1110 STACK SAMPLING
the opacity with a transmissometer, or smoke meter calibrated
accurately to within 2%. Once the observers have “cali-
brated their eyes” to identify emissions of different opacities,
they are given a test consisting of 25 black smoke readings and
25 white smoke readings. An observer first takes a classroom
training session then passes the two sets of smoke observa-
tions and is thus certified to “read smoke.” To pass, the aver-
age error for each color smoke must be no greater than 7.5%
and no individual reading more than 15% from the correct
value. Observers must be recertified every six months, need-
ing to pass the test but not take the classroom training.
The remainder of the method describes the procedure to
be followed by the observer in observing actual smoke emis-
sions and for recording them. The basic idea for performing
visible emission (VE) observations is to try and minimize
the influence of those factors that might bias the results of
the observations or might cause them to be unreliable. The
four major variables that can be controlled by the observer
all relate to position. They are: distance from the emission
point, the viewing angle relative to the direction of the emis-
sion, the angle relative to the sun, and the angle relative to
the wind. The goal is to keep the sun to your back, the wind
to your side, the emission at about eye level, and at a suf-
ficient distance to allow comfortable viewing.
VE observations are momentary glances at the densest
part of the plume every 15 seconds. They are recorded in
5% increments on an observational record sheet that also
contains a diagram of the observer/source/conditions and
other relevant information. Normally, 24 consecutive read-
ings constitute a minimum data set, representing the aver-
age emissions over that 6-minute period. However, more or
fewer observations may be required or allowed by a specific
regulation.
More detailed instructions for operating a certification
program and for making VE observations is included in the
method. However, in almost all cases, a certification course
would be offered by an experienced organization. In the case
of this method, the admonition about experienced personnel
does not apply. Anyone who can pass the test can perform
VE observations.Test Method 9, Alt 1Alternative 1 to Test Method 9 allows for the remote obser-
vation of emission plume opacity by a mobile lidar system,
rather than by human observers, as specified in Method 9.
Lidar is a type of laser radar (Light Detection and Ranging)
that operates by measuring the backscatter of its pulsed laser
light source from the plume. It operates equally well in day-
light or at night.
The lidar unit shall be calibrated at least once a year and
shall be subjected to one of two routine verification proce-
dures every four hours during observations. The calibration
shall be performed either on the emissions from a Method 9
smoke generator or using screen material of known opacity.
The system is considered to be in calibration if the lidar’s
average reading is within 5% for the smoke generator,
or within 3% for the screens. The routine verificationprocedures require either the use of neutral density filers or an
optical generator. In either case, average readings within 3%
are required.
The actual operation of the lidar system is conceptually
straightforward but technically complex. The unit is posi-
tioned with an unobstructed view of a single emission, the
backscatter recorded, and the plume opacity calculated.
However, the actual procedures for accomplishing this must
be performed only by personnel experienced at operating
lidar units and interpreting their results. Detailed criterion
are provided in the method for both.Test Method 10Test Method 10 is used for the determination of carbon mon-
oxide (CO) emissions from stationary sources employing a
nondispersive infrared analyzer (NDIR). Either an integrated
or continuous sample can be extracted from the source for
analysis by this method.
The analyzer compares the infrared energy transmitted
through a reference cell to the same length cell containing a
sample of the stack gases. A filter cell is used to minimize the
effects of CO 2 and hydrocarbons. The use of ascarite traps
further reduces the interference of carbon dioxide. Silica gel
is used to remove water which would also interfere with the
measurement of CO. Both the ascarite and silica gel traps are
placed in an ice bath.
The EPA Test Method 10 write-up contains detailed
instructions for calibrations, leak check procedure, and
sampling, along with a list of references. It should be read
in detail before the Method is attempted. As with all of
these methods, testing should be performed only by trained
and experienced personnel using equipment and analyzer
designed for this purpose.Test Method 11Test Method 11 is used to determine the hydrogen sulfide
(H 2 S) content of fuel gas streams in petroleum refineries.
The H 2 S is absorbed by cadmium sulfate solution to form
cadmium sulfide.
The sampling train is similar to the one used for Method 6.
Five midget impingers are used. The first contains hydrogen
peroxide to absorb SO 2 , which would interfere with the anal-
ysis. The second impinger is empty to prevent carry-over of
the peroxide into the cadmium sulfate absorbing solution in
the last three impingers. If the gas to be sampled is pressur-
ized, a needle valve is used to regulate the flow. If the pres-
sure is not sufficient, a diaphragm pump is used to draw the
sample through the train. A sampling rate of 1 (10%) liter
per minute is maintained for at least 10 minutes. As with
Method 6, the train is purged after sampling is complete.
After sampling, the peroxide solution, containing the
SO 2 trapped as H 2 SO 4 , is discarded. The contents of the third,
fourth, and fifth impingers are quantitatively transferred to
a 500 ml flask. Excess acidic iodine solution is added to the
flask. After allowing 30 minutes for the sulfide to react with
the iodine, the amount of excess iodine is determined byC019_003_r03.indd 1110C019_003_r03.indd 1110 11/18/2005 11:07:16 AM11/18/2005 11:07:16 AM