STACK SAMPLING 1095
every few seconds. A CEMS for SO 2 extracts a small sample
of stack gas and sends it to a monitor outside the stack for
analysis and recording. The CEMS records present an excel-
lent picture of the continuing status of compliance of the stack
gas, reveal any inconsistencies, failures of control devices, or
process upsets, and allow plant operators to act immediately
when any anomaly appears. It is likely that more and more
CEMS will be required as they become available. CEMS are
becoming more available for mercury, particulate matter and
ammonia.
While detailed discussion of CEMS technology, proce-
dures, and operation cannot be presented here, the Performance
Specifications for the CEMs currently required by USEPA
may be found in Appendix B to 40 CFR Part 60, immediately
following the Test Methods.
A third major advance in stack testing is the introduc-
tion of the 300 series methods found in 40 CFR Part 63
Appendix A. In an effort to measure the hazardous air pollut-
ants to demonstrate emission reductions pursuant to MACT
requirements, EPA has introduced a number of 300 series
methods. Some of these methods are specific for pollutants
from specific sources. Method 301, which is undergoing revi-
sion in 2005, is a field validation procedure that will enable
source owners to validate their own test methods in the
absence of a recognized EPA method.
The fourth is the rewrite of the manual methods into a
standard format. The instrumental methods are being rewrit-
ten in 2005, along with some major changes.
The following sections of this chapter will deal with
the selection of sampling and analytical methods, followed
by some general information on protocol and final report
preparation.Selection of a Sampling MethodThe choice of stack sampling methodology is most strongly
dependent on the pollutants to be measured. In many cases,
similar pollutants can be measured with the same or slightly
modified methods, while dissimilar pollutants may require
totally different methods.
The most obvious categorization of pollutants is based
on their physical state: gaseous, liquid, or solid. Gaseous
pollutants are generally easier to sample and can be collected
using one of a few simple train configurations. Liquid and
solid pollutants, usually lumped into a single category called
particulates, require a totally different collection concept.
The following sections will describe the generalized
methods employed for collecting particulate and gaseous
stack samples. These are followed by more detailed descrip-
tions of methods to be used for specific compounds, or groups
of compounds. When the U.S. Environmental Protection
Agency (EPA) has designated a method as a Test Method, it is
required by EPA and by most states for compliance determi-
nations. For convenience, the appropriate EPA designations
are indicated. When a Test Method is used in establishing an
emission limit and is specified by a regulation that method is
called the Reference Method.Particulate SamplingWhen the pollutant of interest is or is attached to solid par-
ticles or liquid droplets at stack conditions, it is necessary to
select a method that physically traps the particles. But the
first step must be the selection of a side stream that is truly
representative of the stack exhaust gases.
A representative sample of the stack exhaust gas will look
and behave like a small-scale version of the actual exhaust
gas. It will contain the same fraction of particulates as the
main stream (including the same ratios of large and small
particles) and will contain a fair share of material from each
part of the stack cross section. (This is necessary because
gases flow faster near the center of a stack and slower near
the walls due to friction).
In addition a representative sample must be taken at a
location that is free of unusual flow patterns, such as cyclonic
flow (in which a significant component of the flow is not
along the axis of the stack) or stratified flow (in which the
particulates are bunched along one side of the stack). This
is because it is very difficult to figure out the actual aver-
age flow rate and particulate rate when the measurements
are all skewed by the flow anomalies. In most cases, flow
abnormalities are caused by recognizable disturbances, such
as bends, fans, expansions, contractions, or shape changes
in the duct. These disturbances, whether upstream or down-
stream from the sampling location, have the potential for
making useful testing very difficult, or even impossible. For
that reason, the first criterion for good particulate testing is
to find a location that is sufficiently far from flow distur-
bances. Extensive testing has shown that a sampling location
8 stack diameters downstream from any disturbance and 2
diameters upstream from any disturbance is sufficiently far.
In this measurement, the term stack diameter is used liter-
ally for circular stacks. For rectangular stacks, an equivalent
stack diameter is calculated.
In some cases, it is impossible to find a location in the
stack or in any straight duct leading to the stack that satisfies
these criteria. It is possible to use a location closer to distur-
bances. However, other provisions must be taken to account
for the possible inaccuracies introduced by the disturbed
flow. All of this is described in detail in Test Method 1.
Once a sampling location is selected, it is necessary to
collect samples of the gas stream that are representative of
the gas flowing by that location. This is achieved by sam-
pling for a short time at each of several points across the
stack cross-section.
In practice, two or more holes, or ports, are cut in the
stack wall and a sampling probe inserted. The probe is
essentially a hollow tube shaped like a shepherd’s crook
with the short end, or nozzle, facing into the gas stream. The
gas stream is then pumped by suction from the main stream
through the nozzle and probe into the collection part of the
sampling train located outside the stack. The probe is held in
one spot, aligned into the main stream, for a specified time.
It is then moved to another point and held for the same time.
This process is repeated along the line between the port and
the opposite wall. The process of moving the probe alongC019_003_r03.indd 1095C019_003_r03.indd 1095 11/18/2005 11:07:14 AM11/18/2005 11:07:14 AM