
Section 3.2.2, 673 ng/dscm at an ESP inlet temperature of 300°F and
1,480 ng/dscm at an ESP inlet temperature of 400°F. At the same ESP inlet
temperature, 300°F, CDD/CDF emissions with DSI were an order of magnitude
lower than with FSI.
4.3 SUMMARY OF PERFORMANCE
Performance of the Dayton DSI/ESP system was evaluated in Section 4.2.
Because the only available data are from Dayton, only the conclusions
reached from the analyses of this single MWC are presented in this section.
Section 4.3.1 evaluates acid gas performance. Section 4.3.2 presents the
evaluations regarding PM performance. Section 4.3.3 presents the discussion
for metals performance, and Section 4.3.4 summarizes CDD/CDF performance.
4.3.1 Acid Gas
At Dayton, with a non-commercial DSI/ESP system, outlet HC1 emissions
were shown to decrease with increasing stoichiometric ratio. HC1 removal
efficiencies of greater than 68 percent were demonstrated across the ESP
during all test runs, and were between 85 and 95 percent during tests
conducted at ESP inlet temperatures of 300 to 310°F. HC1 concentrations at
the ESP outlet during all tests were less than 60 ppm at 7 percent 02. HC1
removal efficiencies across the entire quench spray chamber and DSI/ESP
system may have exceeded 90 percent. Based on these data, 80 percent Nd
removal efficiency is achievable with DSI/ESP systems at temperatures of
400°F or less.
Removal efficiency for SO 2 and outlet $02 emissions at Dayton were
relatively independent of temperature. Analysis of the effect of
stoichiometric ratio showed that performance can be enhanced by increasing
stoichiometric ratio. At ESP inlet temperatures of 300 to 400°F and 160 to
480 lb/hr of lime injection, SO 2 removal efficiencies averaged greater than
50 percent. At only one condition, ESP inlet temperature of 300°F, was an
SO 2 removal efficiency of less than 50 percent obtained (42 percent). The
inlet SO 2 concentrations were unusually low during this condition, and the
outlet SO 2 emissions were similar to the results demonstrated at other test
conditions. This suggests that lime and flue gas mixing constraints
4-12
-------
prevented the system from getting lower SO 2 emissions. Thus, at least
50 percent SO 2 removal efficiency is achievable by a DSI/ESP system.
4.3.2 Particulate Matter
At Dayton, with the DSI/ESP system, outlet PM emission were the same as
measured without any sorbent injection. Thus, the outlet PM emissions of an
existing ESP were not adversely affected when using DSI. Based on the
analysis of ESP performance presented in Section 2.0, a DSI retrofit system
can achieve outlet PM concentrations of 0.03 gr/dscf at 12 percent CO 2 .
4.3.3 Metals
The metals removal efficiencies demonstrated at Dayton with the DSI/ESP
system are the same as was measured at Dayton without sorbent injection.
Therefore, removal efficiencies of greater than 90 percent for arsenic,
cadmium, chromium, lead, and nickel are achievable with a DSI/ESP achieving
0.03 gr/dscf of PM. No mercury removal is expected to occur with a DSI/ESP
system.
4.3.4 COD/COF
The average CDD/CDF emissions at Dayton with DSI/ESP and ESP inlet
temperatures of 300°F were 93 to 99.7 percent lower than measured without
sorbent injection and ESP inlet temperatures of 400 to 575°F. This
reduction cannot be attributed entirely to use of DSI, however, since
emissions of CDD/CDF appear to be dependent on temperature. As shown during
the Dayton tests without sorbent injection, CDD/CDF emissions at 400°F were
more than an order of magnitude lower than at 525°F to 575°F. Therefore,
higher CDD/CDF emissions will probably result with the DSI/ESP system at 350
to 450°F than with the system operating at 300°F.
When compared to CDD/CDF emissions from the Dayton MWC with furnace
sorbent injection at 300°F, the emissions with DSI are 90 percent lower.
Assuming FSI does not affect outlet CDD/CDF emissions when compared to
emissions without sorbent njection, (see Section 3,3.4), DSI can achieve
a 90 percent decrease in COD/COF emissions when compared to emissions at the
same temperature without duct sorbent injection.
4-13
-------
4.4 REFERENCES
1. Radian Corporation. Preliminary Data for October - November 1988
testing at the Montgomery County South Plant, Dayton, Ohio.
4-14
-------
5.0 DUCT SORBENT INJECTION FOLLOWED BY A FABRIC FILTER
Section 5 describes the technology and performance of duct sorbent
injection (DSI) with a fabric filter for PM control. In Section 5.1, DSI/FF
operation and design is described. Section 5.2 describes the MWC facilities
equipped with DSI/FF systems with emissions data and summarizes the
available data. In Section 5.3, performance of DSI/FF systems relative to
the control of acid gas, PM, metals, and CDD/CDF emissions is described.
5.1 PROCESS DESCRIPTION
Duct sorbent injection technology with a fabric filter is very similar
to duct sorbent injection with an ESP described in Section 4.1. However,
because of the performance characteristics of a fabric filter, greater acid
gas and organics removal is achievable than with an ESP.
Fabric filters remove particulate matter by passing flue gas through a
porous barrier (filter bag). Filter bags are arranged vertically or
horizontally into a number of defined compartments. Each compartment can
operate independently of the other compartments. As the flue gas flows
through the filter bags, particulate is collected on the filter surface,
mainly through inertial impaction. The collected particulate builds up on
the bag, forming a filter cake. The presence of unreacted sorbent in the
collected particulate provides acid gas removal. Once excessive pressure
drop across the filter cake is reached for the bags in a given compartment,
that compartment is generally taken off-line, mechanically cleaned, and then
placed back on-line.
Fabric filters are generally differentiated by cleaning mechanisms.
Two main filter cleaning mechanisms are used: reverse-air and pulse-jet.
In a reverse-air fabric filter, flue gas flows through unsupported filter
bags, leaving the particulate on the inside of the bags. The bags are
cleaned by blowing air through the filter in the opposite direction of the
flue gas flow, causing the filter bag to collapse. The filter cake falls
off and is collected in the hopper located below the filter bags. In a
pulse-jet fabric filter, flue gas flows through supported filter bags,
leaving particulate on the outside of the bags. Compressed air is
5-1
-------
introduced at the top of the bag, causing the bag to expand and the filter
cake to fall off.
Fabric filters are capable of achieving very high particulate removal
efficiencies. The PM control effectiveness of the fabric filter depends on
flue gas and filter characteristics, including 1) the air-to-cloth ratio
(expressed as acfm per square foot), and 2) the filter cleaning mechanism.
The air-to-cloth ratio is optimized to give increased surface area without
excess pressure drop. Collection efficiency increases as the air-to-cloth
ratio decreases. Because pulse-jet fabric filters remove more filter cake
than reverse-air units during the cleaning cycle, pulse-jet filters can be
operated at higher air-to-cloth ratios with equal removal efficiencies. 1
The increased particulate loading and changed particulate
characteristics caused by DSI are not expected to significantly decrease the
effectiveness of PM removal by a fabric filter. Larger surface area may be
necessary, though, to ensure the same outlet PM emissions as achieved
without sorbent injection.
5.2 SUMMARY OF TEST DATA
Section 5.2 presents the available emissions data for MWC’s with DSI
systems. A description of each facility and a summary and analysis of the
emission data are provided for each facility.
The effects of stoichiometric ratio, fabric filter inlet temperature,
and uncontrolled acid gas concentrations on acid gas removal are discussed
in each section. The effects of air-to-cloth ratio on particulate removal
is also discussed. The effects of inlet CDD/CDF concentration and fabric
filter inlet temperature on CDD/CDF removal is discussed as well.
5.2.1 Claremont 2 ’ 3
The SES Claremont facility in Claremont, New Hampshire, includes two
identical Von Roll reciprocating grate combustor trains, each designed to
combust 100 tons/day of MSW. Flue gas cooling is achieved by a boiler and
economizer. Following heat recovery, the flue gas flows through a spark
arrester where hot, glowing particles drop out. Located upstream of the
spark arrester is an emergency water quench system to provide extra flue gas
5-2
-------
cooling if necessary. Additional flue gas cooling is accomplished through
the addition of ambient air through an automatically adjustable damper
located after the spark arrester. Downstream of the damper, dry hydrated
lime is injected countercurrently into the flue gas in a venturi duct. The
lime is fluidized with air in the feed duct. The typical lime feed rate is
190 to 220 lb/hr, with a maximum of 346 lb/hr. The design HC1 removal
efficiency is 90 percent.
Immediately following hydrated lime injection, flue gas enters a
Wheelabrator fabric filter at a design flow of 25,000 acfm at 450°F. The
fabric filter has three compartments, each with 225 woven fiberglass bags
with an acid-resistant finish. The bags are cleaned with a pulse-jet system
in which the bags are kept on-line while cleaned. Flue gas exiting the
fabric filter is exhausted through separate 150-foot high stacks.
Compliance testing was conducted at the facility in May and July 1987.
During the May test, flue gas was sampled at the FF outlet and analyzed for
PM, SO 2 , HC1, and NON. Additional simultaneous samples were taken from a
single point at the FF inlet and analyzed for HC1. During the July test,
flue gas at the FF outlet was analyzed for CDD/CDF. The combustor and
DSI/FF system were operated under normal conditions during both test
periods.
Acid gas data are presented in Table 5-1. Nine test runs were
conducted at a FF inlet temperature of 450°F. Six of the test runs, three
on each unit, were performed under a lime feed rate of approximately 190
lb/hr. The other three test runs were performed on Unit 1 at an elevated
lime feed rate of 222 lb/hr. Outlet SO 2 concentrations at a lime feed rate
of 190 lb/hr ranged from 38.1 to 337 ppm at 7 percent 02 averaging 231 ppm
for Unit 1 and 60 ppm for Unit 2. The inlet SO 2 concentration was not
measured. Outlet HC1 concentrations ranged from 7.8 to 176 ppm at 7 percent
02 averaging 104 ppm for Unit 1 and 37 ppm for Unit 2. Removal efficiencies
for HC1 averaged 88 percent for Unit 1 and 94 percent for Unit 2. The inlet
concentration of HC1 for Unit 1 averaged 140 ppm higher over three runs than
Unit 2, which may have led to the higher outlet acid concentrations for Unit
1 and lower HC1 removal efficiency.
5-3
-------
TABLE 5-1. ACID GAS DATA FOR CLAREMONT
aRUfl Number contains the unit number followed
biemperature estim 8 ted from measured value at
fabric fiLter (10 F).
by the run number.
the stack and an assumed temperature drop across the
C
Test
Condition
Run a
Number
FF InLet
Temp 8 ra ure
C F)
Lime
Feed Rate
(Lb/hr)
InLet HC(
Concentrationa
(ppinv, dry
at 7% 02)
Concentration
(ppinv, dry at
O )
SO 2 Nd
HCL
Removal
Efficiency
(%)
Combustor NormaL
DSI/FF z NormaL
1-1
1-2
1-3
451
452
451
194
194
194
519
1,070
774
197
337
160
57.6
176
78.5
88.9
83.6
89.9
Average (Unit 1)
451
194
788
231
104
87.5
2-1
2-2
2-3
41.5
452
448
185
185
185
953
471
502
79.8
62.5
38.1
44.7
35.0
30.1
95.3
92.6
94.0
Average (Unit 2)
448
185
642
60.1
36.6
94.0
Combustor Normal
DSI/FF = High Lime
feed
1-4
1-5
1-6
448
453
454
222
222
222
462
459
424
31.4
50.8
31.6
7.8
33.9
29.4
98.3
92.6
93.1
Average
452
222
448
37.9
23.7
94.7
-------
The effect of lime feed rate was assessed from the data for Unit 1. By
increasing the lime feed rate from 194 to 222 lb/hr, the HC1 removal
efficiency at Unit 1 increased by 7 percent to 94 percent and HC1 and SO 2
emissions decreased by about 80 percent to 24 and 38 ppm, respectively.
However, the HC1 removal efficiency for Unit 2 at a lime feed rate of 185
lb/hr was similar to the HC1 removal observed at Unit 1 with a lime feed
rate of 222 lb/hr despite an inlet HC1 concentration which was 200 ppm
higher on average. Because of the widely varying inlet HC1 concentrations
and unknown inlet SO 2 concentration, conclusions about the effect of lime
feed rate cannot be made. Nevertheless, over 85 percent HC1 removal
efficiency was achieved.
In Table 5-2, particulate data are presented. Six test runs were
conducted; three on each unit. The outlet PM concentrations at Unit 1
ranged from 0.0095 to 0.012 at 12 percent CO 2 and averaged 0.011 gr/dscf.
Outlet PM concentrations for Unit 2 were between 0.0027 and 0.0053 gr/dscf
at 12 percent CO 2 and averaged 0.0043 gr/dscf. The flow rates at Unit 1
were approximately 3,500 acfm higher than at Unit 2, yielding 13 percent
higher air-to-cloth ratios. This corresponds with the higher outlet PM
concentrations measured for Unit 1. The fabric filter at Claremont
operating at 440°F can achieve outlet PM emissions of 0.01 gr/dscf at
12 percent CO 2 , with hydrated lime injection into the duct.
Table 5-3 presents the CDD/CDF data from Claremont. Four test runs
were conducted on Unit 1 at a FF inlet temperature of 437 0 F and three on
Unit 2 at 473°F. At Unit 1, outlet CDD/CDF concentrations ranged from
21.4 to 45.8 ng/dscm at 7 percent 02 and averaged 37.6 ng/dscm. At Unit 2,
outlet CDD/CDF concentrations ranged from 24.4 to 39.9 ng/dscm and averaged
32.3 ng/dscm. Although the FF inlet temperatures for Unit 2 were about 36°F
higher than at Unit 1, there is no difference between CDD/CDF emissions.
5.2.2 Dutchess County 4
The Dutchess County Resource Recovery facility in Poughkeepsie, New
York, consists of two identical Westinghouse-O’Connor rotary waterwall mass
burn combustor trains, each designed to combust 250 tons per day of MSW.
Combustion gases exiting each combustor pass through heat recovery sections
5-5
-------
TABLE 5-2. PARTICULATE DATA FOR CLAREMONT
Test
Condition
Run
Number
a
FF Inlet b
Temp 8 rature
( F)
Flue Gas
Flow
(acfm)
Outlet PM
Concentration
(gr/dscf at
12% C0 2 )
Combustor = Normal
DSI/FF = Normal
1-1
1-2
1-3
440
441
440
31,570
29,300
29,110
0.011
0.012
0.0095
Average (Unit 1)
440
29,993
0.011
2-1
2-2
2-3
433
441
437
24,870
27,390
27,130
0.0048
0.0053
0.0027
Average (Unit 2)
437
26,463
0.0043
aRUfl Number contains the unit number followed by the run number.
biemperature estimated from measured value at he stack and an assumed
temperature drop across the fabric filter (10 F).
5-6
-------
TABLE 5-3. CDD/CDF DATA FOR CLAREMONT
FF
Inlet
Outlet
CDD/CDF
Test
Condition
Run
Number
a
Temp 8 ra ure
( F)
Concentration
(ng/dscrn at 7%
02)
Combustor = Normal
1-1
437
39•4
DSI/FF = Normal
1-2
1-3
1-4
448
430
431
43.9
21.4
45.8
Average (Unit 1)
437
37.6
2-2
2-3
2-4
478
472
469
39.9
32.5
24.4
Average (Unit 2)
473
32.3
aRUfl Number contains the unit number followed by the run number.
biemperature estimated from measured value at he stack and an assumed
temperature drop across the fabric filter (10 F).
5-7
-------
followed by a dry sorbent injection/fabric filter pollution control system.
Following a spark arrestor, hydrated lime and Tesisorb are injected into
the flue gas through a Teller-designed venturi system. The lime injection
rate is manually maintained at 125 to 150 lb/hr, but optimization tests are
being planned which may alter this injection rate. The flue gas temperature
at the injection point is typically about 400°F. Immediately following
sorbent injection, flue gas enters a Zurn reverse-air fabric filter with
fiberglass bags. The operating air-to-cloth ratio of the FF is 2 acfm/ft 2 .
The permitted outlet PM concentration for each unit is 0.015 gr/dscf at 12
percent CO 2 . Flue gases from both units are exhausted through separate
flues in a common stack.
Emission compliance tests were initially conducted at the Dutchess
County RRF during January 31 through February 17, 1989. Additional testing
was conducted on March 15 and 16, and May 24 and 25, 1989. During the
January/February test, continuous SO 2 measurements were taken at the inlet
and outlet to the pollution control system. In addition, flue gas at the FF
outlet was analyzed for PM, HC1, CDD/CDF, HF, NOR, metals (arsenic,
beryllium, cadmium, chromium, lead, mercury, and nickel), and additional
organic compounds at both units. Three runs were conducted for each
pollutant. During the March test conducted to document improved PM removal,
flue gas at the FF outlet was analyzed for PM and HC1 (Unit 2 only) during
three runs and by continuous monitor for SO 2 and CO on both units. The May
test was conducted to demonstrate compliance of the facility with respect to
PM (Unit 2 only) and CO.
In Table 5-4, HC1 data are presented for all the HC1 runs. Outlet HC1
data from the three test runs in February at Unit 1 ranged from 2.4 to 70
ppm at 7 percent 02, and averaged 30 ppm. At Unit 2 in February, four test
runs at the outlet yielded concentrations ranging from 23 to 422 ppm,
averaging 183 ppm. March testing of Unit 2 showed outlet HC1 concentrations
of 70 to 473 ppm for 6 runs and an average concentration of 200 ppm.
Outlet HC1 concentrations were lower for Unit 1 than Unit 2. The
high HC1 values at Unit 2 may be due to clogging of the screw feeder for the
hydrated lime and Tesisorb . Because the feeder is manually operated, there
5-8
-------
TABLE 5-4. HC1 DATA FOR DUTCHESS COUNTY
Test
Conditions
Run
Number
a
FF Inlet
Temp 8 ra ure
( F)
Co
(ppm,
Outlet HG]
ncentration
dry at 7%
02)
Combustor = Normal
1-1
379
70
DSI/FF = Normal
1-2
380
18
(2/89 tests)
1-3
383
2
Average (Unit 1)
381
30
Combustor = Normal
2-1
354
209
DSI/FF = Normal
2-2
358
422
(2/89 tests)
2-3
2-4
377
365
78
23
Average (Unit 2)
364
183
Combustor = Normal
2-1
386
124
DSI/FF = Normal
2-2
399
70
(3/89 tests)
2-3
2-4
2-5
2-6
380
363
385
377
80
200
250
473
Average (Unit 2)
382
200
aRun number contains the unit number followed by the run number.
bTeerature estimated f om a measured value at the stack and an assumed
temperature drop of (10 F) across the fabric filter.
5-9
-------
TABLE 5-5. SO 2 CEM DATA FOR DUTCHESS COUNTY
Test
Condition
Unit
Number Date
FF ntet
Temperature
Timea ( F)
SO Concentration
(p m dry at 7X O,j
InLet 0utLe
so 2
RemovaL
Efficiency
(percent)
Cornbustor = NormaL 1 1/31 930 450 129 66 48.8
DSI/FF NormaL 1030 460 103 94 8.7
(2/89 tests) 1210 430 142 100 29.6
1300 390 109 85 22.0
1400 350 116 113 2.6
1500 350 114 69 39.5
1600 350 138 151 -9.4
1700 360 167 137 18.0
1800 360 133 126 5.3
1900 370 112 158 -41.1
2000 370 103 118 -14.6
2/2 1410 43 188 198 -5.3
1500 NM 70 27 61.4
1600 450 72 26 63.9
Average (Unit 1 , 2/89) 390 121 105 16.4
Combustor = NormaL 2 2/1 1330 MM 118 81 31.4
DS I/FF = Normet 1430 NM 96 62 35.4
(2/89 tests) 1530 NM 87 86 1.2
1820 MM 133 184 -38.4
2/2 950 MM 188 207 -10.1
1120 NM 206 120 41.8
Average (Unit 2, 2/89) MM 138 123 10.2
Combustor = NormaL 1 3/16 1000 MM NM 129
DSI/FF = NormaL 1100 NM NM 104
(3/89 tests) 1200 NM NM 107
NM MM 111
Average (Unit 1, 3/89) MM NM 105
Combustor = NormaL 2 3/15 1700 NM NM 49
DSI/FF = NormaL 1800 NM NM 106
(3/89 tests) 1900 NM NM 128
3/16 1420 MM NM 144
1520 NM NM 109
1620 MM NM 282
Average (Unit 2, 3/89) NM MM 136
astarting time of hour for 1-hour average reported.
bNM = not measured.
5-10
-------
is no means of knowing when the feeder clogs, which is reported to have
happened occasionally.
Table 5-5 presents a summary of all the SO 2 data collected. Each SO 2
data point is a one-hour average of the one-minute averages collected with
the continuous emission monitors. Outlet SO 2 concentrations ‘anged from 26
to 198 ppm at 7 percent 02 at Unit 1 for 14 one-hour averages collected over
two days in February, and averaged 105 ppm. Simultaneous inlet SO 2
concentrations during the February tests at Unit 1 ranged from 70 to 188 ppm
and averaged 121 ppm. At the outlet of Unit 1 during the March tests, three
one-hour averages of 129, 104, and 101 ppm were reported, for an average
concentration of 111 ppm. No inlet data were collected during the March
tests. Removal efficiencies for SO 2 during the February tests ranged from
-42 to 64 percent and averaged 16 percent. There is no certain cause of the
negative removal efficiencies, but they may be due to differences in the
instruments used to collect the data.
At Unit 2, outlet SO 2 concentrations ranged from 62 to 207 ppm for
6 one-hour averages during February and averaged 123 ppm. At the inlet of
Unit 2 during February, So 2 concentrations ranged from 87 to 206 ppm for the
same period and averaged 138 ppm. The SO 2 removal efficiency ranged from
-38 to 42 percent and averaged 10 percent. During March, the outlet SO 2
concentration ranged from 49 to 282 ppm for 6 one-hour averages and averaged
136 ppm.
SO 2 performance was similar for the two units. Although the average
outlet emissions from Unit 1 were about 20 ppm lower than from Unit 2, the
inlet concentrations were also lower by about 20 ppm. SO 2 removal
efficiencies were similar for the two units. Temperature at the FE inlet
did not affect performance for SO 2 , as emissions were widely scattered,
irrespective of temperature. Figure 5-1 graphs outlet SO 2 concentrations as
a function of inlet SO 2 concentration. While there is a wide scatter in the
data, a trend of increasing outlet SO 2 concentration with increasing inlet
SO 2 concentration can be observed. For Unit 1, the lowest inlet SO 2
concentration, 70 ppm, corresponded to the lowest outlet SO 2 concentration,
30 ppm, and the highest inlet SO 2 concentration, 188 ppm, was also
coincident with the highest outlet SO 2 concentration, 198 ppm.
5-11
-------
210
200
190
180
170-
ON
160---
p .-
150 - U
140-
130-
C .
U
110—
100—
C
U, 90-
U Uniti
r\) 0 80—
• UnIt2
• 70 -
U
.
O
50 -
40
30 -
20 —— I -
70 90 110 130 150 170 190 210
Inlet So 2 Concentratlon (ppm at 7% q )
Figure 5-1. Outlet SO 2 concentration as a function of
inlet SO 2 concentration at Dutchess County.
-------
In Table 5-6, particulate data are presented. The outlet PM
concentrations at Unit 1 ranged from 0.004 to 0.015 gr/dscf at 12 percent
CO 2 and averaged 0.0097 gr/dscf. Outlet PM concentrations for Unit 2 were
between 0.030 and 0.037 gr/dscf at 12 percent CO 2 and averaged 0.035 gr/dsci
for the February tests. After the February tests on Unit 2 failed to
demonstrate compliance with permit conditions, repairs and modifications
were made to the Unit 2 fabric filter. Broken bags were replaced, the
cleaning cycle time was reduced, and all particulate conveyors and hoppers
were cleared of built-up ash. For the March tests at Unit 2, outlet PM
concentrations ranged from 0.005 to 0.027 gr/dscf and averaged 0.011 gr/dscf
The outlet PM concentration for the May compliance tests on Unit 2 ranged
from 0.007 to 0.009 gr/dscf at 12 percent CO 2 . Thus, both units were able
to demonstrate outlet PM levels less than the permit level of 0.015 gr/dscf.
Table 5-7 presents the metals data from the three test runs at each of
the two units at Dutchess County. Outlet metals concentrations were
consistently low and similar between the two units for all metals except
mercury. Arsenic was not detected at either unit. Cadmium, chromium, and
nickel concentrations averaged less than 12 ug/dscm at 7 percent 02 at both
units. Lead concentrations averaged 39 ug/dscm at Unit 1 and 49 ug/dscm at
Unit 2. Similar outlet concentrations were observed at both units even
though outlet PM levels were substantially higher at Unit 2 during the same
testing period. The observed outlet concentrations, compared to typical
uncontrolled metals concentrations (Section 1.2), indicate removal
efficiencies of greater than 99 percent for arsenic, cadmium, chromium,
lead, and nickel.
Mercury exhibited higher outlet concentrations than the other metals
and showed wide variation between facilities. At Unit 1, the outlet mercury
concentration averaged 1,080 ug/dscm at a FE inlet temperature of 430°F
while at Unit 2, the average concentration was 85 ug/dscm at a FE inlet
temperature of 365°F. The 65°F lower fabric filter inlet temperature at
Unit 2 than at Unit 1 may have contributed to the difference.
5-13
-------
TABLE 5-6.
PARTICULATE DATA FOR DUTCHESS COUNTY
Test
Condition
Run a
Number
FE Inlet
Temp 8 ra ure
( F)
Flue Gas
Flow
(acfm)
Outlet PM
Concentration
(gr/dscf at 12% C0 2 )
Combustor = Normal 1-1 379 39,700 0.015
DSI/FF = Normal 1-2 380 42,600 0.010
(2/89 tests) 1-3 383 42,000 0.0038
Average (Unit 1) 381 41,400 0.0097
Combustor = Normal 2-1 354 42,600 0.037
DSI/FF = Normal 2-2 358 42,200 0.037
(2/89 tests) 2-3 372 43,300 0.030
2-4 365 42,900 0.034
Average (Unit 2) 364 42,800 0.035
Combustor = Normal 2-1 386 44,300 0.0069
DSI/FF = Normal 2-2 399 44,900 0.027
(3/89 tests) 2-3 380 38,600 0.012
2-4 363 42,800 0.0051
2-5 385 44,400 0.0077
2-6 377 42,700 0.0073
Average (Unit 2) 382 43,000 0.011
Combustor = Normal 2-1 377 48,900 0.0080
DSI/FF = Normal 2-2 382 49,400 0.0072
(5/89 tests) 2-3 389 49,100 0.0087
Average (Unit 2) 383 49,100 0.0079
aRUfl Number contains the unit number followed by the run number.
bTemperature estimated from measured value at he stack and an assumed
temperature drop across the fabric filter (10 F).
5-14
-------
TABLE 5-7. METALS DATA FOR DUTCHESS COUNTY
U.’
UI
Test
Condition
RUfla
Number
FE InLet
Tempera ure
C F)
OutLet PM
Concentration
(gr/dscf at
12% C0 2 )
OutLet Concentration
7%
As
Cd
(ug/dscm
Cr
at
02)
Pb
Hg
Ni
Combustor = NormaL
1-1
426
“
NOd
1.76
3.16
32.7
888
2.54
DSI/FF NormaL
1-2
435
--
ND
2.76
16.8
39.7
1,180
22.8
(2/89 tests)
1-3
428
--
ND
3.64
4.85
44.2
1,160
8.11
Average (Unit 1)
430
0.0097
ND
2.72
8.27
38.9
1,080
11.2
Combustor = Normal
2-1
352
“
ND
1.11
11.3
23.1
35.6
14.5
DSI/FF = NormaL
2-2
367
--
ND
4.49
4.63
75.4
95.6
4.45
(2/89 tests)
2-3
377
--
ND
3.48
3.51
48.9
123
3.46
Average (Unit 2)
365
0.035
ND
3.03
6.48
49.1
84.7
7.47
aRun Number contains the unit number foLLowed by the run number.
biemperature estimated from measured vaLue at the stack and an assumed temperature drop across the FF (10°F).
CParticulate data not coLLected simuLtaneousLy with metaLs. Average PM concentration for the same test period reported.
d - not detected.
-------
However, because the inlet mercury concentrations are unknown and because of
previously observed variability in uncontrolled mercury concentrations, it
cannot be determined how much of the lower mercury concentrations at Unit 2
are a result of FF performance rather than decreased inlet mercury levels.
Table 5-8 presents the CDD/CDF data from the three runs at each unit.
At Unit 1, outlet CDD/CDF concentrations ranged from 3.73 to 5.97 ng/dscm at
7 percent 02 and averaged 4.83 ng/dscm. At Unit 2, outlet CDD/CDF
concentrations ranged from 17.5 to 18.5 ng/dscm and averaged 17.4 ng/dscm.
Unit 1 operated at a FE inlet temperature 10°F less than Unit 2 (378 versus
387°F). There is insufficient information to explain the lower CDD/CDF
concentrations from Unit 1.
5.2.3 Ouebec City 5
The Quebec City, Canada, municipal waste combustion facility contains
four separate waterwall combustors. The combustors were originally built in
1975 with Von Roll reciprocating grates. Waterwall arches were added to
each combustion chamber in 1979. Each unit is designed to combust
250 tons/day of MSW. Emissions were originally controlled by 2-field ESP’s.
Environment Canada, in cooperation with Flakt Canada, LTD., established
an extensive test program to evaluate the capability of a pilot-scale duct
sorbent injection/fabric filter control system to remove PM, SO 2 , HC1, heavy
metals, CDD/CDF, and other organic compounds. Flakt constructed a
pilot-scale DSI/FF facility at the Quebec City Plant equipped with:
(1) a flue gas slipstream from the ESP inlet of Unit 3 to deliver
2,000 ft 3 /min at 500°F to the pilot facility;
(2) a water spray quench chamber where water is sprayed concurrently
with the flue gas to cool the flue gas to less than 230°F;
(3) a sorbent injection chamber with a single, dry hydrated lime
injection nozzle and an internal cyclone at the entrance to the
chamber; and
(4) a pulse-cleaned FE using high-temperature teflon bags with an
air-to-cloth ratio of 4.4 acfm/ft 2 .
5-16
-------
TABLE 5-8. CDD/CDF DATA FOR DUTCHESS COUNTY
Test
Conditions
Run
Number
a
FF Inlet
Temp 8 rature
( F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Combustor = Normal
1-1
378
5.97
DSI/FF = Normal
1-2
380
4.78
(2/89 tests)
1-3
376
3.73
Average (Unit 1)
378
4.83
Combustor = Normal
2-1
378
17.6
DSI/FF = Normal
2-2
379
17.5
(2/89 tests)
2-3
384
18.5
Average (Unit 2)
387
17.9
aRUfl number contains the unit number followed by the run number.
bTempet estimated from a measured value at the stack and an assumed
temperature drop of (10 F) across the fabric filter.
5-17
-------
Testing was conducted with the pilot-scale DSI/FF system in March 1985.
The results of tests with a pilot spray dryer/fabric filter at Quebec City
are reported in Section 7.2. Results of the tests of the full-scale ESP at
Quebec City are reported in Section 2.2.1.9.
The DSI/FF tests were conducted at a single lime feed rate and
temperatures of 230 to 400°F at the FE inlet. Flue gas was sampled
simultaneously at the water quench chamber inlet, FE inlet, and FF outlet,
and analyzed for HC1, SO 2 , metals (arsenic, cadmium, chromium, mercury,
lead, and nickel), CDD/CDF, and other organics. Metals data were not taken
at the mid-point sampling location. PM data were collected only at the
inlet sampling location and are not reported here.
Acid gas data are presented in Table 5-9. Nine test runs were
conducted at average fabric filter inlet temperatures of 400, 285, 250, and
231°F. The constant lime feed rate resulted in stoichiometric ratios
ranging from 1.2 to 1.5. The inlet concentrations of acid gases remained
relatively consistent during each condition. At a FE inlet temperature
setpoint of 392°F, two runs were conducted. Outlet SO 2 concentrations were
90 and 100 ppm, for an average of 95 ppm. SO 2 removal efficiency was
relatively consistent and averaged 23 percent. Outlet HC1 concentrations
were 88 and 122 ppm, for an average of 105 ppm. HC1 removal efficiency was
relatively consistent and averaged 76 percent. At a FE inlet temperature
setpoint of 284°F, three runs were conducted. Outlet SO 2 concentrations
ranged from 13 to 45 ppm and averaged 34 ppm. The SO 2 removal efficiency
varied from 58 to 92 percent and averaged 73 percent. Outlet HC1
concentrations ranged from 23 to 35 ppm and averaged 30 ppm. HC1 removal
efficiencies varied little and averaged 93 percent. At a FF inlet
temperature setpoint of 257°F, two runs were conducted. Outlet SO 2
concentrations were 14 and 7 ppm, for an average of 11 ppm. The average SO 2
removal efficiency was 92 percent. Outlet HC1 concentrations were 9 and
11 ppm. The average HC1 removal efficiency was 98 percent for both runs.
At a FF inlet temperature setpoint of 230°F, two runs were conducted.
Outlet SO 2 concentrations were 4 and 6 ppm. SO 2 removal efficiency was
5-18
-------
TABLE 5-9. ACID GAS DATA FOR QUEBEC CITY PILOT DSI/FF
Test
Condition
Run
Number
FF InLet
Temp rature
( F)
Stoichiometrjc
Ratio
Acid Gas Concentrations
Quench
Chamber
Removal.
(%)
SO 2 HCL
FF
RemovaL
(%)
SO 2 HCL
OveraLL
Acid Gas
RemovaL
(%)
SO 2 HCL
(ppmv.
InLet
SO 2 HCL
dry at 7%
P4idpoint
SO 2 HCL
02)
OutLet
SO 2 HCL
Combustor NgrmaL 5 400 1.6 126 422 112 211 90 88 11 50 20 54 29 77
DSI/FF at 392 F FF 6 400 1.4 122 481 98 251 100 122 20 48 - 3 52 18 75
inLet temperature
Average 400 1.5 124 451 105 231 95 105 15 49 9 53 23 76
Combustor Nç rmat 1 285 1.2 148 458 - - 141 44 35 - - 69 -- 75 70 92
DSI/FF at 284’ F FF 2 287 1.4 158 400 115 142 13 23 27 64 88 83 92 94
inLet temperature 11 284 1.2 107 512 70 139 45 31 35 93 36 78 58 94
“ Average 285 1.3 138 457 92 141 34 30 31 69 62 79 73 93
Combustor = N9rmaL 3 250 1.2 148 470 82 93 14 9 45 80 82 90 90 98
DSI/FF at 257’ F FF 4 250 1.1 108 530 59 55 7 11 46 90 87 79 93 98
inLet temperature
Average 250 1.2 128 500 70 74 11 10 45 85 85 85 92 98
Combustor = N9rmaL 12 229 1.2 128 509 29 4 4 8 77 99 86 -85 97 98
DSI/FF at 230’ F FF 13 232 1.3 129 404 23 28 6 7 82 93 76 75 96 98
inLet temperature
Average 231 1.3 129 456 26 16 5 7 80 96 81 - 5 96 98
-------
consistent between the two runs and averaged 96 percent. Outlet HC1
concentrations for the two runs were 8 and 7 ppm. HC1 removal efficiency
was 98 percent for both runs.
In Figure 5-2, SO 2 removal efficiency is shown as a function of FE
inlet temperature. Generally, SO 2 removal efficiency increased as
temperature decreased. Similar phenomena is observed with HC1, as shown in
Figure 5-3, although the amount of increase is not as great. HC1 removal
efficiencies above 90 percent were demonstrated at FE inlet temperatures of
285°F or less.
The amount of SO 2 and HC1 removal across the water quench chamber
increased as the FF inlet temperature decreased. For SO 2 , the removal
efficiency across the quench chamber increased from about 15 percent at a FF
inlet temperature of 400°F to about 80 percent at a FE inlet temperature of
230°F, while for HC1, the removal efficiency across the quench chamber at
the same temperature increased from 50 to about 95 percent. Acid gas
removal efficiencies across the FE showed some variation with temperature,
but not nearly as much as the removal across the SD. At 400°F at the FF
inlet, SO 2 removal efficiency across the FE was 20 percent for Run 5 and
-3 percent for Run 6. The negative removal efficiency indicates little or
no SO 2 removal, as the midpoint and outlet SO 2 concentrations differed by
only 2 ppm (within analytical error). At the same temperature, HC1 removal
efficiency across the FF averaged 53 percent. At 284°F at the FE inlet, SO 2
removal efficiency was 88 and 36 percent and HC1 removal efficiency averaged
79 percent. At 257 and 230°F FE inlet temperatures, SO 2 and HC1 removal
efficiencies across the FE were consistently about 85 percent except for Run
12 at 230°F. A negative removal efficiency was obtained for HC1 after an
unusually low HC1 concentration was measured at the midpoint (4 ppm).
Because the value was so low, additional removal across the FE could not be
easily measured, because the accuracy of instruments decreases at such low
concentrations. Thus, the negative value probably indicates little or no
HC1 removal across the FF.
Thus, based on analyses of the acid gas data, temperature appears to
significantly affect both SO 2 and HC1 removal efficiency for a DSI/FF
system, although the effect is more pronounced for SO 2 .
5-20
-------
100
L1 [ )
U
90 - [ 1
80
C
I I I I I I
220 240 260 280 300 320 340 360 380 400 420
FF Inlet Temperature (°F)
Figure 5-2. SO 2 removal efficiency as a function of FF inlet
temperature at the Quebec City DSI/FF system.
-------
100 -
98
96 -.
I I I I I I I I I I
220 240 260 280 300 320 340 360 380 400 420
FF Inlet Temperature (°F)
Figure 5-3. HCI removal efficiency as a function of FF inlet
temperature at the Quebec City DSI/FF system.
-------
Metals data for the nine test runs are presented in Table 5-10.
Removal efficiencies for arsenic, cadmium, chromium, lead, and nickel were
above 99.9 percent for all test conditions. Mercury removal was
consistently above 85 percent for all test runs at FF inlet temperatures
below 285°F. At 400°F, no mercury removal occurred. Although outlet
mercury concentrations were greater than the inlet values, there is no
apparent reason for these differences.
CDD/CDF data are presented in Table 5-11 for the eight runs conducted.
Run 2 samples were lost prior to analysis. Outlet CDD/CDF concentrations
ranged from not detected to 9.0 ng/dscm at 7 percent 02 and removal
efficiencies were at least 99.7 percent for all test runs. At temperatures
below 285°F, 35 to 65 percent of the CDD/CDF removal occurred across the
quench chamber. At 400°F, no CDD/CDF removal occurred across the quench
chamber. Removal across the FF was consistently above 99.5 percent. Thus,
decreasing the FF inlet temperature to below 285°F will decrease CDD/CDF
emissions with the DSI/FF system at Quebec City.
5.2.4 Springfield 6
The Springfield Resource Recovery Facility in Agawam, Massachusetts,
consists of three identical Vicon Recovery Systems modular combustors, each
designed to combust 120 tons/day of MSW. Flue gas from the combustor trains
is cooled through heat recovery to about 280°F at a flow rate of
24,000 acfm. Dry hydrated lime is pneumatically injected into the flue gas
downstream of the heat recovery in a reactor chamber which is sited to
provide for about one second of residence time. The design lime feed rate
is 83 lb/hr. The reaction products, fly ash, and unreacted sorbent are
collected in a pulse-jet cleaned fabric filter with 4 compartments of
126 bags each. The bags are made of teflon coated fiberglass and provide a
net air-to-cloth ratio of 2.9 acfm/ft 2 . Flue gases are exhausted through a
common stack.
The initial compliance testing at Springfield in October 1988 is
summarized below. Three test runs were conducted at the common stack for
each of the following pollutants: SO 2 , HC1, PM, trace metals, (arsenic,
cadmium, chromium, lead, mercury, and nickel) and CDD/CDF. The outlet SO 2
concentration averaged 23 ppm, with an average removal efficiency of
5-23
-------
TABLE 5-10. METALS DATA FOB QUEBEC CITY PILOT 051/FT
Outlet PM
FT Inlet Concentration
Teaç 8 r. ure (gr/dscf
Bun ( F) 7% 02)
5 400 - -
6 400 -
400
1 255 -
2 287 -
11 284
Removal Efficiency
1%)
Cr Pb
TOO 99.98
99.95 99.98
99.95 99.98
0uttet PM not measured during metals runs. All concentrations below detection limit
ND not detected.
not measured.
during preliminary teats of the system.
Teat Condition
Coabuator • ormsl
DSI/FF 392 F Inlet
Average
Cosbustor ormaI
OSI/FF 284 F inlet
Aver age
Cosbustor ormaI
051/FT 257 F inlet
Average
Cosbuator normal
DSI/FF • 230 F Inlet
Average
Inlet Concentration
(ueldacm_• _ ____________
As Cd Cr Pb
85 1,007 1,754 35,689
70 1,020 1,979 33,870
78 1,014 1,882 34,780
NNC NM NM NM
174 2,006 2.873 36,942
90 1,028 1,180 31,264
132 1.516 2,027 34,103
3 250 -- 103 1,340 1,918 44,53?
4 250 - - NM NM NM NM
103 1,340 1,918 44,537
229 -- 147 1,277 3,087 61,128
732 - - 272 95.7 1,5.41 30,587
195 1,117 2,309 25,556
Ng Ni
501 626
400 643
651 835
NM NM
269 1,890
351 629
320 1,260
428 1,???
531 NM
450 1,7??
208 950
651 886
445 933
Hg
549
-12.5
339
12
13
Outlet Concentration
— (ua/d,cm at 1% __________
As Cd Cr Pb Hg NI
0.10 1.22 50 b 5.34 776 1.00
0.04 ND 1.03 7.25 451 2.07
0.07 0.61 0.50 6.30 614 1.54
0.12 ND 2.41 9.64 8.0 2.41
0.04 ND 2.02 6.07 $5 1.00
0.04 ND ND 122 31.1 041
0.07 ND 1.46 6.31 15.9 1.27
0.04 0.44 0.64 2.66 0.9 0.44
ND 0.42 2.51 2.51 24.5 4.16
0.02 0.43 1.46 2.59 12.7 2.31
0.02 0.44 0.44 3.94 30.9 1.32
NM NM NM NM 48.6 NM
0.02 0.64 0.44 3.96 39.6 1.32
As
99.9
99.94
99.92
9998
99.95
99.97
99.97
99.97
99.99
9999
Cd
99-9
100
99.93
100
100
100
99.97
99.97
99.97
99.97
Ni
99.9
99.8
999
99.5
99.9
99.9
99.98
99.98
99.9
99.9
99.9
100
99.9
99.98
99.98
99-99
99.99
99.98
100
99.98
99.99
99.99
99.99
99.99
97.1
91.1
99.1
99.8
95.4
97.6
85. 1
92.9
89.0
-------
TABLE 5-il. CDD/CDF DATA FOR QUEBEC CITY PILOT DSI/FF
U,
aNM = not measured
bND
= not detected.
Quench
Test
Conditions
Run
Number
FF InLet
Temp 8 rature
C F)
CDD/CDF Concentration
Chamber
CDD/CDF
RemovaL
(%)
FF
CDD/CDF
Removal
C X)
OveraLl
CDD/CDF
Removal
(%)
Inlet
(ng/dscm at 7%
Midpoint
02)
Outlet
Combustor N 8 rmaL
DSI/FF at 392 F
5
6
400
400
1,820
1,374
1,903
1,373
8.99
5.68
-4.52
0.07
99.5
99.6
99.5
99.6
inlet temperature
Average
400
1,597
1,638
7.33
-4.45
99.6
99.6
Combustor N 8 rmaL
DSI/FF at 284 F FF
inlet temperatureC
1
11
285
284
2,281
2,272
NMa
1,176
NOb
0.97
48.3
99.9
99.96
Average
285
2,277
1,176
0.49
48.3
99.9
99.96
Combustor N 8 rmaL
DSI/FF at 257 F FE
3
4
250
250
1,958
2,766
1,079
934
ND
ND
44.9
66.2
100
100
100
100
inlet temperature
Average
250
2,361
1,007
ND
55.6
100
100
Combustor = N 8 rmaL
DSI/FF at 230 F FF
12
13
229
232
1,018
756
753
413
3.59
1.28
26.0
45.4
99.5
99.7
99.7
99.8
inlet temperature
Average
231
887
456
2.43
35.7
99.6
99.7
CRun 2 samples lost.
-------
83 percent. The outlet HC1 concentration averaged 33 ppm, with an average
removal efficiency of 94 percent. The outlet PM concentration was
0.0016 gr/dscf, with a removal efficiency of 99.8 percent. Outlet metals
were measured from which a removal efficiency was estimated assuming typical
uncontrolled metals concentrations. For arsenic, cadmium, and lead, outlet
concentrations of not detected, 1, and 21 ug/dscm, respectively, indicate
removal efficiencies of greater than 99 percent. Outlet chromium and nickel
concentrations of 10 and 24 ug/dscm, respectively, suggest removal
efficiencies of at least 97 percent. Outlet mercury emissions were
300 ug/dscm, indicating approximately 70 percent removal efficiency. The
total CDD/CDF concentration was not reported in the data summary, but the
2378-TCDD toxic equivalency of 0.09 ng/dscm at 7 percent 02 was given.
5.2.5 St. Croix 7 U
The St. Croix Waste to Energy Facility in New Richmond, Wisconsin,
consists of three identical Cadoux modular, excess air combustors each
designed to combust 38 tons/day of MSW. Flue gas from each of the three
combustors passes through steam boilers, and is then combined and routed to
a dual hydrated lime injection system manufactured by Interel Corporation.
Hydrated lime is injected into the flue gas at a maximum rate of
100 lb/hr upstream of two parallel air-to-air heat exchangers that cool the
flue gas from 400 to 250°F or less. The flue gas then flows upward through
parallel reactor vessels that provide additional contact time before
entering the fabric filters. At the bottom of the union of the reactor
vessels, there is a drum of ceramic balls which break up particulate clumps
as the flue gas passes through. Recirculated fabric filter ash can be
injected into the reactor vessels if desired.
The fabric filters are pulse-jet cleaned horizontal bag units which
provide a surface area of 4,844 ft 2 each at a flue gas flow of about
14,500 acfm each. Air-to-cloth ratio is about 3 acfm/ft 2 . The flue gas
exiting the fabric filters is combined and exhausted through a common stack.
Several tests have been conducted at the St. Croix MWC. In all of
these tests, the combustor was operated normally with very high excess air
rates which required large correction factors to normalize the data. In
5-26
-------
June 1988, compliance testing was conducted at the FF outlet. The DSI
system was operated with the maximum lime injection rate of 100 lb/hr and
maximum ash recirculation ratio, 16 times the flyash and lime loading to th
FE without recirculation. Flue gas was analyzed for SO 2 , HC1, PM, metals
(arsenic, cadmium, chromium, lead, mercury, and nickel), and CDD/CDF.
Simultaneous SO 2 and HC1 measurements were taken at the DSI/FF system to
characterize system performance. In August 1988, additional testing was
conducted with SO 2 and HC1 measured simultaneously at the inlet and outlet
of the OSI/EF system to again characterize system performance. The lime
feed rate was maintained at 100 lb/hr, but the recirculation ratio was
decreased to 3.5. In October 1988, particulate emissions were measured at
the FF outlet because the system had not met the permit requirements for PM
in June. Three test runs to again demonstrate system performance for SO 2
and HC1 were also conducted in October. The lime feed rate was halved
(53 lb/hr) during two runs, and all three runs were without ash
recirculation. Flue gas samples were collected and analyzed for SO 2 and HC1
at the DSI inlet, the two FE inlet locations, and the FE outlet.
In Table 5-12, acid gas data from all three test programs for the
DSI/FF inlet and outlet sampling locations are presented. The data consist
of nine separate test runs. All tests were conducted at the same FE inlet
temperature of 250°F. During testing in June with 100 lb/hr of lime and ash
recycle ratio of 16, outlet SO 2 concentrations ranged from 15 to 58 percent
at 7 percent 02 and averaged 34 ppm. SO 2 removal efficiencies ranged from
64 to 94 percent and averaged 78 percent. Outlet HC1 concentrations ranged
from 0.34 to 1.5 ppm and averaged 0.75 ppm. HC1 removal efficiencies were
all greater than 99.7 percent. In August, with 100 lb/hr of lime and a
recycle ratio of 3.5, outlet SO 2 concentrations ranged from 2 to 8 ppm and
averaged 5 ppm. Corresponding SO 2 removal efficiencies ranged from 89 to 97
percent and averaged 93 percent. HCl was not detected in any of the three
runs, for a removal efficiency of at least 99.97 percent. The single
October test run with 100 lb/hr of lime and no ash recycle resulted in not
detected outlet SO 2 and HC1 concentrations. The removal efficiencies were
at least 98.7 percent for SO 2 and 99.99 percent for HC1. The last two
October test tuns with 53 lb/hr of lime and no ash recycle resulted in
5-27
-------
TABLE 5-12. ACID GAS DATA FOR ST. CROIX
FF InLet
Temp 8 rature Stoichiometric
( F) Ratio
250 2.2
250 3.6
250 2.8
250 2.9
250 3.5
250 3.3
250 2.8
250 3.?
250 2.3
Acid Gas Concentration
(ppm, dry at 7% 0,)
InLet Outtet
SO 2 HCL SO 2 HCL
250 479 15 0.34
117 472 30 0.42
163 507 58 1.5
177 486 34 0.75
74 426 5 NDa
81 524 2
82 529 8 NDa
79 493 5 ND
86 706 NDb NOC
Acid Gas
Remova
Efficiency
(%)
SO 2 HCL
94.0 99.93
74.4 99.91
64.4 99.7
77.6 99.9
92.9 >99.97
97.4 >99.97
89.4 >99.97
93.3 >99.97
98.7 >99.99
Test
Condition
Combustor NormaL
DSI/FF
Lime feed = 100 tb/hr
Ash recycLe rate 16
Average
Run
Number
2
3
N)
Combustor = NormaL 1
DSI/FF 2
Lime feed = 100 Lb/hr 3
Ash recycLe rate = 3.5
Average
Combustor NormaL 1
DSI/FF
Lime feed = 100 Lb/hr
No ash recycLe
Combustor = NormaL 2 250 1.1
DSI/FF 3 250 1.2
Lime feed = 53 Lb/hr
No ash recycLe
Average 250 1.2
8 Not detected. Detection Limit 0.016 ppm at 7 percent 02.
bNot detected. Detection Limit = 1.5 ppm at 7 percent 02.
detected. Detection Limit 0.022 ppm at 7 percent 0 .
81
116
816
669
23
33
NDC
NDC
71.7
71.6
>99.99
>99.99
99
743
28
NDC
71.7
>99.99
-------
outlet SO 2 concentrations of 23 and 33 ppm for an average of 28 ppm. SO 2
removal efficiency was at 72 percent for both runs. HC1 was not detected at
the outlet, for a removal efficiency of at least 99.99 percent.
Acid gas data from the FE inlet taken during the October testing are
presented in Table 5-13. Three test runs were conducted for each unit.
Although not collected simultaneously with the inlet and outlet acid gas
samples, when compared to typical inlet values measured previously at St.
Croix, the data suggest that 60 to 90 percent SO 2 removal occurred across
the heat exchanger. For HC1, 85 to 95 percent of the inlet amount was
removed across the heat exchanger.
Inlet SO 2 concentration appears to affect performance. The highest
outlet SO 2 concentrations were obtained at inlet SO 2 concentrations above
115 ppm. At inlet SO 2 emissions below 90 ppm, four of five runs yielded
outlet SO 2 emissions below 9 ppm. Because the HC1 removal was consistently
near 100 percent, the effects of stoichionietric ratio, and inlet HC1
concentration on performance cannot be evaluated. In Figure 5-4, SO 2
removal efficiency is shown as a function of stoichiometric ratio. In
general, SO 2 removal efficiency increased as stoichiometric ratio increased.
Fly ash recirculation does not appear to affect SO 2 removal efficiency.
Higher SO 2 removal efficiencies were obtained at a recirculation ratio of
3.5 than at a ratio of 16.
Particulate data from the May and October test programs are presented
in Table 5-14. Outlet PM concentrations ranged from 0.010 to 0.020 gr/dscf
at 12 percent CO 2 . The results from the May tests did not meet the permit
requirements for solid plus condensible particulate (not reported here).
Minor changes were made to the system prior to the October 1988 testing
which lowered the average outlet PM concentrations from 0.015 to
0.012 gr/dscf, as shown by the October results.
Metals data from the June testing are presented in Table 5-15. Three
runs were conducted at a FE inlet temperature of 223°F. Low emissions were
measured for all the metals. Based on typical uncontrolled metals
concentrations (Section 1.2), all metals were removed at greater than
5-29
-------
TABLE 5-13. FABRIC FILTER INLET ACID GAS DATA FOR ST. CROIX
Fabric Filter
Inlet
Acid Gas Concentration
Test
Condition
Run
Numbera
FF Inlet
Temp 8 rature
( F)
(ppm, dry at
7% 02)
SO 2
HC1
Combustor = Normal
N-i
NMb
47
41
DSI/FF
N-2
NM
89
34
Lime feed = 100
lb/hr
N-3
NM
14
45
No ash recycle
Average North
NM
50
40
S-i
S-2
S-3
NM
NM
NM
11
8.1
7.8
87
19
119
Average South
NM
9.0
75
aHeat exchanger identification is given first (N=north, S=south) followed by
the run number.
bNM = Not measured.
5-30
-------
100 A
A
95
A
90
C
C)
I-
85
>1
C)
C
a)
A 1 4 1 8 2 2 2 6
Stoichiometric Ratio
Figure 5-4. SO 2 removal efficiency as a function of stoichiometric
ratio at the St. Croix DSI/FF system.
-------
TABLE 5-14. PARTICULATE DATA FOR ST. CROIX
FE Inlet Flue Gas Outlet PM
Test Run Temp 8 ra ure Flow Concentration
Condition Number ( F) (acfm) (gr/dscf at 12% C0 2 )
Combustor = Normal 1 220 27,200 0.011
DSI/FF = Normal 2 220 27,500 0.012
(6/88 tests) 3 220 27,700 0.020
Average 220 27,500 0.015
Combustor = Normal 1 227 28,900 0.010
DSI/FF = Normal 2 227 28,900 0.011
(10/88 tests) 3 227 28,800 0.016
Average 227 28,900 0.012
aTemperature estimated from a measured value a the stack and an assumed
temperature drop across the fabric filter (10 F).
TABLE 5-15. METALS DATA FOR ST. CROIX
Outlet PM
FE Inlet Concentration Outlet Concentration
Test Run Temp 8 ra ure (gr/dscf at (u Jdscm at 7% O ,.L.
Condition Number ( F) 12% C0 2 ) As Cd Cr Pb H Ni
Combustor = Normal 1 223 NMb NDC 3.6 55 26 43 52
DSI/FF = Normal 2 223 NM ND 3.2 17 17 39 29
(6/88 tests) 3 223 NM 6.3 ND 7 12 24 14
Average 223 0015 d 2.1 2.3 26 18 35 32
aTemperature estimated from a measured value a the stack and and assumed
temperature drop across the fabric filter (10 F).
bNM = not measured.
CND = not detected.
dparticuiate not measured simultaneously with metals. Average result from
6/88 test reported.
5-32
-------
94 percent efficiency. Removal efficiencies for cadmium and lead were at
least 99 percent. Removal efficiencies for arsenic, nickel, and mercury
were 96 to 97 percent. Chromium removal efficiency was 94 percent.
CDD/CDF data are reported in Table 5-16. Outlet CDD/CDF concentrations
ranged from 2.8 to 11.8 ng/dscm at 7 percent 02 and averaged 7.7 ng/dscm.
Inlet CDD/CDF concentration was not measured.
5.2.5 Wurzburg 12
The MWC facility at Wurzburg, West Germany, consists of two identical
mass burn, waterwall combustors with Martin GmbH grates, each designed to
combust 330 tons/day of MSW. Flue gas exiting each boiler flows through a
water spray quench chamber, after which, powdered hydrated lime is injected
into the flue gas in a reactor chamber. Particulate matter is removed by a
pulse-jet cleaned fabric filter. The flue gas flow at the FE inlet is
typically 50,000 acfm at 375°F. No other information is available on the
air pollution control system.
Testing was performed at the facility in January 1986 in order to
document emission levels using U.S. EPA test protocols. The combustor and
DSI/FF system were operated under normal conditions during testing. At the
EF outlet, flue gas was analyzed for SO 2 , HC1, PM, metals, CDD/CDE, and
vinyl chloride.
Acid gas data are presented in Table 5-17. Five test runs were
conducted at a FE inlet temperature of 380°F. Outlet SO 2 concentrations for
the final two runs tested were 145 and 199 ppm at 7 percent 02. Outlet HC1
concentrations ranged from 29 to 59 ppm at 7 percent 02 over the five test
runs and averaged 45 ppm. The temperature range is too limited to evaluate
the effect of temperature on HC1 emissions.
In Table 5-18, particulate data for samples collected simultaneously
with the HC1 samples are presented. The outlet PM concentration ranged from
0.0025 to 0.0074 gr/dscf at 12 percent CO 2 over three runs and averaged
0.0042 gr/dscf.
Metals data are presented in Table 5-19. These data represent analysis
of particulate collected over 18 hours. Very low concentrations were
measured for the metals, suggesting removal efficiencies greater than
99 percent for arsenic, cadmium, chromium, lead, and nickel. Mercury was
not measured.
5-33
-------
TABLE 5-16. CDD/CDF DATA FOR ST. CROIX
Outlet CDD/CDF
Test Run FE Inlet Concentration
Condition Number Temperature (oF)a (ng/dscm at 7% 02)
Combustor = Normal 1 223 8.61
DSI/FF = Normal 2 220 2.80
(6/88 tests) 3 220 11.8
Average 221 7.73
aTemperature estimated from measured value of he stack and an assumed
temperature drop across the fabric filter (10 F).
TABLE 5-17. ACID GAS DATA FOR WURZBIJRG
Outlet Acid Gas
Concentrations
FF Inlet (ppmv, dry at
Test Run Temp 8 rature Stoichiometric 7% 02)
Condition Number ( F) Ratio SO 2 HC1
Combustor Normal 1 381 NMa NM 29
DSI/FF = Normal 2 382 NM NM 59
3 365 NM NM 45
4 NM NM 145 42
5 NM NM 199 50
Average 376 NM 172 45
aNM = Not measured.
5-34
-------
TABLE 5-18. PARTICULATE DATA FOR WURZBURG
Test
Condition
FF Inlet Flue Gas Outlet PM
Run Temp rature Flow Concentration
Number ( F) (acfm) (gr/dscf at 12% C0 2 )
Combustor = Normal
1 381 50,400 0.0025
DSI/FF = Normal
2 382 48,100 0.0074
3 365 51,500 0.0026
Average
376 50,000 0.0042
TABLE 5-19. METALS DATA FOR WURZBURG
Test
Condition
Outlet PM
FF Inlet Concentration Outlet Concentration
Run Temp 8 rature (gr/dscf at (ug/dscm at 7% 02)
Number ( F) 12% C0 2 ) As Cd Cr Pb Ni
Combustor = Normal
1 365 0.0042 NDa 5.5 0.50 11 0.23
DSI/FF = Normal
(18-hour test)
aND = Not Detected
5-35
-------
In Table 5-20, CDD/CDF data are presented. Outlet COD/COF
concentrations ranged from 17.2 to 82.0 ng/dscm over three runs and averaged
40.4 ng/dscm. The average FF inlet temperature was 374°F; however, there
was insufficient range in temperature to allow determination of the effect
of temperature on performance. No inlet PM or CDD/CDF data were collected,
preventing analysis of the effects of these parameters.
5.3 SUMMARY OF PERFORMANCE
Section 5.2 discussed DSI/FF performance for individuals systems. This
section combines the data from these facilities to examine the relationship
between key operating parameters (temperature and stoichiometric ratio) and
DSI/FF performance.
5.3.1 Acid Gas
Analysis of acid gas data shows that the fabric filter inlet
temperature significantly affects SO 2 and HC1 removal efficiency with DSI/FF
systems. As shown in Figures 5-5 and 5-6 for SO 2 and HC1, respectively,
removal efficiency across the DSI/FF systems increases as temperature
decreases for individual facilities as well as for the whole data set.
Better than 90 percent SO 2 and 95 percent HC1 removal efficiency is
demonstrated with a DSI/FF system at a FF inlet temperature of 250°F. The
data from St. Croix showing 70 to 99 percent SO 2 removal efficiency
demonstrate the impact of increasing the stoichiometric ratio. At a
temperature between 250 and 300°F, the same levels of performance may be
achievable with a higher stoichiometric ratio than required at lower
temperatures. Several vendors of DSI/FF systems claim that removal
efficiencies of 90 percent for SO and 95 percent for HC1 are readily
achievable using DSI/FF systems.’ ,14,15
5.3.2 Particulate Matter
Outlet PM concentrations averaged 0.01 gr/dscf or lower at all six of
the DSI/FF systems tested.
5.3.3 Metals
Metals removal efficiency data are only available for Quebec City.
However, based on measured outlet concentrations, removal efficiencies were
5-36
-------
TABLE 5-20. CDD/CDF DATA FOR WURZBURG
Test
Condition
Run
Number
FF Inlet
Temperature
Outlet
Concen
(ng/dscm
CDD/CDF
tration
at 7% 02)
Combustor
= Normal
1
375
82.0
DSI/FF =
Normal
2
3
373
373
21.9
17.2
Average
374
40.4
5-37
-------
100
DO
90
80-
70-
60-
________
40- _________ V
30 - 0 Qu.b.c City
0 Springfi.Id
20 - st.Cro ix v 0
10 - utct 1S’ County
0
E
0-
:::
-30 -
-40-
—50——
220 240 260 280 300 320 340 360 380 400 420
FF Inlet Temperature (°F)
igure 5-5. SO 2 removal efficiency as a function of FF
inlet temperature for DSI/FF systems.
-------
100 -
98 +
96
+
94 +
C
9
0
I-
0)
90- +
>
C) +
C
0) 88
C)
86 -
>
o 84-
E +
82-
EJ Qu.b.c City
80 — Springfi.Id
St. Croix
78
+ Ciar.mont
0
76
0
220 260 300 340 380 420 460
FF Inlet Temperature (°F)
Figure 5-6. HCI removal efficiency as a function of FF
inlet temperature for DSI/FF systems.
-------
estimated for the DSI/FF systems at Dutchess County, Springfield, St. Croix
and Wurzburg. For cadmium and lead, removal efficiencies were 99 percent or
greater at all facilities. Four of five sites had arsenic removal at
greater than 99 percent. Two of three runs at the fifth site, St. Croix,
had estimated removal efficiencies of better than 99 percent. Chromium and
nickel were not removed as effectively as the previously listed metals. At
Quebec City, greater than 99 percent removal efficiency was measured for
both metals. Similarly, the measured chromium and nickel emissions at
Wurzburg and Dutchess County suggest removal efficiencies of at least 99
percent. At Springfield, however, estimated removal efficiencies for both
metals are 97 percent and at St. Croix, the estimated removal efficiencies
are 94 percent for chromium and 96 percent for nickel.
Mercury removal efficiency depended on temperature at the FF inlet.
For systems operating below 300°F, mercury removal efficiency ranged from a
measured value of 99 percent at Quebec City, to estimated efficiencies of 93
to 97 percent at St. Croix, and 40 to 70 percent at Springfield. At 400°F
at Quebec City, no mercury removal was observed. At Dutchess County, at
4300 and 365°F, mercury removal efficiencies of zero and 80 percent,
respectively, were estimated.
Removal efficiencies of greater than 99 percent are achievable by
DSI/FF systems for arsenic, cadmium, and lead. Similarly, removal
efficiencies of 96 percent can be achieved for chromium and nickel. For
mercury, DS!/FF systems can achieve 70 percent removal efficiency by
decreasing the FF inlet temperature to below 300°F.
5.3.4 CDD/CDF
Removal of CDD/CDF appears to depend on temperature at the fabric
filter inlet. As shown in Figure 5-7, average CDD/CDF emissions decreased
as the temperature decreased. At less than 300°F, emissions were below 7.7
ng/dscm at 7 percent 02.
5-40
-------
45
40 - • Ou.b.c City x
ad >( Wurzburg
+
N. A St. Croix
— + CIar.mont
I I I +
220 260 300 340 380 420 460
FF Inlet Temperature (°F)
Figure 5-7. Outlet CDD/CDF concentration as a function of
FF inlet temperature for DSI/FF systems.
-------
5.4 REFERENCES
1. Municipal Waste Combustion Study, Flue Gas Cleaning Technology. U. S.
Environmental Protection Agency. EPA/530-SW-87-021d. pp. 2-17.
2. Almega Corporation. SES Claremont, Claremont, NH, NH/VT Solid Waste
Facility. Unit 1 and Unit 2, EPA Stack Emission Compliance Tests, May
26, 27, and 29, 1987. Prepared for Clark-Kenith, Inc. Atlanta,
Georgia. July 1987.
3. Entropy Environmentalists, Inc. Stationary Source Sampling Report,
Signal Environmental Systems, Inc., Claremont Facility, Claremont, New
Hampshire, Dioxins/Furans Emissions Compliance Testing, Units 1 and 2.
Prepared for Signal Environmental Systems, Inc., Claremont, New
Hampshire. Reference No. 5553-A. October 2, 1987.
4. Beachier, D.S. (Westinghouse Electric Corporation) and ETS, Inc.
Dutchess County Resource Recovery Facility Emission Compliance Test
Report, Volumes 1-5. Prepared for New York Department of Environmental
Conservation. June 1989.
5. Flakt Canada Ltd. and Environment Canada. The National Incinerator
Testing and Evaluation Program, Air Pollution Control Technology,
Volume 4. September 1986.
6. McClanahan, D (Fluor Daniel), A. Licata (Dravo), and J Buschmann
(Flakt, Inc.). Operating Experience with Three APC Designs on
Municipal Incinerators. Proceedings of the International Conference on
Municipal Waste Combustion. April 11-14, 1988. pp. 7C-19 to 7C-41.
7. lnterpoll Laboratories, Inc. Results of the June 1988 Air Emission
Performance Test on the MSW Incinerators at the St. Croix Waste to
Energy Facility in New Richmond, Wisconsin. Prepared for American
Resource Recovery. Waukesha, Wisconsin. Report No. 8-2560.
September 12, 1988.
8. Interpoll Laboratories, Inc. Results of the June 6, 1988, Scrubber
Performance Test at the St. Croix Waste to Energy Incineration Facility
in New Richmond, Wisconsin. Prepared for Interel Corporation.
Englewood, Colorado. Report No. 8-2560!. September 20, 1988.
9. Interpoll Laboratories, Inc. Results of the August 23, 1988, Scrubber
Performance Test at the St. Croix Waste to Energy Incineration Facility
in New Richmond, Wisconsin. Prepared for Interel Corporation.
Englewood, Cololrado. Report No. 8-2609. September 20, 1988.
10. Interpoll Laboratories, Inc. Results of the October 1988 Particulate
Emission Compliance Test on the MSW Incinerator at the St. Croix Waste
to Energy Facility in New Richmond, Wisconsin. Prepared for American
Resource Recovery. Waukesha, Wisconsin. Report No. 8-2547. November
3, 1988.
5-42
-------
11. Interpoll Laboratories, Inc. Results of the October 21, 1988, Scrubber
Performance Test at the St. Croix Waste to Energy Facility in New
Richmond, Wisconsin. Prepared for Interel Corporation. Englewood,
Colorado. Report No. 8-2648. December 2, 1988.
12. Hahn, J. L. (Cooper Engineers, Inc.). Air Emissions Testing at the
Martin GmbH Waste-to-Energy Facility in Wurzburg, West Germany.
Prepared for Ogden Martin Systems, Inc. Paramus, New Jersey.
January 1986.
13. Telecon. D. M. White, Radian Corporation, with J. Buschmann, Flakt,
Inc. June 8, 1989. Subject: Achievable acid gas reductions and
retrofit times.
14. Meeting Presentation. R. Ireland, Procedair Industries, with
Industrial Studies Branch, Office of Air Quality Planning and
Standards. March 30, 1989. Subject: Performance of Dry Injection/
Fabric Filter Systems.
15. Meeting Presentation. E. B. Mull, Interel Corporation, with Industrial
Studies Branch, Office of Air Quality Planning and Standards.
April 1989.
5-43
-------
6.0 SPRAY DRYING FOLLOWED BY AN ELECTROSTATIC PRECIPITATOR
Section 6.0 describes the technology and performance of spray dryer
systems followed by an ESP. In Section 6.1, SD/ESP operation and design is
described. In Section 6.2, descriptions of facilities with emissions data
for SD/ESP systems and summaries of the available data from each facility
are provided. In Section 6.3, the performance of SD/ESP systems at
controlling acid gas, PM, metals, and CDD/CDF emissions is discussed.
6.1 PROCESS DESCRIPTION
Spray drying is designed to control acid gases, but also provides
control of organics and volatile metal emissions from MWC’s. In the spray
drying process, lime slurry is injected into the spray dryer (SD) chamber
through either a rotary atomizer or two-fluid nozzles. Rotary atomizers use
centrifugal energy to atomize the slurry. The slurry is fed to the center
of a rapidly rotating disk or wheel where it flows outward to the edge of
the disk. The slurry is atomized as it leaves the surface of the rapidly
rotating disk. Two-fluid nozzles use kinetic energy to atomize the slurry.
High velocity air is injected into a stream of slurry, breaking the slurry
into droplets, which are ejected at near-sonic velocities into the spray
drying chamber. Both of these atomization methods have been used in spray
dryers on MWC’s. Spray dryers with two-fluid nozzles are typically larger
in height than diameter while those with rotary atomizers have a larger
diameter than height. Slurry droplets of comparable size can be obtained
with both two-fluid nozzles and rotary atomizers, minimizing differences in
performance due to atomizer type)
The atomized slurry droplets contact the hot flue gas in the SD
chamber. The moisture in the lime slurry evaporates to cool the flue gas,
and the lime reacts with the acid gases in the flue gas to form calcium
salts. The SD chamber is designed to provide sufficient contact and
residence time to produce a dry product leaving the SD chamber. The
residence time in the chamber is typically 10 to 15 seconds. The
particulate exiting the SD contains fly ash plus calcium salts, water, and
unreacted lime. The simultaneous evaporation and reaction in the spray
drying process increases the moisture and particulate content of the flue
gas.
6-1
-------
Key design and operating parameters that can significantly affect SD
performance are SD outlet temperature, lime-to-acid gas stoichiometric
ratio, and slurry droplet size. The SD outlet temperature is controlled by
the amount of water in the slurry. More effective acid gas removal occurs
at lower temperatures, but the temperature must be kept high enough to
ensure the slurry and reaction products are adequately dried prior to
collection in the ESP. In addition, a minimum SD outlet temperature of
approximately 240°F is required to control agglomeration of PM and sorbent
by calcium chloride. 2
The stoich’iometric ratio is defined as the molar ratio of calcium in
the lime slurry fed to the SD to the theoretical amount of calcium required
to completely react with the HC1 and SO 2 in the flue gas at the inlet to the
SD. At a ratio of 1.0, the moles of calcium are equal to the moles of
incoming HC1 and SO 2 . However, because of mass transfer limitations,
incomplete mixing, differing rates of reaction (SO 2 reacts more slowly than
HC1), and the presence of other acid gases that react with calcium (e.g.
hydrogen fluoride, sulfur trioxide), more than the theoretical amount of
lime is generally fed to the spray dryer. Although not usually measured
during a compliance test, droplet size would be expected to affect SD
performance. Smaller droplet size increases the surface area for reaction
between lime and acid gases and increases the rate of water evaporation.
The amount of lime fed is generally controlled by one of two means. In
one approach, the lime slurry feed rate is controlled by an acid gas
analyzer/controller (generally based on SO 2 ) at the stack. As the outlet
acid gas concentration increases or decreases, the lime slurry feed rate is
accordingly raised or lowered, respectively, to maintain a specified outlet
acid gas concentration. The second approach uses a constant lime slurry
feed rate that is sufficient to react with peak expected acid gas
concentrations. Both systems are currently in use, although the system using
an analyzer/controller is more frequently encountered.
There are no significant mechanical or electrical differences between
ESP’s downstream of SD or £SP’s used alone. The PM control performance of
both is affected by the number of ESP fields, specific collection area, and
6-2
-------
particle size, resistivity, and migration velocity as discussed in
Section 2.0. Operating temperatures of ESP’s installed following a SD are
generally between 250 and 300°F, versus 350°F and above for conventional,
stand-alone ESP’s. A heavy rapper system is employed with an ESP downstream
of a SD because the PM adheres more strongly to ESP surfaces due to the
calcium chloride content.
6.2 SUMMARY OF TEST DATA
Section 6.2 presents the available emission data for MWC facilities
with SD/ESP systems. A description of each facility and a summary and
analysis of the emission data are provided for each facility. All of the
data presented in this section are based on short-term compliance testing
(generally several hours per run). A review of longer-term performance data
for acid gases from the Milibury MWC is presented in Appendix A.
6.2.1 Millbury 3 ’ 4 ’ 5
The Millbury Resource Recovery Facility in Millbury, Massachusetts,
consists of two identical mass burn combustor trains, each designed to
combust 750 tons per day of municipal solid waste (MSW). The MSW is
combusted on a Von Roll reciprocating inclined grate in a Babcock and Wilcox
waterwall combustor. Each boiler is rated to produce 190,000 lbs per hour
of superheated steam.
The combustion gases exiting the combustor enter a spray dryer/ESP
emission control system designed by Wheelabrator Air Pollution Control
Systems. Slaked lime, along with metered dilution water is injected into
the spray dryer vessel through two-fluid nozzles. The system is designed to
independently control the lime and dilution water feed rates. The lime
slurry feed rate is varied automatically to maintain a prescribed outlet SO 2
concentration. Manual adjustments of the set point are made to control SO 2
excursions. Dilution water is added to the lime slurry to reduce the flue
gas temperature, which is typically maintained at about 255°F.
The dry solids from the SD and fly ash are collected in a 3-field ESP.
The specific collection area of the ESP is 330 ft 2 per 1,000 acfm at a flue
6-3
-------
gas flow of about 160,000 acfm. Each ESP field is cleaned according to a
programmed mechanical rapping cycle.
In February 1988, emissions testing was performed by Rust International
Corporation to demonstrate compliance with permit conditions and by EPA to
assess performance of the SD/ESP on CDD/CDF. During the compliance testing,
the combustor and SD/ESP were operated normally. Flue gas at the SD/ESP
inlet and outlet of both units was analyzed for SO 2 and HC1 by manual
methods. At both SD/ESP outlets, flue gas samples were analyzed for PM,
metals (beryllium, lead, mercury, antimony, arsenic, cadmium, chromium,
copper, manganese, molybdenum, nickel, selenium, tin, titanium, vanadium,
and zinc), hydrogen flouride, NON, sulfuric acid mist, non-methane
hydrocarbons, reduced sulfur compounds, vinyl chloride, and volatile organic
compounds. Generally, these runs were conducted at each unit. CDD/CDF was
measured for six runs at the outlet of Unit 2 only. The EPA-sponsored
testing consisted of five sampling runs for CDD/CDF at the inlet to the
SD/ESP of Unit 2 conducted simultaneously with outlet compliance testing for
CDD/ CDF.
Acid gas data for Milibury are presented in Table 6-1. Long-term CEM
measurements of acid gases are described in Appendix A. The first five runs
in the table are from plant CEM data during CDD/CDF sampling. The plant
CEM’s were not certified at the time of this testing, however, and the SO 2
data and stoichiometric ratios are shown for comparison only. The remainder
of the data, three runs for each unit, were collected using manual methods
as part of the compliance test. Outlet SO 2 concentrations during the
six compliance test runs ranged from 41 to 75 ppmv at 7 percent 02. The
average outlet SO 2 concentration from Unit 1 was 23 ppm and the average
outlet SO 2 concentration from Unit 2 was 62 ppm. The corresponding SO 2
removal efficiencies for Units 1 and 2 averaged 73 and 79 percent,
respectively. Concentrations of HC1 at the ESP outlet ranged from 4.1 to
31 ppm at 7 percent 02. The average outlet HC1 concentration from Unit 1
was 23 ppm and the average outlet HC1 concentration from Unit 2 was 6 ppm.
The corresponding HC1 removal efficiencies were 97 and 99 percent,
respectively, for Units 1 and 2. Removal efficiencies for SO 2 and HC1 for
Unit 2 are higher than for Unit 1. This may be due to the difference in ESP
6-4
-------
TABLE 6-1. ACID GAS DATA FOR M!LLBURY
Test
Condition
Run
Numbera
ESP Inlet
Tempegature
( F)
Stoichiometric
Ratio
Acid Gas
Concentration
Acid Gas
Removal
Efficiency
$02
(%)
MCI
(ppmv.
Inlet
dry at 7% 0
SO 2
)
utlet
MCI
SO 2
MCI
Combustor Normal
SD/ESP Normal
2-1
2-2
2-3
2-4
2-5
255
255
255
253
250
1 • 3 b
12 b
16 b
14 b
085 b
108 C
95 c
236 C
215 c
216 C
NMd
NM
NM
NM
NM
171 c
178 C
4460
476 C
420 C
NM
NM
NM
NM
NM
835 e
815 e
807 e
768 e
788 e
Average (Unit 2 -
CDD/CDF tests)
254
13 b
174 c
NM
338 C
NM
80 • 3 d
- -
1-1
1-2
1-3
251
252
253
NA
HA
NA
138
274
206
847
697
767
45.6
74.9
41.2
7.67
31.3
31.0
67.0
72.7
79.9
99.1
95.5
96.0
Average (Unit 1)
252
HA
205
770
53.9
23.3
73.2
96.9
2-1
2-?
2-3
243
243
244
NA
NA
NA
318
254
315
794
704
593
59.0
52.1
73.4
9.73
4.40
4.11
81.5
79.5
76.7
98.8
99.4
99.3
Average (Unit 2)
243
NA
296
697
61.5
6.08
79.2
99.2
aRUfl Number consists of unit number followed by the run number on that unit.
bStoichiometric ratio estimated based on measured Lime feedrate and continuous $02 levels and average inLet MCI
concentration from compliance tests.
C 0 concentration in ppmv, as measured from plant CEll data on Unit 2. Plant CEll’s were not certified at the time of
th test. The 0., data colLected for these runs were on a wet basis and are not used to normalize the data.
d
NM Not Measured.
eRenloval efficiency calculated based on as-measured concentration (not included in overall average).
NA = Not Available. Lime feed rate was not measured.
U,
-------
inlet temperature (243°F for Unit 2 versus 252°F for Unit 1). However, due
to the lack of data on stoichiometric ratio, atomization quality, and other
factors such as relative mixing, the cause of this difference in performance
cannot be evaluated in more depth.
As shown in Figure 6-1, outlet SO 2 and HC1 emissions increased with
increasing inlet SO 2 concentration during testing. Both SO 2 and HCI removal
efficiencies appear relatively independent of inlet SO 2 concentration.
There is no apparent affect of inlet HC1 concentration on outlet SO 2 or HC1
emissions. The SD/ESP system at Millbury demonstrated better than 70
percent SO 2 removal efficiency and 95 percent HC1 removal efficiency at an
ESP operating temperature of 250°F.
The particulate data for Millbury are presented in Table 6-2. Outlet
PM concentrations from three runs at both units ranged from 0.0014 to 0.019
gr/dscf at 12 percent CO 2 . The average PM outlet concentration was
0.0018 gr/dscf for Unit 1 and 0.0083 gr/dscf for Unit 2. Thus, although the
use of spray drying increases PM loading to the ESP, the 3-field ESP at
Milibury operating at 225°F with an SCA of approximately 330 ft 2 /1,000 acfm
yielded outlet PM levels below 0.01 gr/dscf for five of six runs.
Table 6-3 presents the metals emissions data from three runs at each
unit (six Mercury runs at Unit 1). Metal emissions were similar at both
units. Outlet arsenic concentrations averaged 3.6 and 4.6 ug/dscm. Cadmium
and nickel emissions were similar, at 15 to 20 ug/dscm. Outlet chromium
emissions averaged 48 and 98 ug/dscm and lead emissions were about 300
ug/dscm. Compared to typical uncontrolled metals concentrations (Section
1.2), removal efficiencies for arsenic, cadmium, nickel, and lead were 98 to
99 percent. Chromium was removed at an efficiency of 95 percent. The
average mercury emissions of 570 ug/dscm at Unit 1 and 950 ug/dscm at Unit 2
were similar to typical uncontrolled mercury levels, which can range from
250 to 1,000 ug/dscm, suggesting that mercury was not removed by the SD/ESP
system.
The CDD/CDF data from Millbury Unit 2 are presented in Table 6-4.
Outlet CDD/CDF concentrations ranged from 40 to 103 ng/dscm at 7 percent 02
over six runs and averaged 59 ng/dscm. Five of the six runs were between 40
6-6
-------
80 -
Uniti o
U
_ 0 SO
70 0 HCI
Unit2
U SO 2
60- • HCI
E
50-
0
0
— 0
40-
C
0
U
0 0
0• “U-
C,
20 -
4
4 -
0
4-
0
.
0- I I I I I I I I I I I
130 150 170 190 210 230 250 270 290 310
Inlet sq Concentration (ppmv, dry at 7%
Figure 6-1. Outlet SO 2 and HCI concentrations as a function of
inlet SQ 2 concentration at Milibury.
-------
TABLE 6-2. PARTICULATE DATA FOR MILLBURY
Condition
Run
Number
a
ESP Inlet
Temperature
( F)
Flue Gas
Flow (acfni)
Outlet PM
Concentration
(gr/dscf
at 12% C0 2 )
Combustor = Normal
SD/ESP = Normal
1-1
1-2
1-3
254
256
255
167,540
163,280
154,100
0.0026
0.0014
0.0015
Average (Unit 1)
255
161,640
0.0018
2-1
2-2
2-3
240
240
240
155,560
168,540
163,510
0.0194
0.0029
0.0025
Average (Unit 2)
240
162,540
- .0083
aRUfl Number consists of unit number followed by the run number on that unit.
6-8
-------
TABLE 6-3 METALS EMISSIONS DATA FOR MILLBURY
0
a
Run Number consists of unit number foLLowed by the run number on that unit.
bMercury emissions for Unit 1 measured in Nay 1988. Unit 2 and other metaLs resuLts from
cND = Not detected. Considered as zero when caLculating averages.
d
NM = Not measured.
Test
Condition
Runa
Number
ESP InLet
Temperature
(OF)
Concentration
(gr/dscf at
12% C0 2 )
Outlet Concentration
As
Cd
(ug/dscm at
Cr
7%
Pb
02)
Hgb
Ni
Combustor = NormaL
1-1
254
0.0026
3.45
16.5
281
231
467
58.8
SD/ESP = NormaL
1-2
1-3
1-4
1-5
1-6
256
255
244
243
243
0.0014
O.OO 5
NM
MM
MW
4.03
3.18
NM
NM
NM
21.2
15.4
NM
NM
NM
8.9
6.1
NM
NM
NM
341
261
NM
MM
NM
606
709
426
641
542
7A0
MD ’
NM
NM
MM
Average (Unit 1)
249
0.0018
3.55
17.7
98.7
278
565
21.9
2-1
2-2
2-3
240
240
240
0.019
0.0029
0.0025
8.86
2.23
2.70
41.0
12.3
11.2
119
11.8
12.4
714
144
133
947
906
1,009
43.3
ND
ND
Average (Unit 2)
240
0.0083
4.60
21.5
47.7
330
954
14.4
February 1988 test.
-------
TABLE 6-4. CDD/CDF DATA FOR MILLBURY
CDD/
CDF
Removal
Test
Conditions
Run a
Number
ESP Inlet
Temperature
Inlet CDD/CDF
Concentration
(ng/dscm at 7%
02)
Concentration
(ng/dscm at 7% 02)
Efficiency
(%)
Combustor = Normal
103
51.1
SD/ESP = Normal
2-1
2-2
2-3
2-4
2-5
2-6
255
255
255
253
250
250
210
202
136
160
140k
NM ”
58.2
51.4
55.3
40.4
46.6
71.3
62.3
65.4
71.3
Average
253
170
59.2
64.3
aRUfl Number consists of unit number followed by the run number on that unit.
bNM = not measured.
-------
and 58 ng/dscm, with the sixth run at 103 ng/dscm. Inlet CDD/CDF
concentrations ranged from 136 to 210 ng/dscm at 7 percent 02 for five runs
and averaged 170 ng/dscm. Removal efficiencies were between 51 and 71
percent and averaged 64 percent. The range in ESP inlet temperature during
the six test runs was from 250 to 255°F. Thus, the SD/ESP system at
Milibury, operating at an ESP inlet temperature of 250°F with highly
efficient PM and gas control, has demonstrated the capability to reduce
moderate inlet CDD/CDF levels by greater than 60 percent.
6.2.2 Munich 6
The Munich North III MWC facility consists of two mass burn combustors
each designed to burn 530 tons/day of municipal waste and 288 tons/day of
clarified sludge. The emission control system consists of independent
Deutsche Babcock Anlagen (DBA) SD/ESP systems for each combustor. The lower
inlet section of the SD is a cyclonic preseparator where approximately 50 to
70 percent of the fly ash is removed from the flue gas. From the
preseparator section, the flue gas flows upward through the SD where
two-fluid nozzles spray lime slurry into the gas stream. The lime feed rate
and amount of dilution water for the slurry are adjusted by controllers
based on the outlet HC1 concentration and SD outlet temperature,
respectively. Flue gas from the SD is routed through a 2-field rigid-frame
ESP at a flow of approximately 147,000 acfm at 300 0 F. The SCA of the ESP is
unavailable. The ESP exhaust is routed through an ID fan and a concrete
stack.
A test program was conducted at the Munich North facility in May 1984
to demonstrate the ability of the SD/ESP system to control pollutants to
levels acceptable in the U.S. at that time. Target emission levels of
0.02 gr/dscf for PM, 30 to 100 ppm or 70 percent removal for SO 2 , and 30 to
50 ppm or 90 percent removal for HC1 were established, based on regulations
in existence for California, New Jersey, and Connecticut. During these
tests, only MSW was fired. The SD outlet temperature and the set point for
the outlet HC1 concentration were varied during the tests. Flue gas at the
SD/ESP inlet and outlet was analyzed for PM, HC1, and SO 2 . Sampling was
also conducted at the SD/ESP outlet for selected metals including arsenic,
beryllium, cadmium, chromium, lead, and nickel.
6-11
-------
Acid gas data for four test runs at Munich are presented in Table 6-5.
Outlet SO 2 concentrations ranged from 13 ppm at 7 percent 02 at an ESP inlet
temperature of 300°F to 37 ppm at a temperature of 330°F. The SO 2 removal
efficiency was 88 percent at 300°F and about 70 percent for the two runs at
330°F. The SO 2 removal efficiency could not be evaluated for Run 5 at an
ESP inlet temperature of 314°F. Outlet HC1 concentrations ranged from 4.9
ppm at 7 percent 02 at an ESP inlet temperature of 300°F to 44 ppm at a
temperature of 330°F. The HC1 removal efficiency was 99 percent at an ESP
inlet temperature of 300°F and about 94 percent at 300°F. HC1 removal
efficiency could not be evaluated for Run 5 at 314°F.
Figure 6-2 presents a plot of SO 2 and HC1 removal efficiency as a
function of stoichiometric ratio. As the ratio varied from a low of 0.89 at
an ESP inlet temperature of 330°F to a high of 2.1 at 300°F, both SO 2 and
HC1 removal efficiency increased. As shown in Figure 6-3, a plot of SO 2 and
HC1 removal efficiency as a function of ESP inlet temperature, decreasing
the temperature also increased performance. However, increasing
stoichiometric ratio corresponded with decreasing temperature and as a
result, the highest removal efficiency was observed for both SO 2 and HC1 for
Run 2, which had the lowest temperature (300°F) and the highest
stoichiometric ratio (2.1). There were no observable effects of inlet acid
gas concentration on outlet acid gas concentration or removal efficiency.
Particulate data from the four runs at Munich are presented in
Table 6-6. Outlet PM concentrations ranged from 0.0065 to 0.018 gr/dscf at
12 percent CO 2 . The average PM concentration for all four runs was
0.010 gr/dscf with removal efficiencies exceeding 99 percent. There is no
apparent effect of increased inlet PM from the SD or changes in the lime
feed rate on the PM control performance of the ESP. There was no observable
effect on PM performance due to £SP inlet temperature changes, either.
Outlet metals concentrations are presented in Table 6-7. These data
are from a particulate sizing sample collected continuously over 40 hours.
Because combustor and ESP operating conditions varied over that period,
these data should be considered as worst-case only. Based on typical
uncontrolled metals concentrations, arsenic and lead were removed at greater
than 99 percent efficiency. Cadmium was removed at 98.6 percent efficiency.
6-12
-------
TA8IE 6-5. ACID GAS DATA FOR NUNICH
(A)
Test
Condition
Run
Number
ESP Inlet
Tempsrature
( F)
Stoichiometric
Ratio
Acid Gas Concentration
(pp4nv. dry at 7% 0,)
Acid Gas
RemovaL
Efficiency
(X)
SO 2 HC(
SO 2
InLet OutL t
HCL SO 2
HCL
Combustor = NormaL
2
300
2.1
105
588
12.8
4.9
87.8 99.2
SD/ESP = Very low
SD
out tet 0 temperature
(266 F)
Combustor Normal
3
330
1.6
0.9
52
123
622
656
14.7
36.5
34.3
44.1
71.5 94.5
70.4 93.3
SD/ESP = Low SD
4
330
out Let 0 temperature
(293 F)
Average
330
1.2
87
639
25.6
39.2
71.0 93.9
Combustor Normal
5
314
NRd
WRd
NRd
20.4
24.3
- - --
SD/ESP = ormaL
(320 0 F)
aConstant lime slurry feed rate.
bout Let HCL analyzer/controLler used to control Lime sLurry feed rate. Out Let HCI controller set to achieve 30 mg/Nm 3 ,
wet, at 11 percent 02 (approximateLy 32 ppmv, dry, at 7 percent 02)
cOuttet HCL anaLyzer/controLLer used to control time slurry feed rate. Outlet HCL controLLer set to achieve 15 mg/Nm 3 ,
wet, at 11 percent 02 (approximateLy 16 ppmv, dry, at 7 percent 02).
= Not reported. Oxygen data unavaiLable to correct vaLues to 7 percent 02. FLue gas flow rate
unknown, preventing calculation of stoichiometric ratio.
-------
Figure 6-2. SO 2 and HCI removal efficiency as a function of
stoichiometric ratio at Munich.
0
S0
0
HCI
C
S
U
I-
S
a.
>1
U
C
0
U
‘U
S
0
E
— S
S
0
U
4
100
98
96
94 -
92 -
90 -
88 -
86 -
84-
82 -
80 -
78 -
76 -
74-
72 -
70 -
C
C
0
n
0
I I I I I I I
0.8 1 1.2 1.4 1.6 1.8 2
. Stolch lom.trlc Ratio
-------
100 -
98 -
96 -
72 -
290
ro
so 2
0
HCI
0
310
ESP Inlet Temperature CF)
0
0
0
330
Figure 6-3. SO 2 and HCI removal efficiency as a function of
ESP inlet temperature at Munich.
-------
TABLE 6-6. PARTICULATE DATA FOR MUNICH
analyzer controller used to control lime slurry feed rate.
30 mg/Nm wet, at 11 percent 02 (approximately 32 ppmv, dry
ana1yzer controller used to control lime slurry feed rate.
15 mg/Nm , wet at 11 percent 02 (approximately 16 ppmv, dry
a.,
a.,
ESP
Inlet
Flue Gas
PM Concentration
PM
Removal
Test
Condition
Run
Number
Temp 8 rature
( F)
Flow
(acfm)
(gr/dscf atI2% CO .L
Inlet 0utl t
Efficiency
(%)
Combustor = Normal
2
300
155,400
2.74
0.0090
99.7
SD/ESP = Very low
SD
out1 t emperature
(266 F)
Combustor = Normal
3
330
146,400
2.22
0.0065
99.7
SD/ESP = Low SD
4
330
154,800
3.72
0.0082
outl 8 t emperature
(293 F)
Average
330
150,600
2.97
0.0074
99.8
Combustor = Normal
5
314
144,900
NMd
0.0175
-
SO/E P = Normal
(320 F)C
aConstant lime slurry feed rate.
boutiet HC1
to achieve
COutlet HC1
to achieve
Outlet HC1 controller set
at 7 percent °2
Outlet HC1 controller set
at 7 percent 02).
dNM = Not measured. Sample not collected at inlet for Run 5.
-------
TABLE 6-7. METALS EMISSIONS DATA FOR MUNICH
Test Condition
Run
ESP Inlet
Temp 8 rature
( F)
Outlet PM
Concentration
(gr/dscf at
12% C0 2 )
Outlet Concentrationa
As
Cd
(u /dscm
Cr
at
7%
02)
Pb
Ni
Combustor = Normaib
1
300
0.031
1.34
25
3,020
260
1,400
SD/ESP = Normal
aResults based on approximately 40 continuous hours of sampling with a particle sizing train. Particle
fractions and impingers were analyzed for metals.
bjhe system was not kept at a single operating condition during testing and emissions should be
considered as worst—case” only.
-------
The outlet chromium and nickel levels are unusually high. As a result,
removal efficiencies for both chromium and nickel, based on typical
uncontrolled levels, were near zero. The ratio of chromium concentration to
nickel concentration is 2.16, similar to the ratio of chromium to nickel in
18/8 stainless steel (2.25), suggesting that sample contamination may have
occurred from the stainless steel in the sampling train. 7
6.2.3 Portland 8
The Greater Portland Resource Recovery Facility in Portland, Maine,
consists of two L. & C. Steinmuller mass burn waterwall combustors, each
designed to combust 250 tons/day of MSW. The flue gas exiting the conibustor
and heat recovery system flows through a SD/ESP system manufactured by Belco
Pollution Control Corporation. The flue gas first flows through a cyclone
preseparator that removes approximately 50 percent of the fly ash. The flue
gas then flows upward through the SD absorption section. In the absorption
section, lime slurry is injected through two-fluid nozzles to remove acid
gases. The lime slurry feed rate is continuously controlled based on the
outlet SO 2 level and the dilution water rate is continuously controlled
based on the SD outlet temperature. The flue gas flow rate exiting the SD
is typically 56,000 acfm at 290°F. The flue gas then passes through a
5-field rigid ESP. The SCA of the ESP is unavailable.
In September 1988, compliance testing was conducted at both the North
and South units. Both combustors were at normal operating conditions, but
the South unit ESP had one field continuously shorted out and other fields
intermittently shorted out during testing. The North ESP operated normally
throughout testing. At the SD/ESP inlet and outlet, SO 2 and HC1 were
sampled. At the SD/ESP outlet only, PM, metals (lead, cadmium, chromium),
CDD/CDF, and NO were measured. Because of the shorted ESP field on the
South unit, damaged PM and metals samples, and improperly calibrated
equipment for the HC1 samples, only the SO 2 data from both units and the
CDD/CDF data from the North unit are reported here.
Acid gas data for Portland are presented in Table 6-8. Outlet SO 2
concentrations for three runs at each of the North and South units ranged
from 15 to 35 ppm with the exception of Run 1 at the South unit at 112 ppm.
6-18
-------
TABLE 6-8. ACID GAS DATA FROM PORTLAND
Test
Condition
Run
Number
a
ESP Inlet
Temp 8 rature
( F)
Lime Feed
Rate
(lb/hr)
SO
(ppm ,
Concentration
dry at 7% 0 9 L
SO Removal
Ef iciency
(%)
Inlet
Outlet
Combustor = Normal
N-i
285
241
441
31.5
92.9
SD/ESP = Normal
N-2
N-3
285
288
269
304
331
195
34.5
29.2
89.6
85.3
Average (North Unit)
286
271
322
31.7
89.3
S-i
S-2
S-3
315
300
286
292
280
293
437
194
211
112
20.1
14.6
74.4
896
93.1
Average (South Unit)
300
288
281
48.9
85.7
‘.0
aRUfl Number consists of unit identification (N = north, S = south) followed by the run number.
-------
The average outlet SO 2 concentration was 32 ppm at the North unit and 49 ppm
at the South unit. Removal efficiencies for SO 2 were similar for both
units, ranging from 74 to 93 percent and averaging 89 percent at the North
unit and 86 percent at the South unit.
The effect of ESP inlet temperature on SO 2 removal efficiency is shown
in Figure 6-4. Removal efficiencies were lowest at 315°F. During the tests
at 300 and 285 0 F, SO 2 removal efficiencies were similar. There was no
apparent effect of lime feed rate on SO 2 removal of emissions, although the
lime feed rate varied from 240 to 300 lb/hr.
Table 6-9 presents CDD/CDF data for Portland. Outlet CDD/CDF
concentrations at the North unit ranged from 61.9 to 263 ng/dscm at 7
percent 02 over three runs and averaged 173 ng/dscm. There was no observed
effect of ESP inlet temperature on CDD/CDF outlet concentration because
there was only a 3°F variation at the ESP inlet.
6.3 SUMMARY OF PERFORMANCE
Section 6.2 discussed SD/ESP performance for individual SD/ESP systems.
This section combines the data from these facilities to examine the
relationship between key design and operating variables and SD/ESP
performance.
6.3.1 Acid Gas
Acid gas performance for SD/ESP systems was relatively consistent for
the three existing facilities. Millbury, Munich, and Portland each achieved
greater than 70 percent SO 2 removal efficiency during all test conditions.
The corresponding SO 2 emissions were less than 112 ppm at 7 percent °2
During testing at Munich and Portland when ESP inlet temperatures were 300°F
or less, SO 2 removal efficiencies exceeded 85 percent and SO 2 emissions were
32 ppm or less. The stoichiometric ratio at Munich was 2.1. At Millbury,
$02 removal efficiencies averaged 76 percent, and outlet SO 2 emissions
averaged 58 ppm at 7 percent 02, even though the ESP outlet temperature was
below 255°F. The stoichiometric ratio was estimated to be 1.3. HC1 removal
efficiency at Munich and Millbury was above 96 percent for the tests at
300°F or less. The maximum outlet HC1 concentration at these conditions was
31 ppm at 7 percent 02, with averages less than 23 ppm.
6-20
-------
100
98
96
94
H
: _ _
Li North Unit
78 U South Unit
76
74 I
280 290 300 310 320
ESP Inlet Temperature CF)
Figure 6-4. SO 2 removal efficiency as a function of
ESP inlet temperature at Portland.
-------
TABLE 6-9. CDD/CDF DATA FOR PORTLAND
Test
Condition
Run
Numbera
ESP Inlet
Temperature
(°F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Combustor
N-I
285
263
Normal
N-2
285
62
SD/ESP =
N-3
288
195
Normal
Average
286
173
aRUfl Number consists of unit identifcation (N = north, S = south) followed
by the run number.
6-22
-------
Analysis of these data suggests that the most important of the
variables affecting acid gas removal efficiency is SD outlet temperature.
At Munich and Portland, SO 2 removal efficiencies generally increased during
tests conducted at lower temperatures. Similar phenomena were observed for
HC1 at Munich. An approximate stoichiometric ratio of 1.3 at Millbury
yielded performance similar to Munich at a ratio of 1.2, but the temperature
at Munich was 80°F higher. This suggests that there is no effect of
temperatures below 300 0 F on acid gas removal.
Stoichiometric ratio apparently has a secondary effect on acid gas
control with a SD/ESP. At Munich, SO 2 and HC1 removal efficiencies
increased when both ESP outlet temperature decreased and stoichiometric
ratio increased. Whether the temperature or the stoichiometric ratio had a
greater effect could not be ascertained. At Milibury, the stoichiometric
ratio was lower than at the 300°F tests at Munich (1.3 versus 2.1), and both
SO 2 and HC1 removal efficiencies were lower, indicating that the lower ratio
at Milibury may have led to the decreased performance. However, changes in
lime feed rate at Portland of 60 lb/hr (the stoichiometric ratio could not
be calculated because no HC1 data were available), caused no change in SO 2
removal efficiency or outlet concentration. These data suggest that
increasing the stoichiometric ratio up to roughly 2 can increase acid gas
removal across a SD/ESP system. However, further increases in
stoichiometric ratio do not appear to enhance acid gas removal.
Data from SD/FF systems (Section 7.3.1) show that significant acid gas
removal occurs after the spray dryer in the fabric filter. In this case,
increasing the stoichiometric ratio increases the amount of unreacted
sorbent in the fabric filter for secondary removal. An ESP does not provide
similar opportunities for secondary reaction of acid gases due to gas-solid
contact between the flue gas and collected particulate. As a result,
virtually all the acid gas removal in a SD/ESP system must occur in the
spray dryer, and the amount of removal across the spray dryer is apparently
limited.
6-23
-------
Based on these data, average removal efficiencies of 75 percent for SO 2
and 95 percent for HC1 are achievable during short-term compliance-type
tests from SD/ESP systems operating at less than 300°F and stoichiometrjc
ratios of about 2. Outlet SO 2 and HC1 emissions of 60 ppm and 30 ppm,
respectively, are achievable at these conditions. Long-term acid gas
control levels are discussed in Appendix A.
6.3.2 Particulate Matter
Particulate removal was similar for Millbury and Munich. Each unit
averaged 0.010 gr/dscf at 12 percent CO 2 or less. Increasing the lime feed
rate at Munich did not cause outlet PM concentrations to increase. Thus,
average PM emissions of 0.01 gr/dscf at 12 percent CO 2 or less are
achievable by this technology.
6.3.3 Metals
Although uncontrolled metals concentrations were not measured at any of
the SD/ESP facilities, estimated removal efficiencies for arsenic, cadmium,
lead and nickel were 98 to 99 percent. Chromium removal was estimated at 95
percent at Milibury. Although unusually high chromium and nickel emissions
were measured at Munich, these data do not appear to be representative and
may have resulted from contamination by the stainless steel sampling probe.
Measured outlet mercury emissions at Milibury were relatively high and are
comparable to inlet mercury levels measured at other MWC’s, indicating that
little or no mercury removal occurred.
Metals removal efficiencies by SD/ESP systems of 98 percent are
achievable for arsenic, cadmium, lead, and nickel. Chromium removal of
95 percent is achievable. Mercury is not effectively removed by a SD/ESP
system.
6.3.4 CDD/CDF
CDD/COF emissions were measured at Milibury and Portland. Outlet
concentrations were relatively consistent at Milibury, but varied by up to a
factor of 4 at Portland. The outlet CDD/CDF concentrations at Millbury were
between 40 and 58 ng/dscm at 7 percent 02 for five of the six runs and 103
ng/dscm at 7 percent 02 during the sixth run. Removal efficiencies for
6-24
-------
CDD/CDF at Millbury were measured between 51 and 71 percent and did sLow any
dependence on inlet CDD/CDF concentrations. The average ESP inlet
temperature at Millbury was between 250 and 255°F during each test. Outlet
CDD/CDF concentrations at Portland were 62, 195, and 263 ng/dscm at 7
percent 02. The ESP inlet temperature was about 35°F higher at Portland
than at Milibury, suggesting that increased temperature may have affected
the results at Portland. CDD/CDF removal efficiencies across the SD/ESP at
Portland could not be calculated due to lack of inlet samples.
Because of the limited amount of available data, determination of a
consistently achievable performance level is not possible. Based on the
Milibury data, CDD/CDF removal efficiency across a SD/ESP operating at the
gas temperatures of 250 to 255°F is estimated at 50 to 75 percent. The data
also suggest that outlet COD/CDF emissions may increase at higher ESP inlet
temperatures.
6-25
-------
6.4 REFERENCES
1. Masters, K. Spray Drying Handbook, Third Edition. George Godwin
Limited. London. 1979. pp. 24 - 27, 150.
2. Brown, B., et al,. (Joy Technologies, Inc.), Dust Collector Design
Considerations for MSW Acid Gas Cleaning Systems. Presented at:
7th EPA/EPRI Particulate Symposium. Nashville, Tennessee.
March 1988, p. 4.
3. Entropy Environmentalists, Inc. Municipal Waste Combustion
Multi-Pollutant Study, Summary Report, Wheelabrator Millbury, Inc.,
Milibury, Massachusetts. Prepared for U.S. Environmental Protection
Agency. Research Triangle Park, North Carolina. EMB Report No.
88-MINO-07A. February 1989.
4. Entropy Environmentalists, Inc. Emissions Testing Report, Wheelabrator
Milibury, Inc. Resource Recovery Facility, Millbury, Massachusetts,
Unit Nos. 1 and 2, February 8 through 12, 1988. Prepared for Rust
International Corporation. Reference No. 5605-B. August 5, 1988.
5. Entropy Environmentalists, Inc. Stationary Source Sampling Report,
Wheelabrator Milibury, Inc., Resource Recovery Facility, Milibury,
Massachusetts, Mercury Emissions Compliance Testing, Unit No. 1, May 10
and 11, 1988. Prepared for Rust International Corporation. Reference
No. 5892-A. May 18, 1988.
6. Hahn, J. 1., et al., (Cooper Engineers) and J. A. Finney, Jr. and B.
Bahor (Belco Pollution Control Corp.). Air Emissions Tests of a
Deutsche Babcock Anlagen Dry Scrubber System at the Munich North
Refuse-Fired Power Plant. Presented at: 78th Annual Meeting of the
Air Pollution Control Association. Detroit, Michigan. June 1985.
7. Telecon. M.A. Vancil, Radian with J. L. Hahn, Ogden Projects, inc.,
July 1989. Metals results from tests at the Munich MWC.
8. Engineering Science, Inc. A Report on Air Emission Compliance Testing
at the Regional Waste Systems, Inc. Greater Portland Resource Recovery
Project. Prepared for Dravo Energy Resources, Inc. Pittsburg,
Pennsylvania. March 1989.
6-26
-------
7.0 SPRAY DRYING FOLLOWED BY A FABRIC FILTER
Section 7.0 describes the technology and performance of spray dryer
systems with a FF for PM control. In Section 7.1, SD/FF operation and
design is described. Section 7.2 presents descriptions of facilities with
emissions data, and summarizes the available data from each facility. In
Section 7.3, the performance of SD/FF systems relative to the control of
acid gases, PM, metals, and CDD/CDF emissions is discussed.
7.1 PROCESS DESCRIPTION
Spray dryers were originally applied to MWC’s to control acid gas
emissions. A description of the spray drying process and factors affecting
performance are discussed in Section 6.1. Fabric filter design and
operation is discussed in Section 4.1. There is little difference between
FF operation as discussed in Section 4 for duct sorbent injection systems
and FF operations following a SD. The particulate following a SD may
contain more moisture than a FF following a dry injection system, but this
depends on the types of cooling used. Using humidification to cool flue gas
in dry sorbent injection systems will generate a particulate with similar
moisture content to that following a spray dryer. Alternatively, if heat
recovery or mixing with ambient air is used to cool flue gas in a dry
injection system, the particulate will have less moisture than that
following a spray dryer.
7.2 SUMMARY OF TEST DATA
Section 7.2 presents the available emissions data for MWC facilities
with SD/FF systems. A description of each facility and a summary and
analysis of the emissions data are provided for each facility.
The effects of stoichiometric ratio, FF inlet temperature, and
uncontrolled acid gas concentrations on acid gas removal are discussed ‘in
each section. The effect of air-to-cloth ratio on PM removal is also
discussed. Finally, the effects of inlet CDD/CDF concentration and FE inlet
temperature on CDD/CDF removal are discussed.
7-1
-------
7.2.1 Biddeford 1 ’ 2
Maine Energy Recovery Company’s (MERC) York County Waste-to-Energy
Facility in Biddeford, Maine, is designed to combust 300 tons per day of RDF
in each of two identical Babcock and Wilcox “controlled combustion zone”
boilers with traveling grates. The RDF has a nominal top size of 4 inches.
Approximately 105,000 lbs per hour of steam are generated by each unit. In
each combustor, heat for steam generation is recovered in the furnace
waterwalls, superheater, economizer, and combustion air heater sections. At
the air heater exit, the flue gas temperature is approximately 400°F.
Emissions from each boiler are controlled by a cyclone, spray dryer,
and fabric filter system. The combustion gases from the air heater enter a
cyclone-type mechanical dust collector which removes large particulate.
Next, an alkaline spray dryer is used to control acid gas emissions. The
system is controlled to remove SO 2 to an outlet setpoint concentration of
30 ppm. The spray dryer outlet temperature is typically 280 to 300°F.
The lime-to-SO 2 stoichiometric ratio and the flue gas temperature at
the exit of the spray dryer can be controlled separately. The lime, which
is introduced as slurry through a rotary atomizer, is diluted with water
before entering the reaction vessel to achieve the desired SO 2 outlet
concentration and temperature reduction. The SO 2 concentration at the stack
is monitored and used to control the slurry feed rate.
Particulate in the gas stream is collected by the fabric filter. The
fabric filter is a pulse-jet design with fiberglass bags. The design net
air-to-cloth ratio is approximately 5.2 acfm/ft 2 at a flue gas flow rate of
72,000 acfm at 280°F and 15 percent moisture. The pressure drop across the
fabric filter is about 8 inches water column. The fabric filter has six
compartments, with 126 bags in each compartment. Five compartments filter
flue gas while one compartment is being cleaned in a continuous cycle. The
total time to complete a fabric filter cleaning cycle is about 18 minutes.
Flue gas from each fabric filter is exhausted through a 244-foot stack that
is comon to both trains.
In December 1987, emissions testing was performed by EPA to compile
emissions data from an ROE combustor with a SD/FF emission control
system to support regulations development for MWC’s under Section 111 of the
7-2
-------
Clean Air Act. The combustor and SD/FE were at normal operating conditions
during testing. Three runs were performed at the SD/FF inlet of Unit A and
the common SD/FE outlet location. Flue gas samples were collected and
analyzed for CDD/CDE, PM, SO 2 , HC1, NON , cadmium, chromium, arsenic, lead,
and mercury. Additionally, HC1 was measured at the spray dryer outlet
(SD/FE midpoint) during testing. Process data, including the lime slurry
feed rate and lime concentration, were monitored during testing.
Acid gas data from three test runs are presented in Table 7-1. At a
consistent average SD outlet temperature of 278°F, outlet SO 2 concentrations
ranged from 13.6 to 30.5 ppm at 7 percent 02 and averaged 23 ppm. The
corresponding SO 2 removal efficiency ranged from 66 to 89 percent and
averaged 76 percent. Relatively low inlet SO 2 concentrations of 86 to
129 ppm were measured. Outlet HC1 concentrations ranged from 3.4 to 9.7 ppm
at 7 percent 02 and averaged 5.8 ppm. The corresponding removal
efficiencies were between 98.1 and 99.4 percent, averaging 98.9 percent.
Figures 7-1 and 7-2 present graphs of SO 2 and HC1 removal efficiency,
respectively, as functions of stoichiometric ratio. As the stoichiometric
ratio increased from 1.7 (Run 1) to 3.9 (Run 3), SO 2 removal efficiency
increased from 65 to 90 percent and HC1 removal efficiency increased from 98
to 99.4 percent. Thus, especially for SO 2 , but also for HC1, increasing the
stoichiometric ratio increased acid gas removal efficiency. Inlet acid gas
concentrations did not affect acid gas removal at Biddeford.
In Table 7-2, PM data are presented. Outlet PM concentrations ranged
from 0.0095 to 0.019 gr/dscf at 12 percent CO 2 and averaged 0.014 gr/dscf.
The PM removal efficiency averaged 99.5 percent. The SD/FF system at
Biddeford operated with a net air-to-cloth ratio of about 5.7 acfm/ft 2 .
In Table 7-3, metals data are presented. Except for lead, none of the
metals sampled were detected at the SD/FE outlet. The outlet lead
concentrations averaged 159 ng/dscm at 7 percent 02 Removal efficiencies
for all the metals were at least 99.4 percent, roughly equal to the PM
removal efficiency.
7-3
-------
TABLE 7-1. ACID GAS DATA FOR BIDDEFORD
Test
Condition
Run
Number
FF InLet
Tempgrature
( F)
Stoichiometric
Ratio
Acid Gas Concentration
Intermediate Nd
RemovaL
Efficiency
(percent)
Overall Acid
Gas Removal.
Efficiency
(percent)
Inlet
SO 2
(ppmv. dry at 7% 02)
Midpoint Outlet
HCL HCL SO 2
HCL
SD FF
SO 2 lid.
Conibustor NormaL
1
277
1.7
88.7
509 65
30.5
9.68
26.7 85.1
65.6 98.1
SD/FF = NormaL
2
3
278
279
2.8
3.9
86.1
129
634 8.5
603
23.6
13.6
4.46
3.40
90.1 47 • 5 b
99.2 -210
72.6 99.3
89.4 99.4
Average
278
2.8
101
582 24.9
22.6
5.84
72.0 44.2
75.9 98.9
avatue considered questionabLe.
bNegative NC . removal efficiency suggests that Little or no NCL removal occurred across the FF. Midpoint and outLet vaLues within
2 ppm. Considered as zero in evaluating average removaL efficiency.
-------
100 -
95 -
-S
65 j 24 2:8 3:2 36 4
Stoichiometric Ratio
Figure 7-1. SO 2 removal efficiency as a function of
stoichiometric ratio at Biddeford.
-------
100 -
99.9
99.8 -
99.7 -
99.6 -
99.5-
C
0 99.4-
C.)
99.3-
a
99.2-
C . )
99.1-
99-
W 98.9-
0
98.8-
•-1
c; .. 98.7-
98.6-
U
X 98.5-
98.4 -
98.3 -
98.2 -
98.1 -
98 -
1.6 2 2.4 2.8 3.2 3.6 4
Stoichiometric Ratio
Figure 7-2. HCI removal efficiency as a function of
stoichiometric ratio at Biddeford.
-------
TABLE 7-2. PARTICULATE DATA FOR BIDDEFORD
—1
Test
Condition
Run
Number
FF Inlet
Temp 8 rature
( F)
Flue Gas
Flow
(acfm)
PM Concentration
PM
Removal
(grldscf
Inlet
at
12% C0 2 1
Outlet
Efficiency
(%)
Combustor = Normal
SD/FF = Normal
1
2
3
277
278
279
77,200
77,800
83,000
3.23
2.85
3.53
0.0095
0.014
0.019
99.7
99.5
99.4
Average
278
79,300
3.20
0.014
99.5
-------
TABLE 7-3. METALS DATA FOR BIDDEFORD
Test Condition
Run
Number
FF InLet
Temp 8 rature
C F)
OutLet PM
Concentration
(grldscf at
12% C0 2 )
As
Inlet Concentration
(ug/dscm t 7% 0 )
Cd Cr P
Hg
OutLet Concentration
(ug/dscun at 7% 0 j_
As Cd Cr Pb H
Re
As
movat Effici 8 ncy
(Percent)
Cd Cr Pb
Hg
Combustor = NormaL
1
277
0.0095
474
1,016
2,360
26,380
493
NDb ND ND 146 ND
100
100 100 99.5
100 100 99.4
100
100
SD/ESP = NormaL
2
3
278
279
0.014
0.019
527
527
1,058
1,268
2,671
3,205
27,436
28,240
324
351
ND ND ND 155 ND
ND ND ND 177 ND
100
100 100 99.4
100
Average
278
0.014
509
1,114
2,745
27,352
389
ND ND ND 159 ND
100
100 100 99.4
100
8 Removat efficiencies reported as 100 percent when compound not detected at outLet.
bND = not detected. Considered as zero in caLcuLating and removaL efficiency.
-------
Table 7-4 presents the CDD/CDF data. Outlet CDD/CDF concentrations
ranged from 3.5 to 5.2 ng/dscm at 7 percent 02 and averaged 4.4 ng/dscm.
Inlet CDD/CDF concentrations were relatively consistent, ranging from 856 to
987 ng/dscm at 7 percent 02, with an average of 903 ng/dscm. The
corresponding removal efficiencies averaged 99.5 percent. Thus, the SD/FE
at Biddeford operating at 278°F, can reduce inlet CDD/CDF concentrations of
nearly 1000 ng/dscm to 5.0 ng/dscm CDD/CDF or less at the outlet, with a
removal efficiency exceeding 99 percent.
7.2.2 Comerce 3 ’ 4
The Commerce Refuse to Energy Facility, in Commerce, California,
consists of one mass burn waterwall Foster-Wheeler combustor with a Detroit
Stoker grate. The design capacity is 380 tons/day of solid waste, generally
commercial waste. Emissions are controlled by Exxon’s Thermal DeN0 system,
and a Teller/American Air Filter (AAF) spray dryer and fabric filter. The
Thermal DeNO system injects ammonia into the upper combustion chamber to
reduce NO emissions to elemental nitrogen and water. Following heat
recovery, the flue gases then enter a cyclonic separator to remove large
particles before entering the up-flow SD. In the SD, lime slurry is
injected through two-fluid nozzles at a design feed rate of 600 lb/hr of
lime. A residence time of 10 seconds is provided in the SD vessel. The
design flue gas temperature at the SD outlet is 270°F. Tesisorb® is
injected into the flue gas through a venturi after leaving the SD to enhance
collection performance and to assist conditioning of the filter cake. The
FE uses reverse air cleaning with eight compartments each containing 156
fiberglass bags. The design net air-to-cloth ratio is 2 acfm/ft 2 with two
compartments off-line and a flue gas flow of about 85,000 acfm. The flue
gas leaves the FE and exits through a 150-foot high stack.
Several major test programs have been conducted at the Commerce MWC
facility. In May and June 1987, compliance testing was performed at the
facility. The combustor and DeN0 /SD/FF system were generally operated at
normal conditions. The combustor routinely fired mainly commercial refuse.
One run (17) fired residential refuse. Flue gas at the SD/FE inlet and
outlet were analyzed for SO 2 , HC1, HF, PM, CDD/CDF, and metals (antimony,
7-9
-------
— 4
cD
TABLE 7-4.
CDD/CDF DATA FOR BIDDEFORD
CDD/CDF
Test
Condition
Run
Number
FF Inlet
Temp 8 rature
( F)
Inlet COD/CDF
Concentration
(ng/dscm at 7%
02)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Removal
Efficiency
(percent)
Combustor = Normal
1
277
856
4.45
99.5
SD/FF = Normal
2
3
278
279
866
987
5.18
3.51
99.4
99.6
Average
278
903
4.38
99.5
-------
arsenic, beryllium, cadmium, chromium, copper, lead, mercury, nickel,
selenium, silver, thallium, and zinc). Inlet PM and SO 2 samples were not
collected simultaneously with the outlet samples. NO was measured at the
FE outlet. The lime feed rate during this test was not measured, but was
estimated to be at 600 lb/hr, the design rate.
In June 1988, a series of tests were conducted to optimize NO removal
efficiency by the Thermal DeNO system at Commerce. These results, along
with results of other tests of NO removal systems, are presented in a
separate EPA document entitled “Municipal Waste Combustors--Background
Information for Proposed Standards: Control of NO Emissions.” 6
In July 1988, additional testing was conducted at the Commerce facility
as part of the California Waste Management Board’s Waste-to-Energy
Demonstration Program. Emissions were measured while firing commercial
refuse and a mixture of residential and commercial refuse. Three runs were
conducted with each feed type. The objectives of the program were to fully
characterize the incoming waste stream, the air pollution control equipment
performance and emissions, and the ash residue. The DeNO /SD/FF system was
operated normally during testing. At the SD inlet and FF outlet, samples
were collected and analyzed for SO 2 , HC1, PM, metals (arsenic, cadmium,
chromium, lead, mercury, nickel, and others), CDD/CDF, other organics, HF,
and NOR. Data from both test programs are presented below.
Acid gas data are presented in Table 7-5. Outlet SO 2 concentrations
ranged from 0.78 to 8.0 ppm at 7 percent 02 over nine test runs. Removal
efficiencies, recorded during the 1988 testing, were between 93.4 and
99.5 percent and averaged 97.9 percent. Outlet HC1 concentrations ranged
from 3.2 to 11.1 ppm at 7 percent 02 over eight runs. The average removal
efficiency during two test runs in 1987 was 99.1 percent and was
98.9 percent during the 1988 testing. Stoichiometric ratio data are
estimated based on an assumed lime feed rate (600 lb/hr) and the measured
inlet SO 2 and HC1 concentrations for the 1988 tests. No consistent effect
of stoichiometric ratio on performance was observed. Fabric filter inlet
temperatures during the test were relatively constant at 273 to 297°F.
There was no observed difference in SD performance while burning either
residential or commercial refuse.
7-11
-------
TABLE 7-5. ACID GAS DATA FOR COMMERCE
Test
Condition
Run
Number
FF I t
Tempegat re
C F)
Stoichiometric
Ratio
Acid Gas
Concentration
Acid Gas
RemovaL
Efficiency
( ercent)
SO 2 MCI
SO 2
(ppmv.
b
HCL
dry at 7%
SO 2
0 ,)
u let
MCI
Combustor
DeNO /SD/FF
(19 7)
Normal
= Normal
2
3
4
7
8
15
16
283
275
27
NA
NA
NA
NA
NMC
NM
NM
NM
NM
NM
NM
NM
NM
NM
359
187
NM
NM
NM
NM
NM
MM
NM
1,079
710
1.3
1.1
1.6
NM
NM
NM
NM
NM
MM
NM
NM
NM
11.1
6.5
- -
-- --
- -
- . - -
- . -.
-- 99.0
-- 99.1
Average
277
NA
273
895
1.3
8.8
-- 99.1
Combustor =
CommerciaL
DeMO /SD/FF
(1 *88)
Residentiat/
Refuse
= Normal
2
7
11
290
288
297
31 e
50 e
2 9 e
106
99
130
764
415
758
1.1
0.9
3.8
7.0
7.0
8.0
98.9 99.1
99.1 98.3
97.1 98.9
Average
292
37 e
111
646
1.9
7.3
98.4 98.8
Combustor =
Refuse
DeNO /SD/FF
(19 8)
Commercial
= NormaL
16
21
27
290
282
285
3 1 e
6 • 3 e
32 e
161
114
120
627
265
707
0.8
2.6
8.0
8.9
3.2
4.4
99.5 98.6
97.7 98.8
93.4 99.4
Average
286
42 e
132
533
3.8
5.5
96.9 98.9
aTemperature estimated from measured value at stack and an assumed temperature drop across the fabric filter (10°F).
blotal sulfur oxides. SO 2 and SO 3 not separately reported.
cNN = not measured.
dNA = not available.
eStoichiometric ratio calculated assuming the design lime feed rate of 600 lb/hr.
-------
In Table 7-6, particulate data are presented. All outlet PM
concentrations were below 0.0043 gr/dscf at 12 percent CO 2 . The observed
removal efficiencies were at least 99.8 percent for all ten runs. During
two runs in 1987, the PM concentration was measured at the FF inlet and was
about 50 percent higher than at the SD inlet because of the lime injected in
the SD. The air-to-cloth ratio ranged from 2.1 to 2.4. There was no
apparent affect of fuel type on PM removal.
Metals emissions data are presented in Table 7-7. The metals data
presented for the May and June 1987 testing are only estimates because 100
ml of the sampling probe rinse was inadvertently discarded prior to
analysis. Because this error was not discovered until after the sample
fractions were composited, the estimated concentration represents a maximum
value. For volatile metals such as arsenic, lead, and mercury, the actual
values may be 12 to 25 percent lower than reported since only small amounts
of those metals would be expected to be present in the discarded probe rinse
fraction. Outlet metals concentrations were similar for both test
programs. Average arsenic, cadmium, chromium, lead, and nickel
concentrations were less than 6.0 ug/dscm. Removal efficiencies for these
metals averaged above 99 percent for both test programs. Mercury emissions
averaged 570 ug/dscm during the 1987 tests, but during the 1988 tests, the
average outlet concentrations were 39 and 68 ug/dscm (residential and
commercial refuse, respectively). This is similarly reflected in mercury
removal efficiency data. In 1987, zero mercury removal was indicated. In
1988, mercury removal efficiency ranged from 53.3 to 94.6 percent and
averaged 90 percent (residential refuse) and 70 percent (commercial refuse).
It has been suggested that the presence of unburned carbon in the flue gas
may enhance mercury removal. 5 High inlet CDD/CDF concentrations during the
1988 tests, but not the 1987 tests, indicating elevated carbon content on
the fly ash, appears to support this theory.
CDD/COF data are presented in Table 7-8. As described above, much
higher inlet CDD/CDF concentrations were measured in the 1988 tests (233 to
1,010 ng/dscm) than the 1987 tests (28 ng/dscm). However, outlet CDD/CDF
concentrations from the two tests were similar, ranging from 0.7 to
3.5 ng/dscm at 7 percent 02. The average outlet CDD/CDF concentrations
7-13
-------
TABLE 7-6. PARTICULATE DATA FOR COMMERCE
Combustor NormaL 9
DeNO /SD/FF DeNO off
(1987)
Average (1987)
Combustor = Residentiat/ 2
Commercial Refuse 7
DeNO /SD/FF = NormaL (1988) 11
Average
Combustor = CommerciaL 14
Refuse 21
DeNO /SD/FF = NormaL (1988) 27
Average
alemperature estimated from measured vaLue at
bNM = not measured.
CNR = not reported.
Test
Condition
Combustor
DeNO /SD/FF
Normal
= NormaL (1987)
Run
Number
2
3
4
7
8
FF Inlet
Temper 8 ture
F)
Flue Gas
Flow
(acfin)
PM
Concentration
(gr/dscf at 12% C0 2 )
Inlet Midpoint Outlet
283
90,970
NMb
2.84
0.0022
275
89,520
NM
2.43
0.0043
27
NR
89,020
94,320
NM
2.25
NM
NM
0.0022
NM
MR
97,380
1.30
NM
NM
291
88,880
NM
NM
0.0019
PM Removal
Efficiency
(percent)
FF Total
99.9 --
99.8 - -
99.9 99.8
- - 99.97
99.9
-- 99.9
- . 99.9
-- 99.98
- - 99.8
- 99.9
99.9
281 91,680 1.78 2.64
290 103,800 2.87 NM
288 91,500 1.67 NM
297 99,700 1.48 NM
292 98,300 2.01 NM
290 96,800 2.16 NM
282 91,500 0.65 NM
285 100,400 0.87 NM
286 96,200 1.23 NM
the stack and an assumed temperature drop across
0. 0027
0. 0010
0. 0016
0.0016
0.0014
0. 0004
0.0010
0.0008
0.0007
the fabric
fiLter (10 0 F).
-------
TABLE 7 -7. METALS EMISSIONS DATA FOR COMMERCE
Outlet PM
FF Inlet Concentration Inlet Concentration Outlet Concentration
Test Run Temp 8 rature (gr/dscf at ( u /dscm at 7% ____________ ( up/dscm at 7% __________ Removal Efficiency (percent )
Condition Number ( F) 12% Ca 2 ) As Cd Cr Pb Hg Hi As Cd Cr Pb Hg Hi As Cd Cr Pb Hg Ni
Combustor Normal NRb HMC NM NM NM NM NM NM NDd ND ND 5.0 200 ND - -- -- --
DeNO /S0/FF 13 HR NM 190 5.300 800 52.000 670 770 ND ND MD 3.3 940 ND 100 100 100 99.99 -40 100
No nral (1987) 14 a HR MM 250 2,100 660 47,000 240 580 MM NM NM NM NM NM -- -- -- - -
Average (1987) NH NM 220 2,700 730 50,000 450 680 ND ND MD 4.2 570 NO 100 100 100 99.99 -40 100
Combustor 3 MR NM 84 1,615 7,200 6,110 681 9,849 0.02 0.6 3.1 1.5 37 1.7 99.98 99.96 99.96 99.96 94.6 99.98
Residential! 5 HR NM 56 1,912 1,272 24,960 336 1,893 MD 0.4 1.4 0.5 41 0.6 100 99.98 99.9 100 88.0 99.97
Commercial Refuse 9 HR NM 82 1,466 1,882 22,690 341 392 0.42 4.6 2.2 3.6 41 15.6 99.5 99.7 99.9 99.98 88.0 96.0
DeNO /SD/FF Normal
(1*88)
Average 74 1,598 3,451 17,250 453 4,045 0.15 1.9 2.2 1.9 39 6.0 99.8 99.9 99.9 99.98 90.2 98.7
i - f l
Combustor 13 NH MM 2 1,284 551 6,822 278 1,998 ND 0.5 0.5 4.3 51 ND 100 99.96 99.9 99.9 81.8 100
Commercial Refuse 16 MR NM 97 17 473 22,364 159 3,589 0.16 0.3 0.3 2.7 74 0.4 99.8 98.5 99.9 99.9 53.3 99.9
DCNO /$D/FF Normal 18 HR MM 86 1,239 957 19,007 299 503 1.29° 16 8 C 3•4 5 340° 29 e 2 4 C 98 5 e 98 • 6 e 99 6 e 98.2 90 5 e
(1*88) 29 MR MM 89 1,290 243 13,917 308 1,299 ND 0.4 MD 1.6 79 0.3 100 99.97 100 99.98 74.4 90.98
Average 69 958 556 15,528 261 1,847 0.05 0.4 0.3 2.9 68 0.2 99.9 99.5 99.9 99.95 69.8 99.99
Overall Average (1988) NH O.OOl? 71 1,232 1,797 16,310 343 2,789 0.10 1.1 1.3 2.4 54 3.1 99.9 99.7 99.9 99.97 80.0 99.3
concentrations reported for Runs 11, 13, and 14 of the 1987 tests are estimated maximum values. A known amount of the probe rinse from each sample was inadvertently
discarded. The remainder of the rinse was composited with the other sample fractions and the composite was analyzed. For arsenic, lead, and mercury, (which are more
volatile end likely not In the probe rinse) the actual value may be 12 to 20 percent tower than the value reported. For the other, less volatile metals, the actual value
is probably close to the reported value.
b not reported.
CNN not measured.
dHD not detected. Considered as zero in calculating averages and removal efficiencies.
eBaghouse had disconnected bag during test. Results not representative and not included in average.
PN samples not collected simultaneously. Average result given.
-------
TABLE 7-8. CDD/CDF DATA FOR COMMERCE
•-.1
Run
Number
FF Inlet
Tem 3 erature
( F)
CDD/CDF Concentration
CDD/CDF
Removal
Efficiency
(%)
(nci/dscm
Inlet
at
7%
O .L...
Outlet
1
15
265
NMa
0.92
--
16
270
NM
1.09
--
Test ___________________
Condition
1987 TESTS
Combustor Comercial
Refuse
DeNO /SD/FF = Normal
Combustor = Residential
Refuse
DeNO /SD/FF = Norma
Average (1987 Tests)
1988 TESTS
Combustor = Residential!
Commercial Refuse
DeNO /SD/FF = Normal
Average
Combustor = Comercial
Refuse
DeNO /SD/FF = Normal
Average
Overall Average (1988 Tests)
aNM = Not measured.
bRun 1 had low load, unstable combustor
CNR = Not reported.
dlnterferences noted for analysis of these samples. Run 4 inlet considered invalid by
analytical laboratory.
270
28.1
1.83
87.7
1 b
4
8
NRC
NR
NR
233 d
NR
659
254 d
2.37
0.99
891 d
--
99.9
NR
446
9.59
94.5
15
17
22
NR
NR
NR
806
532 d
1,010
3.52
3.12
1.71
99.6
99.4
99.8
NR
783
2.78
99.6
270
545
4.73
95.9
conditions.
-------
ranged from 1.8 ng/dscm in 1987 to 1.7 ng/dscm in 1988 for residential
refuse to 2.8 ng/dscm in 1988 for commercial refuse. Removal efficiency was
87.7 percent for one run with very low inlet during the 1987 test. During
the 1988 tests, removal efficiency was 97 percent. Removal efficiency
increased with increasing inlet CDD/CDF concentration. Above inlet CDD/CDF
concentrations of 530 ng/dscm, removal efficiencies were all above 99
percent. However, outlet CDD/CDF concentrations were relatively independent
of inlet CDD/CDF concentration. The SD/FE system at Commerce can reduce
CDD/CDF levels to less than 5.0 ng/dscm, despite inlet CDD/CDF levels as
high as 1000 ng/dscm.
7.2.3 Long Beach 7
The Southeast Resource Recovery Facility in Long Beach, California
consists of three identical L. & C. Steinmuller GmbH waterwall combustors,
each with a capacity of 460 tons/day MSW. Each combustor uses Thermal DeNO
and flue gas recirculation for N0 control. Other pollutants are controlled
downstream from the boiler with a spray dryer/fabric filter system,
manufactured by Flakt-Peabody Process Systems. In the spray dryer, lime
slurry is injected through a rotary atomizer. with the rate of slurry
addition controlled by an SO 2 monitor/controller at the stack. The amount
of dilution water in the slime slurry is controlled to maintain temperature
at the outlet of the SD. Flue gas existing the SD flows through a
reverse-air FE. Design flue gas flow to each FE is 118,000 acfm at 285°F.
Each FF has 10 compartments of teflon-coated fiberglass bags and a net
air-to-cloth ratio of 1.8 acfm/ft 2 . Ducting is provided to route flue gas
from one FF to another if one unit goes down. Flue gas is exhausted through
a common stack.
In November 1988, testing was conducted on Unit 1 to demonstrate
compliance with permit conditions. The combustor and the air pollution
control equipment operated normally during testing. At the SD/FE inlet and
outlet, samples were collected and analyzed for PM and SO 2 . Particulate was
also measured at the SD outlet. At the SD/FE outlet, HC1, metals (arsenic,
cadmium, chromium, lead, mercury, and nickel), CDD/CDF, NON , and other
7-17
-------
organics were sampled. The NO results are reported in Municipal Waste
Combustors - Background Information for Proposed Standards: Control of NO
6 X
Emissions.
Acid gas data are presented in Table 7-9. Of the three test runs
performed, outlet SO 2 concentrations were 5.61 and 7.95 ppm at 7 percent 02
for Runs 2 and 3. During Run 1, the lime slurry feed stopped, resulting in
an unrealistically high outlet SO 2 concentration. The SO 2 removal
efficiencies for test Runs 2 and 3 were 92.2 and 96.5 percent, respectively.
Outlet HC1 concentrations ranged from 13.4 to 38.2 ppm at 7 percent 02 and
averaged 24.2 ppm. Outlet HC1 sampling was not conducted simultaneously
with SO 2 sampling. Inlet MCi levels were not measured. The fabric filter
inlet temperatures ranged from 303 to 311°F and averaged 307°F during the
SO 2 tests and ranged from 290 to 300°F and averaged 295°F during the MCi
tests. Because of the limited range in fabric filter inlet temperature, the
effect of this key process variable on acid gas control performance of the
SD/FF can not be evaluated. The lime feed rate was not measured during
testing.
Particulate data for Long Beach are presented in Table 7-10. Outlet PM
concentrations ranged from 0.0047 to 0.0076 gr/dscf at 12 percent CO 2 over
three runs and averaged 0.0060 gr/dscf. The corresponding PM removal
efficiency across the SD/FF was between 99.4 and 99.7 percent and averaged
99.6 percent. Concentrations at the FF inlet were very similar for the
three runs, averaging 2.75 gr/dscf at 12 percent C0 2 , representing an
increase in particulate concentration of 53 to 102 percent across the SD.
The flue gas flow rate and resulting air-to-cloth ratio were essentially
constant for each run at approximately 1.7 acfm/ft 2 .
Metals emissions data for Long Beach are presented in Table 7-11.
Arsenic, chromium, lead and nickel concentrations at the FF outlet all
averaged less than 5 ug/dscm. The average cadmium concentration was 18
ug/dscm. Mercury at the FF outlet averaged 180 ug/dscm. Compared to
typical uncontrolled metals concentrations (see Section 1.2), removal
efficiencies for arsenic, cadmium, chromium, lead, and nickel are estimated
to be at least 99 percent. The measured outlet mercury emissions suggest a
removal efficiency of 65 percent based on typical uncontrolled levels.
However, because uncontrolled mercury levels vary widely, it is possible
that the estimated mercury removal may be significantly less.
7-18
-------
TABLE 7-9. ACID GAS DATA FOR LONG BEACH
Acid
Gas
concentrationsbs(2
Test
Condition
Run
Number
FF Inlet a
Tempegature
( F)
Stoichiometric
Ratio
(ppmv,
dry at 7% O ,
j__
SO Removal
Ef iciency
(percent)
InLet
SO 2
Outte
SO 2
HCL
Combustor =
DeNO /SD/FF
X
Normal
- Normal
1
2
3
303 29O
311 295 a
307 300 a
NAd
NA
NA
154
102
158
722 e
7.6
5.6
13.4
21.0
38.2
53.1
92.2
96.5
Average
307 295 a
NA
138
6.8
24.2
944
and FICI measurements were made during separate runs. First temperature is for SO runs.
.1 Se orid temperature is or HCI runs. Temperature estimated from measured vaLue at sta k and assumed
temperature drop of 10 F across the fabric filter.
b
Total sulfur oxides. SO 2 and SO 3 not separately reported.
C 0 and Nd not measured simultaneousLy.
dNA = not available. Lime slurry feed rate not measured.
eoutlet SO 2 concentration from Run 1 is considered anomalous because the lime slurry feed stopped during
the test. Sampling was stopped during this time interval. Results are not included in average.
-------
TABLE 7-10. PARTICULATE DATA FOR LONG BEACH
cD
Test
Condition
Run
Number
FF Inlet
Temp 8 rature
C F)
Flue Gas
Flow
(acfm)
PM Concentration
(gr/dscf at 12% CO
Inlet Midpoint Obttet
PM Removal
Efficiency
(percent)
SD
FF
Overall
Combustor
DeNO /SD/FF
X
Normal
Normal
1
2
3
291
300
302
112,690
113,200
113,820
1.63
1.82
1.29
2.85
2.78
2.61
0.0047
0.0057
0.0076
-74.9
-52.8
-102
99.8
99.8
99.7
99.7
99.7
99.4
Average
298
113,240
1.58
2.75
0.0060
-76.6
99.8
99.6
-------
TABLE 7-11. METALS DATA FOR LONG BEACH
Test
Condition
Run
Number
FF InLet
Temperature
(°F)
OutLet PM
Concentration
(gr/dscf at 12%
C0 2 )
OutLet Concentration
(ug/dscm at 7% 02)
As
Cd
Cr
Pb
Hg
Ni
Combustor = NormaL
1
291
0.0047
ND 8
10.6
4.75
13.9
146
2.92
DeNO /SD/FF = NormaL
X
2
300
0.0057
ND
9.5
1.06
ND
176
1.76
3
302
0.0076
ND
33.8
2.09
ND
219
3.75
Average
298
0.0060
NDb
180
2.63
4.6
180
2.81
8 ND = not detected. Considered as zero for evaLuating averages.
bArsenic was not detected in any of the three runs. The maximum detection Limit was 33 ug/dscm at
7 percent 02.
-------
Table 7-12 presents the CDD/CDF data from Long Beach. Outlet CDO/CDF
concentrations ranged from 1.2 to 9.8 ng/dscm at 7 percent 02 over three
runs and averaged 4.14 ng/dscm. The highest CDD/CDF concentration was
measured at the highest FF inlet temperature, 313°F. However, the range in
temperature, 298 to 313°F, was too limited to evaluate the effect of this
parameter. Based on the above, the SD/FF at Long Beach, operating at less
than 310°F is capable of achieving outlet CDD/CDF concentrations of less
than 10 ng/dscm.
7.2.4 Mid-Connecticut 8
The Mid-Connecticut Resource Recovery Facility in Hartford,
Connecticut consists of three Combustion Engineering spreader stoker-fired
boilers each designed to combust a maximum of 675 tons/day RDF or
236 tons/day coal. Each ROF combustor is designed to produce 231,000 lb/hr
of steam.
The air pollution control system for each combustor consists of a spray
dryer followed by a fabric filter. Slaked pebble lime slurry is introduced
to the spray dryer through a single rotary atomizer. The slurry feed rate
is controlled to achieve the desired SO 2 removal and the dilution water flow
rate is controlled to achieve the desired temperature at the SD outlet.
Upon exiting the SD, the typical flue gas flow is about 190,000 acfm at
280°F. The reverse-air cleaned fabric filter has 12 compartments of
168 teflon-coated glass-fiber bags each and a design gross air-to-cloth
ratio of 1.45 acfm/ft 2 . The net air-to-cloth ratio is 1.74 acfm/ft 2 with
two compartments off-line.
In July 1988, compliance testing was conducted at the facility. The
EPA funded testing at the SD/FF inlet of one of the units during compliance
testing to determine the level of uncontrolled MWC emissions and assess the
performance of the SD/FF system. The combustor and SD/FF were at normal
operating conditions during testing. At the SD/FF inlet and outlet, flue
gas was sampled for PM, metals (including arsenic, cadmium, chromium, lead,
mercury, and nickel), and CDD/CDF. No acid gas measurements were taken.
Particulate data are presented in Table 7-13. Outlet PM concentrations
ranged from 0.0021 to 0.0059 gr/dscf at 12 percent CO 2 over three runs and
averaged 0.0040 gr/dscf. The corresponding PM removal efficiencies averaged
7 -22
-------
TABLE 7-12. CDD/CDF DATA FOR LONG BEACH
Test
Condition
Run
Number
FF
Temp
(
Inlet a
8 rature
F)
Outlet
Conce
(ng 1 dscm
CDD/CDF
ntration
at 7% 02)
Combustor =
1
313
9.75
Normal
2
298
1.47
DeNO /SD/FF
No mal
=
3
305
1.20
Average
305
4.14
aTemperature estimated from measures value at stack and an assumed
temperature drop across the FE (10 F).
7-23
-------
r
TABLE 7-13. PARTICULATE DATA FOR MID-CONNECTICUT
dincludes sootbiowing cycle.
Value not included in average.
Test
Condition
Run
Number
FF Inlet
Temp 8 rature
( F)
Flue Gas
Flow
(acfm)
Inlet PM
Concentration
(gr/dscf at
12% C0 2 )
Outlet PM
Concentration
(gr/dscf at
12% C0 2 )
PM Removal
Efficiency
(percent)
Combustor =
Normal
SD/FF = Normal
1
2
3
274
276
278
159,000
156,000
159,000
2.57
2.25
4 • 78 a
0.0021
0.0041
0.0059
99.9
99.8
99.9
Average
276
158,000
2.41
0.0040
99.9
-------
99.9 percent. Air-to-cloth ratio was relatively constant among the tests at
1.5 acfnh/ft 2 .
Metals data for Mid-Connecticut are presented in Table 7-14. Three
runs were conducted with simultaneous sampling at the inlet and outlet for
arsenic, chromium, lead, and nickel. Cadmium and mercury were additionally
measured at the inlet only. Inlet sampling was performed using the draft
EMSL method and outlet sampling was conducted using EPA Methods 12 and 108.
Three additional runs were conducted with mercury, measured simultaneously
at the inlet and outlet using EPA Method lOlA. Both arsenic and lead were
not detected at the outlet during any runs. Chromium was detected in Run 1
at 115 ug/dscm, but not detected in the other two runs. Mercury
concentrations ranged from 3.7 to 130 ug/dscm and averaged 50 ug/dscm.
Nickel was detected in two runs at 460 and 470 ug/dscm, but not detected in
Run 3. Because arsenic and lead were not detected in any of the three
outlet samples, removal efficiencies of nearly 100 percent were estimated.
Chromium removal efficiency was 88 percent for Run 1 and 100 percent for Run
2. Run 3 had nondetectable chromium levels at the SD/FF outlet, but inlet
chromium samples were not collected during this run. Nickel removal by the
SD/FE was low. Removal efficiencies were 22 and 5.6 percent. However, Run
3 had nondetectable nickel levels at the outlet.
Mercury was removed by the SD/FE at efficiencies of 60.7 to 99.7 percent,
and an average efficiency of 86.3 percent. The lowest mercury removal
efficiency and highest outlet concentration corresponded to the lowest inlet
mercury concentration.
CDD/CDF for Mid-Connecticut data are presented in Table 7-15. Outlet
CDD/CDE concentrations ranged from not detected to 1.39 ng/dscm at 7 percent
02 over three runs and averaged 0.660 ng/dscm. The inlet CDD,”CDF
concentration ranged from 819 to 1,275 ng/dscm and averaged 1,019 ng/dscm.
The corresponding removal efficiencies were at least 99.9 percent. The data
show that the SD/FF at Mid-Connecticut, operating at 270°F with inlet
CDD/CDF concentrations as high as 1275 ny/dscm can achieve an outlet CDD/CDF
concentration of less than 2 ng/dscm.
7-25
-------
TASLE 7-14. METALS EMISSIONS DATA FOR MID-CONNECTICUT
Test
Condition
Run
Number
FF Inlet
Tempzr.ture
C F)
Outlet PM
Concentration
(gr/dscf at
12% C0 2 ) As
Inlet
Concentr.tion
b
Outlet Concentration
Cue/dec.
Cd Cr
• TZ
Pb
0 )
Hg
Ni
A.
ue/dscs
Cd Cr
at 7% 022_.___
Pb Hg Ni
Removal
As Cd
Efficiency
Cr Pb
(Dercent)
Hg Ni
Cosubustor Normal
501FF r Normal
1
2
3
274
276
278
0.0021
0.0041
0.0059
1,284
838
MM
894
1,241
NM
957
895
NM
43,384
31,387
NM
835
1,181
NM
583
499
NM
MDC
ND
ND
MIld 115
NM MD
NM ND
ND MM 457
ND NM 471
ND NM ND
100
100
-•
88.0
100
100
100
.
--
21.6
5.6
Average
216
0.0040
1,061
1,068
921
37,386
1.008
541
MD
NM 38
MD -- 309
100
--
94.0
100
-
13.6
1 (Hg)
2 (Mg)e
3 (Hg)
289
278
285
NM
MM
MM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
331
1,215
1,107
NM
NM
NM
NM
NM
NM
NM NM
NM NM
NM NM
NM 130 NM
NM 4 NM
NM 17 NM
--
-•
.-
60.7
99.7
98.4
-
--
•
Average (Hg tests)
284
NM
NM
NM
NM
NM
884
NM
NM
NM NM
NM 50 NM
86.3
5 lnlet metal. samples for Runs 1, 2, 3 were collected using the draft EMSL metaLs method.
bOutlet metals samples for Runs 1, 2, 3 were collected using a combination Method 12/108.
C 1 , 0 not detected. Considered as zero in evaluating averages.
d ,M = not measured.
eAdditlonat inlet and outlet mercury samples were collected by Method lOlA. Not measured simultaneously with other metals.
-------
Test
Condition
Run
Number
FF Inlet
Temp 8 rature
( F)
Inlet CDD/CDF
Concentration
(ng/dscm at
7% 02)
Outlet CDD/CDF
Concentration
(ng/dscm at
7% 02)
CDD/CDF
Removal
Efficiency
(percent)
Combustor = Normal
SD/FF = Normal
1
2
3
270
266
277
1,275
819
963
1.39
O.5
ND
99.9
99.9
100
Average
271
1,019
0.66
99.9
TABLE 7-15. CDD/CDF DATA FOR MID-CONNECTICUT
N)
aND = not detected. Considered as zero in evaluating averages.
-------
7.2.5 Marion Countv. 9 2
The Marion County Solid Waste-to-Energy Facility in Brooks, Oregon,
consists of two 275-ton/day, mass burn, waterwall Martin GmbH combustors.
Each combustor is equipped with identical, Teller SD/FF systems. The flue
gases leave the boiler economizer and enter a cyclonic separator that
removes large particles before entering the bottom of the SD. Slaked pebble
lime is mixed with water and injected into the SD through an array of five
two-fluid nozzles. The lime slurry feed rate is varied to maintain the SD
outlet temperature between 260 and 300°F. The dry lime feed rate is
approximately 425 lb/hr per unit. The lime concentration in the slurry is
maintained at a rate sufficient to yield a stoichiometric ratio of lime to
HC1 of about 2 to 2.5. Tesisorb is injected into the flue gas through a
venturi located imediately before the FE inlet gas plenum to enhance
collection performance and reduce pressure drop across the FF. The
Amertherm reverse-air FF consists of six compartments of 120 fiberglass bags
each. The net air-to-cloth ratio is 2.3 acfm/ft 2 at a flue gas flow of
about 60,000 acfm. After exiting the FF, the combustion gases are
discharged through a 258-foot high stack.
Three test programs have been conducted at Marion County. The first
test program, conducted in September and October 1986, included compliance
testing at the Unit 1 SD/FF outlet funded by the facility, and simultaneous
testing at the SD/FF inlet funded by EPA. At the inlet and outlet of the
SD/FF, flue gas was analyzed for CDD/CDF, PM, HC1, SO 2 , and metals (lead,
cadmium, chromium, and nickel) for three runs. A second program was
conducted in February 1987. During these tests, SO 2 and CDD/CDF samples
were collected at the SD/FF inlet.
The third test program was conducted in June 1987 on Unit 1 with
funding from EPA. These tests focused on parametric testing of the
combustor and SD/FF. The purposes of the test relative to the SD/FF were
(1) to determine the removal of CDD/CDF by a SD/FE at combustor startup and
shutdown conditions, (2) to evaluate acid gas removal efficiency as a
function of temperature and stoichiometric ratio, and (3) to evaluate the
performance of the SD/FF over the normal operating range of the combustor.
7-28
-------
The combustor parameters which were varied during testing included steam
load, excess air, and combustion air distribution. During the 11 runs with
variations in combustor operation, the SD/FF was maintained at normal
conditions. For the three SD/FF parametric tests, the SD outlet temperature
was varied while the combustor was maintained at normal conditions. Flue
gas was sampled at the SD/FF inlet and outlet using continuous emission
monitors for HC1, SO 2 , NON, THC, and CO. Continuous measurements of HCl and
SO 2 were also taken at the SD outlet. For the shutdown and startup tests,
flue gas was analyzed for CDD/CDF at the SD/FF inlet and outlet.
Acid gas data collected during the compliance and parametric test
programs are presented in Table 7-16. During normal combustor and SD/FF
operation (1986), outlet SO 2 concentrations ranged from 7.8 to 65.5 ppm at 7
percent 02 and averaged 31.3 ppm. The SO 2 removal efficiency during these
runs ranged from 73.8 to 93.3 percent and averaged 85.2 percent. During
normal combustor and SD/FF operation, outlet HC1 concentrations ranged from
4.09 to 42.4 ppm at 7 percent 02 and averaged 17.7 ppm. Removal
efficiencies for HC1 ranged from 93.0 to 99.3 ppm during these runs and
averaged 97.0 ppm.
During parametric testing of the combustor with normal SD/FF operation
(1987), outlet SO 2 concentrations ranged from 9.9 to 386 ppm at 7 percent
02. The corresponding SO 2 removal efficiencies were between 32 and
92 percent. Outlet HC1 concentrations during parametric testing ranged from
11.5 to 84 ppm at 7 percent 02. The corresponding HC1 removal efficiencies
were between 87 and 98 percent. These results show some differences from
the 1986 compliance tests. The average outlet SO 2 concentration from the
1987 tests are approximately five times the value for the 1986 tests, while
the average outlet HC1 concentration is approximately three times as high.
The removal efficiencies for both SO 2 and NC] are higher during the 1986
tests. These differences are probably a result of two times higher inlet
SO 2 concentrations during the 1987 tests and the inability of the spray
dryer to control these high inlet levels. A waste screening program is now
used to remove materials causing high inlet SO 2 levels prior to being
7-29
-------
TABLE 7-16. ACED GAS DATA FOR MARION COUNTY
Test
Conditions
Run
Number
FF Inlet
Te.p ratur.
C F)
Stoichiometric
Ratio
Acid Gas Concentrations
Acid Gas
Removal
Efficiency
(percent)
(Domy. dry •t 1
Inlet Mldoolnt
SO 2 HCL SO 2 Nd
Outlet
SO 2 NCI
SD
SO 2 HCI
SO 2
FF
HCI
Overall
SO 2 HCI
Combustor • NormaL 1 270 2.5 117 555 NM NM 8 4.1 -. -- -- -- 93.3 99.3
SD/FF Normal (1986) 2 272 2.1 181 555 NM NM 21 6.7 - - - - - - 88.5 98.9
3 272 1.9 250 603 NM NM 66 42.4 -- -. 73.8 93.0
Average 271 2.2 183 571 NM NM 31 17.7 - . - - -- 85.2 97.0
Combustor Parametric 1 300 1.0 566 654 448 224 386 84.3 20.9 65.8 13.8 62.3 31.8 87.1
SD/FF Normal (1987) 2 300 1.3 315 674 131 187 104 36.5 58.3 72.2 20.9 80.5 67.0 94.6
3A 300 1.2 428 495 305 183 203 54.6 28.6 63.0 33.5 70.2 52.5 89.0
39 299 1.1 523 704 384 175 264 53.1 26.6 75.1 31.7 69.8 49.6 92.5
4 301 1.9 119 639 30 150 10 11.5 75.1 76.6 66.5 92.3 91.6 98.2
5 299 1.1 424 727 238 106 168 48.6 43.9 85.4 29.1 54.3 60.3 93.3
6A 302 1.4 335 683 230 225 197 74.4 28.2 67.1 17.9 66.9 41.1 89.1
68 300 2.1 277 629 115 90 58 30.2 58.4 85.7 49.9 66.4 79.2 95.2
7 288 1.6 288 670 205 222 148 71.4 29.0 66.8 27.8 67.9 48.7 89.3
8 298 2.4 215 581 158 186 95 42.9 26.5 68.0 39.8 76.9 55.8 92.6
9 299 2.2 171 654 39 196 23 21.0 77.3 70.0 41.2 89.3 86.6 96.8
0
Average 299 1.6 333 646 207 177 151 48.0 43.0 72.3 33.8 72.4 60.4 92.5
Combustor Normal 10 262 1.1 397 845 322 178 110 21.9 18.8 78.9 65.8 88.3 72.3 97.5
SDIFF Low SD outset
temperature (250 F)
Combustor • Normal hA 330 1.1 481 735 533 301 493 160 -10.7 59.0 7.5 46.8 2 • 4 b 78.2
SD/FF High SD ou Let
temperature (330 F)
Combustor Normal 11B 360 1.6 129 817 179 315 167 217 -38.9 61.4 6.5 31.1 29 • 9 b
SD/FF Very high S
outlet temp. (360 F)
8 NN not measured.
bNegetive removal efficiencies indicative of little or no SO 2 removal.
-------
combusted. Another factor may have been that the 1986 tests were conducted
at slightly lower FF inlet temperatures (270 versus 300°F), as discussed
later.
The spray dryer parametric tests were performed to evaluate the effect
of FF inlet temperature on performance. Run 10 was conducted at a very low
FE inlet temperature (250°F). Outlet SO 2 concentration was 110 ppm for a
removal efficiency of 72 percent. Outlet HC1 concentration was 22 ppm for a
removal efficiency of 97.5 percent. Runs hA and 11B were conducted at
elevated FF inlet temperatures (330 and 360°F, respectively). Outlet SO 2
concentrations were 490 and 170 ppm, but both tests indicated approximately
zero removal efficiencies. Outlet HC1 concentrations were 160 and 220 ppm,
for removal efficiencies of 78 and 73 percent, respectively, at 330 and
360°F. Removal efficiencies for both SO 2 and HC1 were lower during the
elevated FF inlet temperature tests as compared to normal SD operations.
Increasing stoichiometric ratio increased removal efficiency for both
SO 2 and HC1, as shown in Figure 7-3 and 7-4, respectively. However, because
the lime feed rate was measured as the total feed to both units and because
the two units were not always operating identically, different lime feed
rates may have been supplied to the units. Thus, the calculated
stoichiometric ratios (based on the assumption of equal lime flow to each
unit) are only estimates. Calculated stoichiometric ratios varied from 1.9
to 2.5 during the 1986 tests, 1.0 to 2.4 during the combustor parametric
tests, and 1.1 to 1.6 during the SD parametric tests. The scatter in the
data are probably due to effects of temperature and inaccuracy in
calculating the stoichiometric ratio.
The effect of FE inlet temperature on SD/FF performance is apparent at
Marion County. At FF inlet temperatures of 330 and 360°F, during runs hA
and 11B, the lowest SO 2 and HC1 removal efficiencies were observed. At 260
to 270°F, during run 10 of the parametric testing and all three runs of
compliance testing, the highest SO 2 and HC1 removal efficiencies were
observed. At approximately 300°F during the 11 parametric combustor
conditions, acid gas removal efficiencies varied between 32 and 92 percent
7-31
-------
100 -
0
0
80- 0
0 0
70 -
0
C
60- 0
0
50- 0 0
40- 0
0
30-
o 20-
E
C ...)
10-
In
0-
-10 -
-20 -
-30-
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
Stoichiometric Ratio
FIgure 7-3. SO 2 removal efficiency as a function of
stoichiometric ratio at Marion County.
-------
100 -
0 0
98— 0
0
0
96 -
0
0
94 -
0
0 V 0
C 92-
0
C)
o 90-
a
0 0
>‘
o 88—
0
72 0 1 2 1 4 1 6 1 8 ‘ 2 2 2 4 ‘ 2.6
Stoichiometric Ratio
Figure 7-4. HCI removal efficiency as a function of
stoichiometric ratio at Marion County.
-------
for SO 2 and between 87 and 98 percent for HC1. The average removal
efficiencies for this temperature were 60 percent for SO 2 and 93 percent for
HC1.
The effect of inlet SO 2 concentration on SO 2 and HC1 removal
efficiencies is shown in Figures 7-5 and 7-6, respectively, for all the
tests. For both SO 2 and HC1, removal efficiency increased with decreasing
inlet SO 2 concentration. This trend is less apparent for HC1 removal
efficiency because HC1 will react with the available lime generally before
SO 2 does, and thus is less prone to be affected by increases in inlet SO 2
concentration. Below inlet SO 2 concentrations of 200 ppm, SO 2 and HC1
removal efficiencies were at least 87 and 97 percent, respectively.
In Figure 7-7, the relationship between SO 2 and HC1 removal efficiency
is shown for all runs. There is an approximately linear relationship
between the two values. At SO 2 removal efficiencies of 90 percent or
higher, HC1 removal efficiency is at least 98 percent.
During the combustor parametric tests with normal SD operation, 20 to
80 percent SO 2 removal occurred across the SD with an additional 15 to 60
percent removal across the FF. The SO 2 removal was similar at the low FF
inlet temperature run, but at the high temperature runs, no removal occurred
across the SD or FF. For HC1, about 65 to 90 percent removal occurred
across both the SD and FF during normal and low temperature operation.
However, at the high FF inlet temperature tests, approximately 60 percent
HC1 removal occurred across the SD and an additional 30 to 45 precent
removal occurred across the FF.
PM data are presented for the 1986 compliance testing in Table 7-17.
Outlet PM concentrations ranged from 0.0013 to 0.0037 gr/dscf at 12 percent
CO 2 over six runs and averaged 0.0023 gr/dscf. The corresponding PM removal
efficiencies for the three runs with simultaneous inlet and outlet data were
between 99.6 and 99.8 percent and averaged 99.7 percent. The net
air-to-cloth ratio was 2.4 acfm/ft 2 , or less during the tests conducted.
Metals data are presented for the 1986 testing in Table 7-18. Removal
efficiencies measured over the course of three test runs for cadmium,
chromium, and lead were all greater than 99.7 percent. Nickel removal
efficiency was measured at 50 percent. Mercury emissions at the FF outlet
are lower than typical uncontrolled values, suggesting a removal efficiency
7-34
-------
100
U -
90 FF Inlet Temperatures
U 300
+ 270
80 + U - 25 O
.
100 200 300 400 500 600
Inlet SO 2 Concentration (ppmv, dry at 7% Q )
Figure 7-5. SO 2 removal efficiency as a function of
inlet SQ 2 concentration at Marion County.
-------
100
+
+ FF Inlet Temperatures
98 U 30O
U 270’
87 0
100 200 300 400 500 600
Inlet SO 2 Concentration (ppmv, dry at 7% Q 2 )
Figure 7-6. HCI removal efficiency as a function of
inlet SO 2 concentration at Marion County.
-------
100
98 -
96
I
.
94
U
. I
92-
C)
0
I-
90
U •.
88-
C
C)
FF Inlet Temperature
- E 82—
C) U 300F
80- + 270°F
I 0 250°F
78- 330°F
76— X 360°
74 -
72 - —- — - — -- ———j ---—-- I I I I I I I -j
-30 -10 10 30 50 70 90
SO 2 Removal EffIciency (percent)
Figure 7-7. HCI removal efficiency as a function of
SO 2 removal efficiency at Marion County.
-------
TABLE 7-17. PARTICULATE DATA FOR MARION COUNTY
Test
Condition
Run
Number
FF InLet
Tempe 5 ature
C F)
FLUe Gas
FLow
(ecfm)
InLet PM
Concentration
(gr/dscf at
(12% C0 2 )
OutLet PM
Concentration
(gr/dscf at
(12% C0 2 )
PM
RemovaL
Efficiency
(percent)
Combustor = NormaL
SD/FF = NormaL
(1986)
1
2
270
272
63,098
58.123
1.05
1.12
0.0037
0.0018
99.6
99.8
3
272
57,029
0.740
0.0013
99.8
4
271
59,921
0.848
WNa
5
272
61,970
0.791
NM
-
6
272
57,476
0.739
NM
--
Average
272
59,603
0.881
0.0023
99.7
8 NM = not measured.
-------
TABLE 7-iL METALS EMISSIONS DATA FOR MARION COUNTY
NM not measured.
bFabrlc filter was by-passed during the test run for five minutes. Emissions measured are higher than normally expected.
= not detected. Considered as zero in evaluating average.
dsercury measured at different times than other metals. Results not affected by the by-pass of the fabric fitter.
eparticutate resuLts not collected simultaneously. Average result given from same test prr’gram.
Average outlet concentrations and removal efficiencies include results from Runs 2 and 3 only.
‘ ,O
Test
Condition
Run
Number
EF Inlet
TempS ature
( F)
Outlet PM
Concentration
(gr/dscf at
12% C0 2 )
Inlet Concentratlona
Outlet
Concentration
‘%
b
Removal Efficiency
CX)
Cd
(ug/dscm at 7%
Cr Pb
0 )
14
Ni
Cd
(ug/dscm at
Cr Pb
Ng
Ni
Cd
Cr Pb Kg
Ni
Combustor = Normal
4
271
--
1 18O
464 16,600
NNa
16.9
18 b
NDb.c
0.13
190 b
15
228 d
296
16 b
3.1
98.1
99.7
100 98.6
99.9 99.9
--
89.7
23.0
SD/FF = Normal
5
272
--
1,210
269 19,900
NM
0.20
22
192
3.1
99.7
99.9 99.9
--
77.4
(1986)
6
272
--
973
533 24,600
MM
15.7
Average
272
00023 e
1.121
422 20.500
NM
12.4
2 . 6 o.ir
i 9
239
3 .i
99.9
--
-------
of approximately 50 percent. However, because inlet mercury concentrations
are highly variable, the actual mercury removal efficiency may differ.
In Table 7-19, CDD/CDF data are presented for the 1986 runs and
transient condition tests. During normal operation of the combustor and
SD/FF, outlet CDD/CDF concentrations for Runs 2 and 3 were 1.86 and 0.665
ng/dscm at 7 percent 02 and averaged 1.26 ng/dscm. Sample recovery problems
prevented analysis of the Run 1 sample. The inlet CDD/CDF concentration was
43 ng/dscm for one run, but sample recovery difficulties prevented analysis
of Runs 1 and 3. The resulting removal efficiency was 96 percent.
During shutdown, inlet CDD/CDF concentrations of 85 and 38 ng/dscm were
measured for two samples collected simultaneously except the second train
did not include a sootblowing. The outlet CDD/CDF concentrations were 3.1
and 0.81 ng/dscm for two samples collected over the entire test, including
the sootbiowing. The removal efficiency was 97 percent based on the
averages of the inlet and outlet concentrations.
During startup, two sample trains were run at the inlet consecutively
over the first and second halves of the 4-hour test. Two concurrent samples
were collected over the full test at the outlet. At the inlet, the CDD/CDF
concentration was 710 ng/dscm during the first half of the test and 160 ng/dscm
for the second half of the test. The only outlet sample which could be
analyzed had a CDD/CDF concentration of 3.4 ng/dscm. Based on the average
of the inlet samples and the single outlet sample, the CDD/CDF removal
efficiency was 99 percent.
CDD/CDF emissions were consistently low despite wide variations in
inlet CDD/CDF concentrations (38 to 712 ng/dscm). Thus, the SD/FF at Marion
County, with inlet CDD/CDF concentrations of 38 to 712 ng/dscm, can achieve
outlet CDD/CDF levels of less than 3.4 ng/dscm.
7.2.6 Penobscot 13 ’ 14
The Penobscot Energy Recovery Facility in Orrington, Maine consists of
two 360-ton/day RDF-fired boilers manufactured by Riley Stoker. Each boiler
produces 66,700 lb/hr of steam. Flue gas emissions from each unit are
controlled by a spray dryer/fabric filter system supplied by General
Electric Environmental Services. Lime slurry is injected into the spray
7-40
-------
TABLE 7-19. CDD/CDF DATA FOR MARION COUNTY
Inlet CDD/CDF Outlet CDD/CDF CDD/CDF
Concentration Concentration Removal
Test Run FF Inlet (ng/dscm at (ng/dscm at Efficiency
Condition Number Temperature 7% 02) 7% 02) (%)
1986 TESTS
Combustor = Normal 2 272 43. 1.86 95.7
SD/IF = Normal 3 272 NR 0.665 --
Average 272 43.0 1.26 95.7
1987 TESTS b
Combustor = Shutdownb lAb 293 84.9 3.01
SD/FF = Normal lB 293 38.3 0.806
Combustor = Startup 1A 301 712 3.36
SD/FF = Normal lB 301 158 NR
aNR = not reported. Analytical difficulties encountered with the sample.
b5hutdowfl test had two trains at inlet and two trains at outlet. Inlet Train A included a
sootbiowing and inlet Train B did not (1/2 hour difference in run length). The two outlet train
samples were collected over the full test.
cRemoval efficiency calculated based on average inlet concentration for two trains and average
outlet concentration from two trains.
dstart_up test had two trains at inlet and two trains at outlet. Inlet Train A was for the first
two hours of the 4-hour test and inlet Train B was run for the last two hours. The two outlet
trains were used over the full test.
eRemoval efficiency was calculated based on average of two inlets and one outlet concentrations.
-------
dryer through a rotary atomizer. Flue gases exiting the spray dryer enter a
pulse-jet fabric filter for PM control. The design flue gas flow at the FE
inlet is 170,000 acfm at 300°F. The FF has six modules of 126 bags each.
The design PM removal efficiency is 99.8 percent to achieve outlet PM levels
of 0.01 gr/dscf at 12 percent CO 2 .
In August 1988, a compliance test was conducted at the facility. The
combustor and SD/FF were operated normally during testing. At the SD/FF
outlet, flue gas was sampled and analyzed for SO 2 , HC1, PM, metals (cadmium,
chromium, and lead), and CDD/CDF. No measurements were made at the SD
inlet.
Acid gas data are presented in Table 7-20. Outlet SO 2 concentrations
ranged from 7.65 to 13.0 ppm at 7 percent 02 over three runs and averaged
11.1 ppm. Outlet HC1 concentrations ranged from 1.06 to 1.42 ppm at 7
percent 02 and averaged 1.18 ppm. The FF inlet temperatures was a
consistent 296°F.
PM data are presented in Table 7-21. Outlet PM concentrations ranged
from 0.00058 to 0.0015 gr/dscf at 12 percent CO 2 over three runs and
averaged 0.0011 gr/dscf. The flue gas flow rate, averaging 172,600 acfm for
the three test runs, was very near the design rate of 170,000 acfm. At this
flue gas flow rate, the FF exceeded its design performance level of 0.01
gr/d s Cf.
Metals emissions data are presented in Table 7-22. Sampling for
cadmium, chromium, and lead was conducted simultaneously with PM sampling.
Based on typical uncontrolled metals concentrations, removal efficiencies
for cadmium, chromium, and lead were all greater than 99.8 percent.
In Table 7-23, CDD/CDF data are presented. Outlet CDD/CDF
concentrations ranged from not detected to 3.87 ng/dscm over three runs and
averaged 2.39 ng/dscm. Inlet CDD/CDF concentrations were not measured.
Nevertheless, at a consistent FF inlet temperature of 295°F, the SD/FF
system at Pensobscot demonstrated outlet COD/COF levels of less than
4 ng/dscm.
7.2.7 Quebec City 15
The Quebec City, Canada, municipal waste combustion facility consists
of four separate mass burn, waterwall combustors. The combustors were
7-42
-------
TABLE 7-20. ACID GAS DATA FOR PENOBSCOT
Outlet Acid Gas
FF Inlet Concentration
Test Run Temperaturea ( pmv, dry at 7% 02)
Conditions Number (°F) SO 2 HC1
Combustor = Normal 1 296 13.0 1.1
SD/FF = Normal 2 297 12.7 1.4
3 296 7.7 1.1
Average 296 11.1 1.2
alemperature estimated from measured value at he stack and an assumed
temperature drop across the fabric filter (10 F).
TABLE 7-21. PARTICULATE DATA FOR PENOBSCOT
FF Inlet Outlet PM
Test Run Tempe aturea Flue Gas Concentration
Condition Number ( F) Flow (acfm) (gr/dscf at 12% C0 2 )
Combustor = Normal 1 296 175,700 0.0011
SD/FF = Normal 2 297 172,200 0.00058
3 296 169,800 0.0015
Average 296 172,600 0.0011
ajemperature estimated from measured value at he stack and an assumed
temperature drop across the fabric filter (10 F).
7-43
-------
TABLE 7-22. METALS DATA FOR PENOBSCOT
Test
Condition
Run
Number
FF Inlet
Temperaturea
(°F)
Outlet PM
Concentration
(gr/dscf at 12%
C0 2 )
Outlet Concen
(u /dscm at
tra
7%
tion
021
Cd
Cr
Pb
Combustor = Normal
SD/FF = Normal
1
2
3
296
294
295
0.0011
0.00058
0.0015
0.57
1.33
0.85
2.38
2.15
1.98
6.9
19.1
6.8
Average
295
0.0011
0.92
2.17
11.0
aTemperature estimated 0 from measured value at the stack and an assumed temperature drop across
the fabric filter (10 F).
-------
TABLE 7-23. CDD/CDF DATA FOR PENOBSCOT
Test
Condition
Run
Number
FF
Temp
Inlet a
e ature
( F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Combustor = Normal
1
295
3.29
SD/FF = Normal
2
3
296
295
3. 7
ND
Average
295
2.39
aTemperature estimated from measured value at he stack and an assumed
temperature drop across the fabric filter (10 1).
bND = not detected.
7-45
-------
originally built in 1975 with Von Roll reciprocating grates. Waterwall
arches were added to each combustion chamber in 1979. Each unit is designed
to combust 250 tons/day of MSW. Emissions were originally controlled by
2-field ESP’s.
Environment Canada, in cooperation with Flakt Canada, Ltd., established
an extensive test program to evaluate the capability of a pilot-scale SD/FE
control system to remove PM, SO 2 , HC1, heavy metals, CDD/CDF, and other
organic compounds. Flakt constructed a pilot-scale SD/FE facility at the
Quebec City plant equipped with:
(1) a flue gas slipstream from the ESP inlet of Unit 3 to deliver
2,000 ft 3 /min at 500°F to the pilot facility;
(2) a SD vessel with a two-fluid nozzle for injecting a lime slurry
and a bottom screw conveyor for removing ash; and
(3) a pulse-cleaned FF using high-temperature teflon bags with an
air-to-cloth ratio of 4.4 acfm/ft 2 . Ash from the FF could be
recirculated to the SD for reinjection with fresh slurry.
Testing was conducted with the pilot-scale SD/FF system in March 1985.
Results from other tests at Quebec City during the same test program are
presented in Section 5.2. These tests were conducted using a dry sorbent
injection system with the same FF at various FF inlet temperatures. In
Section 2.2.1.9, results from testing of the Quebec City combustor with the
existing ESP are presented. The combustor was operated under normal
conditions throughout testing of the pilot-scale SD/FF system. The SD/FE
was operated with and without FF ash recycle, two test runs at each
condition, and at a single SD outlet temperature of 285°F. Flue gas was
sampled at the SD inlet, SD outlet, and EF outlet for PM, HC1, SO 2 , metals
(arsenic, cadmium, chromium, mercury, lead, and nickel), CDD/CDF, and other
organics. PM and metals data were not taken at the mid-point sampling
location.
Acid gas data are presented in Table 7-24. Outlet SO 2 concentrations
ranged from 31.3 to 61.0 ppm at 7 percent 02 for the four test runs. The
average removal efficiencies for SO 2 , with and without fly ash recycle, were
61 and 66 percent, respectively, yielding corresponding outlet emissions of
7-46
-------
-.4
TABLE 7-24. ACID GAS DATA FOR QUEBEC CITY PILOT SD/FF
Test
Condition
Run
Number
FF InLet
rempe ature
C F)
Stoichi-
ometric
Ratio
Acid
Gas Concentrations
Acid Gas RemovaL Efficiency
(percent)
(ppmv. dry at 7%
Inlet Midpoint
SO 2 HCL 502 HCL
02)
Outlet
SO 2 HCL
SD
SO 2 HCI.
FF
SO 2 HCL
Overall
502 HCL
Combustor = NormaL
7
280
1.5
128 338
76.7 168
33.9 27.8
40.2
56.4
55.8 81.2
73.5 91.8
501FF = NormaL
8
283
1.4
100 354
67.7 173
41.5 34.7
23.5
51.1
38.7 80.0
58.6 90.2
No ash recycLe
Average
282
1.5
114 346
72.2 161
37.7 31.3
31.9
53.8
47.3 80.6
66.1 91.0
Combustor = Normal
9
285
1.3
101 461
63.1 152
31.3 35.0
37.7
67.1
50.5 77.0
69.2 92.4
SD/FF NormaL
10
283
1.1
127 552
86.7 176
61.0 54.7
31.8
68.2
29.7 68.8
52.0 90.1
With ash recycle
Average
284
1.2
114 507
74.9 164
46.2 44.9
34.8
67.7
40.1 72.9
60.6 91.3
-------
46 ppm and 38 ppm. Outlet HC1 concentrations were between 27.8 and 54.7 ppm
at 7 percent 02. The HC1 removal efficiency for both SD/FF operating
conditions averaged 91.1 percent.
As shown in Figure 7-8, stoichiometric ratio had some effect on SO 2
removal efficiency. SO 2 removal efficiency generally increased with
increasing stoichiometric ratio. For Runs 9 and 10, which had ash recycle,
the actual stoichiometric ratio is somewhat higher than the value shown,
depending on the amount of unreacted lime in the ash recycle. HC1 removal
efficiency was independent of stoichiometric ratio over the narrow range of
stoichiometric ratios tested. Inlet SO 2 and HC1 concentrations did not
consistently affect performance.
Removal of SO 2 across the SD as well as across the FF was very similar
with and without ash recycle. SO 2 removal across the SD at both conditions
averaged 30 to 35 percent. SO 2 removal across the FF averaged about 40 to
45 percent. Although overall HC1 removal efficiencies were nearly identical
with and without ash recycle, HC1 removal across the SD was higher with ash
recycle (68 versus 54 percent). Conversely, HC1 removal across the FF was
higher without ash recycle (81 versus 73 percent). Whether these
differences are because of the use of recycle or are due to normal variation
of performance cannot be ascertained from the available data.
The use of ash recycle does not appear to significantly change
performance relative to SO 2 and HC1. By using ash recycle, however, the
amount of fresh sorbent required may be lessened.
In Table 7-25, particulate data are presented for the four performance
tests as well as for four characterization tests performed prior to the
performance tests. The characterization tests were used to familiarize
testing personnel with the SD/FF system. It was determined during these
tests that no particulate samples would be collected at the outlet during
the performance tests because insufficient PM was collected for analysis.
Of the four characterization tests, three yielded nondetectable amounts of
PM (<0.0002 gr/dscf) and one had a concentration of 0.0018 gr/dscf.
Metals data are presented in Table 7-26. Greater than 99.9 percent of
the arsenic, cadmium, chromium, lead, and nickel were removed across the
SD/FF both with and without ash recycle. Removal efficiencies for mercury
were consistently near 95 percent.
7-48
-------
100 -
90 El Without ash recycle
With ash recycle
0 ,
C)
0,
a.
8O
C.)
0,
C)
[ 1
> 70-
0
0,
0
U)
60 -
El
50 -- I I
1 1.2 1.4 1.6
Stoichiometric Ratio
Figure 7-8. SO 2 removal efficiency as a function of
stoichiometric ratio at Quebec City.
-------
TABLE 7-25. PARTICULATE DATA FROM QUEBEC CITY PILOT SD/FF
aWN = not measured.
bND = not detected. Considered as zero in evaluating averages.
Test
Condition
Run
Number
FF Inlet
Tempgrature
( F)
Flue Gas
Flow
(acfm)
InLet PM
Concentration
(gr/dscf at
(12% C0 2 )
OutLet PM
Concentration
(grldscf at
(12% C0 2 )
PM
Removal
Efficiency
(percent)
Combustor s Normal
SD/FF Normal
(284 F FF inlet
temperature)
7
8
281
283
4,115
4,213
2.43
2.61
NM 5
MM
“
- -
Average
282
4,164
2.52
NM
--
Combustor NormaL
SD/FF = Ash RecycLe
(284°F FF inlet
temperature)
9
10
285
283
4,125
4,125
3.33
3.30
NM
NM
--
--
Average
Combustor = Normal
SD/FF = Normal
(302°F FF inLet
temperature)
6CTC
7C1
8CT
9C1
284
290
286
284
291
4,125
3,565
3,591
3,530
3,589
3.32
1.61
1.99
1.01
1.84
NM
NOb
ND
0.0018
ND
--
too
100
99.8
100
Average
288
3,568
1.61
0.0005
99.9
C •• = characterization test.
-------
U’
TABLE 7-26. METALS EMISSIONS DATA FOR QUEBEC CUT PILOT SD/U
Test
Condition
Outlet PM
FF Inlet Concentration
Run Tempgrature (gr/dscf at
Number ( F) 12Z C0 2 )
Inlet Concentration
Outlet Concentration
As
Cd
(ugldsc.
Cr
•
Pb
°2
Hg
NI
Ai
(ugJds m
Cd Cr
•1
Pb
7 0 )
N
Ni
As
Removal
Cd
Efficiency (percent)
Cr Pb Hg
Ni
Combustor = Normal
SQ/PP Normal
7
8
280
283
--
--
84
130
1,207
1,242
1.320
1,583
25,520
34,518
139
234
686
742
0.05
0.02
ND 0.44
ND ND
2.4
ND
6.1
13.9
1.84
0.80
99.94
99.98
100
100
100
100
100
100
95.6
94.1
100
99.9
Average
282
107
1,225
1.452
30,019
187
714
0.04
ND 0.22
1.2
10.0
1.32
99.96
100
100
100
94.9
99.9
Combustor = Normal
SQ/PP • Ash Recycle
9
10
285
283
--
--
138
117
1,187
1,117
1,575
1,760
36,548
31,377
331
388
3,937
1,102
0.06
0.01
ND 0.92
ND 0.47
5.1
7.1
17.0
21.6
2.31
1.89
99.96
99.99
100
100
99.97
99.97
99.99
99.98
94.9
94.4
99.99
99.8
Average
284
128
1,152
1,668
33,963
360
2,520
0.04
ND 0.70
6.1
19.3
2.10
99.97
100
99.97
99.98
94.7
99.99
not detected.
Considered as zero in evaluating
average..
-------
CDD/CDF data for the pilot-scale SD/FF system are presented in
Table 2-27. Outlet CDD/CDF concentrations were not detected for three of
four runs and were 2.52 ng/dscm for Run 9 which had ash recycle. Because
CDD/CDF was detected in only one run, the effects of inlet CDD/CDF and
temperature cannot be evaluated. Very little CDD/CDF removal occurred
across the SD at Quebec City. For three of four runs (two without recycle
and one with ash recycle), the CDD/CDF removal efficiency across the SD was
10 to 14 percent. During Run 10, however, the CDD/CDF at the SD outlet was
higher than at the SD inlet. This may be due to desorption of CDD/CDF from
the recycled ash, but also may be within sampling and analytical error.
These data indicate that the overall pilot-scale SD/FF system at Quebec City
when operated 285°F was capable of outlet CDD/CDF concentrations less than 3
ng/dscm and that effective operation of the FF is essential for high levels
of CDD/CDF reduction.
7.2.8 Stanislaus County 16 ’’ 7
The Stanislaus Waste-to-Energy Facility in Crows Landing, California
consists of two identical Martin GmbH mass burn, waterwall combustors, each
capable of combusting 400 ton/day MSW. Each combustor is equipped with
Exxon’s Thermal DeNOx system consisting of ammonia injection into the upper
furnace for NO control.
Emissions are controlled downstream of the boiler with a Flakt spray
dryer/fabric filter system. In the SD, slaked lime slurry is injected
through two-fluid nozzles, with the slurry feed rate controlled according to
the stack so2 concentration and the dilution water flow controlled according
to the SD outlet temperature. A residence time in the SD of 15 seconds is
maintained to dry the slurry. Flue gas exiting the SD flows through the
pulse-jet FF at 94,000 acfm and 285°F. The FF has six compartments of
teflon-coated fiberglass bags (1,596 bags total) and a net air-to-cloth
ratio of 3.2 acfm/ft 2 .
In December 1988, compliance testing was conducted at the facility.
The combustor and air pollution control system operated normally during
testing. At the SD/FF outlet, flue gas was sampled for SO 2 , HC1, PM, metals
(arsenic, cadmium, chromium, lead, mercury, and nickel), CDD/CDF, and other
organics. Nitrogen oxides were also measured to show performance of the
7-52
-------
TABLE 7-27. CDD/CDF DATA FOR QUEBEC CI EY PILOT SD/FF
U,
Test
Condition
Run
Number
Fl Inlet
Tempe ature
( F)
CDD/CDF Concentration
CDD/CDF Removal
Efficiency
(percent)
(ng/dscm at 7%
Inlet Midpoint
O )
Outlet
SD
FF
Overall
Combustor = Normal
7
280
1,954 1,703
NDa
12.9
100
100
SD/FF = Normal
8
283
1,574 1,359
ND
13.7
100
100
No ash recycle
Average
282
1,764 1,531
ND
13.3
100
100
Combustor = Normal
9
285
2,685 2,409
2.52
10.3
99.9
99.9
SD/FF Normal
10
283
1,629 2,213
ND
-35.8
100
100
With ash recycle
Average
284
2,157 2,311
1.26
-12.8
99.9
99.9
aND = not detected.
Considered as zero in evaluating averages.
-------
Thermal DeNO system. These data are presented in “Municipal Waste
Combustors - Background Information for Proposed Standards: Control of NO
Emissions.
Acid gas data are presented in Table 7-28. Outlet SO 2 concentrations
ranged from 0.3 to 7.7 ppm at 7 percent 02 over the course of six runs,
three at each unit. Outlet SO 2 concentations averaged 2.9 ppm at Unit 1 and
5.4 ppm at Unit 2. The lowest SO 2 removal efficiency was obtained at the
lowest inlet SO 2 concentration, but no dependence of SO 2 removal efficiency
on inlet concentration was observed for the other runs. The corresponding
removal efficiencies were above 88 percent for five of six runs with the
sixth run at 63.7 percent. The average SO 2 removal efficiency was
89.8 percent. The average inlet SO 2 levels of 63 ppm are low relative to
other MWC’s. Outlet HC1 concentrations were between 0.6 and 4.1 ppm at 7
percent 02 over six runs. Outlet HC1 concentrations averaged 0.7 Pm at
Unit 1 and 2.6 ppm at Unit 2. Inlet Rd concentration was not measured.
Although outlet SO 2 and HC1 concentrations were higher at Unit 2, there is
no apparent cause for this difference.
The FF inlet temperature was consistent between runs. Although the
stoichiometric ratio was not calculated, the effect of lime slurry rate can
be evaluated. As shown in Figure 7-9, SO 2 removal efficiency increased with
increasing lime slurry feed rate for both Units 1 and 2.
Particulate data are presented in Table 7-29. Outlet PM concentrations
ranged from 0.0011 to 0.0086 gr/dscf at 12 percent CO 2 over the six test
runs and averaged 0.0055 gr/dscf for Unit 1 and 0.0022 gr/dscf for Unit 2.
Inlet PM was not measured. There was no significant variation in
air-to-cloth ratio during testing (3.4 to 3.6 acfm/ft 2 ). Based on these
data, the SD/FF at Stanislaus County with a pulse-jet FF operating at 295°F
with a net air-to-cloth ratio of 3.6 acfm/ft 2 or less, outlet PM emissions
of 0.0086 gr/dscf can be achieved.
Metals emission data are presented in Table 7-30. Metals
concentrations at the outlet were similar at both units. Arsenic and
cadmium concentrations averaged less than 2.5 ug/dscm. Chromium
concentrations averaged about 10 ug/dscrn, while nickel averaged an outlet
concentration of about 20 ug/dscm. The average lead concentrations were
7-54
-------
TABLE 7-28. ACID GAS DATA FOR STANISLAUS COUNTY
aRUfl Number consists of unit number followed by the run number on that
bTota( sulfur oxides concentration
Clemperature monitor at FF inlet ma functioned. Temperature estimated
temperature drop across the FF (10 F).
unit.
Test
Condition
Run
Numbera
FF inlet
Temperature
(°F)
Lime
Rate
Slurry
(Lb/hr)
Acid
Gas
Concentration
Inlet
502 b
(ppmv.
dry
at
7% 0 )
0ut et
SO Removal
Ef iciency
so 2 b
HCI.
(%)
Combustor = Normal
DeNO /SD/FF = Normal
X
1-10
1-11
1-12
292
297
297
303
303
277
57.6
76.6
66.0
0.3
0.8
7.6
0.88
0.69
0.61
99.5
99.0
88.5
Average (Unit 1)
295
294
66.7
2.9
0.73
95.7
2-26
2-27
2-28
291
292
297
317
292
250
89.7
65.5
21.2
3.0
5.6
7.7
2.2
4.1
1.6
96.7
91.5
63.7
Average (Unit 2)
293
292
58.8
5.4
2.6
84.0
U,
U ,
from measured value at the stack and an assumed
-------
100
.
95
-S
C
90—
U
>
U
C
85-
C.)
5 ’..
w
>
0
E
0
- ON 75
Cl)
1 . .71
o
70 • Unit 1
• Unit2
65 -
.
60 -
240 260 280 300 320
Lime Slurry Feed Rate (lb/hr)
Figure 7-9. SO 2 removal efficiency as a function of lime slurry feed rate
at Stanislaus County.
-------
TABLE 7-29. PARTICULATE DATA FOR STANISLAUS COUNTY
Test
Condition
Run a
Number
FF
Temp
(
Inlet
3 rature
F)
Flue Gas
Flow
(acfm)
Outlet PM
Concentration
(gr/dscf at 12%
C0 2 )
Combustor Normal
DeNO /SD/FF = Normal
1-10
1-11
1-12
292 b
297 b
297
101,800
100,500
105,000
0.0045
0.0034
0.0086
Average (Unit 1)
295
102,400
0.0055
2-26
2-27
2-28
291
292
297
103,400
106,100
103,300
0.0027
0.0011
0.0028
Average (Unit 2)
293
104,300
0.0022
aRun Number consists of unit number followed by the run number on that unit.
bTemperature monitor at FF inlet malfunctioned. Temperature estimated from
mea ured value at the stack and an assumed temperature drop across the FF
(10 F).
7-57
-------
TABLE 7-30. METALS DATA FOR STANISLAUS COUNTY
Test
Condition
Run
Number
FF Inlet
Teunp 8 rature
C F)
Outlet PM
Concentration
(gr/dscf at 12%
b
C0 2 )
Outlet Concentration
As
(ua/dscm
Cd Cr
at 7Z0 2 )
Pb
Hg
Ni
Combustor = Normal.
DeMO /SD/FF NormaL
X
1-14
1-16
1-19
297 c
291
296
--
--
--
0.99
2.14
1.65
3.61
1.31
0.23
10.7
7.7
18.1
40.3
8.4
19.6
360
681
457
30.6
15.8
31.0
Average (Unit 1)
295
0.0055
1.59
1.72
12.2
22.8
499
25.8
2-38
2-40
2-42
297
287
287
-
--
•-
2.2k
ND
ND
2.07
2.49
1.76
16.6
10.5
2.3
43.0
36.3
31.0
391
446
550
18.9
18.8
21.2
Average (Unit 2)
290
0.0022
0.75
2.11
9.8
36.8
462
19.6
Overall Average
293
0.0038
1.17
1.91
11.0
29.8
481
22.7
8 Run Number consists of unit number followed by the run number on that unit.
bpN and metals data not collected simultaneously. Average PM only reported.
Clemperature monitor at FF inlet matfunctioged. Temperature estimated from measured value at the stack and an
assumed temperature drop across the FF (10 F).
dM0 = not detected. Considered as zero in evaluating averages.
U,
cx
-------
23 and 37 ug/dscm. Based on typical uncontrolled metals concentrations,
these metals were removed at greater than 98 percent. Outlet mercury
emissions of 462 and 499 ug/dscm at 7 percent 02, for Unit 2 and Unit 1,
respectively, were similar to typical uncontrolled values of 248 to 1,030
ug/dscm (see Section 1.2), suggesting little or no mercury removal.
In Table 7-31, CDD/CDF data are presented. Outlet COD/COF
concentrations ranged from 4.60 to 8.90 ng/dscm at 7 percent 02 for both
units and averaged 6.25 ng/dscm for Unit 1 and 6.53 ng/dscm for Unit 2.
Inlet CDD/CDF concentrations were not measured. Thus, the SD/FF at
Stanislaus County demonstrated achievable outlet CDD/CDF levels less than
10 ng/dscm.
7.3 SUMMARY OF PERFORMANCE
Performance of individual SD/FF systems was evaluated in Section 7.2.
The data are evaluated as a whole in this section. Section 7.3.1 evaluates
acid gas performance, Section 7.3.2 evaluates particulate performance,
Section 7.3.3 evaluates metals performance, and Section 7.3.4 evaluates
CDD/CDF performance.
7.3.1 Acid Gas
Factors affecting acid gas removal by SD/FF systems include
stoichiometric ratio, SD outlet temperature (FF inlet), and inlet SO 2
concentration. Increasing stoichiometric ratio increased both SO 2 and HC1
removal efficiency, as shown in Figures 7-10 and 7-11, respectively. Based
on the available data, a stoichiometric ratio in excess of 3 is generally
necessary to obtain consistently high acid gas removal efficiencies of
90 percent SO 2 and 97 percent HC1.
Also important is the temperature at the SD outlet. Increasing the
temperature decreases SO 2 and HC1 removal efficiencies and prevents SD/FF
systems from achieving consistently good performance as shown in
Figures 7-12 and 7-13, respectively. Other factors, such as inlet SO 2
concentrations, stoichiometric ratio, and type of SD control loop also
affect performance.
Better than 90 percent SO 2 removal efficiency was demonstrated for at
least one run at four of the six facilities with removal efficiency data.
One of the two facilities not achieving 90 percent SO 2 removal efficiency
7-59
-------
TABLE 7-31. CDD/CDF DATA FOR STANISLAUS COUNTY
Test
Condition
Run
Number
a
FF
Temp
(
Inlet
rature
F)
Outlet CDD/CDF
Concentration
(ng/dscm at 7% 02)
Combustor = Normal
DeNO /SD/FF = Normal
1-18
1-21
1-22
305
286
287
4.60
5.26
8.90
Average (Unit 1)
293
6.25
2-37
2-39
2-45
297
292
292
6.17
8.45
4.98
Average (Unit 2)
294
6.53
aRUfl Number consists of unit number followed by run number.
7-60
-------
100— 0
0 0
0
90- A
A
-S
o 80-
0
a
o A R
C
0 +
0
U
60 A +
5 0 A A
•
A Marion County
40 + Qu.b.c City
o
30- I I I
1 3 5 7
Stoichiometric Ratio
Figure 7-10. SQ 2 removal efficiency as a function of stoichiometric
ratio for SDIFF systems at less than 300°F.
-------
100 - ________________________
U 0
99- 0 o
98 -
A
A
96-
A
91
-.. E
90-+
C)
88 -
•
A Marion County
87 - + Qu.b.c City
o Comm.rc.
86 - _____________
85- I
1 3 5 7
Stoichiometric Ratio
Figure 7-11. HCI removal efficiency as a function of stoichiometric
ratio for SD/FF systems at less than 300°F.
-------
100- 0 U 7 _________
0
V A
• Bidd.tord
A 0 Comm.rc.
o Long B.ach
80 — A A Msrion County
+ Qu.b.c City
A • + V Stani.Iau. County
70- + _________
- V
—
C +
C)
+ A
• 50- A
a
>1
o 40-
C
0
30-
20
4 0
E
0 10-
- A
10 -
-20 -
-30- I I I
260 280 300 320 340 360
FF Inlet Temperature (°F)
Figure 7-12. SQ 2 removal efficiency as a function of
FF inlet temperature for SD/FF systems.
-------
100 - ____________________________
AA 9 o
98- U 0 A
A
96 -
A
94 A
A A
0’)- +
C +
78 ____+ A
• BIdd.tord
76 - 0 Comm.rc.
A Marion County
+ Qu.b.c City
74 -
72- I I
260 280 300 320 340 360
FF Inlet Temperature (°F)
Figure 7-13. HCI removal efficiency as a function of
FF inlet temperature for SD/FF systems.
-------
was Biddeford, which had low inlet SO 2 concentrations (86 to 129 ppm) and
low outlet SO 2 concentrations (14 to 31 ppm). At the Quebec City pilot
SD/FF, SO 2 removal efficiency ranged from 52 to 74 percent. The
stoichiometric ratio was consistently less than 1.5 at Quebec City, which
probably contributed to the lower SO 2 removal efficiencies. Three of the
four facilities with HC1 removal efficiency data averaged greater than 95
percent HC1 removal efficiency. The one facility not demonstrating 95
percent HC1 removal efficiency was the Quebec City pilot SD/FF, which
operated with a relatively low stoichiometric ratio. In addition, several
vendors of SD/FF systems claim that 90 percent SO and 95 percent HC1
removal efficiencies are readily achievable.
7.3.2 Particulate Matter
Analysis of PM emissions data from the eight facilities shows that
consistently low PM emissions can be obtained. All eight facilities
averaged less than 0.01 gr/dscf at 12 percent CO 2 . These included four
reverse-air and four pulse-jet cleaned fabric filters.
7.3.3 Metals
Removal efficiencies for arsenic, cadmium, chromium, and lead were
consistently demonstrated at 99 percent or greater. Nickel control was
variable from site to site. At two sites, nickel removal efficiency was
measured at greater than 99 percent. At another two sites, nickel removal
was measured at 14 and 50 percent. At two sites where only outlet nickel
emissions were measured, removal efficiencies of greater than 98 percent are
suggested. There is no apparent reason for the widely varying performance.
Mercury removal varied from site to site. Inlet and outlet mercury
concentrations were measured during five separate test programs at four
facilities, with the average removal efficiencies of 100, 80.0, 0, 86.3, and
94.8 percent obtained. Zero mercury removal efficiency was obtained during
one test program at Commerce. Later tests at Commerce demonstrated an
average of 80 percent mercury removal. Three of these runs were conducted
firing commercial refuse and averaged 70 percent mercury removal efficiency.
The removal of mercury did not appear to depend solely on FF inlet
temperature.
7-65
-------
Outlet mercury concentrations at 10 separate test programs ranged from
not detected to 940 ug/dscm. The average concentration of all these test
programs was 250 ug/dscm. Emissions from seven of the 10 test programs were
less than 240 ug/dscm.
It has been theorized that mercury removal is enhanced by carbon in the
fly ash providing adsorption sites for the mercury. 21 In Figure 7-14,
mercury concentration at the outlet is plotted as a function of inlet PM
concentration. These data suggest decreased n ercury emissions with
increasing inlet PM concentration. In Figure 7-15, mercury emissions are
plotted as a function of inlet CDD/CDF concentration. Recognizing that
increased CDD/CDF concentrations may be associated with increased carbon in
the PM, the graph indicates decreased mercury emissions for higher inlet
CDD/CDF concentrations and thus, more carbon in the PM.
Based on measured removal efficiencies of metals by SD/FE systems,
arsenic, cadmium, chromium, and lead can be removed at 99 percent
efficiency. Nickel removal is generally high (98 percent), but varies. The
level of mercury removal by SD/FE systems is uncertain, but an emission
level of 300 ug/dscm appears achievable.
7.3.4 CDDICDF
Outlet CDD/CDF concentrations during normal conibustor operation were
less than 9.8 ng/dscm for all runs. The average outlet CDD/CDF
concentrations were below 6.4 ng/dscm for the eight facilities. These
rest lts were obtained at inlet CDD/CDF concentrations ranging from 27 to
1,000 ng/dscm and FF inlet temperatures of 270 to 300°F. Thus, spray
dryer/fabric filter systems operating at FE inlet temperatures of 300°F or
less can consistently achieve average outlet CDD/CDF concentrations of less
than 10 ng/dscm at 7 percent 02.
7-66
-------
1000
900 I MId-Connscticut
+ Qu.b.c City
K Bidd.ford
800 A Comm.rcs(1988)
/ Comm.rc.(1987)
Li Long B.ach
E 700
V Marion County
(ft
V
600
C
0
500
400-
C
0
C)
a 0)
.-.J I
V
200-
100
A
A +
+
0 - I I I
0.8 1.2 1.6 2 2.4 2.8 3.2 3.6
Inlet PM Concentration (gr/dscf @ 12% C )
Individual run, during earn. t.st p.rlod at Comm.rcs.
Figure 7-14. Effect of inlet PM on mercury emissions for systems
with spray dryer/fabric filter systems
-------
1000
900 >< + Quebec City
0 Biddeford
800 Commerce(1988)
x Commerce(1987)
v Marion County
700 •
400
C
‘- .l 0
0
300
2O0 x
100
0 I I I
0 400 800 1200 1600 2000 2400 2800
Inlet CDD/CDF Concentration (ng/dscm @ 7% 02)
Figure 7-15. Effect of inlet CDD/CDF on mercury emissions for
MWCs with spray dryer/fabric filter systems
-------
7.4 REFERENCES
1. Kiam, S., G. Scheil, M. Whitacre, J. Surman (Midwest Research
institute), and W. Kelly (Radian Corporation). Emission Testing at an
ROE Municipal Waste Combustor. Prepared for U.S. Environmental
Protection Agency, North Carolina. EPA Contract No. 68-02-4453,
May 6, 1988.
2. Letter with attachments from G. M. Bates, Main Energy Recovery Company
(MERC), to J.R. Farmer, U.S. Environmental Protection Agency, ESD. No
date. Response to information request.
4. McDannel, M. D., L. A. Green, and A. C. Bell (Energy Systems
Associates). Results of Air Emission Test During the Waste-to-Energy
Facility. Prepared for County Sanitation Districts of Los Angeles
County, Whittier, California. December 1988.
J. Donnelly, Joy Technologies, Inc. by 0. M. White, Radian
April 13, 1989.
6. Municipal Waste Combustors--Background
Standards: Control of NO Emissions.
Agency. EPA-45O/3-89-27d August 1989.
7. Ethier, D. 0., L. N. Hottenstein, and E. A. Pearson (TRC Environmental
Consultants). Air Emission Test Results at the Southeast Resource
Recovery Facility Unit 1, October - December, 1988. Prepared for Dravo
Corporation, Long Beach, California. February 28, 1989.
8. Anderson, C. 1. (Radian Corporation). CDD/CDF, Metals, and Particulate
Emissions Summary Report, Mid-Connecticut Resource Recovery Facility,
Hartford, Connecticut. Prepared for U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina. EMB Report
No. 88-MIN-09A. January 1989.
9. Vancil, M. A. and C. L. Anderson (Radian Corporation). Summary Report,
CDD/CDF, Metals, HC1, SO , NO , CO and Particulate Testing, Marion
County Solid Waste-to-En rgy acility, Inc., Ogden Martin Systems of
Marion, Brooks, Oregon. Prepared for U.S. Environmental Protection
Agency. EMB Report No. 86-MIN-03A. September 1988.
3. McDannel, M. D., L. A. Green, and
Associates). Air Emissions Tests
Facility, May 26 - June 5, 1987.
Districts of Los Angeles County.
B. 1. McDonald (Energy Systems
at Commerce Refuse-to-Energy
Prepared for County Sanitation
Whittier, California. July 1987.
5. Interview of
Corporation.
Information for Proposed
U. S. Environmental Protection
10. Letter Report from M. A. Vancil, Radian Corporation to C. E. Riley, EMB
Task Manager, U.S. Environmental Protection Agency. Emission Test
Results for the PCDD/PCDF Internal Standards Recovery Study Field Test:
Runs 1, 2, 3, 5, 13, 14. July 24, 1987.
7-69
-------
11. Anderson, C. L., K. L. Wertz, M. A. Vancil, and J. W. Mayhew (Radian
Corporation). Shutdown/Startup Test Program Emission Test Report,
Marion County Solid Waste-to-Energy Facility, Inc., Ogden Martin
Systems of Marion, Brooks, Oregon. Prepared for U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. EMB Report
No. 87-MIN-4A. September 1988.
12. Anderson, C. L., et. al. (Radian Corporation. Characterization Test
Report, Marion County Solid Waste-to-Energy Facility, Inc., Ogden
Martin Systems of Marion, Brooks, Oregon. Prepared for U. S.
Environmental Protection Agency, Research Triangle Park, North
Carolina. EMB Report No. 86-MIN-04. September 1988.
13. Roy F. Weston, Incorporated. Penobscot Energy Recovery
Facility, Orrington, Maine, Source Emissions Compliance
Incinerator Units A and B (Penobscot, Maine). Prepared
September 1988.
14. Zaitlin, S., State of Maine, Department of Environmental Protection,
Board of Environmental Protection. Air Emission License Finding of
Fact and Order, Penobscot Energy Recovery Company, Orrington, Maine.
February 26, 1986.
15. The National Incinerator Testing and Evaluation Program: Air
Control Technology. EPS 3/UP/2, Environment Canada, Ottawa,
September 1986.
Pollution
17. Hahn, J. L. and 0. S. Sofaer.
Stanislaus County, California
the International Conference
Florida. April 11-14, 1989.
18. Telecon.
June 12,
19. Telecon. D.
Flakt, Inc.
times.
Air Emissions Test Results
Resource Recovery Facility.
on Municipal Waste Combustion.
pp. 4A-1 to 4A-14.
21. Personal communication. Hahn, J., Odgen Projects, Inc. and M. Johnston,
EPA/OAQPS. April 1989.
Company
Test Report
for GE Company.
16. Hahn, J. L. (Ogden Projects, Inc.) Environmental Test Report. Prepared
for Stanislaus Waste Energy Company, Crows Landing, California. OPT
Report No. 177R. April 7, 1989.
from the
Presented at
Hollywood,
0. M. White, Radian Corporation, with J. Donnelly, Joy, Inc.
1989. Achievable acid gas reductions and retrofit times.
M. White, Radian Corporation, with J. Buschmann,
June 8, 1989. Achievable acid gas reductions and retrofit
20. Telecon. 0. M. White, Radian Corporation, with J. Zmuda,
Research-Cottrell, Inc. June 8, 1989. Achievable acid gas
and retrofit times.
reductions
7-70
-------
APPENDIX A. HYDROGEN CHLORIDE AND SULFUR DIOXIDE EMISSIONS DATA
A.1 INTRODUCTION
This appendix reviews emission data from two MWC’s equipped with spray
dryers to (1) investigate relationships between HC1 and SO 2 removal behavior
and (2) assess the variability of SO 2 emissions. The two MWC’s evaluated are
the Milibury Resource Recovery Facility in Milibury, Massachusetts and the
Marion County Solid Waste-to-Energy Facility in Brooks, Oregon. A brief
description for each facility is provided in Section A.2. Section A.3
describes the data collected at each facility. Section A.4 examines the
relationship between HC1 and SO 2 emissions and removal efficiency for these
two data sets. Section A.5 examines the variability in SO 2 emissions at the
Milibury MWC.
A.2 FACILITY DESCRIPTIONS
A.2.1 Milibury ’
The Milibury facility consists of two identical furnace, boiler, and
flue gas treatment systems. Each furnace is designed to process 750 tons/day
of MSW. Acid gas and particulate emissions from each furnace are controlled
by separate spray dryer and electrostatic precipitator (ESP) systems. The
Milibury facility was designed, constructed, and is operated by Wheelabrator
Environmental Systems, Inc. HC1 and SO 2 data were collected from Unit 2 as
part of a two-month test program conducted during July - September, 1988.
A process schematic for Unit 2 is shown in Figure A-i. MSW is charged
to a Babcock and Wilcox waterwall furnace and boiler unit that is equipped
with a Von Roll reciprocating, inclined grate. Each boiler is rated to
produce about 190,000 lb/hr of superheated steam at 825°F and 850 psia. The
combined steam from the two units is supplied to a turbine/generator set
which is rated at 40 megawatts. Auxiliary fuel (natural gas) is used during
startup and shutdown and during low load operation. The furnace flue gases
pass up through the waterwail section of the furnace and then into
superheater and economizer heat transfer passes. After passing through the
spray dryer and ESP, the flue gas exhausts to the atmosphere through a
reinforced concrete stack which is common to both units.
-------
N)
psai* ca.s
50 2 M0 .
0 2 .CPACflY
ELECTW 6TAT1C
ST $(
0
TO
N UUc (S
Figure A-i.
Milibury Resource Recovery Facility process schematic.
-------
The emissions control system was designei by Wheelabrator Air Pollution
Control Systems. Slaked lime, along with metered dilution water, is injected
into the dryer vessel through two-fluid nozzl s. The lime slurry reed rate
is automatically controlled to achieve either an SO 2 emission rate (the unit
is permitted for a maximum SO 2 concentration at the stack of 0.21 lbs per
million Btu of MSW fired) or a percent SO 2 removal efficiency requirement,
whichever is more stringent. Dilution water is added to the lime slurry to
reduce the flue gas temperature. The spray dryer outlet temperature is
typically controlled around 255°F. The dry solids and fly ash from the spray
dryer are collected in a three-field ESP having a design SCA of 333 ft 2 per
1,000 acfm at a flue gas flow rate of 160,000 acfm.
During the emission test program, SO 2 , HC1, oxygen (02), and other flue
gas constituents were measured between the furnace and the spray dryer (inlet
conditions) and between the ESP and fan (outlet conditions). A discussion of
the continuous emission monitors employed and other relevant information
about the test program is contained in the test report. 1
A.2.2 Marion County 2
The Marion County MWC consists of two identical 275-ton/day, mass burn,
waterwall Martin GmbH combustors followed by Teller-designed air pollution
control systems. Figure A-2 presents a process diagram for one of the two
trains. The Marion County MWC was designed, constructed, and operated by
Ogden-Martin, Inc. Testing was conducted on Unit No. 1 in June 1987 to
characterize the performance of the air pollution control system.
The air pollution control system consists of a cyclone, spray dryer, a
dry venturi, and a fabric filter. The flue gases leave the economizer
section and enter the bottom of the quench reactor through a cyclone where
oversize particles are removed. Slaked pebble lime slurry is injected
through an array of five two-fluid nozzles near the bottom of the reactor
vessel. The feed rate is varied to maintain the quench reactor outlet
temperature within an operating range of 125-149°C (258-300°F). The
stoichiometric ratio of lime to HC1 is maintained at approximately 2 - 2.5 to
ensure that peaks in acid gas levels are sufficiently controlled. The system
design did not allow the lime feed rate to be changed independently
A-3
-------
To Atmosphere
Figure A-2.
Marion County Solid Waste-to-Energy Facility process schematic.
-------
of the water feed rate. The lime concentration in the slurry is held nearly
constant. Therefore, as the water feed rate increases so did the lime flow.
A low pressure drop dry venturi located between the spray dryer and the
baghouse was used to inject Tesisorb, a proprietary sorbent, into the flue
gas at a set rate.
An Amerthem reverse air baghouse is located downstream of the dry
venturi for particulate collection. Each baghouse consists of six
compartments with 120 bags in each. The design air-to-cloth ratio is
2.3 acfm/ft 2 at a flue gas flow rate of 60,000 acfm. Particulate, lime, and
Tesisorb cake collected on the fabric is removed every 60 to 70 minutes.
Unspent lime in the filter cake provides an additional opportunity for acid
gas collection.
During the June 1987 test program, HC1, SO 2 , 02, and other parameters
were measured between the economizer and the cyclone (inlet to control
devices) and at the breeching to the outlet stack (control devices outlet).
Gas measurements were also collected between the spray dryer and baghouse,
but were not used in the present analysis. More information about the
continuous emission monitors, testing procedures, and other related aspects
of plant operation is contained in the test report. 2
A.3 DATA BASE DESCRIPTION
A.3.1 Milibury Data Base .
At the Milibury MWC, emission monitoring and process data were collected
on an hourly basis over 62 days during July and August, 1988. During this
period, the combustor and air pollution control systems were operated
“normally” by plant personnel.
The CEM data used in this analysis include hourly average 02 HC1, and
SO 2 measurements collected at both the inlet and outlet of the spray
dryer/ESP system. The HC1 and SO 2 gas concentrations were adjusted to 7
percent 02 using the applicable 02 measurements. Outlet HC1 measurements
collected via CEM were compared to those collected by wet chemistry reference
methods on several occassions during the testing to determine the accuracy of
GEM measurements. These comparisons showed significant scatter in the data
at CEM measurements below 4 ppm and reference method (RM) measurements below
A-5
-------
10 ppm. Since the reliability of these low measurements is questionable, it
was decided for this analysis to adjust all HC1 CEM measurements below 4 ppm
to a value of 10 ppm. This procedure results in a conservatively high
estimate of outlet HC1 emissions. However, because outlet HC1 concentrations
of 10 ppm typically correspond to HC1 removal efficiencies of roughly
98 percent, this conservatism in estimating outlet HC1 emissions is not
expected to materially influence the analysis of HC1 removal effeciencies.
For HC1 CEM measurements above 4 ppm, the following correlation was
developed based comparison of RM and CEM measurements:
HCLRM = 2.33 HC 1 CEM + 2.1.
This correlation (valid up to RM measurements up to 40 ppm) was employed in
the analysis to adjust outlet HC1 CEM measurements above 4 ppm HC1.
No adjustments were made to inlet HC1 CEM measurements or to
inlet/outlet SO 2 measurements (other than 02 basis adjustments) due to
acceptable agreement between CEM and RN measurements.
A.3.2 Marion County Data Base .
Emission data for the Marion County MWC were collected during the
“combustor variation TM phase of the facility characterization test program.
During this testing phase, combustion operating parameters were intentionally
varied to demonstrate their impacts on temperature profile, combustion
efficiency, and other parameters of interest. The spray dryer/fabric filter
system, however, was maintained at normal operating conditions. Thus, the
HC1 and SO 2 data reflect a wide range on spray dryer inlet gas conditions.
During each test condition, emissions data were collected over a two or three
hour period. A summary of the test conditions and combustion operating
characteristics is provided in Table A-i.
The CEM data used in this analysis include hourly average 02, HC1, and
SO 2 measurements collected at both the inlet and outlet of the spray
dryer/fabric filter system. Three hourly average measurements were available
for each test run with the exception of Runs 3A, 6B, and 8, for which only
two hourly average measurements were available.
A-6
-------
ABLE A-i. SUMMARY OF TEST CONDITION PARAMETERS FOR MARION
COUNTY COMBUSTOR VARIATION TESTS
Test Run No.
Combustor Condition
I
Baseline (i.e., normal operation)
2
Baseline (i.e., normal operation)
3A
Low load; low excess air
3B
High excess air
4
Low overfire air
5
High overfire air
6A
Low load
6B
Low load; high excess air
7
Low load; low excess air
8
Low load; low overfire air
9
Low load; high overfire air
A-7
-------
As with the Milibury data, the HC1 and SO 2 gas concentrations for Marion
County were adjusted to a 7 percent 02 basis using the applicable 02
measurements. No other adjustments were made to the data.
A.4 RELATIONSHIP BETWEEN HC1 AND SO 2 EMISSIONS
This section examines the relationship between controlled HC1 and S02
emissions for MWC’s equipped with spray dryers. Two measures of emissions
control are examined using the Millbury and Marion County MWC data: outlet
emission levels and pollutant removal efficiency.
A.4.1 HC1 Removal Efficiency Versus SO., Removal Efficiency
Since both HC1 and SO 2 are acid gases, they are expected to be absorbed
by the alkaline lime slurry used in spray dryers. Theoretically, the
reaction of a strong acid gas such as HC1 with dissolved calcium proceeds
rapidly. The reaction of SO 2 proceeds more slowly and over a narrower pH
range. Moreover, the rate of HC1 absorption is faster than that for SO 2 . 3
As a result, HC1 is expected to be preferentially absorbed relative to SO 2 ,
resulting in higher HC1 removal efficiencies than SO 2 removal efficiencies.
Pollutant removal efficiency is defined as:
C. C
Oil x 100 = % Removal Efficiency
where C 1 is the measured concentration of HC1 or SO 2 at the spray dryer
system inlet and is the measured concentration at the particulate matter
control device outlet. All pollutant concentrations are adjusted to 7
percent 02.
Milibury Results . At the Millbury facility, inlet HC1 concentrations
averaged 770 ppm while inlet SO 2 concentrations averaged approximately 200
ppm. A plot comparing HC1 removal efficiency with SO 2 removal efficiency
for the Milibury data is shown in Figure A-3. Each data point represents a
24-hour average. The data show a general positive correlation between HC1
and SO 2 removal--that is, the highest HC1 removal efficiencies are associated
with the highest SO 2 removal efficiencies; the lowest HC1 removal
efficiencies are associated with the lowest SO 2 removal efficiencies. The
HC1 removal efficiency was above 95 percent in all cases and generally was
between 98 and 99 percent. In contrast, the SO 2 removal efficiency was
A-8
-------
1’
uJ
0
-J
L i i
-J
I
t•• ’1ILLBUF’ [ : 1 AT..A. A1.IA.Li SIS:—24 HR
I-ICL LIflCIDJC’( v 2 U n ZIO K1
5Q2 E .D 1QV4L rrIzIa1;I (
Figure A-3.
c
0 0
0 0
I
4c.
HC1 removal efficiency vs. SO removal efficiency
at Milibury (24-hour averages .
/ \ .iI f’EF I Ei’
•:i
.1 ‘.
q. -,
.pj
P7
I I
-------
always lower, generally between 70 and 90 percent. The lowest SO 2 removal
efficiency of 50 percent was associated with an HC1 removal efficiency of 95
percent. These results clearly demonstrate a predictable relationship
between HC1 and SO 2 removal efficiencies and show that removal efficiencies
are higher for HC1 than for SO 2 .
The two lowest HC1 and SO 2 removal efficiency values in Figure A-3 were
examined as possible outliers. A review of process data from the two days in
question (July 21 and August 15) showed that both days were characterized
by unusually high spray dryer outlet gas temperatures during a significant
portion of the day. A review of the control room operator log showed the
following:
o On July 21, spray dryer outlet gas temperatures were intentionally
set at a higher level to dry out an ash/solids hopper. The hopper
had become moist due to blockage of atomizer air in one of the
spray dryer nozzles.
o On August 15, no obvious reason was found for the higher spray
dryer outlet temperatures although the increase occurred just after
a failure of the automatic control system on No. 1 boiler. The
high outlet temperatures persisted after automatic control was
re-established, however, and did not return to normal levels until
the boiler load was reduced approximately eight hours later.
Spray dryer outlet temperature is an important variable which affects HC1 and
SO 2 absorption. As the temperature increases, HC1 and SO 2 absorption
generally decline. However, since neither of these operating events was
considered to be outside the range of normal operating variability, no data
were excluded as outliers.
A plot of HC1 removal efficiency versus SO 2 removal efficiency was also
examined for 8-hour averages, as shown in Figure A-4. This plot shows the
same general findings as the plot of 24-hour averages, although the data are
more scattered owing to the higher variability associated with shorter
averaging times (see Section A.5 for further discussion).
The positive correlation between HC1 removal efficiency and SO 2 removal
efficiency is noteworthy. Since both HC1 and SO 2 gases are competing for
A- 10
-------
1\.•’i I I__ 1._ B I__I F’
DATA AN•.IALiSIS—$ HR A/G.
F’ E FE’ I ‘•J’ D
hCL nciu jc-r 5Z’2 IIO 1Ci’
ga
‘17
q.
I
Figure A-4.
5 Q2 l D 1C’VAL ErflCio.ici (
HC1 removal efficiency vs. SO 2 removal efficiency
at Milibury (8-hour averages).
0
0
0
DD
0 000
0
0
0
U
4.-)
1•• ° ’
-------
available dissolved calcium, operating conditions could exist durinq which
there is insufficient dissolved calcium available to react with both HC1 and
SO 2 . The fact that HC1 removal efficiencies were high over the full range of
operating conditions indicates there was a sufficient amount of dissolved
calcium available relative to the total concentration of acid gas. In
addition, Figures A-3 and A-4 indicate that the factors which influence SO 2
absorption (e.g., outlet gas temperatures and gas-liquid contact area) also
control HC1 absorption. As a result, HC1 and SO 2 removal efficiencies tend
to move in the same direction in response to changes in these factors.
Marion County Results . Unlike Millbury where data were collected over
an extended period, emission data at the Marion County facility were
collected during a series of 11 relatively short test runs of two to three
hours each. For all test runs, the inlet HC1 concentration averaged 650 ppm
while the inlet SO 2 concentration averaged 330 ppm. Figure A-S shows HC1
removal efficiency versus SO 2 removal efficiency using the averages for each
test run. Like Milibury, the Marion County results show a positive
correlation between HC1 removal efficiency and SO 2 removal efficiency. The
lowest HC1 removal efficiency (87 percent) corresponds to the lowest SO 2
removal efficiency (32 percent). All other HC1 and SO 2 removal efficiency
values were above this level. Similarly, the highest HC1 removal
efficiencies (98 percent) correspond to the highest SO 2 removal efficiencies
(91 percent).
Like Milibury, the results from Marion County confirm the expectation of
higher HC1 removal efficiency relative to SO 2 removal. Similarly, the
observations noted for the Millbury results related to excess calcium and
factors controlling both HC1 and SO 2 removal efficiencies also apply to the
Marion County results.
Conclusions . The results from Millbury and Marion County indicate that
HC1 removal efficiency and SO 2 removal efficiency are positively correlated.
Flue gas and spray dryer operating parameters that coitrol SO 2 removal
efficiency in a lime spray dryer also control HC1 re rwal efficiency. In
addition, HC1 is absorbed preferentially to SO 2 . Th e results indicate
that, once the relationship between HC1 removal effic ency and SO 2 removal
efficiency has been established at a given facili:y. t should be possible to
A-12
-------
M W C DATA A N A LY S I S — I C L
MARION COUNTY’ — TEST RUN AVERAGE
40 60
0
HC1 removal efficiency vs. SO removal efficiency
at Marion County (test run av rages).
80 100
‘ S S02
>-
0
z
w
0
I L-
>
0
ILl
a:
-J
0
I
100
99 —
98 —
97
96 —
95 —
94-
93 -
92 -
91
90 -
89
88 —
87 -
86 -
D
0
0
0
0
0
0
20
Figure A-5.
-------
maintain HC1 removal efficiency above a prescribed level by maintaining the
SO 2 removal efficiency above a corresponding level.
A.4.2 HC1 Removal Efficiency Versus SO 2 Emissions
The objective of this analysis was to determine if an inverse
relationship occurs between HC1 removal efficiency and SO 2 emissions.
Milibury Results . A plot of HC1 removal efficiency versus SO 2 emissions
using 24-hour averages from Milibury is shown in Figure A-6. The data are
generally scattered without a distinct relationship between HC1 removal
efficiency and SO 2 emissions. The HC1 removal efficiencies generally fall
between 97 and 99 percent, while the outlet SO 2 concentrations generally
range from 20 to 60 ppm. During two of the 24-hour periods, higher SO 2
emissions were observed (72 and 97 ppm), yet HC1 removal efficiencies were
also high (98 to 99 percent). These two 24-hour periods represent periods in
which inlet SO 2 concentrations were unusually high (both values were more
than two standard deviations above the mean). The SO 2 removal efficiency
for these periods, however, were near the mean value. These data suggest
that the lack of a distinct relationship between HC1 removal efficiency and
SO 2 emissions may be associated with variations in inlet concentrations of
both HC1 and SO 2 as well as the limited range over which HC1 removal levels
were observed (i.e., between 95 and 99 percent).
A plot of HC1 removal efficiency versus SO 2 emissions was also examined
for 8-hour average values (see Figure A-7). This plot also failed to show a
distinct relationship between HC1 removal efficiency and SO 2 emissions for
the same reasons discussed above. SO 2 emissions are more variable due to the
shorter averaging times. The lowest 8-hour average HC1 removal efficiencies
(90 to 95 percent) are associated with SO 2 emissions of 40 to 60 ppm. The
majority of the data, however, indicate that high HC1 removal efficiencies
(>98 percent) were achieved over a wide range of SO 2 outlet concentrations
(10 to 100 ppm).
Marion County Results . The relationship between HC1 removal efficiency
and SO 2 emissions is more pronounced for the Marion County data (as shown in
Figure A-8) than for the Millbury data, despite the fact that Marion County
averaging times were shorter (2 to 3 hours). This is most likely due to the
A- 14
-------
MILLBU DATA A1IALr SI3—24 HF’ Av’ F’ERl’JL:
PLOT OF HCL Ffl ID’1Ci c’lJrLD 5 ’2
° : 100 0
.“-
4.:
502 MI55IQI• ‘ ppm • 7 :2::
Figure A-6. HC1 removal efficiency vs. SO emissions
at Milibury (24-hour averages .
-------
,.AI’j,.AI_ •, F—IF ’ I:EF:hI,. [ u
I1(L rflClO.Ki v ‘ “.9L T Q2
______ 0
97 o % p°O oDD 00
0
q
0 0
0
z
92
0
I 4 1 F’
C’2 D 4I55IOt r ppm 1=12
Figure A-i. HC1 removal efficiency vs. SO 2 emissions
at Milibury (8-hour averages).
-------
MWC DATA ANALYSIS HCL vs S02
MARION COUNTY — TEST RUN AVERAGE
100
99
98 0
0
96
95 0
U
S .-, 94
>- 0
o 93
z 0
Ui 92
91
U..
90
-J 0
89
‘ s1 >
O 88
87
-J 86
0
1 85
84
83
82
81
80
0 100 200 300 400
Figure A-8. HC1 removal efficiency vs. SO emissions
at Marion County (test run av rages).
-------
wider range of HCI removal efficiencies observed during the Marion County
characterization tests. These data clearly confirm the inverse relationship
expected between HC1 removal efficiency and SO 2 emissions--higher HC1 removal
efficiencies are associated with lower SO 2 emissions and vice-versa. The
data show HC1 removal efficiencies of greater than 95 percent when SO 2
emissions were below 60 ppm. At higher SO 2 concentrations (150 to 400 ppm),
HC1 removal ranged from 87 to 93 percent.
Conclusions . Analysis of emissions data from Milibury and Marion County
indicate that HC1 removal efficiency and SO 2 emissions are inversely related,
although this relationship may be somewhat obscured by variations in inlet
HC1 and SO 2 levels and by the narrow range of HC1 removal levels, as at
Milibury. At both facilities, HC1 removal efficiency remained above 95
percent when SO 2 emissions were maintained near or below 60 ppm. As an
extension of the conclusion drawn above with respect to SO 2 removal
efficiency, this analysis indicates that it should be possible to maintain
HC1 removal efficiency above a prescribed level by limiting SO 2 emissions.
An analysis of variations in uncontrolled HC1 and SO 2 may be necessary,
however, before such a relationship can be established for a given facility.
A.5 SULFUR DIOXIDE EMISSION VARIABILITY
This section discusses a statistical analysis performed to evaluate
variability in SO 2 emissions and SO 2 percent reduction based on the CEM data
from the Milibury facility. The impact of averaging time on emissions
variability was also evaluated. Emission data from Marion County were not
analyzed for variability because the data were collected over short time
periods which are not suitable for time series analysis, an integral part of
the variability assessment.
A.5.1 Statistical Background 4 ’ 5
Sulfur dioxide emissions from any MWC will vary with time. Municipal
waste combustors equipped with a spray dryer system will exhibit a certain
amount of variability in SO 2 emissions due to random fluctuations in the
sulfur content of MSW and in spray dryer operating parameters such as outlet
gas temperature, reagent quality, and liquid and gas flow rates.
A- 18
-------
There are several measures which describe the variability
characteristics of SO 2 data:
o standard deviation and relative standard deviation,
o autocorrel ation,
o probability distribution (normal vs. lognormal),
o length of averaging period and averaging method.
Standard deviation is an indicator of the spread of values; it measures the
variation of measurements from the average. Relative standard deviation
(RSD) is defined as the ratio of the standard deviation to the average or
mean. The greater the standard deviation or RSD, the greater the variability
in observed SO 2 emissions or percent reduction. Sulfur dioxide emission
datasets which have the same mean SO 2 level but different SO 2 variability
levels (standard deviations or RSDs) will have different maximum expected
emissions.
Sequential measurements taken over a period of time are not necessarily
independent observations. When emissions data are collected in sequence,
there is a tendency for observations taken close together in time to be more
alike than those taken farther apart. This association between observations
in a time series is termed “autocorrelation”. Autocorrelation can vary from
-1.0 for inversely related observations to ÷1.0 for data which exhibit an
extreme degree of positive association. Overall, the higher the
autocorrelation factor, the higher the maximum expected emission level. The
effect of autocorrelation on maximum expected emission values is especially
important for longer averaging periods. However, autocorrelation is not as
major a factor as RSD in determining emissions variability.
A first order autoregressive time series model, abbreviated as AR(1), is
often used to project maximum expected SO 2 levels. The model assumes that an
SO 2 emissions value at time t (Yr) is related to the SO 2 emissions value
measured for the previous period according to the equation:
= 1 vt-i + e
where et is random variation with an expected mean value of 0 and a variance
assumed equal to the sample variance. The first order autocorrelation
coefficient (p 1 ) is the estimate of the covariances of the emissions values
A- 19
-------
based on a lag of one time periu . This is the most basic autoregressive
model and has been found in many cases to fit SO 2 emission and reduction time
series measurements adequately. A discussion of the time series technique
applied to data variability analysis is contained in References 4 and 5.
Observed SO 2 emissions data can be described by various probability
distributions. Emissions data have generally been found to be well
represented by either the normal or lognormal distribution. The primary
difference in the two distributions is that the lognormal distribution (in
which the natural logarithms of the data are normally distributed) is
positively skewed; that is, there is an unusually large number of
oh ervations less than the mean.
The length of the averaging period also affects the variability of SO 2
emissions. The averaging period is defined as the period of time (hours or
days) over which emission measurements are averaged. Longer averaging
periods dampen the effects of variability and autocorrelation, resulting in
lower maximum expected emission projections. A block average or a rolling
average may be used to calculate average emission values. Block averages are
computed on a separate, non-overlapping basis. Rolling averages are
calculated by adding the most recent value to the data set and dropping the
oldest value.
The final factor affecting maximum expected SO 2 emission values is the
exceedance frequency. The exceedance frequency is expressed as the number of
times, or percentage of time, that a given SO 2 emission value will be
exceeded over a specified time period. The probability of exceeding a given
emission value can be calculated for a specific averaging method and
exceedance frequency. As an example, consider an 8-hour block averaging
period and a one in one year exceedance frequency. There is a potential of
having three exceedances each day using 8-hour block averages, or a total of
1095 potential exceedances over each year. Thus, one exceedance among the
total potential exceedances results in an exceedance probability of 1/1095 or
0.00091. This exceedance probability, in turn, translates to a standard
normal variate (or Z value) of 3.33. To estimate the maximum expected SO 2
emissions value (in this case, an 8-hour average value) under these
conditions, an emission level is selected for which, in light of the sample
A-20
-------
distribution, standard deviation, ard autocorrelation structure, there will
be a 0.00091 probability in any of the 8-hour block average periods that SO 2
emissions will exceed the selected value. Over an extended period of time,
the number of exceedances of the selected value will average one per year.
The same methodology described above for SO 2 emissions variability
applies to SO 2 removal efficiency variability. The only difference is that
investigators are usually concerned with minimum expected percent SO 2 removal
efficiency rather than maximum expected SO 2 emission levels.
A.5.2 Sulfur Dioxide Emissions Variability
The emission data collected at Milibury consist of hourly measurements
taken over 62 consecutive days. As discussed above, time series analysis is
an integral part of the variability assessment. To evaluate the structure of
a time series, a set of 50 or more consecutive observations with few or no
interruptions in the data is required. Three sets of SO 2 emissions data were
available from the Millbury testing program which satisfied these criteria:
o 15 July to 31 July 1988 (Julian days 197 to 213),
o 4 August to 18 August 1988 (Julian days 217-231) and,
o 1 September to 14 September 1988 (Julian days 245-258).
The first step taken in evaluating the structure of the SO 2 emissions
data was to develop a relative frequency plot of the data, as shown in Figure
A-9. The fact that the sample probability distribution is skewed to the
right suggests that the data fit a lognormal distribution, as opposed to a
normal distribution. Plotting the logarithm of the emissions data on
probability paper, as shown in Figure A-b, confirms that the lognormal
distribution provides a close approximation of the sample distribution
(perfect agreement would result in all the data falling on a straight line).
After selecting the sample probability distribution, the next step was
to calculate the mean, standard deviation, and RSD for each of the three SO 2
emission data sets described above. These statistics are summarized in
Table A-2 using the natural logarithm of the emissions data. The overall
mean of the logarithms of the SO 2 emissions, 3.46, corresponds to a mean $02
emissions value of 32 ppm.
A- 21
-------
.i ‘ i,I i - i r •i r r .
IVI ‘ ‘ L. lvi L L U r I ‘... r-..-- I L .’.•— . I
OUTLD ¶02 s f ELATE flt IJEI4. I
5’:’2 MI i IQt -L ‘ c• oii•rr or r 4i4c; ::J
Figure A-9.
Relative frequency distribution of SO 2 emissions
at Milibury (hourly averages).
0’
yt
1;•
U
Lai
C)
LU
r
Ie
V •
Q,c 1
•1
4. ,
2.5 62,5 1.T2.c ,
-------
I I •$ SI SI •
Figure A-lU. Cumulative probability distribution of SO emissions
at Milibury (logarithms of hourly average ).
In
r’)
I
CWINUIATLV [ PRO AMILITY
-------
TABLE A-2. MILL JRY TEST VARIABILITY STATISTICS
SO , Emissions Data Ln (Hourly SO 2 Emissions)a
First Order
Relative Auto-
Data Period Mean Standard Standard correlation
Set (Julian Days) Deviation Deviation Coefficient
1 197 - 213 3.34 0.59 0.18 0.66
2 217 - 231 3.46 0.65 0.19 0.65
3 245 - 258 3.61 0.49 0.14 0.69
Overall 197 - 258 3.46 0.59 0.17
SO Removal Efficiency Data In (Hourly SO 2 Emissivity)b
Relative Auto-
Mean Standard Standard correlation
Deviation Deviation Coefficient
4 251 - 258 3.11 0.25 0.08 0.63
Natural logarithm of hourly SO 2 emissions data
Natural logarithm of hourly SO , emissivity data; SO 2 emissivity (%) =
100 - SO 2 removal efficiency (t)
A-24
-------
To estimate the autocorrel tion, the log-transformed SO 2 emission values
were fit to a first-order autoregressive time series model, or AR(1). Each
value of the time series is modeled as a regression function of the previous
hour’s value plus a random component. An AR(2) model, which uses the
emission values for the two previous hours was also evaluated. However, a
test of the significance of the second lag value in the AR(2) model showed
that this coefficient is not significantly different from zero at the 95
percent confidence level. Therefore, the use of the AR(1) model is
appropriate for these data.
The AR(1) model generated hourly autocorrelation coefficients between
0.65 and 0.69 for the three subject data sets, as shown in Table A-2. An
overall autocorrelation coefficient cannot be calculated due to time gaps
between the data sets. However, the autocorrelation results from the three
data sets are in very close agreement. Thus, a “best estimate” of the SO 2
emissions hourly autocorrelation coefficient for the entire testing period is
0.67.
The results of the variability analysis are used to estimate the impact
of averaging times and exceedance frequences on maximum expected SO 2 emission
values using the methodology described in Reference 4. In estimating maximum
expected values, the mean and standard deviation for the overall data set
were used in conjunction with the “best estimate” autocorrelation coefficient
of 0.67. The results are summarized in Table A-3. As an example, using a
24-hour block averaging period, the analysis projects that SO 2 emissions from
the Millbury facility will exceed 76 ppm an average of one time in ten years.
This compares with mean SO 2 emissions of 32 ppm. The analysis assumes, of
course, that the facility continues to operate in the same manner as that
observed during the emissions test period.
Inspection of Table A-3 shows that the difference between the maximum
expected SO 2 emission value and the mean SO 2 emission level increases as the
averaging period decreases and as the exceedance frequency decreases. The
first result is due to the higher variability associated with short averaging
times. The second result is due to the shape of the sample probability
distribution, with most emission values clustered near the mean and fewer
emission values distant from the mean.
A-25
-------
TABLE A-3. MAXIMUM IXPECTED SO EMISSIONS AND MINIMUM
EXPECTED SOS, REMOVAL EFFI IENCIES BASED ON
LMILLBURY TEST DATA
Mean SO emissions =
Mean SO 2 remo al efficiency
32 ppm
= 78 percent
Maximum
Minimum
Exceedance Frequency!
Averaging Period
Expected
SO Emissions
( pm SO 2 )
Expected
SO Removal
(p rcent)
One in 10 Years Exceedance Freauencv
391
35
1-Hour Block Average
3-Hour Block Average
233
49
8-Hour Block Average
137
60
24-hour Block Average
76
69
7-Day Rolling Average
45
74
30-Day Rolling Average
38
76
One in 1 year Exceedance Frequency
268
45
1-Hour Block Average
3-Hour Block Average
173
55
8-Hour Block Average
109
64
24-hour Block Average
65
71
7-Day Rolling Average
42
75
30-Day Rolling Average
36
76
One Percent Exceedance Frequency
126
60
1-Hour Block Average
3-Hour Block Average
103
64
8-Hour Block Average
80
68
24-Hour Block Average
58
72
7-Day Rolling Average
40
75
30-Day Rolling Average
36
77
A-26
-------
A.5.3 Sulfur Dioxide Removal Efficiency Variability
To calculate SO 2 removal efficiency, both spray dryer inlet and outlet
SO 2 (as well as 02) measurements are required. Due to gaps in available CEM
data for inlet and outlet SO 2 measurements during early phases of the
Milibury emission test, and the requirement for 50 or more hourly
measurements with few or no interruptions, analysis of removal efficiency
variability with the Milibury data was restricted to the last eight days (a
total of 192 observations) of the test period. During this final period,
four instances of missing °2 removal efficiency data were encountered in
which one or two hourly values were unavailable due to CEM sampling system
malfunctions. These missing values were replaced with the mean; this is a
standard procedure in time series analysis and does not materially affect
autocorrelation coefficient estimation.
As with SO 2 emissions data, the first step in the SO 2 removal efficiency
variability analysis was to select an appropriate probability distribution.
With percent removal data, past experience has shown that the data are easier
to manipulate and evaluate when expressed as SO 2 emissivity. SO 2 emissivity
is a measure of the percentage of the inlet SO 2 concentration which is
emitted by the control device and is given by the expression:
SO 2 emissivity (%) = 100 - SO 2 removal efficiency (%)
The relative frequency plot of the SO 2 emissivity data from the last
eight days is shown in Figure A-li. As with the SO 2 emissions data, the
skewed distribution suggests a lognormal sample probability distribution.
Figure A-12 shows the plot of the natural logarithms of SO 2 emissivity data
on probability paper. The straight line approximation in this figure
supports the selection of the lognormal distribution as a close approximation
of the sample distribution.
Sample statistics for the log-transformed SO 2 emissivity data are also
shown in Table A-2. The logarithmic sample mean of 3.11 for SO 2 emissivity
corresponds to 77.6 percent SO 2 removal efficiency.
As with SO 2 emissions data, an AR(i) model was used to estimate the
autocorrelation coefficient of the log-transformed SO 2 emissivity data. The
AR(2) model was also evaluated, but was rejected due to insignificance of the
A- 27
-------
MWC M I LLB U R C E A DATA
DAI5SMTY V5 1k LATIV( P1 QU!NCY
.T 4 -
0.32 -
/
/
0.28 -
0.26 -
I-.
Q.24
Q.22
Q.2
‘Li
/7
0.18- // /
‘Li
0. 16 // //
Q .14
0.12-
‘Li
0.1
o.oe -
,1 ___
0.04 - ‘1/ 7/ ___
/ /1
0.02 - ,/ / ___
_____________ ______ r- -__i
1 1 1 i T i T
2.5 7, 12.5 17,5 22.5 27.5 2.5 7.5 42.5 47.5 52.5 57.5 62.5 t57..5 72.5 77.5 82.5
502 D 4I WITi (MIOPOtIT OF 4N 3 )
Figure A-Il. Relative frequency distribution of SO 2 emissivity
at Millbury (hourly averages).
-------
j I •$ •) IS • •a
Figure A-12. Cumulative probability distribution of SO emissivity
at Milibury (logarithms of hourly average ).
4.0
‘-S
Ill
S n
r..)
0
2.0
Sn
z
-j
CUPI4UIATIV( PRUBAB1LITY (%)
-------
second lag coefficient. The AR(1) model, considered the best model for
these data, estimated an autocorrelation coefficient of 0.53 for the hourly
emissivity data.
This autocorrelation coefficient, in conjunction with the mean and
standard deviation statistics, was used to estimate the impact of averaging
times and exceedance frequencies on minimum expected SO 2 removal efficiency
in the same manner as described above for maximum SO 2 emission values (see
Table A-3). Using the same example 24-hour block averaging period as
before, the analysis projects that the SO 2 removal efficiency of the Mill-
bury spray dryer/ESP will fall below 69 percent an average of only one time
in ten years. This compares with mean SO 2 percent removal of 78 percent.
The table shows that the same general trends apply to minimum expected
SO 2 removal levels (i.e., difference between projected minimum and the mean)
resulting from changes in averaging times and exceedance frequencies as
discussed above for SO 2 emissions. The factors underlying these trends are
also the same.
REFERENCES
1. Entropy Environmentalists, Inc. Emission Test Report, Municipal Waste
Combustion Continuous Emission Monitoring Program, Wheelabrator Resource
Recovery Facility, Milibury, Massachusetts. Prepared for U.S.
Environmental Protection Agency, Research Traingle Park, NC. January
1989. Emission Test Report 88-MIN-07C.
2. Anderson, C. 1., Vancil N. A., Mayhew, J. W., and Holder, 0. J. (Radian
Corporation). Characterization Emission Test Report, Marion County
Solid Waste to Energy Facility. Prepared for U.S. Environmental
Protection Agency. September 1988. EMS Report No. 87-MIN-04, Volume I.
pp. 4-1 to 4-3.
3. U.S. Environmental Protection Agency. Municipal Waste Combustion Study:
Flue Gas Cleaning Technology. June 1987. EPA/530-SW-87-021d. p. 4-4.
4. Radian Corporation. Determination of Mean SO , Emission Levels Required
to Meet A 1.2 lb SO ,/Million Btu Emission Stahdard for Various Averaging
Times and Compliance Policies. Prepared for the U.S. Environmental
Protection Agency. EPA Contract No. 68-02-3816. Research Trinagle
Park, North Carolina. March 1985.
5. U.S. Environmental Protection Agency. Statistical Analysis of Emission
Test Data from Fluidized Bed Combustion Boilers at Prince Edward Island,
Canada. Publication No. EPA-450/3-86-015. December 1986.
A-30
-------
:; ..
e read fr i . t7 S IcnS c rn r)j fl
1. REPORT NO.
2.
EPA-45O/3-8 -27c
4. TITLE AND SUBTiTLE - T EPORTDA E
Municipal Waste Cornbustors -. Backgr ound In ormat on for ‘\ugust_1989
Proposed Standards: Post-Combustion Tcohnoiogj . PERFJ ING ORGANIZATION CODE
Performance
8. PERFORMING ORGANIZATION REPORT NO
NAME AND ADDRESS I C. PROGRAM ELEMENT NO. —
Pianni;ig and Standarc
Protection A er.:y 11.CONTRACT GRANT NO
North Caroiina 27 ll
68-02-4378
AND ADDRL .S - 13. TYPE OF REPORT AND FERIOD COVERED
Planning and Standards Final
Radiation 14.SPONSORING AGENCY CODE
Protection Agen
Nort Caro 1 i a 27711 200/04
—
evaluates the Lerformance of various air pollution control
new and ex stinc municio l waste combustors (MWC 1 s). The
include electrostatic piecipitators (ESP 1 s), furnace
systems with ESP’s, moderate— and low—temperature duct sorbent
ESPs, or fabric filters (FPs) and spray dryers with
removal capabilities for each of these control devices are
matter, metals (arsenic, cadmium, chromium, lead,
chlorinated dibenzo—p—dioxins and dibenzofurans, and
dioxide and hydrogen chloride.
data for each of the control devices listed, as applied to
The key process parameters affecting control device
discussed. Performance is correlated with these parameters to
achievable removal efficiencies for each pollutant.
KEY WORDS AND DOCUMENT ANALYSIS
.
b.IDENTIFIERSIOPEN ENDED TERMS C. COSATI Field Group
Air Pollution Control 13B
Combustors
—
19. SECURITY CLASS 1 ThisReportl —
Unclassified .
20 SECURITY CLASS Tins page)
I Unclassified
21. NO. OF PAGES
327
22. PRICE
•
3 RECRIENT’S . C
EPA Form 2220—1 (Rev. 4—77)
PREVIOUS EOITION IS OBSOLETE
------- |