EPA-450/3-74-054
REGIONAL AIR
*
POLLUTION STUDY
POINT SOURCE
METHODOLOGY
AND INVENTORY
by
Fred E. Littman
Science Center
Rockwell International
1049 Camino Dos Rios
Thousand Oaks, California 91360
Contract No. 68-02-1081
Task Order No. 16
EPA Project Officer: James Southerland
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, N. C. 27711
October 1974
-------
Point Source Emission Inventory
Point sources are the primary contributors to the emission of many pol-
lutants. A detailed, high resolution inventory was required for RAPS. Point
sources, as defined for the RAPS study, are sources which emit individually
more than 0.01% of the total emissions for the St. Louis AQCR of any pollutant.
Emission data are available on an hourly basis.
The primary requirement of RAPS in the emission inventory field was for
those pollutants which can be used as tracers in the modelling studies. Thus,
Initial emphasis was placed on sources of sulfur dioxide (S02) emissions,
since S02 is closely related to stationary point sources. In time, the inven-
tory was broadened to include all of the "criteria" pollutants.
In addition, a number of specialized inventories were assembled and are
Included in this section.
1. Point Source Methodology and Inventory, Phase I, II and III
Rockwell International - EPA 450/3-74-054.
2. Emission Source Testing Programs
Rockwell International - 6802-2093 T0108B, April 1977.
3. Methodology for Inventorying Hydrocarbons
EPA-600/4-76-013, March 1976.
4. Hydrocarbon Emission Inventory
Rockwell International - 68-02-2093 T0108F, March 1977.
5. Non-Criteria Pollutant Inventory
Rockwell International - 68-02-1081 T054, January 1976.
6. Heat Emissions Inventory
Rockwell International - 68-02-2093 T0108G, April 1977.
7. Sulfur Compounds and Particulate Size Distribution Inventory
Rockwell International - 68-02-1081 T056, April 1976.
-------
SCOPE OF INVENTORY
-------
This report is issued by the Environmental Protection Agency to report technical
data of interest to a limited number of readers. Copies are available free of
charge to Federal employees, current contractors and grantees, and nonprofit
organizations - as supplies permit - from the Air Pollution Technical Information
Center, Environmental Protection Agency, Research Triangle Park, North
Carolina 27711; or, for a fee, from the National Technical Information Service,
5285 Port Royal Road, Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency by
Rockwell International, in fulfillment of Contract Mo. 68-02-1081. The contents
of this report are reproduced herein as received from Rockwell International.
The opinions, findings, and conclusions expressed are those of the author
and not necessarily those of the Environmental Protection Agency. Mention
of company or product names is not to be considered as an endorsement by
the Environmental Protection Agency.
Publication No. EPA-450/3-74-054
-------
This report is issued by the Environmental Protection Agency to report technical
data of interest to a limited number of readers. Copies are available free of
charge to Federal employees, current contractors and grantees, and nonprofit
organizations - as supplies permit - from the Air Pollution Technical Information
Center, Environmental Protection Agency, Research Triangle Park, North
Carolina 27711; or, for a fee, from the National Technical Information Service,
5285 Port Royal Road, Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency by
Rockwell International, in fulfillment of Contract No. 68-02-1081. The contents
of this report are reproduced herein as received from* Rockwell International.
The opinions, findings, and conclusions expressed a
-------
RAPS POINT SOURCE EMISSION INVENTORY METHODOLOGY
Table of Contents
I. Introduction 1
II. The Saint Louis Interstate Air Quality Control Region
(SLIAQCR) 2
III. Sources of AI r Pol lution 4
A. Classification %
B. Pollutants of Interest 4-
B1. Sulfur Dioxide b
B2. Carbon Monoxide 6
B3. Partlculate Matter * 7
BA. Hydrocarbons 8
B5. Oxides of Nitrogen (NOX) 8
B6. Heat Emissions 9
C. Sensitivity Analysis HO
D. Size Distribution of Sources 15'
E. Existing Inventory Data 1*8
IV. Emission Data Acquisition 2A
A. Survey 2$
B. Classification of Sources into Acquisition Groups 27
C. Acquisition of Data 27
C1. Stack Gas Measurements 27
C2. Fuel Consumption and Process Data 33
C3. Operating Data 3-6
V. Handling of Emission Data Al
VI. RAPS Inventory Acquisition Schedule k2
VII. Summary and Conclusions 43
-------
TABLES
I. Qualifications of Selected SMSA's 2
II. Classification of Sources for Emission Inventory 5
III. Values of 0 for Selected'Pairs of (a, 1-C) 13
IV. Point Source Emission Inventory, NEDS December 1973 14
V. Maximum Allowable Error a, for Point Sources of
Various Size Acceptance Material 10%, Confidence
Level 95% 9 = 2.24* 16
VI. Sources of Pollutants in the Safnt Louis AQCR 17
VII. List of Companies Emitting More Than 1000 Tons/Year of SO, 19
VIII. Distribution of Large Sources by SCC Codes -
External Combustion Boilers 30
IX. Distribution of Large Sources by SCC Codes -
Process Heaters 6 Processing Emissions 31
X. Minimum Test Schedule 32
XI. Classification of SO. Sources 33
XII. Classification of CO Sources 34
XIII. Classification of Sources of Particulates 34
XIV. Classification of NOX Sources 35
XV. Classification of HC Sources 35
XVI. Wood River Power Station - Daily Log, Unit 5 37
XVII. Wood River Power Station - Boilers 1, 2, 3 Data 38
XVIII. Wood River Power Station - Units 1, 2, 3 Fuel
Oil Usage Log 39
FIGURES
1. Metropolitan Saint Louis Interstate Air Quality Region 3
Qk
2. Relationship between a. and ^— 12
3. RAPS Inventory Schedule 43
iv
-------
RAPS POINT SOURCE EMISSION INVENTORY METHODOLOGY
1. Introduction
An emission inventory constitutes the starting point for any attempt to .
control emissions to the atmosphere. As long as such controls deal with average
yearly concentrations, inventories giving total annual emissions of the various
sources of pollutants are sufficient. The Regional Air Pollution Study has,
however, as its first goal the validation of atmospheric dispersion models,
which attempt to predict ambient pollutant concentrations on an hourly basis.
Therefore, emission values derived from total annual emissions are largely in-
adequate, and the RAPS emission inventory was conceived to provide the needed
time resolution and accuracy by measuring and recording hourly emissions (or
parameters directly related to hourly emissions) and/or individualized hourly
estimates derived for the principal sources of pollution. Thus, the emission
inventory for the Regional Air Pollution Study (RAPS) at St. Louis is distinguished
from existing emission inventories by two factors: its time and space resolution
and its accuracy.
Although ultimately such an inventory should Include all pollutants of im-
portance, as a matter of priority, emphasis of the data collection will be
placed on the two major pollutants of prime importance for modeling purposes,
S0£ as an indicator of pollution originating from stationary sources, and CO
for mobile sources. Hourly measurement and estimates would provide the needed
time resolution and, at the same time, increase the accuracy of the emission
inventory by updating it. Later, the inventory can be expanded to include hydro-
carbons (or organics), oxides of nitrogen, particulate matter, heat emissions
and others.
Any attempt to obtain measured values for a large number of sources is a
complex and expensive undertaking. Within the usual constraints of air pollution
studies, such an approach is not feasible, and the use of algorithms or models
has been generally resorted to for estimation of emissions. Since such emission
models again describe assumed conditions, their use in the RAPS is less desirable
and they will be used only where it does not impair the overall accuracy of the
inventory, as indicated by a sensitivity analysis.
1
-------
This report proposes an approach to the problem of assembling a "precision"
inventory for the St. Louis Interstate Air Quality Region. It states the nature
of the problem and the rationale for choosing the St. Louis area as a "test chamber";
the pollutants of interest are also discussed briefly. Using an approach sug-
gested by NADB's Weighted Sensitivity Analysis Program, limits were placed on
the scope of the investigation, which were then applied to the actual situation
in St. Louis. The mechanism for the acquisition of data and their preparation
prior to entry into a data bank, as well as a time schedule to accomplish these
aims, are also described.
11. The Saint Louis Interstate Air Qua 1ity Control Region (SLIAQCR)
The St. Louis area was selected on the basis of careful considerations of
the various factors of importance for a regional air pollution study . Standard
Metropolitan Statistical Areas (SMSA's) were used as a basis for the analysis,
and all SMSA's with population in excess of 400,000 were examined. The primary
factors considered in the selection were:
0 Geographic isolation from other SMSA's
0 Location within the Continental climate zone
0 Significant level and density of pollutant emissions
0 Presence of a rural fringe with substantial crop lands
0 Existence of control programs and historical data
The final selection of St. Louis was made by the Assistant Administrator
for Research and Monitoring, EPA, from the four considered sites on the basis of
the following rating (Table 1):
TABLE I
QUALIFICATIONS OF SELECTED SMSA'S
Cri terion
Surrounding area
Heterogeneous emissions
Area size
Con t ro 1 p rog ram
Information
Cl imate
Birmingham Cincinnati
Fai r
Fai r
Good
Poor
Poor
Good
Poor
Fai r
Good
Good
Good
Fai r
Pittsburgh
Good
Fai r
Good
Good
Fai r
Fai r
St. Louis
Good
Good
Good
Good
Good
Good
(l) For details, see: Regional Air Pollution Study - A Prospectus, Part III,
Research Facility, Standford Research Institute, 1972, Contract No. EPA 68-02-0207.
-------
\
100 km
:=*-^in
-------
III. Sources of Air Pollution
A. Class! f i cation^
Virtually every human activity results in some form and degree of air
pollution. For practical purposes, it is convenient to classify the sources of
emission; a general classification in shown in Table II. There, sources are
divided into stationary and mobile, since these present significantly different
problems. Stationary sources are further divided into Point and Area sources.
The division between the two is arbitrary: sources tested individually become
"Point Sources". For the RAPS inventory, sources emitting less than ten tons
of pollutants per year will not be considered, at least Initially, as individual
points but rather assigned to and distributed over the appropriate area. Of
course, even a very small point source can be a major contributor to a given
local or nearby receptor (monitoring station), but the investigation of this
problem constitutes a localized, special situation which needs to be dealt with
separately from the overall inventory. Probably the only way the existence of
a local interference can be determined is by examination of the records of each
station.
The division of sources into combustion and nori-coTnbustion is again a matter
of convenience; however, combustion sources constitute a specific group of emit-
ters which, in some cases, like SO-2 for stationary sources or CO for mobile
sources, constitute the overwhelming fraction of thes-e pollutants.
B. Pollutants of Interest
The RAPS inventory is, initially, emphasiz-ing "cri fetia" pollutants
(for which Air Quality Standards exist) and, of those, primarily S02 and CO,
since these are the ones which will be used at highest priority in the model
validation studies. An attempt will also be made to inventory ftBSt emissions.
Ultimately the inventory will also include lesser pollutants such as trace and
hazardous contaminants.
B. 1 Sulfur Dioxide
Sulfur dioxide (S02) will be the pollutant initially emphasized
in the inventory since It occupies the position of highest priority within the
Regional Air Pollution Study. In St. Louis, virtually all of it (98.9?) is es-
timated to originate from listed point sources'^. Most of the S02 is produced
by the combustion of coal and fuel oil, which average 3% and 1.5 - 2.5fc of sulfur
respectively, although some of it results from ore roasting, steel production, and
(2) Source:NEDS Inventory (1973)
k
-------
S
W
§
M
H
g
W
u
u
3
O
w
T4
XI
1
m
u
(4
3
O
(0
>v
t-
(0
g
•H
+»
a
CO
8r«
p
3 0>
0 +»
W CB
U
tn
0)
y
§
CO
If
d
a
ea
0)
u
3
O
CO
.p
d
•H
O
§
to
1
I*
2!
3 "£
**
CO
d
o
^
+J
tn
5
g
o
y
fo
**
d
o
tn
|
d
o
•H
3
E
8
O
d
0
•H
•M
m
U
d
o
2
_2
§
o
d
0
d
o
•H
n
S
1
o
21
3 n
0 g
ft
•o
c -a
i** 0)
o p a
y C
5 0 +*
> Ml %-t
+* dm o«
O HH -H V« (U
y a +• o fc .*
c: M CO. co n
v, o a> a oh
b h > > * .0
3 f«
W CO
Ifl
m o>
0) -»
IH y
w u -H
4> -H .d
W h 3 0> > -M OJ
•H n £> > d c
y y x o) -H v>
•H -Q M CO g t* **
> M y n (n t£ 3 ^*
o tn U I -H m i ca +* *j
n w 3 G -H to HH "0 dtAfl'
«H«hoaa)<«c y-H-i
3 -H '
w -i
m ^ be
a B r* d
0) -H CO -H
y M -rt fH
2 * fe S
P. 3 W 0
•o n J3
•O C E
2 - -•§
•H t* 0) CO
at J3 m o
-H 4-> 4-1 10 bO
h y m o
•C 03 < C. W
^
h *J -O V
odd d
CS »H i-t JO
h o. u a en h
Q) .rt C ID 3
d h o ^ d **
•H a> m a) -H
y » -o h 4J ho tn
d o o o d d a
0)<*H ft h -M -H 4) bfl CO
*J 1 -rf CO h
dl T3 *-t 1) >i C ^> O
tnacxra y tn d •*-> d i . h
O -rt-H^»l»«OWOJO
t-4 h y (K (U inO>i-l3*" o.
CO Oi ft) fa£
^ t. O bfl ^ O
*-> o coo
a -a o *a c
y -H M n d c <~*
t« d « C aj tq n
0 w>>^M«h
1
m CQ i
•-« ^ a a o h
§*J 0> V D, O
-H W £ JC « « 4J
•-H 0 « 3 0) h-HQ
fci +j o> *-*o>a04->-H
r n tn M EM ^ -H
0 d 0) B
u »H a co
m
4-1
c
3
o
y
3
0
W
h
a
9
d o u
0 -M *H
^ a o c
t. iH -H O
ca 3 d Ui
y y m H
O -H M O
T3 h O M
>> ca
M
a>
3
"O VI
d
to a) vi
h 0
R« to
d « rH M
co o 0) n 04>
> ja > v m > in
h ** •*•> 6 -H Q)
•O « ** (0 S ** >
c y a •-* 0) a -H
o o > co t* -M
M -H h y J«: T3 -H-H
o> OOO ^OTS-P E tu •** ^ > a> ea
h -H 4-» rH
1 S U 3 -g 8
25 23"^
» XK TJt.-H 0-HXBg
a> oooi>tO)ug«'c+J'^>c-"
a ~" a 6
o a. w
•
o
Ad) O
as. s
d eo ~ x-. m o>
> > « g o d
•owe--* a u
t3 -H 0) «H <~* T*
ca c > c w 3 d ffl
a >-t m h y co J<
« O ^ S r* o to
a Ou
•a c
d no
n a *-> -H
h -H B
&tn at
en -ax:
« o a» n B u n
t« -H -p y d
T) 0) ** C5 J3 -r* g D
VOOO XQ)Ufrj33P*"*O,OJ«!
C CB E
O ft W
d
P
3
t-t
£
I
^
>- u
T3 «-
3 ro
4-1 (1)
LO U)
(U
C DC
o
4-1 L.
3 O
— M-
— C
O (D
O. 4-1
CO
(D O (U
C (1) 4J
o a. 3
_ m 4-1
DI o —
a) L. 4J
a: a. >n
3
O
to
-------
petroleum refining operations. The largest contributors are the power generating
stations of the utility companies. The six generating stations in the St. Louis
area produce over 900,000 tons of S02 per year, or about 75% of all the $03
produced by point sources in the area.
Sulfur dioxide is relatively non-reactive in the atmosphere, at least over
the time interval of a few hours, which is likely to be of interest to modelers.
Removal from the atmosphere occurs by several mechanisms, some of which involve
oxidation to sulfur trioxide with subsequent formation of sulfuric acid mist or
sulfates by reaction with basic materials in the atmosphere (e.g., ammonia).
These processes will have to be considered for long-term (2k hours or longer)
model ing.
Available evidence indicates that the ratio of S00 to SO, in ambient
(3)
air is between 50:1 to 100:1. Recent health data indicate that (at least in the
case of elderly patients with heart and lung diseases, as well as asthmatics) it is
the level of suspended sulfates that correlates with adverse health effects rather
than the S02 level. Best estimates indicate that sulfates are about an order
of magnitude more irritating than S02« At this time, it is not clear whether
sulfuric acid mist or sulfates are implicated, and the importance of atmospheric
transformation products of S02 is not certain.
Ambient concentrations of SO. in the St. Louis atmosphere typically range
3 (3)
from 20 to kO micrograms/m (annual average) .
B.2 Carbo.i Monoxide
Carbon monoxide (CO) is closely linked with automotive traffic.
Stationary combustion sources normally generate only relatively minor amounts of
CO. There are, however, a few important industrial sources of CO: the catalytic
cracker regenerators in petroleum refineries, blast furnaces in steel mills, and
certain chemical processes. And, because of the tremendous volume of stack gases
generated by electric utilities, the relatively low concentrations of CO in these
gases do contribute significantly to the overall CO concentration.
Carbon monoxide is chemically inert. It is removed slowly by contact with
certain soil bacteria, which maintain the natural balance of CO in the air, but
the rates of these processes are not significant on the time scale of interest.
(3) Health Consequences of Sulfur Oxides: CHESS 1970-71, U. S. Environmental
Protection Agency, Office of Research & Development, EPA 65011-7^-004.
-------
Carbon monoxide combines with hemoglobin 200 times more readily than oxygen;
it thus prevents the blood hemoglobin from transporting oxygen from the lungs
to the tissues. Exposure to low concentrations (below 100 ppm or 115 mg/m3)
causes headaches and dizziness.. Its actions are most likely to affect persons
living at high altitudes and people with chronic heart and lung diseases.
Cigarette smokers commonly have 5 - 10% carboxy-hemoglobin, an amount that
corresponds to 30 to 60 ppm of CO in ambient air (35 to 70 mg/m^).
Ambient concentrations of CO in the downtown St. Louis area range from 15
o(4)
to 35 mi 1 1 igrams/m-3 .
B.3 Particulate Matter
The fate of particulate matter in the atmosphere is becoming a
major research target. It is a particularly difficult subject because the
characteristics of particles are determined only partially by their chemical
composition and very largely by their size distribution. Thus haziness, by far
the most obvious manifestation of air pollution, is strongly dependent on particle
size. Similarly, the health effect of particulate matter is largely dependent on
particle size, since only particles of a certain size range penetrate into the
lungs and are retained there. The particle size of interest in these areas is
of the order of less than five or six micrometers. Such particles remain afloat
virtually indefinitely and, while their contribution to the total weight of par-
ticulate matter is small, their number is very large.
By contrast, the emission of particulate matter is determined on a weight
basis, whether by sampling or by material balance consideration. Thus the small
number of relatively large particles accounts for most of the mass of particulate
emission. Since particles in excess of 10 ym settle out rather rapidly, these
particles do not contribute much to the ambient concentration of particulates,
nor to their health and visibility effects.
(4) Air Quality Data -_19.72_ Annual Statistics, U. S. Environmental Protection
Agency Office of Air Quality Planning & Standards,
-------
Thus, a really useful inventory of particulate emissions would have to specify
not only the mass but also the size distribution of particulate emission as well
as their chemical composition -- a difficult and expensive task which cannot be
carried out on a routine monitoring basis.
The problem is further complicated by the processes which form particulates —
mainly droplets — In the atmosphere. The formation of SO- leads directly (via
reaction with water vapor) to the formation of a sulfuric acid mist and to the
stabilization of fog; photochemical reactions result in the polymerization of
initially gaseous hydrocarbons, resulting again in particulate droplets. These
products are only indirectly related to emission inventories.
B.4 Hydrpcarbpns
In the air pollution literature, the term "hydrocarbons" is used loosely
to designate gaseous organic compounds. There are two major categories of sources
of hydrocarbons in urban atmospheres: incomplete combustion and evaporation.
Incomplete combustion occurs primarily in internal combustion engines (automobiles),
Evaporation results from the storage and handling of solvents, petroleum products,
etc. Additionally, methane is a normal constituent of the atmosphere, the result
of natural decomposition processes.
Hydrocarbons participate in photochemical reactions leading to "smog", but
their reactivity varies widely. It is therefore important to determine not only
the amount of hydrocarbons present, but also their composition. From a practical
point of view, usually only gross classification is possible on a continuous
inventory basis, such as methane and non-methane hydrocarbons. Further separation
into reactive hydrocarbons (olefins, diolefins, aldehydes, alcohols, etc.) and
stable ones (e.g., parafins) is possible but will require an extensive sampling
program. Complete analyses of samples collected in bags by means of a gas
chromotograph are scheduled for samples of ambient air at St. Louis.
B.5 Oxides of Nitrogen (NOX)
Emission inventories of nitrogen oxides constitute a special problem
since these compounds -- particularly nitric oxide (NO) — are primarily formed
by nitrogen fixation during combustion operations. Their formation during com-
bustion is a complex function of the time-temperature relationships in the com-
bustion chamber, the amount of excess air present, and even the chamber con-
figurations. Any nitrogen compounds present in the fuel also contribute to the
8
-------
formation of nitrogen oxides. Because of this, the nitrogen oxide concentration
in flue gases cannot be calculated from a theoretical basis but must be determined
experimentally for, at least, each typical situation.
In addition to combustion sources, there are specific point sources emitting
nitrogen oxides, usually N0_, such as nitric acid plants. The NEDS inventory does
not show any such sources in the > 100 tons/year category in the St. Louis area.
As mentioned previously, the importance of oxides of nitrogen and hydro-
carbons as pollutants is primarily as participants in photochemical reactions
where N0« acts as primary light absorber. These compounds will therefore be of
importance to RAPS only if and when a study of photochemical reactions in the
atmosphere is planned.
B.6 Heat Emissions
The large amounts of energy produced and consumed by a city event-
ually are converted into heat, resulting in a "heat island" which has an effect
on atmospheric stability and thus affects modeling efforts. A heat emission
inventory is required for a comprehensive understanding of this effect in much
the same way as a pollutant emission inventory forms the basis for an understanding
of the fate of the pollutants.
Point sources contribute significantly to the heat emission inventory, since
a sizeable portion of the energy consumed is wasted as sensible heat of the stack
gases. Even in highly efficient power plants, about 20 per cent of the energy con-
sumed is wasted at the plant. In some industrial operations, such as blast furnaces,
essentially all of the heat of combustion is released to the atmosphere at the plant.
Actually, in a self-contained area such as St. Louis, not only the waste
heat turns up as heat emissions, but virtually all of the converted energy as
well. Except for minor amounts of energy stored as chemical energy (e.g., in
a primary aluminum plant) or radiated into space as visible light, all other
forms of energy, whether electrical or mechanical, are converted into heat and
released into the atmosphere, spread out over the inhabited area. Thus, as a
first approximation, the total Btu content of the fuel used at St. Louis can be
assumed to be released at either point or area sources. The amount of heat re-
leased by point sources can be calculated directly from fuel consumption and
-------
known conversion efficiencies of the power boilers; it can be verified by stack
analysis and measurement of gas volume and temperature, from which the
sensible heat above ambient can be calculated.
Since fuel consumption figures will be obtained in any case, a program to
calculate heat emission from point sources will be initiated. Significant point
sources, defined similar to pollution point sources, will be treated individually,
All other sources will be assigned to grid squares, whose total emission can
be estimated, given the daily total of heating or cooling degree days, average
wind speed, the day of the week and month of the year. Point sources can be
classified into industrial, commercial and residential sources; power generation
will be treated separately.
C. Sensitivity Analysis
An important aspect of every inventory is its accuracy. While no
inventory can be better than the numbers supplied by the data acquisition process,
a statistical estimate of the overall quality and probable error would help place
the uncertainties on a quantitative basis.
As a first approach to this problem, the National Air Data Branch
of EPA commissioned a study which produced a Weighted Sensitivity Analysis Program.
While this program does not supply any estimates of the absolute accuracy, it
does help evaluate the maximum permissible error of any part of the inventory,
given a maximum permissible error for the whole system. In doing so, it keeps
the inventory at an equivalent level of accuracy and points out areas where
accuracy has to be improved to provide a desired overall accuracy. In addition,
it also provides an approach to establish confidence levels for the emission
inventory.
The basic theoretical development proceeds as follows . The linear model:
22 22
<* Q ' I . <£ \
k=l
where Q = total amount of pollutant emitted
100 0 = percentage error associated with Q
0. = amount of pollutant emitted by subclass k
100 a. = percentage error associated with 0_,
is postulated as an appropriate model to analyze the propagation of errors through
the emission inventory.
(5) See F. H. Pi tto et al, Weighted Sensitivity Analysis of Emission Data,
Fed. Syst. Div., IBM, EPA Contract #68-01-0398 (1973).
10
-------
If each subclass contributes to the error an amount proportional to its
relative physical contribution, it can be shown that
The analysis demonstrates that to obtain a predetermined level of precision
for a source class, not all subclasses need to be measured with the same
precision; the greater the ratio of Q:0_. becomes, the greater becomes the
allowable value of a. . Conversely, 0_, approaches the value of 9 as the ratio
approaches unity (Figure 2). *>
The authors also developed a method for predicting the confidence level
for the inventory; that is, the probability that the actual overall error will
not exceed 0, using Chebyshev's theorem^'. The results for selected pairs
of (a and 1-c) are shown in Table III, where a - 29 and 1-c is the confidence
level .
A two-step procedure thus is suggested. First, establish the overall
allowable error G, either from user's (modeler's) requirement or as a trade-
off between confidence level and acceptable error interval; secondly, compute
the values of a. for the components of interest.
Applying these considerations to our case suggests that, in the absence
of any definite information about the modeler's requi refrients for the accuracy
of emission data, a fairly stringent set of conditions would be a confidence
level of 95 per cent and an acceptance interval of 10 pt£r cent (these conditions
are probably stricter than the accuracy of the emission data). This would lead
to a permissible maximum error 0 of 2.24%
Using the emission values in the latest NEDS inventory, as shown in
Table IV, we can now calculate the allowable error for source classes of
various sizes, such as 100 tons/year, 1000 tons/year, etc. For example,
the allowable error for a 100 tons source of S0_ would be
(6) Miller, I. and J. E. Freund, Probability and Statistics for Engineers
Prentice-Hall, 1965.
11
-------
0 = 5*
100
90
80
70
60
50
ko
30
20
10
1 1
1 1 1 1 1 1 1 t
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Source: Reference (k)
RELATIONSHIP BETWEEN a^ AND 0_K/0_
Figure 2
12
-------
Confidence Level
4-1
c
0)
o
c
CO
•M
Q.
<0
o
o
1-C
90%
95%
99%
1.58%
1.12%
0.5%
10%
3.16%
1.0%
20%
6.32%
2.0%
TABLE III. VALUES OF 0 FOR SELECTED
PAIRS (a, 1-C)
Source: Reference
-------
TABLE IV
POINT SOURCE EMISSION INVENTORY
NEDS, DECEMBER 1973
S0x
NO
X
CO
HC
Part.
S°x
N0x
CO
HC
Part.
STL
112,
34,
1,
9,
12,
Co.
206
979
178
255
129
Randolph
473
128
2
6
,599
,964
,993
901
,399
STL
12
2
39
15
6
City
,798
,647
,433
,965
,527
Mon roe
—
—
— 2
--
359
St.
1
Madi
202,
51,
,787,
47,
271,
Charles
15,392
60,086
1,093
328
321
son Cl
630
423
726(7)
694
338
Jefferson
138,384
214
63
3
835
in ton Bond
270
114
4
2
277
Frankl in
111,132
28,492
1,161
303
1,244
Grand Total
1,
2,
Wash.
9
5
2
I
900
(Point
187,296
310,993
836,270
78,476
323,952
St.
20
4
2
4
23
Clai r
,816
,069
,617
,024
,623
Sources)
NEDS, December 1973
14
-------
"IOOT =
- 2 2k ]'187-296
" *24 100
The data are tabulated in Table V.
The very large a. for S02, CO and eve^ NO , even for the relatively
stringent statistical conditions, suggests that there is probably no. need
to obtain measured hourly values for 100 tons/year sources of these pollutants,
since in most instances NEDS can be relied upon to provide data of this accuracy,
Thus the collection of hourly data can be limited to sources of I'tiOO ton/year
and larger. As indicated in Table III (Section C) , thi's will red'uce the
number of SO- sources to be measured to 62, the stationary CO Sources to 13,
and the NO sources to 26. The remaining sources can ttien be modeled as dis-
J\
cussed in Section IV C3.
D. Size Distribution of Sources
The situation in St. Louis lends itself to a direct attack on the
problem of direct measurement of emissions because of the relatively limited
number of major point sources. In terms of S0_, the current National Emission
Data System (NEDS) inventory lists about 300 sources emitting oven ten tons
of S02 per year. Of these, only about 62 emit in excess of 1000 tons/year,
an additional 120 over 100 tons/year. (Since sources emitting le'ss than 100
ton/year are likely to contribute less than 10% of the amount specified by the
National Air Quality Standards to ambient concentrations, they are usually
lumped with area sources.) The 62 largest sources, representing 15 companies,
are concentrated at 20 locations. Thus, the sheer physical magnitude of the
problem of collecting hourly data for the major sources of pollution appears
to be manageable within a reasonable budget.
The situation for other pollutants is somewhat similar. The data are
summarized in Table VI.
Thus, if direct measurements of emission will be limited to sources emit-
ting in excess of 1000 tons/year, we need to obtain data from 62 sources at 20
15
-------
TABLE V
MAXIMUM ALLOWABLE ERROR ok FOR POINT SOURCES OF VARIOUS SIZE
ACCEPTANCE INTERVAL 10%, CONFIDENCE LEVEL 95%, 0 = 2.24%
TOTAL P.S.
EMISSIONS, Q ALLOWABLE ERROR a|< FOR POINT SOURCE OF
POLLUTANT TONS/YR.* 100 T/Yr. 1000 T/Yr. 10,000 T/Yr.
S°2
CO
NO
X
HC
PART
1,187,296
1,684,794
310,993
78,474
323,952
244%
230%
125%
63%
127%
77%
92%
40%
20%
40%
24%
29%
-
-
-
*Source: NEDS December 1973, except for CO, which is given in NEDS at
2,836,270, apparently in error.
16
-------
cc
CO
o
CO
UJ
co co
< H-
O
O.
u.
o
CO
LU
cc.
=>
o
PARTI CULATE NO • HYDROCARBONS
y\
0
o
POLLUTANT S02
(0
4->
O
o
A
CM
O
A
CO
O
A
O
o
A
CNJ
O
A
CO
O
A
to
4->
O
0
A
C-g
O
A
CO
O
A
"flj
4-1
O
O
A
CM
O
A
CO
O
A
Tons /Year
cr\
r^
C>J
ur\
so
o
CO
ff\
(M
LT»
vO
ro
vO
LA
CM
PO
OO
vO
CM
r^
CM
ir\
vD
oo
r^.
oo
CM
CM
CM
vO
oo
f>
J-
o
ro
CM
CM
CM
vO
I/)
(U
u
L.
3
O
CO
4-
O
0
Z
NO
IA
-a-
CM
CM
CM
O
CM
ir\
CM
CM
%
0^
oo
o
4-1
TO
u
o
—1
<4-
O
*
o
z
UA
CM
CM
0)
C
-------
locations for S02, 13 sources at 9 locations for CO, 28 sources at 12
locations for particulates, and so on. Many of these sources overlap, thus
further reducing the data collection (but not the data recording) problems.
For example, of the 26 major sources of NO , 21 are also major emitters of
P\
S02« The extent of the overlap is shown in Table VII, which lists all major
sources of pollutants in matrix form.
E. Existing Inventory Data
Air pollution studies have been conducted in the St. Louis area for
many years, and several emission inventories have been developed. In 1964,
an "Interstate Air Pollution Study, Saint Louis-East Saint Louis Metropolitan
Areas" was undertaken by the U.S. Public Health Service. Questionnaires were
sent out to determine fuel use and combustible waste disposal practices in
the area as well as manufacturing activities. A revised emission inventory,
still based on 1963 data, was published in December 1966 as Phase II of the
Interstate Study.
After the Metropolitan Saint Louis Interstate Air Quality Control Region
had been established, the first comprehensive inventory was taken in 1968,
to serve as a basis for the Implementation Planning Program (IPP). Since
then, four more inventories have been compiled:
IBM Emission lnventory-1970
. DAQED Emission Inventory-1971
. NATO Emission Inventory-1971
NEDS Emission lnventory-1973
In addition, the following traffic and transportation inventories exist:
Streets and highways
Railways and vessels
The emission inventories in current use by the Missouri and Illinois
regulatory agencies were recently (Summer 1973) acquired and transferred to
the NEDS files.
18
-------
TABLE VII
SOURCES OF POLLUTANTS IN THE ST. LOUIS AO.CR
EMITTING IN EXCESS OF 1000 TONS/YEAR
Source Name
All led Chemicals
Alpha Cement Co.
Alton Box Co.
Amoco
Anheuser-Busch Co.
Anl in Corp.
Chrysler Corp.
Clark Oil Co.
Columbia Quarry
East St. Louis Stone
Ford Motor Co.
Point
No.
01
01
01
02
03
01
02
03
04
05
06
01
01
01
02
01
02
03
04
01
02
01
01
02
Pol lutant
S°2
X
X
X
X
X
X
X
X
X
X
X
CO
X
X
Particu-
lates
X
X
X
X
X
X
NO
X
X
X
HC's
X
X
X
X
X
X
X
X
X
X
X
Proposed
Stack
Samp] ing
X
X
X
X
19
-------
TABLE VI I CONTIlUJEU
Source Name
' GMAC
Granite City Steel
Highland Electric .Co
1 1 1 inois Power Co.
Laclede Steel
Point
No.
01
02
03
04
01
02
03
Ok
05
01
01
02
03
04
05
oe
07
OS
' 09
10
11
01
02
03
Pollutant
S02
.
X
X
X
X
X
X
X
X
X
CO
X
X
X
X
X
X
X
X
X
Particu-
lates
;;
X
X
X
X
X
X
X
X
N0x
X
X
X
X
X
X
X
X
X
X
X
X
HC's
X
X
X
X
X
Proposed
Stack
Sampl inp
•X
X
20
-------
TABLE VI I CONTINUED
Source Name
Mississippi Lime Co.
Mississippi Portland
Cement
Monsanto Chemical Co
Municipal Incinerator
NL Ti taniuni Uj v.
PPG Glass
St. Joseph Lead Co.
Shell Oil Co.
Point
No.
OK
01
01
02
03
Ok
Op
06
07
08
09
01
02
01
02
03
04
01
01
02
01
02
03
04
Pollutant
S02
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
V
CO
X
X
X
X
Particu-
lates
X
X
X
X
X
X
X
N0x
X
HC's
X
X
Proposed
Stack
Sampl ing
X
X
X
X
X
X
X
X
05
21
-------
TAbLE V! I COtJTIIlUtJ
Source Name
Shell Oil Co.
Socony
Stol le Quarry
Texaco.
Union Electric
Point
No. •
06
07
08 •
09
10
11
12
01
01
02
01
01
02
03
04
05
Ob
07
05
03
10
]]
12
13
14
Pollutant
S°2
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CO
Part icu-
lates
X
X
X
X
X
X
X
NO
X
X
X
X
X
X
X
X
X
X
X
X
HC's
X
X
X
X
X
Proposed
Stack
Sampl ing
X
X
X
22
-------
TABLE VI I CONTINUED
Source Name
Union Electric
TOTALS
Point
No. •
15
16
17 ,
96
Pollutant
so2
X
X
X
62
CO
13
Particu-
lates
28
N0x
26
HC's
23
1
TOpOStJU
Stack
Sampl ing
17
.
23
-------
These inventories are described in detail in SRI Report "A Regional
Air Pollution Study Preliminary Emission Inventory" (1971*) EPA No. 68-02-1026.
Most of these inventories-are only of historical interest. Current data
are contained in the National Emission Data System (NEDS) inventory, ad-
ministered by the Federal EPA,and similar inventories kept by the Illinois
EPA and the Missouri agencies.
The NEDS inventory contains information on annual emissions of the five
"criteria" pollutants (particulates, SO., NO , hydrocarbons [HC] and CO) from
£- x\
stationary point and area sources, as well as a listing of selected industrial
materials emitted by chemical process, food, agriculture, chemical and mineral
products industries, petrochemical operations, wood processing, and incinerators,
From the point of view of the Regional Air Pollution Study, the NEDS
inventory has two major uses: it contains emission data for those sources for
which detailed data are unavailable, and it provides a basis for an analysis
of the problem of obtaining measured data. It therefore can serve as an in-
terim data base for the St. Louis study until the RAPS inventory becomes
operational.
IV. Emission Data Acquisition
A series of sequential steps leads to the eventual acquisition and re-
cording of point source inventory data for the RAPS inventory. The steps are:
Survey
Classification of Sources into Acquisition Groups
Acquisition of Data by:
1) Stack analyses
2) Fuel consumption or process data
3) Derivation from operational data
Transformation of Data and Entry into Computer Bank
(7) APTD-1135, "Guide for Compiling a Comprehensive Emission Inventory",
March 1973.
2k
-------
A. Survey
Classically, data for emission inventories are acquired by the use of
questionnaires which are either mailed out or prepared by the interviewer or
inspector on a one-time basis. -
The requirements of the RAPS inventory for hourly measured data for a
period of a year far exceed the normal reporting routine and require special
arrangements with the management of the various facilities. Thus, personal
contact with the appropriate corporate office by mail, phone and, ultimately,
in person was considered essential to obtain the necessary cooperation. The
request was made for access to data which would provide a basis for calculating
hourly emissions (using OMB approved NEDS questionnaires as a starting point).
Such data could be
stack concentration measurements
fuel consumption records
process data
steam production records
These data, coupled with the necessary secondary information, such as stack
gas volume, concentration of sulfur in fuel or in process materials, etc.,
will permit the calculation of the weight estimates of pollutant (e.g., S02)
emitted per hour.
As a sample, appropriate officials of nine of the 15 companies emitting
more than 1000 tons/year (shown in Table 2) were contacted and interviewed.
These included Union Electric, Illinois Power, St. Joseph Lead, Alton Box Board,
Laclede Steel, Monsanto Chemicals, Anheuser-Busch, Shell Oil Company, and Amoco
Oil. These nine companies are responsible for over 90% of the total S0«
emitted from point sources in the St. Louis area (based on NEDS data). All
but two agreed to supply the necessary hourly data to RAPS; this includes
the utility companies, who are emitting approximately half of all the S0? in
the area. Thus, even if the percentage of cooperation of the smaller companies
should drop off, it appears that measured data for at least 90% of the total
emission of SO^ will be available. The accuracy of these data is of the order
of the finance accounting procedures used by the companies, which is higher
than that of chemical analysis.
25
-------
This type of a survey will be continued to include the remainder of the
major sources. There are two levels at which the initial information has to
be gathered:
1. Management level
2. Operational level
At the management level, an "agreement in principle" is required; usually
operational personnel is present at these meetings since they will later on
be involved. After an agreement is reached, the details of the data acquisition
are worked out with operational personnel.
The following information is then secured from information gathered at the
operating level:
1. Source Description: address, location (by UTM coordinates),
type of operation (SIC and SCC Codes), etc. Most of this
information is available in the NEDS printout but has to be
verified (particularly location, which should be to ± 0.01 Km).
2. Needed Data: pollutant concentration in stack (rarely
available), quantity and type of fuel burned, amount of steam
produced, fuel analysis, process information (weight and analysis),
etc. All of these data should be on an hourly basis; if they are
not, the time interval should be noted, as well as time related
variabi1i ty.
3. Col lection of Data: data collection will be arranged so as to
minimize the effort required by the affec.ted companies. A mes-
senger circuit will be set up to pick up data once a week (or
other agreed-on time interval). If required, the data may have
to be reproduced; in some cases, mail arrangement will be worked
out.
For point sources emitting less than 1000 tons/year, as well as those major
sources where detailed data are not available, hourly emissions have to be
derived by a model, as discussed further on under B3. For these sources, the
following information js necessary.
1. Source description - as above.
2. Work schedule
3. Maximum process and space heating loads
**. Monthly and shift fuel weighting
5. Fuel analysis data
26
-------
B. Classification of Sources into Acquisition Groups
The division of sources of pollutants into major (those emitting more
than 1000 tons per year) and minor (emitting between 100 and 1000 tons/year), which
is based on sensitivity analysis discussed above, produces two broad categories.
Data from sources in Category 1, .the major sources, will be collected on an hourly
basis to the extent that they are available. Data from all other sources, that ,
is the minor ones and those of the larger ones where detailed data are not avail-
able, will be derived by a model ing or algorithmic procedures.
In Group 1 are the utilities and the majority of sources emitting over 1000
tons/year of pollutant, as determined by the initial survey of sources. The
data available from these sources permit a direct calculation of the weight of
pollutants emitted any given hour.
Although sources in Group II contribute only a minor portion of the overall
pollutant load, they may be of considerable significance locally. Under certain
conditions it may become necessary to obtain measured emission data from some of
these sources as a special project.
C. Acquisition of Data
Cl. Stack Gas Measurements
The RAPS emission inventory should ideally contain direct statements
of weight of pollutants emitted from each major source as a function of location
for every hour. The most direct way to acquire this information would appear to
be to monitor stack emissions.
In actuality, emissions (in terms of weight of pollutant) cannot be directly
measured. Stack gas analyzers only provide a measure of the concentration of the
pollutant, thus requiring another measurement — stack gas volume — before the
weight of the emitted pollutant can be determined. Stack gas volume, in turn, is
not measured directly, but rather is determined by measuring the gas velocity by
traversing the cross-section of the stack. From the average velocity and the
known dimension of the stack, the volume of the stack gases can be calculated.
In addition, the molecular weight of the sampled gas has to be determined to ob-
tain therpass flow rate. Thus, the seemingly direct and straight-forward approach
to the determination of pollutant emissions by stack analysis actually consists
of a number of measurements, manipulations and calculations, each of which con-
tributes to the accuracy of the final figure.
27
-------
In the case of SO , there is an alternative approach since all of the sulfur
' y\
is contained in the fuel and is either emitted in the stack gases or remains in
the residue (ash). The distribution of SO-iSO, in stack gases has been found
to be about 98-99*1. Therefore, the amount of S0_ emitted can be determined
quite accurately from fuel consumption and analysis figures, and the SO. inventory
for RAPS will be obtained in this manner. With other pollutants, there is no
choice, and stack sampling is the only way to obtain the desired information.
Stack sampling methods for compliance purposes have been standardized. EPA
methods are described in CFR Title AO (Protection of Environment) as an appendix
to paragraph 60.85. The methods are:
Method 1: Sample and Velocity Traverses for Stationary Sources
Method 3: Gas Analyses for C0», Excess Air and Dry Molecular Weight
Method A: Determination of Moisture in Stack Gases
Method 5: Determination of Particulate Emissions from Stationary Sources
Method 6: Determination of SO^ Emissions from Stationary Sources
Method 7: Determination of NO Emissions from Stationary Sources
Method 8: Determination of Sulfuric Acid Mist and S0? from
Stationary Sources.
Method 9: Visual Determination of Opacity of Emissions from
Stationary Sources.
For the purposes of the Regional Air Pollution Study, continuous instrumental
monitoring would seem to be preferable to the wet chemical analyses employed by
the EPA methods. A mobile van containing instrumentation for the determination
of S09, CO, HC, NO and possibly particulates is expected to be used in the later
£- s\
stages of the emission inventory.
Stack sampling is time consuming and expensive; for this reason, it should
be used only to provide a primary calibration of emission factors which are
used in conjunction with more readily accessible data, such as fuel consumption
or processing rates. The most extensive collection of emission factors is con-
tained in EPA's "Compilation of Air Pollutant Emission Factors" (AP-^2) which
is in almost universal use. Nevertheless, emission factors contained there
are averages and vary widely in accuracy. They are rated for estimated accuracy
-------
on a scale ranging from "A" to "E", depending on the number and quality of field
measurements on which they are based.
To insure the accuracy of the RAPS emission inventory, some stack testing
has to be performed. Such testing should include at least one example in each
SCC category; if budgetary constraints permit, a considerable number of important
(8)
sources should be sampled individually (the SRI report suggests a total of 65
stack tests). Tables VIM and IX show the distribution of the major sources by
SCC categories.
By combining similar sources and matching categories with actual sources in
the St. Louis AQCR, the following minimum schedule (Table X) was determined if
at least one installation of each type is to be represented.
The total of 19 stack tests should really be considered as a "Phase 1"
program, to be supplemented by further tests based on inspection and review of
existing faci1i ties.
As discussed above, SCL is the one pollutant for which adequate data can
be obtained with only minimal stack testing, at least for those facilities which
do not have any stack gas cleaning (scrubbing) equipment. At present, none of
the boilers are equipped with such scrubbers; experimental work is being con-
ducted with a "Catox" unit at the Wood River power plant.
Though fuel consumption and process data are potentially capable of providing
quite accurate SO- emission figures, there is a hitch: sampling for sulfur analysis
is not usually done adequately. Practices vary widely; some plants have continuous,
automatic samplers, but these are located at the coal-pile end of the conveyor
system. Since there are usually storage bins in the boiler-house itself, there
is an 8 to 12 hour lag between the sample and the material burned. Most plants
sample only intermittently — once a shift, once a day, even once for each barge.
Fortunately, the sulfur analysis of coal does seem to be fairly constant (about
+_ 10%). A statistical evaluation of the sampling procedures will be performed;
when possible, the time lag will be incorporated in the calculations.
There is a way to get good coal samples, and that is to sample at the pul-
verizing mill, immediately ahead of the injection point into the furnace. If the
sampling becomes a problem, it may be necessary to attempt sampling at that point.
Data for NO will have to be based almost wholly on stack testing. The EPA
/\
emission factors span a range of 3 to 55 pounds of NO per ton of coal,
s\
(8) A Regional Air Pollution Study (RAPS), Stanford Research Institute,
Preliminary Emission Inventory
23
-------
TABLE VI I I
DISTRIBUTION OF LARGE SOURCES IN THE ST. LOUIS AQCR BY SCC CODES
External Combustion Boilers
SCC Code
1-01-002-01
02
03
08
1-02-002-01
02
04
09
004-01
02
Description
External Combustion Boilers, Elect. Gen., Bitum. Coal, Pulv., wet
> 100 x 10° Btu/hr.
Pul v. , dry
Cyclone
Stoked
External Combustion Boilers, Industrial, Bitum. Coal, Pulv., wet
> 100 x 106 Btu/hr. _ . .
Pulv., dry
Stoked
Stoked
Residual Oil
Residual Oil
Number
7
14
4
1
1
3
5
5
4
1
30
-------
. TABLE IX
DISTRIBUTION OF LARGE SOURCES IN THE ST. LOUIS AQCR BY SCC CODES
Process Heaters & Processing Emissions
SCC Code
Description
Number
3-01-023-99 Industrial Process Chemical Mfg.
999-99
H2SO/|-Contact 2
Miscellaneous 2
-03-010-01
Lead Smelter
-04-004-03
-06-001-03
001-04
002-01
999-98
Secondary Metal
Petroleum Ind.
Lead Smelter
1
Process Heater,Oil 2
,Gas *
Fluid Crackers 4
, Ml seellaneous 1
31
-------
TABLE X
MINIMUM TEST SCHEDULE
A. -POWER GENERATION
Equipment
Ext. Combust. Boiler
n n n
n n n
n n n
Fuel
Bitum. Coal
Minimum No.
Firing Mode of Tests
n
n
Stoked
Pulverized
Cyclone
2
2
2
Oil
Suggested
Location
Monsanto
Wood River-Labadie
Sioux-Baldwin
Shell Oil
B. INDUSTRIAL SOURCES
Industry
Chemical Industry
Prim. 6 Sec. Metals
Petroleum
'.ype
Sulfuric Acid
Mi seellaneous
Heaters
Crackers
Others
Minimum No.
of Tests
1
2
2
2
10
Suggested
Location
N.L.
Monsanto-Anlin
St. Joseph
Shel1-Amoco
Shel1-Amoco
Amoco
32
-------
40 to 105 pounds of NO per 103 gallons of oil.
J\
C2. Fuel Consumption and Process Data
From the point of view of sampling methodology, there is no real
difference between emission data based on fuel consumption and data based on pro-
cessing of, as an example, a sulfide ore. In both cases, the hourly weight of
consumed material determines the amount of gaseous discharge. Process data are
more complex, though, since the amount of residual sulfur may be more significant.
and variable then in the case of a simple boiler operation. Of course, if a
recovery operation is part of the process (e.g., a sulfuric acid plant), then
stack sampling may become the more reliable source of data. The pattern of
hourly variations may still have to be determined by process data unless con-
tinuous monitors are available.
An analysis of the NEDS inventory shows that the 62 point sources emitting
in excess of 1000 tons of S02 per year fall into the following categories (Table XI)
TABLE XI
CLASSIFICATION OF S02 SOURCES
SCC Code
1-01
1-02
3-05
3~xx
Category
Boilers, Electric Generation
Boilers, Industrial
Petroleum Processing
Other Industrial
Number
27
19
11
5
Thus, almost 75% C»6) of the 62 sources (including all of the large ones) are
boilers; another M% are concentrated in the petroleum industry.
Carbon monoxide, another combust ion-related pollutant, has quite a different
distribution (Table XI I).
33
-------
TABLE XI I
CLASSIFICATION OF CO SOURCES
SCC Code
1-01
3-01
-03, -04
06
5-01
Category
Boilers, Electric Generation
Chemical Process
Metal. Processing
Petroleum Processing
Incinerators
Number
1
2
6
2
2
Here the largest sources are metal processing (blast furnaces, etc.), petroleum
processing (cat. cracking) and certain chemical processes.
Even particulate emissions are largely related to boilers; almost half of
the emission sources are boilers; another 2$% comes from the mineral industry
(quarries, cement plants, etc.). The breakdown is shown in Table XIII. The
overlap of pollutants from different sources has been indicated in Table VII.
TABLE XIII
CLASSIFICATION OF SOURCES OF PARTICULATES
SCC Code
1-01
3-03
3-05
3-06
Category
Boilers, Power
Metal Industry
Mineral Industries
Petroleum Processing
Number
13
k
7
k
-------
TABLE XIV
CLASSIFICATION OF NO SOURCES
SCC Code
1-01
1-02
3-05
3-06
Category
Boilers, Elect, generation
Boilers, Industrial
Industrial - Cement
Industrial - Petroleum
Number
22
2
1
1
TABLE XV
CLASSIFICATION OF HYDROCARBON SOURCES
SCC Code
1-01
2-01
3-01
3-03
3-06
*»-03
4-02
5-01
Category
Boilers, Elect, generation
Internal Comb., Turbine
Chem. Industry
Primary Metals - Cooking
Petroleum Industry - Processing
- Evaporation
Surface Coating - Evaporation
Municip. Incinerator
Number*
(0)
(0)
(1)
(1)
W
(11)
(6)
(0)
9
2
k
2
kk
19
2
^Bracketed numbers are sources in excess of 1000 tons/year; unbracketed
are sources greater than 100 tons/year.
35
-------
The significance of the high percentage of power boilers lies in the fact
that data pertaining to boiler operations are usually well kept and more readily
available than data about process operations. Since SO- emissions can be calcu-
lated readily from fuel consumption and analysis figures, and the emissions of
other pollutants are closely related to fuel consumption and operating conditions,
the acquisition of hourly fuel consumption data will go a long way toward the
creation of an hourly emission inventory. For this reason, considerable emphasis
will be placed on the acquisition of hourly fuel consumption (and related data),
particularly in the early stages of the RAPS inventory effort.
The actual data obtainable cover a wide range of formats, from computer
printouts of hourly fuel consumption to strip and circular charts, and even
entries in log books. As an example, the logs of the Wood River Power Station
of the Illinois Power Company for 29 April 197^ are attached. Table XVI is a
computer printout, giving actual weight of coal used per hour, as well as information
on boiler efficiency, BTU/lb. of coal etc. for Unit 5. On the other extreme,
Tables XVII and XVIII show the oil and gas usage respectively of units 1, 2, and
3. Here data are available only per eight-hour shift and consist of single meter
readings.
C3. Operating Data
This group includes all those point emission sources which are
either minor (emitting less than 1000 tons/year) or for which no detailed hourly
data are available. Here it is necessary to fall back on the annual data recorded
in the NEDS inventory or the corresponding inventory of the local enforcement
/q\
agency. An approach similar to Roberts v;7/ will be used to approximate the
emission patterns by determining the relative amounts of fuel used for space
heating and for process purposes and allocating each. Space heating require-
ments are distributed in accordance with degree-days (deviation of mean daily
temperature from 65°F in degrees) while process loads are determined from
appropriate month, day and shift factors. From these data, both the sulfur
and heat emissions can be calculated.
Roberts assumes that when the temperature, T, is between -10 and 55°F, there
is a linear relationship for the space-heating thermal load L . This Is expressed as
(9) For detailed description of such a system, see: Roberts, J. J., et al.,
Chicago Air Pollution Systems Analysis Progress, Argonne National Lab.,
ANL/ES-CE007.
36
-------
HWJI/OJt-tf H 1MV-W
VOHWM 3410
OWO3
;$ uvs
UH/SMOl »n 1VOD
tWO M01J
TM-«i
•or% AX3OMJ3
% V svo srru
4. d»«i ino HIV ss H v
J. d«3JL ino HIV vs H v
J. dW3i Nl HIV 85 H V
1.
M3V8
HMx/rue TJ -H -anal
j. H3iooa -*u 110 •euni
J. dW31 'H 11 10H
i. =W31 'H "a O10O
ISd SS3Hd 'H '« CTIO3
J. YJW31 wtf3_lS KIIVH
ISd •S3JU
HIS
ISd 'H 'N33 "d ~\
SMVAW >130 "d T
ISd 'H N33 'd -H
SdlVV N30 'd "H
SHVAW SSOU9
MW SSOB3 N30 d 1
Mtt 13M
Mn SSOHO
-T ° « V -V *•"* ^~ *~ O
3 5831858 R
-
•« t- o
o O f-
I 8
CO O\ •" O »*- (*•» j»
o o o o o o oo o c
o o c o o o o
o o " o o o o
or >_
uj 3
§ Q "»
Ell
« J3
Q 2j
P 9 .
Q
O* JN
O O
0 0
" «
g I?
«»•* f\t
.3- -»
t: y
rv ni
O_ —
UN UN
c" "K
R •£
T— r—
K. -5^
ft «
So
0 0
gr m
to R
S "8
o 3
(V CM
.a- ^»
0 O*
s s-
f 8
.=• O
o *-
cu
Of CM
O UN
3 fr,
*_ ON
P- CO
CN —
,» UN
O CM
Nt VJ
OX- 0,
S S
0 0
O 0
5? &
UN UN
ON O
S S
y <±)
UN UN
cxj ru
CM_ fM_
01 01
UN V3
UN UN
UN UN
OJ O3
S -£
S <£
c? !ft
»* UN
0 O
CU —
3 s
Q* H
Ov ^K
Ol ITN
o o1
OJ (M
UN fe
(\J (M
•» ON
UN UN
OJ (M
O —
ni m
fc s;
C —
c a
CT> 0.
O O
O O
o "8
\o *•—
o o
*N u\
C 0
_» jy
OJ fM
a y
" ^
f— i —
^
y
\u
nj
ON
CM
UN
0
rv.
UN
CO
is
*
o
o
f— .
O
<5
R
S
s
I
"r*i
g
„
CM
iTN
O
0
*v
«, ON
0 0
0 0
£ 3
*N *>
S. £
trv u^
••
UN UN UN UN UN UN UN UN iO UN WN IT
wfu n*^CMcu CUCM (M ry rn »r
Jjr_,
^^s^v-oSS^^^v..
^S^oS^Sir.^oi;
fgCU CM niCVCMCMOJCMOJCMOi
j— t*\ 1,*% O — JT ^ ^p NU jy cy 1-
iTN |T\ VN UN UN UN UN WN UN UN ITv U"
UN L-N "f^ ^W WNvO IP^VOO UNCMtT
,— O •" f-
£2 fc_ *^1 r— ^S ^i 15 SN ^o\ i^ f> L-
^ CO ^»" V^ f"~ CO ON CO CO CN ITN tJ
O O r— C?N o O O O Q O -t> -=r
Csr-vONCNO ONONCNO\3>CN^
BN UN «N «N «J UNUNUNUNUNUNU"
oooooooooooo
*~ CN Ol CO CO O_ CO *W Ol OJ C* O.
tB~^tBtSi:Si:B
^•OQCM — r— j»t— — Of~ "~
O C- O t'% O Ov IK O — O r— CJ
o *^NO I/NO O^g-^r«->jrC5 c\.
0*0*0 Sf"S £J S 5' *5 S 3 -5
SgSgRS5PS£5:g
rw f'j cj r\jryr\iryc\lojrij
e
s
HOURS
t
1
cc
1'
vt ~
s z
S M
X
^-
ca *"
tn
2? o
o
O
T
CO »T
^ CD
cc tr
X X
— (U
VO £
S X
CN 0
•3
3 3
M/t I JON
MILHON
0 °
UN rrl
sl.
r- O
" -5
in u>
'•^ ^
§§
<«^ j
?. If
' 4? '
O c\
^- f*
s
cc <
W M
*" Ul
X 2
P 55
iO e
fM 1*1
* *fw ^
-3- OO
-f 15 TOO
OTO-1
C*"! CO
cm
0 0
X X
O JT
cv *°
cc a:
UN 0
— O
£ -
e
3 °
0 w
I S
< -J
t- <
f- U
S" §)
CM CO
- g
UN -1
CO S3
^ a
CM ol
*" 1
(A
.<•? !;
•• i
§i
« *
r •>
^ »-
S r
8*
C Z)
h-
3
5
O
t-
T
en -
"~ I!
2 S
cc sc
X =
•*. ",
C o
tt or
X X
o
e cc
(- H-
O
u^* c\
a: cr
X X
o c
It.
_J •
0 u
S "
S 5
0 0
e I
R
j «
}r
in H
w <
o\ c
u>
0
z
X
UJ
—I
CO
<
37
-------
TABLE XVI I.
ILLINOIS POWER COMPANY
WOOD RIVER POWER STATION
BOILERS 1, 2, 3 DATA
2 9 1974
DATE.
COAL SCALE
Shift 12 MM - 8 AM
1. Kd0. 01 8 AM
2, Rcfj. of 12 MN
3. Different*
il.lft 8 AM - 4 PM
4. Rcig. ert 4 PM
5. Rrfg, at 8 AM
6. Difference
,hifr 4 PM - 12MN
7. R'Jf . «f 12 fc'.N
5. t;"a ^
GAS
M«t«r
Reeding
Start
M«ter
Reading
End
FEEDWUER
Meter
Reading
Start
77 ?
Meter
Reading
End
$(, J
Meter
Reading
Start
STEAM
Meter
Reading
End
06
4 W
Meter
Difference
HEATING STEAM
12 MN
8 AM
8 AM
4 PM
4 PM
12 MN
SLOWDOWN UNITS 1, 2, & 3
12 MN
6AM
8 AM
4 PM
-
4 PM
12 MN
38
-------
TABLE XVIII.
ooooooooooooooooo
3TOT1
ooooooooooo o o o o o
O ro o CO o o cp n o cy) o ro o o} o
cibooc> ooo*~r~rN
-------
Then the total thermal load is
L - LS + LP,
p
where L , the process load, is determined from the appropriate month, day,
C 0
and shift factors. The amount of load due to coal, L , and due to oil, L ,
is then determined, and the SCK emission due to each source is calculated as
follows:
. / .. v L (therms/hr) x 105(Btu/therm)
L uons/nrj - 12oOO(Btu/lb) x 2000(lb/ton)
SO^lb/hr) = C(tons/hr) x 38 x %SC;
0(. al/hr) = L (therms) x 105(Btu/therm)
l8000(Btu/lb) x 8000(lb/kgal)
S0°(lb/hr) = 0(kgal/hr) x 157.0 x *S°.
Thus, the total S0_ emission is
so2 = so^ + so°.
When the ambient temperature is such that a dual-fuel interruptible plant
is probably receiving natural gas, the amount of SO- produced is correspondingly
reduced.
To facilitate data storage according to a uniform and consistent format,
each plant is assumed to have four stacks. For plants having less than four
stacks, zeros are filled in for nonexistent stacks. The following parameters
are associated with each stack:
1. SO^ emission in pounds per hour
2. Heat emission in therms per hour.
These parameters are determined by weighting the total SO- and heat emissions
for the plant by the percentage emitted from each stack. The heat emission, H,
is assumed to be 15% of the thermal input.
-------
V. Handling of Emission Data
As indicated in Section IV-C2, emission (or emission related) data will
be provided in many different forms, ranging from computer printouts to strip
or circular charts. The raw data wi11 have to be read off the original records
and tabulated in appropriate form before entry into the RAPS computer bank.
The format for the RAPS emission data storage has not yet been finalized.
A data handling system (System 2000) will be used which is capable of storing
data elements of variable length in repeating groups. The repeating groups define
the structure for storing multiple sets of data values and link the hierarchical
levels.
Data preparation forms will be designed to aid the data clerk in the struc-
turing of the data and to make it easier to use the correct syntax. It will not
be possible, however, to depend only on well-designed data preparation forms for
data quality because the content as well as the form of the data must be verified.
Data verification can be carried out in part by the data management system,
probably using a preliminary storage file which can be verified, proofread and
corrected before the data manager decides that it is accurate enough to merge
into the main file.
A detailed instruction sheet has to be prepared for each data sheet for
the guidance of the data clerk. This sheet has to stipulate the units (if not
indicated on the original record) and specify the manipulations, if any, which
have to be performed to obtain hourly data which can be fed into the computerized
RAPS inventory. In order to avoid human error as far as possible, only a minimum
of handling will be carried out. For example, data will be recorded in whatever
unit it is supplied and the units made part of the record. Transformation into
standardized units can then be performed by the retrieval program to meet the
specific needs of the user.
For most actual sources, the values stored will be consumption or other
source data, rather than measured values of emission. The format will accom-
modate emission or consumption data. For those sources for which there are no
direct emission data, emissions must be calculated using emission factors or
models applied to the stored data. The emission inventory software system will
-------
be capable of assessing the consumption data element, refer to the appropriate
code, look up the emission factor of model, and compute the emission values for
each specified set of pollutants.
VI. RAPS Inventory Acquisition Schedule
The acquisition of the RAPS emission inventory comprises the following
elements: 1) Survey and arrangement for data collection from measured sources;
2) Data acquisition and processing; 3) Acquisition of data from smaller
sources; 4) Source testing.
The survey of sources from which hourly data should be obtained should be
accomplished in about three months; this will include detailed arrangements which
wi11 spel1 out:
which sources will be observed.
what data will be forthcoming.
the necessary factors to transform available data into mass emission
uni ts.
the mechanism of data collection.
During the next three-month period, data will be collected and their
transformation into machine readable data accomplished. Data collection will
continue for at least a year, possibly longer.
Data from lesser sources will be collected concurrently, beginning with
about month 6 or 7. Appropriate algorithms will be designed to provide
hourly emission values. ''
A source testing program will be set up (jjnder another task order) to pro-
vide verification of the emission factors and other assumptions used in the
program. This should be an on-going effort, utilizing a mobile test unit and
providing calibration data on 20 to 60 sources. The more sources are tested,
the more reliable the inventory will become; this effort is limited mostly by
budgetary considerations. A minimum program was outlined in Section C-l (p. 27).
This schedule is shown graphically in Figure 3.
-------
Months
1 . Survey & arrangement
for data collection
from measured sources.
2. Data acquisition and
process i ng.
3- Acquisition of data
from smaller sources.
4. Source testing.
1, 2, 3
1 k\ 5I 6
^ 7, 8, 9
10, 11, 12
1
1
1 13I 1Z»I 15
July Oct. Jan.'75 April
FIGURE 3. RAPS INVENTORY SCHEDULE
July
Acqufs ftfon of data (Step 2) wflf start about I October, Iy7^, using data
obtained from Union Electric Company to check out the data entry system. By
1 November data from Illinois Power Co. will be added. Additional data,
including all major sources (but limited to S0_ only) will be included
gradually. By 1 January 1975 the inventory of hourly S02 emissions from all
major sources should be operational. Data from lesser sources will then be
incorporated.
All incoming data will be entered on coding forms and visually checked for
discrepancies. The tabulated data will then be transferred onto machine readable
cards for delivery to NADB. Cards will be sent to NADB once a month.
VII. Summary and Conclusions
The history and purpose of the Regional Air Pollution Study in St. Louis
was reviewed from the point of view of emissions of air pollutants and their
inventories. Based on NEDS data on the size and distribution of the principal
sources of air pollution and on a Weighted Sensitivity Analysis supplied by
NADB, an emission inventory program methodology was designed to provide hourly
data on criteria pollutants, with initial emphasis on sulfur dioxide. The
criteria for choices were the estimated requirements of the most critical users
of the data, the investigators working on dispersion model verification.
-------
The methodology envisages a two-level approach: the measurement of
hourly emissions or emission related data for the principal sources, defined
as those emitting in excess of 1000 tons of pollutants per year, and a simu-
lation of hourly emissions for smaller sources, based on yearly outputs and
appropriate information on the consumption or production cycle.
The successful accomplishment of these goals should provide an emission
inventory of a much higher accuracy than has heretofore been available.
-------
TECHNICAL REPORT DATA
{Please read Instructions on the reverse before completing)
1. REPORT NO.
2.
3. RECIPIENT'S ACCESSIOWNO.
4. TITLE AND SUBTITLE
Regional Air Pollution Study Point Source Methodology
and Inventory
5. REPORT DATE
October 1974
6. PERFORMING ORGANIZATION CODE
7. AUTHOH(S)
Fred E. Littman
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADSRESS
Science Center
Rockwell International
1049 Camino Dos Rios
Thousand Oaks, California 91360
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-1081
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final-2/74-10/74
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
An emission inventory constitutes the starting point for any attempt to
control emissions to the atmosphere. As long as such controls deal with average
yearly concentrations, inventories giving total annual emissions of the various
sources of pollutants are sufficient. The Regional Air Pollution Study has,
however, as its first goal the validation of atmospheric dispersion models,
which attempt to predict ambient pollutant concentrations on an hourly basis.
Therefore, emission values derived from total annual emissions are largely in-
adequate, and the RAPS emission inventory was conceived to provide the needed
time resolution and accuracy by measuring and recording hourly emissions (or
parameters directly related to hourly emissions) and/or individualized hourly
estimates derived for the principal sources of pollution. Thus, the emission
inventory for the Regional Air Pollution Study (RAPS) at St. Louis is distinguished
from existing emission inventories by two factors: its time and space resolution
and its accuracy.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
SMSA
Emission Inventory
Point Sources
Pollutants
SCC
AQCR
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
45
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |