United States
Environmental Protection
Agency
Office of
Research and Development
Washington, DC 20460
EPA/625/6-79/005
June 1979
&EPA Handbook
Continuous Air
Pollution Source
Monitoring Systems
-------
-------
EPA/625/6-79/005
June 1979
Handbook
Continuous Air Pollution
Source Monitoring Systems
Center for Environmental Research Information
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Printed on Recycled Paper
-------
ACKNOWLEDGEMENTS
This handbook was prepared for the Environmental Research Information Center, U.S.
Environmental Protection Agency, by Northrup Services, Inc., Research Triangle Park' N C
Norm J. Kulujian was the EPA Project Officer. James A. Jahnke, PhD, and G. J. Aidina
were the principal authors. Technical reviewers included Gerald F. McGowan of Lear
Siegler Inc.. Dale A. Burton of Duke Power Company, James Steiner of Acurex Corporation,
and several continuous monitoring experts within the Agency.
NOTICE
This is not an official policy and standards document. The opinions, findings, and conclusions
are those of the authors and not necessarily those of the Environmental Protection Agency.
Every attempt has been made to represent the present state of the art as well as subject
areas still under evaluation. Any mention of products or organizations does not constitute
endorsement by the United States Environmental Protection Agency.
-------
PREFACE
The emissions standards for stationary sources established by the United States Environmental
Protection Agency (EPA) apply to both power and process industries. These standards
have forced the development of many methods of emission control over the past decade. In
addition, methods to monitor emissions from both controlled and uncontrolled sources have
been developed. Included in these monitoring methods is continuous source monitoring
instrumentation, which has become sophisticated and reliable enough to provide a deter-
mination of the actual level of emissions and a continuous record of the performance of a
control device.
The purpose of this handbook is to provide the environmental engineer in industry or in
government with a background in continuous monitoring instrumentation. The handbook
covers continuous monitoring requirements established by the Federal Government, details
of available instrumentation, and methods of using monitor data. The material presented
is intended for the engineer who may be familiar with process or control equipment operation
but who has had little previous experience with monitoring instrumentation.
The handbook also is intended to serve as a guide for the application of Federal regulations,
for the selection of monitoring instrumentation, and for the utilization of monitoring systems
for improving and optimizing source process operations. Since the field of instrumentation
progresses ra-pidly, efforts must be made to keep abreast of new developments and to supple-
ment the material in this handbook with information from the current literature.
in
-------
ABSTRACT
This handbook provides the detailed information necessary to develop a continuous emissions
monitoring program at a stationary source facility. Federal and State EPA requirements
are given, including design and performance specifications and monitoring and data reporting
requirements. Discussions of extractive sampling techniques and in-situ methods are
presented, along with explanations of the analytical techniques used in currently marketed
instrumentation. Methods for monitoring opacity, pollutant gases, and combustion gases,
such as oxygen and carbon dioxide, are described. A detailed explanation of the EPA
Performance Specification Test is given along with an explanation of the statistical procedures
used to evaluate newly installed systems. Selection procedures for monitoring systems and
specific instrumentation are included as a guide to the industrial engineer. Photographs of
existing instruments and monitoring systems are presented along with explanatory diagrams
to familiarize the reader with the equipment. References are given for each topic discussed
in the handbook. The handbook serves as a basic tool for continuous source monitoring.
enabling the reader to refer to original research and development work for the initiation
of a continuous monitoring program.
IV
-------
CONTENTS
Chapter pa»e
PREFACE ii
ABSTRACT iv
TABLE OF CONTENTS v
LIST OF FIGURES x
LIST OF TABLES xiii
1 INTRODUCTION 1_1
2 REGULATIONS AND MONITORING REQUIREMENTS 2-1
2.1 Introduction 2-1
2.2 New Sources — Part 60 2-1
2.3 Existing Sources - Part 51 2-6
2.4 References 2-9
2.5 Bibliography 2-9
3 INTRODUCTION TO THE ANALYTICAL METHODS 3-1
3.1 Emission Monitoring 3-1
3.2 Monitoring and the Properties of Light 3-2
3.2.1 The Wave Nature of Light 3-3
3.2.2 The Interaction of Light with Matter - Absorption 3-6
3.2.3 The Interaction of Light with Matter - Scattering 3-7
3.2.4 The Interaction of Light with Matter -
The Beer-Lambert Law 3-8
3.3 References 3-10
3.4 Bibliography 3-10
4 CONTINUOUS MONITORS FOR OPACITY MEASUREMENTS 4-1
4.1 Opacity and Trans m ittance 4-1
4.2 The Transmissometer 4-2
4.3 Design Specifications 4-5
4.3.1 Spectral Response 4-5
4.3.2 Angle of Projection 4-7
-------
CONTENTS-Continued
Chapter Page
4.3.3 Angle of View 4-8
4.3.4 Calibration Error 4-9
4.3.5 System Response Time Test 4-9
4.3.6 Sampling Criteria 4-9
4.3.7 System Operation Check 4-9
4.4 Installation Specifications 4-10
4.5 The Performance Specification Test 4-12
4.6 Data Reporting Requirements 4-13
4.7 Opacity Monitor Selection 4-17
4.8 Bibliography 4-17
5 CONTINUOUS MONITORS FOR THE MEASUREMENT OF GASES 5-1
5.1 Introduction 5-1
5.2 Extractive Analyzers 5-2
5.2.1 Extractive Analyzers - Spectroscopic Methods of Analysis 5-3
5.2.1.1 Nondispersive Infrared Analyzers 5-3
5.2.1.2 Nondispersive Ultraviolet Analyzers (NDUV) -
Differential Absorption 5-8
5.2.2 Extractive Analyzers - Luminescence Methods of Analysis 5-11
5.2.2.1 General 5-11
5.2.2.2 Fluorescence Analyzers for SO2 5-13
5.2.2.3 Chemiluminescence Analyzers for NOX and NO2 5-16
5.2.2.4 Flame Photometric Analyzers for Sulfur
Compounds 5-18
5.2.3 Extractive Analyzers - Electroanalytical Methods
of Analysis 5-18
5.2.3.1 General 5-18
5.2.3.2 Polarographic Analyzers 5-20
5.2.3.3 Electrocatalytic Analyzers for Oxygen 5-25
5.2.3.4 Amperometric Analyzers 5-29
5.2.3.5 Conductimetric Analyzers 5-29
5.2.4 Extractive Analyzers — Miscellaneous Methods 5-29
5.2.4.1 Paramagnetic Analyzers for Oxygen 5-29
5.2.4.2 Thermal Conductivity Analyzers 5-31
5.3 Bibliography 5-31
VI
-------
CONTENTS-Continued
Chapter Page
6 EXTRACTIVE SYSTEM DESIGN 6-1
6.1 Introduction 6-1
6.2 Gas Stream Parameters 6-2
6.3 Sampling Site Selection 6-3
6.3.1 SO2/NOX Monitors 6-3
6.3.2 O2/CO2 Monitors 6-3
6.3.3 General Comments 6-3
6.4 Analyzer(s) Selection 6-4
6.5 Design of the Sampling Interface 6-4
6.5.1 General 6-4
6.5.2 Sampling Probe 6-5
6.5.3 Coarse Filters 6-5
6.5.4 Fine Filters 6-7
6.5.5 Gas Transport Tubing 6-8
6.5.6 Sampling Pump 6-9
6.5.7 Moisture Removal 6-12
6.5.8 Sampling Interface/Monitor Calibration 6-15
6.5.9 Dilution Systems 6-16
6.5.10 Controlling the Sampling Interface/Monitor System 6-16
6.6 Bibliography 6-17
7 1N-SITU MONITORING SYSTEMS 7-1
7.1 Introduction 7-1
7.2 Terminology 7-1
7.3 In-Situ Cross-Stack Analyzers 7-3
7.3.1 Differential Absorption Spectroscopy 7-3
7.3.2 Gas-Filter Correlation Spectroscopy 7-8
7.3.3 Advantages and Limitations 7-11
7.4 In-Situ, In-stack Analyzers: Second-Derivative Spectroscopy 7-12
7.5 Bibliography 7-20
8 MEASURING, RECORDING, AND REPORTING REQUIREMENTS 8-1
8.1 Introduction 8-1
8.2 Measuring Requirements 8-2
8.3 Recording Requirements and Systems 8-3
8.3.1 Requirements 8-3
8.3.2 Recording Systems — Continuous Analog Recording 8-4
vn
-------
CONTENTS-Continued
Chapter Page
8.3.3 Recording Systems - Intermittent Digital Recording 8-6
8.3.4 Recording Systems - Data Processors g-7
8.4 Reporting Requirements g_9
8.5 References g_16
8.6 Bibliography g.j5
9 EQUIPMENT SELECTION 9-1
9.1 Introduction 9_1
9.2 Vendors of Recording Instrumentation 9-25
9.3 Bibliography 9_2g
10 APPLICATIONS OF CONTINUOUS MONITORS 10-1
10.1 I ntroduction 1 o_ \
10.2 Advantages of Monitoring Data to the Source 10-1
10.3 Advantages of Monitoring Data for the Regulatory Agency 10-2
10.4 Continuous Monitoring: Aid to Manual Source Sampling 10-3
10.5 Bibliography 10-3
11 THE PERFORMANCE SPECIFICATION TESTS 11-1
11.1 Introduction l_l
11.2 Performance Specification Test I - Transmissometer Systems 1-1
11.2.1 General 1-1
11.2.2 Transmissometer Design Criteria 1-1
11.2.3 Performance Specification Test I 1-3
11.2.4 Zero and Calibration Drift Tests 1-4
11.3 Performance Specification Test 2 - SO2/NOX Systems 1-8
11.3.1 General l_g
11.3.2 Monitor Location and Installation 1-9
11.3.3 Specification Test Procedures I-IO
11.3.4 Calibration Error Test Procedures 1-11
11.3.5 Response Time Test 1-13
11.3.6 Field Relative Accuracy Test 1-14
11.3.7 Instrument Zero Drift and Calibration Drift - 2 Hours
and 24 Hours 1-16
11.3.8 The Operational Test Period 1-19
11.4 Performance Specification Test 3 - O2 or CO2 Monitors 1-20
VIII
-------
CONTENTS-Continued
Chapter Page
11.4.1 Introduction 11-20
11.4.2 Monitor Location and Installation 11-20
11.4.3 O2 or CO2 Monitor Calibration Gases 11-21
11.4.4 Instrument Calibration Check 11-21
11.4.5 Response-Time Test 11-22
11.4.6 Zero and Calibration Drift — 2-hour and 24-hour 11-23
11.5 Bibliography ' 11-24
12 QUALITY ASSURANCE 12-1
12.1 Introduction 12-1
12.2 Calibration Gas Evaluation 12-1
12.3 Instrument Performance Evaluation 12-2
12.4 EPA Inspection Procedures 12-3
12.4.1 Level-One Inspections - (Office Evaluation of Quarterly
Reports) 12-3
12.4.2 Level-Two Inspection — (Field Inspection) 12-3
12.4.3 Level-Three Inspection 12-6
12.5 Bibliography 12-7
APPENDIX A - BIBLIOGRAPHY A-I
APPENDIX B - CALCULATIONS FOR THE PERFORMANCE
SPECIFICATION TEST B-l
APPENDIX C - F-FACTORS C-l
APPENDIX D - PERFORMANCE SPECIFICATIONS - APPENDIX B
TITLE 40 PART 60 - FEDERAL REGISTER D-l
APPENDIX E - CONVERSION FACTORS AND USEFUL INFORMATION E-l
IX
-------
LIST OF FIGURES
Figure No. page
3-1 Types of Source Emission Monitors 3-1
3-2 Wavelength 3.3
3-3 The Electromagnetic Spectrum for Continuous Monitoring 3-5
3-4 Light Absorption Processes 3_6
3-5 Light Scattering Effects 3.7
3-6 Light Absorption 3-8
3-7 Calibration Curve for the Beer-Lambert Relation 3-10
4-1 Single-Pass Transmissometer System 4-2
4-2 Double-Pass Transmissometer System 4-3
4-3 Double-Pass Transmissometer Installed at EPA Source Simulator
Facility, Research Triangle Park, NC 4-4
4-4 Retroreflector Assembly at the Facility 4-4
4-5 Electromagnetic Spectrum and Factors That Affect Opacity
Measurements 4^
4-6 Paniculate Attenuation of Incident Light 4-7
4-7 Angle of Projection 4^
4-8 Angle of View 4_g
4-9 Lear Siegler RM41-P Showing Instrument "Zero" Reflector 4-10
4-10 Transmissometer Siting 4_H
4-11 Lear Siegler RM41-P Portable Transmissometer 4-12
4-12 Relation Between Emission Opacity at Plume Exit and Monitor
Opacity in Duct 4-15
4-13 Two Ducts Entering Common Exit Stack 4-16
5-1 A Lorentzian Absorption Curve 5-3
5-2 Simplified Schematic Diagram of a Nondispersive Infrared Analyzer 5-5
5-3 Operation of the "Microphone" Detector of an NDIR Analyzer 5-5
5-4 Internal View of a Beckman NDIR Analyzer 5-6
5-5 Operation of a "Negative Filter" NDIR Analyzer 5-7
5-6 Internal View of a Bendix NDIR Analyzer 5-8
5-7 The Ultraviolet-Visible Spectrum of SO2 and NO2 5-9
5-8 Operation of a Differential Absorption NDUV Analyzer 5-11
5-9 A DuPont NDUV Analyzer at an Industrial Site 5-12
5-10 Internal View of a DuPont Analyzer Showing Measurement Cell
and Aspirator 5_12
5-11 Fluorescence Spectrum of SO2 5-13
5-12 Operation of the SO2 Fluorescence Analyzer S-15
5-13 Internal View of a TECO Fluorescence Monitor 5-15
-------
LIST OF FIGURES-Continued
Figure No. Page
5-14 The Chemiluminescent Emission Spectrum of NO2 5-16
5-15 Operation of a Chemiluminescence Analyzer 5-17
5-16 Operation of a Flame Photometric Analyzer 5-19
5-17 Instrument Panel of a Meloy Flame Photometric Analyzer 5-20
5-18 Operation of an Electrochemical Transducer 5-21
5-19 Details of the Polarographic Process 5-23
5-20 A Portable Inspection System Using a Polarographic Analyzer 5-24
5-21 An Industrial SO2 "Alarm" Monitor Using a Polarographic Analyzer 5-25
5-22 Example of a Typical "Concentration" Electrochemical Cell 5-26
5-23 Operation of an Electrocatalytic Oxygen Analyzer 5-27
5-24 A Lear Siegler In-Situ Electrocatalytic Oxygen Analyzer Installed on a
Power Plant Stack 5-28
5-25 Operation of a "Magnetic Wind" Paramagnetic Oxygen Analyzer 5-30
6-1 Porous Cylinder Used as External Coarse Filter 6-6
6-2 Actual Porous Cylinder Installed in a Stack Gas Stream 6-6
6-3 Internal Coarse Filter 6-7
6-4 Surface Filter 6-7
6-5 Depth Filter 6-8
6-6 Schematic of Pump Placement — Position A 6-10
6-7 Actual Sampling System with Pump in Position A 6-10
6-8 Schematic of Pump Placement — Position B 6-11
6-9 Actual Sampling System with Position B Pump Location 6-11
6-10 A Refrigerated Chiller Manufactured by Hankinson 6-12
6-11 Interior of Typical Condenser Used for Moisture Removal 6-13
6-12 Schematic Diagram of Permeation Dryer 6-14
6-13 Corrugated Stainless Steel Enclosed Permeation Tube Dryer 6-15
7-1 Types of In-Situ Monitors 7-2
7-2 Operation of In-Situ Differential Absorption Analyzer 7-3
7-3 Mounted EDC Cross-Stack In-Situ Analyzer 7-6
7-4 Internal View of Analytical Section of the EDC Analyzer 7-6
7-5 Internal View of the EDC Light Source Assembly 7-7
7-6 Differential Absorption Spectrometer Installed at Research Triangle
Park Source Simulator Facility 7-7
7-7 Operation of a Cross-Stack Gas-Filter Correlation Spectrometer 7-8
7-8 Absorption Principles of a Gas-Filter Correlation Analyzer 7-9
7-9 The Contravez-Goertz Cross-Stack GFC Monitor 7-10
7-10 The Lear Siegler In-Stack In-Situ SO2-NO Analyzer 7-13
XI
-------
LIST OF FIGURES-Continued
Figure No. Page
7-11 Second-Derivative Spectrometer Installed and Operating at a Steam
Generating Facility 7.13
7-12 Operation of the Second-Derivative In-Stack Monitor 7-14
7-13 Ultraviolet Light Wavelengths Scanned by Spectrometer Moving Mask 7-15
7-14 Scanning a Broad Band Absorption 7-16
7-15 Scanning an Absorption Peak 7_17
7-16 First and Second Derivatives of Linear Absorption 7-18
7-17 First Derivative of an Absorption Curve 7_jg
7-18 Second Derivative of an Absorption Curve 7-19
8-1 Possible Methods of Measuring-Record ing-Reporting 8-2
8-2 Data from Typical Data Processor Designed for Continuous Source
Monitoring Applications g_g
8-3 Suggested Format for Quarterly Excess Emissions Report 8-12
XII
-------
LIST OF TABLES
Table No. Page
2-1 Industries Required to Monitor Emissions 2-1
2-2 Summary of NSPS Continuous Emission Monitoring Requirements 2-2
2-3 Contents of Part 60 — Appendix B Outline 2-5
2-4 Continuous Monitoring Requirements - Differences Between New and
Existing Sources 2-8
3-1 Principles Used in Emission Monitors 3-3
4-1 Opacity Monitor Performance Specifications 4-13
5-1 Infrared Band Centers of Some Common Gases 5-4
8-1 Measuring Requirements 8-2
8-2 Recording Requirements 8-3
9-1 Opacity Monitors - Selection Procedures 9-2
9-2 Gaseous Monitors — Selection Procedures 9-7
9-3 Vendors of Double-Pass Transmissometers 9-12
9-4 Vendors of Single-Pass Transmissometers 9-13
9-5 Principal Continuous Source Monitor Manufacturer Summary
(July 1978) ' 9-14
9-6 Oxygen Analyzer Summary 9-18
9-7 In-Situ Monitor Summary 9-19
9-8 List of Instrument Manufacturers 9-19
9-9 Manufacturers of Strip Chart Recorders 9-26
9-10 Manufacturers of Data Logging Equipment 9-26
9-11 Manufacturers of Continuous Monitor Data Processors 9-27
1-1 Neutral Density Filters for Transmissometer Calibration Error
1-2 Opacity Monitors Performance Specifications
1-3 Values for to.975
1-4 24-Hour Transmissometer Zero Drift Data
1-5 Span and Calibration Gas Values
1-6 Performance Specifications for SO2/NOX Systems
1-7 Performance Specifications for 02 or CO2 Monitors
1-2
1-3
1-6
1-7
i-n
1-20
12-1 Level-Two Inspection Check List 12-4
xtu
-------
-------
CHAPTER 1
INTRODUCTION
The application of continuous monitoring techniques to the measurement of pollutants
emitted from stationary sources has become an area of growing interest in industry and
government. The.promulgation of regulations dealing with source level continuous monitors
on October 6, 1975, has developed a growing need for information on monitoring systems
and their ability to meet the performance specifications defined in the regulations. It is also
becoming apparent to personnel in industries affected by the regulations that continuous
monitors can save money and improve plant performance in addition to providing continuous
source emissions data.
This handbook is intended to provide a background in the field of continuous monitoring to
individuals actively engaged in industrial air pollution control programs. Topics in this
handbook cover studies ranging from the Code of Federal Regulations to details of instrument
operation. A survey is made of presently available instrumentation, and guidelines are given
for the selection of monitors. The advantages and limitations of several types of monitoring
system designs are reviewed so that the environmental engineer can make informed decisions
for a given application.
The performance requirements defined by the Environmental Protection Agency (EPA) and
the Code of Federal Regulations for installed monitoring systems are discussed in detail.
Siting requirements, drift and calibration limitations, the definition of relative accuracy.
and the statistical methods established by EPA for instrument evaluation are all elements
of the Performance Specification Test.
It should be kept in mind throughout the reading of this handbook that the intent of the
promulgated continuous monitoring regulations was to ensure that a source operator would
utilize some type of instrumentation system that could monitor the performance of an air
pollution control device. The cost of modern air pollution control equipment is considerable,
but all too often, an instrument that could monitor the effectiveness of such equipment is
considered unnecessary. However, in many cases modern analytical instruments have been
found to increase process efficiency and decrease control equipment operating costs.
Continuous source monitors were not originally intended to be a tool for the enforcement
of compliance to the new source emission standards (except in the use of primary copper,
zinc, and lead smelters - see Ref. 3, Chapter 2). To prove or disprove source compliance,
the manual EPA reference methods must still be performed. Several States, however, are
developing enforcement programs utilizing continuous monitoring data. Further develop-
ments in this regard are expected on the Federal level, as well as from the States.
-------
The technology of source monitoring has advanced rapidly in the past few years along with
a steady improvement in instrument reliability. The major concern now is proper application
and maintenance. It is the purpose of this handbook to provide a background for selecting
and designing an adequate monitoring system for a source application. It is hoped that the
guidelines given here will enable those involved in continuous source monitoring to gather
reliable, valid emissions data.
1-2
-------
CHAPTER 2
REGULATIONS AND MONITORING REQUIREMENTS
2.1 Introduction
Specific source categories are required by law to install and maintain continuous emissions
monitoring systems (1)(2)(3). The United States Government publishes the Code of Federal
Regulations (CFR) once a year, which is supplemented daily by the Federal Register. It is
in these two publications that the regulations concerned with stationary source emissions
and emission monitoring may be found. The Federal regulations establish standards and
monitoring requirements for new sources. Individual States, however, are required by the
Clean Air Act of 1970 to draft regulations for existing sources. It is important that the
environmental engineer keep abreast of the CFR and the Federal Register to determine how
a facility is to comply with the regulations.
2.2 New Sources - Part 60
Regulations concerning new stationary sources are found under Part 60 of Title 40 of the
CFR. Title 40 is composed of five volumes dealing with the protection of the environment.
Part 60 deals with the standards of performance for new stationary sources or New Source
Performance Standards (NSPS). Table 2-1 gives those source categories required by Part 60
for continuous monitoring of either opacity or some type of gaseous pollutant.
TABLE 2-1
INDUSTRIES REQUIRED TO MONITOR EMISSIONS
Fossil-Fuel-Fired Steam Generators Sulfuric Acid Plants
Nitric Acid Plants Petroleum Refineries
Primary Copper Smelters Iron and Steel Plants
Primary Zinc Smelters Ferroalloy Production Facilities
Primary Lead Smelters Kraft Pulp Mills
The Federal Register published on October 6. 1975, is the document in which the EPA
performance specifications for continuous monitoring systems were promulgated. Several
minor points have been revised by subsequent publication in the Federal Register, but
pages 46250 to 46271 of the October 6, 1975 document, cover the agency's position on
continuous monitors for new sources. Limits of drift and acceptable error limits for monitors
are given; but more importantly, the October 6, 1975, Federal Register establishes the
position that the EPA does not approve specific brands of instrumentation or specific
2-1
-------
analytical methods (in the case of gas monitoring) for source monitoring systems. This is
in contrast to the policy of approving specific instrument models for continuous ambient
air analysis. What has been established, however, is the Performance Specification Test.
EPA provides latitude in continuous monitoring system design and application to allow
sources to handle individual problems. The installed system must prove that it meets
specified, minimum requirements for instrument location, drift, accuracy, etc. These points
will be covered in subsequent chapters of this handbook.
Table 2-1 gives the source categories required to install some type of monitoring system.
The primary question deals with what is to be monitored. A partial listing is given in the
October 6, 1975, document. A complete summary of the requirements for a given category
is given in the CFR. As stated previously, the NSPS come under Part 60 of Title 40 of the
CFR. Each source category that is currently regulated under these standards is assigned a
subpart letter. For example, sulfuric acid plants come under Subpart H. A summary of the
emission parameters required to be monitored, emission limits, and applicability dates for
each new source category is given in Table 2-2.
TABLE 2-2
SUMMARY OF NSPS CONTINUOUS
EMISSION MONITORING REQUIREMENTS
Source
Category
Fossii-fuel-fired
steam generators
(Subpart D)
>73 megawatts
heat input rate
August 17, 197!
Affected
Facility
Coal-fired
boilers
Oil-fired
boilers
Monitoring
Required
Opacity
SO2
NOX
(only if emissions
>70% of
standard)
O2 or CO2
for conversion
factors
Opacity
SO2
NOX
O2 or CO? for
conversion factors
Emissions
Standards
20%
520 nanograms/ joule
(1.2 lb/106 Btu)
300 nanograms/joule
(0.70 lb/10* Btu)
20%
340 nanograms/joule
(0.20 lb/IO<> Btu)
130 nanograms/joule
(0.30 lb/106 Btu)
Averaging
Time
6 minutes
3 hours
3 hours
6 minutes
3 hours
3 hours
2-2
-------
TABLE 2-2
SUMMARY OF NSPS CONTINUOUS
EMISSION MONITORING REQUIREMENTS-Continued
Source
Category
Nitric acid plant
(Subpart G)
August 17, 1971
Sulfuric acid
plant
(Subpart H)
August 17, 1971
Petroleum
refineries
(Subpart J)
June 11, 1973
Affected
Facility
Gas-fired
boilers
Process
equipment
Process
equipment
FCC catalyst
regenerator
Fuel gas
combustion
device
Claus recovery
units oxidation
on reduction
control systems
followed by an
incinerator
Reduction
control system
without an
incinerator
Monitoring
Required
NOX
(only if emissions
>70% of
standard)
O2 or CO2 for
conversion factors
NOX
S02
Opacity
Carbon
monoxide
H2S
S02
Reduced sulfur
compounds
H2S calculated
as SO2
Emissions
Standards
86 nanograms/joule
(0.20 Ib/lO* Btu)
1.5 kg/ metric ton of
acid produced
(3.0 Ib/ton)
2.0 kg/ metric ton of
acid produced
(4 Ib/ton)
30% except one
6 min. period/hr.
0.050% by volume
230 mg/dscm
0.1 gr/dscf)
0.025% by volume
at 0% O3 on a dry
basis
0.030% by volume
at 0% O2 on a dry
basis
0.0010% by volume
at 0% 02 on a dry
basis
Averaging
Time
3 hours
3 hours
3 hours
6 minutes
1 hour
3 hours
12 hours
12 hours
12 hours
2-3
-------
TABLE 2-2
SUMMARY OF NSPS CONTINUOUS
EMISSION MONITORING REQUIREMENTS-Continued
Source
Category
Primary copper
smelters
(Subpart P)
October 16, 1974
Primary zinc
smelters
(Subpart Q)
October 16, 1974
Ferroalloy
production
facility
(Subpart Z)
October 21, 1974
Iron and steel
plants
(Subpart A A)
October 24, 1974
Kraft pulp mills
(Subpart BB)
February 23. 1978
Affected
Facility
Dryer
Roaster,
smelting furnace
on copper
converter
Sintering
machine
Roaster
Electric
submerged arc
furnace
Electric arc
furnace control
device
Recovery
furnace
Any and all
process equip-
ment (exceptions
noted in
43 FR 7568
2/23/78)
Monitoring
Required
Opacity
SO2
Opacity
S02
Opacity
Opacity
Opacity
Total reduced
sulfur (TRS)
Oxygen
Emission
Standards
20%
0.065% by volume
20%
0.065% by volume
15%
3%
35%
5 ppm corrected to
10% oxygen
Averaging
Time
6 minutes
6 hours
6 minutes
2 hours
6 minutes
6 minutes
6 minutes
12 hours
12 hours
Once it has been determined that a continuous monitoring system is required for measuring
opacity or the concentration of a specific gas, the type of system or instrument that will
best satisfy the EPA performance specifications and the needs of the plant must be selected.
The performance specifications give the general characteristics expected of an instrument
and clearly define procedures for checking the installed instrument performance. These
methods are given in Appendix B of Part 60 of the CFR and are also included in the
appendix of this handbook. An outline of the contents of Appendix B is given in Table 2-3.
2-4
-------
From Table 2-3, it can be seen that, at present, there are performance specifications for
opacity monitors (transmissometers), SO2 and NOX monitors, and CO2 and O2 monitors.
Each specification discusses the installation requirements, the levels of performance expected
of the instrument system over a 1-week operational test period, and the statistical methods
of analyzing the data obtained over the test period. The specifications for opacity monitors
include a number of design characteristics such an instrument must possess. Design
specifications are not given for the gaseous analyzers.
TABLE 2-3
CONTENTS OF PART 60 - APPENDIX B OUTLINE
PART 60 - APPENDIX B
Performance Specifications
(Added)
Page 46257
Performance Specification I
Transmissometer Systems
Performance Specification 2
Monitors of SO2 and NOx
Performance Specification 3
Monitors of CO2 and O2
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0 -
10.01-
(Page 46259)
Principle and Applicability
Apparatus
Definitions
installation Specifications
Optical Design Specifications
Determination of Conformance
with Design Specifications
Continuous Monitoring System
Performance Specifications
Performance Specification
Calculations, Data Analysis,
and Reporting
References
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.01-
(Page 46263)
Principle and Applicability
Apparatus
Definitions
Installation Specifications
Continuous Monitoring System
Performance Specifications
Performance Specification
Test Procedures
Calculations, Data Analysis,
and Reporting
References
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0 "-
(Page 46268)
Principle and Applicability
Apparatus
Definitions
Installation Specifications
Continuous Monitoring System
Performance Specifications
Performance Specification
Test Procedures
Calculations, Data Analysis,
and Reporting
References
2-5
-------
In summary, Part 60 of the CFR incorporates the requirements for the continuous monitoring
of designated new stationary sources. The manner in which continuous monitoring systems
are expected to perform after being installed on a source is given in Appendix B of Part 60.
Existing sources are regulated by the States, and the continuous monitoring requirements
for existing sources must be established by each State.
2.3 Existing Sources — Part 51
Since existing sources far outnumber new sources, continuous monitoring requirements for
existing sources have a greater impact on an entire industry. There is a significant difference,
however, between regulations for new and existing sources. As part of the Clean Air Act,
the States must regulate existing sources. The Federal Government regulates new sources.
The States, however, may not arbitrarily set standards and regulations. They must follow
certain minimum requirements established by EPA.
The Federal requirements that each State must follow when drafting regulations for con-
tinuous source emission monitors are found in Part 51 of Title 40 of the CFR (4). Regulations
and specifications are given for new sources in Part 60, but only requirements for the
preparation, adoption, and submittal of State regulations are given in Part 51. Once the
State regulations are approved by EPA, they become part of the State Implementation
Plan (SIP). The SIP is a continually evolving document that establishes the procedures
through which a State plans to meet the ambient air quality standards set by EPA and
other goals established by the Clean Air Act.
The Federal requirements for individual State continuous emission monitoring regulations
are given in Appendix P of Part 51. The requirements were published in the October 6,
1975, Federal Register at which time, 1 year was given for the States to submit continuous
emissions monitoring regulations as part of their SIP's. No State met that deadline;
however, most States have either submitted some regulations or are actively developing
them. The date when these regulations are approved is important to the affected industries.
Each existing source that is required to install continuous monitoring systems must do so
within 18 months from the date of approval. Regulated industries must be aware of and
meet this deadline. There are presently four source categories for which States must
draft continuous monitoring regulations: fossil-fuel-fired steam generators, sulfuric and
nitric acid plants, and petroleum refineries.
Several exemptions are allowed for existing sources which do not arise for sources covered
by the New Source Performance Standards. The exemptions were allowed so that undue
hardship would not be placed on existing facilities or on those that will be retired within
5 years (5 years after inclusion of the source category in Part 51 Appendix P). Also, States
are only required by EPA to monitor sources that have an emission standard for SO2, NOX,
or opacity in its SIP for that source category. The aim in Part 51 is to have States develop
regulations that will be fair to existing sources. Sulfuric acid and nitric acid plants would
be required to install monitors if they have production capacities greater than 300 tons per
2-6
-------
day. Catalyst regenerators at petroleum refineries need to monitor opacity only if they have
a feed capacity of greater than 20,000 barrels per day.
In contrast to new fossil-fuel-fired steam generators that burn oil or coal, existing fossil-fuel-
fired plants will be required by the States to monitor SO2 emissions only if a flue gas
desulfurization (FGD) system is used. EPA, however, is currently considering extending
SO2 monitoring to existing sources that do not have FGD systems. Also, for existing plants,
NOX emissions are to be continuously monitored if the plant is within an Air Quality Control
Region (AQCR) that has a control strategy for nitrogen oxides, if the source has a heat input
rate of greater than 1000 x I06 Btu/hr, and if the source emits nitrogen oxides at levels
greater than 70 percent of the State NOx standard. A summary of the differences for
monitoring requirements between new and existing sources is given in Table 2-4.
Once it has been established that an existing source must install a continuous monitoring
system, the instrument specifications, the data reporting requirements, and the Performance
Specification Test Requirements are the same as those for new sources. In fact, in Part 51,
which gives the minimum requirements for the State regulations, it is stated that each State
plan must incorporate, as a minimum, the contents of Appendix B Part 60 (which gives
the Performance Specifications for monitoring systems on new sources, as discussed earlier).
Existing sources may have continuous monitors already in use which may not meet the
EPA Performance Specifications of Appendix B. This case is covered by a grandfather
clause that requires monitors installed before September 11, 1974, to demonstrate an accuracy
of only ±20 percent with respect to the reference method. These older monitors are to
undergo a complete Performance Specification Test 5 years after approval of the SIP con-
tinuous monitoring regulations.
The States are allowed some degree of latitude on a case-by-case basis in making exceptions
or in permitting alternative monitoring requirements for an existing source. Examples of
special cases would be the presence of condensed water in the flue gas stream, infrequent
operation of a facility, or difficulties in installing a continuous monitoring system because
of physical limitations at the facility.
In summary, the intentions of requirements of Part 51 are that the continuous monitoring
regulations of the States satisfy the following points:
• Allow the utilization of existing instrumentation where possible.
• Reduce installation costs where possible.
• Reduce maintenance costs where possible.
• Reduce the number of monitors required where possible.
• Encourage new technology.
2-7
-------
TABLE 2-4
CONTINUOUS MONITORING REQUIREMENTS -
DIFFERENCES BETWEEN NEW AND EXISTING SOURCES
Source Category
Fossil-fuel-fired steam
generators
Suliuric acid plants
Nitric acid plants
Petroleum refineries -
catalyst regenerators
for fluid bed catalytic
cracking units
Electric arc furnaces
Primary copper, zinc,
and lead smelters
Ferroalloy produc-
tion facilities
Kraft pulp mills
Portland cement kilns
and cylinder coolers
Pollutant
Opacity
SO2
NOX
O2 or CQ2
SO2
N02
Opacity
Opacity
Opacity
S02
Opacity
Opacity
TRS
02
Opacity
New Sources
>250 million Btu/hr
>250 million
Btu/hr and
if emissions >70%
of standard
(If SO2 or NOX
monitor required)
All sources covered
by NSPS
All sources covered
by NSPS
All sources covered
by NSPS
All sources covered
by NSPS
All sources covered
by NSPS
All sources covered
by NSPS
All sources covered
by NSPS
Possible future
requirements
Existing Sources
>250 million Btu/hr
SO2 only if flue gas
desulfurization used and
>30% capacity factor
NOX only where control
strategy required and if
>70% of standard and
if >IOOO X 106 Btu/hr
heat input
O2/CO2 only if State
requires data for con-
verting to emissions
standard
>300 ton/ day production
>300 ton/ day production
and only where control
strategy required
>20,000 bbl day
Possible future
requirements
Possible future
requirements
Possible future
requirements
Possible future
requirements
Possible future
requirements
2-8
-------
2.4 References
1. U.S. EPA, "Standards of Performance for New Stationary Sources," Code of Federal
Regulations, 40 CFR, Part 60.
2. U.S. EPA, "Requirements for Submittal of Implementation Plans Standards for New
Stationary Sources - Emission Monitoring," Federal Register 40, FR 46240-46271,
October 6, 1975, and Revisions:
41 FR 44838, 10/12/75, "Approval of Alternate Monitoring Requirements" (Definition
of the Wet F Factor).
42 FR 5936, 1/3/77 " Revision to Emission Monitoring Requirements and to Reference
Methods" (Use of CO2 Monitors, After Wet Scrubbers, Clarification of Data Recording
Requirements for Opacity Monitors, Other Clarifications).
42 FR 26502, 5/23/77 "Compliance with Standards and Maintenance Requirements"
(Use of Continuous Monitoring Data as Evidence).
3. U.S. EPA, "Continuous Monitors and Primary Smelters - Use in the Compliance
Test," 41 FR 2338 see 60.166a2.
4. U.S. EPA, "Requirements for Preparation, Adoption, and Submittal of Implementation
Plans," Code of Federal Regulations, 40 CFR, Part 51.
2.5 Bibliography
Chaput, L. S,, "Federal Standards of Performance for New Stationary Sources, for New
Stationary Sources of Air Pollution — A Summary of Regulations," Journal of the
Air Pollution Control Association, V. 26, No. 11:1055-1060, 1976.
Donovan, P. C, "Emissions Monitoring from Stationary Sources," Proceedings, Continuous
Monitoring of Stationary Air Pollution Sources, APCA Specialty Conference, APCA,
1975, pp. 13-23.
Floyd, J. R., "The Implementation of the NSPS Continuous Emission Monitoring Regula-
tions in EPA, Region VIII," Paper 78-35.1 presented at the 71st Meeting of the Air
Pollution Control Association, Houston, Texas, June 26-30, 1978.
Jaye, F., Steiner, J., and Larkin, R., "Resource Manual for Implementing the NSPS
Continuous Monitoring Regulations Manual 1 - Source Selection and Location of Con-
tinuous Monitoring Systems," EPA-340/l-78-005a, April 1978.
2-9
-------
Kendall, D. R., "Estimation of Compliance of Gaseous Pollutant Emissions from Routine
Continuous Monitoring Data," TAPPI. V. 59, No. J;123-126, January 1976.
Lillis, E. J., and Schueneman, J. J., "Continuous Emission Monitoring: Objectives and
Requirements," Journal of the Air Pollution Control Association, V. 25, No. 8, August 1975.
Smith, G. W., "Federal Emission Monitoring Regulations," Proceedings of the'Workshop
on Sampling, Analysis, and Monitoring of Stack Emissions, NTIS PB-252-748 April 1976
pp. 1-16.
Smith G. W., "New Federal Requirements for Continuous Source Monitoring for the
Electric Power Industry," Paper presented at the U. of Texas Conference on Air Quality
Management in the Electric Power Industry, January 29, 1976.
Wolback, D. D., and James, R. E., "Texas Experience with Company Owned Monitors
and EPA Monitoring Requirements," Air Pollution Measurement Accuracy as it Relates
to Regulation in Compliance, APCA Specialty Conference, APCA, 1976, pp. 292-302.
U.S. EPA, Compliance Status of: Major Air Pollution Facilities — Stationary Source
Enforcement Series EPA-340/1-76-010, December 1976.
U.S. EPA, Standards of Performance for Ne\v Stationary Sources — A Compilation as
of August 1976, EPA-340/1-76-009, August 1976.
U.S. EPA, Conference Report and Responses to Key Questions anil Issues, Continuous
Emissions Monitoring Conference, Dallas, Texas, February 15-17, 1978, EPA-340 1-77-025,
Stationary Source Enforcement Series, December J977.
2-10
-------
CHAPTER 3
INTRODUCTION TO THE ANALYTICAL METHODS
3.1 Emission Monitoring
Federal or State regulations will dictate whether an opacity monitor, gas monitors, or both
are required on a given source. Many sources will be required to monitor opacity only.
In such cases, instrument selection is relatively easy, since there is only one measurement
principle that will satisfy the EPA opacity monitor design specifications. On the other
hand, selection of gas monitors is more difficult, since EPA has established no design
specifications in this case. A gaseous emission monitor can be approved if it performs
according to EPA specifications once it is installed on the source. Any chemical or physical
monitoring method can be used so long as it accurately monitors emissions (accurate,
relative to the reference methods for determining pollutant gas concentration, being defined
in 40 CFR Part 60 Appendix B).
There are many instruments marketed for monitoring emissions from stationary sources.
Opacity monitors may be either single-pass or double-pass systems (these will be discussed
in Chapter 7). Gas monitoring systems may be either extractive systems, in-situ systems,
or remote monitoring systems. These divisions are shown in Figure 3-1.
SOURCE EMISSION MONITORS
OPACITY MONITORS
GASEOUS EMISSION MONITORS
SINGLE-PASS
SYSTEMS
DOUBLE-PASS
SYSTEMS
EXTRACTIVE
SYSTEMS
IN-SITU
SYSTEMS
REMOTE
SYSTEMS
-\ CROSS-STACK
FIGURE 3-1
TYPES OF SOURCE EMISSION MONITORS
Extractive gas monitors were the first type of instruments to be incorporated into continuous
gas monitoring systems. Many of the first extractive systems used modified ambient air
analyzers, or they adapted an ambient air analyzer to source applications with the use of a
3-1
-------
gas dilution system. Many problems were found with this type of approach. Systems
were later designed to deal directly with the problems of extracting, sampling, and analyzing
pollutant gases at source level concentrations.
The in-situ gaseous emission analyzers are the second generation of instruments designed
for source monitoring. The analysis is performed on the gas as it exists in the stack or
duct (hence, in-situ) generally by some advanced spectroscopic technique. These analyzers
are installed either across a stack (cross-stack) or employ a probe inserted into the flue
gas stream (in-stack). These two types of in-situ analyzers do not extract or modify
the flue gas.
The remote monitoring instrument is the third and latest generation of the source monitoring
techniques. These instruments use laser and other spectroscopic methods to monitor
emissions at distances from 500 to 1000 meters away from the source. At the present
time, remote systems are used by government agencies and their contractors for research
into specific emission problems. Performance specifications have not been written for remote
monitoring systems, but such systems soon may find some utility in enforcement cases.
Reference 2 of this chapter provides an excellent overview of remote sensing techniques.
The analytical techniques used in continuous source monitors encompass a wide range of
chemical and physical methods. These vary in range from chemical methods as basic as
coulometric titration to the measurement of light produced in a chemiluminescent reaction.
Principles of physics as basic as light scattering are utilized as are the more complicated
methods of detecting light absorption by second-derivative spectroscopy. A summary of
the principles of chemical physics that are used in currently marketed emission monitoring
systems is given in Table 3-1.
Before each of these methods is discussed in detail, it is necessary to review some basic
principles of chemical physics. In the next section, the characteristics of the interaction
of light with particulates and gases will be discussed.
3.2 Monitoring and the Properties of Light
The majority of instruments developed for continuous emission monitoring utilize some
phenomenon arising from the interaction of light with matter. Opacity monitors measure
the effects of light scattering and absorption; a nondispersive infrared analyzer measures
the amount of light absorbed by a pollutant molecule; and a chemiluminescence analyzer
senses the light emitted in a chemical reaction. A better understanding of the details of
instrument operations can be gained by reviewing some of the properties of light and by
examining the nature of light scattering and absorption.
3-2
-------
TABLE 3-1
PRINCIPLES USED IN EMISSION MONITORS
Opacity Monitors
Gaseous Emission Monitors
Extractive
Systems
In-Situ
Systems
Visible light
scattering and
absorption
Absorption Spectroscopy
Nondispersive infrared
Differential absorption
Luminescence Methods
Chemiluminescence (NOX)
Fluorescence (SO2)
Flame photometry
Electroanalytical Methods
Polarography
Electrocatalysis (Oa)
Amperometric Analysis
Conductivity
Paramagnetism (O2)
Cross-Stack
Differential absorption
Gas-filter correlation
In-Stack
Second-derivative
Spectroscopy
Electrocatalysis (O2)
Note: (Methods followed by the gas in parenthesis indicate that the technique is currently
commercially applied only to that gas)
3.2.1 The Wave Nature of Light
Light has a wave nature; it is composed of oscillating electric and magnetic fields. Light
waves, or electromagnetic waves as they are better termed, are characterized by their
wavelength or frequency. Figure 3-2 shows a typical oscillating electric field as a function
of distance at a frozen instant in time.
UJ
/\/\/\
LENGTH
FIGURE 3-2
WAVELENGTH
3-3
-------
The length between successive oscillations of a wave is called the wavelength (X). The
period of time that it takes a wave to go through an oscillation cycle is called the frequency (v).
Since light waves travel at a speed, c = 3 X 1010 cm/sec, the following relationship exists
between wavelength and frequency:
speed of light
Frequency = wavelength
c cm/sec
v =— = I/sec (cycle/second)
X cm i \ j i >
Literature describing continuous monitoring instruments often uses wavelength to characterize
the spectral region of light that is used in the analytical method. Another term often
used for the same purpose is the wavenumber. The wavenumber is expressed as:
_ c/X 1
u-~c—7
(I/cm or cm" ; the number of wavelengths per centimeter)
The units of v are those of cm , called reciprocal centimeters or wavenumbers. The
wavenumber, u, is essentially a measure of frequency, differing from v by the constant
factor of the velocity of light.
The light used in continuous monitoring instrumentation ranges from ultraviolet light, with
a wavelength of 200 nanometers, to infrared light, with a wavelength of 6000 nanometers.
Figure 3-3 shows the regions of the electromagnetic spectrum used in continuous moni-
toring methods.
It should be noted that different spectraf regions often use different units for the expression
of wavelength. For example, angstroms have historically been used in the ultraviolet
to identify wavelengths, whereas in the infrared, micrometers and wavenumbers are commonly
used by spectroscopists. This difference in units arose from the independent development
of each field of spectroscopy. Different unit scales have been placed on Figure 3-3 for
easy reference. It should be noted that in the ultraviolet and visible region, angstroms
and nanometers are most commonly used (1 nm = 10 angstrom = I0~9 meter: see Appendix E).
In the infrared, micrometers (Mm) tend to be used interchangeably with reciprocal centimeters
(wavenumbers) when characterizing light. To change between the two designations, obtain
the reciprocal of the wavelength expressed in jum and multiply by 10 to obtain wavenumbers
in units of cm"1.
For example, if
X = 5 jum I/A =—— x 104-^-= 2000 cm"1
5 m cm
3-4
-------
T
z
cc
0-
CC
0 CM J"
fNQ 0
^ 0 - z
OdDON ujrtE'9-
HICIM m Lurfo-b
oivjni uj « y y
Od9 rOS uJrfQ'f-
DdO IDH uJTtfrg-g-
"
T r
f
L
J
-|U LL Z
T Z ^
w
1
]o Oo
Jo CQ u
cc a
1 <
O
0 0
I CC
D
-1 ~ >-
CMrC*^ 1
•s^X (j J i
,*fc
wC"
~. UJ ^
>• u
cc z
03 LU O
5 « i1^
O ^ '
0) 2
W 5
ii i
i g i
DC 3; W
IP | w
•— tO+ON m
LUU 009 1 uj
1 AilOVdO I LU
UJUOOS^- p
(AHON) ON uju 9£fr 3
i QOH13W UJU 03*- £
Odd Zoo UJU frfiP r
*J<" Us "°6 £
^
/
,
4lP
^__— , (^ —
• —
• —
,
/
/-
---
r^\
|c
__>
""- ''O
TO
—r"
F'
"~* **~^
^>
— -—
I>
^N
^A
V"
5
J
z
o
H
w
z
^f
CC
i —
2 ^
tr <
< z
E O
~ t
*£ cc
LU i
Z J,
Is.
w^ r
° I e
Z < 0
/= -J LO
1- LU CN
Q3H
39NVUQ
MO113A
N33H9
131OIA
> LO O LO O
> 1** IT) CM
/w*,e*
(S ^^
(AOQN) 2OS UJU 083- uj O "( O*
2P ON uju g-932 g £ \VZ
»P ^OS UJU G'Q 17 * ^ £ X^
^r»o UJU o-crT^ ^ W j>*<^Q
NOI1V1IOX3 <
Cx1 £...
M . j^. *^
-OO3C °^
-OOOE I
-0083 £
-0093 1
—00*3 uj
-0033 uj
-0003 <
-0081 >
'o o o v 'o
UJ3}
I'
lN3IOIdd3OO
NOiidHOSav
^>
z
0 „
1- T
< £
> > r^-
— £ o
— <£.
OOZ.
T u *"
LU Z Q
m 2 P
— u ^
T 3|
00*
fc y z
"j ZO
O OP
1 gs
CC uj <
hr _j DC
=: uj i—
-j iu r-
i s«
§ §§g
5 ^ o 4
D 5 O 0
3 N W O
u 5 — w
? — UJ S
o £_
CO O
to
If,
^
-
c
a.
CM
_
E
c
§£
3.
^
cx> _
§
^ ?
^•fc DC
<^ u
LU
Q
(U
tr =
^ m CC
t s °° /*l
v ^ O
o ^ ^_
^3 z ^
§ o
^i ^^
/* 5 •=
^ <<*> \ w
Q» i I =>
/- %> i i 0
' . c'v (/) — ^^
^•^ < S ^
"— uj co 2
%, K < uT
*tX U LU K
*y z o: 2
" o
~~ C_J
f
''^ m c
CL ' O
4^ CO U_
*?/ X LJJ ^
^^ § § E
\ S E §
°
S fo UJ
m _l
LJJ
LU
** X
CM
orf c
Q _ *
3-5
-------
3.2.2 The Interaction of Light with Matter - Absorption
Light carries energy. This is obvious to anyone who has sat on a beach and absorbed
the rays of the sun. Early in this century. Max Planck deduced that parcels of light
(photons) carry energy according to the relation:
x -h -hc
Ae — hu — —
A
where Ae is the amount of energy, and h is Planck's constant. This expresses the fact
that light of shorter wavelengths (e.g., light in the ultraviolet) will carry with it more energy
than light of longer wavelengths (e.g., light in the infrared). In monitoring instruments,
therefore, light of different wavelengths will have different effects on a pollutant molecule.
This is an important point to realize, especially when the effects of interfering molecules
need to be considered.
Molecules are made up of atoms, and an atom is made up of a nucleus and electrons. If
light energy strikes a molecule, the atoms and electrons can do certain things, if the energy
is of the right value. For example, refer to Figure 3-4 for the SO2 molecule.
ROTATION VIBRATION ELECTRONIC
(a) (b) (c)
FIGURE 3-4
LIGHT ABSORPTION PROCESSES
Light of low energy (long wavelength) will cause a molecule to rotate as shown in Fig-
ure 3-4a. Light of somewhat higher energy may cause the atoms to move back and forth
in one of the normal modes of vibration of the molecule as shown in Figure 3-4b. Light
having the correct wavelength in the ultraviolet region of the electromagnetic spectrum may
have enough energy to excite an electron in the molecule and make it jump into a new orbit.
This is an electronic transition. The science of chemical physics has shown that the energy
must be exactly the right value to cause a rotational, vibrational, or electronic transition.
When one of these excitations does occur, the light is said to be absorbed. In absorption,
energy has been lost from the light beam and has been transferred to the molecule.
3-6
-------
In Figure 3-3, the electromagnetic spectrum for continuous monitors shows the regions of
the spectrum in which molecules absorb light energy. In the infrared and near-infrared
region, rotational transitions occur at the longer wavelengths, and vibrational transitions
occur at the shorter wavelengths (higher energy). Generally, the vibrations of the molecule
will be coupled with the rotations to produce distinct absorption spectra. There are CO.
CO2, NO, and SO2 monitors that operate in the region of 3 to 6 ^m, based upon the
excitation of vibrational-rotational energy states by the absorption of infrared light.
In the visible and ultraviolet region of the electromagnetic spectrum, electronic transitions
occur where the electrons in the pollutant molecules become excited and jump into a new
energy state because of the impinging light. The technology associated with the measurement
of ultraviolet light is quite advanced, and in the region of 200 nm there are few inter-
fering species. As a result, a number of monitors have been developed to measure SO2 in
this region of the ultraviolet range.
3.2.3 The Interaction of Light with Matter - Scattering
There is another way to remove energy from a beam of light other than by absorption.
Light can be scattered in different directions if it impinges upon aerosols or particulates.
The mechanism of light scattering is somewhat complex, but the details become important
when monitoring the opacity of a flue gas.
Depending upon the size of a particle, light scattering can be described in macroscopic or
microscopic terms. For large particles, where the wavelength of light is smaller than the
size of the particle (0.5 ,um for the wavelength and 1.0 nm or greater for the particle),
the macroscopic phenomena of reflection, refraction, and diffraction describe the scattering
process. Figure 3-5 shows these effects.
DIFFRACTION
INTERNAL
REFLECTION
FIGURE 3-5
LIGHT SCATTERING EFFECTS
3-7
-------
• Reflection is a change in the direction of light after striking the surface of a
particle.
• Refraction occurs after light enters the particle; its speed and direction change
because of the optical characteristics (refractive index) of the material. Once light
has entered the particle, it can also undergo internal reflection.
• Diffraction is a bending of light around an object caused by the interference of
light waves near the surface of the object.
For visible light and particles having a size near 1 pm or larger, the light will be primarily
scattered in the forward direction. The transmission of light through a plume containing
many particles will be reduced, because the light will scatter before emerging from the plume.
Particles smaller than about 0.1 pm will scatter visible light by a process called dipole or
Rayleigh scattering. In this case, the light interacts with the electrons, oscillating them in
the electromagnetic field. An accelerating electron will emit electromagnetic radiation (in
virtually all directions) at the same frequency at which it is oscillating. This is dipole
scattering; visible light interacts with the small particle and is scattered equally forward and
backward. As a result of this phenomenon, small particles are very effective in scattering
light. This phenomenon is important when studying the transmission of light through a
flue gas.
3.2.4 The Interaction of Light with Matter - The Beer-Lambert Law
The continuous emission monitors that utilize light in the measurement process apply the
Beer-Lambert law. (Consider Figure 3-6.)
LAMP
DETECTOR
LIGHT ABSORPTION
3-8
-------
The Beer-Lambert law states that the transmittance of light through a medium that absorbs
or scatters light is decreased exponentially by the product o-c/, or
T _ t, i _ -ad
i — i; IQ — e
where:
T = transmittance of light through the flue gas
Io = intensity of the light energy entering the gas
I = intensity of the light energy leaving the gas
a — attenuation coefficient
c = concentration of the pollutant
/ = distance the light beam travels through the flue gas.
The attenuation coefficient, a, is dependent upon the wavelength of the radiation and also
upon the properties of the molecule or particle. In the case of particulates, a characterizes
the effects of scattering and absorption. The coefficient tells how much a molecule will
absorb light energy at a given wavelength. If no absorption occurs, a will be zero, and the
transmittance would equal 100 percent. If an electronic or vibrational-rotational transition
occurs at some wavelength, a will be a large number, and the reduction of light energy
across the path / will depend upon the pollutant concentration, c. and the original intensity,
I0, of the light beam.
Utilizing this principle, an instrument for determining the concentration of a pollutant in
a flue gas can be designed. All that is needed is light having a wavelength that will cause
a transition in the molecule of interest and a light detector. 10 is determined by taking a
reading from the detector when no pollutant is in the duct or sample cell. The concentration
is obtained from the Beer-Lambert law if a and / are known. Generally, a calibration
curve is generated with known gas concentrations rather than using a theoretical value for
a (see Figure 3-7).
The complexity of modern monitoring instrumentation arises from the need to analyze one
specific pollutant in a sample containing many types of gases. The cross-stack and in-stack
gas monitors must also be designed to eliminate the effects of paniculate matter in reducing
the light transmission. There are, of course, problems in choosing and designing light
sources, detectors, and optical assemblies, as well as with the electronic circuitry. The
ability of an instrument manufacturer to solve the problems of specificity and design are
reflected in the operation of the monitor itself.
The approaches which instrument companies have taken in designing source level pollutant
monitors will be discussed in detail in the next two chapters. Opacity monitors, compared
to the gaseous emission monitors, are relatively simple and will be discussed in Chapter 4.
3-9
-------
-LOG I/I,
CALIBRATION
GAS 1
CALIBRATION
GAS 2
CALIBRATION |
GAS 3 <
t
CONCENTRATION
OF UNKNOWN
FIGURE 3-7
CALIBRATION CURVE FOR THE BEER-LAMBERT RELATION
Spectroscopic and nonspectroscopic techniques, such as polarography and electrocatalysis,
will be examined in the subsequent chapter.
3.3 References
1. U.S. EPA, "Standards of Performance for New Stationary Sources," Code of Federal
Regulations, 40 CFR, Part 60 Appendix B.
2. Ludwig, C. B., and Griggs, M., "Application of Remote Techniques in Stationary
Source Air Emission Monitoring," EPA-340/1-76-005, June 1976.
3.4 Bibliography
Willard, H. H., Merritt, L. L., and Dean, J. A., Instrumental Methods of Analysis, D. Van
Nostrand Company, Inc., Princeton, New Jersey, 1966.
Williamson, S. J., Fundamentals of Air Pollution, Addison-Wesley Publishing Co., Reading,
Massachusetts, 1973.
Conner, W. D., and Hodkinson, J. R., Optical Properties and Visual Effects of Smoke-
Stack Plumes, U.S. Dept. of Health, Education and Welfare, 1967 - PHSP No. 999-AP-30.
Barrow, G. M., Molecular Spectroscopy, McGraw-Hill Book Company, Inc., New York, 1962.
Kauzmann, W., Quantum Chemistry - An Introduction, Academic Press Inc., New York,
1962.
Instrumentation for Environmental Monitoring, LBL - I Vol. I: Air, Lawrence Berkeley
Laboratory, University of California, Berkeley, 1972.
3-10
-------
CHAPTER 4
CONTINUOUS MONITORS FOR OPACITY MEASUREMENTS
4.1 Opacity and Transmittance
The regulations require that an opacity monitor or transmissometer be installed on all
new coal- and oil-fired steam generators with a capacity greater than 73 megawatts. The
regulations covering opacity were made primarily to provide the plant operator with a
means of checking the operation of the source control equipment. However, the opacity
monitor is not, as yet, considered by EPA to be an enforcement tool for new sources,
since the visible emissions observer (EPA Reference Method 9) is still used to enforce
opacity standards. Data from opacity monitors may be used as evidence (see 40 CFR 60.11
and 42 FR 26205 5/23/77) in cases where there is a question of an opacity violation.
Opacity monitors on existing sources may be used for compliance purposes, depending on
the State regulations. In addition, the opacity monitor can serve as a process control
instrument by optimizing combustion conditions or control device efficiency.
The term "transmissometer" comes from a combination of "transmission" and "meter."
As mentioned in the previous chapter, when light passes through a plume or flue, some
of the light will be scattered and absorbed by paniculate matter in the plume. The absorbed
and scattered light will not reach the detector on the other side of the flue gas and will
be lost to observation. The transmission of the light through the gas is, therefore, decreased.
A transmissometer is essentially a meter that gives a quantitative value of the decrease in
light transmission.
If light is not able to penetrate through a plume, the plume is said to be opaque — the
opacity of the plume is 100 percent. Transmittance and opacity can be related in the
following manner:
Percent Transmittance = 100 — Percent Opacity
Therefore, if a plume or object is 100 percent opaque, the transmittance of light through
it is zero. If it is not opaque (zero percent opacity), the transmittance of light will be
100 percent. A plume from a stationary source rarely will have either zero or 100 percent
opacity, but some intermediate value. In the New Source Performance Standards (NSPS),
the opacity limits have been established for a number of stationary sources. The following
new sources are required to perform continuous monitoring for opacity, and to maintain
the opacity within the standard shown:
4-1
-------
Fossil-Fuel-Fired Steam
Generators
Petroleum Refineries
(Catalytic Cracker)
Ferroalloy Production Facilities
(Submerged Electric Arc Furnaces)
Iron and Steel Plants
(Electric Arc Furnaces)
Primary Copper, Lead, and Zinc
Smelters
Kraft Pulp Mills
(Recovery Furnace)
Percent Opacity Limit
20
30
15
4.2 The Transmissometer
20
35
A transmissometer may be constructed using either a single-pass system (Figure 4-1) or a
double-pass system (Figure 4-2). In the single-pass system, a lamp projects a beam of light
LIGHT SOURCE
COLLIMATING LENS
DETECTOR
COLLIMATING
LENS
ROTARY
BLOWER
FIGURE 4-1
SINGLE-PASS TRANSMISSOMETER SYSTEM
4-2
-------
LIGHT
BEAM
SPLITTER" DETECTOR H
RETRO-
REFLECTOR
ROTARY
BLOWER
FIGURE4-2
DOUBLE-PASS TRANSMISSOMETER SYSTEM
across the stack or duct leading to the stack, and the amount of light transmitted through
the flue gas is sensed by a detector. Such instruments can be made rather inexpensively;
however, they often do not satisfy specific EPA criteria for system zero and calibration
checks. The double-pass system shown in Figure 4-2 houses both the light source and light
detector in one unit. By reflecting the projected light from a mirror on the opposite side
of the stack, systems can be easily designed to check all of the electronic circuitry, including
the lamp and photodetector as part of the operating procedure. Most transmissometer
systems include some type of air purging system or blower to keep the optical windows
clean. In the case of stacks with a positive static pressure, the purging system must be
efficient or the windows will become dirty, leading to spuriously high readings. Figures 4-3
and 4-4 show a typical installation of a double-pass transmissometer.
As mentioned in Chapter 1, EPA does not recommend specific manufacturer models. Since
most stationary sources have unique monitoring problems, the Performance Specification
Test is used as a procedure for assuring that the instrument will operate properly once
mounted on a stack or duct. In addition, the transmissometer itself must satisfy several
design specifications. In order for a specific opacity monitoring installation to be approved,
it must meet these criteria.
4-3
-------
FIGURE 4-3
DOUBLE-PASS TRANSMISSOMETER INSTALLED AT EPA SOURCE
SIMULATOR FACILITY, RESEARCH TRIANGLE PARK, NC
FIGURE4-4
RETROREFLECTOR ASSEMBLY AT THE FACILITY
4-4
-------
4.3 Design Specifications
There are essentially seven design criteria that must be met by an opacity monitor:
1. Spectral Response - The system must project a beam of light with the wavelength
of maximum sensitivity lying between 500 and 600 nm. Also, no more than
10 percent of this peak response can be outside of the range of 400 to 700 nm.
2. Angle of Projection - The angle of the light cone emitted from the system is
limited to 5 degrees.
3. Angle of View - The angle of the cone of observation of the photodetector
assembly is limited to 5 degrees.
4. Calibration Error - Using neutral density calibration filters, the instrument is
limited to an error of 3 percent opacity.
5. Response Time - The transmissometer system must detect and identify 95 percent
of the value of a step change in opacity within 10 seconds.
6. Sampling - The monitoring system is required to complete a minimum of one
measuring cycle every 10 seconds and one data recording cycle every 6 minutes.
7. System Operation Check - The monitor system is to include a means of checking
the "active" elements of the system in the zero and calibration procedures.
Check the opacity monitoring instrument specifications before purchasing to assure that it
satisfies these minimum requirements. Failure to do so may mean that the monitor will
not be accepted by EPA.
There are several reasons for establishing these design specifications; the most important
is that there is no widely available independent method of checking the opacity. Instead,
it is assumed that if the system is designed correctly and if it can be checked with filters
for accuracy, it should be able to give correct flue-gas opacity readings. The rationale
behind each of the design specifications follows.
4.3.1 Spectral Response
The transmissometer is required to project a beam of light in the visible or photopic
region - that portion of the electromagnetic spectrum to which the human e£e is sensitive
(Figure 4-5).
4-5
-------
SPECTRAL CHARACTERISTICS
PHOTOPIC TUNGSTEN FILAMENT
SPECTRAL RESPONSE
100
INCANDESCENT LIGHT 2500° K
LLJ
ULTRAVIOLET
500
VISIBLE
1000
- INFRARED
1500 2000 2500
LIGHT
WAVELENGTH IN NANOMETERS
FIGURE 4-5
ELECTROMAGNETIC SPECTRUM
AND FACTORS THAT AFFECT OPACITY MEASUREMENTS
There are three reasons for specifying this region.
1.
2.
It was originally hoped to correlate the opacity readings of the transmissometer
with those of the visible emissions observer performing EPA Method 9. If the
transmissometer does project light in this region, generally the reading will be
comparable. However, problems of background light contrast, acid aerosol
formation, etc., may cause the readings of visible emissions observer to differ
from those of the transmissometer.
Water and carbon dioxide absorb light at wavelengths higher than 700 nm. If the
transmissometer projected light in this region (as some earlier systems did in fact),
any water vapor or carbon dioxide in the flue gas would take away some of the
light energy by absorption processes; a high opacity reading would result (see
absorption regions in Figure 4-5). For example, since this would unduly penalize
the operator of a fossil-fuel-fired boiler, filters or special optics are required to
limit the spectral response of the transmissometer.
4-6
-------
3. Particles less than 0.5 microns in size will scatter light more effectively if the light
has a wavelength in the region of 550 nm rather than at higher wavelengths
(Figure 4-6).
HI
103-
102-
101-
550 NANOMETER WAVELENGTH LIGHT
1000 NANOMETER WAVELENGTH LIGHT
0.1 0.2 0.3 0.4 0.5 1.0 2.0 3.0 4.0 5.0
PARTICLE DIAMETER IN MICRONS
10.0
FIGURE4-6
PARTICULATE ATTENUATION OF INCIDENT LIGHT
Industrial sources utilizing paniculate control devices emit particulates in the lower size
ranges. Consequently, shorter light wavelengths are needed to provide meaningful opacity
measurements.
4.3.2 Angle of Projection
The ideal transmissometer would have a collimated laser-sharp beam projected across the
stack. When a beam diverges, particles outside of the transmissometer path absorb or
scatter the light. Thus, light energy would be lost outside of the path, which would appear
as higher opacity readings. Since constructing sharply collimated instruments is expensive,
specifications have been given to limit beam divergence to 5 degrees, as shown in Figure 4-7.
The procedure for checking the angle of projection is to draw an arc with a 3-meter
radius, then measure the light intensity at 5-cm intervals for 26 cm on both sides of the
center line, both horizontally and vertically.
4-7
-------
26 cm
3 METER
ARC
COLLIMATING
OPTICS
F1GURE4-7
ANGLE OF PROJECTION
4.3.3 Angle of View
The reason for specifying the angle of view of the detector assembly is similar to that for
the projection angle specification. In this case, if the angle of view were too great, the
detector could possibly pick up light outside of the transmissometer light path. It would,
therefore, "see" more light energy than it should, and the transmissometer readings would
be lower than true (Figure 4-8).
26 cm
DETECTOR
26 cm
(3m)
FIGURE 4-8
ANGLE OF VIEW
4-8
-------
The angle of view may be checked by using a small nondirectional light source to find
out where, on an arc of 3-meter radius, a signal will appear. Generally, however, the
projection and detection angles are determined by the instrument manufacturer.
4.3.4 Calibration Error
Transmissometers are calibrated with neutral density filters corresponding to a given percent
opacity. The calibration error test is the best method for checking the accuracy of the
instrument. For that reason, before an instrument is placed on a duct or stack, the
instrument response to calibration filters should be within 3 percent of the predetermined
filter values.
4.3.5 System Response Time Test
The regulations require a transmissometer system to measure opacity every 10 seconds. An
approvable transmissometer must reach 95 percent of a calibration filter value within
10 seconds after being slipped into the light path in order to satisfy this design specification.
4.3.6 Sampling Criteria
EPA regulations specify that an approvable transmissometer must be able to complete a
minimum of one measuring cycle every 10 seconds (40 CFR 60.13e). Also, some provision
must be made in the monitoring system to record an averaged reading over a minimum of
24 data points every 6 minutes.
These specifications were made so that the opacity monitor would provide information
corresponding to the behavior of the paniculate control equipment and to the data obtained
by the visible emissions observer. (EPA Method 9 requires the reading of 24 plumes at
15-second intervals. See also, the discussion on page 4-17.)
4.3.7 System Operation Check
The system operation check often has not been recognized by instrument vendors as one
of the design criteria for transmissometer systems. In 40 CFR 60.13e3, it is stated that:
.. .procedures shall provide a system check of the analyzer internal optical surfaces
and all electronic circuitry including the lamp and photodetector assembly.
This means that when calibrating or zeroing the instrument, the lamp, photodetector,
etc., used should be the same as that used in measuring the flue gas opacity. Most
single-pass opacity monitors would not be acceptable under EPA design specifications,
since a zero reading could not be obtained unless the stack was shut down. Figure 4-9
shows the automatic zeroing mirror on a double-pass transmissometer.
4-9
-------
FIGURE 4-9
LEAR SIEGLER RM41-P SHOWING INSTRUMENT "ZERO" REFLECTOR
4.4 Installation Specifications
After an approved transmissometer has been selected by the source operator, the instrument
must be installed and checked for proper operation on the source itself. There are several
points that must be considered when installing a transmissometer:
• It must be located across a section of duct or stack that will provide a representative
measurement of the actual flue gas opacity.
• It must be downstream from the particulate control equipment and as far away as
possible from bends and obstructions.
• It must be installed in the plane of the bend if located in a duct or stack following
a bend.
• It should be installed in an accessible location.
• It may be required to demonstrate that it is obtaining representative opacity
values at its installed location.
These installation specifications are designed so that the transmissometer will measure the
actual flue gas opacity or "an optical volume which is representative of the particulate
4-10
-------
matter flowing through the duct or stack." Figure 4-10 shows some of the problems in
particulate matter flow distribution occurring in an exhaust system.
FIGURE 4-10
TRANSMISSOMETER SITING
Particulate matter may settle in ducts or stratify in the flue gas stream depending upon the
construction of the exhaust system. In Figure 4-10 the plane of the bend is formed by the
stack and the duct (in this case, the plane of the paper).
If a transmissometer were located perpendicular to this plane, such as at point A, Figure 4-10,
a large portion of the particulate matter would not be seen. A transmissometer located
at B would be in the plane of the bend and would be sensing a cross-section of the total
particulate flow. Location C would not be appropriate for an opacity monitor, since
the monitor would not be in the plane formed by the horizontal duct and the breeching
duct. A monitor at location C also would not satisfy criterion 1 or 2, since settling of
particulate matter might not provide a representative sample, and the location is close to
two bends in the exhaust system. Location D would be one of the most ideal points
for monitoring, since the transmissometer would be more accessible and might be more
4-11
-------
carefully maintained than if it was in location B. Location D comes after the control
device and does not follow a bend. The only problem that might arise is the settling of
paniculate matter in the duct and possible re-entrainment to give unrepresentative opacity
readings. An examination of the opacity profile over the depth of the duct might be
necessary to place the monitor at this point.
A portable in-situ, double-pass monitor, such as that shown in Figure 4-11, may be found
useful in examining possible installation sites. Proper monitor siting is very important to
the source operator, since an inappropriate choice for the location of a monitor may cause
measurement problems and may be costly, particularly if resiling were necessary.
FIGURE 4-11
LEAR SIEGLER RM41-P PORTABLE TRANSMISSOMETER
4.5 The Performance Specification Test
Before an opacity monitoring system can be used for EPA reporting requirements, it must
undergo the Performance Specification Test. Since most sources differ in operational
design and construction, a given monitor might perform well at one source, but might
produce unacceptable data at another. Also, since differences in paniculate stratification,
vibration, temperature, etc., affect operation, the opacity monitor must pass the performance
test at the location for which it was intended; design specifications alone are not sufficient
for approval (in contrast to ambient air monitors). A brief description of the test is given
here; specific test details are given in Chapter 11.
For the Performance Specification Test, the opacity monitors must undergo a I-week
conditioning period and a 1-week operational test period. In the conditioning period, the
monitor is merely turned on and is run in a normal manner. This is essentially a burn-in
period for the new instrument to eliminate those problems that one might expect for a new
device. In the operational test period, the monitor is run for 1 week without any corrective
maintenance, repair, or replacement of parts other than that required as normal operating
procedure. During this period, 24-hour zero and calibration drift characteristics are deter-
mined. If the instrument is poorly designed or if it is poorly mounted, these problems
4-12
-------
should become evident from the drift data, and corrective action would have to be taken.
Only zero, calibration drift, and response time data are necessary for the performance test.
The acceptable limits for these parameters are given in Table 4-1.
TABLE 4-1
OPACITY MONITOR PERFORMANCE SPECIFICATIONS
Conditioning Period — 1 week
Operational Period - 1 week
Zero Drift (24 Hr) - ^2% opacity
Calibration Drift (24 Hr) - <2% opacity
Response Time - 10 seconds
4.6 Data Reporting Requirements
After an opacity monitoring system has passed the Performance Specification Test, it may
be used to monitor the source emissions. New sources required to monitor opacity are
required to report excess emissions on a quarterly basis. Since opacity standards are
based on the opacity of the plume at the stack exit, the in-stack transmissometer data
must be corrected to the pathlength at the stack exit.
A term used in opacity monitoring called optical density (O.D.) is related to opacity in
the following manner:
O.D. = loglft i r-
&iu ! - opacity
This is a useful expression since, by considering the properties of paniculate scattering
and absorption, a linear relationship between paniculate concentration and optical density
results. The Beer-Lambert law for the transmittance of light through an aerosol states that
7 - e-naQL
or
(|_0):=e-naQL
4-13
-------
where:
T = transmittance
n = number concentration of particles
a = projected area of the particles
Q = particle extinction coefficient
L = light path through the aerosol
O = opacity
If the logarithm is taken of both sides, then
log(I - O) = 0.434 naQL
where:
0.434 is the conversion factor between the natural and base 10 logarithm
and
log ir^y =KcL
where:
K = a constant describing the characteristics of the particle scattering
L = the pathlength
c = the concentration (being proportional to n)
This merely states that O.D. = Kc, or that the optical density is proportional to the
particulate concentration.
If the diameter of the stack exit differs from the transmissometer pathlength, a relationship
between the two can be derived from a consideration of the optical density. Refer to
Figure 4-12.
4-14
-------
L! - EMISSION OUTLET
PATHLENGTH
L2 - MONITOR PATHLENGTH
OT = EMISSION OPACITY
O2 - MONITOR OPACITY
0, - 1 - (1 - 02) L1
FIGURE 4-12
RELATION BETWEEN EMISSION OPACITY AT PLUME EXIT
AND MONITOR OPACITY IN DUCT
Assuming that the concentration of the paniculate matter is the same at LI as it is at
L2,* the optical density across each path will be
O.D.i = log
(I -Oi)
= KcLi
Dividing the two
O.D.2 = log
(1 - 02)
= KcL2
O.D.i = -" O.D.2
L2
*Note that the velocities will change in order that the volumetric flowrate can remain the same.
4-15
-------
Taking the antilogarithms and solving for opacity, it is found that
Oi = 1 - (1 - Oz)
U/L2
Optical density is a useful parameter for calculating stack exit correlations for other cases.
For instance, if two ducts fed into a single stack and two transmissometers were used to
monitor the opacity (Figure 4-13), the following expression can be derived:
O.D.iAivi -^ + O.D.2A2V2~
ML2
O.D.3 -
where AI and A2 are the cross-sectional areas of each duct at the point of measurement, and
vi and V2 are the flue gas velocities in each duct. If the areas and velocities of each duct are
identical, this simplifies to
O.D.i
O.D.3 -
+ °-D-2 f1
* L2
The opacity at the stack exit can then be obtained from the optical density, O.D.3.
FIGURE 4-13
TWO DUCTS ENTERING COMMON EXIT STACK
4-16
-------
To satisfy the NSPS continuous monitoring regulations, the opacity must be measured
every 10 seconds. The data must be averaged and recorded every 6 minutes, with a
minimum of 24 equally spaced data points being used in the average. Dividing 24 into
6 minutes gives a measuring time of 15 seconds. This does not correspond to the minimum
required measuring time of 10 seconds. The discrepancy arises because a visible emissions
observer performing EPA Method 9 is required to average 24 plume opacity observations
at 15-second intervals, and the continuous opacity monitor reporting requirements were
made to correspond to EPA Method 9.
The transmissometer system must be able to record the average of at least 24 equally
spaced opacity readings taken over a 6-minute period. Any readings in excess of the
applicable standard (e.g., 20 percent opacity for a coal-fired boiler) must be reported. Also,
a report of equipment malfunctions or modifications must be made. Although the recorded
data do not have to be reported to EPA unless excessive emissions occurred, the data
must be retained for a minimum of 2 years.
4.7 Opacity Monitor Selection
The plant operator who selects an opacity monitor for a given application must consider
many factors. If the monitor must satisfy the NSPS continuous monitoring regulation's,
one of the first things to check would be whether the instrument satisfies the design
specifications established by the EPA. Additional criteria would be the capability of
satisfying the Performance Specification Test and the frequency-of-repair record.
Cost is always a major factor. Reliable transmissometers that require more frequent routine
maintenance than a top-of-the-line instrument are available at relatively lower cost. One
of the major factors in this consideration is the availability of an instrument technician
at the plant who could periodically check the monitor.
The vendors of opacity monitors are divided into those who market single-pass instruments
and those who market the double-pass systems. Most of the double-pass systems will
satisfy the EPA design criteria, and there are now a number of these monitors that have
passed the Performance Specification Test after having been placed on sources. The
single-pass instruments are less expensive than the double-pass systems, but most are
incapable, by virtue of their design, of meeting the EPA system zero and calibration
checks unless the stack is cleared every 24 hours. Single-pass systems, however, may be
applied in situations where there are less stringent requirements, such as process monitoring
or bag breakage in fabric filter paniculate control systems. Refer to Chapter 9 for a
detailed outline of selection procedures for opacity monitors and a list of vendors.
4.8 Bibliography
Avetta, E. D., "In-Stack Transmissometer Evaluation and Application to Paniculate Opacity
Measurement," EPA Contract No. 68-02-0660, Owens, Illinois NTIS PB-242-402, January
1975.
4-17
-------
Beutner, H. P., "Measurement of Opacity and Participate Emissions with an On-Stack
Transmissometer," Journal of the Air Pollution Control Association, V. 24, No. 9, Sep-
tember 1974.
Beutner, H. P., "Measurement of Opacity and Paniculate Emissions with the Lear Siegler
On-Stack Transmissometer," No. 73-169, 66th Annual Meeting of the APCA, June 24, 1973.
Beutner, H. P., "Monitoring of Paniculate Emissions from Cement Plants, Rock Products,"
May 1974.
Buhne, K. W., "Investigations into the Directional Dependence of Photoelectric Smoke
Density Measuring Instruments," Staub-Reinhalt, V. 31.
n
Buhne, K. W., and Diiwel, L., "Recording Dust Emission Measurements in the Cement
Industry with the RM4 Instrument," Staub-Reinhalt, V. 32:19, 1972 (in English).
Buhne, V. K. B., and Jockel, W., "Ortliche and Zeitliche Verteilung des Staubgehaltes in
Tauchgaskenalen grosser Dampfkesselanlagen," Staub-Reinhalt, V. 37:189-194, 1977.
Conner, W. D., "Measurement of the Opacity and Mass Concentration of Paniculate
Emissions by Transmissometry, Chemistry and Physics Laboratory," EPA-650/2-74-128,
November 1974.
Conner, W. D., "A Comparison Between in-Stack and Plume Opacity Measurements at
Oil-Fired Power Plants," Energy and the Environment — Proceedings of the Fourth National
Conference, AICHE, Dayton, Ohio, 1976, pp. 478^83.
Conner, W. D., and Hodkinson, J. R., "Optical Properties and Visual Effects of Smoke-
Stack Plumes," EPA Publication AP-30 (May 1972 - 2nd Printing).
Cristello, J. C., and Walther, J. E., "An Evaluation of an On-Stack Transmissometer, as a
Continuous Paniculate Monitor," APCA Article, 67th Annual Meeting, Denver, Colorado,
1974.
Dliwel, L., "Comparative Studies of Different Measuring Principles for the Continuous
Monitoring of Paniculate Emissions from Lignite Fired Boilers," Proceedings Second Int.
Clean Air Congress, Edited by H. M. England and W. T. Berry, Academic Press, New
York, 1971, pp. 437-496.
Ensor, D. S., and Pilat, M. J., "Calculation of Smoke Plume Opacity from Paniculate
Air Pollution Properties," Journal of the Air Pollution Control Association, V. 21:496, 1971.
Hamil, H. F., et al., "Evaluation and Collaborative Study of Method for Visual Deter-
mination of Opacity of Emissions from Stationary Sources," EPA-650/ 4-75-009, 1975.
4-18
-------
Haville, D., "A Single-Pass Photoelectric Opacity Measurement System," Proceedings,
Continuous Monitoring of Stationary Air Pollution Sources, APCA Specialty Conference,
APCA, 1975, pp. 154-170.
Hood, K. T., and Coron, A. L., "The Relationship Between Mass Emission Rate and
Observed Plume Appearances from Kraft Recovery Furnaces," 74-AP-08, Regional APCA
Meeting, Boise, Idaho, November 17, 1974.
Hood, K. T., and Coron, A. L., "The Relationship Between Mass Emission Rate and
Opacity," TAPPI, V. 60, No. 1:141-145, January 1977.
Hurley, T. F., and Bailey, D. L. R., "The Correlation of Optical Density with the Concen-
tration and Composition of Smoke Emitted from a Lancashire Boiler," J. Inst. Fuel,
V. 31:534-540, 1958.
Larssen, S., Ensor, D. S., and Pilat, M. J., "Relationship of Plume Opacity to the
Properties of Particulates Emitted from Kraft Recovery Furnaces," TAPPI, V. 55:88, 1972.
Lester, D. J., "Opacity Monitoring Techniques," Proceedings of the Workshop on Sampling,
Analysis and Monitoring of Stack Emission, NTIS PB-252-748, April 1976, pp. 31-48.
Lukacs, J., "Continuous Source Mass Emissions Monitoring — An Operations Guide,"
Proceedings, Continuous Monitoring of Stationary Air Pollution Sources, APCA Specialty
Conference, APCA, 1975, pp. 48-53.
McKee, H. C., "Texas Regulation Requires Control of Opacity Using Instrumental Measure-
ments," Journal of the Air Pollution Control Association, V. 24, No. 6, June 1974.
Molloy, R. C., "Smoke Opacity Monitoring Systems: Pollution Control and Energy
Conservation," ASHRAE Journal, September 1976, pp. 27-32.
Nader, J. S., "Current Technology for Continuous Monitoring of Paniculate Emissions,"
Journal of the Air Pollution Control Association, V. 25, No. 8, August !975.
Nader, J. S., Jaye, F., and Connor, W. D., "Performance Specification for Stationary-
Source Monitoring Systems for Gases and Visible Emissions," EPA Report 650/2-74-013,
January 1974.
Peterson, C. M., and Tomaides, M., "In-Stack Transmissometer Techniques for Measuring
Opacities of Particulate Emissions from Stationary Sources," EPA Report R2-72-099,
April 1972.
Pilat, M. J., and Ensor, D. S., "Plume Quality and Paniculate Mass Concentrations,"
Atmos. Environ., V. 4, No. 2:163-173, 1970.
4-19
-------
Reisman, E., Gerber, W. D., and Potter, N. D., "In-Stack Transmillometer Measurement
of Paniculate Opacity and Mass Concentration," EPA-650/2-74-120, November 1974.
Schneider, W. A., "Opacity Monitoring of Stack Emissions - A Design Tool with Promising
Results/ The 1974 Electric Utility. . .Generation Plan Book, McGraw-Hill, New York, 1974.
Sem, G. J., and Borgos, J. A., "State of the Art: 1971 Instrumentation for Measurement
of Paniculate Emissions from Combustion Sources," Vol. IV, NT1S PB-231-9I9/AS, Nat.
Tech. Inform. Serv., Springfield, Virginia, 1973.
Slowinski, Z., "New Construction of an Optical Dustmeter," Staub-Reinhalt, V. 37, No. 6:
232-234, 1977.
Williamson, S. J., Fundamentals of Air Pollution - Appendix C, "Light Scattering by
Small Particles," Addison Wesley Publishing Co., Reading, Massachusetts, 1973.
Woffinden, and Ensor, "Optical Method for Measuring the Mass Concentration of Paniculate
Emissions," EPA Contract No. 68-02-1749, Meteorology Research, Inc., EPA-600/2-76-062,
March 1976.
4-20
-------
CHAPTER 5
CONTINUOUS MONITORS FOR THE MEASUREMENT OF GASES
5.1 Introduction
Sources required to install continuous-gaseous emission monitors are faced with the problem
of selecting instruments that will give data that are representative of the actual source
emissions. The problems that are encountered in a sulfuric acid plant will be different
from those found at a primary smelter. Even within a given source category, the plant
design often will dictate the choice of a monitoring system. For example, an in-situ
sulfur dioxide monitor may work well on a coal-fired power plant that uses an electro-
static precipitator to control paniculate emissions, but it may encounter problems when
placed on a facility operating at high temperatures.
The proper selection or evaluation of a gas analyzer requires a knowledge of the practical
differences between extractive and in-situ systems and a knowledge of the operating principles
of the analyzers themselves. The remainder of this chapter and the next four chapters
present a basis for such an evaluation. These chapters cover:
• Extractive Monitoring Instruments
• Extractive System Design
• In-Situ Monitoring Systems
• Recording Systems
• Selection Procedures
The extraction of a sample gas from a stack or duct presents a number of problems
for the first class of continuous analyzers. To obtain accurate results, a sample repre-
sentative of the exhaust gas constituents first must be selected before entering the monitor
itself. The sample must be processed by removing paniculate matter, condensing water
vapor, and, in some cases, removing specific gases that interfere in the analytical method.
In-situ monitors, on the other hand, do not require the removal of particulates or water
vapor. The analytical methods used in this class of monitor have been chosen to avoid
these interferences. In-situ monitors do, however, have limitations in their application.
If a stack or duct contains entrained water in the form of liquid droplets, light scattering
problems and absorption of the pollutant gases in the liquid may cause the instrument
values to differ from those obtained by the EPA reference method. The choice of the
type of system (either extractive or in-situ) to be used in a given application often will
depend upon features of the plant design. The choice of a specific instrument will depend
5-1
-------
upon variables ranging from practical considerations, such as cost, to purely analytical
factors, such as the scientific principle that will give the most accurate concentration
data for a given pollutant.
The selection of a monitor also is dependent upon the EPA criteria for the Performance
Specification Test. A gaseous-emissions monitoring instrument must meet the following
specifications after it is installed on the source:
SO2 and NOX
O2 or CO2
Accuracy
Calibration error
Zero drift (2 hr & 24 hr)
Calibration drift
(2 hr & 24 hr)
Response time
Operational period
20%
5%
2% of span
2.5% of span
15 min (max)
168 hrs
—
-
<0.4% & <0.5% O2 or CO2
<0.4% & <0.5% O2 or CO2
10 min (max)
168 hrs
The Performance Specification Tests 2 and 3 for gases will be discussed in Chapter II.
Gaseous emission monitors, both extractive and in-situ, can be characterized by the principles
of chemical physics used. The methods used in source level analyzers can be grouped into
three major categories:
Absorption
Spectrometers
Luminescence
Analyzers
Elect roanalytical
Methods
Extractive monitors utilize methods from all of these categories, whereas in-situ systems
generally use spectroscopic absorption methods. An exception to this is an in-situ electro-
catalytic cell that monitors oxygen. There are a few special methods that do not fit into
this classification: paramagnetism is used in oxygen analyzers and thermal conductivity
is used in a few SO2 monitors. These methods will be discussed separately.
5.2 Extractive Analyzers
As mentioned in Chapter 2, extractive analyzers have had a longer developmental period
than the in-situ monitors. In the past, either existing ambient air monitors or common
laboratory instruments were modified for source-level monitoring applications. Problems
tended to arise with the inevitable dilution systems and delicate nature of some of these
systems. Many of these earlier problems now have been solved. Extractive analyzers are
now designed specifically to monitor at-source-level concentrations and are constructed
to withstand the rigors of a plant environment.
5-2
-------
5.2.1 Extractive Analyzers - Spectroscopic Methods of Analysis
5.2.1.1 Nondispersive Infrared Analyzers
Nondispersive Infrared (NDIR) analyzers have been developed to monitor SO2, NOX,
CO, CO2, and other gases that absorb in the infrared, including hydrocarbons. An
NDIR analyzer is basically an instrument that does not disperse the light that is emitted
from an infrared source. Not dispersing the light means not breaking up the emitted
radiation into its component wavelengths with a prism or diffraction grating. Dispersive
instruments, or dispersive absorption spectrometers, are most often found in the chemistry
or physics laboratory where they are commonly used to identify molecular compounds
from their infrared absorption spectra by continuously scanning over many wavelengths.
NDIR instruments utilize a broad band of light that is centered at an absorption peak
of the pollutant molecule, such as that shown in Figure 5-1.
100
TRANSMISSION
AS A FUNCTION
OF FREQUENCY
0
0
ABSORPTION
AS A FUNCTION
OF FREQUENCY
100
FREQUENCY
FIGURE5-1
A LORENTZIAN ABSORPTION CURVE
This broad band is usually selected from all the light frequencies emitted by the infrared
source, by using a bandpass filter. Table 5-1 gives the band centers for several of the gases
found in source emissions.
In a typical NDIR analyzer, such as that shown in Figure 5-2, infrared light from a lamp
or glower passes through two gas cells — a reference cell and a sample cell. The reference
cell generally contains dry nitrogen gas, which does not absorb light at the wavelength
used in the instrument. As the light passes through the sample cell, pollutant molecules
5-3
-------
TABLE 5-1
INFRARED BAND CENTERS OF SOME COMMON GASES*
Gas
NO
NO2
S02
H2O
CO
CO2
NH3
CH4
Aldehydes
Location of Band Centers (jum)
5.0-5.5
5.5-20
8-14
3.1
5.0-5.5
7.1-10
2.3
4.6
2.7
5.2
8-12
10.5
3.3
7.7
3.4-3.9
Wave Number
(cm'1)
1800-2000
500-1800
700-1250
1000-1400
1800-2000
3200
2200
4300
850-1250
1900
3700
950
1300
3000
2550-2950
Table from LBL-1, "Instrumentation for Environmental Monitoring*
SAMPLE T
CELL n I !
DETECTOR
REFERENCE CELL
FIGURE 5-2
SIMPLIFIED SCHEMATIC DIAGRAM OF A NONDISPERSIVE INFRARED ANALYZER
5-4
-------
will absorb some of the infrared light. As a result, when the light emerges from the
end of the sample cell, it will have less energy than when it entered. It also will have
less energy than the light emerging from the reference cell. The energy difference is
then sensed by some type of detector, such as a thermistor, a thermocouple, or micro-
phone arrangement. Figure 5-3 shows one of the more common commercial arrangements
for this type of system.
BEAM SAMPLE SAMPLE
CHOPPER IN EXHAUST DETECTOR SENSOR
INFRARED
SOURCE
FIGURE 5-3
OPERATION OF THE "MICROPHONE" DETECTOR OF AN NDIR ANALYZER
Infrared radiation passes through a reference and a sample cell. The microphone type
detector that is used consists of two chambers separated by a thin metal diaphragm,
each chamber being filled with gas of the species being measured. When the infrared
radiation strikes a pollutant molecule, the molecule will absorb light energy and will
move faster. This greater agitation for a number of molecules produces heat. This
heating, in turn, will increase the pressure in each chamber of the detector cell; however,
the light that passed through the sample cell will have lost some of its energy to the
pollutant molecules in the sample gas, and the sample chamber will not be heated as
much as the reference chamber. As a result, a pressure difference will develop, and the
diaphragm will be distended.
The greater the amount of the pollutant gas in the sample, the greater the displacement.
The displacement is detected as a capacitance change by the instrument electronics and
is ultimately processed to give a reading for the concentration of pollutant in the sample.
The rotating wheel chopper is used to create an alternating signal in the detector and,
hence, will make the signal easier to detect and amplify. Figure 5-4 shows a typical
configuration of a double-beam NDIR analyzer.
5-5
-------
FIGURE 5-4
INTERNAL VIEW OF A BECKMAN NDIR ANALYZER
A common problem with analyzers that use a detecting arrangement, as shown in Figure 5-3,
is that gases that absorb light in the same spectral region as the pollutant molecule will
cause a positive interference in the measurement. For example, water vapor and CO2 will
interfere in the measurement of CO using this arrangement (see also Table 5-1). These
gases must be removed by some scrubbing system before the sample gas enters the
analyzer. A unique solution to this problem is to put the detector cells in series instead
of in parallel, as shown in Figure 5-5.
The front chamber of the detector will absorb the infrared radiation primarily at the
frequencies in the center of an absorption band, such as that shown in Figure 5-1. Since
the front cell takes away energy from the light beam at the center frequencies, the rear
measuring chamber will absorb more of the energy in the outer edges of the band than
it will from the center. The geometries and gas concentrations of each measuring chamber
are chosen so that the pressure in each will be the same as when no pollutant molecules
are in the sample cell. Once pollutant molecules are introduced into the sample cell,
the amount of energy reaching the detector will be reduced; however, most of this reduction
5-6
-------
SENSOR
r-MOTOR
DIAPHRAGM
5 0
I*
INFRARED
SOURCE
BEAM CHOPPER
(-SAMPLE IN
SAMPLE CELL
BANDPASS
FILTER
FRONT
MEASURING
CHAMBER
'REAR
MEASURING
CHAMBER
DETECTOR
FIGURE 5-5
OPERATION OF A "NEGATIVE FILTER" NDIR ANALYZER
will arise from absorption at the band center, and the front chamber of the detector
will be less affected by the incoming radiation. The front chamber therefore, will be
cooler than the rear chamber, causing a pressure difference and a distention of the thin-
metal diaphragm. This method is often called negative filtering.
Interfering species generally will not have an absorption band that coincides exactly with
that of the species of interest. In such a case, absorption will occur relatively evenly
over the region, and the interference will be minimized. Several monitors have been
constructed utilizing this principle and need less supportive apparatus to remove such
species as water and CO2- A photograph of the Bendix 8501 single-beam analyzer,
which utilizes this method, is shown in Figure 5-6.
The advantages of the NDIR-type analyzers are their relatively low cost and the ability
to apply the method to many types of gases. Generally, a separate instrument is required
for each gas, although several instruments have interchangeable cells and filters to provide
more versatility. Problems associated with the method are those that arise from inter-
fering species, the degradation of the optical system caused by corrosive atmospheres,
and in some cases, limited sensitivity. The microphone type detectors are sensitive to
vibration and often require both electronic and mechanical damping, for example, by
placing the instrument on a foam insulation pad.
5-7
-------
FIGURE 5-6
INTERNAL VIEW OF A BENDIX NDIR ANALYZER
5.2.1.2 Nondispersive Ultraviolet Analyzers (NDUV) - Differential Absorption
Several available nondispersive systems use light in the ultraviolet and visible regions
of the spectrum rather than in the infrared. To analyze for SO2, these instruments
utilize one of the narrow absorption bands of the ultraviolet absorption spectrum (Figure 5-7).
NO2 may be determined by taking advantage of its absorption spectrum in the visible
region. The instruments that are designed to work within these regions do so in a
manner somewhat different from the NDIR method discussed previously. Essentially,
the analyzers measure the degree of absorption at a wavelength in the absorption band
of the molecule of interest (280 nm for SO2 and 436 nm for NO2, for example). This
is similar to the NDIR method, but the major difference is that a reference cell is
not used. Instead, a reference wavelength, in a region where SO2 or NO2 has minimal
absorption, is utilized. The rationale behind this method comes from the Beer-Lambert
law (which was introduced in section 3.2.4):
- I0e
-aC(
5-8
-------
2.0
1.8
1.6
S 1-4
| 1.2
CD
§ 1.0
m 0.8
<
0.6
0.4
0.2
0.0
NO.
578
250 300 350 400 450 500
WAVELENGTH (cm)
550 600
FIGURE 5-7
THE ULTRAVIOLET-VISIBLE SPECTRUM OF S02 AND N02
In the NDIR method, using a reference cell where o-eference = 0, two light intensities
are compared:
I , - i -Sample/ , _ , -aCreference /
'sample ~ 'oc 'reference ~~ 'oc
therefore
Ireference = Io = constant
I0 remains constant, and Isample can be related to it to obtain a concentration measure-
ment. In the ultraviolet system, on the other hand,
Creference — Csample ^ 0
The absorption coefficients are wavelength dependent, and the reference wavelength is
chosen so that ^reference = 0.
i . _ i _-OtCsample' i , _ i -aret'erence Csample /
•sample — lot 'reference — 'oc
^reference ~ 0
5-9
-------
therefore;
(reference = ID = constant
The ratio between Isample and I0 can then be taken to obtain a relation for the concentration.
(sample = loe"a(^mPle'
(reference IQ
and
. 'sample
, ^reference
Csample = :
al
where,
/ = known
or = known or instrument calibrated to account for the value
Isample /(reference = detected by the instrument
This method of analysis is often termed differential absorption, since measurements are
performed at two different frequencies. This method is not limited to extractive monitoring
systems, but it also is used in both in-situ analyzers and remote sensors.
Figure 5-8 shows a schematic of one of the more typical NDUV monitors. Instead
of using a reference cell (as in the ND(R systems), the instrument uses a reference wave-
length at 578 nm. Light from the mercury discharge lamp passes through the sample cell
to a beam splitter. The beam splitter, actually a semitransparent mirror, directs the
light to two separate photomultiplier tubes. Narrow bandpass filters allow light of only
the specified wavelengths to reach each of the photomultipliers. The reflected beam
passes through a 578-nm filter and is used to generate the reference signal in the detector.
The transmitted beam, however, passes through a 280-nm filter for an NO2 monitor.
Since SO2 will absorb light at 280 nm (NO2 at 436 nm), the amount of light or energy
reaching the phototube will be less than that reaching the reference phototube. The
resultant photomultiplier signals are amplified and processed to give a reading for the
pollutant concentration. Nitric oxide (NO) does not absorb in the spectral region covered
by the instrument and first must be quantitatively converted to NO2 for subsequent analysis.
This is done sequentially by stopping the flow in the NO2 sample cell, pressurizing it
with O2, and waiting approximately 5 minutes for the NO to be converted to NO2 by
the excess oxygen. The NO is then determined from the difference in the readings
before and after the reaction with oxygen.
5-10
-------
MEASURING
PHOTOTUBE
SEMITRANSPARENT MIRROR
{BEAM SPLITTER)
SAMPLE CELL
SO2/NOX
CALIBRATION FILTER 1
OPTICAL FILTER
SAMPLE CELL
SO2/NOX
IN OUT
ELECTRONICS
RECORDER
REFERENCE
PHOTOTUBE
FIGURE 5-8
OPERATION OF A DIFFERENTIAL ABSORPTION NDUV ANALYZER
The extractive analyzers using differential absorption have proven to be reliable in monitoring
source emissions. Several of the instrument models currently available are well built,
since they were designed for in-plant environment (Figures 5-9 and 5-10).
The differential absorption SO2 analyzers are somewhat more sensitive than are the NDIR
counterparts. The sequential nature of the NOx analysis may limit the utility of the
method in some cases. As with all extractive monitoring systems, particulate matter
should be removed before entering the analyzer. It is not necessary, however, to remove
water vapor in some of these systems (DuPont, specifically). A heated sample line
and heated cell prevent condensation in the analyzer. Since water does not absorb
light in this region of the ultraviolet spectrum, no interference occurs.
5.2.2 Extractive Analyzers - Luminescence Methods of Analysis
5.2.2.1 General
Luminescence is the emission of light from a molecule that has been excited in some
manner. Photoluminescence is the release of light after a molecule has been excited by
ultraviolet, visible, or infrared radiation. The emission of light from an excited molecule
created in a chemical reaction is known as chemiluminescence. The atoms of a molecule
5-11
-------
FIGURE 5-9
A DUPONT NDUV ANALYZER AT AN INDUSTRIAL SITE
FIGURE 5-10
INTERNAL VIEW OF A DUPONT ANALYZER SHOWING
MEASUREMENT CELL AND ASPIRATOR
5-12
-------
even can be excited to luminescence in a hydrogen flame. These three types of luminescent
processes are used in source monitoring applications. Monitors utilizing the effects of
luminescence can be very specific for given pollutant species and can have greater sensitivity
than some of the absorption or electrochemical methods. Monitors that use each of these
luminescent processes will be discussed in this section.
5.2.2.2 Fluorescence Analyzers for SO2
Fluorescence is a photoluminescent process in which light energy of a given wavelength
is absorbed and light energy of a different wavelength is emitted. In this process, the
molecule that is excited by the light energy will remain excited for about 1(T to 10
second. This period of time will be sufficient for the molecule to dissipate some of this
energy in the form of vibrational and rotational motions. When the remaining energy
is reemitted as light, the energy of the light will be lower, meaning light of a longer wave-
length (shorter frequency) will be observed. The fluorescence spectrum for SO2, shown
in Figure 5-11, illustrates this point.
0.3
0.2
0.1
0.0
QQ
CC
SO2 ABSORPTION SPECTRUM
S02 FLUORESCENCE EMISSION SPECTRUM A„„__.„,
* ABSORPTION
n
BANDPASS FLUORESCENCE
FILTER EMISSION
200
250 300 350
WAVELENGTH (nm)
400
FIGURE 5-11
FLUORESCENCE SPECTRUM OF S02
5-13
-------
The basis behind the fluorescence technique is to irradiate the molecule with light at a
given wavelength (usually in the near ultraviolet) and to measure the emitted light at a
longer wavelength.
Commercially available instruments contain either a continuous or a pulsed ultraviolet
light source (see Figure 5-12). The light from the source is filtered to a narrow region
that is centered near 210 nm in the near ultraviolet range where the SO2 molecule
will be excited. The fluorescent radiation is measured at right angles to the sample
chamber with a photomultiplier tube. Another filter is used to select only a portion of
the fluorescent radiation for measurement, since interferences can occur over the range
of the fluorescence emission spectrum. Figure 5-13 is an internal view of a TECO
fluorescence monitor. Fluorescence monitors have been applied successfully to monitoring
ambient air. Using these instruments in source monitoring requires attention to the
problem of quenching. In the process,
S02 + hr - SO2 - S02 + hi/
The excited SO2 molecule (SO2) may collide with another molecule before it can release
its extra energy as light. The energy, instead, will be lost in the collision and will make
the molecules move faster after the collision. Water, CO2, O2, N2, or any other molecule,
for that matter, can quench the emission of the radiation. The problem is, however,
that each of these molecules has a different quenching efficiency. If one changes the
composition of the background gases in a sample, such as having 5 percent O2 and
10 percent CO2 in a combustion gas, the SO2 reading obtained would be different from
that obtained if the background gas were ambient air containing 21 percent oxygen.
However, it so happens that the quenching effect of CO2 is approximately the same as
that of oxygen. A decrease of oxygen in a flue gas .generally means a relative increase
in CO2. The errors due to the differences range from 5 to 10 percent of the SO2 concen-
tration. The SO2 values can be corrected by knowing the CO2 and oxygen percentages
by means of a nomograph supplied by the instrument manufacturer.
The SO2 fluorescence monitors are customarily calibrated using SQ2 in air mixtures.
It has often happened that a technician will take a convenient cylinder of span gas
having SO2 in nitrogen instead of in air. Spanning the instrument with such a mixture
will cause the subsequent SO2 readings to be approximately 30 percent lower than the
true values. Ideally, the best way to span fluorescence analyzers for source application
would be to use a span gas with a composition similar to that of the stack effluent.
Fluorescence monitors, outside of the quenching problem, have no other significant inter-
ference problems. Particulates and water must be completely removed .from the sampling
stream before entering the sampling chamber or the instrument will be easily fouled.
Permeation tube dryers (see the following chapter) generally are used in the instrument
itself to eliminate any remaining water vapor that is not removed by the extractive system.
5-14
-------
210nm BANDPASS
FILTER
SAMPLE OUT
350 nm BANDPASS FILTER
ELECTRONICS RECORDER
TUBE
y
L.T1PL1EF
E
i
I
,
•;:::»:•:•::>:•:
<^ S
^ ::i;:
^*^fc •'''
|
FIGURE 5-12
OPERATION OF THE S02 FLUORESCENCE ANALYZER
FIGURE 5-13
INTERNAL VIEW OF A TECO FLUORESCENCE MONITOR
5-15
-------
5.2.2.3 Chemiluminescence Analyzers for NOX and NO2
Chemiluminescence is the emission of light energy that results from a chemical reaction
It was found in the late 1960's that the reaction of NO and ozone, O3, will produce
infrared radiation from about 500 to 3000 nm.
NO + 03 - NO2 + 02
NO2 - NO2 + hv (light)
Figure 5-14 shows the emission spectrum observed in this reaction. Monitors that measure
NO concentrations by observing the chemiluminescent radiation select only a narrow
region of the total emission; a filter is used to select light in the region from about
600 to 900 nm.
5 BANDPASS FILTER
j-
z
z 100 +
o
CO
CO
UJ
LII
LLf
CC
50- •
0
400
1200 2000 2800
WAVELENGTH nm "
FIGURE 5-14
THE CHEMILUMINESCENT EMISSION SPECTRUM OF NO2
Nitrogen dioxide (NO2) does not undergo this reaction and must be reduced to NO
before it can be measured by this method. Most commercial analyzers contain a converter
that catalytically reduces NO2 to NO.
NO2
Heat
Catalyst
*> NO +
The NO produced is then reacted with the ozone and the Chemiluminescence measured
to give a total NO + NO2 (NOX) reading. Figure 5-15 shows a schematic typical
of this class of instruments.
5-16
-------
NO2 TO NO
CONVERTER
FLOW CONTROL
SAMPLE IN
STEP1
NO + N02
STEP 2
NO +
NO {CONVERTED
FROM N02)
DETECTOR
CONTROL
iREACTION CHAMBER]
SAMPLE EXHAUST
FIGURE 5-15
OPERATION OF A CHEMILUMINESCENCE ANALYZER
Ozone is generated by the ultraviolet irradiation of oxygen in a quartz tube. The
ozone is provided in excess to the reaction chamber to ensure complete reaction and
to avoid quenching effects. Since the photomultiplier signal is proportional to the number
of NO molecules, not to the NO concentration, the sample flowrate must be carefully
controlled. The NO2 to NO converter chamber is generally made of stainless steel or
molybdenum to effect the catalytic decomposition. A few monitors on the market will
switch the sample gas automatically in and out of the converter to give almost continuous
readings for both NO and NO2.
The chemiluminescence method has been proven reliable (it is now an approved EPA
method for ambient air NOX analysis). Molecules, such as 02, N2, and CO2, quench
5-17
-------
the light radiation of this chemiluminescent reaction as in the fluorescence measurement
technique. The quenching problem has. however, been uniquely solved by choosing a
flowrate of ozone into the sample chamber much greater than that of the sample flowrate.
The resulting dilution gives a relatively constant background gas composition and the
effects caused by different quenching efficiencies of different molecules are minimized. The
only serious interference is ammonia, which will oxidize to NO in stainless steel converter
chambers. This is not usually a problem when the monitor is placed on a combustion
source, but care should be taken in other applications. Molybdenum converters operated
at lower temperatures will not oxidize such nitrogen compounds as ammonia.
5.2.2.4 Flame Photometric Analyzers for Sulfur Compounds
Another luminescence technique used to detect gaseous pollutants is that of flame photom-
etry. Flame photometric analyzers are primarily used in ambient air sampling, but have
been applied to stationary source sampling by using sample dilution systems.
Flame photometry is a branch of spectrochemical analysis in which a sample is excited
to luminescence by introduction into a flame. Instead of using an ultraviolet or visible
light source to excite the SO2 molecule, as in photoluminescence, a hydrogen flame
is used in the flame photometric method to excite the sulfur atom. The excited atom
will in turn emit light in a band of wavelengths centered at about 394 nm, which
is then detected by a photomultiplier tube, as shown in Figure 5-16. The method
is specific to sulfur, not to sulfur dioxide. Compounds, such as H2S, SO3, and mercaptans,
will contribute to the ultraviolet emission to give a measure of the total sulfur content
of the sample stream. With the use of scrubbers or chromatographic techniques, selective
determinations could be made of each of these compounds.
A disadvantage of flame photometric analyzers is the required hydrogen for the flame.
Facilities that have strict regulations concerning the use of hydrogen and hydrogen cylinders
may find it inconvenient to utilize this method. There are currently only a few manufacturers
of source-level flame photometric analyzers. The analyzers manufactured by Meloy contain
a dilution system within the analyzer. Meloy has recently completed an EPA development
contract for an H2S fuel-gas monitor using this method. The HaS monitor has proven
successful in field experiments and may soon become available commercially from Meloy.
An instrument panel of a Meloy analyzer is shown in Figure 5-17.
5.2.3 Extractive Analyzers - Electroanalytical Methods of Analysis
5.2.3.1 General
The instruments discussed in previous sections rely on spectroscopic, electro-optical tech-
niques to monitor particulates and gases. Another class of instruments based upon
5-18
-------
EXHAUST
FILTER
PHOTOMULTIPLIER
TUBE
SAMPLE
H2
ELECTROMETER
FIGURE 5-16
OPERATION OF A FLAME PHOTOMETRIC ANALYZER
eiectroanalytical methods of measurement has found great utility in source monitoring
applications. There are four distinct types of eiectroanalytical methods used in source
monitoring. These are:
Polarography
Electrocatalysis
Amperometric
Analysis
Conductivity
A number of monitors based on polarographic and electrocatalytic methods are available
for source monitoring applications. Polarographic analyzers have been developed for
a number of gases and can be inexpensive and portable, ideal for inspection work. Com-
plete continuous source-monitoring systems also are available from manufacturer* of these
instruments. The electrocatalytic or high temperature fuel-cell method, as it is often
called, is used to monitor oxygen only. Both extractive and in-stack monitors are available
using this technique. The methods of amperometric analysis and conductivity are less
widely used and are subject to a number of interferences. Descriptions of these methods
are given here, since a few instruments employing them are still marketed.
5-19
-------
SULFUR GAS ANALYZER
FIGURE 5-17
INSTRUMENT PANEL OF A MELOY FLAME PHOTOMETRIC ANALYZER
The principles behind the polarographic and electrocatalytic methods are somewhat more
difficult to understand than the spectroscopic principles discussed earlier. A combination
of classical electrochemistry and modern fuel-cell technology has provided the theoretical
bases for their development. An understanding of the underlying operational principles,
however, is important for their evaluation.
5.2.3.2 Polarographic Analyzers
Polarographic analyzers have been called voltammetric analyzers or electrochemical trans-
ducers. With the proper choice of electrodes and electrolytes, instruments have been
developed utilizing the principles of polarography to monitor SO2, NO2, CO, O2,
H2S, and other gases.
The transducer in these instruments is generally a self-contained electrochemical cell in
which a chemical reaction takes place involving the pollutant molecule. Two basic tech-
niques are used in the transducer: 1) the utilization of a selective semipermeable membrane
5-20
-------
that allows the pollutant molecule to diffuse to an electrolytic solution, and 2) the measure-
ment of the current change produced at an electrode by the oxidation or reduction of the
dissolved gas at the electrode. For SO2, the oxidation that takes place is:
S02 + 2H20 - S04"
_2
4H
2e~
E°298
Figure 5-18 shows a schematic of a typical electrochemical transducer.
SAMPLE IN
SEM1PERMEABLE
MEMBRANE
THIN FILM
ELECTROLYTE
SENSING ELECTRODE
BULK ELECTROLYTE
REFERENCE
ELECTRODE
SAMPLE OUT
OUTPUT
FIGURE 5-18
OPERATION OF AN ELECTROCHEMICAL TRANSDUCER
The generation of electrons at the sensing electrode produces an electric current that
can be measured. There are two reasons why this type of system may be termed
polarographic or voltammetric. In typical polarographic analyzers used in chemical lab-
oratories, the electric current in the system is related to the rate of diffusion of the
reacting species to the sensing electrode. It turns out that if the rate at which the
5-21
-------
reactant reaches the sensing electrode is diffusion controlled, the current will be directly
proportional to the concentration of reactant. This is known as Pick's law of diffusion:
- nFADc ,
i = — = kc
where:
i = current in amps
n = number of exchanged electrons per mole of pollutant
A = exposed electrode surface area
F = Faraday constant (96,500 coulombs)
D = diffusion coefficient of the gas in the membrane and film
c = concentration of the gas dissolved in the electrolyte layer (moles//im3)
d — thickness of the diffusion layer in cm.
This effect is characteristic of polarographic analyzers.
The other reason why this type of system is termed polarographic is that a retarding
potential can be maintained across the electrodes of the system to prevent the oxidation
of those species that are not as easily oxidized. There is a difference between the electro-
chemical transducers used for source monitoring and those used in the chemical laboratory.
In the laboratory instruments an external potential is applied to the system until the
decomposition potential of a given species is reached and an oxidation-reduction reaction
occurs. By varying the potential, both qualitative and quantitative information can be
obtained about the composition of a solution. The polarographic analyzers used in
source monitoring, however, act much like a battery. The oxidation-reduction reaction
occurs at the sensing electrode, because the counterelectrode material has a higher oxidation
potential than that of the species being reacted. In the cell, the sensing electrode has
a potential equal to that of the counterelectrode minus the iR drop across the resistor.
The sensing electrode is electrocatalytic in nature and, being at a high oxidation potential,
will cause the oxidation of the pollutant and a consequent release of electrons. This
can be seen from the example given in Figure 5-19.
The reaction that takes place at the counterelectrode is:
PbO2 + S04 + 4H+ + 2e~- PbSO4 + 2HaO E = I.68v
The half-cell potential of l.68v is in contrast to +0.17v for the oxidation of SO2 to
SO4 . Similar oxidation-reduction reactions occur for different pollutants and electrode
electrolyte systems. Figure 5-19 shows that the operation of these systems involves
5-22
-------
SAMPLE FLOW-W) » n
DIFFUSION — -»
DISSOLUTION
POROUS
SENSING
ELECTRODE
ABSORPTION
BULK
ELECTROLYTE
MEMBRANE
REACTION
POLAROGRAPHIC
ANALYZER
ELECTRONS
COUNTER ELECTRODE
FIGURE 5-19
DETAILS OF THE POLAROGRAPHIC PROCESS
1) diffusion of the pollutant gas through the semipermeable membrane, 2) dissolving of
the gas molecules in the thin liquid film, 3) diffusion of the gas through the thin liquid
film to the sensing electrodef4) oxidation-reduction at the electrode, 5) transfer of the
charge to the counterelectrode, and 6) reaction at the counterelectrode. The electron current
through the resistor then can be picked off as a voltage and suitably monitored.
The cells themselves come in a number of configurations, depending upon the manufacturer;
various claims are made about the response and selectivity of the instrument related to
the cell design. These systems are small and portable and compared to practically all
other source monitoring instruments, they are the least expensive. These two factors
make them ideal for source inspection, as warning detectors or even as dosimeters. An
example of such an inspection system is given in Figure 5-20.
If this method is used for continuous monitoring, a turn-key system should be purchased
from the vendor. Figure 5-21 shows an SO2 alarm monitor developed for industrial
application. The vendor will design and build a monitoring system to satisfy a given
5-23
-------
FIGURE 5-20
A PORTABLE INSPECTION SYSTEM USING A POLAROGRAPHIC ANALYZER
need using the experience gained over the many years of developing extractive systems.
Attempts by inexperienced technicians to save money by building monitoring systems with
inexpensive instruments and components usually result in innumerable problems and often
failure.
The polarographic analyzers in their earlier development were temperature sensitive, but
temperature compensation devices are now generally provided to avoid this problem. The
electrolyte of the cells generally will be used up in 3 to 6 months of continuous use.
The cells can be sent back to the company and recharged or new ones can be purchased.
It is extremely important that the sample gas be conditioned before entering these analyzers.
The stack gas should come to ambient temperature and the paniculate matter and water
vapor should be removed to avoid fouling the cell membrane.
With proper use, polarographic analyzers can be a valuable tool to an air pollution
agency's inspection program or to a source operator wishing to check pollutant levels at
various plant locations. Complete systems also are available for continuous monitoring,
but should be designed carefully so as to give accurate emission data.
5-24
-------
FIGURE 5-21
AN INDUSTRIAL S02 "ALARM" MONITOR USING A POLAROGRAPHIC ANALYZER
5.2.3.3 Electrocatalytic Analyzers for Oxygen
A new method for the determination of oxygen has developed over the past several
years as an outgrowth of fuel-cell technology. These so-called fuel-cell oxygen analyzers
are not actually fuel cells, but simple electrolytic concentration cells that use a special
solid catalytic electrolyte to aid the flow of electrons. These analyzers are available
in both extractive and in-situ (in-stack) configurations. This versatility of design is making
them popular for monitoring diluent oxygen concentrations in combustion sources.
In basic electrochemistry, one of the common phenomena studied is the flow of electrons
that can result when two solutions of different concentrations are connected together.
As an example, Figure 5-22 reviews this effect.
The electron flow results from the fact that the chemical potential is different on each
side and that equilibrium needs to be reached. There are two half-reactions that take
place in this example:
5-25
-------
c
ANODE
Ag = Ag+ + e
NO-
A- A9
DILUTE
AgN03 SOLUTION
CATHODE
Ag+ + e~ = Ag
CONCENTRATED
^AgNO3 SOLUTION
MEMBRANE POROUS TO NO3~
FIGURE 5-22
EXAMPLE OF ATYPICAL "CONCENTRATION" ELECTROCHEMICAL CELL
Ag = Ag+ (in dilute solution) + e
Ag+ (in concentrated solution) + e~ = Ag
The tendency for metallic silver to be oxidized to silver ion in a dilute solution of a
silver salt is greater than if it were in a concentrated solution. The transfer of electrons
effectively results in a transfer of silver ion from a more concentrated to a more dilute
solution. In this case, a porous membrane is placed between the two solutions to allow
the passage of nitrate ions (NO3~) to balance charges. The electromotive force (EMF),
or output voltage, that results from a concentration cell is described by the Nernst equation:
CMC RT i C*
EMF = —— In
4F C2
where:
R
T
= gas constant
= absolute temperature
F = Faraday constant
ci and C2 = concentrations of solutions.
5-26
-------
The instruments designed to continuously monitor oxygen concentrations utilize, instead,
different concentrations of oxygen gas expressed in terms of partial pressures. A special
porous material, zirconium oxide, serves both as an electrolyte and as a high temperature
catalyst to produce oxygen ions. A schematic of the electrocatalytic sensing system is
shown in Figure 5-23.
POROUS
GAS
OUT
Zr02 POROUS ELECTROLYTE ^ ELECTRON CURRENT J
0 « Q
PREF{02) > PSAMPLE (02)
FIGURE 5-23
OPERATION OF AN ELECTRO CATALYTIC OXYGEN ANALYZER
When sampling combustion gases, the partial pressure of the oxygen in the sample side
will be lower than the partial pressure of oxygen in the reference side, which is generally
that of air. When such a cell is kept at a temperature of about 850° C, oxygen molecules
on the reference side will pick up electrons at the electrode-electrolyte interface. The
porous ceramic material of ZrO2 has the special property of high conductivity for oxygen
ions. This occurs because the metal ions form a perfect crystal lattice in the material,
whereas the oxygen ions do not, resulting in vacancies. Heating the zirconium oxide
causes the vacancies and oxygen ions to move about. The oxygen ions migrate to the
electrode on the sample side of the cell, release electrons to the electrode, and emerge as
oxygen molecules. The EMF from this process, expressed in terms of the oxygen partial
pressures, is given as
=RI Pref(Q2)
4F Psample (02)
5-27
-------
This EMF can be measured. If the temperature is well stabilized and the partial pressure
of the oxygen on the reference side is known, the percentage of oxygen in the sample
can be easily obtained.
This phenomenon is used in some high temperature fuel cells. The oxygen analyzers
employing this technique, however, do not utilize fuels in the measurement and actually
cannot be called fuel cells. One problem with the method is that carbon monoxide,
hydrocarbons, and other combustible materials will burn at the operating temperature of
of the device. This will result in a lowering oxygen concentration in the sample cell,
which, however, would be insignificant for concentrations of the combustible materials
on the ppm level.
A number of manufacturers are presently marketing oxygen analyzers. Both extractive
and in-situ type systems have been developed, providing the source operator with versatility
in application. The in-situ system shown in Figure 5-24 employs a ceramic thimble to
eliminate particulates from the sample side of the cell. It should be noted that a constant
supply of clean dry air for the reference side of the cell is required. Calibration gases
can be injected into the measuring cavity contained within the ceramic thimble to check
the instrument operation.
FIGURE 5-24
A LEAR SIEGLER IN-SITU ELECTRO CATALYTIC OXYGEN
ANALYZER INSTALLED ON A POWER PLANT STACK
5-28
-------
5.2.3.4 Amperometric Analyzers
Amperometric analysis is a technique used in a few instruments developed for both
ambient and source monitoring. These analyzers (also called coulometric analyzers) measure
the number of coulombs required to produce a chemical reaction. Typically, amperometric
analyzers measure the current in an electrochemical reaction, such as, 2Br — Br2 + 2e~.
Sulfur dioxide will affect this reaction in the following manner:
SO2 + 2H2O + Br2 - H2SO4 + 2HBr.
The instrument measures the change of current flow caused by the change in the rate of
Br2 generation caused by the presence of SO2. However, amperometric instruments are
susceptible to interferences from compounds other than those of interest. Problems with
the necessary chemicals and associated plumbing also have made the application of these
systems somewhat limited in terms of continuous source monitoring. The technique,
however, is useful for the measurement of SO2, H2S, and mercaptans.
5.2.3.5 Conductimetric Analyzers
Conductimetric analyzers sense the change in the electrical conductivity in water when
a soluble substance is dissolved in it. This change of conductivity is proportional to
the concentration of the substance added and can be measured easily. The method,
however, is not entirely specific, since both SO2, NOX, and acid gases will change the
conductivity of water. Interfering gases, therefore, have to be removed before analysis.
Calibrated Instruments,' Inc. (Mikrogas-MSK), produces a Conductimetric analyzer that
absorbs SO2 in a hydrogen peroxide solution.
5.2.4 Extractive Analyzers - Miscellaneous Methods
5.2.4.1 Paramagnetic Analyzers for Oxygen
Molecules will behave in different ways when placed in a magnetic field. This magnetic
behavior will be either diamagnetic or paramagnetic. Most materials are diamagnetic and
when placed in a magnetic field will be repelled by it. A few materials are paramagnetic;
they are attracted by a magnetic field. Paramagnetism arises when a molecule has one
or more electrons spinning in the same direction. Most materials will have paired electrons;
the same number of electrons spinning counterclockwise as spinning clockwise. Oxygen,
however, has two unpaired electrons that spin in the same direction. These two electrons
give the oxygen molecule a permanent magnetic moment. When an oxygen molecule is
placed near a magnetic field, the molecule is drawn to the field and the magnetic moments
of the electrons become aligned with it. This striking phenomenon was first discovered
by Faraday and forms the basis of the paramagnetic method for measuring oxygen
concentrations.
5-29
-------
There are two methods of applying the paramagnetic properties of oxygen in the commercial
analyzers. These are the magnetic wind or thermomagnetic methods and the magneto-
dynamic methods:
• Magnetic Wind Instruments (thermomagnetic) - The magnetic wind instruments
are based on the principle that paramagnetic attraction of the oxygen molecule
decreases as the temperature increases. A typical analyzer utilizes a cross-tube
world with filament wire heated to 200° C (see Figure 5-25).
-GAS OUT
MAGNETIC FIELD
CROSS TUBE
AS
FIGURE 5-25
OPERATION OF A "MAGNETIC WIND" PARAMAGNETIC OXYGEN ANALYZER
A strong magnetic field covers one half of the coil. Oxygen contained in the
sample gas will be attracted to the applied field and enter the cross-tube. The
oxygen then heats up and its paramagnetic susceptibility is reduced. This heated
oxygen will then be pushed out by the colder gas just entering the cross-tube. A
wind or flow of gas will therefore continuously pass through the cross-tube. This
gas flow will, however, effectively cool the heated filament coil and change its
resistance. The change in resistance detected in the Wheatstone bridge circuit
can be related to the oxygen concentration.
Several problems can arise in the thermomagnetic method. The cross-tube
filament temperature can be affected by changes in the thermal conductivity of
the carrier gas. The gas composition should be relatively stable if consistent
5-30
-------
results are desired. Also, unburned hydrocarbons or other combustible materials
may react on the heated filaments and change their resistance.
• Magneto-dynamic Instruments - The magneto-dynamic method utilizes the para-
magnetic property of the oxygen molecule by suspending a specially constructed
torsion balance in a magnetic field. Here, a dumbbell-shaped platinum ribbon
is used. Since platinum is diamagnetic, the dumbbell will be slightly repelled
from the magnetic field. When a sample containing oxygen is added, the
magnet attracts the oxygen and field lines surrounding the dumbbell are changed.
The dumbbell then will swing to realign itself with the new field. Light reflected
from a small mirror placed on the dumbbell then can be used to indicate the
degree of swing of the dumbbell, and hence, the oxygen concentration.
All of the commercial paramagnetic analyzers are extractive systems. Water and particulate
matter have to be removed before the sample enters the monitoring system. It should
be noted that NO and NQz are also paramagnetic and may cause some interference in
the monitoring method if high concentrations are present.
5.2.4.2 Thermal Conductivity Analyzers
Thermal conductivity analyzers operate on the principle that different gases will conduct
heat differently. When a sample gas flows over a heated wire, the wire will be cooled
and the resistance of the wire will change accordingly. If the composition of the sample
gas changes, the cooling rate and the resistance of the wire will change to give an indication
of the gas composition. A Wheatstone bridge circuit is generally used to detect the
resistance changes in the heated wire.
Thermal conductivity analyzers have been used to monitor CO2, SO2, and other gases
in process gas streams. A disadvantage to the method is that a flow of reference gas must
always be maintained. Changes in the composition of the gas stream other than those
due to changes in the pollutant level will interfere in the measurement. Scrubbing systems
or some other methods would be necessary in such cases for accurate measurements.
5.3 Bibliography
Allen, J. D., "A Review of Methods of Analysis for Oxides of Nitrogen," J. of fast.
of Fuel, March 1973, pp. 123-133.
Allen, J. D., Billingsley, J., and Shaw, J. T., "Evaluation of the Measurement of Oxides
of Nitrogen in Combustion Products by the Chemiluminescence Method," J. of/nst. of Fuel,
December 1974, pp. 275-280.
Barrett, D. F., and Small, J. R., "Emission Monitoring for SO2 and NOX from Stationary
Sources," presented at 7th. National Meeting, American Institute of Chemical Engineers,
New Orleans, Louisiana.
5-31
-------
Cheremisinoff, P. N., and Young, R. A., "New Developments in Air Quality Instru-
mentation," Pollution Engineering, February 1975, pp. 24-29.
Corning Glass Works, "The Oxygen Sensor: Key to Furnace Control," Plant Energy
Management, January/February 1978, pp. 18-19.
Feldman, J., "Continuous Stack Analyzer for Multicomponent Analysis," Air Quality
Instrumentation, V. 2:147-154, Instrument Society of America, Pittsburgh, 1974,
Heyman, G. A., and Turner, G. S., "Some Considerations in Determining Oxides of
Nitrogen in Stack Gases by Chemiluminescence Analyzer," Paper 13.18, presented at the
22nd Annual ISA Analysis Instrumentation Symposium, May 9-12, 1976, San Francisco,
California.
Hodgeson, J. A., McClenny, W. A., and Hanst, P. L., "Air Pollution Monitoring by
Advanced Spectroscopic Techniques," Science, V. 182:248-258, October 1973.
Homolya, J. B., "Current Technology for Continuous Monitoring of Gaseous Emissions,"
Journal of the Air Pollution Control Association, V. 24, No. 8, August 1975.
Hollowell, C. D., McLaughlin, R. D., and Stokes, J. A., "Current Methods in Air Quality
Measurements and Monitoring," IEEE Trans, on Nuclear Sci., NS.22/No.2, pp. 835-848.
1975.
Huntzicker, J. J., Isabelle, L. M.» and Watson, J. G., "The Continuous Measurement
of Paniculate Sulfur Compounds by Flame Photometry," Paper 76-31.3, presented at 1976
APCA meeting, Portland, Oregon.
Jahnke, J. A., Cheney, J. L., and Homolya, J. B., "Quenching Effects in SO2 Fluorescence
Monitoring Instruments," Environmental Sci. and Technology, V. 10:1246, 1976.
Jahnke, J. A., "Gaseous Emission Monitors," Continuous Monitoring for Source Emissions:
Course Manual, Air Pollution Training Institute, Research Triangle Park, North Carolina,
August 1977.
Kikuchi, M., et al., "Mitsibushi Stack Gas Analyzer, Model SA-302," Mitsubishi Denki
Giho, V. 48:459, 1974.
Koltoff, I. M., and Miller, C. S., "Polarographic Determination of Sulfite," J. Amer.
Chem. Soc.. V. 63:2818, 1941.
Nader, J. S., "Developments in Sampling and Analysis Instrumentation for Stationary
Sources," Journal of the Air Pollution Control Association, V. 23, No. 7, July 1973.
5-32
-------
Nader, J. S., "Source Monitoring," Chapter 15, Air Pollution, V. 4, A. C. Stern, Ed.,
Third Edition, Academic Press, New York, 1977.
Neuberger, E., "Reliable Oxygen Measurement Optimizes Fuel Cost," ISA Paper 74-721.
Okabe, H., Splitstone, P. L., and Ball, J. J., "Ambient and Source SO2 Detection
Based on a Fluorescence Method," Journal of the Air Pollution Control Association,
V. 23:514, 1973.
Parts, L., et al., "A Review of Instrumental Techniques for Monitoring Nitrogen Oxides
Emissions from Stationary Sources/* Air Quality Instrumentation, V. 2:204-215, Instrument
Society of America, Pittsburgh, 1974.
Robertson, D. J., Groth, R. H., and Gardner, D. G., "Interferences and Oxygen Errors
in NDIR Analyses for CO and CO2 (old Beckman models),* Paper 77-27.3, presented at
the 70th Annual Meeting of APCA, Toronto, Canada, June 20-24, 1977.
Rollins, R., *A Continuous Monitoring System for Sulfur Dioxide Mass Emissions from
Stationary Sources," Paper 77-27.5, presented at the 70th Annual Meeting of APCA,
Toronto, Canada, June 20-24, 1977.
Rosenthal, K.., and Bambeck, R. J., "Continuous Monitoring of Stack Gases," Air Quality
Instrumentation, V. 2:179-183, Instrument Society of America, Pittsburgh, 1974.
Ross, D. T., Pocock, R. E., and McGandy, E, T., "Electrochemical Oxygen Analyzer with
Dry Jet Sampling System," Preprint Paper No. 75-60.4, Annual Meeting Air Pollution
Control Association, Pittsburgh, Pennsylvania, 1975.
Saltzman, R. S., and Williamson, J. A., "Monitoring Stationary Source Emissions for
Air Pollutants with Photometric Analyzer Systems," Air Quality Instrumentation, V. 1:169-
177, Instrument Society of America, Pittsburgh, 1972.
Seymour, S. J., "Gas Analysis Instrumentation," Instrumentation Technology, July 1975.
Shaw, M., and Shaw, M. D., "Membrane Polarographic Sensors in Air Pollution Analysis,"
Proceedings, Continuous Monitoring of Stationary Air Pollution Sources, APCA Specialty
Conference, APCA, 1975, pp. 54-63.
Stevens, R. K., and Herget, W. F., "Analytical Methods Applied to Air Pollution Measure-
ments," Chemistry and Physics Laboratory, National Environmental Research Center, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina; Ann Arbor
Science Publishers, Inc., P.O. Box 1425, Ann Arbor, Michigan, 1974.
5-33
-------
Wallace, D. A., "Alarm Level Monitor for SO2 Emissions from Stationary Sources,"
EPA-600/2-77-130.
Warner, P. O,, Analysis of Air Pollutants, Wiley & Sons, New York, 1976.
Williamson, J, A., "If You Have to Monitor SO2 and NOX, Then Choose Your Instruments
Wisely," Instruments & Control Systems, Buyers Guide Issue, 1975.
Williamson, J. A., Jr., "Oxidation of Nitric Oxide to Nitrogen Dioxide for Photometric
Measurement of NOX on Emission Source Monitoring," Air Quality Instrumentation,
V. 2:109-116, Instrument Society of America, Pittsburgh, 1974.
Wolf, P. C, "Continuous Stack Gas Monitoring Part One: Analyzers," Pollution Engineering,
June 1975, pp. 32-36.
Young, R. A., and Cheremisinoff, P. N., "New Developments in Industrial Pollution Control
Measurement and Instrumentation," Pollution Engineering, February 1976, pp. 22-28.
Zolner, W., Cieplinski, E., and Helm, D., "Source Level SO2 Analysis via Pulsed Fluo-
rescence," in Analysis Instrumentation, (W. V. Dailey, J. F. Comb, T. L. Zinin, eds.)
Proc. 20th Annual ISA Analysis Instrumentation Symposium, 1974. Available from Thermo-
Electron Corp., Waltham, Massachusetts.
5-34
-------
CHAPTER 6
EXTRACTIVE SYSTEM DESIGN
6.1 Introduction
The discussion of continuous-gaseous emissions monitors up to this point has covered
only the measurement principles used in the currently available commercial extractive
analyzers. An extractive analyzer cannot provide reliable monitoring data without a
properly designed sampling interface. The total extractive system must perform several
functions:
• Remove a representative gas sample from the source on a continuous basis.
• Maintain the integrity of the sample during transport to the analyzer (within
specified limits).
• Condition the sample to make it compatible with the monitor analytical method.
* Allow a means for a reliable calibration of the system at the sampling interface.
The design of the sampling interface, including the components used in its construction,
will depend on the characteristics of both the source gas stream and the monitoring
instrument. Emphasis is placed here on the design of the minimum system that will
present a minimum capital investment and low operating and maintenance costs.
The procedure recommended for designing an extractive monitoring system includes the
following steps:
* Study Federal regulations to determine which pollutant gases must be monitored.
• Review specifications and operating characteristics of several analyzers that could
monitor these gases.
• Determine the gas stream parameters at the most feasible sampling sites for
the given source.
• Select the best sampling site.
• Select an analyzer most compatible with the sampling site and gas parameters.
• Design a gas sampling interface that will provide the analyzer with a properly
conditioned and representative gas stream sample.
-------
This chapter will discuss primarily the information necessary for making evaluations and
decisions about sampling interface design.
6.2 Gas Stream Parameters
The gas temperature and velocity profile at all contemplated sampling sites should be
determined first. This must be done for any potential sampling site that is not located
8 or more duct diameters downstream of a disturbance to the gas stream. The Federal
regulations require that a representative gas sample be extracted. A temperature and
velocity profile of the gas stream for locations less than 8 duct diameters downstream of
a flow disturbance may give some indication whether or not gas stratification exists.
The paniculate loading in the gas stream and the character of the particulates should be
evaluated. All extractive systems will require the filtration of particulates from the gas
sample stream. The paniculate character and loading will affect decisions for coarse and
fine filtration systems, sampling pump location in the extractive system, and maintenance
scheduling. The reactivity of the particulates toward sulfur dioxide and/or oxides of
nitrogen may need to be evaluated (this generally has not been a major consideration;
it is worth noting, however). The presence of acid mist and/or water droplets in the
gas stream also will effect sampling interface design.*
The water vapor content of the stack gas should be determined. The amount of water
vapor present in the stack gas is an important consideration in designing the sample
conditioning system for the analyzer. The water vapor content of the stack gas, the
analyzer requirements, and sample-gas flowrate are needed to calculate the water removal
and drainage needs of the sampling interface. This will assist in making decisions on
whether it is necessary to dry the gas stream and on the type of system to use.
The duct absolute pressure may be an important parameter in terms of pump and system
valve requirements. It also can be a factor in determining calibration gas injection into
the system.
The use of a single analyzer stream to monitor multiple sources requires that all preceding
considerations be evaluated for all of the sources to be monitored. It is also necessary
to determine the ability of the analyzer to monitor all possible pollutant concentrations
from the various sources.
*Note: Acid mists and/or entrained water droplets are special cases that may require
extra care in sampling system design. It is recommended that these situations be discussed
with Agency personnel for approval prior to installation.
6-2
-------
6.3 Sampling Site Selection
6.3.1 SO2/NOx Monitors
Obtaining a representative sample of the pollutant gases is the most important item to
consider when selecting a sampling site for an extractive SQ2/NQx sampling system.
The gas sample analyzed must be directly representative or able to be corrected to be
representative of the total emissions from the source. A representative sample of the
stack gas may be taken at a point where the gases are not stratified. Nonstratified means
that the difference between the pollutant gas concentration at any point in the duct more
than 1 meter from the duct wall and the average pollutant concentration in the duct is
less than 10 percent of the average pollutant concentration. The effluent gases generally
are assumed to be nonstratified if the sampling site is located 8 or more duct diameters down-
stream of any air in-leakage or confluence of different gas streams. This general case does not
apply to -sampling locations upstream of an air preheater in a steam generating facility.
A sampling location less than 8 duct diameters from air in-leakage must be proven to be
consistently representative or corrected to be consistently representative of the total emissions
from the facility. It must be shown that the point of average pollutant gas concentration
does not shift with changes in the operation of the facility. As a result, a gas concentration
profile study is essential for sampling locations being considered for continuous monitoring
applications which do not satisfy the 8 duct-diameter criterion.
6.3.2 O2/CQ2 Monitors
An 02/CQz monitor is used to convert continuous monitoring pollutant concentration
data to units of the applicable standard. The 02/CO2 monitor must, therefore, be
located at a point where measurements can be made that are representative of the pollutant
gases sampled by the SQ2/NOX monitor(s). The Q2/CO2 monitor sampling point location
conforms best with this requirement when it is at approximately the same point in the
duct as the SQz/NOx system. The Qz/CQz gas sample may be extracted from a different
duct location if the stack gas is nonstratified at both locations and there is no air
in-leakage between the two sampling points. If the O2/CQ2 monitor sampling point
is at a different location from the SQ2/NOX sample point and stratification exists in
that duct, a multipoint extractive probe must be used for sampling. This is also true
for the extractive monitoring system when the Qz/CCfc and SQz/NOx monitors are not
of the same type (i.e., one is extractive and the other in-situ).
6.3.3 General Comments
The final sampling site selected for continuous monitoring applications must meet the
guidelines given in the Federal Register. Several other factors for installation of the
extractive sampling interface also must be considered. These include accessibility to the
monitor and the interface, system response time, and overall system design.
6-3
-------
The sampling site should be easily accessible to plant personnel. A continuous monitor
and the sampling interface will need maintenance and calibration by the plant. All
components of the system should be as accessible as circumstances will permit. The
monitor should be located near the sampling site. This may mean constructing a housing
for the monitor to protect it from the environment. The plant may find such an arrangement
inconvenient, choosing instead to put the monitor in the plant control room with the
sampling interface extended to supply the monitor with the gas sample. The response
time of the system for long sampling interface connections must then be considered.
An analyzer placed in a control room, away from the sampling site, may require a
slightly more complex extractive system. This situation is not prohibitive in cost or
operation and may be the best arrangement for a given operation.
The final site selection requires an evaluation of all aspects of accessibility, maintenance,
response time, convenience, and gathering of representative data. The decision may involve
some trade-offs.
6.4 Analyzers) Selection
The analyzer selected for continuous monitoring must be compatible with the source
characteristics, sampling site, intended location for installation, and sampling interface.
The engineer involved in installing the continuous monitoring system and having performed
an examination of gas stream characteristics and site location will now have a basis for
choosing the analyzer best suited for the source. The requirements of the analyzer and
source gas parameters then will determine the design of the sampling interface.
6.5 Design of the Sampling Interface
6.5.1 General
The design of a sampling interface requires that the system deliver a conditioned, continuous
gas sample to the gas analyzer. A number of different interface designs may be able to
perform this task at a given source. The actual system designed for a specific source
generally will incorporate a variety of trade-offs based on source/analyzer requirements
and financial restraints. A system typically will include the following components:
• In-stack sampling probe
• Coarse in-stack filter
• Gas transport tubing
• Sampling pump
• Moisture removal system
6-4
-------
• Fine filter
• Analyzer
• Calibration system
• Data recorder
6.5.2 Sampling Probe
Representative gas sampling requires samples that will demonstrate the total pollutant gas
emissions from a source. The temperature and velocity traverse across the duct may
indicate a necessity for a multipoint probe to extract samples from numerous points
across the entire duct. Several research studies have shown that although gas concentration
cannot be assumed to correspond directly to temperature and velocity gradients in a duct,
these measurements are excellent indications for positioning gas sampling probes. This
research has shown that a representative gas sample may be extracted from a grid of
equal areas laid out in the duct. A temperature and velocity traverse is then performed
in each row of the grid. The multipoint gas sampling probe is then positioned across
the row that indicated temperature and velocity readings closest to the average reading
in the duct.
Gas sampling requires that particulates, which can harm the analyzer and shorten the
operating life of the sample pump, be removed from the gas stream. Directing the probe
inlet counter current to the gas flow helps prevent many large particulates from entering
the system. Particulates that enter the probe can be removed by coarse and fine filters.
6.5.3 Coarse Filters
The coarse filter should be located at the probe tip in the stack, where it then can prevent
particulate matter from plugging the sampling probe and will not require heat tracing to
prevent moisture condensation. There are two general types of coarse in-stack filters:
external or internal.
The external coarse filter is a porous cylinder (see Figures 6-1 and 6-2). The cylinder
is typically constructed of sintered 316 stainless steel, though it may also be glass, ceramic,
or quartz. It is essential that the porous cylinder be protected by a baffle to prevent
excessive particulate buildup on the leading edges. These porous cylinders have an
expected utility of approximately 2 to 3 months before they become clogged with particulate,
depending on the sampling rate. Although they can be regenerated by back flushing,
they will eventually need replacement. The nominal cost (~$25) suggests that it may be
easier to replace the filter on a routine basis rather than install costly automatic back-
flushing equipment.
6-5
-------
POROUS
CYLINDER
, i,\ • , fij^^^^^^^^^^^^
,4i:\r ••..••;;• A
. .. \ ; .,„, . , BAC
BAFFLE'
STACK
WALL
SAMPLING INTERFACE
STACK GAS '' i / '":; f '' •'
FIGURE 6-1
POROUS CYLINDER USED AS EXTERNAL COARSE FILTER
FIGURE 6-2
ACTUAL POROUS CYLINDER INSTALLED IN A STACK GAS STREAM
The internal filter is housed within a tube (Figure 6-3). The gas enters a probe nozzle,
passes through the filter, and proceeds into the sampling interface.
Filter material is available from a number of manufacturers. It has been shown experi-
mentally that a Western Precipitation Alundum thimble permits high paniculate loading
with a minimal pressure drop. Other filters and filter holders have lower paniculate
loading capacities. Glass wool filters have been used in some experiments; however,
they have a higher pressure drop than the Alundum thimble. The internal filter arrange-
ment is preferred because it allows easier injection of calibration gases (see pump
configurations).
6-6
-------
NOZZLE
STACK WALL
SAMPLING INTERFACE
FIGURE 6-3
INTERNAL COARSE FILTER
6.5.4 Fine Filters
The majority of extractive stack gas analyzers require almost complete removal of all
particles larger than 1 micron from the gas stream. This is best accomplished by including
a fine filter near the analyzer inlet. Fine filters are divided into two broad categories:
surface filters and depth filters.
Surface filters remove particulates from the gas stream using a porous matrix (Figure 6-4).
The pores prevent penetration of particulates through the filter, collecting them on the
PARTICULATE
CAKE
GAS
STREAM
SURFACE
FILTER
+* CLEAN GAS
ALL PARTICLES
< 1 MICRON
FIGURE 6-4
SURFACE FILTER
6-7
-------
surface of the filter element. Surface filters can remove particulates smaller than the
actual filter pore size with paniculate cake buildup and electrostatic forces acting to trap
smaller particles. These filters perform well on dry, solid particulates without excessive
pressure drop. A surface filter will foul quickly if it becomes wet or if the paniculate
is gummy.
Depth filters collect particulates within the bulk of the filter material. A depth filter
may consist of loosely packed fibers or relatively large diameter granules (Figure 6-5).
These filters perform well for gummy solids or moist gas streams and dry solids. In the
case of malfunction, their flexibility can protect the analyzer from damage. Glass wool
packed to a density of 0.1 gm/cm3 and a bed depth of at least 2 inches can act as an
inexpensive depth filter for normal gas flowrates. These filters must be carefully packed
to avoid channeling.
CLEAN GAS
ALL PARTICLES
< 1 MICRON
1 GAS STREAM
FIGURE 6-5
DEPTH FILTER
6.5.5 Gas Transport Tubing
The gas tubing or sample lines transport the extracted gas sample from the stack through
the interface system and into the analyzer. When selecting sampling lines, it is important
to consider:
• Tube interior-exterior diameter
• Corrosion resistance
• Heat resistance (for lines near high temperature areas or heat tracing)
• Chemical resistance to gases being sampled
• Cost
The gas tubing must be sized to ensure an adequate gas flowrate with a reasonable
pressure drop and good system response time. A flowrate of 2 standard liters per minute
(enough to supply two gas analyzers) through a 6.35-mm OD (1/4 inch) tubing exhibits
a pressure drop between 1 and 3 mm Hg per 30.48-meter length. This pressure drop
6-8
-------
is quite acceptable for most sampling pumps. The response time (t) for a sampling
line volume (V) can be calculated at a flowrate (F) in the equation:
V
t - — (assuming no axial dispersion or wall effects)
r
At a flowrate of 1 standard liter minute, the response time for a 30.48-meter tube
section at 25° C and pressure drop of 152 mm Hg is only 30 seconds. These data indicate
that 6.35-mm OD tubing is acceptable for sampling lines.
Teflon® and stainless steel exhibit excellent corrosion and heat resistance in addition to
being chemically inert to stack gases and acid mist. The corrosion resistance of stainless
steel is enhanced by keeping gases above the dew point. These materials are commercially
available in heat traced form. Teflon® is recommended for out-of-stack heat traced lines;
stainless steel is a good material for in-stack lines. Polypropylene and polyethylene lines
exhibit good chemical resistance (except to nitric acid). Plastic lines are a good, economical
choice for sampling lines that carry dry gas and are maintained above the freezing point
without heat tracing. A reliable, effective, and economical sampling line system probably
would incorporate stainless steel. Teflon®, and plastic.
6.5.6 Sampling Pump
A diaphragm or bellows pump upstream of the analyzer is superior to other pump types
for gas handling. The primary advantages offered are:
• No shaft seal is required; these pumps are not subject to seal failure air in-leakage.
• No internal lubrication is required.
• These pumps are relatively inexpensive.
• Adequate suction and discharge pressures are developed at flowrates well above
those needed for gas sampling systems.
A pump positioned upstream of an analyzer may be located in either position A (see
Figures 6-6 and 6-7) or B (see Figures 6-8 and 6-9).
There are operational trade-offs that must be considered. The pump positioned in A of
this portable system offers the highest condensate removal potential (based on the mole
fraction of water vapor being equal to its partial pressure divided by the total pressure
at condenser temperature). Pump position A also minimizes the chances of air leakage
and allows the use of a simple ball-float trap for water removal from a condenser trap.
It also allows the analyzer to operate at pressures and temperatures below those which would
occur using pump position B. Pump position B protects the pump from moisture and
particulates. This extends pump life and may be an overriding advantage for this position.
6-9
-------
BYPASS
SAMPLE
MOISTURE
REMOVAL
SYSTEM
PUMP
ANALYZER
FINE
FILTER
FLOW
INDICATOR
FIGURE 6-6
SCHEMATIC OF PUMP PLACEMENT — POSITION A
FIGURE 6-7
ACTUAL SAMPLING SYSTEM WITH PUMP IN POSITION A
6-10
-------
SAMPLE
BYPASS
ANALYZER
FLOW
INDICATOR
PUMP
FIGURE 6-8
SCHEMATIC OF PUMP PLACEMENT — POSITION B
FIGURE 6-9
ACTUAL SAMPLING SYSTEM WITH POSITION B PUMP LOCATION
The diagrams for positions A and B both show a bypass system which connects pump
suction and discharge to protect the pump from excessive wear when operated at low
flowrates. A pump throttled down for low flow produces a high pressure drop across
the pump which can greatly reduce its expected life span.
Some sampling interface systems may place the pump downstream of the analyzer, pulling
the sample through the system. This could allow the use of an aspirator pump without
moving parts. Pressure drop at the analyzer would be higher, but for some analyzers
with built-in pressure regulators, this may be a preferrable arrangement. Downstream
pumps increase the potential for air leaking in and, in the case of aspirator pumps, require
a source of large quantities of compressed air, steam, or water.
6-1
-------
6.5.7 Moisture Removal
Stack gases may contain significant quantities of water vapor. A limited number of
analyzers are not affected by the presence of water vapor in the sample (e.g., a differ-
ential absorption ultraviolet instrument). These analyzers do, however, require that gases
be kept above the dew point to protect against condensation and corrosion within the
analyzer. Other analytical methods that are affected by water vapor require moisture
removal. Generally, the gas is dried to a low constant level of moisture content for
both stack gases and calibration gases. Refrigerated condenser traps or permeation dryers
are commonly used for moisture removal.
A refrigerated condenser receives the hot stack gases, then rapidly cools the gas to drop out
moisture (Figures 6-10 and 6-11). The refrigerated condenser must provide enough cold
surface area to remove the latent heat of vaporization and to cool the gas stream within a
minimum residence time in the condenser. This greatly reduces the possibility of pollutant
gas absorption in the condensate. The cooling requirements for the condenser are directly
FIGURE 6-10
A REFRIGERATED CHILLER MANUFACTURED BY HANKINSON
6-12
-------
FIGURE 6-11
INTERIOR OF TYPICAL CONDENSER USED FOR MOISTURE REMOVAL
proportional to the gas flowrate. Cooling requirement calculations probably will be made
by the manufacturer for use by the condenser purchaser; however, calculates and pro-
cedures are given in the literature.
The moisture dropped out of the gas stream must be trapped and removed periodically.
This mTy * done'by automatic va.ving or manual drainage. The ~t of water
collected over time at the analyzer flowrate (assume 100 percent removal) should be
calculated in order to decide on whether Or not an automat.c system ,s necessary. The
approximate water trapped may be calculated by
where:
VLC=
F =
Hv =
K. =
milliters of water collected per hour
flowrate (standard liters/ minute)
percent water vapor by volume in stack gas
a constant - 1.333 liters water vapor/ml at standard conditions
6-13
-------
The permeation dryer (Figures 6-?? *nH A m ~«-
-
Less prone to corrosion - no materials
contact with wet gases
• No possibility of sample loss in condensate
• No condensate trap required
• Competitively priced
HIGH PRESSURE
WET FEED
INLET
r
LOW PRESSURE WET PURGE GAS OUTLET
SHELL
HEADER
HIGH PRESSURE
DRY PRODUCT
OUTLET
HEADER
PERMEABLE
TUBE PACK
LOW PRESSURE
DRY PURGE GAS
IllllllllllllllimiiilC EXPANSION VALVE
FIGURE 6-12
SCHEMATIC DIAGRAM OF PERMEATION DRYER
6-14
-------
FIGURE 6-13
CORRUGATED STAINLESS STEEL ENCLOSED PERMEATION TUBE DRYER
6.5.8 Sampling Interface/Monitor Calibration
The entire sampling interface and monitor must be calibrated as a unit. The calibration
gases should enter the continuous gas monitoring system as near as possible to the same
entrance point for the stack gas. This is essential to check the entire system. The
analyzer should be calibrated at the same gas flowrate, pressure, temperature, and operating
procedure used in monitoring the stack gas. Flooding the coarse filter with calibration
gas at the probe inlet or using a check valve that allows calibration gas injection directly
behind the coarse filter are the best methods for accomplishing this calibration. Calibration
in this manner assures that any leaks, blockage, or sorption of gases taking place in
the system will be discovered. The importance of this method cannot be overemphasized.
Automatic gas injection systems are easily constructed with electric solenoid valves.
The calibration gases must be checked with triplicate runs of the reference method pro-
cedure for that gas. All runs of the reference method must agree with the average for
the three runs within 20 percent or they must be repeated. The gas analysis should be
repeated every 6 months. Although many manufacturers certify a longer shelf life, experience
has shown that manufacturer calibration gas certification is subject to error.
6-15
-------
EPA is currently studying the option of using National Bureau of Standards (NBS)
calibration gases or gases traceable to NBS standards, instead of requiring reference method
analyses. NBS gases are relatively accurate and stable but are more expensive than
commercial gases.
6.5.9 Dilution Systems
Several gas dilution systems are commercially available (Meloy, Thermoelectron, Hastings).
These systems dilute the stack gas sample with known volumes of inert carrier gas. This
reduces gas-handling problems by decreasing the temperature and moisture content of the
gas entering the analyzer. A dilution system is useful in adapting ambient air instruments
to source monitoring applications. These dilution systems may add unnecessary complexity
to the sampling system, increasing initial costs and maintenance costs in addition to slowing
system response time. The inherent problems involved in maintaining precise dilution
ratios also may reduce the overall measurement accuracy.
6.5.10 Controlling the Sampling Interface/Monitor System
The best system does not require elaborate control mechanisms. The necessary controls
are easily installed and maintained by plant personnel. The suggested controls include
the following:
• Temperature control at the cold end of the heated sample line. This is to
ensure that the gases are above freezing to protect the lines from fracture or
blocking. Temperature should also be controlled at the refrigerated condenser
to maintain moisture removal efficiency.
• Pressure control is needed at the pump discharge to protect the pump. The
pressure drop across the fine filter should be monitored to protect the analyzer
and to ensure proper system function (most analyzers are sensitive to pressure
changes).
• Gas flowrate control should be installed to make certain the analyzer receives
the correct gas flow. This is not critical, since most analyzers are relatively
insensitive to minor flowrate change.
• Calibration gas valving should automatically inject calibration" gases once every
24 hours. This can be accomplished with a simple electric solenoid valve. The
calibration gases should flow through the sampling system at the same condition
of temperature, pressure, and flow as does the stack gas.
6-16
-------
6.6 Bibliography
Brooks, E. F., Guidelines for Stationary Source Continuous Gas Monitoring System,
EPA Contract No. 68-02-1412, TRW Systems Group, November 1975.
Brooks, E. F., et al., Continuous Measurement of Total Gas Flowrate from Stationary
Sources, EPA-650/2-75-020, February 1975.
Brooks, E. F., et al., Continuous Measurement of Gas Composition from Stationary
Sources, EPA-600/2-75-012, U.S. Environmental Protection Agency, Office of Research
and Development, Washington, D.C., July 1975.
Brooks, E. F., and Williams, R. L., Flow and Gas Sampling Manual, EPA-600/2-76-203,
July 1976.
Chapman, R. L., "Continuous Stack Monitoring," Environmental Science & Technology,
V. 8, No. 6:520-525, June 1974.
Felder, R. M., Miller, G. W., and Ferrell, J. K., "Continuous Stack Monitoring Using
Polymer Interfaces," Chemical Engineering Progress, June 1978, pp. 86-88.
Gregory, M. W., et al., "Determination of the Magnitude of SO2, NO, CO2 Stratification
in the Ducting of Fossil Fuel Power Plants," Paper 76-35.6 presented at the 1976 APCA
Meeting, Portland, Oregon.
Hedley, W. H., Dilution Device for Coupling Monitoring to Source Emissions, EPA-650/
2-74-055, United States Environmental Protection Agency, Washington, D.C., 1974.
Homolya, J. B., "Coupling Continuous Gas Monitors to Emissions Sources," Chem. Tech.,
July 1974, pp. 426-433.
Homolya, J. B., "A Review of Available Techniques for Coupling Continuous Gaseous
Pollutant Monitors to Emission Sources," Analytical Methods Applied to Air Pollution
Measurements, Ann Arbor Science Publishers, Inc., Ann Arbor, Michigan, 1974.
Marcot, R. V., "A New Approach to Sample Preparation in Multi-Parameter Turn-Key
Systems for Process Control and Stacjc Monitoring," Air Quality Instrumentation, V. 2:293-
304, Instrument Society of America, Pittsburgh, 1974.
McNulty, K. J., et al., "Investigation of Extractive Sampling Interface Parameters," EPA-750/
2-74-089, Environmental Protection Technology Series, Environmental Protection Agency,
Research Triangle Park, North Carolina, October 1974.
6-17
-------
PAT Report, "Sampling Hot Gases for Analysis," Environmental Science and Technology
V. 12, No. 2:138-139, February 1978.
Treece, L. C., Felder, R. M., and Ferrell, J. K., "Polymeric Interfaces for Continuous
SO2 Monitoring in Process and Power Plant Stacks," Environmental Science and Technology
V. 10, No. 5:457^61, May 1976.
Wolf, P. C., "Continuous Stack Gas Monitoring Part Two: Gas Handling Components
and Accessories," Pollution Engineering, July 1975, pp. 26-29.
Wolf, P. C., "Continuous Stack Gas Monitoring Part Three: Systems Design," Pollution
Engineering, August 1975, pp. 36-37.
Wyss, A. W., and Stroud, B. D., "Design and Operation of a Sampling Interface for
Continuous Source Monitors," Paper 77-27.2, presented at the 70th Annual Meeting of the
Air Pollution Control Association, Toronto, Canada, June 20-24, 1977.
6-18
-------
CHAPTER 7
IN-SITU MONITORING SYSTEMS
7.1 Introduction
The problems and expense associated with extractive monitoring systems have led to the
development of instrumentation that can directly measure source-level gas concentrations
in the stack. The so-called in-situ systems do not modify the flue gas composition and
are designed to detect gas concentrations in the presence of paniculate matter. Since
particulate matter causes a reduction in light transmission, in-situ monitors utilize advanced
electro-optical techniques to eliminate this effect when detecting gases. These techniques are:
Differential
Absorption
Gas Filter
Correlation
Second Derivative
Spectroscopy
Also, as discussed earlier, an electrocatalytic analyzer has been designed to monitor oxygen
concentrations in-situ.
7.2 Terminology
There are a number of terms used to categorize the different types of in-situ monitors,
as shown in Figure 7-1.
Cross-stack in-situ monitors measure a pollutant level across the complete diameter or a
major portion of the diameter of a stack or duct. Stratification effects are lessened by
the use of cross-stack instruments, since an average reading is taken over a relatively
long sample path. There are two types of cross-stack monitors: single pass and double
pass. Single-pass and double-pass transmissometers have been discussed earlier, and the
distinction holds for in-situ gas monitoring systems.
• Single-pass systems locate the light transmitter and the detector on opposite ends
of the optical sample path. Since the light beam travels through the flue gas
only once, these systems are termed single pass.
• Double-pass systems locate the light transmitter and the detector on one end of
the optical sample path. To do this, the light beam must fold back on itself
by the use of a retroreflector. The light beam will traverse' the sample path
twice in going from the instrument housing to the retroreflector and back to the
instrument. Double-pass systems are easier to service than single-pass systems,
since all of the active components are in one location.
7-1
-------
CROSS-STACK
IN-STACK
SINGLE-PASS
DOUBLE-PASS
DOUBLE-PASS
POINT, OR SHORT PATH SYSTEMS
FIGURE 7-1
TYPES OF IN-SITU MONITORS
In-stack, in-situ systems monitor emission levels by using a probe that measures over a
limited sample pathlength. All of the commercial optical in-stack monitors are double-pass
systems (the in-stack electrocatalytic oxygen monitor discussed earlier is not an optical
system). The pathlength may vary from 5 cm to a meter. A retroreflector. usually
made of quartz, is located at the end of the probe. The in-stack systems are also
termed short-path monitors. The siting of such systems should follow the same guidelines
as those given for extractive systems. The location should be chosen carefully so that
consistent levels of emissions can be accurately monitored.
There are currently only three vendors of in-situ.optical gaseous emission monitors. Environ-
mental Data Corporation (EDC) uses the technique of differential absorption to monitor
CO2, SO2, and NO and the gas filter correlation technique to monitor CO. Contraves Goerz
markets an instrument that measures SO2, NO, CO2, and CO levels by the gas-filter
correlation method. Lear Siegler, Inc. (LSI), utilizes second derivative spectroscopy to
measure SO2 and NO levels. The following discussion of each of these methods is intended
to provide the reader with a background in these new technologies so that informed
evaluations may be made of the commercially marketed systems.
7-2
-------
7.3 In-Situ Cross-Stack Analyzers
7.3.1 Differential Absorption Spectroscopy
The technique of differential absorption Spectroscopy used in the EDC cross-stack gas
monitor is similar to that used in the NDUV extractive analyzers discussed in Section 5.2.
A diffraction grating is used in this analyzer to obtain a narrow band of radiation over
which the pollutant molecule will absorb energy. A grating disperses light from an
ultraviolet lamp and light of the appropriate wavelength is detected: one wavelength for
monitoring the pollutant level, another to serve as a reference wavelength (Figure 7-2).
LIGHT
SOURCE
BLOWER
MONOCHROMETER
SYSTEM
DIFFRACTION GRATING
PHOTODETECTOR
CHOPPER
FIGURE 7-2
OPERATION OF IN-SITU DIFFERENTIAL ABSORPTION ANALYZER
The ratio of the intensities, I/10 produces a signal that is related to the pollutant concentration:
I/Io = e-*c/
where:
I = intensity of light at the measuring wavelength
I0 = intensity of light at the reference wavelength
7-3
-------
a = absorption coefficient at wavelength A.
c = concentration
/ = measuring pathlength
In the differential absorption technique obtaining a ratio of intensities is important in the
case of in-stack monitors. Particulates in the flue gas will attenuate the amount of light
energy passing through the optical path. This is the principle of measurement in the
opacity monitors. If the light attenuation is the same for the light energy at the measuring
wavelength and at the reference wavelength, each intensity would be reduced by a constant
factor.
In - KI
wp
I0p - KIowp
Ip _ KlWp _ Iwp
KI
°wp
Io
wp
where:
K = fraction of light attenuated by particulates in the gas stream
Ip = light intensity at measuring wavelength with paniculate attenuation
Iwp — light intensity at measuring wavelength without particulate attenuation
I0p = light intensity at reference wavelength with particulate attenuation
,,,r.
wp
- h'ght intensity at reference wavelength without particulate attenuation
This satisfies the requirement demanded of all in-situ monitors that particulates not interfere
in the analytical method. Interference caused by broad-band absorption of water vapor or
other molecular species should similarly cancel out if the measuring and reference wavelengths
do not differ too greatly.
Optical depth, used in in-situ and remote monitoring, is defined as the concentration
of the gas times the optical measuring pathlength. The Beer-Lambert relation for light
absorption gives a dependence on the pathlength /. A cross-stack monitor set up to measure
pollutants on a stack of a given diameter (di) would give different readings if moved to
another stack of diameter (d2) and a correction was not made for the new diameter.
To create some type of unit related to the pathlength, the optical depth is defined as
the equivalent concentration of the pollutant in a l-meter path expressed in terms of
7-4
-------
ppm-meters. A vendor might specify an optical depth range from 0 to 20,000 ppm-meters
for a cross-stack instrument. Such an instrument located on a source with a 10-meter
stack diameter would have a concentration range from 0 to 20,000/10 = 2,000 ppm. An
optical depth is effectively a compression of the real pathlength into a 1-meter length to
give an equivalent concentration in ppm-meters. The specification of optical depth values
is also important when selecting calibration cells for a cross-stack monitoring system.
The absorption wavelength used for SO2 monitoring in the EDC analyzer is 309 nm, with
a reference wavelength of 310 nm. Nitrogen oxide absorption is detected at a wavelength
of 226.5 nm with a reference wavelength of 228 nm. The EDC monitor also detects
CO2 using the differential absorption method, although in this case band-pass filters are
employed instead of a diffraction grating. Narrow band-pass filters are chosen to distinguish
light at 2 y,m for the CO2 absorption and 2.1 pm for the reference channel. The method
for CO2 is similar to that used in the DuPont extractive analyzer for SO2 in the ultraviolet,
except that in the EDC, infrared radiation is used in the analysis. CO is detected by
the gas-filter correlation technique in the EDC system. This method will be discussed later.
The optical systems in cross-stack analyzers are designed to eliminate the effects of paniculate
matter. Figure 7-3 shows a typical stack-mounted system. The analyzer box contains
the major electronics, monochrometer subassembly, and the calibration and zero assemblies
as indicated in Figure 7-4. Most of the components are fixed on sliding mounts that
can be easily moved in and out of the box for servicing. Figure 7-5 shows the lamp
assembly of the EDC system. Figure 7-6 shows the protective housing for the lamp
assembly. An EDC analyzer can be purchased to monitor opacity and up to four gases.
Four 2-inch-square light beams pass through the windows of the light source assembly of a
typical EDC analyzer. A single ultraviolet beam will pass from the lamp assembly to the
analyzer box to detect SO2 and NO. The IR beam will split in two before passing
through the flue to the analyzer box. Separate channels are used to monitor CO2 and CO.
The EDC analyzer is a single-pass system for the measurement of pollutant gases. The
opacity channel, however, is a double-pass system, sending a beam of visible light from
the analyzer assembly through the fourth 2-inch-square hole of the light source assembly
and back again to the analyzer box.
Further information on this system may be obtained from:
Environmental Data Corporation
608 Fig Avenue
Monrovia, California 91019
7-5
-------
FIGURE 7-3
MOUNTED EDO CROSS-STACK IN-SITU ANALYZER
FIGURE 7-4
INTERNAL VIEW OF ANALYTICAL SECTION OF THE EDC ANALYZER
7-6
-------
FIGURE 7-5
INTERNAL VIEW OF THE EDC LIGHT SOURCE ASSEMBLY
FIGURE 7-6
DIFFERENTIAL ABSORPTION SPECTROMETER INSTALLED AT
RESEARCH TRIANGLE PARK SOURCE SIMULATOR FACILITY
7-7
-------
7.3.2 Gas-Filter Correlation Spectroscopy
The gas-filter correlation (GFC) method is used in the EDC analyzer to monitor CO and
is exclusively used in an analyzer produced by Contraves Goerz Corporation (originally
developed and marketed to monitor CO2, CO, SO2, and NO). The method shows
potential in both in-situ and remote emissions monitoring.
There are a number of optical configurations that can be designed into a GFC system.
The essential feature of such a system, however, is the gas-filter cell (Figure 7-7).
LIGHT
SOURCE
BEAM
ALTERNATOR
NEUTRAL FILTER
DETECTOR
GAS-FILTER
CORRELATION
CELL
FIGURE 7-7
OPERATION OF A CROSS-STACK GAS-FILTER CORRELATION SPECTROMETER
First, consider a down or zero condition where there is no pollutant gas1 in the stack.
Light, generally in the infrared, is emitted from a lamp and passes through the empty
stack to an analyzer where it is split into two separate beams. One beam passes through
a neutral filter and the other through the gas-filter correlation cell. This cell contains
enough of the gas being analyzed so that most of the energy contained in the individual
absorption lines of the gas will be removed. Light of wavelengths not absorbed by the
specified gas is not removed and passes on to the detector. This results in a reduction
in light energy after the beam traverses the correlation cell.
In most GFC instruments, a neutral density filter is chosen to reduce the amount of light
energy in the other beam by an equal amount. The neutral density filter reduces the
energy from all of the wavelengths in the beam before it reaches the detector. The gas-filter
7-8
-------
cell only cuts out energy at the absorption wavelengths. With the proper choice of a
neutral density filter and gas concentration in the correlation cell, the amount of energy
reaching the detector from each beam is the same, and the system is said to be balanced
[Figure 7-8(a)].
NEUTRAL BEAM
(a)
GAS-FILTER CELL BEAM
-100—-
50
(b)
g
t
GC
O
V)
CO
<
._. 0
-100
50
,__ 0
•-•100
NOS02
IN STACK
- SO2 IN STACK
PARTICULARS
IN STACK
WAVELENGTH
WAVELENGTH
FIGURE 7-8
ABSORPTION PRINCIPLES OF A GAS-FILTER CORRELATION ANALYZER
Next, consider the condition where pollutant gas is in the stack. The beam again traverses
the stack, but in this case pollutant molecules are present and absorb light energy at
wavelengths corresponding to their absorption spectra. Since the gas-filter correlation cell
was chosen to absorb energy at these same wavelengths, the absorption is already complete
in the correlation cell beam, and the detector will see the same signal as it did when the
stack was clean. The beam passing through the neutral density filter, however, will have
less energy than previously, since light was selectively absorbed by the pollutant gas in
the stack. The difference in energy between the two beams can be related to the pollutant
concentration and is monitored at the detector [see Figure 7-8(b)].
Particulates will reduce the intensity equally in each of the beams. If the two signals
are ratioed, the effect of particulate matter will cancel out. Note that paniculate inter-
ference is equal in both graphs of part (c) of Figure 7-8. Molecules with spectral patterns
near that of the pollutant molecule being measured will not affect the measurement if
they do not "correlate" or overlap with the pollutant spectral pattern. If there is some
overlap, some interference will result.
7-9
-------
The GFC method has been found to be a very sensitive and specific method in the infrared.
The ability to monitor a large number of absorption lines provides greater sensitivity, in
some cases, than can be obtained with the differential absorption technique using only
filters. The GFC method is an NDIR method; the light is not dispersed.
The instrument manufactured and marketed by Contraves Goerz uses two detectors instead
of one, as shown in Figure 7-9.
STACK
IR
SOURCE
CHOPPER BEAM GFC
MIRROR SPLITTER CELL
NEUTRAL
FILTER
DETECTOR B
DETECTOR A
SIGNAL
ALTERNATOR
COMPARATOR
SIGNAL
OUT
FIGURE 7-9
THE CONTRAVES GOERZ CROSS-STACK GFC MONITOR
A reference IR source is placed in the analyzer portion of this single-pass system to
detect concentration levels in a slightly different manner than described previously. When
the light from the stack infrared source passes through the flue gas and is divided between
the correlation cell and neutral filter, the higher signal coming from the detector after
the correlation cell is electrically attenuated. The attenuated signal is adjusted automatically
to the same value as the signal given by the other detector after the neutral density
filter. A rotating mirror then switches to the light from the IR reference source in the
analyzer and blocks out the light coming from across the stack. This time the signal
received from each detector will be different. The signals are electronically subtracted
to give a signal related to the gas concentration. This two-step procedure is employed
to eliminate any effects related to differences in the sensitivity of the two detectors and
also to provide a means for a zero and calibration check. The Contraves Goerz system uses
only one correlation cell containing all of the four gases, CO, CO2, SO2, and NO. Full
advantage is taken of the spectral characteristics of these molecules to prevent problems of
interference in the measurement.
7-10
-------
More information may be obtained on this system from:
Contraves Goerz Corporation
610 Epsilon Drive
Pittsburgh, PA 15238
7.3.3 Advantages and Limitations
The currently marketed cross-stack gas analyzers in principle present many advantages
over extractive monitoring systems. A cross-stack system may allow greater flexibility
in site selection, since an average sample reading is taken over a relatively long path.
It should be noted, however, that gas stratification in a duct or stack is a two-dimensional
phenomenon, not one-dimensional. A cross-stack monitor can linearly average concen-
trations over its measuring path, but does not properly weigh the contributions of stratified
areas to the measurement. For severe cases of stratification, the problem of obtaining
representative concentration .values may be comparable to the problems encountered by
point monitors. Quartz or glass cells, used in cross-stack optical systems for calibration,
reduce the time and expense that result with span gas cylinders and the associated plumbing
of extractive systems. The calibration cells need only be certified by the manufacturer
and are not required to be checked periodically, as are span gas cylinders.
One of the principal marketing features of cross-stack analyzers is that a single instrument
can monitor a number of gases and even opacity. The cost of such a monitor can be
comparable to the purchase price of three or four separate instruments combined in an
extractive system. The operating costs of in-situ monitors can be less than those of
extractive systems, since zero and span gases are not required for the 24-hour checks.
There'are also fewer separate components in an in-situ system, so problems with chillers,
heat-traced lines, valves, and pumps are avoided.
The Code of Regulations gives an alternative method that a single-pass, cross-stack monitor
can use to perform a system zero check. Three or more calibration cells are inserted
into the system operating in the measuring cross-stack mode. The upscale readings given by
these known cells then can be extrapolated to a zero value. A graph showing this
extrapolation is reported by the source operator. The zero drift values for 2 and 24 hours
must be within 2 percent of span before the instrument can be accepted by EPA. A
problem has arisen in some cases that it is difficult to distinguish zeroing and upscale
calibration checks by this method.
There are, however, a number of disadvantages associated with the cross-stack monitors.
An in-situ cross-stack monitor can monitor only one flue or stack at a time. Costs
might be prohibitive if a number of stacks must be monitored. In such a case, multiple
probes and sampling lines leading into a single extractive system might be the better
choice. Problems with optical misalignment, vibration affecting the optical systems, and
7-11
-------
failure of electronic components also can occur. It is common among vendors of these
instruments to offer service packages where the systems are periodically checked by a
company serviceman. A service package generally will ensure that a system will continue
to function, but the cost involved may bring the operating expenses to a level comparable
to that of an extractive system.
7.4 In-Situ, In-stack Analyzers: Second-Derivative Spectroscopy
At the present time, only one instrument is manufactured that monitors SO2 and NO
in-stack. This is the Lear Siegler (LSI) second-derivative, stack-gas monitor. Although
the second-derivative technique is somewhat more complicated than those discussed earlier,
an understanding of the method is necessary if a source operator or agency observer
has to make an evaluation of different monitoring systems.
The LSI in-situ monitor is shown in detail in Figure 7-10, and in a typical source application,
mounted to a stack wall, in Figure 7-11.
This monitor analyzes the gas in-situ; the gas is not extracted, but is monitored as it
exists in the flue gas stream. The tip of the probe contains the measuring chamber,
which senses across a distance of 10 cm. The instrument therefore does not measure
cross-stack. It is an in-stack point monitor or short-path monitor. Care should be
taken when siting such a system, since a representative location is required to be monitored
by the EPA. The guidelines given for siting of the probe of an extractive system could
be followed in choosing the location of an in-stack monitor, although EPA has not
published any specific siting criteria for this technique outside of the general criteria for
representative measuring.
The probe of this system consists of a ceramic thimble surrounding the measuring chamber.
The thimble and a metal V bar in front of the thimble prevent particulates from entering
the chamber. The filtering action of the thimble prevents particulate matter from fouling
the optical surface of the retro reflector shown in Figure 7-12. Gas diffuses into the
measuring cavity and the pollutant can be monitored. Ultraviolet light is sent from the
analyzer section, down the length of the probe, through the measuring cavity, to the
retroreflector. A quartz corner cube reflector is used in this case, and the light is bounced
back to the analyzer section. The pollutant gas only occupies the small measuring cavity
and not the entire length of the probe assembly.
The technique of second-derivative spectroscopy (SDS) utilizes the spectral absorption
features of a molecule in a manner somewhat different from the methods discussed for
the cross-stack monitors. A diffraction grating selects the specific absorption wavelengths,
but instead of just sitting on a specific wavelength as is done in differential absorption
techniques, a scanner -or moving slit scans back and forth across the central wavelength.
7-12
-------
FIGURE 7-10
THE LEAR SIEGLER IN-STACK IN-SITU S02-NO ANALYZER
FIGURE 7-11
SECOND-DERIVATIVE SPECTROMETER INSTALLED AND OPERATING
AT A STEAM GENERATING FACILITY
7-13
-------
UV LIGHT MODULATED
BY GAS ABSORPTION
MIRRORS
NO CHANNEL
SCANNER
ENTRANCE
SLIT
STACK
RETROREFLECTOR
RETURNED
LIGHT
SO2
CHANNEL
ULTRAVIOLET
LIGHT SOURCE
SEQUENTIAL
SHUTTERS
DUAL EXIT SLITS
GRATING
DETECTOR
T
POROUS
FILTER
STACK
GAS
DIFFUSION
ABSORPTION
CHAMBER
FIGURE 7-12
OPERATION OF THE SECOND DERIVATIVE IN-STACK MONITOR
In this instrument, light at 218.5 nm, corresponding to the maximum of an SO2 absorption
peak in the ultraviolet, is utilized. The scanner modulates the light at wavelengths from
217.8 to 219.2 nm, across the width of the absorption peak (Figure 7-13).
The results of this scanning are seen at the detector of the instrument. ^Before looking
at the signal that such a scan of the absorption peak would produce on a detector,
consider the detector signal produced by a scan of a broad band absorption (Figure 7-13).
Here, there is no strong absorption peak, but a gradual decrease in transmission (increased
absorption) as the light varies from the lowest scanned wavelength to the highest scanned
wavelength.
The moving mask scans over the wavelengths of light separated by the diffraction grating
and then goes back over the same wavelengths. One cycle, back and forth, will take
0.09 second (a period of 0.09 = 11 cycles per second). The resultant signal seen at
the detector will be in the form of a sine wave or an alternating current, with a period
t = 0.09 second and frequency of 11 cps.
7-14
-------
0
%
ABSORPTION
100
217.8 nm
218.5
FREQUENCY
219.2 nm
FIGURE 7-13
ULTRAVIOLET LIGHT WAVELENGTHS
SCANNED BY SPECTROMETER MOVING MASK
In the following case there is no broad band absorption, but instead, a sharp absorption
peak caused by the presence of an SO2 molecule (Figure 7-14). Following the same
argument, where the slit moves back and forth in a time period of 0.045 second there
is an extra hump in the detector signal (Figure 7-15). Although the mask scans the
wavelengths at a frequency of 11 cycles per second, maxima will appear at the detector
signal at double the frequency, or 22 cycles per second. Since the amplitude of the
peaks seen at the detector are related to the amount of light absorption, the amplitude
is related to the amount of pollutant gas in the optical path.
Electronically, the concentration of a pollutant is determined by tuning in on the frequency
which is double that of the frequency of movement of the scanner, much like tuning
a radio. A radio station produces a signal at a given frequency and a dial is adjusted
to receive that station. A station with a strong transmitter will produce a louder signal
than a weaker station. In the second-derivative method, the instrument is tuned to a
frequency of 2f, where f is the scanning frequency of the mask. A strong signal from
the detector indicates strong absorption and a high concentration of SO2. A weak signal
at this frequency indicates a lower concentration of SO2-
This discussion has so far dealt with the mechanical aspects of the detection method.
The question arises, however, what does this have to do with second derivatives? Taking
a derivative of a function is equivalent to determining the slope. For example, for a
broad-band absorption curve similar to that of Figure 7-16(a), the first derivative gives
a constant negative value and the second derivative gives a value of zero, since the
slope of Figure 7-16(b) is zero.
7-15
-------
O co
C/> Z
mo
ofc
IS
• a
3<
So
<= S
CD
2
UJ
QL
W
CD
LU
OC
CL
QC
O
O)
m
<
Q
<
oo
Q
<
O
QC
m
O
7-16
-------
LA
LLJ
CC
D
(D
ill
0_
O
DC
O
CO
QQ
(D
O
CO
7-17
-------
dA
0
(b)
H-
-+-
A,
dll
0
(c}
FIGURE 7-16
FIRST AND SECOND DERIVATIVES OF LINEAR ABSORPTION
For an absorption peak, the curvature and, therefore, the slope change, depending upon
the wavelength. The first derivative, evaluated at a given wavelength, will reflect the
curvature of the absorption peak [Figure 7-17(a)]. The second derivative indicates the
curvature of the first-derivative curve [Figure 7-17(b)]. Since the slope of b changes
often, the second-derivative curve is much more complicated, as shown in Figure 7-18.
(a)
dA
FIGURE 7-17
FIRST DERIVATIVE OF AN ABSORPTION CURVE
7-18
-------
t
FIGURE 7-18
SECOND DERIVATIVE OF AN ABSORPTION CURVE
In the mechanical method used in second-derivative spectroscopy, the actual detector output
appears much like that shown in Figure 7-15. The amplitude of the detector signal at
the frequency of 2f is proportional to d2I/dX2 evaluated at XQ. The second-derivative
source monitoring instruments constructed by LSI do not produce curves like that of
Figure 7-18. They only produce the value of d2l/dX2 evaluated at Xo. The signal at
the detector is given by
i2,
S 4
where 6 equals the distance from X_x to X+x (in the example this would be 1.8 nm). By
an expansion of Beer-Lambert's law
dX2
dX2
where:
a = absorption coefficient
c = concentration
/ — optical pathlength of gas of interest
The resultant expression for the signal is
S = - c
4 dX2
I
7-19
-------
or
-= Kc/
where K is constant.
This is the actual instrument output that is proportional to both gas concentration and
optical pathlength.
By dividing S by I, problems caused by variations in the source intensity, optical misalign-
ment, and broad-band absorption from other gases or particulates are avoided. This
results because a change in I by a constant factor will induce an identical change in S.
Determining the ratio of the two cancels out the effect.
The second-derivative in-stack monitor is of course limited to monitoring one stack at a
time. Vibration also can be a problem, since extreme cases can affect the optical system.
One of the most common problems in this and similar electro-optical systems is the failure
of electronic components. The complicated circuitry of such systems in some cases may
lead to a higher probability of component failure. A significant feature of the LSI system
is that zero and span gases can be used to flood the sample chamber to a pressure greater
than the stack static pressure. This provides an alternate method to the use of calibration
cells if desired. The calibration cells may be used for daily span checks and would save
the expense of span gas and the associated plumbing systems. The LSI second-derivative
source monitor also may be modified to measure ammonia concentrations. More information
may be obtained on the analyzer from:
Lear Siegler, Inc.
Environmental Technology Division
74 Inverness Drive East
Englewood, CO 80110
7.5 Bibliography
Burch, D. E., and Gryvnak, D. A., "Cross-Stack Measurement of Pollutant Concentration
Using Gas-Cell Correlation Spectroscopy," Anal. Methods AppL Air Pollut. Meas., R. K.
Steven and W. F. Herget, eds., Ann Arbor Science, Ann Arbor, Michigan, 1974, p. 173.
Byerly, R., "New Developments in the Measurement of Gaseous Pollutants in Air," IEEE
Transaction on Nuclear Science, NS-22, April 1975, pp. 856-869.
Cooke, M. J., Cutler, A. J. B., and Raask, E., "Oxygen Measurements in Hue Gases
with a Solid Electrolyte Probe," March 1972, pp. 153-156.
7-20
-------
Fyock, D. H., "Test of the Environmental Research and Technology Stack Gas Analyzer
at the Conemaugh Generating Station/ Paper No. 75-60.6, 68th Annual Meeting of the
Air Pollution Control Association, Boston, Massachusetts, June 15-20, 1975, Boston,
Massachusetts.
Hager, R. N., Jr., "Derivative Spectroscopy with Emphasis on Trace Gas Analysis,**
Analytical Chemistry, V. 45, No, 13:1131A-U38A, November 1973.
Hager, R. N., Jr., and Anderson, R. C, "Theory of the Derivative Spectrometer," J. Opt.
Soc. Am., V. 60:1444, 1970,
Herget, W. F., Jahnke, J. A., Burch, E. E., and Gryvnak, D. A., "Infrared Gas-filter
Correlation Instrument for In-Situ Measurement of Gaseous Pollutant Concentrations,"
Applied Optics, V. 15:1222-1228, May 1976.
Huillet, D. F., "The Monitoring of SO2, NO, CO and Opacity with an In-stack Ispersive
Spectrometer," TAPPl, V. 58, No. 10:94-97, 1975.
Klasens, H. A., "Analyze Stack Gases via Sampling or Optically, in Place," Chemical
Engineering, November 21, 1977, pp. 201-205.
Lord, H., "In-Stack Monitoring," Environmental Science and Technology, V. 12, No. 3:
264-269, March 1978.
Lord, H. C., Egan, D. W., Paules, P. E., and Holstrom, G. B., "Instantaneous, Con-
tinuous, Directly On-Stream Boiler Flue Gas Analysis," presented at Instrument Society
of America, 24th Annual Power Industry Symposium, New York City, May 17, 1971.
Polhemus, C., and Hudson, A., "A Performance Analysis of Lear Siegler's In-Situ SO2/NO
Monitor," Paper 76-35.5 presented at the 69th meeting of the Air Pollution Control
Association, Portland, Oregon, June 27-July I, 1976.
Polhemus, C., "The Design and Performance of a Spectrometer for In-Situ Measurement
of SO2 and NO," ISA Analysis Instrumentation Proceedings, Volume H, 1976.
Williams, P. T., and Palm, C. S., "Evaluation of Second Derivative Spectroscopy for
Monitoring Toxic Air Pollutants," NT1S Report No. SAM-TR-74-19, September 1974.
7-21
-------
-------
CHAPTER 8
MEASURING, RECORDING, AND REPORTING REQUIREMENTS
8.1 Introduction
The development of a continuous emissions monitoring system extends beyond the choice
of a set of opacity and gas analyzers. The analyzers, themselves, must measure emissions
within specified time periods. The measurements, however, then must be recorded in
some manner. After the data are recorded, they must be converted into units of the emissions
standard, such as IDS/10 Btu.
Calculated emission values that are in excess of the standard must then be reported on
a quarterly basis to the EPA Administrator. In addition, the EPA regulations of 40 CFR
60.7 require the reporting of the following:
• Time and magnitude of excess emissions
• Nature and/or cause of excess emissions
• Corrective and/or preventative action taken to prevent their recurrence
• Zero/span calibration values
• Normal measurement data
• Log of inoperative periods
• Repair and maintenance logs
• Performance, test, calibration data
A complete emissions monitoring system, therefore, requires some means of recording
the analyzer data. Strip-chart recorders have been used most often, but data loggers
and computer systems are beginning to become popular. Data processors have been
developed specifically to reduce the time necessary to evaluate and report excess emissions.
A summary of the process of measuring-recording-reporting is given in Figure 8-1.
A data reporting system may encompass anything from the manual reduction of raw strip chart
data and compilation of associated data to the near fully automatic preparation of complete
excess emission reports, including most of the mentioned data requirements. The choice
of the data reduction and reporting system may be the most important factor in the
overall emission monitoring system, since it greatly affects the amount of manual effort
involved in meeting the NSPS requirements.
8-1
-------
MEASURING OF OPACITY
AND/OR CONCENTRATION
BY THE ANALYZERS
RECORDING DATA
GIVEN BY ANALYZERS
ANALYZERS
PRODUCING
INSTANTANEOUS,
SEQUENTIAL OR
INTEGRATED SIGNALS
ANALOG CHART RECORDER
DATA LOGGERS
REPORTING EXCESS
EMISSIONS BY EVALUATION
OF RECORDED DATA
MANUAL REVIEW
DATA PROCESSOR
LJNK WITH IN-HOUSE
COMPUTER
DEDICATED
EMISSIONS
MONITORING DATA
PROCESSOR
FIGURE 8-1
POSSIBLE METHODS OF MEASURING-RECORDING-REPORTING
This chapter will review some of the techniques and problems involved in completing an
emissions monitoring system. A discussion is given in Appendix C on the F-factor method
used by the EPA in converting concentration data into the units (lb/106 Btu) required
by the New Source Performance Standards.
8.2 Measuring Requirements
The measuring requirements for continuous emissions monitors are important, since they
can influence the choice of the recording system. The requirements for systems applied
to new sources (NSPS) are given in Table 8-1.
TABLE 8-1
MEASURING REQUIREMENTS
Opacity — Completion of one cycle of operation (sampling
Monitors and analysis) every 10 seconds
SC>2, NOX, - Completion of one cycle of operation (sampling
CO2, O2 and analysis) every 15 minutes
8-2
-------
The data generated by the monitoring instrument give much more information than is
actually required. The measuring periods given in Table 8-2, for gases, allow for the
use of sequential analyzers or systems designed to sample from more than one stack or
duct. The DuPont UV 463 analyzer is an example of a sequential system, since it operates
on 5-minute cycles to convert NO to the measured NO2. Monitoring systems that come
under the State plans (existing sources) may have measuring requirements different from
those given above.
The actual data that can be used to satisfy these measuring requirements may be of
three types:
• Instantaneous values taken at the end of each time period
• Values obtained by integrating data over each time period
• Values obtained by averaging a number of data points over each time period
The method used often will be determined by the type of gas and opacity analyzers
purchased and by the recording method. The measuring requirements are tied in with the
recording requirements. A consideration of both often will dictate the choice of the complete
monitoring system.
8.3 Recording Requirements and Systems
8.3.1 Requirements
All of the data that an emissions monitor may produce do not need to be recorded.
The NSPS requirements for recorded emissions data are given in Table 8-2.
TABLE 8-2
RECORDING REQUIREMENTS
Opacity - An average of a minimum of 24 equally spaced
data points taken over a 6-minute period is to
be recorded every 6 minutes.
SO2, NOX, — An average of a minimum of 4 equally spaced
CO2, O2 data points taken over an hour is to be recorded
every hour.
Since a monitor may produce a continuous trace on a strip chart for a 6-minute or 1-hour
period, a larger amount of data may be obtained than is actually used. The regulation,
however, specifies only the minimum number of points that need to be averaged and
8-3
-------
recorded. It is often easier to design systems that integrate the continuous data over the
averaging periods. These would be acceptable under the regulations.
The recording requirements were established to coincide roughly with the type of data
obtained from the manual EPA reference methods. Since it was hoped to correlate in-stack
opacity data with the visible emissions data obtained by an observer using EPA Method 9,
the same averaging and recording requirements were given. It should be noted that by
dividing 24 into 6 minutes (360 seconds), the recording requirements give a 15-second
measuring time. Opacity monitors are, however, required to complete one cycle of measure-
ment every 10 seconds as discussed in Section 9.2. This inconsistency is not particularly
important, since an average of 36 data points would serve just as well to satisfy the
regulation. Integrating systems for the analyzer generally are available as an option for
some opacity monitors.
EPA Method 6 for SO2 specifies a 20-minute sampling time (for fossil-fuel-fired steam
generators). EPA Method 7 for NOx specifies 4 grab samples to be taken at 15-minute
intervals. The continuous monitoring regulation for SO2 and NOx analyzers of an average
of 4 data points taken over each 1-hour period corresponds roughly with these reference
methods.
The recording requirements for monitoring systems on existing sources covered by a State
plan may be somewhat different than those given in Part 60 of the Code of Federal
Regulations. The State averaging periods are chosen to correspond to the averaging
period specified by the State compliance test method. Further information should be
obtained from the State if the compliance test methods differ from the Federal methods.
8.3.2 Recording Systems - Continuous Analog Recording
There are a variety of methods used to record data from analytical devices. The strip-chart
recorder is encountered most frequently in continuous source monitoring applications.
However, the availability of low-cost digital recording devices provides alternatives for the
recording and processing of emissions data.
A continuous analog record is obtained by using some type of chart recorder. The voltage
or current signal from the source analyzer is fed into the recorder and a driving mechanism
produces a trace of the signal strength as a function of time. The types of analog
recorders most often encountered in engineering applications are either circular-chart or
strip-chart recorders.
The circular-chart recorder, although used extensively in process control applications has
some disadvantages in recording emissions data. First, the chart length for a single
chart is limited. For instance, the length of a trace at 50 percent of full scale on a 12-inch
diameter chart would be only 21 inches. A 20 percent opacity trace would give an even
8-4
-------
smaller effective chart length and would require frequent changing of the chart paper.
Second, the curved lines of the circular-chart paper make it difficult to compare and interpret
data. Time resolution becomes poorer at smaller values of the measured parameter. On
the other hand, circular-chart recorders present all of the data at a glance for a given time
period. They have been developed to be used in many types of field situations and may be
obtained at relatively low cost.
Strip-chart recorders provide greater versatility in monitoring applications. There are over
100 companies marketing some form of strip-chart recorder in the United States today.
This method dominates the recording field for several reasons:
• It is the only feasible way of making long-term, easy-to-read records of high-speed
phenomena.
• It is highly efficient in its ability to pack information into available space.
• Proper arrangements make it easy to analyze multiple-trace data.
• On a per-square-foot basis, the paper is almost always much less costly than that
for circular charts.
There are a number of factors that should be considered when evaluating a recorder
for a given systems application. An evaluation should consider;
• Type of unit — Portable, rack-mountable, or table model
• Type of signal input (volts, millivolts, amps) accepted and range
• Type of pen (capillary, felt tip, heated stylus, electrically charged stylus)
• Type of paper take-up (take-up roll, folded paper)
• Appropriate chart speed
• Accuracy as percent of full scale
• Response time consistent with EPA emissions monitoring requirements
• Chart supply provisions - At the required chart speed, number of days' supply of
chart paper that can be stored in the unit
• Maintainability — Features that may enhance the serviceability and reliability of
the instrument
8-5
-------
Since the recorder is a part of the continuous monitoring system, the response time,
drift, and accuracy requirements established in the EPA performance specifications must
be considered when choosing the recorder itself. If a recorder is chosen that has poor
response time and limitations in recording accuracy, the overall monitoring system will
suffer. There are many factors that contribute to the relative inaccuracy (relative to the
EPA reference method) of a monitoring system. The recording system does not need
to be one of these factors if a proper choice of the system is made initially.
8.3.3 Recording Systems.- Intermittent Digital Recording
The analog chart recorders give a continuous record of the signal produced by an analyzer.
The digital recorder or data logger, however, selects some value (either an instantaneous
or integrated value) after a given time period and records it. For this reason, a digital
system may be characterized as recording data over intermittent periods. These periods
may be short, a tenth or hundredth of a second or less; but for too short a period, the
printed data produced might be unmanageable.
It should be noted that a data logger is not a computer or a microprocessor. A computer
can process data, convert it into emission rates, and record it in specified formats. Data
loggers merely record data at specified intervals. There are two options available on digital
recorders that extend their utility. These are an alarm monitoring capability and the
ability to print out by exception. A data logger, therefore, could be set to send off an
alarm or print out data once a specified value is reached. It could not, however, compute
the emission rate by the F-factor method and print it. A microprocessor or computing
system would be necessary in this case.
There are several advantages to digital instruments (1):
• They produce a permanent printed output record that can be readily understood
without having to interpret tracings, as in the case of a chart recorder.
• They can be modified to provide values that can be read directly (ppm, percent).
• Data can be read quickly with less chance of misinterpretation.
• Since the data have already been converted into digital form, the data logger
can be interfaced easily with a computer.
• The data may be duplicated easily, without the problems of shrinking or distortion
that may occur with a strip-chart record.
There are a few significant disadvantages with digital recording systems, such as the following:
• They are more complex and more difficult to troubleshoot than an analog recorder.
8-6
-------
• It is more difficult to detect trends from data given by a digital recorder.
• It is more difficult to compare digital data, either between different instruments
or over different time periods.
• It is difficult to troubleshoot intermittent or peculiar causes of failure, since the
data are averaged and an instantaneous signature of the system is not available
as with a strip-chart recorder.
The difficulty of detecting trends has been overcome in some systems by recording the
digital data on cassette tape. The tape can be read off on a computer and the data
for the time period of interest then can be graphed automatically with a plotter. This
method provides a convenient means of storing the continuous monitoring record. Cassette
tapes are easily handled and cataloged and detailed graphs need only be reproduced
when desired.
8.3.4 Recording Systems - Data Processors
The most convenient method of handling continuous monitoring data is with a data
processor. Several firms involved in the manufacture of stack monitors have seen the
need for the instrumentation that will rapidly average and compute data in terms of the
emission standard. An example of the type of data that can be produced by a computerized
system is given in Figure 8-2.
Formats of course can vary, but it is important to eliminate the manual task of reducing
the data. The preparation of the required EPA quarterly reports then becomes much easier.
There are two data processing methods that generally are used in continuous monitoring
systems. These are:
• Analyzer - Analog-to-digital (A/D) - Large general purpose computer or data
processing system
• Analyzer - Dedicated continuous monitor data acquisition system
The first method utilizes the plant computer or data processing system. The analog signals
from the flue gas monitors first must be converted into digital form by use of a data
logger or an A/ D converter; however, the data processing system may already have this
feature as a part of its software. The digital signals then are sent to the computer,
which is programmed to accept them and perform the necessary calculations for the
resultant printout. There are several problems with this method. First, the plant computer,
designed or purchased for process applications, may not have enough storage or programming
facility to accommodate the continuous-monitoring requirements. Second, when the computer
is down, the continuous-monitoring data may be lost or difficult to retrieve. The in-house
8-7
-------
EMISSIONS SUMMARY REPORT
HOURLY REPORT
06/18/75
OP AC IT YT X 14
(6 MIN AVGS) 14
HOURLY AVG-
I INDICATED
ZERO CAL
SPAN CAL 40
CORRECTED 14
LE--/MBTIJ
KLE/HR
NOTES
ij-SCFM*10E3 V
40J.06 '•$'£.
HEAT-MDTU/HR
1722.5
DAILY REPORT
(36/16/75
OP AC
6MIN, INT
PERIOD NUMBER
0 0 ti 0 - # 3 (3 0 0
0301-0600 0
(3 6 w I - !? 9 f? 0 (3
09(?1 - 12^0 1
1201-1500 Z
1501-1500 5
1801-210(3 1
2101-2400 0
,770
.770
OPAC
--
0244
. 136
.846
_-
--
CN
-IND
. -3V£
T
500
ITY
AV.
21
23
25
22
TIME:
.770. 14.770 11.
14.770 14.770
PART 802
544.92
65,185
1177.7
.01 39 533. (32
.0278 1.1351
.0479 1.96(32
V-AVG DIA-FT
37.807 20
P H20-V.
417 10
802
MAG 3HR.INT AV.MAG
X NUMBER LE/METU
1 1.24
0
0
.6 1 1.32
.4 1 1.65
.6 1 1 . 63
.7 0
0
0600-07012
474 18.
14.770
NO
070 ^ o
-Jl^Jt-J-J
22.705
1150
389.80
.3878
.672
MW
30.5
823
14.770
02
3.7246
2.0214*
8.872*
4. 1383
--
__
CB
NO
3HR. INT
NUMBER
0
0
0
1
1
1
0
0
AV.MAC-
LB/MBTU
_ _
--
--
0.92
0.84
0.75
..
™ fc
FIGURE 8-2
DATA FROM TYPICAL DATA PROCESSOR DESIGNED FOR
CONTINUOUS SOURCE MONITORING APPLICATIONS
8-8
-------
computer capability and existing utilization will have a direct bearing on any decision to
extend its use to continuous-emissions-monitoring applications.
The second method involves making a major purchase for a system that will be dedicated
to processing only the continuous monitoring data. A number of systems are available
on the market that are designed to do this. These systems can average automatically
opacity and gaseous emissions data, can compute emissions by the F-factor method using
the input of the pollutant gas and diluent gas monitor, and can provide a summary
report of the data on an hourly, weekly, or monthly basis. The systems can generate an
alarm signal and also may record the data on magnetic tape. The Emissions Summary
Report shown in Figure 8-2 is an example of the type of output that can be produced.
The dedicated systems may save time and money in the long run. Many source operators
will first purchase the gas analyzers and rely on strip-chart output for the data-recording
requirements. If the monitoring system is working properly and the data are reliable,
consideration is given to a data processor in order to reduce the amount of time spent
analyzing what can amount to volumes of data. Many operators have found it convenient
to keep the chart recorders to provide an easily interpreted record of the trends occurring
during the source operation. Cross checks then can be made between the two systems;
if either malfunctions, the data may not be lost.
8.4 Reporting Requirements
Continuous-emissions-monitoring data, obtained for the purpose of satisfying the EPA
regulations, must be reported on a quarterly basis. The originally proposed regulations,
which appear in the Federal Register, September II, 1974, required that all of the data
were to be reported to the EPA office. This proposal received a large number of comments
from both agency and industry personnel. It was generally felt that the amount of data
would be excessive and that the expense and manpower involved would be unjustified.
Changes were subsequently made in the promulgated October 6, 1975, regulations, effectively
requiring the reporting of only excess emissions.
Excess emissions are defined in the Subpart of the Code of Federal Regulations dealing
with the affected industry. In Subpart D 40 CFR 60, for example, excess emissions are
defined for fossil-fuel-fired steam generators. Under 40 CFR 60.45g:
• Opacity. Excess emissions are defined as any 6-minute period during which the
average opacity of emissions exceeds 20 percent, except that one 6-minute average
per hour of up to 27 percent opacity need not be reported.
• Sulfur dioxide. Excess emissions for affected facilities are defined as any 3-hour
period during which the average emissions (arithmetric average of three contiguous
1-hour period) of sulfur dioxide as measured by a continuous monitoring system
exceed the applicable standard under § 60.43.
8-9
-------
• Nitrogen oxides. Excess emissions for affected facilities using a continuous
monitoring system for measuring nitrogen oxides are defined as any 3-hour period
during which the average emissions (arithmetric average of three contiguous I-hour
periods) exceed the applicable standards under § 60.44.
Excess emissions may be defined differently for sources other than fossil-fuel-fired steam
generators. Reference should be made to the appropriate subparts of the Code of Federal
Regulations for the promulgated emission standards and average times (e.g., Subpart J for
petroleum refineries, see Table 2-2 for a reference listing of the CFR subparts).
Since a majority of the sources affected by the continuous-monitoring requirement will be
coal- and oil-fired power plants, a few additional comments on the excess emissions defined
in Subpart D are appropriate.
• The definition of excess emissions for opacity appeared December 5, 1977, in
42 FR 61537. Note that an exception of one 6-minute period at a level of
27 percent opacity is allowed. Prior to this promulgation, 2 minutes of soot
blowing at 40 percent opacity were allowed. This soot blowing allowance was
retained, but expressed in terms of the 6-minute average, i.e.,
-no, it w •* 2 X 40% + 4 X 20%
27% allowable opacity = *-
6
• Standards for SO2 and NOX emissions from fossil-fueled steam generators are
expressed in lbs/10" heat input. The emissions are to be calculated by using
the F-factor method. The F-factor method essentially reduces the amount of data
necessary to compute the emissions rate. A thorough explanation of the method
is given in Appendix C.
• Only excess emissions are required to be reported. All of the data produced by
a continuous source analyzer need not be converted into units of the standard.
Only the data reported as being in excess of the standard has to be expressed in
these terms. However, there should be some means to single out excess emissions
from the unreduced data. A source with varying emission and excess air rates
may have to convert all of the data into units of the standard. See the preamble
on page 46250 of the October 6, 1975, Federal Register for further clarification.
Reports of excess emissions determined from a continuous monitoring system are to be
reported on a quarterly basis. Each period of excess emissions also requires an explanation
of the reasons for the high values. These may be identified as startups, shutdowns,
or malfunctions of the affected facility. Also, if there were no periods of excess emissions
during the reporting quarter, a report to that effect must be made.
8-10
-------
The problem of monitoring equipment malfunctions is a matter of serious concern to the
continuous monitoring program. It is obvious that an improperly operating continuous
monitor serves neither the source operator nor the control agency. In order to keep aware
of the instrumental problems that inevitably develop, occasions of instrument downtime,
repair, or significant readjustment also must be documented and explained in the quarterly
report. Many agencies are now developing inspection programs for these systems in an effort
to insure that reliable emissions data can be obtained.
The source operator also must maintain a file of all of the continuous monitoring data,
including records of the Performance Specification Test, adjustments, repairs, and calibration
checks. The file must be retained for at least 2 years and is required to- be maintained in
such a condition that it can be easily inspected by a field enforcement officer.
The details of the reporting requirements are given in 40 CFR 60.7. Although no specific
format for the quarterly reports is required, the EPA Office in Region 8 has developed a
form for their use that includes the points required by the Code of Federal Regulations
(Figure 8-3). Although other formats may be suitable, this format could serve as a guide
for these reports.
-------
QUARTERLY EXCESS EMISSIONS REPORT
-------
Part 4. Conversion factors (not for diluent monitor report}
a. Diluent measured {02 or CO2l
b. F-Factor value used
j. Published or developed
ii. F, Fc, or Fw
c. Basis for gas measurement data (wet or dry)
d. Zero and Cal values used, by instrument:
Opacity (%) S02 (ppm) NOX (ppm) Diluent (% or ppm -
circle one)
Zero
Cal
Part 5. Continuous Monitoring System operation failures
See Table II: Complete one sheet for each monitor, including diluent:
attach separate narrative per instructions.
Part 6. Certification of report integrity, by person in 1-g, above:
THIS IS TO CERTIFY THAT TO THE BEST OF MY KNOWLEDGE,
THE INFORMATION PROVIDED IN THE ABOVE REPORT IS
COMPLETE AND ACCURATE.
NAME
SIGNATURE
TITLE
DATE
'Suggested Format for Subpart D sources in: Colorado, Montana, North Dakota,
South Dakota, Utah, Wyoming
FIGURE 8-3
SUGGESTED FORMAT FOR QUARTERLY EXCESS EMISSIONS REPORT—Continued
8-13
-------
OPACITY: Week6
TABLE la • Excess Emissions Summary by Week1
Day
Excess Emission
Range Category
A
B
C
0
E
S02
Excess Emission
Range Category
A
B
C
D
E
N0x
Excess Emission
Range Category
A
B
C
D
E
Percent of
Emission Limit
100-125
126-150
151-175
176-225
>225
Week6 Limit
Percent of
Emission Limit
10M08
109-120
121-135
136-155
> 155
Week6 Limit
Percent of
Emission Limit
101-108
109-120
121-135
136-155
> 155
Number of
6-Minute Periods
During Day2
Number of
3-Hour Periods
During Week2
Number of
3-Hour Periods
During Week2
q
Reason Codes
Reason Codes^
*i
Reason Codes
1 Format to be used in automatic data-handling systems; TaWe I (2) to be used in manually-prepared reports
to show each excess emission.
2 As defined in 60.45{g).
3 List in descending order the three most frequent codes, by number, followed in parenthesis by the number
of occurrences of the reason.
6 Begin Sunday morning at midnight; list date of the Sunday starting the week.
7 List the day of the week; e.g., Tuesday.
FIGURE 8-3
SUGGESTED FORMAT FOR QUARTERLY EXCESS EMISSIONS REPORT-Continued
8-14
-------
TABLE I
Excess Emissions (by pollutant)
Time Magnitude*
Date From-To Pollutant (%O2 or COj) U./106 BTU
'as defined in the instructions from the applicable section of the Federal Register; attach narrative of causes, etc.
TABLE II
Continuous Monitoring System Operation Failures
Time" Effect on
Date From-To Instrument Instrument Output
"attach narrative of causes, etc.
FIGURE 8-3
SUGGESTED FORMAT FOR QUARTERLY EXCESS EMISSIONS REPORT—Continued
8-15
-------
8.5 References
1. Quinn, G. C, "Recording Instruments - A Special Report," Part I, Power, December
1977, pp. sl-s28, Part 2, Powtr, January 1978, pp. s9-sI8.
2. Floyd, J. R., "The Implementation of the NSPS Continuous Emission Monitoring
Regulations in EPA, Region VIIi;" Paper 78-35.1, presented at the 71st Annual Meeting
of the Air Pollution Control Association, Houston, Texas, June 25-30, 1978.
8.6 Bibliography
McGowan, G. F., "Discussion of Alternative Emission Measurement Schemes for Wet
Scrubber Applications," Unpublished Monograph - Contact G. F. McGowan of Lear
Siegler, Inc.
8-16
-------
CHAPTER 9
EQUIPMENT SELECTION
9.1 Introduction
The selection of source monitoring instrumentation often has been a problem for those
needing to comply with air pollution regulations. Fortunately, instrumentation developed
over the past several years is workable and reliable.
One of the major problems in the past with continuous monitoring systems was not with
the monitor itself but with the system. Practically all of the instruments marketed
performed within specifications in the laboratory, but when set up in an extractive system
or placed across a stack, many problems would arise. These problems still occur, but
with the experience that has been gained, most of the problems now can be solved.
Although source monitors sold today are more carefully constructed and have better
specifications than their predecessors, the same limiting performance factors relate to
system and the application. For this reason, lists of instruments and their specifications,
which periodically appear in air pollution magazines (and in this handbook), should be
used only as a guide in selecting a monitor.
In this chapter, selection guides are given for both opacity and gas monitors (Tables 9-1
and 9-2). These guides are intended to answer a number of questions when evaluating a
monitor. No single instrument could meet all of the criteria implied by these questions.
The purchaser of the equipment should evaluate which features are important or unimportant
for the particular application.
The lists of vendors in this chapter (Tables 9-3 through 9-8) were compiled from the
trade literature, responses from information requests, and personal communications with
vendors at instrument shows. It is felt that these lists and tables represent the state of
the market as of March 1979. There are many other companies that manufacture gas
analyzers. However, many of these are not designed for measurements at source level
concentrations or source conditions. Only those companies whose instruments are designed
specifically for source applications and who are seriously competing in this market are
listed here. The competitive nature of this market leads to frequent changes in product
lines and the occasional demise or reorganization of a firm. Ideally, these tables should
be updated on a 6-month basis to remain current. In any case, the tables should be
viewed only as guidelines to the market.
9-1
-------
TABLE 9-1
OPACITY MONITORS - SELECTION PROCEDURES
Choice of Instrument Method
Determine the Type of System Required
• Single Pass
• Double Pass
• Multi-Parameter (opacity & gases)
• Stack Diameter & Optics Requirements - Operational Distance
• Cross-stack Permanent
• In-stack Portable
• Breech Pipe System
• Automatic Features
Choice of the Specific Instrument
EPA Requirements
1. Does it meet EPA requirements for the following:
Peak Spectral Response 500-600 nm
Mean Spectral Response 500-600 nm
Angle of View ^5°
Angle of Projection <5°
2. Does it or will it satisfy EPA performance requirements in terms given in 40 CFR 60
Appendix B?
Calibration Error <3% opacity
Zero Drift (24 hr.) <^2% opacity
Calibration Drift (24 hr.)
-------
TABLE 9-1
OPACITY MONITORS - SELECTION PROCEDURES-Continued
4. Is it necessary that the monitor meet the Performance Specifications for the given
application?
5. What experience can the vendor demonstrate under the relevant EPA requirements
and in other applications?
Design Characteristics
1. Cost
2. Blower System
a. Is it required? (is the stack pressure positive or negative? how does pressure
vary during startup and shutdown?)
b. Air flow rate capacity (ft. /min.)
c. Filters — type, number of stages, filter capacity?
d. Shutters - for system or power failure?
e. Design of instrument purge system (i.e., is air supplied by blower actually effective
in keeping optics clean)?
3. Optical Assembly
a. Sensitivity to ambient light (chopper-modulator)?
b. Collimation method (self-correcting?)
c. Calibration System
1. Is same lamp and detector used in zero check?
2. Is same lamp and detector used in span check?
3. Is automatic zero calibration correction available?
d. Does it have alignment viewing port or sight glass for aligning system?
e. Will the assembly adapt to the stack diameter?
f. Maintenance Indicators — does the system monitor its operating condition and
alert operator to required maintenance?
9-3
-------
TABLE 9-1
OPACITY MONITORS - SELECTION PROCEDURES-Continued
4. Electronic System
a. Analog or digital display?
b. Easy to wire and set up?
c. Warning system available? Remote, visible, audible? Alarm set-point?
d. Automatic zero and calibration checks?
e. Remote manual calibration available?
f. Automatic counter-timer for recording excess emissions?
g. Linear response, units and measurement ranges?
h. Automatic stack exit correlation available?
i. Automatic optical density display?
j. Output requirements (mA or mV)?
k. Computer interface?
1. Quick disconnect cables or hard wired?
m. Recorder module available? Separate recorder required?
n. Sensitivity to line voltage fluctuations?
o. What options are available to enhance applications flexibility?
Environmental Requirements
1. Can it operate in a corrosive environment?
2. Will the readings remain stable under varying ambient temperatures?
3. Is the instrument constructed ruggedly enough for the proposed location?
4. Is an explosion-proof cabinet available?
5. Does the instrument satisfy space and weight requirements?
6. Does the instrument severly restrict or modify the flue gas flow?
7. Can the system operate under positive or negative static pressure?
8. Can the instrument withstand vibration if located on or near the stack?
9-4
-------
TABLE 9-1
OPACITY MONITORS - SELECTION PROCEDURES-Continued
9. Can the instrumental parts in or near the stack withstand high temperatures and stack
temperature fluctuations?
10. Can the optical system of the transmissometer account for alignment changes brought
about by changes in stack or duct temperature?
Maintenance and Operational Considerations
1. What type of warranty is available?
2. Is there a guaranteed maintenance-free period? Or guaranteed operational period?
3. Is field installation supervision available?
4. Is user training available?
5. Will the instrument manufacturer assist in the performance specification tests?
6. Are leasing or service contracts available?
7. Are service contracts required?
8. Is a special operator or special maintenance required?
9. Is cleaning or the replacement of parts required on a regular basis? If so - cost and
labor? How available are parts?
a. What are the sizes (costs) of the basic replacement modules?
b. What test points, diagnostics, decal information, troubleshooting procedures are
provided to simplify diagnosis and repair?
c. Do schematics reference commonly available component descriptions or only
factory part numbers?
d. How complete is technical manual provided with the equipment?
e. Where are most of the electronics and repairs likely to be encountered - on stack
or in the control room?
10. Is the instrument easy to service? Is access to previous service records available?
11. What is the lamp life expectancy? (>20,000 hrs.?)
12. How often do lenses or filters need cleaning or replacement? How easy is this to do?
9-5
-------
:'• - TABLE 9-1
OPACITY MONITORS - SELECTION PROCEDURES-Continued
13. Would the instrument's location be accessible for servicing?
14. What adjustments or manual checks are required during operation?
15. Is this a new model instrument? When was the last design change?
16. Vendor Characteristics:
a. What is the vendor's commitment to this product line?
b. What is the probability that the vendor will be able to service the equipment over
its full design life?
c. What research and development capability does the vendor have in this product
line?
d. What is the availability of applications assistance?
e. How many service people are actually trained on this product line and where
are they located?
f. What is the company's ability to guarantee compliance with EPA requirements?
9-6
-------
TABLE 9-2
GASEOUS MONITORS - SELECTION PROCEDURES
Choice of Monitoring System Extractive or In-Situ
Determine the Type of System Required
1. Is the system needed to monitor a single stack or multiple stacks?
2. How many pollutants are required to be monitored?
3. Can a single monitor be used or are several instruments needed?
4. How severe is gas stratification? Cyclonic flow?
5. What response time is required?
6. How representative will the sample be? (Will conditioning affect the sample?)
7. What are the location requirements for an extractive or in-situ system?
Choice of PhysicalrChemical Method of Analysis
First: Determine the following:
1. Type of operation and details of the process.
2. Stack gas composition.
3. Pollutant concentrations and variation.
4. Stack gas temperature and ambient temperature (variations of both).
5. Stack gas velocity and volumetric flowrate.
6. Stack static pressure.
7. Moisture content and dew point of stack gas.
8. Particulate loading.
9. Stack dimensions and possible locations for the monitoring system.
10. Abrasion and corrosion problems.
9-7
-------
TABLE 9-2
GASEOUS MONITORS - SELECTION PROCEDURES-Continued
Ask the following questions:
I. Is the method sensitive to interferences at levels characteristic of the source?
a. H2O vapor, CO2-
b. Entrained mist or condensed vapors.
c. Heavy paniculate loading.
2. Will variable temperature be a problem?
3. Will high and/or varying stack gas velocity and pressure be a problem?
4. Will high temperatures be a problem?
5. Will the method be accurate under conditions experienced at the source?
6. Can the method be adapted to the siting requirements?
7. Will it require so much conditioning of stack gas that accuracy or validity will be
questioned?
Choice of the Specific Instrument
EPA Requirements
1. Does it or will it satisfy EPA performance requirements in terms of
SO2 & NOx O2 or CO2
Accuracy
Calibration error
Zero drift (2 hr. & 24 hr.) 2% of span <0.4% & <0.5%
Calibration drift (2 hr. & 24 hr.) 2.5% of span <0.4% & <0.5%
Response time 15 min. (max) 10 min. (max)
Operational period 168 hrs. 168 hrs.
9-8
-------
TABLE 9-2
GASEOUS MONITORS - SELECTION PROCEDURES-Continued
2. Has the instrument ever undergone a performance specification test?
3. How extensive has previous use been? Documentation?
4. Is the response linear over the operating range?
5. What is the signal-to-noise ratio?
6. Would the system be reliable for long term continuous operation?
Design Characteristics
1. Cost? Immediate and long term.
2. Is the design and operation simple?
3. Does it have multi-gas measuring capability?
4. Does it read in the correct range - is auto-ranging capability available?
5. Is the calibration method simple and convenient? Are adjustments simple and con-
venient? How can the operator verify calibration of the system as it is installed?
6. Does it read out directly in concentration? Does it correct for dilution of air?
7. Can the instrument be easily installed with few alterations to the existing facility?
8. What are the sampling volume requirements? (Will it rob gas from other monitors
in an extractive system?) (Will it take too much span gas each time it is calibrated?)
9. What are the power requirements?
10. Does it require compressed air or an air blower?
II. Is the instrument sensitive to power fluctuations?
12. What is the warm-up time after shut-off?
13. Does it have a display analog or digital?
14. Does it have automatic zero and calibration?
15. Is there a warning device? Remote, visible, audible?
16. Is the recorder output compatible with your recorder on data handling system? Is an
interface needed?
17. Does the instrument have good accessibility for repairs and service?
9-9
-------
TABLE 9-2
GASEOUS MONITORS - SELECTION PROCEDURES-Continued
18. What automatic features are available to alert operator of operating condition or the
need for maintenance?
19. What options are available to enhance the flexibility of application of the instrument?
Environmental Requirements
1. Can it operate in a corrosive environment?
2. Will the readings remain stable under varying ambient temperatures?
3. Is the instrument constructed ruggedly enough for the proposed location?
4. Is an explosion-proof cabinet available?
5. Does the instrument satisfy space and weight requirements?
6. Does the instrument severely restrict or modify the flue gas flow?
7. Can the system operate under positive or negative static pressure?
8. Can the instrument withstand vibration if located on or near the stack?
9. Can the instrumental parts in or near the stack withstand high temperatures? And
stack temperature fluctuations?
Maintenance and Operational Considerations
I. What type of warranty is available?
2. Is there a guaranteed maintenance-free period?
3. Is field installation supervision available?
4. Is user training available?
5. Will the instrument manufacturer assist in the performance specification test? What
is the company's ability to guarantee compliance with EPA requirements?
6. Are leasing or service contracts available?
7. Are service contracts required?
8. Is a special operator or special maintenance required?
9-10
-------
TABLE 9-2
GASEOUS MONITORS - SELECTION PROCEDURES-Continued
9. Is cleaning or the replacement of parts required on a regular basis? If so — cost and
labor? How available are parts?
10. How often do lenses and/or filters need cleaning or replacement?
11. Is the instrument easy to service? Is access to previous service records available? How
many service people are actually trained on this product line and where are they located?
12. Would the instrument's location be accessible for servicing?
13. What adjustments or manual checks are required during operation?
14. Is this a new model instrument or has it been a basic design used for a long time?
Are you going to be stuck with a "one of a kind" instrument?
15. When was the last model change?
16. What is the vendor's commitment to this product line?
17. Are wiring diagrams supplied from the manufacturer? If so, are they for your model
and accurate? How complete is the technical manual provided with the equipment?
18. Is the instrument safe? Does it use hydrogen? Are the reagents safe?
19. How susceptible is the instrument to plugging?
20. Will the extractive materials of construction, etc., withstand corrosion?
21. Can the pumps withstand corrosion?
22. What is the size (cost of the basic replacement modules)?
23. What test points, diagnostics, decal information, troubleshooting procedures are provided
to simplify diagnosis and repair?
24. What is the probability that the vendor will be able to service the equipment over its
full design life?
25. What research and development capability does the vendor have in this product line?
9-11
-------
TABLE 9-3
VENDORS OF DOUBLE-PASS TRANSMISSOMETERS
Cost Range $8000-$ 16000
Environmental Data Corp.
608 Fig Avenue
Monrovia, CA 91016
Lear Siegler, Inc.
74 Inverness Drive East
Englewood, CO 80110
Research Appliance Co.
Chemed Corp.
Route 8
Gibsonia, PA 15044
Dynatron Inc.
57 State Street
North Haven, CT 06473
Same
Instrument
Same
Instrument
Contraves Goerz Corp.
301 Alpha Drive
Pittsburgh, PA 15238
Western Precipitation Div.
Joy Manufacturing Co.
P.O. Box 2744 Terminal Annex
Los Angeles, CA 90051
Esterline Angus (Marketing Durag Instrument-Germany)
Box 24000
Indianapolis, IN 46224
Datatest, Inc.
1117 Cedar Avenue
Croyden, PA 19020
9-12
-------
TABLE 9-4
VENDORS OF SINGLE-PASS TRANSMISSOMETERS
Bailey Meter
29801 Euclid Avenue
Wickliffe, OH 44092
Cost Range $800-$4000
Leeds & Northrop
Sumneytown Pike
North Wales, PA 19454
Cleveland Controls, Inc.
1111 Brookpark Road
Cleveland, OH 44109
Photomation Inc.
270 Polatis Avenue
Mountain View, CA 94042
De-Tec-Tronic Corp.
2512 N. Halsted Street
Chicago, IL 60614
Preferred Utilities Manufacturing
II South Street
Danbury, CT 06810
Electronics Corp. of America
1 Memorial Drive
Cambridge, MA 02142
Reliance Instrument Manufacturing
164 Garibaldi Avenue
Lodi, NJ 07644
HABCO
85 Nutmeg Lane
Glastonbury, CN 06033
Robert H. Wager
Passiac Avenue
Chatham, NJ 07928
9-13
-------
TABLE 9-5
PRINCIPAL CONTINUOUS SOURCE MONITOR
MANUFACTURER SUMMARY (JULY 1978)
Extractive
Instrument
Vendor
Beckman
Bendix
Esterline
Angus
Horiba
Infrared Ind.
Leeds and
Northrop
MSA
Teledyne
SO2
X
X
X
X
X
X
X
Gases Measured
NO NO2 CO2
Nondispersive
X X
X X
X X
XXX
X X
X
X X
X
CO
Monitors
Measurement Turnkey
Range Systems
Approximate
Cost in
Thousands
of Dollars
Infrared Instruments
X
X
X
X
X
X
X
Various ranges Yes
in ppm or
percent
0.5 ppm - 50% Yes
2 ppm - 100%
10 - 2000 ppm
200 ppm - I0%*
0 - 1000 ppm* Yes
0 - 2000 ppm* Yes
0 - 1000 ppm* Yes
3-5.4
3-4 (I)
5
3-5
1-2
5.5 (I)
3-4(1)
1I-13(S)
Extractive Differential Absorption Instruments
CEA
DuPont
Esterline
Angus
Teledyne
Western
X
X
X
X
X
X
X X
x -
X X
X
X
2-50,000 ppm
I ppm - 100% Yes
Yes
2 ppm - 100% Yes
75-5000 ppm Yes
3-6
13-23(5)
12-14 (S)
12-22 (S)
(S) - System cost estimate by manufacturer on request
(I) - Instrument cost only; does not reflect system cost
* — Other ranges available upon request
9-14
-------
TABLE 9-5
PRINCIPAL CONTINUOUS SOURCE MONITOR
MANUFACTURER SUMMARY (JULY 1978)-Continued
Instrument
Vendor SO2
Approximate
Cost in
Gases Measured Measurement Turnkey _, , ,.
NO NO2 CO2 CO Range Systems Dollars
Fluorescence Instruments
Research
Appliance X
Corp.
Thermo
Electron X
Corp.
1-5000 ppm 6
1-10,000 ppm 6-7
Chemiluminescence Instruments
Beckman
Bendix
McMillan
Electronics
Meloy
Monitor
Labs
Scott
Source Gas
Analyzers
Inc.
Thermo
Electron
Corp.
X X 0-9% 6.2
X X Yes (S)
X X
X X 0-2000 ppm Yes 5.4
X X
X X 0-10,000 ppm 4-5
X X' 0-3000 ppm
X X 5-10,000 ppm 5-6
Flame Photometric Instruments
Meloy X
Process
Analyzers, X
Inc.
Tracer X
25-10,000 ppm Yes 3
5 ppm - %
9
9-15
-------
TABLE 9-5
PRINCIPAL CONTINUOUS SOURCE MONITOR
MANUFACTURER SUMMARY (JULY I978)-Continued
Instrument
Vendor SO2
Gases Measured
NO NO2 CO2 CO O2
M easurement Turnkey
Range Systems
Approximate
Cost in
Thousands of
Dollars
Polarographic Instruments
Beckman
IBC/
Berkeley X
Dyna- X
sciences
InterScan
Corp. X
Teledyne
Theta
Sensors X
(MRI)
Western
Precipi- X
tator (Joy)
X
X X
XX XX
X X
X
X X
X X X X
0-25%
0-10,000 ppm
0.01-200,000 Yes
ppm
Yes
0-25%
1-20,000 ppm
0-1000 ppm
1-1.5
2-5.5 (S)
2-8 (S)
1 d)
-------
TABLE 9-5
PRINCIPAL CONTINUOUS SOURCE MONITOR
MANUFACTURER SUMMARY (JULY 1978)-Continued
Approximate
Cost in
Instrument Gases Measured
Vendor SO2 NO NO2 CO2 CO O2
Measurement Turnkey
Range
Systems
Thousands of
Dollars
Amperometric Instruments
Barton
ITT X
Inter-
national X
Ecology
Systems
0-1000 ppm
0-10,000
ppm
Paramagnetic Instruments
Beckman
MSA
CEA
SCOTT
Leeds
and
Northrop
Taylor-
Servomex
X 0-25%
X 0-25%
X
X 0-100%
X
X 0-100%
Yes
1-1.5
(S)
1-1.5
9-17
-------
TABLE 9-6
OXYGEN ANALYZER SUMMARY
Analysis Method
Vendor Paramagnetic Polarographic
Astro
Beckman X
Cleveland
Controls X
Corning
Dyna-
sciences X
Dynatron
Esterline
Angus X
Gas Tech X
Hays-
Republic
Joy X
Lear
Siegler
Leeds and
Northrop X
Lynn X
MSA X
Scott X
Taylor-
Servomex X
Teledyne X
Thermox
Theta
Sensors X
Westing-
house
Electro-
catalytic
X
X
X
X
X
X
X
X
X
Sampling Type
In-Situ Extractive
X X
X
X X
X
X
X
X
X
X X
X
X
X
X
X
X
X
X
X
X
X
9-18
-------
TABLE 9-7
IN-SITU MONITOR SUMMARY
Approximate
Method Measure- Cost in
Cross- mem Thousands
Vendor SO2 NO CO2 CO O2 Opacity In-stack stack Range of Dollars
Gases Measured
CEA
Contra ves
Goerz
Dy natron
Environ-
X
X
0-25%
TABLE 9-8
LIST OF INSTRUMENT MANUFACTURERS
30
mental X X X X X
Data Corp.
Lear
Siegler XX X
Westing-
house X
X 0-5000
ppm
X 0-500;
0-1000;
0-1500
ppm
X
20-40
4.5-17
MANUFACTURERS OF ND1R MONITORS
Positive Filtering Instruments
Negative Filtering Instruments
Beckman Instruments, Inc.
2500 Harbor Boulevard
Fullerton, CA 92634
Calibrated Instruments, Inc.
731 Saw Mill River Road
Ardsley, NY 10502
Bendix Corporation
Process Instruments Div.
P.O. Drawer 831
Lewisburg, WV 24901
Esterline Angus
19 Rozel Road
Princeton, NJ 08540
9-19
-------
TABLE 9-8
LISTS OF INSTRUMENT MANUFACTURERS-Continued
Positive filtering Instruments
Negative Filtering Instruments
CEA Instruments (Peerless)
555 Madison Avenue
New York, NY 10022
Horiba Instruments, Inc.
1021 Duryea Avenue
Santa Ana, CA 92714
Infrared Industries
P.O. Box 989
Santa Barbara, CA 93102
Leeds & Northrop
Sumneytown Pike
North Wales, PA 19454
MSA Instrument Division
Mine Safety Appliances
201 Penn Center Boulevard
Pittsburgh, PA 15208
Teledyne - Analytical Instruments
333 West Mission Drive
P.O. Box 70
San Gabriel, CA 91776
MANUFACTURERS OF EXTRACTIVE DIFFERENTIAL
ABSORPTION ANALYZERS
Teledyne - Analytical Instruments
333 West Mission Drive
P.O. Box 70
San Gabriel, CA 91776
CEA Instruments
555 Madison Avenue
New York, NY 10022
Western Research and Development Ltd.
Marketing Department
No. 3, 1313 - 44th Ave. N.E.
Calgary, Alberta T2E GL5
DuPont Company
Instrument Products
Scientific & Process Div.
Wilmington, DE 19898
Esterline Angus
19 Rozel Road
Princeton, NJ 08540
9-20
-------
TABLE 9-8
LISTS OF INSTRUMENT MANUFACTURERS-Continued
MANUFACTURERS OF FLUORESCENCE SOURCE ANALYZERS
Thermo Electron Corporation
Environmental Instruments Div.
108 South Street
Hopkinton, MA 01748
MANUFACTURERS OF CHEMILUMINESCENCE ANALYZERS
Beckman Instruments, Inc.
Process Instruments Division
2500 Harbor Boulevard
Fullerton, CA 92634
Bendix Corporation
Process Instruments Division
P.O. Drawer 831
Lewisburg, WV 24901
McMillan Electronics Corporation
7327 Ashcroft
Houston, TX 77036
Meloy Laboratories, Inc.
6715 Electronic Drive
Springfield, VA 22151
Monitor Labs
4202 Sorrento Valley Boulevard
San Diego, CA 92121
Scott Environmental Systems Division
Environmental Tectonics Corporation
County Line Industrial Park
Southampton, PA 18966
Source Gas Analyzers, Inc.
7251 Garden Grove Boulevard
Garden Grove, CA 92641
Thermo Electron Corporation
Environmental Instruments Division
108 South Street
Hopkinton, MA 01748
9-21
-------
TABLE 9-8
LISTS OF INSTRUMENT MANUFACTURERS-Continued
MANUFACTURERS OF FLAME PHOTOMETRIC ANALYZERS
Tracer, Inc.
Analytical Inst.
6500 Tracer Lane
Austin, TX 78721
Meloy Laboratories, Inc.
6715 Electronic Drive
Springfield, VA 22151
Process Analyzers, Inc.
1101 State Road
Princeton, NJ 08540
MANUFACTURERS OF POLAROGRAPHIC ANALYZERS
Dynasciences (Whitaker Corp.)
Township Line Road
Blue Bell, PA 19422
IBC/Berkeley Instruments
2700 DuPont Drive
Irvine, CA 92715
Western Precipitation Division
Joy Manufacturing Company
P.O. Box 2744 Terminal Annex
Los Angeles, CA 90051
(Portable models - not designed
for continuous stack application)
Beckman Instruments, Inc.
Process Instruments Division
2500 Harbor Boulevard
Fullerton, CA 92634
(02 only)
Gas Tech Inc.
Johnson Instrument Division
331 Fairchild Drive
Mountain View, CA 94043
(02 only)
InterScan Corp.
20620 Superior Street
Chatsworth, CA 91311
Theta Sensors, Inc.
Box 637
Altadena, CA 91001
(will provide systems)
Teledyne Analytical Instruments
333 West Mission Drive
San Gabriel, CA 91776
(O2 only - micro-fuel cell)
Lynn Products Company
400 Boston Street
Lynn, MA 01905
only)
9-22
-------
TABLE 9-8
LISTS OF INSTRUMENT MANUFACTURERS-Continued
MANUFACTURERS OF ELECTROCATALYTIC OXYGEN ANALYZERS
Westinghouse Electric Corporation
Computer and Instrument Division
Orrville, OH 44667
(in-situ)
Lear Siegler, Inc.
Environmental Technology Division
74 Inverness Drive East
Englewood, CO 80110
(in-situ)
Dynatron. Inc.
Barnes Industrial Park
Wallingford, CT 06492
Teledyne Analytical Instruments
333 West Mission Drive
San Gabriel, CA 91776
Astro Resources Corp.
Instrument Division
P.O. Box 58159
Houston, TX 77573
Mine Safety Appliances
Instrument Division
201 Penn Center Boulevard
Pittsburgh, PA 15235
(extractive)
Thermox Instruments, Inc.
6592 Hamilton Avenue
Pittsburgh, PA 15206
Cleveland Controls, Inc.
1111 Brookpark Road
Cleveland, OH 44109
Corning Glass Works
Ceramic Products Division
Corning, NY 14803
(designed for glass furnaces)
Hays-Republic
Milton Roy Company
4333 So. Ohio Street
Michigan City, IN 46360
MANUFACTURERS OF AMPEROMETRIC ANALYZERS
Barton ITT
Process Instruments and Controls
580 Monterey Pass Road
Monterey Park, CA 91754
International Ecology Systems
4432 North Kedzie Avenue
Chicago, IL 60625
(combined colorimetric method)
9-23
-------
TABLE 9-8
LISTS OF INSTRUMENT MANUFACTURERS-Continued
MANUFACTURERS OF CONDUCTIMETRIC ANALYZERS
Calibrated Instruments, Inc.
731 Saw Mill Road
Ardsley, NY 10502
MANUFACTURERS OF PARAMAGNETIC ANALYZERS
Cleveland Controls, Inc.
1111 Brookpark Road
Cleveland, OH 44109
Scott Environmental Systems Division
Environmental Tectonics Corp.
County Line Industrial Park
Southampton, PA 18966
Taylor Servomex-Sybron Corp.
Analytical Instrument Division
Rochester, NY 14604
Mine Safety Appliances Co.
201 Penn Center Boulevard
Pittsburgh, PA 15235
Beckman Instruments, Inc.
Process Instruments Division
2500 Harbor Boulevard
Fullerton, CA 92634
Leeds and Northrop
Sumneytown Pike
North Wales, PA 19454
MANUFACTURERS OF CONDUCTIVITY ANALYZERS
Leeds and Northrop
Sumneytown Pike
North Wales, PA 19454
Esterline Angus
19 Rozel Road
Princeton, NJ 08540
MANUFACTURERS OF IN-SITU MONITORS
Cross-Stack
Environmental Data Corporation
608 Fig Avenue
Monrovia, CA 91016
Contraves Goerz Corporation
610 Epsilon Drive
Pittsburgh, PA 15238
9-24
-------
TABLE 9-8
LISTS OF INSTRUMENT MANUFACTURERS-Continued
MANUFACTURERS OF IN-SITU MONITORS-Continued
In-Stock
Lear Siegler, Inc.
Environmental Technology Division
74 Inverness Drive East
Englewood, CO 80110
Oxygen Monitors Only
Westinghouse Electric Corporation
Computer and Instrument Division
Orville, OH 44667
Dynatron. Inc.
Barnes Industrial Park
Wallingford, CT 06492
Cleveland Controls, Inc.
Ill I Brookpark Road
Cleveland, OH 44109
Corning Glass Works
Ceramic Products Division
Corning, NY 14803
Hays-Republic
Milton Roy Company
4333 So. Ohio Street
Michigan City, IN 46360
9.2 Vendors of Recording Instrumentation
Although there are many manufacturers of strip chart recorders, there are a few firms
that tend to dominate sales in industrial applications. A sampling of these is given in
Table 9-9.
There are over 50 manufacturers of data-logging type equipment. Those most frequently
encountered in air pollution applications are given in Table 9-10. Vendors of data processors
designed for continuous stack emission data are given in Table 9-11.
9-25
-------
TABLE 9-9
MANUFACTURERS OF STRIP CHART RECORDERS
Beckman Instruments
Process Instruments Division
2500 Harbor Boulevard
Fullerton, CA 92634
Esterline Angus Instr. Corp.
An Esterline Company
1201 Main-Box 24000
Indianapolis, IN 46224
The Foxboro Co.
38 Neponset Avenue
Foxboro, MA 02035
Gulton Industries Inc.
M easurement-C ontrol
Systems
Gulton Industrial Park
East Greenwich, RI 02818
Hewlett-Packard
Scientific Instruments Division
1601 California Avenue
Palo Alto, CA 94304
Honeywell Inc.
Process Control Division
1100 Virginia Drive
Fort Washington, PA 19034
Leeds & Northrop Co.
Sumneytown Pike
North Wales, PA 19454
Westinghouse Electric Co.
Product Info Center
Westinghouse Building
Gateway Center
Pittsburgh, PA 15222
TABLE 9-10
MANUFACTURERS OF DATA LOGGING EQUIPMENT
Acurex Corporation
Autodata Division
485 Clyde Avenue
Mountain View, CA 94042
Datel Systems, Inc.
1020 Turnpike Street
Canton, MA 02021
Doric Scientific
Division of Emerson Electric
3883 Ruffin Road
San Diego, CA 92123
Esterline Angus
Instrument Corp.
Box 24000
Indianapolis, IN 46224
Monitor Labs, Inc.
4202 Sorrento Valley Boulevard
San Diego, CA 92121
Zonics, Inc.
6862 Hayvenhurst Avenue
Van Nuys, CA 91406
9-26
-------
TABLE 9-11
MANUFACTURERS OF CONTINUOUS MONITOR DATA PROCESSORS
Acurex Corporation
485 Clyde Avenue
Mountain View, CA 94042
(Autodata Nine CSM)
E.I. DuPont DeNemours & Co., Inc.
Instrument Products Division
Wilmington, DE 19898
(463 Emission Monitoring System
Data Processor — Part of Total
Monitoring DuPont Systems)
Bendix
Process Instruments Division
P.O. Drawer 831
Lewisburg, W VA 24901
(Model 9000 Analyzer Control)
Electro Scientific Industries, Inc.
13900 NW Science Part Drive
Portland, OR 97229
(Model ESI-6000)
Esterline Angus
P.O. Box 24000
Indianapolis, IN 46224
Environmental Data Corp.
608 Fig Avenue
Monrovia, CA 91016
(EDC-3110 Data Systems)
Lear Siegler, Inc.
Environmental Technology Division
74 Inverness Drive East
Englewood, CO 80110
(DP-30 Data Processor)
Dynatron, Inc.
Barnes Industrial Park
Wallingford, CT 06492
Dynasciences
Tpwnship Line Road
Blue Bell, PA 19422
(Model 6043 Prog. Data Acquisition
System)
(Computing Limits Reporter
Model 6000)
9-27
-------
9.3 Bibliography
There have been a number of articles published that deal with the evaluation or comparison
of source monitors. Unfortunately, many of these publications are misleading. A number
of them have been written by instrument vendors wishing to publicize their product line.
Others have been written with a particular bias in mind, where the author may be attempting
to show that no monitor operates properly, or that monitors can do more than they are
capable of. Another problem is that the experiments performed to evaluate a number of
systems often have been poorly designed or have suffered because of a lack of funding and
time. The evaluation of these systems under field conditions is an expensive and tedious
procedure; very few good field comparison tests have actually been done. Also, many of
the articles referenced in this section are now dated. The work described involves monitors
that are no longer marketed, and conclusions reached for a given model of an instrument
are no longer applicable, since the model may have been modified since the test.
These references, however, can give a broader perspective to the reader of some of the
problems that may be encountered in selecting, installing, and operating a continuous
monitoring system. Specific articles, applicable to the reader's application, should be
obtained to give insights into possible problems. The articles, however, should be read
critically.
American Laboratory, Laboratory Buyer's Guide Edition* November 1977.
"Air Quality Monitoring Equipment Reference Guide," Pollution Equipment News, V. 10
No. 3. June 1977.
Bambeck, R. J., and Huettemeyer, "Operating Experience with In-Situ Plant Stack Monitors,"
paper presented at 1967 Annual Meeting of APCA, Denver, Colorado, June 1974.
Bogatie, C. F., and Saltzman, R. S., "A Survey of Continuous Monitoring of Stack Emissions
from Kraft Mills," TAPPI. V. 58, No. 10:81-84. October 1975.
Chand, R., and Marcote, R. V., "Evaluation of Portable Electrochemical Monitors and
Associated Stack Sampling for Stationary Source Monitoring," presented at the 68th National
Meeting of the American Institute of Chemical Engineers, Houston, Texas, February 28-
March 4, 1971, (Addresses: Dynasciences Corp., Environmental Products Division,
9100 Independence Avenue, Chatsworth, CA 91311; American Institute of Chem. Engrs.
345 E. 47th Street. New York, New York 10017.)
Driscoll, Becker, McCoy, Young, and Ehrenfeld, "Evaluation of Monitor Methods and
Instrumentation for Hydrocarbons and Carbon Monoxide in Stationary Source Emissions,"
Research Corp., EPA Contract No. 68-02-0320, EPA-R2-72-106, November 1972.
9-28
-------
Green, M. W., et al., Beckman Instruments, Inc., EPA Contract No. 68-02-1743, EPA-
600/2-76-171, June 1976.
Homolya, J. B., "A Comparison of In-Situ and Extractive Measurement Techniques for
Monitoring SO2 Emissions from a Stationary Source," Science of the Total Environment,
V. 3:349, 1975.
Homolya, J. B., "Monitoring Systems of Gaseous Emissions: Evaluation of Commercially
Available Systems and New Developments," Energy and the Environment Proceedings of
the Fourth National Conference, pp. 574-577, AICHE, Dayton, Ohio, 1976.
Irwin, G. B., "Legal Aspects of Vendor Design," Journal of the Air Pollution Control
Association, V. 26, No. 8:748-752, August 1976.
Jahnke, J. A., Cheney, J. L, and Homolya, J. B., "Quenching Effects in SO2
Fluorescence Monitoring Instruments," Environmental Science & Technology, V. 10, Decem-
ber 1976.
Jaye, F. C, "Monitoring Instrumentation for the Measurement of Sulfur Dioxide in
Stationary Source Emissions," Report EPA-R2-73-I63, Environmental Protection Agency,
Office of Research and Monitoring, Washington, DC, February 1973.
Kretzchmar, J. G., Loos, M., and Bosnians, F., "Comparison of Three Instrumental Methods
to Monitor Nitrogen Dioxide," Science of the Total Environment, V. 7:181-187, 1977.
Instrumentation for Environmental Monitoring, LBL-1 V. 1: Air, Lawrence Berkeley
Laboratory, University of California, Berkeley, 1972.
McRanie, R. D., "Experience with Continuous Stack Monitoring Systems for SO2, NOX
and O2," SRI-Proceedings of Workshop on Sampling, Analysis and Monitoring of Stack
Emissions, Stanford Research Institute, NT1S PB-252-748, April 1976.
McRanie, R. D., Craig, J. M., and Layman, G. O., Evaluation of Sample Conditioners
and Continuous Stack Monitors for the Measurement of Sulfur Dioxide, Nitrogen Oxides
and Opacity in Flue Gas from a Coal-Fired Steam Generator, Southern Services, Inc.,
Birmingham, Alabama, February 1972.
Osborne, M. C., "Survey of Continuous Source Emission Monitors: Survey No. 1 -
NOX and SO2," EPA-600/4-77-022.
Osborne, C., "Survey of Continuous Gas Monitors to Emissions Sources," Chem Tech,
July 1974, pp. 426-431
9-29
-------
Parts, P. L., Sherman, P. L., and Snyder, A. D., Instrumentation for the Determination
of Nitrogen Oxides Content of Stationary Source Emissions, report of the Monsanto
Research Corp., Dayton Laboratory, Dayton, Ohio 45407, prepared for the EPA Office
of Research and Monitoring, EPA document APTD-0847, October 1971. (Address-
APTIC, EPA, Research Triangle Park, NC 27711.)
Pollution Engineering, Environmental Yearbook and Product Reference Guide, V. 9 No I
January 1977.
Pollution Equipment News, 1978 Catalog and Buyer's Guide. November 1977, V. 6.
Quick, D. L., Field Evaluation of SO2 Monitoring Systems Applied to //2-S04 Plant
Emissions, Volumes I & II, EPA Contract No. 68-02-1292, Scott Environmental Technology,
EPA-650/2-75-053a (Vol. I) and EPA-650/2-75-053b (Vol. II), July 1975.
Repp, M., Evaluation of Continuous Monitors for Carbon Monoxide in Stationary Sources
EPA-600/2-77-063, March 1977.
Sem, G. J., et al., State of the Art: 1971, Instrumentation for Measurement of Paniculate
Emissions from Combustion Sources. Volume I: Paniculate Mass - Summary Report.
Clearinghouse Report NTIS PB-202-655, Springfield, Virginia.
Shen, T., and Stasiuk, W. N., "Performance Characteristics of Stack Monitoring Instruments
for Oxides of Nitrogen," Journal of the Air Pollution Control Association, V. 25, No I
January 1975.
Shikiya, J. M., and MacPhee, R. D., "Multi-Instrument Performance Evaluation of Con-
ductivity-Type Sulfur Dioxide Analyzers," JA PCA, V. 19:943, 1969.
Snyder, A. D., et al., "Laboratory and Field Evaluation of Stationary Source Instrument
for^Oxides of Nitrogen Emissions," Air Quality Instrumentation, V. 2, Instrument Society
for America, Pittsburgh, 1974, pp. 184-203.
Wolbach, D. C, and James, R. E., "Texas Experience with Company Owned Monitors
and EPA Continuous Monitoring Requirements," Proceedings: Air Pollution Measure-
ment Accuracy as it Relates to Regulation Compliance Specialty Conference, Air Pollution
Control Association, New Orleans, Louisiana, October 1975.
9-30
-------
CHAPTER 10
APPLICATIONS OF CONTINUOUS MONITORS
10.1 Introduction
A continuous source monitor can provide both industry and regulatory agencies with
numerous tangible benefits. A properly installed and operated continuous monitoring
system can yield a large amount of data on source emissions. This information is beneficial,
since it establishes a reliable foundation upon which important decisions can be made.
The following sections illustrate and explain the various advantages of accurate continuous
monitoring data for both the plant and the agency.
10.2 Advantages of Monitoring Data to the Source
A well-designed and maintained continuous monitoring system supplies the source operator
with valuable information. It provides data on the operation of the industrial process
in addition to data on the process emissions to the atmosphere. It can be viewed as a
money-saving process control tool for evaluating source operating efficiency. The monitor
allows the source operator to assess process variables so that emissions can be held to a
minimum, maximizing product output with reduced fuel consumption. A good example
would be the use of a sulfur dioxide monitor at a sulfuric acid plant. The analyzer
would give the operator a continuous record of SO2 lost to the atmosphere. The plant
engineer then would be able to use this information to regulate the process and to evaluate
process efficiency with a possible reduction in fuel usage or raw material loss.
The continuous monitoring system also can be used to determine maintenance needs for
process and emissions control equipment. The monitor data can be useful in determining
the operating efficiency of key process equipment and emissions control systems. The
maintenance necessary for these pieces then can be performed according to indicated need
rather than a rigid periodic schedule. This could extend the operating life of some equipment
or indicate the necessity of maintenance for equipment that is not operating properly under
the routine schedule. The upkeep of process and control equipment can be done, there-
fore, on a more cost-effective basis. For example, opacity monitors have been used to
diagnose and tune electrostatic precipitators, and continuous SOa monitors have been used
to tune flue gas desulfurization processes.
A continuous monitor also is applicable to process and control equipment design and
evaluation. The continuous analyzer gives the most accurate data on source operation
and emissions available. This information is a valuable factor in evaluating the performance
of emission control equipment at the source. Analyzer data yield a record of control
equipment efficiency over a much longer period of time than a manual stack test. The
data are truer indicators of the equipment's effectiveness in reducing emissions. The user
would be able to evaluate manufacturer contractual obligations thoroughly for equipment
10-1
-------
performance and perhaps avoid the many past problems that have arisen. Designers
would be able to use these data to improve control and process equipment design. Monitor
data analysis also may help a source operator reach optimum conditions for all source
operations. The potential benefits derived from using monitor data for these purposes
have not yet been fully evaluated.
The plant's managerial staff could find that an unanticipated benefit from a continuous
monitoring system is a reduction in problems associated with meeting emission regulations.
The monitor could give a valid record of source emissions that would aid in making
decisions for source regulatory compliance. This would help relieve the ambiguous situations
sometimes created by manual stack testing. It also could act as a safeguard in refuting
false charges of high emissions from a regulatory agency or local citizens (see Ref. 2 of
Section 2.4). These points alone would be sufficient justification for monitor installation
at a source.
10.3 Advantages of Monitoring Data for the Regulatory Agency
The continuous monitor data gathered from sources greatly improve the enforcement
abilities of a regulatory agency. A continuous monitoring system can provide a full-time
data base for use by the agency in evaluating source compliance with regulations. These
data reflect the actual operating conditions at the source and its emissions to the atmosphere.
Continuous data, therefore, yield a much more accurate assessment of source compliance
than the information generated during short-term manual stack testing at optimum source
conditions. The agency is then able to develop more realistic, fairer regulations on
emissions to the atmosphere. This situation ultimately would benefit the industrial source
and the agency.
The regulatory agency can use the data to assist several other agency functions, for example:
• As a screening tool to identify the need for source inspections, visible emissions
observations, etc.
• As a valuable source of information on nuisance conditions i
in an area.
• To pinpoint the nuisance and prove that a nuisance condition exists or to settle
disputes between industry and community factions on the origin of the nuisance.
• For developing regional emission control strategies that evaluate the short-term
impact of source emissions on air quality.
• To augment visible emissions observations.
• To provide input to episode control plans, improving episode control capabilities.
10-2
-------
• To improve agency ability to relate effects of source emissions to ambient air
quality with dispersion modeling.
10.4 Continuous Monitoring: Aid to Manual Source Sampling
A continuous monitoring system gives the plant and the agency a useful measure of
manual stack test validity. The compliance tests require that data be obtained using
manual reference methods. These methods may be subject to error from sampling mistakes
or source fluctuations. A continuous emissions monitoring system can serve as a cross-
check of manual reference method data. These two sets of data on source emissions can:
• Confirm validity of sampling results
• Point out possible errors in either manual or instrument measurements
• Help locate where errors may have taken place
• Assist in correcting problems in manual or instrument measurements
The continuous monitor is required for new sources; therefore, it can serve the above
purposes with no additional cost.
10.5 Bibliography
Adams, D. F., and Koppe, R. K., "Direct GLC Coulometric Analysis of Kraft Mill Gases,"
Journal of the Air Pollution Control Association, V. 17:161, 1967.
Anson, D., et al., "Carbon Monoxide as a Combustion Control Parameter," Combustion,
March 1972, pp. 17-20.
Barden, J. D., and Lucero, D. P., "Monitoring Industrial Sulfur Scrubbers by Flame
Photometry," in Power Generation, Ann Arbor Science Publisher, Inc., Ann Arbor, Michigan,
1976, pp. 247-259.
Cheney, J. L., Fortune, C, Homolya, J., and Barnes, H. M., "The Application of an
Acid Dewpoint Meter for the Measurement of Sulfur Trioxide/Sulfuric Acid Emissions,"
Energy and the Environment, Proceedings of the 4th National Conference, AICHE, Dayton,
Ohio, 1976, pp. 507-511.
DeSourza, T. L. C., Lane, D. C., and Bhatia, S. P., "Analyzing Sulphur Compounds
in Draft Furnace Stack Gases," TAPPI, V. 76, No. 6:73-76, June 1975.
Dieck, R. H., "Gas Turbine Emission Measurement Instrument Calibration/ Calibration in
Air Monitoring, ASTM Tech. Pub. 598, ASTM, Philadelphia, Pennsylvania, 1976, pp. 16-39.
10-3
-------
Elliot, T. C, "Monitoring Boiler Stack Gases," Power, April 1975, pp. 92-94.
Flashberg, L. S., Johnson, E. S., and Bambeck, R. J., "Automatically Correlating Flue
Gas Measurements Using Carbon Dioxide," Paper No. 75-60.5, 68th Annual Meeting of
Air Pollution Control Association, Boston, Massachusetts, June 15-20, 1975.
Frenkel, D., "Tuning Electrostatic Precipitators," Chem. Eng,, V. 85, No. 14:105-108, 1978.
Fuller, W. F., Apple, G. D., Reader, J. R., Jr., and Seago, J. L., "Combine Measurement
of SO2, NOX, and O2 for Power Plant Applications," ISA AID, 1975.
Gilbert, L. R, "Precise Combustion-Control Saves Fuel and Power," Chemical Engineering
June 21, 1976, p. 145.
Hays, L., Resnick, L., and Wakeman, J., "Cost-Benefit Criteria for Gaseous Emission
Monitors," Proceedings, Continuous Monitoring of Stationary Air Pollution Sources,
APCA Specialty Conference, APCA, 1975, pp. 64-82.
Hougen, J. O., "Boiler Control System Design," CEP, June 1978, pp. 83-85.
Hyatt, J. R., and Wood, G. M., "On-Line Monitoring of Stack SO2 and Paniculate
Emissions," SRI-Proceedings of Workshop on Sampling, Analysis and Monitoring of Stack
Emissions, NTIS PB-252-748, April 1976.
Jahnke, J. A., Cheney, J. L., Rollins, R., and Fortune, C. R., "A Research Study of
Gaseous Emissions from a Municipal Incinerator," Journal of the Air Pollution Control
Association, V. 27, No. 8, August 1977.
Karels, G. G., et al., "Use of Real-time Continuous Monitors in Source Testing," presented
at APCA Annual Meeting, June 15-20, 1975, Paper 75-19.5, NTIS PB-230^934/AS GPO.
Lang, R. S., Saltzman, R. S., and DeHaas, G. G.s "Monitoring Volatile Sulfur Compounds
in Kraft Sulfite Mills," TAPPI, V. 58, No. 10:88-93, October 1975.
Lieberman, N., "Instrumenting a Plant to Run Smoothly - Looking at Instrumentation
from the Operators Standpoint," Chemical Engineering, September 12, 1977, pp. 140-154.
Lillis, E. J., and Schueneman, J. J., "Continuous Emission Monitoring: Objectives and
Requirements," Journal of the Air Pollution Control Association, V. 25, No. 8:804-809,
August 1975.
Lord, H., "CO2 Measurements can Correct for Stack-gas Dilution," Chemical Engineering,
January 31, 1977, pp. 95-97.
-------
McShane, W. P., and Bulba, E., "Automatic Stack Monitoring of a Basic Oxygen Furnace,"
Paper 67-120, 60th Annual Meeting of the Air Pollution Control Association, Cleveland,
Ohio, June 1967.
Manka, D. P., "Automatic Analysis of Sulfur Compounds and Hydrogen Cyanide in Coke
Oven Gas," Adv. Instrum., V. 29, No. 3:701, 1974.
Monroe, E. S., "Equations for Determining Excess Air from Oxygen Analysis of Combustion
Gases," J. Inst. Fuel, March 1972, pp. 167-169.
NCASI, "The Relationship of Paniculate Concentration and Observed Plume Characteristics
at Kraft Recovery Furnaces and Lime Kilns," Nat. Council for Air & Stream Improvement,
Tech. Bulletin No. 82, March 1976.
NCASI, "Application of Light Transmissometry and Indicating Sodium Ion Measurement
to Continuous Paniculate Monitoring in the Pulp and Paper Industry," NCASI Tech.
Bulletin No. 79, Mary 1975.
OTCeefe, "How Much Stack Instrumentation for Industrial Boilers?;' Power, August 1977,
pp. 112-113.
Paules, P. E., Holstrom, G. B., and Lord, H. C, "Instantaneous, Continuous, Directly
On-Stream Boiler Flue Gas Analysis," published in Instrumentation in the Power Industry,
V. 14, Instrument Society of America, Pittsburgh, Pennsylvania, 1971, p. 10.
Saltzman, R. S., and Hunt, E. B., Jr., "A Photometric Analyzer System for Monitoring
and Control of the (HaS) (SO2) Ratio in Sulphur Recovery Plants," presented at the
18th Annual ISA Analysis Instrumentation Symposium, San Francisco, California, May 3-6,
1972.
Saltz, J., "Baghouse Performance Monitor by Opacity Particulate Measuring Techniques,"
Proceedings of the 8th Annual Industrial Air Pollution/Contamination Control Seminar,
King of Prussia, Pennsylvania, April 1978, pp. 7.1-7.8.
Sarli, V. J., Eiler, D. C., and Marshall, R. L., "Effects of Operating Variables on Gaseous
Emission," Air Pollution Measurement Accuracy as it Relates to Regulation Compliance,
APCA Specialty Conference, APCA, 1976.
Shaw, J. T., "Progress Review No. 64: A Commentary on the Formation, Incidence,
Measurement and Control of Nitrogen Oxides in Flue Gas," J. of the Inst. of Fuel,
April 1973, pp. 170-178.
Schen, T. T., "Online Instruments Expedite Emissions Test," Chem. Eng., V., 82:109,
May 26, 1975.
10-5
-------
Thoen, G. N., DeHaas, G. G., and Austin, R. R., "Instrumentation for Quantitative
Measurement of Sulfur Compounds in Kraft Gases," TAPPI, V. 51:246, 1968.
Villalobos, R., and Houser, E. A., "On-line Analysis of SO2 in Flue Gas and Other
Process Streams by Gas Chromatography," Air Quality Instrumentation, V. 2:28-41, Instru-
ment Society of America, Pittsburgh, Pennsylvania, 1974.
Wheaton, W. L., "Field Evaluation of a Prototype H2S Monitor for Refinery Fuel Gas
Lines," Paper 77-27.6 presented at the 70th Annual Meeting of APCA, Toronto, Canada
June 20-24, 1977.
Winiski, J. W., "Continuous Sulfur Dioxide Measurement at Kamloops," Pulp and Paper
Canada, V. 77, No. 5:57-58, May 1976,
Zegel, W. C, "The Value of Continuous Monitoring to the User," Proceedings, Continuous
Monitoring of Stationary Air Pollution Sources, APCA Specialty Conference APCA
1975, pp. 199-204.
10-6
-------
CHAPTER II
THE PERFORMANCE SPECIFICATION TESTS*
11.1 Introduction
Continuous monitoring instruments installed at an affected facility must pass the Performance
Specification Test requirements given in Part 60, Appendix B, of the Code of Federal
Regulations. These tests evaluate the performance characteristics of opacity, SO2,
NOX, and O2 or CO2 continuous monitors. This discussion presents the major considerations
involved in carrying out the specification tests and the methods of data calculation.
Appendix B of 40 CFR 60 is reproduced in Appendix D of this handbook. Some
practical suggestions and comments for performing the specification tests also are included
to aid in planning and conducting such a test.
11.2 Performance Specification Test 1 - Transmissometer Systems
11.2.1 General
Transmissometers installed at an affected facility must meet design and specification require-
ments given in Performance Specification Test I. Instruments installed prior to the proposal
date of the CFR subpart addressing the source category may be requested by the EPA
Administrator to demonstrate acceptability for continuous monitoring applications.
11.2.2 Transmissometer Design Criteria
Performance Specification Test 1 establishes required EPA instrument design criteria for
transmissometers. These have been discussed in Chapter 4 of this handbook.
If measurements of the design specifications are made by personnel at the facility, all
measurements must be recorded and reported to the Administrator. The projection and
view tests need not be performed by the source operator if the analyzer and optics are
certified by the manufacturer to conform to design specifications. Results of manufacturer
tests must be reported to the Administrator by the source operator.
The transmissometer specification test includes design criteria for instrument calibration error.
The calibration error tests in Performance Specification Test 1 are especially important,
since there is no practical manner in which the relative accuracy of a transmissometer can
be determined after it has been installed. The .other Performance Specification Tests check
the relative accuracy of monitoring instruments against the values of reference methods
3, 6, or 7 obtained by manually sampling the stack gas. The opacity reference method 9
The Performance Specification Tests are currently undergoing revision by EPA. Further
information on these revisions may be obtained by referring to Stack Sampling News,
Volume 7, No. 2, February 1979, pp. 2-3 and references therein.
-------
is not used to check relative accuracy of transmissometer instruments. The instrument
calibration error test then becomes very important.
The transmissometer calibration error test may be performed by the manufacturer prior to
shipment or by the source before the instrument is installed on the stack. Proper notification
of this test must be given to the agency in either case. The response time of the instrument
also is determined during the calibration error test. The instrument must be set up for
the monitor pathlength to be used on the stack. AH of the manufacturer's written instructions
for initial operation and calibration are to be performed. The calibration test is then
performed using three neutral density filters corresponding to low-, mid-, and high-span
range as specified for the facility in the Part 60 subparts. The table of filter opacity
and optical density required- for a given span range as found in the Federal Register is
given in Table II-I.
TABLE IM
NEUTRAL DENSITY FILTERS FOR TRANSMISSOMETER CALIBRATION ERROR
Span Value*
(percent opacity)
50
fin
7f*
ftn
on
inn
Calibrated Filter Optical Densities with
Equivalent Opacity in Parentheses
Low-
range
0.1 (20)
O.I (20)
0.1- (20)
0.1 (20)
0.1 (20)
O.I (20)
Mid-
range
0.2 (37)
0.2 (37)
0.3 (50)
0.3 (50)
0.4 (60)
0.4 (60)
High-
range
0.3 (50)
0.3 (50)
0.4 (60)
0.6 (75)
0.7 (80)
0.9 (87-1/2)
•That span value given in the Subpart of Part 60, Title 40 of the CFR for a specific
source category.
It should be noted that the table is written around a single-pass instrument and should
be interpreted in terms of stack exit opacity. The filters used in the test must be certified
by the manufacturer to be within ±3 percent of the recorded filter value. It is recommended
that all filters be checked for true opacity on a well-collimated photopic transmissometer
(view and projection angle of I degree). The most stable filters are made of stainless
steel wire mesh. These supply a true neutral density without decomposition. The glass
or gelatin filters can decompose, changing opacity value in addition to permitting the
transmittance of infrared radiation. All filters used in the calibration error test must be
large enough to block the entire optical volume of the transmissometer.
11-2
-------
Instrument calibration error is determined by inserting each of the three filters (low-,
mid-, and high-range) into the middle of the optical path of the transmissometer. A total
of five nonconsecutive readings for each filter is made with data recorded in percent
opacity. The calibration error for each of the 15 readings is calculated:
Transmissometer opacity reading - Known filter opacity reading = xi difference (+ or -)
. i = 1 to 15
The data gathered from this test will be used in calculating the sum of the absolute mean
difference and the confidence interval for each of the three filters. A complete explanation
of these calculations will be given in the performance specification calculation section.
The instrument response time tests are made at this time using the high-range neutral
density filter. After proper instrument zero and span have been completed, the high-range
filter is inserted in the instrument optical path. The time needed for the instrument to
reach 95 percent of the filter range value is recorded, and the reading is allowed to stabilize
at the full filter span value. The high-range filter then is removed from the optical path
and the time required for the instrument to come to within 95 percent of zero is recorded.
This procedure is repeated for a total of five sets of span and zero readings. The mean
response time is calculated:
Sum of the 5 upscale and 5 downscale times
TT = mean response time
The mean response time must be no greater than 10 seconds.
11.2.3 Performance Specification Test 1
The Performance Specification Test evaluates the monitoring system location on the stack,
the operating characteristics, and the data recording ability. The performance specifications
for a continuous opacity monitor are given in Table 11-2.
TABLE 11-2
OPACITY MONITORS PERFORMANCE SPECIFICATIONS
Calibration error ^3% opacity*
Zero drift (24 hr) <2% opacity*
Calibration on drift (24 hr) <2% opacity*
Response time 10 sec (maximum)
Operation test period 168 hr
*Expressed as the sum of the absolute value and the 95-percent confidence interval.
11-3
-------
Instrument response time and calibration error for an opacity monitor are determined
before stack installation. The other specifications are evaluated while the instrument is
operating on the stack. Proper installation of the opacity monitor must be made following
Federal Register guidelines before the drift tests are performed.
An opacity monitor must be located across a duct or stack section that provides a particulate
matter flow through the instrument optical volume representative of the particulate matter
flow through the entire duct. The transmissometer must be downstream of all particulate
control devices and as far as possible from bends or obstructions in the duct. The most
suitable location would be at least 8 duct diameters downstream of any flow disturbance.
A transmissometer located after a bend is to be located in the plane defined by the bend
(when possible). The instrument pathlength should include the entire duct diameter. A
shorter monitor pathlength requires extra caution in locating the instrument for repre-
sentative readings. The Administrator may require that the owner or operator of the opacity
monitor demonstrate that the instrument location provides representative opacity readings.
The monitor optical and zero alignments are checked after the monitor is installed at the
selected location. The optical alignment of the transmissometer light source and reflector or
photodetector (for single-pass instruments) should be checked following the manufacturer's
instructions. Instrument zero alignment is determined with a clean stack after the monitor
mounting is mechanically stable (i.e., no duct thermal expansion or contraction). The
internal instrument zero is then balanced to coincide with the actual zero check performed
across the clean stack. These optical and zero alignment checks must be performed once
a year.
Final instrument alignment is made after the facility has returned to normal operation.
The optical alignment is rechecked as specified by the manufacturer. If the alignment has
shifted, it must be readjusted. Any shifts in opacity measurements attributable to realignment
are to be recorded and reported to the Administrator. A situation in which alignment
shift can occur may not produce significant shifts in opacity measurements if the plant
operating parameters change within a constant and adequately narrow range. The Adminis-
trator may require the transmissometer to be moved or mounted on more stable structures
if these problems become significant.
In the Performance Specification Test, the instrument is conditioned for 168 hours after
installation. During the conditioning period, the instrument is operated in a normal
manner. Necessary repairs and adjustments are allowed so that the instrument may be
brought into proper operation.
11.2.4 Zero and Calibration Drift Tests
The zero and calibration drift tests are conducted during the 168-hour operational test
period. The opacity monitoring system is operated continuously for 168 hours after the
completion of the conditioning period. The instrument must monitor the effluent at all
1-4
-------
times during the operational test period except when it is being zeroed and calibrated.
The instrument electronic components, including light source and photodetector, simulated
zero, and up-scale opacity calibration, must be checked at 24-hour intervals. The trans-
missometer must be zeroed and calibrated on a daily basis. The zero setting (on the recorder
only) must be offset 10 percent to account for possible negative drift. All exposed optical
surfaces must be cleaned and optical alignment must be checked each 24 hours. Note
that this conflicts with the requirement of Paragraph 9.2.8 of Performance Specification 1,
which disallows cleaning during the specification test if it is not part of normal operating
procedure. (A decision will most likely be required for this situation by the agency
representative observing the test.) Manufacturer recommendations may be followed for
these procedures provided they meet or exceed the Federal specifications. Automatic
instrument corrections (no operator intervention) are allowed at any time.
The data recorded at each 24-hour interval includes the zero and upscale calibration
readings after system calibration (set at the same values each 24-hour interval). The zero
reading is recorded after 24 hours of instrument operation but before cleaning or adjustment
of the instrument zero. The instrument calibration reading is recorded after cleaning the
optical surfaces and resetting the zero adjustment but before resetting the calibration
adjustment. These data are used in calculating individual differences (xi) used in the
calculation of the 95-percent confidence interval and absolute mean value.* The zero
drift xj values are calculated:
(Zero readings after \ / Zero reading before \ _
cleaning optics 1 Icleaning optics 24 hours later/ l
Calibration drift xj values are:
/Calibration reading after\ /Calibration reading after cleaning\
cleaning optics and - °Ptics and 2ero adjustment (+ or _,
\ zero adjustment / I but before calibration I
* / \ adjustment 24 hours later /
/
The sum of the absolute mean value and the 95-percent confidence interval is computed
separately for zero drift and calibration drift, then reported. The absolute mean value
is found:
•42
Xi
*See Appendix B for a further discussion of the statistical methods used in the Performance
Specification Tests.
11-5
-------
where:
n = number of data points
xi = difference (+ or -)
x = absolute mean value
and the 95-percent confidence interval is:
C.I.95 =
n vn-
where:
n = number of data points
xi = difference (+ or -)
to.975 = the t value derived in the t test corresponding to the probability that a
measured value will be within 95 percent of the true value
C.I.95 = 95-percent confidence interval estimate of the mean value
TABLE 11-3
VALUES FOR to.975
n to.975
1 Z. /UO
Z.JOD
2.JOG
n to.975
10 2.262
1 2.228
12 2.201
1J — 2.179
2.160
lj — 2.145
16 • 2.131
11-6
-------
For example, the zero drift is computed by taking the zero reading differences (xj) and
adding them together according to their sign value (see Table 11-4).
TABLE 11-4
24-HOUR TRANSMISSOMETER ZERO DRIFT DATA
24-Hour Interval
1
2
3
4
5
6
7
Sxi
xi Value
-1
0
-2
+1
0
+2
-I
-1
Xi2
1
0
4
1
0
4
I
11
Then using the 2x; in the equation for the absolute mean value:
NK^=THO.'43
The confidence interval is then calculated:
C.I.95 =
tp.975
2.447
V? (11) - (1) = 1-244
The sum of the absolute mean value and the 95-percent confidence interval is reported
as zero drift:
|x| + C.1.95 = 0.143 + 1.244 = 1.387 zero drift
The preceding calculations also are made for the calibration error tests (each of the three
filters used) and calibration drift using the xj values found in each test. The statistical
basis for these calculations is given in Appendix B.
11-7
-------
11.3 Performance Specification Test 2 - SO2/NOX Systems
11.3.1 General
Performance Specification Test 2 gives the installation requirements, test procedures, and
data calculation methods for evaluating the acceptability of a continuous SO2/NOX
monitoring system. This performance specification procedure does not prescribe specific
instrument design criteria. However, reference methods 6 and 7 are conducted concurrently
with the monitor evaluation to determine the monitor's relative accuracy.
Performance Specification Test 2 requires the advance preparation of instrument calibration
gas mixtures and the reference method test equipment. The calibration gases used during
the Performance Specification Test are:
• Zero gas certified by the manufacturer to contain less than 1 ppm of the pollutant
gas to be measured. Ambient air may be used as zero gas (depending on instrument
requirements).
• Sulfur dioxide (SO2) gas mixtures may be in air or nitrogen.
• Nitrogen dioxide (NO2) gas mixtures must be in air.
• Nitric oxide (NO) gas mixtures must be in an oxygen free (<10 ppm) inert gas.
The two gas concentrations for the performance test must be approximately 50 percent and
90 percent of the span values as given in the subparts. A listing of calibration gas concen-
tratiojis given in the subparts for a fossil-fuel steam generating facility is shown in Table 11-5.
Refer to the subparts for calibration values at other types of facilities.
All calibration gas concentrations must be checked by triplicate reference method gas
analysis. The three test analyses results are averaged. Individual tests must agree within
±20 percent of the average gas concentration or the analysis must be repeated. The
analyses are to be performed on the calibration gas no more than 2 weeks prior to use
in the specification tests.
The recommended equipment for reference method testing of SO2/NOX stack gas concen-
tration is:
• Method 6 - Reference Method for SO2 (Reference Method 8 is used to measure
SO2 in Sulfuric Acid Plants)
— Calibrated control console — pump, flow meter, and dry gas meter
— Three heated gas sampling probes
— Data sheets
11-8
-------
TABLE 11-5
SPAN AND CALIBRATION GAS VALUES
Gas
Monitored
S02
NOX
02
C02
Type of
Fossil Fuel
Liquid
Solid
Gas
Liquid
Solid
"
Span
Value
1000 ppm
1500
500 ppm
500
1000
1.5-2.5
normal source
cone.
1.5-2.5
normal source
cone.
Calibratior
50% of
Span Value
500 ppm
750
250 ppm
250
500
0.50 Span
Value
6,50 Span
Value
Gas Value
90% of
Span Value
900 ppm
1350
450 ppm
450
900
0.90 Span
Value
Or ambient
if span >21%
0.90 Span
Value
— Minimum of 11 midget impinger sampling trains
— Fresh reagents
— 80 percent isopropyl alcohol (tested for oxidant contamination)
— 3 percent hydrogen peroxide (H2O2)
• Method 7 - Reference Method for NOX
- Minimum 30 gas sampling flasks (2 liter)
- Three heated gas sampling probes
- Pressure gage (0 to 30-inch Hg)
- Pump (capable of >30-inch Hg suction)
— Data sheets
Performance Specification Test procedures do not specifically state that the reference methods
be conducted for a source compliance test. It is recommended however, that reference
methods rather than equivalent methods be used whenever possible.
11.3.2 Monitor Location and Installation
The SO2/NOX continuous gas monitor must be located at a sampling point where measure-
ments can be made which are directly representative or can be corrected to be representative
11-9
-------
of the total emissions. The Federal Register establishes requirements for the location and
installation of extractive and in-situ-type monitors at a source. The requirements are:
• Stack gases may be assumed to be nonstratified at any point greater than 8 duct
diameters downstream of air in-leakage.
• This assumption on stratification may not be made for sampling locations upstream
of an air preheater at a steam-generating facility.
• For sampling points located where the gas is assumed or demonstrated to be
nonstratified, extractive or in-situ monitors may sample at a point of average
concentration.
• Extractive sampling points must be no closer than 1 meter to the stack wall.
• Multipoint extractive sampling probes may be located at any points necessary to
obtain consistently representative gas samples.
• Sampling locations at which gases are stratified must employ extractive sampling
systems or in-situ sampling locations that obtain results that are consistently
representative or can be corrected to be representative of the total emissions
from the affected facility.
• The extractive type of system may accomplish this requirement by using a
multipoint sampling probe. The in-situ monitor must be located so that its
optical path will view a representative gas sample.
• It must be demonstrated that sampling at stratified gas locations gives consistent
readings for several plant operating conditions (i.e., points of average concen-
tration do not shift with operating changes).
• Pollutant and diluent gas monitoring systems should be of the same type — both
extractive or in-situ. If the systems are of different types, the extractive system
must use a multipoint sampling probe.
• Temperature, velocity, and gas concentration traverses of the stack gas may help
to characterize gas stratification. If no stratification is shown at a point less than
8 duct diameters downstream of air in-leakage, procedures for sampling non-
stratified gas may be used.
11.3.3 Specification Test Procedures
The test procedures given in Performance Specification Test 2 are designed to evaluate
the continuous monitor operating performance. The required instrument performance
specifications are given in Table 11-6.
11-10
-------
PERFORMANCE SPECIFICATIONS FOR SO2/NOx SYSTEMS
<20% of the mean value of the reference
of each (50% of span, 90% of span)
calibration gas mixture or internal calibra-
tion cell value
2% of span
2% of span
2% of span
2.5% of span
15 min maximum
168 hr minimum
Zero drift (2 hr)*
Zero drift (24 hr)*
Calibration drift (2 hr)*
Calibration drift (24 hr)*
Response time
Operational period
•Expressed as sum of absolute mean value plus 95-percent confidence interval of a series
of tests.
An i
mcmmng in. extractive- and in-situ-type monitors. The
caHbratJadjustmcnt a, 24-hour interval and automat.c ms.rument
operations.
1134 Calibration Error Test Procedures
11-11
-------
n
a means of troubleshooting, he em iddtionn- ^ at "" faci"ty Provides
field operations. t0 g'Vmg a truer evaluation of instrument
itioning
manufacturer's wri.. lral0n "" P6"^ foUowi»g the
Average calibration gas
concentration (ppm)
X 100 = Calibration Error
Example 50-percent calibration gas:
Reference Method Average Concentration = 550 ppm
Calibration Error Test Data (extracted from table of 15 readings)
Difference (\j)
Instrument Reading
Minus Calibration Gas Value
Reading
No.
^•^^^—^^•^^^•^^^
2
5
7
8
12
•" —
Calibration Gas
Concentration (ppm)
— ••—
550
550
550
550
550
— ~ 1
Instrument
Readings (ppm)
558
560
562
547
548
+8
+ 10
+ 12
-3
-2
64
100
144
9
4
11-12
-------
The absolute mean value x is calculated:
n
i\=±
i=\
|x| =j(+25) = |5|
The 95-percent Confidence Interval (C.I.95) is calculated:
C.1.95
2.776
CI-95 = e /rS (321) - (625) = 8.690
The calibration error is:
|x| + C.1.95
[SO2] ppm
5 + 8.690
x 100
550
x 100 = 2.489%
11.3.5 Response Time Test
The monitor response-time test and all other specification test procedures are conducted
during the 168-hour operational test period. The response-time test evaluates the response
time of the entire continuous monitoring system as installed. All system parameters, such
as gas flowrate, sample line size, pump rates, and pressures, must be operated at normal
system settings as given in the manufacturer's written instructions. Calibration gas injection
must not change the normal system operating pressure.
The response-time test measures the time it takes the monitor to reach 95 percent of the
final stable response when calibration gas and zero gas are quickly switched into the
system. The test is carried out by injecting .zero gas into the sampling interface allowing
the monitor to reach a stable reading. Calibration gas of known concentration is then
quickly switched into the system (50 percent or 90 percent gases may be used). (The
90-percent gas is recommended over the 50-percent gas for response-time test.) The time
from injection to the time the instrument shows 95 percent of final stable response is
recorded. Zero gas is reinjected into the sampling interface after the upscale reading has
stabilized. The time from zero-gas injection to the time to reach 95 percent of final
11-13
-------
stable instrument response is recorded. The entire procedure is repeated three times. An
in-situ, nonextractive monitor would perform these functions by inserting the highest
calibration gas cell available. The average of the three upscale readings and average of
the three downscale readings are taken with response time reported as the slower time.
The upscale and downscale times must not differ more than 15 percent of the slower time:
% deviation from
slower system
average response
11.3.6 Field Relative Accuracy Test
/average upscale \ /average downscaleN
V response time J \ response time J
(slower time)
X 100
The relative accuracy of the continuous monitoring system is determined by comparing
the instrument pollutant concentration measurements to the manual reference method
analysis of the pollutant concentration in the stack gas. The Performance Specification
Test 2 regulations require that nine consecutive sets of reference method data be taken with
no more than one data set per hour. The sulfur dioxide data would then consist of
nine applicable reference method tests: one reference method performed each hour for
9 hours. The oxides of nitrogen tests require a total of 27 NOx reference method tests,
which are divided into nine sets, three tests per set. The three individual NOx reference
method tests are to be performed concurrently or within a 3-minute interval; no more than
one set (three tests) are to be made in 1 hour. The results of the three tests are averaged
and the average is used in the calculations. The probe tip for the reference method tests
described must be as close to the sampling location (extractive or in-situ) of the monitor
as possible. The analyzer must continuously monitor stack gas pollutant concentrations
during reference method testing. The average analyzer pollutant concentration measurements
for each test period are determined by integrating or averaging the monitor data for each
period. All data for the reference methods and the continuous monitor must be given on a
consistent base (wet or dry). A moisture correction factor must be applied to data that
are not on a consistent base. The moisture correction factor'is determined by running
concurrent reference method 4 tests with the other reference method tests. Reference
method data are converted to parts per million in the following manner:
Dry Basis at Standard Conditions:
gm/dm3 x 24.06 x IP3
PPm — molecular weight (gm)
Wet Basis at Standard Conditions:
gm/dm3 x 24.06 x 103
ppm~ molecular weight (gm) (NBws)
where Bws is the moisture fraction of the stack gas.
11-14
-------
The difference between each test period instrument pollutant concentration (ppm) and
the reference method concentration (ppm) is used in calculating the sum of the absolute
mean value and 95-percent confidence interval for each pollutant (NOX and SO2).
/instrument test period\ /reference method\ ,, ,
-1 v 1 = xi (+ or -)
^ average (ppm) j ^ (ppm) J
The sum of the absolute mean value and the confidence interval divided by the average
reference method gas concentration gives the relative accuracy of the instrument:
x + C.I.95
average reference method (ppm)
X 100 = % relative accuracy
The following example illustrates the calculation of relative accuracy (all data have been
expressed in ppm on a dry basis).
CALCULATION OF RELATIVE ACCURACY
Test
No.
1
2
3
4
5
6
7
8
9
SO2
Sample (ppm)
417
430
429
429
429
446
450
436
410
430.7
NOx Sample {ppm)
Average of 3 tests
614
572
624
614
723
709
696
699
758
667.7
Average reference
method values
Analyzer 1 hour
Average (ppm)
SO2
398
409
404
400
404
429
438
422
434
NOx
602
602
638
655
744
744
744
744
744
Difference (xi)
(ppm)
S02
-19
-21
-25
-29
-25
-17
-12
-14
+24
15.3
NOx
-12
+30
+14
+41
+21
+35
+48
+45
-14
23.1
mean difference
2
*1
SO2
36!
441
625
841
625
289
144
196
576
4098
NOx
144
900
196
1681
441
1225
2304
2025
196
9112
£xj2
11-15
-------
I. 95-percent Confidence Interval
A. S02
_ 2.306
CI-95 = 9 /£VV9 X (4098)-(15.3)2= 17.342
B. NOX
CI-95 = 9 /£r V9X (9112)-(23.1)2 = 25.857
II. Relative Accuracy
A. SO2
% Relative Accuracy = 100 x 15'3 + 17'342 = 7579
430.7
B. NOX
% Relative Accuracy = 100 x 23A + 25'857 = 7 332
7.579
11.3.7 Instrument Zero Drift and Calibration Drift - 2 Hours and 24 Hours
The instrument 2-hour and 24-hour zero and calibration drifts are evaluated during the
168-hour operational test period. The zero and calibration drift tests examine the analyzer
ability to hold its calibration over a period of time. The system drift is evaluated at
2-hour intervals for a total of 15 data sets and at 24-hour intervals for the duration of
the operational test period. The system zero and calibration may be adjusted only at
24-hour intervals (or at shorter periods if the manufacturer's written instructions require it).
The zero and calibration drifts are determined by injecting zero and calibration gases for
,an extractive monitoring system. The 24-hour zero and calibration drift are noted as part
of the 24-hour calibration procedure. The 2-hour drift requires 25 data sets. The Federal
Register does not require consecutive 2-hour readings to be done on an around-the-clock
basis; however, the data for the 2-hour zero and calibration drift should be collected
over a maximum of three 24-hour periods. It is recommended that the 15 data sets
be collected over the shortest possible time span. The 2-hour checks cannot overlap. All
data must be recorded as ppm.
The in-situ or nonextractive monitor may determine the zero and calibration drift by
producing a mechanical instrument zero and checking the calibration with a certified
11-16
-------
calibration gas cell. An alternative is to insert a series of three calibration gas cells into
the detector-radiation source path and calculate the zero point from the upscale measurements.
A graph must be kept as a record of the procedure used. The zero check using this
three-gas-cell method is a special method for cross-stack monitors. It requires a thorough
evaluation of the instrument linear response and a careful interpretation of the manufacturer's
written instructions. The in-situ instrument calibration may be checked using a certified
calibration gas cell equivalent to a 50-percent span concentration. All data are recorded
as ppm.
The 2-hour and 24-hour zero and calibration drifts are reported as the sum of the absolute
mean value and 95-percent confidence interval as a percent of the instrument span. The
difference values (xi) for the 2-hour zero drift are calculated from the consecutive readings:
zero
_. v \ /Zero readmgX
Time x \ / , \
,. 1 [ 2 hours later 1 ,, ,
reading - |_r_ ._._._ J = x; (+ or -)
(ppm) I
iafter injecting)
zero
gas /
(for zero drift)
The 2-hour calibration drift is corrected for corresponding zero drift. The \i values are
determined by subtracting the change in zero from the change in calibration:
A Span - A Zero = xi (+ or -) (for calibration drift)
The calculations for 2-hour drift are illustrated in the following example:
Instrument Span = 0 - 1000 ppm
Time
X
Zero Reading
0 ppm >
Calibration Reading
850 ppm
Time
X+2 Hours
Zero
Reading
+6 ppm
Zero Drift
A Zero (xi)
+6 ppm
Calibration
Reading
859 ppm
A Span
9 ppm
Calibration Drift
A Span ~ A Zero (xi)
9-6 = 3 ppm
This procedure would be carried out for all of the 15 data sets taken during the 2-hour
drift test. The xj values for zero drift and calibration drift are then used in computing
the sum of the absolute mean value and 95-percent confidence interval. The individual
zero and calibration drift are values then given by:
|x| + C.I.95
Instrument Span 1000 (ppm)
x 100 = Zero drift or Calibration drift
11-17
-------
The 24-hour zero and calibration drift xi values are calculated:
Zero drift
:Zero (ppm) *
after adjustment
Calibration drift
/Zero (ppm) 24 hours later
after zero gas injection but
I prior to zero adjustment / (zero drift)
Calibration (ppm) value
after zero and
calibration adjustment
Calibration (ppm) value
24 hours later after
zero adjustment but
iprior to calibration adjustment
The numerical operations are illustrated in the following example:
Instrument Span = 0 - 1000 ppm
DATA FOR 24-HOUR ZERO AND CALIBRATION
(calibration drift)
Data
Set
1
2
3
4
5
6
7
Time
X
x+24h
X2
x2+24h
X3
x3+24h
x4
X4+24h
X5
X5+24H
*6
x6+24h
X7
x?+24h
Zero
Reading
(ppm)
0
+5
0
-4
0
+6
0
-5
0
-5
0
-5
0
-6
Calibration
Reading
(ppm)
950
-
950
-
950
~
950
-
950
-
950
-
950
-
Zero Drift
(ppm)
A Zero (Xj)
-
+5
-
-4
-
+6
-
-5
-
-5
-
-5
-
-6
Sxi -14
Calibration Reading
(ppm)
After Zero Adjustment
-
959
_
943
-
944
-
948
-
947
-
943
-
945
Calibration Drift
(ppm)
A Span
-------
I. 24-Hour Zero Drift
A. Absolute Mean Value x
|x| =^=2.000
B. 95-Percent Confidence Interval
2447 i
C.I.95 = ? •/^jN/7(188) - (14)2 = 4.776
C. Drift as Percentage of Span (Ds)
100 X ZOW*4'776 = 0.678%
II. The same operations are performed for calibration drift xi values
A. Absolute Mean Value
|x| = 3.000
B. 95-Percent Confidence Interval
C.I.95 = 5.205
C. Drift as Percentage of Span
Ds = 0.821%
III. AH the above data and operations are also performed for the NOx monitor.
11.3.8 The Operational Test Period
The SO2/NOX continuous monitor must meet all of the specifications given in Performance
Specification Test 2. The monitor also must continuously analyze the stack gas pollutant
concentration during the 168-hour operational test period. During this time, no corrective
maintenance, repairs, replacements, or adjustments may be made to the monitor other than
those clearly specified as routine and expected in the manufacturer's written instructions.
The operational test period is successfully completed after all of the parameters have been
checked with the analyzer operating without corrective maintenance. If the analyzer fails
during the operational test period, the specification test must be repeated during another
168-hour test period. It is not necessary to repeat any test that successfully met speci-
fications during the first operational period.
11-19
-------
11.4 Performance Specification Test 3 — O2 or CO2 Monitors
11.4.1 Introduction
The O2 or CO2 analyzer performance requirements are given in Performance Specification
Test 3 (PST 3). The monitor response time, calibration, and zero and calibration drifts
are evaluated in the operational test period. The regulations have placed increased
emphasis on the location of the O2 or CO2 monitor when it is used to convert pollutant
concentration data to units of the standard. It is important to note restrictions on the
placement of CO2 gas analyzers after limestone scrubbers have been removed (see Federal
Register - January 31, 1977).
The performance specification requirements given in the Federal Register for an O2 or
CO2 monitor are reproduced in Table 11-7.
TABLE 11-7
PERFORMANCE SPECIFICATIONS FOR O2 OR CO2 MONITORS
Zero drift (2 hr)* ^0.4% O2 or CO2
Zero drift (24 hr)* <0.5% O2 or CO2
Calibration drift (2 hr)* <0.4% O2 or CO2
Calibration drift (24 hr)* ^0.5% O2 or CO2
Operational period 168 hr minimum
Response time 10 min maximum
•Expressed as sum of absolute mean value plus 95-percent con-
fidence interval of a series of tests.
The regulations do not require specific design criteria for O2 or CO2. There are a number
of different types of analyzers available commercially. The most frequently encountered
are paramagnetic or zirconium oxide cells for O2 analysis and nondispersive infrared
instruments for CO2 analysis.
11.4.2 Monitor Location and Installation
The O2 or CO2 monitor must be located and installed at a point that permits measure-
ments of the diluent gas concentration that are directly representative of the total effluent
emitted. An O2 or CO2 monitor that is used for converting data to units of the standard
is to be located at a point where the stack gas is nonstratified, with no air leaking in,
or at a position in which gas stratification has been characterized and the sampling interface
11-20
-------
has been designed to give representative data. A review of the location requirements
given for SO2/NOx monitors and transmissometers provides details of these requirements.
The O2 or CO2 monitor used in converting data to units of the standard must be located
at a sampling point where its measurements are representative of the effluent gases sampled
by the pollutant monitoring system. This requirement is best fulfilled by installing the
O2 or CO2 monitor at a location near the SO2/NOx monitor such that approximately
the same extractive point(s) or in-situ path is sampled for both monitor types. If the
pollutant monitor and O2 or CO2 monitor are located at different points on the duct,
the installation details given in Performance Specification Test 2 must be carefully followed.
It is recommended that the sampling locations be verified as representative and equivalent
by temperature, velocity, and pollutant traverses of the stack gas. A portable gas monitor
used to characterize pollutant concentration at traverse points across the duct in conjunction
with temperature and velocity traverses is considered essential for satisfying questions of
monitor location.
11.4.3 O2 or CO2 Monitor Calibration Gases
The Federal Register states that known concentrations of O2 or CO2 corresponding to
50 percent and 90 percent of instrument span (as given in the subparts of the CFR for
each affected facility) be used to calibrate the instrument. The manufacturer's instructions
should be followed for the type of inert carrier gas required for the O2 or CO2 and
instrument zero gas. If the O2 analyzer span range is greater than 21 percent O2, ambient
air may be used as the calibration gas. The calibration gas mixtures must be analyzed
by triplicate-reference-method 3 tests no more than 2 weeks prior to use in the speci-
fication test.
11.4.4 Instrument Calibration Check
The wide variation in analyzer designs may make procedures for this test and the following
tests of instrument response and drift inapplicable. The EPA Administrator must approve
any alternative procedures employed for the tests.
The calibration of the instrument is checked by establishing a calibration curve. Zero,
mid-range, and 90-percent span gases are injected into the analyzer. The data are plotted
as instrument response versus gas value. The graph must be consistent with the expected
response curve described by the manufacturer or additional measurements must be made to
verify the instrument accuracy. This test should be performed with the analyzer installed
as intended for use in the field.
The O2 or CO2 analyzer is operated for a 168-hour conditioning period after the calibration
check. The monitor may be maintained and repaired as needed to prepare it for the
168-hour operational test period. During the operational test period, the monitor must
11-21
-------
continuously monitor the stack gas without corrective maintenance or repair. Manu-
facturer's written instructions for routine procedures are permitted. Any other type of
repair or corrective maintenance required by the analyzer will cause the operational test
period to end. The operational test period must then be repeated. Performance speci-
fications successfully completed during the first operational test period do not require
repetition.
11.4.5 Response-Time Test
The response-time test is performed during the 168-hour operational test period with the
monitor installed as intended for use at the affected facility. The entire sampling interface
is included in the test with careful attention made to assure that system flowrates, line
diameters, pumping rates, and pressures are the same as in normal operating procedure.
The test for response time must be repeated for each sampling point - if the analyzer
is used to sample more than one source, the response time must be determined for
each system.
The system response time is determined by injecting zero gas into the sampling interface
(or as close as possible) to establish a stable output reading. A known concentration
of calibration gas at 90 percent of span is quickly switched into the system. The time
required for the instrument to reach 95 percent of the final stable response is recorded.
The system is allowed to stabilize at the upper span reading, then zero gas is reinjected
into the sampling interface. The time needed for the instrument to reach 95 percent of
the final stable zero is recorded. The procedure is carried out for three sets of upscale-
downscale tests. A nonextractive system is evaluated by switching the highest available
calibration gas concentration into and out of the sample path and by recording the
95-percent upscale and downscale response times.
The system response time is calculated from the time intervals required for 95 percent
of final stable response. The mean of the three upscale response times is found by:
upscale response
5 = mean upscale response time
and the mean of the three downscale times is calculated:
downscale response , ,
= mean downscale response time
11-22
-------
The upscale-downscale response should not deviate more than 15 percent of the slower time:
(mean upscale response time) - (mean downscale response time)
100 X - — < 15%
slower time
The system response time is reported as the slower time.
11.4.6 Zero and Calibration Drifts - 2-hour and 24-hour
The 2-hour and 24-hour zero and calibration drift tests are performed for the O2 and
CO2 monitor in the same manner as for the SO2/NOX analyzer. The 2-hour drifts are
determined from 15 sets of zero and calibration readings taken at 2-hour intervals. (Readings
need not be consecutive but must not overlap.) Zero gas is introduced into the system
and analyzer zero output is recorded. The change in readings between consecutive 2-hour
measurements is the xi values for zero drift calculations:
A Zero = [Time x zero reading (ppm)] - [Time x + 2-hr zero reading (ppm)] = x, (+ or -)
The calibration drift xi values are determined by injecting mid-range calibration gas
(for nonextractive monitors a calibration gas cell functionally equivalent to 50 percent of
span is used) at 2-hour intervals and correcting the calibration drift values for corresponding
zero drift:
A Span = [Time x calibration reading (ppm)] - [Time x + 2-hr calibration reading (ppm)]
A Span - A Zero - xi (+ or -) for calibration drift
The respective xi values for zero and calibration drifts then are used to calculate separately
the absolute mean difference and 95-percent confidence interval for the individual drifts.
The 2-hour zero drift or calibration drift is then expressed as the sum of the respective
absolute mean difference and 95-percent confidence interval:
'x| + C.I.95 = zero or calibration drift
The 24-hour zero and calibration drifts are determined by taking the difference between
instrument setting at calibration and the values 24 hours later.
24-hour zero drift
zero setting \ / zero setting 24 hours \
after adjustment - later before adjustment = difference (xi)
(ppm) (ppm) /
11-23
-------
24-hour calibration drift
calibration \
setting after I -
^adjustment (ppm)/
/calibration concentration^
24 hours later after zero 1 _..«
adjustment but before 1 ~ dlfference
\ calibration adjustment /
The sum of the absolute mean value and 95-percent confidence interval is then expressed
as the respective zero or calibration drift:
|x| + C.I.95 = 24-hour zero or calibration drift
11.5 Bibliography
Baladi, E., "Acquisition, Installation, Performance Testing, and Operation of a Continuous
Monitoring System of an Existing Steam Generator Stack," Paper 76-35.3 presented at
1976 APCA Meeting, Portland, Oregon.
Baladi, E., Midwest Research Institute, Manual Source Testing and Continuous Monitoring
Calibrations at the Lawrence Energy Center of Kansas Power and Light Company, EPA
Contract No. 68-02-0228, EPA Report No. 73-SPP-3, May 7, 1976.
Barnes, H. M., and Homolya, J. B., "Data Requirements for NOX Emission Monitoring
from Fossil-Fuel Fired Steam Generators," J. Environ. Sci. Health-Environ Sci Eng
All (2), 1976, pp. 107-119.
Bonam, W. L., and Fuller, W. F., "Certification Experience with Extractive Emission
Monitoring Systems," in Calibration in Air Monitoring ASTM Special Tech. Publication 598,
Proceedings of Symposium, August 1975.
Homolya, J. B., "Data Output Requirements for Monitoring SO2 Emissions from a
Stationary Source," Paper presented at Instrument Society of American Conference &
Exhibit, Houston, Texas, October 1973.
Howes, J. E., "Qualification of Source Test Methods as Reference Methods," Calibration
in Air Monitoring, ASTM Tech. Pub. 598, ASTM, Philadelphia, Pennsylvania, 1976
pp. 80-95.
Jacquot, R. D., and Houser, E. A., "Qualification Testing of an Infrared Analyzer System
for S02 and NOX in Power Plant Stack Gas," Paper 72-730, Proceedings of the 27th
Annual Conference and Exhibit of the ISA, October 9-12, 1972, New York, New York.
11-24
-------
Jaye, F., Steiner, J., and Larkin, R., "Resource Manual for Implementing the NSPS
Continuous Monitoring Regulations Manual 2 - Preliminary Activities for Continuous
Monitoring System Certification," EPA-340/I-78-005b, April 1978.
Kendall, D. R., and Bartok, R. H., "Evaluation of Continuous SO2 Source Monitoring
Systems via EPA Performance Specification Procedures," Journal of the Air Pollution
Control Association, V. 27, No. 9:872-879, September 1977.
Lord, H. C., "Verification of In-Situ Source Emission Analyzer Data," presented at
ASTM/NBS/EPA Symposium, Calibration in Monitoring, Boulder, Colorado, August 1975.
Lukacs, J., and Beamish, M. C., "Comparative Operating Data from Manual and Automatic
Source Emission Methods," Paper 75-60.3 presented at the 68th Annual Meeting of the
Air Pollution Control Association, Boston, Massachusetts, June 15-20, 1975.
Nader, J. S., et al, Performance Specifications for Stationary Source Monitoring Systems for
Gases and Visible Emissions, NERC Chemistry and Physic Lab. NTIS PB-209-190, January
1974, EPA-650/2-74-013.
Polhemus, C., and Hudson, A., "A Performance Analysis of Lear Siegler's In-Situ SO2/NO
Monitor," Paper No. 76-35.5, 69th Meeting of the Air Pollution Control Association,
Portland, Oregon, June 27-July 1, 1976.
11-25
-------
-------
CHAPTER 12
QUALITY ASSURANCE
12.1 Introduction
A continuous monitoring system will provide valid, reliable data when properly maintained
and operated. The investment involved in purchasing a continuous monitoring system
is only the first step in supplying emissions data from a source. The system must be
maintained at regular intervals to ensure that it is operating within prescribed limits.
A monitoring system that is not well maintained becomes a possible legal liability and is
not cost effective for the user. It is in the interest of both the user and the regulatory
agency to have the monitoring system provide good data on source emissions. This is
required by the Federal regulations; however, a user benefits from the data by avoiding
legal conflicts, protecting its monitoring investment, in addition to gaining a reliable process
control monitor.
The continuous monitors presently available on the commercial market will supply valid
data when an organized program of routine maintenance and quality assurance is carried
on. All monitors will need some maintenance. If the maintenance program is organized
following manufacturer's instructions at the time of installation, then reasonably followed
throughout the operational life of the analyzer, the monitor will provide useful emissions
information. A system of routine maintenance will prove much less costly or bothersome
than a neglected system that frequently breaks down. The legal problems and repair
headaches that follow using a poorly maintained system soon become much more trouble
than originally employing a good maintenance program. The following paragraphs deal with
quality assurance procedures for keeping a continuous monitor in good operating condition.
12.2 Calibration Gas Evaluation
The continuous monitoring system will yield valid data only after it has been properly
calibrated. An analyzer delivers an output signal proportioned to the pollutant concentration
in analysis. The actual concentration is obtained by calibrating the instrument against a
known concentration of pollutant. The significance of the instrument calibration for
obtaining data on pollutant emissions is obvious. The calibration of an analyzer is
performed with gases purchased from commercial suppliers.
A commercial calibration gas manufacturer product must always be checked to ensure
that the stated gas concentration in the cylinder is accurate. The gas concentration marked
on a cylinder should never be assumed to be correct. The cylinder concentration must
be checked using the reference method gas analysis, which is required for the Performance
Specification Test procedures. The regulation subparts must be read carefully to determine
the required calibration gas for an affected facility. This must correlate with the requirement
stated by the instrument manufacturer. The gas is then analyzed no more than 2 weeks
12-1
-------
prior to use in the Performance Specification Test by the applicable reference method.
A check also is required at 6-month intervals by a reference method test. The regulations
allow the cylinder to be checked at less frequent intervals if the shelf life is guaranteed
by the manufacturer to exceed a 6-month stability. It is recommended, however, that
all gases be analyzed every 6 months.
As mentioned earlier, EPA is developing a protocol for the use of National Bureau of
Standards (NBS) and NBS traceable gases and forgoing the reference method analyses.
A comment to this effect is made in 40 FR 46251 October 6, 1975; however, it has not yet
been incorporated into the body of the regulations.
Reference methods 3, 6, and 7 describe wet chemical analysis of carbon dioxide, oxygen,
sulfur dioxide, and oxides of nitrogen. These methods have shown good results when
properly performed. The details of the procedures are given in the August 18, 1977,
Federal Register. It is good practice to cross check new cylinders by using them in the
analyzer calibrated with the existing gases. The data from the already analyzed calibration
gases, new cylinder gases, and reference methods then may be correlated. This provides
a thorough check of all gases and may indicate problems not previously anticipated.
The calibration gases must be checked by the reference methods even if they are traceable
to NBS reference gases. The term traceable to NBS is not an absolute assurance of
accuracy. It can have several meanings. The gas manufacturer may be able to trace
all its gases to NBS by calibrating a bulk gas against an NBS cylinder. This is a general
procedure involving much less expense than checking every cylinder shipped against an
NBS gas. There could conceivably be some problems in diluting and filling cylinders
from this bulk gas, yet it would be traceable to a reference standard. The real reference
standard is the original NBS cylinder, which must be carefully stored and used before
deterioration. It should be clear that all calibration gases should be checked by reference
method gas analysis to assure that the cylinder concentration is correct and has not changed
by reaction in the cylinders. Finally, no cylinder may be expected to give good results
if the interior pressure is less than or equal to 100 psi. If the cylinder pressure is low,
change it before problems arise.
12.3 Instrument Performance Evaluation
A complete Performance Specification Test is a complex, expensive undertaking. It is
used to fully evaluate system operation to within given limits in the Federal regulations.
The evaluation of a continuous monitoring system after the initial Performance Specification
Test need not involve repetition of the tests. The plant owner or operator of an affected
facility should make efforts for proper maintenance of the continuous monitoring system.
It has been possible for regulatory agencies to adopt a three-level form of monitoring
inspection procedures. The user and the agency would benefit; the user would receive
guidance and comment on system up-keep and maintenance; the agency would be able to
perform its duties efficiently with less direct enforcement proceedings.
12-2
-------
12.4 EPA Inspection Procedures
12.4.1 Level-One Inspections - (Office Evaluation of Quarterly Reports)
A level-one inspection is intended to identify problem areas. The agency inspector would
carry out routine evaluation of a plant's quarterly emissions report. The inspector would
check;
• Reports of periods and magnitudes of excess emissions
• Nature and cause of each period of excess emissions
• Periods during which continuous monitoring system was inoperative
* Record of calibration checks, adjustments, and maintenance performed on the
monitoring system.
These administrative evaluations save agency manpower and expense. Problem areas in
the monitoring system and reports should present themselves upon thorough evaluation of
the above items. The agency then can contact the plant operator to assist in checking
out the monitoring system, possibly avoiding a situation calling for more extensive agency
action. This could be considered an exchange of information rather than an adversary
confrontation.
The quarterly reports probably will not be very lengthy, yet the experienced inspector can
gain insights into the operation of the continuous monitoring system. The calibration
procedures and routine maintenance indicated for the analyzer can suggest to the knowl-
edgeable inspector whether the instrument is well cared for. The emissions records should
illustrate the normal operating parameters of the plant and its control equipment. If the
records show frequent excess emissions, the plant may have a faulty monitor or control
systems; data that are important to the plant and the agency. The calculation methods
may indicate operator understanding of the regulations. The experienced agency inspector
then may be able to decide upon the necessity of helping plant operators to understand
the intent of the regulations and necessary calculations fully. The quarterly report is a
good indicator of compliance when in the hands of a trained inspector.
12.4.2 Level-Two Inspection - (Field Inspection)
The level-two inspection procedures are initiated when the quarterly report has given
indications that the inspection is warranted. The inspector may feel that the quarterly
report or other indicators require a site inspection of the affected facility. The operations
of the plant process and its monitor system may need review to satisfy the inspector that
all regulations are being properly carried out. The inspector is required to do this type
of inspection to protect the plant and the environment.
12-3
-------
A second-level continuous monitor inspection includes an examination of all major parts
of the monitoring program. The inspector should check instrument zero and calibration
procedures, determining the proper methods of instrument operation from the manufacturer's
written instructions. The plant operator should explain the instrument maintenance log
to the inspector, in addition to routine system maintenance procedures. The strip-chart
or data-logging recorder is checked to ensure that a good record is being made of the
instrument output. The storage and retrieval system for these records should be reviewed
for reliability and accuracy. The inspector then will want to review data conversion and
emission calculations methods with the plant operator.
The specific areas a level-two inspector would be examining are outlined in the check
list presented in Table 12-1. This list will aid in determining the efficiency and proper
operation of the monitoring system. It may be used in decisions concerning the step
to a level-three inspection.
TABLE 12-1
LEVEL-TWO INSPECTION CHECK LIST
I. Monitor Zero and Calibration Time Needed
A. Performed every 24 hours
B. Performed by experienced personnel
C Procedure follows manufacturer written instructions
D. Calibration gases analyzed within Fast 6 months
I. Tank pressure above 100 psi
2. Calibration checks entire sampling interface
E. Data log kept
I. Each entry dated and signed
2. Neat, orderly appearance
3. Up-to-date
4. Maintenance record included
5. Manufacturer recommended maintenance
followed
6. Unusual trends in instrument performance
evident
7. Inoperative Monitor time recorded
a. Source of problem recorded
b. Corrective procedure given
c. Problem repetitive
Yes
No
N/A
Needs
Revision
Yes
No
12-4
-------
TABLE 12-1
LEVEL-TWO INSPECTION CHECK LIST-Continued
F. General inspector comment on procedures
1. Acceptable and effective
2. Could be improved
3. Inadequate
Yes
N
N/A
Needs
Revision
Yes
No
II. Maintenance Procedures Outlined for All Plant
Personnel Involved in Monitor Program
III. Data Recording System
A., Type
1. Strip-chart recorder
2. Data-logging system
B. Electronic interface
I. Instrument output checked
a. Output signal at zero
b. Output signal at span
C. Strip-chart record clear
1. Zero offset 10% at recorder
2. Pollutant concentration readily identified
3. Inking system in good order
4. Plenty of spare charts and pens
5. All pertinent chart data (date, speed,
instrument settings, etc.)
6. Charts easily identified for record retrieval
7. Chart shows cyclic nature of process
8. Chart indicates problem exists in monitoring
system
IV. Data Handling and Calculation
A. All monitor data recorded
I. Any data discarded
2. Any data omitted in averaging monitor
readings
B. Data conversion to units of the standard
1. O2 F-factor; wet or dry
2. CO2 F-factor; wet or dry
12-5
-------
TABLE 12-1
LEVEL-TWO INSPECTION CHECK LIST-Continued
3. F-factor determined by chart
4. F-factor determined by fuel analysis
5. F-factor determined for each batch of fuel
6. Operator maintains record of fuel batch
combusted
C. Excess emissions recorded in units of standard
1. Frequent excess emissions
2. Frequent plant breakdowns
D. Calculations
1. Conversion factors correctly derived
2. Equations and methods clear
3. Any noticeable errors
Yes
No
N/A
Needs
Revision
Yes
No
V. General Check Points
A. System seems to operate well
B. Plant maintenance seems appropriate
C. Instrument drift appears normal
D. Records procedures adequate
E. Calibration materials and instrument spare parts
easily available
F. Instrument site properly located for representative
readings
G. Instrument easily accessible
12.4.3 Level-Three Inspection
The level-three inspection procedures involve a complete evaluation of the source continuous
monitoring system. The level-three inspection becomes necessary after level-two procedures
have indicated unsatisfactory monitoring performance at the source. The inspector may
feel that performance of the entire monitoring program can be assessed only by direct
comparison with agency results obtained by an experienced sampling team. The sampling
team would perform manual reference method testing, a possible portable instrumental
gas analysis, or a process operation inspection. The results from these tests then would
be correlated to previous source monitoring data. The level-three inspection includes a
12-6
-------
thorough evaluation of the monitoring system. A Performance Specification Test would
be performed with agency inspectors observing all procedures.
The results of a level-three inspection may determine the necessity of legal action against
the plant operator if the regulations are not being followed. The level-three inspection is
performed when other inspections indicate significant problems in the system. When legal
action may be necessary, a level-three inspection need not be viewed as the first step in
such a proceeding until all results have been carefully examined.
12.5 Bibliography
Acurex-Aerotherm, "Resource Manual for Implementing the Continuous Monitoring Regula-
tions," developed under EPA Contract 68-01-4142, Division of Stationary Source Enforce-
ment, 1977.
Chapman, R. L., "Calibration of Stack Gas Instrumentation," Calibration in Air Monitoring,
ASTM Tech. Pub. 598, ASTM, Philadelphia, Pennsylvania, 1976, pp. 5-15.
Clements, J. B., Midgett, M. R., and Margeson, J. H., "Evaluation of Air Pollution
Measurement Methodology," Air Pollution Measurement Accuracy as it Relates to Regula-
tion Compliance, APCA Specialty Conference, APCA, 1976, pp. 271-279.
Hughes, E. E., "Role of the National Bureau of Standards in Calibration Problems
Associated with Air Pollution Measurements," Calibration in Air Monitoring, ASTM Tech.
Pub. 598, ASTM, Philadelphia, Pennsylvania, 1976, pp. 223-231.
James, R. E., and Wolbach, C. D., "Quality Assurance of Stationary Source Emission
Monitoring Data," Inst. of Electrical and Electronics Engineers, Inc., V. 36, 1976.
James, R. E., "Quality Assurance of Data from SO2 and NOX Stack Monitors Required
by EPA New Source Performance Standards," Proceedings - Quality Assurance in Air
Pollution Measurement, APCA Specialty Conference, March 11-14, 1979, New Orleans,
Louisiana, APCA, 1979, pp. 419-429.
Jaye, F., Steiner, J., and Larkin, R., "Resource Manual for Implementing the NSPS
Continuous Monitoring Regulations — Manual 4 — Source Operating and Maintenance
Procedures for Continuous Monitoring Systems," EPA-340/ I-78-005d, April 1978.
Jaye, F., Steiner, J., and Larkin, R., "Resource Manual for Implementing the NSPS
Continuous Monitoring Regulations - Manual 3 - Procedures for Agency Evaluation of
Continuous Monitor Data and Excess Emission Reports," EPA-340/ l-78-005c, April 1978.
12-7
-------
Lee, W. G., and Paine, J. A., "Stability of Nitric Oxide Calibration Gas Mixtures in
Compressed Gas Cylinders," Calibration in Air Monitoring, ASTM Tech. Pub. 598, ASTM,
Philadelphia, Pennsylvania, 1976, pp. 210-219.
Licata, A., Kurtz, A. J., and Egdall, R. S., "Possible Errors and Uncertainties in Correcting
Emission Data to Selected Conditions/ Air Pollution Measurement Accuracy as it Relates
to Regulation Compliance, APCA Specialty Conference, APCA, 1976, pp. 218-235.
Logan, T., and Midgett, R., "Quality Assurance Programs to Support the Use of Continuous
Emission Monitors for Direct Compliance," Proceedings - Quality Assurance in Air Pollution
Measurement, APCA Specialty Conference, March 11-14, 1979, New Orleans, Louisiana,
APCA, 1979, pp. 413-418.
Midgett, M. R., "How EPA Validates NSPS Methodology," Environmental Science &
Technology. V. II, No. 7, July 1977.
Reeves, J. B., "Statistical Implications of the Environmental Protection Agency Procedure
for Evaluating the Accuracy of Sulfur Dioxide and Nitrogen Oxide Monitors of Stationary
Sources," Calibration in Air Monitoring, ASTM Tech. Pub. 598, ASTM, Philadelphia,
Pennsylvania, 1976, pp. 118-128.
Wechter, S. G., "Preparation of Stable Pollution Gas Standards Using Treated Aluminum
Cylinders/ Calibration in Air Monitoring, ASTM Tech. Pub. 598, ASTM, Philadelphia,
Pennsylvania, 1976, pp. 40-54.
Wohlschlegel, P., "Guidelines for Development of a Quality Assurance Program,"
Volume XV - Determination of Sulfur Dioxide Emissions from Stationary Sources by
Continuous Monitors, EPA-650/4-74-005o, March 1976.
12-8
-------
APPENDIX A
BIBLIOGRAPHY OF PROCEEDINGS AND REVIEW BOOKS
Proceedings - Continuous Monitoring of Stationary Air Pollution Sources, APCA Specialty
Conference, March 20-21, 1975, St. Louis, Missouri, APCA, 1975.
Proceedings of the Workshop on Sampling, Analysis and Monitoring of Stack Emissions,
Southern Research Institute, prepared for EPR1 (workshop October 2-3, 1975, Dallas, Texas,
NT1S PB-252-748).
Calibration in Air Monitoring - Collected Papers from ASTM Symposium, presented at
University of Colorado, August 5-7, 1975, ASTM Special Technical Publication 598.
Continuous Emissions Monitoring, Dallas, Texas: February 15-17, 1977, Conference Report
and Responses to Key Questions and Issues, EPA-340/1-77-025, December 1977.
A Workshop Meeting for Field Inspection Procedures for the Evaluation of Continuous
Emission Monitoring Systems, Engineering-Science under Contract 68-01-4146, Task Order 7
to DSSE/EPA, Engineering-Science, 7903 Westpart Drive, McLean, Virginia 22101.
Continuous Emissions Monitoring Workshop Manual, Entropy Environmentalists, Inc.,
under Contract 68-01-4148, Task Order 34 to DSSE/EPA, Entropy Environmentalists, Inc.,
P.O. Box 12291, Research Triangle Park, North Carolina 27709,
Stevens, Robert K., and Herget, William F., Analytical Methods Applied to Air Pollution
Measurements, Chemistry and Physics Laboratory, National Environmental Research Center,
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina; Ann Arbor
Science Publishers, Inc., P.O. Box 1425, Ann Arbor, Michigan, 1974.
Instrumentation for Environmental Monitoring, LBL-1 Vol. 1: Air, Lawrence Berkeley
Laboratory, University of California, Berkeley, 1972.
-------
-------
APPENDIX B
CALCULATIONS FOR THE PERFORMANCE SPECIFICATION TEST
A. Calculations
The "Confidence Interval for the Mean Estimated Value" is what is calculated in the
Performance Specification Test. For example, the requirement that an SO2 monitor
have a 2 hr zero drift of 2 percent of span means that the computed "Confidence Interval
for the Mean Estimated Value" divided by the instrument span value be <2 percent.
One calculates the performance parameter (PP) by using the formula
|x| + |Cl95l
pp = -L- — x 100
where:
n
1 V
|x| — the absolute mean value |x] = — 2-, (Aj — Wj)
n j-j
A; = analyzer data - 2-hr or 24-hr period
Wj = concurrent wet chemistry data - 2-hr or 24-hr period
where:
CI95 = the two-sided 95-percent confidence interval estimate of the average
mean value
B-l
-------
where:
= Ai - Wi
to.975 - The t value derived in the t test corresponding to the probability that a
measured value will be within 95 percent of true value.
(Note: to.975 = tl-a/2 = t]..05/2 = to.975)*
The values of to.97S are obtained from Table B-l.
TABLE B-l
VALUES FOR to.975
n
2
3
4
5
6
7
8
9
tO.975
12.706
4.303
3.182
2.776
2.571
2.447
2.365
2.306
n
10
11
12
13
14
15
16
to.975
2.262
2.262
2.201
2.179
2.160
2.145
2.131
RV is either a reference value, a calibration value, or a span value. In the case of opacity
monitors, values of RV are not divided into |x| + |ci95J to obtain PP.
This is all that is needed in the calculation of the performance parameters. The major
problem in doing the calculations results from misunderstanding the fact that the absolute
mean value, |x|, and the absolute value of the confidence interval |CI95| must be added
together.
It should be noted, that since |ci9s| must be added to |x| to obtain PP, an increase of
the number of samples will decrease the contribution of CI95 to PP. This is due to
the fact that n is in the denominator of CI95 and that to.975 decreases with n.
*This means that for a 5-percent level of significance, there is a 2-1/2-percent probability
of obtaining t greater than that in Table B-l and a 2-{/2-percent probability of obtaining
t smaller than the negative of the tabulate value.
B-2
-------
Also, for low values of calibration or span gases, PP values may be correspondingly higher.
Also in the case of a source with low emission values for SO2, the PP for accuracy of
an SO2 monitor would be higher than if the same monitor were placed at a source emitting
higher concentrations of SO2-
B. Rationale Behind the Confidence Interval Calculation
The use of the t test and the confidence interval calculation in obtaining performance
parameters is an attempt to place a numerical value on the correlation between two
parameters, e.g., the SO2 value given by the instrument and the SO2 value determined
by the manual reference method. If the values correlate well, the value of t will be low,
i.e., both the monitor and the reference method would each be measuring the same thing,
each measuring it relatively accurately.
Essentially, this statistical method arises out of regression analysis. Say, for example,
the monitor gives a value X and the reference method gives a value Y. A least squares
regression method attempts to show that there exists a relationship such as
Y = A + BX
where A and B are numbers.
In Figure B-l(a) the parameter X and Y -would be highly correlated; in Figure B-I(b)
the parameters would be independent. What the t values (or t test) then give us, is the
value of t that one would obtain if there was a 95-percent probability that X and Y
correlated. This is the 95-percent confidence level.
(a)
(b)
FIGURE B-1
B-3
-------
The confidence interval for the "mean estimated value of a number" is given by
Clmean = |x| + tS
where | x | is the mean estimated value, S is the standard deviation of the mean.
Here, t is obtained from the table and S is computed. Therefore,
CI
_!_ A(Xi)2 -
n V „ n-
= lx| + |CI95|
and
Clmean
100
B-4
-------
APPENDIX C
F-FACTORS
The New Source Performance Standards require that emissions from an affected facility
be reported to the Administrator in terms of process rate. A process rate standard is
written in units that relate pollutant emissions to the production rate of the industry:
pollutant emission weight
process production rate
The emission rate (E) is given in units such as:
• pounds/million Btu
• grams/million calories
• pounds/ton
The F-factor is used to calculate the emission rate in the units of the standard. It reduces
the amount of data necessary to complete the emission rate calculation. The relationships
that make possible emission rate calculation using an F-factor are explained in this section.
A table of F-factors and summary to types of F-factors is included for easy reference.
Definition: The F-factor for making emission rate calculations is developed from
a chemical and combustion analysis of the fuel burned to operate a production
process. The F-factor is the ratio of the theoretical volume of dry gases (Vt)
given off by complete combustion of a known amount of fuel, to the gross
caloric value of the burned fuel (GCV).
_ volume dry combustion gases _ Vt
gross calorific value GCV
The values of the constituents in the F-factor are determined by a fuel analysis. There
are two types of fuel analysis, proximate and ultimate analysis.
Proximate analysis - a fuel analysis procedure that expresses the principal characteristics
of the fuel as:
1. Percent moisture 4. Percent fixed carbon 6. Heating value (Btu/lb)*
2. Percent ash 5. Percent sulfur 7. Ash fusion temperature
3. Percent volatile matter
(Total 1-5 = 100 percent)
*Gross Caloric Value (GCV) - Also termed the "high heating value." The total heat obtained
from the complete combustion of a fuel referenced to a set of standard conditions. The
GCV is obtained in the proximate analysis as the "heating value."
C-l
-------
Ultimate analysis - the determination of the exact chemical composition of the fuel
without paying attention to the physical form in which the compounds appear. The
analysis is generally given in terms of percent hydrogen, percent carbon, percent sulfur,
percent nitrogen, and percent oxygen.
The data generated in an ultimate analysis of a given fuel allow the calculation of an Fd
factor based on the composition of the fuel constituents. The individual chemical components
are included in the theoretical volume as uncombined elements. Each contributes to the
total Vt based upon the percentage present in the fuel. An F-factor can then be calculated
for any fuel when the percent composition of each constituent is known:
Fd = 227.0(%H) + 95.7(%Q + 35.4(%S) + 8.6(%N) - 28.5(%O) metric
GCV units
Fd = 1Q6 6.34(%H) + 1.53(%C) + 0.57(%S) + 0.14(%N) - 0.46(%O) English
GCV units
The F-factor is developed from theoretical calculations on the combustion of a fuel. The
preceding equations account for only a stoichiometric amount of oxygen - enough oxygen
to completely oxidize the fuel to its combustion products. An industrial facility burning
large quantities of fuel adds a stoichiometric amount of air (oxygen and nitrogen) and some
excess air to assure complete combustion of fuel. The volume of the combustion products
is related to the heat input of the fuel and the excess air in the expression:
f^. Excess
vs
X
Air Correction
OH
Term
where:
Qs = volumetric flowrate of dry combustion gas
QH = heat input rate
The stoichiometric oxygen present would be consumed for combustion of the fuel. The
remaining oxygen present in the combustion gases is, therefore, an excess. The percentage
excess is then calculated using the percent oxygen in air and the percent oxygen found in
the combustion gases:
Percent excess air - -Q'9~J° 2 (dry basis)
20.9
C-2
-------
H is then possible to show that combustion gases at an actual combustion facility can be
corrected to the theoretical combustion gas volume:
•
Q 20.9 - %o2 Vt
QH 20.9 GCV
which is dimensionaily consistent if we consider QH in terms of heat input per pound
of fuel per hour:
Qs 20.9 - %Q2 = Vt
QH 20.9 GCV
DSCF(s)
Hr 20.9 - %02 = DSCF(t)
106Btu/lb 20.9 !06Btu/lb
DSCF(S) DSCF(t)
106Btu/lb 106Btu/lb
The importance of the F-factor becomes obvious if we now write the equations:
_2i - Vt x 20.9 _p 20.9
QH GCV 20.9 - %O2 20.9 - %O2
which illustrates that by using the Fd-factor generated in the laboratory for a given fuel
and correcting for percent excess air in the combustion gases, it would not be necessary
to determine the stack gas volumetric flowrate or the fuel feed rate of a combustion source.
The emission rate could then be calculated from the pollutant concentration in the stack gas,
F-factor, and the percent excess air.
CsQs
E =
E - CsFd
Ih
E =
QH
20.9
20.9 - %02
lb DSCF(t) f 20.9
DSCF(s) 106 Btu [20.9 - %O2_
The Fd-factor may be used for emission rate calculations if the percent oxygen (%O2ws)
and the pollutant concentration (Cws) are determined on a wet basis and if the moisture
content (BwS) of the stack gas is known. The emission rate is then corrected to a dry
basis for reporting to the Administrator by the equation:
90 Q
— /~< r- . *•«• -*
— l_W<;
20.9(l-Bws)-%02ws
C-3
-------
Different Types of F- Factors
The Fd-factor is derived for dry gas volumes and determining excess air by measuring
percent oxygen in the flue gas on a dry basis. There are other F-factors which, have been
developed for measurements on a wet basis or for excess air determinations made by
measuring percent carbon dioxide in the flue gas.
Fc: Carbon Dioxide F-Factor
Fc=^M%Q (metric units)
GCV
F _ 321X1Q3(%C)
re -- - (English units)
100
E - CsFc %C02
Fw: Wet F-Factor
F = 347.4(%H) + 95.7(%C) + 35.4(%S) + 8.6(%N) - 28.5(%O) + 13.4(%H2O)
GCV (metric units)
F _ 1Q6 [ 5.56(%H) + 1.53(%C) + 0.57(968) + 0.14(%N) - 0.46(%Q) + 0.21(%H2Q)]
w GCV (English units)
The wet F-factor, Fw, may be used in the expression:
_ _ _ 20.9
~.(*(\ , r» \m\
20.9 (1 - Bwa) - %O2W
where:
CWs = the concentration of the pollutant given on a wet basis
%O2W = the percent oxygen on a wet basis
Bwa = the ambient air moisture fraction
may be determined by a number of methods given in 41 FR 44838 October 12, 1976.
This method may be used in systems which measure gas concentrations on a wet basis,
such as the Lear Siegler and DuPont systems. For wet scrubber applications, a determination
of Bws must be made, then utilized in the expression given on the previous page.
C-4
-------
F-Factor for Combination Fuels
A source may utilize a combination of fossil fuels. The F-factor for a combination of
fuels is determined from the general expression:
n
Fm — 2* Xi Fj
i=l
where:
Fm - F-factor for the fuel mixture (gaseous, solid, liquid fuel)
Xi = fraction of total heat input from each type fuel
Fi = the applicable F-factor for each fuel in the mixture
which states that the F-factor for the mixture is the sum of the products of the fraction
of heat input for each fuel multiplied by the applicable F-factor for that fuel. All F-factors
must be on a consistent basis - 02 or CO2; wet or dry. The example shows the calculations.
A combustion source burned the following combination of fuels to produce process steam:
Fd Fc
Fraction Heat Input Fuel Type DSCF/106 Btu SCF/106 Btu
10% Natural gas 8740 1040
10% Butane 8740 1260
20% Oil 9220 1430
60% Bituminous Coal 9820 1810
The combination Fd-factor is:
DSCF
Fd = Fm - (0.10) (8740) + (0.10) (8740) + (0.20) (9220) + (0.60) (9820) - 9484
106 Btu
The combination Fc-factor is:
Fc = Fm = (0.10) (1040) + (0.10) (1260) + (0.20) (1430) + (0.60) (1810) - 1602 S,CF
106 Btu
The plant engineer may obtain the F-factor values used in calculating the combination
F-factor from Table C-l. These values have been determined for various categories of fuel
from fuel analysis data taken of a large number of samples. The Fd-factors for a fuel
C-5
-------
category have shown that they may be calculated to within ±3 deviation; the Fc-factors
have been calculated to be within ±5.9 percent deviation. The Federal Register gives the
engineer the option to use these factor values, or with approval from the Administrator, to
develop F-factors from fuel analysis for the fuel as received at the plant.
Compilation of F-Factors and Emission Rate Calculations
Tables C-l, C-2, and C-3 give a summary of F-factors, units, and emission rate calculations.
The subscript w indicates a wet basis expression; all others are on a dry basis.
TABLE C-l
F-FACTOR CALCULATION EQUATIONS
Fd -
227.0%H + 95.7%C + 35.4%S + 8.6%N - 28.5%O]
GCV
(metric units)
Fd =
106 [3.64%H + 1.53%C + 0.57%S + 0.14%N - 0.46%Q]
GCV
(English units)
Fc =
20.0%C
GCV
(metric units)
Fc =
321 X 1Q3%C
GCV
(English units)
lit
_ 347.4%H + 95.7%C + 35.4%S + 8.6%N - 28.5%O + 13.4%H2O=
- i .1 . **
GCV
w
(metric)
Fw -
106 [5.56%H + 1.53%C + 0.57%S + 0.14%N - 0.46%Q2 + 0.21%H2Q*]
n
GCV,
XiFi (consistent basis)
(English)
*Note: The %H2O term may be omitted if %H and %O include the unavailable hydrogen
and oxygen in the form of
C-6
-------
TABLE C-2
F-FACTORS FOR VARIOUS FUELS a-b
Fuel Type
Coal
Anthracite
Bituminous
Lignite
Oil
Gas
Natural
Propane
Butane
Wood
Wood Bark
Fd Fw
DSCF WSCF
I06 Btu 106 Btu
10140(2.0) 10580(1.5)
9820(3.1) 10680(2.7)
9900 (2.2) 12000 (3.8)
9220 (3.0) 10360 (3.5)
8740 (2.2) 30650 (0.8)
8740 (2.2) 10240 (0.4)
8740(2.2) 10430(0.7)
9280(1.9)*
9640(4.1)
Fc
SCF
106 Btu
1980(4.1)
1810(5.9)
1920(4.6)
1430(5.1)
1040 (3.9)
1200(1.0)*
1260(1.0)
1840(5.0)
1860(3.6)
Fo
1.070(2.9)
1.140(4.5)
1.0761 (2.8)
1.3461 (4.1)
1.79(2.9)
1.10(1.2)*
1.479 (0.9)
1.5 (3.4)
1.056(3.9)
a Numbers in parentheses are maximum deviations (%) from the midpoint F-factors.
b To convert to metric system, multiply the above values by 1.123 X IO"4 to obtain scm /106 cal.
Note: All numbers below the asterisk (*) in each column are midpoint values. All -others
are averages.
C-7
-------
TABLE C-3
F-FACTORS AND EMISSION RATE CALCULATION
Factor
Fd
Units
DSCF
106 Btu
Measurement
Required For
Excess Air
Determination Calculations
(drv 20 9
*VnO ^ * F r1 PI ^u"
'0*J u • \ fc ~~ t-S^d JG?n ^J?"
basis) 20.9 - %O2
Comments
Cs determined on
dry basis
Fc -DSCF (dryor _ _JOO_
106 Btu %C°2 wet basis) E ~ QFc %CO2
WSCF
106 Btu
%02
(wet
basis)
E —
20.9
20.9 (1- Bwa) -9602
Fd
DSCF
106 Btu
%02
(wet
basis)
E = CwsFd
20.9
20.9(1-Bws)-%02
F - 20-9 Fd _ 20.9 - %O2
° 100 Fc %C02
Cs on dry or wet
basis consistent with
CO2 measurement
The "wet" F-factor,
Cws and %O2 on
wet basis
Bwa = average
moisture content of
ambient air
Fd used to calculate
E with %O2 and CWs
on a wet basis and
gas moisture content
known
Miscellaneous
factor useful
for checking
Orsat data
*Note: The wet F-factor, Fw, may not be used in any application which involves the addition or
removal of moisture from the combustion effluent. As a result, it is not suitable for wet
scrubber applications without additional correction.
Note also: Bwa = Amount of moisture in ambient air, which value can be established by any of the
following four methods.
a) Fixed constant value of 0.027 allowed
b) Continuous measured value
c) Monthly value based on previous history
d) Annual value based on previous history
C-8
-------
Sulfuric Add Conversion Factor
A sulfuric acid manufacturing operation must establish a conversion factor for converting
continuous monitoring data to units of the standard (Kg/metric ton, lb/short ton). The
factor must be determined a minimum of three times per day by measuring the sulfur
dioxide concentration entering the gas conversion unit. The Reich test is generally used
for measuring SO2 at the plant. The conversion factor is calculated:
CF = K l-000-9.015 (r)
r - s
where:
CF = conversion factor (kg/metric ton per ppm, Ib/short ton per ppm).
K = constant derived from material balance. For determining CF in metric units,
K = 0.0653. For determining CF in English units, K = 0.1306.
r = percentage of sulfur dioxide by volume entering the gas converter.
Appropriate corrections must be made for air injection plants subject to
the Administrator's approval.
s = percentage of sulfur dioxide by volume in the emissions to the atmosphere
determined by the continuous monitoring system.
The continuous monitoring data are then multiplied by the conversion factor to give units
in the standard:
vy Kg/metric ton .. , . ( (metric
monitor ppm X -JL_ = Kg/metric ton ^
Nitric Acid Conversion Factor
A nitric acid manufacturing operation must establish a conversion factor for converting
continuous monitoring data to units of the Applicable Standard (kg/metric ton, lb/short ton).
The conversion factor must be established by measuring plant emissions in terms of NO2
concurrent with reference method tests of the emissions. The conversion factor is determined
from monitor data taken only during the reference method test. It is calculated by dividing
the average reference method NO2 data by the NO2 data average measured by the monitor.
The ratio is expressed in units of the standard:
Average gm/m3 x 0.602 = Kg/ton (metric units)
Reference Method
„„ Reference Method Kg/ton „ (i ,
CF = n—r ~ = Kg/ton/ppm
Monitor ppm ° 'r
C-9
-------
The conversion factor multiplied by monitor concentration data then yields units of the
standard:
monitor
ppm X Kg/ton/PPm =
Other Uses of F-Factors
If values for Qs, the stack gas volumetric flowrate, and QH, the heat input rate, are obtained,
as they often are, several cross-checks can be made by comparing various calculated
F-factor values with the tabulated values.
Equations that can be used to do this are given below:
-ft %CQ_QSW%CQ2W
" 100 -~~
If, after calculating Fd, Fc, or Fw, a large discrepancy exists between the calculated value
and the corresponding value in the table, the original data for Qs, QH, and the O2 or CO2
data should be checked. This is essentially an easy way of conducting a material balance
check.
Using a tabulated value for Fd, Fc, or Fw and the data obtained during the stack test for
Qs and %O2 or %CO2, a value for QH may be obtained from the equations.
If ultimate and proximate analyses are available, they may be used to calculate an F-factor
using one of the equations. The calculated value can then be checked with the tabulated
values and should be within 3 to 5 percent agreement, depending on the type of fuel and
type of F-factor.
The Fo-factor may be used to check Orsat or continuous monitoring O2, CO2 data in
the field.
C-10
-------
The Fo-factor is the ratio
= 20.9 Fd
0 100 Fc
and is equal to
,, 20.9 - %02
%C02
the %O2 and %CO2 being obtained or adjusted to a dry basis. A value differing from
those tabulated would necessitate a recheck of the O2, CO2 data.
Errors and Problems in the Use of F-Factors
The following factors may contribute to errors in reporting emissions by using F-factors:
• Deviations in the average or "midpoint** F-factor value itself.
• Errors in the Orsat analysis and the consequent %O2 and %CO2 values.
• Failure to have complete combustion of the fuel (complete combustion is assumed
in the derivation of all of the F-factor methods).
• Loss of CO2 when wet scrubbers are used - affecting the Fd, Fc, and Fw
factors. Addition of CO2 when lime or limestone scrubbers are used - affecting
the Fc factor.
• Variations in Bwa for the wet F-factor method, Fw-
The deviations in the F-factors themselves have been found to vary no more than about
5 percent within a given fuel category. Since the F-factors given are averaged values,
differences in the ultimate analysis between fuel samples could easily account for the
deviation. The most significant problem in the use of the F-factors, however, is in the
excess air correction. An error of a few percent in the O2 or CO2 concentration could
cause a relatively large error in the value of E.
Since the F-factor method has been developed assuming complete combustion of the fuel,
incomplete combustion will cause an error. However, if the CO is determined in the flue
gas, some adjustment can be made to minimize this error.
(%C02)adj = %C02 + %CO
(%02)adj = %02 - 0.5 %CO
C-ll
-------
By making these adjustments, the error amounts to minus one-half the concentration of CO
present. However, this does not account for any unburned combustible matter. The F-factor
methods would count the calorific value of this unburned fuel towards the heat input, and
a positive bias would result in the calculated emission level.
C-12
-------
APPENDIX D
TITLE 40 PART 60 APPENDIX B
CODE OF FEDERAL REGULATIONS 1977
PERFORMANCE SPECIFICATIONS
D-I
-------
So-
o*"Ei.»*««s " s
scSS^ie'Si
list! JflSsW i
5ilsS§|aSsfli.*
-S.-3«>«8B4iSMS*-3
SH3?sffilsS|3fl
his
Ba|ft2»aB1a''a'OHS
SSS-sBExisfs"
!l?^--
5 a
lisa
as
X
w
e
*
o>
c
o
5S=
k 8S
J3&
a™
cfi
EC
»S
h
II
lilfH
MB*er>a
o
iiilff
SSS3S§
a
a
!i«H piiii Jfiii lii: ^tiii
SoS^*- §7?E£«_ dS^Sw B=1S °«SSB
|S^e°
^ Q. ** ft ^
^"3
51 PI
s|i3£
5s 5^*;
SH^O-S,
^?gs-
l^o'fi
** f fl-2
Illia
'+»».£5^
I -I B ** =4
O I I ^
• M Ssa
8l6!i
<£*!*
S?
i c * *s w ni>._2 P S> *, a. ^ •> « e O 5 *
**aji5sssowoS-
BSwSS§S.s_s-s
css^5S||«
WHJJI
fllll^gSl^^si^i
|tl|2£|Ss|^og!|i
i-SeaifJfrfsa^l-i.
Ihll Ulj1iill!!S
Egg
• J3O O O
^Iflljl:!! fl!?!«" I61HIJirf*
slliill^I^llllllf-llllili!
D-2
-------
*» a e
tf D •• r!
^ii
•°5
*lgill«i3ii^!;psi3il
*31iiU?«S9illia*ISiM
ra«j».. ~*38.«sI!>3,.i.ceS'Df
98|
iPiHi^^S'flwS*
*-,§•£ fl§*2*3283«S>SiB
lEIS.SPt*
fcsfiu's
Ss|!||8jl|5
*5?Is3g1&S3
!i\\H us
«ssg »s * £g =s
fl 1 ? S ;
I
-------
H
i 5
i s
^ AI
B ^^
II ^B^>^
^™ ^™
fe it. •
O ^
• ^
> M f
! CO.
o £
1 ^~ 0
! « <-
• iS
3
w ai
E •>
•oi£
*
• **
:
V C
yr jv
^
4J
V-
U
«
o.
XO
TJ**
^
1
gl
n^ C
8c «
w o.
t \
•»->
T>
i-
4T
M
10
|V
CO
Ol
o
^.
CU
«•>
V
IA
Ol
Z
t»
^~
S
— 1
o
•*-
f*}
V
V
C
£
at
*^
**•
o
c
•o
E:
•— n
E o
« ai L.
u ** j_
C C 41
£ "" c:
- Ol Of O
*t* t) *—
•^ ttl M
•w -o t.
^~ ^3
wo 7
E . O «J
O»
C
vz
Ol
(-
C"
^
r^
Ot '»—
™- — v
JC H- 3
«- C "3
O O >
T» ** at
" J 3
• -^ o
X r> tn
o n ja
t^ CM n
w
e
9
!|iil I1
Sl«89*
rialii*
I*is2"|.
^laPl-
•sasi§&5
SlliS^.
Ifir
-I'te"
;HS
s°g?
SaS
&^
S1HS8
fe 2 » 9<9
??i <
^i?§«
sISa^
3552
1^-3 *•
I2!i:ig
a55oS£ IlStSlf
s^igs |l?s»«?
-M-- - v_ o o a ^-
Ss-alSif
««;£«£«
IwS.oS
"8SS|
S|js«
e«*A«i>eio*
Niss?il:s
iS]|jiSlM«
5s£
«"£,
D-4
-------
et!itfl
"liiil
3sa££
S5*|8
B«0 H O „*•«>
3«3||8«
8) o 6.5 C
g*3
«il!
C5? *
I Us
0*0+*
** E ^*
ai*5|
'd«aa."Pa
IHSW
S&«!fS
§&sls
i5!i
III1"
iSvS&*jS5i
Hil«5!ll
:-!s
•111!
i
P
^Piileil!
.it38*s~Ssfe*S£9«|
a.
i
o
5
e
I
^
;
J
I
I
)
1
•
j
1
j
i
i
.
i
1
I
n
1
w
1
^
|
_
i
]
it
*J
i
ft
i
,
!
„
i
.
!
S
i
i
*
IM Ttll
j
i
s
~
7
I
"•
|
9
C
£
£
•
CT
c
V
e
S.
1
*> *-
« *- '
o
c
3 C
« ^
•OMB
S-o c
01 5 S.
« o>
"11
t. 3
d~
•E
£•—
"C £
a «
rj
Ei
V
v q of
•s^S
V U M
a, a
Zon
«!%>
"el
*-sM
_ *""
,
o
5
c
•r
f
J
,
*J
a
i
|^
(
1
in
u
*
«,
c
C
!
•o
e
•
a
i
i*.
^
o
e
<
•
i
»
S
€1
*
>
*
3
J
*>
M
,2
*>
o
c
o
A
i
£
V
t
•**
e
£
u.
D-5
-------
J E
o Q
B &
8 §
I!
o S
| 8
! S
I §
||
liil
:I,
HI
P^I!1
S -::** i5
. C vw C
5=
. A
£-
l«
S5
*S£
S-°
s
SB
ss
&s"
E*
s «
fi£
sg
l£!
tsS-^5
I*
x; **
C a
S *>
HP
4 at TI - C c S
8Sse,.eo,SS
5s5cBc+"
£C *io = >,«
., « tj *» »> « ft
S c
C.C 0 =
* •«-•=
I *
03*
oS,p^«
eE&_«--«s-«-,gg*«5
liM-ilUHt
s£i
t->b
•'?•«
£ a a
. .
S = t _°K?-2e c
23
0
slsis?
a*S£
D-6
-------
*g£o:* |s H^Q
flg£$G> |3 "a**
5 fl e ffiSSV* B^ga?
i i i li
m
am
s«a
8;i
l^lfili
iMttHE ,
''HUSH ills
M**!i!i! hit
iiiH&i *s?»
*-*8lslil -Kit
^3|i|g' RS8§-|i
lgiB||glgI|.
|
UliliiiiFiM
w
e
en
o
c
Ul
is8i«
*Ǥ
t|
Sa
««
M
V 2
5 Sg
.2 •§§
0 (C
Q. a v
ti 5
s s^s
«
^fl^"3
•3 * 4> „
as o »
•gsiS ^ .»
"i!
M .?» $ >
11 II 11
us O O * P^t*-^1
a c.
: i i ; i
i i i i
i' t i
! I ! •
•dvocr-tc
o
; «
i •
< i
! i
o — «r-
*ri
-------
If
5 3
J.
' £S
H Ol
IF
!i
Lv
il
•a-
11
P
_
e't?
a-
*J*^
J~
111
3* - - «
•
i
i
ii
i
i-
ss*
ffS
1
Is
. p:
i"
i
i
*
U
1
•I
£
»-
£
^
|
E
S ,
i"
a*
s
B
m
k
•
•
IS
if
I .
n
? -S
u o
i =
1.
i
S
1
J «f
s"
o
o
e
f
*
£
(
7
*
JI
** **
IS 2
^ V
a _
f
£5
h
S
a.
1
5-
c
»
e
S
I
e
»
E
I
1
•
a
o
6
*"• 5-
T T
e j
!Z »*
_ o
e 01
P **^
i*. «
4!
S
3
X B~
- ^
!J
i3
v. S
1 I
-
«
,
o
w
I
C •
E (
»-T
M 4
*
S °
e e
B !u
^ M
a c
- 41
»— u
« E
J O
• ^
^
r-
«
*»
w
^
•»
e
^
u
2
2
u»
- 1 1
f i *l
- 1 1
s -u«
-
+
•
c
*.
*-
5
c
J
•* I
* S 1
1 = i
I i 1
"* 2 i
I 5 S
H
1
I1
K
n
O V
5 S
s
M
i i
o
J-
=1
M
•O
«•
1
c
o
*-
2
* S
o
w
;**
« V
1—
i 5
f *O
-5^
5
#»
S
i
2
|
Ul
g
^f
«
,A
U
Ctf
r*
C
ot
b*
D-8
-------
I
8
s i
u e
S.
S
i— N f>
»• OJ «
S
€ *
I ?
J6 a
§IPg!
rf^*"
a A o ti
« C H 3
Q • 3 S A *
HI
s i s
siili* sgsst
lilB*«
ssilS
as g.« S
.f!M
i-in
• q w A
g3a
** »" o fl *»
S5S3CHS 5
I8ass|*«f
a2
i
t? M T- s a Ti « :
O^a ™*-PS5ft°*tJ«
§siisS iSl'Sgtllg
Bashg Missis3^
II
>>g
si
<
>.
B
1
C
e
ti
*
?
£
"5
•onment
,J
>
I
t*
£
a
o
u
c
o ^,
*J*J C
i_t- a.
3 "
^
1
en
Ct-5
trt fw I
°=
O *J O
J-<»- S-
a S3
Ot
41-5
*** S
Q£
W d
SI?
i— »
c
CL f
12.
~ ' H *
w +
ISI
*
0 • •
* g g
•~ *"
X « X
4J
3. o a.
V) VI
*J +J 4O 4^
if e a. c
T E B
£ 5 c p
fc
s
^
O
"g
-------
'^:
II
S
a
*~ B ^
— >•« • " -D
sells.!
o o
C «
fe '
:i
ii
a«3
as
5^
c «
•£
5 3
58
3
'£
B,«
Is
• *
r E
SH
3J^5^
Sft^^S
« 1-3
;iS^|
srs£8
•c«°S
c S w— »
« a-o o
M^g =
o c t g«
£B.|S-
aS«Is
llts =
e!hi
Is*
a =Ii*
u MM * O<
Sl^33f -
-= ^- *S 9
S^I'ofl
S.*9?V°a
•-|3S=F
og2=-s£
^S.assi
•> « n » *i -Q
^i^»±-£3
iS"J«S"
Is
5 ?S:
g£ S * • -
us—*o •
i--2 ff;!§
isfinSi
c « « a « C -3
Illlfl
•= :ii
•ft 5*1
HsfS^sPis*!
»•=" S^ff
•c S
** a,™
- £*!?cc;i'1Be«*6J£ — ~c —
9°-"Mi^glIs»flsl»
E uS™r;^« E*<'™ 5S™
£3Hg£ i
: = ^ = £&»SS.5fi
I
1
t
o
a
S
^
n
E
|
|
• _
«
I
5
8
B
•e
a
o
o
•n
1
S
ft
.
|S
•o a
mS
** tf
0 3
5*
Is
!£
M
«8
£ w
-S^l
it^i
j I
"1
K
C
^
a
n
e-
K
M
=HT
— 1C
|
in
£
•>
»
|
*
S
e
Ii
if
*•=
£§
*-2
0 >
si
_ c
s«
fi5
2o
l£
€S
II n
*"»•»
w
c
«'
_5
c
•
•
E
Ii
M
5
c
I
•
V
*-
o
k.
£
E
3
C
1!
C
if
B -.
S3
f <•
-------
c
i
C
o
*>
111
*o
c
5
u
£
g
£
4
_•
a
a
a
I I
** 4
etf **
I
CSV-
4) 4
S
~
|c
1
*
4 4r e
_.°M.?f
,
O
ll
e
S"
;„
j
+
'•i
* C
*j m
*- a
a e
M
c "
bfi J
** S^
i"1" *
k- 4 3
a w *-
sil
isiu .
e
UI
o.
o
~i o e > —> „ 16*^
fc.5 — -nfflSi-v.
-3 -3 « C O
^iPH'iit!^!
|flll|l|ji|lil|!!
llsilililllllill!
!ffa!i|MMia °*
5 .
c«
t. o u ,= Ji
a> a.» *> a °
o, a "• » tf
Oh *« w *• d
•4 n ^
u.u liiii"frtfc-S'^**r'IMff^
«^XviOa>*1itiO<00sC v
S5£££§S||S?^IS?
||f-|l|5gS.S,|»| = |
D-H
-------
& «
&, »
a.
a
^
^^
2
*
9
c
o
w
0
^
^^
5
e
E
e
o
•y
c
Ul
I
k
&
O
u
c
o
-u ** c
•a *- W p
w C
of
V
3
O
M
*
L.
O
J^
1
SI
<*-.
^—
0
c
o
*J
10
t.
^
a
o
c
*
o
. Ol
M
cu
f*>
1
en
1Z
D-12
-------
APPENDIX E
CONVERSION FACTORS AND USEFUL INFORMATION
E.I International Metric System - Le Systeme International d'Unites (SI Units)
Base Units of the International Metric System (SI)
Name of the Unit Symbol
| Quantity
Length
Mass
Time
Temperature
Electric current
Luminous intensity
Amount of substance
E.2 Recommended Decimal Multiples and Submultiples and
the Corresponding Prefixes and Names
meter
kilogram
second
Kelvin
ampere
candela
mole
m
kg
s
K.
A
cd
mol
Factor
10»
106
102
10
lo-1
10-3
10-9
10-12
10
-18
Prefix
tera
giga
mega
kilo
hecto
deca
dec!
centi
mi Hi
micro
nano
pico
femto
atto
Symbol
T
G
M
k
h
da
d
c
m
M
n
P
f
a
Meaning
One trillion times
One billion times
One million times
One thousand times
One hundred times
Ten times
One tenth of
One hundredth of
One thousandth of
One millionth of
One billionth of
One trillionth of
One quadrillionth of
One quintillionth of
E-i
-------
E.3 Some Derived Units of the International Metric System (SI)
Quantity
Frequency
Force
Pressure
Energy
Power
Quantity of electricity
Electrical potential or
electromotive force
Electric resistance
Electric conductance
Electric capacitance
Magnetic flux
Magnetic flux density
Inductance
Luminous flux
Illumination
Wave number
Name of the Unit
hertz
newton
pascal
joule
watt
coulomb
volt
ohm
Siemens
farad
weber
tesla
henry
lumen
lux
tymbol Equivalence
Hz
N
Pa
J
W
C
V
n
s
F
Wb
T
n
1m
Ix
V
I Hz =
1 N =
1 Pa =
1 J
1 W =
I C =
1 V =
i n =
1 S =
1 F =
I Wb =
1 T =
i n =
1 1m =
1 Ix =
Is-'
1 kg X m X
1 N X m-2
I N X m
1 J X s'1
I A X s
1 W X A'1
1 V X A'1
i n-1
1 C X V1
V X s
Wb X m-2
Wb X A"1
cd X sr
Ix X m-2
E.4 Some Suggested SI Units for Air Pollution Control
Volume flow: Litres per second (1/s)
Velocity (gas flow): Meters per second (m/s)
Air to cloth ratio: Millimeters per second (mm/s)
Pressure: Kilopascals (kPa)
E.5 Conversion from ppm to g/m3 at STP
Tstd = 273.15° K
Pstd = 1 atm
g
dscm
ppm X M
.w.(-
\g-
$_}
mole /
22.414
foers x 10_3 m3 /293.15QK
g-mole I031 \273.15°K
X 106 ppm
E-2
-------
E.6 Conversion Factors
Equivalents
Energy* Heat, and Work:
\ Btu = 252.0 cal
I Btu = 0.2520 kg-cal
1 therm = 100,000 Btu
1 Btu = 778.2 ft-lb
1 Btu = 1055 Joules
I cal = 4. IS? Joules
1 hp-hr = 2544 Btu
I kwh = 3412 Btu
1 hp-hr = 1,980,000 ft-lb
1 kg-m - 7.233 ft-lb
Power and Heat Flow:
1 kw - 1.341 hp
J hp = 550 ft-lb/sec
1 hp = 42.41 Btu/min
1 Btu/sec = 1.055 kw
I kw = 3412 Btu/hr
1 hp = 2544 Btu/hr
Multiply
Btu
cal
Btu
kg-cal
therm
Btu
Btu
ft-lb
. Btu
Joules
cal
Joules
hp-hr
Btu
kwh
Btu
hp-hr
ft-lb
kg-m
ft-lb
kw
hp
hp
ft-lb/sec
hp
Btu/min
Btu/sec
kw
kw
Btu/ hr
hp
Btu/hr
Btu/min
kw
Btu/min
Ib/hr steam
Mega watts
Boiler Hp
Boiler Hp
by
252.0
0.003968
0.2520
3.968
100.000
0.00001
778.2
0.00)285
1055
0.0009478
4.187
0.2388
2544
0.0003930
3412
0.0002931
1,980,000
0-0000005051
7.233
0.1383
1.341
0.7457
550
0.001818
42.41
0.02358
t.055
0.9478
3412
0.0002931
2544
0.0003930
0.01757
56.92
0.001
0.454
1360
33,479
9,803
to obtain
cal
Btu
kg-cal
Btu
Btu
therm
ft-lb
Btu
Joules
Btu
Joules
cal
Btu
hp-hr
Btu
kwh
ft-lb
hp-hr
ft-lb
kg-m
hp
kw
ft-lb/ sec
hp
Btu/min
hp
kw
Btu/sec
Btu/hr
kw
Btu/hr
hp
kw
Btu/min
Ib/hr (steam)
kg/hr (steam)
kg/hr (steam)
Btu/hr
kw
E-3
-------
Equivalents
Heal Flux:
\ cal/hr sq cm = 3.687 Btu/hr sq ft
1 watt/sq cm = 3170 Btu/hr sq ft
Thermal Conductivity.
Btu ft _ Btu in.
hr sq ft °F hr sq ft °F
1 B'U ft 14 SB ^ Cm
" hr sq ft °F ' hr sq cm °C
j watt cm Btu ft
' sq cm °C """ hr sq ft °F
Heal Content:
'TT-"»—
Ib gm
. Btu cal
lb°F gm°C
Multiply
cal/hr sq cm
Btu/hr sq ft
watts /sq cm
Btu/ hr sq ft
Btu ft
hr sq ft °F
Btu in.
hr sq ft °F
Btu ft
hr sq ft °F
cal era
hr sq cm °C
watts cm
sq cm °C
Blu ft
hr sq ft °F
heat content in Btu/lb
heat content in cal/gm
specific heat in Btu/lb °F
by
3.687
0.2712
3170
0.0003154
12
0.0833
14.88
0.0672
57.79
0.01731
0.556
1.80
1
to obtain
Btu/hr sq ft
cal/hr sq cm
Btu/hr sq ft
waits /sq cm
Btu in.
hr sq ft °F
Btu ft
hrsq ft °F
cal cm
hr sq cm °C
Btu ft
hr sq ft °F
Btu ft
hr sq ft ° F
watts cm
sq cm °C
heat content in cal/gm
heat content in Btu/lb
specific heat cal gm °C
E-4
-------
E.7 Conversion Between Different Units
We have gathered below quantities of the English and engineering systems of units that
are commonly found in the literature on air pollution. Our intention is to list them in
such a way that their equivalent in the MKS system of units can be found quickly.
Quantities which are listed in each horizontal line are equivalent. The quantity in the
middle column indicates the simplest definition or a useful equivalent of the respective
quantity in the first column.
1 acre 1/640 mi2 4.047 x 103 m2
1 Angstrom (A) 10'8 cm 10~'° m
1 atmosphere (atm) 1.013 X 106 dyn/cm2 1.013 X 105 N/m2
1 bar (b) 106 dyn/cm2 105 N/m2
1 barrel (bbl) 42 gal, U.S.A. 0.159 m3
I boiler horsepower 3.35 x I04 Btu/hr 9.810 x 103 W
1 British Thermal Unit (Btu) 252 cal 1.054 x 103 J
1 Btu/hour 1.93 x 106 erg/sec 0.293 W
1 calorie (cal) 4.184 X 10'7 erg 4.184 J
1 centimeter of mercury (cm Hg) 1.333 X 104 dyn/cm2 1.333 X 103 N/m2
1 cubic foot, U.S.A. (cu ft) 2.832 X 104 cm3 2.832 x 10'2 m3
1 dyne (dyn) 1 g-cm/sec2 10"^ N
1 erg 1 g-cm2/sec2 10"^ J
1 foot, U.S.A. (ft) 30.48 cm 0.3048 m
1 foot per minute (ft/min) 1.829 x 10~2 km/hr 5.080 x 10'3 m/sec
1 gallon, U.S.A. (gal) 3.785 x 103 cm3 3.785 x 10^ m3
E-5
-------
E.8 Pressure
\To
Frorn\
mm Hg
in Hg
in H2O
ft H2O
atm
Ib/in2
Kg/cm2
mmHg
1
25.40
1.868
22.42
760
51.71
735.6
in Hg
0.03937
1
0.07355
0.8826
29.92
2.036
28.96
in H2O
0.5353
13.60
I
12
406.8
27.67
393.7
ft H2O
0.04460
1.133
0.08333
1
33.90
2.307
32.81
atm
0.00132
0.03342
0.00246
0.02950
I
0.06805
0.9678
lb/in2
0.01934
0.4912
0.03613
0.4335
14.70
I
14.22
Kg/cm2
0.00136
0.03453
0.00254
0.03048
1.033
0.07031
1
E.9 Volume
\. To
From^x.
cm •*
liter
m^
irP
ft3
cm 3
1
1000
I x JO"6
16.39
2.83 x 10"4
liter
0.001
1
1000
0.01639
28.32
m3
1 x 10'6
0.00 1
1
1.64 x 10-6
0.02832
in3
0.06102
61.02
6.10 x lO"4
I
1728
ft3
3.53 x ID'5
0.03532
35.31
5.79 X lO-4
1
E.IO Temperature
°C = 5/9 (°F-32)
°K = °C+273.2
°F = 9/5 °C+32
°R = °F+459.7
E-6
-------
E.ll Conversions - Dust Loadings in Flue Gas (Approximate)
Grains per standard cubic foot
Grains per standard cubic foot
Grains per cubic foot of 500° F flue gas
Pounds per 1000 Ibs. of flue gas
Pounds per 1000 Ibs. of flue gas
Pounds per 1000 Ibs. of flue gas
Pounds per 1,000,000 Btu input
Pounds per 1,000,000 Btu input
Pounds per 100 Ibs. of Type 0 Waste
Pounds per 100 Ibs. of Type 1 Waste
Pounds per 100 Ibs. of Type 2 Waste
Pounds per 100 Ibs. of Type 3 Waste
Pounds per 100 Ibs. of Type 0 Waste
Pounds per 100 Ibs. of Type I Waste
Pounds per 100 Ibs. of Type 2 Waste
Pounds per 100 Ibs. of Type 3 Waste
Pounds of flue gas per hour
Standard cubic feet per minute
X 1.87
X 2.20
X 3.45
X 0.53
X 0.29
X 1.18
X 0.45
X 0.85
X 1.01
X 1.30
X 1.84
X 2.80
X 0.54
X 0.70
X 0.90
X 1.50
X 0.22
X 4.50
= Ibs. per 1000 Ibs. of flue gas
= Ibs. per 1,000,000 Btu input
= Ibs. per 1000 Ibs. of flue gas
— grains per standard cubic foot
= grains per cubic foot of 500° F flue gas
= Ibs. per 1,000,000 Btu input
= grains per standard cubic foot
- Ibs. per 1000 Ibs. of Hue gas
= Ibs. per 1000 Ibs. of flue gas
= Ibs. per 1000 Ibs. of flue gas
= Ibs. per 1000 Ibs. of flue gas
= Ibs. per 1000 Ibs. of flue gas
= grains per standard cubic foot
= grains per standard cubic foot
= grains per standard cubic foot
= grains per standard cubic foot
= standard cubic feet per minute
= Ibs. of flue gas per hour
NOTE:
Grains is a measure of weight; 7000 grains = I pound. In these Standards, all expressions
of particulate emissions (dust loadings) are given with the total flue gases (products of
combustion) corrected to 50 percent excess air.
All factors are based on properties of flue gas approaching those of dry air.
For ease of calculations any small differences are ignored and "corrected to
50 percent excess air" and "corrected to 12 percent CO2" are considered equal.
Standard cubic feet is air at 70° F, and 29.92 inches of mercury.
E-7
-------
E.12 Useful Information
Energy equivalences of various fuels:
Bituminous coal - 22.4 X 10$ Btu/ton, 1971-1973
- 21.9 x 106 Btu/ton, 1974
Anthracite coal - 26.0 x 106 Btu/ton
Lignite coal - 16.0 x 10*5 Btu/ton
Residual oil - 147,000 Btu/gal
Distillate oil - 140,000 Btu/gal
Natural gas - 1,022 Btu/ft3
I Ib. of water evaporated from and at 212°F equals:
0.2844 Kilowatt-hours
0.3814 Horsepower-hours
970.2 Btu
1 cubic foot air weighs 34.11 gm.
Avogadro's Number
Gas-Law Constant R
Miscellaneous Physical Constants
6.0228 X 1023 Molecules/gm-mole
1.987
1.987
82.06
10.731
0.7302
Cal/(gm-mole) (°K)
Btu/(Ib-mole) (°R)
(cm3) (atm)/(gm-mole) (
(Ib) (in2)/(lb-mole)
(ft3) (atm)/(lb-mole) (°
E-8
-------
Weight of O2, N2 and Air
Pounds Tons
SCF Gas
i POUND
1 TON
1 SCF GAS
f Oxygen
1 Nitrogen
f Oxygen
I Nitrogen
( Oxygen
1 Nitrogen
1.0
2000.0
0.08281
0.07245
0.0005
1.0
0.00004141
0.00003623
12.08
13.80
24,160
27,605
1.0
Air Density
Gram Mole
1 Ib. Mole
= 1.293 g/lat 1 atmO°C
= 0.0808 Ib/ft3 at atm 32° F
= 22.414 liters at 0°C, 1 atm
= 359.05 ft2 at 32°F, 1 atm
1 Boiler Horsepower:
= 33,475.3 Btu/hr heat to steam
= 34.5 Ibs steam evaporated per hr from and at 212°F
— 44,633 Btu/hr fuel input for 75 percent overall efficiency
- approximately 10 sq ft of boiler heating surface (basis for boiler ratings)
— 139.4 sq ft of equivalent direct radiation
Overall Boiler Efficiency, Percent:
= 0° steam/hr) X (heat content of delivered steam - heat content of feed water)
(fuel input rate) X (gross heating value of the fuel)
E-9
-------
Over-Rate Firing (Common Practice):
Cast iron sectional boilers
Steel fire tube and brick set horizontal return tube (HRT)
Scotch marine boilers (conventional type)
Water tube boilers (small)
Water tube boilers (large, power type)
1 Sq Ft of Equivalent Direct Radiation (EDR):
= 240 Btu/hr for steam heating
= 150 Btu/hr for hot water heating (open system)
= 180 Btu/hr for hot water heating (closed system)
125 percent rating
150 percent rating
200 percent rating
300 percent rating
600 percent rating
• U.S.GOVOlNMEVrf'WVTINGOfHCE; t 991-5..B. is ?<• 0597
E-10
-------
-------
m
T3
fc
03
cn
CD
o
o
Ql
3 :±
0) O
75
r ro
£ £
CO
O
o
> me
S 3
Q) (O
T)
O
5'3
OJ
03
o
-•o
ro
m
3
CD
W
Q)
B
m
9
O
O
CO
> 03
oc
------- |