United States
Environmental
Protection Agency
&EPA
Office of Solid Waste
and Emergency Response
Washington, DC 20460
Office of Air
and Radiation
Washington, DC 20460
Off ice of Research
and Development
Washington, DC 20460
EPAOOO-SW-00-000
June 1987
Municipal Waste
Combustion Study
Combustion Control of
Organic Emissions
DRAFT
-------
MUNICIPAL WASTE COMBUSTION STUDY:
COMBUSTION CONTROL OF MSW COMBUSTORS TO
MINIMIZE EMISSION OF TRACE ORGANICS
Final Report
by
W. R. Seeker, W. S. Lamer and M. P. Heap
Energy and Environmental Research Corporation
18 Mason
Irvine, CA 92718-4190
EPA Contract 68-02-4247
Project Officer: James Kilgroe
Combustion Research Branch
Air and Energy Engineering Research
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Prepared for:
U.S. Environmental Protection Agency
Office of Research and Development
Washington, D.C. 20460
May 1987
-------
This document has been approved for publication by the Office of Research and
Development, U.S. Environmental Protection Agency. Approval does not signify
that the contents necessarily reflect the views and policies of the
Environmental Protection Agency, nor does the mention of trade names or
commercial products constitute endorsement or recommendation for use.
n
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the significant contribution of a number
of individuals and organizations. Mr- Walter Niessen of Camp, Dresser,
McKee, Inc. actively participated in data gathering and aided significantly
to the ideas for combustion control strategy. We also wish to thank the
representatives from the manufacturers with which we had detailed
discussions. Without their willingness to share data and discuss combustion
control strategy, this study would not have been possible. Finally, the
authors want to acknowledge Jim Kilgroe (EPA-ORD) and Steve Greene (EPA-OSW)
both for their financial support and for their technical guidance.
PREFACE
This document was prepared by Energy and Environmental Research
Corporation as a task report under EPA's Fundamental Combustion Research
Program III (Contract Number 68-02-4247). The FCR III program contract
Project Officer was Mr. Jim Mulholland while Mr. James A. Kilgroe has served
as Task Officer-
The work described in this document is part of a comprehensive Municipal
Waste Combustion Study being carried out by EPA. This comprehensive
evaluation is the result of a combined effort involving EPA's Office of
Research and Development, Office of Solid Waste and Emergency Response and
Office of Air and Radiation.
n i
-------
TABLE OF CONTENTS
Section Page
ACKNOWLEDGEMENTS AND PREFACE ii
1.0 EXECUTIVE SUMMARY 1-1
1.1 Scope of the Study 1-1
1.2 Combustion Control of Trace Organic Emissions 1-2
1.3 A Combustion Control Strategy 1-4
1-4 Recommended Research 1-6
2.0 INTRODUCTION 2-1
2.1 Background 2-1
2.2 Program Overview 2-5
2.3 References 2-8
3.0 MUNICIPAL WASTE COMBUSTION PROCESSES . . 3-1
3.1 Municipal Waste Characteristics 3-2
3.2 Combustion Control of Pollutant Emissions from
Municipal Waste Combustion Facilities 3-5
3.2.1 Basic Combustion Concepts 3-6
3.2.2 Criteria Pollutants 3-12
3.2.3 Emission of Organic Compounds 3-12
3.3 Combustion Control of Starved Air Systems 3-17
3.4 RDF Systems 3-19
3.5 References 3-20
4.0 POTENTIAL FOR AIR EMISSIONS 4-1
4.1 Organic Emissions 4-1
4.2 NOX Formation and Control 4-8
4.3 Particulate and Trace Metals 4-14
4.4 Acid Gases 4-18
4.5 References 4-18
IV
-------
TABLE OF CONTENTS (CONTINUED)
Section Page
5.0 CURRENT PRACTICES IN MASS BURN TECHNOLOGY 5-1
5.1 Mass Burn Technologies 5-1
5.2 Deutsche Babcock Anlagen 5-5
5.3 Steinmueller 5-10
5.4 Von Roll 5-17
5.5 W+E Environmental Systems Ltd 5-23
5.6 Martin GmbH 5-27
5.7 Volund 5-34
5.8 Riley Takuma 5-37
5.9 Detroit Stoker 5-42
5.10 Combustion Engineering - De Bartolmeis 5-46
5.11 Westinghouse O'Connor 5-50
5.12 Basic Environmental Engineering Inc 5-54
5.13 Enercon/Vicon 5-58
5.14 References 5-61
6.0 CURRENT PRACTICES IN RDF COMBUSTORS 6-1
6.1 Types of RDF 6-1
6.1.1 RDF-1 or MSW 6-1
6.1.2 RDF-2 or Coarse RDF (c-RDF) 6-3
6.1.3 RDF-3 or Fluff RDF (f-RDF) 6-7
6.1.4 RDF-4 or Powdered RDF (p-RDF) 6-11
6.1.5 RDF-5 or Densified RDF (d-RDF) 6-11
6.2 Current RDF Projects - 6-12
6.3 Firing Systems in Current RDF Projects 6-12
6.3.1 Detroit Stoker RDF Firing Systems 6-12
6.3.2 Combustion Engineering RDF Firing System 6-19
6.4 References 6-22
-------
TABLE OF CONTENTS (CONTINUED)
Section Page
7.0 CURRENT PRACTICES IN STARVED AIR (TWO-STAGE) COMBUSTORS 7-1
7.1 Starved Air Technologies 7-1
7.2 Consumat Systems 7-4
7.3 Synergy 7-10
7.4 References 7-12
8.0 COMBUSTION CONTROL OF ORGANIC EMISSIONS FROM MUNICIPAL WASTE
COMBUSTORS (MWCs) 8-1
8.1 Design and Operating Problems - Failure Modes 8-1
8.1.1 Mass Burn Waterwall Failure Modes 8-2
8.1.2 Refuse Derived Fuel Spreader Stoker Failure Modes. . 8-8
8.1.3 Small, Multi-Staged Modular Unit Failure Modes . . . 8-11
8.2 Combustion Strategy for Minimizing Emissions of Air
Pollutants 8-13
8.3 Temperature 8-15
8.4 Combustion Air 8-21
8.4.1 Total Air Requirements 8-21
8.4.2 Primary Air Requirements 8-23
8.4.3 Distribution of Overfire or Secondary Air 8-24
8.4.4 Verification of Appropriate Air Distribution .... 8-30
8.5 Combustion Monitoring Requirements 8-32
8.6 System Control 8-34
8.7 Minimization of Hydrocarbon Species and Other Pollutants. . 8-35
8.8 References 8-37
9.0 HYDROCARBON CONTROL STRATEGY - SUMMARY 9-1
9.1 Combustion Practices for Trace Hydrocarbon Emission
Control of municipal waste combustors 9-2
9.2 Research Recommendations 9-7
9.2.1 Combustion Control Guideline Definition and
Verification 9-8
9.2.2 Mechanisms of PCDD/PCDF Formation and Destruction. . 9-11
9.2.3 Tradeoffs in Other Pollutants 9-11
vi
-------
LIST OF FIGURES
Figure Page
2-1 Program approach 2-6
3-1 The LNC Steinmueller mass burn design features 3-7
3-2 Estimate of fuel bed thickness and accumulated inert layer
(Run 18) 3-9
3-3 Gas compositions from fuel bed probe 2 (Run 18) 3-10
3-4 Relationships of CO and 02 for appropriate operating regions. . 3-13
3-5 Adiabatic equilibrium species distribution 3-16
3-6 Components of typical starved-air modular combustor 3-18
4-1 Hypothetical mechanisms of PCDD/PCDF formation chemistry. . . . 4-3
4-2 Thermal decomposition characteristics of selected organics
under dilute (nonflame) conditions for 1 second
(Bellinger et al. 1977) 4-6
4-3 Impact of temperature and fuel nitrogen on NOX emissions for
excess air conditions (calculated using EER kinetic set). . . 4-10
4-4 Application of reburning and de-NOx schemes for NOX control
of mass burn MSW furnaces 4-13
4-5 Transformation of mineral matter during combustion of metal
containing waste 4-17
5-1 Mass burning waste power plant at Widmer and Ernst
at Bielefeld-Hertford, Germany 5-2
5-2 Deutsche Babcock Anlagen mass burn furnace design features. . . 5-6
5-3 Deutsche Babcock furnace geometry selected based on refuse
characteristics (DBA, 1986) 5-9
5-4 Relationship of CO and 02 for appropriate operating regions
(DBA, 1986) 5-11
5-5 The l&C Steinmueller mass burn design features 5-13
5-6 Steinmueller in-furnace testing and cold-flow modeling 5-16
5-7 Refuse combustion plant with von Roll two-pass boiler and
flue gas scrubber 5-19
5-8 Details of Von Roll grate and feeding devices 5-22
-------
LIST OF FIGURES (CONTINUED)
Figure Page
5-9 W+E combustion system and overthrust grate design 5-24
5-10 Design features of Martin Refuse Combustors 5-29
5-11 Adaptation of secondary air injection to changed refuse
conditions in Martin Systems 5-32
5-12 Martin Combustion Control System for mass burn combustors . . . 5-33
5-13 Volund System mass burn design features 5-35
5-14 Cross sectional schematic of combustion zone on a
Riley-Takuma mass burn plant 5-39
5-15 Cross-section of boiler with current design Detroit Stoker
firing system 5-44
5-16 Design features of DeBartolomeis grate 5-48
5-17 Boiler cross-section of CE design using db grate and
EVT boiler 5-49
5-18 Typical plant configuration using Westinghouse/O'Connor
combustor 5-51
5-19 Cross-section of Westinghouse/O'Connor combustor 5-52
5-20 Process flow diagram of Basic Environmental Engineering
modular mass burn technology 5-55
6-1 Albany, New York, Solid-Waste Energy-Recovery System
(ANSWERS) 6-4
6-2 Cross-section of Albany, New York boiler plant firing c-RDF . . 6-5
6-3 Ames, Iowa, Resource Recovery System for production of
fluff RDF (f-RDF) 6-8
6-4 Fluff RDF production system. Illustration is for system
developed by Combustion Engineering 6-9
6-5 Trommel Screen for RDF size segregation 6-10
6-6 Side sectional view of Detroit Rotograte Stoker equipped
with Detroit air swept refuse distributor spouts 6-15
6-7 Detroit Stoker Hydrograte^M with water cooled, vibrating,
continuous-discharge ash-discharge grates 6-17
6-8 Detroit air swept refuse fuel distributor spout arranged
with motorized rotary air damper 6-18
6-9 Combustion Engineering continuous ash discharge type RC
Stoker for RDF 6-20
6-10 Combustion Engineering pneumatic RDF distributor 6-21
vii'i
-------
LIST OF FIGURES (CONTINUED)
Figure Page
7-1 Evolution of U.S. vendor companies supplying starved-air
waste-to-energy systems, including persons notably
influencing system designs 7-3
7-2 The standard Consumat module for energy-from-waste 7-5
7-3 Internal transfer rams in primary chamber of typical
Consumat facility 7-6
7-4 Theoretical temperature of the products of combustion,
calculated from typical MSW properties, as a function of
refuse moisture and excess air or oxygen (Hasselriis, 1986) . 7-7
7-5 Synergy two-stage municipal waste combustion process 7-11
7-6 Synergy two-stage combustion saturate steam system 7-13
8-1 Mass burn municipal waste combustor failure modes 8-3
8-2 Boiler sectional side of NASA/Hampton mass fired waste-
to-energy facility 8-7
8-3 Failure modes of RDF spreader-stoker systems 8-9
8-4 Temperature distributions in an operating mass burn
municipal waste combustor 8-16
8-5 Required temperature for destruction of intermediate
organics 8-18
8-6 Thermal decomposition characteristics of selected
hydrocarbons 8-20
8-7 Theoretical temperature of the products of combustion,
calculated from typical MSW properties, as a function of
of refuse moisture and excess air or oxygen (8,9) 8-22
8-8 Typical designs of overfire air systems 8-27
9-1 Research program to establish design guidelines based upon
"Good Combustion Practice" 9-9
-------
LIST OF TABLES
Table
Page
1-1 Good Combustion Practices for Minimizing Trace Organic
Emissions From Mass Burn Municipal Waste Combustors (MWCs). . . 1-6
1-2 Good Combustion Practices for Minimizing Trace Organic
Emissions from RDF Combustors 1-7
1-3 Good Combustion Practices for Minimizing Trace Organics
Emissions from Starved-Air Municipal Waste Combustors (MWCs). . 1-8
2-1 Summary of Organic Emission Ranges from Full-Scale Municipal
Waste Combustor Facilities 2-4
2-2 Manufacturers Fact Finding 2-7
3-1 Current and Forecast Composition of Disposed Residential and
Commercial Solid Waste (Weight Percent) 3-3
3-2 Ultimate Analysis of Typical MSW as Presented by Hickman et
al. (1984) 3-4
4-1 Status of NOX Control Options for Municipal Waste Combustors
(MWCs) 4-12
4-2 Metals Present in MSW 4-16
5-1 Design Features of Deutsche Babcock Anlagen Systems 5-7
5-2 Design Features of Steinmueller Mass Burn Systems 5-14
5-3 Design Features of Von Roll Mass Burn Systems 5-20
5-4 Design Features of W+E Mass Fired System 5-25
5-5 Design Features of Martin Refuse Municipal Waste Combustors
(MWCs) 5-30
5-6 Design Features of Volund Mass Burn Systems . 5-36
5-7 Design and Operating Features of Riley-Takuma Technology 5-40
5-8 Design Features of Detroit Stoker Mass Burn Municipal Waste
Combustors 5-45
5-9 Design Features of Basic Environmental Engineering Small
Modular Mass Burn Technologies 5-56
6-1 ASTM Classification of Refuse Derived Fuels 6-2
6-2 Active RDF Projects 6-13
x
-------
LIST OF TABLES (CONTINUED)
Table Page
7-1 Selected Data on Small-Scale U.S. Systems Using the
Starved-Air Design 7-2
8-1 Overfire Air Jet Configurations for Large Mass Burn
Municipal Waste Combustors (MWCs) 8-26
9-1 Good Combustion Practices for Minimizing Trace Organic
Emissions from Mass Burn Municipal Waste Combustors (MWCs). . . 9-4
9-2 Good Combustion Practices for Minimizing Trace Organic
Emissions from RDF Municipal Waste Combustors (MWCs) 9-5
9-3 Good Combustion Practices for Minimizing Trace Organic
Emissions from Starved-Air Municipal Waste Combustors (MWCs). . 9-6
-------
1.0
EXECUTIVE SUMMARY
1.1
Scope of the Study
This report is an assessment of combustion control of organic emissions
from municipal waste combustors (MWCs). The information presented in this
report was developed during a comprehensive, integrated study of municipal
waste combustion. An overview of the findings of this study may be found in
the Report to Congress on Municipal Waste Combustion. Other technical
volumes issued as part of the Municipal Waste Combustion Study include:
• Municipal Waste Combustion Study:
Report to Congress
EPA/530-SW-87-021A
• Municipal Waste Combustion Study:
Emissions Data Base for Municipal
Waste Combustors
EPA/530-SW-87-021B
Municipal Waste Combustion Study:
Combustion Control of Organic Emissions EPA/530-SW-87-021C
Municipal Waste Combustion Study:
Flue Gas Cleaning Technology
EPA/530-SW-87-021D
• Municipal Waste Combustion Study:
Costs of Flue Gas Cleaning Technologies EPA/530-SW-87-021E
Municipal Waste Combustion Study:
Sampling and Analysis
EPA/530-SW-87-021F
Municipal Waste Combustion Study:
Assessment of Health Risks Associated
with Exposure to Municipal Waste
Combustion Emissions
EPA/530-SW-87-021G
1-1
-------
• Municipal Waste Combustion Study:
Characterization of the Municipal
Waste Combustion Industry EPA/530-SW-87-021H
• Municipal Waste Combustion Study:
Recycling of Solid Waste EPA/530-SW-87-021I
The objectives of this study were:
1. To determine the current state of combustion control of municipal
solid waste combustion technology
4
2. To formulate a combustion control strategy based upon "best
engineering practice" that will minimize the emission of trace
organics from waste-to-energy plants
3. To define the research which is necessary to develop and verify
this combustion control strategy.
Although the focus of this study was concerned with the best combustion
practices which will minimize the emissions of organics, including
polych 1 orinated dibenzo(pjdioxin and furans (PCDDs/PCDFs), the inter-
relationship with other pollutants such as particulate matter, metals, NOX.
other organics, and carbon monoxide was also considered. The study focused
on the design of new units and the operation and monitoring of new and
existing units from the viewpoint of the combustor/boiler subsystem. No
consideration was given to the impact of the design and operation of air
pollution control devices upon the emission of trace organic species because
it was covered in the volume entitled "Flue Gas Cleaning Technology"
(EPA/530-SW-87-021E).
1.2 Combustion Control of Trace Organic Emissions
Combustion control of organic species from three types of municipal
solid waste combustors, was considered in this study:
1-2
-------
1) Mass burning excess air (mass burning)
2) Starved air two-stage (modular)
3) Refuse-derived fuel firing (RDF)
RDF can be burned in fluidized bed combustors but they represent only a small
fraction of the waste-to-energy systems either in service or planned. Thus,
this technology was not considered.
Municipal waste, unlike fossil fuels, is extremely heterogeneous. Its
composition varies with the seasons and is affected by weather thus, its
calorific value can vary widely and many systems are therefore designed to
handle wastes with heating values ranging from 3,800 to 6,000 Btu/lb. The
size of MSW also varies and this affects the feed to the combustion device.
The three incinerator types listed above have different design philosophies
which enable them to take account of variation in fuel input properties which
is imposed by the characteristics of MSW.
Mass burn excess air municipal waste combustors (MWCs) burn minimally
treated MSW in a thick bed supported by a pusher grate. Sufficient air is
provided in the bed region to burn the fuel, although combustion of the
volatile gases is completed above the bed. Variation in fuel properties is
counteracted by controlling the feed rate, grate speed and the amount and
distribution of air which is supplied through and above the grate. In a
starved-air two stage system the unit is divided into two distinct sections.
The first section receives the waste and is operated with only 40 percent of
the air needed for combustion, thus, it acts as a gasifier. The fuel rich
effluent is burned in the second section which may contain heat exchange
surfaces. Heat extraction is minimized in the first, fuel rich section.
Pre-processing to produce a refuse derived fuel (RDF) is the third method
used to take account of the heterogeneous nature of MSW. RDF can be burned
alone, on a spreader stoker, in a fluidized bed or in combination with coal
in stoker, pulverized coal or cyclone boilers or in combination with other
fuels such as wood chips.
1-3
-------
Thermodynamic equilibrium considerations indicate-that under excess air
conditions and the characteristic temperatures typical of municipal waste
combustors (MWCs), emissions of organic species should be so low that they
can be considered to be zero. However, measurements have been made showing
that some plants have significant emissions of trace organic species some of
which are toxic. The basis for combustion control of emissions of organics
from municipal waste combustors (MWCs) is to provide conditions whereby the
combustion products can approach equilibrium. This requires that all
combustion products are affectively mixed with oxygen at a temperature which
is sufficiently high to allow the rapid destruction of all organic species.
Hydrocarbons, some of which may be toxic or which may be precursors to
the formation of toxic hydrocarbons, can be formed during the combustion of
MSW. The fuel is not completely mixed with air because of the heterogeneous
nature of the fuel. Thus, fuel-rich pockets will exist and under these
conditions hydrocarbons can be formed. However, kinetic considerations
indicate that they can be destroyed rapidly in the presence of oxygen at
elevated temperatures. The goal of combustion control is the complete
destruction of all hydrocarbon species in the combustion system of municipal
waste combustors (MWCs). Approaching this goal will minimize emission of
potentially toxic species as well as other species which may be precursors
and capable of forming toxic compounds downstream in cooler regions of the
boiler or the air pollution control devices.
1.3 A Combustion Control Strategy
Strategies for good combustion can minimize the emission of hydrocarbon
pollutants from the combustor/boiler subsystem of a waste-to-energy system.
These practices will also contribute to the operability of the plant because
they will reduce furnace corrosion rates and improve efficiency levels.
Thus, while it may not be possible to entirely eliminate trace organic
emissions it should be possible to operate cost-effective, waste-to-energy
systems with minimal emissions of potentially toxic organics (fractional part
per trillion emission concentration range).
-------
This study synthesized "best combustion practices" from: theories on
the basic mechanisms of PCDD/PCDF formation and destruction in combustion
systems; information provided by manufacturers on the design of waste-to-
energy systems, and operating/emissions data from specific plants. The
practices are designed to:
1. Limit the formation of hydrocarbons
2. Maximize the destruction of these same compounds prior to the exit
of the combustor/boiler should they be formed
As such, these practices restrict conditions which promote the formation of
hydrocarbons. In addition, they ensure that the environment experienced by
these hydrocarbons will promote the destruction of these compounds. Thus,
the conditions within the combustor environment that satisfy these goals are:
• Mixing of fuel and air to minimize the existence of long-lived,
fuel-rich pockets of combustion products
• Attainment of sufficiently high temperatures in the presence of
oxygen for the destruction of hydrocarbon species
• Prevention of quench zones or low temperature pathways that will
allow partially reacted fuel (solid or gaseous) from exiting the
combustion chamber.
The development of "best combustion practices" to minimize emissions of
trace organics from municipal waste combustors (MWCs) involves all of these
elements:
t Design. Design the system to satisfy several criteria which will
ensure that temperatures and the degree of mixing within the
combustor are consistent with the minimization of formation and the
maximization of destruction of the species of concern.
1-5
-------
• Operation Control . Operate the system in a manner which is
consistent with the design goal and provide facility controls which
prevent operation outside on established operating envelope.
t Aerification. Monitor to ensure that the system is continually
operated in accordance with the design goals.
It is suggested that if all of these elements are satisfied, then the
emission of hydrocarbons from the combustor of a municipal waste combustor
(MWC) will be minimized.
This project has identified the components of each of these elements
which make up the best combustion practices and that are expected to be
important in the control of PCDDs and PCDFs trace organic emissions from
municipal waste combustors (MWCs). In addition, preliminary recommendations
have been made on the values of the individual components. Identification of
the elements was based upon the current design practices that have been shown
to restrict successfully trace organic emissions. The combustion control
components are summarized in Table 1-1 for mass burn systems. It must be
explained that the values presented in Tables 1-1, 1-2 and 1-3 are
preliminary targets and require considerable verification to ensure their
appropriateness. It is possible, for example, that the goals of the strategy
can be satisfied by focusing solely on the verification elements. In this
manner, an innovative scheme would be allowed provided that it could be
demonstrated that the system satisfies the goals of flue gas CO, excess
oxygen, furnace temperature and in-furnace CO concentration uniformity-
1.4 Recommended Research
The "best combustion practices" defined above for mass burn MSW, RDF and
starved-air modular combustors were derived from an analysis of the available
information which includes little direct evidence relating to the
appropriateness of the preliminary target values recommended in Tables 1-1,
1-2 and 1-3. Further work is needed to better define and verify these
recommendations. Further work is required in three major areas:
1.-6-
-------
TABLE 1-1. GOOD COMBUSTION PRACTICES FOR MINIMIZING TRACE ORGANIC
EMISSIONS FROM MASS BURN MUNICIPAL WASTE COMBUSTORS
Element
Design
Operation/
Control
Verification
Component
Temperature at fully mixed
height
Underfire air control
Overfire air capacity (not
an operating requirement)
Overfire air injector
design
Auxiliary fuel capacity
Excess air
Turndown restrictions
Start-up procedures
Use of auxiliary fuel
Oxygen in flue gas
CO in flue gas
Furnace temperature
Adequate air
distribution
Recommmendati ons
1800°F at fully mixed height
At least 4 separately adjustable
plenums. One each under the drying
and burnout zones and at least two
separately adjustable plenums under
the burning zone
40% of total air
That required for penetration and
coverage of furnace cross-section
That required to meet start-up
temperature and 1800°F criteria under
part-load operations
6-12% oxygen in flue gas (dry basis)
80-110% of design - lower limit may
be extended with verification tests
On auxiliary fuel to design
temperature
On prolonged high CO or low furnace
temperature
6-12% dry basis
50 ppm on 4 hour average - corrected
to 12% C02
Minimum of 1800°F (mean) at fully
mixed height across furnace
Verification Tests (see text Chapter
8 and 9)
1-7
-------
TABLE 1-2. GOOD COMBUSTION PRACTICES FOR MINIMIZING TRACE ORGANIC
EMISSIONS FROM RDF COMBUSTORS
Element
Design
Operation/
Control
Verification
Component
Temperature at fully mixed
height
Underfire air control
Overfire air capacity
(not necessary operation)
Overfire air injector
design
Auxiliary fuel capacity
Excess air
Turndown restrictions
Start-up procedures
Use of auxiliary fuel
Oxygen in flue gas
CO in flue gas
Furnace temperature
Adequate air
distribution
Recommmendations
1800°F at fully mixed height
As required to provide uniform bed
burning stoichiometry (see text)
40% of total air
That required for penetration and
coverage of furnace cross-section
That required to meet start-up
temperature and ISOOpF criteria under
part-load operations
3-9% oxygen in flue gas (dry basis)
80-110X of design - lower limit may
be extended with verification tests
On-auxiliary fuel to design
temperature
On prolonged high CO or low furnace
temperature
3-9$ dry basis
50 ppm on 4 hour average - corrected
to 121 C02
Minimum of 1800°F (mean) at fully
mixed height
Verification Tests (see text Chapter
8 and 9)
1-8
-------
TABLE 1-3. GOOD COMBUSTION PRACTICES FOR MINIMIZING TRACE ORGANIC
EMISSIONS FROM STARVED-AIR COMBUSTORS
Element
Design
Operation/
Control
Verification
Component
Temperature at fully mixed
height
Overfire air capacity
Overfire air injector design
Auxiliary fuel capacity
Excess air
Turndown restrictions
Start-up procedures
Use of auxiliary fuel
Oxygen in flue gas
CO in flue gas
Furnace temperature
Adequate air distribution
Recommmendations
1800°F at fully mixed height
80 percent of total air
That required for penetration and
coverage of furnace cross-section
That required to meet start-up
temperature and 1800°F criteria
under part-load conditions
6-12% oxygen in flue gas (dry
basis)
80-110% of design - lower limit
may be extended with verification
tests
On auxiliary fuel to design
temperature
On prolonged high CO or low
furnace temperature
6-12% dry basis
50 ppm on 4 hour average -
corrected to 12% C02
Minimum of 1800°F at fully mixed
plane (in secondary chamber)
Verification Tests (see text Chapter
8 and 9)
1-9
-------
1. Guideline definition and verification
2. Mechanisms of PCDD/PCDF formation from MSW
3. Tradeoffs in other pollutants
1-'
-------
2.0
INTRODUCTION
2.1
Background
This report is an assessment of combustion control of organic emissions
from municipal waste combustors (MWCs). The information presented in this
report was developed during a comprehensive, integrated study of municipal
waste combustion. An overview of the findings of this study may be found in
the Report to Congress on Municipal Waste Combustion. Other technical
volumes issued as part of the Municipal Waste Combustion Study include:
• Municipal Waste Combustion Study:
Report to Congress
• Municipal Waste Combustion Study:
Emissions Data Base for Municipal
Waste Combustors
EPA/530-SW-87-021A
EPA/530-SW-87-021B
• Municipal Waste Combustion Study:
Combustion Control of Organic Emissions EPA/530-SW-87-021C
Municipal Waste Combustion Study:
Flue Gas Cleaning Technology
EPA/530-SW-87-021D
Municipal Waste Combustion Study:
Costs of Flue Gas Cleaning Technologies EPA/530-SW-87-021E
Municipal Waste Combustion Study:
Sampling and Analysis
EPA/530-SW-87-021F
Municipal Waste Combustion Study:
Assessment of Health Risks Associated
with Exposure to Municipal Waste
Combustion Emissions
EPA/530-SW-87-021G
2-1
-------
• Municipal Waste Combustion Study:
Characterization of the Municipal
Waste Combustion Industry EPA/530-SW-87-021H
• Municipal Waste Combustion Study:
Recycling of Solid Waste EPA/530-SW-87-021I
It is expected that the number of waste-to-energy plants within the
United States will dramatically increase in the next decade. Some estimates
indicate that the amount of waste being burned will increase by a factor of
three by 1990. "Characterization of the Municipal Waste Combustor Industry"
volume (EPA/530-SW-87-021H) provides a detailed analysis of current and
projected MSW-to-energy capacity in the U.S.
One major concern that stems from proliferation of waste-to-energy
plants is trace emissions of potentially toxic organic compounds. Of
particular interest are the cogeners of PCDD and PCDF species, with major
concern centering on 2,3,7,8 Tetrachlorodibenzo-p-dioxin (TCDD).
As part of EPA's comprehensive evaluation of municipal waste combustion,
an emissions data base has been established and documented in the volume
entitled "Emission Data Base for Municipal Waste Combustors." Table 2-1 has
been extracted from that data base to illustrate the wide variability in PCDD
and PCDF emissions from various facilities from which field test data is
available. As shown, the emission rate for total PCDDs and total PCDFs vary
by four to five orders of magnitude. Data forming the high end of
theseaission range come from field tests performed on the facility at
Hampton, Virginia (Haile, 1984); Scott Environmental Services, 1985; Nunn,
1983; Howes, 1982) and at Philadelphia, Northwest (NeuTicht, 1985). Both of
these facilities represent older system designs where the primary design
concern was waste volume reduction. Data forming the low end of the emission
range come from field tests performed on new facilities at Tulsa, Oklahoma
(Seelinger, 1986), at Marion County, Oregon (Ogden Projects, 1986) and at the
Westchester facility at Peekskill, New York (New York State Department of
Environmental Conservation, 1986). Each of these new facilities' designs is
2-2
-------
TABLE 2-1. SUMMARY OF ORGANIC EMISSION RANGES FROM
FULL SCALE MUNICIPAL WASTE COMBUSTION
FACILITIESa
Dr\ 1 T 1 1 +• a n-f-
rO 1 1 U uail u
2,3,7,8 TCDD, ng/Nm3
2,3,7,8 TCDF, ng/NM3
TCDD, ng/Nm3
TCDF, ng/Nm3
PCDD, ng/Nm3
PCDF, ng/Nm3
Emission
Mass Burn
0.018-62.5
0.168-448
0.195-1,160
0.322-4,560
1.13-10,700
0.423-14,800
Concentration F
Starved Air
<0. 278-1. 54
58. 5C
1.24-43.7
15.0-345
77.2-1,550
118-1,760
vangeb
RDF-Fired
0.522-14.6
2.69C
3.47-258
31.7-679
64.4-2,840
164-9,110
a Results from commercial-scale facilities only-
b All concentrations are reported in units corrected to 12 percent
C02.
c Data available for only one test.
2--3
-------
based on extensive research and development efforts by the manufacturer to
produce municipal waste combustors (MWCs) with high thermal efficiency and
low trace organic emission rates.
The broad range in emission rate data illustrated in Table 2-1 and the
success of modern, wel1-operated facilities in significantly reducing trace
organic emission rates strongly suggests that municipal waste combustor
design and operation are critical components in developing an overall organic
emission control strategy. More direct evidence on the importance of
facility design and operation is available through studies being conducted at
the municipal waste combustor facility in Quebec, Canada. In those studies,
an older facility was modified to reflect current low emission design
philosophy. Exhaust emission measurements were obtained before (1984 tests)
and after the facility modification (1986 tests). Preliminary results
(Finkelstein, 1986) are indicated below:
1984 Tests 1986 Tests
Total PCDD 800-3980 ng/Nm3 12-205 ng/Nm3
Total PCDF 100-1100 ng/Nm3 49.3-336 ng/Nm3
All of these data are corrected to 12 percent C02- In the modified facility
tests (1986 tests) the indicated data range reflects operation of the
facility in a "fine tuned" versus a "poor combustion" mode. Thus, the
combination of combustion system design and operational tuning was sufficient
to reduce total PCDD/PCDF from a high of nearly 4000 ng/Nm3 to 12 ng/Nm3-
For this facility, combustion control was sufficient to reduce the PCDD/PCDF/
emission rate from near the top end to near the bottom end of the range
indicated in Table 2-1. The goal of the current study is to gather and
evaluate available information to help identify those components of
engineering design, operation, and monitoring practices which minimize trace
organic emissions from municipal waste combustor systems.
2-4
-------
2-2 Program Overview
This volume of the comprehensive report series on municipal waste
combustion is the culmination of a program having three major objectives:
• To determine the current state of combustion control of municipal
waste combustion technology;
• To recommend design and operating approaches which will minimize
emissions of trace organics; and
• To define the research which is necessary to develop and verify
these design and operating approaches.
The focus of this volume is to present information on the combustion
practices which will minimize the emissions of organic compounds including
PCDDs and PCDFs. However, the interrelationship with other pollutants such
as particulate matter, metals, NOX> and carbon monoxide must also be
established so that conditions are not followed that minimize emissions of
PCDD/PCDF while increasing the emissions of other potentially harmful
compounds. The focus of this volume is on the design of new units and the
operation and monitoring of both new and existing units.
In Figure 2-1, the approach followed in the study that resulted in this
volume is shown. The approach was designed in an attempt to make use of the
knowledge base, which resides with those engineers who design and construct
waste-to-energy facilities. Detailed, face-to-face meetings were held with
fifteen of the major organizations associated with waste-to-energy operations
in the United States and Europe. Four other U.S. manufacturing firms were
contacted by telephone. Table 2-2 provides a listing of those organizations
who contributed to this study with their technical expertise in a series of
individual fact-finding meetings. Archival combustion literature and
literature on PCDD/PCDF formation mechanisms made up the remainder of the
input data base.
-2-5
-------
MANUFACTURER DESIGN
APPROACH AND
KNOWLEDGE BASE
ENGINEERING ANALYSIS
- SYSTEMS
- FAILURES
- CASE STUDY
DESIGN/OPERATING
GUIDELINES
- REVIEW PROPOSALS
- RECOMMENDATIONS
RESEARCH NEEDS
- UNCERTAINTIES/
ISSUES
- EXPERIMENTAL
APPROACH
LITERATURE DATA
COMBUSTION
CONTROL
FINAL
REPORT
Figure 2-1. Program approach.
2-6
-------
TABLE 2-2. MANUFACTURERS FACT FINDING
• EUROPEAN MANUFACTURERS
VOLUND (DENMARK)
DEUTSCHE BABCOCK (GERMANY)
STEINMULLER (GERMANY)
VON ROLL (SWITZERLAND)
WIDMER AND ERNST (SWITZERLAND)
MARTIN (GERMANY)
• OTHERS
DANISH EPA
PROF. HUTZINGER (UNIV. OF BAYREUTH)
ANRED (FRENCH AGENCY FOR WASTE
RECOVERY AND DISPOSAL)
AMERICAN BOILER MANUFACTURERS ASSN.
AMERICAN MANUFACTURERS
RILEY STOKER *
.COMBUSTION ENGINEERING
WESTINGHOUSE/O'CONNOR
DETROIT STOKER
FOSTER WHEELER
BABCOCK & WILCOX *
CONSUMAT INCINERATION SYSTEMS
JOHN BASIC*
ENERCON *
CONTACTED BY TELEPHONE ONLY
-------
The second phase of the program was the analysis and interpretation of
the available data on design and .operating of municipal waste combustion
equipment. This phase included an analysis of current practices in different
types of municipal waste combustor systems and potential failures that might
occur in both design and operation that could lead to the emission of trace
organics. The engineering analysis phase also included an examination of
case studies of incinerator facilities of older design and operating
practices which were found to have higher PCDD/PCDF emissions than is common
for modern units.
From the engineering analysis phase, recommendations were made on design
and operating approaches that would reduce PCDD/PCDF/furan emissions and
prevent system failure. This program phase included a review and critique on
the draft guidance prepared by PEER Consultants (1986). Finally
recommendations were made on research approaches to address those
uncertainties and issues that cannot be addressed with the present knowledge
base.
The remaining chapters cover overviews of the municipal waste combustion
process and potential for air emissions (Chapters 3 and 4; current practice
in different types of combustion systems (Chapter 5, 6 and 7; and approaches
to combustion control to minimize PCDD/PCDF emissions (Chapter 8 and 9).
2.3 References
Finkel stein, A., et al . Presentations by Environment Canada at
Municipal Solid Waste Incineration Research and Planning Meeting.
Durham, North Carolina. December 9-11, 1986.
Haile, C. L., et al. Assessment of Emissions of Specific Compounds from
a Resource Recovery Recovery Municipal Refuse Incinerator (Hampton,
Virginia). EPA-560/5-84-002. June 1984.
Howes, J. E., et al. Characterization of Stack Emissions from Municipal
Refuse-to-Energy Systems (Hampton, Virginia; Dyersburg, Tennessee; and
2-8
-------
Akron, Ohio). Prepared by Battelle Columbus Laboratories for U.S.
Environmental Protection Agency/Environmental Research Laboratory.
1982.
Neulicht, R. Emission Test Report: City of Philadelphia Northwest and
East Central Municipal Combustors. Prepared for U.S. Environmental
Protection Agency/Region III by Midwest Research Institute. October
1985.
New York State Department of Environmental Conservation. Emission
Source Test Report - Preliminary Test Report on Westchester RESCO.
January 8, 1986.
Nunn, A. B., III. Evaluation of HC1 and Chlorinated Organic Compound
Emissions from Refuse Fired Waste-to-Energy Systems (Hampton, Virginia;
and Wright-Patterson Air Force Base, Ohio). Prepared for U.S. EPA/HWERL
by Scott Environmental Services. 1983.
Ogden Projects. Tulsa Waste-to-Energy Tests Show Low Dioxin. Coal and
Synfuels Technology. September 1, 1986, p. 6.
PEER Consultants, Inc. Design and Operating Guidance to Minimize
Dioxins and Other Emissions from Municipal Waste Combustors. Draft
Report to EPA Office of Solid Waste Under EPA Contract 68-02-6940.
May 19, 1986.
Scott Environmental Services. Sampling and Analysis of Chlorinated
Organic Emissions from the Hampton Waste-to-Energy System. Prepared for
The Bionetics Corporation. May 1985.
Seelinger, R., et al. Environmental Test Report (Walter B. Hall,
Resource Recovery Facility, Tulsa, Oklahoma)- Prepared by Ogden
Projects, Inc. for Tulsa City County Health Department. September 9,
1986.
2-9
-------
3.0 MUNICIPAL WASTE COMBUSTION PROCESSES
The current report considers combustion control of organic species from
three types of municipal waste combustors, namely:
• Mass-burning excess air (mass burning)
• Starved air two-stage (modular)
• Refuse derived fuel firing (RDF),
The mass-burning excess air and starved-air two stage categories include the
combustion systems which burn raw municipal solid waste. Descriptions of th'e
hardware and design features offered by various mass-burn and modular system
suppliers are presented in Chapters 5 and 7 respectively- Processes have
been developed which convert raw municipal solid waste into a more uniform
fuel termed refuse derived fuel (RDF). RDF may be burned in a wide variety
of furnaces including furnaces designed to fire coal or other fossil fuels
such as wood or bagasse. RDF may also be co-fired with traditional fossil
fuels. The processes to produce RDF and hardware configurations being built
for combustion of RDF are presented in Chapter 6. Fluidized bed combustion
systems are being built to burn RDF; however, as illustrated elsewhere in the
comprehensive municipal waste combustion study, fluidized bed systems
represent only a small portion of the current and projected refuse-to-energy
market. Accordingly, this technology is not included in this report.
This chapter provides a brief review of the physical and chemical
processes which occur during municipal waste combustion. The discussion
illustrates the relationship between combustion process design and the
emission rate of major pollutants of concern:
• Polycyclic organic matter (including PCDDs and PCDFs)
• Particulate matter
t Nitrogen oxides
3-t
-------
Throughout the discussion, any reference to emissions rate or exhaust
concentration will refer to conditions at the boiler exhaust, upstream of any
flue gas air pollution control device (APCD).
3-1 Municipal Waste Characteristics
An understanding of municipal waste combustion must begin with an
appreciation for the non-uniformity of MSW and the implication of that
variability on combustion system design. Typical wastes might include paper
products, food scraps, tin and aluminum cans, glass bottles, a wide range of
plastic items, cloth items, floor sweepings, etc. During the summer months
there will be plastic bags .of yard clippings, and in the fall, dead leaves
might replace grass clippings. Thus, a municipal waste combustor burning
minimally treated waste must be able to handle a heterogeneous feed.
Characterizations of the constituency of MSW represent the average over
large volumes of waste and may also include temporal averaging. Table 3-1
presents a characterization of MSW in the United States as reported by
Franklin Associates (1986). Table 3-2 presents an ultimate analysis of MSW
as reported by Hickman, et al., (1984). The higher heating value of the MSW
described by both Franklin Associates and by Hickman, et al., was on the
order of 4500 Btu/lb. This heating value is typically used to estimate mean
operating conditions for mass burning municipal waste combustion facility
designs, Riley Stoker (1986) and Detroit Stoker (1986); note that system
performance guarantees are typically based on burning 4500 Btu/lb MSW, but
that systems are designed to accommodate waste with average heating values
which range from 3800 to 6000 Btu/lb. During rainy periods the heating value
of MSW will decrease while dry periods or holiday seasons tend to result in
MSW with increased heating value.
The municipal waste delivered to a resource recovery facility will have
items of all sizes and shapes. Many municipal waste combustors remove bulky-
oversize items such as refrigerators but the characteristic size is
sufficiently large that unprocessed MSW must be burned on a grate. Feeding
the MSW at a controllable rate and achieving uniform distribution across the
3-2
-------
TABLE 3-1. CURRENT AND FORECAST COMPOSITION OF DISPOSED RESIDENTIAL
AND COMMERCIAL SOLID WASTE (WEIGHT PERCENT)
Component
Paper and Paperboard
Yard Wastes
Food Wastes
Glass
Metals
Plastics
Wood
Textiles
Rubber and Leather
Miscellaneous
TOTAL
Year
1980
33.6
18.2
9.2
11.3
10.3
6.0
3.9
2.3
3.3
1.9
100.0
1990
38.3
17.0
7.7
8.8
9.4
8.3
3.7
2.2
2.5
2.1
100.0
2000
41.0
15.3
6.8
7.6
9.0
9.8
3.8
2.2
2.4
2.1
100.0
3-3
-------
TABLE 3-2. ULTIMATE ANALYSIS OF TYPICAL MSW AS
PRESENTED BY HICKMAN ET AL. (1984)
Ultimate Analysis
Moisture
Carbon
Hydrogen
Oxygen
Nitrogen
Chlorine
Sulfur
Inorganics (ash)
25.2
25.6
3.4
20.3
0.5
0.5
0.2
24.4
100.0
3-4
-------
grate is difficult but necessary for proper operation. Mass burning
facilities use cranes or front-end loaders to fill a loading chute.
Hydraulic rams or pusher grate sections are used to load the MSW into the
municipal waste combustor. Municipal waste combustion facilities use a thick
fuel bed which tends to damp out effects due to the variability in fuel
heating value and the inability to feed waste at a precise rate.
Once the MSW is in the furnace it is slowly moved by grate action toward
the ash dump. For the MSW to burn it must be exposed to combustion air and
heated. Combustion air is introduced both through the grate (underfire or
undergrate air) and through air jets located above the bed (overfire air).
To achieve fuel burnout the grate must be designed to agitate or stir the
thick fuel bed as it moves the MSW from entrance to exit. Each of the grate
manufacturers have developed their designs to accomplish this objective as
discussed in Chapter 5.
3.2 Combustion Control of Pollutant Emissions from Municipal Waste
Combustion Facilities
Chapter 4 provides a discussion of the potential air emissions from
municipal waste combustors. The "emissions data base for municipal waste
combustion" study provides a summary of available field test data from
municipal waste combustion facilities including emissions data on criteria
pollutants, acid gases, trace metals, and various hydrocarbons including
toxic organic pollutants. An evaluation of PCDD and PCDF emissions from
resource recovery facilities is available in a recent report provided to EPA
by Camp Dresser and McKee (1986).
Details of the combustion process will impact directly the emission rate
of several criteria pollutants (CO, NOX and total particulate matter) and
unburned hydrocarbons. In addition, they may impact the emission rate of
trace metals. The following subsections will review a limited number of
basic combustion concepts and then discuss the relationship between
combustion processes and emissions. The focus of this discussion will be
municipal waste combustion in mass burning, excess air facilities.
3-5
-------
3-2.1 Basic Combustion Concepts
Figure 3-1 is a schematic of the grate region of a mass burning excess
air municipal waste combustor- The waste-fuel enters the grate from the
charging hopper (1) on the left and is moved toward the ash dump on the
right. Table 3-2 presented an analysis for an MSW containing 24.4 weight
percent ash. Ideally, the combustible constituents would be burned and the
ash dumped from the system for landfilling. Air must be supplied to sustain
the burning process. To completely burn the waste the theoretical air
requirement is determined by the formula
»
Wa = 11.5 C + 34.5 (H-0/8) + 4.32S
where Wa is the mass of air per mass of fuel at stoichiometric conditions and
C, H, and 0 and S are the weight fractions of carbon, hydrogen, oxygen and
sulfur in the fuel, respectively. For one pound of MSW defined in Table 3-2,
3.25 pounds of air (approximately 42.8 cubic feet at 70°F and 1 atmosphere)
are required to convert all of the carbon to C02, all of the hydrogen to
water, and all of the sulfur to S02- If this quantity of air and MSW were
mixed and burned to completion there would be a zero oxygen concentration in
the exhaust. For a variety of reasons, including the variability of fuel
characteristics, municipal waste combustion facilities are operated with
significantly more than the theoretical air requirements, typically 70 to 100
percent excess air at full load. If a facility were burning the MSW defined
in Table 3-2 with 100 percent excess excess air, the extra 3.25 pounds of air
per pound of fuel would serve as a diluent to the combustion products. The
pound of MSW burned would still release only 4500 Btu of heat and thus the
temperature of the combustion gas would be decreased relative to that at
lower excess air conditions.
The MSW defined in Table 3-2 contained approximately 25 percent
moisture. As illustrated in Figure 3-1, many municipal waste combustors
provide an arch over the grate region where the MSW enters the municipal
waste combustor. Heat from the hot arch wall is transferred to the fuel bed,
raising the MSW temperature enough to begin driving off the moisture. This
3-6
-------
Figure 3-1. The L&C Steinmueller mass burn design features,
3-7
-------
initial grate section is referred to as the drying grate. Undergrate air
provided to this section will help to drive off the moisture, particularly if
the air is preheated slightly. The important point is that the volatile
matter released by the MSW in the initial grate section is mainly water and
thus limited net heat release occurs in this region of the furnace.
The second section of grate shown in Figure 3-1 is referred to as the
burning grate. Because much of the MSW moisture will be driven off in the
drying grate region, the fuel entering the burning grate will have a
significantly increased heating value. The fuel is still in a solid state
and the bed may be upwards of a meter thick. Williams, et al ., (1974)
describe a series of pilot scale studies designed to define the controlling
phenomena in thick burning beds of solid waste. That report also reviews
several attempts to model the process. As described in that report (also see
Niessen) et al . , (1972)) the top layer of the fuel bed is ignited by
radiation received from the hot combustion gas and the refractory lining of
the lower furnace. This is followed by slow propagation of the ignition wave
down into the fuel bed. The experimental data indicated that the underfire
air flow rate and the fuel consumption rate were essentially in
stoichiometric proportion but that the gas composition at the top of the fuel
bed generally corresponded to the equilibration of the water-gas shift
reaction:
C02 + H2 = CO + H20
Figure 3-2 is taken from the Williams et al., report and shows the temporal
variation in ignition front location, bed thickness and depth of accumulated
inert. These data were obtained using a stationary bed of waste fuel and do
not include the effect of fuel bed agitations by motion of the grate. The
data do illustrate, however, many of the basic features of bed burning.
Figure 3-3, compiled from data taken during the same experiment, illustrates
gas composition measured by a sampling probe located within the fuel bed, 12
inches above the grate. As shown, the ignition front reaches the sampling
location in approximately 2000 sec. Prior to that time the sampling probe
only detects underfire air- As the ignition front moves closer to the grate,
3-8
-------
I
10
LJ
K
<
cr
CD
LJ
>
o
CD
Ld
U
Ul
5
Location of Top of Bad
Location of
Ignition Front
Depth of Accumulated
Inert —
Active Burning
Depth
0.8 -
0
0
100O 2000 3OOO
TIME (SEC)
4000
Figure 3-2. Estimate of fuel bed thickness and accumulated inert
layer (Run 18).
-------
CO
I
O
-
00
- 90
20
CD
cc
Q
O
to
O
o.
2
O
u
*• Time Ignition Front
4 ,-
O
Ignition Fron t s
at Grate
to
CD
2000
3OOO
4000
ELAPSED TIME(SEC)
Figure 3-3. Gas compositions from fuel bed probe 2 (Run 18).
-------
the local oxygen concentration falls with a commensurate increase in C02, CO,
H2 and CH4. After approximately 2500 sec, the local 02 concentration falls
to zero indicating that the active burning region is below the sampling
location.
The data trends shown in Figure 3-3 are consistent with viewing the MSW
fuel bed as a gasifier. The process begins with reactions producing char,
C02 and H20 plus the vaporization of any free moisture. This is followed by
endothermic reactions (reactions for which heat must be added) consisting of:
C + H20 ^ CO + H2 and
C + C02 ^ CO + CO
Note in Figure 3-3 the continual increase in CO and H2 concentration until
approximately 3300 sec. while the local C02 concentration is seen to fall.
In an actual MSW municipal waste combustor, propagation of the ignition
front and gas concentration profiles will be far more complex than described
above. The bed material in the experiments by Williams, et al. (1974) was
either simulated or real RDF and not MSW. Further, in actual municipal waste
combustors the grate action is designed to mix and aerorate the MSW bed which
results in dispersion of the ignition front. Regardless, the pilot-scale
results illustrate several of the critical features of municipal waste
combustion in the burning grate region. One of the critical features
illustrated is the need for overfire air. The underfire air added to the
burning grate coupled with grate agitation combined to define the rate at
which the MSW is consumed. The gases leaving the fuel bed, however, contain
a significant concentration of ^2* CO, and unburned hydrocarbons. Additional
air must be mixed with that effluent to complete the conversion of
combustibles to C02 and H20. The complete conversion is essential for
maximizing combustion efficiency and for minimizing pollutant emissions. A
portion of that air will be provided by underfire air which short circuits
the fuel bed by channeling. The majority, however, must be provided by the
overfire air supply.
3-11
-------
3-2.2 Criteria Pollutants
Combustion process details can have a significant impact on the emission
of several criteria pollutants including CO, NOX and total particulate.
Discussion of how combustion conditions impact on emission and control of
NOX, particulate matter and trace metals is contained in Chapter 4.
As noted in the previous section, substantial concentrations of CO will
be released from the fuel bed. Controlling emission of CO from the facility
exhaust requires that a controlled quantity of air be mixed with that
effluent. Carbon monoxide is the most refractory species in the oxidation
chain from hydrocarbon to C0£ and H20. The dominant kinetic step for CO
oxidation is:
CO + OH —» C02 + H
Adding too much air to the effluent from the bed will depress the local gas
temperature and the concentration of the OH radicals. If too little air is
added, the probability is increased that pockets of gas from the burning
grate will exit the hi gh. temperature regions without ever being mixed with
oxygen. Figure 3-4 is taken from data provided in a meeting with Deutsche
Babcock (see Section 5.2) and illustrates the existence of an appropriate
operating range for minimum CO emissions.
3.2.3 Emission of Organic Compounds
One of the major objectives of the current report is to develop a
framework for defining municipal waste combustor design and operating
practices which minimize the emission of trace organic compounds. The
organics of particular concern are the polychlorinated dibenzo-p-dioxins
(PCDDs) and the polychlorinated dibenzofurans (PCDFs). Available field test
data were outlined previously in Table 2-1 and indicated that PCDD emissions
ranged from 1.13 to 10,700 ng/Nm3 while PCDF emissions ranged from 0.423 to
14,300 ng/Nm3. In developing a control strategy for these compounds, it is
important to recognize that these emissions rates, including those at the
3-12
-------
X i—I
Oh-
z-a:
OCC
OCJ
co 2:
ceo
CJ
3 6 9 12
OXYGEN CONCENTRATION
A - INSUFFICIENT AIR C+|02-*CO
B - APPROPRIATE OPERATING REGION
C - "COLD BURNING"
Figure 3-4. Relationships of CO and CL for
appropriate operating regions.
3-13
-------
high end of the range, represent trace concentrations. The measured PCDD
concentrations range from approximately 1 ppb (molar basis) to a fraction of
a part per trillion.
Combustion begins with a fuel which is itself an organic and proceeds by
destroying organics through a complex series of oxidation reactions. In the
earlier discussion of basic combustion concepts, an equation was presented
for calculating the amount of air theoretically required to completely burn a
given mass of hydrogen fuel. That equation is based on determining the
quantity of air required to oxidize fuel carbon to C02, fuel hydrogen to h^O
and fuel sulfur to SC^. The actual burning process is propagated through a
complex series of chemical reactions. Exhaust emission of organic compounds
represents a failure to complete the oxidation reactions. Chapter 4.0 will
discuss current theories/understanding of how PCDDs and PCDFs are formed and
destroyed during municipal waste combustion. Generally, however, emission of
organics will occur because of a failure to supply sufficient oxygen to all
of the reacting gases or termination of the oxidation reaction through some
quenching process.
As noted previously, definition of an acceptable PCDD or PCDF emission
rate does not currently exist. Considering the fact that dioxin emission
concentrations in the part per trillion range may be of concern, it is
important to determine if there is a lower limit on emission reduction by
combustion control. That limit, if any, may be examined through equilibrium
calculations which predict the concentration of reaction product species in
the absence of mixing or chemical kinetic limitations. Equilibrium
calculations depend on the elemental composition of the mixture (moles of C,
H, 0, N, S, Cl, etc.), the reaction temperature, and the thermodynamic
properties of product species.
Kramlich et al., (1984) studied the equilibrium product distribution for
various chlorinated benzene/air mixtures. Since equilibrium calculations
depend on elemental composition of the mixture rather than the chemical
structure of the reactants, results from the Kramlich study should reflect
limiting condition trends for municipal waste combustion. Sample results are
3-14
-------
presented in Figure 3-5 as a plot of species concentration versus percent
theoretical air assuming that the mixture is at the adiabatic flame
temperature. The curve labeled THC represents the predicted total
hydrocarbon concentration (sum of concentration of all hydrocarbon species
present). For mixtures with at least 55 percent of the theoretical air
requirements (and no heat loss) the equil ibrium hydrocarbon concentration
will be less than 1 ppt. As the mixture ratio becomes more fuel-rich, the
presence of light hydrocarbons such as CH4 (and later Cz^Z' etc.) are
predicted. Part per trillion concentrations of benzene and tolune are not
predicted until the mixture contains less than 30 percent of the theoretical
air requirements. Prediction of ppt equilibrium concentrations of PCDD, PCDF
or precursors such as chlorobenzenes and chlorophenols are restricted to even
more fuel rich conditions.
Equilibrium calculations may also be used to evaluate other limiting
case characteristics. Consider, for example, the situation when a pocket of
organic gas does exist in the municipal waste combustor and is mixed into an
oxidizing gas stream. That situation may be described by the overall
reaction:
Kp -
PC02
porganic
where P represents the partial pressure of a given constituent and Kp is the
equilibrium constant. The equilibrium constant is fundamentally related to a
measurable thermodynamic property called the Gibbs Free Energy (AG) by:
Kp = EXP (AG/RT)
where T is temperature and R is the Universal gas constant. If the organic
compound is taken as a furan, G would be approximately 492 Kcal/mole (Stull,
et al., 1969). For typical municipal waste combustor exhaust gas C02, H20
and 02 concentrations and for an assumed reaction temperature, the
equilibrium concentration of furan can be calculated. For temperatures as
3-15
-------
10
I I
0
cr
10
10'
-2 -
2 10~6
10
-8
10
-10
0 20
OVER ENTIRE RANGE
I | | I
60 80 100 120 140 160 180 200
PERCENT THEORETICAL AIR
Figure 3-5. Adiabatic equilibrium species distribution.
3-T6
-------
low as 1000K, that equilibrium furan concentration is less than ID'100 mole
fraction.
The above discussion illustrates two significant aspects of combustion
control for trace organics. First, at reasonable temperature levels and
under excess oxygen conditions, there is no thermodynamic barrier to
achieving essentially zero emission level of trace organic. (There may,
however, be mixing or kinetic barriers). Second, semi-volatile organic
compounds are not thermodynamically stable under high temperature conditions,
unless the mixture fuel/air ratio is very fuel-rich. The fact that organics
such as PCDD and PCDF are emitted from municipal waste combustors indicates
that chemical kinetic or mixing factors have prevented the gases from
reaching the equilibrium condition. The chemical kinetic and mixing aspects
will be considered in greater detail in later portions of this report.
3.3 Combustion Control of Starved Air Systems
Figure 3-6 illustrates the basic components of starved-air systems which
are typically small (less than 100 ton/day capacity per unit). MSW is
generally fed to the system with a front-end loader and charged into the
primary chamber with a hydraulic ram. The typical solids residence time in
the first stage chamber is on the order of 10 to 12 hours which helps to damp
out variations in fuel heating value and the cyclic nature of the charging
process.
The primary zone functions as a gasifier with only about 40 percent of
the theoretical air requirements. The relatively large volume of this
chamber coupled with low air flow results in low gas velocity and minimal
particulate entrainment.
In mass burn, excess air systems, the effluent from the fuel bed
contains significant concentrations of CO, H£ and unburned hydrocarbon.
There is, however, an overall excess air condition in the lower portion of
the municipal waste combustor. In modular, starved air systems, the overall
primary zone effluent is fuel-rich. Completion of the combustion process is
3-17
-------
CONTROLLED AIR
SECONDARY
SR >100%
HYDRAULIC RAM
CHARGING
WASTE HEAT
RECOVERY
CONTROLLED AIR TO
PRIMARY CHAMBER
SR<100%
Figure 3-6. Components of typical starved-air modular combustor.
3-18
-------
totally dependent upon the efficiency of air mixing in the secondary chamber.
The small modular systems have no heat removal from either the primary or
secondary chamber. Energy recovery is accomplished in a downstream waste
heat boiler.
3.4 RDF Systems
As noted at the beginning of this chapter, there are two basic
approaches to dealing with the problem of variability in MSW. The mass burn
and starved-air systems accept the variability of MSW as a fuel
characteristic to be controlled in the combustion system design. The-
alternate approach is to pre-process MSW to generate a more uniform fuel.
The objective is to modify the fuel characteristics sufficiently for the
refuse-derived fuel( RDF) to be burned in boilers designed for fossil fuels,
e-g. spreader stokers and suspension fired units. Also, RDF might be burned
in cyclone boilers, pulverized coal-fired boilers or fluidized bed furnaces.
Processed fuel might be burned separately or it might be burned in a co-
firing configuration using the RDF as a supplement to the baseline fossil
fuel.
To function properly, the RDF characteristics should be matched to
combustion system requirements. In its simplest form, RDF is produced by
passing the raw waste through a shredder after first removing bulky items
such as refrigerators. This results in a fuel which can be burned in
spreader stokers. RDF production may also include metals removal which
produces a salable by-product. Often that process will be coupled with
screening and secondary shredding. Such a fuel is better suited for firing
in spreader stokers than simple "shred-and-burn" operations. Production of
refuse-derived fuel is a relatively recent engineering innovation; thus it is
not surprising that RDF systems have been plagued with reliability problems.
Some systems have produced a fuel product with less than the desired
characteristics which creates significant difficulty achieving uniform fuel
feed to the boiler.
3-19
-------
3-5 References
Camp Dresser & McKee, Inc., "Dioxin Emissions from Resource Recovery
Facilities and Summary of Health Effects," Draft Report prepared for
U.S. EPA Office of Solid Waste, November 19, 1986.
Detroit Stoker, 1986. Discussions and data presented by Detroit Stoker
personnel (T. A. Giaier and D. C. Reschley) to W. S. Lanier and
W. R. Seeker in Monroe, Michigan, September 9, 1986.
Franklin Associates, Ltd., "Characterization of Municipal Solid Waste in
the United States, 1960.to 2000," (prepared for the U.S. Environmental
Protection Agency) Washington, D. C., July 11, 1986, pp. 1-8.
Kramlich, J. C., M. P. Heap, W. R. Seeker, and G. S. Samuelsen.
20th Symposium (International) on Combustion, p. 1991, The Combustion
Institute.
Niessen, W. R., A. F. Sarofim, C. M. Mohr, and R. W. Moore, "An Approach
to Incinerator Combustible Pollutant Control," Proceedings National
Incinerator Conference, ASME, New York, pp. 248-259, 1972.
Stull, R. D., E. F. Westrum, and G. C. Sinke; The Chemical
Thermodynamics of Organic Compounds, Wiley, 1969.
Telecon. Riley Stoker Corp., with Lanier, W. S., EER Corp., August 17,
1986. Conversation about Riley Stoker's MSW System design practices.
G. C. Williams, A. F. Sarafim, J. B. Howard, and J. B. L. Rogers.
"Design and Control of Incinerators," Final Report to Office of
Research and Monitoring, U.S. Environmental Protection Agency under
Grant No. EC-00330-03, 1974.
3-20
-------
4.0 POTENTIAL FOR AIR EMISSIONS
During the combustion of municipal waste, a number of pollutant species
can be produced. The pollutants include both criteria pollutants and other
pollutants as follows:
• Nitrogen oxides, NOX
0 Carbon monoxide, CO
• Acid gases (HC1, S02, HF, H2S04)
• Particulate matter
t Metals (As, Be, Cd, Pb, Hg, Ni, etc.)
• Toxic Organics
The focus of this study is on the combustion control of emissions of organics
such as chlorinated dibenzo-p-di oxi n (PCDD) and chlorinated dibenzofurans
(PCDF). However, the control of PCDD and PCDF emissions should not be
pursued without consideration of how control scenarios will influence the
emission of other pollutants. In some instances, the control schemes may in
fact adversely impact the system's ability to control the emissions of
another pollutant. The following sections will provide information on
formation and control schemes of the variety of potential air pollutants.
4.1 Organic Emissions
There are a number of organic compounds that have been measured in
effluents from municipal waste combustors (MWCs). In addition, other
hydrocarbons including a broad range of aromatic and chlorinated organics
might potentially be emitted. The principle emphasis at present is being
placed on the chlorinated congeners of dibenzo-p-dioxi n (PCDD) and
dibenzofuran (PCDF). However, PCDD and PCDF comprise only one of many types
of organics emitted from municipal waste combustion (MWC) facilities that may
be eventually of concern. For this reason, control approaches to prevent
emissions of trace organics in general, not just PCDD and PCDF, are included
in the subsequent chapters although the emphasis is clearly directed towards
PCDD and PCDF.
4-1
-------
As illustrated previously in Table 2-1, a wide range of PCDD and PCDF
emission rates have been measured from various refuse fired facilities
throughout the world. A comprehensive review of available emission data from
municipal waste burning facilities has recently been compiled for EPA by
Midwest Research Institute and is included as a volume entitled "Emission
Data Base for Municipal Waste Combustors."
There are many different theories concerning the formation of PCDD's and
PCDF's from MSW combustion systems (Germanus, 1985, Niessen et al ., 1984,
1986). The best-supported theories are illustrated in Figure 4-1. The first
theory involves the breakthrough of unreacted PCDD/PCDF present in the raw
refuse. A few measurements have indicated the presence of trace quantities
of PCDD/PCDF in the refuse feed that, if not burned in the furnace, could
account for the levels of emissions (Lustenhouwer et al., 1980). The trace
levels of PCDD/PCDF found in the solid waste feed would have to be completely
unreacted to account for the emissions levels of these species which is
unlikely in any combustion environment (Germanus, 1985).
A more plausible theory involves the conversion of species referred to
as precursors which are of similar structure. For example, relatively simple
abstraction and combination reactions can convert chlorophenols and
polychl orinated biphenyls to PCDD/PCDFs. These precursors can be in larger
abundance in the refuse and can be produced by pyrolysis in oxygen-starved
zones. There is direct evidence of gas phase yields of PCDD/PCDF from PCB
fires (Axelrod, 1985) and lab and bench-scale studies on PCB, chlorinated
benzene and chlorinated phenols (Olie et al., 1983, Hutzinger et al., 1985).
The third -mechanism involves the synthesis of PCDD/PCDF from a variety
of organics and a chlorine donor. Again, the simplest mechanisms involve
those species that are structurally related to PCDD/PCDF but a full spectrum
of plausible combustion intermediate chemistry could be proposed to lead to
precursors and eventually to PCDD/PCDFs. For example, the analysis for
intermediates formed during the combustion of complex fuels such as coal and
wood indicate yields of (unchlorinated) PCDD and PCDF species (Hites and
Howard, 1978) that could be chlorinated when a suitable chlorine donor is
4-Z
-------
I.
PRESENCE IN REFUSE
Cl
Unreacted
PCDD/PCDF
II.
Evidence: Occasional PCDD/PCDF contamination in refuse
FORMATION FROM RELATED CHLORINATED PRECURSORS
III,
Furan
Evidence: PCDD/PCDF on soot from PCB fires
Lab and bench studies of PCB, Chlorinated Benzene
and Chlorinated Phenols yielded PCDD/PCDF
FORMATION FROM ORGANICS AND CHLORINE DONOR
PVC \
Lignin J
+ Chlorine donor
Nad , HC1, Cl2
PCDD/PCDF
IV.
Evidence: Lab scale tests of vegetable matter, wood, lignin,
coal with chlorine source yielded PCDD/PCDF
SOLID PHASE FLY ASH REACTION
Precursor
+ Cl Donor
low
temp
PCDD
Evidence: Lab scale demonstrating potential for ash catalysis
reactions of PCDD's to other homologues.
Figure 4-1. Hypothetical mechanisms of PCDD/PCDF formation chemistry.
4-3
-------
available. Complex fuels which contain lignin have been shown to yield
higher levels of PCDD/PCDF than non-lignin fuels (Niessen, 1984).
The final hypothetical mechanism involves catalyzed reactions on fly ash
particles at low temperatures. The lab scale evidence indicates that the
chlorination of related-structure precursors on fly ash particles to form
PCDD/PCDF can occur but that the rates are relatively slow (Eiceman et al.,
1982; Rhgei and Eiceman, 1982). Recent work (Vogg, Metzger and Stieglitz,
1987) has shown the importance of this mechanism as well as quantifying the
temperature dependence.
In summary, current theories on formation of PCDD/PCDF involve either
unburned PCDD/PCDF from the refuse or related-structure precursors, or
precursors that are likely to be formed in oxygen-starved zones of the
furnace. The precursors can be converted to PCDD/PCDF even at low
temperatures on fly ash. Vogg et al. (1987) indicate that the optimal
temperature for such catalytic reactions is approximately 600°F. Hence it
is important to destroy the PCDD/PCDF and the precursors in the furnace.
Therefore, the destruction mechanisms are equally as important as the
formation mechanisms. Assuring complete destruction will likely be the
more appropriate combustion control method rather than the prevention of
formation.
The destruction of PCDD/PCDF is expected to be similar to the oxidation
of other aromatic chlorinated hydrocarbons primarily involving free radicals
(such as OH) which attack the structure or unimolecular decomposition (Shaub
and Tsang, 1983). The details of the mechanism are uncertain and more
fundamental research is required. Direct experimental measurements of the
thermal decomposition characteristics of non-chlorinated PCDD and PCDF
species are available from data at the University of Dayton Research
Institute (UDRI) (Duvall and Rubey, 1977). These studies using the Thermal
Decomposition Analytical System (TDAS) provide a definition of the thermal
requirements for destruction under highly diluted reactant conditions. These
conditions, sometimes referred to as nonflame, provide a conservative
indication of the temperature to which the test species must be subjected in
4-4
-------
order to achieve destruction. The thermal decomposition behavior of (non-
chlorinated) PCDD/PCDF and hypothetical precursors are shown in Figure 4-2 at
a specific residence time.
The thermal decomposition data indicate that relatively low temperatures
are required to destroy PCDD and PCDF and that the temperature requirements
are similar to the decomposition temperature required for non-substituted
biphenyl. Chlorinated biphenyls have higher decomposition temperatures and a
similar increase in the temperature requirements are expected for PCDD/PCDF
due to the expected similarity of the decomposition mechanisms of PCB and
PCDD/PCDF. These data indicate that 800°C (~1500°F) is a characteristic
temperature for high levels of destruction (>99.9 percent) of PCB and by
analogy for destruction of PCDD/PCDF. This temperature is in general
agreement with the Dow Chemical conclusions that TCDD decomposition occurs
above 800°C (Bumb, et al . 1980, Stehl et al. 1973). Other potential
precursors to the formation of PCDD/PCDF can have somewhat higher
decomposition temperatures but are still low relative to combustion
temperatures.
The UDRI decomposition data (Rubey et al ., 1983 and Dellinger et al.,
1983) also indicate that temperature is more important than residence time.
The data for the thermal decomposition temperature generally correlates with
the residence time t as
E
a
T = —
R
-tA \ /[02J
vln(fr) / \0.21
-1
= Thermal Decomposition
Temperature
where [03] is the oxygen mole fraction, b is the reaction order with respect
to oxygen, R is the universal gas constant, fr is the fraction of the species
remaining and Ea and A are constants which depend on the test species in
question. Thus, the destruction of the species depends predominantly on the
temperature and effectively there exists a threshold temperature above which
the compound will be rapidly destroyed. Effective combustion control schemes
must ensure that all PCDD/PCDF and precursors experience the required
4-5
-------
100
HEXACHLORO-
'BENZENE
CD
10
CJ
cc
.1 _
1000
DIQXIN
FURANS
1100 1200 1300 1400 1500 1600
TEMPERATURE (°F)
Figure 4-2.
Thermal decomposition characteristics of selected
organics under dilute (nonflame) conditions for
1 second (Dellinger et al. 1977).
4-6
-------
threshold temperature if they are to be destroyed. For PCDD/PCDF and
potential precursors, the threshold temperature is near 925°C (1700°F).
Another possible control scheme involves removal of PCDD/PCDF from the
flue gas. PCDD/PCDF are condensible at flue gas temperatures and will
deposit on fly ash particles very efficiently. Because of the higher surface
areas of smaller particles, the condensation will occur preferentially on
smaller particles. Hence, a removal system must efficiently remove the finer
particles in order to effectively remove PCDD/PCDF- Data on full scale tests
have clearly shown the effectiveness of a high efficiency baghouse with dry
injection flue gas treatment (Martin, 1986) in removing PCDD/PCDF from the
flue gas.
Thus, in summary, the current theories concerning the emission of PCDD/
PCDF and other similar organics include any or all of the following
possibilities:
• Lack of destruction of PCDD/PCDF originally present in the feed
refuse
• Conversion of precursors present in the feed stock or formed in the
combustion processes to PCDD/PCDF and lack of destruction of the
synthesized PCDD/PCDF
t Lack of destruction of precursors in the combustion system and
conversion of the precursors to PCDD/PCDF on fly ash particles at
low temperatures.
Combustion control of the emission of PCDD/PCDF must ensure that both
PCDD/PCDF present in the feed and formed in the combustion of precursors are
destroyed in the high temperature combustion zone. This implies ensuring
that all furnace gases experience sufficiently high temperatures to destroy
both PCDD/PCDF and potential precursors. Since even trace emissions of PCDD/
PCDF are of concern, both spatial and temporal variations must be considered
in the combustion control scheme so that there are no pathways which at any
4-7
-------
time are not sufficiently hot. The presence of oxygen lowers the required
temperature; ensuring complete mixing at uniformly high temperatures will
ensure low PCDD/PCDF emissions.
4.2 NOX Formation and Control
Air emissions of NOX from municipal waste combustors (MWCs) are highly
variable. Measurements have been made on a number of plants as summarized by
O'Connell et al (1982), Rigo et al (1982), Russel and Roberts (1984) and MRI
(1986). Discussions with manufacturers have further indicated the high
variability of NOX emissions. The reported levels range from less than 0.05
to near 1.0 Ib/MMBTU with average values for all facilities reported at
around 0.25 Ib/MMBTU (130 ppm at 12 percent C02). The sources have indicated
that there is no clear trend in NOX emissions from different types of
municipal waste combustors (MWCs).
For most of the country, there are no NOX emissions standards for
municipal waste combustors (MWCs). Even though there are no national limits,
municipal waste combustion (MWC) facilities will generally have NOX
regulations imposed locally. Manufacturers of facilities indicated that
plants under permitting evaluation in several locations around the country
were having to meet NOX emission limits imposed by either the state or local
environmental agencies.
In the south coast air basin of California (Los Angeles area), for
example, all refuse-to-energy facilities exceeding 50 tons/day are subject to
New Source Review. In addition the district restricts NOX emissions to less
than 225 ppm (-3 percent, 02 dry, 15 min average). This requirement is
expected to dictate special consideration to ensure compliance. The New
Source Review requires a preconstruction review of the new facilities and if
net cumulative emissions exceed 100 Ib/day then facilities must have Best
Available Control Technologies (BACT), emissions offsets and not cause a
violation or make measurably worse an existing violation of a national
ambient standard. Other local districts are also requiring NOX limits on
their own.
4-8
-------
The NOX formed in municipal waste combustion (MWC) facilities occurs by
two separate pathways, thermal fixation of molecular nitrogen ("thermal NOX")
and conversion of nitrogen from the fuel. The thermal fixation mechanism
involves high temperature reactions of free radicals of nitrogen and oxide.
The controlling mechanism have been determined to be Zeldovich reactions of
the form:
0 + N2 = NO + N
N + 02 = NO + 0
These reactions are strongly temperature dependent as shown in Figure 4-3.
This figure represents a computer simulation of the NOX formed for a specific
condition and residence time ( ) and clearly indicates the strong temperature
dependence of thermal NOX formation.
A comparison of the actual NOX levels with those expected by thermal
mechanisms at the maximum temperature expected in the furnace indicates
another NOX source must be operational. For example, less than 100 ppm of
thermal NOX would be expected from thermal processes for conditions which
exist in normal furnace designs. This extra source of NOX could be
attributed in part to the conversion of nitrogen bound in the refuse feed.
This mechanism was first discovered by burning coals and fuel oils in
nitrogen-free oxidizers where the only nitrogen source was in the fuel.
These studies have indicated that the conversion of the fuel-bound nitrogen
is highly dependent on the local oxygen availability to volatile species, the
amount of fuel bound nitrogen and the chemical structure of the nitrogen and
fuel in the refuse. The conversion efficiency can vary from near 50 percent
for highly mixed conditions to near 5 percent for oxygen starved-staged
combustion conditions for coal bound nitrogen (Chen et al , 1982). As shown
in Figure 4-3, the conversion of fuel nitrogen is not expected to be
temperature dependent.
Traditionally, NOX emissions from municipal waste combustion (MWC) have
not been controlled. However, with local air regulations, there may be a
need to control NOX to attain standards. Also, the general tendency to
4-9
-------
10,000
Q_
Q_
1000
100
o 10
1.0
0.1
3140 2813
T(°F)
2509 2310
2112 1941
THERMAL NO
0.5% FUEL N
l
MAX
EXPECT^
ED
ADIA-
BATIC
TEMP.
30% EXCESS AIR
r= 0.5 SEC.
FUEL N
0.45 0.50 0.55 0.60 0.65 0.70
103/T(K~1)
Figure 4-3. Impact of temperature and fuel nitrogen on NOX
emissions for excess air conditions (calculated
using EER kinetic set).
-------
produce higher temperatures and better mixing for PCDD/PCDF control will lead
to higher NOX emissions. Manufacturers of MSW combustion equipment have
indicated that NOX emissions from modern-designed plants will probably be
higher than those from older plants.
Only a few of the potentially applicable control schemes have been
demonstrated at full-scale for MSW combustion devices. In Table 4-1 is
provided a summary of the status of the control schemes for NOX and their
potential impact on control of organic emissions such as PCDD/PCDF. Flue gas
recirculation is a demonstrated technology for thermal NOX which involves the
dilution of combustion air to lower the flame temperature. This scheme has
been demonstrated by Volund as an effective means of controlling NOX in its
refractory-lined high temperature municipal waste combustors. Flue gas
recirculation lowers flame and furnace temperatures which is expected to be
counter to the control requirements for PCDD/PCDF. The significance of the
detrimental impact on PCDD/PCDF emissions due to reduced bulk temperatures
has yet to be determined.
Reburning with an auxiliary fuel such as natural gas is currently being
developed for fossil fuel fired furnaces as a retrofittable combustion
control scheme for NOX. The process could potentially be applied to MSW
municipal waste combustors as shown in Figure 4-4. The SR-j terms in this
figure refer to the stoichiometric ratio with SR < 1.0 representing fuel-rich
conditions. Enough reburning fuel should be injected at a location low in
the furnace to create a hot, slightly oxygen-starved zone. The overfire air
is injected above the reburning zone to complete the combustion process.
Reburning can be combined with urea or ammonia injection to optimize the NOX
reduction. In'addition to NOX reduction, reburning has the potential for
destruction of PCDD/PCDF due to the high temperatures and high concentration
of flame radicals that exist in the reburn zone (Overmoe et al, 1985).
Reburning with natural gas holds the promise of combined NOX and PCDD/PCDF
control but currently is only a concept that has not been tested above bench-
scale.
4-1-1
-------
TABLE 4-1.
STATUS OF N0y CONTROL OPTIONS FOR MUNICIPAL WASTE COMBUSTORS
A
APPROACH
COMPATIBILITY WITH
ORGANIC CONTROL
BENEFITS/
DISADVANTAGES
STATUS
FLUE GAS RECIRC.
DETRIMENTAL
INEXPENSIVE
EFFECTIVE FOR
THERMAL ONLY
DEMONSTRATED FOR SOME
SYSTEMS (VOLUND, VICON)
I
—k
INJ
THERMAL deNO* BY
NH3 INJECTION
NO IMPACT
70-80% EFFECTIVE
NH3 SLIP
FURNACE INJECTION
FULL SCALE INSTALLATION
(COMMERCE A JAPAN)
TESTING UNDERWAY
REBURNING WITH
NATURAL GAS
BENEFICIAL
POTENTIAL PCDD/PCDF CONTROL
50% EFFECTIVE
CONDITIONS NOT WELL
SUITED FOR REBURNING
MAY REQUIRE MODIFICATION
OF AIR FLOWS
CONCEPTUAL ONLY
DEMONSTRATION UNDERWAY
FOR FOSSIL FIRED BOILERS
SELECTIVE CATALYTIC
REDUCTION
NO IMPACT
CATALYSIS POISONING
EXPENSE
90% EFFECTIVE
FULL SCALE
(JAPAN)
INSTALLATION
-------
STOICHIOMETRY RATIOS
(SR)
SR3 = 1.9
SR2 = 0.9
SR1 = 1.1
OVERFIRE AIR
REBURNING FUEL
Figure 4-4.
Application of reburning and de-NOx schemes
for NOX control of mass burn municipal waste
combustors.
4-T3
-------
Other NOX control schemes for municipal waste combustors (MWCs) involve
post-combustion zone control. For example, thermal DeNOx involves the
injection of ammonia in the upper furnace, to achieve selective reduction of
NOX. There are a few examples of the application of this technology in Japan
and recently in the U.S. for municipal waste combustors (MWCs) (e.g., Hurst
and White, 1986). The NH3/NO reactions are extremely sensitive to
temperature so that the injection location must be carefully selected. Also,
there is generally some slip of NH3 which does not completely react that can
cause odors, fouling and the production of visible plume. New tests at full-
scale installations should indicate the viability of this technology for
municipal waste combustors (MWCs) in the near future.
Another post-combustion control scheme involves selective catalytic
reduction (SCR) which enhances the reaction of NO and NH3 to form N£. The
use of a catalyst enables the reactions to take place at lower temperatures
over a broader temperature window and there is little NH3 slip. The process
achieves very high NOX reductions (typically 80 percent). There are numerous
full-scale installations of SCR on oil- and gas-fired boilers principally in
Japan although there is too little experience with SCR with MSW combustion
effluents to know if catalyst poisoning is to be a factor.
In summary, the levels of NOX emissions from MSW combustion facilities
are generally on the order of 100-300 ppm (12 percent C02). However, the
emissions vary widely due to differences in nitrogen content in the raw
refuse and the thermal environment in different municipal waste combustors
(MWCs). Trends to higher temperatures and more uniform mixing for PCDD/PCDF
control are expected to increase NOX emissions in newly designed plants.
State and local regulatory agencies will likely dictate NOX emission
standards in the future which will require the implementation of separate NOX
control schemes. Unfortunately, few control schemes are available which have
been evaluated at full-scale.
4-14
-------
4'3 Partlculate and Trace Metals
Fine fly ash particles are produced as a natural consequence of
combustion of heterogenous refuse. The refuse feed can contain 30 to 50
percent noncombusti bles by mass. Some of the finer ash material can become
entrained in the flow and be carried out of the furnace. Also certain
inorganic compounds present in the burning bed are volatile at combustion
temperature and will leave the combustion zone as a gas before recondensing
downstream. Finally, carbonaeous particulate matter or soot is always
present during the combustion of solid fuels and if it is not burned out it
will also escape the furnace.
A number of measurements have been undertaken on municipal waste
combustion (MWC) facilities and several survey programs are available which
summarize the results on particulate emissions (see e.g., MRI, 1986;
O'Connell et al , 1983; Rigo et al, 1982; and ISWA, 1986). Uncontrolled
particulate loadings have been found to range up to 3 Ib/MMBTU depending on
the refuse and combustor characteristics. There is some evidence that
indicates that starved air units have the lowest uncontrolled emissions while
spreader stoker units are the highest. Control of particulate is necessary
to meet NSPS. The volume entitled "Flue Gas Cleaning Technology" provides a
thorough review of downstream air pollution control devices.
Of particular concern for particle emissions is the emission of heavy
metals in the form of fine particulate. In Table 4-2, a summary is provided
of the typical metal concentrations expected in municipal solid waste. In
Figure 4-5 is shown the transformation of these metals during the combustion
process. The ash included in the refuse can either remain in the bed and be
rejected together with other residuals, be entrained into the flow or be
vaporized and re-condense downstream. The residual matter generally includes
the incombustible fraction of the refuse and such metals as iron, aluminum,
copper and zinc along with the other inerts for example, calcium and silica.
Roughly about 1-3 percent of the incombustible fraction is entrained as fly
ash in the 1-20 micron range. A smaller fraction is vaporizable. Within the
bed, the locally-reducing high temperature conditions can lead to the
4-T5
-------
TABLE 4-2. METALS PRESENT IN MSW
MUNICIPAL SOLID WASTE
ELEMENT
Ag
AT
Ba
Ca
Cd
Co
Cr
Cu
Fe
Hg
K
Li
Mg
Mn
Na
Ni
Pb
Sb
Sn
Zn
AVG.
3
9000
170
9800
9
3
55
350
2300
1.2
1300
2
1600
130
4500
22
330
45
20
780
CONCENTRATION (ppm)
RANGE
<3-7
5400-12000
47-450
5900-17000
2-22
<3-5
20-100
80-900
1000-3500
0.66-1.9
920-1900
<2-7
880-7400
50-240
1800-7400
9-90
110-1500
20-40
<20-40
200-2500
Data from Law and Gordon, 1979
4^16
-------
VOLATILE
INORGANIC
[Na, Zn, Ba, Hg]
INTERNAL
REDUCING
ENVIRONMENT
HETEROGENEOUS
CONDENSATION OR
ADSORPTION
ASH
PARTICLE
I ,
^4
COAGULATION^. •..';•
v> •••'!*
Hg, Pb, Fe, Mg, OXIDATION
CHLORINATION AND VAPORIZATION
... NUCLEATION . TX
(e.g., PbCl )
CHAR PARTICLE
DURING COMBUSTION
ENTRAINED
PARTICLES
RESIDUAL FLY ASH
(1-20 urn)
BURNING
BED OF
SOLID WASTE
FUME
(~0.05 /urn)
RESIDUALS
Figure 4-5. Transformation of mineral matter during combustion of metal containing waste,
-------
formation of oxides and chlorides of metals which can be volatile. Also,
some metals such as sodium, mercury and to a lesser extent, cadmium, have
high vapor pressures as the pure metal. These materials can volatilize and
re-condense either homogeneously as a fume or heterogeneously on the fly ash.
The condensation mechanisms favor the finer particles due to their higher
surface area to mass ratio. Thus, the finer particles that can escape
particulate control devices, i.e., submicron, are enriched in these volatile
species (sometimes more than 100-fold over the refuse incombustibles).
The actual partitioning of the metals among the residuals, fly ash and
fume depends on the waste composition as well as the combustion environment.
The impact of the design and operating variables on the partitioning can not
currently be predicted with confidence. Equilibrium metal s partitioning
analysis are currently being developed under EPA support which will lead to a
better understanding of the fate of metals. However, it is clear that design
and operating conditions defined for PCDD/PCDF control will influence
particulate and metals emissions. For example, higher velocities through the
bed will increase the entrainment of particles. Changes in bed stoichiometry
for proper air distribution will influence the vaporization of volatile
metals. Also, temperature increases will favor vaporization of the metals.
Therefore, the impact of design and operating conditions for PCDD/PCDF
control on the emission of metal enriched fume requires further
investigation.
4.4 Acid Gases
The acid gases of interest as pollutant emissions from MSW combustion
facilities are-S02, HC1, ^04 and HF. The emissions of these species is a
direct function of the amount of elemental sulfur, chloride and fluoride in
the feed refuse. For example, municipal waste has been found to have 0.12
percent sulfur on average and 30-60 percent is converted to S02 (O'Connel et
al, 1983). The balance of the sulfur is retained in the residual ash or is
absorbed on fly ash. In the same manner, roughly half of the chloride is
emitted as HC1 . Most state and local environmental protection agencies are
requiring acid gas control. The control technologies all involve flue gas
4-T8
-------
scrubbing either with dry, semi-dry or wet alkaline sprays. Combustor design
and operating conditions for control of PCDD/PCDF are expected to have no
impact on the emission of or the ability to control acid gas emissions.
4.5 References
Axel rod, D.,. MD, "Lessons Learned from the Transformer Fire at the
Binghampton (NY) State Office Building." Chemosphere 14 (6/7) p.
775-778, 1985.
Bumb, R. R. "Trace Chemistries of Fire: A Source of Chlorinated
Dioxins", Science, 210, 385, 1980.
Chen, S. L., M. P. Heap, D. W. Pershing, and G. B. Martin. "Influence
of Coal Combustion on the Fate of Volatile and Char Nitrogen During
Combustion." Nineteenth Symposium (Int.) on Combustion/The Combustion
Institute Pittsburgh, p. 1271, 1982.
Del linger, B. et al., "Laboratory Determination of High Temperature
Decomposition Behavior of Industrial Organic Materials". Proceedings of
75th APCA Annual meeting, New Orleans, 1982.
Duvall, D. S. and W. A. Rubey, "Laboratory Evaluation of High
Temeprature Destruction of Polychlorinated Biphenyls and Related
Compounds," EPA 600/2-77-228, 1977.
Eichman, G. A. and H. 0. Rhgei. Chemosphere, 11, p. 833, 1982.
Germanus, D. "Hypothesis Explaining the Origin of Chlorinated Dioxins
and Furans in Combustion Effluents." Presented at the Symposium on
Resource Recovery, Hofstra University, Long Island, New York, 1985.
Hites, R. A. and J. B. Howard. "Combustion Research on Characterization
of particulate Organic matter from Flames". EPA-600/7-78-167, 1978.
4-T9
-------
Hurst, B. E. and C. M. White. "Thermal DeNOx: A Commercial Selective
Noncatalytic NOX Reduction Process for Waste to Energy Applications".
Hutzinger, 0., M. J. Blumich, M. V- D. Verg. and K. Olie. "Sources and
Fate of PCDD and PCDFs: An Overview", Chemosphere, 14 (6/7) p. 581,
1985.
Law, S. L. and G. E. Gordon. "Sources of Metals in Municipal
Incinerator Emissions". Environ. Sci. Tech., 13, p. 433, 1979.
Lustenhower, J. W. A., K. Olie, and 0. Hutzinger. "Chlorinated Dibenzo-
p-Dioxin and Related Compounds in Incinerator Effluents. A Review of
Measurements and Mechanisms of Formation," Chemosphere, 9, p. 501, 1980.
Martin, GMbH. Data presented to W. R. Seeker and W. R. Niessen, Munich,
July 1986.
MRI 1986. Emission inventory for EPA, 1986. See Appendix A of the
Comprehensive Report.
Niessen, W. R. "Production of PCDD and PCDF from Resource Recovery
Facilities, Part II", 1984 National Waste Processing Conference ASME
proceedings, p. 358, 1984.
Niessen, W. R. "Dioxin Emissions from Resource Recovery Facilities and
Summary of Health Effects." Report prepared for EPA OSW, 1986.
O'Connell; W. L., G. C. Statler and R. Clark. "Emissions and Emission
Control in Modern Municipal Incinerators". 1982 National Waste
Processing Conference. ASME Proceedings, p. 285, 1982.
Olie, K., M. V. D. Berg, and 0. Hutzinger. "Formation and Fate of PCDD
and PCDF Combustion Processes", Chemosphere 12 (4/5) p. 627, 1983.
4-20
-------
Overmoe, B. J., S. L. Chen, D. W. Pershing, and G. B. Martin.
"Influence of Coal Combustion on the Fate of Volatile and Char Nitrogen
During Combustion." Nineteenth Symposium (Int.) on Comb./The Combustion
Institute, Pittsburgh, p. 1271, 1982.
Rhgei , H. and G. A. Eiceman. "Adsorption and Thermal Reactions of
1,2,3,4-TCDD on Fly Ash from Municipal Incinerator." Chemosphere, 11
(6) p. 569, 1982. See also Chemosphere, 13: p. 421, 1984 and
Chemosphere 14 (3/4) p. 259, 1985.
Rigo, H. G., J. Raschka, S. Worster. "Consolidated Data Base for Waste-
to-Energy Plant Emissions." 1982 National Waste Processing Conference.
ASME Proceedings, p. 305, 1982.
Rubey, W. A., J. Torres, D. Hall, J. L. Graham, B. Dellinger.
"Determination of the Thermal Decomposition Properties of 20 Selected
Hazardous Organic Compounds", EPA Cooperative Agreement Report, 1985.
Russel , S. H. and J. E. Roberts. "Oxides of Nitrogen: Formation and
Control in Resource Recovery Facilities". 1984 National Waste
Processing Conference. ASME Proceedings, p. 441, 1984.
Shaub, W. M. and W. Tsang. "Dioxin Formation in Incinerators". Envir.
Sci. Tech., 17, p. 721, 1983.
Stehl , R. H., R. R. Patenfuss, R. A. Bredeweg, and R. W. Roberts. "The
Stability of Pentachlorophenol and Chlorinated Dioxins to Sunlight, Heat
and Combustion" in Chlorodioxin - Origins and Fate. E. H. Blair, ed.
ACS Washington, D. C., p. 119, 1973.
Vogg, H., M. Metzger, and L. Stieglitz. "Recent Findings on the
Formation and Decomposition of PCDD/PCDF in Solid Municipal Waste
Incineration." Proceedings of Emissions of Trace Organics From
Municipal Solid Waste Incinerators, Specialized Seminar, Part 1,
Session 2, Copenhagen 20-22, January 1987.
4-21
-------
World Health Organization, Regional Office for Europe. March 17-21,
1986. Notes from Meeting on Working Group on Risks to Health from
Dioxins from Incineration of Sewage Sludge and Municipal Waste.
4-2-2
-------
5.0
CURRENT PRACTICES IN MASS BURN TECHNOLOGY
5.1
Mass Burn Technologies
In Figure 5-1 is shown an example of a large, modern, mass burning
facility. The facility consists of the following basic components:
• Waste storage, handling and feeding system (1,2,3)
• Combustion Systems (4,5)
• Boiler and Generators (6,11)
• Air Pollution Control Equipment (7,9,10)
• Residuals Handling (12,13,14)
The major distinguishing features of mass burn technologies is the lack of
virtually any processing of the refuse and the large capacities (>100 TPD).
The combustion system consists generally of a grate on which the solid waste
burns in a layered bed that is progressively moved down the grate. The air
required for combustion of the waste is introduced both underneath the grate
and directed through the bed (referred to as primary air) and sbove the bed
through secondary air jets.
Most of the major organizations which manufacture mass burn technologies
were contacted as part of compiling information presented in this volume.
These included the following:
Manufacturer
American Licensee
Volund (Denmark)
Waste Management
Deutsche Babcock (Germany)
Steinmueller (Germany)
Browning Ferris & American Ref-Fuel
Dravo Energy Resources, Inc.
Von Roll (Switzerland)
Signal Resco
5-1-
-------
1 Waste bunker
2 Crane
3 Charging hopper
4 Grate
5 Combustion chamber
6 Steam boiler
7 Electrostatic precipitator
8 Flue gas fan
9 Wet scrubber
10 Stack
1 1 Turbine-generator
1 2 Fly ash conveying system
1 3 Residue discharging system
14 Residue bunker
1 5 Primary air system with
prehealer
1 6 Secondary air system
un
ro
Longitudinal section of the
Waste power plant Bielefeld-Herford,
Federal Republic of Germany
Combustion capacity: 3 X 385 t/24h
Steam production: 3 X 52.4 t/h
Figure 5-1.
Mass burning waste power plant at Widmer and Frnst
at Bielefeld-Hertford, Germany.
-------
American Licensee
Blount
Odgen Martin
American-Japanese Joint Technology
Manufacturer
Widmer & Ernst (Switzerland'
Martin (Germany)
Riley/Takuma
Detroit Stoker
(Grate Supplier)
Combustion Engineering
(with de Bartolomeis)
Foster Wheeler
(Boiler Supplier)
Babcock and Wilcox
(Boiler Supplier)
Westinghouse/01Conner
Enercon/Vicon
The topics discussed with these manufacturers were their perception on PCDD/
PCDF formation mechanisms, design approaches to prevent PCDD/PCDF formation,
and design and operating guidelines that would be useful for minimizing PCDD/
PCDF emissions.
The manufacturers all indicated that modern mass burn designs were both
capable of and were achieving what they considered to be low PCDD/PCDF
emission levels. The current design philosophy relied predominantly on
optimizing the combustion zone performance by attempting to maximize
combustion efficiency and minimize furnace non-uniformities. Flue gas CO and
62 were treated as an adequate indicator of combustion efficiency. However,
5-3
-------
flue gas concentration of CO was not generally considered to be directly
relatable to PCDD/PCDF emissions but rather was used as an indication that
the systems, once tuned and adjusted properly, were being maintained in the
appropriate operating range.
The key techniques employed to minimize non-uniformities were to
optimize the mixing across the furnace generally using secondary air injected
at high velocities over the grate region, to optimize the grate and waste
feed to obtain uniform bed coverage, and to adjust the combustion zone air
distribution to natch the air with the burning characteristics of the solid
waste. There has been much emphasis on combustion uniformity and techniques
to minimize non-uniformities due to the manufacturers desire to minimize
fire-side corrosion due to reducing zones. To a lesser extent, some
manufacturers expressed the need for sufficient time at temperature or at
least sufficient temperature in the upper furnace region. Also, some
manufacturers indicated a design philosophy of lowering flue gas temperatures
combined with using baghouses to remove particulate matter onto which PCDD/
PCDF and other non-volatile hydrocarbons may have condensed.
Individual designer/manufacturers implement the philosophy in different
ways. The next sections will highlight the approaches followed by some of
the major manufacturers of large mass burn waste-to-energy facilities. It
should be indicated that many of the designs originated in Europe where
waste-to-energy plants are much more common and the concern for PCDD/PCDF
emissions has been factored into the designs for some time. The resulting
technologies employ very sophisticated combustion systems and controls along
with utility type boiler furnaces. Compared to other types of municipal
waste combustion technologies, modern mass burn waste-to-energy systems can
be considered to be second or even third generation in terms of designs and
operation to minimize PCDD/PCDF emissions. The systems described here
represent the most current practice that has been pursued for combustion
control of PCDDs/PCDFs. However, it must be pointed out that these
descriptions are current design practices and are not indicative of all
systems currently operating in the United States.
5-4-
-------
5.2 Deutsche Babcock Anlagen
The modern Deutsche Babcock Anlagen (DBA) mass burn technology is the
result of a significant amount of research and testing of different system
configurations. The design features of the DBA mass burn system are
highlighted in Figure 5-2 and Table 5-1. One unique feature of the DBA
system is the use of a roller grate that carries the refuse down to the next
roller grate. Stirring actions occur at the transition between the rollers
where the burning refuse can tumble and expose new combustible material.
DBA did not have any direct cause and effect data for PCDD/PCDF
emissions in their design but did have evidence of low emissions with the
current design and operating approach. Measurements are available on
seventeen DBA plants primarily in Germany for CDD/CDF analysis on ESP dust,
residuals, scrubber water and flue gas (DBA, 1986). These data indicate flue
gas emissions of less than 10~3 ng/m3 of 2,3,7,8 TCDD for many of the plants.
The average total CDD/CDF emission level was 22.5 ng/m3 for five completely
tested facilities.
The DBA design philosophy was oriented towards optimizing combustion
efficiency and maximizing uniformity of mixing. The general design criteria
included hot combustion temperature, exhaust CO levels of less than 100 mg/
Mm3, high excess air levels (9-10 percent 02) and prevention of a reducing
environment above the secondary air injection point.
A key element of the DBA approach is the use of overfire or secondary
air injection schemes which serve to mix all of the furnace gases above the
grate. DBA has performed a large number of experiments on different
configurations in order to find optimum furnace configurations and injection
schemes for overfire air. Experiments have included both cold-flow furnace
modeling (water and air) and field in-furnace measurements. The newest
design has overfire air injection at a nose in the furnace in which the
effluent from the grate is pinched to allow a smaller mixing distance for
overfire air. The DBA design uses 25 percent of the air as overfire air and
injects it at 100 m/sec. This results in a penetration depth of the overfire
5-5
-------
OLD GEOMETRY
NEW GEOMETRY
I
O
Figure 5-2. Deutsche Babcock Anlagen mass burn furnace design
features.
-------
TABLE 5-1. DESIGN FEATURES OF DEUTSCHE BIBCOCK
ANLAGEN SYSTEMS
• ROLLER GRATE (1)
a FURNACE NOSES (6) FOR MINIMAL
SECONDARY JET PENETRATION
8 SECONDARY AIR (3)
- 20-25% OF TOTAL AIR
- 100m/sec VELOCITY
- 70% PENETRATION ACROSS
FURNACE
I SIDE WALL AIR FLOW (4)
(ASPIRATED WALL)
I TIME AT TEMPERATURE ABOVE
SECONDARY INJECTION
- TA LUFT STANDARD
- SIC CLADDING OF LOWER
FURNACE (5) END OF
FLAME TIP.
5-7
-------
air jets to approximately 70 percent of the distance across the furnace nose.
DBA indicated that the overfire air requirements as suggested in previous EPA
recommendations of 50 percent of total air flow was felt to be inappropriate.
Such high levels would result in poor burnout. DBA expressed the opinion
that operation with 20 to 25 percent design air flow capacity air flow as
overfire air was appropriate.
DBA has examined the impact of the orientation of the furnace throat
relative to the grate and has developed specialized furnace geometries for
different refuse characteristics. Three of the configurations are portrayed
in Figure 5-3. The configuration names refer to the relative direction of
gas flow over the bed to the direction of movement of the solid waste. For
example, in the parallel flow configuration, the volatiles released from the
thermal decomposition of the solid move in the same direction as the solids
due to the presence of a hood or arch over the early region of the grate. In
the contra flow configuration, the gas flow is generally in the direction
opposite to the direction or movement of solids. The configuration which is
considered to be the most flexible is the center flow arrangement which is
between the other two extremes. In this configuration the zone of volatile
thermal decomposition is arranged just below the furnace throat. The
volatile flame can freely develop and overfire air can be added to uniformly
mix the material as it enters the radiant furnace region. DBA recommends
contra flow configuration for refuse of low calorific value i.e. refuse
containing high amounts of water and ash due to the "high pre-drying effects
and preparation of the refuse for ignition by recycling hot gas across the
first rollers". Parallel flow configurations are reserved for special high
volatile MSW or installations that have stringent space limitations. Such
special configurations are currently under construction at sites in Germany
(e.g. Duesseldorf).
In the DBA design the lower portion of the municipal waste combustor is
refractory-clad. The insulation consists of silicon carbide that is directly
attached onto the water tubes. The SIC refractory extends up beyond the nose
to halfway up the furnace itself. This ensures that sufficient time at
temperature is achieved, such that the DBA designs can meet the current
5-8
-------
I. PARALLEL FLOW
O
II. CONTRA FLOW
III. CENTRE FLOW
I SPECIAL REFUSE
(HIGH V.M.)
II LOW CALORIFIC
REFUSE (HIGH
WATER OR ASH)
III HIGHEST FLEXIBILITY
Figure 5-3. Deutsche Babcock furnace geometry selected based on refuse characteristics (DBA, 1986)
-------
German requirements of temperature above the last air injection point
(800°C). DBA has examined the previous EPA criteria suggested for
"qualifying maximum volumetric heat release" rate and has concluded that the
definition of the lower furnace volume is difficult and portions of the
volume do not actively participate in the combustion process. The DBA design
has a heat absorption rate in the lower furnace region of approximately half
of the heat release rate in the upper furnace region. The upper furnace heat
release absorption is roughly 48 kcal/m2-
To monitor furnace operation, DBA uses flue gas measurements of carbon
monoxide and oxygen. Carbon monoxide is maintained below the German TA Luft
standard of 100 mg/Nm3 (~80 ppm)and oxygen is generally between 9-10 percent
(dry). The stack gas oxygen concentration range suggested by the DBA
operating experience is shown in Figure 5-4. The DBA system, as any other
mass burn systems, has an optimum operating envelope of excess air as shown
in Figure 5-4 as the conditio'n of minimum carbon monoxide. Excess oxygen
must be maintained at levels that ensure that all zones within the furnace
have sufficient oxygen even with fluctuations in the volatile content of the
refuse. However, too high of excess oxygen will excessively cool the
combustion zone due to dilution. Finally, it is also a DBA operating
practice to only operate at full design load with only slight variations
above and below the design conditions.
5.3 Steinmueller
The L&C Steinmueller Corporation is providing the combustion system to
Dravo, Inc. who is constructing two systems in the United States in Portland,
Maine (500 tpd)'and Long Beach, California (1380 tpd). Two other systems are
currently in the permit stage in Montgomery County, Pennsylvania (1200 tpd),
and Huntsville, Alabama (690 tpd).
The personnel at Steinmueller did not have additional information on
direct cause and effect relationships between PCDD/PCDF formation and system
design and operation. However, their current practice was found to have
average 2,3,7,8 TCDD emissions of less than 0.02 ng/m3 for several plants in
5-10-
-------
x>
CD)
Zi
oct
OC_>
CD 2:
ceo
CCCJ
CJ
3 6 9
OXYGEN CONCENTRATION
A - INSUFFICIENT AIR C+i02-*CO
B - APPROPRIATE OPERATING REGION
C - "COLD BURNING"
Figure 5-4.
Relationships of CO and 0^ for
appropriate operating regions
(DBA, 1986).
5-11
-------
Germany (Stei nmuel 1er, 1986). Total CDD levels were found to be on average
161 ng/m3. The Steinmueller approach primarily relied on an optimized firing
design with high combustion efficiency and uniform mixing condition. The gas
cleanup system is also utilized by Steinmueller as a backup. Either dry
scrubbers with baghouses or wet scrubbers with ESPs were used for the
Steinmueller installations in Germany. There was concern expressed about
whether baghouses with municipal waste combustors was the best application of
the technology. For their U.S. projects, Steinmueller/Dravo is using dry
scrubbers with either an ESP or a baghouse.
The Steinmueller firing system is portrayed schematically in Figure 5-5
and the key features are provided in Table 5-2. The design employs dual ram
feeders onto a forward-push block grate. The grate blocks move in a
reciprocating motion and push the refuse down the inclined grate. The
radiant furnace was constrained with front and rear wall noses and was
centered over the grate. The lower furnace was clad with silicon carbide
refractory for corrosion control and to lower heat absorption rates.
A key aspect of the Steinmueller design philosophy is to achieve uniform
high temperatures within the combustion volume. Their current design is
capable of meeting the German standard of 800°C above the last air injection
location. In order to achieve the temperature, the refractory cladding is
added to lower the heat absorption rate. Even more importantly, the
Steinmueller systems use an optimized furnace configuration and overfire air
injection scheme to ensure mixing of volatiles with air and to maintain a
high temperature combustion process. The front nose acts to redirect the
volatiles into the hot gases from the burnout portion of the grate. Numerous
high velocity secondary air jets at the furnace throat are then used to
ensure uniform mixing before entering the radiant furnace. The Steinmueller
design employs 80 mm diameter secondary jets injected at a velocity of 80 m/
sec with a pressure drop of 600 mm of water- The distribution of air was as
follows: 60 percent primary, 40 percent secondary, with an operating range
of 80 to 90 percent total excess air. This ratio was maintained at all
loads. Steinmueller identified as a key problem as the operation of the
5-12
-------
I I
Figure 5-5.
The L&C Steinmueller mass burn
design features.
5-13
-------
TABLE 5-2. DESIGN FEATURES OF STEINMUEILER MASS BURN SYSTEMS
• FORWARD PUSH BLOCK GRATE (2)
I CENTER FLOW FURNACE WITH THROAT
I SECONDARY AIR (A)
- 80 mm DIA.
- VELOCITY 80m/sec
- PRESSURE DROP ~ 600 mm w.g.
- 40% OF TOTAL AIR
a CLADDING REFACTORY ON LOWER FURNACE (SIC)
I SEPARATE.PLENUM CONTROL OF PRIMARY AIR (3)
I CONTROLLED AND UNIFORM FUEL BED DEPTH
• FIVE ADJUSTABLE UNDERFIRE AIR' ZONES AND
GRATE SECTIONS TO CONTROL BURNING RATE
5-14
-------
system at low loads where much poorer mixing was found. Under low load
conditions Steinmueller has measured higher CO and oxygen levels.
Steinmueller has relied heavily on extensive cold-flow modeling and in-
furnace field measurements to optimize the design and operation of their
systems in order to ensure high temperatures and uniform mixing conditions.
In Figure 5-6 are provided some representative examples of in-furnace
measurements which illustrate the impact of design and operational changes.
In Figure 5-6a is shown the detrimental impact of lower load operation. At
full load the temperatures in the radiant furnace are between 800 and 1000°C
(1470-1830°F) and the plane of CO concentration at 0.1 percent is uniformly
spread across the furnace. The furnace CO will continue to oxidize resulting
in stack emissions of less than 80 ppm. At lower load operation the upper
furnace mixing is much less uniform and the temperatures have dropped by
almost 100°C (180°F). The surface of constant CO concentration is shown to
be strongly skewed to the front wall of the furnace indicating poor mixing at
this reduced load condition. Also, exhaust CO will increase to 196 ppm with
spikes as high as 660 ppm. In Figure 5-6b are shown cold flow modeling
results on the impact of different combustion air distributions on the mixing
patterns. A mixing parameter ("mischparameter") of unity across the furnace
indicates a high degree of mixing since all parts of the flow field have an
equal concentration of the gas tracer. Uith normal air distributions through
the various underfire plenums and overfire air jets, the furnace uniformity
is excellent. In the extreme case of no overfire air, the mixing is
extremely poor indicating the important role that the secondary air plays in
mixing.
Figure 5--6b also indicates the air control capabilities of the
Steinmueller system. The operator has control of the underfire air to five
separate underfire air plenums and separately to overfire air jets on the
front and rear walls. In this manner the operator has the ability to put the
air where combustion is occurring depending on the particular refuse
characteristics. The system must be "tuned" to the refuse characteristics at
startup of the facility; Steinmueller relies on in-furnace profiles of carbon
monoxide as the indicator of the "mixedness" condition at unit startup. High
5-15
-------
(a) Impact of Load
(b) Impact of Overfire Air
FULL
LOAD
NORMAL
OVERFIRE
AIR
• Y/a-o.23
D Y/B-0.3
X Y/B-0.73
I9X 33X I6X 23X I9X
en
I
60%
LOAD
24. BX
,-S -,
QH 3
5*2
55,
a _
™ I
*0
$
a
$2
M
S1
xo
s
^!fel^>
i
*;-«^
•*+
rmi
NO OVERFIRE AIR
LESS FURNACE
CONSTRICTION
Y/B-O.23
D Y/B-O.5
X Y/B-O.73
O ARBITRARY
I I .4X23.7X17. IX 5.7X 2 . 9X
Figure 5-6. Steinmueller in-furnace testing and cold-flow modeling.
-------
CO peaks in the furnace indicates that insufficient air is being put at that
certain point. Thus, Steinmuel1er operational success relies on the
combustion air scheme as follows:
• High velocity, multiple jets of overfire air for furnace mixing
0 Multiple underfire plenums for air control with independently
controlled stoker sections (speed)
• Tuning of air flows using CO profiling as indicator of unmixedness
0 Even distribution of air in the fuel bed
Finally, Steinmueller uses carbon monoxide continuous monitoring in the flue
gas as an indicator that combustion efficiency is being maintained at a high
level. Steinmueller reports that their systems generally have no trouble
meeting the TA Luft (German) CO standard of 100 mg/Nm3 (11 percent 02) on a
30 minute rolling average.
5.4 Von Roll
Von Roll is one of the larger manufacturers of municipal and industrial
combustion plants in the world with more than 170 plants either operating or
under construction on five different continents. Signal Resco has the
American license for this Swiss (Zurich) technology and currently has a
number of plants in operation or under construction in the United States.
Signal Resco uses special Babcock and Wilcox boiler designs integrated with
the Von Roll grate.
The Von Roll philosophy to prevent PCDD/PCDF and other trace organic
emissions is to optimize the refuse combustion system. Von Roll has little
cause and effect data available but they do have performance data which
indicates that the Von Roll system can be operated with low PCDD/PCDF
emissions. Test data supplied by Von Roll (1986) for the Neustadt MSW
5-T7
-------
combustion plant (225 TPD) indicated total PPDD/PCDF of 88 ng/Nm3 before the
air pollution control device and 13.8 ng/Nm3 in the exhaust.
The Von Roll personnel were concerned about not only the possibility of
in-furnace formation mechanisms but also downstream mechanisms that might
occur as a result of catalytic reaction of hydrocarbon precursors on fly ash
particles. They indicated that some recent work by Vogg (1986) demonstrated
that catalytic transformation could occur in the temperature range of 200 to
300°C (390-570°F). Yon Roll expressed the concern that such downstream
mechanisms could lead to the presence of PCDD/PCDF on residual fly ash. The
Von Roll approach is to design and operate the combustion system such that
the destruction of all organics i.s achieved. If all hydrocarbons are
destroyed in the combustion zone then even downstream mechanisms are
prevented due to the lack of precursors to be reacted.
The Von Roll criteria for a good combustion environment are as follows:
1. Uniform bed layer on grate
2. Proper air distribution through the bed including very high
pressure drops across the grate and multiple plenums
3. Load following and control procedures
4. Avoid slag buildup on the side walls (they use an air aspirated
side wall design in Europe although not in the United States)
5. High injection velocity for the secondary air with numerous air
jets
6. The amount of secondary air is held constant at 30 percent of the
total combustion air
The features of the Von Roll design are provided in Figure 5-7 and Table
5-3. Von Roll believes that one of the most important features of the design
5-18
-------
in
i
Figure 5-7. Refuse combustion plant with Von Roll two-pass boiler and flue gas scrubber
-------
TABLE 5-3. DESIGN FEATURES OF VON ROLL MASS BURN SYSTEMS
t GRATE SYSTEM (5)
- HIGH PRESSURE DROP
- PUSH BLOCK
- SELF CLEANING SLOTS
• CENTER FLOW FURNACE
I PRIMARY AIR
-2x5 PLENUM
- SEPARATE CONTROL
• SECONDARY AIR
- 30% TOTAL AIR
- 50 m/sec
- 5 cm dia WITH 1 m
SEPARATION
• ASPIRATED AIR
SIDE WALL (8)
I SiC CLADDING IN
LOWER FURNACES
• ESP PARTICULATE
CONTROL
5-20
-------
is the grate (Figure 5-8). The grate is designed to have a high air pressure
drop. Von Roll feels that with this feature uniform air distribution through
the grate can be achieved even with variable thicknesses of refuse on the
bed. The underfire air is divided into four to six plenums along the grate
with separate air feed control to each region. Larger systems will have two
plenums side by side. The air passes through air slots that are self-cleaned
by the movement of the grate blocks. The grates are forward-push
reciprocating blocks with a steep inclination angle. For lower calorific
refuse a modified grate is utilized wherein steps are present along the grate
to ensure breakup of clumps of refuse.
The second feature of the Von Roll system that is important is the
mixing level within the furnace. Overfire air jets are used as the primary
means to ensure adequate furnace mixing. The penetration of the secondary
air jets across the furnace which is dependent upon velocity and size of the
jets, is crucial to the successful performance of the system. Von Roll has
performed full-scale tests in order to optimize the design and operation of
the overfire jets. The current Von Roll design is provided in Table 5-3
and the orientation of the jets relative to the furnace nose is shown in
Figure 5-7. An array of complex injection schemes has been developed with
multiple rows and nozzle diameter to achieve adequate penetration and
coverage of the flow. In-furnace profiling of carbon monoxide concentrations
performed during system start-up is used as the indicator of proper air
distribution and mixing. The various air flows are adjusted to minimize CO
peaks. The CO profiling and air adjustment is performed both at unit startup
and annually. Von Roll relies on exhaust CO concentration measurements as an
indicator that the system, once tuned, is being maintained in the proper
operating envelope. However, Von Roll does not use exhaust CO as an
indicator of trace organic or PCDD/PCDF emissions. According to Von Roll,
the typical operating range for CO is in the range of 50 ppm.
The final important feature of the Von Roll system is the combustion
control system. All waste-to-energy systems must have combustion controls
that respond to changes in steam demand and account for the variability of
the fuel characteristics of the refuse. In the Von Roll system the steam
5-21
-------
Von Roll refuse feeding device
Von Roll combustion grate system
Figure 5-8. Details of Von Roll grate and feeding devices,
5-22
-------
production rate is monitored and control of the ram feeder frequency is
modulated along with the primary air to the middle region of the grate
(burning region) to maintain the correct steam rate. Von Roll systems also
include furnace temperature monitoring using thermocouples in the roof of the
radiant furnace (previously corrected for radiation by comparison to suction
pyrometry measurements). Furnace temperatures are used to control the
secondary air flow rates. For example, if steam production rate goes down,
then primary air is increased in the middle grate region; if temperature goes
down, then secondary air flow is decreased. The grate movement rate is not
automatically controlled but is manually adjusted depending on the refuse
burnout characteristics. Preheat air is also started manually for wet
refuse. Finally exhaust oxygen is measured but is not used in the automatic
control system.
The Von Roll system also uses automatic control for auxiliary burners.
Secondary burners will automatically fire when CO levels exceed 80 mg/Nm^.
It is a TA Luft (German) regulation that there is capacity for 60 percent of
the load as auxiliary fuel with startup at high CO and during system startup.
5.5 W+E Environmental Systems Ltd.
W+E Environmental Systems Ltd. (Widmer and Ernst) is a wholly-owned
subsidiary of Blount Inc. (Montgomery, Alabama) and is located in Zurich,
Switzerland. The W+E design is portrayed in Figure 5-9 and details of the
design are provided in Table 5-4. The W+E design has a unique grate. The
grate is horizontal with the reciprocating blocks pushing the refuse over the
next block in what is termed an "overthrust" motion. The double motion
overthrust tends to first drop ignited particles as the block moves out from
under the layer and then pushes the ignited particles back underneath the
non-burning waste layer. The ignited material is constantly pushed downwards
by intensely rotating and stirring the waste layer and forcing ignition to
start at the bottom of the bed. The air slot design causes a high pressure
drop across the grate and ensures uniform air flow. High air velocities at
the slots also prevents blockage of the slots. The relative movement of
blocks acts to continually clean the grates and acts to keep them open. The
-------
r-6
1_
Motion sequence
of the grate bars
Figure 5-9. W+E combustion system and overthrust grate
design.
5-24
-------
TABLE 5-4. DESIGN FEATURES OF W+E MASS FIRED SYSTEM
I GRATE (3)
- HORIZONTAL
- "OVERTHRUST" MOTION
- SLOT AIR - SELF CLEANING
I CONTRA FLOW FURNACE WITH MINIMAL
CONSTRICTION
• PRIMARY AIR (9)
- SEPARATE PLENUM CONTROL
- HIGH PRESSURE DROP
- 80 - 90% OF PRIMARY TO FIRST
2 ZONES
8 SECONDARY AIR INJECTORS (10)
- 80 m/sec
- 70 mm dia
- FRONT AND BACK WALLS INTERLACED
- 30% OF TOTAL AIR
- OFFSET VORTEX
• SIC CLADDING TO 10 m
5-25
-------
underfire (primary) air is introduced through four to five separately
controlled plenums along the grate.
Another unique design feature of the W+E system is the overfi re air
injection scheme. As with most of the other mass burn systems the overfire
air injection is primarily used for furnace mixing and flame height control.
High velocity air jets on the front and rear walls are employed to achieve
jet penetration across the furnace and coverage of the entire furnace flow.
The front and rear wall jets are not directly opposed but rather they are
staggered to achieve an "interlacing" of the air streams. In addition, some
configurations include directing the jets to a firing circle which creates a
vortex motion in the region adjacent to the overfire air jets. W+E has
suggested that this rolling vortex zone arrangement lowers particulate
emission apparently because it serves as a "centrifugal bottle" to particle
carryover- W+E has also examined the impact of furnace design such as
furnace noses at the overfire air injection point. No enhancement in mixing
was achieved with these noses alone; however, the noses provided less
penetration differences for the overfire air jets.
Thirty percent of the total combustion air was used in the overfire
injection with an 80 to 100 percent overall excess air. The overfire air
velocity has recently been increased from 50-60 m/sec to approximately 80 m/
sec (at full load conditions). This change was introduced to improve low
load operation where overfire air jet velocities will decline. W+E indicated
a belief that such overfire air design improvements have direct impacts on
PCDD/PCDF emissions. W+E quoted test results which indicated a five-fold
decrease in PCDD/PCDF emission levels with these modifications with no
measurable change in CO emissions (W+E, 1986).
The proper air distribution into underfire plenums and overfire air jets
is established at startup by in-furnace CO profiling and air adjustments to
minimize CO concentration peaks. Once the system is tuned to the
characteristics of the refuse, then W+E relies on exhaust measurements of CO
as an indicator of continuing performance. Under optimal conditions CO
levels as low as 20-30 ppm could be achieved. Based on W+E experience, CO is
5-26
-------
used only as an indicator of insufficient air or bad combustion conditions.
For exhaust concentration of CO less than 250 ppm, W+E believes that no
correlation exists between exhaust CO and PCDD/PCDF emission. However, some
correlation was suggested to exist if CO was greater than 250 ppm, i.e. if
the system was clearly being operated in a failure mode. W+E has some
experience with the use of total organic carbon as an indicator of PCDD/PCDF
emission. W+E suggests that total organic carbon is directly related to
PCDD/PCDF emissions.
The current W+E furnace configuration (see Figure 5-9) could be
classified as a "contra flow" arrangement (compare to Figure 5-3). W+E are
currently performing detailed studies of the necessity of such design
features. Specifically, W+E are investigating all three types of
configurations, parallel, center and contra, to determine whether such design
modifications will improve performance and the cost impacts.
The W+E facilities employ an automatic load control system. The current
systems have both control and upset loops. The primary control loop monitors
the steam production rate and controls the ram feeder, grate speed and air
flows. Oxygen is monitored and if the value falls below a set value
(typically 6 percent) alarms will sound and feeding of the grate will be
stopped. If furnace temperature falls below 800°C then auxiliary burners
will be started automatically. Finally if load falls below 60 percent of
design, then the system will shut down. Future control loops will focus on
the use of furnace temperature monitoring and control of primary air. This
new emphasis is to allow extensions to lower load operation.
5.6 Martin GmbH
Martin GmbH (Germany) with its partners Ogden Martin (Paramus, NJ) and
Mitsubishi Heavy Industries (Japan) have constructed 130 refuse burning and
energy recovery facilities. This includes 250 operating units with a
combined burning capacity of over 75,000 tons/day. Martin Systems is one of
the largest manufacturers of mass burn municipal solid waste-to-energy
plants. Martin systems have been tested for PCDD and PCDF emissions in both
5-2-7
-------
Europe (e.g. Wurzburg, Stockholm and Munich plants) and the United States
(e.g. Chicago, Marion County and Tulsa). Some of these data are available in
the comprehensive municipal waste study series. The total PCDD/PCDF
emissions vary from 50 to 150 ng/Nm3 for systems with ESP particulate control
and less than 5 ng/Nm^ for the system with a dry lime dust baghouse
combination (Martin, 1986).
The design features of Martin refuse combustors are shown schematically
in Figure 5-10 and design details are provided in Table 5-5. The Martin
design philosophy for the minimization of trace organic emissions is
described in Martin technical literature (Martin and Schetter, 1986) and
relies on optimization of the combustion process. The Martin approach to
minimize organic emissions focuses on the prevention of any hydrocarbons
leaving the combustion zone and in that manner minimize downstream formation
of PCDD/PCDF by eliminating PCDD/PCDF precursors. The approach is based not
on one feature alone but rath'er to the combined aspects of the Martin
combustion system.
One of the main features of the Martin system is the grate shown in
detail in Figure 5-10. The grate is inclined downwards from the feed end at
an angle of 26 degrees. The grate blocks are "reverse acting" i.e. they are
pushing counter to the direction of overall refuse movement. This action
forces burning refuse back underneath freshly fed material and promotes more
complete burnout.
The grate underfire air is introduced through gaps at the sides of the
head of grate blocks which are 2 mm wide. The gap area is small enough so
that the grate-has a high pressure drop and therefore there is uniform air
distribution through the refuse layer regardless of the refuse bed thickness.
The underfire air is divided into 5 to 6 zones along the length of the grate
through the use of separately controlled plenums. For larger systems, the
air plenums are doubled with side-by-side plenums. The underfire air to each
plenum is individually controlled by dampers and flow orifices for each grate
section.
5-Z8
-------
© Initial pnase of reluse drying and
volatiliuuton, by means of flame
radiauon
-------
TABLE 5-5. DESIGN FEATURES OF MARTIN REFUSE COMBUSTORS
I GRATE
- REVERSE ACTING
- AIR SLOT, SELF CLEANING
- STEEP INCLINATION
• CONTRA FLOW FURNACE WITH
MINIMAL CONSTRICTION
• PRIMARY AIR
- INDIVIDUAL PLENUMS
- HIGH GRATE PRESSURE DROP
I SECONDARY AIR
- 20-40% OF TOTALAIR
- FRONT AND BACK WALLS
- 2-4" (50-100mm) DIA
- 450 mm w.g.
• SIC CLADDING OF LOWER FURNACE
5-50
-------
An important aspect of the Martin approach is the "penetration of air
into all volatilization products" and the use of secondary air nozzles to
"increase the degree of turbulence in the flame area". The current secondary
air nozzle design should actually show in Figure 5-10 two rows of air nozzles
on the front wall, below the front arch, for refuse with higher volatile
content. The adaptation of the secondary air injection scheme from older
designs burning lower volatile refuse to designs for higher volatile matter
is shown in Figure 5-11. The amount of overfire air is between 20 and 40
percent of the total combustion air (100 percent excess air).
Another important aspect of the Martin approach designed for low
emissions are combustion control systems which attempt to reduce the
disturbances resulting from changes in refuse characteristics and limits load
variations. New Martin facilities have automatic combustion control systems
which consist of two independent loops (see Figure 5-12). The first loop
monitors wet flue gas 02 (Zirconium oxide probe) and controls the refuse ram
feeder and grate speed to control MSW feed rate. The second control loop
monitors steam production rates and controls underfire to maintain desired
steam production rate. Development work is continuing on the combustion
control system including monitoring on furnace temperature and controlling
secondary air. Finally the Martin combustion control approach is to limit
load variations to 85-110 percent of design full load.
The newest Martin design also employs air pollution control devices
(APCD) as a removal technique for PCDD/PCDF species. For example, Martin
plants at Wurzburg, Stockholm and Marion County, Oregon have dry scrubber/
baghouse combinations. This APCD technology has been operating successfully
for three years and available data indicates significant reduction in PCDD
and PCDF emissions. For example, total PCDD/PCDF emissions data from
Stockholm at the baghouse inlet were 74 ng/Nm3 while after the baghouse, the
emissions were less than 5 ng/Nm3 (Martin, 1986).
5-31
-------
Furnoct "QS found'
Configuration at lirrn o<
Isl meosucment series
Configuration at limt of
2nd measurement series
to
ro
Front wall overfire air
Numb* of nozzki It
Anglt of mdnaUjn lo horizontal -70*
NozzU moulh oVyrxlif (Omm
of nonkj.
7ond 6
Anglt of nchnalcn to horizorid -IS*
Malik mautfi damtUf 45and SSrrvn
of nozzlfS
Angle of ncfnofon U> horamtal-
NoizJ« moulh
-------
en
U)
Feed woltr
{[J Conliol ol rilust throughput
(T) Conltol of pflrrxvy air
Figure 5-12. Martin Combustion Control System for mass burn
combustors.
-------
5.7 Volund
Volund, located in Denmark, has an installed capacity of over 29,000 T/d
in 129 operating plants in Scandanavia, Europe, Japan and the orient, and in
the United States. In the United States, Waste Management has the technology
license and has a new operating plant in Tampa using the Volund system.
Volund provides a combustion system that is distinct from other European mass
burn technology- Two separate concepts are available depending on the refuse
characteristics as shown in Figure 5-13. For difficult to burn materials
such as vegetables, coarse pieces of wood, and high water content refuse, a
rotary kiln is added at the exit of the traditional grate region. In
addition, the lower furnace is constructed of high grade, low conductance
refractory not SiC clad water tubes as in other European designs. The
combustion zone also includes a refractory arch which tends to keep radiation
heat losses from the burning zone to a minimum. Hence, the combustion zone
temperature is somewhat elevated over waterwall systems. For these reasons,
Volund Systems can achieve impressive solids burnout levels.
The features of the Volund System are provided in Table 5-6. The Volund
grate is a forward-push, tilted design with longitudinal fixed and movable
beams. Underfire air is introduced between the beams and is controlled by
three separate plenums. The Volund System has a relatively low air pressure
drop across the grate and relies on the even distribution of refuse on the
grate to ensure uniform air flow. The secondary air is added into the
primary zone over the grate region using overfire air jets and aspirated side
walls.
Volund System designers indicated that good mixing in the various zones
is a key factor to controlling the emissions of trace organics. Volund's
approach, after extensive water-table flow modeling studies, has been to
position the refractory arch above the grate which splits and guides the
buoyantly rising gases such as to stimulate mixing. The Volund furnace is
oriented above the discharge area of the grate. Mixing is augmented by the
introduction of overfire air but not dependent upon it.
5-34
-------
Figure 5-13. Volund System mass burn design features,
5-35
-------
TABLE 5-6. DESIGN FEATURES OF VOLUND MASS BURN SYSTEMS
I GRATE
- FORWARD PUSH
- GRATE BARS
- AIR SLOTS BETWEEN BEAMS
• PRIMARY AIR
- LOW PRESSURE DROP AC'ROSS GRATE
- MULTIPLE PLENUM CONTROL, GENERALLY 4
I SECONDARY AIR
- INTO PRIMARY COMBUSTION ZONE
l EXCESS AIR 80-120%
I FURNACE CONFIGURATION
- BLOCK REFRACTORY LINED LOWER FURNACE
- REFRACTORY ARCH WALL
- ASPIRATED AIR SIDE WALL
- UPPER FURNACE WATER WALL
i APCD CONCEPT
- FGR FOR NOX
- ESP FOR PARTICULATE
- HC1 WET, SEMIWET, DRY SYSTEMS
• TIME AT TEMPERATURE
2.2 SECONDS > 1000°C
5-56
-------
The higher temperatures achievable in the Volund System could
potentially increase NOX formation. Volund can incorporate a flue gas
reci rcul ation technique to moderate temperatures in the combustion zone to
effectively reduce thermal NOX. For particulate and acid gas removal, Volund
offers a variety of air pollution control device configurations. These
configurations include wet scrubbers alone with flue gas reheat, semi-dry
(spray dryers) with either ESP or baghouses, and dry injection of calcium
compounds again with either baghouses or ESP.
Volund has undertaken a series of emission tests at one of the
representative plants to examine the impact of startup and load changes and
the relationships between temperature and carbon monoxide to PCDD/PCDF
emissions. These data clearly indicated lowering of 2,3,7,8 TCDD emissions
as time progressed after startup; almost an order of magnitude lower 2,3,7,8
TCDD emissions were achieved in 10 hours after startup versus 1 hour even
though the average temperature increased only slightly in the same period.
For example, the emissions of 2,3,7,8 TCDD dropped from 2.5 ng/Nm3 at 1 hour
after startup to 0.4 at 5 hours and 0.33 ng/Nm3 at 10 hours after startup
(Volund 1986). However the minimum temperature measured did increase
significantly (~150°C, 270°F) in the same period indicating the minimum
temperature pathway must be considered in PCDD/PCDF destruction. With
decreasing load to 70 percent, the 2378 TCDD and total PCDD/PCDF increased by
a factor of 2 (from 0.17 ng/Nm3 to 0.39 ng/Nm3 for 2,3,7,8 TCDD). These data
have also indicated to Volund that CO in the flue gas of a well designed
plant can be used as a measure for monitoring and control since it closely
tracked PCDD/PCDF emissions. For example, just after ignition, CO levels
were near 1000 ppm without auxiliary fuel start-up while after 10 hours
exhaust CO were-below 100 ppm (Volund, 1986).
5.8 Riley/Takuma
The Riley Stoker Corporation is a major United States manufacturer of
boilers and combustion hardware. Their first major entry into the MSW area
was through the facility at Braintree, Mass, which began operation in 1971.
That facility incorporated a Riley boiler as well as a Riley horizontal
5-37
-------
traveling grate. That grate design was a direct adaptation of the Riley
stokers for coal and wood-fired boilers. Turner (1982) provides a thorough
review of the Braintree facility's operating history.
Riley Stoker is currently involved in at least five (5) MSW projects
(Riley, 1986). Two of these projects involve use of Riley Stoker boilers
coupled with grates from Ogden/Martin. The general design and operating
features of facilities using Ogden/Martin grates were discussed earlier in
this chapter. Another project - Jackson County, Mich. - represent the
combination of a Riley boiler and the Takuma grate. Riley has licensed the
Takuma grate which is a Japanese developed design.
The Riley/Takuma approach to minimization of PCDD/PCDF and other trace
organics is to optimize the refuse combustion system. Riley personnel
expressed an opinion that the published data base is not sufficient to
correlate PCDD/PCDF emissions with operating performance parameters (e.g.,
stack CO emissions or maximum qualifying volumetric heat release rate). They
feel that the basic design features of the Riley/Takuma firing system,
coupled with a sophisticated automatic combustion control (ACC) system, will
provide burning conditions sufficient to minimize unburned hydrocarbon
species emissions. The system is designed to maintain at least a one-second
residence time above 1800°F as required by several State emission regulatory
bodies.
Basic features of the Riley/Takuma firing system are illustrated in
Figure 5-14 and Table 5-7. As shown, there are four grate sections including
a feeder grate, a drying grate, a firing grate, and a finishing grate.
Underfire air is provided to the drying, combustion and finishing grates
through plenums located under each grate section. Each plenum is equipped
with dampers and venturi sections to modulate and measure flow to the
particular plenum. Normally, the total underfire air flow will provide 125
to 145 percent of the theoretical air requirements with the majority of the
underfire air proportioned through the combustion grate- They hope to
achieve a uniform (spatial and temporal) release of volatile matter from the
burning grate.
5-38
-------
REFUSE CHARGING
HOPPER & CHUTE
FURNACE WATER
TUBE PANEL
FEEDER
HYDRAULIC
OIL CYLINDER
DRYING
GRATE
RIDDLING HOPPER
AND GRATE
AIR PLENUM
COMBUSTION
GRATE
DOUBLE DUMPER
RIDDLING CONVEYOR
BURN-OUT GRATE
HYDRAULIC
OIL CYLINDER
ASH DISCHARGER
CAST IRON
BLOCK OVERLAY
REFRACTORIES
OVERLAY
AUXILIARY
FUEL BURNER
MAIN ASH
CHUTE
Figure 5-14.
Cross sectional schematic of combustion zone on a
Riley-Takuma mass burn plant.
5-3S
-------
TABLE 5-7. DESIGN AND OPERATING FEATURES OF RILEY-TAKUMA
TECHNOLOGY
• A GRATE SYSTEM
- FEEDER GRATE
- DRYING GRATE ( UNDERFIRE AIR
rnMRiKTTHM PDATC? PROPORTIONED AND
- COMBUSTION GRATE ( MONITORED BY ACC
- BURNOUT GRATE
I 80-90 PERCENT EXCESS AIR AT MCR
- 70 - 80% OF TOTAL AIR THROUGH
GRATE
I REFRACTORY CLADDING 30' ABOVE GRATE
• ACC SENSES EXHAUST 02 AND FEED
- TRIMS 02 WITH OFA FLOW f
- MAINTAINS T > 1800°F WITH
AUXILIARY BURNER
i THREE LEVELS OF OFA PORTS
I . CO < 200 PPM
i WILL OPERATE AT 55% WASTE FEED RATE
ON 3800 BTU/# WASTE
- SYSTEM DESIGNED AT 100% - A,500
5-40
-------
The Riley/Takuma design provides a center flow furnace configuration by
placing an arch over both the drying and finishing grates. The boiler is a
straight wall configuration above the combustion grate. Silicon carbide
refractory cladding is provided to a height approximately 30 feet above the
grate. Overfire air ports are provided at three locations and are designed
to achieve complete coverage of flow from the lower furnace. Overfire air
quantities usually amount to approximately 20 to 30 percent of the total
combustion air. Overall, the system is designed to operate at 80 to
90 percent excess air.
As stated above, Riley/Takuma attempts to achieve PCDD/PCDF emission
control by maintaining at least one-second residence time above 1800°F- This
is accomplished through use of a sophisticated automatic combustion control
system in conjunction with auxiliary burners. Measurements are made of
exhaust oxygen concentration and the furnace exit gas temperature (i.e.
temperature at convective section inlet). The ACC will modulate the overfire"
air flow to maintain constant 03 concentration. This is primarily a trimming
operation on the overfire air. Correlations of required furnace exit gas
temperature to assure one-second residence time above 1800°F (as a function
of load) will be developed and incorporated in the microprocessor of the ACC.
That minimum condition will be maintained by modulation of the fuel-charging
rate, grate speed and underfire air flow rate, as well as through the use of
an auxiliary burner -
The auxiliary burner is located in the lower furnace region near the
grate in order to maintain the time-at-temperature requirement. This
required positioning has generated operational problems associated with
burner openings getting covered with slag. Riley is currently working on an
advanced auxiliary burner design to circumvent that problem.
A potentially important aspect of MSW units designed by Riley/Takuma is
the provision for operation at lower than full load. Riley personnel
indicated that their systems could be operated at loads as low as 55 percent
of rated input (TPD) burning waste with 3800 Btu/lb heating value. It should
5-4-1
-------
be recalled, however, that the ACC will still maintain the time at
temperature criteria through use of the auxiliary burner.
Further inquiry into operational control at reduced load conditions
indicated that the overall excess air level was allowed to increase. Exhaust
02 concentration versus load curves were not provided. There is obviously
concern over maintenance of overfire air jet penetration and mixing at
reduced load. Riley personnel noted that there were three levels of overfire
air ports and that levels could be shut down at reduced load. That procedure
should maintain jet penetration. However, shutting down a row of overfire
air jets is accomplished manually and is not automatically controlled through
the ACC.
It should be noted that the MSW facility being constructed at Olmstead
County, Minnesota will be the first Riley/Takuma system in operation in the
United States. Even though there is significant operating experience on
Takuma grates in Japan, it is reasonable to expect that Riley will fine-tune
its design and control strategy through experience gained at Olmsted.
5.9 Detroit Stoker
Turner (1982) reported that in 1982 there were nine mass-burn, waterwall
refuse combustion plants in operation in the United States and that four of
the plants employed Detroit Stoker grates. The combination of Detroit Stoker
grates with various boiler manufacturers (Foster Wheeler, Babcock and Wilcox,
and Keeler) reflects the historic contracting procedure in the United States.
That procedure generally involves selection of an engineering firm to design
the plant with major component selection based on competitive bids.
The preceding sections of this chapter have discussed design and
operating philosophy of various firms marketing complete resource recovery
systems. Detroit Stoker is not a MSW system supplier. Typically, they
supply fuel feed, grates and air supply hardware (underfire and overfire) to
the boiler manufacturer- Detroit Stoker then works with the boiler designer
and control system contractor (this could be a third party in the American
5-42
-------
contractual approach) to integrate the various components into an overall
system.
Significant testing has been undertaken of the Hampton plant in Virginia
over the last several years. The PCDD/PCDF emissions from this plant which
employs a Detroit Stoker grate are significantly higher than other nass burn
waterwall systems (see Table 4-1 and the data base volume entitled "Emission
Data Base for Municipal Waste Combustors." The reasons for these higher
emissions are discussed in some detail in Chapter 8 of this report. Current
Detroit Stoker hardware design is significantly different from that used at
Hampton (Detroit Stoker, 1986). Figure 5-15 and Table 5-8 illustrates the
current design. Major changes have occurred in the distribution of underfire
air, addition of flow constrictions and modifications to the overfire air
system design. Each of these areas are discussed below.
The underfire air system now consists of separate plenums under each
grate section. The grates themselves are constructed as standard width
modules. Thus, depending upon the unit size the grate may consist of a 3 x 1
or 3 x 2 module configuration with either 3 or 6 underfire air plenums. Air
flow to each is individually adjustable. Air pressure in the underfire
plenums is typically on the order of 3 to 3-1/2 inches water gauge which is
low by European design standards. Detroit Stoker is convinced, however, that
this pressure level is sufficient to achieve uniform air distribution across
the drying, firing and finishing grates. Establishment of the underfire air
flow distribution is based on visual observation of the flame. Normally, 70
to 75 percent of the underfire air is directed through the firing grate.
As the combustion gases leave the grate region they are directed through
a constricted flow region formed by "noses" on both the front and rear boiler
walls. Clearly, construction of the throat region is the responsibility of
the boiler manufacturer. This is a critical area requiring close
coordination between the boiler and grate manufacturer. Overfire air ports
are located on both the front and rear walls. The number, location and angle
of the overfire air ports are adjusted for each unit design following design
5-43
-------
Key
1. Refuse Charging Hopper-
Charging Throat
Charging Ram
Grates
Roller Bearings
Hydraulic Power Cylinders
Vertical Drop Off
Overfire Air Jets
Combustion Air
Automatic Sifting Removal
Figure 5-15. Cross-section of boiler with current design Detroit Stoker firing system.
-------
TABLE 5-8. DESIGN FEATURES OF DETROIT
STOKER MASS BURN COMBUSTORS
I FUEL FEED, GRATE AND OFA
BY DSC
• BOILER NOT BY DETROIT STOKER
I MULTIPLE GRATE AIR PLENUMS
WITH INDIVIDUAL CONTROLS (9)
I PUSHER TYPE GRATE DESIGN
I FLOW RESTRICTIONS TO LIMIT
COLD REGION BY-PASS
I MULTIPLE ROWS OF OFA PORTS
- 40-50% TOTAL AIR
- DESIGN CONTROLS FLAME
HEIGHT
I REFRACTORY CLADDING (SIC)
15-20 FEET ABOVE GRATE
I DESIGNED TO PROVIDE 2:1
TURNDOWN
5-45
-------
criteria for controlling flame height. Typically 30 to 40 percent of the
total combustion air is supplied as overfire air.
The lower furnace region - to an elevation above the overfire air ports
- is covered with approximately 1 inch of silicon carbide refractory to
prevent tube corrosion. Detroit Stoker feels that the combination of
insulated walls, flow constriction and careful design of the overfire air
jets will lead to high combustion efficiency. As in the case with
essentially all manufacturers, "good combustion" is equated with low PCDD/
PCDF emission. They do not, however, have emission data to establish the
validity of that belief. Detroit Stoker recommends that the boiler
manufacturer provide at least one second residence time above 1800°F.
Startup and shutdown rely on auxiliary burners. These burners are
either gas or oil-fired and are located well above the grate. Exact
location, size, fuel, and controT sequence for the auxiliary burners is the
responsibility of the boiler manufacturer.
Detroit Stoker grates are designed to operate with a 2:1 turndown ratio
but Detroit Stoker prefers to have base-loaded systems. The lower limit for
turndown is generally established by furnace temperature. This condition
will also be characterized by deterioration in overfire air mixing as well as
increased CO and unburned hydrocarbon emissions. Detroit Stoker suggested
that a row of overfire air ports could be dropped from service at low load
but that is not an integral part of their suggested control strategy-
The Detroit Stoker design for overfire air mixing strategy has evolved
through experience gained in numerous MSW projects. They are currently
considering cold-flow modeling studies to refine their current designs.
5.10 Combustion Engineering - De Bartolomeis
Over 40 percent of the world's thermal electric power is produced by
equipment of Combustion Engineering (C-E) design. They offer complete system
designs and hardware for resource recovery applications and will supply
-------
either mass burning or RDF configurations. Further discussion of C-E's RDF
system offering will be presented in Chapter 6. Combustion Engineering has
recently been selected as the system supplier for a 600 ton/day MSW facility
to be built in Chattanooga, Tennessee. This will be the first C-E designed
mass burn facility in the United States.
The grate system for C-E MSW systems is obtained through an exclusive
North American license with the Italian firm, De Bartolomeis (db).
Figure 5-16 illustrates the db grate design which consists of alternate
moving and stationary steps in an overlapping configuration (Combustion
Engineering, 1986). The three-part step design includes a scraper which
serves the dual function of cleaning the lower surface plate and forms the
minimum free area passage for underfire air.
Each grate section is equipped with its own plenum to allow control of
underfire air. It should be noted that various grate sections are shown to
be configured in a continuous path, without steps between grate sections.
The C-E personnel indicated that the grate motion provided sufficient
aeration of the fuel and felt that steps between grate sections could result
in uneven burning. Unlike most other grate designs, the slope of db grates
can vary from horizontal to as much as 21 degrees. C-E sales literature
indicates the slope of the grate is based on waste composition. For the
typical ranges of heating values found in the United States, the slope is 8°-
Very low heating values (<4000 Btu/lb avg. HHV) would require steeper slopes.
Figure 5-17 illustrates the db grate integrated into a complete resource
recovery system. For MSW, C-E uses a boiler based on a design developed by
Energie und Verfahrenstechnik GmbH (EVT). EVT is partially owned by
Combustion Engineering. Detailed information on overfire air mixing
considerations, splits between underfire and overfire air, combustion control
strategy, startup procedures, etc. were not made available.
5-47
-------
en
ji'
CO
Supporting Bo*
Surface Plate
Supporting Roll
-Operating
Shan
Hydraulic
Cylinder
Enlarged Area Showing Air Flow
Figure 5-16. Design features of DeBartolomeis grate.
-------
en
-Cv
10
Figure 5-17. Boiler cross-section of CE design using db grate and EVT boiler,
-------
5.11 Westinghouse/0'Connor
Westi nghouse Corporation is a relatively recent entrant in the list of
companies competing in the combustion portion of the resource recovery
industry. That entry was accomplished through a 100 percent purchase of the
O'Connor Combustor Corporation (Westinghouse, 1986). Prior to the
Westinghouse buyout, O'Connor had built a MSW facility in Sumner County
(Gallatin), Tennessee as well as facilities at five locations in Japan.
Current projects being designed or under construction include facilities in
York County, Pennsylvania; Bay County, Florida; Dutchess County, New York;
and in Bloomington, Indiana.
The O'Connor combustion system.represents a unique approach for burning
municipal solid waste. A typical plant configuration is illustrated in
Figdre 5-18 and shows that the heart of the system is a water-cooled rotary
combustor- The rotary cylinder consists of alternating watertubes and
perforated steel plates. MSW is metered into the combustor via a feed chute
and ram feeder. Preheated combustion air is divided into six zones and
enters the combustor through the perforated plates forming the walls of the
cylinder. The rotary combustor turns slowly (10-20 RPH) and is oriented with
a slight (approximately 6°) downward tilt. The rotary section terminates
within a waterwalled boiler allowing the residue to fall into an ash removal
system.
Combustion air is drawn from the waste pit and passes through an air
preheater at the boiler exhaust. Typical air preheat temperature is 450°F.
In the Gallatin facility a portion of this air was drawn off by a second fan
and directed to three overfire air port elevations in the boiler - one below
the rotary combustor dump and two elevations above the rotary combustor. As
shown in Figure 5-19, the main combustion air is distributed axially down the
rotary combustor. Each axial section is subdivided into two zones. With
counterclockwise rotation, air passing through the right zone is forced
through the burning material bed. Air admitted to the left zone will enter
the rotary combustor in a region which is effectively above the burning bed.
5-50
-------
Gas Clean-Up Equiprmvr
Combustion Air
Waterwall
Boiler
Forced
Draft Fan
Westingnouse
O'Connor
Water-Cooled
Rotary Combustor
Ram
Feeding
Fly Ash
Conveying System
Ash Conveying
System
Ash Removal System
Figure 5-18. Typical plant configuration using Westinghouse/O'Connor combustor.
-------
Figure 5-19. Cross-section of Westinghouse/O'Connor combustor
5-52
-------
Westinghouse personnel refer to those two zones as underfire and overfire air
respectively.
In comparison to the grate firing systems discussed in earlier sections
of this chapter, there are several other unique features of the O'Connor
combustor. Typical grate firing systems operate at 80 to 100 percent excess
air while the Westi nghouse/0'Connor system operates at approximately
50 percent excess air (both at 100 percent load operation). Traditional
grate systems for MSW firing provide silicon carbide cladding to the lower
waterwalls to prevent corrosion. There is no refractory on the waterwalls of
the O'Connor rotary combustor.
As noted previously, the Gal latin, Tennessee facility was equipped with
"overfire air" ports at three elevations in the boiler. In that facility,
the region from the rotary combustor to the ash pit of the boiler was
equipped with a stationary grate referred to as an "afterburning grate."
This particular facility has been subjected to extensive field testing by
Cooper Engineers, Inc. as part of a study sponsored by the California Waste
Management Board. A 1984 report (Cooper Engineers, 1984) states that "it was
found from earlier Cooper Engineers testing that the boiler overfire air and
grate air was ineffective, so these systems were no longer used and all of
the combustion air is now introduced through the combustor as underfire and
overfire air." This has been confirmed in discussions with Westinghouse
personnel . The test report by Cooper Engineers includes data on criteria
pollutants and heavy metals, but does not include information on PCDD/PCDFs/
furans and other air pollutants. It is not known whether the boiler overfire
air system had an impact on those pollutant emissions.
All of the existing facilities using the O'Connor combustor system were
constructed prior to the Westinghouse purchase of O'Connor. For a variety of
technical and marketing reasons, Westinghouse has entered into a long-term
contractual arrangement with Sumner County. They have made a variety of
major system repairs including replacement of the stacks which collapsed due
to corrosion. It is also reported that the facility has a new superintendent
who is a Westinghouse employee. The significance of this arrangement is that
5-53
-------
Westlnghouse intends to use the Gal latin facility as both a show place
facility and as a test bed for evaluating issues such as PCDD and PCDF
emissions.
5.12 Basic Environmental Engineering Inc.
Basic Environmental Engineering located in Glen Ellyn, Illinois builds
multistaged mass burn combustors for smaller scale applications (Basic
Environmental Engineering, 1986). Basic offers systems with sizes ranging
from 4 MMBtu/hr to 56 MMBtu/hr (10 to 150 T/d). The construction of these
units is done in a "unitized" modular fashion as opposed to field erected
which is common for larger mass burn systems. The primary zone of the Basic
system is not starved air so that the technology can be classified as small
modular mass burn combustion. There is currently little information
available on trace organic emissions from Basic Environmental Engineering
Systems and hence it is uncertain whether current design practices are
sufficient; however, the design and operating approach employed by Basic is
worthy of consideration because Basic is one of the leaders in advanced small
unit systems. Also many of the features of the Basic technology will likely
have favorable impact on trace organic emissions. The process flow diagram
of a typical Basic system is provided in Figure 5-20 and design features are
summarized in Table 5-9. The combustion system consists of three zones as
fol 1ows:
• Pulsed hearth primary operated at near stoichiometric conditions
• Fired afterburner secondary with secondary air addition
0 Final air addition in tertiary
The primary zone includes a ram feeder to transfer the solid waste onto the
first of two pulse hearths. (Basic uses one, two or three pulse hearths
depending on model size.) Primary combustion air is introduced through the
hearth at approximately stoichiometric conditions. The primary chamber is
constructed of membrane water wall. The firing density in the primary is
designed to 12,000 Btu/hr/ft3 and Basic personnel believe that this is a key
parameter that should not be exceeded. Higher volumetric charging rates are
5-5*
-------
Figure 5-20. Process flow diagram of Basic Environmental
Engineering modular mass burn technology.
-------
TABLE 5-9. DESIGN FEATURES OF BASIC ENVIRONMENTAL
ENGINEERING SMALL MODULAR MASS BURN
TECHNOLOGIES
• THREE STAGE COMBUSTION
- PULSED HEARTH PRIMARY
- FIRED AFTERBURNER SECONDARY
- "THERMAL EXCITER" TERTIARY
• PRIMARY
- MEMBRANE WATER WALL
- PULSED HEARTH
- STOICHIOMETRIC AIR
• SECONDARY
- FIRED AFTERBURNER - TEMPERATURE CONTROL
- AIR INTRODUCTION WITH "THERMAL EXCITER"
CYCLONIC OUTWARD INJECTION
- REFRACTORY LINED
• TERTIARY
- FINAL AIR ADDITION WITH EXCITER
- REFRACTORY LINED
I EXCESS AIR
- TYPICALLY 80%
• MONITORING & CONTROL
- TEMPERATURES OF EACH ZONE
- STEAM PRODUCTION
- CO (20-30 PPM)
• TIME AT TEMPERATURE
- IN REBURN ZONES 1 SEC ABOVE 1800°F
NOX ~ 35 PPM
5-56
-------
expected to lead to poor burnout and excessive carbon and volatile carryover
to the secondary chamber. Higher charge rates also lift larger particle
sizes which require greater residence time to destroy the carbonaceous
particles before leaving the high temperature furnace zone.
The secondary or "stage two" chamber includes a fired afterburner and
additional air injection. The first afterburner is positioned at the exit of
the transition duct between the chambers and can be fueled on oil or natural
gas. The design capacity of the afterburner is 20 percent of the total Btu
input of the system for normal municipal solid waste and is increased to 40
or even 50 percent when the refuse includes: a) large amounts of chlorinated
plastics or b) when the furnace is operated at lower furnace input firing
rates, but the waste fuel is primarily long chain hydrocarbons such as
plastics and rubber. The afterburner is fired before startup for heating the
system (to 1400°F) and is cycled on and off to maintain temperature. For
refuse with higher moisture content (less than 3800 Btu/lb) the afterburner
is continually fired but varied and adjusted automatically to setpoint
temperatures. Additional air is added into the second stage after the
afterburner using a unique injection scheme called a "thermal exciter". The
air is injected outward from a refractory-lined closed cylinder positioned
along the center axis of the secondary chamber. The air flows through small,
high-velocity jets and are oriented to achieve, vortex flow in the secondary
chamber.
The tertiary chamber consists of a second "thermal exciter" for final
air addition to an overall excess air level of approximately 80 percent. The
chamber exit temperature is monitored and is used as the control variable for
the fired afterburner. The temperature is maintained at between 1600 and
1800°F depending on refuse characteristics. The maintenance of the exit
temperature of the final combustion chamber ensures that the gases experience
at least one second above that temperature in the combined second and third
chambers. In other words, typical design conditions would also ensure
greater than 1 second above 1800°F under excess air conditions.
5-57
-------
The Basic combustion monitoring and control system is relatively
straightforward. Temperatures of each chamber exit are used as the principal
control variables and the air flows and afterburners are used to maintain the
temperature. Basic is currently investigating the use of oxygen and carbon
monoxide monitoring for system control. The Basic technology is capable of
achieving high turndowns (up to 4:1) without requiring additional auxiliary
fuel to maintain the destruction temperatures of the reburn zones that it
required at full design conditions. Example: if at full load the waste fuel
burned did not require auxiliary fuel, then down to 4:1 it would not need it.
Basic engineers would not advise lower load operation without auxiliary fuel
being required. If a client desires to operate to 3:1 turndown, the design
supplied to the client will be changed from the standard control of
60 percent turndown without use of auxiliary fuel to maintain temperature.
The distributed air addition and moderated uniform temperatures likely
keeps NOX emissions from being excessive. Basic estimates NOX levels of less
than 45 ppm compared to other modular systems in the 170 ppm range. The
exhaust carbon monoxide level is also low, typically in the range of 20-30
ppm. Although CO is not typically measured unless requested by the client,
Basic believes that this is an indicator of the maintenance of good
combustion conditions. Most currently installed Basic systems do not have
air pollution control devices; however if the waste stream has greater than 2
percent by weight heavy metals or 1/2 percent of plastics with halogens or
halogen salts, then baghouses, electrostatic precipitators and/or scrubbers
are recommended.
5.13 Enercon/Vicon
Enercon Systems, Inc. is a small corporation specializing in the
engineering, design and construction of industrial energy systems. They have
developed and patented a multistage controlled excess air mass burning
technology. Vicon Recovery Systems, Inc. specializes in the design,
construction and operation of MSW facilities. Vicon is the exclusive
licensee of the Enercon technology for MSW. The Enercon/Vicon technology is
used at the MSW facility in Pittsfield, MA which is the site of a continuing
5-5$
-------
extensive emission performance evaluation (evaluation developed under the
auspices of ASME with project sponsorship from numerous state government
agencies, DOE, Ontario Ministries of Energy and Environment, the Vinyl
Institute, and the Association of Plastics Manufacturers in Europe).
The Enercon/Vicon technology utilizes modular construction but extends
the "modular" concept to significantly larger scale than the John Basic
Technology discussed in the previous section. The Pittsfield, MA facility,
for example, consists of three municipal waste combustors (or modules), each
rated at 120 tons/day; two furnaces are typically on line, with one standby.
The Enercon/Vicon system technology relies on refractory lined chambers
without heat extraction. Burning occurs in the primary chamber which is
equipped with a recuperative combustion air liner to minimize heat loss from
the zone. Combustion gases exist the primary chamber through a passage
referred to as the "mixing throat" and enter the refractory lined secondary
combustion chamber. This chamber simply provides high temperature residence
time to maximize trace organic burnout and oxidation of CO to
The incinerator module referred to as the tertiary chamber is actually
an insulated transfer duct carrying hot combustion gases from one or more
secondary chamber(s) to one or more waste heat boilers. At the Pittsfield
facility, there are three incinerator modules feeding into a common tertiary
chamber. The combustion products are fed to two waste heat boilers designed
to operate with 1400°F boiler inlet temperature.
Typically, a front end loader is used to fill the charging hopper of the
incinerator. When filled, the hopper door is closed, the fire door opened
and a hydraulic ram actuated to shove the MSW into the primary chamber. A
fresh MSW charge is added at approximately ten minute intervals. -The design
volumetric heat release rate in the primary zone is 10,000 Btu/hr per cubic
feet. Water cooled, hydraulic rams are used to move the MSW along the
stepped series of refractory hearths. Stoking rams are actuated on
approximately a five minute cycle.
5-59
-------
Oxidizer is added to the primary chamber in three ways. Undergrate gas
can be a mixture of recirculated flue gas (underfire FGR) and fresh air drawn
from the tipping room. Visalli et al. (1986) report that in the Pittsfield
facility, undergrate oxidizer is 100 percent FGR. As noted previously, the
primary chamber is equipped with a recuperative air passage. This preheated
air is injected at the head end of the primary chamber above the burning bed
of MSW. The third source of oxidizer to the primary zone is recirculated
flue gas, added through high velocity, overfire jets. The flow rate of each
oxidizer gas source (and distribution) is controllable.
Oxygen and gas temperature monitors are located within the secondary
chamber and provide feedback to control primary zone operation. 'The
temperature sensor provides a control signal to modulate the quantity of
overfire FGR while the 02 monitor is used to modulate the quantity of primary
combustion air. Typically, the temperature in the secondary chamber is
controlled to 1800°F while the excess air is cut to approximately 50 percent.
FGR modulation controls gas temperature by controlling the amount of dilution
but has minimal impact on 03 level. Auxiliary burners are provided to assist
in maintaining temperature for particularly wet MSW. Oxygen concentration is
impacted by modulation of the preheated combustion air since that oxidizer
stream contains 21 percent 02 (versus 4 to 8 percent for the FGR streams).
It should be noted that the excess air level maintained by Enercon/Vicon is
significantly less than that provided in typical mass burn, waterwall
systems.
Additional recirculated flue gas is added to the tertiary chamber to
control the boiler inert temperature. This final dillution occurs after the
combustion products have remained in the secondary chamber for at least one
second of a temperature of 1800°F.
As noted previously, the Pittsfield, MA facility is undergoing extensive
performance evaluation testing. Only limited data is publicly available.
However, based on the reportable information, this technology is capable of
operation at very low CO and NOX levels. Over the operating temperature
range (secondary zone) of approximately 1600 - 1800°F, CO emissions were
5-60
-------
found to be less than 30 ppm. Even though the Enercon/Vicon system does not
remove heat from either the primary, secondary or tertiary chambers, the
system operates at relatively low temperature (due to extensive use of FGR)
and thus the NOX emission levels were typically in the 100 ppm range.
In addition to the Pittsfield facility, Enercon/Vicon is providing
several new MSW facilities. A 600 tpd facility (5 units at 120 tpd each) is
in advanced shakedown at Pigeon Point, Delaware. This facility will burn a
50/50 mixture of RDF and MSW (not necessarily combined). A 240 tpd (2 @
120 tpd) facility is an advanced construction stage at Rutland, Vermont while
a 360 tpd (3 @ 120 tpd) facility is under construction at Springfield, Mass.
Finally, a 420 tpd (3 @ 140 tpd) facility is being built at Wallingford, Ct.
With the exception of the Wallingford facility, each of the Enercon/Vicon
systems use 120 tpd modules. Each of these 120 tpd units consists of six
hearths rated at 20 tpd each. To increase the rating to 140 tpd, a seventh
hearth has been added to the Wallingford facility.
5.14 References
Basic Environmental Engineering, 1986. Discussion and data supplied to
W. R. Seeker in telephone conversation with A. J. Matlin on October 16,
1986.
Combustion Engineering, 1986. Discussion and data supplied to W. R.
Seeker and W. S. Lanier by Combustion Engineering in Windsor, CT on
September 8, 1986.
Cooper Engineers, 1984. Air Emissions Tests of Solid Waste Combustion
in a Rotary Combustor/Boi 1 er System AT Gallaton, Tenn. Report to West
County Agency of Contra Costa County Waste Co. Disposal - Energy
Recovery Project.
DBA, 1986. Discussion and data supplied by DBA to W. R. Seeker and
W. R. Niessen by Deutsche Babcock in Keyersfeld, Germany on August 1,
1986.
5-6T
-------
Detroit Stoker, 1986. Discussion and data supplied to W. R. Seeker and
W. S. Lanier by D. Reschly and T. Gaerer in Monroe, MI on September 9,
1986.
Martin, 1986. Discussions and data supplied by Martin personnel
(D. Kreuch, J. Horn, H. Weiand, G. Schetter, J. Martin, D. Sussman) to
W. R. Seeker and W. R. Niessen in Munich, August 6, 1986.
Martin & Schetter, 1986. "Reduction of Pollutant Emissions from Refuse
Incinerators by Means of Optimized Combustion Conditions." 8th Members'
Conference of the IFRF, Nordwijkerhaut, The Netherlands, May 28-30,
1986.
Riley, 1986. Discussion and data supplied to W. S. Lanier in telephone
conversation with Riley personnel on August 17, 1986.
Steinmueller, 1986. Discussion and data supplied by Steinmueller
personnel (H. Pollack, P. Daimer, 0. Kaiser) to W. R. Seeker and
W. R. Niessen in Duesseldorf, August 1, 1986. Also report "Feurung
stechnische Moglictikeiten zur Schadstoffreduzierung bei
Mullverbrennungsunlagen", K. Leikert and H. Pollack.
Turner, 1982. "Mass Burning in Large-Scale Combustors" Thermal
Conversion Systems for Municipal Solid Waste, Noyes Publications,
Park Ridge, NJ.
Visalli, Joseph R. et al., "Pittsfield Incinerator Research Project;
Status and Summary of Phase 1 Report; Plant Characterization and
Performance Testing," Paper presented at 12th Biennial National Waste
Processing Conference, ASME, Denver, Colorado, June 1-4, 1986.
Vogg and Steiglitz. Presented at the 5th International Symposium on
Chlorinated Dioxins and Related Compounds. September 16-19, 1985.
Bayreuth, West Germany.
5-62
-------
Volund, 1986. Discussion and data supplied by Volund personnel to
W. R. Niessen, July 30, 1986. See also "Emission Test at a Danish
Energy from Waste Plant." M. Rasmussen, Volund Report No. 32V1-8,
October 1985
Von Roll, 1986. Discussion and data supplied by Von Roll personnel (W.
Staub and A. Scharsach) to W. R. Seeker and W. R. Niessen in Zurich,
August 4, 1986. Also Von Roll Report on the testing at Neustadt/
Ostohstein, 1984.
W + E Umwel ttechnik AG, 1986. Discussions and data supplied by M.
Sudobsky and M. Zweifer of WE Umwel ttechni k, AG to W. R. Seeker and W.
R. Niessen in Zurich, August 5, 1986.
Westinghouse, 1986. Discussions and data supplied to W. R. Seeker and
W. S. Lanier by Suh Lee, D. Bechler and S. Winston in Pittsburgh, PA,
September 10, 1986.
5-63
-------
6.0 CURRENT PRACTICES IN RDF COMBUSTION
The chemical and physical characteristics of a fuel are important
parameters in the design of any combustion system. The previous chapter
described firing systems used for burning raw municipal solid waste in water
wall boilers. Due to the wide variability in MSW characteristics, those
combustion systems are significantly different from systems designed to burn
traditional solid fuels such as coal, wood, hogged fuel, etc. An individual
charge of MSW may physically vary from small scraps of paper to discarded
refrigerators. The volatility of an MSW charge may have contributions from
discarded propane tanks and bottles of solvents as well as contributions from
rain-soaked yard waste and discarded firewall insulation. The ash
constituents will include easily melted items such as aluminum cans as well
as a liberal amount of sand and glass. The melted aluminum tends to solidify
on grate openings sealing air passages, while sand and glass tend to erode
the grate. Mass-fired MSW systems are specifically designed to accommodate
those variations. An alternate design approach is to process the MSW to
produce a less difficult fuel. The degree of processing can vary from simple
removal of bulky items and shredding to extensive processing to produce a
fuel suitable for co-firing in pulverized coal-fired boilers. Preprocessed
MSW, regardless of the degree of processing is broadly referred to as refuse
derived fuel or RDF.
6.1 Types of RDF
The American Society for Testing and Materials (ASTM), through its
Committee E-38.01 on Resource Recovery Energy has established classifications
defining different types of RDF- The various classifications and resultant
fuel descriptions are listed in Table 6-1 and discussed below.
6.1.1 RDF-1 or MSW
As noted in Table 6-1, removal of bulky items such as discarded water
heaters, refrigerators, etc. produces RDF-1. Producing RDF-1 is accomplished
by hand removal of the bulky items on the tipping floor or by careful pit
6-1.
-------
TABLE 6-1. ASTM CLASSIFICATION OF REFUSE DERIVED FUELS
Type
of RDF
Description
RDF-1
(MSW)
RDF-2
(c-RDF)
RDF-3
(f-RDF)
RDF-4
(p-RDF)
RDF-5
(d-RDF)
RDF-6
RDF-7
Municipal solid waste used as a fuel in as-discarded form, without
oversize bulky waste (OBW).
MSW processed to coarse particle size, with or without ferrous
metal separation, such that 95 wt % passes through a 6-in.-square
separation, such that 95 wt % passes through a 6-in.-square mesh
screen.
Shredded fuel derived from MSW and processed for the removal of
metal, glass, and other entrained inorganics. The particle size
of this material is such that 95£ by weight (wt %) passes through
a 2-in.-square mesh screen. Also called "fluff RDF."
Combustible-waste fraction processed into powdered form, 95 wt %
passing through a 10-mesh (0.035-in.-square) screen.
Combustible waste fraction densified (compressed) into the form of
pellets, slugs, cubettes, briquettes, or some similar form.
Combustible-waste fraction processed into a liquid fuel (no
standards developed).
Combustible-waste fraction processed into a gaseous fuel (no
standards developed)".
6-2
-------
management by the crane operator. This type of fuel retains the majority of
fuel variability characteristics found in raw municipal waste. Clearly, the
combustion system requirements to burn RDF-1 are the same as for raw MSW.
6.1.2 RDF-2 or Coarse RDF (C-RDF)
The second type of refuse derived fuel described in Table 6-1 is
referred to as RDF-2, "Coarse RDF," or c-RDF- A primary shredder is used to
reduce the "particle" size such that 95 weight percent passes through a 6
inch square mesh screen. Flail mills, hammer mills or rotary cutters may be
used for this primary shredding operation. Preparation of coarse RDF may
also include removal of ferrous metals. Several types of magnetic devices
have been developed for that purpose. Since the extracted metals generally
contain loose paper or other materials, an air separation device is often
used to clean the ferrous metal.
The decision on including ferrous metal separation in the waste
preprocessing is largely an economic issue. Metal removal generates a
potentially saleable by-product and also reduces the mass of material
processed in the municipal waste combustor. In the Albany, N.Y., Solid-Waste
Energy-Recovery System (ANSWERS) waste is processed in one facility and then
transported by truck to the boiler facility. In this instance ferrous metal
separation also reduces costs associated with fuel transportation. Another
important consideration is that ferrous metals can also be separated from the
boiler's bottom ash. That is a practice employed in several MSW systems
(e.g., the Signal Resco facility at Baltimore, MD). Ferrous metals recovered
from the ash will have first passed through the boiler leaving a cleaner
product for sale".
The ANSWERS operation described in Figures 6-1 and 6-2 may be used to
illustrate two important RDF considerations. First, by preprocessing the
waste it may be economical to store and transport the RDF. Though long-term
storage and/or long distance transport would generally be restricted to more
highly processed forms of RDF, these factors are an important consideration
in developing RDF processing technology. The second important consideration
6-3
-------
en
i,
Vlbrellng-
Pan Feeder*
Picking ^
Olatlona •
Magnetic Refuee
Separators 8*iredder»
Tipping
Room
Untreated Refute
Cor proceeding)
Stationary
Compactor, %£„.„
Ferroua-Melal Removal
(for recycling)
Fu«l Product
((or delivery to
generating plant)
Figure 6-1. Albany, New York, Solid-Waste Energy-Recovery System
(ANSWERS).
-------
01
, f
in
Figure 6-2. Cross-section of Albany, New York boiler plant
firing c-RDF.
-------
was that the coarse RDF produced at the ANSWERS processing facility
(Figure 6-1) was to be burned in a boiler (Figure 6-2) equipped with a
spreader stoker and traveling grate firing system. The two RDF boilers at
ANSWERS can fire 100 percent RDF. The firing system design, however, is an
adaptation of hardware designs developed for coal- and wood-firing systems.
Thus, preprocessing of the waste to produce c-RDF allowed the use of existing
American boiler technologies as an extension of experience with a variety of
industrial waste fuel and wood residues.
Initial operation of the ANSWERS facility produced RDF with a much
smaller particle size distribution than anticipated in the boiler design. As
a result, a significant portion of the fuel was entrained into the gas flow
instead of falling to the grate. Hasselriis (1983) reports that "too much of
the RDF carried up and out of the furnace" and that the incompletely burned
material collected in the boiler hopper as char. Further, bridging tended to
occur in the hopper. Subsequently, the shredder grates have been changed at
the processing facility to produce a larger mean size RDF. Using the larger
sized RDF, excessive carry-over can be avoided if the boiler firing rate is
held to less than 80 percent of rated capacity- Other factors such as boiler
fuel feed design, boiler height and volume, overfire air distribution and
grate heat release rate can help compensate for this problem.
The above described initial experience at ANSWERS illustrates that
significant operational problems can occur if c-RDF, to be burned in a
spreader stoker firing system, has too high percentage of small sized
particles. It should also be recognized, however, that the normal size
distribution for c-RDF (95 weight percent passes 6 inch square screen) allows
relatively large items to enter the firing system. Note that 95 percent
smaller than any given mesh size implies the possibility of 5 percent larger
than that size. Accordingly, it may be very difficult to actually achieve
uniform fuel feed rate in boilers firing coarse RDF. One approach to this
problem is a higher degree of waste pre-processing.
6-6
-------
6.1.3 RDF-3 or Fluff RDF (f-RDF)
Fluff RDF or RDF-3 production involves removal of ferrous and non-
ferrous metals, removal of glass and other entrained inorganics, as well as
size reduction such that 95 weight percent of the mass will pass through a
2 inch square mesh screen. This type of fuel is also referred to as f-RDF.
Figure 6-3 illustrates the RDF production facility at Ames, Iowa which
generates RDF-3. Size reduction is accomplished with primary and secondary
stage shredding. Flow to the secondary shredder occurs after magnetic
separation of ferrous metals and size segregation in disk screens. Underflow
from the first disk screen (less than 1.5 inch) is fed to a second screen.
Underflow from the second screening (less than 0.375 inch) consists primarily
of fine glass, grit and finer fibers which are disposed of in a landfill.
The overflow from secondary screening is combined with the secondary shredder
output and fed to an air classifier. The heavy fraction from the air
classifier is discarded to landfill (after a secondary magnetic separation)
while the lighter fraction is pneumatically transported 600 feet to a large
storage bin next to the Ames Municipal Electric Co.
In the above described Ames system the RDF yield is approximately 70
percent of the raw MSW input and the resultant fuel has an ash content
typically less than 10 percent. The fluff RDF product was originally co-
fired with pulverized coal in a modified Combustion Engineering tangential
boiler. The main boiler modification was addition of a bottom burning dump
grate to improve burnout of bottom ash. Current operation co-fires the RDF
with pulverized coal (PC) in a B&W boiler specifically designed for f-RDF co-
firing. Note that the CE unit was originally designed to fire pulverized
coal rather than the PC/RDF mixtures.
Figure 6-4 illustrates a slightly different procedure for producing
fluff-RDF. This procedure uses trommel screens as opposed to disk screens
for size segregation of the shredded material. General operation of the
trommel screen is illustrated in Figure 6-5 and shows that the single trommel
should accomplish the same task as the scalpering disk screen and fine disk
screen combination employed at Ames. The RDF production design shown in
6-7
-------
Raw Inleed
Conveyor .
Primary
/Shredder
Magnetic
/ Separation
Vibrating
Pan
Scalping
Disk Screen
I
00
Separation-
Zone Air - ""
Classifier
Fine Glass, Gril, & Finer Fiber
Magnetic
Separation
/ Heavy Fraction
Q ^ to Landfill
Iron Fraction to Storage
Figure 6-3. Ames, Iowa, Resource Recovery System for production
of fluff RDF (f-RDF).
-------
Ferrous Metal
Municipal
Solid Waste
Receiving
Residue
Fuel Storage
Figure 6-4. Fluff RDF production system. Illustration is for system
developed by Combustion Engineering.
-------
Municipal
Solid
Waste with
Ferrous Metal
Removed
en
J
Residue
Size: Under 1/2"
Non-Combustibles
Heavies
Sire: 1/2" to 2"-4"
Combustibles &
Non-Combustibles
Oversize
Combustibles
Landfill Processing
* *
Landfill Fuel
Storage
Secondary
Shredding
Fuel Storage
Figure 6-5. Trommel Screen for RDF size segregation.
-------
Figure 6-4 is being employed in four new facilities with scheduled operation
to begin in 1987, 1988 and two in 1989. For each of those planned
installations the fluff-RDF will be the primary fuel and will be burned in a
spreader stoker, travelling grate boiler system. That combustion system will
be discussed in Section 6.3.2.
6.1.4 RDF-4 or Powdered RDF (p-RDF)
The best known example of p-RDF is ECO-FUEL™; however, it is reported
that Combustion Equipment Associates, the ECO-Fuel producer, is no longer
in business. As defined in Table 6-1, p-RDF involves processing the waste
into a powder form with 95 weight percent passing through a 10-mesh screen.
The mechanical requirements for producing such small size RDF are similar to
those for producing fluff-RDF but include a milling step after air
classification.
Milling operation will generally require predrying of the waste. In the
ECO-Fuel process hot exhaust gas from a process heater is substituted for air
in the "air classifier" converting that device to a dryer as well. Drying
continues during the ball milling process. The resultant fuel is designed to
have a moisture content on the order of 2 to 3 percent and an ash content on
the order of 10 to 12 percent. The higher heating value of RDF-4 is expected
to be 7500 Btu/lb as compared to approximately 5800 to 6000 for fluff RDF.
The use of p-RDF would be for co-firing in pulverized- or cyclone-fired coal
fired utility boilers.
6.1.5 RDF-5 or Densified RDF (d-RDF)
There have been a number of demonstration programs which produced
pelletized or briquetted RDF for use in spreader stoker boilers with
travelling or vibrating grates. As noted by Hasselriis (1983), these short
test burns have shown that d-RDF can be successfully burned, with little
effect on the boiler performance. The main impact observed was a drop in
boiler efficiency due to the higher moisture content of d-RDF relative to
coal. The main difficulty with d-RDF is the cost of fuel production and
6-f1
-------
transportation relative to coal. At the current time there are no known
boiler systems continuously operating on d-RDF.
6.2 Current RDF Projects
As of October, 1986 there were 31 active resource recovery projects
which are sufficiently advanced to have an estimated construction completion
date (Mcllvane 1986). Table 6-2 provides a listing of those active projects,
noting the location, the projected operational date, facility size and the
combustion system supplier. As noted in this table, three of these projects
will use fluidized bed combustor to burn the RDF. It is also noted that
these new RDF projects tend to be very large scale operations. With the
exception of the Franklin Project, each project is larger than 500 tons per
day. The Detroit Project is planned for 4000 tons per day.
6.3 Firing Systems in Current RDF Projects
The majority of the current RDF projects will utilize boiler systems
from Combustion Engineering or from Babcock and Wilcox. Each of these
manufacturers is involved with 5 new RDF facilities. Combined, these two
manufacturers represent 81.4 percent of the active RDF projects (tonne per
day basis). For the projects using B&W boilers "at least three (San Marcos,
Palm Beach and Biddeford) will employ combustion systems supplied by Detroit
Stoker. The CE facilities will use firing systems designed by CE. The
following sections describe those firing systems. As will be shown there are
major differences in the two firing system design philosophies.
6.3.1 Detroit Stoker RDF Firing Systems
Detroit Stoker company has manufactured hardware for burning non-
traditional fuels such as wood, sawdust, bagasse, etc. since early 1940's.
Firing systems developed for burning RDF result in semi-suspension burning
where the fuel is injected through wall ports into the furnace. The fuel
partially burns in the suspension phase with larger material falling to the
grate for burnout on the fuel bed. Figure 6-6 shows the Detroit Stoker air-
6-12
-------
TABLE 6-2. ACTIVE* RDF PROJECTS
Project Name and
Location
Bay Area Resource Recovery
Redwood City, California
San Marcos Resource Recovery
San Marcos (Oceanside). CA
Mid Connecticut Hartford Project
Harford, Connecticut
Wilmington-New Castle Project
Wilmington, Delaware
Collier County Resource Recovery
Naples, Florida
Palm Beach Solid Waste Authority
Palm Beach County, Florida
Honolulu Resource Recovery (H-Power)
Honolulu, Hawaii
Biddeford Resource Recovery
Biddeford, Maine
Detroit Resource Recovery
Detroit, Michigan
Mankato Project 1 and 2
Mankato, Minnesota
Redwing Resource Recovery 1 and 2
Red Wing, Minnesota
Penobscot Resource Recovery
Bangor, Orrington, Maine
Projected
Facility Startup
1990
1989
1987
1986
1989
1989
1989
1987
1989
1987
1987
1988
Size and
Boiler Mfgr-
3600 TPD
C-E(D
800-1600 TOP
2000 TPD
C-E
720 TPD(3)
Vicon/ENERCON
600-900 TPD
Westinghouse/
Goetaverkin in
circulating
fluid bed.
3000 TPD
B&W
1800-2000 TPD
C-E
600 TPD
B&W
3300 TPD
C-E
2@470 TPD each
B&W
20470 TPD each
B&W
721 TPD
KTI
Active implies that project has progressed to the point of having
projected facility start-up data. It does not imply that construction
has necessarily begun.
6-13
-------
TABLE 6-2. ACTIVE RDF PROJECTS (CONTINUED)
Project Name and
Location
Projected
Facility Startup
Size and
Boiler Mfgr.
Erie Resource Recovery
Erie, Pennsylvania
Lubbock Resource Recovery
Lubbock, Texas
Petersburg Resource Recovery
Petersburg, Virginia
Southeastern Tidewater Energy Project
Portsmouth, Virginia
Franklin Project
Franklin, Ohio
1987
1987
1986
1987
1987
600 TPD
Uestinghouse/
Goetaverken in
circulating
fluid bed.
500 TPD
Westinghouse/
O'Connor
650 TPD
2000 TPD
C-E
150 TPD
110 TPD RDF to
be burned in
fluid bed
combustor
' Combustion Engineering
(2) Babcock and Wilcox
(3) Combined MSW/RDF system
-------
CTl
tfl
Figure 6-6.
Side sectional view of Detroit Rotograte Stoker equipped
with Detroit air swept refuse distributor spouts.
-------
swept fuel distributor system coupled with a travelling grate (Detroit
Rotograte Stoker). Figure 6-7 shows the same RDF injection system coupled
with the Detroit Hydrograte system.
The traveling grate illustrated in Figure 6-6 is based on the design
developed for stoker coal firing. The grate forms a continuous loop driven
by sprockets at the front and rear of the boiler. As the grate travels from
rear-to-front the ash layer thickness increases. Ash is dumped from the
grate at the front end of the boiler. The Detroit Hydrograte™ illustrated
in Figure 6-7 has an inclined, water cooled grate which is intermittently
vibrated to shift the fuel bed (and ash) toward the discharge at the front
end of the grate. This grate design is still under development for use with
RDF systems but has not be.en incorporated into actual operating systems.
With both grate systems, there is a single plenum for underfire air.
Openings in the grates are designed to provide a nearly uniform spatial
distribution of underfire air.
A critical component of the design approach is that the fuel injection
hardware can provide a thin bed of RDF which is uniformly distributed on the
grate. The ash layer thickness will increase from the rear to the front of
the boiler but the burning RDF layer thickness will be essentially constant.
If that goal is achieved, spatially uniform underfire air addition would
result in spatially uniform heat release and excess air levels in the
furnace. To accomplish that objective, Detroit Stoker has developed a system
comprised of the "Detroit Refuse Feeder" and the Detroit air swept, rotor
distributor spout. The Detroit air swept, rotary distributor spout is
illustrated in Figure 6-8. A stream of air impinges on the RDF as it falls
from the refuse feeder, injecting the fuel into the boiler. A rotating
damper is used to modulate the flow of distributor air- With the damper
fully open, RDF is blown toward the rear of the grate. With the damper
closed, RDF tends to fall near the front of the grate.
A significant portion of the RDF will burn in the suspension phase. The
suspension burning and fuel injection scheme allows the RDF to impact the
lower furnace walls. To prevent slagging the lower furnace walls in boilers
6-16
-------
Figure 6-7. Detroit Stoker Hydrograte with water cooled, vibrating,
continuous-discharge ash-discharge grates.
6-T7
-------
Fuel
Outlet
Inlet
Rotating
Damper
Fuel Feed
Figure 6-8.
Detroit air swept refuse fuel distributor spout
arranged with motorized rotary air damper.
6-1 a
-------
designed for RDF firing are bare-tubed, not refractory clad as in mass-burn
systems.
As shown in Figures 6-6 and 6-7, Detroit Stoker incorporates several
elevations of overfire air ports on both the front and rear walls of the
boiler. One elevation of overfire air ports is provided below the fuel
injector level. Unlike the systems for MSW firing the furnace walls are
straight. Thus, the overfire air must penetrate across the entire plan view
dimension of the furnace. Design of the overfire air jet system has evolved
over the years. Detroit Stoker is currently considering development of flow
modeling capabilities to optimize their designs.
6.3.2 Combustion Engineering RDF Firing System
The RDF firing system employed by CE is based on a spreader stoker/
horizontal travelling grate design. Figure 6-9 illustrates the overall
firing system configuration while Figure 6-10 illustrates RDF distributor
hardware. As shown, RDF is injected into the furnace by impinging a high
pressure air jet on the fuel as it falls from the feed shoot. The
distribution air nozzle is adjustable which provides control over the spatial
distribution of RDF on the grate.
The grate itself travels from the rear to the front of the boiler with
ash dumping at the front of the boiler. As shown in Figure 6-9, the CE
design incorporates multiple undergrate air compartments with a siftings
screw conveyor in each compartment. The air flow to each undergrate air
plenum is individually controllable. Thus, the air flow distribution through
the grate can be adjusted to match the RDF flow pattern developed by the fuel
feed system.
Overfire air addition is accomplished through a tangential entry system
characteristic of CE utility boilers. The tangential overfire air ports are
located well above the fuel injection elevation. The design is obviously
influenced by CE's utility boiler design philosophy, and by CB experience
with burning other waste fuel on traveling grates but has also been optimized
6-T9
-------
RDF
Distributors
Grate Surface
Drive Shan
Undergrate
Air Compartment
Tangential
Overfire Air
Water
Seal
Idler Shaft
Sifting Screw
Conveyor
Figure 6-9. Combustion Engineering continuous ash discharge type
RC Stoker for RDF.
6-20
-------
Fuel Flow
Adjustable Nozzle
5% Total Combustion
Air at 30" w.g. Pressure
Wear Liner
Adjustable Retention
Plate
Figure 6-10. Combustion Engineering pneumatic RDF distributor.
6-2T
-------
through extensive cold flow modeling studies. The first of the CE-designed
RDF facilities is scheduled for completion in 1987.
6.4 References
Hasselriis, F. "Burning Refuse-Derived Fuels in Boilers". Part IV of
Thermal Conversion Systems for Municipal Solid Waste, H. L. Hickman,
Noyes Publication, N. J. 1983.
Mcllvane Company. "Waste Burning Projects and People", The Mcllvane
Company, Northbrook, Illinois, October 1986.
6-22
-------
7.0 CURRENT PRACTICE IN STARVED AIR (TWO-STAGE) COMBUSTORS
The literature on municipal waste combustion tends to use the terms
"small system," "modular combustor" and "starved-air combustor"
interchangeably. In the current study a distinction is drawn between one-
and two-stage combustion systems regardless of system size or modular nature
of the manufacture. Single-stage systems which operate under excess air
condition through the entire combustor were discussed in Chapters 5 and 6.
The current chapter discusses two-stage systems where the first stage
operates under fuel-rich (starved-air) conditions.
7.1 Starved Air Technologies
The term "modular combustor" implies that the unit is constructed at the
manufacturer's facility and shipped as a module(s) to the installation site.
Initial development of this type of MSW system came as an advancement to the
small batch-fed municipal waste combustors used to burn waste from hospitals,
stores, restaurants, etc. The first facility to combine a capability for
continuous operation and energy recovery was installed in 1976 by Consumat
Systems, Inc. in North Little Rock, Arkansas. Table 7-1 is taken from a
report by Hopper (1983) (with updates) providing a select list of starved-air
systems in operation in the United States. As illustrated by the data in
this table, Consumat is the dominant system vendor for starved-air systems.
Other significant suppliers included in this table are Synergy/Clear Air,
Environmental Control Products (ECP), and Scientific Energy Engineering
(SEE). The Hopper report was published in 1983 and lists sixteen
manufacturers of mass-burning starved air systems. Figure 7-1 was provided
in the Hopper" report to describe the evolution of those U.S. vendor
companies. A cursory examination indicates that in 1986 the proliferation of
companies listed in Figure 7-1 have begun to contract. The Mcllvaine
Company's Market Research report on waste burning projects (Mcllvane, 1986)
indicates that there are eleven active projects (in planning or under
construction) involving starved air system greater than 40 TPD. Six of those
systems are Consumat projects. Clear Air, Inc. and Ecolair, Inc. each have
two active projects while Synergy is involved in one project.
7-T
-------
TABLE 7-1. SELECTED DATA ON SMALL-SCALE U.S. SYSTEMS USING THE STARVED-AIR DESIGN
Location
Si loam Springs, Ark.a
Blytheville, Ark.a
Groveton, N.H.
North Little Rock, Ark.
Salem, Va
Jacksonville, Fla.c
Osceola, Ark.
Genesee, Mich.
Durham, N.H.
Auburn, Maine
Dyersburg, Tenn.
Windham, Conn.
Crossville, Tenn.
Cassia County, Idaho
Batesville, Ark.
Park County, Mont.
Waxahachie, Texas^
Miami Airport, Fla.
Portsmouth, N.H.
Red Wing, Minn.
Cattaraugua County, N.Y.d
Miami , Okla.
Oswego County, N.Y.
Pasagoula, Miss.
Oneida County, N.Y.
Tuscaloosa, Ala.
Hampton County, S.C.
Carthage, Texas
Center, Texas
Barron County, Wisconsin
No. of
Modules
2
(2)
2
4
4
1
2
2
3
4
2
4
2
2
(2)
2
2
2
4
2
3
3
4
3
4
(4)
(3)
1
1
2
Capacity '
( ton/day) ,
Each
Module
10.5
(36)
12
25
25
48
25
50
36
50
50
25
30
25
50
36
25
30
50
36
37.5
36
50
50
50
(75)
75
40
40
40
Date of
Startup,
Past or
Projected
6/75
8/75
Unknown
8/77
9/78
1978
1/80
2/80
9/80
4/81
8/81
8/81
12/81
1982
1982
1982
1982
1982
1982
1983
1983
1983
1983
1983-84
1983-84
1984
11/85
2/86
10/86
10/86
Capital Cost
$106)
Unknown
Unknown
Unknown
1.45
1.9
Unknown
1.1
2
3.3
3.97
2
4.125
1.11
1.5
1.2
2.321
2.1
Unknown
6.25
Unknown
5.6
3.14
Unknown
6+
Unknown
13
Unknown
Unknown
Unknown
Unknown
System Vendor
Consumat
Consumat
ECP
Consumat
Consumat
SEE
Consumat
Consumat
Consumat
Consumat
Consumat
Consumat
Env. Services Corp.
Consumat
Consumat
Consumat
Synergy/Clear Air0*
Synergy/Clear Air^
Consumat
Consumat
Synergy/Clear Air**
Consumat
Consumat
Unknown
Unknown
Consumat
Consumat
Consumat
Consumat
Consumat
Energy Market
Allen Canning Co.
Chrome Plating Co.
Groveton Paper Mill
Koppers
Mohawk Rubber
Unknown
Crompton Mill sa
Unknown
University of New
Hampshi re
Pioneer Plastics
Colonia Rubber
Kendall Co.
Crossville Rubber
J. R. Simplot
General Tire &
Rubber
Yellowstone Park
International
Aluminum Co.
Miami Airport
Pease Air Force Base
S.B. Foote Tanning
Cuba Cheese
B.F. Goodrich
Armstrong Cork
Unknown
Griffis AFB
B.F Goodrich
Westinghouse
Tyson Foods
Holly Farms
Twin Tower Cheese Cn.
Northern States Power
ro
a System now shut down and equipment removed.
c System now shut down.
d Synergy and Clear-Air are now separate companies, each marketing its own system
0 Updates as reported by Consumat.
-------
GORDON HOSKINSON
HO&KIN
Gl ORGI HOWVI RS
SON I IO
HOSKINSON SOUlHlRN
MtCftOIRON
tCf OIV
l»
• •1
ICP
|!9»tl
ANDf ftltN
Ci
AA
KIM
MIDUCftOft INC
1
I
|
cnots
-^
I
cgt
CIIA
N AIM
i"
UCWIllIN
• v»
ClI
1
r
AIR
HI HQ
AH t
»'t
1
INC
Horn
If -
m -^
Mom
VON noit
t,NC
»U»
1
HOVHINSON flOtt
IHl
1 lOWt RS IVI (H
SMOKAlROl
1
CARRIER COA1IECM
WASH COMauSllON IIMOKAIROII
•«- — — — — — 1 RNII Ul Ml
• —
COM
119
«UIE» HOSKINSON ^
HOSKINSON
& ASSOC
"""
AUGER
CONIROl "^ ""
i
(CIINIIMC ENERGY
INGINEIHING ISEEI
CCP .
|I9
I9|
SCUNIIIIC
(NERGV
ENGINEERING
HO^KINSUN
-Ml-
(19701
1
• EllEVCO
• EUEV CO
^- - -
PANV
—
1
C(N( WHIIt
CONSUUAt
tvttlus INC
IH7II
COUIRO
IIA'E 10 >l
GENE WHIIE
SUNIEAM ICOH
4IROI
IEARIT »0 II
CONiuMAI
ItiMM*
0
ENE WHIIE
U S IMEITJNO
FUHNACt CO
(SMOKATROl)
I ' *?•)
^ INTERNAIIONAL
SUNUAM
ICOMIHOI
(19791
1
tNI ERNAIIONAI
WASIE
INOUSlRlEi
U S SMftliNG
{SMOKAIRO1I
UISCCILANCOUS
OTHERS
• RUL(
CCONOIHCMM
LAMB CAHGATE
SlUOHOt
WAJHtuRN t
GRANGE R
Figure 7-1. Evolution of U.S. vendor companies supplying starved-air
waste-to-energy systems, including persons notably
influencing system designs.
-------
As part of the current study interviews were conducted with Consumat and
Synergy. The Consumat interview was conducted at their manufacturing
facilities in Richmond, Virginia. Synergy was interviewed by telephone. The
interview format and topics of discussion were similar to those used in the
excess air mass burn technology evaluation.
Both Consumat and Synergy are small manufacturing firms without
extensive research capability. Recent developments have primarily involved
improvements to the mechanical and operational aspects of their product line.
Both of these companies build the starved-air systems in a modular fashion.
Standard designs have been developed for various size requirements and are
offered as multiple modules to meet the demands of a particular project.
7.2 Consumat Systems
Consumat Systems, Inc. has developed a wide range of standardized,
starved air, municipal waste combustor designs with continuous ratings from
8.6 to 100 tons of municipal solid waste per day- Figure 7-2 illustrates the
standard Consumat module. A front-end loader is used to place a batch charge
of MSW into the automatic loader. A hydraulic ram and charging gate assembly
injects the MSW into the .lower chamber. This lower chamber can be thought of
as the first stage in the two-stage system design. The fuel is slowly moved
from the front to rear of the first stage by a series of hydraulic transfer
rams which are shown in the photograph in Figure 7-3. Holes in the center of
the transfer rams are used to provide a controlled quantity of air to the
primary chamber. Under normal operating conditions it will require
approximately 12 hours for the solid waste charge to traverse from the first-
stage entry to ash dump at the end of the chamber.
The quantity of air introduced to the first-stage defines the rate at
which the mass burns as well as the quantity of gaseous effluent from the
first stage. Figure 7-4 presents results from a theoretical calculation
showing the variation in adiabatic flame temperature with percent theoretical
air. Note that peak temperatures occur when the fuel/air ratio is near
stoichiometric conditions. Under excess air conditions (more than
7-4
-------
© © © © ©
The above cutaway view of the stan-
dard CONSUMAT- energy-from-waste
module shows how material and hot
gas flows are controlled to provide
steam from solid waste A front end
loader (1) pushes the waste to the
automatic loader (2). The loader then
automatically injects the waste into
the gas production chamber (3) where
transfer rams [4) move the material
slowly through the system. The high
temperature environment in the gas
production chamber is provided with
a controlled quantity of air so that
gases from the process are not burned
in this chamber but fed to the upper
or pollution control chamber (5). Here
the gases are mixed with air and con-
trolled to maintain a proper air fuel
ratio and temperature for entrance into
the heat exchanger (6) where steam
is produced A steam separator (7) is
provided to ensure high quality steam.
In normal operation gases are dis-
charged through the energy stack (8)
When steam is not required or in the
event of a power failure, hot gases are
vented through the dump stack (9).
The inert material from the combus-
tion process is ejected from the ma-
chine in the form of ash into the wet
sump (10) and conveyed (11) into a
closed bottom container (12) which
can then be hauled to the landfill for
final disposal.
Figure 7-2. The standard Consumat module for energy-from-waste,
7-5
-------
Figure 7-3. Internal transfer rams in primary
chamber of typical Consumat facility.
7-6-
-------
4000
UJ
Q
0.
UJ
I-
LL)
_J
LL
Q
LLJ
h-
<
O
o
3000 -
2000 -
1000 -
,0% Moislure
15%
% Excess Oxygen: 0
10
12 14
50 % SA 100
150
200
250
300
% Excess Air:
50
100
150
250
Figure 7-4.
Theoretical temperature of the products of combustion,
calculated from typical MSW properties, as a function
of refuse moisture and excess air or oxygen
(Hasselriis, 1986).
-------
100 percent of the theoretical air requirements) gas temperatures fall with
increasing air addition due to dilution effects. Temperatures also falls as
the air flow is reduced below stoichiometric conditions since the full
heating value of the fuel is not released. In the Consumat System design,
first-stage air flow is substoichi ometric and is controlled to maintain a
first-stage exhaust gas temperature set point. The temperature set point is
typically in the 1200 to 1400°F range which corresponds to first-stage
operation at approximately 40 percent theoretical air.
The first stage essentially functions as a gasifier producing a hot fuel
which is transferred to the second stage. In the second stage, additional
air is added through a series of wall jets. These jets are oriented in
opposed pairs in at least three axial locations in the secondary chamber. It
is important to note that the temperature of the fuel-rich gas leaving the
first stage (1200-1400°F) is above the autoignition point for the gas
composition. Thus, completing the burnout process is a matter of getting air
to the first-stage effluent.
There is no heat extraction from either the first or second stage of the
combustor. Heat recovery does not occur until a location downstream of th.e
secondary chamber. The quantity of air added to the second stage is adjusted
to maintain a given combustor exit gas temperature. The temperature level is
typically in the 1800°F to 2200°F range. In the absence of waterwall heat
extraction, this is equivalent to between 80 and 150 percent excess air at
the second stage exit. Thus, approximately 80 percent of the total
combustion air is introduced as secondary air.
As a result of state and local regulatory requirements relative to PCDD
and PCDF emissions, Consumat has recently increased the physical size of the
secondary combustion chamber. The length and diameter of the chamber have
been adjusted to provide a minimum of one second residence time downstream of
final air addition at a temperature above 1800°F. In addition, an auxiliary
burner is provided in both the primary and secondary chamber. Control of the
secondary chamber auxiliary burner is used to assure that the time-at-
temperature requirement is maintained.
7-8
-------
The control system incorporated into the Consumat System design is now
automated. The main parameter being controlled is the first-stage exit gas
temperature. Three system parameters are used to hold that temperature
constant. These include, in order of priority, the primary zone air flow
rate, the frequency of fuel loading and, finally, a water quench is available
if the temperature should reach excessive levels. As noted earlier, the gas
temperature at the second stage exit is used to control the quantity of
secondary air addition and the operation of the secondary zone auxiliary
burner.
With the exception of furnace exit gas temperature control, operation of
the Consumat combustion components is decoupled from heat recovery operation.
The overall system is designed to operate at full load. If for some reason
steam production is not required, or if a power failure should occur, exhaust
gases from the secondary chamber are vented through a dump stack upstream of
the boiler. In some cases, Consumat utilizes either a steam condenser or a
steam vent for excess steam rather than by-passing the flue gas.
Available pollutant emission data from Consumat Systems indicates
performance commensurate with large, modern waterwall MSW grate systems
(Consumat 1986). Consumat Systems achieve total particulate emission levels
of 0.08 grains per dry standard cubic foot (corrected to 12 percent C02)
without the use of any APCD. The PCDD and PCDF emissions have been measured
from three different Consumat systems and the average 2,3,7,8 TCDD toxic
equivalent emission rate (EPA Method) from these three facilities is 8.0 ng/
Nm3 (see the data base volume entitled "Emission Data Base for Municipal
Waste Combustors." This compares with an emission average of 6.0 ng/Nm3 for
modern mass burn/water wall systems. Typical CO levels (corrected to 12
percent C02) are found to be in the 20 to 50 ppm range.
The probable reason for the low particulate emission rate is low gas
velocities which occur in the first stage. As noted earlier, only about 20
percent of the total combustion air is added in the primary zone. Due to the
large diameter of this chamber, the resultant gas velocity is not sufficient
to entrain particles from the bed for transport to the secondary zone.
7-9 -
-------
7.3 Synergy
Information on the two-stage combustion system being offered by Synergy
Systems, Corporation was obtained through a telephone interview with
Mr. William McMillen and Mr. George Hotti (Synergy, 1986). Figure 7-5
illustrates the Synergy two-stage combustion system which has an external
appearance similar to the Consumat system. There are, however, many
differences in the design and control of the Synergy and Consumat Systems.
One of the obvious differences is that the Synergy system is equipped with a
reciprocating grate instead of transfer rams. The grate was designed by
Mr. George Hotti based on his extensive prior experience with Von Roll in
Switzerland. Air is provided to the primary zone through slots between grate
rows and air holes at the end of grate plates. The distribution of underfire
(or primary) air is achieved through a series of five undergrate plenums,
each of which is individually adjustable.
The amount of air flow added in the primary zone was reported to be
approximately 20 percent of the total air. The primary zone temperature was
estimated to be on the order of 1600 to 1700°F. The primary zone grate
system provides an extremely long (8 to 12 hour) solid residence time. The
bed thickness at the end of the grate was estimated to be on the order of
18 inches to 2 feet.
The primary (and secondary) zone is refractory-lined and is designed to
minimize wall heat loss. Effluent from the primary zone flows into a
cylindrical secondary zone with a tangential entry. The secondary zone is
separated into two regions by a refractory ring. Secondary air is added in a
co-flowing (tangential) direction in the first portion of the secondary
chamber and in the refractory ring. Additional secondary air is added
through radial air inlets in the downstream half of the secondary stage.
Auxiliary burners are provided in both the primary and secondary portion
of the municipal waste combustor. The primary zone burner is used to preheat
that chamber during system start-up. The auxiliary burner in the secondary
7-10..
-------
SECONDARY AIR
! I I
SECONDARY BURNER
PRIMARY AIR
Figure 7-5. Synergy two-stage combustion process.
7'1
-------
chamber is used for preheating as well but its main function is to assure
that furnace exit gas temperature is maintained at the required level.
Synergy has recently developed a control scheme based on continuous
measurement of the gas flow from the primary zone exit. These measurements
will be used to control primary zone air flow rate.
Synergy feels that it is important to control both the stoichiometry in
the primary zone and to preclude temperature peaks in the secondary zone.
The control scheme they have developed is designed to accomplish that
objective.
Figure 7-6 illustrates a.Synergy two-stage combustion system with energy
recovery and shows that no heat is extracted until the waste heat boiler.
The size of the secondary chamber has been set to provide a full two seconds
residence time above 1800°F (after final air addition). This expensive
design change has been made to accommodate anticipated emission regulations.
The municipal waste combustion facility currently being installed at Perham,
Minnesota will be the first Synergy system incorporating all of the above
features. Accordingly, there is not yet any operational data on CO, NOX,
trace metals PCDDs or PCDFs from this type system.
7.4 References
Consumat, 1986. Discussions and Data presented by Consumat personnel
(C. Ziegler, C. Stout, W. Wiley) to VI. S. Lanier and W. R. Niessen,
October 14, 1986.
Hasselriis, F. "Minimizing Trace Organic Emissions from Combustion of
Municipal Solid Waste by the Use of Carbon Monoxide Monitors." 1986
National Waste Processing Conference, ASME Proceedings, p. 129, 1986.
Hopper, R. "Small Scale Systems". Part III in Thermal Conversion
Systems for Municipal Solid Waste, H. L. Hickman, et al ., Noyes
Publications, Park Ridge, NJ, 1983.
i-n-
-------
CONDf N5ME
LEGEND
, - TWO STAGE INCINERATOR
2 - RESIDUE EXTRACTOR
3 - WATER TUBE BOILER
4 - ECONOMIZER
5 - WATER SOFTENER
6 - DEAERATOR-STORAGE TANK
7 .- BOILER FEED PUMP
Figure 7-6. Synergy two-stage combustion saturate
steam system.
-------
Me II vane. "Waste Burning Projects and People". Market Survey by the
Mcllvane Company, Northpark, IN, October 1986.
Niessen, W. R. "Dioxin Emissions from Resource Recovery Facilities and
Summary of Health Effects", OSU Report, 1986.
Synergy, 1986. Discussions by telephone between G. Hoth and W. E.
McMillen with W. S. Lanier, October 17, 1986.
7-14 ..
-------
8.0 COMBUSTION CONTROL OF ORGANIC EMISSIONS FROM MUNICIPAL WASTE
COMBUSTORS
Tests of modern mass burn Municipal Uaste Combustors (MWCs) have
indicated emission levels of chlorinated dibenzo-p-dioxin (PCDD) and furans
(PCDF) on the order of 1 ng/Nm3- That is equivalent to an exhaust gas mass
fraction of a part per trillion. However, older designs such as the mass-
burn refractory system at Philadelphia and even the newer mass-burn waterwall
facility at Hampton, Va. have been found to have emissions as much as four
orders of magnitude higher than those from modern mass-burn systems. At
least part of the high PCDD/PCDF emissions from the Hampton facility may be
traced to system control and operation. Thus, even though modern municipal
waste combustors can be designed and operated with low emissions of trace
organics, they can be designed and operated improperly giving rise to higher
emission levels. The purpose of this chapter is to document these combustion
practices which, when adhered to, are expected to minimize the emission of
trace organics from municipal waste combustors without undue impact upon the
emission of other pollutants.
8.1 Design and Operating Problems - Failure Modes
The design or operating conditions which result in higher emissions of
hydrocarbons (including species such as PCDDs and PCDFs) will be referred to
as failure modes. A clear definition of the potential failure modes will
provide insight into how those conditions can be prevented and help to define
"good combustion practice". Design goals can then be recommended which will
help avoid the failure. In this manner, organic emissions can be minimized.
The establishment of the dominant failure modes must be accomplished by
careful consideration of the cause and effects of system parameters on PCDD/
PCDF emissions. In this study, failure modes were established in two ways.
First, the experience base associated with the designers and manufacturers of
municipal waste combustors was investigated. Each manufacturer contacted was
asked to identify failure modes for their particular system. Few of the
manufacturers had specific cause and effect data. Most relied upon their
8-T
-------
understanding of how PCDD/PCDF and other orgam'cs might be formed or failed
to be destroyed and how their specific system operated. The second approach
relied on an examination of combustors which have been reported to have PCDD
and PCDF emissions on the high end of the range for the current data base. A
comparison of these cases with current practices revealed key differences
which likely contributed to the higher emissions, and which are therefore
considered to be failure modes. The following subsections will discuss the
potential failure modes which have been identified in this study. It should
be noted that those potential failure modes do not necessarily apply to every
design within a given class of combustor.
•
8.1.1 Mass Burn Waterwall Failure Modes
The potential failure modes identified by the manufacturers of mass-burn
waterwall combustors are summarized in Figure 8-1. Starting with the input
of the refuse onto the grate, the failure modes are as follows:
1. Non-Uniform Introduction of Waste: Improper introduction of the
refuse onto the grate will result in clumping or poor distribution
on the grate and hence improper burning.
2. Insufficent Control of Primary Air Distribution: Combustion air
requirements along the bed vary, thus it is important that
underfire air be supplied to several independently-controlled
undergrate plenums.
3. Insufficient Primary Air Pressure^ Proper distribution of the
underfire air requires that a pressure drop be taken across a known
plane - the grate. Consequently, it is desirable that the grate
resistance be sufficiently high to distribute the air flow control.
4. Load Control Systems Allow Closure of Underfire Air: Most modern
furnace load control systems use primary air to modulate the
combustion intensity of the bed. For example, a high heating value
fuel charge will cause a step change in the amount of heat released
8-2
-------
oo
i
CO
IMPROPER LOAD
CONTROL OF AIR
LOW LOAD
OPERATION
CO CONCENTRATION
PEAKS
POOR SECONDARY
AIR JET PENETRATION
OR COVERAGE
VOLATILES
NOT CONTROLLED
UNSTEADY
NON UNIFORM
FEED
LOAD CONTROL ALLOWS
CLOSURE OF UFA DAMPER'
TOO LOW OR HIGH
02 LEVELS
HIGH CO LEVELS
TOC LEVELS ABOVE
ZERO (AMBIENT)
TOO MUCH LOWER
FURNACE HEAT REMOVAL
INSUFFICIENT OFA
NO CONTROL OF PRIMARY
AIR DISTRIBUTION
LOW AIR PLENUM
PRESSURE
Figure 8-1. Mass burn municipal waste combustor failure modes.
-------
on the grate. The heat release rate can be lowered by dropping the
primary air flow. If this control technique is used to the
extreme, then the control system can cause oxygen-starved
conditions to exist in the entire lower furnace.
5. Control of Volatile Heat Release: Whether this item is a failure
mode is a point of controversy with different manufacturers. Some
manufacturers believe strongly that the volatiles released from the
bed must be redirected into hotter combustion zones by the furnace
configuration. Others believe that this is not critical and merely
a refinement.
6. Insufficient Overfire Air: In almost all mass burn designs,
overfire air is used to control flame height and to mix the furnace
gases before they enter the upper furnace. If insufficient
overfire air is employed, then mixing of the furnace gases and the
combustion air may be inadequate allowing the existence of f uel -
rich pockets of combustion products.
7. Inadequate Secondary Air Jet Design: All manufacturers stressed
the importance of furnace mixing to prevent PCDD/PCDF emissions and
generally the designs relied on overfire air jets to promote
mixing. Poor coverage of the furnace flow (e.g. improper locations
or insufficient jets) or low jet penetration (e.g. insufficient jet
momentum) will result in poor mixing of the overfire air and the
gases rising from the grate.
8. Excessive Lower Furnace Heat Removal: After mixing with air, all
furnace gases should be hot enough to ensure that all hydrocarbons
are destroyed. If there is too much heat removal in the lower
furnace, then the temperature at the fully mixed height will be too
low to ensure complete destruction of the pollutants.
9. Low Load Operation: Mass burn systems are not tolerant of load
changes. At very low load, temperatures may fall to the point
8-4
-------
where efficient destruction of organics cannot be maintained.
However, even at moderately low loads, conditions can deteriorate
due to other failures such as insufficient overfire air, poor
overfire jet penetration and lower temperatures. Some
manufacturers suggested that loads less than 85 percent of the
design are inappropriate.
10. Low or Hi gn 03 Level s: If exhaust excess oxygen level is too low,
then oxygen-starved zones may exist within the furnace or a charge
of volatile refuse may momentarily lower the entire furnace volume
to oxygen-starved conditions. If the oxygen level is too high,
then the excess air levels will excessively cool the furnace giving
low temperatures which will affect the destruction of hydrocarbon
species.
The fractional part per trillion PCDD/PCDF emission concentration
measured for some modern mass-burn, waterwall combustors is, very likely, the
result of the strong emphasis placed upon the attainment of uniform
combustion conditions for the mitigation of fireside corrosion problems. In-
furnace testing and system re-designs have been undertaken to minimize mal-
distribution of combustion gases particularly in Europe. Because fireside
corrosion rates are dependent upon gas-phase stoichiometry this emphasis on
the attainment and verification of uniform conditions will most probably
minimize the emissions of trace organics by preventing low temperatures or
long life-times for fuel-rich products.
A few MSW systems have been found tc have noticeably higher PCDD/PCDF
and unburned carbon emission levels than those measured in other systems.
These systems were generally designed before there was concern over PCDD/PCDF
emissions. Their design objectives were generally to ensure refuse
throughput and constant steam production, without consideration being given
to the impacts of design on organic emissions. One such system is being
investigated currently to determine how it might be modified and operated to
minimize emissions of these species. The failures of this system to minimize
8-5
-------
emissions and the proposed modifications will serve as a specific example of
potential failure modes of mass burn systems.
The NASA/Hampton refuse-fired steam plant in Hampton, Virginia, began
operation in September, 1981. The plant consists of two combustors each with
a design rating of 100 T/d. The combustors (Figure 8-2) employ six-foot
wide, two-section reciprocating grates with a vertical drop-off between
sections. Underfire air is controlled separately to the two sections of the
grate and overfire air is added under the front wall nose, on the back wall
and on the side walls. The boiler furnaces are top-supported, natural-
circulation, single-pass systems. Silicon carbide refractory is used in the
lower furnace to a height of two feet above the last air injection point.
The primary combustion control system modulates ram feeder frequency and
inlet dampers on the underfire air fan in order to maintain steam pressure.
A secondary loop modulates the grate reciprocating rate to maintain upper
furnace gas temperature.
Measurements of PCDD and PCDF emissions from the Hampton facility
generally established the high end of the emission data base range shown
earlier as Table 2-1. The manufacturer has made major changes in its design
of more recent mass-burn systems and is currently investigating modifications
to the design and operation of Hampton. Several failure modes have been
identified by the manufacturer of the grate used at Hampton as probable
causes for the non-optimal emission performance. The primary problem
identified to date is associated with a design oversight in the automatic
control system. A detailed discussion of the control system developed by
NASA for the Hampton facility is included in a report by Taylor, et al.
(1981). Briefly, the control scheme is designed to maintain a constant steam
production rate. Steam production is controlled by adjusting the modulation
rate of the burning grate and by adjusting a damper on the forced draft (FD)
fan supplying underfire air. With the facility operating at a given steam
production set point, the primary function of the control system is to
account for the variability in MSW heating value and volatility. The control
scheme does an excellent job of maintaining a constant steaming rate. When a
pocket of hi_gh heating value MSW fuel begins to burn, steam production will
8-6
-------
Charging
Hopper
Water
Chargi
Municipal Refuse-Fired
Boiler
Capacity: 100 Tons/Day
Figure 8-2.
Boiler sectional side of NASA/Hampton mass
fired waste-to-energy facility.
8-7
-------
start to rise. The controller will compensate by closing down the FD fan
damper. The control system design oversight was failure to establish a lower
limit for closure of that damper. Failure occurred when the damper
completely closed off the underfire air (for short time periods) leaving the
refuse bed to smolder under starved-air conditions. The manufacturer reports
that plant data shows the FD fan damper continually modulating from fully
open to fully closed.
Compounding the problems caused by the automatic controller is the fact
that this facility was constructed with an insufficient overfire air
capacity. Although the original design called for 30 percent of the total
combustion air to be supplied as overfire air, recent tests indicate that
actual overfire air capability is only 15 percent. This construction (not
design) flaw exacerbated the control system problem and allowed fuel-rich
pockets of gas to escape the high temperature region of the furnace.
Furthermore, although stack gas Q£ level was monitored continuously, the
temporary fuel-rich operation was not detected because of significant air in-
leakage. This in-leakage, driven by the induced draft fan, masked the f uel -
rich operation of the main boiler.
The Hampton facility was the first MSW installation in the U.S. to
incorporate a complete automatic control scheme (Taylor et al. 1980) and it
should be noted that the failure modes noted above are probably correctable.
A lower limit can be set on closure of the FD fan damper, and the system can
be brought up to its original overfire air design specifications. Well -
documented tests before and after modification could provide a data base upon
which to build design guidelines.
8.1.2 Refuse Derived Fuel Spreader Stoker Failure Modes
Refuse Derived Fuel (RDF) fired on spreader stokers have some potential
failure modes that are similar to those of mass-burn units. However, the
unique features of RDF spreader stokers can also lead to different problems.
A summary of potential failure modes that could occur is provided in Figure
8-3. These include the following:
8-8
-------
LOW TEMPERATURE
PATHWAYS
00
OFA DISTRIBUTION
NOT COMPLETE
HIGH CO LEVELS
INSUFFICIENT 0,
-v-
CARRY OVER OF
UNBURNED RDF
RDF DISTRIBUTOR
NOT UNIFORM
•FUEL RICH ZONES IN
UPPER FURNACE
RDF PROPERTIES NOT
MATCHED TO SPREADER
INCORRECT UNDERFIRE AIR DISTRIBUTION
Figure 8-3. Failure modes of RDF spreader-stoker systems.
-------
!• RDF Properties Not Matched to Spreader- Spreader designs are
generally matched to the characteristics of the RDF in order to
disperse the fuel adequately on the traveling grate. Changes in
refuse size and density can adversely impact the distribution of
the RDF.
2. RDF Distribution Not Uniform. The performance of the spreader is
crucial to the successful operation of the combustor. The spreader
must spread the fuel uniformly on the grate or there may exist
zones which are oxygen-starved.
3. Incorrect Underfire Air Distribution. In a fashion similar to mass
burn systems, if the underfire air is not distributed in the same
region as the combustibles, then oxygen-starved zones can occur-
Incorrect air distribution can result if there is too low a
pressure drop across the grate and no ability to control the air to
separate plenums.
4. Incorrect Overfire Air Distribution. If overfire air is not mixed
efficiently with the furnace gases, then oxygen-starved conditions
and non-uniform temperatures can exist in the furnace which can
result in the escape of unburned organics.
5. Excessive Heat Removal. Excessive lower furnace heat removal by
waterwalls will result in low combustion gas temperatures which may
be too low to ensure the destruction of organic intermediates.
6. Carryover of Unburned RDF. The unique feature of spreader stokers
is that RDF is burned in partial suspension. This can lead to an
additional potential failure mode if there is significant carryover
of unburned RDF which may contain precursors to toxic organics.
7. Low-Load Operation. RDF systems are susceptible to failure at
lower loads due to lower temperatures and poorer mixing conditions
similar to mass burn systems.
8-TO-
-------
9- Low or High Oxygen Levels. Just as with mass burn systems, oxygen
levels can be either too high or too low.
The measurements on PCDD/PCDF emissions from RDF waterwall systems are
limited. However, there is some indication that emissions from RDF firing
are higher than those from mass-burn systems (Camp, Dresser and McKee, 1986).
It should be emphasized, however, that the indicated variations may be due to
data base limitations as well as variations in the application of downstream
air pollution control devices. The key differences in RDF spreader stoker
systems that could account for potentially higher emissions are as follows:
• Semi-suspension firing of RDF allows carryover of unburned
materials. This unburned material may be a precursor or allow the
formation of precursors of toxic organics in the cooler regions
downstream of the injection point. Destruction may not be possible
because residence times are less and temperatures are lower.
• Some overfire injection schemes do not appear to be designed to
achieve complete coverage and penetration of the flow. Mixing and
uniformity are achieved by fuel dispersion rather than by stirring
caused by the overfire air jets.
• Unclad steam tubes in the lower furnace increase heat extraction
rates and lower combustion temperatures. This may prevent the
destruction of trace quantities of organic species.
8.1.3 Small, Multi-Staged Modular Unit Failure Modes
Small, modular starved-air combustion systems have design strategies
which are significantly different than their large mass-burn excess air
counterparts. Many or all of the failure modes listed for mass-burn
waterwall systems could also apply to starved-air systems. However, the
equipment generally sold eliminates or modifies many of these potential
failure modes.
8-1-1
-------
Non-steady, non-uniform mass feed and overcharging were identified as
potential failure modes for excess air systems. Typical solids residence
time in starved-air systems is on the order of 10 to 12 hours. Consequently,
the effects associated with any non-uniformity in the fuel feed are damped to
a large extent because of the mass of the fuel bed.
A major concern in mass-burn and RDF systems is inadequate control of
both the amount and distribution of underfire air. In contrast, starved-air
systems are designed to generate a fuel-rich gas in the primary chamber,
which is then combusted further in the secondary chamber. Thus modulation of
the primary air flow is less important in a starved-air system than in mass-
burn or RDF system. Primary air flow rates will, however, affect waste
throughput and ultimately the composition of the residue.
Perhaps the most critical aspect of starved-air systems with respect to
the emission of trace organic compounds is the design of the overfire air
system. Fuel-rich pockets of gas flow from the primary to the secondary
furnace zone. Failure to mix that primary zone effluent with a sufficient
quantity of air could allow trace organics or their precursors to exit the
secondary furnace, providing the opportunity for an increase in their
emissions.
It is possible to quench second-stage burnout reactions through
excessively fast heat extraction. Current design philosophy minimizes heat
loss for approximately one second (mean gas residence time) after secondary
air injection. Thus, excessive reactant quenching is avoided.
PCDD and PCDF emission levels from starved-air systems generally fall in
the low end "of the range of emissions from MSW systems. Thus, it appears
that the current design philosophy for starved-air systems is capable of
maintaining relatively low emissions of these compounds. However, failure
modes can be envisioned which would impact emission levels adversely.
8-T2
-------
8-2 Combustion Strategy for Minimizing Emissions of Air Pollutants
Definitive cause and effect data on the relationships between design and
operating procedures and emissions of organics such as PCDD/PCDF from
municipal waste combustors do not exist. A significant body of data has been
developed by Environment Canada in their tests on the Quebec City combustor.
These tests clearly demonstrate that combustion control significantly reduces
trace organic emissions. Further analysis of that data may provide some of
the required cause and effect information.
It is clear that PCDDs/PCDFs or at least their precursors are organics
which form as intermediates in the combustion process. The vast bulk of
these intermediates will be destroyed as a natural consequence of combustion,
leaving products which consist of primarily C02 and h^O. However, because of
the potential toxicity of some of the combustion intermediates, it is
essential that they be destroyed as completely as possible. Conditions which
favor the formation of intermediates should be avoided. Combustion
strategies can be developed around the simple principle of minimizing the
emission of all combustion intermediates. The available data on low
temperature, catalytic reactions to form PCDDs/PCDFs rely on emission of
organics such as phenols from the radiant section of the combustor (Vogg et
al., 1987).
The strategy for minimizing the emission of combustion intermediates is
based upon a desire to provide an environment which will ensure the
destruction of effectively all gaseous organic species. This environment
requires the presence of oxygen at a sufficiently high temperature.
Furthermore, it is necessary that, to the extent possible, all combustion
gases experience the same environment. Thus, mixing is important because it
will ensure uniform temperatures and composition. Mixing should be rapid as
well as complete in order to maximize residence times for destruction of
intermediates. The attainment of the appropriate temperatures will depend
upon the balance between heat released and heat absorbed. All of the above
will maximize combustion efficiency. Exhaust CO concentration is a good
8-13
-------
measure of combustion efficiency because high levels of CO are normally
associated with poor air/fuel mixing.
The simplistic view of optimization of combustion by the "three Ts",
time, temperature and turbulenci, is not directly valid in this context. For
example, as will be discussed further in the next section, the gas-phase
residence tide should not be considered solely as a necessary reaction time
but rather as a mixing time. Time is required for air and intermediates to
mix but, once mixed at sufficient temperature, the destruction reactions take
place virtually instantaneously. There is no need to hold the mixed gases at
this temperature for a longer time to destroy trace organics. Further,
turbulence on its own is not sufficient to ensure the necessary mixing. Two
separate highly-turbulent stream tubes in the furnace will not mix despite
their high turbulence level unless they come into contact. Thus, mixing of
the furnace gases with air requires that the turbulent air jets be dispersed
throughout the combustion gases. Finally, the definition of a mean
temperature is inappropriate. The trace species that escape the furnace may
experience temperature pathways significantly different from the mean
temperature level. Concepts based strictly on mean times above temperatures
(such as qualifying maximum volumetric heat release rates) do not treat the
low temperature escape mechanisms.
Specific elements of a general combustion strategy can be formulated to
minimize the emission of trace organics from municipal waste combustors.
These elements must first consider the nature of the combustion process in
municipal waste combustion facilities and the practical implications of
current systems. There are also several restrictions that must be placed on
the operating requirements. These must account for impact on the following:
• Availability of equipment
• Emission of other species such as NOX and metals
Economics of waste-to-energy systems
•
The elements of a combustion control strategy based upon good combustion
practices include the following:
8-14 -
-------
• Temperature
• Combustion air (primary and secondary)
• Combustion monitoring
• Automatic combustion control
• Limits on operating conditions
• Verification that operating conditions were adhered to
These elements and the establishment of performance criteria are discussed in
the next sections. It should be noted that the following represents an
initial attempt to develop such a strategy- It has not been verified and
must necessarily be modified as the data base is expanded.
8.3 Temperature
There is a broad distribution of temperatures within a municipal waste
combustor that make mean conditions difficult to define. In Figure 8-4 is
shown temperature profiles in the Steinmueller mass burn combustor at
Stapefeld, for both full load and partial (60 percent) load operation. Mean
times at temperature defined by maximum volumetric heat release rates do not
characterize conditions in the furnace adequately. Three-dimensional heat
transfer, flow and combustion models are generally required to predict, even
crudely, the temperatures experienced by the combustion gases. The
Steinmueller data indicate that a greater than 200°C (360°F) temperature
variation exists across any plane in the mid-portion of the furnace. The
typical temperature variation is from 1000°C (1832°F) at near center-line to
800°C (1472°F) near the walls. At low loads the temperatures are generally
lower (peak 900°C) and the distribution is skewed due to changes in flow and
combustion conditions.
Ensuring that organics such as PCDDs/PCDFs are minimized from municipal
waste combustors requires a knowledge of minimum temperature pathways. The
formation mechanisms of PCDDs and PCDFs are extremely difficult to define
reflecting the complexity of the reactions that occur during the combustion
of refuse. The critical step necessary to prevent emissions of the
intermediate organics is to ensure that the minimum temperature experienced
8-15
-------
00
I
0V
FULL
LOAD
60%
LOAD
Figure 8-4. Temperature distributions in an operating mass-burn combustor.
-------
by the organics, after they have mixed with oxygen, is sufficient for their
rapid destruction. Thus, it is not the mean temperature that must be
defined, but the minimum temperature experienced by the combustion gases.
Combustion control of organics requires that the combustion gases are
mixed with oxygen. Only in the presence of oxygen are the reactions, which
complete the combustion of organics to harmless species such as CC>2 and 1^0,
rapid. The requirements for ensuring adequate mixing are therefore at least
as important as the temperature criteria. The mixing requirements for
combustion air are discussed in the subsequent sections.
Municipal waste combustion systems are generally characterized by
defining a mean temperature at some (often arbitrary) location in the radiant
section. Using that characterization approach, the mean temperature must be
high enough to ensure that the minimum temperature is sufficient to destroy
organic species. The location chosen for the characteristic temperature is
also important. The current evaluation has selected that location as the
"fully mixed height" - the point beyond the final air addition location where
complete mixing should have occurred. In mass burn, waterwall designs and
RDF fired systems, overfire air jets are used above the grate region to mix
air with the gases leaving the burning refuse bed. For most systems, with
good engineering practice for overfire jet design, the fully mixed height
should be on the order of one meter (or less) above or beyond the last
overfire jet. For less traditional systems such as the Westinghouse/O'Connor
combustor the fully mixed height might be in the radiant furnace just above
the rotary portion interface with the stationary furnace wall (assuming that
there is no additional overfire air addition). For the Volund system the
fully mixed height would be above the refractory arch after final air
injection and after the split flows have merged. For other, non-traditional
designs, the location selected for the characteristic temperature should be
based on sound engineering analysis.
The mean temperature must be defined which ensures that the minimum
temperature at the fully mixed height is sufficient to destroy PCDDs/PCDFs
and their precursors. Figure 8-5 is an illustration of the conditions that
8-17 -
-------
00
00
DISTANCE
TO MIX OFA
MINIMUM
1GAS TEMP
REQUIRED
TO DESTROY
° DIOXIN AND
PRECURSORS
FURNACE
CROSS-SECTION
ISOTHERMS
Figure 8-5. Required temperature for destruction of intermediate organics.
-------
exist in a mass burn combustor. At the fully mixed height above the overfire
air jets, there is a distribution of temperatures, as shown by the
hypothetical cross-sectional isothermals. With good grate and overfire air
jet design, the range of temperatures will be kept to a minimum. However, a
200°C temperature variation could still exist at this plane (e.g. see Figure
8-4). Thus, a mean temperature must be selected such that the minimum
temperature is greater than the thermal decomposition temperature required to
fully destroy intermediate hydrocarbons.
The thermal decomposition data obtained by UDRI (Dellinger, 1982) can be
used as the basis for indicating the temperature which will be required to
destroy organics. In Figure 8-6 is provided some of the reference data for a
range of organic intermediates including PCDDs/PCDFs, benzene and chlorinated
hydrocarbons. PCDD/PCDF species are generally unstable above 1300°F although
potential precursors such as chlorophenols are stable to 1500°F.
Hexachlorobenzene is one of the most stable species; it does not totally
decompose below 1650°F although such species (totally chlorinated organics)
are unlikely candidates as PCDD and PCDF precursors. A temperature of
1800°F, which is quite often associated with municipal waste combustion
requirements is roughly 150°F above the thermal decomposition temperatures of
the most stable hydrocarbon intermediates. Thus a mean temperature of 1800°F
at the fully mixed height, should be adequate to ensure that the minimum
temperature is sufficient to destroy the most thermally stable hydrocarbons
at very short residence times. In particular, such a design guideline is
adequate for PCDD and PCDF species along with their potential precursors such
as chlorophenols.
It is possible to demonstrate the adequacy of a particular design by
estimating the mean temperature at the mixed height for the design oxygen
levels, the design refuse bed but with clean waterwalls. The use of clean
walls is recognized to be conservative since the existence of an ash layer
will increase the resistance to heat transfer and therefore raise furnace gas
temperatures; however, the characteristics of the layer cannot be defined
accurately. Experimental data could be used to establish appropriate values.
8-t9-
-------
UDRI THERMAL STABILITY
1000 1100 1200 1300 1400 1500 1600
TEMPERATURES (°F)
THEORETICAL CALCULATIONS
- FORMATION/DESTRUCTION FROM
CHLOROPHENOLS (NBS)
10'
to"
O
Q
r
600
i
900
I
1200
I
1500
1600
fEMPERMURE (°F>
Figure 8-6. Thermal decomposition characteristics of selected hydrocarbons.
-------
8.4 Combustion Air
An excess of combustion air is essential for the complete oxidation of
combustible material. Combustion will not be completed, and harmful
intermediate species may form ar,d escape the furnace if oxygen is not
provided in sufficient quantities at the appropriate location. Four factors
concerning combustion air are important; they are as follows:
• The total amount of air
• The distribution of primary air
• The distribution of overfire air
• The verification of appropriate air distribution
Each of these aspects associated with combustion air will be discussed in the
following sub-sections.
8.4.1 Total Air Requirements
The first aspect involves the amount of total combustion air which is
necessary for complete combustion. There can be both too much, or too little
air. With too little air, the furnace will be oxygen-starved either
throughout the furnace or in localized zones. The total air flow must be
sufficient (1) to ensure that there are no localized zones which are oxygen-
deficient and (2) to dampen out surges due to excessively volatile refuse.
In other words, there must be sufficient excess oxygen available to dampen
out spatial and temporal variations in fuel composition to avoid the
possibility of oxygen-deficient zones. On the other hand, too much air can
excessively cool the furnace. Adiabatic flame temperatures (see Figure 8-7)
are hottest at stoichiometric (zero percent excess oxygen) conditions and
fall off with increasing excess air due to the dilution of furnace gases with
air that must be heated.
The range of excess air that is appropriate can be established based
upon current practice and experience. Mass-burn waterwall systems have been
found to operate best with flue gas oxygen levels between 6 and 11 percent.
8-21
-------
4000
CD
ro
PO
LLJ
Q
Q.
5
Q
UJ
<
O
3000 -
2000 .
o 1000 .
,0% Moisture
15%
30%
50
% SA 100
150
200
250
300
% Excess Air:
50
100
150
250
Figure 8-7. Theoretical temperature of the products of combustion,
calculated from typical MSW properties, as a function
of refuse moisture and excess air or oxygen (8.9).
-------
Operation with less than 6 percent may be acceptable for certain situations
such as extremely well- mixed mass burn system. It is reported that the
Westinghouse/0'Connor system and the Enercon/Vicon systems operate at lower
excess air levels and still achieve relatively low trace organic emissions.
Similar levels (6-11 percent) are current practice for small modular two-
stage systems. These excess oxygen levels correspond to excess air levels of
between 40-100 percent. RDF-fired systems are typically designed to operate
at lower excess air levels, taking advantage of the reduced variability in
fuel characteristics.
8.4.2 Primary Air Requirements
The second aspect relating to combustion air is the proper distribution
of primary or underfire air. For mass-burn and RDF systems, primary air is
introduced through the grate and into the bed of burning refuse. For multi-
zone systems, the primary air is introduced to the first zone. The key
requirement for the primary air system is to get the primary air to the
location where burning is taking place in the refuse bed.
For mass burn grate systems, the primary air should be introduced
through a number of separately controllable plenums underneath the grate. In
current practice, between four and six primary air control regions are
employed along the grate in order to have the capability of distributing the
air to the active burning zones. A high pressure drop across the grate is
desirable to ensure that the bed region supplied by each plenum is completely
covered with air and there is minimal bypassing around thicker regions of
refuse.
Both the quantity and the distribution of underfire air are important
parameters. The quantity of underfire air will directly impact the burning
rate of the MSW on the system grate or hearth. Therefore, the quantity of
undergrate air - along with grate speed - may be used as a quick response
trim parameter for the steam production rate. Over long averaging times, the
steaming rate is obviously controlled by the MSW feed rate while the quantity
and distribution of undergrate air will 1) control the location of MSW
8-2-3
-------
burnout and 2) the mass averaged excess air of combustion products leaving
the lower furnace- Generally, the distribution of underfire air should be
such that good carbon burnout is achieved and the main flame is concentrated
on the burning grates. Air flows to the drying grates and finishing grates
are generally minimized, consistent with moisture content in the fuel and low
organic content in the bottom ash.
For RDF fired spreader-stoker units, primary air should be introduced
through the traveling grate to correspond to the distribution of RDF. For
example, some RDF distributors attempt to spread the refuse uniformly over
the grate and thus, the air must be distributed uniformly also.
Alternatively, some distributors throw the refuse toward the end of the
traveling grate and the refuse is carried to the other end by the grate
movement. In this instance, separately-controlled underfire air plenums are
necessary to ensure proper primary air distribution. Similar to mass-burn
systems, a high pressure drop across the grate may aid in the uniformity of
air flow through the refuse bed.
Finally, for smaller multi-stage starved-air combustors, primary air
control is not as critical as in larger units. The primary units are small
enough that, in general, only the amount of primary air needs to be
controlled. Individually controlled underfire air plenums are useful but not
necessary. However, uniformity of air flow across the bed is still crucially
important and can be accomplished most effectively by high grate pressure
drops.
8.4.3 Distribution of Overfire or Secondary Air
The most critical requirement for complete combustion of the unburned
material exiting the primary zone is mixing with air before the temperature
drops below the level required to destroy the organic intermediates. In
almost all current municipal waste combustion designs, secondary air is used
to aid the lower furnace mixing. The correct design of the overfire air
injection scheme is critical to the prevention of unmixed zones and
minimizing the emission of trace organics. With proper overfire air
8-24-
-------
injection schemes, particle carryover can be reduced, the flame height can be
controlled and the furnace gaseous concentrations can be made to be more
uniform.
For any combustor type, the overfire air must be introduced in such a
way as to cover the entire furnace flow. Only with complete coverage of the
flow can the adequate mixing of the furnace gases be ensured. Mixing is a
function of how well the injected overfire air contacts the gases rising from
the burning refuse bed. In order for "good" mixing to occur, the overfire
air jets must penetrate into the gas flow at well distributed locations and
they must entrain as much of the furnace gases as possible within a short
distance downstream from the injection location. For mass burn systems, the
normal overfire air configuration involves rows of high velocity air jets on
the front and rear walls. Table 8-1 and Figure 8-8 provide a summary of
typical overfire air jet designs for large mass-burn systems as provided by
manufacturers for new systems. The configurations all use front and rear
walls and multiple high velocity jets to completely cover the furnace flow.
Refuse-derived-fuel-fired spreader stoker designs also employ overfire air
jets to completely cover the furnace flow. For example, Detroit Stoker
systems have OFA jets on all four walls below the height of the fuel
distributor while Combustion Engineering designs use tangentially-directed
high velocity jets above the distributor level. Again, the intent is to
achieve complete coverage of the flow with penetrating high velocity jets.
Current designs of small multistage mass-burn systems and water cooled
rotary combustion units also use overfire air to mix the furnace gases.
Again, the design and operating goal is to achieve flow coverage and
penetration. However, different overfire air injection schemes can be used
because of different furnace configurations and smaller dimensions. Many of
these systems employ duct injectors which achieve the coverage and
penetration by injecting air at high velocities into the flow or from the
center line through smaller multiple holes.
8-25 ..
-------
TABLE 8-1. OVERFIRE AIR JET CONFIGURATIONS FOR
LARGE MASS BURN INCINERATORS
Manufacturers
Deutsche Babcock
(Browning Fern's)
Steinmuel 1 er
Von Roll
(Signal Resco)
Widmer & Ernst
(Blount)
Martin
(Ogden Martin)
Detroit Stoker
Riley Takuma
Combustion
Engineeri ng-db
Volund
OFA
% of Total Air
20 - 25
40
30
30
20 - 40
40
30 - 40
20
40 - 50
OFA
Configuration
Rows on Front and
Rear Walls
Side Wall Aspiration
Rows on Front and
Rear Walls at Nose
Rows on Front and
Rear Walls at Nose
Front and Rear Wall s
Interlaced
Offset Vortex
Front and Rear Wall s
Offset
2 Rows Each on Front
and Rear Walls
Rows on Front and Rear
Walls at 3 Levels
1 Row Front Wall
2 Rows Rear Wall
1 Row Side Walls
Perforated Ceramic
Plates-Side Walls
1 Row Back Wall
Controlled Air at
Feed and Exit of
Rotary Kiln
Jet Velocity
and Configuration
100 m/sec
70% penetration
80 m/sec
600 mm w.g.
80 mm dia.
50 m/sec
50 mm dia.
1 m separation
80 m/sec
70 mm dia.
50 - 100 mm dia.
450 mm w.g.
(Data Not
Available)
(Data Not
Available)
High Velocity
750 mm w.g.
Varies
8-26
-------
II
Figure 8-8. Typical designs of overfire air systems,
8-27
-------
The need for secondary air to completely cover and penetrate the flow
implies a number of considerations for the design and operation of the
injectors. The injector parameters of importance include the following:
0 Quantity of overfire air
t Number and location of injectors
0 Injector velocity
0 Injector spacing
0 Injection angle
0 Injector shape
0 Air supply pressure drop
These parameters are related to each other and to the design of the furnace
configuration. The optimum overfire jet configuration would involve numerous
high-velocity jets with correspondingly high pressure drops. However, the
pumping energy requirements (as indicated by the pressure drop and quantity
of air) may become excessive and the auxiliary power requirements will be too
high. Thus, good engineering practice involves achieving penetration and
coverage with minimal pumping energy (i.e. low pressure drop). Since the
design of the overfire air system may not be exact, it is appropriate to
allow for a range of air flow, velocity and potentially even jet direction in
initial hardware design. This would permit on-line optimization to take
place. The overfire air design must be tailored to the overall furnace
design and operation.
Quantity of Secondary Air
The amount of overfire or secondary air is vitally important. For a
given steam production rate and overall excess air level, the quantity of air
available as overfire air must be balanced with the amount of air used as
underfire air. With too little overfire air, the mixing cannot be achieved
since insufficient momentum is available to achieve cross stream mixing. On
the other hand, excessive overfire air can result in thermal quenching due to
the introduction of too much cold air- Current practice has established
20-40 percent of the total air as overfire air as appropriate. Extensive
8-2&
-------
testing using cold-flow modeling in-furnace sampling, and emissions testing
(especially on mass-burn systems) have verified the suitability of these
values. Thus, it appears that the design of overfire air supply with
40 percent of total air capacity and ensuring that the system is operated
with at let-st 20 percent overfire air is appropriate to minimize emissions of
trace organic species.
Injector Diameter and Velocity
The diameter of the overfire air injectors and the initial jet velocity
can be determined based on the required quantity of overfire air and the
furnace configuration. Design procedures exist to determine jet penetration
depths and trajectories. The appropriate design goals would include:
• The trajectory of the jet should cross the center!ine of the
furnace duct before reaching two injector diameters downstream.
• The depth of penetration of the jet should be at least 90 percent
of the furnace depth (in opposed wall or centerline injectors the
penetration should reach 45 percent of the depth).
Demonstration that the appropriate jet diameter and velocity has been
selected can be based on either design equations or other data such as cold-
flow modeling or furnace flow measurement data.
Number of Injectors and Location
Once the velocity and diameter of OFA injectors have been determined to
achieve appropriate furnace flow penetration, the number and location of
nozzles must be examined to ensure adequate flow coverage. The number of
nozzles is of course related to the quantity of air available for use as
overfire air. There are a number of appropriate configurations that would be
acceptable for achieving the coverage depending upon the furnace
configuration. However, a cross-sectional view of the furnace showing the
8-29
-------
location of injectors is generally sufficient to illustrate the coverage of
the furnace flow.
Air Pressure
The pressure of overfire air supply, AP, is related to the jet velocity,
V by:
+ 460\
V(ft/sec) = 66.3\/ MP (in w.g.)
V y 520 /
where T is the temperature of overfire air. No friction losses are accounted
for in this equation, thus actual jet velocities will be 7-10 percent lower
than predicted by the above equation.
8.4.4 Verification of Appropriate Air Distribution
The appropriate design of an municipal waste combustor will allow
control of the various combustion air streams and will be sufficiently
flexible to allow the air to be distributed to the burning zones. The
current report has attempted to underline the fact that an MSW or RDF
combustor is a combustion system and that the various subsystems must be
carefully integrated. Recognizing these systems aspects, an important
question arises as to how appropriate adjustment of the subsystems can be
verified. Current practice involves use of a variety of indicators
including:
• Visual inspection
• Exhaust concentration measurements and
• In-furnace probing/mapping experiments.
The use of visual inspection is often extremely effective in adjusting air
flow distribution (at least coarse distribution). Such visual inspection can
be used to guide adjustment of underfire air distribution among the various
plenums beneath the grate sections. Exhaust concentrations of oxygen, and CO
8-30-
-------
coupled with visual observation of the flame can be used to adjust overfire
air quantity and distribution. Clearly, experimental verification that there
are acceptably low exhaust concentrations of the spectrum of PCDDs/PCDFs is
proof that the combustor is sufficiently adjusted. To achieve that
acceptable emission level (a level yet to be defined) almost certainly
implies that an appropriate quantity of overfire has been injected and mixed
with the effluent from the lower furnace. A number of combustion system
designers including Steinmueller/Dravo, Martin Systems and Von Roll have used
in-furnace profiling of CO concentration to develop their overfire air
injection scheme and to establish an appropriate air flow distribution in
particular furnaces. The rationale is that CO concentration is a direct
indication of furnace mixing and that peaks in concentration of CO at the
fully mixed height are indicators of poor air distribution. Figure 8-4
illustrate a relatively flat CO concentration profile at full load and a
highly skewed in-furnace CO profile at 60 percent load.
In establishing a new MSW system a combination of the above
verifications is required. Visually guided adjustments and exhaust gas
measurements are essential components. In-furnace CO profiling may not be
necessary for every new unit but is strongly suggested as a diagnostic tool
but it is strongly suggested as a diagnostic tool for correcting air
distribution in facilities which do not achieve sufficiently low trace
organic emission.
If in-furnace profiling is to be performed the CO measurements must be
taken over the entire measurement plane with sufficient spatial resolution to
cover the entire furnace flow. Twenty-five to fifty sampling locations
should be adequate and a variation in CO concentration greater than 50
percent (e.g. from 1 percent to 1.5 percent when corrected to the same oxygen
concentration level) may be indicative of the need to adjust air flows,
although it is essential that this criteria be established more precisely by
extensive field testing and engineering analysis. Despite the current lack
of precision in this technique, CO concentration profiling can provide an
indication of combustion uniformity.
8-31
-------
8-5 Combustion Monitoring Requirements
In the previous sections, the goals for design and operation were
discussed which would be consistent with proper combustion system operation.
In addition to these requirements, the system must be monitored continuously
and controlled so that the system functions in the proper operating mode. In
other words, continuous monitoring and control is required to help prevent a
failure condition. This is particularly challenging when burning the ever-
changing supply of refuse which has a relatively low fuel value.
Fluctuations in the volatile and moisture content of the refuse must be
accounted for properly if the design is to be implemented correctly in
practice. If the system is designed properly and tuned accordingly, then the
monitoring and control system must only ensure that this system remains
within the proper operating envelope.
Three continuous monitoring schemes may be used to ensure that the
system stays in compliance. These include the following:
• Minimum flue gas carbon monoxide concentration (corrected for 63
concentration)
• Minimum and maximum level of flue gas oxygen concentration
• Minimum furnace temperature
These monitoring variables, on their own, are not sufficient to ensure the
system is not emitting PCDDs/PCDFs. In particular, they do not provide
information on the degree of uniformity within the furnace. This is
crucially important to ensuring destruction of intermediates. For example,
low exhaust mean carbon monoxide levels are indicative of the mean
conditions, not the isolated pathways which could lead to part per trillion
emissions of PCDD/PCDF species. On the other hand, once the system has been
proven to have well-mixed conditions by in-furnace profiling or by other
means, then flue gas CO levels might be used to indicate that no significant
8-32-
-------
change has taken place. Field measurements are required to establish the
effectiveness of this type of monitoring.
The levels of these monitoring parameters must be specified along with
appropriate averaging periods. Discussions with manufacturers of new MSW
combustion equipment have provided an indication of the currently achievable
level of control and the appropriate target. Currently, many of the
manufacturers already use some or all of these measurements for monitoring
and control. For carbon monoxide, it was generally agreed that 100 ppm
(corrected to 12 percent C0£) was achievable and appropriate. Many systems
regularly achieve continuous CO levels in the range of 20-40 ppm. However, s
even well-behaved systems can have momentary spikes in CO that can exceed 100
ppm. These momentary spikes due,, for example, to a small burst of volatiles
are not necessarily of concern since a well-designed system will dampen out
surges and destroy intermediates. Also, field and laboratory measurements
have suggested that CO is a conservative indicator of failure. However, if
this CO excursion is too high, or if it persists, then corrective action must
be taken.
Recent data from mass burn systems have shown that well designed and
operated units can achieve low PCDD/PCDF levels from the combustion zone.
For example, the new units at Marion County, Tulsa and Wurzberg and the
modified older design, Von Roll facility at Quebec City, were found to have
low PCDD/PCDF and other trace organic emissions over a four hour average
(sampling time). These same units were found to have a rolling average
exhaust CO levels of less than 50 ppm over the same averaging period. Thus,
this level of CO is proven to be acceptable and achieveable with modern
design and operating practice.
It has been pointed out that the total combustion air flow should not be
too high or too low if the emission of trace organic species is to be
minimized. Current practice has recommended that an exhaust 02 range of
between 6 and 12 percent is appropriate. When the proper excess air level
has been established for a particular unit, then the control system should
operate within a more restrictive range. However, for major corrective
8-33
-------
action such as a system shutdown, it is appropriate to establish levels
outside the 6 to 12 percent range.
Gas phase furnace temperature monitoring is required to ensure no low-
temperature excursions occur. Measurements in the furnace above the fully
mixed height using suction pyrometry to shield radiation can indicate that
the system is being continuously operated in the same thermal environment.
In conjunction with other design and operating procedures which ensure
uniformity, this measurement prevents thermal failure modes. The most
desirable locations are low in the furnace but pyrometer survivability is low
in this region. For this reason an upper furnace location which is related
to the fully mixed height temperature of 1800°F is a reasonable compromise.
Simultaneous measurement at both locations during tune-up would allow the
upper appropriate furnace temperature requirements to be defined.
8.6 System Control
The final element of a strategy to minimize emissions of trace organics
is the control of the combustion system relating to the combustion process.
A control system is required in order to maintain a desired steaming rate
with changing refuse characteristics. The control schemes used by the
various manufacturers are described in Section 5. Many of the new systems
use multiple loop automatic control schemes which change refuse charging
rates, grate feed rates, underfire and overfire air flows in response to
changes in steam pressure and other monitoring variables such as flue gas
oxygen or furnace temperature. The key goal of these automatic control
systems is to avoid failure modes while maintaining steaming rate. For
example, one notable failure discussed in Section 8.1 is the closure of
underfire air flows during a steam excursion which drives the system to an
oxygen-starved condition. Such control system flaws can be prevented by
monitoring other variables such as flue gas oxygen.
In addition to maintenance of steam pressure, the other aspect of
control is the system response to excursions in monitored parameters.
Specifically, one possible response to a CO excursion or for low temperatures
8-34
-------
is to initiate auxiliary fuel combustion via an auxiliary burner. These
auxiliary burners should burn a clean fossil fuel such as natural gas or
light fuel oil with sufficient capacity to dampen out rapidly the results of
the excursion. In this manner the inventory of refuse that remains in the
furnace at the start of the excursion will be burned under appropriate
conditions. A complete halt to refuse feeding should not be necessary unless
the excursion persists. However, the feed rate would have to be reduced so
that the steam rate noes not exceed the desired level.
The other need for auxiliary fuel is for startup. As previously
mentioned, low temperature has been specifically identified as a failure
mode. To avoid this failure mode on startup, auxiliary fuel should be used
to raise the furnace temperature to the operating point before waste is
added. The capacity of the auxiliary fuel system should be consistent with
these needs; current practice suggests that capacity equal to 60 percent of
the heat input is appropriate.
Finally, low load operation is of concern because the combustion
conditions may deteriorate as the firing rate is reduced. Lower firing rates
may result in lower temperatures and poorer mixing due to slower overfire air
injection velocities and lower combustion intensity- A specific load level
below which performance is unacceptable has not been established, and is most
likely system-specific. In fact, special design and operating procedures can
be developed to extend the range of operation for any particular system. For
example, steam jets might be used to supplement overfire air jets during low
load operation. Most manufacturers expressed preference for operating above
80 percent but below 110 percent rated capacity. If lower load operation is
desired, then additional testing such as in-furnace CO profiling, temperature
profiling or stack emission testing to verify low exhaust organics
concentration at the lowest firing rate may be appropriate.
8.7 Minimization of Hydrocarbon Species and Other Pollutants
The previous sections have described a strategy for the design and
operation of municipal waste combustors to minimize the emission of trace
8-35 -
-------
organics. There is concern that its implementation could result in higher
emissions of other pollutants. In particular, the impact on emission of
acids, particulate matter, NOX and metals must be evaluated. It is expected
that there will be little or no impact on emission of acids and particulate
matter but ther3 may be adverse impacts on NOX and metals emissions.
The design and operating techniques for minimizing organic emissions is
not compatible with combustion-zone NOX emission control. High-temperature,
well-mixed excess air conditions favor the formation of NOX from both thermal
fixation of molecular nitrogen and the conversion of fuel nitrogen. The
tradeoff between NOX and CO has been observed for some time for large
furnaces where optimum conditions for CO are not compatible with combustion
control of NOX. A similar tradeoff exists between NOX and hydrocarbons.
However, NOX can be controlled through post-combustion controls, as discussed
in Chapter 4.
Combustion modification techniques such as air staging and flue gas
recircul ation which are designed to moderate flame temperatures have been
shown to be effective for NOX control. For example, Volund Technology
employs flue gas recirculation in its refractory-lined combustors to control
NOX by lowering the primary zone temperature. However, it is presently
unclear whether such control strategies can be applied universally or will be
compatible with control strategies for hydrocarbons. An alternative to
combustion modification are downstream processes which would not interfere
with optimizing the combustion zone for control of organic pollutants. NH3
injection above the combustion zone is currently being installed on new
systems in California (e.g. City of Commerce). Other NOX control strategies
that potentially could be applied include reburning which may also aid in
hydrocarbon control and selective catalytic reduction. More research is
required to define the impact of combustion control techniques for
hydrocarbon control on NOX and NOX control strategies.
The emission of heavy metals may also be adversely impacted by design
and operating practices for hydrocarbon control. The partitioning of metals
between residuals, fly ash and condensible fume depends strongly on the
8-36
-------
temperature and stoichi ometry that the metal bearing refuse experiences in
the burning refuse bed. Data was presented earlier which indicates that
higher temperatures favor vaporization of metals and that stoichi ometry
impacts the condensible fraction. Hence, changes in temperature to higher
values and modification of the air distribution might impact metal
fume emissions adversely- The magnitude of these effects has not been
established. Post-combustion flue gas cleaning may be used to reduce levels
of metals emissions, as discussed in MWC study entitled "Flue Gas Cleaning
Technology."
8.8 References
Del linger, B. et a!., "Laboratory Determination of High Temperature
Decomposition Behavior of Industrial Organic Materials". Proceedings of
75th APCA Annual Meeting, New Orleans, 1987.
Haile, C. L., Blair and Stanley. "Emissions of PCDD and PCDF from a
Resource Recovery Municipal Incinerator." USEPA/RTP, 1983.
Haile, Blair, Lucas and Walker. "Assessment of Emissions of Specific
Compounds from a Resource Recovery Municipal Refuse Incinerator."
Washington, D.C.: USEPA, Office of Pesticides and Toxic Substances, May
1984.
Hasselriis, F. "Minimizing Trace Organic Emissions from Combustion of
MWS by Use of Carbon Monoxide Monitors." Proceedings of the Twelfth
Biennial Conference, 1986 National Waste Processing Conference. ASME,
New York, N.Y., p. 129, 1986.
La Fond, R. K., J. C. Kramlich, W. R. Seeker, and G. S. Samuelsen:
Evaluation of Continuous Performance Monitoring Techniques for Hazardous
Waste Incinerators, Jrnl. Air Pollut. Control Assoc. 35, 658 (1985)
8-37-
-------
9.0 HYDROCARBON CONTROL STRATEGY - SUMMARY
The previously presented information about municipal waste combustion
points to the conclusion that good combustion engineering practices will
minimize the emission of trace organic pollutants from the combustor/boiler
subsystem of a waste-to-energy system. Not only will these practices
minimize emissions, they will also contribute to the operability of the plant
because they will reduce corrosion rates and could improve combustion
efficiency. In addition, air pollution control devices may reduce trace
organic emission levels even lower (see the volume entitled "Flue Gas
Cleaning Technology."
This report has synthesized what are referred to as "good combustion
practices" from: current theories on the basic mechanisms of PCDD/PCDF
formation and destruction in combustion systems, information provided by
manufacturers on the design of waste-to-energy systems, and operating/
emissions data from specific plants. The practices are designed:
1. To limit the formation of hydrocarbons
2. To maximize the destruction of these same compounds and their
precursors prior to the exit of the combustor/boiler should they be
formed
As such, good combustion practice attempts to preclude conditions which
promote formation of PCDDs/PCDFs and their precursors and to ensure that the
environment experienced by the gaseous products of waste combustion will
destroy both PCDDs/PCDFs and their precursors if those compounds have been
formed. These conditions are:
• Mixing of fuel and air to prevent the existence of long-lived fuel -
rich pockets of combustion products
• Sufficiently high temperatures in the presence of oxygen for the
destruction of hydrocarbon species
9-1 -
-------
• Prevention of quench zones or low temperature pathways that would
allow partially-reacted fuel (solid or gaseous) from exiting the
combustion chamber
The appropriate design of a waste-to-energy system alone is not
sufficient to ensure that the optimum combustion environment will exist in
practice. The combustion system must also be operated continually in an
appropriate manner and respond properly to changes in refuse combustion
characteristics in order to maintain proper conditions. Only by examining
all aspects of design and operation of these systems can the emissions of
trace organics be minimized.
This final section of the report will review these elements and indicate
the integration of the elements into a combustion control strategy for trace
organics such as PCDDs and PCDFs. Finally, the remaining uncertainties and
issues will be identified along with recommendations on research approaches
to address them.
9.1 Combustion Practices for Trace Hydrocarbon Emission Control of
Municipal Waste Combustors
Currently, there are no definitive data on the mechanism of formation or
destruction of PCDD/PCDF in municipal waste combustion facilities. The
available data indicates, however, that trace quantities of PCDD and PCDF
species can be formed in the process of burning heterogeneous fuels such as
MSW. Several theories propose that PCDD/PCDF form more easily from certain
similar structure precursors such as chlorophenols and some species in lignin
which are either present in the feed or are themselves formed during the
combustion process. However, once formed, both the PCDD/PCDF and their
precursors can be destroyed if exposed to the appropriate conditions. The
intent of good combustion design practices for municipal waste combustors is
not only to minimize the formation of PCDD/PCDF or their precursors but also
to ensure that escape of precursors from the combustion zone is minimized
because conditions might exist downstream which will allow these precursors
to form PCDD/PCDF (e.g. by fly ash-catalyzed reactions). The appropriate
9-2
-------
conditions for the destruction of PCDD/PCDF and their precursors are that the
species experience a given temperature after being mixed with sufficient
oxygen. The challenge for the designer is that no pocket of material can
escape the system which has not experienced these conditions.
The development of "good combustion practices" to minimize emissions of
hydrocarbons from municipal waste combustors involves three separate
elements:
• Design. Design the system to satisfy several criteria which will
ensure that temperatures and the degree of mixing within the
combustor are consistent with minimizing formation and maximizing
destruction of the trace organics of concern.
• Operati on Control . Operate the system in a manner which is
consistent with the design goal and provide facility controls which
prevent operation outside on established operating envelope.
• Verification. Monitor to ensure that the system is continually
operated in accordance with the design goals.
If all three of these elements are satisfied, then the emission of trace
organics from the combustor of a municipal waste combustors should be
minimized.
This project has identified the components of each of these elements
which make up good combustion practices, and that are expected to be
important in the control of PCDD/PCDF emissions from municipal waste
combustors. In addition, preliminary recommendations have been made on the
values of the individual components. Identification of the elements was
based upon the current design practices associated with combustion systems
that have generally shown low PCDD/PCDF emissions. The combustion control
elements and preliminary component value recommendations are summarized in
Table 9-1, 9-2, and 9-3 for mass burn combustors, RDF-fired systems, and
starved-air combustors, respectively.
9-3 -
-------
TABLE 9-1. GOOD COMBUSTION PRACTICES FOR MINIMIZING TRACE ORGANIC
EMISSIONS FROM MASS BURN MUNICIPAL WASTE COMBUSTORS
Element
Design
Operation/
Control
Verification
Component
Temperature at fully mixed
height
Underfire air control
Overfire air capacity (not
an operating requirement)
Overfire air injector
design
Auxiliary fuel capacity
Excess air
Turndown restrictions
Start-up procedures
Use of auxiliary fuel
Oxygen in flue gas
CO in flue gas
Furnace temperature
Adequate air
distribution
Recommmendations
1800°F at fully mixed height
At least 4 separately adjustable
plenums. One each under the drying
and burnout zones and at least two
separately adjustable plenums under
the burning zone
40% of total air
That required for penetration and
coverage of furnace cross-section
That required to meet start-up
temperature and 1800°F criteria under
part-load operations
6-12? oxygen in flue gas (dry basis)
80-110% of design - lower limit may
be extended with verification tests
On auxiliary fuel to design
temperature
On prolonged high CO or low furnace
temperature
6-12% dry basis
50 ppm on 4 hour average - corrected
to 12% C02
Minimum of 1800°F (mean) at fully
mixed height across furnace
Verification Tests (see text Chapter
8 and 9)
9-4
-------
TABLE 9-2. GOOD COMBUSTION PRACTICES FOR MINIMIZING TRACE ORGANIC
EMISSIONS FROM RDF COMBUSTORS
Element
Design
Operation/
Control
Verification
Component
Temperature at fully mixed
height
Underfire air control
Overfire air capacity
(not necessary operation)
Overfire air injector
design
Auxiliary fuel capacity
Excess air
Turndown restrictions
Start-up procedures
Use of auxiliary fuel
Oxygen in flue gas
CO in flue gas
Furnace temperature
Adequate air
distribution
Recommmendati ons
1800°F at fully mixed height
As required to provide uniform bed
burning stoichiometry (see text)
40$ of total air
That required for penetration and
coverage of furnace cross-section
That required to meet start-up
temperature and 1800PF criteria under
part-load operations
3-9% oxygen in flue gas (dry basis)
80-110? of design - lower limit may
be extended with verification tests
On auxiliary fuel to design
temperature
On prolonged high CO or low furnace
temperature
3-9% dry basis
50 ppm on 4 hour average - corrected
to 12% C02
Minimum of 1800°F (mean) at fully
mixed height
Verification Tests (see text Chapter
8 and 9)
9-5.
-------
TABLE 9-3. GOOD COMBUSTION PRACTICES FOR MINIMIZING TRACE ORGANIC
EMISSIONS FROM STARVED-AIR COMBUSTORS
Element
Design
Operation/
Control
Verification
Component
Temperature at fully mixed
height
Overfire air capacity
Overfire air injector design
Auxiliary fuel capacity
Excess air
Turndown restrictions
Start-up procedures
Use of auxiliary fuel
Oxygen in flue gas
CO in flue gas
Furnace temperature
Adequate air distribution
Recommmendations
1800°F at fully mixed height
80 percent of total air
That required for penetration and
coverage of furnace cross-section
That required to meet start-up
temperature and 1800°F criteria
under part-load conditions
6-12% oxygen in flue gas (dry
basis)
80-110% of design - lower limit
may be extended with verification
tests
On auxiliary fuel to design
temperature
On prolonged high CO or low
furnace temperature
6-12% dry basis
50 ppm on 4 hour average -
corrected to 12% C02
Minimum of 1800°F at fully mixed
plane (in secondary chamber)
Verification Tests (see text Chapter
8 and 9)
9-6
-------
It must be emphasized that the values presented in these tables are
preliminary targets and require verification to ensure their appropriateness.
As such, they can be viewed as an initial basis for developing test plans for
field evaluation and pilot-scale R&D studies. Further study may also
determine that some of the elements are more important than others in
ensuring good combustion.
As with any general principles, the specific design of individual
systems must be considered in applying the recommendations in these tables.
In particular, several combustion systems, such as the mass burn refractory
technologies of Volund and Enercon/Vicon and the mass burn rotary technology
of Westinghouse/O'Connor, incorporate differences from the typical mass-burn
approach that make it infeasible to directly apply some of the
recommendations in Table 9-1. For such systems, parameters such as "fully
mixed height" will have to be defined based on technology-specific
engineering analysis rather than on the general one meter rule suggested for
traditional mass burn systems.
Of course, the final determinant of the performance of each system is
the measured level of trace organics emitted from the system. Whether these
levels indicate acceptable performance will depend on emission levels
established in the facility's permits, state standards or guidance, and any
feder-al guidance or regulation that may be established in the future. If
emission measurements indicate that the performance of a system needs
improvement, in-furnace CO profiling can be used to determine appropriate
adjustments to the air distribution. Specifically, a flat in-furnace CO
concentration indicates sufficient air adjustment and furnace mixing. A
precise definition of "flat" considering spatial and temporal variations is
not yet available and should be developed as part of future test programs.
9.2 Research Recommendations
The "good combustion practices" defined above were derived from an
analysis of the available information which includes little direct evidence
relating to the appropriateness of values recommended in Tables 9-1, 9-2, and
9-7
-------
9-3. Further work is needed to better define and verify these details for
control of hydrocarbon pollutants. This work is required in three major
areas:
1. Definition and verification of target components and levels
2. Mechanisms of PCDD/PCDF formation and destruction from MSW and RDF
fired systems
3. Tradeoffs with control of other pollutants
Figure 9-1 summarizes the flow of information required to establish more
specific design and operating recommendations for municipal waste combustors
to minimize the emission of trace organic species. This report has used an
existing data base to define "good combustion practices" for municipal waste
combustors. In order to define these practices more specifically, better
(more comprehensive) field data must be obtained, our understanding of the
mechanisms of formation and destruction of the species of interest must be
broadened, and the tradeoffs with other pollutants must be established.
Furthermore, since it is not possible to test all municipal waste combustors,
some generalization procedure must be available to extrapolate the data that
is collected on operating combustors to the total population.
9.2.1 Combustion Control Guideline Definition and Verification
There is little information relating directly the emissions of the
pollutants of concern with the design and operational parameters of municipal
waste combustors. Thus, the definition of "good combustion practices" must
be considered preliminary and further definition and verification is
essential . The approach recommended is to examine directly the impact of
design and operating conditions on PCDD/PCDF emissions while monitoring key
performance parameters and other pollutant emissions from representative
municipal waste combustors operating under a wide range of operating
conditions. To establish the required data base these conditions should
include testing beyond the normal systems operating envelope.
9-8-
-------
EXISTING
DATA BASE
MECHANISTIC STUDIES
H DEFINITION OF
BEST COMBUSTION
PRACTICE"
I
IDENTIFICATION OF
RESEARCH NEEDS
i
FIELD EVALUATION
\
OTHER POLLUTANTS
GENERALIZATION PROCEDURE
DESIGN AND
OPERATING GUIDELINES
Figure 9-1.
Research program to establish design guidelines based upon "Good
Combustion Practice".
-------
The tests on operating combustors should be run on units which are
representative of new combustion systems i.e. mass-burn waterwall, RDF-
spreader stoker, two-staged/starved-air, RDF fluidized bed. The selection
criteria for the units to be tested are as follows:
• Representative of the units within a particular combustion system
class
0 Capable of flexible operation in order to generate the information
on cause and effect
• Availability of engineering design and operating data to help data
interpretation
• Availability of in-furnace and stack sampling access
t Willingness of owner/operator to participate in a program of this
type
The focus of these field tests would be to determine the effect of design and
operation on both the emissions of PCDD/PCDF and other pollutants. In
addition, it will be necessary to establish the in-furnace conditions (e.g.
CO profile) which lead to high and low emissions of PCDD/PCDF. Test series
should define the baseline conditions, and the impact of furnace load,
combustion air distribution, firing auxiliary fuel, excess air levels,
startup procedures and, if possible, fuel type. Field tests are expensive
and their utility will be expanded if procedures are available for on line
evaluation of combustion performance factors. Measurements will not be
possible over the total operating range. Spatial and temporal fluctuations
may make the data difficult to interpret. Thus, it is essential that
engineering analysis tools be available to interpret and rationalize the
data.
9-10
-------
9-2.2 Mechanisms of PCDD/PCDF Formation and Destruction
The combustion control strategy would be more valid if it were based
upon a better knowledge of the factors controlling the formation and
destruction of PCDD/PCDF. Several chemical mechanisms have been proposed
(see Chapter 4) but many fundamental physical/chemical issues remain that
should be addressed before a controlling mechanism can be accepted. These
issues include the following:
t Identify conditions which favor formation of PCDD/PCDF from refuse
combustion
• Define the impact of refuse properties on PCDD/PCDF formation
• Define the temperature necessary to destroy PCDD/PCDF species as a
function of gas-phase environment
• Examine the role of downstream condensation and formation
mechanisms
Both laboratory scale gas phase kinetics studies and bench scale refuse
burning studies would be suitable for addressing these issues.
9.2.3 Tradeoffs in Other Pollutants
As indicated in Chapter 4, the combustion control strategy for PCDD/PCDF
may have detrimental impacts on other pollutants. For example, maximizing
temperature and improving mixing may promote NOX formation. Also, improving
the air distribution can change conditions in the burning refuse bed which
can impact the vaporization of heavy metals. These changes could result in
the concentration metals in the small particulate fume that is not easily
controlled by air pollution control devices.
For heavy metals, the key issue is the fate of metals in the refuse as a
function of the combustion conditions. The research needs include a data
9-T1
-------
base on the fate of metals, control of metals by particulate control devices
the Teachability of metals from residuals, and predictive methodologies for
metal partitioning. Carefully controlled bench-scale studies on burning
refuse could provide a significant contribution to our understanding of the
fate of metals during combustion. In the area of NOX, development of NOX
control schemes that are not detrimental to PCDD/PCDF control are necessary
if more stringent NOX levels are enforced by local authorities.
9-12
------- |