EPA-R2-73-163
February 1973 ENVIRONMENTAL PROTECTION TECHNOLOGY
Monitoring Instrumentation
For The Measurement
Of Sulfur Dioxide
In Stationary Source Emissions
, \
S — *•---" _ ~
o
Office Of Research And Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
EPA-R2-73-163
Monitoring Instrumentation
For The Measurement
Of Sulfur Dioxide
In Stationary Source Emissions
By
Fredric C. Jaye
Based on Information Obtained from Contract with
TRW Systems Group
One Space Park
Redondo Beach, California 90278
Contract No. EHSD 71-23
Program Element No. 1A1010
Project Officer: Fredric C. Jaye
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared For
Office of Research and Monitoring
Environmental Protection Agency
Washington, B.C. 20460
February 1973
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
-------
Foreword
In March of 1970, the Division of Process Control Engineering
issued a Request for Proposal to provide an "Evaluation of sulfur
dioxide monitoring equipment for stationary sources." The purpose
of this program was to provide the data and evaluation of perfor-
mance under actual industrial application which would permit the
intelligent selection of monitoring equipment for a specific appli-
cation .
On 2 September 1970, TRW Systems Group, Redondo Beach, Califor-
nia initiated work under contract EHSD-71-2J with Dr. Robert M. Statnick,
Division of Process Control Engineering, as the EPA project officer.
With the reorganization of EPA and the subsequent transfer of pollu-
tant measurement development work to the Division of Chemistry and
Physics, National Environmental Research Center, Dr. Fredric C. Jaye
replaced Dr. Statnick as the EPA project officer in March 1971.
The contract was terminated on 29 September 1971 for the con-
venience of the government at the end of the field test phase (Task G).
The data analysis was performed by members of the Stationary
Sources Emissions Measurements Methods Section of the Division of
Chemistry and Physics.
This final report was prepared by members of that section and
with the exception of task H and conclusions is largely derived from
the monthly progress reports submitted by TRW.
in
-------
ACKNOWLEDGEMENTS
A considerable effort was put forth by the project personnel
to prepare the preliminary specifications and instrument evaluations
in the attachments. This section is to acknowledge the individuals
responsible for the first draft specifications.
APPENDIX TITLE TRW PROJECT PERSONNEL
B-l Proposed General Operating Specifica- J. Eynon
tions for S02 Stack Gas Monitoring
System
B-2 Specific Design, Manufacture and Use G. I. Gruber §
Specifications A. A. Lee
D-l NDIR Instrumentation for Stack Gas E. Tagliaferri
Monitoring
D-2 Electrochemical Instruments for Monitoring R. R. Sayano §
SO-> in Stack Gases E. T. Seo
-------
TABLE OF CONTENTS
INTRODUCTION ' 1
TASK A - DATA SEARCH AND ACQUISITION 4
TASK B - FORMULATION OF INSTRUMENT SPECIFICATIONS 11
Appendix B-l - Proposed General Operating Specifications
for S02 Stack Gas Monitoring 15
Appendix B-2 - Specific Design, Manufacture and Use
Specifications 21
TASK C - ACQUISITION OF DATA ON EXISTING INSTALLATIONS 35
TASK D - PRELIMINARY EVALUATION OF INSTRUMENTATION 38
Appendix D-l - NDIR Instrumentation for Stack Gas
Monitoring 43
Appendix D-2 - Electrochemical Instruments for Monitoring
S02 in Stack Gas 55
TASK E - DESIGN OF TEST PROCEDURES AND EXPERIMENTAL PLAN .... 65
Appendix E-l - Preliminary Sulfur Dioxide Analysis
Methods Evaluation 79
TASK F - PROCUREMENT AND INSTALLATION OF S02 MONITORING
EQUIPMENT 86
TASK G - FIELD TEST PROGRAM 91
TASK H - DATA EVALUATION 108
CONCLUSIONS AND RECOMMFJMDATIONS . 123
-------
INTRODUCTION
The objective of this program was to evaluate and rank commer-
cially available sulfur dioxide monitoring instrumentation for fixed
combustor sources. To accomplish this objective the project was
organized into eight tasks which provided for a logical review of
available instruments and specifications in terms of established use-
criteria, selection and procurement of the most promising instruments
and finally a field test program where the instruments were installed
and operated continuously under real, fixed combustor stack conditions.
Instrument performance was rated on criteria including accuracy,
specificity, reliability, maintainability, ruggedness and other selected
performance criteria. The ranking'of instruments was necessarily
dependent on the order of importance the potential user places on the
established performance criteria.
Task titles with brief content descriptions were as follows:
Task A - Data Search and Acquisition - Obtain from manu-
facturers and other sources available information on
commercial SO continuous monitors suitable for monitoring
stationary sources.
Task B - Formulation of Instrumentation Specifications -
Formulate specifications that continuous sulfur dioxide
analyzers must meet for R/D studies and stationary power
plants.
Task C - Acquisition of Data on Existing Installations -
Obtain field evaluation data to aid in assessment of
instrument applicability, minimize costs of field testing,
obtain pertinent data on sampling and handling techniques
and desirable instrumentation modifications.
-------
Task D - Evaluation of All Instruments - Compare vendor
specifications and user reported performance data with
the specifications developed in Task B and rank instru-
ments.
Task E - Design of Test Procedure and Experimental Plan -
Generate detailed: l) design plans, 2) instrument pro-
cedures, 3) operation plan, U) control analysis, and
5) document formats.
Task F - Procure and Install SO,, Monitoring Equipment -
Procure the selected SO monitoring equipment, sampling
manifolds, and install them.
Task G - Field Test Program - Conduct the field test
program for six months.
Task H - Data Evaluations and Conclusions - Provide a
detailed evaluation of the applicability and limitations
of the specific instrumentation tested.
At the project kickoff meeting held on 7 October 1970 at TRWs
Redondo Beach Facility and attended by Jim Dorsey and Robert Statnick
of EPA along with Arnie Grunt, Paul Testerman, Pete White, Al.Lee,
Fred Harpt, Jim Eynon, Jerry Gruber, Don Wolpert, and Norm Garner of
TRW, Project Objectives, tasks and technical approach were reviewed
and restated. ,
During this meeting the following project philosophy and points
of emphasis were discussed.
If at all practical the field testing of instruments
will be accomplished on a pulverized, coal-fired boiler
of at least 50 megawatts or larger generation capacity.
The rationale for the decision is that the majority of
power generating plants do burn coal which may give rise
to unique and significant interferences from the metals
and other coal constituents.
-------
Fuel must, in general, be consistent in type and composition,
as would be the case from a single mining area.
The S0? concentration of interest will range from 200 to
700 ppm.
Where practical, the instrument manufacturers sampling and
pretreatment methods will be utilized.
-------
TASK A
Data Search and Acquisition
Persuant to Bureau of the Budget survey clearance No. 158-S71010,
letters of inquiry were sent to all identified and potential manufac-
turers and representatives of SO monitoring instruments (over 220).
From this data search, a list of 18 companies claiming to manufacture
or sell SO monitoring instruments suitable for stationary source
monitoring was compiled. These companies are listed in Table A-l.
Table A-2 contains a number of companies who indicated that they were
in the development or marketing survey stage as of 1 January 1971.
Of these companies several units have since become available in
the commercial market. These companies are:
Environmental Data Corporation
Melloy Laboratories
Theta Sensors
A number of gas chromatography equipment manufacturers had indi-
cated that while they do not market a G.C. system specifically designed
and produced for SO monitoring, that process G.C. systems could be
produced to accomplish the task. Table A-3 shows these manufacturers
and some brief comments furnished by them.
As a result of several months of survey effort the following
sulfur dioxide monitoring equipment for stationary source use was
considered as commercially available as of February 1, 1971. See
Table A-k.
At the conclusion of this identification task, sufficient
information and commercially available monitoring equipment were
available to begin Task B, Formulation of Instrument Specifications.
-------
TABLE A-l
MANUFACTURERS OR VENDORS OF SO STACK GAS MONITORS
AERO-VAC CORPORATION
P. 0. Box M*7
Troy, N. Y. 12l8l
(518) 271*-5850
ANALYTICAL INSTRUMENT DEVELOPMENT
250 S. Franklin St.
West Chester, Pa. 19380
(215) 692-1*575
ATLAS ELECTRIC DEVICES CO.
l+lll* No. Ravenswood Ave.
Chicago, 111. 60613
(312) 327-1+520
BARTON, ITT
580 Monterey Pass Rd.
Monterey Park, Calif.
(213) 283-6501
9175^
BECKMAN INSTRUMENTS, INC.
2500 Harbor Blvd.
Fullerton, Calif. 9263!+
871-1*81+8
BENDIX CORP.,
PROCESS INSTRUMENTS DIV.
P. 0. Drawer 1*77
Ronceverte, W. Va. 21*970
CALIBRATED INSTRUMENTS, INC.
17 W. 60th St.
New York, N. Y. 10023
(212) 21*5-5590
CANADIAN RESEARCH INSTITUTE
85 Curlew Dr.
Don Mills, Ontario, Canada
E. I. DU PONT COMPANY
1007 Market St.
Wilmington, Del. 19898
(212) 61+9-9600
DYNASCIENCES CORPORATION
9601 Canoga Avenue
Chatsworth, Calif. 91311
(213) 31*1-0800
ENVIROMETRICS INC.
13311 Beach Avenue
Marina Del Rey, Calif. 90291
(213) 391-8268
INTERTECH CORPORATION
262 Alexander St.
Princeton, N. J. OQ'ykO
(609) 1*52-8600
KIMOTO ELECTRIC CORP. LTD.
1*2 Funahashi-Cho
Tennoji-Ku
Osaka, Japan
MINE SAFETY APPLIANCE CO.
201 N. Braddock Ave.
Pittsburgh, Pa. 15208
(1*12) 21*1-5900
OLSON HORIBA
1021 Duryea Ave.
Irvine Industrial Complex
Santa Ana, Calif. 92705
5l+0-787l+
SCIENTIFIC INDUSTRIES, INC.
150 Herricks Rd.
Mineola, N. Y. 11501
(516) 746-5200
SCOTT AVIATION/DAVIS INSTRUMENTS
Charlottesville, Va.
(703) 973-5366
VAN WATERS & ROGERS/WILL SCIENTIFIC
1363 S. Bonnie Beach Place
Los Angeles, Calif. 9005!+
(213) 269-9311
-------
TABLE A-2
SULFUR DIOXIDE MONITORS CURRENTLY UNDER DEVELOPMENT
AS OF 1 JANUARY 1971
Manufacturer
Anacon Inc.
62 Union St.
Ashland, Mass.
Bacharach Instrument Co.
625 Alpha Drive
Pittsburgh, Pa.
Burhans-Sharpe
P. 0. Box 3906
Seattle, Washington
Combustion Equipment Association
6l Taylor Reed Drive
Glenbrook, Connecticut
Environmental Data Corporation
608 Fig Avenue
Monrovia, Calif.
Environmental Instruments Co.
13U6 Willow Road
Menlo Park, Calif.
General Monitors, Inc.
3019 Enterprise St.
Costa Mesa, Calif.
Gow-Mac Instrument Co.
100 Kings Road
Madison, N. J.
Kimoto Electric Corp., Ltd.
^2 Funahashi-Cho
Tennoji-Ku
Osaka, Japan
Melloy Labs
6631 Iron Place
Springfield, Virginia
Comments
Currently have instrument in
development.
Purchased rights for manufacture
and sale of U.V. in-stream analyzer
from Weyerhaeuser Co.
Hold the rights to manufacture and
sell Barringer SO in-stack monitor.
Has one NO in-stack analyzer on a
power plant.
Developing a mass spectrometer
analyzer system for air and stack
gases.
Developing solid state SO and NO
monitors which are based upon changes
in the chemical activity of solid
state sensors.
Manufacture U.V., coulometric and
conductometric instruments.
Flame photometric total sulfur
analyzer.
-------
TABLE A-2 (Continued)
Manufacturer
Monitor Labs, Inc.
10^51 Roselle Street
San Diego, Calif.
Research & Development Products
1808A Harmon Street
Berkeley, Calif.
Rockland Instrument Corporation
P. 0. Box 205
Pearle River, New York
Scientific Research Instruments
6707 Whitestone Road
Baltimore, Maryland
* Theta Sensors
1015 No. Main St.
Orange, Calif
2000 Corporation
5899 So. State St.
Salt Lake City, Utah
Wilkens-Anderson Co.
1*525 W. Division St.
Chicago, 111.
Wilks Scientific Corp.
1UO Water St.
So. Norwalk, Connecticut
Comments
An emission spectrophotometer that
measures radiation from heated SO
in the I.R. region.
Developing an SO monitor.
Have a reflectance colorimetric
tape system under development.
Developing a mass spectrometer
monitoring system.
Developing a fuel cell type
monitor.
Have a wet chemical analyzer under
development.
In the early stages of developing
a stack gas monitor.
Developing an I.R. in-stack
monitor.
*Now commercially available.
-------
TABLE A-3
MANUFACTURERS OF GAS CHRQMATOGRAPHIC INSTRUMENTS EXPRESSING
INTEREST IN SO STACK MONITORS
Analytical Instrument Development Inc.
250 S. Franklin St.
Westchester, Pa. 19380
Antek Instrument Inc.
P. 0. Box 7903
Houston, Texas 77007
Applied Automation Inc.
3838 S.E. State St.
Bartlesville, Oklahoma 7^003
Beckman Instrument Inc.
2500 Harbor Blvd.
Fullerton, Calif. 9263^
Chemical Data Systems Inc.
Oxford, Pa.
Process Analyzer Inc.
6^00 W. Freeway, Suite ^00
Houston, Texas 77036
Tracor, Inc.
6500 Tracor Lane
Austin, Texas 78721
Unico Environmental Instruments
P. 0. Box 590
Fall River, Mass. 02722
Carle Instruments, Inc.
llUl E. Ash Ave.
Fullerton, Calif.
Produce portable G.C.
Developing G.C. No field
test. No users.
Developing G.C. No field
test. No users.
Process chromatograph used
on stacks with customer
developed technology.
Developing G.C. No field
test. No users.
Process chromatograph. No
field test.
Process chromatograph. No
field test. No users.
Developing G.C. No field
test. No users.
Column 10$ didecyl phthalate
on chromosorb T. UO-60 mesh.
Can determine CO , SO , and
TT A C. tL.
-------
TABLE A-h
UV-Vis
DuPont
NDIR
Intertecha
Bendixa
Beckman
M.S.A.
Horiba, Ltd.
Electroanalytical
Dynasciences a
SS330/CS1000
Enviro Metrics S-64a
Calibrated Instrumentsa
INSTRUMENTS COMMERCIALLY AVAILABLE
AS OF 1 FEBRUARY 1971
Comments
Scientific Industries Model 70
ITT Barton Model
Only currently available UV-Vis
analyzer suitable for stack gas
monitoring. Good reports have
been received from users.
Good reports have been received on
this instrument. Appears to be of
good design and construction.
No users located to date. Some new
concepts, etc. were utilized in
design and construction. Appears to
be an excellent instrument.
Most user information satisfactory.
No user information obtained to date.
In general M.S.A. has a reputation
for good construction.
Instrument to be marketed in the U.S.
in the near future. Distributor will
send information as soon as received
from Japan.
Fuel cell type.
Same as above.
Conductometric type.
Conductometric type.
Coulometric type.
-------
10
TABLE A-k (Continued)
Other Comments
Aero-Vac Model 170-200 Mass spectrometer.
Environmental Data In stream spectrometric type.
Instruments recommended for inclusion in the field evaluation.
Based on our search and verbal contacts with the vendor we were
unable to ascertain that an SO in-stack monitor had been produced.
-------
n
TASK B
Formulation of Instrument Specifications
Prior to any laboratory or field testing of SO monitoring
equipment, it was necessary to establish those criteria which are
considered important in the evaluation of the usability of such
equipment and to propose target goals for each area. As of the
time of initiation of this program, the only widespread utilization
of SO monitoring equipment was in the Federal Republic of Germany
and some use of ambient air monitors here in the United States.
Therefore, we were required to formulate these general requirements
on the basis of limited information and rely heavily on engineering
judgment.
Table B-l is a brief summary of the external environmental
conditions based on a cross section of United States weather. This
information was taken from "Climate Atlas of the United States"
(June 1968), U. S. Department of Commerce. The table clearly indi-
cates that all instrumentation will have to be protected and that
the sampling system must be protected, insulated and have sufficient
heating power to keep the sample in gaseous form.
General Instrument Specifications -
The following instrument generalizations, exclusive of sample
probe and preconditioner were presented for consideration:
Selectivity-measure SO content and relate the content
measured to the original stream SO content—define inter-
ferences and required correction or sample pretreatment.
Routine operation - unmaintained for extended periods
of time (up to 1 week excluding recording device) with
specified stability, accuracy and reproducibility (see
below).
Calibration - capable of calibration and standardization
by a power plant instrument technician, without special
training or unique equipment.
-------
12
Maintenance/Repair - capable of being maintained by a
qualified power plant instrument technician or be subject to
readily available contract maintenance service. Parts avail-
ability must be on a maximum 2^-hour basis from nearest supply
point, for any and all parts required for repair. Shelf life
on parts stock in power plant instrument warehouse shall be
minimum of two years. Special charts and inks are considered
as "parts." Mean time to repair the unit should have a goal
of 2 hours while maximum time to repair shall have a goal of
8 hours.
Reliability - long term reliability of SOg monitor
system must be compatible with the monitoring needs of the
power plant, i.e., instrument must operate most of the year
with minimum repair/replacement; normal annual preventive
maintenance/refurbishment shall be accomplished during the
annual down time of the power unit.
Output signal - instrument outputs shall be linear analog
or equivalent signals capable of immediate interpretation in
SO cone, units (vol.).
Recording - linear analog or equivalent in SO vol. cone.
units by fractional hour, with minimum 2k hour chart capability
+1.5$ of indicated instantaneous value. Accessibility for chart
replacement/notation via hinged instrument case required.
Operating Specifications
Operating specifications for specific instruments may have to
be developed for each general type depending on principle. Discus-
sions with power plant engineers and TEW project personnel indicated,
however, that certain tentative common operating specifications
may be delineated at this time. For example, operating accuracy is
derived from the combination of absolute accuracy, reproducibility,
drift, interference and many other factors, and must be within
-------
13
certain known limits if the final measurement is going to be meaning-
ful. The following preliminary instrument operating specifications
were developed to meet the need for a recorded SO concentration
value that relates meaningfully to the effluent SO concentration.
Accuracy - ±2% of span as measured against standard
referee chemical test procedure.
Reproducibility - 1.0$ of span
Zero Drift - 2.U hr +0.8$ of span
7 day ±2.2$ of span
Span Drift - 2k hr ±0.8$ of span or ±0.002$ of SO (which ever
is larger)
7 day +2.2$ of span or ±0.006$ of SO (which ever
is larger)
Sensitivity - sufficient to detect a 5 ppm change in SO
concentration when the instrument is operating
in the 200-700 ppm range(s).
Temperature, Voltage,
and Pressure Variations - sufficient stability to meet above
specifications when reasonably
protected from adverse power plant
environment. (Nominal Operating
Conditions--
Temperature 70° + 20 F,
Relative Humidity 0-90$,
Voltage at Stated Value + 20$).
Utilizing these preliminary specifications, the intended use
environment, the manufacturers' specifications and generally accepted
state-of-the-art, the TRW project team generated two more detailed
specifications. Appendix B-l provides a set of general operating
specifications. Appendix B-2 provides a set of guidelines with re-
quirements for design, fabrication, performance and testing that would
be required in a controlling document for manufacture.
-------
TABLE B-l
U. S. CHmdtic Conditions by Region
Location:
National Area and Cities
West Coast
Seattle, Wash.
Los Angeles. Calif.
{Inland areas incl. mountain
areas)
Southwest
Las Vegas, Nevada
Flagstaff, Arizona
El Paso, Texas
Northwest
Boise, Idaho
Mountain States
Denver, Colo.
Western Plains
Blsmarch, N. Dakota
Wichita, Kansas
Midland, Texas
Middle States (Midwest to
Gulf Areas)
Minneapolis, Minn.
Chicago, Illinois
Pittsburgh, Pa.
St. Louis, Missouri
New Orleans, La.
Aooalachlan Area
New England
Caribou, Maine
Boston, Mass.
Eastern Seaboard
Hew York, N.Y.
Washington, D.C.
Charleston, S.C.
Miami, Florida
Mean Mean
inual Days Annual Days
F or Greater 32°F or Less
0-5 10-20
20-120 20-210
2-90 1P-210
2-20 90-240
0-5 150-24Q
20-90 60-180
5-90 10-18"
2-20 120-150
2-10 10-1RO
5-60 0-90
Extreme Daily Ten'oerature °F and R.H.
January
High
67
90
76
65
77
63
76
60
74
83
58
67
75
77
83
75
51
72
72
79
82
83
Low
3
28
8
-30
-8
-17
-29
-45
-15
-8
-34
-20
-18
-22
13
-15
-32
-13
-6
-14
10
31
R.H.-,
90
65
40
55
(JO
80
60
70-80
60-70
60
70-80
70-80
70-80
7Q-80
80
70
80
8D
70
70-80
80
80
High
100
109
117
93
109
111
104
114
117
107
108
105
103
115
102
103
95
10'
106
106
104
95
July
Low
46
49
56
32
56
40
42
32
53
71
44
49
42
51
Pf
45
40
50
52
5?
PI
65
Average Annual
Fastest Wind
Total -Precipi tation. Inches
R . H . 5
90
70
20
40
50
40
50
65
55
50
70
70
70
65
80
80
70
80
80
80
50
80
(as rain)
100
17
5
1"
12
?5
16
18
20-25
22
28
36
40
34
62
«
40
43
45
45
46
60
Snow
0-12
<12
0
12-400
0-60
<6
24
22
24-200
24
36-400
60
1-36
12-24
a
0-1 OD
36-60
36-60
36
12-24
0
12-100
24-100
100
36-60
0-60
36
12-24
< 1
0
Direction
S
N
SW
NW
U
W
V
N
WSV
W
NE
Nu
?W
SE
-
NU
^
SE
SE
-
E
Velocity
M.P.H.
65
19
7?
70
61
65
7?
100
67
92
87
73
82
9B
-
76
87
113
78
76
132
-------
15
APPENDIX B-l
PROPOSED GENERAL OPERATING SPECIFICATIONS
FOR
SO STACK GAS MONITORING
by J. Eynon
1. Scope
This document is intended to delineate the desired general
operating specifications for SO stack gas monitoring instrumenta-
tion as a guide for the evaluation and selection of such equipment.
The equipment is for continuous monitoring of SO in the effluent
gases from the combustion process in electric utility power plants
using fossil fuels. It is to be comprised of a sampling system,
including any necessary sample conditioning equipment, and a measur-
ing system including an analyzer and recorder.
2. Measuring System
Proposed operating specifications for the measuring system are as
follows:
2.1 Range
0 to 3000 ppm SO in one step is required; provision for selecting
a shorter span, such as 0 to 1000 ppm SO is a desirable option.
2.2 Sensitivity
Sufficient to detect a change of 10 ppm in SO concentration
at each of the span of concentrations covered by the analyzer.
2.3 Readability
Scale of the readout device shall be direct reading in SO
concentration and have sufficient resolution to permit concen-
tration to be read within 10 ppm.
2.U Limit of Error
After the span of the measuring instrument has been set under
fixed conditions of temperature, pressure, and line voltage
using a zero gas and a span gas of known SO concentration, the
-------
16
calibration of the scale is to be checked using four gases of
known SO concentrations corresponding to approximately 25$,
50$, 75$, and 100% of full scale. The readings should agree with
these known S02 concentrations with a limit of error of ±1$.
Over a period of 7 days of unadjusted operation, the readings
of the analyzer as compared with values obtained by the standard
referee chemical test procedure, shall have a maximum error
of ±5$.
2.5 Zero Drift
Definition: The change with time in instrument output at a
constant concentration of 0.00 ppm SO over a given period of
unadjusted continuous operation of the analyzer.
Specification:
2k hour period: ± 1% of span
7 day period: + 2$ of span
2.6 Span Drift
Definition: The change with time in instrument output at a
constant concentration of SO over a given period, exclusive
of zero drift.
Specification:
2k hour period: + 1% of span
7 day period: + 2% of span
2.7 Noise
For a constant value of SO concentration in the sample gas,
short term fluctuations in the recorded readings shall not
exceed the equivalent of 0.5$ of full scale during a period
of 10 minutes.
2.8 Repeatability
Definition: Differences in successive readings obtained for a
gas sample containing a constant amount of SO when it is applied
to analyzer alternately with a gas sample containing no SO .
-------
17
Specification: +.5% of span.
2.9 Selectivity
Analyzer must be selective for SO and discriminate against
variations of other constituents in the background gas from
fossil fuel combustion. These constituents typically include,
but are not limited to the following:
Constituent Concentration Range
SO 0 to 3000 ppm
CO 0 to 50 ppm
C02 12 to 18$
H20 k to lH
N0x (as N02) 50 to 800 ppm
SO 0 to 100 ppm
02 1 to 5$
N Balance
Response of analyzer to such variations must not be greater
than 1$ of span or 10 ppm SO .
2.10 Ambient Pressure Influence
Variations in atmospheric pressure of ± 1 inch of mercury from
the value at which the SO monitor is calibrated shall not cause
more than ±0.3$ variation in the reading.
2.11 Ambient Temperature Range
The equipment, exclusive of the sampling probe located in the
stack, shall be capable of operating over an ambient temperature
range from 0°C to 50°C (32° to 122°F).
2^12 Ambient Temperature Influence
Variations in ambient temperature shall not cause variations
in readings of SO monitor in excess of +0.1$ of reading per
degree Centigrade, assuming maximum rate of temperature change
is 10°C/hour (l8°F/hour).
-------
18
2.13 Power Requirement
The equipment shall operate on 120 volts + 20$, AC, 50 to 60 Hz.
2.lk Line Voltage Influence
Variations in line voltage from a nominal voltage of 120 volts
shall not cause variations in reading of the SO Monitor in
excess of 0.01$ per volt.
2.15 Lag Time
Definition: The time interval from a step change in SO input
concentration at the instrument inlet to a readable change in
recorded output.
Specification: Not more than 10 seconds.
2.16 Response Time
Definition: The time interval from a step change in SO input
concentration at the instrument inlet to a recording of 9°$ of
the ultimate recorded output.
Specification: Not more than 1 minute.
2.17 Local Indication
An indicator, calibrated to read directly in SO concentration,
shall be provided in the monitor location.
2.18 Remote Indication
A DC output signal shall be provided to operate a potentiometric
type recording instrument. Minimum output signal level should
be 0 to 10 millivolts, corresponding to the full span of SO
concentration measured by the analyzer.
2.19 Sample Temperature
The monitor should be capable of operating with the sample cell
at a temperature in excess of the expected naximum dew point of
the sample (See Section j).
-------
19
2.20 Construction
2.20.1 The construction and finish of the equipment are to be
suitable for the industrial environment of an electric
power plant.
2.20.2 Components of the equipment which will contact the sample
gas are to be constructed of materials which are inert
to the constituents of the sample gas, i.e., which do not
alter the SO content of the sample and are corrosion
resistant.
2.20.3 Modular construction insofar as possible is preferred so
that faults may be quickly located and corrected by re-
placement of the faulty module.
2.20.!+ Electronic circuits preferable shall be of modern solid-
state design with modular printed circuit boards. All
components shall be conservatively rated; and fabrication
and workmanship shall conform to good commercial practice.
2.20.5 All necessary equipment for adjustment and control of
sample flow rate and reagent supply system (if reagents
are required) are to be furnished with the analyzer.
2.20.6 Provision shall be made so that zero and span test gases
may be connected to analyzer conveniently either by a
valving system or by disconnection and reconnection of a
simple fitting.
2.21 Mounting
The instrument shall be suitable for either flush (panel) or
wall mounting.
-------
20
3. Sampling System
The sampling system including any necessary sample conditioning
equipment shall not alter the SO content of the sample and shall
permit operation of the analyzer with sample gases having typical
composition and temperature as follows :
SO 0 to 3000 ppm
Temperature 200° F to 900° F
Fly-ash content up to 5 grains per standard cubic foot
C02 12 to
CO 0 to 50 ppm
H0 k to
NO (as NO) 50 to 800 ppm
.JC
SO 0 to 100 ppm
02 1 to %
N Balance
HC 0 to 1500 ppra
-------
21
APPENDIX B-2
SPECIFIC DESIGN, MANUFACTURE AND
USE SPECIFICATIONS
by G. I. Gruber and A. A. Lee
INTRODUCTION
The following document is to serve as criteria for the selection
of SO Monitoring Systems for fixed combustion sources and may serve
as a guide for the manufacture of such instruments.
1. SCOPE
1.1 Scope
This specification establishes the requirements for the design,
fabrication, performance, and testing of a fixed location SO gas
concentration, sample measuring and recording instrument system.
(SO fixed combustor stack gas monitor) The system is to be used
to continuously determine quantitatively the concentration of S02 gas
in exit gas from stationary power plants using fossil fuels.
2. APPLICABLE DOCUMENTS
2.1 Government Documents
The following documents are intended as a guide to good practice
in design, fabrication, testing and manufacturers literature preparation
for the purposes of this specification.
SPECIFICATIONS
Military
MIL-C-5015 Connectors, Electrical AN Type
MIL-S-68?2 Soldering Process, General Specification for
-------
22
STANDARDS
Federal
Fed Std No. 596 Colors
Military
MIL-STD-15
MIL-STD-16
MIL-STD-U5U
MIL-STD-202
MIL-STD-1^72
MS33586
Electrical and Electronic Symbols
Electrical and Electronic Reference Designation
STD General Requirements for Electronic
Equipment
Test Methods for Electronic and Electronic
Components Parts
Human Engineering Design Criteria for
Military Systems, Equipment and Facilities
Metals, Definition of Dissimilar
Other
(a) Emissions From Coal Fired Power Plants: A comprehensive
summary U. S. Department of Health, Education, and Wel-
fare, NAPCA, Cuffee and Gerstle 196?
(b) Performance characteristics of Instrumental Methods for
Monitoring Sulfur Dioxide, JAPCA Volume 19, No. 8,
August 1969, Rodes, Palmer et. al U.S. Department of
Health, Education, and Welfare Public Health Service.
3. REQUIREMENTS
3.1 Qualification
The unit described by this specification shall be a product which
has passed all the examinations and tests specified in the following.
3.2 Materials
The unit to be constructed of materials of the corrosion-resis-
tant type or suitably treated to resist corrosion caused by NO . SO ,
SO , or H SO, , salt spray, or atmospheric conditions likely to be
met in storage or normal service.
-------
23
3.2.1 Dissimilar Metals
Unless suitably protected against electrolytic corrosion,
dissimilar metals shall not be used in intimate contact with each
other. Dissimilar metals are defined in Standard MS33586.
3.2.2 Design and Construction
The monitor system may be a combination pneumatic-electrome-
chanical/ electrochemical/ optical-electronic instrument. It shall
consist of the following major systems:
a) Sampling probe, b) sampling lines and valves, c) sample gas
pump, d) analyzer, e) recorder, f) calibration and back purge
system as required by the sensing technique employed.
3.2.2.1 Derating of Electronic Parts
All electronic parts should be derated as follows:
(a) Capacitors: Minus 25 percent of working voltage.
(b) Resistors: Minus 25 percent of working power dissipation
plus an additional minus 25 percent for precision resistors
(tolerances less than plus or minus 1 percent).
(c) Transistors: Minus 25 percent of working power or current.
(d) Transformers and Inductors: Minus 10 percent of voltage
and 10 percent of current ratings.
(e) Relay and Switch Contracts: Minus 10 percent of a maximum
current - allow additional derating for elevation above
sea level.
(f) Diodes and Rectifiers: Minus 25 percent of current or
voltage, or minus 25 percent of power, whichever is
greater. In all cases use lowest possible operating
temperature.
Derated operating values of electronic parts are not to be exceeded
whenever the equipment is subjected to the service conditions specified
herein. As a guide, even under the maximum load conditions, the
recommended safe load rating of the part is not to be exceeded.
-------
3.2.2.2 Power
The system should be designed to operate from a 120V+ 15 per-
cent 6o±l H, AC power source.
3.2.2.3 Indicators
The instrument shall be equipped with a readout meter. The
meter shall be installed on the face of the detector case in such a
position that the operator's line of sight to the meter will not be
obstructed during normal use including calibration. The meter scale
shall have at least 100 divisions (preferably 200) between 0 and 1000
ppm. (Scale markings shall be 0 to 1000-ppm.)
3.2.2.4 Analyzer Integration
The monitor shall be integrated into a unit that will provide
for simultaneous reading of the meter and the recorder.
3.2.2.4.1 S0_ Monitor
2
The analyzer shall be used to detect SO gas from power
plant stacks. The analyzer shall be calibrated with the use of fixed
concentration sources of S0? gas and shall cover its range in a minimum
of 4 calibration points.
Range App SO^ Concentration
2 ' *"^"~™~~~—^~~
Low Range 0-200 ppm
STD Range ^50 ppm
High Range 700 ppm
Span 1000 ppm
3.2.2.5 Interferences
The SO analyzer should not indicate a total error reading
greater than 5 ppm or 1% of span whichever is greater for all consti-
tuents other than SO of the stack gas. The composition normally
found in a stack gas include but are not limited to the following
gaseous substances:
-------
25
App. Range
NO 100-700 ppm
NO 0-100 ppm
CO 0-50 ppm
co2 12-1856
H20 k-lk%
o2 1-16$
SO 200-1000 ppm
0-100 ppm
N2
Other characteristic stack gas substances are shown in tables
1, 5, and 7 of Reference (a) (under 2.0).
3-2.2.6 Calibration of Monitor or Analyzer
The unit shall be designed so that calibration and preventive
maintenance may be performed on it with a minimum disassembly, other
than removal of minor fittings.
3.2.2.7 Calibration Adjustments
Calibration adjustments should be of the locking screw type,
or where adjustments are mounted internally on sub-chassis, circuit
boards, etc., adjustments nay be of the trimmer type employing a
worm drive screw mechanism capable of retaining its setting when sub-
jected to shock and vibration. External adjustments which must be
operated by the user should have control knobs and shall be equipped
with hand operated locking devices. Such controls should be conveniently
located.
3.2.2.8 Soldering
Soldering shall be in accordance with good practice.
(Ref. MIL-S-6872)
3.2.2.9 Connectors
All external electrical connectors shall be of good quality
and of the lock-on type. (Ref. MIL-C-5015)
-------
26
3.2.2.10 Wire and Wire Harness
Wire used to accomplish electrical interconnection between
components or parts within the systems may be routed individually
or as a part of wire harness. Such wire shall meet the requirements
of good commercial practice. All wires shall meet Underwriters Labora-
tory requirements for insulation.
3.2.2.11 Jumpers
Jumpers not exceeding h inches in length may be fabricated
from uninsulated solid copper wire having a cross sectional area not
less than the cross sectional area specified for AWO-2U wire. Where
required to provide adequate dielectric strength, such jumpers may
be covered with an insulating sleeve.
3.2.2.12 Self-tapping Screws
Thread-cutting screws shall not be used in the design;
however, self-tapping screws may be used to mount identification
plates.
3.2.2.13 Calibration
The monitor system should be capable of automatic calibration
or the necessary equipment for calibration should be supplied with the
system. In the case of an automatically calibrating system the use of
automatic feature for the long term stability checks (2.k hour and
7 day) etc. will be allowable.
3.2.2.14 Sampling Flow Rate
The unit shall be capable of providing a sampling flow rate
of U liters (NTP) per minute minimum, or twice that required for the
specific instrument, which ever is greater.
3.3 Shelf Life
The instrument and stored spare parts shall have the capability
of a shelf storage life of not less than 2 years. To place the stored
detector in use shall require only cell activation and detector
calibration.
-------
27
3.4 Maintenance
The unit shall be capable of operation for' 1 month on a 2k hour
per day basis without any maintenance and with weekly unit calibration.
Replenishment of normal operating supplies is not considered mainte-
nance. The unit shall be designed and constructed to permit maximum
accessibility to parts requiring maintenance or replacement with a
minimum use of special tools. It shall be capable of being maintained,
repaired, or replaced with a minimum use of special tools. The unit
shall insofar as possible be of modular construction, with each module
capable of "off-the-shelf" replacement, and shall be capable of trouble
diagnosis to the module level with vendor supplied maintenance equip-
ment.
3.5 Human Engineering
The analyzer and arrangement of components therein shall conform
to the requirements of good human engineering practice. (Ref. MIL-STD-1^72)
3.6 Performance
The unit shall have the following performance characteristics:
3.6.1 System Performance
The total system performance, from probe inlet through the
recorder system should be such that the instrument readout versus
actual system SO input should not vary more than 3$ during any
24 hour period or more than 5$ during a seven day period.
3.6.2 Monitor
The monitor shall be capable of reliably detecting SO with
a minimum sensitivity of 5 ppm by volume. They shall cover the range
of 0-1000 ppm in 3 steps. (Other steps optional) Readings shall
have a maximum error of less than 5 percent or 5 ppm, whichever is
greater for a period of 7 days of unadjusted operation. Linearity
shall be defined by "best fit" straight line starting at zero ppm.
Wo reading shall deviate from this line more than ±5 percent of the
readings or 5 ppm, whichever is greater. Instruments that do not
-------
28
have a linear response must be provided with a direct method of
converting response to ppm.
3.6.3 Stability
Span stability shall be within .+ 0.8% of Span for 2k hours and
+ 2% of Span for 7 days of unadjusted operation.
3.6.U Response and Sensitivity
The monitor shall respond to a 5 ppm change in SO concentration
range of 200-700 ppm. Response, is defined as the time for the indicated
valve of SO to reach 90% of the true value when switching from zero
gas to 75% span gas at the instrument inlet. The time shall not exceed
90 to 120 seconds. Mayimum time for recovery (return to 7.5% of span
when zero gas is introduced) shall be less than 90 to 120 seconds.
3.6.5 Inlet and Sample Handling System
The monitor system should be offered with a probe and sample
handling system made of material inert to H SO, etc. which is capable
of: a) having the probe operate within the stack environment and
b) sample transfer and conditioning outside the stack atmosphere so as.
to not influence the SO concentration of the sampled stack gases.
3-7 Environmental Conditions
The unit shall be capable of withstanding the environmental
conditions indicated in Table 1. The inclusion of internal heaters
is permitted. In addition, the unit shall be capable of satisfactory
operation in an environment of 120 degrees F (50°C) ambient conditions
for an 8 hour period.
3.7.1 Case
The case of the unit shall be constructed and insulated in such
a manner that the unit will operate within its specified accuracies
for a minimum of 2k hours at a temperature of 32°F (0°C). This may
be accomplished by the use of an internal heater, or any other accep-
table means.
-------
29
3.8 Intel-changeability and Replaceability
Assemblies, components, and parts having identical part numbers
shall, where practicable, meet the requirements for an interchange-
able item as defined in J.8.1. Where interchangeability is not prac-
ticable, the requirements for a replacement item defined in 3.8.2
shall apply.
3.8.1 Interchangeable Item
When two or more items possess such functional and physical
characteristics as to be equivalent in performance and durability
and are capable of being exchanged one for the other without alter-
ation of the items themselves or of adjoining items except for
adjustment, and without selection for fit or performance, the items
are interchangeable.
3.8.2 Replacement Item
An item which is functionally interchangeable with another item,
but which differs physically from the original part in that the instal-
lation of the replacement part requires operations such as drilling,
reaming, cutting, filing, shimming, etc. in addition to the normal
applications and methods of attachment, is known as a replacement
item.
3-9 Interference
The unit shall comply, in the 50 kc to 10,000 me range, with the
requirements for susceptibility to rf interference. (Ref: Specifica-
tion MIL-I-26600.)
3.10 Weight
The weight of each component part of the monitor unit, probe,
recorder, etc., shall not exceed ko pounds.
3.11 Finish
The exterior surfaces of the unit shall meet good standards for
finishes of industrial process instrumentation.
-------
30
3.12 Identification and Marking
The supplier shall attach identification plates to the unit.
The plates shall contain the following information: Manufacturer's
name, Model number, Part number, serial number.
3.12.1 Electrical and Electronic Reference Designations
Electrical and electronic reference designations shall be
in accordance with the commercial abbreviated unit numbering method.
(Ref: MIL-STD-16)
3.12.2 Electronic Symbols
Electronic symbols used on drawings in operation and service
instructions and on the components of the unit shall be in accordance
with good practice. (Ref: MIL-STD-15)
3.13 Workmanship
The monitor shall be constructed and finished in a thoroughly
workmanlike manner. Particular attention should be given to the neat-
ness and thoroughness of soldering, wiring, marking of parts and assem-
blies, welding, brazing, riveting, plating, painting, alignment of
parts, tightness of assembly screws and bolts, and freedom of parts
from defects, burrs, and sharp edges.
3.l4 Operating Manuals
Operation and maintenance manuals shall be supplied with the
unit which give detailed instructions for operation, preventive main-
tenance, calibration trouble shooting, and repair. Detailed sche-
matics of all elements shall be included so that a qualified operator
can acquire a knowledge of the system.
Details of the maintenance manuals shall be such that a qualified
power plant instrumentation technician shall be capable of trouble
shooting using vendor supplied equipment to the repairable module
level within 2 hours minimum repair time and a maximum repair time of
8 hours.
-------
31
4. QUALITY ASSURANCE PROVISIONS
4.1 Responsibility for Inspection
Unless otherwise specified in the contract or purchase order,
the supplier is responsible normally for the performance of all
test requirements.
4.2 Test Procedures
The manufacturer should prepare a test procedure for the imple-
mentation of these tests. These procedures should include a descrip-
tion of the methods and devices to be used to establish and control
the required test conditions and a detailed step-by-step procedure
for conduct of the test.
4.3 Test Samples
The customer may furnish test samples to the supplier, other-
wise accurate gas blends should be utilized by the supplier.
4.4 Component Measurement Tolerances
Instruments used to measure test results should be accurate to
within 10 percent of the tolerance allowed on the quantity being
measured; however, the greatest accuracy required of an instrument
under these conditions is no more than 0.5 percent. This require-
ment does not apply to vapor samples.
4.5 Test Equipment Calibration
The supplier should furnish the customer with evidence of current
calibration of test devices and instruments used in the tests speci-
fied herein.
4.6 Test Data
The supplier should furnish the customer with a certified copy
of the test data.
4.7 Test Conditions
Unless otherwise specified, tests shall be conducted under the
following ambient conditions:
-------
32
(a) Temperature + 59°F to + 95°F
(b) Humidity - 0$ to 90$ relative humidity
(c) Pressure - 10.0 psia to 15-2 psia
^.8 Testing
Testing should consist of the following examinations and tests
as described in U.10.
(a) Examination of Product (^.10.1)
(b) Functional Test (^.10.2)
(c) Environmental Temperature Test (U.10.3)
(d) Environmental Shock and Vibration Test (U.10.U)
(e) Post Shock and Vibration Functional Test (U.10.5)
^.9 Instrument Test Samples
Samples for the full test series of ^.8 should consist of 2 or
more units representative of the production item. Production testing
of all items should be determined by the suppliers Quality Control
System.
k.10 Test Methods
^.10.1 Examination of Product
The unit should be examined for workmanship, identification,
finish, and ease of operation.
U.10.2 Functional Test
While operating at ambient temperature, incremental concen-
trations of SO , to cover the full range of operation of the instrument,
shall be introduced at the system inlet and recordings made of the
instrument readings. The outputs of the unit shall be as specified
in J.6.
I*.10.3 Environmental Temperature Test
-------
33
^.10.3.1 Preparation for Test
The unit placed in a temperature chamber and stabilized at
the required temperature prior to beginning of each step of the test.
The stabilization to be determined by the thermocouple method or some
suitable alternate methods. Stabilization will be considered established
when the readings of the thermocouple are within the specified temper-
ature for three successive readings taken five minutes apart.
U.10.3.2 Test
The unit should first be stabilized within the ambient tem-
perature range of +59°F to +95°F and operation checked. The temperature
should then be increased and stabilized at a temperature of 50°C (l22°F),
then decreased in steps to 0°C (32°F). At each of the temperature
increments, the unit should be operated throughout its sensing range
for a period of at least 3 hours. The outputs of the unit shall be
as specified in 3-6.
4.10.U Environmental Shock and Vibration Test
The unit shall be tested to values listed in Table 1. (Ref:
KEL-STD-202, Method 201.)
^.10.5 Post Shock and Vibration Functional Test
After the shock and vibration tests, a repeat of the functional
test described in ^.10.2 should be performed.
5. PREPARATION FOR DELIVERY
5.1 Preservation and Packaging
The unit shall be preserved and packaged in accordance with
best commercial practice.
6. NOTES
None
-------
TABLE 1
ENVIRONMENTAL CONDITIONS
34
Environment
Temperature
Humidity
(Relative)
Vibration
Shock
Salt Fog
Fungus
Altitude
Non-Operating (Transportation
Storage and Handling) in Packaged
Condition, except as noted otherwise
High: 50°C (122°F)
Low: 0°C (32°F)
0 to 90% rh with conditions such that
condensation takes place in form of
water or frost
Vibration as encountered by equipment
requiring shipment by all modes of
common carrier. This vibration is
+1.3g from 2 to 27 cps and 3g from
27 to 500 cps
Drops on three mutually perpendicular
faces on a concrete floor as follows:
Cross Weight, Ibs.
50
100
150
200
600
3000 or over
Height, Inches
30
21
18
16
14
12
50 hours exposure to a 20% salt fog
solution
As encountered at maximum temperature
and humidity
Sea Level to 15,000 feet
Operating Under
Sheltered Environment
Same as Column "A"
5 - 98% rh
intermittent
Protected from plant
vibrations etc., which
could effect operation
of the instrument
None
No Requirement
Same as Column "A"
Sea Level to 6,000 ft.
-------
TASK C
Acquisition of Data on Existing Installations
The original plan for the program envisioned an intensive effort
to contact users of SO monitoring equipment and to obtain their
comments and operating experience in the field before completing the
design of the field test program. This worthwhile idea and plan
met with two major snags. First, the required Bureau of the Budget
clearance for the survey was not received until the program was nearly
completed. Second, there are very few SCL monitoring equipment users
in the U. S. In retrospect, a total of nine user questionnaires
were usable, a number of which would not have required Bureau of the
Budget approval at all. Table C-l, essentially, summarizes the respon-
ses of these equipment users with the footnote added at the end of the
program as hindsight. Most of the users of SO monitoring instruments
are located on the East Coast to Midwest regions which in general are
the same regions where high sulfur coal is produced. Many power com-
pany personnel stated that they were waiting for the regulating legis-
lation defining emission limits and approved means of monitoring before
they take action on either removal equipment or monitoring instrumen-
tation .
-------
TABLE C-l
Manufacturer
ITT Barton
Number of
Returned
Questionnaires
Sectarian
Beckman 315A
DuPont
Model ^00
Dynasciences
EnviroMetrics
Intertech NDIR
User
Comments
Instrument was utilized for evalua-
tion only on a pulp and paper mill.
Sampling proved to be a major prob-
lem; however, instrument performance
was excellent.
Used for measurement of SO in flue
gas. Overall instrument performance,
etc. was rated as good. Estimated
repeatability 2$; estimated accuracy
+10$. Scrubber system should be
improved to reduce interferences.
Used for measurement of SO in flue
gas. Overall instrument performance
rated as fair to poor. Estimated
accuracy 95+ %•
Used for SO measurement in flue gas.
Overall performance was rated as
excellent for analyzer unit and fair
for sampling system. Estimated repeat-
ability ±1$. Estimated accuracy ±1$.
Used for SO measurement in flue gas.
Overall rating was poor. Low stability,
routine functions, etc. required 2 hrs
operator time per day. Estimated
repeatability + 5$. Estimated accuracy
+ 10$.
Used for SO measurements in flue gas.
Overall rating was fair to good.
Estimated repeatability ± 5$ at 1000
ppm S02.
Used for SO measurements in a power
plant. Overall rating was excellent
requiring little attention and main-
tenance. Estimated repeatability
+ 2$. Estimated accuracy + 5$.
-------
37
TABLE C-l (Continued)
Beckman Chromatograph
620D
Leeds and Northrup
MDIR
Used for SO , CO , and 0 in
power plant stacK gas ana cat-ox
converter. This unit utilized user
developed technology and columns.
Overall rating excellent. Average
down time 25$. Estimated repeat-
ability + 1$. Estimated accuracy
± 2-3%.
S02 analysis of MgSO waste gases
from pump mill. Overall rating poor.
*It should be noted that at the time of this survey (early 1971) these
units were the only units on this list which were available as
"complete field tested systems."
-------
TASK D
Preliminary Evaluation of Instrumentation
According to the original contract plan, after completing the
extensive survey and digesting large amounts of user provided infor-
mation, the project staff was to evaluate and rank the existing
instrumentation according to user information, Task B preliminary
specifications, and any other engineering or scientific judgments
which could be brought to bear. However, Tf-sk C provided very
little data, so the staff was forced to make their ranking and eva-
luation judgments on the basis of very meager and often very
subjective data.
Since the rankings and judgments were arrived at in this
manner, the following sections were presented in order to fulfill
the contract terms and show the effort involved. The remainder
of this section on Task D are to be considered as tentative, largely
unsubstantiated, and subjective information and comment.
Based on the current information, a preliminary ranking of
instruments was performed. To a large extent the rankings, brief
comments and problem areas included below, were derived from the
information in the attachments.
Ultraviolet - Visible Optical Instruments
At the time the DuPont models ^00 and U60 were the only can-
didate UV-VIS SO stack gas monitors. The others had either with-
drawn or had not supplied sufficient information for evaluation.
Non-Dispersive Infrared HDIR (Attachment D-l)
A major potential problem exists with the NDIR instruments,
i.e., the stack gas sample must be conditioned to remove H 0. Such
conditioning also removes approximately 5$ S0p. This value can vary
with sample conditioner design and operating parameters.
-------
39
Ranking of instruments for this and other reasons is at best
arbitrary without laboratory testing under controlled, comparative
conditions. Minimum laboratory testing consists of the following:
a) base line drifts
b) span drifts (at 75$ of span)
c) physical inspection of instruments
d) sensitivity to SO
e) sensitivity to HO
The NDIR instruments were ranked as follows:
1. Bendix Unor
2. Intertech Uras
3. Beckman Model 315A
k. M.S.A. Lira Model 300
The above ranking was based upon design and theoretical operating
considerations from limited information. It may be expected that the
ranking would change as other assessment criteria, e.g. user informa-
tion on reliability, accuracy, etc. are applied.
Flame Photometric
This instrument (the only one of its type) is under development by
Melloy Laboratories as a potential SO stack gas monitor. A few of
the potential application problems include plugging by particulates,
the large volume of dilution gas required and potential interference
by CO . There were no commercially available F.P.D. analyzers.
Melloy Laboratories had constructed one ''production prototype/'
The unit was being used for display and was not available for testing
on this program. It had been reported verbally that definite
interferences occur in the measurement of SO when high concentra-
tions of CO were present. The interference is most likely caused
by CO from the dissociation of CO in the flame.
The major problems for this type of detector system include:
a) dilution of UOOO to 1 requirement
b) reported negative interference from CO
-------
Mass Spectrometer SO Stack Gas Monitors
AeroVac, Troy, N.Y. produces the only known process type mass
spectrometer that may be suitable for stack gas analysis. AeroVac
proposes to utilize a Teflon membrane to separate SO from the
flue gas. As the instrument was under development and not commer-
cially available, it was recommended that no additional evaluation
be conducted until the vendor had the monitor ready for marketing.
It was strongly urged that a laboratory evaluation of the dilution
system, potential interferences and overall performance be performed
on an off-the-shelf production unit when one becomes available.
Gas Chromatographic (GC) SO Monitors
This class of instruments, at the time, had questionable utility
as a stack gas monitor for SO . With the information available, it
was not possible to recommend that a GC be included in the test program;
however, the status of G.C. systems will be monitored for any signi-
ficant change.
Gas Chromatograph (G.C.) - Evaluation of commercially available
process G.C. systems was not appropriate at the time as the critical
factors of sample pretreatment, automatic sample injection, column
and materials technology to prevent spurious reactions and reliable
analyses were not proven for this application. To date only one user
(utilizing nonvendor supplied technology) reported satisfactory results.
Electrochemical Instruments (Attachment D-2)
Numerous electrochemical instruments were available on the market
for the purpose of monitoring SO . Of these, at least 6 systems were
recommended by their manufacturers for stack monitoring applications
as opposed to ambient air monitoring. In the detailed evaluation
presented in Attachment D-2, the principles of operations and poten-
tial interferences for conductometric, coulometric and fuel cell
instruments are discussed.
-------
Conductometric Instruments were ranked as follows:
1. Mitrogas-MSK-SO E (Calibrated Instruments)
2. Scientific Industries Model 70 (they sell their own air
dilution system)
Interferences in the above systems can be caused by NO and
X
other stack gas constituents. Reaction kinetics and catalysts which
may be present could pose application limitations.
Coulometric Instruments were ranked as follows:
1. ITT Barton Model 286
2. Beckman Model 906A (a Teflon membrane is utilized for
SO separation and dilution). It is to be
expected that Nop, etc. will interfere to some
extent. The Becfenan instrument is basically
an air monitor and had not been tested on com-
bustion stacks.
Galvanic or Fuel Cell Type Detectors - Dynasciences and EnviroMetrics
offered SO- analyzers. These units are small, compact and appear to
be easy to operate. User reports indicated the analyzers do not
have the expected long term reliability apparently because of cell
poisoning.
-------
As a result of the Task D "evaluation,'' the following equipment
was selected for inclusion in the test program. The major criteria
was: did a specific piece of equipment best typify a given analysis
method or technique?
TABLE D-l
Equipment Selected for Field Testing
Manufacturer
Dynasciences
Calibrated Instruments
ITT Barton
Bendix
Intertech
DuPont
Melloy Labs
MSK-S02-E
1*00
UNOR 6
URAS-2
1*60
prototype
Technique
Electrochemical - "fuel cell"
Electrochemical - conductometric
Electrochemical - coulometric
NDIR "new type detector"
NDIR
Ultra violet absorption
Flame photometric detector
The decision not to include several other instruments or
manufacturers was generally based on delivery times (Beckman, Envi-
ronmental Data Systems), cost or duplication of other techniques.
-------
APPENDIX D-l
NDIR INSTRUMEHTATION FOR
STACK GAS MONITORING
1. Introduction
This report summarizes the results of a study of five non-
dispersive infrared (NDIR) SO monitors. The devices were
designed to continuously monitor the concentration of SO,
2
in stack gases from coal burning power plants. All units use
as their principle means of detecting SO the selective ab-
sorption of infrared radiation by SOg molecules, and consist
of an IR source, a chopper, a sample chamber, a reference
chamber, and a detector unit.
The radiation from the source traverses the sample chamber
and the reference chamber. Beneath each chamber is a sealed
unit filled with SO gas; a thin diaphram separates the unit
beneath the reference chamber from that beneath the sample
chamber. When no SO is in the sample chamber, equal amounts
of energy are absorbed by the SO gas in each unit. When S0g
gas is introduced into the sample chamber, it absorbs energy,
and more energy enters the chamber beneath the reference cell,
creating a pressure differential which is detected by the
diaphram.
2. Units Under Consideration
The units considered were:
l) Leeds and Northrop - IR gas analyzer
2) Bendix Unor
3) Intertech Uras
k) Beckman Model 315A
5) MSA - Lira Model 300
-------
Unit number 1, although of a potentially superior design,
has been withdrawn from the market by the manufacturer and will
not be discussed further.' All of the remaining units were of a
two cell, electro pneumatic type.
3. Features Common to All Detectors
All of the units suffer from problems associated with the
method of detection.
J.I Ply Ash and Particulate Matter -
The manufacturers of all units require that the gas sample
be filtered to remove fly-ash and particulate matter. This is
because the measurement to be made is an absorption measurement,
and any thing which reduces transmission gives a positive indi-
cation of SO .
In addition, it is clear that particulate matter collecting
in the windows of the sample cell in a two-cell system will
introduce a systematic error into the reading. At the same time,
it will decrease the signal to noise ratio and therefore, affect
the accuracy of the measurement.
Hence the degree to which NDIR detectors will be useful
depends critically upon the filtrations system ability to uni-
formly and consistently remove participates from the sample.
3.2 Water Vapor -
All of the instruments under discussion work on the prin-
ciple of absorption of radiation by SO in the sample gas.
C T
The primary S00 absorption band in the 1^00 cm region is
2 _1
overlapped by a water system which exists in the 1200 cm to
2000 cm regions. Hence, water vapor in the system will
effect the measurements and therefore must be removed from
the sample or be a known quantity. A method suggested by
most manufacturers involves the use of a cold trap held at
0°C. This has two effects: it removes a large percentage of
-------
the water vapor and in principle at least, the remaining
water vapor content is known and can be accounted for in the
measurement. This is not as straight forward as it appears,
however, since the design of the cold trap is all important to
the efficiency with which the water vapor is removed. In
addition, the vapor pressure of water at 0°C is ^.6 torr or
about 6000 ppm. Variations in the temperature of + 1°C
result in changes in the vapor pressure of + 0.3 torr or
about kOO ppm. Since we are concerned here with SO concen-
trations in the order of 200 to 1000 ppm, we not only require
that the temperature be held very constant in the trap but
that the rejection ratio of HO over SO by the detector
must be very high if the measurements are to be valid. Beckman
indicates that the rejection ratio for their detector is 100:1,
i.e., 100 ppm HpO gives the same signal as 1 ppm SO . If
this is representative for the other detectors then an effi-
cient trap held at 0 + 1°C will be sufficient, as this will
result in an error of + U ppm S02, which is about the required
sensitivity of the instrument.
An additional problem exists with respect to water vapor.
At 0°C, the solubility of SO in water is very high, being about
22.8 gms/100 cc. This means that SO will be taken out of the gas
when the water is removed. The amount of SO removed cannot
le calculated but must be measured, as it depends upon the trap
design, flow rates, amount of precipitable water present, etc.
The purpose of this report is to discuss the NDIR detectors,
and the consideration of cold trap design constitutes a digression
which introduces a host of additional problems. Suffice it to
say that it is a matter of concern for all such detectors, and
unless a manufacturer can specifically demonstrate he has over-
come the problem, the NDIR detectors discussed below are of
questionable use as consistently accurate SO monitors where a
wide variation in water vapor concentration exists.
-------
The methods used to compensate for water vapor in the
instruments consists tf positive filtering and negative fil-
tering. In positive filtering, the radiation is prefiltered in
the wavelength region of the interferent before traversing the
sample and reference chambers. Removing radiation in the water
band does reduce the sensitivity to water vapor but also affects
the SO sensitivity by reducing the amount of signal in the
SO band as well.
Negative filtering is done in various ways, but in one
model (URAS2) S0? gas is introduced in the path of the radia-
tion traversing the reference chamber. This primarily removes
radiation from the center of the absorption band in the refer-
ence side and the difference between the radiation in the center
of the absorption band and the flanks is formed in the detection
system. Such a system is less sensitive to interference from
water vapor. If a known amount of water vapor is present, an
additional filter is placed in the reference side which absorbs
radiation equivalent to that absorbed by the water in the sample
side.
3.3 Pressure and Temperature -
In ail of the instruments discussed below, the degree of
absorption is proportional to the total number of S0? molecules
in the sample path. Assuming the volume then gives a calibration
of concentration vs. signal level. However, changes in the
ambient temperature and pressure will affect the readings. The
errors introduced are 0.3$ per °C change and 1.3% per 10 mm hg
change. Most instruments compensate for changes in the baro-
metric pressure and some for temperature. These are noted in
the instrument discussion. No mention is made where the infor-
mation was unavailable.
-------
1*7
3.^ Microphonics -
All of the instruments discussed below are thermo pneumatic
and are therefore subject to microphonics induced by vibration.
In this respect it is important to test the instruments prior to
use, as this is the only practical way of determining the sound-
ness of design, care of construction, etc. of a unit.
3-5 Other Errors -
All of the instrument manufacturers recommend frequent
recalibration using prepared sample gases with known SO
concentrations. Obviously then, all of the instruments are
susceptible to calibration errors in direct proportion to the
accuracy of the composition of the sample gases.
In addition, care in filling the detectors and contaminants
introduced therein constitutes an error source to which all of
the units are susceptible.
h. Discussion of Units
The units considered will be discussed below in the order
in which they were ranked. The ranking was partially subjective
in that it was done solely on the basis of design considerations.
In those applications where one specification (e.g. low span
drift) is desired above all others, a different ranking would
result. In addition, a unit of potentially superior design is
of no use if it is poorly constructed, has a low time between
failures, is difficult to repair etc. Since the units were not
available for test, no consideration could be given to these
aspects. It is strongly recommended the units be tested prior
to selection of one of them for on-site measurements. This will
be discussed further in Section 5 below.
U.I Bendix Unor -
This unit is shown schematically in Figure 1. It is a two
chamber/two detector device which utilizes the fact that, for a
-------
spectral absorption line with a Lorentz shape, the absorption
coefficient is given by:
E(\>) = Pb
II[(v-vQ)2+ b2]
where VQ is the line center frequency, P is the power, and
b is the half width. For radiation passing through a sample
of gas, the energy in the center frequencies of the absorption
line is preferentially absorbed over energy from the wings of
the line. In the Unor the design is such that the center energy
is mostly absorbed in the top detector, and the flank energy
in the bottom.
With no SO in the sample chamber, the absorbed energy is
equal in both detectors. When SO is introduced in the sample
chamber, the center frequency energy is absorbed in the sample
chamber instead of in the top chamber creating a pressure im-
balance, and therefore a signal in the microphone.
This unit has some advantages over the remainder of the
units in that it is less sensitive to broad band interference,
since such interference affects the wings and center of the line,
and therefore both detectors, equally.
It has a single IE source, in contrast to double IR sources
of other units; this also is considered an advantage since there
is just one less thing to malfunction during long periods of
unattended operation.
For some reason, temperature compensation is not avail-
able with this unit, although pressure compensation is.
U.2 Intertech URAS (Figure 2) -
The main advantage of this unit over the remaining units
is what appears to be flexibility in design. The manufacturers
seem -.'.ailing to alter the system quite drastically as might be
necessary to suit the applications. It has capability of posi-
tive filtering for interferents, and, particularly for water
-------
vapor problems, can be obtained set up with negative filtering.
In the latter case the detection technique is changed from
a simple energy detector (depicted above) to a scheme similar to
to Unor which depends upon preferential absorption of center
frequencies of the absorption line. This is shown in Figure 3
below. Operation in this mode is not completely satisfactory,
however, since in order to balance the two chambers, a neutral
density filter (a wire mesh) must be inserted into the sample
chamber path. This reduces the available energy and therefore
decreases the signal/noise ratio.
U.3 Beckraan 315A (Figure 4) -
This unit is a simple energy balance detector. It has a
slightly wider range of temperature operation than the other
units, in that it is capable of operation at temperatures ran-
ging from -20 to +120°F.
The manufacturer recommends daily recalibrations, which is
somewhat more frequent than other manufacturers; however, this
may be due more to a lack of candor on the part of other manu-
facturers than to a lack of stability in this unit.
k.k MSA Lira Model 300 (Figure 5) -
This unit is simply a UNOR scheme with a single chamber.
The temperature range of operation of the unit is from +30 to
+120°F, which may be inadequate for stack monitoring in colder
climates. The smallest range of operation available off-the-
shelf is 0-5000 ppm which is a bit gross for the measurements
under consideration here. The manufacturer states that smaller
ranges are available but there is no information as to cost,
delivery times, effect of the change on the accuracy of the unit,
etc.
The chopper frequency is 2 GPS, which is a bit low. The
optimum chopper frequency for thermo-pneumatic detectors is
about 10 to 15 GPS.2
-------
50
And lastly, the electronics may be a bit outmoded, being of
tube design rather than solid state. While tube designs can be
made which are every bit as stable and capable of long MTBF as
solid state designs, they usually are not. They also require
more power and are more susceptible to heat problems.
In defense of the unit, it appears to work under actual con-
ditions. The unit has been in service for years, is very rugged
and is apparently reliable. In the final analysis this may be
the qualifying factor and the MSA unit may prove to be superior
to the other units which, although of potentially superior
design, may not in fact deliver that superiority in the finished
product.
5. Recommendations and Conclusions
It is strongly recommended that a laboratory program be
carried out to test these instruments prior to selection of one
of them for stack installation. There is no way of determining
the care of circuit design, ruggedness of construction, relia-
bility, and susceptibility to interference by reading the manu-
facturer's literature. As stated above in the discussion of the
MSA unit, shoddy workmanship or poor design can easily negate
the minor inherent advantages possessed by units such as the
Unor.
Such a program need not be extensive. The minimum test
program should probably include the following:
l) Turn units on and measure zero drift as a function of
time. If environmental chambers are available, drift
can be checked as a function of temperature and pressure
variation.
2) Test sensitivity and reproducibility of SO measurement
using sample gases of known composition over the range
of operation of the instruments.
-------
51
3) Tests to determine the effects of water vapor on the
measurement.
Item 1 (without the environmental test equipment) and Item 2
require almost no special laboratory equipment and can be done
at very little cost. The cost and time involved in Item J is
as little or as great as desired, since a simple experiment
would suffice in giving a rough idea of the effect of water
vapor on the SO measurement. For example, SO gas of known
concentration in dry air can be run through a detector. The
same gas is then passed over a water reservoir held at known
temperature (and therefore vapor pressure), the wet gas run
through the detector, and the resultant measurement compared with
the previous one. This is a crude measurement but it can be
performed on all of the instruments uniformly and the instrument
response noted. Any differences in the ability of the instruments
to reject water vapor interference ;should show up immediately.
If more time and money are available, more elaborate schemes
could be carried out wherein the water vapor content is accurately
measured before and after the gas is run through cold traps and
the effects on the measurements nob ed.
6. Conclusions
A survey has been made of five instruments designed to
measure S0? concentrations in stack gases of coal burning power
plants. The instruments have been ranked according to design.
All such instruments, generally called NDIR detectors, are only
fnndit.innB.lly suitable for use as continuous monitors
since all suffer interference from particulate matter, water
vapor and vibration.
It is concluded that the proper ranking of the instruments
can only be done after the instruments have been subjected to
laboratory tests to measure their performance.
-------
Figure Dl-1
Bendix Corporation UNOR-2 Infrared Analyzer
.11
1 Radiation
2 Sync'ircnou
3 Rotating c
4 Sample cei
& Reference
6 Analysis c
7 Beam co:iui
d Detection
9 Front mcas
10 Rear niedst.
1 1 Connectviy
\2 Connecting
13 Di apnrag'1:
Id Keas urine
IS Ouiajt :-,e:
16 So corner
1 7 Gas pui'.u
19 Sofety ft 1
2\) Safety fil
ource
motor
opcer
ell
a^oer
er
namber
ring cndnber
ing cnat:«nef
cnannel
cianne 1
f capaci tor
"pi i f i er
r
er in ire o-is inlet
er in me ^a$ outlet
/
—J
lizi
Figure Dl-2
Intertech URAS-2 Infrared Analyzer
123 4 5 -
1 Radiation emitter
2 Mechanical chopper
3 Filter
4 Measuring gas cells (top: sample
cell, bottom: reference cell)
5 Sensor
6 Capacitance-type detector
7 Amplifier
8 Meter or recorder
-------
Figure Dl-3
Intertech URAS-2 In The Negative Filter Mode
123 4 j!
1 Radiation emitter
2 Mechanical chopper
3 Filter
4 Measuring gas cells
5 Sensor
6 Capacitance-type detector
7 Amplifier
8 Meter or recorder
Figure Dl-4
Beckman Model 315A Infrared Analyzer
53
INFHAREO SOUKCE
O O'HLR MOLECULES
CONTROL UN I
-------
Figure Dl-5
Mine Safety Appliances LIRA Model 200
LIRA OPTICAL SYSTEM
RECORDER
(OPTIONAL)
L I
INFRARLO SOURCES
Figures Dl-1 through Dl-5 from "Instrumentation for Determination
of Nitrogen Oxides Content of Stationary Source Emissions," APTD
PB 204877, Volume I, February 1972.
-------
55
APPENDIX D-2
ELECTROCHEMICAL INSTRUMENTS FOR MONITORING S02 IN STACK GASES
Prepared By
R. R. Sayano
E. T. Seo
Science Staff
Chemistry and Chemical Engineering Laboratory
December 1970
-------
1. INTRODUCTION
Commercially available SO- analyzers based on the use of electro-
chemical principles were evaluated for use in continuous monitoring,
of SO- in stack gases. The primary requirement was for "off-the-
shelf" instruments capable of continuous ii-stack monitoring at coal-
or oil-fired stack gases containing 200-700 ppm SO- (but with a range
extendable to 3000 ppm SO ) . The evaluation included a review of:
basic electrochemical principles involved, theoretical and practical
limitations, possible interference from other stack gas constituents,
operating procedures, maintainability, and physical construction. A
tentative criteria for selection and further evaluation were established
and instruments rated according to the criteria. Recommendations for
modifications and/or improvements of these instruments were also
considered.
The instruments were of the following three types: conductimetric,
coulometric, and "fuel cell." Following instruments were evaluated:
Conductimetri c:
Mikrogas-MS.K-SO -Ej
Model 70 Stack Monitoring
System
Series 11-7000 Series S02
Air Pollution Monitor
Series 9000 PPM Gas
Analyzer Systems
Model 902
Coulometric:
Model 286 Recording Sulfur
Ti trator
Model-906A Diffusion Stack
Moni tor
DW97CO S02 Monitor
Model 200 Ic'.-c.-s'rr i c
3 0 2 Analyzer
Titrilog II
Calibrated Instruments.
Scientific Industries
Davis Instruments
Devco Engineering Inc.
Instrument Development Co.
ITT Barton
Beckman Instruments, Inc.
Phillips E-lectronic Instruments
Process Analyzers Inc.
-------
57
"Fuel Cell"
Series NS-200 Multigas EnviroMetrics Inc.
Analyzer
Model SS-330 Dynasciences
In all these analyzers, the SO- gas passed into water or a
reagent solution where it either is converted to an ionically conducting
species, reacts" with a coulometrically generated titrant, or reacts
at an electrode.
Since electrical measurements can be made with great accuracy,
sensitivity, reproducibi1ity, and relative ease, the amount of SO-
detected can be very low if the technique is specific to S0_. Detection
at the 10 molar level is achieved with present day electrochemi al
techniques. In addition, the electrical signals are amenable to
remote control and signal processing techniques. The overall construction
of electrochemical systems are usually simple, and can be made small
and often portable.
Theoretically by the proper selection of electrolytes and electro-
chemical conditions, the detection system can be made highly selective
and sensitive to a specific species with response time within a
fraction of a second.
In reality, the procedures encounter difficulty because of sample
handling, low rates of reaction, and lack of specificity. Drawbacks
of the instruments evaluated in this study are:
conduct imetr i c
sample gas must be passed into deionized water or H_0_
solution to give an ionically conductive solution
other constituents of stack gases can form interfering
jonic species
reagent (viz., H C?) concentration changes due to depletion
or evaporation
temperature changes affect conductivity and reaction rates
(requires ..all ~_-± '~-~~ t-a ted cell)
-------
58
coulometric
sample gas must be passed into electrogenerated titrant
solution for reaction,
reagent lost through evaporation
temperature changes affect electrode potential and reaction
rates
any other constituent of stack gas that can undergo
oxidation or reduction in coulometry cell will cause
interference
"fuel cell"
sample gas must pass through membrane (increases response time)
temperature effects electrode process and diffusion rate
membrane characteristics changes with time and changes in
envi ronment
reagents within fuel cell become depleted
nonspecificity of membrane to SO- causes interference
-------
59
2. SELECTION A;;D EVALUATION CRITERIA
Commercially available SO- analyzers (air monitoring instruments)
based on electrochemical techniques were evaluated for use in continuous
in-stack monitoring of S0_. Primary emphasis was placed on "off-the-
shelf" units capable of monitoring S0_ in the range 200-700 ppm.
If a gas sampling or conditioning unit for adapting an air monitoring
instrument for stack-gas use was not available as standard equipment,
the instrument was not rated as a candidate. Such an instrument may
be useable later when modifications or stack gas sampling are made
available. Information required for preliminary selection and
evaluation was obtained from a basic understanding of an instruments
operating characteristics, manufacturer specification sheets and
instruction manuals, or by direct contact with a manufacturer's
representat i ve.
The following parameters were established as possible criteria
for further evaluation of the instruments. The nominal values given are
those required or can be met by the unit.
Range: overall dynamic range for S0_ detection (nominally 200-700 ppm)
Accuracy: percentage (or ppm) variation from known standard (nominally
±5% of full scale)
Zero Drift: variation in output with time for zero S02 input
(nominally less than 3% per day)
Span Drift: variation in output with time for constant SO- input
Response Time: time required for output to reach 90* steady state
value for step change in SO (nominally 20 sec to 5 min)
Lag Time: time required for first detectable output for step change
in S02
Ambient Condition: operating condition of total unit (nominally
temperature 32 to 100 °F, relative humility, R.H., 0 to 100%)
Flow Rate: sample gas flow rate to analyzer
Interference: percentage (or ppm) change in 50_ reading caused by
possible interfering sc^cies present in equal amount as S0_
-------
6o
Sensitivity: minimum detectable level of SO- (nominally less than
21 full scale)
Electronic Drift: output drift with time with zero S0~ input
attributable to electronic system (nominally negligible
compared to zero or span drift)
Maintenance Cycle: period of unattendent operation of unit before
manual maintenance is required (nominally 3 mo)
-------
61
3. CONDUCT!METRIC INSTRUMENTS
Conductimetric instruments measure the change in conductivity of
a solution resulting from the addition of SO-. In some instruments,
SO- is simply dissolved in deionized water and the increase in
conductivity of the solution is due to the reaction
S02 + H20 = H2SO- = H+ + HSO ~
Other instruments use a solution of H 0_ v/hich oxidizes SO- (or H SO )
to the more highly ionized sulfuric acid.
The increase in solution conductivity in either case is directly
proportional to the amount of SO- in the sample gas. Schematically,
the sample gas is continuously exposed to deionized water or the H_0-
solution which then flows into the measuring cell consisting of two
fixed-geometry inert electrodes separated by a fixed distance. Flow
rate of the solution is set so that all the SO- is equilibrated or
reacted before entering the measuring (conductivity) cell. To prevent
polarization of the electrode, alternating current is used and the con-
ductance is measured using a bridge circuit. Stable readings are
obtained if the cell is maintained at constant temperature* (e.g.,
20 °C) or if a temperature compensation circuit is used. The
spent solution is either dumped into a holding tank or recycled after
it is passed through deionization columns. In the case where deionized
water is used (no H-0-), CO- can cause interference. Correction for
the CO- interference is made by measuring the conductance of a S02~free
sample gas. In addition NO can (thermodynamically) react with water
to yield HfJO,, a strong acid. However, the rates of such reactions is
probably slow enough that the interference is small. Instruments
capable of monitoring SO- in stack gases are:
Mrkrogas-MS;<-SO--E. Calibrated Instruments
Model 70 Stack Monitoring Scientific Industries
Syste.-n
-------
62
J». COULOMETRIC INSTRUMENTS
Coulornetric instruments are based on the principle of coulometric
titration where elect regenerated halogen (bromine or iodine) serves
as the titrant. Basically, the measuring cell contains, for example,
an aqueous solution of Br and Br . A fixed 8r./Br ratio establishes
a constant potential at an indicator electrode. When S0_ is introduced,
+ + 2Br"
The concentration of Br. decreases, causing a shift in potential. This
potential shift is detected and current is supplied to a pair of
generating electrodes to re-establish the initial Br concentration.
The current required in the process is proportional to the amount of
S0_ in sample gas. Since the role of S0_ is to cause a change in Br-
concentration by reduction, any other species present in stack gas
which either oxidizes Br or reduces Br_ will interfere with the S0_
readings. For example, if NO and N0_ are present with SO-, the formation
of HMO. is thermodynamically favored and UNO. can be reduced by iodide
or S0_. Once again, since the rates of these reactions are probably
low, the extent of interference could be small. The nominal interference
due to NO. (v;here the concentration of NO = S0_ = 1 ppm) is reported
to be less than 5'°. Even when the NO. and SO concentrations are about
300 pprn each, preliminary tests (telecon v/ith ITT Barton) have shown
that the interference is less than 10%.
Off-the-shelf instruments recommended for possible use in stack-gas
monitoring are:
Model 286 / or kOO ITT Bartoa
Diffusion Stack "onitor Beckman Instruments
Model SCoA
Other instruments considered v;ere:
PW 9700 Phillips
c 1 £Ct r' i C -.cv I CcS
Titrilog II Process Analyzer Inc.
-------
5. "FUEL CELL" INSTRUMENTS
The "fuel cell"-type instruments are based on a completely sealed
sensor functioning as an electrochemical transducer. Basically, the
SO- diffuses through a membrane and dissolves in a thin layer of
electrolyte. An oxidation or reduction occurs at the sensing electrode,
resulting in a current which fjows to the counter electrode through an
external load.' The amount of current indicates the S0.» content of
the sample gas. Some of the drawbacks of this type of instrument are:
o Membrane is not highly specific to S02 (N0_ is a major interfering
species); and
o Membrane surface must be kept free of condensate and particulates.
In one unit (Envi ro,~etnic Inc.), the N0~, interference is subtracted
from the total "SO '' reading by using an additional sensor which
responds only to H0_. However, this requires continuous calibration
of two sensors and that both sensors to be operative at all times.
-------
6. FUTURE WORK
From this preliminary study in the selection and evaluation of
electrochemical monitoring instruments for stack gas S0_, future work
in the following areas is recommended:
o Examine the rates of chemical reactions involving interfering
species and develop procedures that take advantage of reaction
rate differences to minimize the effect of the interfering
species.
o Laboratory testing of selected instruments using simulated
stack-gas conditions to establish quantitative criteria for actual
in-stack evaluation; and
o Design of gas sampling or conditioning system for isolating
S0_ from interfering spec-ies in stack gas.
-------
TASK E
Design of Test Procedures and Experimental Plan
The design of hardware, operating procedure and test plans
will be discussed under the following categories:
Sampling Manifolds - These include probes, heavy particulate filters,
main heat traced transfer lines, mixing zone (to insure homogeneous
gas stream) common sampling point for individual transfer lines to
each instrument. Initially Intertech probes were installed in the
duct between the dust collector and the stack'(of Unit 1, Moapa
Power Plant). Dekoron or a similar type of heat traced Teflon tubing
was planned to be initially utilized as transfer lines. The final
design of the transfer system (line sizes, number of filters, etc.)
was determined after the instruments are selected, filters tested
and required flow rates determined. Evaluation of at least two types
of particulate filters was planned. These include Intertech carbo-
rundum and V/estern Precipitation fiberglass types.
Manuals for the following instruments and equipment have been
assembled.
Supplier Manual
Bendix IR215 for CO General operation manual
Intertech Corp. Self balancing CMR converter
Electronic temperature controller
Operating manual for URAS-2 SO
analyzer
Westinghouse Operating instructions for oxygen
analyzer
DuPont Temporary instruction manual for
k6o SO analyzer
Dynasciences Operating instructions for air
pollution monitor
-------
66
Supplier Manual
Calibrated Instruments Operating instructions for MSK-
S°2-E1
Bendix UNOR-2 Installation and operating instructions
Servicing instructions
Barton ITT Training manual
The above manuals vail be used for:
Installation procedures
Calibration procedures
Operation and maintenance information
Flue Gas Velocity Mapping - Figures E-l through E-3 illustrate the
originally proposed probe assembly and method of probe introduction.
By utilizing the instruments provided in the Intertech trailer, SO ,
CO and 0 could be concurrently determined while performing the
velocity traverse.
Calibration System - Figure E-U is a sketch of the originally proposed
calibration and back flush systems. As illustrated individual instru-
ments could be manually calibrated and an electronically controlled
aero-span check system would operate through a common sampling point
manifold.
Sampling System - Figures E-5 and E-6 represent the Intertech probe
and filter assembly and one approach to sampling "on-the-stack,!:
respectively.
Sampling System
An alternative sampling system was suggested for consideration by
the EPA Project Officer in which a large volume or portion of the flue
gas could be diverted to the trailer. Figure E-7 is illustrative of
such a system. This system was considered for inclusion and has the
following advantages and disadvantages on Unit 1 of the MOAPA site.
-------
67
Advantages -
a) easy to install probe for each instrument
b) ease of checking and replacing filters
c) high velocity fan would provide excellent mixing
and uniform sample in manifold
d) sampling conditions could be modified with relative
ease
e) technicians and engineers do not have to leave ground
level to service system, install or change filters, etc.
f) all of the system, excluding bracing and blower, can
be obtained locally
g) short sample lines to instruments could be used
Disadvantages -
a) a high horsepower, efficient blower is required to
insure that velocity in the pipes and manifold are
sufficiently high to insure that particular build-up
is minimized
b) a blower suitable for high dust loading and capable
of producing necessary pressure increase for the
desired high flow rates would require k to 6 weeks
delivery time
c) the transfer pipe would require additional external
supporting to insure structural strength to withstand
the high velocity gusts of desert winds
d) particulate buildup could be a potential problem if
there were any stagnant points in the transfer system
e) the zero and span check systems are more complex
After consideration of these disadvantages, the THW staff decided
to use the primary approach illustrated in Figures E-5 and E-6.
Sample Conditioning Systems - Sample conditioning systems are those
pieces of hardware and equipment that are necessary for preparing
and/or modifying stack gas physical or chemical characteristics prior
to introduction into specific instruments. This equipment included
any required pumps, filters, moisture removal traps, etc. The follow-
ing instruments required preconditioning or pretreatment systems as
per vendor specifications.
-------
68
UV-VTS DuPont Model k60 - This system came complete with the
necessary sample pretreatment equipment which is comprised of
heated lines and cell, negative pressure regulation and air
aspirated pump.
KDIR - In general this class of instruments required efficient
small particle filtration, water removal, sulfuric acid demisting
and pumping.
Electro-analytical - The galvanic cell type instruments (Dyna-
sciences and EnviroMetrics) required reduction in the water
content of the gas, fine particle filtration and pumping. The
coulomctric and conductometric instruments required that the
gas stream be cooled and maintained at a precise flow using an
ambient heat exchanger and pump/regulator equipment. These
instruments may suffer from interferences by other species in
the gas stream. If the vendor provides specific pretreatment
equipment to eliminate interference, the equipment was incor-
porated in the sample conditioning system.
Calibration Systems - Two types of calibration were planned. These
include standard systems for individual instruments and a drift check
system where zero and 800 ppm SO gases were sequentially introduced
at the common sampling point manifold, thus allowing a simultaneous
check of all instrument operation on the manifold.
Initial Data Acquisition and Test Plan
The operational procedures for the various instruments to be
tested'were those provided by the instrument vendors.
The following initial plan for data acquisition was planned
to permit acquisition of meaningful and adequate data for the sta-
tistical evaluation of each instruments' performance compared to each
other, for ranking and to the criteria specified in Task B.
-------
Accuracy - initially duplicate, parallel trap systems will be
operated a minimum of five* time/day, during first three weeks,
on the order of three days/week (90 tests). This number will be
reduced for the duration of the program to once/week and three
tines/day if adequate.
Data rate for instruments during the accuracy test will be
30 sec. per instrument (if changes in concentration cannot ade-
quately be followed by this data rate, it will be increased --
conversely, if changes are sluggish data rate can be decreased.)
Precision-Reproducibility-Response Time - these are all determined
from zero drift-span drift and linearity calibration tests. Zero
and span drift will be run daily and calibration weekly; all data
points will be obtained in triplicate on a random basis.
Selectivity - specificity will be determined in the accuracy test
but in addition, variations SO values will be correlated with
changes in CO, CO , 0 , power, etc. in plant. The EPA supplied
trailer contained equipment for SO , CO, CO and 0 measurements
on a continuous basis.
Standard Daily Operation - for the remaining criteria tests of
environmental effects, maintenance reliability, etc. the instru-
ments will be operated continuously but at a slower data acqui-
sition rate.
The test procedure and experimental plan was reviewed at the
21 May project meeting. The plan remained as previously discussed
except for the following changes.
Routine Data Collection - Data on all instruments will be
collected every two hours. This data collection will consist of
*In order to detect a 2a change between sample means at 90% con-
fidence, a minimum sample size of k points is required. A population
of five was chosen to provide a margin which will permit discarding
of an outlier.
-------
70
10 data points for each instrument and will be collected at the
maximum rate of the Esterline-Angus data acquisition printer
(2.5 channels/sec).
Zero-Span - Zero and span checks will either closely precede or
follow as many as practical of the above 2 hr. data collection
periods. The data rate for this phase of the test will remain
at a slower rate for longer periods of time to ensure that
individual instrument responses and other important factors are
recorded.
These changes were manually initiated and the" procedure was later
automated. An external trigger timer and relay system was installed
for each of the above modifications. Esterline-Angus provided recom-
mendations on timers and the proper integration with the printer.
The computer program MONSO was written in CDC Fortran IV to
provide a historical data storage and retrieval and statistical analy-
sis capability for the SO Monitor calibration data. The program is
designed to accept as input the following information: date, times,
millivolt (ADC) readings and gauge readings for the zero, span and
calibration data and to add these data to a file containing past
information. The retrieval and analysis function included the ability
to list a particular date's results or each date's results (the
individual data points), the ability to compute mean and standard
deviations for any, each, or all calibrations, and the ability to
compute linear or log linear calibration parameters and associated
errors for any, each, or all calibrations.
Evaluation of Referee Wet Analysis Methods
A brief eveluation of three fundamental wet analysis procedures
for quantitative determination of SO in flue gas was completed In
TRW's laboratory, i.e. l) precipitation titration with barium,
2) colorimetric measurement utilizing barium chloranilate, and
-------
71
3) acidimetric titration with standard base. The precipitation
titration utilizes barium perchlorate titrant with the end point
detection being visual (both thorin and Sulfanazo III indicators
are being evaluated, as well as a specific ion electrode technique).
Nonetheless, difficulties with end point detection will signi-
ficantly impact accuracy evaluation and consequently the study of
other methods of end point detection. The colorimetric method which
eliminates the,major drawback of the titrimetric method is based on
the original work of Bertolacini and Barney at Standard Oil Co.
of Indiana (A.C. Vol. 29, No. 2, February 1957). It is not as
desirable as the titrimetric from the standpoint of speed and ease
of analysis. The third method is the acidimetric method of Berk
and Burdick developed at the Bureau of Mines and has been included
only for laboratory evaluation and comparison where acidic inter-
ferences will not be encountered.
A discussion of the experiments that were conducted and results
are attached as Appendix E-l. To summarize, the results indicated
that the barium perchlorate titration with thorin indicator yields
slightly better accuracy and precision than does the same titration
with Sulfanazo III. Nonetheless the three operators in TRW's labo-
ratory expressed the (unwarranted) opinion that the Sulfanazo III
end point was easier to detect. The barium chloranilate colorimetric
method may be suitable for routine laboratory usage, however, be-
cause of the time required for sample work-up it was not used in
the field test phase of this program.
-------
72
O.D. = 3/4"
WELDED
WELDED
Figure E-i High Particulate Loading Velocity Mapping Probe
-------
73
\\\\v
WELD
SOFT RUBBER SEAL
A\\\\
,5" PIPE
Figure E-'4 Details of Gasket Assembly for Mapping Stack Gas
-------
74
10 FEET
TO TEMP.
READOUT DEVICE
FOAM RUBBER OR
SIMILAR MATERIAL
THERMOCOUPLE
1/4" TEFLON
OR S. S. SAMPLE TUBE
UPSTREAM
Figure E-3 Details of Integrated Flue Gas Mapping Probe
-------
HEAT TRACED TRANSFER LINE
TO SAMPLE PROBES
NEEDLE VALVE FLOW
CONTROLLER
CARBON FILTER
ELECTRICALLY OPERATED
VALVE
Figure E-M Proposed Calibration and Backpurge Systems
-------
76
FILTER
GASKET'
EXTENSION WELDED
TO EXISTING PROBE
FILTER ASSEMBLY
<
ROD REMOVED AND
LONGER UNIT
INSTALLED
EXISTING
FLANGE
EXISTING PROBE
Figure E-S Modified Intertech Filter Assembly
-------
TO
TRAILER
PUMP
(OPTIONAL)
HEAT TRACED TRANSFER
LINES
SAMPLE
PROBE
u
u
u
u
Figure f-L Sample Probe Installation "On-Stack"
-------
78
SAMPLE PORT
FLUE GAS
SAMPLE LINE
(3-INCH HEAT TRACED
ALUMINUM PIPE)
OPTIONAL ACCESS
FOR FILTER
HIGH VOLUME
BLOWER
TRAILER
RETURN PORT
STACK
RETURN LINE (3-INCH)
SAMPLE MANIFOLD
WITH PROBE INSERTION PORTS
FIGURE E-T OPTIONAL FLUE GAS DIVERSION SYSTEM
-------
79
APPENDIX E-l
PRELIMINARY SULFUR DIOXIDE ANALYSIS METHODS EVALUATION*
INTRODUCTION
The tests described in this report were conducted in order to
select a referee wet method for SO. to be used in a fiel^ "est of S0_
monitoring instrumentation for stack gas analysis. The gas sampling
apparatus shown in. Figure 1 was assembled and utilized in the laboratory
tests so that sampling parameters would be considered as well as the
analytical test procedures in the evaluation. The barium perchlorate
titrimetric method utilizing Sulfanazo III and thorin' as visual end
points, an acid-base titrimetric method and the barium chloranilate
colorimetric method were evaluated.
Instructions for analysis of sulfate using the barium perchlorate-
thorin method were taken from the Shell method, while the procedure for
sulfate analysis by barium chloranilate was that proposed by Bertolacini
and Barney and modified by Kanno.
The upper range that was examined by other investigators for the
thorin method was 6000 ppm as sulfur dioxide, while the lower limit was
25 ppm. The useful and linear concentration range given by Bertolacini
and Barney for the chloranilate method is 1 pg/ml as sulfate in the
solution measured in a 5 cm cell to 400 -vig/ml in a 1 cm cell. Kanno
proposed this method for the analysis of sulfur dioxide contaminant in
ambient air. The dynamic range quoted for this technique is suitable for
use as a referee stack gas monitoring procedure; however, our experience
leads to the conclusion that frequent standardization is necessary to
obtain working curves (calibration curves are not linear). These conclu-
sions are discussed in the following sections.
•
An in-house volumetric method using barium perchlorate with Sulfanazo
III visual end point was modified according to work done by Dragusin and
Gavriliuc for analysis of sulfur dioxide for sampling by absorption scrub-
bers. The same range for sulfur dioxide determeination quoted for the
thorin method is applicable here also.
* This work was accomplished prior to the selection of the barium perchlorate-
thorin or method 6, New Source Performance Standards, 23 December 1971.
-------
80
EXPERIMENTAL
Experimental conditions, reagent concentrations, titration media,
apparatus, procedures and calculations are presented very briefly for
general information; details of methodology can be found in those papers
referenced in the introduction.
Barium Perchlorate - Thorin Method [Ba(C10.) = .01M]
a) Use 40 ml scrubber solution sample, run blank concurrently.
b) Add 4 times sample volume of 80% IPA.
c) 2-3 drops thorin (0.2% in HO).
d) Titrate with standard Ba(C10,)2 from yellow to light pink.
Blank: 40 mis of 3% H20 solution.
e) Calculations of ppm SO. by Ba(C10,) titrations:
[Net mis Ba(C10 ) x M of Ba(C10.) ] (24.47 ml/ramole) (103) (dilution
S02, ppm = _ factor)
liters of gas at standard conditions (25°C and 760 mm Hg)
Dilution factor = total scrubber sample volume
aliquot taken for test
Barium Perchlorate - Sulfanazo III Method [Ba(C10,) = .01M]
a) Pipet 40 mis of sample to one flask. Pipet 40 mis of
blank to another flask.
b) Add 40 mis of acetone to both
c) Add 3-5 drops Sulfanazo III to both
d) Titrate to light blue endpoint (blank preparation as Thorin
procedure).
e) Calculations of ppm SO. by Ba(C10,)2 titrations:
[Net ml Ba(C10.)0 x M of Ba(C10,)„] (24.47 ml/mmole)(103) (dilution
S02' ppm = U 4 2 factor)
liters of gas at standard conditions (25°C and 760 mm Hq)
Acid-Base Titration with Phenophthalein (NaOH = 0.02 M)
a) Use 30 mis scrubber solution sample (30 mis 3% H00 served
as blank). l 2
b) Add 5-10 drops phenophthalein indicator.
c) Titrate to first light pink color.
d) Calculation
S02, ppm = [millim°^es Na°H] x (24.47 ml/mmole) x (103) x (dilution factor)
liters of gas at standard conditions (25°C and 760 mm Hg)
-------
81
Barium Chloranilate Method
a) Standardization:
1. Weigh out 931.0 mg of sodium sulfate [45.1% as 30n] and make
up volume in a 1000 ml volumetric flask with deiohized water.
Concentration of solution, S02 per milliliter equals 45.1% x
931.0 rag/liter = 419.9 ug/ml.
x 2A-47 >jl/mole - 160-6
2. Pipet 15 ml of solution No. 1 into a 100 ml volumetric flask
and dilute with deionized water. This yields a solution con-
taining, 24.09 yl of SO /ml.
3. Pipet 5 ml, 10 ml, 15 ml and 20 ml of above solution into
separate 100 ml volumetric flask (don't dilute to mark).
This yields solutions containing 1.2, 2.4, 3.6, 4.8 ul of SO /
ml respectively.
4. Add to each of the above flasks and to an empty volumetric
flask (blank) 10 ml of a KHP buffer and 50 ml of 95% ethanol.
5. Add dilute HC1 to obtain pH 4.0 on final solution.
6. Add ^0.3 g of barium chloranilate and dilute to volume.
7. Shake for 10 minutes on auto shaker.
8. Filter through Whatman No. 40 paper or equivalent and run
samples on DK-2A or similar spectrophotometric at 530 mu
versus a blank prepared in the same manner.
9. Operating Conditions for Beckman DK-2A
• Measure absorption at 530 my versus a blank in the
reference cell.
• Tungsten lamp.
• Photomultiplier detector.
e 1 cm or 10 cm cells (depending on concentration) .
b) Sample analysis:
Take suitable aliquot from scrubber sample solution and treat
as above.
c) Calculation:
Read directly from standard curve, correcting for dilution
and volume of gas .
-------
82
Stack Sampling Apparatus
The apparatus for sampling stack gas was adapted from the stan-
dard thorin indicator method for use in the laboratory. Figure 1
shows the components of this absorber train. A standard certified
875 ppm S02/N2 gas blend was substituted for the probe. The 875 ppm
sulfur dioxide/nitrogen blend was drawn through the 50 ml of H202
scrubber solution into the calibrated evacuated volume while monitor-
ing the pressure with a transducer. For this standard gas blend 3.6 1
to 14.4 1 volume samples were used. It was found that flows faster
than 1.0 liter/minute could result in S02 loss therefore, the rate
used here was 0.5 liters/minute through the modified ASTM lamp sulfur
scrubbers.
The following formula was used to correct gas volume to standard
conditions (Z5°C and 760 mm Hg).
r
V = V
298 Ff \ _ / 298 Pi
" -» f f\ I I m I ^ T *•» **
rf+273 760 ) U.+273
\ / 2
"H
Where: V = vol. of gas sample at standard conditions in liters
V = vol. of calibration tank in liters
T = final temperature, °C
T. = initial temperature, °C
P.. = final pressure, mm Hg
P. = initial pressure, mm Hg
RESULTS AND DISCUSSION
Repeated standardization for the barium chloranilate method over a
wide concentration range (0-400ug/ml sulfur dioxide) resulted in non-linear
and non-reproducible curves. It was found that pH was even more critical
than expected and careful adjustment to pH 4.0 with dilute HC1 was necessary.
Because of these problems a working curve was prepared for the concen-
tration range from 0-10ug of sulfur dioxide per ml (0.4ul/ml) resulting
again in a non-linear curve at pH 4.0 (See Figure 2). Reproducibility of
the curve from day to day was not good.
A Bausch & Lomb 340 Colorimeter and a Beckman Model DK2A Spectrophoto-
meter were both evaluated for use with the barium chloranilate method. The
B&L 340 was totally inadequate for this analysis.
Critical pH, necessity for filtering, mixing time, and the difficulty
in reproducing the calibration curve makes the barium chloranilate method
impractical for use in the field test program.
-------
83
The following tables give results obtained from the volumetric pro-
cedure. Statistical evaluation of the data indicates that there is essen-
tially no major difference in the precision of the thorin and Sulfanazo III
methods. The mean values of the thorin and acid-base titrations are not
significantly different from the certified value (875 ppm) but the mean
value is significantly different for the Sulfanazo titration. This bias
maybe related to selection of the point of color change.
COMPARISON OF VOLUMETRIC TECHNIQUES FOR SULFATE ANALYSIS
(PPM S02)
Volume of Sample Acid-Base
(liters at STP* Thorin Sulfanazo III (phenophthalein)
3.63 896 863 897
14.4 893 852 878
14.4 860 831 833
14.4 872 879
*Certified from vender as 875 ppm SO
Thorin Sulfanazo III Acid-Base
Sigma value 17.2 20.2 32.9
Mean 00 - 880 856 869
The speed and simplicity of these methods suggests that one be used
as the referee technique. Given the above data, the barium perchlorate-
thorin method appears to be the preferred technique.
REFERENCES
1. Bertolacini, R. J. , and Barney II, J.E. Anal. Chem., 29, 281,
(1957)
2. Kanno, S., Int. J. Air Poll. 1., 231-33 (1959)
3. Dragusin, I. and Gavriliuc Angela, Revue Roumaine de Chimie,
12, 1239-43 (1967)
-------
=0=
ADAPTER
NEEDLE
VALVE
SO2/N2 CALIBRATION GAS
r
i
i
, ,„ . Oj
|ICE „
[BATHj]_ |
ABSORBER I ABSORBER II
DRYING
TUBE
I
TRANSDUCER
J
ROTAMETER
VACUUM
\
VACUUM TANK
Figure f'.-- Apparatus for Sampling SO /N Simulated Flue Gas
oo
-------
BARIUM CHLORANILATE CALIBRATION CURVE
5.0
4.0
CN
o
Z
O
i
3.0
-5
Z.
Z
o
u
1.0
BECKMAN DK-2A
10 CM CELLS
pH 4.0
530 mu
.10
.20
.30
.40
.50
ABSORBANCE
Figure eL-\ Barium Chloranilate Calibration Curve
oo
tn
-------
86
TASK F
PROCUREMENT AND INSTALLATION OF S02 MONITORING EQIPMENT
The scheduled installation of power to the trailers by the Nevada
Power Co. was delayed several times because of the more critical
demand on their personnel to trouble shoot and maintain their power
plants operating. In addition to the manpower delays, TRW had to
purchase a transformer for use as the power company did not have
adequate service to supply their power needs. Finally, 2^0 volt, sin-
gle phase, 120 amp service was installed to both trailers. Transient
noise and power fluctuations including 60 cycle ground loop noise, 60
and 120 cycle spiking noise were observed. These line problems were
remedied as required by rewiring, use of Sola and isolation transfor-
mers and filters, but only to the extent that is reasonable for field
installation. If an instrument was unduly susceptible to minor line
noise or was generating noise itself, these observations were documented
an an instrument deficiency.
Sampling System
The Intertech stack probe, designed for high particulate loading,
(i.e., the unit that incorporates a shield on the in-line or in-stack
filter), was tested in conjunction with the Intertech NDIR in the
Unit 1 breaching duct. The probe and heat traced Teflon lines were,
equilibrated at 300°F prior to drawing the sample stream to the instru-
ment. The probe became completely clogged (packed) with particulate
matter in a relatively short duration (^ several hours). A back
pressure of JO psig gave no indication of backpurge flow. Only
the partial shield was used (protecting the filter from direct
impingement by the flue gas stream) and the effective filtration area
of the filter may have been diminished in previous testing. When
this unexpected failure occurred, it became quite apparent that the
Project Officer's suggested approach, described as the contingency
sampling plan in Task E, should be immediately implemented.
-------
87
The details of the complete sampling system incorporating the
flue gas stream diversion approach is shown in Figure F-l. The heat
traced insulated sample line from the duct is 3" Aluminum Schedule
!*0 conduit pipe. Originally an order was placed for an American Air
Filter Co. blower (5) with 2 h.p. motor and a 9K industrial exhauster
for 300°F abrasion resistant service.
Since the manufacturer was unable to supply the unit as specified
by the required delivery date, the order was cancelled in favor of
the following blower purchased from Smallcomb Electric Co. with a 5-
day delivery guarantee:
Size 6, Type CI Clarage blower, Arrangement 9, THEW, 38^5 RPM,
complete with drive and variable pitch pulley on 220 VAC single
phase 1-1/2 HP motor specifications: jUo CFM at 10: water static
pressure.
The Clarage blower had 6" round inlet and discharge openings which
required adaptation to the 3" sample line. The adapters were fabri-
cated at the TRW Redondo Beach facility.
In order to ensure a homogeneous gas stream at the sample manifold
the minimum pump capacity to give 37.5 ft/sec (^2250 ft/min) linear
velocity or a sufficiently high Reynolds number (>30,000) through the
Y aluminum pipe was calculated. The CFM of 128 ft^/min could be
handled by a Q.25 h.p. motor.
Continuing downstream, to the sample manifold ( \6/ ), a section
V
of 3" IPS pipe k ft. long was modified to provide 10 outlets for
either single instrument sampling or multiple-parallel filtering of
sample gas for as many instruments as would be installed in the down-
stream instrument manifold. Outlets from the stack gas manifold
were provided with a 3/^" straight pipe thread required to fit the
Intertech filter housing thereby allowing primary filtering at that
point. Fabrication of the manifold has been completed. This design
of parallel filtering through 10 filters that were designed for in-
stack operation was a fail-safe system from the standpoint of parti-
culate clogging. It gave long term continuous flow through service
-------
88
with a minimum of back purge required.
As a precaution, however, a second type of filter unit purchased
from AMF Cuno Division of AMF was procured for evaluation in the
event expanding the number of parallel filters becomes necessary.
The description of this filter equipment is as follows:
Type 1B1 Filter Housing (cast iron)
Microscreen cartridge 200 mesh (^75 microns)
Porolkean cartridge 20 micron
Mircorscreen cartridge 2 micron
Teh cartridges are constructed of stainless steel and are cleanable
by ultrasonic methods for reusability. Viton gaskets were specified
for high temperature service. Use.of these filters was not required
as the primary design using Intertech filters operated in a satisfac-
tory manner during the test period.
Flue Gas Velocity Mapping and Installation of Trailer
The flue gas in the upper half of the duct work of Unit 1 was
mapped vertically downward from each of the four access sampling ports
to determine relative flow velocities and respective flow directions.
The probe was that shown in Figure E-l. It was found that on a conti-
nuous downward traverse, a.11 flow directions were parallel to the
main mass flow in the duct and the area of msximum velocity was approx-
imately four feet below sample ports B and C (Figure F-2) at the
mid-line of 11 feet. The relative velocity was 1/2 of the maximum.
The temperature of the flue gas was 108°C (2^0°F). The velocities
were not calculated as this information was not relevant because the
unit was operating at 75% of the full power rating due to a missing
water preheater.
-------
Probe
Duct
Stack
Heat-traced Sample Line
Stack Gas Return Ltne
Blower
6 Sample Manifold (See (T) )
7 Valve Timers (2)
8 Blowback Compressed Air
System
Standby Blowback N
10,11 Zero and Span Case
12 Calibration Cases
13. 16 Vacuum Pumps
Chemistry
15 Typical Instrument
17 Digital Printer
Figure M Gas Diversion and Instrument Installation Design
CD
-------
90
D.
22 FEET
4 FT
11 FT
DEPTH TRAVERSED
AREA OF
MAXIMUM
VELOCITY
0
RELATIVE VELOCITY
OF MAXIMUM
DEAD ZONE
FIGURE F-a RELATIVE VELOCITY PROFILE OF UNIT 1
-------
TASK G
Field Test Program
The field test program was started on the 10th of May, 1971
when the nominal installation of all the monitors was accomplished
by TRW personnel. The Bendix UNOR-6 monitor was delivered and
installed in mid June. The remaining time in May was spent in the
usual sort of trial operation and error debugging, of which the
following is typical.
After one week of continuous operation, on 21 May 1971, unusually
low SO values on the order of 90 ppm were noted on the instruments
along with abnormally high oxygen and low carbon dioxide values,
approximately 15$ and 7.5$» respectively. Shortly after this was
noted a heat trace line short caused melting of two Kynar tees.
These tees are temperature rated well within the normal operating
temperature for our system (300°F upper limit for continuous use as
compared to 200°F for the sample transfer line). Subsequently, the
entire system was refurbished with new fittings, heat tracing and
insulation. The short problem was identified as a specific friction/
pressure contact at the base of the 10 port manifold which was corrected
in the rewrap.
The relatively routine operation began about the first of June
1971. The Esterline-Angus data recording system (model D-2020)
was set to record values from each instrument at hourly intervals.
TRW recorded 10 points at 6 second intervals each hour, the data
tables show the mean value of the 10 millivolt readings and the
standard deviation. In this manner, one .could evaluate short term
noise, voltage spikes, etc. A very stable instrument whould show a
very small standard deviation of the readings. In the case where
manual wet chemical samples were taken, the printer recorded 10
readings from each instrument over the 10 minute sampling period for
91
-------
92
the wet method. In this manner, the mean values for instruments and
wet analyses should be directly comparable. The sample line back-
purge, zero and span gas valves and data system were connected to a
series of timers which permitted semi-automated operation.
The data acquisition phase ran from 1 June 1971 to 31 August 1971
with periodic interruptions due to operating failures in the power
plant boiler. The plant was down from 12 July to 2h July 1971.
The sampling, filtering, calibration and sample delivery system
operated exceptionally well throughout the entire test period requir-
ing only minor maintenance. (Tighten fittings, oil pump motors.)
-/
Table G-l contains a brief abstract of the maintainance log during
the field test period.
Figures G-l thru G-6 show the individual analyzers and sample
conditioning equipment for each unit as they were installed during
the test program. All analyzers and sampling equipment were fed with
a common 1/2" heated line containing the stack effluent gas at stack
conditions of temperature, CO and water content, but filtered to
remove particulate matter. Each analyzer was equipped with whatever
sample conditioning was suggested in the instruction manuals.
Table G-l contains an abstract of the maintainance log kept by TRW.
It should be noted here that the wet chemical analysis did not
agree with the vendor's certificate of analysis'' provided with the
bottle. This seems to be a common problem which ha.s been noted by EPA
equipment users and research facilities for some time: Calibration
gas vendors are apparently capable of blending gases to within quite
reasonable tolerances •'- 5 to 10$ at the 200 ppm level, but are unable
to provide a good certified analysis of exactly what is present.
Table G-2 illustrates this problem.
Table G-3 contains the typical millivolt readings from the
instruments. The channel assignments were:
-------
93
Channel 1 - Calibrated Instruments MSK-SO -E
Channel 2 - DuPont 460
Channel 3 - Intertech URAS-2
Channel 6 - Dynasciences SSJ30/CS1000
Channel 7 - ITT Barton UOO
Channel 8 - Bendix UNOR-6
The notes and "true value" columns contain annotations as to function
being performed (span-gas, zero gas, calibration gas), and the "true
value'' as indicated by the wet chemical analysis of the stack gas
at that time or the wet chemical analysis of the span or calibration
gas bottle in use at that time.
Rather than burden the report with the mass of hour by hour
routine data obtained over the three months of testing, Table G-3
illustrates typical data obtained during the period. The actual data
used in each analysis phase will be detailed in concise tables under
Task H, Data Reduction.
-------
TABLE G-l
ABSTRACT OF MAINTAINANCE LOG
DuPont SO Analyzer - Instrument was fully operational before June.
Malfunction problems were encountered in mid June were related to
modifications made during field installation. Output range was
changed from 5 millivolts to 10 millivolts by factory service
representatives in late June.
Intertech URAS-2 SO Analyzer - Instrument made operational by TRW
before June. 22 June 1971 defective tubes in amplifier replaced.
This maintenance item for the URAS-2 NDIR has been included as part
of the installation discussion because the EPA owned unit was pre-
viously used on another program and the unit not refurbished prior
to the current test. No other operational problems were encountered.
Calibrated Instruments - Instrument installed by TRW in May.
l6 June factory representative checked instrument and made recommen-
dations .
Bendix UHOR-2(NDIR) - Received 2 June and installed by TRW on
June 8. June 22 factory representative placed instrument in oper-
ation. June 28 analysis of data revealed that Bendix not operational.
Repaired .lU July 1971 by factory service personnel.
Dynasciences SO Analyzer - Non-operative June 2 needed cooling for
C.
water knockout. 7 June TEW supplied cooling system installed.
ITT Barton - 7 June instrument rendered operational by TRW person-
/
nel. No valid operating data obtained for June because of extremely
slow response time, etc. Although several calls were placed to
vendor to assist and evaluate in the field installation, no positive
response was obtained.
-------
95
TABLE G-2
CALIBRATION GAS ANALYSIS (ppm
Requested Level
200
400
600
900
1000
900
Vendor certified
Level
205
400
660
824
1080
880
Wet chemical
Analysis Level(a)
mean + S . D .
216 + 2.4
413 ± 3.3
646 ± 7.6
812 ± 3.2
1036 + 4.2
859 + 3.2
Number of
Analyses
12
9
12
12
12
3
a. The analytical technique used is discussed in appendix E-l.
-------
96
TABLE G-3
Typical Routine Data
M.V. Means and Standard Deviations as a Function of Time
Julian
Date Time
161
162
Zero
162
Zerob
Span
Zero
162
Span
162
162
Span
162
Zero
165
Span
10:11
9:00b
10:05
10:45
10:52
10:57
11:03
11:09
11:40
12:08b
12:16
12:20
12:23
17:15
Channel
1
95.53
.419
29.62
.383
1.92
.0919
27.62
.563
1.966
.058
94.6
.777
1.87
.115
22.7
2.65
98.02
1.7
0
0
93.0
.354
4.44
.445
23.86
.219
0
0
2
3.295
.0425
1.216
.150
.134
.051
1.20
.1077
.0267
.0012
3.55
.156
.007
.093
1.14
.065
3.286
.559
1.362
.104
3.984
.1374
.162
.025
.032
.058
3.336
.0396
3
8. .331
.0120
3.252
.056
.198
.0187
3.02
.0778
.133
.0115
8.64
.051
.263
.032
2.896
.0776
8.63
.132
2.924
.0483
8.62
.0055
.262
.030
2.48
.011
8.67
.0141
67 8 True Value*
57.29 ppm S02
.191
33.52 229±2
.597
4.77
.116
35.7
.259
6.17
1.17
55.6 871
1.06
9.1
1.01
35.4
1.432
58.04 871
1.124
35.9 280+3.3
.187
57.22 871
.3033
162
.856
31.46
.0548
62.62 871
1.931
-------
97
Julian
Date
166
166
166
166
166
Zero
Span
166
167
167
167
167
168
Zero
Span
168
Time
9:00
11:04
12:00
14:00
16:00
17:19
17:25
17:30
8:34
11:00
15:01
16:06
10:05
10:13
10:18
13:00
Table G-3 (Continued)
Channel
1 2
.22
.0224
_
25.98
.164
1.8
.1
89.38
.606
27.4
.19
25.8
.27
25.3
.1
26.2
.2
25.8
.2
23.5
.1
1.76
.05
90.84
.89
24.0
0
2.868
.00837
.334
.0344
.19
.042
.084
.032
.074
.0643
.794
.12
3.02
.08
.448
.05
.99
.05
1.32
.05
1.16
.1
1.08
.08
.952
.02
0
.03
4.05
.06
1.25
.04
3
2.71
.007
2.71
.03
2.876
.006
3.072
.413
.10
.01
8.47
.02
3.07
.035
2.81
.01
2.61
.01
2.65
.03
2.63
.02
2.54
.01
.66
.05
8.73
.04
2.46
.01
678 True Value
34.18 Ppm S02
.0837
32.78
.1643
47.74
1.014
32.3
.141
29.86
.134
3.42
.13
52.9 871
.33
33.82
.11
3.04
.15
29.04
.1
16.9
1.6
21.0
.07
23.7
.02
4.18
.32
39.64 871
.45
24.2
.1
-------
98
Table G-3 (Continued)
Channel
Julian
Date Time
172 11:18
172 13:00
172 14:14
172 15:00
173 8:43
173 ll:llb
179 15:00
179 17:00
179 19:00
180 9:00
Zero 9:50
.0837
Span 10:05
180 11:01
180 13:01
180 15 : 00 b
Zero
1
25.1
.1
25.7
.2
24.2
.2
25.4
.2
29.84
.167
30.46
.089
30.22
.217
28.82
.228
1.78
.0853
92.86
.513
30.0
30.66
.048
31.34
.152
1.94
.0894
2
1.14
.06
1.02
.05
.87
.09
.33
.03
.102
.04
1.21
.04
.77
.0583
.824
.0456
.956
.0477
.878
.0327
.212
.0466
3.4
.056
.964
1.022
.055
.982
.0795
.0417
.0337
3
2.59
.01
2.82
.01
2.94
.01
3.03
.01
.12
.03
.248
.025
2.936
.0089
3.124
.0114
3.072
.0082
2.74
.0071
.172
.02
8.21
.0783
2.848
2.936
.0055
3.04
.0122
.054
.0688
6
1.15
.02
1.72
.02
1.66
.01
.92
.02
.02
.122
2.76
.17
1.12
1.19
.02
1.11
.02
.22
.04
7 8 True Value
24.4 ppm S0?
.1
0
0
23.78
.09
23.5
.2
25.3
.09
27.62 262±8.6
.08
25.44
.114
25.42
.277
25.6
.100
27.42
.1095
4.4
50.56 871
1.49
29.46
27.36
.0894
27.08 288±6
.130
3.70
.158
-------
99
Table G-3 (Continued)
Julian
Date
Span
180
181
181
181
181
182
182
182
Time
15:19
17:00 b
9:00
31.2
13:00
15:00
9:00
13:00
15:00
1
94.52
.432
32.12
.259
30.68
.295
.984
.1414
30.8
.173
31.3
.235
31.42
.148
31.22
.164
.396
2
3.448
.0492
.954
.0573
.13
.0506
2.866
.0397
.850
.0469
.944
.023
.914
.0602
.986
.077
.0482
3
8.456
.0365
3.178
.053
2.784
.0152
1.70
.0089
2.88
.0173
2.92
.0141
2.776
.0055
2.736
.0089
.0114
6
3.57
.21
.17
.01
.77
.02
26.2
.03
1.66
.04
1.64
.02
2.08
.02
.31
.03
.02
7 8 True Value
ppm SCL
57.88 871
2.21
27.5 27814.0
.300
27.3
.1225
.100
25.38
.130
25.44
.0894
24.88
.0837
26.0
.100
.100
Channel Identification
1. Calibrated Instruments
2. DuPont
3. Inter tech
6. Dynasciences
7. ITT Barton
8. Bendix
Simultaneous sample - Analysis by barium perchlorate - thorin titration
-------
DATA SUMMARY
TIMt
CHANNELS
3
UATt 182
ZERO
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
MV
'10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
11.
11.
11.
11.
11.
11.
11.
11.
11 .
11 .
11.
11.
11.
11.
11.
31
31
31
31
31
45
45
45
45
45
00
00
00
00
00
14
14
1 4
14
14
30
30
30
30
30
0
0
0
0
0
78
78
78
78
7b
25
25
26
25
25
98
98
98
98
99
49
49
49
C9
4
-------
CALIBRATION 880 PPM
MV 11.42 99.00 3.51 8.03 1.57 54.90 3.92
MV 11.42 99.10 3.50 8.01 1.57 54.80 3.91
MV 11.42 99.00 3.56 8.00 1.62 54.90 2.7O
MV 11.42 99.20 3.50 8.02 1.62 55.20 2.74
MV 11.4?. 99.30 3.57 8.04 1.58 55.30 3.48
03
t—
m
CTl
Ji.
n
o
3
rt
H-
3
n>
o.
-------
STATISTICS SUMMARY
TIME
QATE 182
CHANNELS
236
ZERO
ACTUAL
MV MEANS
MV SIGMAS
ACTUAL
MV MEANS
MV SIGMAS
CORRECTED
MV MEANS
MV SIGMAS
ACTUAL
MV MEANS
MV SIGMAS
CORRECTED
MV MEANS
MV SIGMAS
ACTUAL
MV MEANS
MV SIGMAS
0.00
0.000 .809
CALIBRATION
78.6* 2.51
.230 .036
78.6*
.230
CALIBRAT ION
.25.88 .76
.130 .036
25.88
.130
CALI8RAT ION
98.78 *.2*
.295 .055
-.23 -.98
.016 .258
660 PPrt
6.11 .07
.011 .130
6.35 1.05
.020 .2B9
190 PPM
l.*2 .*£
.007 .01*
1.65 I.*L
.016 .2i>9
1080 PPM
9.78 L.b^
.008 .dJ*
3.08
.0*5
*7.68
.277
**.60
.201
30.50
.122
27. *2
.130
51. 1*
.089
loll
.559
1.06
o*3*
-.05
.708
1.08
.508
-.02
.756
2.60
.1*3
H
>
O3
m
Cl
i
.b.
"n
o
3
It
!-••
3
C
re
Cu
o
ro
-------
CORRECTED
MV MEANS
MV SIGMAS
ACTUAL
MV Mf-ANS
MV SIGMAS
CORRECTED
MV MFANS
MV SIGMAS
ACTUAL
MV MEANS
MV SIGMAS
CORRECTED
MV MEANS
MV SIGMAS
98.78
.295
CALIBRATION
49.48 1.64
.130 .009
49.48
.130
CALIBRATION
99.12 3.53
.130 .034
99.12
.130
10.01
.018
400
3.41
.008
3.64
.018
880
8.02
.016
8.25
.023
2.60
.261
PPM
1.07
.,019
2.05
.259
PPM
1.59
.026
2.57
.250
48.06
.100
47.78
.342
44. 70
.345
55.02
.217
51.94
.221
1.49
.577
1.98
.477
.88
.735
3.35
.602
2.24
.822
DO
t—
m
C~>
i
-t^
"n
o
3
r+
H-
O
O.
* — i
-------
MV R6AD INGS
GAS CONC
190
CALIBRATION CURVES
CHANNEL
3 6
660
MEAN
LOWER
UPPER
71.95
57.01
86.89
3.12
2.97
3.27
6.15
5.92
6.38
2.09
1.50
2.65
A3. 73
r-1
33.10
54.36
.94
-.72
2.60
MEAN
LOWER
UPPER
1080
MEAN
LOWER
UPPER
400
MEAN
LUWER
UPPER
880
MEAN
LOWER
UPPER
30.98
15.48
46.49
108.56
93.09
12L4.03
*9.29
34.18
64.39
91.12
76.03
106.22
1.28
1.13
1.44
4.76
4.61
4.91
2.10
1.96
2.25
3.98
3.83
4.13
1.71
1.47
1.95
10.12
9.88
10.26
3.69
3.46
3.93
8.23
7.99
8.47
1 .48
.88
2.07
2.62
2.02
3.21
1.75
1 .16
2.33
2.36
1 .78
2.94
33.62
22. 59
44.66
52.76
4i. 76
63. 77
38. 14-
2/.39
48.89
48. 46
37.72
59.21
.04
-1.69
1.76
1.75
.03
3.47
.44
-1.24
2.12
1.37
-.31
3.05
D3
t—
m
en
-^
o
3
3
C
to
O-
o
-------
105
TASK H
Data Evaluation
The field test program ended on 31 August 1971 after approximately
90 days of on-line testing were completed. Tables G-3 and G-4 show
typical "field" and "calibration" data obtained during the period. For
simplicity and ease of reference, the data actually used for the analysis
programs are reproduced in the appropriate tables when a particular
analysis is discussed.
Table H-l contains the data used in computing thp accuracy of the
units under test. The reference value given in column 2 is the value of
SO concentration in ppm (wet basis) given by triplicate reference method
analyses of duplicate samples taken from the sampling manifold (Fig. F-l,
point 14) at the same time the instrument readings were taken. The
indicated instrument values were the mean values of 10 readings taken at
1 minute intervals during the 10 minute wet chemical sampling period.
The reference method used was described in Appendix E-l. The values indi-
cated for Channel 3 were corrected for water vapor.
Table H-2 shows the analysis of the monitor's accuracy using the data
in Table H-l. Calculated were the average error, E, in ppm, the variance
around E, the standard deviation, and the 95% confidence interval according
to the following:
(1) mean E = ^ Z EI
,„ . , 2,
(2) variance (s ) =
N 7
n(n-l)
(3) standard deviation = V s^ = ^/variance
(4) confidence interval CI(95%) = E + t(—-
'
-------
106
where
n = number of observations
s. = value of an individual observation
t = percentile of the t distribution for t(l-o^2) and df = n-1
From Table A-4, Handbook 91, Experimental Statistics, National
Bureau of Standards, 1963, a= .05
The data produced by instruments connected to Channels 6 and 7 of the
data system were completely anomalous at the times at which wet chemical
samples were taken. Therefore, we are unable to indicate paired data
points for these channels.
Table H-3 contains the data used to compute "calibration drift."
There are two types of data analysis which can be used on the data in
Table H-3. A linear regression vs time should show the effect of any
systematic long term degradation effects as it does for Channels 1, 6 and 8.
In the case of the instrument on Channel 8, the change in slope probably
corresponded to the degradation of the sample cell. The same effect
was noted in a similar program studying NO monitoring equipment.^
In the cases where no linear correlation can be established, the
data is treated as normally distributed data around a central mean and
the usual standard deviation, variance and confidence intervals are com-
puted. Eimutis, in the NO monitor evaluation program, showed that the
A
distribution is not truly normal but the error associated with this assump-
tion are not large. The methods used are the same as for accuracy on
Tables H-l and H-2. The "calibration drift" analysis results are shown
in Table H-4.
Table H-6 shows the results of a simple correlation of the millivolt
readings for the various channels with each other and with the reference
value. The data used was taken on days 214, 215, 216 and comprises 18 data
sets including five reference tests and 1 each zero and span set. It
should be noted that the monitors correlate far better among themselves
(.9 or better) than with the reference method. Similar performance was
a. Instrumentation for the 'Determination of Nitrogen Oxides Content of
Stationary Source Emission, Vol. II, APTD 0942, NTIS PB 209190, Monsanto
Research Corporation, February 1972.
-------
107
was noted in the NO monitor evaluation program conducted by Monsanto
Research Corporation. A tentative explanation may be that despite our
taking 10 readings per channel at one minute intervals that the average
of these1readings still does not adequately represent the 10 minute inte-
grated sample taken in the reference method even though the overall
accuracy obtained by comparing mean values is reasonably good. The low
correlation of channel 8 with the reference value may be in part ascribed
to a shifting zero point of the instrument during this period. The
millivolt readings obtained during this time actually were smaller than
the "zero" reading obtained prior to the reference tests.
Table H-7 contains data which illustrates that which we choose to
call short term stability or rather short term "instability." The data
are not the same as reproduceability, as these are consecutive millivolt
readings for each monitor at about 10 second intervals with no change in
pollutant level. Hence the difference from reprodu c i bility. Repro-
duceability is the band around the mean values of successive measurements
of the same level generally with some change in level intervening. Here
we have a steady level and continuous readings which we record at 10
second intervals. Therefore we call this measurement "stability." Since
these measurements were made using calibration gases of known steady
composition there should be no source variations to consider here. Thus
we can assign all the deviation in the readings to the stability or
lack of stability in the basic instrument desgin.
The four tests already described constitute the major aim of our work.
These are however, several factors which are of lesser importance but
we have the data to investigate them.
In a laboratory when using a similar monitoring system for R and D work
it is not unusual to calibrate the instrument at several points over the
span. In the field, a user is more likely to put one calibration gas into
the system near full scale and twist the span knob until he gets the
same number. The obvious question is what type of error is introduced
by this latter practice? To attempt to answer the question we treated
the calibration data from one day in two ways as shown in Table H-9. We
computed a calibration curve slope based on the millivolt readings
-------
108
produced by a 1036 ppm calibration gas and used this value as a single
point calibration. We then computed a least squares regression line
based on all five calibration gas values and used this curve to predict
a value for each mv reading to compare with the value given by the single
point calibration curve. As can be seen from the final error figures in
Table H-10, in some cases a single point calibration can lead to substan-
tial errors in measurement. The instrument connected to channel 1 as a good
case in point, having a departure from linearity above about 90% of span.
After recomputing the linear regression calibration to exclude the 1036 ppm
point the values changed from 10.00 ppm/mv slope and -70.2 ppm intercept
to 8.66 ppm/mv slope and -14.8 ppm intercept. The error from the four
point calibration curve fell from 2.9% of span to /02% of span or .2ppm.
A similar single point calibration based on the 859 ppm calibration gas
changed the slope from 10.48 ppm/mv to 8.66 ppm/mv which will change the
single point calibration curve from 18% of span to less than 2% of span.
The next logical question is what has been proved by this exercise. We
consider that the following items have been illustrated:
a. The linearity of most properly operating equipment is within
±2% over most of their operating range.
b. The major departures from linearity occur at the extremes of
the range (below 10% and above 90% of span).
c. If an instrument is known to be functioning correctly a single
point calibration may be sufficient.
d. When an instrument is not known to be functioning correctly or
is functioning incorrectly, a single point calibration can lead
to extremely large errors.
-------
109
TABLE Hrl
ACCURACY DATA VS REFERENCE METHOD
(a)
DAY
162
173
180
180
190
216
216
216
216
REFERENCE
JVALUE
280
262
288
278
227
289
241
241
331
216 ; 284
224 ' 245
224 " 240
230 183
230 213
1
231.2
273.6
281.4
-
219.8
214.8
219.8
2.
326.3
260.2
227.0
219.0
214.4
239.3
231.3
221.1
227.8 ^ 232.5
225.8 215.5
238.8 253.4
238.8 246.6
237.5
191.3 240.9
DATA CHANNEL
3(b).
269.8
234.2
300.9
315.4
254.2
268.0
266.4
264.8
273.8
273.8
274.4
274.4
-
-
6 ; 7
231.4
;
8(c)
254
204
i
i i
i i
i
!
1
207 ;
204
314
86.8,
d
96.2,
d
30. 5d
38.65,
d
a. all data ppm SCL
b. output corrected for 4% H~0 content of stack effluent
c. unit placed in service day 215 after repair
d. shown for illustration, not used in error analyses
-------
no
TABLE H-2
ACCURACY ANALYSIS
i! "" i
i,
:j
^Average error, E
3.
: „ . 2
; Variance s
Standard deviation
Number of samples
t. 975
Confidence interval (95%)
E/span(b)
95%CI/spanf,,
CHANNEL
1
- 31.7
959.5
30.97
11
2.228
±20.80
-3.2%
+ 2.1%
i i
1 :
2
:
-16.9
1908
43.68
: 14
2.160
25.21
-1.7%
±2.5%,
1 i
: 3 . ,
i
i
+5.3
: 826 :
i 28.74 !
12 ;
2.201'
18.26 '.
-.5%
+1.8% :
8
-40.6
2507
t
50.07:
5
2.776
(
62.16
-4.1%
+6.2%;
a. values are ppm SCL
b. span set at 1000 ppm SO-
-------
in
TABLE Hr3
CALIBRATION DRIFT DATA
JULIAN
_DATE_
148
160
182
196
215
223
231
236
ELAPSED Channe
TIME £ -)
0
12
34 -16.2
48 24.1
11 Channel 2
Tl_ "\ ^
34.0
59.1
j 8.67 -1.2
10.50 -23.0
67 14.0 10.69 12.9
75 18.7
83 9.5
88 16.5
238 90 18.1
i
239 91 17.4
243 :• 95
11.08 -19.9
11.12 10.4
219.62
221.65
Channdl 3
I
58.2
87.7
-236.29 49.9
103. 37J 37.6
115.66 43.5
106.60
97.85
-
3.4
10.47 -20.9 121.94 -15.4
I 10.67 97.1 128.13
10.72 24.3 129.57
I !
| - 31.9
T j
130.13
3.6
36.9
-22.5
S
92.72
Channe'l 8
I
-
91.34
97.67
_ . S. -
-
(d)
92.18
90.90
-
87.69
88.00
90.84
129.4
149.3
151.29
140.0
138.6
79.15 133.5
84.94
173.8
-
173.58
165.85
165.03
159.5
161.0
158.96
152.94
... I
(a) I = Intercept of PPM vs MV curve for 5 calibration gases (ppm)
(b) S = Slope of PPM vs MV curve for 5 calibration gases (ppm/mv)
(c) output span voltage changed from 5MV to 10MV Fullscale
(d) Data for this date completely anamolous, no curve fit possible
-------
TABLE H-3 (Continued)
112
DAY
148
160
182
196
215
223
231
236
238
239
243
ELAPSED I Channbl 6 | Channel 7 j
TIME ! I | Si I S
0
12
34
48
67
75
83
88
90
91
95
3.7
-
-839.1
-
-4.0
-11.4
7.0
139.38
-
716.61
-
I
1
260.0
-726.0
-4081.9
-861.8
-
-139.1
8.46 ;
29.87 '
110.92
34.70
I
18.42 '
38.8 28.54 ;
156.29 ' -23.2 17.8
162.03 . -2.4 , 23.34
165.25 -2.5 26.12
-------
113
TABLE H-4
CALIBRATION DRIFT ANALYSIS
CHANNEL
_ __ ... .1 2 . .3
Linear Regression analysis of slope data (Table H-3)
Correlation coefficient .895 .663 .657 .87 (a) .877
' Slope (PPM/MV)/day .045 .537 -.093 1.55 - -.599
; Span Voltage 100MV 10MV 10MV 6.5MV - 6.5MV
95% Confidence interval about slope .103 - - - -1.4 (ppm/mv)
'"Correlation with time" (b) linear random random linear - linear
'1
Analysis of non-time correlated channels
Mean, x (ppm)
Variance
Standard deviation (ppm)
; Number of samples
t. 975
1
Confidence interval (95%)
115.5
111.32
10.55
11
2.228
7.08
89.54 - - -
22.49 '••
4.74 - - -
10 - - -
2.262 -
3.4 ;
"Calibration drift" %span/24 hr 1.03% .7% : .3% ; 1% - ' .9%
'a. anomalous data
b. linear means simple correlation coefficient of 0.9 or greater.
-------
TABLE H-5
LONG TERM ZERO DRIFT
114
Date
162
162
166
168
180
180
189
194
194
195
215
218
218
222
223
224
225
228
228
228
229
229
Timew
100.. 5
1052
1103
1220
1719
10:13
9:50
1510
1748
1347
1515
10:11
1539
0806
1147
1923
1633
1635
0953
1352
1412
1555
0852
1558
Ch 1
1.92
1.966
1.87
4.44
1.8
1.76
1.78
1.94
1.94
-
1.48
1.94
1.86
0.4
0.58
1.7
1.9
1.9
1.7
1.8
1.8
1.8
-
.7
Ch 2
.134
.0267
.007
.162
.794
0
.212
.0417
.002
.08
-.004
.03
.16
.18
.08
.27
-.004
-.06
-.15
-.01
.08
-.04
.08
-.23
_ .
Ch 3
.198
.133
.263
.262
.10
.66
.172
.054
.106
-.04
0
.07
-.26
-0.27
.54
.99
-.26
-.27
1.18
-
-
-
, — ...
Ch 6
.02
.22
-.084
-
-
-
-
-
-
-
-
-
-
-
-
-
Ch 7 Ch 8
4.77
6.17
9.1
162
3.42
4.18
4.4
3.70 !
2.88
.54 -.82
.57 ' -.81
.01 -.95
8.26 -1.21
1.7 1.01
1.1 1.04
-
1.20 1.02
1.06
1.01
.96
1.08
.77
.94
1.04
.. ._ ...
-------
115
TABLE H-5 (Continued)
1! !
ll !
Date ii Time : Ch 1
..jj .......
230 1121
230 1448 . 2.0
235 1500 I 0.42
235 2055 ' 1.9
237 1038 . 0
237 1508 1.8
243 'i 1339 0
243 '. 1410 2.0
mean zero point 1.62mv
variance .67
standard deviation .82
95% confidence
interval 1.76mv
number of points 29
calibration slope
ppm/ mv 11
zero offset 1.8
zero band 1 . 9
~r- i r i !
1 i i i
Ch 2 ! Ch 3 Ch 6 Ch 7 Ch 8 !
- ; " : ' i
-.06 ' -.11 1.09 |
'• , \
-.11 ; .20 ! 1.19 '
.3 -.10 ;
.25 .27 . . !
-.11 0 .6 1.01 '
.05 -.27 .24 1.01 '
.34 .22 -.09 .8
.16 -.24 -.08 .82
r
:
-
.08 .12 -.003 2.86 .58 mv
.03 .13 .014 7.06 .68
.18 .36 .12 2.65 .82 mv
.36 .74 .33 5.6 1.73 mv
32 28 5 19 18
120 90 160 28 160 ppm/mv
.96 1.1 -.04 8. 9.2 % of span
4.3 6.7 5.3 15.7 27.7 % of span
a. Time is 24 hour clock
-------
116
TABLE H-6
MONITOR CORRELATIONS
Data taken from days 214, 215, 216
i 1 ! ! i
Monitor to monitor correlation
1
^
2 .9796
3 .9425
6(a)
7 .8550
8 .8378
i :
i Channel '
2 i 3 6 (a) 7 i 8
i i
| • ;
.9796 ! .9425 - .8550 .8378
.9384 - ' .9138 ; .8145
.9384 - - .8604 ! .7050 |
.9138 .8604 i - ' - .9014(b)
.8145 ', .7050 - .9014O)
; ' i
i Correlation of monitors to Reference method :
Channel ;
! 1
i .8281
2 3 67 8
.2890 .9533 - .1667 :
(a) unit not operational '
(b) 4 data points including 1 zero and 1 span
-------
117
TABLE H-7
SHORT TE'RM STABILITY
| ! Channel ! | i
me Level ! 1
zero
it
IT
it
ii
Std. Dev.
646
Std. Dev.
859
Std. Dev.
216
mv
0
0
0
0
0
0
78.60
78.90
78.80
78.30
78.60
.23
99.00
99.10
99.00
99.20
89.30
.130
25.80
25.90
! 26.10
. 25.80
25.80
Std. Dev. .130
Average S.D. ,016mv
Span mv 100
deviations, % span .16%
2 !
[
i
-.19 I-
-.10 !
-.19 !
-.13
.809
2.47
2.48
2.51
2.52
2.56
.036
3.51
3.50
3.56
3.50
3.57
.034
.71
.81 !
.76
.76 :
.77 :
.036 ;
.035
.229
5.0
.7%
3 6
-.23 •• -.84
-.22 ' -.72
-.23 : -1.40
-.22 ! -1.01
-.26 . -.93
.016 ; .258
6.12 .81
6.11 .46
6.11
6.13
6.10
.011
8.03
8.01
8.00
8.02
8.04
.016
1.42
1.42
1.43
1.42
1.41
.007
.67
.71
.72 i
.13 ;
!
1.57 •
1.57 i
1.62 :
1.62
1.58
.026 ;
.41 :
.40
.45
.41 •
.41
.019 .
.0125 .108
10.0 2.0
.12% 5.4%
7
3.10
3.10
3.00
3.10
3.10
.045
47.80
47.50
47.40
47.60
48.10
.277
54.90
54.80
54.90
55.20
55.30
.217
30.70
30.50
30.50
30.40
30.40
.122
.165
60.0
.27%
8
.61
1.56
1.61
.41 :
1.35
.559 .
1.51
1.29
.52 '
.68 :
1.31
.434
3.92
3.91 ,
2.70
2.74
3.48
.602
.53
1.65
1.49
1.09
.63
.508
.525
5
10.5%
-------
118
DATA
CHANNEL
Table H-8
IDENTIFICATION OF MONITORING INSTRUMENTATION
1 Calibrated Instruments MSK-S02-E2
2 Dupont Model 460
3 Intertech Model URAS-2
6 Dynasciences Model SS 330
7 ITT Barton Model 400
8 Bendix UNOR-6
-------
TABLE H-9
119
!j ! P T Channel \
ji PPM
" jj
span
Zero (a)
i 646
j
1 2 | 3 6 ' 7 8
| !
i
0 0 i -.23
78.64 2.51 i 6.35
; 216 ' 25.88 .76 ; 1.65
1
1036 98.78 ; 4.24 \ 10.01
\ 413 • 49.48 1.64 \ 3.64
; ;
;
-.98
1.65
3.08
44.60
1.40 27.42
2.60 48.06
2.05 44.70
1.11
-.05
-.02
1.49
.88
859
99.12 3.53 , 8.25 . 2.57 51.94 2.24
regression : • : .
slope 10.00 235.5 97.52 523.66 28.89 ppm/mv
intercept -70.2 36.7 50.79 -.441.6 -618 ppm
single point
slope 10.48 244.3 103.5 398.5 21.56
ppm/mv
-------
TABLE H-10
120
PPM
Value Remarks
646 Single point pred.
Lin. regression
pred.
Channel
1 23 6 7
824 613.2 657.2 657 961.6
716.2 627.8 670.0 422 670.4
216 Single point pred. 271.2 185.7 170.8 558 591.2
Lin. regress, pred. 188.6 215.7 211.7 291.5 179
413 Single point pred. 518.6 400.6 376.7 816.9 963.7
Lin. regress, pred. 424.6 422.9 405.8 631.9 673
859 Single point pred.
Lin. regress, pred.
1039
921
Single point calibration vs
Error, x
V
SD
N
x , % span
Linear regression (5
Error, x ppm
V
SD
N
129
2748.5
52.4
4
13%
point)
29.1
1567.5
39.6
4
862.4
868.0
"True
18.0
214.9
14.7
4
1.8%
853.9
855.3
Value ppm
18.8
522.7
22.86
4
1.9%
1024.1
904
II
230.5
23739.
154.1
4
23%
calibration vs "true value
.17
127.5
11.29
4
2.5
156.3
12.5
4
28.9
25622
160
4
1119.8
882
375.6
11859.8
108.9
4
37.6%
It
67.5
12962
113.9
4
x, % span
2.9%
.02%
.25%
2.!
6.7%
-------
121
CONCLUSIONS AND RECOMMENDATIONS
Conclusions
The conclusions which may be drawn from the data acquired during
the program are to be used for several purposes.
The original objective of the program was to provide the Division
of Control Systems with the technical information necessary to select
instrumentation for in-house and contract SC^ control technology develop-
ment programs.
For this immediate need, on the basis of data from the field test
program, the following equipment is recommended for use in S02 control
technology development programs:
a. Calibrated Instruments Co. model MSK-S02-Ej conductometric
S02 analyzer installed according to schematic 521-51/2.
b. DuPont model 460 ultraviolet S02 analyzer when calibration
gas standards are used instead of the optical calibration
filter for calibration purposes.
c. Intertech model URAS-2 non-dispersive infrared S02 analyzer
. when used with model 7651 probe and filter, model 7865 sample
conditioning unit and model CMR 5869 auto zero and calibration
units. The CMR 5869.unit shall be set at 8 hour intervals.
These recommendations are made on the basis of:
a. Proven reliability and performance.
b. The availability of the unit as a complete system.
The specifications suggested in appendix B-l and B-2 were found to
be too rigid for the type of field service which is found in this appli-
cation. Therefore, the judgements as to the performance of these systems
were made in a largely subjective manner rather than by comparison with
fixed criteria. We do not feel that this approach diminishes the validity
of the conclusions, but rather indicates that we aimed our sights at too
high a level of performance before obtaining the field experience.
The second objective of this program was to define the state-of-the-
art in S02 monitoring techniques for the Environmental Protection Agency,
in order that EPA could make basic decisions as to the availability of
-------
122
equipment for possible monitoring requirements; to provide a basis for
future Research and Development programs; to provide an information base
for providing "guidance on the selection and use of continuous monitoring
instrumentation as required by Part 60, Chapter I, Title 40, Code of
Federal Regulations."
The conclusions from interpretation of the data in this report con-
cerning the definition of the state-of-the-art in S02 monitoring are:
(a) There is equipment available which is adequate for the monitoring
purpose required by the New Source Performance Standards of
December 23, 1971.
(b) The equipment which does operate properly is that which is
available as a complete system consisting of probe, particu-
late filtration, sample conditioning and analyzer equipment.
The Calibrated Instruments and Intertech units listed on the
previous page are German built units which have 4 to 6 years
of operating experience in stack S02 monitoring.
(c) No particular measurement technique is superior to all others.
Acceptability depends on a particular implementation of the
idea in the analyzer and the supporting sampling equipment.
For example, the failure of the Dynasciences unit on channel 3
was probably due to our original insufficient temperature con-
ditioning of the sample and the degradation of the sensing cell
under a hot gas stream. If the unit was retested using the
sampling systems since developed and made available by the
manufacturer, the unit would probably exhibit reasonable per-
formance .
(d) A major operational problem encountered with most of the units
was the variance or instability of the "zero" setting (Tables
H-5 and H-7). The consistency of the "zero" of the unit on
channel one was due to the self-zero or balancing of the unit
each measurement cycle. The units on channels 2 and 3 contained
-------
123
zero correction circuitry which operated on a longer time cycle
(30 minutes to 4 hours), and thus, exhibited a larger "zero band"
around the automatic set point. Channels 6, 7, and 8 had no
automatic zero correction circuitry.
The following are recommendations for future research and development
programs to improve stationary source monitoring systems.
(a) Development of improved particulate and t^O vapor removal devices
to be used with existing analyzer equipment.
(b) Development of automatic zero and calibration drift correction
devices for use on existing equipment.
(c) Improvement in the zero and calibration stability of existing
units.
(d) Development of improved calibration devices and techniques.
(e) Development of extractive sampling monitoring systems which do
not require removal of particulate and h^O vapor or are more
tolerant of these substances and do not require complete removal.
(f) Development of in situ monitoring systems which do not require
sampling and conditioning systems.
(g) A thorough investigation of the losses and conversions of pollu-
tants in sampling and sample conditioning systems while removing
particulate matter and water vapor.
It is further recommended that a periodic re-examination of the state
of the art in SC^ monitoring be performed. A number of new analysis tech-
niques as well as major product improvements have appeared on the commercial
market even as this program was underway.
-------
124
Manufacturers should be encouraged to use analytically verified cali-
/
bration gases as primary standards rather than electrical offsets or
optical absorption filters. The latter are very convenient but have proved
less than completely reliable in actual service and do not allow functional
checking of the sampling system.
The manufacturers are also encouraged to provide complete "turnkey"
installation service with their monitoring systems. The installation
difficulties encountered by even the experienced TRW personnel indicates
that normal plant operating personnel would have a very difficult time
installing and putting into operation most units. This is another reason
for treating the monitoring equipment as a complete "system."
U. B. GOVERNMENT PRINTING OFPICEi 1973 746769/4164
-------
BIBLIOGRAPHIC DATA '• RcPort N°- 2-
SHEET EPA-R2-73-163
4. title and Subtitle
Monitoring Instrumentation for the Measurement of Sulfur
Dioxide in Stationary Source Emissions
7. Author(s)
Fredric C. Jaye, EPA
9. Performing Organization Name and Address
TRW Systems Group-
One Space Park
Redondo Beach, California 90258
12. Sponsoring Organization Name and Address
Environmental Protection Agency
Office of Research and Monitoring
Washington, D. C.
3. Recipient's Accession No.
5- Report Date
February 1973
6.
8. Performing Organization Rept.
No.
10. Project/Task/Work Unit No.
17205
11. Contract/Grant No.
EHSD 71-23
13. Type of Report & Period
Covered
Final Sept . 1970-Sept .
14.
15. Supplementary Notes
Prepared by EPA Project Office
16. Abstracts
This report covers work done by TRW Systmes group to assess the state of the
art in automatic S02 monitoring systems for fossil fuel fired combustion units.
The results indicate that adequate monitoring equipment is available to meet EPA
New Stationary Source Performance Standards monitoring requirements as long as
the units are dealt with as total systems-(i.e. probe, sample conditioning, and
analyzer as an integral unit.) Equipment selections were made on an arbitrary
basis from that equipment available in December 1970 and was not intended to be
monitor vs. monitor comparative or all inclusive in any manner. No approval,
certification, recommendation, or acceptance is made or implied other than for the
very specific purposes clearly stated in the report.
17. Key Words and Document Analysis. 17o. Descriptors
Key Words:
monitoring instrumentation, air pollution measurements systems, stack monitoring,
source monitoring.
17b. Identifiers/Open-Ended Terms
17e. COSATI Field/Group
18. Availability Statement
release unlimited
19. Security Class (This
Report)
UNCLASSIFIED
20. Security Class (This
Page
UNCLASSIFIED
21. No. of Pages
22. Price
FORM NTIS-33 1REV. 3-72)
USCOMM-OC I4932-P72
-------
INSTRUCTIONS FOR COMPLETING FORM NTIS-35 (10-70) (Bibliographic Data Sheet based on COSATI
Guidelines to Format Standards for Scientific and Technical Reports Prepared by or for the Federal Government,
PB-180 600).
I. Report tjumber. Each individually bound report shall carry a unique alphanumeric designation selected by the performing
organization or provided by the sponsoring orgar^ation. Use uppercase letters and Arabic numerals only. Examples
FASEB-NS-87 and FAA-RD-68-09.
2. Leave blank.
3. Recipient's Accession Number. Reserved for use by each report recipient.
4* Title and Subtitle. Title should indicate clearly and briefly the subject coverage of the report, and be displayed promi-
nently. Set subtitle, if used, in smaller type or otherwise subordinate it to main title. When a report is prepared in more
than one volume, repeat the primary title, add volume number and include subtitle for the specific volume.
5- Report Dote. I'.ach report shall carry a date indicating at least month and year. Indicate the basis on which it was selected
(e.g., date of issue, date of approval, date of preparation.
6> Performing Organization Code. Leave blank.
7. Authors). Give name(s) in conventional order (e.g., John R. Doc, or J.Robert Doc). List author's affiliation if it differs
from the performing organization.
8. Performing Organization Report Number. Insert if performing organisation wishes to assign this number.
9. Performing Orgonizotion Name and Address. Give name, street, city, state, and zip code. List no more than two levels of
an organizational hierarchy. Display the name of the organization exactly as it should appear in Government indexes such
as USGRDR-I.
10. Project/Task/Work Unit Number. Use the project, task and work unit numbers under which the report was prepared.
11. Contract/Grant Number. Insert contract or grant number under which report was prepared.
12- Sponsoring Agency Nome and Address. Include zip code.
'13* Type of Report and Period Covered. Indicate interim, final, etc., and, if applicable, dates covered.
14* Sponsoring Agency Code. Leave blank.
15. Supplementory Notes. Enter information not included elsewhere but useful, such as: Prepared in cooperation .with.. . .
Translation of ... Presented at conference of ... To be published in ... Supersedes . . . Supplements-.
16. Abstract. Include a brief (200 words or less) factual summary of the most significant information contained in the report.
If the report contains a significant bibliography or literature survey, mention it here. . . ;
17. Key Words and Document Analysis, (a). Descriptors. Select from the Thesaurus of Engineering and Scientific Terms the
proper authorized terms that identify the major concept of the research and are sufficiently specific and precise to be used
as index entries for cataloging.
(b). Identifiers and Open-Ended Terms. Use identifiers for project names, code names, equipment designators, etc. Use
open-ended terms written in descriptor form for those subjects for which no descriptor exists.
(c). COSATI Field/Group. Field and Group assignments are to be taken from the 1965 COSATI Subject Category List.
Since the majority of documents are multid isc iplinary in nature, the primary Eield/Group assignments) will be the specific
discipline, area of human endeavor, or type of physical object. The application(s) will be cross-referenced with secondary
Field/Group assignments that will follow the primary posting(s).
18. Distribution Statement. Denote releasabj 1 ity to the public or limitation for reasons other than security for example "Re-
lease unlimited". Cite any availability to the public, with address and price.
19 & 20. Security Classification. Do not submit classified reports to the National Technical
21. Number of Pages. Insert the total number of pages, including this one and unnumbered pages, but excluding distribution
list, if any.
22. Price. Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
FORM NTIS-35 IREV. 3-72) USCOMM-DC 14932-P72
------- |