EPA-450/2-74-023
NOVEMBER 1974
EPA RESPONSE TO REMAND
ORDERED BY U.S. COURT OF APPEALS
FOR THE DISTRICT OF COLUMBIA
IN PORTLAND CEMENT
ASSOCIATION V. RUCKELSHAUS
(486 F. 2d 375, JUNE 29, 1973)
3, -^ ^- UJ
V**r
-------
-------
EPA-450/2-74-023
EPA RESPONSE TO REMAND
ORDERED BY U.S. COURT OF APPEALS
FOR THE DISTRICT OF COLUMBIA
IN PORTLAND CEMENT
ASSOCIATION V. RUCKELSHAUS
(486 F. 2d 375, JUNE 29, 1973)
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
November 1974
-------
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.
Publication No. EPA-450/2-74-023
11
-------
EPA Response to Remand Ordered by
U. S. Court of Appeals for the
District of Columbia in Portland Cement Association
v. Ruckelshaus (486 F.2d 375, June 29, 1973)
On June 29, 1973, the U. S. Court of Appeals for the District
of Columbia remanded to EPA the new source performance standards for
Portland cement plants (40 C.F.R. 60.60 et seq.) promulgated by EPA
under §111 of the Clean Air Act. EPA has reviewed its standards
for these plants pursuant to the remand and has concluded that the
standards other than the opacity standard should not be revised.
The bases for these standards are set forth in full below.
This description of the bases for the standards responds to concerns
expressed by the Court in its opinion explaining the bases for the
remand. In order to assure that all of the Court's concerns are
addressed, Appendix I is provided which lists the questions raised
by the Court and EPA's responses to these questions.
-------
EPA sought and received comments from the public and
the affected industry in developing the challenged standards
prior to proposal and during the proposal period prior to
promulgation. Criticisms of these standards were again
raised by petitioners and fntervenors in connection with
the legal challenge to these standards and were also carefully
reviewed. EPA has, of course, carefully considered the
comments as well as the order of the U.S. Court of Appeals
for the District of Columbia Circuit in responding to this
remand. However, to further insure that the affected industries
had an opportunity to comment on EPA's response to the remand
and to insure that a reviewing court would have the benefit
of EPA's response to such comments, a proposed remand response
was submitted for comment to the petitioner and intervenors
in this case.
At the request of Appalachian Power Co. and Essex Chemical
Corporation, petitioners in another case in which new source
performance standards were remanded, Essex Chemical Corp. v.
EPA and Appalachian Power Co. v. EPA, 486 F.2d 427 (C.A.D.C. 1973),
copies of the remand response were also submitted to these
companies for comment. The remanded issues in the latter
decision included an issue raised in this case, i.e. reliability
of opacity readings, and EPA believed it appropriate to obtain
-------
the comments of these companies with respect to this issue.
Comments were received from both companies.
The comments received were carefully reviewed and have
persuaded the agency that no revision should be made to the
standards other than the opacity standard. Because it was
not found necessary or desirable to revise the standards in
any other respect as a result of the comments received, the
proposed response as submitted to petitioners is set forth
as a final response. (Minor corrections have been made to the
text of the response on pages 8, 26, and 39.) The comments
received with respect to issues other than opacity and EPA's
response to these comments are set forth in Appendix II.
Because substantial revisions were deemed necessary to EPA's
draft remand response on the issue of opacity, EPA's
discussion of opacity and justification for the revised
opacity standard are set forth in full in Appendix III of
this response.
Copies of documents referenced herein are available from
Emission Standards and Engineering Division, Environmental
Protection Agency, Research Triangle Park, North Carolina 27711,
Attention: Mr. Don R. Goodwin.
Actina Administrator
-------
-------
CONTENTS
Section Page
HISTORY OF CONTROL TECHNIQUES 1
COMPARISON OF WET PROCESS AND DRY PROCESS EMISSION CONTROLS 4
TESTING 7
CONTROL EQUIPMENT VENDOR STATEMENTS 15
OPACITY STANDARD 16
COSTS 29
ADVERSE ENVIRONMENTAL IMPACTS 30
SUMMARY. 34
APPENDICES
I. EPA RESPONSES TO SPECIFIC ISSUES RAISED BY THE COURT 37
REFERENCES FOR BODY OF RESPONSE AND APPENDIX I 59
II. EPA RESPONSE TO PETITIONERS' COMMENTS REGARDING EPA
RESPONSE TO REMAND 61
REFERENCES FOR APPENDIX II 82
III. PORTLAND CEMENT REMAND RESPONSE 83
PART A: OPACITY STANDARD 85
REFERENCES FOR APPENDIX III, PART A 124
PART B: PREAMBLE AND AMENDMENTS TO REGULATIONS 127
TECHNICAL REPORT DATA SHEET 152
vn
-------
-------
HISTORY OF CONTROL TECHNIQUES
Prior to December 1970 vrhen the Clean Air Act was
amended, EPA and its predecessor agencies had acquired
extensive knowledge of portland cement operations. A
study of the industry had been made by the U. S. Depart-
ment of Health, Education, and Welfare and in 1967 a re-
port was published entitled, "Atmospheric Emissions from
the Manufacture of Portland Cement", PHS Publication No.
I/
999-AP-17. Engineers from EPA's predecessor agencies had
inspected numerous plants and discussed control technology
with vendors and knowledgeable individuals within the
cement industry and had provided technical assistance to
State and local air pollution control agencies regarding
air pollution problems associated with cement plants.
Further information regarding cement plants was developed
by EPA engineers for publications concerning cement plant
control technology, e.g., AP-51, "Control Techniques for
U
Particulate Air Pollutants", 1969. In addition, a report
dated September 30, 1970, relating to the establishment of
national emission standards for portland cement manufacturing
plants had been prepared under contract by the Research
3/
Triangle Institute and Pedco-Environmental, Incorporated.—
These references were used to obtain general knowledge
of cement plant operation, control and nature and quantity
1
-------
of emissions. They were also used for their enumeration
of cement plants in the United States,
It was recognized that at portland cement plants,
there are several sources of dust,- with some lucre difficult
and more costly to control than others. Kilns were
recognized as the most difficult and costly to control.
Exhaust gases from kilns are hot, contain heavy concentra-
tions of very fine dust, and often large fractions of water
vapor. Clinker coolers which are used»to cool the product
from the kiln were considered easier to control because
they contain more coarse particulates and less moisture.
Other equipment in cement plants was generally operated at
ambient temperature, and therefore easier to control. For
instance, grinders and conveyors often produced heavy con-
centrations of very fine particulates, but since this dust
is handled at room temperature, it can be collected in
fabric filters or baghouses which the cement industry acknow-
ledged to be extremely effective in controlling dust.
For those sources operated at ambient temperature,
there has never been an argument with the cement industry
as to how they should be controlled. The industry acknow-
ledged that baghouses were the most effective devices and
since they collected a larger fraction of their product—
cement—than any other available device, the industry
-------
readily used such devices. In the case of clinker coolers,
many cement plants had used only centrifugal collectors
to remove the coarse particulate from exhaust gases. However,
in 1970, it was observed that several plants were installing
more effective devices, usually electrostatic precipitators
or baghouses. EPA observations indicated that both electro-
static precipitators and baghouses appeared to be effective
in controlling particulate emissions.
In the case of cement kilns, by 1970 industry had
generally supplemented centrifugal collectors with electro-
static precipitators or baghouses. (Centrifugal collectors
are not capable of collecting the very fine particulates
that are released from the kilns and would not therefore
be acceptable in most communities in the United States.)
There are many more kilns in the United States equipped
with precipitators than with baghouses but baghouses have
been applied to both wet and dry process kilns. Precipitators
can be designed to achieve extremely high efficiencies but the
higher the efficiency the more costly the unit.
Because precipitators can be purchased which will operate
over a range of efficiencies, plant operators that wished
to control emissions but not necessarily to the high
efficiency that baghouses achieve generally selected lower
efficiency, and thus lower cost, precipitators, As public
-------
pressure and local regulations caused plant operators
to install higher efficiency equipment, more plant
operators chose to install baghouses. The best controlled
kilns observed and -tested by EPA were controlled by baghouses.
Later data which will be discussed below show that precipitator
controlled plants can also meet the standards.
COMPARISON OF WET PROCESS AND DRY PROCESS EMISSION CONTROLS
Both precipitators and baghouses are used on wet
process as well as dry process kilns. The only difference
between controlling emissions from the two processes is that the
gases from the wet process kiln contain 25-40 percent moisture
by volume whereas those from a dry process kiln contain less
than 5 percent moisture. At one time, plant operators feared
that this additional moisture would clog bags and cause
shutdowns. It was also feared that baghouses could not be
used on wet process plants in cold climates because of con-
densation. This apprehension is why there are fewer bag-
houses on wet process plants than on dry process plants.
However, at actual installations these problems have been
resolved by insulating and if necessary heating the exhaust
emissions to keep the moisture from condensing. There are
presently wet process plants with baghouses operating
successfully in such cold and moist climates as Thomaston,
-------
Maine (Dragon Cement Company), Lake Oswego, Oregon
(Oregon Portland Cement Company), and Northampton,
Pennsylvania (Universal Atlas Company).
The existence of the greater moisture content in -wet
process gases means that a greater volume of emissions
must be controlled at such plants. Thus, controlling emis-
sions from wet process kilns will cost more, both for heat-
ing and insulation as well as treating the additional
volume of gases, than will controlling emissions from dry
process plants. The cost of insulating a wet process bag-
house or precipitator would depend on the external tempera-
ture range. Nevertheless, the differential cost of insulating
a wet as opposed to a dry process plant is not significant
compared to other dust control costs.
The greater volume of gas from a wet process plant re-
quires an equivalent increase in baghouse or precipitator
size. For a 10,000 bbl per day plant, the cost of a wet
process baghouse would be approximately $890,000 as opposed
to approximately $730,000 for a dry process baghouse. Corre-
sponding costs for an electrostatic precipitator are esti-
mated as $750,000 and $650,000 for wet and dry process plants,
i/
respectively.~
However, these additional costs for a wet process plant
are largely offset by the fact that a wet process plant
does not need to control ,-iaissions from the grinding
-------
process before the raw material is put in the kilns. This
emission control equipment for a 10,000 bbl a day dry process
4/
plant costs approximately $90,000 .
Achieving the same mass particulate limit per ton of
cement produced is more difficult for a wet process plant
than a dry process plant because of the greater volume of
emissions and because of inherent limitations of dust
collectors. Baghouses and precipitators can only reduce
concentrations of particulates of gas emissions down to
certain levels. Since wet process kilns emit greater
volumes of gas than dry process kilns of equivalent size,
a given concentration of particulates emitted will result in
a greater total amount of particulates being emitted from a
wet process kiln pej: unit of cement produced.
For example, if a baghouse or precipitator can reduce
concentrations of particulates down to a certain number of
grains per actual cubic foot of gas emitted, the total emis-
sions of particulates from a wet process plant per unit of
cement produced will be 25-40 percent higher than the total
emissions of particulates from a dry process plant because
of the 25--40 percent greater volume of gases which wet process
plants emit. Thus, the highest level of particulate control,
expressed in pounds per ton of dry feed, which a wet process
plant can achieve would result in about 25-40 percent more
emissions than that from a dry process plant. As will be
discussed later, EPA took this into account in. setting the
standard.
6
-------
TESTING
While EPA was knowledgeable in 1970 regarding applica-
ble control devices for cement plants, they possessed very
little source test information with regard to the emission
of participates from cement plants. Host of the test data had
been collected using the American Society of Mechanical
Engineers (ASME) Power Test Code. This is considerably less
specific and as a result different from the EPA particulate
sampling method which had been utilized to a large degree
a_/
with incinerators and combustion sources. It was recog-
nized that an appreciable effort would be required to survey
and select plants demonstrating best control technology
and to evaluate the emissions from these sources using
the EPA particulate sampling method.
Early in 1971, EPA engineers contacted the Portland
Cement Association, the Bureau of Mines, and knowledgeable
vendors of air pollution control equipment. They procured
pertinent information regarding what appeared to be the
better controlled plants and arranged inspection tours to
evaluate the apparent efficiency of the air pollution control
equipment at various plants. During these inspections, EPA
engineers observed that those plants—both dry and wet
process—equipped with baghouses appeared to operate with
less particulate emissions than did those plants at which
a_yThe ASME code is broadly defined such that it covers several
techniques including even the EPA method with and without
impingers. However, most ASME tests have been conducted
with a different filter and under different conditions than
EPA Method 5 such that results are not normally comparable
to the EPA method.
-------
the kilns were equipped with precipitators. Nevertheless, some
precipitator-equipped kilns also exhibited what appeared to be
good control of particulates. The basis for the initial judgment
of the effectiveness of control was the amount of visible emissions
emanating from stacks and leaks in ductwork, etc., as well as
observation of the quantity of dust deposited on surfaces of roofs,
floors, equipment, etc., within the plant. Four plants were inspected
where kilns equipped with baghouses had minimum dust emissions and
appeared to be operated in an extremely clean manner. These four
plants were operated by Ideal Cement at Tijeras, New Mexico; Dragon
Cement at Northampton, Pennsylvania; Oregon Portland Cement at Lake
Oswego, Oregon; and Coplay Cement Company at Nazareth, Pennsylvania.
There were four plants where kilns equipped with precipitators also
appeared to be well operated and with minimum dust emissions. These
kilns were at plants operated by the Ideal Cement Company in Seattle,
Washington; the Maule Industries plant at Hialeah, Florida; the Lone
Star Cement plant at Greencastle, Indiana; and the Ideal Cement plant
at Castle Hayne, North Carolina. Cement plants equipped with both
types of control devices were observed which exhibited heavy concentrations
of dust in and around the buildings and in the gases being discharged
from stacks serving the kilns. The latter instances appeared to be the
result of inadequate maintenance and/or dust collection equipment
deficiencies.
8
-------
As a result of plant visits, EPA selected plants for its
testing program of kilns, clinker coolers, and finish mills.
Five kiln installations were tested as well as three clinker
coolers, one rav/ mill, and three finish mills. These included
a wet process kiln with a baghouse at Harleyville, South
Carolina which appeared to be so poorly maintained that
results would be questionsable. Nevertheless, with the time
constraints in the Clean Air Act, it was deemed necessary
to test these installations to gather information for set-
ting our standards. It is emphasized that the selection of
test sites was a compromise between what appeared to be best
available control and the ability to conduct representative
source tests. Four plants had to be discounted because the
kiln control devices were not constructed in a manner that
would allow them to be tested. These were the Ideal Cement
Company in Tijeras, New Mexico; the Universal Atlas plant in
Northhampton, Pennsylvania; the Lone Star Cement Company in
Greencastle, Indiana; and the Coplay Cement Company in Nazareth,
Pennsylvania. These included open discharge baghouses and
precipitators in which the configuration of the ductwork
made testing impossible. The kilns that were tested before
proposal were operated by Dragon Cement Company in Northampton,
Pennsylvania (dry process with baghouse, Emission Testing Report
57
ETB Test Number 71-MM-05); by Ideal Cement Company in
Seattle, Washington (wet process with precipitator, ETB Test
£/
Number 71-MM-03); by Maule Industries in Hialeah, Florida
y
(wet process with precipitator, ETB Test Number 71-MM-01);
-------
and by Giant Portland Cenent in Harleyville, South Carolina
8/
(wet process v;ith baghouse, ET3 Test Dumber 7l-HM-07)T
The test data .obtained from these kilns include the
amount of particulates collected in both the front half and
the back half of the EPA sampling train. At the time EPA
proposed these standards, it was anticipated that compliance
with the standards would be determined by measuring particu-
lates captured in the total train (front half plus the back
half of the sampling trciin) . However, the standard as pro-
mulgated requires that compliance be determined by measuring
only the particulates collected in the front half of the
train. This change in sampling techniques required for
compliance in no way brings the test data into question since
EPA obtained data on the catch in the front half of the train.
The only effect of the change is that back half test results
are no longer used for compliance purposes. All particulate
emission data referred to in this Notice refers to the front
half catch unless otherwise noted.
The test results for the four plants showed the follow-
ing averages:
Dragon Cement Company 0,07 Ibs/ton
Ideal Cement Company 0,88 Ibs/ton
10
-------
Maule Industries 1.15 Ibs/ton
Giant Portland Cement 0.52 Ibs/ton
As these results show, the Dragon Cement dry process
kiln equinped with a baghouse was operating well below the
k/
standard but none of the other three plants, all of which
were wet process operations, met the eventual limit of .3
pound of particulate per ton of feed. However, the control
devices on these plants did not appear to be operating at
optimum effectiveness during the test period. For example,
the baghouse at the Giant Portland Cement Company plant in
Harleyville, South Carolina was actually a 24-compartment
unit with 12 separate stacks. EPA tested each of the 12
stacks and averaged the results'. P.esults showed emissions
well in excess of the proposed standard; however, visible
emissions from some stacks looked better than others indicating
that the bags in some of the compartments were in better condi-
tion than in others. There was considerable evidence of poor
maintenance in that dust was observed on the roof of the
baghouse and on other horizontal surfaces in proximity to
the baghouse discharges.
Taking into account that the only difference between
the emissions of wet and dry process plants is the higher
moisture content of the wet process emissions, EPA estimated
b_7Subsequent tothe EPA tests, the plant operators conducted
comparison tests using a somewhat different test ^rocedure.
Results were similar, averaging 7 percent lower than EPA
measurements. The company, Martin Marietta Cement, has stated
that they believe the two test procedures give the same re-
sults and requested that EPA certify its particulate test meth
equivalent to Method 5. No action was taken on the request
because EPA had not established guidelines for determining
equivalency and felt that more comparative testing was needed
before the method could be deemed equivalent.
11
-------
from the dry process test that wet process plants could
achieve the standard by a reasonable margin. The Dragon
test shoved that particulate concentrations in gases could
be controlled to approximately .005 grains per actual
cubic foot of air emitted. If the volume of air emitted
from a wet process kiln were 40% greater per ton of feed
due to greater moisture content, the same concentration,
.005 grains per actual cubic foot of air emitted, would
result in 40% more pounds of particulate per ton of feed
or, based on the Dragon test, would be about 0.1 pounds
per ton of feed. This would be well below the standard.
Although it was EPA's judgment based on this estimate
and statements of control equipment vendors that wet as
well as dry process plants could be controlled to meet the
standards, EPA decided to conduct another test on a wet
process plant. A wet process kiln equipped with a baghouse
at the Oregon Portland Cement installation at Lake Oswego,
97
Oregon, was selected and a test was run in August of 1971
shortly after the standards were proposed. During the test,
it was observed that there were visible emissions from the
stack at somewhat greater opacities than had been observed
during the initial inspection. The process was upset during
testing and EPA engineers believe that bags were broken. Subse-
quent evaluation of the data showed that emissions were con-
siderably greater than emissions from
12
-------
the Dragon Cement Company—averaging .60 pound per ton of
feed. Discussions v:ere held with company personnel and it
was agreed that efforts would be made to repair the system
such that it would operate in a more representative manner.
Such repairs were made although it must be emphasized that
the repairs were still not of the type that would be con-
sidered good operating and maintenance practice. In most
instances, the broken bags were merely blocked or tied off
in the system such that they no longer served as filter
media. The EPA engineer supervising the test estimated
15 percent of the bags were tied off during the test. This
effectively removed some of the fabric filter area from the
system and meant that the remaining bags were required to
perform the task that the entire baghouse would perform
under well maintained conditions. This decreased filter
area tended to cause greater wear on the bags and could have
resulted in greater particulate emissions. Nevertheless,
tests on this plant were run in October of 1971, which
showed average emissions of 0.28 Ibs/ton (Emission Testing
Report ETB Test Number 71-MM-15)?/ Thus, at the time EPA
promulgated the standard both a wet and a dry process plant
had shown in tests that they could meet the standard.
13
-------
Since the standard was promulgated, summary test data
on other Portland cement plant kilns have been obtained by
EPA from a State air pollution control agency. One test
was conducted by a private testing firm commissioned by the
plant operator and the other test was conducted by plant
personnel. -The data may be obtained from EPA. They show
the following:
Test Date Name of Company Emissions in
Tbs/ton of dry feed
4/18-25/72 Lehigh Portland Cement 0.125
Company, Union Bridge,
Md., dry process, electrical
precipitator
7/11-14/73 Marquette Cement Manufacturing 0.087
Company, Hagerstown, Md., dry
process, electrical precipitator — -
Thus, two additional portland cement plant kilns appear
capable of meeting the standard. To date, therefore, both
a wet and dry process plant using a baghouse have been shown
by EPA tests to have achieved the standard and data submitted
to a State air pollution control agency indicates that two
additional dry process plants using precipitators have
achieved the standard.
The EPA tests which showed emissions above the standard
do not show that those sources could not meet the standard.
They only showed that they were not meeting it when tested,
whether due to inadequate maintenance, inadequate control
equipment, temporary upset, or some other reason.
14
-------
CONTROL EQUIPMENT VENDOR STATEMENTS
Control equipment vendors have stated that their
equipment will permit the attainment of the standards.
Mr, T, A, Brown of the Koppers Company in a letter to EPA
dated September 30, 1971, said that he saw no problem meet-
ing the standard with electrostatic precipitators.
Wheelabrator-Frye in a letter dated April 7, 1972, said it
would guarantee Medusa Portland Cement Company that its
precipitators and baghouses would limit emissions below the
standard, Buell Environ-tech in a letter dated May 10,
1972, and Kaiser Engineers in a letter dated May 10, 1972,
said their precipitators would limit emissions from wet
process plants to the standard.
EPA believes that these statements are useful supple-
mentary evidence confirming that the standards can be met.
While considerations of self interest could lead a control
equipment vendor to advocate a stringent standard, it would
be contrary to the self-interest of such a vendor to repre^-
sent that his equipment could meet the standard if it could
not. It would be especially contrary to the" self-interest
of the vendor to represent or guarantee to a potential
purchaser of his equipment that it could meet the standard
since the vendor would be liable for failure to satisfy his
obligations. Furthermore, all of these vendors have good
15
-------
reputations and they would not have achieved nor could
they maintain their reputations if they misrepresented
the capability of their equipment. Because of these
factors and because the vendor statements supplement the
source tests rather than serve as a sole basis for the
standard, EPA believes that the expenditure of time and
resources needed to independently verify the vendors'
statements is not warranted.
One of the factors which several of the vendors and
many cement plant operators expressed doubt about was
whether the standard could be met at all times including
periods of startup, shutdown, and upset. EPA has promul-
gated regulations which make it clear that EPA's mass emis-
sion limits apply only during performance tests rather than
at all times. The standards do not apply during periods of
startup, shutdown, and malfunction. (38 F.R. 25864,
October 15, 1973).
OPACITY STANDARD
EPA promulgated opacity limits as well as mass emis-
sion limits because opacity limits provide an effective and
easily utilized means of determining relative mass emissions
as well as wether control equipment is being properly main-
tained and operated. EPA personnel observed during testing
that two plants which met the mass emission standard did
not exceed 10 percent opacity during emission test periods.
16
-------
An additional cement kiln was observed that had no visible
emissions, but it could not be tested for mass emissions
due to physical limitations. A summary of mass particulate
emissions and observed opacities for cement kilns investi-
gated by EPA is shov;n in Table 1.
17
-------
Table 1. PARTICULATE MATTER AND OPACITY OF EMISSIONS
FROM CEMENT KILNS
Cement Plant
Measured Parti-
culate Emissions,
Ibs/ton feed,
from kiln
Opacity of
Visible
Emissions, %,
from kiln
Dragon Cement Company 0.0705
Northhampton, Pa.
Oregon Portland Cement
Lake Oswego, Oregon
Test No. 1, August 1971 0.56
Test No. 2, October 1971 0.272
Giant Portland Cement 0.574
Harleyville, S. C.
Ideal Cement Company 0.883
Seattle, Washington
Ideal Cement Company
Tijeras, New Mexico
5-10
10 - 30
0-10
0-35
20
0
18
-------
EPA has given adequate consideration to the problems
of measuring opacity by use of the trained and qualified
observer. Use of a human observer to measure stack opacity
entails essentially the same considerations that would
be given to use of a mechanical instrument, i.e., the
calibration, accuracy, and precision of the instrument.
Calibration of any instrument is achieved by comparing
the test instrument's output against a known standard.
•
Similarly, a human observer must be qualified or "calibrated"
by comparing his observations of 25 white and 25 black
plumes with the measurements of a calibrated smoke indicator
or light transmission meter. See Appendix to 40 C.F.R. Part
60, Test Method 9 (36 FR 24895, December 23, 1971).
EPA has evaluated the accuracy and the precision of
the qualified observer. Precision of an instrument refers
to the ability to consistently produce essentially the
same measurement or response when subjected, to the same set
of test conditions. A precise instrument is not necessarily
accurate because interferences or improper test conditions
may produce a bias that would cause the measurement to be
consistently in error when compared with the "true" value
of the quantity being measured. In the document
entitled, Optical Properties and Visual Effects of Smoke
Stack Plumes (Public Health Service Publication No. 999-AP-30.),
19
-------
a description is given of tests designed to evaluate the
precision and accuracy of the qualified observer (section-
title: Evaluation of Plumes by Trained Observers, pages
23-29). The evaluation showed that the. precision of the
qualified observer was reasonably good, i.e., "Group assess-
ments showed good agreement for similar sun-plume-viewer
geometries" (p. 25), however the accuracy of the observation
was greatly affected by the position of the observer with
respect to the sun and plume due to light scattering effects
of the plume (Figures 16, 17, and 19, and Tables 1 and 2). This
light scattering interference was essentially nullified
by positioning the observer with his back to the sun as
is shown in Figure 16 (East: 9:30, 10:00, 10:30 a.m.) and
Figure 17 (East: 9;45, 10:15, 10:45 a.m.). Figures 16
and 17 show that a qualified observer with the sun at his
back was able to determine the opacity of both white and
black plumes with very good agreement (accuracy) with the
in-stack light transmission meter measurements.
Thus petitioners miscited test data in Table 2 of
the Optical Properties report as indicating that opacity
tests are unreliable. Data in this table were developed
without regard to the placement of observers with respect
to the sun and plume. As discussed previously, the sun
must be to the observer's back to accurately determine
20
-------
opacity of a plume. The EPA test method for opacity
specifically requires that "The qualified observer stands
at approximately two stack heights, but not more than a
quarter of a mile from the base of the stack with the sun
to his back." (40 CFR Part 60 Appendix, Test Method 9).
EPA periodically conducts smoke reading courses for
the purpose of qualifying candidates as smoke observers in
accordance with the requirements of Test Method 9. In the
smoke school, an observer is shown and"assigns opacity read-
ings to 25 different black plumes and 25 different white
plumes. The readings, which are recorded on data sheets,
are assigned in even 5 percent increments. In order to
qualify, no single reading may be in error by more than 15
percent opacity; the average error may not exceed 7.5 per-
cent opacity in each category.
In order to assess the accuracy and precision with
which qualified observers read plumes, the results of 50
individuals were selected at random from candidates in nine
smoke school courses conducted in 1972 and 1973. This
provided a total of 2500 observations for analysis. These
results are available in a report titled Average Observational
19/
Error Associated with Smoke Plumes at Levels of Known Opacity. —
The analysis of these observations considered two
parameters:
21
-------
3. the average error ;aade in observation of plumes
of known opacity, and
2. the precision of individual observations
of plumes of knov:n opacity.
The results of these analyses are described below.
For each smoke plume which was observed the error or
difference between, the observed opacity and the actual
opacity (observed opacity minus actual opacity) was recorded.
For plumes of 0%, 5%, 10%, and 15% opacity, these errors
were tabulated separately for the respective plume. Above
this level plumes were combined into 10% opacity intervals
for convenience of calculation. Thus a 5% error associated
with a plume of either 20% or 25% was tabulated under the
actual opacity of "20% or 25%."
Two calculations were made using the above data: the
"average error," and the "average absolute error." The
following example illustrates these calculations:
Data for a plume of
actual opacity =10%
•-5% error = 2
0% error = 31
+5% error = 31
+10% error = 23
+15% error = _6_
93 total observations
2(-5%) + 31(5%) + 23(10%) + 6(15%)
Average error =93 = +5.0%
2(5%) + 31(5%) + 23(10%) -t 6(15%)
Average absolute error = 91= 5.2
22
-------
The "average error" indicates the overall bias of the
group with respect to a particular plume. A positive value
would indicate that the average observation is greater than
the actual plume, and vice-versa. As example, a -5% error
and a +5% error would result in an "average error" of 0%.
The "average absolute error" indicates the average
absolute accuracy attained by the observers. As example,
a -5% error and a +5% error would result in an average
absolute error of 5%.
The results of these calculations are shown in Table 2.
Table 2. AVERAGE OBSERVATIONAL ERROR
Average Absolute
Actual
Opacity*
20
30
40
50
60
70
80
90
0
5
10
15
or 25
or 35
or 45
or 55
or 65
or 75
or 85
or 95
100
Number of
Observations
12
12
93
96
331
409
606
393
235
185
120
96
12
Average Error
% Opacity
+ 3.
+2.
-1-5.
+2.
+ 2.
+0.
-0.
-0.
-0.
-1.
-2.
-2.
-5.
3
5
0
8
5
1
1
1
2
2
4
6
8
Error
% Opacity
3
2
5
5
5
5
5
6
6
6
6
4
5
.3
.5
.2
.4
.0
.1
.5
.0
.2
.1
.0
.6
.8
*Measured with a photoelectric transmissometer located in the
smoke generator stack.
A review of Table 2 shows that on the average, the smoke
readings can be made with good accuracy. With two exceptions,
23
-------
the average accuracy is within j^ 3.3 percent. At two
levels of opacity, 10% and 100%, the average error was
noted to be 5% and 5.8% respectively.
No test method is invariably perfectly accurate.
Water vapor does not cause major problems in opacity testing.
When stack emissions contain appreciable water vapor, the
vapor becomes visible when the gases are cooled below the
dew point and water condenses in fine droplets. Some water
plumes are detached while others are visible as released
c-/
from the stack. However, since ambient air is normally
much drier than stack emissions' containing visible water
vapor, dilution causes the gases at some point to become
unsaturated. At this point, condensed water evaporates and
visually disappears from the plume. Trained smoke observers
can discriminate between condensed water vapor and visible
pollutants by determining the opacity at or after that point
in the plume where the water vapor dissipates.
Under relatively warm and dry conditions, the visible
water plume evaporates and becomes invisible within a
short distance of the stack. In cold weather, air has
much less capacity to carry water vapor, the water plume
The phenomenon of~a "detached plume " is often observed
when a warm moist gas stream is released into a cooler
atmosphere, e.g., when hot, wat exhaust gases from a vet
process kiln are released to the atmosphere. Here the
gases are clear at the point of discharge but become
opaque as the wet gases mix with cooler air in the atmos-
phere .
24
-------
from a given source is usually much more pronounced and
has a longer path. Due to the character of the dust and
atmospheric conditions existing at a given point in time,
i.e., humidity, cloud cover, rain, snow, etc., it may not
be possible in some cases to evaluate opacities of the
particulate fraction with sufficient precision to warrant
enforcement action. In such cases, it is necessary to make
multiple observations under varying atmospheric conditions.
Opacity and water vapor phenomena are shown photographically
in the EPA report entitled, "A Report on Contaminated Water
W
Vapor Plumes". The proper place to deal with such small
variations is at the time of enforcement. For example, the
agency would not take, and the courts would not sustain,
enforcement action based upon opacity observations that
do not exceed the standard by more than the average opacity
error associated with the particular opacity standard. It
is important to note in this connection that under §113(c)(1)(C)
of the Clean Air Act criminal sanctions can only be imposed
on persons who "knowingly" violate §111 standards. Also
orders issued under §113(a)(4) must take into account "the
seriousness of the violation and any good faith efforts to
comply with applicable requirements."
25
-------
Opacity standards have been tested by time, experience,
and litigation. California enacted a statutory opacity
requirement in 1947. California Health and Safety Code
24242, Stats. 1947, ch. 63251. In an analysis prepared
for EPA by L. J. Duncan of The Mitre Corporation titled
Analysis of Final State Implementation Plans - Rules and
16/
Regulations, (APTD-1334, July 1972), Table XVI (p. 51-52)
shows that every State has equivalent opacity restrictions
applicable to particulate emissions. These range from 0%
opacity ("no visible discharges") to 40% opacity during
normal operations with most restrictions being 20% opacity.
Both Maryland (in two Air Quality Control Regions) and the
Dis-trict of Columbia have no visible discharge limitations.
Various courts have upheld the validity of opacity
requirements. See, e.g. State v. Fry Roofing Co., 495 F.2d
751, 4 ERC 1116 (Oregon Court of Appeals 1972) ; People v.
Plywood Manufacturers, 291 F.2d 587 (Superior Court of Los
Angeles, California, 1955).
Opacity standards are a necessary supplement to mass
emission standards. Opacity standards help assure that
sources and emission control systems continue to be properly
maintained and operated so as to comply with mass emission
standards. Particulate testing by EPA Method 5 and most
other techniques requires an expenditure of $3,000 to $10,000
26
-------
per test including about 300 man-hours of technical personnel.
Furthermore, scheduling and preparation are required such
that it is seldom possible to conduct a test with less than
two weeks notice. Therefore, Method 5 particulate tests
can be conducted only on an infrequent basis.
If there were no standards other than mass emission
standards, operators could inadequately operate or maintain
pollution control equipment at all times except during periods
of performance testing. Operators are aware that two weeks
or longer are necessary to schedule a typical stack test.
If only small repairs were required, e.g., pump or fan repair
or replacement of fabric filter bags, the operator could
delay such remedial action and continue to pollute until
shortly before the test is conducted. For some types of
equipment such as scrubbers, the operator could reduce the
energy input (the pressure drop through the system) when
stack tests weren't being conducted and could release signi-
ficantly more particulate than normal. Therefore, EPA has
required that operators properly maintain air pollution con-
trol equipment at all times (40 CFR 60.11(d)) and meet
opacity standards at all times except periods of startup,
shutdown, malfunction, and as otherwise provided in the
applicable standard (40 CFR 60.11(c)).
Cost applicable to retests only. Initial performance test
is usually more costly because of need to erect platforms,
scaffolding, or other sampling facilities.
27
-------
Opacity of emissions is indicative of whether control
equipment is properly maintained and operated. However, it
is established as an independent enforceable standard, rather
than an indicator of maintenance and operating conditions
because information concerning the latter is peculiarly
within the control of the plant operator. Furthermore, the
time and expense required to prove that proper procedures
have not been followed are so great that the provisions of
40 CFR 60.11(d) by themselves (without opacity standards)
would not provide an economically feasible means of insuring
on a day-to-day basis that emissions of pollutants are within
allowable limits. Opacity standards require nothing more
than a trained observer and can be performed with no prior
notice. However, in some cases, including times when opacity
standards may not be violated, a full investigation of
operating and maintenance conditions will be desirable. Accord-
ingly, EPA has requirements for both opacity limits and
proper operating and maintenance procedures.
For the foregoing reasons, EPA believes that opacity
standards are such an inclispensible air pollution control
tool for the States and Federal government, and that the range
of uncertainty in applying such standards is so small that
this type of standard should continue to be used by EPA as
well as the States.
28
-------
COSTS
The only cost issue which the Court remanded to EPA
was whether the costs of a standard which precluded the
use of wet process kilns and precipitators were reasonable.
EPA's standard will not preclude use of wet process kilns
nor of precipitators. This issue was discussed above
along with -the relative costs of control equipment for wet
and dry process kilns.
For purposes of this remand, it is sufficient to state
that EPA continues to believe that the economic costs
associated with the standard are reasonable. This conclu-
sion is based primarily on the reports: "The Financial
Impact of Air Pollution Control Upon the Portland Cement
4/
Industry" by J. R. Elias and J. M. Dement~and "The Cement
Industry; Economic Impact of Pollution Control Costs" by
13/
the Boston Consulting Group.
29
-------
ADVERSE ENVIRONMENTAL IMPACTS
The new source performance standards for Portland cement
plants are not expected to result in significant adverse
environmental impact. The possible areas of adverse environ-
mental impact are: 1) increased pollution from generation^
of electricity required to operate precipitators, 2) water
pollution from disposal of collected particulate matter.
Neither of these is a serious problem.
The electric power rate required to operate a precipitator
that would meet the new source standard on a typical wet
process kiln would be approximately 115 kilowatts. This
should be compared with the total electric power rate consump-
tion of a typical cement plant which approximates 5300
kilowatts and with the 500,000 kilowatt output of a typical new
steam generating electric power plant. The additional
115 kilowatts will thus have little incremental impact on
pollution created by electric power generation generally
and that created by electric power generation for Portland
cement plants specifically.
The above figures were developed on the following
assumptions: A typical cement plant producing 2.5 million
bbl of cement annually would exhaust gases at a rate of
230,000 cfm. (Robert Elias and John Dement, "Financial
Impact of Air Pollution Control Upon the Portland Cement
I/
Industry"). In order to meet the new source performance
30
-------
standards of 0.3 pound per ton kiln
feed (1.6 ton feed per ton product), based on an emission
factor of 235 pounds per.ton of cement product, a precipitator
would have to have a collection efficiency of approximately
99.92 percent. A study conducted by Southern Research
Institute shows that precipitators recently installed on two
wet process plants which have collection efficiencies of
99.92 and 99.95 percent had power rates of 440 watts per
1000 cfm and 520 watts per 1000 cfm respectively. (A manual
of Electrostatic Precipitator Technology, Contract CPA
22-69-73 for NAPCA, Figure 18.10, Part II - Application
Areas)Ji/At a power rate of 500 watts per 1000 cfm, a
precipitator handling 230,000 cfm at a collection efficiency
of 99.92 percent would require 115 kilowatts.
The total electrical consumption (excluding fuel) of
a cement plant is 120 kilowatts per metric ton product.
For the typical plant producing 44-45 metric tons per hour
(2.1 million bbl per year), the total plant energy consumption
approximates 5300 kilowatts. (Southern Research Institute,
"Development Document for Effluent Limitations Guidelines
and Standards of Performance: Cement Manufacturing Industry,"
15/
Contract 68-01-0599) .~~
State implementation plan requirements would require
up to 81 kilowatts to be consumed in the precipitator of
the typical cement kiln even if there were no
31
-------
applicable .now source star.d^ras. State implementation
plans require controls on tho emission of particulate
m'atter by cement plants ranging from 0.36 to 0.89 Ib. per
ton feed. ("Analysis of Final State Implementation Plans -
16/
Rules and Regulations," APTD-1334). These standards require
collection efficiencies of 99.76-99.9U percent. The
power rates associated with these efficiencies are between
125-350 watts per cfm. A precipitator with an approximate
size of 230,000 cfm would thus require between 29 and 81
kilowatts to meet State Implementation*?lan standards.
Although higher efficiency dust collectors will increase
the amount of material collected, the new source performance
standards will not substantially increase the land disposal
problems faced by cement plants. Furthermore, it will
require the collection of dust which would otherwise be
vented to the atmosphere and indiscriminately deposited on
land in the vicinity of the cement plant. The present
practice in the cement industry is to dispose of the collected
kiln dust that cannot be returned to the kiln. The waste
dust is hauled or slurried to an unused part of the quarry
from which the raw material came or is landfilled. Table
3 summarizes the methods employed by 80 cement plants to
dispose of kiln dust.
32
-------
Table 3. METHODS OF DISPOSAL OF COLLECTED KILN DUST
BY CEMENT PLANTS ^
Method Number of Percent of 80
Plants Reporting Plants Reporting
all dust returned to kiln 27 34
surface piling (dry) 29 36
returned to quarry (dry) 11 14
leached 9 11
slurried and discarded 7 9
some sold or hauled away 8 10
by contractor
(a) Source: Southern Research Institute, "Development Document
for Effluent Limitations Guidelines and Standards of Per-
formance: Cement Manufacturing industry," Contract 68-01-
0599. 15/
(bj Percentage total is greater than 100 because some plants
report more than one method.
State implementation plans require controls on the
emis: sion of particulate matter by cement plants ranging
from 0.36 to 0.89 Ib. per ton feed. Compliance with these
standards will require cement plants to dispose of substantial
amounts of material. The amount of additional dust collected
to meet the new source performance standards of 0.30 Ib.
per ton feed will not be significantly greater than that
collected to meet the state implementation plan standards.
33
-------
for similar reasons, the increased amount of material
collected by higher efficiency dust collectors will not
substantially increase the risk of water pollution. Present
land disposal methods could permit dust with a high
alkaline content to-cause potential water pollution problems
from rainfall runoff or by percolation into streams and
rivers. If alkalai dust were not controlled but instead
were released to the atmosphere, it would also find its
way into streams and rivers. However, containment and
treatment of runoff from kiln-dust piles will eliminate
the danger of water pollution resulting from alkalai cement
dust. The Southern Research Institute study previously cited
15/
(EPA Contract 6 8-01-059 9^"indicates that technology exists
to treat the contamimited water resulting from leaching
and describes that technology.
In any event, the increase in the collection of
particulate due to new source performance standards as
compared to the collection of particulates resulting
from the application of state implementation plan
requirements will be insignificant.
SUMMARY
EPA thus concludes that the standards are achievable
and have been achieved by both wet and dry process kilns,
that the economic costs are reasonable and that the adverse
environmental impact is minimal. EPA further concludes
34
-------
that the opacity standard is a valid and necessary standard
which meets the requirements of §111 of the Clean Air Act.
Therefore, EPA, after reconsidering the standards pursuant to the
remand of the U. S. Court of Appeals for the District of Columbia,
has concluded that the standards should not be changed.
35
-------
-------
APPENDIX I
EPA RESPONSES TO SPECIFIC ISSUES RAISED BY THE COURT
37
-------
-------
APPENDIX I
This appendix contains responses to specific issues
raised by the U.S. Court of Appeals for the District of
Columbia in the Portland Cement case. References are to
the Slip Opinion.
Issue 1
On page 21, the Court asks EPA to respond to petitioner's
comments concerning possible adverse environmental impact
of the standards.
Response
These impacts are discussed in the text above.
Issue 2
On page 23 and 24 and again on page 41, the Court asks
EPA to consider the justification for and economic impact
of standards that would terminate the use of wet process
kilns and precipitator control of both wet and dry process
emissions.
Response
EPA's standards do not preclude the use of either wet
process kilns or precipitators. Test results show both have
met the standards. This is discussed in the text above,
along with economic considerations.
39
-------
Issue 3
In footnote 55 the Court s?id:
"If the sair.e technology is now available and
in use for incinerator?, steam power plants and
cement plants, the Administrator on remand may wish
to offer sorr.e further explanation of the difference
in standards set simultaneously for the three
industries."
Response
At the time the cement, incinerator and steam generator
standards were promulgated, EPA had data showing that bag-
houses would enable both wet and dry process Portland cement
kilns to achieve concentration levels of .03 grains per dry
standard cubic foot (gr/dscf) or less. EPA judged but
had not confirmed that a precipitator could achieve the
same levels. EPA also suspected but had inadequate data
to confirm that incinerators and steam generators could
use baghouses to achieve concentration levels approximating
those achieved by cement kiln baghouses.
Because different types of sources emit gases of
different composition at different velocities, it cannot
be assumed that a device that removes particulates will
be equally effective in removing every type of particulate
from every type of source. Furthermore, different types
of sources nay emit substances which will interfere with
the operation of the control equipment.
40
-------
Thus EPA based its incinerator, steam generator and
cement standards primarily on applying baghouses to cement
plants and assumed less efficient precipitators would be
used for incinerators and steam generators, although EPA
had indications that baghouses could be applied to incinerators
and steam generators and that precipitators could achieve
emission concentrations similar to those of baghouses.
As note'd in the text above, precipitators on cement
kilns do achieve the standards. As data is gathered show-
ing precipitators, baghouses, or other dust collectors on
incinerators and steam generators also can achieve similar
concentration levels, the standards for these sources will
be revised. Any data the cement industry has on the emission
levels which can be achieved by incinerators and steam
generators should be submitted to EPA to facilitate the
review of the standards for these two source classes.
41
-------
ISSUE 4
On page 37 tha Court directs EPA to consider conten-
tions presented in briefs to this Court, though not pre-
viously raised, unless EPA explains why they are not
material.
RESPONSE
The above text and these responses to questions raised
by the Court effectively respond to the points raised
by petitioners.
ISSUE 5
In footnote 73 the. Court says it is uncertain whether
the test was referred to in the Background Document as not
available at the time this document was published was of
one of the two kilns tested at the Dragon Plant or v/as
located at still another plant.
RESPONSE
The missing test was that of the Giant Portland Cement
plant in Harleyville, South Carolina, a wet process plant
with a baghouse.
ISSUE 6
On page 40, the Court questions whether a single
test is an adequate basis for inferring that all new
cement plants would be able to meet the proposed standards.
42
-------
RESPONSE
At the tir.3 the standard was promulgated, EPA had
tested two plants which met the standard (The Dragon and
Oregon plants). Each test consisted of three repititions
of two hour sampling periods or six total hours of testing.
EPA has since obtained test results from two other plants
which show emissions well below the standard: The Lehigh
Portland Cement Co. and Marquette Cement Manufacturing Co.
tests referenced in the text above.
The Court in raising this issue referred to a letter
from Mr. Kreichelt which was critical of EPA's basing the
standard on tests which showed the standard was met at
only one plant. As three additional plants have since been
tested which meet the standard, Mr. Kreichelt's criticism
is no longer applicable. However, even if it were, the
following considerations should be taken into account in
evaluating his criticism. First it was based on the assump-
tion, correct at that time, that EPA's test method would
include particulates collected in the impinger portion of
the sampling train. Upon this assumption, the one test
result (Dragon Cement) referred to by Mr. Kreichelt showing
the standard could be met would show an average of .184
pounds per ton compared with the standard of .3 pounds per
ton. However, the deletion in the promulgated standard of
43
-------
the requirement that impinger catch must be included means
that the Dragon plant in fact tested at .07 pounds per ton
or less than one-fourth of the standard.
Second, as noted, test results have since become
available which show three more plants, including a wet
process plant, have met the standard.
Third, even if EPA did not have these additional test
results, the Agency is not constrained to accept Mr. Kreichelt's
definition of what is "achievable." The greater the number
of tests required to prove a standard can be met before the
standard can be promulgated, the greater the likelihood the
new source performance standards would reflect only the
technology and maintenance requirements which existing
plants have chosen to apply. However, EPA was testing to
determine the best levels of control being achieved at
existing plants. Engineering judgment was then applied by
EPA to determine whether new plants could be constructed to
perform at control efficiencies equal to or better than that
observed at the tested existing plants. EPA was not attempt-
ing to determine the statistical average performance of
existing plants.
ISSUE 7
On page 41, the Court questions whether validity of
standard is affected by the different testing method
44
-------
required by EPA in the promulgated standards and that
used by EPA in its tests in developing the standards.
The Court first refers to the elimination of the require-
ment in the p: ilqated standard that the impinger catch
of the sampling train be included in determining the parti-
culate emissions.
RESPONSE
In EPA's proposed standards the particulate sampling
«
method required that both the front half (probe, cyclone
and filter) and back half (impinger) catches of the sampling
train be measured. When EPA tested the Dragon and Oregon
plants, it calculated the pounds of particulate per ton of
feed by adding the amounts collected in both the front
and back half of the sampling train. However, EPA separately
measured both front and back half catches. (See Summaries
of Test Data for Standards of Performance for New Stationary
1Z/
Sources, August 1971.) Thus, the change in test methods in
no way invalidated EPA's test results. It only rendered
some of the data valueless. Based only on the probe,cyclone
and filter catch data, the Dragon Cement plant emitted an
average of 0.07 pounds per ton during the test of that plant
(See Summaries of Test Data and Emission Testing Report,
5/
ETB Test Number 71-MM-05r and the Oregon plant emitted an
average of 0.28 pounds per ton (See Emission Testing Report,
i/
ETB Test Number 71-MM-15).
45
-------
ISSUE 8
On page 42, the Court asks the Administrator to explain'
why compliance with the standard is based on a two-hour
test while the data base was developed on 30-minute tests?
RESPONSE
There were four ports to be sampled at Dragon
Cement; during each of the three tests each port was
tested for 30 minutes for a total sampling time of 2
hours, the time required for compliance test. Thus,
there are inevitable time periods during which no
sampling will be done as it takes time to correctly
position and adjust the sampler at each traverse point.
The Sources subject to this standard will also have such
gaps in their own tests, which gaps will not invalidate
their tests.
The general provisions of 40 CFR 60 are being revised
to clarify the fact that a run or repetition may be either
continuous or intermittent, consistent with good engineering
practice. This revision does not alter the test method; it
only makes clear the fact that the normal practice of inter-
rupting runs for such purposes as moving sampling equipment
or sampling cyclic processes is, in fact, expected and is
an accepted procedure which is within the scope of the test
method.
46
-------
issu:: 9
On page 42, the Court raises the question on the
basis of a handwritten note made by the EPA contractor, as
to whether the tested plant was operating at maximum per-
formance during testing. The contractor had noted, "Bag-
house is undersize and production is held back due to this."
Compliance tests under the regulation require, however,
that "All performance tests shall be conducted while the
affected facility is operating at or above the maximum
production rate. . . ."
RESPONSE
The handwritten note of the EPA contractor refers to
a presurvey note of an EPA engineer. That note referred
to the potential production capacity of the plant as re-
ported by plant personnel and did not indicate that the
Dragon Cement kiln was being operated at an abnormally low
production rate.
A cement plant's production rate is related to many
variables in the system. A few of the variables that could
influence or limit the production rate are: the capacity
of the grinders, the size of the kiln, the size of the fans
or pumps in the system or the size of control equipment
installed to control particulate emissions from the kiln.
In addition,.variables such as kiln temperature, pressure,
Btu content of fuel, excess air for combustion and moisture
content of the feed could all directly or indirectly affect
the production rate of the cement plant.
47
-------
The baghouca install-: u at. Dragon Cement was sized ror
the production rate desired by the company taking into
account the othr.r variables -in the system. During the test
period, the plant v:as op-orated at its normal maximum product ion
rate.
ISSUE 10
On page 42 and 43, the Court restated petitioner's
contention that EPA erred in the Dragon test in (a) not
using observed measured values, (b) substituting lower
reading from another test run for a reading based on con-
tamination of the sampler by solids, and (c) pouring
liquid obtained from one run into the beaker used to hold
liquid from another run. The Court went on to say
"However, deviations from prescribed measurement techniques
are not necessarily significant as to testing results,
and if petitioners press this point on remand they must
establish that such test deviations bear significant conse-
quences." Nevertheless, in the interest of providing a
complete response to the Court's remand EPA will respond
to this issue.
RESPONSE
Petitioner's comments were based on comments made bv
EPA on page 7 of the Dragon Test Report, ETB Test No. 71-MII-05.
EPA has to test existing plants as they are. The Dragon
plant had allowed solids to build up in the duct downstream
of the baghousG. These solids, which EPA was unaware of
48
-------
until after Test Run 1, contaminated the probe and cyclone
and also reduced the cross section of the duct. In order
to correct for these solids, EPA had to compute the cross
section of the duct through v:hich air could pass.
The catch of the front half of the sampling train
is measured by adding together the probe and cyclone catch
and the filter catch. The probe and cyclone capture large
particles and the filter catches smaller particles. It
was observed that the probe and cyclone in Test Run 1
captured 98.1 mg of particulates while the probe and cyclone
catch in Run 2 was 15.9 mg and'in Run 3 was 18.0 mg. It
was also observed that at one sampling point in the duct the
gas velocity was zero. This led to an inspection of the
duct which showed that solids v/ere in fact deposited on
the bottom. The probe and cyclone had obviously collected
some of these particles. Therefore, the catch of the probe
and cyclone was assumed to be 17.0 mg, the average of Runs
2 and 3. The actual filter catch in Run 1, 38.6 mg, was
used. Thus a part of the Run 1 results were calculated
rather than measured. However, the possible error intro-
duced in this one portion of one Test Run at the Dragon
plant was so small that in the opinion of EPA it does not
substantively prejudice the results of the Dragon test
and certainly not the standard, which was based on other
tests in addition to the Dragon tests.
49
-------
The mistake in pouring liquid obtained in one run in the
wrong beaker is now irrelevant since the liquid was impinger water.
The promulgated standard deleted the requirement that impinger
catch be added in ascertaining compliance.
ISSUE 11
On pages 43-44, the Court directs EPA to respond to criticisms
expressed by Kir. Striker to the effect that EPA underestimated the
stack gas flow rate, and thus the amount of particulates emitted, by
a factor of two.
RESPONSE
Before responding to specific allegations of error made by
Mr. Striker, it is important to note that even if Mr. Striker were
correct in his allegations, and EPA believes that he is not, the
tests would still show that the Dragon Cement Plant met the EPA
promulgated new source performance standard. Mr. Striker concluded
that the "correct emission of particulate matter is in the neighborhood
of .404 pounds per ton of kiln feed." But this figure was reached
by using 0.2 Ibs/ton based on the entire train for the first, and
maximum, of three tests. After changing to a front-half catch only,
(the terms of the promulgated new source performance standard), the
result from this test was 0.11 Ibs/ton and even if the error that
Mr. Striker claims had been possible and had occurred, test results
from this maximum test would be 0.22 Ibs/ton and average test results
would change from .07 libs/ton to .14 Ibs/ton, well below the standard
of 0.3 Ibs/ton.
50
-------
It is this average of three tests that is to be used for
compliance with the new source performance standard. This plant
would still be well below new source performance standards of .30
Ibs/ton.
In any event, EPA has thoroughly considered Mr. Striker's
specific allegations of error and disagree with him. Mr. Striker's
allegations are well summarized by the Court as follows:
Finally, engineer Striker claims significant errors of
measurement were made in determining the measurement of the
cubic feet of stack gas per minute, and a resulting under-
statement of the true volume of calcining carbon dioxide
included in total stack gas. He states that commonly "35%
(plus or minus 1%) of raw feed is converted into carbon
dioxide in the burning process." He then notes that an
accurate measure of raw feed is the volume of calcining
carbon dioxide appearing in stack gas, which in turn depends
on an accurate measure of the volume of stack gas. His own
calculations, based on EPA data showing a stack flow rate
of 51,187 cubic feet per minute of dry gas, indicate that
there were 2,153 cubic feet per hour of stack gas in the
test attributable to calcining carbon dioxide coming from
the raw feed and that "as a matter of basic chemistry"
2153 cubic feet of calcining carbon dioxide "comes only
from 22.11 tons of raw feed." This was at variance from
the kiln rate of 44.03 tons of raw feed per hour reported
in the test. He concludes that the error lay not in the
measurement of the raw feed, but in the test data reported
on the stack gas volume -- flow rate of 51,187 — which in
his judgment requires more sophisticated equipment for
recording than does the raw feed which is easily measureable.
Having corrected the stack gas figure, he states his opinion
that the EPA assumption of emissions satisfying its ultimate
0.30 standard was in error. He concluded:
It is my personal opinion that the particulate
matter emissions of .202 pounds in test 1 per ton of
kiln feed reported ... is grossly erroneous and
that the correct emission of particulate matter is
in the neighborhood of .404 pounds per ton of kiln
feed.
51
-------
We are not competent to decide if Mr. Striker's methodology
and conclusions are correct. We can note, however, that
he claims that as a matter of "basic chemistry" two test
values, for feed and gas volume, cannot co-exist. This is
certainly the type of criticism EPA should be required to
discuss on remand.
The possible error in measurement of dry stack gas volume has
been gone over repeatedly by EPA engineers and Mr. Striker. Basically
the situation is this: Mr. Striker claims that the CCL does not
balance in the system, i.e., that the calculated CCL from combustion
plus COp driven off from the calcium carbonate is not equal to the
COp calculated from Orsat (gas composition) readings and stack gas
flow measurement. Mr. Striker's C0? calculations on the CCL balance
appear to be essentially correct. Mr. Striker's error lies in his
assumption of the cause of this CCL discrepancy. There is no conceivable
source of error in measurement of the stack gas flow rate that could
even approach 100%. Moreover the gas flow rate was measured consistently
during twelve traverses at a total of 180 points. Nowhere does Mr. Striker
say where such an error could occur.
A flow rate as suggested by Mr. Striker was impossible because of
the plant design. Neither the fan nor the baghouse could have possibly
handled such a volume. This has been confirmed with the plant management.
Further evidence of the correctness of EPA measurements (and the
52
-------
incorrectness of llr. Striker's assertions) is contained in
later test resul-s from the Dragon Cement plant cited by
Edward Klein of Martin-'Iarietta (owners of Dragon Cement Co.).
These results show flow rates very close to those measured
20/
by EPA at full production (as in EPA tests). The summary of
these test results supplied by Mr. Klein is available from
EPA. Thus Mr. Striker's allegations would violate the basic
laws of nature had they been correct. Mr. Striker assumes
that everything measured was correct except for a 100% error
in stack gas volume. Under that assumption our number when
expressed as pounds per ton of feed should have been twice
that stated. As shown above, this could not be the case.
There are two other possible sources of this CC^ error:
feed rate and Orsat (gas composition) analyses. Mr. Striker
agrees that kiln feed rate is not in error. The plant was
operating steadily at full production rate. A check with.
plant personnel shows our feed rate and their records to
be in agreement. It should be pointed out that Orsat
analysis for oxygen, carbon dioxide, and other combustion
gases were only incidential to particulate testing. They
in no way affect results. Kiln feed is the chief variable
of cement production and careful control is maintained.
Any substantial variation would have been noticed by the
plant.
Thus, the only possibility is that the error was in
the Orsat analysis (% of each gas in effluent). Calculations
53
-------
performed by EPA experts on this indicate that the CC>2
absorbing solution in the Orsat was exhausted and CO- in
the stack gases was not measured as CO£ but rather as
oxygen. Checks with EPA emission testing engineers who
supervised the teat confirm that this was the source
of discrepancy in the CO- balances,
In order to determine the pounds of particulate
per ton of feed, it is necessary to sample the amount of
particulates per unit of gas emitted and then multiply
this figure by the total amount of gas emitted during
the period a measured amount of feed is consumed. This
is what EPA did in calculating the emission rates from
the plants tested. If, Mr. Striker's allegations were
correct, the emission rate would indeed need to be changed.
The error in the Orsat analysis explains the apparent
violation of the laws of "basic chemistry" but does not
necessitate any change in the test results reported.
ISSUE 12
On page 44 and 45 and in footnote 90, the Court
suggests that if their study of the matter is correct, the
Oregon wet process kiln test results to be compared with
the standard, i.e. front half catch only, are .247, .309
and .261 Ibs/ton rather than .535, .361 and .291 Ibs/ton.
RESPONSE
The court is correct. The average test results for
the Oregon test which should be compared with the standard
is .28 Ibs/ton.
54
-------
ISSUE 13
On pages 45 to 47, the Court expresses concern that
problems at the Oregon plant caused interruptions in the
sampling. The Court suggests that a provision allowing
for malfunctions could be taken into account in consider-
ing this issue on remand.
RESPONSE
The testing of the Oregon plant is described in the
test above. During the testing period«at Oregon Portland
Cement plant, a test was begun on 10/7/71 and discontinued
due to kiln shutdown. The test was resumed and completed
on 10/8/71. Two additional tests were run on 10/8/71 with
only short interruptions to change ports as required by
the test method. On October 5 and 6, 1971, two earlier
attempts were made at measurement of emissions but conditions
prevented meaningful measurements. The important things
to note are that process operation was extremely upset
during these attempts and did not represent normal operating
conditions and that the manner in which the early attempts
at testing were-performed was an attempt to make a test at
normal operating conditions. The second and third tests
recorded were at normal operating conditions (not artificially
ideal conditions) to be expected during nornal cement plant
operation. No stop-start testing, as described in the
55
-------
petitioners' quote, was used (except as noted in the response
to Issue 8). The first test recorded had only one
interruption (other than port changes) due to the kiln
shutting down. The circumstances described above are
9/
in ETB Test Report Number 71-MM-15, (pages 16 and 47),
the same report quoted by petitioner.
The. fact that the. Oregon plant was not being properly
operated or maintained does not limit EPA to standards
based on tests conducted during periods of malfunction.
EPA attempts to determine the best levels that a plant
can achieve. In any event, the testing at the Oregon
Portland Cement plant was consistent with the manner of
conducting performance tests under the proposed revision
to §60.8 (c) of Title 40, Code of Federal Regulations. 38
Fed. Reg. 10820-21 (May 2, 1973). That section provides
that compliance tests shall be based upon representative
performance, which shall not include periods of startup,
shutdov/n, and upset. EPA has promulgated regulations
which make it clear that EPA's mass emission limits apply
onxy during performance tests rather than at all times.
(38 Fed. Reg. 28565, October 15, 1973).
ISSUE 14
On page 48, the Court directs EPA to state with
specificity what literature is relied on to support the
standard.
56
-------
PESPONSE
This is discussed in the text above. As indicated in
the text, a principal literature source was a government
publication, resulting from a study conducted in cooperation
with representative companies manufacturing portland cement
entitled, "Atmospheric Emissions from the Manufacture of
Portland Cement". This principal literature source provided
basic information regarding cement planj-.s, including process
details, process equipment, types of control equipment
utilized, and documentation that baghouses could be used
for the control of particulates from both wet and dry process
kilns, i.e. that baghouse control is "adequately demonstrated."
In addition, the names of cement companies - location and
size of cement plants were listed and this information was
utilized by locating the best controlled plants to be
considered for inspection and possible source testing in
conjunction with the development of new source performance
standards. Source test data included in the report were
not relied on by EPA to show that th_e standard was being
met.
ISSUE 15
In footnote 95, the Court states that if vendor repre-
sentations are to be relied on, more might be required than
mere comments.
57
-------
RESPONSE
The relevance of vendor statements is discussed in
the text above. The tests EPA conducted were sufficiently
supportive of EPA's standards that vendor representations,
although relevant to confirm the standard can be met, are
not of such crucial importance that the time and effort
required to go beyond their professional representations
is warranted.
ISSUE 16
On pages 49-52, the Court directs EPA to reconsider
the opacity standard.
RESPONSE
This is discussed in the text above.
58
-------
BODY OF RESPONSE AND APPENDIX I
1. Atmospheric Emissions from the ^Manufacture of Portland
Cer.ent, U. S. Dlir/?, PI-:S, national Center for Air Pollu-
tion Control, Cincinnati, Ohio, PHS Publication I\To.
999-AP-17, 1367.
2. Control Techniques for Particulate Air Pollutants, U.S.
DHEW, PHS, CPZHS, National Air Pollution Control Admin-
istration, V7ashington, D. C., MAPCA Publication No. AP-
51, 1969.
3. Establishment of National Emission Standards for Stationary
Sources, Volume VI, Portland Cement Manufacturing Plants,
Contract No. CPA-70-164, Task Order No. 2, Prepared for
the National Air Pollution Control Administration, Durham,
N. C., by Research Triangle Institute and PEDCo Environ-
mental Specialists, Incorporated, 1970.
4. The Financial Impact of Air Pollution Control Upon the
Portland Cement Industry, Internal document, Environ-
mental Protection Agency, Raleigh, North Carolina, 1971.
5. Emission Testing Report, Environmental Protection Agency,
ETB Test Number 71-I1M-05, Dragon Cement Company,
Northampton, Pennsylvania.
6. Emission Testing Report, Environmental Protection Agency,
ETB Test Number 71-MI1-03, Ideal Cement Company, Seattle,
Washington.
7. Emission Testing Report, Environmental Protection Agency,
ETB Test Number 71-MM-01, Kaule Industries, Hialeah,
Florida.
8. Emission Testing Report, Environmental Protection Agency,
ETB Test Number 71-MM-07, Giant Portland Cement Company,
Harleyville, South Carolina.
9. Emission Testing Report, Environmental Protection Agency,
ETB Test Number 71-MM-15, Oregon Portland Cement Company,
Lake Oswego, Oregon.
10. Emission Testing Report, Lehigh Portland Cement Corpany,
Union Bridge, Maryland, dry process, electrical precipita-
tor, Union Bridge No. 4 kiln, Tested by Resources Research,
Inc., 1972.
59
-------
11. Emission Testing Popart, Xarquette Ceir.ent Manufacturing
Company, llr.gors - ov, n, Marylcnd- dry process, electrical
precipitator, TooLcd by Marquette Cement Manufacturing
Company, 1972.
12. Optical Properties and Visual Effects of Smoke-Stack
Plumes, U. S. DHL?!,7, PKS, Bureau of Disease Prevention
and Environmental Control, Cincinnati, Ohio, PHS Publica-
tion No. 999-AP-30, 1967.
13. The Cement Industry:: Economic Impact of Pollution Con-
trol Costs, Volume I, Contract No. EQC-204, Prepared
for the Council on Environmental Quality and the Environ-
mental Protection Agency, Washington, D. C., by the
Boston Consulting Group, 1971.
14. A Manual of Electrostatic Precipitator Technology,
Part II—Application. Areas, Contract No. CPA 22-69-73,
Parepared for the National Air Pollution Control
Administration, Cincinnati,'Ohio, by Southern Research
Institute, pp. 609-640, 1970.
15. Development Document for Effluent Limitations Guide-
lines and Standards of Performance, Cement Manufacturing
Industry, Contract No. 68-01-0599, Draft report prepared
for the U. S. Environmental Protection Agency, Washington,
D. C., by Southern Research Institute, 1973.
16. Analysis of Final State Implementation Plans--Rules
and Regulations, Contract No. 68-02-0248, Prepared
for the U. S. Environmental Protection Agency, Research
Triangle Park, N. C,, by the Mitre Corporation, Office
of Air Programs, Publication No. APTD-1334, 1972.
17. Standards of Performance for New Stationary Sources,
Summaries of Test Data, Internal document, U. S. Environ-
mental Protection Agency, Durham, N. C., 1971.
18. A report an Contaminated Water Vapor Plumes, Internal
document, U. S. Environmental Protection Agency, Durham,
N. C., 1973.
19. Average Observational Error Associated With Smoke Plumes At Levels
Of Known Opacity, Internal Document, U. S. Environmental Protection
Agency, Durham, N.C. 1973.
20. Letter From Martin Marietta Cement Company (Formerly Dragon Cement
Company), Confirming Gas Flow Rates As Measured By EPA During
Test No. 71-MM-05, Dated August 24, 1972.
60
-------
APPENDIX II
EPA RESPONSE TO PETITIONERS'COMMENTS
REGARDING EPA RESPONSE TO REMAND
61
-------
-------
APPENDIX II
Portland Cement Association Comments
Adverse Environmental Impact
Portland Cement Association (PCA) objected to the draft response
to the issue of the environmental impact associated with the electricity
required to operate the control technology. In particular, PCA objected
that the Agency should not have compared the electricity required to
operate the control technology needed to meet EPA standards with the
electricity required to meet State pollution control standards. PCA
believed that State regulations "do not furnish a valid basis to excuse
EPA from its statutory duty because most of those regulations, were in
effect, mandated by EPA itself." EPA did not mandate specific State
particulate emission limitations.
Prior to the date that new source performance standards and
implementation plans were proposed, many progressive State and local
air pollution control agencies had promulgated rules and regulations
that were as strict or stricter than those required by EPA to assure
that national ambient air quality standards would be met. For example,
the State of Maryland had promulgated regulations— that allowed no
visible emissions and mass emission standards that are as stringent
as EPA new source performance standards.
After the Clean Air Act was signed into law, EPA was required to
review State implementation plans to determine whether they would achieve
and maintain national ambient air quality standards, including standards
63
-------
for particulates. In some cases, EPA promulgated emission limits where
the State plans did not adequately achieve or maintain the national
ambient standards. However, the State was free to determine which
emission sources should be reduced to meet the ambient standards or
to achieve more stringent ambient levels. It is, therefore, incorrect
that EPA "mandated" specific State particulate emission limitations.
Furthermore, EPA did not limit its consideration of the adverse
environmental impact of the electricity required to operate the control
equipment to a comparison of the impact of EPA's standards with the
impact of State standards. EPA in its remand response compared the
approximately 115 kilowatts required to operate a precipitator that
would meet the new source performance standards with the total electric
power rate consumption of a typical cement plant which approximates
5,300 kilowatts and with the 500,000 kilowatt output of a typical new
steam generating electric power plant. EPA believes that this comparison
alone indicates that a detailed discussion of the incremental and environmental
impact associated with the additional 115 kilowatts required to meet the
new source performance standards is unwarranted. Furthermore, EPA would
note that the 115 kilowatt increment is based upon the assumption that
emissions would otherwise be totally uncontrolled. It is nearly inconceivable
that States and/or local governments would permit Portland cement plants to
be operated without any particulate control whatsoever but for EPA's require-
ment of such control.
PCA also objects that EPA's consideration of the environmental impact
of water pollution from the run-off from kiln dust piles has been
inadequately considered. First, it should be noted that water
64
-------
pollution resulting from particulate matter is not a question of whether
water pollution occurs, but where it occurs. Uncontrolled emissions
deposit the particulates throughout the countryside where it then washes
into water supplies. On the other hand, removing particulates from exhaust
gases enables the source to collect the particulate matter and thus control
water run-off of the particulate matter should it occur.
Second, much of the particulate matter collected by emission control
devices can be used by Portland cement plants. Thus, the total amount of
particulates disposed of will be less if collected by emission control
devices than if vented uncontrolled into the atmosphere.
The technology for control of water run-off from dust piles is
similar to that required for coal pile drainage, yard drains, and run-off
from construction. These streams must be wholly or partially detained during
storm events in order to minimize overall treatment costs. Diking, ditching,
and holding technology can practically be utilized in conjunction with
facilities for sedimentation or filtration for solids removal and for pH
8/
control."" EPA does not believe that water run-off from piles of particulate
emissions from Portland cement plants is a major problem. To the extent
there is a problem, it is the judgment of the Administrator that the
problem of water run off from collected piles of particulate matter is
less than the problem of uncontrolled releases of particulate matter into
the atmosphere.
Cost-Benefit Analysis
PCA argues that EPA should do a cost benefit analysis. Such an
analysis has not been made because reasonable means for doing so do not
yet exist, in the opinion of EPA. An emission limitation on a specific
type of emission source will have different effects on ambient levels in
different places. Ambient levels are affected not only by emission
65
-------
levels of specific sources but also the placement of such sources
relative to other sources and to meteorological conditions around
each of such sources. Furthermore, sources vary in their geographical
relationships to populations, wildlife, vegetation and other factors
which would be impacted by the emission source. Thus relating the
cost of control to the benefits of the control at least at this time
is a practical impossibility.
We believe that Congress intended §111 to be used to establish
emission limitations which are the lowest that can be achieved by that
industry. Where the cost of meeting such standards would be so great
that the industry could not bear the costs and survive, such standards
could not be implemented by the industry regardless of technological
feasibility. There are also cases where a reduction in emissions that
would be achieved by a particular control technique would be so grossly
disproportionate to the cost of the control technique that it would not
be required. However, these determinations are necessarily judgmental
and are left specifically to the Administrator in §111. EPA does not
believe that cost benefit analysis can be established for standards
under §111 nor does it believe that Congress intended that standards be
established or justified on the basis of such an analysis.
Although EPA has not conducted a cost benefit analysis for these
standards, we have considered PCA's criticisms of the standards based
on cost benefit analysis. PCA relies heavily on a conceptual chart
published in an HEW publication: Control Techniques for Particulate
Air Pollutants, Pub. No. AP-51. This chart shows a smooth curve relating
"Capital Costs" (although no units are shown on the chart) to efficiency
of the control techniques. The chart shows the curve rising asymptotically
as efficiency approaches 100 percent. However, this chart does not represent
66
-------
the realities of control of participate emissions in the Portland cement
industry. The primary method of control relied on by EPA at the time the
standards were developed was control by baghouses. Baghouses when properly
designed, installed, and maintained operate at nearly the same collection
efficiency, thus the relationship between cost and efficiency is somewhat
inelastic. It is well established that properly designed and maintained
baghouses can limit emissions to .03 grains per actual cubic foot (ACF)
of emission air and gases. EPA tests confirmed that baghouses already
installed on Portland cement plants were capable of achieving .03 gr/acf
if properly maintained and operated. In fact, EPA believes and recent
tests confirm that baghouses can achieve higher removal efficiencies.
However, at some point of control where emission levels are less than
.03 gr/acf the removal efficiencies will not increase to any great extent
regardless of the cost of the equipment. It is only coincidential that
the standard for Portland cement plants results in costs that fall below
the level where costs rise sharply compared to increased removal efficiency.
This type of cost-effectiveness analysis was neither the basis for setting
the Portland cement standard nor any other new source performance standard.
In general, a standard will not be established at a level at which the cost
of compliance would result in the inability of new plants to compete with
established firms in the industry. This type of analysis is not based upon
cost effectiveness but rather upon the economic structure of the industry as
a whole.
PCA commented that if cost benefit comparisons were made between the
standards for Portland cement plants on the one hand and coal-fired power plants
and incinerators on the other, they would indicate that the Portland cement
standards should be loosened and the others should be tightened to achieve a
cost benefit equality. The particulate standards for coal-fired power
plants and incinerators were based on the use of precipitators. Although
67
-------
there were indications at the time the standards were promulgated that
precipitators could be used on Portland cement plants and baghouses
could be used on coal-fired power plants and incinerators, EPA based
its standards on the methods, which had been established as applicable
to the source subject to the standard. Furthermore, EPA based its
particulate standards for power plants and incinerators on emission
levels being achieved at existing plants. Because the removal efficiencies
being achieved by properly maintained and operated baghouses on Portland
cement plants were superior to those being achieved by power plants and
incinerators, the effective emission concentration limits were more
restrictive for Portland cement plants.
It now appears that pre'Cipitator technology may achieve the same
levels of control, i.e., below .03 gr/acf, required of the Portland
cement industry and that baghouses can be applied to the other two
industries. EPA will consider whether the particulate standards for
the other two industries should be tightened. However, EPA does not
believe that inequalities between the effective emission concentration
limits between these three types of sources should be eliminated by
loosening the standards applicable to Portland cement plants as urged by
PCA to permit such plants to emit an additional "10,000 tons per year"
(PCA's figure) of particulate matter to the atmosphere.
Cost Estimate
PCA commented that EPA underestimated the cost of achieving the
standards. PCA particularly objected that EPA did not use cost figures
supplied by the industry. E!PA does not agree with the industry's cost
estimates. The costs presented by PCA did not speak specifically to the
costs resulting from the standard of performance for new cement plants.
The estimates reflect the total costs necessary to meet State emission standards
68
-------
and the Federal standard. It is not possible to separate the two cost
estimates in the submittal by PCA. In addition, PCA spoke only to the
investment cost for pollution control and did not present any costs for
operating and maintaining the equipment. As a result, the cost estimates
submitted by PCA were of no value in estimating the costs resulting from
the standard of performance.
A study of the economic impact of pollution control costs upon the
cement industry was performed by the Boston Consulting Group and
2/
published in 1971.- One of the conclusions reached was as follows:
"The combined effects of plant size advantage, required
pollution control expenditures on existing plants, and
labor cost pressure will motivate a rate of plant construc-
tion in excess of the rate of demand growth, which itself
will exceed that of the past decade. The consequence will
be a substantial increase in plant construction, virtually
all of which will be in plants with kiln sizes of 2.5 million
barrels and above."
Although the costs used in this study were supplied by EPA, the report
goes on to state:
"If the industry estimates of total and anticipated pollution
control costs were used, instead of the EPA figures, the fore-
going conclusions would not be changed, but would be intensified."
In summary, the Boston Consulting Group study forecasts an increase in
new construction, not a decrease.
PCA further argues that inflation will change the cost figures.
It is true that inflation will change the absolute magnitude of the cost
figures. It is also true, however, that cement prices have risen at a
rate at least equal to the overall rate of inflation. The existence
of inflation is not a factor that invalidates the cost impact originally
presented since it tends to affect all items, revenues as well as costs,
69
-------
and leaves the relative impact unchanged.
PCA also criticizes the Boston Consulting Group study on the basis
that it considered a model plant rather than "widely ranging
variations in raw material availability fuel costs and availability,
and regional marketing considerations." PCA argues that EPA should show
that capital will be obtainable to finance pollution control equipment.
Raising capital does not appear to be much of a problem for cement
plants, even for pollution control equipment. The following comments are
taken from the January 1974 volume of Pit and Quarry, an edition that
highlighted past and future developments in the cement industry: "No
new plants went into operation in 1973. There were, however, eight major
expansions involving new kilns, at least three of which might be considered
as a new plant alongside an older one. In all, seven new kilns and 17
grinding mills are known to have been installed during the year." Obviously,
capital was available for these expansions,
Specifically with regard to pollution control equipment, the Pit and
Quarry article mentioned above states that "1973 was another record year
for the installation of dust collectors. At least 34 electrostatic precipitators,
17 baghouses, and 2 gravel bed filters are known to have been installed, most
of them on kilns, but some also on coolers. These figures are far above
those in 1972 and 1971, which were record installation years. The rate
of installation of smaller collection units was also undoubtedly ahead of
any previous year." In light of the above, we find it difficult to believe
that capital availability is a problem in the cement industry.
EPA believes it has adequately considered cost even though it has not
conducted an analysis of the costs for pollution control for each and every
projected new plant and has not shown where and how the capital for pollution
70
-------
control equipment is to be raised by such plants. The cost of pollution
control is one of several factors that might affect a company's decision
to locate a plant at a given site. EPA does not believe that §111 demands
the finely tuned cost analysis demanded by PCA. Furthermore, the agency
believes that a plant by plant analysis is impractical and that the model
plant representing average facilities is the best approach.
Demonstration of Achievability
PCA criticizes the testing procedures which EPA used to determine that
the standard can be achieved. PCA admits the standards can be achieved.
They state that:
"Petitioner does not contend that the .03 gr/scf Standard
cannot be achieved. Indeed, as stated in Section II above, it
is believed that efficiencies in excess of 99.9 percent can be
achieved with existing technology and available equipment applied
to many traditional stationary sources of particulate pollutants,
including cement plants." (PCA Comments to EPA draft response,
p. 13, note 7.)
This should be dispositive of this issue. However, EPA has considered
each specific criticism of PCA on this issue and for purposes of developing
a complete response, will address each issue here.
Apparently the bases of PCA's criticisms are that some tests showed
the standards were not being met, that each test that showed the standards
could be met had some defect or was not adequately verified by EPA and in
particular that EPA discontinued testing one plant (the Oregon plant) that
was poorly maintained until broken bags were tied off.
PCA apparently construes §111 as requiring EPA to set a standard of
performance for new sources that is the lowest emission level achieved
at any existing plant regardless of how poorly the pollution control
equipment is maintained or operated. The purpose of §111 is clearly to
require new sources to use the best pollution control equipment which
new sources can obtain. In order to determine what emission levels
71
-------
new sources can obtain, EPA is required to look at the best equipment
used by existing plants, not the average and certainly not the worst.
Furthermore, §111 standards are not required to permit new sources to
use inadequate maintenance simply because some existing plants are
inadequately maintained.
Turning now to specific criticisms by PCA of EPA's testing, PCA
argues that at the Dragon cement plant test the admitted presence of
settled dust in the duct at the test point raises further doubts as
to the accuracy of the results obtained. The Dragon cement plant test
was conducted by expert source test personnel. The settled dust in the
duct was recognized. The layer of dust reduced the cross-sectional
area and this was considered in calculations for the total volumetric
flow rate.
PCA objects that the location of the test point in the Dragon cement
plant did not comply with the EPA test procedures set forth in Federal
Register Vol. 36, No. 247, Section 60.85.
The test conducted at Dragon Cement plant was conducted before test
procedures were detailed in the Federal Register for new plants. It must
further be realized that testing had to be conducted at existing plants
to show what levels of emission control were achievable by new plants.
EPA could not construct well-controlled plants with proper duct configurations
solely for standard-setting purposes. It is conceded that ductwork at the
plant was not ideal; however, in order to compensate for this, sampling
was conducted at a larger number of traverse points than would normally
be used at a test location with a suitable ductwork configuration and
a precision low-range inclined manometer was used for the velocity
measurements. This type of manometer permitted a higher standard of
precision by expanding the scale to improve readibility and sensitivity
72
-------
of the velocity head measurements. In addition, three separate test
runs were conducted, each of which indicated compliance with the standard.
If results of each test run had varied widely, then there could have
been serious doubts concerning the test results. This was not the case.
PCA objects that the Orsat analysis which EPA admits was inaccurate
affects the results of the source test conducted at the Dragon Cement
plant. While petitioner is technically correct, the inaccurate Orsat
analysis introduced only a small error due to the slightly different
density of the gases. It appears that the actual carbon dioxide content
in Test No. 1 was 15.4 percent instead of 7.5 percent as reportedr There-
fore, the dry molecular weight of the gases would have been 30.93 pounds
per cubic foot rather than 29.82 pounds per cubic foot. This factor
could have introduced an error of no more than 1.8 percent in reported
emission levels. EPA relied on measured values as the critical parameters
§/
in the Dragon test. The fact that one measured value which was not critical
and which is strongly suspect to be in error, would lead to calculated values
which disagree with the measured values does not support excluding the Dragon
test from consideration of whether the standard can be achieved.
PCA objects that the Oregon Portland Cement plant had an estimated
15 percent of the bags tied off thus creating an unrepresentative operating
condition. EPA was faced with the necessity of testing existing Portland
cement plants to set a standard for new sources. The record shows that
most existing plants were poorly maintained. Although the plant operators
had been notified that their plants were going to be source tested for new
source performance standards, both the Oregon Portland Cement plant and the Giant
Portland Cement plant operations were being operated with holes in their bag-
house bags at the time that source testing was scheduled. Abnormalities
a/ As noted on page 1-12 of tiie proposed response to remand order bv U.S.
Court of Appeals for the District of Columbia, even if Mr. Striker were
correct, the particulate emission test results for the maximum test would
be 0.22 Ibs/ton, 30 percent less than the standard.
73
-------
in their baghouse operation were detected by visible emissions. In
both cases, EPA engineers recognized these excessive emissions, stopped
the testing, and requested plant maintenance personnel to enter the
baghouse and tie off bags that were leaking excessive emissions of dust
through holes in the bags.
In the case of the Oregon Portland Cement plant, it was estimated
that 15 percent of the bags were tied off; however, even with this
relatively large number of bags tied off, the baghouse control system
was capable of ,being operated (with the kiln in normal operation) within
the limits of the proposed mass emission and visible emission standards.
PCA argues further on this point that:
"Special repairs made for the purpose of this test . . .
present an ethical issue similar to the Ford Motor Company
maintenance procedures for which Ford was prosecuted. ..."
(PCA Comments, p. 15, note 9).
EPA believes that the ethical distinctions between adjusting equipment
at EPA's request and unauthorized tampering with cars undergoing certification
tests are self evident.
PCA criticized reliance on the Marquette and Lehigh tests in Maryland
because they were performed on new facilities just brought up to production
rate. PCA believes that these were ideal, rather than representative,
conditions and therefore shouldn't be relied on. It is pointed out that the
cover letter from the State of Maryland states that the equipment for both
plants was installed new in 1970. The source tests were not conducted until
1972.
Source tests for compliance with new source performance standards
will be conducted on new or substantially modified facilities. There
is no reason that control equipment cannot be installed, maintained,
and operated as necessary to meet the standard long after the plant has
74
-------
been in operation. Tests of the Dragon and Oregon plants confirm this.
PCA argues that EPA did not conduct the source tests on the Marquette
and Lehigh plants. They claim EPA shouldn't rely on the report of another
agency. The test reports for tha Marquette and Lehigh Portland Cement
plants were official documents submitted by the companies to a State air
pollution control agency. The test reports state that the tests were conducted
following EPA test methods and test results indicated compliance with the
NSPS. Tests to determine compliance with the standards in question would
be conducted in essentially the same manner.
Subsequent to November 26, 1973, the U. S. Environmental Protection
Agency has received additional source test data and information indicating
that additional Portland cement plant kilns and clinker coolers are capable
of being operated in compliance with new source performance standards.
It is pointed out that this data and information has not been presented
to the petitioners because of the time element involved; however, because
of its relevancy, the data are summarized in Table I.
Table I
Emissions in Lbs/Ton
Test Date Name of Company of Dry Feed
6/25-27/74 United Cement Company^/ 0.209
Artesia, Mississippi,
kiln, wet process,
electrical precipitator
8-20-74 United Cement Company,- 0.04
Artesia, Mississippi,
clinker cooler, cyclone and
wet scrubber.
5/23-24/74 Santee Portland Cement Company-' 0.107
Holly Hill, South Carolina,
kiln, wet process, two
electrical precipitators
-------
Emission in Lbs/Ton
Test Date Name of Company of Dry Feed
6/12-14/74 Monarch Cement Company- 0.196
Humboldt, Kansas, kiln
No. 4, dry process,
baghouse
6/12-14/74 Monarch Cement Company i/ 0.0174
Humboldt, Kansas, clinker
cooler, No. 4 kiln system,
baghouse
76
-------
NORTHWESTERN STATES PORTLAND
CEMENT COMPANY COMMENTS
Comments on Affidavit of Mr. Striker
Mr. Striker commented rather extensively on the Orsat analysis used
at the Dragon Cement plant test. Both Mr. Striker's comments on this
issue and EPA's response-are highly technical and are not included here.
Copies of both of these documents may be obtained from the Emission
Standards and Engineering Division, Research Triangle Park, North
Carolina 27711, Attention: Mr. Don R. Goodwin. It will only be noted
here that EPA disagrees with Mr. Striker's conclusions. Furthermore,
even if Mr. Striker's conclusions were correct, the Dragon test still
met the standard as did the tests taken at three other plants, as noted
in the text of this response.
Mr. Striker comments as follows: "The standard is inferentially based on
the premise that all cement plants are essentially alike although this
is not the case. Every plant is custom designed and a
unique operation." Nearly every stationary source is a "unique operation".
Availability of local materials and local labor force as well as siting
and other unique constraints vary with each plant. It would be infeasible
and Congress did not intend for EPA to set a different standard for each
new Portland cement plant. However, it is the judgment of EPA that all
new plants can meet the standard regardless of raw materials used.
Control equipment installed to minimize particulate emissions from Portland
cement plant kilns are not off-the-shelf type items. Each unit must be
carefully engineered and consideration must be given to all variables
that can affect the efficient operation of the equipment.
77
-------
Mr. Striker hasn't shown that the four plants which were found
to meet the particulate standard were unique in any respect. Neither
has he shown that new plants which install equally effective controls
won't meet the standard. Operators have reasonable options in
choosing sites and raw materials for new plants. If operators were
in fact convinced that there were factors which adversely affected
particulate emissions, these could be avoided in the building of new
plants.
Mr. Striker claims that independent source tests and summary test
data submitted to EPA by Dragon Cement Company are of little value.
It is EPA's opinion that sumrrary test data submitted by Dragon Cement
Company showing the results of a test conducted independently by the
Company which tend to supper*, the standard are of great value. Further-
more, the volume of gases, the dust concentration, and emissions reported
in the summary agree favorably with EPA test results.
Mr. Striker believes the Dragon Cement test should be free from all
doubts and should stand up under critical inspection from any standpoint.
The Dragon Cement results are as free from doubt as almost any other
source test. Examination shows that the configuration of the ductwork
was not ideal but that EPA modified the sampling procedure as much as
possible to offset these conditions. In addition, all reasonable analysis
including information and tests supplied by the operating company confirm
that the exhaust volume rate is as measured in the EPA test,not the much
larger value predicted by Mr. Striker. We believe the operating company
is much more knowledgeable of the kiln system than Mr. Striker.
Mr. Striker suggests the Lehigh Portland Cement tests on No. 4 kiln
at Union Bridge, Maryland, appear erroneous. Mr. Striker has provided
78
-------
no basis for this conclusion. As in other instances, it appears that
estimates were used to make theoretical determinations that do not
agree with the measured values. Mr. Striker has submitted no factual
information to prove that his determinations are correct.
79
-------
Comments of Mr. Photlacles
Mr. Photiades commented that manufacturer's warranties are limited.
The inference of this comment is that such warranties should not be
greatly relied on to support the standard. As EPA noted in the text of
its response, vendor representations were not of critical importance in
supporting the standard. However, since a vendor has a reputation
to maintain as well as a financial obligation to meet the strict terms of
a warranty, representations of a reputable vendor that its equipment will
meet the standard is of some value in supporting the standard even if the
warranty of its equipment is limited. Vendors usually limit their warranties
not because of doubt about the reliability of the equipment so much as due
to the fact that they have no control over operation and maintenance of
the equipment.
Mr. Photiades also comments that inflation and other factors have
caused prices to increase since the time EPA promulgated the standard.
Mr. Photiades cites bids made on a Northwestern baghouse to support
his contention. EPA cannot disagree that costs have risen due primarily
to inflation over the 1971-1974 period; however, it is also pointed out
that the costs of other cement plant components as well as the revenues from
&/
the cement sold have likewise risen over the same period. No information
or data have been presented that EPA estimates of costs as shown in
the document entitled, "The financial Impact of Air Pollution Control
Upon the Portland Cement Industry" by R. J. Elias and J. M. Dement are
grossly in error. The bids made on one baghouse neither prove the study
is invalid nor that the standard snould be changed, particularly when
the normal variability in capital expenditures is considered. It is
also important to note that Mr. Photiades did not provide sufficient
b/The average cement price rose from $3.52/bbl in 1971 (Minerals Yearbook,
~I972 Edition, page 259) to $5.79/bbl in 1974 (Engineering News-Record,
9-5-74).
80
-------
detail in his estimate to enable EPA to separate the financial impact
resulting from the State Implementation Plans from the impact resulting
from new source performance standards.
Comments of Mr. Shepard
Mr. Shepard, counsel for Northwestern criticizes EPA's reliance
on the Oregon test because EPA stopped the test until broken bags
were tied off. He believes that this prejudiced the test because
in his opinion broken bags represent "normal operating procedure"
which must be reflected in the test and therefore the standard. He
also expressed concern that EPA did not state that tests were not
under normal operating procedures until he saw the draft EPA response
dated November 26, 1973.
EPA did not and does not consider inadequate maintenance or kiln
operation fixed at a lower than normal feed rate as normal operating
procedure which must be permitted under the standard. As for Mr. Shepard's
comment that he didn't know that EPA discontinued the test until broken
bags were tied off, even if that were material he has had an opportunity
to comment on this issue as his clients' comments on the draft of this EPA
response to the remand were solicited. Indeed he has commented on the issue,so it
is not clear how his clients were prejudiced by any alleged failure of EPA
to supply information at some point in the past.
Mr. Shepard also suggests that EPA's request to the Oregon operator
to tie off its bags is equivalent to Ford Motor Company's unauthorized
tampering with cars undergoing Federal pollution control tests for which
Ford was fined $7,000,000. As noted earlier in this Appendix (page 74),
the distinction between the activities is self-evident.
81
-------
REFERENCES FOR APPENDIX II
1. State of Maryland, 43P04 regulations governing the control of air
pollution in area III (as amended through December 23, 1971).
2. The Cement Industry - Economic Impact of Pollution Control Costs,
Boston Consulting Group, November 1971, Vol. 1, p. 11.
3. Emission Testing Report, United Cement Company, Artesia, Mississippi,
kiln, wet process, electrical precipitator, tested by company per-
sonnel, 1974.
4. Emission Testing Report, United Cement Company, Artesia, Mississippi,
clinker cooler, cyclone arid wet scrubber, tested by company person-
nel, 1974.
5. Emission Testing Report, Santee Portland Cement Corporation, Holly
Hill, South Carolina, kiln No. 2, wet process, two electrical
precipitators, tested by Pollution Control-Walther, Inc., 1974.
6. Emission Testing Report, Monarch Cement Company, Humboldt, Kansas,
kiln, dry process and clinker cooler, two baghouses, tested by the
Fuller Company, 1974.
7. Memorandum from Roger T. Shigehara and Clyde E. Riley to Robert T.
Walsh, dated April 23, 1974, corrected July 26, 1974.
8. Development Document for Effluent Limitations Guidelines and New
Source Performance Standards for the Cement Manufacturing Category,
U.S. Environmental Protection Agency, EPA-440/l-74-005-a, 1974.
82
-------
APPENDIX III
PORTLAND CEMENT REMAND RESPONSE
83
-------
-------
PORTLAND CEMENT REMAND RESPONSE:
APPENDIX III, PART A: OPACITY STANDARD
In the case of other points at issue in this Remand Response,
it has been necessary only to supplement the November, 1973 Draft
Response to deal with the comments of the Portland Cement Association
and others on that Draft Response. In order to respond fully to the
comments with respect to opacity, it has been necessary to prepare the
rewritten, extensive response found in this Appendix III. The response
contained herein should therefore be considered to supercede that portion
of the Draft Response dealing with the opacity standards, and the
present Appendix III constitutes the Administrator's full and complete
position with regard to this issue.
EPA promulgated opacity limits as well as mass emission limits
primarily because EPA believes that opacity limits provide the only
effective and practical method for determining whether emission
control equipment, necessary for a source to meet the mass emission
limits, is continuously maintained and operated properly. It has
not been EPA's position that a single, constantly invariant and
precise correlation between opacity and mass emissions must be iden-
tified for each source under all conditions of operation. Such a
correlation is unnecessary to the opacity standard, since the
Administrator consistently develops opacity standards for each class
of sources at a level no more restrictive than the corresponding mass
emission limitation with due consideration given to all conditions
85
-------
123
of operation. ' ' Any source which is meeting the mass limit will there-
fore also be meeting the opacity limit; if a source is not meeting
the opacity limit, this will be due to the failure of that source to
properly maintain and operate its emission control equipment. And if
the equipment necessary to meet the mass emission standard is not
properly maintained and operated, the source will also fail to meet
the mass emission standard.
Opacity standards are a necessary supplement to mass emission
standards, since there is no reliable instrument to directly
measure mass emissions currently available. If only mass emission
standards were in effect, operators would not be compelled to assure
proper maintenance and operation of pollution control equipment at
times other than during periods of performance testing. Two weeks
or longer are necessary to schedule a typical stack test, due to the
preparation involved. If relatively minor repairs were all that
were needed in a given case, such as pump or fan repair or replace-
ment of fabric filter bags, a plant operator could delay such
repairs and continue to pollute until shortly before the stack test
was conducted. For some types of control equipment, such as
scrubbers, the operator could reduce the input of energy (the pressure
drop through the system) thereby partially or totally defeating the
air pollution control system when stack tests were not being
conducted and through such action could release significantly more
particulate matter than permissible under the mass emission standards.
86
-------
Furthermore, stack tests can only be conducted with limited frequency,
since they often cost up to $10,000 to perform. Therefore, EPA has
required that operators properly maintain and operate air pollution
control equipment at all times [40 CFR 60.11(d)J and meet opacity
standards at all times except during periods of startup, shutdown,
malfunction, and as otherwise provided in the applicable standard [40
CFR 60.11(d)]. Opacity is established as an independent, enforceable
standard, rather than merely as an indicator of maintenance and
operating conditions, because information regarding those conditions
is peculiarly within the control of the operator. The time and expense
required to prove that proper procedures have not been followed are so
great that the provisions of 40 CFR 60.11(d) by themselves (without
opacity standards) would not provide a practical means for ensuring
that on a day-to-day basis the pollution control equipment is main-
tained and operated in the way necessary to meet the mass emission
limitations. Opacity standards, in contrast, require only a qualified
observer and can be performed with no prior notice. In some cases, of
course, a full investigation of operating and maintenance conditions
will be desirable. Accordingly, the opacity standards and maintenance
requirements were both promulgated, and work in tandem to guarantee
that proper maintenance and operation of pollution control equipment,
the sine qua non of continuous compliance with emission limits, can in
fact be required and monitored.
Two primary aspects of the opacity standards, raised by Petitioner
Portland Cement Association, were identified by the Court of Appeals as
87
-------
the bases of its concern: 1) the inherent reliability of opacity
standards, i.e. whether measurements of opacity can be made with
reasonable accuracy, and 2) the achievability of the opacity standard
set for portland cement plants.
RELIABILITY OF OPACITY STANDARDS
Accuracy of Opacity Measurements in Detecting Violations
The Portland Cement Association (PCA) contends that opacity
measurements cannot be made with reasonable accuracy. PCA indicates,
in particular, that individual observer readings under certain con-
ditions are often in error by 10 percent opacity or more, and therefore
that "equivalent opacity—at any level—is an inherently inadequate
and invalid measure to determine compliance with, or violation of an
anti-pollution law." In support of this contention, PCA argues that
observed opacity may be affected by such factors as: dark nights,
contaminated water vapor or detached plumes, lack of contrasting back-
ground (e.g., white plume viewed against an overcast sky), and wind
velocity.
The PCA and other commentators contend also that EPA smoke school
requirements and procedures are inadequate to ensure that qualified
observers will make accurate readings. The procedure whereby training
plumes are shown in advance of certification runs, and under which
observers are allowed to use fixed (non-sky) contrasting backgrounds
are noted as specific deficiencies.
In response to the questions raised by the Court and the PCA,
EPA has reviewed available data concerning the accuracy of qualified
88
-------
observers; ' ' has reviewed the theory of equivalent opacity,
particularly concerning the effect of variable viewing conditions on
r c
accuracy; ' has reviewed the smoke school certification procedures;
and has conducted extensive field tests involving both specially
generated plumes and industrial plumes. On the basis of this extensive
review, EPA concludes that the accuracy of Reference Method 9 has been
well established, that its error tolerance is reasonable and is within
limits considered normal by the scientific and engineering community,
and that when that error tolerance is accounted for in assessing the
results of opacity readings, the method is valid and reliable for
determining compliance with prescribed opacity standards. EPA also
concludes that the smoke school certification procedures are adequate
to assure proper qualification and accuracy of qualified observers. The
basis for these conclusions is set forth below.
PCA contends that plumes of the same actual opacity will appear
to have somewhat different opacities under dissimilar conditions of
viewing. EPA agrees that this is true. A plume will appear most
visible to an observer when the plume is viewed against a contrasting
background. ' A white plume therefore will be more visible when
viewed against a black background or a blue sky than it will when
viewed against a hazy or overcast sky. Indeed, a white plume viewed
against a white background may well be invisible. PCA concludes that
because a plume will appear of lesser opacity under less contrasting
conditions, EPA's opacity standards, which limit allowable opacity
89
-------
even under conditions of maximum apparent opacity, are inaccurate
and unreliable. It is with this illogical conclusion that EPA
disagrees. Cement plant operators are in no way prejudiced by the
fact that under less contrasting conditions the opacity of their plume
may be read at less than its actual value; obviously, this is to their
advantage. Their only real concern, therefore, is with the accuracy
of opacity readings under viewing conditions of highest apparent opacity.
Since opacity appears greatest under conditions of high background
contrast, it is under such conditions that the potential for positive
observer error (observed opacity greater than true opacity) is
greatest. This is verified by the available data, which show that
positive observer error (average of 25 readings) can approach + 7.5
percent under relatively high contrast conditions such as are
encountered at a smoke school, when a training plume is observed
4 7
against a contrasting background. ' Against a less contrasting
background, a plume will be less visible and observer errors will tend
to be negative (observed opacity less than true opacity). A colla-
borative test performed at. a power plant where a white plume was
observed on an overcast day confirmed that this downward bias can be
significant, to the extent that the probability of positive errors
even in single readings is essentially non-existent under such
4
conditions.
It is clear therefore that in assessing the reliability of the
opacity standards, the possibility of positive errors is of primary
concern, since such errors could in theory result in the determination
90
-------
of a violation when true opacity was below the standard. While
negative errors are undesirable from an enforcement point of view,
they do not place a complying source in jeopardy. It is only necessary
therefore to quantify the positive error, which may occur under high
contrast viewing conditions, to demonstrate that the method does
measure opacity with the reasonable accuracy required of an enforceable
standard. As the Court of Appeals notes in Amoco Qi1 Company v. E.P.A.,
501 F2d 72£, (6ERC 1481, May 1, 1974, slip op. at pp. 38-39):
Petitioners also contend that any test methods adopted
will involve a certain capacity for statistical error.
We fail to appreciate the force of this argument. The
possibility of statistical measurement error, which is
often unavoidable where regulations set quantitative
standards, does not detract from an agency's power to
set such standards. It merely deprives the agency of
the power to find a violation of the standards, in
enforcement proceedings, where the measured departure
from them is within the boundaries of probable measure-
ment error. (Emphasis in original.)
The following data provide a basis for such quantification and
lead to the conclusion that, using an average of 25 readings obtained
under the conditions prescribed in Method 9, qualified observers are
able to consistently assign opacity values under relatively ideal
8 9
conditions, with a positive error seldom exceeding 7.5 percent opacity. '
(a) October 1973 report titled "Average Observation Error Asso-
n
dated with Smoke Plumes at Levels of Known Opacity". This report in
addition to showing that each observer met the +_ 7.5 percent require-
ment, shows that when the readings of all observers are averaged, the
largest positive error associated with a given opacity level was 5.0
91
-------
percent opacity which occurred when the true opacity was 10 percent.
For plumes between 15 percent and 90 percent opacity, the average error
did not exceed + 2.8 percent opacity. This is a clear demonstration of
the inherent accuracy of the technique of determining plume opacity
through observation by qualified observers.
(b) In a test conducted November 19-20, 1973, nine qualified
observers were shown plumes generated by a smoke generator. The plumes
were presented in runs of 25 each during which each observer assigned
opacity using the procedures set forth in Method 9. A total of 133
sets of 25 observations using black smoke and 170 sets of 25 observa-
tions using white smoke were made. The results of this test are as
follows. For white plumes, 99 percent (168 of 170) of the sets of
observations were made with an error of less than + 7.5 percent opacity.
Ninety-five percent of the sets of observations were made with an error
of less than +5.8 percent opacity. For black plumes, the accuracy was
even better: 99 percent of the sets of observations were made with an
error of less than + 5 percent opacity; 95 percent were made with an
error of less than +2.7 percent opacity.
(c) A test was conducted at a power plant in Charlotte, North
Carolina, on September 30, and October 1, 1974. In this test, 8
qualified observers made a total of 168 sets of 25 observations and
92
-------
assigned opacity according to the procedures of Method 9.* Each run
consisted of 25 individual readings. All observers had been certified
according to Method 9. Two of the 8 observers had not had previous
field experience and two had only limited field experience. The power
plant plume (which was monitored with an in-stack transmissometer) was
white in appearance and was viewed against a constrasting clear blue
sky (no condensed water vapor was present in the plume). The plume was
held relatively constant during some runs and was varied during other
runs in order to simulate the various conditions which could be
encountered in the field. The test results showed that 99 percent
(166 of 168) of the observations (average of 25 readings) were made with
an error of less than + 6.2 percent opacity; 95 percent with an error of
less than + 4.7 percent opacity, and 90 percent with an error of less
than +2.6 percent opacity.
(d) A test was conducted at a sulfuric acid plant on January 29-30,
1974. The plant was equipped with an in-stack transmissometer.**
* This test was conducted with the 8 observers divided into two groups
of 4 each. One group, the control group, made all readings from an
optimum viewing location. The second group made readings under differing
conditions. Only the readings which were made according to Method 9
criteria are reported here, except that 32 sets of readings are also
included which did not conform precisely to Method 9. In these sets,
observers were located at an optimum location and assigned opacity in
1 percent increments, not 5 percent increments as specified in the
method. The only other data collected but not included in the totals
shown here, was read with the sun outside the 140° sector to the observer's
back. This is not acceptable under Method 9.
** Not required at sulfuric acid plants by EPA standards.
93
-------
Observer readings were made according to Method 9 against a contrasting
background and were compared with the transmissometer readings. The
results are as follows: A total of 298 sets of 25 readings were made.
Two sets of readings were made with an average error of + 8 percent
opacity. A total of nine sets of readings were made with an average
error of + 5 percent opacity or greater. The remaining 289 sets of
readings were made with errors of less than + 5 percent opacity.
In summary, a total of 636 runs of 25 readings each were made
using white plumes and 133 runs of 25 readings each using black plumes
for the specific purpose of quantifying the degree of positive error
possible in Method 9. The readings were made against contrasting
backgrounds, the condition under which a plume is most visible and is
most conducive to positive observer errors. The results show that
single observers, on the basis of single sets of 25 readings, are able
to assign opacity with errors of less than +7.5 percent with good
consistency (630 of 636 for white plumes; 133 out of 133 for black
plumes). Errors in excess of + 7.5 percent can occur, although very
infrequently. In short, on the basis of this extensive research of
the very sort demanded by PCA, observer readings of opacity have been
shown to be within an error tolerance of + 7.5 opacity with a frequency
of approximately 99.3 percent (763 of 769 total sets of observations).
This indicates clearly that opacity observers do in fact meet the +7.5
percent opacity accuracy limit which is specified explicitly in the
introduction to (amended) Reference Method 9.
94
-------
Finally with regard to observer accuracy and the inherent relia-
bility of opacity readings, much of the concern of the Court of Appeals
is related to PCA's contention, based on an HEW study, that at low
levels opacity readings cannot be made reliably. The Court notes that
PCA's claim that Table 2 on page 28 of that study shows that the six
trained observers rated a plume known to have 0 percent opacity at more
than 0 percent opacity, and three rated it at more than 20 percent.
The Court notes further that a plume known to be 20 percent opacity was
rated higher than 20 percent by five of the six observers, based also
on Table 2.
A close reading of Table 2 shows, however, that Petitioners have
wholly misconstrued the nature of that test and the data presented in
Table 2. In fact, the study specifies that
"tests of trained observers show that the observers that
viewed white plumes on a clear day facing the sun assessed
the plumes at a higher Ringelmann number (lower trans-
mittance) than did observers that viewed the plumes with
the sun to their backs. For darker, plumes the effect was
less pronounced.Group assessments showed good agreement
for similar sun-plume-viewer geometries'^ (emphasis added).
EPA agrees that viewing plumes while facing the sun may produce inaccurate
o
readings. It is for this very reason that Reference Method 9 requires
observations to be made with the sun at the viewer's back.
When Table 2 is examined in this regard, the "good agreement" for
apparent similar geometries divides the observers into two groups. The
first, consisting of Inspectors 1, 2, and 3, had similar readings among
95
-------
themselves. The second group, consisting of Inspectors 4, 5, and 6,
also had good agreement among themselves. The levels of the second
group are significantly higher than those of the first group, indicating
that the second group was viewing into the sun, while the first group (1,
2, and 3) was viewing with the sun to their backs. This probability is
confirmed by the report itself, which states that no restriction was
placed on viewing direction. Since only Inspectors 1, 2, and 3 appa.-
ently made their observations with the sun at their backs, it is only
their results that are at all relevant to the question of observational
reliability or accuracy. These results, for the low levels of opacity
concerning the Court, are reproduced below:
Table 2. AP-30. Evaluation of White Plumes
In-stack transmittance, percent 75%
Equiv. Ring. No.
Insp. No. 1
Insp. No. 2
Insp. No. 3
1.00
1.03
1.13
0.97
90%
0
0.03
0.16
0.17
It is at once apparent that the observations of Inspectors 1,2,
and 3 reflect a very high degree of accuracy, contrary to PCA's conten-
tions. Closer examination of this table, moreover, shows that the
data itself is presented in a way not corresponding with EPA's
standards. It will be noted that each column is headed by two
figures: the "equivalent Ringelmann No." and the "In-stack Trans-
mittance, Percent." Transmittance, of course, represents the
percentage of light not blocked by the smoke, so the 75 percent
mittance is a 10 percent opacity. As the Court notes, the presently
96
-------
accepted equivalency between Rinaelmann numbers and opacity of white
smoke is 20:1, i.e. Ringelmann No. 2 = 40 percent opacity, Ringelmann
No. 1 = 20 percent opacity, Ringlemann No. .5 = 10 percent opacity,
Ringelmann No. 0 = 0 percent opacity, etc. The Table 2 column headings,
however, reflect a different equivalency: there, Ringelmann No. 1
corresponds to 25 percent opacity (75 percent transmittance), and
Ringelmann 0 corresponds to 10 percent opacity (90 percent transmittance),
Since EPA's opacity standards are predicated upon and expressed in terms
of the transmission and obscuration of light, it is therefore only the
transmittance figures which are at all relevant to evaluating EPA's
standards.
The discrepancy between the equivalency used in Table 2 and the
presently accepted Ringelmann-opacity equivalency exists because during
the middle nineteen-sixties when the data were obtained, there was little
interest in reading low opacity plumes, since visible emissions stan-
dards were generally at about forty percent opacity. The tests were
conducted with smoke readers at a smoke school where^the Ringelmann
Chart was considered the primary standard, and the school's smoke genera-
tor transmissometers were calibrated in terms of Ringelmann numbers (a
procedure not used today). This nonlinear calibration resulted in
white plumes below thirteen percent opacity being assigned a Ringlemann
number of zero. Black plumes below five percent opacity were also
assigned a Ringelmann number of zero. It is clear therefore that this
equivalency (now replaced by the standard 20:1 ratio) was not intended
-------
to deal with low opacity plumes, since all six inspectors in fact
reported seeing smoke at "what was supposed to be Ringlemann zero.
As a result of these flaws, the data in Table 2 can only be
partially converted to opacity percentages for purposes of evaluating
observer accuracy and are not completely reliable for this purpose as
the observations were not made in accordance with current procedures
specified by Method 9. Since a definite ratio of 25:1.00 was used to
calibrate the observers at the 25 percent level, those figures
can be expressed as opacity percentages. The conversion for the 10
percent opacity column is less exact, but the fact that all observers
did observe smoke with a known 10 percent opacity plume allows the use
of the standard ratio (Ringelmann .5=10 percent opacity) as a close
approximation for purposes of conversion. The results of this conversion
appear below.
Table 2. AP-30. Evaluation of White Plumes6
In-stack transmittance, percent 75% 90%
Opacity, percent 25% 10%
Inspector No. 1
Ringelmann observed 1.03 0.03
Equivalent opacity 25.75% 0.6%
Inspector No. 2
Ringelmann observed 1.13 0.16
Equivalent opacity 28.25% 3.2%
Inspector No. 3
Ringelmann observed 0.97 0.17
Equivalent opacity 24.25% 3.4%
It is apparent from this more properly presented data that Table
2 in fact refutes, rather than supports, PCA's claims of unreliably
98
-------
high readings at low levels rf opacity. Indeed, the column which PCA
treated as 0 percent opacity is in fact 10 percent opacity; in this
column all three inspectors rated the opacity at well bejow its actual
value. Likewise, the column which PCA treated as 20 percent is actually
25 percent opacity. In that column, all observers rated the opacity at
values very close to the actual level, and well within the 7.5 percent
opacity tolerance to which observers are limited by Method 9.
Because of the fundamental differences in the method of relating
and calibrating Ringelmann numbers with equivalent opacity between the
method used at the smoke school where the above data were gathered and
the present approach used by EPA (as reflected in Method 9), the
author of AP-30 states that the data in Table 2 cannot be used to
assess the reliability of EPA's method or opacity readings in general.
The above discussion is therefore presented solely to demonstrate that,
contrary to PCA's belief, those data do not in any way indicate the
unreliability of EPA's use of qualified observers. The Agency is not
relying on those data as part of its demonstration of the accuracy of
Method 9.
In accordance with the Amoco case discussed above, this error
tolerance of + 7.5 percent opacity must be accounted for in the enforce-
ment process, so that no source which is in fact in compliance with the
applicable opacity standard is cited for a violation due merely to
normal observational error. EPA has consistently maintained that it
does in fact take this error tolerance into account in the enforcement
99
-------
process. In an earlier draft response to this remand (November, 1973),
for example, the Administrator stated that "the agency would not take,
and the courts would not sustain, enforcement action based upon opacity
observations that do not exceed the standard by more than the average
opacity error associated with the particular opacity standard." PCA,
however, has been unsatisfied by this assurance, apparently fearing that
EPA in its enforcement decisions would ignore the established error
tolerance. To provide the certainty desired by PCA and others, EPA is
therefore now amending the applicable regulations (40 CFR 60, Appendix
A, Reference Method 9) by adding an introductory section which discusses
the concept of visible emission readings and describes the effect of
variable viewing conditions. This section presents pertinent portions
of the above data and notes specifically that the accuracy of the method
must be taken into account, when determining possible violations of
applicable opacity standards. A copy of the revised regulation appears
in Part B of Appendix III of this Response.
EPA is also now amending the provisions of Reference Method 9 to
make clear that the determination of an opacity level reportable as a
violation involves averaging 24 readings taken at 15-second intervals.
This will make clear that a single high reading will not be cited as a
violation. It will also ensure, as PCA requests, that such individual
readings, which may result under normal plant operations from brief
events such as shaking of fabric filter bags in baghouses or rapping
of electrodes in electrostatic precipitators, will not result in
100
-------
determinations of violation. This amendment to Reference Method 9 is
set forth in Part B of Appendix III. (This approach is a satisfactory
means of enforcing opacity standards in cases where the violation is a
continuing one and time exceptions are not part of the applicable
opacity standard. However, the opacity standards for steam electric
generators in 40 CFR 60.42 and fluid catalytic cracking unit catalyst
regenerators in 40 CFR 60.102 and numerous opacity standards in State
Implementation Plans specify various time exceptions. Many State and
local air pollution control agencies use a different approach in
enforcing opacity standards than the six-minute average period specified
in this revision to Method 9. EPA recognizes that certain types of
opacity violations that are intermittent in nature require a different
approach in applying the opacity standards than this revision to Method
9. It is EPA's intent to propose an additional revision to Method 9
specifying an alternative method to enforce opacity standards. It is
our intent that this method specify a minimum number of readings that
must be taken, such as a minimum of ten readings above the standard in
any one hour period prior to citing a violation. EPA is in the process
of analyzing available data and determining the error involved in
reading opacity in this manner and will propose this revision to Method
9 as soon as this analysis is completed. The Agency solicits comments
and recommendations on the need for this additional revision to Method
9 and would welcome any suggestions particularly from air pollution
control agencies in how we might make Method 9 more responsive to the
needs of these agencies.)
101
-------
PCA also contends that the wind velocity during a given seu of
observations can affect the opacity levels read by observers. EPA
agrees that this is true. However, as is the case with lack of contras-
ting background in viewing a plume, the result of high winds is a
negative error in reading the opacity of the plume; this results from
the dissipation of the plume caused by the wind. Since a negative
error in reading the opacity of the plume cannot prejudice an operator,
this consideration is not relevant to the reliability of the method for
assessing violations of the opacity standards.
PCA contends that the viewing angle of the observer with respect
to the sun and the plume can produce erroneously high readings; if the
observer reads the plume while facing the sun, PCA says, such high
reading may result. EPA agrees, as was noted above, and for this reason
Reference Method 9 has always required the observer to read the plume
with the sun at his back. Reference Method 9 is now being amended to
include even more specific criteria concerning observer position with
respect to the sun. The method will now require that the sun shall be
within a 140° sector to the observer's back. This amendment appears in
Part B of Appendix III.
The final aspect of opacity observations with which PCA and the
Court of Appeals were concerned is the procedure employed by qualified
observers to evaluate plumes containing condensed (visible) water vapor,
or "steam plumes." When stack emissions contain appreciable water vapor,
the vapor becomes visible when the gases are cooled below the dew point
102
-------
and water condenses in fine droplets. Some steam plumes are detached
***
while others are visible as released from the stack. However,
since ambient air is normaTly much drier than stack emissions containing
visible water vapor, dilution causes the gases at some point to become
unsaturated. At this point, condensed water evaporates and is no
longer visible in the plume. PCA contends that these plumes cannot be
read properly, and the Court of Appeals also notes concern over the way
in which steam plumes are to be read. This issue was addressed in part
24
in the report titled "A Report on Contaminated Water Vapor Plumes." As
EPA has previously noted, instruction is given in the smoke school as
to how such plumes are to be read; the above report illustrates how
qualified observers can readily distinguish the point at which all visible
water vapor has evaporated from the plume. It is this point at which
the opacity reading is made. Since the plume has been somewhat dissi-
pated by that point, the reading is often lower than that which would
have been found at the lip of the stack, had the plume been free from
visible water vapor. This error is negative, however, and in no way
prejudices the plant operator. EPA agrees, however, that the prior
language of Method 9 did not sufficiently make this procedure clear.
Method 9 is therefore now being amended to remedy that deficiency;
*** The phenomenon of a "detached plume" is often observed when a warm,
moist gas stream is released into a cooler atmosphere, e.g. when hot,
wet exhaust gases from a wet process kiln are released to the atmosphere.
Here the gases are clear at the point of discharge but become opaque as
the wet gases mix with cooler air in the atmosphere.
103
-------
when the new guidelines are adhered to (as they were in the past, though
not formally articulated in the Method), any errors due to the presence
of water vapor can be eliminated. These amendments appear in Part B of
Appendix III.
In addition to proposing specific procedures, EPA is currently
investigating techniques which could be employed in the smoke school to
further enhance an observer's capability to accurately identify
contaminated water plumes. This investigation is being undertaken by
a contractor—Pacific Environmental Services, Inc.--and by EPA staff
as part of an overall program to develop further guidelines concerning
12
opacity training and field observation. Procedures used by State and
local regulatory agencies to observe contaminated water vapor plumes
are being reviewed, and the use of such training aids as movies and
apprenticeship training will be considered.
The basic question involved is not whether an observer can accurately
assign opacity to a contaminated water plume, but rather is whether an
observer can distinguish between condensed water vapor and visible
particulate matter. To accurately evaluate wet plumes pursuant to
revised Method 9, it is necessary only for an observer to distinguish
between condensed water vapor and visible particulate air pollutants.
In no case is the observer required to assign an opacity to a plume
1 o
which contains visible (condensed) water droplets.
As noted in the previously cited report, water plumes present an
entirely different appearance to an observer than a dust plume. A
104
-------
trained air pollution inspector is familiar with both types of plumes,
and has no difficulty distinguishing between the two. While much of
his ability in this area stems from smoke school training, his background
in inspecting industrial processes is also a factor. Inspectors
employed by control agencies are required to become familiar with a
broad spectrum of sources and to evaluate opacities from them under
all weather conditions. For instance they know which sources employ
scrubbers, which cement plants employ wet and dry process systems and
generally which sources tend to produce wet plumes. They also know that
extreme weather conditions can create visible water plumes at sources
which release only negligible particulate matter, e.g. natural gas-fired
boilers and water heaters. Thus, when a trained air pollution inspector
evaluates a wet plume source he relies on all of this background in
making a judgment. In following Method 9 he first identifies that
portion of the plume which contains condensed water. As explained in
the previously cited report, fluffy water plumes may trail 100 feet or
24
more from the stack. When the water plume disappears, he evaluates the
opacity of the remaining dust plume which, if of violating opacity, will
persist for a much longer distance. At this point, when water droplets
evaporate and cease to be visible, the dust plume has been diluted
considerably below the concentration at the stack exit. Only if the
dust plume is still of greater than allowed opacity at the point of
complete evaporation can an inspector consider a citation of violation.
His normal procedure is to enter the plant and evaluate the process
105
-------
operation before in fact issuing the notice of violation. Thus it is
obvious that the procedure allows a great deal of latitude to the
operator of a wet plume process. Far from citing unjustly for violative
visible emissions, the air pollution inspector will often be forced to
ignore periods of marginal violations simply because the visible dust
plume is masked by visible water droplets.
The revisions in Reference Method 9 make clear that the above
Q
procedure is to be followed. For attached steam plumes, the observer
is to make his readings at the point where all visible water vapor has
evaporated and only particulates remain in the plume. He is to record
the reading made at that point in the plume, and is not (as PCA fears)
to extrapolate back to any hypothetical value at the lip of the stack.
For detached steam plumes, the observer is to read the opacity value as
usual at the lip of the stack, providing that no visible water vapor is
present there, i.e. providing that there is a gap between the lip of the
stack and the point of initial condensation into steam.
106
-------
EVALUATION OF SMOKE SCHOOL PROCEDURES
The purpose of a smoke school is two-fold. It provides a
means for training an observer, and for certifying that an observer
is, in fact, capable of making opacity observations with reasonable
accuracy. In this sense, the school serves to calibrate (or re-calibrate)
observers, and rejects candidates who are not capable of accurately
reading opacity. The question of whether the school adequately serves
these purposes cannot primarily be answered by examining the procedures
used. Rather, it is necessary to examine the accuracy of observers
who have been trained and certified in the school, i.e., the results
of those procedures. If, as the data discussed earlier demonstrate,
qualified observers can read plumes with reasonable accuracy, the
procedures and instruction used at the smoke school must, by
definition, be accomplishing the desired results.
With regard to PCA's criticism concerning the ability of an
observer to accurately read emissions under various field conditions
which were not encountered in the training school, it is necessary
to refer again to the factors affecting apparent opacity. As discussed
above, observations made under conditions of high contrast are most
conducive to positive errors. To the extent possible, observers employ
contrasting backgrounds during qualification. It is true that this
procedure may tend to limit the accuracy of an observer under less than
ideal conditions. However, since a plume will present an appearance of
lower opacity under such conditions, the resulting bias will be negative.
As discussed above, this effect is confirmed by theory and by field
observation.
107
-------
PCA has noted the fact that smoke school trainees are given
repeated opportunities to pass the certification test. Thus, an
observer may be certified on one run even though the certification
criteria were not met on previous runs. Presumably, he could
also fail to meet the criteria on subsequent runs. Petitioners
are correct in this description—particularly when new or relatively
inexperienced persons are being tested. It does not follow, however,
that the smoke school procedures are deficient or inadequate in this
respect. First, as noted above, the adequacy of the smoke school
cannot be determined solely by reviewing the procedures; evaluation
can also be made by testing the ability of observers who have
passed the certification test. Secondly with respect to the
certification procedures, it should be recognized that the +7.5%
criterion is particularly restrictive in that it involves the absolute
error, not simply the average taken over 25 readings. Thus, negative
and positive errors are not compensating (i.e., a -5% error and
a +7 error result in an average absolute error of +6%). This is
unlike the field observation procedure where negative and positive
errors are compensating (a -5% error and a +7 error result in an
average error of +1%). During an EPA smoke school conducted on
October 9, 1974, a record was kept for all observers who attended and
Q
were certified. This record includes both runs during which the
observers passed the certification criteria and during which observers
failed to meet the criteria. These data show that even on these
runs when observers failed to qualify, over two-thirds of them met
108
-------
Q
7.5% average error limit imposed by Method 9. Since the Reference
Method 9 involves an average of 24 readings as a means of determining
plume opacity, failure to continuously meet the more stringent
certification requirements is relatively unimportant as long as
the average reading is consistently within the prescribed criterion
of less than +7.5% average opacity. For example, the requirement
that no single reading be missed by more than +15 percent is somewhat
arbitrary, and could be relaxed to +20 percent to reduce the number
of observers failing the certification test. EPA has concluded,
however, that the stringent certification criterion serves the intended
purpose of forcing a trainee to learn to assign opacity within a
reasonable range of precision and therefore should not be relaxed.
EPA has further concluded, in summary, that while any changes
recommended by the contractor Pacific Environmental Services will
be carefully considered, the training program does in fact effectively
train observers to read plumes under field conditions with
reasonable accuracy.
Summary
EPA believes that the above data have convincingly demonstrated
that the opacity standards are inherently reliable.**** The data show
that observers can, with a very high confidence level, read plumes
to within the certified error tolerance of 7.5%. That tolerance
is taken into account in the enforcement process, as provided
**** A New Jersey court has recently reached the same conclusion. See
New Jersey v. Lloyd A. Fry Roofing Co., Docket no. C-3682-72
New Jersey v. LLoyd A. Fry Roofin;
(Superior Court, Sept. 30,~1974).
109
-------
explicitly by the (amended) regulations. EPA further believes that
the effectiveness of the training program is well proven by the
resulting accuracy of its graduates under actual field conditions.
In those few situations where the language of the regulations or
Reference Method 9 was insufficiently clear or inadequate, those
regulations are being or have been amended to remedy such deficiencies.
The Administrator concludes, therefore, that since the concerns of
the Court of Appeals and the Portland Cement Association in this
regard have all been satisfactorily dealt with, the Agency should
continue to use opacity standards for new sources under Section 111
of the Clean Air Act. EPA does, however, recognize that some
operators of Portland cement plants may install in-stack transmissometers
14
for the purpose of continuous monitoring of particulate emissions.
EPA will therefore accept as probative (though not conclusive)
evidence the results of suci continuous monitoring by trans-
mi ssometer which indicate that the opacity at the time readings
were taken by Reference Method 9 was not in excess of the standard,
provided that the source shall meet the burden of proving that the
instrument used meets (at the time of the alleged violation) the
14
applicable technical specifications, has been properly maintained
and (at the time of the alleged violation) calibrated, and that the
resulting data have not been tampered with in any way. The applicable
regulations are being amended to reflect this provision and appear in
Part B of Appendix III.
no
-------
EPA will of course continue to review the degree of accuracy
and need for opacity measurements as new data and information are
acquired. For example, at the request of the Department of Commerce,
EPA solicited such information in the preamble to new source
performance standards recently proposed (39 FR 37466). However, at
this time the data and information available to EPA support the
position taken in this response.
ACHIEVABILITY OF THE OPACITY STANDARDS FOR PORTLAND CEMENT PLANTS
Under Section 111 of the Clean Air Act, EPA has established
10 percent opacity as the standard which may not be exceeded by
new kilns at Portland cement plants. PCA contends that some new
plants will or may be unable to achieve compliance with this
visible emissions limit. EPA policy is that opacity limits should
be set at such levels that any source which meets the mass emission
limits will also meet the opacity limits. EPA has conducted an
intensive review of its opacity standards, including conducting new
tests and gathering new data, and has concluded that although the 10%
opacity standard is achievable by almost all new Portland cement plants,
that standard should nevertheless be relaxed to 20% opacity to accomo-
date certain extreme circumstances. EPA further concludes that at this
new level, practical circumstances under which any combination of the
factors cited by PCA or the Court as having a possible effect on opacity
(transmittance) of smoke plumes that could cause a new source to be
unable to meet the opacity standard is exceedingly rare. The
111
-------
reasons for these conclusions are set forth below.
PCA contends that the opacity standards for Portland cement
plants are unachievable for two primary reasons, both related to
the theory of transmittance of light through a plume. PCA states,
first, that opacity varies with particle size and shape, so that
a given mass concentration of particles -- which could be composed
of various combinations of different size and shape particles ~
could result in differing opacities. EPA agrees that this correctly
states the theory of plume transmittance (opacity) as it relates
to particle dimensions. PCA concludes from this statement of theory,
disregarding the results of the application of the theory to operation
of Portland cement plants, that such plants might have widely varying
opacities while still meeting the mass emission limits. It is with
PCA's conclusion that EPA disagrees. The relevant question is what
variations in opacity are i.r\_ fact produced by the particle size-shape
distributions actually present in Portland cement plants under normal
operation and with proper maintenance and operation of pollution control
equipment. It is this question which is next addressed.
It is known from various studies and experience that uncontrolled
gases vented from cement kilns contain heavy concentrations of coarse
as well as fine partielate, i.e., in the range of 5 to 15 grains
15
per standard cubic foot. Standards of performance for new sources
requiring best available technology, necessitate the installation of
high efficiency collectors. Collectors of 99.4 to 99.8 percent
efficiency are necessary to reduce exit concentrations to approximately
112
-------
0.03 grains per standard cubic foot which is necessary to meet
the mass emission standard.' .These collectors, fabric filters or
electrostatic precipitators, are more effective in removing coarse
particles than fine particles. Thus emissions from collectors, i.e.,
visible emissions, tend to be made up of much greater fractions
of fine particulates than coarse particulates.
In typical high efficiency collector exhaust gases there are
generally few particulates larger than 40 microns diameter. ' The
predominant number of particles are between 0.5 and 10 microns with
217
the average size being about 2-4 microns. ' Maximum light scattering
is generally acknowledged to be caused by particles in the size range
of 0.2 to 2.0 micron. ' Available data indicate that the size dis-
tribution of particulates released from well controlled cement kilns
are similar within a narrow range (approximately 2 to 6 microns) from
one kiln to another, and therfore from one plant to another.
What the above data and studies indicate, in short, is that the
size of particles emitted by plants with such control equipment varies
only within a very narrow range. This variability in average size
is theoretically not sufficient to cause more than a +_ 5 percent
variation in opacity for typical cement kilns. These data are
wholly incompatible with PCA's speculation that widely varying parti-
cle sizes might produce much larger variations in opacity at cement
plants meeting the mass emission standard.
The shape of particles emitted from well-controlled
Portland cement plants may vary as PCA contends. However, we disagree
with their conclusions that such variations in shape would affect the
113
-------
opacity of emissions perceived by a qualified observer or in-stack
transmissometer. Any irregularly shaped particles would be randomly ori-
ented thereby smoothing or averaging out any significant variations of
opacity that could possibly result. In addition, it is likely that the
particles will be abraded or will condense into spherical shapes for the
following reasons: 1) feed materials are ground to less than 75 microns
in almost all cases; 2) further attrition and size reduction occurs
as the feed material passes through the high temperature rotating
kiln; and 3) alkaline material tends to vaporize and condense as
a fume.15'16'17
Also, the recent microscopic examination of particles from
the exhaust gases of a precipitator controlled kiln showed that
2
almost all particles were approximately spherical in shape.
Given that particle size and shape do not vary to the extent
necessary to substantiate PCA's speculation, it is apparent that
normal operations even from plant to plant will not produce vari-
ations sufficient to significantly affect the opacity observed
at the point of emission from the stack. Therefore, if the opacity
standards are set in accordance with the opacity actually produced
by normal operations, then with proper maintenance and operation
of control equipment the very slight variations which may occur
from plant to plant or from time to time within an individual plant
will not cause the source to exceed the opacity standard. And if,
for any reasons, some unusual design feature of a given plant
caused it to exceed the opacity standard but not the mass emission
114
-------
standard, the variance procedure described below would elimi-
nate any prejudice to such source.
The second major factor which PCA claims affects opacity
(transmittance) to the extent that some plants may exceed the
opacity standard while meeting the mass emission standard is
the diameter of the duct involved; i.e., the path length through
which the light must pass. EPA agrees that the greater the path
length through which the light must pass, the more light will be
attenuated, and the greater will be the opacity. EPA again dis-
agrees, however, with PCA's conclusion, which is again in disregard
of the actual circumstances of plant design and operation, that
some plants will therefore be unable to meet the opacity standards,
though meeting the mass standard. EPA believes that the proper
issue in this regard is the actual range of stack diameters
resulting from the engineering and economic necessities of plant
design. This question is addressed below.
Stack diameter affects opacity directly because it governs the
length of the path through which the light is attenuated before it
reaches the observer viewing the plume. The larger the diameter
of the stack, and therefore the plume, the more light will be
attenuated as it passes through the plume, and the greater will
be the opacity of the plume. Stacks venting high efficiency control
equipment, such as that serving kilns at portland cement plants, are
designed to ventilate the process to the degree necessary with the
minimum pressure drop and at the lowest cost. These engineering
and economic factors result in the construction of stacks in which
115
-------
the gas velocities approximate 30 to 100 feet per second, since at
higher velocities friction losses become excessive, necessitating
much larger horsepower fans. At lower velocities, the size of the
stack required would result in disproportionately large capital
costs. Thus, stack diameter tends to be directly related to and
characteristic of plant size, varying only within a relatively
limited range.
Stack gas exit velocity is controlled by stack diameter.
Exit velocity is directly proportional to the volume of gases
vented from a kiln and inversely proportional to the cross-sectional
area of the stack. For plumes of a given cross-section, velocity
has no effect upon opacity, since opacity is a function of path
length and is proportional to the concentration and characteris-
tics of particles in the gas stream. The velocity at which the
gas stream moves through the viewer's line of vision has no
effect on opacity.
It follows, therefore, that an extremely large diameter
stack could theoretically be installed at a cement plant, even
though its cost would be much greater than normal. In the course
of EPA's survey of the cement industry, the largest size kiln
stack that has come to our attention is 15 foot in diameter. In
the unlikely event that larger stacks were used or other abnormal
conditions of design or operation were encountered that could
cause a violation of an applicable opacity standard, EPA would evaluate
opacity during the performance tests for the mass emission standard.
If it were shown necessary, a suitable opacity level greater than
116
-------
20 percent would be established for that specific kiln installation.
Although such an eventuality is highly unlikely, appropriate lan-
guage is being promulgated in the general provisions of the regulation
to accomodate such situations and appears in Part B of Appendix III.
EPA concludes, on the basis of the actual stack diameters
employed by the industry, that stack diameter will not cause a
plant meeting the mass emission standard to exceed the opacity
standard. To deal with the uncertainties cited by PCA as to
how plumes from baghouses with stub stacks are to be read, EPA is
revising Method 9 to specify that a row of such stub stacks, or
any non-circular stack, is to be read along its shortest axis
(perpendicular to its longer axis). This ensures that the
shorter path length will be used to determine opacity, and will
eliminate any concern that a long row of stub stacks, common on
baghouses, might be read along its long axis, producing higher
opacity.
Given that particle size-shape and stack diameter do not pro-
duce opacity variations sufficient to cause a source which might
otherwise meet the opacity standard, and which is meeting the mass
standard, to exceed the opacity standard, the only remaining issue
is whether well-controlled portland cement plants can, in fact,
meet the opacity standard. EPA has accumulated data for portland
cement plant kilns that are vented to high efficiency control
equipment that is properly designed, operated and maintained. These
data are summarized in Table I. Opacity readings were also taken of
117
-------
Table I
Name of the Company,
Date of Test and
Type Control Equipment
Dragon Cement Co.
Northampton, Pa.
4/29-30/71
Baghouse
Oregon Portland Cement Co.
Lake Oswego, Oregon
10/7-8/71
Baghouse
Santee Portland Cement Co.
Holly Hill, S. C.
5/23-25/74
Electrical Precipitator
United Portland Cement Co.
Artesia, Miss.
6/25-27/74
Electrical Precipitator
Monarch Cement Co.
Humboldt, Kansas
6/12-14/74
Baghouse
Marquette Cement Co.
Neville Island, Pa,
6/19-21/74
Electrical Precipitator
Particulate Emissions
Test Duration
2,19-23
Ibs/ton
of dry
feed
6 hours
6 hours
8 hours -
2 minutes
8 hours -
23 minutes
0.070
0.272
0.107
0.209
12.5 minutes 0.196
9 hours -
0 minutes
0.339
Stack Dia.
11'0"
4'4"
(velocity
cone)
12'0"
1V3"
7'0'
12'6'
Max.
Average Single
Opacity Reading
<5
3.8
18"
<5
*Based on transmissometer data,
118
-------
several additional well controlled portland cement plant kiln stacks
i g
by an EPA contractor. Mass emission tests were not conducted in
these instances, however, in light of the petitioners contention
that the mass standards are overly restrictive, it would seem
unlikely that these plants would be operating significantly below
the mass emission standard. A summary of these installations is
shown in Table II.
Table II
Opacity of Visible Emissions from Cement Kilns Controlled by Fabric Filters
18
Name of Company
Date of Test and
Control Equipment
Whitehall Cem
Cementon, Pa.
7/29 & 8/1/74
Baghouse
Duration
of Test
(hours, min. )
8h 2rn
Stack Size
(feet)
15' dia.
(1 stack)
Average
Opacity
(percent)
0
Maximum
6-Minute
Average
Opacity
0.2
Maximum
Single
Reading
5
Cop!ay (Egypt) 7h 15m
Cement Co.
Egypt, Pa.
7/29-31/74
Baghouse
Hercules Cement Co. 5h 16m
Stockerton, Pa.
7/29 & 8/2/74
Penn-Dixie Cement Co.* 7h 55 m
Nazareth, Pa.
7/29-30/74
Baghouse
4'10" dia. each.
(7 stacks)
3.8
13' dia.
(1 stack)
2'xT each.
(12 stacks)
0
0
0
0
0
0
5'xll1 each with
center divider.
(8 stacks)
10-15
15
25
*These observations were taken across the shorter axis of Jthe eight stacks (40' path-
length). If the observations had been taken across the longer axis of any one stack, the
pathlength would have been 11 feet. The estimated maximum 6-minute average opacity
under these latter conditions would be 4.4 percent rather than 15 percent. Like-
wise, the maximum single reading would be 7.5%. Reference Method 9 now requires
these observations to be made across the 11' axis.
119
-------
With the exception of Marquette Cement Co. all of the kilns
identified in Table I were- operating below the mass and opacity
standards promulgated in 1971 (i.e., 0.3 Ib/ton of dry feed and
10%). The data for Marquette Cement Co. indicate that the kiln
was not in compliance with the mass emission standard (0.339
versus 0.3 Ib/ton of dry feed). The opacity data were collected
with an in-stack transmissometer. The stack diameter is 12'6", the
average opacity during the mass emission tests was 18%, and the
average size of emitted particles was determined to be approximately
2
2 microns diameter. If the stack diameter were 15', which is the
largest diameter stack EPA has observed on cement plants, the average
opacity would have been 21%. The mass emission tests for Marquette
Cement Co. consisted of three separate runs. The results of the three
runs are:
Table III
2 3
Opacity and Mass Emissions for a Kiln '
Particulate Average opacity Average opacity
Emissions, (percent) at (percent) extrap-
Run Ibs/tons of actual stack con- olated to 15' diameter
Number dry kiln feed ditions (12'6") stack conditions
1
2
3
0.386
0.339
0.292
22
17
15
25
20
18
The particulate matter emissions from Marquette Cement Co. were
analyzed to determine the particle size distribution. This analysis
indicated that the average particle diameter was approximately 2 microns
120
-------
which is very low for port!and cement plants. These data indicate
that for a plant with an unusually large stack diameter and
with particulate emissions having very small average particle diameter
properties, a 20% opacity standard is more reasonable than a 10% standard.
The achievability of a 20 percent standard is amply demonstrated by
the data in Tables I and II which illustrate that most portland cement
kilns have no difficulty in meeting even a 10 percent opacity standard.
The Marquette Cement Company test results illustrate that even under
the combination of conditions that would produce the maximum obscuration
of light (i.e., large stack diameters, small average particle sizes,
and concentrations near the mass emission standard), a 20 percent
standard (Table III, run number 3) would be achieved. It should be
additionally noted that the combination of conditions for maximum
opacity, even though accommodated by the 20 percent standard, will
occur infrequently. Therefore, it is concluded that a plant with a 15'
diameter stack which meets the mass emission standard will also meet a
20% opacity standard and when it exceeds the mass emission standard
will exceed the opacity standard. For stacks smaller than 15 feet in
diameter, a portland cement kiln that violates a 20% opacity
standard will also violate the mass emission standard by a sub-
stantial amount. Data in Table II show that in several cases,
opacity of emissions from well controlled cement kilns are
appreciably less than 20% even where larger stacks are employed
e.g., 0.2% from a 15 foot diamater stack and zero from a 13 foot
diameter stack. Based on these data and EPA's policy that opacity
standards be established at levels which require proper operation
121
-------
and maintenance of control systems, the opacity standard for
Portland cement plant kilns is being changed from 10 to 20 percent.
This change appears in Part B of Appendix III.
In addition to considerations of the average opacity standards,
EPA has evaluated the need for possible time exemptions in the
opacity standard. It is recognized that during normal operation of
Portland cement kilns, process variations, shaking of bags in bag-
houses, rapping of electrodes in electrical precipitators, etc.,
can produce mass emissions that are not perfectly constant. It
follows that slight variations in opacity can also occur under
these conditions. However, with the clarification to Method 9
whereby a minimum of 24 opacity readings must be taken at 15
second intervals over a 6 minute period and with the averaging
of observations for determining compliance with the opacity
standard, it is apparent that such instantaneous opacity readings
will not constitute a violation of the standard. Changing the opacity
standard for portland cement kilns from 10 to 20 percent further
ensures that such temporary incidents will not cause the opacity
standard to be exceeded. Since the change in the opacity standard
merely represents a relaxation of the standard, it will be appli-
cable to all sources which commenced construction subsequent to
the promulgation of the original, more stringent level. To an even
greater degree than the 10% standard, this 20% standard is
significantly less stringent than the mass emission limit for
Portland cement kilns.
122
-------
As discussed in this remand response, the change of the
opacity standard from 10 percent to 20 percent was made to insure
that no plant that is meeting the mass emission standard is cited
for violating the opacity standard. It does not change the
degree of particulate control required of new portland cement
plants. Accordingly, this relaxation of the opacity standard
does not require an environmental impact statement.
123
-------
APPENDIX III
References
1. K.W. Buhne and L. Duwel. Recording Dust Emission Measurements in
the Cement Industry with the RM4 Instrument. Staub. Vol.32, No. 19.
1972 (in English).
2. Preliminary Test Results; of Marquette Cement Manufacturing
Company for Task Number 11. Cottrell Environmental Sciences.
EPA Contract No. 68-02-0239. August 26, 1974.
3. Robert L. Ajax. Memo to Jack R. Farmer, Credibility of Emission Test
Data Contained in the Marquette Cement Manufacturing Company's Report.
October 24, 1974.
4. Average Observational Error Associated with Smoke Plumes at Levels of
Known Opacity. EPA, OAWP, OAQPS, ESED Report. October 1973.
5. W. D. Conner. Memo to R. L. Ajax, Measurement of Opacity by
Transmissometer and Smoke Readers. October 16, 1974.
6. Optical Properties and Visual Effects of Smoke Stack Plumes. PHS
Publication No. 999-AP-30.
7. Henry F. Hammill, Richard E. Thomas, and No!lie F. Swynnerton.
Evaluation and Collaborative Study of Method for Visual
Determination of Opacity of Emissions from Stationary Sources:
Interim Report. October 1974.
8. Federal Register: 39 FR 32857, September 11, 1974.
9. Evaluation of EPA Smoke School Results. EPA, OAWM, OAQPS, ESED Report.
October 9, 1974.
10. Portland Cement Association v. Ruckelshaus. 486 F^.2d 375, No. 72-1073.
June 29, 1973.
11. Coons, J. D., et al. Development, Calibration, and Use of a Plume
Evaluation Training Unit. Journal of Air Pollution Control Association.
Vol. 15. May 1965. pp. 199-205.
12. Task Order: Preparation of Guideline for Evaluation of Visible
Emission Opacity by Trained Observers. EPA Contract No. 68-02-1390,
Task 2, June 5, 1974.
13. Federal Register: 39 FR 32854, September 11, 1974.
14. Federal Register: 39 FR 32852, September 11, 1974.
15. R. Emmet Doherty. Current Status and Future Prospects - Cement Mill
Air Pollution Control. Proceedings, The Third National Conference
on Air Pollution, Washington, D. C. December 12-14, 1966.
124
-------
16. A Manual of Electrostatic Precipitator Technology, Part II--
Application Areas. Southern Research Institute. Contract No.
22-69-73. 1970. pp. 609-640.
17. Particulate Pollutant System Study, Volume II—Fine Particulate
Emissions. Midwest Research Institute. Contract No. CPA 22-69-104.
1971.
18. Visible Emission Testing at Four Portland Cement Plants. Scott
Environmental Technology, Inc. Contract No. 68-02-1400. 1974.
19. Emission Testing Report (tested by company personnel). United
Cement Company, Artesia, Mississippi. 1974.
20. Emission Testing Report (kiln No. 2). Santee Portland Cement
Corporation, Holly Hill, South Carolina. Pollution Control-
Walther Inc. 1974.
21. Emission Testing Report. Monarch Cement Company, Humboldt, Kansas.
Fuller Company. 1974.
22. Emission Testing Report. Dragon Cement Company, Northampton,
Pennsylvania. EPA, OAWM, OAQPS, ESED Report. March 1972.
23. Emission Testing Report. Oregon Portland Cement Company,
Lake Oswego, Oregon. EPA, OAWM, OAQPS, ESED Report. March 1972.
24. Contaminated Water Plumes. EPA, OAWM, OAQPS, ESED Report. 1973.
125
-------
-------
PORTLAND CEMENT REMAND RESPONSE
APPENDIX III, PART B
PREAMBLE AND AMENDMENTS
TO REGULATIONS
127
-------
-------
Title 40--Protection of the Environment
CHAPTER I — ENVIRONMENTAL PROTECTION AGENCY
Subchapter C--Air Programs
PART 60—STANDARDS OF PERFORMANCE FOR NEW STATIONARY SOURCES
Opacity Provisions
On June 29, 1973, the United States Court of Appeals for the
District of Columbia in Portland Cement Association v. Ruckelshaus,
486 F. 2d 375 (1973) remanded to EPA the standard of performance for
Portland cement plants (40 CFR 60.60 et. se£.) promulgated by EPA
under section 111 of the Clean Air Act. In the remand, the Court
directed EPA to reconsider among other things the use of the opacity
standards. EPA has prepared a response to the remand. Copies of
this response are available from the Emission Standards and Engineering
Division, Environmental Protection Agency, Research Triangle Park,
N. C. 27711, Attn: Mr. Don R. Goodwin. In developing the response,
EPA collected and evaluated a substantial amount of information which
is summarized and referenced in the response. Copies of this informa-
tion are available for inspection during normal office hours at EPA's
Office of Public Affairs, 401 M Street, S.W., Washington, D. C. EPA
determined that the Portland cement plant standards generally did
not require revision but did find that certain revisions are appro-
priate to the opacity provisions of the standards. The provisions
promulgated herein include a revision to §60.11, Compliance with
Standards and Maintenance Requirements, a revision to the opacity
129
-------
standard for Port!ant cement plants, and revisions to Reference
Method 9. The bases for the revisions are discussed in detail in
the Agency's response to the remand. They are summarized below.
The revisions to §60.11 include the modification of paragraph (b)
and the addition of paraqraph (e). Paragraph (b) has been revised
to Indicate that while Reference Method 9 remains the primary and
accepted means for determining compliance with opacity standards in
this part, EPA will accept as probative evidence in certain situations
and under certain conditions the results of continuous monitoring by
transmissometer to determine whether a violation has in fact occurred.
The revision makes clear that even in such situations the results of
opacity readings by Method 9 remain presumptively valid and correct.
The provisions in paragraph (e) provide a mechanism for an owner
or operator to petition the Administrator to establish an opacity
standard for an affected facility where such facility meets all
applicable standards for which a performance test is conducted under
§60.8 but fails to meet an applicable opacity standard. This provision
1s intended primarily to apply to cases where a source installs a
very large diameter stack which causes the opacity of the emissions
to be greater than if a stack of the diameter ordinarily used
in the industry were installed. Although this situation is considered
to be very unlikely to occur, this provision will accommodate such
a situation. The provision could also apply to other situations
where for any reason an affected facility could fail to meet opacity
standards while meeting mass emission standards, although no such
situations are expected to occur.
130
-------
A revision to the opacity standard for Portland cement plants
is promulgated herein. The revision changes the opacity limit
for kilns from 10 percent to 20 percent. This revision is based
on EPA's policy on opacity standards and the new emission data
from Portland cement plants evaluated by EPA during its reconsidera-
tion. The preamble to the standards of performance which were pro-
mulgated on March 8, 1974 (39 FR 9308) sets forth EPA's policy on
opacity standards: (1) opacity limits are independent enforceable
standards; (2) where opacity and mass/concentration standards are
applicable to the same source, the mass/concentration standards are
established at a level which will result in the design, installation,
and operation of the best adequately demonstrated system of emission
reduction (taking costs into account); and (3) the opacity standards
are established at a level which will require proper operation and
maintenance of such control systems. The new data indicate that
increasing the opacity limit for kilns from 10 percent to 20 percent
is justified, because such a standard will still require the design,
installation, and operation of the best adequately demonstrated
system of emission reduction (taking costs Into account) while
eliminating or minimizing the situations where it will be necessary
to promulgate a new opacity standard under §60.11(e).
In evaluating the accuracy of results from qualified observers
following the procedures of Reference Method 9, EPA determined that
some revisions to Reference Method 9 are appropriate. The evaluation
showed that observers trained and certified in accordance with the
procedures prescribed under Reference Method 9 are consistently able
to read opacity with errors not exceeding +7.5 percent based upon
131
-------
single sets of the average of 24 readings. The revisions to
Reference Method 9 include the following:
1. An introductory section is added. This includes a dis-
cussion of the concept of visible emission reading and describes
the effect of variable viewing conditions. Information is also
presented concerning the accuracy of the method noting that the
accuracy of the method must be taken into account when determining
possible violations of applicable opacity standards.
2. Provisions are added which specify that the determination of
opacity requires averaging 24 readings taken at 15-second intervals.
The purpose for taking 24 readings is both to extend the averaging
time over which the observations are made, and to take sufficient
readings to insure acceptable accuracy.
3. More specific criteria concerning observer position with
respect to the sun are added. Specifically, the sun must be within
a 140° sector to the observer's back.
4. Criteria concerning an observer's position with respect to
the plume are added. Specific guidance is also provided for reading
emissions from rectangular emission points with large length
to width ratios, and for reading emissions from multiple stacks. In
each of these cases, emissions are to be read across the shortest
path length.
5. Provisions are added to make clear that opacity of contami-
nated water or steam plumes is to be read at a point where water does
not exist in condensed form. Two specific instructions are provided:
132
-------
one for the case where opacity can be observed prior to the
formation of the condensed water plume, and one for the case
where opacity is to be observed after the condensed water plume
has dissipated.
6. Specifications are added for the smoke generator used
for qualification of observers so that State or local air pollution
control agencies may provide observer qualification training
consistent with EPA training.
In developing this regulation we have taken into account the
comments received in response to the September 11, 1974 (39 FR 35852)
notice of proposed rulemaking which proposed among other things
certain minor changes to Reference Method 9. This regulation repre-
sents the rulemaking with respect to the revisions to Method 9.
The determination of compliance with applicable opacity standards
will be based on an average of 24 consecutive opacity readings taken
at 15 second intervals. This approach is a satisfactory means of
enforcing opacity standards in cases where the violation is a con-
tinuing one and time exceptions are not part of the applicable
opacity standard. However, the opacity standards for steam electric
generators in 40 CFR 60.42 and fluid catalytic cracking unit catalyst
regenerators in 40 CFR 60.102 and numerous opacity standards in State
implementation plans specify various time exceptions. Many State
and local air pollution control agencies use a different approach
in enforcing opacity standards than the six-minute average period
specified in this revision to Method 9. EPA recognizes that certain
133
-------
types of opacity violations that are intermittent in nature
require a different approach in applying the opacity standards
than this revision to Method 9. It is EPA's intent to propose
an additional revision to Method 9 specifying an alternative
method to enforce opacity standards. It is our intent that this
method specify a minimum number of readings that must be taken,
such as a minimum of ten readings above the standard in any one
hour period prior to citing a violation. EPA is in the process
of analyzing available data and determining the error involved in
reading opacity in this manner and will propose this revision to
Method 9 as soon as this analysis is completed. The Agency solicits
comments and recommendations; on the need for this additional revision
to Method 9 and would welcome any suggestions particularly from air
pollution control agencies on how we might make Method 9 more
responsive to the needs of these agencies.
These actions are effective on (date of publication). The Agency
finds good cause exists for not publishing these actions as a notice
of proposed rulemaking and for making them effective immediately
upon publication for the following reasons:
(1) Only minor amendments are being made to the opacity standards
which were remanded.
(2) The U, S. Court of Appeals for the District of Columbia
Instructed EPA to complete the remand proceeding with respect to the
Portland cement plant standards by November 5, 1974.
(3) Because opacity standards are the subject of other litiga-
tion, it is necessary to reach a final determination with respect to
134
-------
the basic issues involving opacity at this time in order to properly
respond to this issue with respect to such other litigation.
These regulations are issued under the authority of sections 111
and 114 of the Clean Air Act, as amended (42 U.S.C. 1857c-6 and 9).
Date Administrator
135
-------
Part 60 of Chapter I, Title 40 of the Code of Federal Regula-
tions Is amended as follows:
1. Section 60.11 is amended by revising paragraph (b) and
adding paragraph (e), reading as follows:
* * * * *
(b) Compliance with opacity standards in this part shall be
determined by conducting observations in accordance with Reference
Method 9 in Appendix A of this part. Opacity readings of portions
of plumes which contain condensed, uncombined water vapor shall not
be used for purposes of determining compliance with opacity standards.
The results of continuous monitoring by transmissometer which
indicate that the opacity at the time visual observations were
made was not in excess of the standard are probative but not
conclusive evidence of the actual opacity of an emission, provided
that the source shall meet the burden of proving that the instrument
used meets (at the time of the alleged violation) Performance
Specification 1 in Appendix B of this part, has been properly
maintained and (at the time of the alleged violation) calibrated,
and that the resulting data have not been tampered with in any way.
* * * * *
(e)(l) An owner or operator of an affected facility may
request the Administrator to determine opacity of emissions from the
affected facility during the initial performance tests required by §60.8.
136
-------
(2) Upon receipt from such owner or operator of the written
report of the results of the performance tests required by §60.8,
the Administrator will make a finding concerning compliance with
opacity and other applicable standard.,. If the Administrator finds
that an affected facility is in compliance with all applicable
standards for which performance tests are conducted in accordance
with §60.8 of this part but during the time such performance tests
are being conducted fails to meet any applicable opacity standard,
he shall notify the owner or operator and advise him that he may
petition the Administrator within 10 days of receipt of notification
to make appropriate adjustment to the opacity standard for the
affected facility.
(3) The Administrator will grant such a petition upon a demon-
stration by the owner or operator that the affected facility and
associated air pollution control equipment was operated and main-
tained in a manner to minimize the opacity of emissions during the
performance tests; that the performance tests were performed under
the conditions established by the Administrator; and that the
affected facility and associated air pollution control equipment
were incapable of being adjusted or operated to meet the applicable
opacity standard.
(4) The Administrator will establish an opacity standard for
the affected facility meeting the above requirements at a level at
which the source will be able, as indicated by the performance and
opacity tests, to meet the opacity standard at all times during
which the source is meeting the mass or concentration emission standard.
137
-------
The Administrator will promulgate the new opacity standard in
the Federal Register.
2. In §60.62, paragraph (a)(2) is revised to read as follows:
§60.62 Standard for particulate matter
* * * * *
(a)
(2) Exhibit greater than 20 percent opacity.
3. Appendix A—Reference Methods is amended by revising
Reference Method 9 as follows:
138
-------
METHOD 9—VISUAL DETERMINATION OF THE
OPACITY OF EMISSIONS FROM STATIONARY SOURCES
Many stationary sources discharge visible emissions into the
atmosphere; these emissions are usually in the shape of a plume.
This method involves the determination of plume opacity by qualified
observers. The method includes procedures for the training and certi-
fication of observers, and procedures to be used in the field for
determination of plume opacity. The appearance of a plume as viewed
by an observer depends upon a number of variables, some of which may
be controllable and some of which may not be controllable in the field.
Variables which can be controlled to an extent to which they no longer
exert a significant influence upon plume appearance include: angle of
the observer with respect to the plume; angle of the observer with respect
to the sun; point of observation of attached and detached steam plume; and
angle of the observer with respect to a plume emitted from a rectangular
stack with a large length to width ratio. The method includes specific
criteria applicable to these variables.
Other variables which may not be controllable in the field are
luminescence and color contrast between the plume and the background
against which the plume 1s viewed. These variables exert an influence
ypon the appearance of a plume as viewed by an observer, and can affect
the ability of the observer to accurately assign opacity values to the
observed plume. Studies of the theory of plume opacity and field studies
have demonstrated that a plume is most visible and presents the greatest
apparent opacity when viewed against a contrasting background. It follows
139
-------
from this, and is confirmed by field trials, that the opacity of a
plume, viewed under conditions where a contrasting background is present
can be assigned with the greatest degree of accuracy. However, the
potential for a positive error is also the greatest when a plume is
viewed under such contrasting conditions. Under conditions presenting
a less contrasting background, the apparent opacity of a plume is less
and approaches zero as the color and luminescence contrast decrease
toward zero. As a result, significant negative bias and negative errors
can be made when a plume is viewed under less contrasting conditions.
A negative bias decreases rather than increases the possibility that a
plant operator will be cited for a violation of opacity standards due to
observer error.
Studies have been undertaken to determine the magnitude of positive
errors which can be made by qualified observers while reading plumes under
contrasting conditions and using the procedures set forth in this method.
The results of these studies (field trials) which involve a total of 769
sets of 25 readings each are as follows:
(1) For black plumes (133 sets at a smoke generator), 100% of the
sets were read with a positive error* of less than 7.5% opacity; 99% were
read with a positive error of less than 5% opacity.
(2) For white plumes (170 sets at a smoke generator, 168 sets at a
coal-fired power plant, 298 sets at a sulfuric acid plant), 99% of the
sets were read with a positive error of less than 7.5% opacity; 95% were
read with a positive error of less than 5% opacity.
*For a set, positive error = average opacity determined by observer's
25 observations - average opacity determined from transmissometer's
25 recordings.
140
-------
The positive observational error associated with an average
of twenty-five readings is therefore established. The accuracy of
the method must be taken into account when determining possible
violations of applicable opacity standards.
1. Principle and applicability.
1.1 Principle. The opacity of emissions from stationary sources
is determined visually by a qualified observer.
1.2 Applicability. This method is applicable for the determination
of the opacity of emissions from stationary sources pursuant to
§60.11{b) and for qualifying observers for visually determining
opacity of emissions.
2. Procedures. The observer qualified in accordance with paragraph 3
of this method shall use the following procedures for visually
determining the opacity of emissions:
2.1 Position. The qualified observer shall stand at a distance
sufficient to provide a clear view of the emissions with the sun oriented
in the 140° sector to his back. Consistent with maintaining the above
requirement, the observer shall, as much as possible, make his observations
from a position such that his line of vision is approximately
perpendicular to the plume direction, and when observing opacity of
emissions from rectangular outlets (e.g. roof monitors, open baghouses,
noncircular stacks), approximately perpendicular to the longer axis of
the outlet. The observer's line of sight should not include more than
one plume at a time when multiple stacks are involved, and in any case
141
-------
the observer should make his observations with his line of sight per-
pendicular to the longer axis of such a set of multiple stacks (e.g.
stub stacks on baghouses).
2.2 Field records. The observer shall record the name of
the plant, emission location, type facility, observer's name and
affiliation, and the date on a field data sheet (Figure 9-1). The
time, estimated distance to the emission location, approximate wind
direction, estimated wind speed, description of the sky condition
(presence and color of clouds), and plume background are
recorded on a field data sheet at the time opacity readings are
initiated and completed.
2.3 Observations. Opacity observations shall be made at the
point of greatest opacity in that portion of the plume where
condensed water vapor is not present. The observer shall not look
continuously at the plume, but instead shall observe the plume
momentarily at 15-second intervals.
2.3.1 Attached steam plumes. When condensed water vapor is
present within the plume as it emerges from the emission outlet, opacity
observations shall be made beyond the point in the plume at which con-
*t
densed water vapor Is no longer visible. The observer shall record the
approximate distance from the emission outlet to the point in the plume
at which the observations are made.
2.3.2 Detached steam plume. When water vapor in the plume condenses
and becomes visible at a distinct distance from the emission-outlet, the
opacity of emissions should be evaluated at the emission outlet prior to
142
-------
the condensation of water vapor and the formation of the steam plume.
2.4 Recording observations. Opacity observations shall be recorded
to the nearest 5 percent at 15-second intervals on an observational record
sheet. (See Figure 9-2 for an example.) A minimum of 24 observations
shall be recorded. Each momentary observation recorded shall be deemed
to represent the average opacity of emissions for a 15-second period.
2.5 Data Reduction. Opacity shall be determined as an average of
24 consecutive observations recorded at 15-second intervals. Divide the
observations recorded on the record sheet into sets of 24 consecutive
observations. A set is composed of any 24 consecutive observations.
Sets need not be consecutive in time and in no case shall two sets
overlap. For each set of 24 observations, calculate the average by
summing the opacity of the 24 observations and dividing this sum by 24.
If an applicable standard specifies an averaging time requiring more than
24 observations, calculate the average for all observations made during
the specified time period. Record the average opacity on a record sheet.
(See Figure 9-1 for an example.)
3. Qualifications and testing.
3.1 Certification requirements. To receive certification as a
qualified observer, a candidate must be tested and demonstrate the
ability to assign opacity readings in 5 percent increments to 25
different black plumes and 25 different white plumes, with an error
not to exceed 15 percent opacity on any one reading and an average
error not to exceed 7.5 percent opacity in each category. Candidates
shall be tested according to the procedures described in paragraph 3.2.
Smoke generators used pursuant to paragraph 3.2 shall be equipped with
a smoke meter which meets the requirements of paragraph 3.3.
143
-------
The certification shall be valid for a period of 6 months, at which
time the qualification procedure must be repeated by any observer in
order to retain certification.
3.2 Certification procedure. The certification test consists of show-
ing the candidate a complete run of 50 plumes--25 black plumes and 25 white
plumes—generated by a smoke generator. Plumes within each set of 25
black and 25 white runs shall be presented in random order. The
candidate assigns an opacity value to each plume and records his
observation on a suitable form. At the completion of each run of 50
readings, the score of the candidate is determined. If a candidate fails
to qualify, the complete run of 50 readings must be repeated in any
retest. The smoke test may be administered as part of a smoke school
or training program, and may be preceded by training or familiarization
runs of the smoke generator during which candidates are shown black and
white plumes of known opacity.
3.3 Smoke generator specifications. Any smoke generator used for
the purposes of paragraph 3.2 shall be equipped with a smoke meter
installed to measure opacity across the diameter of the smoke generator
stack. The smoke meter output shall display in-stack opacity based upon
a pathlength equal to the stack exit diameter, on a full 0 to 100 percent
chart recorder scale. The smoke meter optical design and performance
shall meet the specifications shown in Table 9-1. The smoke meter shall
be calibrated as prescribed in paragraph 3.3.1 prior to the conduct
of each smoke reading test. At the completion of each test, the zero
and span drift shall be checked and if the drift exceeds +1 percent
144
-------
opacity, the condition shall be corrected prior to conducting any
subsequent test runs. The smoke meter shall be demonstrated, at the
time of Installation, to meet the specifications listed 1n Table 9-1.
This demonstration shall be repeated following any subsequent repair
or replacement of the photocell or associated electronic circuitry
including the chart recorder or output meter, or every 6 months, which-
ever occurs first.
TABLE 9-1 SMOKE METER DESIGN AND PERFORMANCE SPECIFICATIONS
Parameter Specification
a. Light source Incandescent lamp operated at
nominal rated voltage
b. Spectral response Photopic (Daylight spectral
of photocell response of the human eye--
reference 4.3)
c. Angle of view 15° maximum total angle
d. Angle of projection 15° maximum total angle
e. Calibration error +_3% opacity, maximum
f. Zero and span drift +_ 1% opacity, 30 minutes
g. Response time <_ 5 seconds
3.3.1 Calibration. The smoke meter is calibrated after allow-
ing a minimum of 30 minutes warmup by alternately producing simulated
opacity of 0 percent and 100 percent. When stable response at 0
percent or 100 percent 1s noted, the smoke meter is adjusted to
produce an output of 0 percent or 100 percent, as appropriate. This
calibration shall be repeated until stable 0 percent and 100 percent
readings are produced without adjustment. Simulated 0 percent and 100
145
-------
percent opacity values may be produced by alternately switching the
power to the light source on and off while the smoke generator is not
producing smoke.
3.3.2 Smoke meter evaluation. The smoke meter design and per-
formance are to be evaluated as follows:
3.3.2.1 Light source. Verify from manufacturer's data and
from voltage measurements made at the lamp, as installed, that the
lamp is operated within +_ 5 percent of the nominal rated voltage.
3.3.2.2 Spectral response of photocell. Verify from manu-
facturer's data that the photocell has a photopic response; i.e.,
the spectral sensitivity of the cell shall closely approximate the
standard spectral-luminosity curve for photopic vision which is
referenced in (b) of Table 9-1.
3.3.2.3 Angle of view. Check construction geometry to ensure
that the total angle of view of the smoke plume, as seen by the
photocell, does not exceed 15°. The total angle of view may
be calculated from: o = 2 tan" d/2L, where e= total angle of view;
d = the sum of the photocell diameter + the diameter of the limiting
aperture; and L = the distance from the photocell to the limiting
aperture. The limiting aperture is the point in the path between the
photocell and the smoke plume where the angle of view is most
restricted. In smoke generator smoke meters this is normally an
orifice plate.
3.3.2.4 Angle of projection. Check construction geometry
to ensure that the total angle of projection of the lamp on the smoke
146
-------
plume does not exceed 15°. The total angle of projection may be
calculated from: e = 2 tan~ d/2L, where 9 = total angle of projection;
d = the sum of the length of the lamp filament + the diameter of the
limiting aperture; and L = the distance from the lamp to the limiting
aperture.
3.3.2.5 Calibration error. Using neutral-density filters of
known opacity, check the error between the actual response and the
theoretical linear response of the smoke meter. This check is
accomplished by first calibrating the smoke meter according to 3.3.1
and then inserting a series of three neutral-density filters of
nominal opacity of 20, 50, and 75 percent in the smoke meter pathlength.
Filters calibrated within +2 percent shall be used. Care should be
taken when inserting the filters to prevent stray light from affecting
the meter. Make a total of five nonconsecutive readings for each
filter. The maximum error on any one reading shall be 3 percent opacity.
3.3.2.6 Zero and span drift. Determine the zero and span
drift by calibrating and operating the smoke generator in a normal
manner over a 1-hour period. The drift is measured by checking the
ze^ and span at the end of this period.
3.3.2.7 Response time. Determine the response time by producing
a series of five simulated 0 percent and 100 percent opacity values
and observing the time required to reach stable response. Opacity
values of 0 percent and 100 percent may be simulated by alternately
switching the power to the light source off and on while the smoke
generator is not operating.
147
-------
4. References.
4.1 Air Pollution Control District Rules and Regulations,
Los Angeles County A1r Pollution Control District, Regulation IV,
Prohibitions, Rule 50.
4.2 Weisburd, Melvin I., Field Operations and Enforcement Manual
for Air, U. S. Environmental Protection Agency, Research Triangle Park,
N. C., APTD-1100, August 1972, pp. 4.1-4.36.
4.3 Condon, E. U., and Odishaw, H., Handbook of Physics,
McGraw-Hill Co., N. Y., N. Y., 1958, Table 3.1, p. 6-52.
148
-------
II
CO
^
O
Qi
UJ
OO
CO
o
Oi
UJ
>•
a;
CD
o
o;
UJ
ct:
UJ
CQ
C5
o
O-
or
tf
c_>
Q
U_
o
CO
O
O
a.
+J
QJ
QJ
QJ
'r-
QJ
i-
.
CL. cC
p: <_>
O O
<_) —I
a:
UJ
ca
'Z.
t—
oo
UJ
UJ UJ I—
I— CL. -ZL
~ o
Q t— CJ
r i T - -T--T- - - T-- - j
>-
1—
CJ
ct
CL.
O
UJ
c£
cv
UJ
^^
ej^
U_
0
CC
f^
y
^
— 1
VO
-^i
4->
•r—
O
CX
0
a
^_
H
QJ
ro1
CL'
«;
E
^3
CO
ID
c:
UJ
i
i
s_
ro
4-1
CO
S-
4-> QJ
QJ J2l
to E
—^
2: i
1
1
j
-^
O
0.
o
S"5
o
+J
E
o
s-
TZI
QJ
cn
c
ro
*~
t/i
en
JZ
•f—
•a
ro
QJ
/•v"
jj
ro
4->
^
QJ
O
C
ro
•f—
i —
CL
E
0
CJ
c
•r—
4->
O
W
J
^^^
CO
ro
5
QJ
U
i.
Z3
O
l/l
QJ
i
QJ
-a
ro
E
CO
ro
0
.J3
ro
™t
r—
ro
>
QJ
QJ
E
•r-
4_}
QJ
_L_>
CATION
to Discharge
o
_l QJ
rv ci
uj ro
> +->
(X CO
UJ •*—
CO 0
CO
o
from Discharge
cr
o
• r-
4J
O
QJ
t-
• r~
CD
Observation Po
4—
O
4->
JC
cn
•i —
a>
DESCRIPTION
o
— »
o
c£
CD
CJ
^£
CO
co
z
C5 d
>-. O
1— ••-
t— « 4-*
Q U
-^ QJ
0 i-
CJ ••-
/•y
uj ~o
2: c
h- -^-
UJ
3:
T3
QJ
QJ
CL
co
-a
c;
'T—
3:
emperature
i—
4J
C
QJ
• r^
JD
E
ro
QJ
u
in
z
o
»— «
l—
i~\
•^
0
CJ
>-
*^^
t/>
o
4->
QJ
T3
Z3
O
(J
a<
M
-M
t/>
ro
o
i.
QJ
>
O
•Z.
O
p^
Q-
rv
CJ
UJ
C3
UJ
— i
— i
CL.
S_
O
1 —
o
aj
•s*
QJ
(/i
•r—
o
149
-------
COMPANY
LOCATION
TEST NUMBER
DATE
FIGURE 9-2 OBSERVATION RECORD
OBSERVER
PAGE
OF
TYPE FACILITY
POINT OF EMISSIONS
Hr.
L .,..,1 !.,-.
Hin.
0
1
2
3
4
5
6
7
8
9
10
n
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
0
Seconds
15
30
4b
STEAM PLUME
(check if anplicable)
Attached
Detached
COMMENTS
150
-------
FIGURE 9-2 OBSERVATION RECORD
(Continued)
PAGE
OF
COMPANY
LOCATION
TEST NUMBER
DATE
OBSERVER
TYPE FACILITY
POINT OF EMISSIONS
Hr.
Min.
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
Seconds
0
15
30
45
STEAM PLUME
(check if applicable) u
Attached
Detached
COMMENTS
151
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-450/2-74-023
3 RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
EPA Response to Remand Ordered by U.S. Court of Appeals
for the District of Columbia in Portland Cement Assoc-
iation v. Ruckelshaus (486 F. 2d 375, June 29, 1973).
5 REPORT DATE
November 1974
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, N. C. 27711
10. PROGRAM ELEMENT NO.
11 CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
On June 29, 1973, the U.S. Court of Appeals for the District of Columbia remanded
to EPA the standards of performance for Portland cement plants promulgated under
section 111 of the Clean Air Act, directing the Agency to reconsider the standards
and provide additional explanation of the action taken. EPA has reviewed the standards
for these plants pursuant to the remand and has concluded that the standards other
than the opacity standard should not be revised. This document is a detailed technical
justification and explanation of the actions taken by the Administrator in
promulgating the standards for Portland cement plants. Included is the justification
for revising the opacity standard.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Ai r pol 1 uti on
Opacity standards
Visual observations
Portland cement plants
Particulate matter
Emi ssions
b.IDENTIFIERS/OPEN ENDED TERMS
Air pollution control
c. COSATI Field/Group
13. DISTRIBUTION STATEMENT
nlimited
19 SECURITY CLASS (This Report)
Unclassified
21. NO OF PAGES
152
20 SECURITY CLASS (TMspage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
752
-------
-------
m
£ °
' -n
o 2
2 Q
>
O
3)
m
U)
VI
ff <
« -i 5
oj » o
" 23-1
=• 3: = 3
-loo?
2
«|?s
•o 3 o o
" = SS
jr- — - — • (Tl
- SJ o n
— 2 2 -i
N)
§ §
m
31
r <^
m
Z
n
"Td
C
da
r
HH
n
>
H
»—i
O
21
^
O
•
w
^d
>
4^
^n
O
K)
i
-4
^
\
O
tsJ
V>i
— CD
CL m rt
CD 5 3"
2.^
CD
K> CD
CL ^
Q
—\ —*»
S<§
co O
CO CL
_ O
CD 0
C^3
tts
' co
CL CD
— n
O CD
E<
CD CD
§5-
J
CO
Tl
m
O
X
O-o
?0
Z >
m -< O
o
(S)
1 X*
So°
"mS
O m
—t vi
o?
>0
o
m
Z
n
------- |