8 RESOURCE MANUAL
FOR IMPLEMENTING
THE NSPS CONTINUOUS
8 MONITORING REGULATIONS
Manual 1 - Source Selection
LU
g and Location of Continuous
Monitoring Systems
CO
>»
(0
c
o
75
35
(0
in
o
o
CO
00
& U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF ENFORCEMENT
OFFICE OF GENERAL ENFORCEMENT
WASHINGTON, D.C. 20460
-------
EPA-340/1-78-005a
RESOURCE MANUAL
FOR IMPLEMENTING
THE NSPS CONTINUOUS
MONITORING REGULATIONS
Manual 1 - Source Selection and Location
of Continuous Monitoring Systems
by
F.Jaye, J. Steiner, and R Larkin
Acurex Corporation/Aerotherm Division
485 Clyde Avenue
Mountain View, CA 94042
Contract No. 68-01-3158
EPA Project Officer: Louis Paley
Division of Stationary Source Enforcement
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Division of Stationary Source Enforcement
Research Triangle Park, North Carolina 277 n
April 1978
-------
STATIONARY SOURCE ENFORCEMENT SERIES
The Stationary Source Enforcement series of reports is issued by the Office
of General Enforcement, Environmental Protection Agency, to assist the
Regional Offices in activities related to enforcement of implementation
plans, new source emission standards, and hazardous emission standards to
be developed under the Clean Air Act. Copies of Stationary Source
Enforcement reports are available - as supplies permit - from the U.S.
Environmental Protection Agency, Office of Administration, Library
Services, MD-35, Research Triangle Park, North Carolina 27711, or may be
obtained, for a nominal cost, from the National Technical Information
Service, 5285 Port Royal Road, Springfield, Virginia 22161.
REVIEW NOTICE
This report has been reviewed by the Division of Stationary Source
Enforcement and approved for publication. Approval does not signify
that the contents necessarily reflect the views ana policies of the
Environmental Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for
use.
-------
TABLE OF CONTENTS
Section Page
A INTRODUCTION 1-1
B MONITOR SELECTION 1-1
1. Regulations 1-1
a. Regulations Affecting Selection of Monitoring
Equipment 1-2
b. Regulations Affecting Selection of Data
Handling Equipment 1-2
2. Major Equipment Types and Operating Principles ... 1-6
a. Opacity Monitor Principles 1-6
b. Gaseous and Diluent Monitor Principles 1-8
3. Monitoring System Selection Criteria 1-23
a. Prior Experience of Users and Agencies 1-24
b. Specific Criteria for Gaseous and Diluent
Monitoring Systems 1-25
c. Specific Criteria for Opacity Monitoring
Systems 1-28
d. Maintenance Availability and Requirements ... 1-32
e. Gas Conditioning Requirements ... ... 1-32
f. Selection Criteria for Data Handling
Equipment 1-33
g. Meeting Regulation Requirements T-33
h. Level of Automation of Functions 1-3^
4. Equipment Provisions for an Agency Observing and
Verifying Data 1-34
C MONITOR LOCATION 1-40
1. Regulations 1-4C
2. Installation Specification for Opacity Mom tori nc
Systems 1-41
a. Representative Effluent 1-42
b. Multiple Feed Points 1-45
c. Stratification 1-46
d. Condensed Water 1-47
e. Installation and Maintenance Considerations . . 1-48
-------
TABLE OF CONTENTS (Concluded)
Section Page
C (Continued)
3. Installation Specification for Gaseous and
Diluent Emission Monitoring 1-49
a. Representative Effluent 1-49
b. Multiple Feed Points 1-52
c. Stratification 1-52
d. Dilution 1-54
e. Accessibility for Agency Inspection and Source
Maintenance 1-55
D CONTINUOUS MONITORING SYSTEM SELECTION AND LOCATION
CHECKLISTS 1-56
REFERENCES: SOURCE MONITORING EVALUATION PROGRAMS ... 1-64
IV
-------
LIST OF TABLES
Table Page
1-1 SUMMARY OF NSPS REGULATIONS AFFECTING SELECTION OF
CONTINUOUS MONITORING EQUIPMENT ..... 1-3
LIST OF ILLUSTRATIONS
Figure Page
1-4 SINGLE-ENDED OPACITY MONITORING SYSTEM ......... 1-7
1-5 DOUBLE-ENDED OPACITY MONITORING SYSTEM ......... 1-9
1-6 NDUV ANALYZER ...................... 1-10
1-7 NDIR ANALYZER ...................... 1-12
l-8(a) A VARIABLE WAVELENGTH NDIR ANALYZER ........... 1-14
l-8(b) ANOTHER NONDISPERSIVE ANALYZER ............. 1-16
l-8(c) SINGLE-BEAM, DUAL WAVELENGTH NDIR ............ 1-17
l-8(d) A DISPERSIVE IR ANALYZER ................ 1-17
l-9(a) CHEMILUMINESCENT ANALYZER .............. 1-19
l-9(b) FLUORESCENCE ANALYZER .................. 1-21
1-10 ELECTROCHEMICAL ANALYZER ............... 1-22
1-11 MONITORING SYSTEM OUTPUTS ................ 1-29
1-12 TYPICAL ANALYZER CONDITIONING SYSTEM ......... 1-34
1-13 VELOCITY AND PARTICULATE DISTRIBUTION IN A RECTANGULAR
DUCT .......................... I-43
1-14 FLOW RECIRCULATION AND PARTICLE DEPOSITION IN A
RECTANGULAR DUC1 ..................... T'44
1-15 STRATIFICATION DUE TO POOR INITIAL MIXING S0?
CONCENTRATION (ppm) IN A RECTANGULAR DUCT . ...... 1-53
-------
A. INTRODUCTION
Manual 1, Source Selection and Location of Continuous Monitoring
Systems, of the "Resource Manual for Implementing the NSPS Continuous
Monitoring Regulations" presents all NSPS regulations for selecting and
locating opacity and gaseous monitors for affected sources. In addition,
the regulations have been presented in a form that will be of practical
value to the observer and the affected source operator. Specific problem
areas such as stratification, dilution, water vapor interferences and multiple
feed points, are discussed.
Since a good observer must have an understanding of the operating
principles of opacity and gaseous monitors, this Manual contains discussions
and schematic diagrams of the most common types of monitors in use today
Discussions of criteria for selecting monitoring systems, equipment provisions
for Agency observation and a referenced checklist are also included.
Manual 1 is one of a series that comprise the "Resource Manual for
Implementing the NSPS Continuous Monitoring Regulations." The other manuals are:
Manual 2 Preliminary Activities for Continuous Monitoring System
Certification (Installation, Notification and Performance
Evaluations)
Manual 3 Procedures for Agency Evaluation of Continuous Monitor Data
and Excess Emission Reports
Manual 4 Source Operating and Maintenance Procedures for Continuous
Monitoring Systems
B. MONITOR SELECTION
1. Regulations
The regulations for selecting monitoring equipment are found in the
subparts of Part 60 that deal with specific classes of sources. These
regulations do not specify a certain brand of monitor or a specific operating
principle. In most cases, the regulations specify the emission to be monitored
and identify the Performance Specification that the continuous monitoring
r;y,fp
-------
a. Regulations Affecting Selection of Monitoring Equipment
Table 1-1 presents a summary of the NSPS regulations which will have
an effect on which continuous monitor a source may select.
b. Regulations Affecting Selection of Data Handling Equipment
Although the regulations do not place particular requirements on the
data recorder component of measurement systems, many requirements do affect
the selection of a data recorder, the operating procedures to be used, and
the way recordings are read and reported. In describing apparatus, most
subparts of Part 60 include general statements on data recorders. For example,
60.152 (amended) Paragraph 19, Appendix B, Paragraph 2,2 reads:
Data Recorder. Analog chart recorder or other suitable device with
input voltage range compatible with the analyzer system output. The
resolution of the recorder's data outputs shall be sufficient to
allow completion of test procedures in this specification.
More specific data recorder requirements are also included in Part 60
Appendix B - Performance Specifications. For example, Paragraph 4.3 reads:
Data Recorder Output. The continuous monitoring system output shall
permit expanded display of span calibration on a 0- to 100-percent
opacity scale.
And 60.45(a) contains the following statement:
— the continuous monitoring system shall be spanned at 80- or 90-
or 100-percent opacity.
Since the data recorder is a subsystem of the continuous monitoring
system, there are general performance specifications for the analyzer, inter-
face, and recorder. The relevant portion of the definition of a continuous
monitoring system is in 60.2(y):
"continuous monitoring system" means total equipment... to analyze,
and to provide a permanent record of emissions or process parameters.
1-2
-------
'tt^_
UJ
Q_
t— 4
i
o-
UJ
C_J
^•^
1— 1
O
H-
f i
o
oo
—•^
Q
_„ )
^^
h-H
1—
o
O
u_
O
g^
^-^
^"^
*""*
l_
0
UJ
i
1 1 1
oo
C3
*
\ t
r"^
LU
U-
OO
o
1 — I
1 —
<
— 1
o
UJ
oo
CL.
OO
1 1
o
^_
Q;
S;
ZD
c/l
^_
1
LjJ
i
1
CQ
1 —
4-> -o
00 o
CU -C
OJ
oi s:
o
C OJ
CO O
E £
0 S.
4- 01
i- 4-
OJ OJ
C
o
•r—
CO OO
1- co
_O o
••-
CO
<-J
OJ
C 3
(0 t—
CL co
CO ;»
^
C
rO
•4-J
3J
,—
O
Q_
CD
4-)
^
•r- Q;
Ql U
c
cr: QJ
»-•• CU
'- OJ
OJ
Ll- 1
,
|
-o :>,
Ol 4->
u , —
CU -r- ,
4- l_>
4- 10
*X ' '
;>^
Ol 1-
0 O '
s- cn
3 01 ,
O 4->
o I
1
ua r-
-o -o
, 0 0
1 -c-c
•)
4J
•r- CM X
U O O
Q.
CD
— -
T3 -Q U
UO LO LT5
«3- ^ ^r
OOO
VI
5^
QJ
o
-Q
O)
•r—
14—
o
s^
o
4-)
tJ
g_j
C
cn
E
OJ
l/l
r~~ ko l^~
•O T3 -a
0 ,00
j= 1 -c -c
01 a; cu
CM
o loo
~z. co z
a. 0 a. o.
Q- O O- 0.
0 '"" O O
0 1 0 O
cn O O
0 r- ,—
CO
Jfc,
X — CM
0 OOO
Z >
O X
CO >,^-
o
0 O
0 0
. — CO
II II
;>^
+J
•r- CM X
OOO
O.
O
, — «
' —
o
O 4-> r—
O 4-> CO (O
r- 0 S- >
OJ *J O
i -i-> in i-
J3 -r- O.
§3 C O.
OO -r— CO
1
^
4->
•r- CM
O O
CO CO
O-
o
s~~*
• — •
as JQ
tn LO
•y *t
O O
r— i—
•r- OJ
OO •!—
O O
4- .0
1
c -o
O Ol
c s-
i — 4-
• ^
00 I —
00 OJ
O 3
U- U-
ro
•a
0
2
C T3
•r- C C
CO cO
CM t- O.
O -r- OJ m
(_> co Ol
C 4-
t- 1- co O
O O S-
CM CM-r- O
O z E cn
c;
0
CO CM
S- A
^•^
CO
0 S-
CO
O- L-
00 O
4- 00
O co
cn
&£
O
O!
CM
CD
s_
O
CM
O
-o
cn
*»•
o'
oo
t_
CU
O
a
^
r^ oo
-O 13
0 0
-C -C
cu cy
X CM
0 O
Z CO
t i.
a. o.
o o
o o
co O
X CM
O O
Z CO
CO CT3
ro ro
r- oo
d O
l£J vO
4-> 40
C C
cu cu
Q. CL
3 3
cr o-
ai oi
00 l/l
CU Cu
o o
£ 2
o- a.
4-J
i/l C
C i —
(O CL
0- T3
"C O
i- co
*O CJ
•r-
O L.
•c- 3
i- 4-
4-1 ^»
•r- 3
Z CO
U3
T3
, O
1 J=
£
1 CJ^
CO
O O-
CO Q. <
£
1 O C
0 t
o ^~-
--*) '
•r- CM -r
O O t
CO CO 11
CL C
O C
*— *r-~*
•— ro
*-^^~* *-
•H-H -H
o o r-
r— r— C<
O O C
IO IO k
CO S^
>, 0
CO CO
4-> I.
co eu
u c
cu
<_) CT
<_> QJ
Lu S_
b
00 C
CU I
•r~ rt
S- c
CU i
c :
•r- 4.
4-
CU <.
S- i
^
3 C
Ol -r
O 4-
S- C
*-" a
OJ i-
D- U
VO
-o
, o
1 -c
«
1 o^1
CO
J3
g
>2 O
o >
o
1) r— >1
= J^
3 1
§0
CM
0
*"> ^>
r- CNJ
J DO
a 03 uo
X Q-
3 O
*•—• «-^— *
. — C\J
-. ^-^-— v
3 -Q -Q
- £ £
J i — i —
D CD O
D ^O '-O
cn
c
4-> QJ
r- O-
s&
oo o
- « OJ
S- CU 4J
O) U S-
t- +-> co OJ
OJ t/l C >
>-, CO !- C
S- O 3 0
O Q£ 4- O
T
LJ
J
3
3
S-
OJ
J Q.
O.
3 O
U
;
. u1 oi
J E r—
J *r— CU
J CL, 00
CM
"••»
1
>^
OJ
.^
1 '
O
Oi
o.
00
OJ
oo
CU
<4-
.f-
00
00
0
4-
-o
O
00
-o
C
CO
-a
^
cr
r—
1/1
O
OJ
00
CO
Ol
E
O
4-
-o
>
'£
•S
g_
c
^J
CO
OJ
x:
^
o
o
c
o
4_1
o
CO
s_
4-
01
^*
*^
CO
cr
Ol
M
X)
c
CO
>^
i
X
CO J
C
0
CO
s-
4J
C
OJ
O
C
0
u
OJ
03 .
•i- Ol
X j—
O co
•i- u
"O oo
3 ^1
4- 3
r— 4—
O
CO
4-> C
CO OJ
O
4-> S-
OJ Ol
l/l CL
OJ O
-Q Ol
S C
E ca
e o
OJ CM
4->
OO C
>> OJ
S
Ol 4->
C CU
S-
O 4J
4-> 3
•c- a.
C 4->
§3
O
oo E
3 CU "O
O 4-> O
3 oo £
C >, 4J
•<-' "" E
C CU
O -C 4-*
O 4-> 00
OJ
CU C 4~>
w o
4- C C
0 -r- T3
c E —
CO CL
0.0 E
OO 4-» O
o
cu >,
x: i_ o;
•4-* IO f~
00 4->
* y/1
CO) 00
O O co
•r- OJ
4-> C 03
CO Ol
34- oo
r— -I- 3
CO
> qj oo
cu E ••-
OJ r— i.
OOO
C > 4->
CO •!-
E>> c
-Q 0
0 S
4- 4-J
^- C oo
CJ Ol 3
Q- O O
i- 3
CU CU ~
JT CL r~
Ltl C
Ol^- O
c • o
-.- o
S- OJ
34- x:
O O I—
Q 0
1-3
-------
-o
0)
-o
3
O
c
o
o
1 —
J'
-^-
UJ
SI
Q.
1 — 1
13
o-
UJ
C3
•z.
\ — c
ct.
o
I—
1 — 1
-f
0
2:
oo
ID
O
IZ5
~Z^
1 — 1
1 —
^L
O
o
u_
o
"Z.
o
f— 1
CJ>
1 1 1
1
_J
UJ
oo
CD
^
i i
i —
o
UJ
1 1
\ 1 .
u_
-
C£
o
cu -c
111
V A
o
c cu
IS
§2
it- d
oo >
4->
c
a
4->
3
r—
*0
rt
i.
f- CU
CT1 cu
T3
cu
u.
"O >>
ai *->
•t-> -^
U F—
O) -^
•4- 0
<4- 10
< u.
>>
at u
u o
U CT
3 CU
0 -U
co id
o
IO
•a
o
1 J=.
+J
£
CM
1 0
00
_a
a< 3
O i—
o o
r— >
1 >>
-d
o
CO »*
o
CM
O
>,
+? 0
•r- CM
u o
(O OO
a.
o
,
CM
(O (O
tn ID
i^^ r^-
0 C3
IO IO
cu
c
.c
(.J
n3
E
o>
c
s- cu
CU 4->
4-) in
C fl
•r- O
OO C£
i.
cu
4->
1
in
u
c
hj
>,
C
IO
1
s_
Q.
1
1
»«
0
0
1
o
oo
>!
4->
o
It)
CL
o
^ — s
^_^
iO
LD
co
0
i£>
>>
t?
o o>
4-> C
(O •«-
S- i.
cu cu cu
O J3 4J
ITS t. C
c: cu •*-
s- > in
*+- t- cu cu
0 C
4-> l/l IO ••—
1/1 1/1 C ^
1
t/1
-a
>
c
g
s.
a.
l£>
-a
o
.c
4->
^
CM
O
OO
E
3
cf
•M
1
C
O
4->
•a
i_
4->
c
cu
o
c
o
u
cu
•a •
-i— a;
X i —
O id
•r- O
T3 1/1
J_ (
3 i—
It- 3
r— *^-
3
1/1 14-
o
10
4-1
4-> C
id cu
o
4-> (-
ai cu
, — i/> _i.
o
>
>>
.a
a«
o
CM
O
O
CM
O
OO
, — *
CM
ID
in
CO
o
iO
cu
c
-C
u
01
c
L
cu
4->
C
oo
cy o
J3 ,TJ
10 C
E 10
E 0
CU CM
4-1
in c
>, 0)
1/1 CU
J
01 4J
C CU
•r- J3
t.
O +->
4-> 3
•rw Q.
C 4->
§3
O
in E
3 5 -0
O •!-> O
3 in .c
C >> 4->
•i- i^ CU
4-> E
c cu
O -C 4-J
U 4J l/l
CU
CU C 4->
-C ••-
4-i IO CU
4-> U
i)- C C
o ••- 10
(O -i-
C E i—
10 a.
a. o E
l/l 4-* O
U
"" ?^ „,
^ w CO
4-3 fd -C
in 4->
» tn
c cu 1/1
o u CU in
CU S T-
CU i-~ ^~
o o o
C > •!->
E >> C
o £=
it- 4J
I- C U1
CU Ol 3
ex u o
S_ 3
cu cu c
5^'43
in c
oi>— o
c • o
••-o
s-
-------
Part 60 specifies a minimum number of data points rather than an
integrated average to compute average emissions rates. Since most of the
processes affected have continuous production, the readings or samples can be
relatively infrequent.
60.13(a) — all continuous monitoring systems shall be in continuous
operation and shall meet minimum frequency of operation requirements
as follows:
(1) — for measuring opacity of emissions shall complete a minimum of
one cycle of operation (sampling, analyzing, and data recording)
for each successive 10-second period.
(2) — for measuring oxides of nitrogen, sulfur dioxide, carbon dioxide,
or oxygen shall complete a minimum of one cycle of operation
(sampling, analyzing and data recording) for each successive 15-
minute period.
For opacity, readings are averaged over a 6-minute period. For other
pollutants, readings are averaged over a 1-hour period or combination of
1-hour periods.
Section 60.2 definitions:
(r) one-hour period means a 60-minute period commencing on the hour.
(x) six-minute period means any one of 10 equal parts of a 1-hour period.
Other time periods are specified for particular sources. For example,
Subpart D: Standards of Performance for Nitric Acid Plants Section 60.73:
Emission Monitoring:
(e) For the purpose of reports required, 60.7(c), periods of excess
emissions that shall be reported are defined as any 3-hour period
during which the average nitrogen oxides emission (arithmetic
average of any three continuous 1-hour periods) as measured by
a continuous monitoring system exceeds the standard under 60.72(a).
The required frequency of reporting, units of measurement, and other
process parameters to be reported may also influence the selection of a data
recorder and data reduction method.
1-5
-------
2. Major Equipment Types and Operating Principles
The major types of continuous monitoring equipment and their basic
operating principles are described briefly below, A thorough review of all
available instrumentation is not within the scope of this manual. However,
the Lawrence Berkeley Laboratory of the University of California (Berkeley)
published a two-volume study* of available instrumentation which is updated
periodically. This document should be referred to for more specific
information.
a. Opacity Monitor Principles
Systems for monitoring opacity are based on the principle of trans-
missometry, or directly measuring the attenuation of visible radiation
(light) caused by particulate matter in a stack effluent. Light with specific
spectral characteristics (visible wavelengths) is projected from a lamp
across the stack into a light sensor. The particulate matter absorbs and
scatters the light beam, decreasing the intensity seen by the sensor. The
amount of attenuation of the light beam is defined as the opacity of the
emission. Transparent stack emissions which tlo not attenuate visible light
will have a transmittance of 100 or an opacity of 0. Opaque stack emissions
which obscure all of the visible light from the source will have a transmittance
of 0 and an opacity of 100 percent. Since the transmittance is a function
of the path lengths, transmittance must be adjusted to the stack exit diameter.
Typical opacity monitors on the market are usually either "single" or
double-ended" systems. In the single-ended system (dual path) (Figure 1-4),
the light source, sensor, and all electronics are located in a single unit
on the side of the stack, the light beam is sent, then reflected from the
opposite side of the stack by a mirror, making two passes through the stack.
Since the complete system is on one side of the stack, it is relatively simple
to make differential as opposed to absolute measurements of light intensity.
Voltage variations, detector variations, and electronic drift are all
cancelled out in the differential measurement.
*Instrurnentation for Environmental Monitoring - AIR, May 1, 1972, Environmental
Instrumentation Group, LBL, University of California, Berkeley, CA 94720.
1-6
-------
1 Tnhf" cnnrvo
and photo-
electric
detector
Air purge
system
^
/
/
Stack
Light beam
4
t
^mnkp nr Hiict
— ^w .
\
\
ji
J Reflector
Air purge
system
(0
D
<
Figure 1-4. Single-ended opacity monitoring system.
1-7
-------
In a double-ended (.single path) system (Figure 1-5), the light source
and detector are on opposite side of the stack and the light travels once
across the stack. These types of transmissometers are inexpensive and have
been popular throughout the last 35 years. However, because of their align-
ment sensitivity, difficulty in zero and span calibration, and sensitivity
to voltage and light variations, many of these monitors cannot be used for
EPA monitoring purposes. Since they are inexpensive, they will probably be
useful as simple warning devices, indicators of poorly performing control
devices, etc., where quantitative accuracy is not most important. However,
some manufacturers of single-end monitors meet NSPS by using a method that
creates a zero opacity condition for the full light beam in order to obtain
absolute online calibration of all the optics on both sides of the duct.
This method also employs a slotted mounting pipe to maintain alignments.
b. Gaseous and Diluent Monitor Principles
Many varieties of equipment currently are available for continuously
monitoring gaseous pollutants. However, after sorting out all the publicity,
there are only four or five major types of equipment. The operating prin-
ciples used in each major type are discussed below.
(I) Ultraviolet Absorption
The nondispersive ultraviolet analyzer shown in Figure 1-6, is a split-
beam photometer. It monitors the concentration of a pollutant by measuring
the difference in the amount of radiation absorbed at two different wave-
lengths (in the ultraviolet and visible). Radiation from the light source,
usually a gas discharge lamp, is partially absorbed as it passes through a
sample of gas. As radiation leaves the sample, it is divided in two beams by
a semi transparent mirror. Each beam then passes through an optical filter
which removes all wavelengths except the one to be measured. The filtered
beam then strikes a phototube.
The beam splitter is set up to reflect and transmit so that the
intensities of the radiation striking each phototube will be nearly equal
during normal operation of the analyzer.
1-8
-------
Light
source
Ai r purge
system
<
Stack
Light
til
beam
ti
\
Photo elec-
tric detec.
Air purge
system
Figure 1-5. Double-ended opacity monitoring system.
1-9
-------
S-
cu
Isl
-------
The beam reaching one phototube is in a wavelength which is absorbed
strongly by the pollutant being measured; thus, this beam actually measures
the concentration. The beam directed to the second phototube is in a wave-
length which is absorbed weakly or not at all by the pollutant and is used
as a reference. If the concentration of the absorbing pollutant changes,
the intensity of radiation reaching the first phototube varies in accordance
with the Lambert-Beer law. The radiation to the second phototube varies very
little, or not at all.
In each phototube, the current flow is proportional to the intensity
of radiation striking the phototube. This current is fed to an amplifier that
outputs a dc voltage that varies logarithmically with the current and, hence,
with the intensity of the radiation. Voltage output from the reference beam
amplifier is subtracted from that of the measuring beam amplifier. This
difference in voltage output varies linearly with the concentration of the
pollutant being measured.
Nondispersive UV analyzers are somewhat more sensitive than non-
dispersive infrared (NDIR) analyzers and are not subject to the interferences
(e.g., HnO. ^2) that affect NDIR analyzers. Nondispersive UV analyzers for
measuring SCL may have interference from N0o'> however, this interference can
be filtered out electronically.
(II) Infrared Absorption
Two main types of infrared sensors commonly used are nondispersive and
dispersive. The term "nondispersive" refers to the lack of a light-dispersing
element, such as a prism or grating that selects a particular analytical wave-
length of light. The traditional Luft-type nondispersive infrared (NDIR)
analyzer has been used to make process and environmental measurements for
over 20 years. A schematic of this type of analyzer is shown in Figure 1-7.
In the NDIR analyzer, two identical infrared-radiation emitters
serve as matched sources of broad-band infrared energy. Radiation from these
sources is modulated by a motor-driven chopper disk and passed through filers
and measuring cells into an energy receiver. The reference cell is filled
with a gas, such as nitrogen, that does not absorb infrared energy. Another
1-11
-------
1 1 1 1
e
O
en
s«,
« Kg
a
2
UjUUH-^
eZgcu
>.
-------
cell contains the gas being analyzed. The amount of infrared energy passing
through the analysis cell that will be absorbed by the gas depends on the
concentration and wavelength band of the gas.
The energy receiver consists of two chambers separated by a membrane.
Both chambers are filled with mixtures of the gas to be measured. As the
gas absorbs infrared energy, it heats up and pressure increases. If the
two chambers are exposed to equal amounts of energy, the membrane will
remain in a neutral position and there will be no output from the instrument.
However, since the gas in the analysis cell absorbs more energy, pressure on
that side of the chamber increases and the membrane deflects. This deflec-
tion is detected as a change in electrical capacitance between the membrane
and a fixed electrode and, therefore, as a change in ac voltage at a resistor.
This ac voltage change is displayed as dc current using a measuring amplifier
on an indicating instrument.
If the gas to be tested contains components with absorption bands that
slightly overlap those of the pollutant being measured, identical filter cells
are filled with the interfering components to screen them out. This "positive
filtering" makes it easier to measure the pollutant.
The Luft-type NDIR analyzer generally uses an extractive sampling
system to transport the sample of stack gas to the analyzer. Particulate
matter and water vapor that interfere with the species measurement are
removed. Generally, Luft-NDIR analyzers are used to monitor SCL, CO and CO^.
A schematic of another nondispersive analyzer is shown in Figure l-8(a).
In contrast to the Luft-type analyzer, which looks at a broad spectral region
and which must be sensitized for each particular gas using a detector cell,
this NDIR spectrometer can be set at any wavelength within its range. In
addition, unlike the Luft-type analyzer, the adjustable analyzer is not limited
to a single, preselected gas. However, when many absorbing gases are present,
it may be difficult to locate a wavelength in a spectral region where other
gases do not absorb.
1-13
-------
Patntengtti can be waned with an outside adiustmg knob
Irom *i meter to more than 20 meters
SAMPLE CELL
OUT IN
ANALYZER
PATHLENdTH
AOJUSTMgNT
ELE' : TRONIC
.
HIGH CONCENTRATION CELL LOCATION I
Figure l-8(a). A variable wavelength NDIR analyzer.
1-14
-------
The analyzer shown in Figure l-8(a) uses a narrow-band pass filter
which scans the spectral region between 2,5 and 14 microns. It could be
used, for example, to measure S0?, which has an absorbance maximum at 7.4
microns. Since it has a variable pathlength from 3/4 to 20 meters, the
analyzer has a large sensitivity range. This and similar analyzers can
measure many gases simultaneously. However, other gases, water vapor, or
particulates may cause absorption problems in the spectral region of the
pollutant.
Another type of nondispersive infrared analyzer is the gas cell
correlation spectrometer shown in Figure l-8(b). EPA researchers are
investigating this type of analyzer as an in situ analyzer for various
pollutants. The analyzer uses a broad band IR source and detector and one
or two cells filled with varying concentrations of the gas to be measured
to provide the wavelength selection. This technique has shown potential for
rejecting interference. Thus, it can be used to improve low level sensitiv-
ities, or to simplify or eliminate requirements for sampling systems. In
addition, by switching selection cells, several pollutants can be monitored
by the same unit.
Figure l-8(c) shows a single-beam, dual wavelength nondispersive
infrared analyzer. In this analyzer, a single IR radiation beam is passed
through the sample gas and filtered into two wavelengths by a chopper. The
wavelengths are chosen so that one wavelength has no absorption, while the
other has maximum absorption. The ratio of their absorptions is sensed by
a conventional thermoelectric IR detector.
The dispersive infrared analyzer (Figure 1-8(d)) uses a broad band IR
radiation source and a broad band detector similar to the gas cell correlation
and filter-type analyzers. However, the wavelength is selected by using a
prism or diffraction grating which physically disperses the light at different
angles. Light at 7.4 microns (a typical S02 wavelength) may be deflected
at 22.2°, whereas light at 4.3 microns (C02) may be deflectd only 9°. By
placing sensitive detectors at the proper positions, several components can
be measured concurrently, limited only by the space available. However,
this approach has several drawbacks. First, the analyzer's ability to select
a specific wavelength depends on the angle and length of the dispersing
1-15
-------
Correlation
ceil
o
Source
ft!
Attenuator
Band pass
_S-mple_cell '"T Detector
Amplifier
and
meter
Figure l-8(b). Another nondispersive analyzer.
1-16
-------
Sample
analysis
electronics
Chopper
filter
Thermalelectric
detector
Recorder
Sample
out
1
\1
1
/ 1
/
T
\
Sample
in
Light
source
\
Meter
Figure l-8(c). Single-beam, dual wavelength NDIR.
CHOPPER
MOTOR ;
PILOT LAMP—^i (|
-TIMING BELT DRIVE
TT
-T-,3
\-t
•-ILUI Law -U [| (H . I i I
PHOTODUODOOE^ TfS, I* i V '\ ' ^
* '* i '' / ' . i '/
CHOPPER' * 7 'v \ .'/
PILOT
r
SAMPLE CELL
cTH*o°PPE>'«EOUE*¥ -U^C' VGRAT,NG
Jj9 SOURCE
_ «—
ENTRANCE SLITS
OPTICAL FILTER
SEALED
SECONDARY REDUCING INSTRUMENT
CELL VALVE s^ /PURGE INLET
*^r
;,F2 WINDOWS EACH END
'" SEALED
S£CONOAf?r CELL PURGE EXIT, TO
EXTERNAL SAMPLE INDICATOR 8 VENT
Figure l-8(d). A dispersive IR analyzer.
1-17
-------
path (L). To be very selective, it is necessary to use very long paths (a
meter or two) or a very narrow slit or mask on the detector for good
resolution. But long pathlength makes optical alignment difficult, and the
narrow slit decreases the light to the detector, thus decreasing the
signal level.
Until recently, the dispersive IR approach was used primarily for
large, expensive laboratory equipment. While dispersive stack analyzers
are commercially available, other new types of analyzers, such as the gas cell
correlation, the new chemiluminescent, or the fluorescent analyzers, are
expected to play major roles in the future.
(Ill) Chemiluminescence
Although NDIR analyzers have been used to measure nitric oxide (NO)
for 20 years, chemiluminescence has been used in approximately 90 percent
of the installations since it was introduced commercially. This is true
even though chemiluminescence is two or three times more expensive.
Chemiluminescence is a chemical and optical monitoring technique.
In this technique, a gas molecule reacts with a reagent to form an excited
molecule that spontaneously decays, producing photo-emissions. Sensitivity,
rapid response time, and instrument stability make the Chemiluminescence
method suitable for continuous monitoring, although care must be taken to
maintain the monitored stream at constant flowrates. A typical NO and NO
/\
analyzer is shown in Figure l-9(a). In this analyzer, NO molecules combine
with 0- to form an excited molecule (NOp). A photomultiplier detects photo-
emissions decay and sends the signal to a sample-and-hold circuit. To detect
N02, an N02 to NO converter is connected into the gas stream and the NO
analysis is run again. After the second pass, a circuit subtracts the first
reading from the second to obtain the N02 reading. Few interferences have
been observed, but high concentrations of C02 or water vapor may partially
quench the Chemiluminescence, In addition to monitoring NO and NOV,
A
Chemiluminescence devices are also used to detect 0.,.
1-18
-------
in
s.
3
1-19
-------
Other gas molecules, such as carbon monoxide and sulfur dioxide, have also
been measured by fluorescence (Figure l-9(b)). In the fluorescent analyzer,
as in the chemiluminescent analyzer, the gas molecules are excited to a
higher energy level, and their photo-emissions during decays are measured.
However, where the chemiluminescent analyzer uses a chemical reaction to
produce the excited molecule, the fluorescent analyzer irradiates the gas
molecules with a pulse of light. The pulsed, ultraviolet fluorescence
SO,, analyzers recently introduced into the market may be as successful as
the chemiluminescent NO monitors.
A
As with chemiluminescence, fluorescence analysis is highly specific
and has few interferences. However, since high concentrations of COo and
water vapor may partly quench the fluorescence, most units use some type of
water vapor removal.
(IV) Electrochemical
Analyzers using electrochemical transducers measure the current
induced by the electrochemical oxidation of the pollutant at a sensing
electrode. Sensors are available for measuring Op, SO,,, CO, H^S, NO, and NOp.
Figure 1-10 shows a simplified schematic of an electrochemical transducer.
In this analyzer, the pollutant diffuses through the semipermeable membrane
into the transducer at a rate proportional to the concentration. At the
sensing electrode, the pollutant undergoes an electrochemical oxidation or
reduction which causes a current directly proportional to the partial pressure
of the gas being monitored. Since electrons are produced at the sensing
electrode, this electrode is at a lower potential than the counterelectrode.
Thus, an electron current can flow from the sensing electrode through the
amplifier to the counter electrode, and the current will be proportional to
the concentration of the pollutant.
Selectivity of the transducer is determined by the semipermeable
membrane, the electrolyte, the electrode materials, and the retarding
potential. This potential is adjusted to retard oxidation of species that
are less readily oxidized than the pollutant of interest.
Selecting a specific chemical compound for measurement is not always
possible. For example, the SOp sensor is sensitive only to S02» but the
1-20
-------
I/)
o
S-
o
O)
(U
N
(O
O)
o
c
ts > cr>
I/I IB -r-
3 -U
£X i—
1-21
-------
GAS CONTAINING POLLUTANT
• XT1* "
SELECTIVE MEMBRANE
_ ^^kX_Zi. i_\t_V •i-i-Yk^-i
THINJFILM ELECTROLYTE
SENSING ELECTRODE ,
BULK ELECTROLYTE
COUNTER ELECTRODE
ELECTROCHEMICAL
TRANSDUCER
RETARDING
POTENTIAL
AMPLIFIER
Figure 1-10. Electrochemical analyzer.
1-22
-------
NO (NO-NCL) sensor is also sensitive to SCL- Thus, SCL in the gas stream
must be removed with a scrubber solution, or measured by another sensor and
cancelled out electrically. Electrochemical sensors have been used primarily
in portable instruments for quick spot checks of pollutant concentrations,
although several vendors offer continuous source monitoring systems based
on these sensors.
(V) Paramagnetic
Very few gases are paramagnetic, oxygen being one. Analyzers for
oxygen, using this principle, have no particular advantages or disadvantges
over other types, such as the electrochemical analyzer. They do, like most
other types of analyzers, require a clean (particulate-free) and dry sample.
In one model paramagnetic analyzer, a test body suspended on a platinum-
iridium wire in a nonuniform magnetic field experiences a torque proportional
to the magnetic susceptibility of the gas surrounding the test body. The
torque is counteracted by an opposing torque produced by passing a current
through a coil around the test body. By means of a light source reflected
from a mirror mounted on the test body, and a differential photocell and
amplifier, the current is continuously and automatically maintained at the
value needed to balance the initial torque. This current is thus proportion
to the magnetic susceptibility of the sample gas. The output is not affected
by the density of thermal conductivity of the sample gas or by any catalytic
effects. Errors caused by variations in other sample gas components are
small and predictable.
3. Monitoring System Selection Criteria
In choosing a monitoring system for a specific source, both economic
and technical factors must be considered. Economic considerations include
costs for installation, calibration, maintenance, data processing, and the
overall system. Technical considerations include the ability to measure and
record accurately for a particular source, the levels and types of inter-
ferences present, sample and ambient temperature and pressure extremes,
accessibility of sample locations, and availability of maintenance personnel.
Sources of information on these monitoring systems are plentiful.
1-23
-------
a. Prior Experience of Users and Agencies
Several groups within EPA have evaluated prototype and commercialized
monitoring systems on a number of sources, These studies have been widely
distributed through meetings, papers, and reports. Reference 8 summarizes
some of the more useful EPA and industrial reports that are readily available.
These reports are referenced at the end of this section.
The earlier reports (1971-1973) show that the monitoring equipment
then available generally performed adequately when equipped with a good
sample conditioning system. Unfortunately, at that time good sampling systems
were sold by only one or two vendors. Subsequently, both vendors and the
EPA began to develop better sample conditioning systems, as well as systems
that do not require sample conditioning (in situ or new extractive).
In opacity monitors the window purge system constantly blows clean
air on optical windows to minimize particulate fouling. EPA has found that
the overall design of the window purge system is one of the major differences
between opacity monitors that operate correctly and those that do not. Most
commercial vendors with field experience in installing monitoring systems
now have an adequate sampling system that will operate correctly if maintained
properly. If a vendor is willing to give several references (companies using
his equipment on a particular source), chances are good that his system will
operate correctly, if properly maintained, for that type of source. However,
potential customers will do well to follow up on these references.
In all monitoring systems, the key words are "if properly maintained."
Almost every problem found by EPA over the last 2 years was related to lack
of daily checks and weekly maintenance. Typically, a monitor would encounter
a minor problem such as a plugged filter, a stuck valve, or a closed shutter
would be out of operation until the next scheduled evaluation 2 or 3 weeks
later, even though 5 minutes of maintenance would have corrected the problem.
Therefore, an effective monitoring program requires daily checking and main-
tenance.
1-24
-------
b. Specific Criteria for Gaseous and Diluent Monitoring Systems
In selecting a monitoring system for a, specific plant, several general
questions should be answered:
1. What are the concentrations of the pollutants of interest?
If a facility is normally controlled at 20 to 110 ppm but can upset
to 1000 ppm, a monitoring system with a wide range should be used.
Furthermore, as previously delineated, the states require some
specific performance characteristics. Optical absorption tech-
niques (NDIR, UV, IR), chemiluminescent (for NO ) and fluorescent
A
(for SOp) analyzers all have wide ranges. Although some optical
absorption techniques have difficulty maintaining linearity over
a 20:1 range change, they may be more appropriate at pollutant
concentrations in the percentage ranges than chemiluminescent or
fluorescent methods.
2. Does the stack have any stratification?
For gases with severe stratification or other flow problems, an
extractive monitor with multipoint probes or a long-path
averaging in situ system mounted in the plane of the bend is most
effective.
3. Does the source produce high concentrations of water vapor (3 to
8 percent)?
Most infrared techniques and the pulsed fluorescent technique are
adversely affected by high water concentrations in the gas stream.
However, the ultraviolet absorption technique is relatively
unaffected by water vapor, and, therefore, would be a good choice.
In addition, when sampling streams with high concentrations of
water, care must be taken to prevent water from condensing and
causing corrosion in the system. Plastic (PTFE, PVC), stainless
steel, or even glass may have to be used in severe cases. To
alleviate the problem, early use of knockout traps, dilution,
etc. is recommended.
1-25
-------
4. Does the gas stream contain a high concentration of participate?
If possible, monitoring locations with high participate loadings
should be avoided. In situ monitors correct for opacity by
measuring at a neutral wavelength and adjusting for the net
absorption by taking a ratio of the intensity of the two wave-
lengths. The success of this correction varies among
manufacturers.
Extractive monitors generally remove particulate with a filter at
the probe inlet. High particulate loadings usually require the
filter to be backpurged more frequently. This in-stack filter is
required for final cleanup of the sampled gas.
5. Does the gas stream contain entrained water or other liquid
droplets?
The presence of entrained water droplets, acid mist droplets,
etc. can blind filters used for removing particulate. These
drops also corrode sample lines and attack valves in extractive
systems. To remove excessive quantities of liquid in extractive
systems, water traps, dilution, and heating of the stream can be
used.
Theoretically, liquid droplets (acting as particulate to light)
should not effect measurements from in situ systems, for reasons
discussed in the Subsection 2(d). Most in situ analyzers have
optics external to the stack, which makes them particularly
beneficial in certain plants, such as pulp and paper mills.
However, in some in situ systems in which the optics are exposed
to the gas stream, lenses, windows, and mirrors can become coated
with liquid.
6. Will multiple sources or emissions points be monitored?
As mentioned earlier, Section (g) of Part 60.13 permits a single
system to be used for monitoring multiple emission points if all
sources must meet the same standard and all sources are combined
into a common duct. However, even if multiple streams are not
combined into a common stream, a single measurement system can still
be used. Most extractive measurement systems can be equipped with
1-26
-------
multiple sampling probes. Measurements can be made with these
probes by automatically switching between them on a time-sharing
basis. The approach is acceptable as long as the minimum require-
ment (i.e., 5 min. for SCL) one measurement cycle is completed
for each measurement point in each sampling period. Effective
sample cleanup and conditioning is extremely important where
multiple emissions are measured, because the long sample lines
increase chances for deposition, condensation, plugging, and
corrosion. If an in situ monitor is used, each monitor can measure
only the stack on which it is installed. However, multiple
systems can easily share a data recording system.
Will the temperature and pressure at the monitoring location
present problems?
Temperature and pressure changes can cause problems in systems
that measure pollutant concentration by optical absorption in a
fixed cell length. Becuase the gas may expand or contract in
the optical pathlength, errors of 0.3 percent per °C change in
temperature and 1.3 percent per 10 mm Hg change in barometric
pressure can occur. The analyzer must be given a sample at a
known temperature and pressure or be equipped with compensating
circuits to measure these variables and correct for them. Most
extractive sampling systems send samples with steady temperature and
pressure to the analyzer. However, the Luft NDIR and electro-
chemical extractive analyzers are more affected by ambient
temperature variations than other types of analyzers.
Will vibration and line voltage affect the monitor's performance?
Vibration and line voltage fluctuations will certainly affect the
performance of the monitor and adequate precautions must be taken
to eliminate these variables. Voltage fluctuations will be a
particular problem for magnetic type recorders installed near
electrostatic precipitators.
1-27
-------
9. Do other gaseous species interfere with the pollutant being
measured?
Pollutants other than that being measured can interfere with
certain types of monitors (e.g., NO interferes with SO electro-
X X
chemical analyzers, 02 interferes with pulsed fluorescent instru-
ments). The presence of interfering species should be taken into
account when selecting an analyzer.
10. What are the monitoring system output requirements?
The regulations do not place any particular requirements on the
"data recorder" portion of any measurement system except as noted
previously and in paragraph 4.3 of Performance Specification 1
(transmissometers). This paragraph requires an expanded opacity
display on a standard 0- to 100-percent scale. Most vendors
make available a wide variety of data output, display, and
recording devices. Figure 1-11 shows a recent advertisement by
an opacity monitor vendor. The analog and digital display modules
are equipped with alarm indicators and lights that can be set at
any level by the plant operators. To insure that positive action
is taken when upset or excess emissions occur, sources should
consider locating an additional set of alarms and realtime
displays in the office of a responsible plant official. Too
frequently, because his prime responsibility is to the process
equipment rather than to pollution control, the operator in a
control room will merely reset or turn off the alarm. Therefore,
the operator should set the alarm for a high level of pollutants,
but not the highest level permitted. Then, if the alarm sounds,
the operator has time to prevent excess emissions from occuring.
c. Specific Criteria for Opacity Monitoring Systems
i
As with gaseous and diluent monitoring systems, opacity monitoring
systems have definite questions which need to be answered before a source
can select a specific instrument.
1-28
-------
s_
cu
o
rO
4->
oo
u c -*:
••- c
3 O
O- 0
I/I
Ol
-Q
•r- (T3
T3 0
fl I
OJ I
f
OJ
01 c
•r- (O
o
OJ
(U
O)
S- E
3 01
CL-M
to
S- >>
en
o ~
(U
(J
s_ +J
0) -r-
U
03
03
•!->
00
S_
O
•»->
.7. 3 «>
3 S-
O S-
o
o
i- i/1
cu o c
•M 4-> O
O «O •»""
O 1/1
•r- S_
1-
O)
O
O
0)
£5-*
0)0.0
M CO
O)
o
to
cu
•r™
L.
O
V)
CO
a»
u
u
§
Q.
O
in
-u
3
Q.
4->
3
O
(U
(/I
en
ra
4->
~°
i.
ai
-o
S-
o
o
CD
O
>>
'-5
ra
•1—
en
Q
a.
1-29
-------
Does the stack gas have particulate stratification?
For gas streams with excessive particulate stratification (as
indicated by a sampling traverse of an EPA Method 5 train), it is
imperative that the tranmissometer be located such that the light
beam passes through representative effluent. Sharp bends and
horizontal rectangular ducts are typical locations that encourage
stratification and should be avoided.
Does the stack gas have a high particulate concentration?
If a source has a high particulate loading, they should be certain
to install a transmissometer with an adequate optical window purge
system - on each end of the optical path if a reflector system is
used.
Does the gas stream contain condensed water, acid mist or other
matter in droplet form?
Because condensed matter attenuates visible light in the same
manner as particulate, transmissometers cannot be used in gas
streams of this nature. An alternative mounting location must
be found, where the gas stream temperature is above the dewpoint,
or an alternative parameter will have to be agreed upon by the
source and Administrator to monitor and characterize control device
efficiency.
Will multiple sources or emission points be monitored?
If a source has one or more affected facilities subject to the
same emission standard, they may install a transmissometer on each
effluent or on the combined effluent. If the facilities are subject
to different standards, they must have separate opacity monitors.
If an affected facility releases effluent through two or more
locations, then each one must be monitored separately unless
installation of fewer systems is approved by the Administrator.
Will the temperature at the monitoring location present problems?
Fluctuations in stack temperature will have an adverse effect
on transmissometer alignment particularly in steel stacks and
ducts. Opacity monitors have a narrow field of view and thermal
1-30
-------
expansion of steel stacks and ducts caused by load cycling and
other non-steady-state conditions will put alignment out of
tolerance. Monitors should be adaptable to mounting on a rigid
portion of a stack or duct or on a totally independent structure
that is not prone to thermal expansion.
Will vibration and line voltage affect the monitor's performance?
Again, due to the critical optical alignment required by opacity
monitors, precautions must be taken to eliminate vibration.
Transmissometers will tolerate some voltage fluctuations
(typically 115 VDC _+ 10 percent) but a source would do well to
eliminate the possibility of severe fluctuations.
What are the monitoring system output requirements?
The regulations do not place any particular requirements on the
"data recorder" portion of any measurement system except as noted
previously and in paragraph 4.3 of Performance Specification 1
(transmissometers). This paragraph requires an expanded opacity
display on a standard 0- to 100-percent scale. Most vendors make
available a wide variety of data output, display, and recording
devices. Figure 1-12 shows a recent advertisement by an opacity
monitor vendor. The analog and digital display modules are equipped
with alarm indicators and lights that can be set at any level by
the plant operators. To insure that positive action is taken
when upset or excess emissions occur, sources should consider
locating an additional set of alarms and realtime displays in the
office of a responsible plant official. Too frequently, because
his prime responsibility is to the process equipment rather than
to pollution control, the operator in a control room will merely
reset or turn off the alarm. Therefore, the operator should set
the alarm for a high level of pollutants, but not the highest level
permitted. Then, if the alarm sounds, the operator has time to
prevent excess emissions from occuring.
1-31
-------
d. Maintenance Availability and Requirements
Service and availability of parts for monitoring systems vary widely.
Several vendors of monitoring systems maintain nationwide networks of equip-
ment parts and service centers; others provide regional sales offices with
a limited stock of parts. Some vendors sell a periodic maintenance program,
whereas others will provide the entire monitoring function, including
maintenance, data reporting, etc., for a fixed monthly fee.
With one or two exceptions, the vendors would prefer the source owner
to assume responsibility for maintenance. In general, a source with an
instrumentation shop that can repair process instrumentation such as strip
chart recorders, flowmeters, transmitters, and related process instrumentation
can maintain a continuous monitoring system. For sources that do not have
this capability, the purchase of a complete service package could be highly
desirable.
The sources's commitment and attitude will also play an important
role in successful monitoring. As mentioned several times before, keeping
a system online requires daily functional checks (which should not take more
than 15 to 20 minutes). There also must be a commitment to service the
monitoring equipment routinely. Unfortunately, such attention is generally
not given to nonproductive (frequently perceived as useless) equipment.
Furthermore, the equipment must be reasonably accessible to make this
commitment more firm. Where possible, emission monitoring equipment should
be included in the process control instrumentation that is monitored and
logged hourly. In addition, the group that services process instrumentation
should service the monitoring equipment. Operators or service personnel
should be encouraged to use and benefit from continuous monitors and develop
professional responsibility for operating the equipment effectively.
e. Gas Conditioning Requirements
Most extractive continuous monitoring systems require some sample gas
conditioning. Depending on the operating principle of the analyzer, partic-
ulates, water, and other condensibles must be removed and a particular sample
temperature must be maintained. For example, in nondispersive infrared
analyzers, particulate matter and water vapor interfere with measuring the
1-32
-------
species and must be removed. In addition, condensed water vapor damages
internal components and also must be removed.
In some types of industrial processes, if condensed tars from the gas
stream are not removed initially, they may form in the analyzer. Also,
condensed sulfuric acid mist can cause severe corrosion. For example, in
monitoring kraft pulp mills, elemental sulfur can form when sulfur dioxide,
hydrogen sulfide, and liquid water react. These problems can usually be
solved by a complete sample conditioning system, such as the one shown in
Figure 1-12. This system includes particulate removal filters, permeation
driers for water removeal, plumbing for purge, span, and calibration gases,
and temperature controllers.
f. Selection Criteria of Data Handling Equipment
Again, the regulations have no particular requirements on data handling
equipment. In fact, a continuous monitoring system operates perfectly well
without one. However, due to the data required to be reported, data handling
equipment can be a great aid to affected sources.
Paragraph (c) of Part 60.7 requires that the owner or operator shall
report the following information, as well as other information, quarterly:
magnitude of excess emission, conversion factor used, the start and stop
time of emission and as well as key process and control system data. This
data can be collected, utilized and stored automatically. A broad spectrum
of equipment and approaches, including different techniques of averaging
data, can be used. The simplest approach is to collect the raw data on time-
calibrated strip chart recorders. At the end of each month an engineer
calculates the emissions for hi£ source, and then tabulates the data into
the required report.
A second approach would be to equip the opacity monitor or gaseous
monitors with a time-integrating device that calculates the specified average
emission value. The integrated value may then be recorded on a strip chart
recorder. Paragraph (h) of 60.13 specifies the procedure to be used if an
integrating device is included.
A third approach would use one of the automatic data loggers available
from several suppliers. These recorders log data, along with the date and
1-33
-------
Stack
Calibration
gases
Air in
Parallel
permeation
drier
Air out
Probe with Balston
fiberqlass cylindrical
tube filters
Heated
^f sample
line
Vent
Purge valve
Select valve
«_ Pump
0—
. \
Electrical
signal
Electrical
signal
Electrical
| ' signal
i
Figure 1-12. Typical analyzer conditioning system.
1-34
-------
time, in digital form, on a small 12-column printer. More advanced units
incorporate microcomputers which can perform the entire sequence of averaging,
timing, calibrating zero and span functions, converting emission factors, and
recording data.
These data handling approaches may be adpated for any system or
operator functions and procedures. For example, instead of printing out
each 6-minute and hourly reading as it occurs, the system may store the data.
It will print or display only when readings approach preset limits or excess
emission levels, or during calibrations or on demand. Thus, the system can
provide a warning system.
Prime considerations in choosing between these approaches are the mean
time between failures of teletypes, printers, and recorders and the attention
an operator must give to the system, even though it may be operating properly.
Automated data loggers can store and process all readings on a central data
processing computer or on the process control data acquisition computer.
Thus, the data can be further processed, analyzed, and output less frequently.
In addition, automated systems can be used effectively when report preparation
and data processing from several plants are handled by one center.
Automated, integrating hardware eliminates potential human errors in
reading and averaging and decreases the amount of data reduction required
later. The reference methods are all based on integrated averages of samples
collected over time. In contrast to paper chart data, which is usually
visually averaged, most automated data collection systems record instantaneous
measurements. Integrated averaging can be added to the strip chart or auto-
mated data collection hardware, or included as part of the data reduction.
However, in many steady rate processes, the instantaneous reading is a good
representation of the integrated average.
The accuracy of the data recorder is inherent in its basic design.
The specification for each analyzer will indicate its accuracy. Typically,
an accuracy specification of 1/2 or 1 percent of full-scale is satisfactory
for use with emission monitors.
1-35
-------
The resolution of analog data recorders is inherent in the design of
strip chart recorders. Important factors that influence resolution include:
1. Chart width (typically 4-1/2-inch or 10-inch)
2. Chart division (typically 100 divisions full-scale)
3. Chart speed (1 to 4 inches per hour)
4. Type scale (log or linear)
5. Amount of damping
The resolution of digital or multipoint recorders, though not as
obvious, is still an important factor in selection. To reduce costs,
amplifiers and analog-to-digital (A/D) systems are often shared by several
monitors. Thus, the expense of high resolution equipment can be shared.
The specifications will indicate the resolution. Important factors include
the averaging time and the frequency of analog to digital conversion, typically
referred to as the number of counter bits or LSB (Least Significant Bit) of
the counter.
For example, the LSB of an eight-bit (28) A/D system is one part in
256. In contrast, a 10-bit A/D has an LSB of one part in 1024. Resolution
at full scale for the eight-bit A/D is 1/2 of 1/256 or about 1/2 of + 0.4
percent, whereas the 10-bit resolution is 1/2 of 1/1024 or about 1/2 of
;t 0.1 percent.
NOTE
More than four display or print characters (numerals)
and/or decimal places will not increase the resolution
of even the 10-bit A/D system (which has maximum count
of 1024).
Using the wrong combination of equipment can degrade accuracy and
resolution. For example, if an analyzer with 0- to 1-volt otuput is
connected to a 0- to 5-volt recorder, the mid-scale readings will be at the
10-percent point on the chart. Thus, the recorder would be off as much as
10 percent from the true reading, yet still be within the recorder specifica-
tions (1 percent of full-scale accuracy). Proper mating and selection of
scales is important. For resolution and accuracy, critical readings should
1-36
-------
be recorded between 50 and 75 percent of full scale. Therefore, sources may
request approval for monitoring systems that do not meet the specification.
For example, some opacity monitors are set with full scale at 40 percent in
order to obtain good mid-scale readings while the baghouse or precipitators
are working. If the control device fails, these monitors would likely not
read the actual emission value but would read 40 percent (the full-scale
limit).
Most types of data handling equipment, from multipoint recorders to
computer-controlled data acquisition systems, scan a number of analyzers
with one set of electronic amplifiers, an analog to digital converter, and
a single recorder for output. The frequency at which the scans are made
can be selected. Even though opacity readings are taken once every 10 seconds,
data systems may sample much more frequently in order to average out extreme
readings and noise. High speed data systems at many chemical plants read
100 or more sensors per second. At this rate, the sample time is very short;
thus, the noise peak part of the signal from an analyzer can be easily read.
Average times of from l/60th of a second or 1 hz or more are better suited
for opacity monitoring. A continuous, integrated averager using 6-minute
averaging periods would make the readings fully compatible with the reference
method.
When using systems that rapidly scan all the analyzers, all individual
readings need not be checked. However, one must be certain the software or
mechanical programmer is scanning frequently enough and averaging its readings
properly for the periods required (6 minutes or hourly). This is most easily
checked by observing output during periods of change from zero to span
calibration. It is also important (for purposes of correlating process
with pollutant date) to average similarly, know (specifically) time of
pollutant readings, and to permit easy collation of pollutant and process
data. In addition, the outputs must always be identified for point/source
of origin of signal, pollutant, time, day and most recent span/gain,
calibrations, etc.
Accessories available with these data loggers range from 12-column
printers to magnetic tape recorders, computer interfaces and typewriter-
style printers. Now that the monitoring regulations have been promulgated
1-37
-------
and the reporting requirements specified, we can expect that data logger
vendors will make custom units available specifically for pollutant monitoring.
These units will probably cost from $7,000 to $20,000, depending on the output
required (column printer, magnetic tape, etc.).
The most automated approach incorporates a minicomputer into the
system. This system could automate all routine functions of data acquisiton,
calculation, and reporting as well as complete routine control of the monitoring
system functions such as purge, zero, calibrate, and routine function checking.
Although no such systems are marketed currently, custom systems of this type
can be built on special order. One specific requirement of the purchase
specifications is that the analyzer meet the performance specifications of
the EPA Continuous Emission Monitoring regulations when outputs are connected
to the data recorder and data reduction system the plant intends to use.
Reporting requirements will be discussed in more detail in later
sections. Suffice it to say that a facility should select data handling
equipment carefully. The level of data handling equipment that should be
used will generally depend on the complexity of the monitoring situation
(single versus multistack, one pollutant or many, etc.). However, simple,
inexpensive systems will require considerable manpower to produce the required
reports. An automated system will reduce manpower requirements but increase
initial costs.
g. Meeting Regulation Requirements
The regulations place several key performance requirements on each
system. Most of the requiremements -- such as accuracy, zero and span drift,
range, and time response -- are simply defined and easily tested. Most
vendors will state that their equipment meets these specifications. However,
the source should make sure these statements and requirements are incorporated
into a guarantee or warranty in the purchase documents. Manufacturer's
literature is no substitute for demonstrated field experience.
In addition, vendors should be asked specific questions about how this
equipment is designed to meet specific requirements. For example, paragraph
(d) of 60.13 requires that the owner or operator must check the zero and span
drift of a monitoring system at least once in each 24-hour period. Minimum
1-38
-------
procedures are outlined in items (1) and (3) of the paragraph. Regardless
of the vendor's representation that his equipment does not drift, the drift
must be checked daily. Therefore, a system must be able to make these checks
either manually or automatically when commanded by the operator.
Except for the Design Specifications for opacity monitors, all require-
ments placed on monitoring systems by Performance Specifications 1, 2, and 3
are checked specifically during the performance evaluation at each source.
Thus, whether the monitor can or cannot meet the specifications must be
specifically verified and demonstrated at each facility.
The only required vendor representation that becomes a matter of record
is covered in Paragraphs 6.2 and 6.3 of Performance Specification 1. This
paragraph allows a manufacturer to certify that his equipment complies with
the requirements of Section 6 of the specification. In this case, the vendor
must furnish a certification of testing, a description of test procedures,
and the test results. This certification should contain a test procedure
as detailed as that in Section 6.3 of Performance Specification 1 (opacity).
h. Level of Automation of Functions
Automation of many of the self-check functions in most monitoring
systems is a relatively recent development. It has been stimulated greatly
by the EPA stationary source monitoring requirements. We can expect that
competition will force manufacturers to rapidly automate the more routine
functions of both measurement and data reduction. In particular, systems
may be controlled with microprocessors or microcomputers similar to these
used in the new generation of electronic ovens.
4. Equipment Provisions for an Agency Observing and Verifying Data
There are no provisions on any monitors that will accomodate an Agency
observer's equipment for routine checking. There are however, provisions and
equipment for checking and maintenance that is available. This equipment
would include any automatic or manual zero, span, or alignment check equip-
ment, such as the optical alignment sight on the Lear Siegler RM-4 transmis-
someter or the span check filter on the DuPont 460 series gas montiors.
1-39
-------
In extractive gas monitoring systems, although not a requirement, it
should be easy for the source to provide an extra valve or fitting for the
inspector to withdraw a sample from the probe or to introduce a calibration
or test gas mixture into the probe. Similar fittings should be provided on
the calibraion gas cylinders as well.
Systems with automated data recording can generally be checked with
the source's own manual output devices (i.e., strip chart recorders) or with
the observer's portable indicating or recording meters at the interface to
the analyzer. If use of agency device is desired, the device must be
electrically compatible with the source's system. During performance tests,
the data flow to the final report should be described and calculations made
in sufficient detail to demonstrate that the calculations correlate with the
readings from the portable meters. During subsequent evaluations, the same
procedure, test points, and equipment should be used to verify the validity
of both the entire monitor system and of the automatic data processing sub-
system.
It should be noted that Part 60.8(d) of the regulations requires safe
sampling platforms, adequate sampling ports, safe access to sampling platforms,
and utilities for sampling and testing. These facilities are required to
evalutate the performance of the continuous monitoring system when it is
first installed. They also would allow an agency to conduct comparison
tests between the installed monitor and the sampling technique used as
reference after the installation has been operational for several months.
Providing for these facilities during initial construction is considerably
less expensive than retrofitting them.
C. MONITOR LOCATION
1. Regulations
Paragraph (f) of Section 60.13 of the regulations requires that the
monitoring systems be located to measure representative emissions or process
parameters from the source.
(f) All continuous monitoring systems or monitoring devices shall be
installed such that representative measurements of emissions or
1-40
-------
process parameters from the affected facility are obtained.
Additional procedures for location of continuous monitoring systems
contained in the applicable Performance Specifications of Appendix
B of this part shall be used.
Each performance specification includes specific installation require-
ments for the type of measurement (opacity, SCL, etc.) and the type of equip-
ment to be used (in situ, extractive). Subsection B, which follows, discusses
the requirements in Performance specification 1 for opacity monitors. Sub-
section C discusses the corresponding requirements in Performance Specification
2 for measuring gaseous emissions.
2. Installation Specification for Opacity Monitoring Systems
The installation specifications in Section 4 of Performance Specification
1 include the following:
4. Installation Specification
4.1 Location. The transmissometer must be located across a section
of duct or stack that will provide a particulate matter flow
through the optical volume of the transmissometer that is
representative of the particulate matter flow through the duct
or stack. It is recommended that the monitor path length or
depth of effluent for the transmissometer include the entire
diameter of the duct or stack. In installations using a shorter
path length, extra caution must be used in determining the
measurement location representative of the particulate matter
flow through the duct or stack.
4.1.1 The transmissometer location shall be downstream from all
particulate control equipment.
4.1.2 The transmissometer shall be located as far from bends and
obstructions as practical.
4.1.3 A transmissometer that is located in the duct or stack following
a bend shall be installed in the plane defined by the bend
where possible.
4.1.4 The transmissometer should be installed in an accessible
location.
1-41
-------
a. Representative Effluent
The problems in measuring opacity come from trying to measure one
phase of material (liquid or solid) suspended in another phase (gas). This
results in nonhomogeneous mixing, stratification due to gravity, segregation
due to aerodynamic or centrifugal forces in bends, etc. In measuring parti-
culate loadings using EPA Method 5, the effect of this maldistribution can
be corrected for by an extensive traversing scheme. When installing an
opacity monitor, however, it is not possible to provide a traversing mechanism
or equivalent, so that special care must be taken in examining and selecting
the monitoring site.
As a general rule, the transmissometer used to measure opacity should
be installed in a round, vertical stack rather than a horizontal, rectangular
duct whenever possible. This is because rectangular ducts tend to have a
less uniform particulate concentration than round ones. As shown in Figure
1-13, the distribution of particulate matter from the top of a horizontal
rectangular duct varies widely. Settling due to gravity causes some of this
stratification, and large particles, high or low gas velocities, or small
radius of curvature can make stratification worse. The use of flow
straighteners and the number and abruptness of bends in the duct also affect
the distribution of particulate.
Besides horizontal rectangular ducts, sources should also avoid sharp
bends in ducts without corrective devices. As Figure 1-14 shows, with sharp
bends and high velocities, the gas flow can separate from the duct wall for
several diameters, forming recirculation zones and causing dust to deposit
in the ductwork. These conditions can also occur in ducts that diverge too
rapidly.
If the Agency believes that the transmissometer being used by a source
does not measure a representative portion of the emission stream, the owner
of a source may be asked to demonstrate that the selected measuring location
is representative for maximum or varying operating rates. This is clearly
stated in Section 4.1.5:
4.1.5 When required by the Administrator, the owner or,operator of
a source must demonstrate that the transmissometer is located
1-42
-------
o
3
(O
en
C
rt>
-P
o
QJ
C
O
S-
-!->
I/)
O)
-(->
tO
S-
03
Q.
to
o
o
O)
3
CD
1-43
-------
in
§.
O)
a
\
j-
0)
Q.
(O
E
O
O
3
-o
s-
ta
c
(O
-M
U
2!
I/)
o
Q.
(U
-a
01
s-
-------
in a section of duct or stack where a representative partic-
ulate matter distribution exists. The determination shall
be accomplished by examining the opacity profile of the
effluent at a series of positions across the duct or stack
while the plant is in operation at maximum or reduced
operating rates or by other tests acceptable to the
Administrator.
When a source installs the transmissometer, ports should be provided
at several different positions on a rectangular duct so that an opacity
profile can be made of the duct. For a round stack, two ports at 90° from
each other will usually be sufficient. The transmissometer must not be
installed in ports used for particulate and gaseous sampling. It should,
in fact, have its own separate set of ports. This will prevent the need
for removing the transmissometer during manual sampling for gases or partic-
ulates. Care must be taken in locating these ports to avoid flow disturbances
caused by other sample ports and sampling apparatus. The regulations require
the source to install the transmissometer in an "accessible location." Far
too frequently, "accessible" is understood to mean "convenient," and little
attention is paid to whether or not the location will give a representative
sample. In fact, what is most important is that the transmissometer be
located in a representative section of duct or stack. Once this location
is established, safe, convenient access should be provided (OSHA approved
elevators, manlifts, ladders, sheltered platforms, railings, etc.). This
will ensure that the instrument gets regular inspection and maintenance.
b. Multiple Feed Points
Paragraph (g) of Section 60.13 requires that each separate emission
stream be monitored, with the following exceptions:
(g) When the effluents from a single affected facility or two or more
affected facilities subject to the same emission standards are
combined before being released to the atmosphere, the owner or
operator may install applicable continuous monitoring systems on
each effluent or on the combined effluent. When the affected
facilities are not subject to the same emission standards, separate
1-45
-------
Continuous monitoring systems shall be installed on each effluent.
When the effluent from one affected facility is released to the
atmosphere through more than one point, the owner or operator
shall install applicable continuous monitoring systems on each
separate effluent unless the installation of fewer systems is
approved by the Administrator.
Take, for example the core of a coal fired power plant with two or
more boiler feeding into a single, common stack. If all boilers are subject
to NSPS, then only the common stack need be monitored. However, if any one
of the boilers feeding that stack is not subject to NSPS, then each of the
other boilers that are subject to NSPS must be monitored separately. If
individual monitors are installed on three boilers, the installation standards
in Performance Specification 1 will apply.
c. Stratification
In stratification, flowstream components move in layers or concen-
tration bands (i.e., nonuniform pollutant or flow distributions) in the
duct. For example, as shown in Figure 1-13, the larger particulate will tend
to move toward the bottom of a large horizontal duct because of higher
settling rates. Similarly, large particulate is generally forced toward the
outside of a bend in a duct by centrifugal forces. These conditions can occur
in most duct configurations. Unless action is taken to remix these "layers,"
the particulate is likely to remain stratified. This will make it difficult
to select a representative monitoring point. Because stratification involves
mainly large rather than fine particulate — and since due to population
densities, large particulate has only a minor effect on opacity -- stratific-
cation is not a critical factor in opacity monitoring.
Short of intuitive judgement by people experienced with locating
transmissometers, there is only one reliable way to determine if particulate
stratification exists in a duct: manual testing methods for gas velocity
and for particulate concentrations on a point-by-level basis must be performed
on a duct cross section. If velocities vary significantly across the duct,
it is likely that stratification exists. Particulate sampling at numerous
points in the duct (as defined by EPA Reference Method 1) will, of course,
best show if stratification exists.
1-46
-------
d. Condensed Water
Condensed water is the most serious problem in making accurate opacity
measurements. The opacity of an effluent is a direct measurement of the
attenuation of visible radiation by particulate matter in the stream. Since
water droplets attenuate light in the same way as solid particles, opacity
cannot be accurately measured by conventional monitoring systems when
condensed water vapor is a major constituent of the effluent.
A source can tell how much condensed water vapor is in the effluent
by determining moisture content (by volume) using two methods. They are:
(1) EPA Reference Method 4 -- Part 60, Appendix A and (2) psychrometry --
assuming that the effluent is a saturated vapor. If Method 4 yields a higher
moisture content than the psychrometry method, the difference can be assumed
to be condensed water vapor.
Paragraph (i) of Part 60.13 allows a source to request and the
Administrator to approve alternative monitoring requirements when interference
caused by liquid water (not a high concentration of water vapor) or other
substances make accurate opacity measurements impossible.
(i) Upon written application by an owner or operator, the Adminis-
trator may approve alternatives to any monitoring procedures or
requirements of this part including, but not limited to the
following:
(1) Alternative monitoring requirements when installation of a
continuous monitoring system or monitoring device specified
by this part would not provide accurate measurements due to
liquid water or other interferences caused by substances with
the effluent gases.
In some cases it may be possible to move the monitor to another
location in the flue system where the stack gas temperature is above its
dewpoint. This will eliminate attenuation of visible radiation by condensed
water droplets. However, this solution is probably not possible in most
installations. For instance, very few wet scrubbers are equipped with flue
gas reheat capabilities to eliminate the visible steam plume. Furthermore,
even though a stack may be above its dewpoint in the hot summer months, it
1-47
-------
will often be below its dewpoint in the winter. This problem is particularly
severe in stacks with relatively low temperatures (<200°F) and uninsulated
stack walls. Consequently, in a case where opacity measurements are affected
by water droplets, the facility and the Agency will have to agree upon the
monitoring of some other parameter to characterize control device efficiency
(i.e., scrubber pressure drop).
e. Installation and Maintenance Considerations
Since opacity monitors must have a narrow field of view, installation
and alignment are often a problem. Field experiments performed by EPA
show that the thermal expansion and contraction of steel stacks and ducts
caused by unit load cycling, startup/shutdown, or weather conditions is
sufficient to throw the alignment of the transmissometer out of tolerance.
To minimize the effects of misalignment, 40 FR4626 Appendix B, Performance
Spec. 1, 8.2.1, requires annual optical and zero alignment during a plant
shutdown. Concrete or brick stacks do not generally experience this problem
due to a negligible thermal response.
A monitor, or any of its optical components (reflectors, sources),
should be located on a rigid section of duct or stack or even mounted
independently so that the alignment will not be seriously affected by thermal
expansion. An independent mounting may be required when the stack is
vibrating heavily or is quite hot (<400°F), since radiaiton from the walls
will affect the measurement.
The monitor should be located in a clean, sheltered, readily accessible
area. The easier it becomes for a maintenance technician to check blower
filters, dirty windows, dirty retroreflectors, alignment, etc., the more
likely it is that maintenance will be done on a regular basis. The most
serious operating problem with opacity monitors is generally keeping the
optical windows and other components clean. Many, but not all, manufacturers
offer high volume blowers that blow 40 to 60 cfm of clean air across the
optical windows. The use of this type of blower is almost mandatory. Note
that positive pressure stacks may seriously affect blower capacity.
1-48
-------
3. Installation Specification for Gaseous and Diluent Emission Monitoring
Systems
The primary specifications for installing gaseous emission measuring
equipment are contained in Section 4 of Performance Specification 2:
4. Installation Specification. Pollutant continuous monitoring
systems (S09 and NO ) shall be installed at a sampling location
C, A
where measurements can be made which are directly representative
(4.1), or which can be corrected so as to be representative (4.2)
of the total emissions from the affected facility.
As explained in the following paragraphs, there are several important
differences between this specification and the corresponding specification
for opacity monitoring. In addition, further regulations in Part 60 are very
specific in regards to individual sources or groups of sources.
a. Representative Effluent
In contrast to opacity monitoring, gaseous emissions monitoring is
concerned only with measuring the gaseous phase. Solids or liquids may be
eliminated or separated from the gas stream during this measurement if
required. Also, since once the gas stream is mixed, it will remain homo-
genous, stratification caused by gravity, aerodynamics, or centrifugal forces
which affect particulate measurements rarely affect gas measurement. As a
result, the major problems with gaseous monitoring are poor initial mixing
of the various gases in the stream and in-leakage of air into the ducts.
Paragraph 4.1 of Performance Specification 2 deals with the problem
of in-leakage:
4.1 Effluent gases may be assumed to be nonstratified if a sampling
location of eight or more stack diameters (equivalent diameters)
downstream of any air in-leakage is selected. This assumption
and data correction procedures under Paragraph 4.2.1 may not be
applied to sampling locations upstream of an air preheater in a
steam generating facility under Subpart D of the part. For
sampling locations where effluent gases are either demonstrated
1-49
-------
(4.3) or may be assumed to be nonstratified (eight diameters),
a point (extractive systems) or path (in situ systems) of average
concentration may be monitored.
A number of EPA programs have shown that considerable air leakage
occurs in air preheaters in steam generating facilities. Therefore, gas
concentrations should only be assumed to be homogeneous if there are eight
or more duct diameters downstream of the air preheaters. This assumption
is not applicable to sampling locations upstream of an air-preheater, nor
is it applicable to paragraph 4.2.1 of Performance Specification 2, because
a diluent monitor may not be installed upstream of the largest, source of air
in-leakage -- the air-preheater. A multipoint probe coupled to an extractive
system or an in situ system must be used for gaseous monitoring as explained
in Paragrpah 4.2.2 below. The Administrator may also require additional
data to prove that the monitoring system data is consistently representative
for various process operating conditions.
4.2 For sampling locations where effluent gases cannot be assumed to
be nonstratified (less than eight diameters) or have been shown
under paragraph 4.3 to be stratified, results obtained must be
consistently representative (e.g. a point of average concentra-
tion may shift with load changes) or the data generated by
sampling at a point (extractive systems) or across a path (in
situ systems) must be corrected (4.2.1 and 4.2.2) so as to be
representative of the total emissions from the affected facility.
Conformance with this requirement may be accomplished in either
of the following ways:
4.2.1 Installation of a diluent continuous monitoring system (02
or CCL as applicable) in accordance with the procedures under
paragraph 4.2 of Performance Specification 3 of this appendix.
If the pollutant and diluent monitoring systems are not of
the same type (both extractive or both in situ), the extractive
system must use a multipoint probe.
1-50
-------
4.2.2 Installation of extractive pollutant monitoring systems using
multipoint sampling probes or in situ pollutant monitoring
systems that sample or view emissions which are consistently
representative of the total emissions for the entire cross
section. The Administrator may require data to be submitted
to demonstrate that the emissions sampled or viewed are
consistently representative for several typical facility
process operating conditions.
An extractive or in situ monitoring system should be installed in its
own sampling port(s) rather than in ports used for manual particulate and
gaseous sampling. Separate ports are required for the performance test of
the monitoring system or for concurrent manual sampling. As with opacity
monitors, adequate sheltered sampling platforms, ladders, electrical power
outlets, etc., should be installed to provide convenient access to the
monitoring location.
Prior to a late 1975 study by the Emission Measurement Branch of the
Office of Air Quality Planning and Standards, it was commonly thought that
alkaline scrubber fluids in most flue gas desulfurization systems alter the
COp concentration, thereby affecting downstream COp measurements. The Federal
Register Part 60.45 (b) (3) (i-iii) reflected this assumption, since if CO-
was used to correct the emission measurements to the units of the standard,
the source was required to locate the (^ monitor upstream of the scrubber.
In response to the EMB study findings, 42FR5936-1/31/77 deleted this require-
ment on the location of C02 monitors where flue gas desulfuriztion is used.
However, since there is a possibility of COp production from reactions in a
limestone scrubber, when this type of FGD device is used a 1-percent increase
in the F factor is required.
\+
If a different fuel from that used in the boiler is fired directly
into the flue gas for any purpose (e.g., reheating), the pollutant (e.g.,
S02), opacity, CO,,, or Op continuous monitoring system must be installed
downstream of this location.
1-51
-------
b. Multiple Feed Points
As noted in Section (g) of Part 60.13, a single monitoring system
may be used to measure the combined effluent from several facilities if all
facilities are subject to the same standard. If there are multiple emission
points from one facility, separate monitoring systems are required unless
an alternative method is specifically approved by the Administrator. In
the case of gases, extractive systems that can sequentially sample multiple
locations are available. As long as the basic cycle time of these systems
is consistent with the 15-minute maximum (Section (e)(2) or 60.13), a single
system may be used to sample several feed points. Multipoint systems are
available from most major extractive equipment manufacturers.
c. Stratification.
Stratification of gaseous pollutants has not been as universally
accepted as it has been for particulate matter. In fact, it is commonly
believed that there is no gas stratification in ducts with turbulent gas
flow. In reality, stratification of pollutant species due to poor initial
mixing and air infiltration commonly occurs. Figure 1-15 shows an example of
S02 stratification in a large rectangular duct at a coal fired power plant.
Paragraphs 4.2 and 4.3 of Performance Specification 2 place specific require-
ments on the sampling locations where stratification has been demonstrated
or may be presumed to exist:
4.2 For sampling locations where effluent gases cannot be assumed to
be nonstratified (less than eight diameters) or have been shown
under Paragraph 4.3 to be stratified, results obtained must be
consistently representative (e.g., a point of average concentra-
tion may shift with load changed) or the data generated by
sampling at a point (extractive systems) or across a path (in
situ systems) must be corrected (4.2.1 and 4.2.2) so as to be
representative of the total emissions from the affected facility.
Conformance with this requirement may be accomplished in either
of the following ways:
4.2.1 Installation of a diluent continuous monitoring system (Op
or C02 as applicable) in accordance with the procedures under
1-52
-------
d.
O-
C
o
ro
OJ
o
C
o
o
OJ
o
CD
C
rO
s-
o
o
Q.
O •*->
+-> O
n
(1) TJ
T> S-
03
-t-> C
fO fO
O -t->
•i- O
4- Ol
+J
ro ro
S-
Lf>
I
O)
01
•r-
u.
1-53
-------
Paragraph 4.2 of Performance Specification 3 of the appendix.
If the pollutant and diluent monitoring systems are not of
the same type (both extractive or both in situ), the extrac-
tive system must use a multipoint probe.
4.2.2 Installation of extractive pollutant monitoring systems using
multipoint sampling probes or in situ pollutant monitoring
systems that sample or view emissions which are consistently
representative of the total emissions for the entire cross
section. The Administrator may require data to be submitted
to demonstrate that the emissions sampled or viewed are
consistently representative for several typical facility
process operating conditions.
4.3 The owner or operator may perform a traverse to characterize any
stratification or effluent gases that might exist in a stack or
duct. If no stratification is present, sampling procedures
under Paragraph 4.1 may be applied even though the eight diameter
criterion is not met.
4.4 When single point sampling probes for extractive systems are
installed within the stack or duct under Paragraphs 4.1 and 4.2.1,
the sample may not be extracted at any point less than 1.0 meter
from the stack or duct wall. Multipoint sampling probes installed
under Paragraph 4.2.2 may be located at any points necessary to
obtain consistently representative samples.
As explained in Paragraph 4.3, a plant operator may perform manual
traverses to show that the gas stream is not stratified at a particular
sampling location. Traverses using a continuous 02 or CCX, analyzer can
quickly determine the extent of stratification.
d. Dilution
Paragraph (d) of Part 60.45 requires the installation (generally in
close proximity to pollutant monitor probe inlet) of either a carbon dioxide
or an oxygen continuous monitoring system at each steam generating facility
(exceptions are noted in Section I(B)(1)). The data from these systems is
1-54
-------
used to correct the emission values from the S0? and NO measurement system
^— /A
for the dilution effects of excess air and air in-leakage from preheaters,
etc. The use of these corrections is discussed in Section III.C.5. Oxygen
and carbon dioxide measurement systems are subject to Performance
Specification 3.
41FR44838 cites the approval of alternative monitoring procedures
provided for by 60.13(i)(3). Restrictions on wet basis diluent and pollutant
measurements have been removed. Specifically, approval has been given for
alternative emission data reduction procedures to be used with wet-basis
stack gas pollutant and oxygen monitoring data from fossil fuel-fired steam
generators. The approved alternative data reduction procedures have been
detailed in 41FR44838.
e. Accessibility for Agency Inspection and Source Maintenance
The regulations for inspecting continuous monitoring systems and
monitoring devices are contained in Section 114 of the Clean Air Act
amendments of 1970:
Section 114(a)
(2) the administrator or his authorized representative, upon
presentation of his credentials --
(a) shall have a right of entry to, upon, or through any premises
in which an emission source is located or in which any records
required to be maintained under Paragraph (1) of this section
are located, and
(b) may at reasonable times have access to and copy any records,
inspect any monitoring equipment or method required under
Paragraph (1), and sample any emissions which the owner or
operator of such a source is required to sample under
Paragraph (1).
Unlike Performance Specification 1, which states, "4.1.4 The trans-
missometer should be installed in an accessible location," Performance
Specification 2 (S0? and NO ) and Performance Specification 3 (C0? and 0?)
£ /\ C- f—
1-55
-------
do not place any particular requirements on accessibility of the systems.
However, as discussed in the opacity maintenance section, the key to an
effective monitoring operation is maintenance, daily operating checks and
weekly periodic maintenance. The EPA evaluation programs are full of
examples where systems down for a week or more could have remained operational
with 5 minutes of maintenance. Furthermore, almost all problems could have
been easily recognized by even a casual daily inspection that included a
zero and span calibration check. This experience clearly shows that a
monitoring system must be readily accessible to receive the daily main-
tenance and care necessary to operate effectively. Therefore, even though
the performance specifications for gaseous monitors do not specifically
require an accessible location, the evaluator should strongly recommend one.
D. CONTINUOUS MONITORING SYSTEM SELECTION AND LOCATION CHECKLISTS
This section contains checklists to be used as an aide to the agency
observer when assisting with and evaluating the selection and location of a
source's continuous monitoring system. Each question in the checklists has
been referenced to the appropriate part of the Manual text.
1-56
-------
SELECTION OF MONITORING EQUIPMENT CHECKLIST
Opacity
Subsection
References Yes No N.A. Notes
t Is the system double or single B.2.a
ended?
• Is the measurement path length B.2.a
the same as the stack exit
diameter?
• How is the unit calibrated? B.2.a
• Does the unit have an alignment B.2.a
check?
0 Does the unit have automatic B.2.a
calibration features?
t Does the unit have any type of B.2.a
malfunction?
• Does the unit have automatic B.2.a
zero adjustment features?
• Does the unit have any type of B.2.a
malfunction indicators? (lamps,
meters, alarms)
• Is the unit certified to meet B.2.a
the spectral response and field
of new requirements of Perfor-
mance Specification 1?
t What window cleaning is B.2.a
required?
t Does the unit have window purge B.2.a
accessories?
1-57
-------
SELECTION OF MONITORING EQUIPMENT CHECKLIST (Continued)
Subsection
References Yes No N.A. Notes
• Does the unit have a visual B.Z.a
output for the operator?
(meter, strip chart)
• Is the unit equipped with any B.2.a
excess emissions indicators?
(lights, bells, horns)
Gaseous Emission Monitoring Equipment
• Which gas is to be measured? B.3.b.i
• What concentrations? B.3.b.i
t Do the emission levels vary B.S.b.i
widely (3/1 or greater) with
time or plant input?
• What are the controlled versus B.3.b.ix
uncontrolled emissions?
• Are appreciable amount of water B.3.b.iii
vapor present (about 1 percent)?
• Is there any condensed water or
other chemical droplets?
• Is multiple stack sampling B.S.b.vi
desirable?
• Do several streams feed into B.S.b.vi
one stack?
• Does the stack also have parti- B.S.b.iv
culate emissions?
• How well are particulate B.S.b.iv
emissions controlled?
1-58
-------
SELECTION OF MONITORING EQUIPMENT CHECKLIST (Continued)
Subsection
References Yes No N.A. Notes
• Does the sampling system B.S.b.iii
require water removal?
• Were specific tests made to B.S.b.ii
select the sampling location?
• Does the temperature or pres- B.S.b.vii
sure vary at the sampling
location?
• Does the monitoring system com- B.S.b.viii
pensate for power line
variations?
t Does the unit have automatic or B.S.a
manual "zero" capability?
• Does the unit have any type of B.S.a
output display for the operator?
• Does the unit read out in con- B.S.e
centrations, corrected or
uncorrected for dilution air?
• Is there an excess emissions B.S.c
warning device?
• When were calibration gas B.3.g
cells last checked?
• Has the instrument been cali- B.S.g
brated over all potential
operating ranges?
3. Data Recording and Reduction Equipment
• What is the frequency of system B.S.g
analysis and recording?
1-59
-------
SELECTION OF MONITORING EQUIPMENT CHECKLIST (Concluded)
Subsection
References Yes No N.A. Notes
What is the method used for B.3.g
averaging?
- Six-minute averages?
- Hourly averages?
How is time of S02 readings B.3.g
correlated with CO readings for
excess air conversion?
How is the date and time B.3.g
recorded?
What is the data recorder B.S.g
resolution and accuracy?
What is the chart speed? B.S.g
Does the recorder display have
a zero offset or is the chart
zero 5 percent of scale?
What are the units of the B.S.g
recordings?
Is the software and system data B.S.g
flow well documented and
demonstratable?
How are frequency and time of B.S.g
process and monitoring system
malfunctions documented?
Who is responsible for reducing B.S.g
data and preparing reports?
1-60
-------
MONITOR LOCATION CHECKLIST
Sampling Location
Subsection
References Yes No N.A. Notes
• Does more than one unit share C.2.b
a stack?
• What standards apply to each C.Z.b
unit?
• How are combined effluent C.2.b
records kept?
2. Opacity Measurements
• Does the monitor view across C.2
the entire duct or a portion?
• Is the unit downstream of all C.2
particulate control equipment?
• Is the unit too near any duct- C.2
ing bends, obstructions,
transitions?
• If the unit is near a bend, C.2
does it measure in the plane
of the bend?
• Is the unit readily accessible? C.2
• Is dilution air added to the C.2
stack?
• Is the monitor at the end of a
long straight duct?
t Is the dust velocity belo-. 50
feet per second?
-------
MONITOR LOCATION CHECKLIST (Continued)
Subsection
References Yes No N.A. Notes
• Does the flow velocity vary C.2.a
with plant output variations?
• Does the opacity vary with C.2.a
plant output?
• Are condensed water or other C.2.d
liquid mists present?
• Is stratification evident? C.2.C
3. Gaseous Measurements
t Is the monitor downstream of C.3
all control equipment?
t Is the sampling location eight C.3.a
or more diameters downstream of
any air in-leakage?
• Is dilution air added to the C.3.a
stack?
• Does an extractive monitor use C.S.c
single or multiple point
sampling?
• Does an in situ monitor view the C.3.a
entire cross section or a
limited portion?
• Does the gas concentration, C.3.a
dilution, or in-leakage vary
with plant output?
• Is the monitor readily accessi- C.3.a
ble?
1-62
-------
MONITOR LOCATION CHECKLIST (Concluded)
Subsection
References Yes No N.A. Notes
• Is a flue gas scrubber system C.3.a
installed?
t Is C02 or oxygen monitored? C.3.a
• Are different fuels fired? C.3.a
• Is stack reheating used? C.S.a
• Are multiple sources fed into C.3.b
one stack?
t Does the monitoring system C.3.b
measure more than one stack?
• If so, is each stack sampled C.3.b
at least once every 15 minutes?
1-63
-------
REFERENCES: SOURCE MONITORING EVALUATION PROGRAMS
1. Evaluation of Monitoring Methods and Instrumentation for Hydrocarbons
and CO in Stationary Source Emissions: EPA-R2-72-106.
2. Monitoring Instrumentation for the Measurement of S0? in Stationary
Source Emissions: EPA-R2-73-163.
3. Instrumentation for the Determination of NO Content of Stationary
X
Source Emissions: APTD-0847 and APTD-0942.
4. Measurement of the Opacity and Mass Concentration of Particulate
Emissions by Transmissometry: EPA-650/2-74-128.
5. Performance Specifications for Stationary Source Monitoring Systems for
Gases and Visible Emissions: EPA-65Q/2-74-103.
6. Continuous Measurement of Gas Composition from Stationary Sources:
EPA-600/2-75-053 a,b.
7. Field Evaluation of S02 Monitoring Systems Applied to hLSO, Plant
Emissions: EPA-650/2-7b-053 a,b.
8. Instrumentation for Environmental Monitoring: Lawrence Berkeley
Laboratory, University of California.
9. Evaluation of Sample Conditioners and Continuous Stack Monitors for the
Measurement of SC0, NO , and Opacity in Flue Gas from a Coal-fired
t*. )\
Steam Generator: Sourth.^rn Services, Inc., Birmingham, Alabama.
1-54
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
340/1-78-005 A
3. RECIPIENT'S ACCESSI OI* NO.
4. TITLE AND SUBTITLE
Resource Manual for Implementing the NSPS Continuous
Menitoring Regulations. Manual 1 - Source Selection
and Location of Continuous Monitoring Systems
5. REPORT DATE
April 1. 1978
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
F. Jaye, J. Steiner, R. Larkin
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORG MMIZATI ON NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Acurex Corporation/Aerotherm Division
485 Clyde Avenue
Mountain View, CA 94042
11. CONTRACT/GRANT NO.
68-01-3158
12 SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
EPA Office of Enforcement
Division of Stationary Source Enforcement
Washington, D.C. 20460
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Manual 1 - Source Selection and Location of Continuous Monitoring Systems - is one
of a series of four manuals that comprise the "Resource Manual for Implementing
the NSPS Continuous Monitoring Regulations." The other manuals are:
Manual 2 - Preliminary Activities for Continuous Monitoring System
Certification (Installation, Notification and Performance Evaluation
Manual 3 - Procedures for Agency Evaluation of Continuous Monitor Data and
Excess Emission Reports
Manual 4 - Source Operating and Maintenance Procedures for Continuous
Monitoring Systems
Manual 1 presents NSPS regulations for selecting and locating opacity and gaseous
monitors for affected sources. Specific problem areas such as stratification,
dilution, water vapor interferences and multiple feed points, are discussed.
Criteria for selecting monitoring systems and checklists for agency inspection
and review of equipment installations are included.
17.
KEY WC^DS AND DOCUMENT ANALYSIS
DESCRIPTORS
Stationary Source
Continuous Emissions Monitoring
New Source Performance Standards
b.IDENTIFIERS/OPEN ENDED TERMS
Continuous Emission
Monitoring
COS -Tl Field/Group
13
B
14 D
-I
13. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (This Report,
Unclassified
21 NO. OF PAGES
64
[20 SECURITY CLASS iTIns pafcj
\ Unclassified
PRICE
EPA Form 2220-1 (9-73)
-------
U*4GE^
C
CD
O
H
O
Z
Z
o
•
m
•
CO
oo
i
o
o
01
Q)
111 If
lljjfll
PSaSS
I" <0 3 3
m o o o
lS 55
O i ;
— —
m
O
=>
o
o
------- |