&EPA
           United States
           Environmental Protection
           Agency
           Office of Air Quality
           Planning and Standards
           Research Triangle Park NC 27711
EPA-450/2-78-042d
October 1978
           Air
Stack Sampling
Technical Information
A Collection of
Monographs and Papers
Volume IV

-------
                                  EPA-450/2-78-042d
Stack Sampling Technical  Information
A Collection of Monographs and Papers
                  Volume IV
             Emission Standards and Engineering Division
             U S ENVIRONMENTAL PROTECTION AGENCY
                Office of Air, Noise, and Radiation
              Office of Air Quality Planning and Standards
             Research Triangle Park, North Carolina 27711

                     October 1978

-------
This report has been reviewed by the Emission Standards and Engineering
Division,  Office of Air  Quality Planning and Standards, Office of Air, Noise
and Radiation,  Environmental Protection Agency,  and approved for publica-
tion. Mention of company or product names does not constitute endorsement
by EPA. Copies are available free of charge to Federal employees, current
contractors and grantees, and non-profit organizations - as supplies permit
from the Library Services Office, MD-35, Environmental Protection Agency,
Research  Triangle Park, NC 27711;  or may be obtained, for a fee, from the
National Technical Information Service, 5285 Port Royal Road, Sprinqfield
VA 22161.
                     Publication No. EPA-450/2-78-042d
                                  n

-------
                                 PREFACE


     The  Clean Air Act of  1970 requires the Administrator of the
Environmental Protection Agency to establish national emission standards
for new stationary sources  (Section  111) and hazardous air pollutants
(Section  112).  The development of these emission standards required the
concurrent development of reference  test methods and procedures.  The
reference test methods and  procedures are published in the Federal Register
along with the appropriate  regulations.

     From time to time, questions would surface concerning the methods and
procedures.  In many cases, specific studies would be needed to provide
informed, objective answers.  The papers and monographs resulting from these
studies were usually distributed to  people involved in emission measurement;
a major method of distribution has been the Source Evaluation Society
Newsletter.

     To provide a readily available  resource for new and experienced personnel,
and to further promote standardized  reference methods and procedures, it has
been decided to publish the  papers   and monographs in a single compendium.
The compendium consists of  four volumes.  The Table of Contents for all
four volumes is reproduced  in each volume for ease of reference.

     Congratulations and sincere appreciation to the people who did the
work and  took the time to prepare the papers and monographs.  For the most _
part the work was done because of personal commitments to the development
of objective, standardized methodology, and a firm belief that attention
to trie details of stack sampling makes for good data.  The foresight of
Mr. Robert L. Ajax, the former Chief of the Emission Measurement Branch and
now the Assistant Director, Emission Standards and Engineering Division, in
providing the atmosphere and encouragement to perform the studies is
gratefully acknowledged.  The skill  and dedication of Mr. Roger Shigehara,
in providing personal supervision for most of the work, is commended.
                                          Don R. Goodwin
                                             Director
                                      Emission Standards and
                                       Engineering Division
                                       m

-------
                               VOLUME I

                           TABLE OF CONTENTS


 Method for Calculating Power Plant Emission Rate                      i

      by R. T. Shigehara, R. M. Neulicht, and W. S. Smith
                    •

 Emission Correction Factor for Fossil  Fuel-Fired Steam               in
 Generators (C02 Concentration Approach)

      by R. M. Neulicht


 Derivation of Equations  for Calculating Power Plant  Emission          20
 Rates  (02  Based Method - Wet and  Dry Measurements)

      by R.  T.  Shigehara  and R.  M.  Neulicht


 Summary of F  Factor Methods  for Determining  Emissions  from            29
 Combustion  Sources

     by R.  T.  Shigehara,  R.  M.  Neulicht, W.  S.  Smith,
                and  J.  W.  Peeler


 Validating  Orsat Analysis Data  from Fossil-Fuel-Fired Units          44

     by R.  T. Shigehara,  R.  M.  Neulicht, and W. S. Smith


A Guideline for Evaluating Compliance Test Results                   55
 (Isokinetic Sampling Rate Criterion)

     by R. T. Shigehara

-------
                              VOLUME II

                           TABLE  OF  CONTENTS



A Type-S  Pi tot  Tube Calibration  Study                                 1

      by Robert  F.  Vollaro


The Effect  of Aerodynamic Interference Between a Type-S              24
Pi tot Tube  and  Sampling  Nozzle on the Value of the
Pi tot Tube  Coefficient

      by Robert  F.  Vollaro


The Effects of  the Presence  of a Probe Sheath on Type-S              30
Pi tot Tube Accuracy

      by Robert  F.  Vollaro


An Evaluation of Single-Velocity Calibration Technique as            48
a Means of Determining Type-S Pitot Tube Coefficients

     by Robert  F.  Vollaro


Guidelines for  Type-S Pitot  Tube Calibration                         63

     by Robert  F.  Vollaro                                             :
The Effects of  Impact Opening Misalignment on the Value of           89
the Type-S Pitot Tube Coefficient

     by Robert  F. Vollaro


Establishment of a Baseline Coefficient Value for Properly           95
Constructed Type-S Pitot Tubes

     by Robert  F. Vollaro


A Survey of Commercially Available Instrumentation for the          104
Measurement of  Low-Range Gas Velocities

     by Robert  F. Vollaro


The Use of Type-S Pitot Tubes for the Measurement of Low            122
Velocities

     by Robert  F. Vollaro
                                 vi

-------
                              VOLUME III
                          TABLE OF  CONTENTS

Thermocouple Calibration Procedure  Evaluation                         1
     by Kenneth Alexander

Procedure for Calibrating and Using Dry Gas Volume Meters            10
As Calibration Standards
     by P. R. Westlin and R. T. Shigehara

Dry-Gas Volume Meter Calibrations                                    24
     by Martin Wortman, Robert Vollaro, and Peter Westlin

Calibration of Dry Gas Meter at Low Flow Rates                       33
     by R. T. Shigehara and W. F. Roberts

Calibration of Probe Nozzle Diameter                                 41
     by P. R. Westlin and R. T. Shigehara

Leak Tests for Flexible Bags                                         45.
     by F. C. Biddy and R. T. Shigehara

Adjustments in the EPA Nomograph for Different Pitot Tube            48
Coefficients and Dry Gas Molecular Weights
     by R. T. Shigehara

Expansion of EPA Nomograph (Memo)                                    60
     by R. T. Shigehara

EPA Nomograph Adjustments (Memo)                                     63
     by R. T. Shigehara
Graphical Technique for Setting Proportional Sampling                65
Flow Rates
     by R. T. Shigehara
                                 vi i

-------
                               VOLUME  IV
                           TABLE OF CONTENTS

 Recommended Procedure for Sample Traverses in Ducts  Smaller           1
 Than 12 Inches in Diameter
      by Robert F. Vollaro

 Guidelines for Sampling in Tapered Stacks                           24
      by T. J.  Logan and R. T. Shigehara

 Considerations for Evaluating Equivalent Stack Sampling              28
 Train Metering Systems
      by R. T.  Shigehara

 Evaluation of  Metering Systems for Gas-Sampling Trains               40
      by M. A.  Wortman and R.  T.  Shigehara

 An  Evaluation  of the Current  EPA Method 5 Filtration                 49
 Temperature-Control  Procedure
      by Robert F. Vollaro

 Laboratory Evaluation of Silica  Gel Collection  Efficiency            67
 Under Varying  Temperature and Pressure  Conditions
      by  Peter R. Westlin and Fred C. Biddy

 Spurious  Acid  Mist Results Caused  by  Peroxides  in Isopropyl          79
 Alcohol  Solutions Used in EPA Test Method 8  (Memo)
      by Dr.  Joseph E.  Knoll

 Determination  of Isopropanol  Loss  During Method 8 Simulation         80
 Tests  (Memo)
      by Peter  R.  Westlin

 Comparison of  Emission Results from In-Stack  Filter Sampling         82
 and  EPA Method 5 Sampling
      by Peter  R.  Westlin and  Robert L.  Ajax

EPA Method 5 Sample Train  Clean-Up Procedures                        93
     by Clyde  E.  Riley

                                 viii

-------
               RECOMMENDED PROCEDURE FOR SAMPLE TRAVERSES
               IN DUCTS SMALLER THAN 12 INCHES IN DIAMETER
                           Robert F.  Vollaro**
                             INTRODUCTION
     In source sampling, stack gas velocity Is usually measured with a
Type-S pi tot tube.  In many field applications, the pi tot tube is  attached
to a sampling probe, equipped with a nozzle and thermocouple.    This combi-
nation is called a pitobe assembly.  Most conventional pitobe  assemblies*
have a cylindrical sampling probe of 1-inch diameter, but, occasionally,
an assembly has an external cylindrical  sheath of about 2-1/2  inches in
diameter, encasing the probe, pi tot tube and thermocouple.  When a pitobe
assembly is used to traverse a duct that is 36 inches or less  in diameter,
the pitobe assembly can "block" a significant part of the duct cross section,
as illustrated in the projected-area models, Figures la and Ib.  This reduction
in the effective cross-sectional area of the duct causes a temporary, local
increase in the average velocity of the flowing fluid.  In most pitobe
assemblies, the impact opening  of the Type-S pi tot tube lies in approximately
the same plane as the probe sheath (Figure 2) and, whenever appreciable sheath
blockage exists, velocity head (&P) readings made with the pi tot tube tend
to reflect the local increase in gas velocity, and are not truly representa-
                                                1  2
tive of the mainstream velocity.  Recent studies '   have shown that, for
sample traverses in ducts having diameters or equivalent diameters between
12 and 36 inches, blockage effects are not particularly severe, and a simple
*Designed according to the specifications outlined in APTD-0581  (Reference 3),
 or allowable modifications thereof.
** Emission  Measurement  Branch,  ESED,  OAQPS,  EPA,  RTP, NQ January 1977

-------
                  (a)
                             ESTIMATED
                             ' SHEATH
                             BLOCKAGE
 **"•   I
DUCT AREAj
x 100
                                 (b)
                                                                                                               PO
Figure 1. Projected-area models for typical pitobe assemblies; shaded area represents approximate
average sheath blockage for a sample traverse.

-------
            SAMPLING
             PROBE
                TYPE-S   '
              PITOTTUBE
STATIC PRESSURE
    OPENING,
                           I
1
                         APPROXIMATE
                        PLANE OF PROBE
                       SHEATH BLOCKAGE
                    IMPACT PRESSURE
                       OPENING
       FLOW
     DIRECTION
Figure 2. Type-S pilot tube, attached to a sampling probe, showing that the pitot impact
opening and probe sheath lie in approximately the same plane.

-------
adjustment in the value of the Type-S pi tot tube coefficient (C )  can  be
made to compensate for the pseudo-high AP readings (Figure 3).   When the
duct diameter (Ds) is less than  12 inches, however, probe sheath blockage
effects intensify, and the adjustment technique illustrated in  Figure  3 no
longer applies.  Therefore, alternative methodology must be used in order
to obtain representative sample  traverses in ducts of this size.  The
purpose of this paper is to propose a method by which satisfactory sample
traverses can be conducted when  DS is between 4 and 12 inches.

                  PROPOSED METHOD FOR SAMPLE TRAVERSES
                        WHEN 4 in. <_ DS < 12 in.
METHODOLOGY
     To conduct representative sample traverses in ducts having diameters
between 4 and 12 inches, it is recommended that the arrangement illustrated
in Figure 4 be used.  In Figure  4, velocity head (AP) readings  are taken
downstream of the actual sampling site.  The purpose of the straight run of
duct between the sampling and velocity measurement sites is to  allow the flow
profile, temporarily disturbed by the presence of the sample probe, to redevelop
and stabilize.  The pitot tube and sampling nozzle shown in Figure 4 are
different from those of a conventional pitobe assembly;3 construction  details
of these components are discussed below.
     A.  Pitot tube.
         A standard pitot tube shall be used, instead of a Type-S, to  monitor
stack gas velocity.  When DS is  less than 12 inches, a Type-S pitot tube can
begin to block a significant part of the duct cross section and yield

-------
      _
2 1/2-in. CYLINDRICAL MODEL.
   USE FOR ASSEMBLIES
 WITH EXTERNAL SHEATHSv
o
5
u

o
u- O
u£ 3
 CO
 o
                    1-in. CYLINDRICAL MODEL.
                      USE FOR ASSEMBLIES
                   WITH NO EXTERNAL SHEATHS
        i	          1234
                              DECREASE IN PITOT TUBE COEFFICIENT. percent
         Figure 3.  Adjustment of Type-S pitot tube coefficients to account for sheath blockage
         (12 in. 
-------
   FLOW
DISTURBANCE

-* 	 2DS-

1 (


	 1
> i


^ 	 o us 	 ^
^ t 	 '
V\
M- 0 D


I
1 i

s
0,
"s

S


   STANDARD1
    PITOT
     TUBE
                             TEMPERATURE
                               SENSOR
•SAMPLING
  PROBE
                                                                                                                   CTl
   FLOW
DISTURBANCE
                       Figure 4. Recommended sampling arrangement, when 4 in. < Ds < 12 in.

-------
                                    7
pseudo-high AP values.   Cross-section blockage is not a serious  problem
with a standard pi tot tube, however, for two reasons:  (1)  the impact and
static pressure openings of a standard pi tot tube, unlike those  of a
Type-S, follow a 90° bend, and are located well  upstream of the  stem of
the tube (compare Figures 2 and 5); and (2) when properly aligned, the
sensing head of a standard pitot tube is parallel, not perpendicular, to
the flow streamlines in the duct.
     The preferred design for the standard pitot tube is the hemispherical-
nosed design (Figure 5).  Pitot tubes constructed according to the criteria
                                                             4  5
illustrated in Figure 5 will have coefficients of 0.99 ±0.01 '   .  Note,
however, that for convenient tubing diameters (dimension "D" Figure 5), the
static and impact sensing holes of the hemispherical-type pitot tube will
be very small, thus making the tube susceptible to plugging, in particulate
or liquid droplet-laden gas streams.  Therefore, whenever these conditions
are encountered, either of the following can be done: (1) a "back purge"
system of some kind can be used to clean out, periodically, the static and
impact holes; or  (2) a modified hemispherical-nosed pitot tube (Figure 6),
which features a shortened stem and enlarged impact and static pressure holes,
can be used instead of the conventional hemispherical type.  It has recently
been demonstrated that the coefficients of the conventional and modified
hemispherical-nosed tubes are essentially the same.
     B.  Sampling nozzle.
         The sampling nozzle can either be of the buttonhook or elbow design.
The nozzle shall meet the general  design criteria specified in Section 2.1.1
of  the  revised version of  EPA Method  5, except that  the entry plane of the

-------
                 T
                                                r~3D
                          STATIC PRESSURE OPENINGS -
                                  ~0.1 D
                              IMPACT PRESSURE
                                 OPENING
                                    0.4D
                                                                       16 D
                                                                                     oo
                                                                       80
Figure 5. Hemispherical-nosed standard pitot tube.

-------
                                                      4STATIC HOLES
                                                           3/8 D
                                                     IMPACT OPENING
                                                          1/20
Figure 6. Modified hemispherical-nosed pitot tube.

-------
                                    10
nozzle must be at least 2 nozzle diameters (i.d.) upstream of the probe
sheath blockage plane  (see  Figure 7).
PROCEDURES
     The following procedures shall be used to perform sample traverses
using the arrangement  illustrated in Figure 4:
     A.  Location of sampling site.
         Select a sampling  site that is at least 8 duct diameters downstream
and 10 diameters upstream from the  nearest flow disturbances; this allows
the velocity measurement site to be located 8 diameters downstream of the
sampling location and  2 diameters upstream of the nearest flow disturbance
(see Figure 4).  For rectangular stacks,  use an equivalent diameter, calcu-
lated from the following equation,  to  determine the upstream and downstream
distances:
          De  =  L^tf                                      (Equation l)

      Where:
          D  = Equivalent diameter
          L  = Length of cross section
          W  = Width of cross section
 If a sampling site located 8 diameters downstream and 10  diameters  upstream
 from the nearest disturbances is not available, select a  site that  meets
 these criteria as nearly as possible.  Under no circumstances, however, shall
 a sampling site be chosen which is less than 2 diameters  downstream and 2.5
 diameters  upstream from the nearest disturbances; this guarantees a minimum

-------
                         PLANE OF
                          SHEATH
                         BLOCKAGE
Figure 7.  Recommended sampling nozzle design for use when 4 in. < Ds < 12 in.

-------
                                    12
of 2 diameters of straight run between the sampling and velocity measure-
ment sites, and 0.5 diameters between the velocity measurement site and
the nearest flow disturbance.
     B.  Number of traverse  points.
         The correct  number  of traverse  points shall be determined from
Figure 8.  To  use Figure  8,  proceed  as follows:  first, determine the three
distances, MA",  "B",  and  "C",  and express each distance in terms of duct
diameters; second, read from Figure  8 the number of  traverse points
corresponding  to each of  these three distances;  third, select the highest
of the  three  numbers  of traverse points, or a greater  number, so that  for
circular ducts the  number is a multiple  of  4; for  rectangular ducts, the
number  should be chosen so that  it  is one of those shown  in
Table 2.
      C.  Location of traverse points, circular cross sections.
         For circular stacks, locate the traverse points on 2 perpendicular
 diameters, according to Table 1  and the example of Figure 9a.   Any traverse
 point located less than 1/2 inch from the stack wall will not be acceptable
 for use as a  sampling  point; all  such traverse  points shall  be "adjusted"
 by relocating them to  a  distance of 1/2 inch from the wall.   In some cases,
 this relocation process  may involve combining two adjacent traverse points
 to form a single "adjusted" point;  thus, in some  instances, the number of
 points  actually used for sampling may be less than  the number of traverse
 points  obtained from Figure 8.
      D.  Location of traverse points,  rectangular cross  sections.
           For rectangular stacks, divide the cross section into as many
 equal  rectangular  elemental areas as traverse points  (as

-------
NUMBER OF DUCT DIAMETERS BETWEEN VELOCITY MEASUREMENT SITE AND NEAREST DISTURBANCE  "
                                  DISTANCE C
                                                                             10
      NUMBER OF DUCT DIAMETERS BETWEEN SAMPLING SITE AND NEAREST DISTURBANCE
                                ! DISTANCE A
                                     OR
     NUMBER OF DUCT DIAMETERS BETWEEN SAMPLING AND VELOCITY MEASUREMENT SITES.
                                 DISTANCE B
         Figure 8.  Minimum number of traverse points, 4 in. < Ds < 12 in.

-------
      Table 1. LOCATION OF TRAVERSE POINTS
     IN CIRCULAR STACKS (PERCENT OF STACK
DIAMETER FROM INSIDE WALL TO TRAVERSE POINT)
Traverse
point
number
on a
diameter
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Number of traverse points on a diameter
2
14.6
85.4






















4
6.7
25.0
75.0
93.3




















6
4.4
14.6
29.6
70.4
85.4
95.6


















8
3.2
10.5
19.4
32.3
67.7
80.6
89.5
96.8
















10
2.6
8.2
14.6
22.6
34.2
65.8
77.4
85.4
91.8
97.4














12
2.1
6.7
11.8
17.7
25.0
35.6
64.4
75..0
82.3
88.2
93.3
97.9












14
1.8
5.7
9.9
14.6
20.1
26.9
36.6
63.4
73.1
79.9
85.4
90.1
94.3
98.2










16
1.6
4.9
8.5
12.5
16.9
22.0
28.3
37.5
62.5
71.7
78.0
83.1
87.5
91.5
95.1
98.4








18
1.4
4.4
7.5
10.9
14.6
18.8
23.6
29.6
38.2
61.8
70.4
76.4
81.2
85.4
89.1
92.5
95.6
98.6






20
1.3
3.9
6.7
9.7
12.9
16.5
20.4
25.0
30.6
38.8
61.2
69.4
75.0
79.6
83.5
87.1
90.3
93.3
96.1
98.7




22
1.1
3.5
6.0
8.7
11.6
14.6
18.0
21.8
26.2
31.5
39.3
60.7
68.5
73.8
78.2
82.0
85.4
88.4
91.3
94.0
96.5
98.9


24
1.1
3.2
5.5
7.9
10.5
13.2
16.1
19.4
23.0
27.2
32.3
39.8
60.2
67.7
72.8
77.0
80.6
83.9
86.8
89.5
92.1
94.5
96.8
98.9

-------
                15
Figure 9a. Cross section of circular stack divided into
12 equal areas, showing location of traverse points.
        o     o  I   o

        o  !  o  I   o  I

     —H--i
        o     o  I   o  ,
           I      I       I
Figure 9b. Cross section of rectangular stack divided
into 12 equal areas, with traverse points at centroid
of each area.

-------
                                    16
determined in Section "B" above), according to Table 2.  Locate
a traverse point at the centroid of each elemental area, according  to  the
example of Figure 9b.
     E.  Sampling.
         Sample at each non^adjusted traverse point for the time interval
specified in the method being used (e.g., Method 5).  If two successive
traverse points have been relocated to a single "adjusted" traverse point,
sample twice as long at the adjusted point as at non-adjusted points,  taking
twice as many readings, but record the data as though two separate  points
had been sampled, each for half of the total time interval.  During the
sample run, velocity head (AP) readings shall be taken at points downstream
of, but directly in line with, the sampling points.  The sampling rate
through the nozzle shall be set based upon the AP readings; if a nomograph
is used, be sure when setting it to use the correct value (~ 0.99)  of the
pi tot tube coefficient.
            ALTERNATIVE SAMPLING STRATEGY  (STEADY-FLOW  ONLY)
     If the average total volumetric  flow  rate in a duct is constant with
time, it is unnecessary to monitor stack gas velocity during a sample run.
Thus, whenever  time-invariant flow is believed to exist in a stack (e.g.,
for a steady-state process),  the following  traverse procedures can be used
in lieu of those outlined  in  the preceding  sections:
     A.  Location of  Sampling-Velocity  Measurement  Site.
         When steady  flow  is  believed to exist in a duct,  the sample and
velocity traverses can  be  conducted  non-simultaneously; therefore, the
sampling and  velocity measurement  sites  need  not  be separate.  Rather, a

-------
                              17
Table 2.    CROSS-SECTIONAL LAYOUT FOR RECTANGULAR STACKS
     No. of traverse                        Layout
          points
            9                               3x3
           12                               4x3
           16                               4x4
           20                               5x4
           25                               5x5
           30                               6x5
           36                               6x6
           42                               7x6
           49                               7x7

-------
                                    18
single location can be used for both sampling and velocity measurement
(see Figure 10).
         Select a sampling-velocity measurement site that is at least
8 duct diameters downstream and 2 diameters upstream from the nearest
flow disturbances.  For rectangular stacks, use an equivalent diameter
(Equation 2) to determine the upstream and downstream distances.  If a
sampling-velocity measurement site located 8 diameters downstream and
2 diameters upstream from the nearest disturbances is not available,
choose a site that meets these criteria as nearly as possible.  Under  no
circumstances, however, should a sampling-velocity measurement site be
chosen that is less than 2 diameters downstream and 0.5 diameter upstream
from the nearest disturbances.
     B.  Number of Traverse  Points.
         The correct number  of traverse points shall be determined from
Figure 11.  To use Figure  11, proceed as  follows:  first, determine the
distances  "A" and "B"  and  express each distance in terms of duct diameters;
second,  read from Figure 11  the  number of traverse points corresponding to
each of  these distances; third,  select the higher  of these two numbers of
traverse points, or a  greater number, so  that  for circular  ducts the number
is  a multiple of 4 and,  for rectangular ducts, the number is one of  those
 shown in Table 2.
      C.   Location of  Traverse Points,  Circular Cross  Sections
          For circular stacks, locate the  traverse  points  on 2  perpendicular
 diameters, according  to Table 1  and the example  of Figure 9a.   Any  traverse
 point located  less  than 1/2 inch from the stack  wall  will  be  unacceptable

-------
   FLOW
DISTURBANCE
                     SAMPLING-VELOCIT'i
                       MEASUREMENT
                           SITE
                                                                                       FLOW
                                                                                    DISTURBANCE
                                                                                                            vo
        Figure 10. Recommended sampling arrangement; 4 in. < Ds < 12 in.; steady-flow only.

-------
     NUMBER OF DUCT DIAMETERS UPSTREAM (FROM NEAREST FLOW DISTURBANCE).
            ,  .                 DISTANCE A
                                                       2.0
                                                                    FLOW
                                                          \   / DISTURBANCE


                                                        A   T   SAMPLING
                                                                  AND
                                                                                                ro
                                                                                                o
     NUMBER OF DUCT DIAMETERS DOWNSTREAM  (FROM NEAREST FLOW DISTURBANCE),
                               DISTANCE B
Figure 11.  Minimum number of traverse points; 4 in. < Ds < 12 in.; steady-flow only.

-------
                                    21
for use, either as a velocity traverse  point or as  a  sample  point;  all
such points shall be "adjusted" by relocating them  to a distance  of 1/2
inch from the wall.  In some cases, this relocation process  may involve
combining two adjacent traverse points  to form a single "adjusted"  point;
thus, the number of traverse points actually used will sometimes  be less
than the number of points obtained from Figure 11.
     D.  Location of Traverse Points, Rectangular Cross Sections.
         For rectangular stacks, divide the cross section into as many
equal rectangular elemental areas as traverse points (as
determined in Section "B" above), according to Table 2.
Locate a traverse point at the centroid of each elemental area, according
to the example of Figure 9b.
     E.  Preliminary Velocity Traverse.
         Perform a preliminary velocity traverse of the duct.  Take velocity
head UP)  readings at each traverse point, using a standard pi tot tube
(designed  as shown in Figure 5 or  Figure 6).   Calculate the average velocity
                                                                          o
in  the duct, using Equation 2-2  in  the  December 23, 1971 Federal Register.
     F.  Sampling
         Sample  at each  non-adjusted traverse  point for the time interval
specified  in the method  being  used (e.g.,  Method 5).   If two successive
traverse points  have  been  relocated to  a single  "adjusted" traverse point,
sample twice as  long  at  the adjusted point as  at non-adjusted  points,  taking
twice  as many  readings,  but  record the  data  as though two separate  points
had been sampled,  each  for half of the  total  time  interval.  Time-invariant

-------
                                    22
flow is assumed; therefore, the sampling rate at each point shall  be  set
based on the AP reading obtained at that point during the preliminary
velocity traverse.
     G.  Post-Test Velocity Traverse.
         Perform a second  velocity  traverse of the duct, at the end of the
sample run.  Calculate  the average  velocity in the duct (Vg) avg., using
Equation 2-2 of the  December  23, 1971  Federal Register.8  If the value of
(V  ) avg.  is within  +. 10 percent of the value obtained in the preliminary
  5                        /-
traverse,  the assumption of time-invariant flow  is valid, and the results
are acceptable.   If  the difference  between the pre-test and post-test values
of  (V  ) avg. is greater than  + 10 percent, reject the results and repeat the
run, monitoring velocity during sampling,  as shown in Figure 4.

-------
                                    23
                               REFERENCES

     1.   Vollaro, R.  F.   The Effects of the Presence of a Probe Sheath on
Type-S Pitot Tube Accuracy.  Environmental Protection Agency, Emission
Measurement Branch.   Research Triangle Park, N. C.  August 1975.
     2.   Vollaro, R.  F.   Guidelines for Type-S Pitot Tube Calibration.
Environmental Protection Agency, Emission Measurement Branch.  Research
Triangle Park, N.C.   September 1975.
     3.   Martin, Robert M.  Construction Details of Isokinetic Source-
Sampling Equipment.   Environmental Protection Agency, Research Triangle
Park, N.C.  Publication No. APTD-0851, April 1971.
     4.   Perry, Robert H., Cecil H. Chilton, and Sidney D. Kirkpatrick (ed.).
Chemical Engineers'  Handbook, Fourth  Edition.  New York, McGraw-Hill  Book
Company, 1963.
     5.   Fluid Meters, Their Theory and Application.  New York, Published
by the American Society of Mechanical Engineers.  5th Edition,  1959.
     6.   Vollaro, R. F.   Evaluation of Modified Prandtl-Type Pitot Tube.
Interoffice memorandum.   Environmental Protection Agency.  Emission Measurement
Branch.   Research Triangle Park, N.C.  November 28, 1975.
     7.   Shigehara, R. T.  Adjustments in  the  EPA Nomograph  for Different
Pitot Tube  Coefficients and Dry  Molecular  Weights.  Environmental Protection
Agency.   Emission Measurement Branch.  Research Triangle  Park,  N.C.   August 1974,
      8.   Standards of Performance  for New Stationary  Sources.   Federal
Register.   36 (247).  December  23,  1971.

-------
                                   24
                  GUIDELINES FOR SAMPLING IN TAPERED STACKS
                        T. J. Logan and R. T. Shigehara*
 INTRODUCTION
      Tapering of the inside diameter of stacks is occasionally done when
 designing natural  draft stacks, when there are special flow or structural
 considerations,  and for pressure recovery.   These tapers  seldom exceed a
 few degrees.   Although guidelines for the selection of a  sampling site to
 aid in  the extraction of a representative sample  are given in Method 1 of
 the December 23, 1971, Federal  Register,  no mention is made about tapered
 stacks.   The purpose of this paper is to provide  the necessary background
 on  how  to deal with tapered stacks.
 BASIC CONSIDERATIONS
      In order to obtain a representative sample,  the particles must be
 extracted at an  isokinetic flow rate.  The  condition of isokineticity de-
 mands that the particles and gases flow directly  into the sampling nozzle
 and that  the  velocity be accurately  measured.   Therefore, two factors must
 be  considered:   (1) the effect  of the taper on flow conditions within the
 stack and (2)  the  effect of the taper on velocity determination and parti-
 culate matter collection.
 Effect of Taper  on Stack Flow Conditions
      About the only information related to  this area was the work done with
 venturi meters.  The American Society of Mechanical  Engineers research on
             2
 fluid meters   indicates that beyond  a convergent  included angle of 21  degrees
 and a divergent  included angle  of 15 degrees,  gas  separation from the walls
* Emission Measurement Branch, ESED, OAQPS, EPA, RTP, NC,  November  1974

-------
                                   25
is expected to occur.   This is undesirable as  eddies  would  be  formed,
causing particles and  gases to flow in undeterminable directions.
     From a physical  standpoint,  convergent angles of 15  or 21  degrees
would not likely occur in stacks  due to the tremendous increase in  velocity.
If the larger stack diameter D is used, a tapered stack meeting the minimum
2.5 D requirement of Method 1  would cause an increase in  velocity of about
8.6 times at the outlet for a 15-degree included angle and  186 times for
the 21-degree included angle.   Such an increase would require  considerable
additional power and would be impractical and  uneconomical.
     One builder of chimmeys3 related that convergent stacks generally  do
not exceed 0.5 in/ft.   This corresponds to an  included angle of about 4.8
degrees for convergent stacks.  Divergent stacks are  normally  designed  at
about 5 to 15 degrees.
     Based on the above, the 15-degree included angle can be considered the
maximum limit for both convergent and divergent stacks, with the under-
standing that the 15-degree angle will be very unlikely in  convergent stacks.
The purpose for making this statement is to form the  limit  and basis for
evaluating the effect of the taper on the velocity determination and the
particulate matter collection.
Effect of 15-degree Included Angle on Velocity and Particulate Concentration
     Convergent or divergent stacks with an included  angle of  15 degrees
would cause a maximum 7.5-degree angle of attack on the pi tot  tube  and  par-
ticulate sampling probe nozzle.  Data presented by Grove and Smith   show
that a 7.5-degree angle will result in velocity measurements with a type-S

-------
                                     26
pi tot tube being biased  3.5  percent  high.  This higher apparent velocity
also causes particulate  sampling  to  be in error because isokinetic sampling
requires that the sample gas  velocity  be made equal to the stack gas velocity,
which is in error since  it is measured by the misaligned pitot tube.  In ad-
dition to the sampling rate  being over-isokinetic, the misalignment of the
probe nozzle with the stack  gas stream results in a reduction of 0.85 percent
in the effective nozzle  area.
     The magnitude of the effect on  the particulate concentration by being
over-isokinetic and having a  reduced nozzle area is a function of particle size.
For particles of less than 1  micrometer, the concentration will not be af-
fected.  However, with the larger particles of greater than 50 to 75 micro-
meters, the sampled concentration will  be low; a bias of about 4.3 percent will
occur (about 3.4 percent from being  over-isokinetic and 0.86 percent from the
reduced nozzle area).  In a  practical  case, where there is a distribution of
particle sizes, the error will  be considerably less than the 4.3 percent, and
for well-control led sources  where the  majority of the particles are charac-
teristically small  (<2 micrometers), the error will be near zero.
     For pollutant mass  rates,  the error of the higher measured volumetric
flow rates will cancel out the  errors  of the lower measured concentrations,
with the true concentration  being between the maximum limits of +3.5 and -0.8
percent.
RECOMMENDATIONS
     Based on the above  discussion,  the following guidelines, which should not

-------
                                      27
cause maximum errors greater than 4.3 percent in measured  concentration  or
3.5 percent in mass rate determinations,  are recommended (actual  errors  for
small particle sizes will be from 0 to -0.8 percent):
     1.  Consider all stacks with the total included angle of <15 degrees
         as straight stacks.  If this angle is exceeded, consider the taper
         to be a flow disturbance and modify the stack with a straight sec-
         tion of at least 2.5 D.
     2.  Use the maximum diameter at point of upstream or downstream dis-
         turbance and Method 1 for determining the sampling point location
         and number of sampling points.

REFERENCES
1.  Standards of Performance for New Stationary Sources. Federal  Register.
    Vol. 36, No. 247, December 23, 1971.
2.  Fluid Meters, Their Theory and Application, Report of ASME Research
    Committee on Fluid Meters.   (5th Ed.).  American Society of Mechanical
    Engineers. New York. 1959.
3.  Personal Communication with  Richard Lohr, Vice-president, International
    Chimney Corp., Buffalo, N. Y.
4.  Grove, J. D. and W.  S.  Smith. Pi tot Tube Errors Due to Misalignment and
    Nonstreamlined Flow. Stack Sampling News. 1 (5):7-ll, November 1973.

-------
                                     28
                CONSIDERATIONS FOR EVALUATINfi EQUIVALENT STACK
                         SAMPLING TRAIN METERING SYSTEMS
                                             *
                               R.  T.  Shigehara
 Introduction
     The  basic  purpose of sampling train equipment  is to collect a repre-
 sentative sample from a point (small  area)  within a stack cross-section
 or, when  conducting a sample traverse,  to collect a series of such sam-
 ples.   To accomplish this, the sampling train must  (1) maintain either
 isokinetic or proportional sampling rate, depending on whether particulate
 or  gaseous pollutants are being sampled, (2)  efficiently collect reprodu-
 cible  samples of the pollutant at known levels,  and (3) accurately mea-
 sure the  sample gas volume.  Thus, conventional  sampling trains incor-
 porate some means of gas metering to regulate the  sampling flow rate and
 to  measure the sample gas volume.
      "Method 5 - Determination of Particulate Emissions from Stationary
 Sources,  Section 2.1.6" specifies the above requirements.  It states,
 "Metering system - Vacuum gauge, leak-free pump, thermometers capable of
 measuring temperature to within 5° F, dry gas meter with 2% accuracy, and
 related equipment, or equivalent, as required to maintain  an  isokinetic
 sampling  rate and  to determine sample volume."
      There are many different workable  metering techniques or systems.
 Individual stack samplers  and control agencies usually have  their own
 ideas  as  to  which  mode  is  the best.  The purpose of this paper is to pro-
 pose  criteria  to evaluate the different stack sampling train metering
 techniques or  systems.
* Emission Measurement Branch, ESED, OAHPS, EPA,  RTP,  NC, September 1974

-------
                                 29
Criteria for Isokinetic Sampling - EPA Particulate  Test  Train
     The EPA particulate test train1'2 will  be used as a baseline
reference for this discussion on the development of criteria  for
evaluating the different stack sampling train metering techniques  or
systems.  The EPA train uses the pi tot tube-orifice meter-dry gas
meter system for setting isokinetic rates and for determining sample
gas volume.  In this system, the pi tot tube is attached  to the probe
so that the gas velocity at each of the sampling points  can be constantly
monitored.  The observed pi tot  tube manometer reading is related to the
orifice meter manometer reading by an equation such that the flow rate
through the sampling train can  be  adjusted  to isokinetic conditions.
To perform these  calculations,  the EPA train utilizes a  nomograph, which
requires  as little  as  5 to 10 seconds  to  determine and  adjust the sampling
rate  after a  new  velocity reading or  a change  in stack  flow has been
observed.  The  nomograph  is  only  a type  of  aid.  Graphical techniques  or
electronic calculators can  also be used  to  yield the same  result.   The
dry gas meter is  used  to  measure  the  sample gas volume  and to measure
 the gas sampling  flow  rate  independently from the  orifice  meter.
      For particulates, it is the "condition" of isokineticity that ensures
 the extraction of a representative point sample, not the "means"  by which
 the desired sampling rate is achieved.  However, it is  insufficient to
 simply state that all  metering systems that have at some time demonstrated-
 capability of obtaining isokinetic conditions are equivalent.  All techniques,
 null balance, pi tot tube-rate  meter,  pi tot tube-volume meter-timer, and
 others,  rely on  the knowledge, experience, and conscientiousness of the

-------
                                  30
operator and require error-free equipment performance during  sampling.
Some assurance that isokinetic conditions were maintained throughout  the
actual sampling run is needed.
     The EPA pi tot tube-orifice meter-dry gas meter system provides
sufficient proof of isokineticity:
     1.  All the components can be calibrated against a standard.  The
dry gas meter and orifice meter can be calibrated against a wet test
meter  (secondary standard) or a spirometer  (primary standard).   With
periodical calibrations, the dry  gas meter  can maintain an accuracy of  about
1% in  volume measurement.3'4  The type-S pi tot tube can be calibrated
against a standard type  pi tot tube which generally has a calibration
coefficient between 0.98 and l.OO.5  If one would send a standard pitot
tube  to the National  Bureau of Standards, a certified calibration for
velocity ranges from  6  to 100 fps or 6 to 155 fps can be obtained.
      2.  The pitot tube  attached  to the probe allows the velocity to  be
monitored and  isokinetic sampling to be maintained throughout the entire
test  run.   In  this manner  the  sampling can  be conducted under normal
everyday process  conditions;  it  is  not limited  to steady-state conditions
of  gas velocity.
      Some  limitations of the pitot  tube  for monitoring and measuring
velocity  should be recognized:    (1)  It  still  relies  on the operator  to
properly  orient the pitot tube into the  direction of flow, correctly set
up  the manometer, and accurately read  the velocity  pressure  head; (2) the
pitot tube also has a lower velocity limit, usually reported at about 10 fps
This limitation is caused mainly by the difficulty  in reading  the manometer

-------
                                 31
scale; (3) in addition,  the pitot tube is  dependent  on  the density of  the
flowing gas stream.   Thus, "great"  changes in  temperature, pressure, and
gas composition (particularly moisture) may cause difficulty in  determin-
ing the gas velocity and in setting isokinetic rates.   Shigehara et al.
show a method of analysis to determine how much variation  in the para-
meters can occur before  "significant" errors result.
     One method of reducing the problem of setting isokinetic rates when
the gas density or composition of the stack effluent changes "significantly"
with time is to place the orifice meter immediately after  the filter,  which
is heated to stack temperature and does not allow moisture to condense.
This method eliminates the problem of changes in composition, but adds the
variable of total pressure at the orifice.  However, this  does not solve
the problem of determining velocity.
     The problems of low velocities and great changes in gas density have
not yet been adequately solved.  Until better means are specified, we
can only attempt to increase  the sensitivity of  the manometer for low
velocities and for great changes in gas density, to evaluate the source
conditions and use techniques that, in our opinion, would  provide adequate
results.
      3.   The  pitot tube-dry  gas  meter combination allows an overall
average and  individual  average point  deviations  from isokinetic condi-
tions to  be  calculated  for each  test  run.   This  is  helpful  in that it
 permits acceptance or rejection  of a  run  based  on per  cent  of isokineticity
 actually  obtained.   The Federal  Register   allows an overall  average
 deviation of 10% from isokinetic.   Smith  et al.   have  shown a calculation

-------
                                   32
 method by which a deviation of 20% can be tolerated with the assurance
 that the sample concentration will be within 10% of the true concentration.
      4.  The  orifice meter-dry gas meter combination provides  a  cross-
 check of flow rate and sample gas volume.

      However, although both the dry gas meter and the orifice meter  can
 yield accurate results, there is no means for checking against  improper
 use or malfunctions under the actual  operating conditions  of the  sampling
 train if the components are used separately.

      Summarizing,  it is the condition of isokineticity that  produces a
 representative point sample.  Any means that provides this condition
 could theoretically be considered equivalent.   However,  as improper
 uses  or errors do  occur,  "sufficient  proof"  can be  defined as:
      1.   All  components be calibrated against  a standard.
      2.   Velocity  be monitored constantly and  simultaneously with sampling.
      3.   A  check of isokineticity actually obtained be provided.
 Null  Balance  Probe System
      This system is  deceptively  simple  in  principle.  Also called static
balance,  zero  pressure, and  isokinetic  probes,  the  pressure null balance
probe is a  nozzle  specifically designed to measure  the static pressure  of
the stack gases flowing around and within  the probe nozzle.  When  both  static
pressures are equal,  isokinetic conditions are said to exist.  Cooper8
summarized as follows:

-------
                                    33
     "However, numerous problems have  been observed in attempting to
accurately maintain true isokinetic sampling conditions because the exis-
tence of equal pressures at outer and  inner probe walls does  not neces-
sarily mean that equal  velocities exist at both points.  Differences in
frictional flow losses  between inner and outer surfaces caused by turbu-
lence and surface nonuniformities, progressive coating and possible plug-
ging of the inner static tap by particles, and possible differences of
static tap location may all produce these conditions.   Parker ("Some
Factors Governing the Design of Probes for Sampling in Particle- and
Drop-Laden Streams," Atmospheric Environment 2;477-490, September 1968) found
that null balance systems had limited  usage for large  probes  greater than
3/4 inch diameter.  Toynbee and Parkes ("Isokinetic Sampler for Dust Laden
Gases," International Journal of Air and Water Pollution 6_:113-120, 1962)
postulated that by a slight expansion  of the rear section of  the probe the
inner frictional losses could be reduced inside the nozzle, and the
system could be used over the velocity range from 600  to 2500 fpm.  However,
subsequent comments by Nonhebel in the same issue stated that the plug-
ging problems associated with the inner static taps could not be overcome.
Work by Dennis ("Isokinetic Sampling Probes," Industrial and  Engineering
Chemistry 49_:294-302, 1957) and Hemeon and Haines ("The Magnitude of
Errors in Stack Dust Sampling," Air Repair 4_:159-164,  November 1954)
indicated that it was not always possible to assure isokinetic sampling
conditions, and found the errors at different velocities for two nozzle
sizes when departing from nozzle conditions."

-------
                                  34
     This excerpt illustrates that even with careful  calibration  for
the specific source and conditions, one cannot be positive that iso-
kinetic conditions existed throughout the entire run.  Although the
feasibility of such a system has been demonstrated under controlled
conditions, it suffers from the lack of proof of isokineticity for the
actual operating conditions as provided by the EPA metering system.  In
order to provide sufficient proof of isokineticity, the null balance
probe system must incorporate a pi tot tube and a dry gas meter.  This
is what Wilson and  Falgout9 did to  show that their null probe design
was workable.
Dry Gas Meter as a  Rate  Measurement Device
      The  volume meter (dry gas  meter),  in addition to  measuring the  total
sample  volume, could  serve as a rate meter for  setting isokinetic  rate  by
timing  the needle  travel.  However, since the  needle travel must be
observed for one  or more whole revolutions to  obtain a reasonably  accurate
rate value, the  rate is only an average,  and changes are  possibly  delayed
one or more minutes past the time they occur.   Thus, its  application is
 limited to sources where velocity is "fairly"  constant.  There is  also
 the disadvantage of not  having a cross check of volume and rate  under
 actual operating conditions as with the  orifice meter-dry gas meter
 combination.

 Proportional Samp!ing
      The  same criteria  apply to  proportional sampling  as to
 isokinetic sampling.   It is  the  condition of proportionality that counts,
 not  the  means by which  proportional  sampling  is achieved.

-------
                                  35
     Like the EPA participate sampling train, the same  pi tot  tube-
orifice meter-dry gas meter system can be used to regulate  and  check
proportionality.  But because of the lower sampling rates used  for
gaseous trains, a rotameter is normally used instead of an  orifice
meter.
Total Gas Sample Volume
     The usual  means for measuring the qas sample volume are  dry na§
or rate meters such as orifices and rotameters.  Cyclones,  venturi  meters,
evacuated containers, critical orifices, and mass flow  rates  are  also
used.  Whatever the means, it is the total gas volume that  is desired.
Integrating volume meters such as the dry gas meters, when  sized  properly,
readily provide the desired result.    As mentioned previously, the dry
gas meter can maintain an accuracy of about 1% in volume measurement when
                                                               3  4
calibrated periodically against a wet test meter or spirometer. '
     Rate meters can also be used to measure the sample gas volume.
However, they measure instantaneous flow, which is subject  to density
changes of the gas stream.  Therefore, other variables  such as  time,
temperature, pressure, and pressure drop must be carefully  recorded
during the test run so that an integrated total volume  can  be calcula-
ted or obtained graphically.  The same is true with dry gas meters  if
they are placed before the pump,  because  pressure could vary considerably,
at times, during the test run as  particulate matter builds  up on the
filter material.

-------
                                   36
     The EPA sampling train  places the dry gas meter and orifice meter
behind the gas pump with the orifice meter open to the atmosphere.   There
are several practical advantages with this placement, which requires that
the pump be leak proof.  The advantages are:
     1.  The dry gas meter is  subjected to a fairly constant pressure--
the only variation coming from the orifice meter pressure drop, which
is no more than 10 in. of water.  The orifice meter is at a relatively
constant atmospheric pressure;  therefore, there is no need to record
or to observe for all practical purposes pressure and meter readings
extra carefully.
     2.  The dry gas meter need not be calibrated under the expected
range of negative pressures  that would occur if it were placed before the
pump to compensate for the leakage around the meter diaphragm valves,
particularly under high vacuums.
     3.  It is not necessary to have special gas meters that can
withstand the high vacuums.
Condensers
     Condensers are generally  an integral part of a metering system.  Their
main purpose is to prevent moisture from condensing within the pump and gas
metering devices.  They also serve as a means for the determination of the
average moisture content over  the sampling duration.
     The EPA test method gives a clear procedure for determining moisture
when the gas stream does not contain water droplets.  (If liquid droplets
are present, the gas stream  is assumed to be saturated).  The probe and
filter holder are heated to  a  minimum of 225°F so that moisture contained

-------
                                37
in the sample will  remain in gaseous form until  the  gas  has  passed  the
filter.  Following the filter is a series of four Greenburg-Smith impingers
which are immersed in an ice bath.  The first two impingers  each  contain
100 ml of water.  This chilled water acts to condense and  trap the  water
vapor contained in the hot gases coming from the filter  holder.   The third
impinger is empty and acts as a trap to collect any  entrained water which
might be carried over from the first two impingers.   Finally, the fourth
impinger contains approximately 200 grams of silica  gel.  The silica gel
adsorbs most of the moisture which remains in the gas stream; for a 1-hour
sampling run, less than 3% passes through if the temperature at the third
impinger is kept below 70°F and less than 15 in. Hg  vacuum.     The water
collected in the first three impingers is easily measured  volumetrically,
and the weight change in the silica gel gives the amount of moisture
collected there.  The amount of moisture in the gas  stream thus measured,
and the sample gas volume as measured by the dry gas meter are then used
to determine the moisture content.
      The choice of equipment is not important as long as the moisture
collected and  leaving the condenser and  gas  sample volume can be measured
accurately.   For long sampling  runs  (3  to 4  hours), condensation coils
may  be better  than or as effective  as  the EPA method.   Temperature  and
pressure must  be measured at the  exit  of the condenser  to account  for the
moisture still  remaining  in the gas  stream.  However, because at 10 in.
Hg.  vacuum and 70°F, the  amount of moisture at  saturation conditions is
about 3.7% by volume,  the  silica  gel  should still be  used to protect the
 pump and metering devices.

-------
                                38
Summary
     Flow rate regulation and sample volume systems have been discussed.
The basic purpose of these systems is to ensure that a representative
point sample  is collected and that the sample gas volume is accurately
measured.  However, representativeness is not a direct measurement. Thus,
individual measurements  that ensure  representativeness must be compared
against  a  standard.   In  the  absence  of any  standard, the question of which
result is  right when  two sampling trains yield different values can never
be  answered.   When a  standard is not available dnd if  an evaluation is
desired, design  and/or performance criteria which have been  scientifically
or  arbitrarily derived must be used.

      Since it is the condition of isokineticity or proportionality that
 is important in the extraction of representative point samples, any
 technique that provides  these conditions can be used.  However, since they
 are a vital  part  of obtaining representative samples, checks under actual
 operating conditions  must  be provided.  In this  regard, the pltot tube-
 rate meter-volume meter system offers clear advantages.

-------
                                   39
                               References
 1.   "Standards  of  Performance  for  New  Stationary Sources," Federal
     Register, Thursday,  December 23, 1971.

 Z.   W.  S.  Smith, et  al., "Stack Gas Sampling  Improved and Simplified
     with  New Equipment," Presented at  the 60th Annual Meeting of APCA,
     Cleveland,  Ohio, June 11-16, 1967.

 3.   American Society of  Heating, Refrigerating and Air Conditioning
     Engineers Handbook of Fundamentals, ASHRAE, Inc., N.Y., N.Y., 1967.

 4.    "Source Sampling,"  Institute  for  Air Pollution Training, Office of
      Manpower Development,  USDHEW, NAPCA, P.  0, Box 12055, RTP, N.C. 27711.

 5.   J.  H.  Perry, C.  H. Chilton, &  S. D, Kirkpatrick, Chemical Engineering
     Handbook. 4th  edition,  McGraw  Hill, N.Y., 1969.

 6.   R.  T.  Shigehara, W.  F.  Todd, and W. S. Smith, "Significance of Errors
     in  Stack Sampling Measurements," Presented at the Annual Meeting of
     APCA,  St. Louis, Missouri, June 14-19, 1970.

 7.   W.  S.  Smith, R.  T. Shigehara,  and  W. F. Todd, "A Method of Interpreting
     Stack Sampling Data," Presented at the Annual Meeting of APCA, St. Louis,
     Missouri, June 14-19, 1970.

 8.   H.B.H. Cooper, Jr. and  A.  T. Rossano, Jr., Source Testing for Air
     Pollution Control, Environmental Science  Services, 24 Danbury Rd.,
     Wilton, Conn.  06897

 9.   K.  D.  Wilson and D.A. Falgout, "A  New Approach to Isokinetic Null
     Probe Design," presented at the 65th annual meeting of APCA,
     Miami  Beach, Florida, June 18-22,  1972.

10.   "Standard Method for Sampling  Stacks for  Particulate Matter," ASTM
     Designation D  2928-71,  American Society for Testing and Materials,
     1916  Race St., Philadelphia, Pa.,  19103,  1971.

11.   W.  L.  Johnson, Emission Measurement Branch, Emission Standards and
     Engineering Division, Office of Air Quality Planning and Standards,
     Environmental  Protection Agency, Durham,  N. C. (unpublished data),
     1973.

-------
                                                                              an
Published in "Stack Sampling Hews,"  March 1975, Vol. 2, No. 9.
         EVALUATION OF METERING SYSTEMS FOR GAS-SAMPLING TRAINS
                    M. A. Wortman & R. T. Shigehara
                 U. S. Environmental Protection Agency
INTRODUCTION
     In the December 23, 1971,  Federal Register,  several types of gas-
sampling trains are specified.   Each uses basically the same types of
components in its metering system,  i.e. flow control valve, diaphragm pump,
rotameter, and dry gas meter, but differs in the sequence in which they are
arranged.  The different sequences  are summarized as follows:
     1.  Method 3 (Integrated Gas Sampling Train):  Flow control valve,
         diaphragm pump, and rotameter.  A flexible bag follows the
         rotameter in this train.
     2.  Method 4 (Moisture Sampling Train):  Flow control valve, diaphragm
         pump, dry gas meter, and rotameter.
     3.  Method 6 (S02 Sampling Train):  Diaphragm pump, flow control
         valve, rotameter, and  dry  gas meter.
                         2
     A recent publication  reported an adverse effect on the calibration of
dry gas meters in particulate sampling trains utilizing diaphragm pumps
with bypass valve systems.  Although the gaseous sampling train metering
systems do not use a pump bypass valve, questions were raised on whether or
not this same effect would also be  present in the smaller gas-sampling trains,
Thus, tests were conducted to determine the effect, if any, of the position
of the control valve in relation to the pump and metering devices on the
calibration of the dry gas meter.
     During the course of the test  program, certain problems with the leak
check procedure and the diaphragm pump were encountered.  The purpose of
this paper is to report these findings and the results of this test.

-------
                                   41

PROCEDURE
Test Equipment
     The test train components used were the same as those specified by
Method 6, as published in the December 23, 1971, Federal  Register.   A wet
                   3
test meter (0.05 ft /rev.) was connected to the inlet of  the metering
system.  A drying tube was inserted immediately after the wet test  meter
to protect the rotameter, dry gas meter, and pump from moisture condensation.
Schematics of the two sampling train arrangements used to determine the
effect of valve position are shown in Figure 1.
Test Procedure
     The test was conducted in the following manner:
     1.  A leak check was first conducted.  This leak check consisted of
         plugging the inlet to the metering system (before the drying tube),
         leaving the control  valve fully open, turning on the pump, and
         noting the travel of the dry gas meter dial.   If any leaks were
         indicated, they were corrected before any test was conducted.
     2.  Using the rotameter as a flow rate indicator, the following infor-
         mation was gathered:  rotameter reading, wet test meter reading
         and temperature, dry gas meter readings and temperature, barometric
         pressure, and running time.  From the raw data,  two values were
         computed:  (1)  the calibration factor (F), which is the ratio  of dry
         gas meter volume to wet test meter volume, and (2) average standard
         flow rate (Q) obtained by dividing the wet test  meter volume,  after
         being corrected for moisture content, by the running time.

-------
     o
Thermometers
   Wet  test
   meter
                                                     Rotameter
                     i
                       Silica gel  tube
                                                        Valve
                                     Diaphragm
                                     pump
Dry  gas
meter
     O

                                                              P
                                                              O
                                                                                              ro
Figure   1

Initial  test sampling  train arrangements

-------
                                   43

 TEST RESULTS
      Four  different  sampling trains were tested, each with the valve before
 and  after  the  pump.   During one of the tests, with the valve placed after
 the  pump and closed  completely, movement of the wet test meter dial was
 noted.  A  leak  check of the pump with a mercury manometer revealed a leak,
 which was  not  detected by the normal leak check procedure.  (This leak was
 occurring  where the  diaphragm was connected by two screws to the connecting
 rod.)  Plots of the  calibration factor, F, versus the flow rate, Q, for
 pumps with and without leaks are shown in Figure 2.
     During these tests, it was also noted that with the valve placed before
 the  pump,  the  rotameter readings were greatly affected due to the pulsating
 motion of  the diaphragm.  But there was less of an effect on the calibration
 factor over a wider  flow range with this arrangement than with the valve
 placed after the pump.  The calibration factor was also less affected by
 leaks, when present^with this arrangement.   Since these were desirable
 characteristics, steps were taken to reduce the effect of the pulsations.
This was easily accomplished by placing a surge tank between the pump and
the rotameter or by using the dry gas meter as a surge tank, i.e., inter-
changing the position of the dry gas meter and rotameter.   The results  are
shown in Fiqure 3.   Using the dry gas meter as the surge tank, however,
caused the control  response of the rotameter to be sluggish.   Therefore,
the surge tank placed before the rotameter was selected, and the final  train
shown in Figure 4 was used for subsequent tests.
     After ensuring that all  systems were leak free,  this  time using  the
manometer or the wet test meter procedure for the leak check,  the tests
were rerun.  The results are shown in Figure 5.

-------
                                                   44
1 ?fl


1.10


1.00




0.90

0.80
1 1
A A TRAIN NO. 10 WITH
—
. A A
A A A A ^
	 0 0
ft O
O °
O
o
	 0
o . ,
	 1 	 1 — 1


LEAK








—


    1.30
    1.20
   1.10
   1.00
o
£ 0.90
o

-------
    12
   10
LLJ    c

I
O
c!    4
                                                                O DRY GAS METER AS SURGE TANK


                                                                * SURGE TANK
                                  4              6             8

                                        ROTAMETER SETTING

                                  Figures.  Rotameter calibration.
10
12
                                   en

-------
          Thermometers
                                         Valve
                     Silica gel  tube
                                                                                                CTl
      Wet test
      meter
Diaphragm
pump
Dry  gas
meter
Figure 4

Final  test  sampling  train  arrangement

-------
      1.00




      0.90




      0.80
                                                                                  TRAIN NO. 10

                                                                                     o
      1.10




      1.00
K    0.90
o
|    i-io
t-
O£

3    1.00

o


      0.90
                                                                                  TRAIN NO. 7
_   o   o
o    °        000°            °
1 1 1
-o o o ° 0 ° o ° ooo
1 1 1
TRAIN NO. 6
      1.00
      0.90
      0.80
                                              o     O
                                                                                            TRAIN NO. 1
                          Z
                                                                               10
                                                                      12
                                                  FLOW RATE (Q), dscfh


                          Figure 5.  Calibration factor versus flow rate for Figure 4trains.

-------
                                   48
SUMMARY AND RECOMMENDATIONS
     The results of these tests showed that a constant dry  gas  meter cali-
bration factor could be obtained whether the control  valve  was  placed
before or after the pump.  However, the placement of the valve  before  the
pump provided a constant calibration factor over a wider flow range and
was not as greatly affected by leakages from within the pump.  It is
recommended that the metering system shown in Figure 4 be used  for gaseous
sampling.
     The present leak check procedure was found to be inadequate.  It  is
suggested that leak checks be conducted by either of the following two
procedures:   (1) connect a wet test meter at the inlet of the sampling train,
turn on the pump,  pinch off the  line after the pump, and note wet test meter
dial (suggested for laboratory),  or  (2) connect a vacuum gauge (mercury
manometer,  bourdon gauge, or  similar) at the inlet, turn on  the pump,  pinch
off the line  after the  pump,  turn off the pump after maximum vacuum is
reached,  and  note  gauge reading  (suggested for field use).   Any movement
of the wet  test meter dial  or vacuum gauge reading denotes a leak and
must be  corrected.

REFERENCES
1.   Standards of  Performance for New Stationary  Sources.   Federal Register.
     Vol.  36,  No.  247.  December 23, 1971.   p.  24882-24895.
2.   Smith,  W. S.   When Your Valves Float.   Stack Sampling  News.   Vol. 7,
     No.  1.   January, 1974.   p. 5-8.

-------
                  AN EVALUATION OF THE CURRENT EPA METHOD 5
                  FILTRATION TEMPERATURE - CONTROL PROCEDURE
                                                 a
                               Robert F. Vollaro

Introduction
     Method 5, promulgated in the December 23, 1971 Federal Register , re-
quires the use of probe and filter holder heating systems during isokinetic
sampling.   Prior to sampling, these heating systems are adjusted as follows:
(1)  the probe heater is set to provide a gas temperature of about 250°F* at
the probe outlet; probe heater settings are obtained from Figure 21 of the
                                2
sampling train operations manual , APTD-0576 (Figure 2 of this report); (2)
the sample box thermostat is set to provide a temperature of approximately
250°F* around the filter holder.  Although it is not explicitly stated in
Method 5,  one of the primary reasons for making these temperature adjustments
is so that filtration will take place at 250°F +_ 25°F*.
     Recently, however, some observers have expressed concern over the ade-
quacy of the above filtration temperature control procedure, particularly
whether probe heater setting estimates made from the APTD-0576 reference curves
will actually provide probe outlet temperatures around 250°F under field test
conditions.  Among the reasons given are:  (1)  the curves give no specific
probe heater setting guidelines for sources with temperatures above 250°F or
below 80°F; (2)  the temperature of the gases surrounding a sample probe during
an actual  traverse will seldom be 80°F, which is the temperature base from
which the curves are derived; and (3)  the curves are strictly applicable only
to gas streams of low moisture content.  These comments fail to note that the
*Unless otherwise specified by a particular regulation.
a  Emission Measurement Branch,  ESED, OAQPS, EPA, RTP, NC,  July 1975

-------
                                    50
reference curves were not originally  intended to provide exact filtration
temperature control; their original purpose was to furnish approximate
guidelines by which moisture could  be prevented from condensing ahead of the
impingers.  There is, nevertheless, question as to whether probe outlet tem-
peratures around 250°F can be  generated with confidence, even with the sample
box set at 250°F.
     In light of the above question,  experiments were conducted, under a num-
ber of simulated field test conditions, to evaluate the present means of fil-
tration temperature control.   This  paper  reports the results of these experi-
ments.
Experimental Set-up
     The  Method 5 sampling train  configuration  used in  the experiments is
shown in  Figure 1.  The  components  of the train met the design specifications
outlined  in the source  sampling equipment construction manual, APTD-0581 ,
except for the modifications  necessary to facilitate temperature monitoring
at  the probe  inlet, probe outlet, and inside the back half of  the filter holder.
Chromel-alumel  thermocouples,  insulated from the metal  parts of the  train,
were  used to  monitor  temperature  in these experiments.
Filtration Temperature  vs.  Probe  Outlet Temperature
      Preliminary  experiments were conducted to establish  a relationship  between
probe outlet  temperature and filtration temperature, at constant  sample  box
setting.   At  each of three different box settings* (220,  240,  and 260°F),  the

*  Note  that  the  term "sample box setting," as used  in  this  report,  refers to
the average  temperature inside the box during a sample  run.   During  sampling,
the box  temperature changed continually with time, rising and  falling in 5-
minute cycles between thermostatically controlled upper and  lower limits.

-------
 PROBE INLET
THERMOCOUPLE
                                     SAMPLE BOX
                                    THERMOCOUPLE
                       PROBE OUTLET
                       THERMOCOUPLE
                                                       FILTRATION TEMPERATURE
                                                           THERMOCOUPLE
                        HEATED SAMPLE
                           PROBE
      THERMOMETER
                                            HEATED
                                            SAMPLE  f
                                             BOX
f


»
i:
•- N.
* V t
H
                             THERMOMETERS (   )

                                        r*T
                        CALIBRATED
                          ORIFICE
                                                    '100 ml OF WATER J  ICE
                                                    I                BATH
                                                                   CONTROL
SILICA GEL
  VACUUM
  GAUGE
                    MANOMETER  '-
                                      Figurel. Sampling train configuration.

-------
                                            52
    400



u.
o


K  300




cc
m


|  200
    100
ca
o
K
              3-ft PROBE (5-minWARMUP)
                 INLET. 250 °F


        INLET. 150 °F
        INLET AMBIENT, 80 °F
                                      4-ft PROBE (10-minWARMUP)
                                 —       INLET, 250 °FV



                                 INLET. 150 °F
                                                              INLET AMBIENT. 80 °F
20       40       60       80    0       20




                 POWERSTAT SETTING, percent
                                                               40      60
                                                                                80
    400
             5-ft PROBE (10-min WARMUP)
                                      6-ft PROBE (15-min WARMUP)
                                                	       INLET. 250 °F>



                                                INLET, 150 °F
               20       40
                 60       80    0       20       40      60       80



                  POWERSTAT SETTING, percent
                     oc
                     in
                         400
                         300
                         200
                     o  100
                                    I        I         I

                                  7-ft PROBE (15-min WARMUP)
              —      INLET. 250 °F



              INLET, 150 °F
                                                INLET AMBIENT,
                            0       20       40       60


                                 POWERSTAT SETTING, percent
                                              80
                          Figure 2.Probe temperatures.

-------
                                53

temperature of the gas at the probe outlet was  varied  from  100  to 450°F while
holding the sampling rate constant at 0.75 cfm.   The results of these  exper-
iments are presented graphically in Figure 3.
     Figure 3 shows that at constant sample box temperature, filtration temper-
ature is a linear function of probe outlet temperature,  requiring a  2.3°F
change in probe outlet temperature to effect a  1°F change in filtration temper-
ature.*  Figure 3 also shows that, with the sample box set  at  its customary
250°F, it is necessary for the probe outlet temperature  to  be  maintained be-
tween 230° and 350°F, in order for the filtration to take place at  250 +_ 25°F.
Probe Outlet Temperatures
     Further experiments were conducted, under a number of  simulated field
test conditions, to determine whether heater setting estimates made from  the
APTD-0576 curves (See Figure 2) would provide the necessary probe outlet  temper-
atures to keep the filtration temperature between 225  and  275°F. Temperature
was monitored during each run at the probe inlet, at  the probe outlet, inside
the sample box, and inside the filter holder, just behind  the  glass frit (See
Figure 1).  A constant sample rate of 0.75 cfm was maintained  for all  experi-
ments.  The following test cases were considered:
     Test Case I—Possible Underheating.  In this experiment,  cold  air at 37°F
was drawn through a 3-foot sample probe.  The sample box temperature was set
at 260°F, and the temperature of the gases surrounding the probe was 37°F.  In
the absence of specific  guidelines from APTD-0576 for probe inlet temperatures
below  80°F, the probe heater was  set according to the "closest available" probe

*This  value will, of course,  be a  function of  sample box design and the
path  length that  the  gas must go  through  (e.g.,  if a cyclone is used); however,
a  similar  relationship should exist  for different configurations.

-------
                                                  54
    500
    450
    400
u.   350
    300
    250
2

a
UJ
fe
    200
    150
    100
     50
                                                                    DESIRED OPERATING LIMITS
                                                         • 220°F DATA


                                                         A 240°F DATA


                                                         • 260°F DATA
               50      100      150      200      250     300      350     400


                               STEADY-STATE PROBE OUTLET TEMPERATURE, °F
                                                                                  450      500
                   Figure 3.  Filtration versus probe outlet temperatures (dry air).

-------
                                  55
inlet temperature curve,  namely the 80°F curve.   Case  I was designed  to  simulate
sampling from an ambient  source with a  short probe  on  a cold day.   Its purpose
was to determine if the probe heater was capable of heating cold  sample  gases,
having only a short residence time in a probe set in cold  surroundings,  to  an
acceptable probe outlet temperture.
     The results of this  experiment are presented in the Appendix (See Table I).
The data show that after  a few minutes, the filter  temperature had risen above
225°F; it continued to climb slightly thereafter, reaching a steady-state value
of about 235°F.  These results indicate that, even  when a  cold gas stream
(T«80°F) is sampled with a short probe set in cold surroundings, setting the
probe heater by the appropriate 80°F inlet curve of APTD-0576  is  satisfactory.
Very little reduction in heater performance occurs, and a  steady-state value
of filter temperature safely within the range 250 + 25°F  is rapidly established.
     Test Case II--Possible Overcooling.  Hot sample air  at 475°F was drawn
through an 8-foot probe set in 80°F surroundings; the  sample box  thermostat was
set at 240°F.  In the absence of APTD-0576 guidelines  for sources hotter than
250°F, the probe heater powerstat was arbitrarily set  at  25 percent.  Case II
was designed to simulate the testing of a very hot source (T>300°F) with a
long sample probe.  More specifically, Case  II represents the outset of  the
sample traverse, when points close  to the near stack wall  are tested (i.e., when
a  good part of the probe is outside the stack), and overcooling of the sample
gas can occur before  it enters  the  filter box.
     After a few minutes of the Case II sample run  (See Table II in the Appendix),
it was noted that  the sample gases  were cooling  from 475°F at the probe inlet

-------
                                  56

to 175°F at the outlet.  During this same time span, the filtration  temperature
reached only 190°F.  For the remainder of the test, the probe heater setting
was gradually increased, at 10-minute intervals, until filtration temperatures
consistently above 225°F were obtained.  When 225°F was reached, the powerstat
setting was at 75 percent of maximum.  Thus, the Case II data indicate the
importance of proper probe heater calibration if the desired level of probe
outlet temperature is  to be achieved at  the outset of the traverse of a very
hot stack; random guessing at the powerstat setting to be used will  not suffice.
     Test Case  III—Possible Overheating.  A 20-inch diameter incinerator
duct in which hot  (520°F) combustion gases were  flowing was  sampled with a
3-foot probe.   The probe was  inserted  as far as  it would go  into  the duct,
leaving about 16  inches of  it  exposed  to the ambient  (40°F)  air.  The sample
box thermostat  was  set at255°F.  Again,  in  the  absence  of an APTD-0576 guide-
line,  the  probe heater was  arbitrarily set  at  20 percent.   Case  III was designed
to simulate  that stage of the sample traverse  of a very hot (T>300°F) stack
when  points  close to the far wall  are tested and a good part of the probe  is
 inside the stack, surrounded by hot gases.   The purpose of  this test was  to
 check for possible overheating.
      The Case III data (See Table III in Appendix) show that although the temper-
 ature was leveling out, overheating of  the filter occurred  after 16 minutes  of
 sampling.  After 17 minutes, the probe  heater was shut off to try and bring  the
 filtration temperature back below 275°F.  The filtration temperature did drop
 to 274°F; however, had the ambient temperature been higher  than 40°F or the

-------
                                  57
stack gas temperature higher than 520°F,  overheating would most  likely  have
continued, and to achieve lower temperatures an adjustment in  the  sample  box
temperature would have become necessary.
     To determine the severity of filter overheating,  had an arbitrary  powerstat
setting higher than 20 percent been chosen, Case III was repeated.  This  time,
the probe heater setting was gradually increased, at  12-minute intervals, from
25 percent to 80 percent of maximum.  These data are  shown  in  Table IV  (See
Appendix).  Filtration temperatures well  in excess of 300°F  occurred at the
higher powerstat settings.
     Test Case IV—Effect of Moisture.  During Test Case III,  when incinerator
gases were sampled,  the sample box was set at 255°F.   However, a plot of fil-
tration  temperature  versus probe outlet temperature (See Figure 4) produced a
data line well above the 225°F region of Figure 3.  It was assumed that the
high moisture content of the combustion gases caused the difference.  To check
this assumption, four test  runs were  performed,  in which moist air  (estimated
at 5 to  10 percent)  at  different  temperatures  (228, 270, 293, and 468°F) was
sampled.  The sample box was maintained at about 250°F.
     The results of  these  tests  are plotted in  Figure 4.  They confirm that
moisture in  the  sample  stream  can alter  the relationship between  the probe
outlet and filtration temperatures.
 Conclusions
      An evaluation of the present means  of controlling  filtration temperature
 in the EPA Method 5 train has demonstrated that:

-------
                                              58
   500
   450
   400
S-  350
    300
S 250
u  200
§  150
    100
     50
                      CASE III DATA; SAMPLE BOX@25S°F
                                                                     1	
                                                              DESIRED OPERATING
                                                                   LIMITS     ~~
                                                      	i	
 SAMPLE BOX @ 260°F
.SAMPLE BOX @ 240°F
 SAMPLE BOX @220°F
               50      100      150      200     250     300     350     400     450      500

                           STEADY-STATE PROBE OUTLET TEMPERATURE, °F
                       Figure 4.  Filtration versus probe outlet temperatures.

-------
                                  59
     1.  At constant sample box setting,  filtration  temperature  is  a  linear
function of probe outlet temperature.   High (>5 percent)  sample-stream  moisture
content (or presence of water droplets) can,  however,  alter this relationship.
     2.  The APTD-0576 reference curves provide reasonable estimates  of
probe outlet temperature when applied  to the  sampling  of  streams similar to
those upon which they are based, i.e., streams of low  moisture content  with
temperatures between 80°F and 250°F.  The use of the 80°F "inlet curves  for
ambient streams with temperatures as low as 37°F has been shown to  be satisfac-
tory (Case I); by analogy, it can be inferred that the 250°F inlet  curves will
apply reasonably well to low-moisture  streams with temperatures up  to about
300°F.
     3.  For very hot stacks (T>300°F) and for stack gas  streams of high moisture
content (or containing water droplets), it has been demonstrated that the prac-
tical value of the APTD-0576 reference curves diminishes  considerably.   In
these cases, sample gas overheating or overcooling at the probe outlet can
occur  (depending on the probe heater setting and the temperature of the gases
surrounding the main body of the probe) and can cause the filtration temper-
ature  to be outside the desired operating limits.

-------
                                  60
                            REFERENCES
1.   "Standards of Performance  for New Stationary Sources," Federal  Register.
    December 23, 1971.
2.   Rom, Jerome, J.. Maintenance, Calibration, and Operation of Isokinetic
    Source-Sampling Equipment,  Environmental Protection Agency.  Publication
    No. APTD-0576.  Research Triangle Park, N. C. 27711.  March, 1972.
3.   Martin, Robert  M.. Construction  Details of Isokinetic Source Sampling
    Equipment.  Environmental  Protection Agency.  Publication No. APTD-0581.
    Research Triangle Park, N.  C. 27711.  April, 1971.

-------
      61
APPENDIX

-------
                   62
        TABLE II: CASE II DATA
Date: 1/7/75
Case:    II
Sample Rate: 0.75 cfm
Sample Box Setting:  240 °F
Probe Heater Setting: 25%
Operator: R. Vollaro
Time
(Minutes)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Probe Inlet
Temperature
(°F)
455
457
466
470
472
474
476
477
477
477
477
479
476
476
475
477
477
477
477
477
476
475
475
475
474
Probe Outle
Temperature
(°F)
153
157
161
164
167
170
172
174
177
179
182
188
193
195
197
198
200
201
203
204
208
213
217
221
225
Sample Box
Temperature
(°F)
243
220
241
260
262
231
216
242
267
256
230
245
269
247
226
223
258
265
241
223
234
256
262
240
219
Filtration
Temperature
(°n
181
183
183
185
188
192
193
192
193
196
198
198
199
203
204
204
204
205
208
209
209
208
211
215
216
Powerstat
Setting
(%-)
25
25
25
25
25
25
25
25
25
25
40
40
40
40
40
40
40
40
40
40
60
60
60
60
60

-------
      63
  TABLE II
(Continued)
Time
(Minutes)
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Probe Inlet
Temperature
r°n
474
475
474
473
473
472
471
472
472
471
471
472
475
472
471
Probe Outlet
Temperature
(°n
228
229
231
234
235
241
244
250
254
257
261
262
265
268
269
Sample Box
Temperature
(°F)
231
257
259
238
222
232
256
257
235
220
235
263
254
234
222
Filtration
(Temperature
f«n
216
216
219
222
223
222
222
225
229
230
230
231
234
236
237
Powerstat
Setting
f*)
60
60
60
60
60
75
75
75
75
75
75
75
75
75
75

-------
                   64
Date:
Case:
TABLE III: CASE IIIA DATA
12/6/74
   IIIA
Sample Rate: 0.75 cfm
Sample Box Setting: £5jj_
Probe Heater Setting:   20%
Operators: R. Vollaro and R. Mobley
Time
(Minutes)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
'robe Inlet
Temperature
(°n
508
509
510
513
514
511
513
514
516
515
514
517
519
521
522
524
529
534
535
532
532
533
532
533
535
Probe Outlet
Temperature
cn
259
272
280
286
291
296
299
302
304
305
306
307
309
310
311
314
314
313
309
306
305
304
304
304
304
Sample Box
Temperature
cn
272
269
245
225
236
266
273
247
226
253
276
283
256
234
251
277
274
249
228
235
265
282
256
235
236
Nitration
Temperature
f«F)
223
235
245
252
255
257
262
266
267
267
269
272
275
275
275
275
276
278
278
276
274
274
275
276
274
Powerstat
Setting
(*)
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
0
0
0
0
0
0
0
0

-------
                 65
     TABLE IV:  CASE 11 IB DATA
        1/27/75
        IIIB
Date:
Case:  	
Sample Rate: 0.75 cfm
Sample Box Setting:   255 °F
Probe Heater Setting:   25%
Operator: R. Vollaro
Time
(Minutes)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Probe Inlet
Temperature
?°n
450
453
454
455
456
457
456
457
458
462
463
466
469
470
470
467
472
473
476
477
476
478
479
485
486
3robe Outlel
Temperature
ifon
256
277
281
284
288
292
295
299
301
304
305
308
313
316
320
322
326
327
330
332
333
335
336
338
344
Sample Box
Temperature
?-n
279
248
225
243
269
269
243
220
256
278
262
237
234
256
278
264
238
225
262
282
257
231
237
275
266
Filtration
Temperature
7°F)
212
230
243
251
256
262
266
268
269
271
274
277
277
277
279
283
286
286
286
288
291
293
292
292
295
Powers tat
Setting
m
25
25
25
25
25
25
25
25
25
25
25
25
40
40
40
40
40
40
40
40
40
40
40
40
60

-------
   66
 TABLE  IV
(Continued)
Time
(Minutes)
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
'robe Inlet
Temperature
(°F)
493
490
490
492
491
492
492
491
492
496
497
500
502
503
503
509
506
508
507
511
511
511
509
Probe Outlet
Temperature
(°F)
352
357
361
365
367
371
372
373
375
376
379
395
404
413
417
424
428
431
432
434
435
436
438
Sample Box
Temperature
(°F)
244
224
248
275
273
247
227
253
273
274
247
228
261
286
269
243
224
256
275
283
269
251
227
Filtration
Temperature
(°F)
298
300
301
304
307
310
312
312
312
315
316
317
319
325
328
333
335
337
338
339
341
343
344
Powerstat
Setting
w
60
60
60
60
60
60
60
60
60
60
60
80
80
80
80
80
80
80
80
80
80
80
80

-------
                                     67
               LABORATORY  EVALUATION OF SILICA GEL COLLECTION
        EFFICIENCY  UNDER VARYING  TEMPERATURE AND PRESSURE CONDITIONS
                     Peter R.  West!in  and  Fred C. Biddy*

Introduction
     The impinger section  of the  EPA  Method 5  sampling  train   is  intended
to collect moisture from  sample gases  for  determination of  moisture con-
tent.  The final stage of the collection  train is an impinger with silica
gel.  Laboratory experiments were conducted in order to determine the
effectiveness of the silica gel impinger as a  moisture  collector under
various sampling conditions of temperature and pressure.
Equipment Set-up
     Figure 1 shows the sampling train as  it was used in the experiments.
The moisture and heat source was a flask of water heated by a rheostat con-
trolled hotplate.  The sample  entered the  train through a flow control-valve
used to simulate resistance  through the sampling train.  The first and
second  impingers of  the train  were each filled with  100 mill inters  (ml) of
water.  The third  impinger was a dry  impinger with wet-bulb  and  dry-bulb
thermometers  attached to  the center  tube.   In order  to reach the gas  velocity
necessary to  obtain  correct wet=bulb  temperature readings, a 2.2-cm  (0.87-in.)
diameter  orifice was placed in this  impinger,  and  the  thermometer tips were
 located in the orifice opening.   The fourth impinger contained approximately
 200 grams (g) of silica  gel for each run.  The silica  gel  was grade 42 and a
 6-16 mesh size indicating type.  Following the impinger section was another
 thermometer for measuring the temperature of the gas stream leaving the silica
 gel impinger.  A  standard EPA Method 5 meter box was used to draw and measure
  *  Emission Measurement Branch,  ESED,  OAOPS,  EPA,  RTP, NC, July  1975

-------
                                          68
   FLOW RESTRICTING
        VALVE
STEAM
WATER-FILLED   DRY IMPINGER    SILICA GEL
  IMPINGERS         WITH         IMPINGER
                                          ORIFICE
                                                                              TO METER
                                                                                BOX
                                                                     ICE BATH
                      Figure 1.  Laboratory moisture sampling train.

-------
                                  69
the volume of sample and measure the vacuum in the volume meter.
Procedure
     At the beginning of the test run, the flow control  valve and the meter
box pump were adjusted to attain the desired meter vacuum and flow rate.
Flow rate was maintained between 1.1 and 1.3 standard cubic meters per hour
(scm/hr)(0.52 to 0.61 scf/min).  Temperature in the impingers was controlled
and maintained with an ice bath.  Readings of meter volume, meter temperatures,
train vacuum, wet-bulb and dry-bulb impinger temperatures, and exit-gas tem-
perature were recorded at 5-minute intervals during the 2-hour runs.   Adjust-
ments to flow rate and train vacuum were made as necessary.  The  vacuum in
the third impinger was determined so that the moisture content of gas enter-
ing the silica gel impinger could be calculated.  This sample vacuum, reported
in Table 1, was varied from 107 to 460 millimeters of mercury (mm Hg)(4.2 to
18.1 in. Hg).
    The moisture entering the silica gel impinger, the moisture entering  the
meter box, and the moisture collection efficiency of the silica gel were
calculated as follows:
    1.  Moisture fraction entering silica gel:

              PS -0.00066(1 + 0.00116b T^HPjMT^ - Twb)
        Bwi                       fT	

    Where:
        B  . = volume fraction of moisture, %/100

-------
                                 Table 1.  MEASURED AND CALCULATED VALUES RELATIVE TO
                                       MOISTURE COLLECTED IN SILICA GEL IMPINGER
                               IN EPA METHOD 5 SAMPLING TRAIN DURING LABORATORY TESTS3



Run
A
B
C
D
E
F
G
H
I
J
K

a i
Sample
vacuum,
mm Hg
107
109
107
107
107
224
226
226
460
460
460
T.
db
Inlet
temp. ,
°C
9.2
15.2
26.6
21.6
32.6
9.9
19.3
33.8
9.7
21.2
33.7
T
'e
Exit
temp. ,
°C
14.2
18.5
30.0
24.1
34.9
14.5
25.0
35.2
16.8
27.0
35.8
V
m
Meter
vol ume
scm
2.490
2.257
2.458
2.494
2.491
2.671
2.496
2.470
2.520
2.546
2.536
e
sg
Collected
H20 vol.,
scm
0.024
0.034
0.068
0.052
0.063
0.035
0.059
0.076
0.037
0.060
0.065
V + e
m sg
Total
volume,
scm
2.51
2.29
2.53
2.55
2.55
2.71
2.56
2.55
2.56
2.61
2.60
B .x 100
Wl
Inlet
H20,
%
1.3
2.0
4.0
3.0
5.7
1.7
3.2
7.5
3.0
6.4
13.2
e.
i
Inlet
H20 vol . ,
scm
0.032
0.046
0.102
0.076
0.145
0.047
0.081
0.190
0.078
0.167
0.343
eo
e
Exit
H20 vol . ,
scm
0.007
0.012
0.034
0.024
0.082
0.012
0.022
0.114
0.041
0.106
0.278
ee
^
m
Exit
H20,
%
0.3
0.5
1.3
0.9
3.2
0.4
0.9
4.5
1.6
4.1
10.7
e
100 x I9-
i
Collection
efficiency,
%
78
74
67
68
43
75
73
40
48
36
19
a  Symbols above columns refer to calculations section of text.

-------
                                71
    P_ = saturated vapor pressure  at  T. , mm  Hg
     S                                WD
    T .= wet-bulb temperature,  °C
    T..= dry-bulb temperature,  °C
    P. = absolute impinger pressure,  mm Hg
2.  Volume of moisture collected by silica  gel:
    esg= 1.342 x Iff3 (Msg)
Where:
    e  = moisture gas volume in silica gel, scm
    M  = mass of water collected  in silica  gel,  g
Note:  Standard temperature and pressure are  21°C  (70°F) and  760
    mm Hg (29.92 in. Hg)
3.  Total moisture volume entering silica gel:

    e1 = Bwitesg + VJ
Where:
    e. = moisture volume entering  silica gel, scm
    V  = standard dry-gas meter volume, scm
4.  Collection efficiency of silica gel:
    E  = 1

Where:
    E  = collection efficiency, %

-------
                                      72
     5.   Moisture exiting the silica gel impinger:
         ee = ei   esg
     Where:
         e  = moisture volume exiting silica gel, scm
Discussion of Results
     Table 1 shows the results of 11 test runs at three different train
vacuums.  Note that the calculated values represent averages over each of
the complete 2-hour runs.  As can be seen from this table, the moisture
collection efficiency of the silica gel decreased as the inlet temperature
and the exit-gas temperature increased.  A result of this relationship
is that the  percent of moisture  in the  sample gas entering the meter box in-
creased from 0.3 percent at  14.2°C  (58°F) exit temperature to 3.2 percent
at 34.9°C  (95°F) exit temperature at the same train vacuum of 107 mm Hg
(4.2 in. Hg).
     Also  shown  in these results is  the effect of sample train vacuum on
collection  efficiency.   For  example, looking at  runs A, F, and I, the  inlet
temperatures are approximately equal at 9.5°C  (49°F) while the train vacuum
varies  from 107 mm Hg  (4.2 in.  Hg)  to  460 nm  Hg  (18.1  in. Hg).   The silica
gel  moisture collection efficiency  decreases  from 78 percent  at  107 mm Hg
 (4.2 in.  Hg) to 48  percent at 460 mm Hg (18.1  in. Hg).  Moreover, the  amount
of moisture exiting  the silica gel  increases  from 0.3  percent to 1.6  percent
over the same  conditions.   Figure 2 shows  graphically  the effect of both
 exit temperature and sample vacuum on  the  moisture  content  in the exit gas
 from the silica gel  impinger.

-------
                                       73
100
200
         300
SAMPLE VACUUM, mm Hg
                                                             400
  Figure 2. Silica gel exit moisture content versus sample vacuum and exit temperature.

-------
                                   74
     Note that the exit gas from the silica gel impinger has a temperature
several degrees higher than the inlet temperature of the silica gel  im-
pinger.  This temperature difference was somewhat influenced by temperature
of the room, but also by the exothermic reaction that occurred when  moisture
was adsorbed by the silica gel.
                                                          2
     Data from similar EPA experiments reported by Johnson  substantiate
the findings of this study.  For example, at an impinger temperature of
21 °C (70°F) and a vacuum of about 254 mm Hg (10 in. Hg), Johnson found
that the moisture collection efficiency was from 60 to 70 percent, comparing
favorably with the 70 percent  predicted by the curve in Figure 3.  Other
values, difficult to compare because temperature conditions and pressures are
different, in general, show trends similar to the results noted here.  Table
2 shows the results of Johnson's study.  In addition, further experiments by
Johnson showed that adding one or two more silica gel impingers did little
to decrease the moisture content of the final exit gas.  This implies that
the moisture collection ability of silica gel in the EPA Method 5 train is
limited by temperature and pressure conditions.
     The moisture content  in the exit gas of the silica gel also affects
the dry gas meter volume.  A 5 percent increase in moisture content produces
a similar increase  in volume.   In source sampling results, this means a 5
percent error in the  isokinetic calculations and an error in the emission
calculations.
Conclusions
     The moisture content  of  the  gas entering  the meter box can be greatly

-------
                                       75
10
                      200
                                                                400
                                            300
                                   SAMPLE VACUUM, mm Hg
Figure 3. Silica gel moisture collection efficiency versus sample vacuum and exit temperature.
                                                                                     500

-------
                   76
Table 2.  Results of W. L. Johnson's Study of
                                            2
Moisture Collection Efficiency of Silica Gel
Run
A
B
C
D
E
F
G
Vacuum
mm Hg
508
508
508
381
254
152.4
127
Impinger
temp . ,
°C
-
18.3
16.7
21.7
21.7
21.7
16.7
Exit gas
temp. ,
°C
27.8
21.1
20.0
21.7
20.6
22.8
20.0
Collection
efficiency,
%
38
52
61
46
60
52
84

-------
                                  77
affected by the temperature and vacuum of the  sampling  train  during  Method
5 testing.   The moisture in the sample gas is  incorrectly measured as  "dry"
gas by the volume meter, and this value is carried through the isokinetic
calculations as well  as the concentration calculations.   A volume measure-
ment error due to moisture in the sample gas directly affects the isokinetic
calculations; a 3 percent increase in moisture content of the sample gas
produces a similar error in the isokinetic results.
     Method 5, as written, stipulates that the sample temperature as it
exits the silica gel  impinger exit gas be held below 21°C (70°F) and that
the sample train vacuum be held under 381 mm Hg (15 in. Hg).   These tests
show that at these limits the  "dry" gas volume error would be less than 2
percent, and a similar error would appear in the isokinetic determination.
     It is noted in the text that the temperature of the wet gas in the
third dry impinger was 4° to 7°C  (7° to 13°F) less than the temperature of
the exit gas from the  silica gel.  This difference is influenced by the heat
of adsorption  of the silica gel and ambient conditions.  These tests were
run under steady-state ambient temperature conditions and therefore do  not
reflect results  that may  be obtained under field conditions.  A  better  field
 indicator of acceptable temperature  limits for the sample gas would be  the
dry-bulb temperature in the dry impinger  preceding the  silica gel.  A limit
 of 15.6°C  (60°F)  in  the impinger would  meet the  intentions of the present
 EPA Method  5  specifications.
 References
 1.  Title  40 -- Protection of  the Environment,  Part  60 -- Standards of  Per-
     formance for New Stationary Sources.   Federal  Register.  36_ (247):  24888,
     December 23, 1971.

-------
                                    78
2.  Johnson, William L.,  "Moisture Collection Efficiency of Silica Gel  in
    Stack Sampling Trains,"  Environmental Protection Agency, National
    Environmental Research Center, Research Triangle Park, North Carolina
    unpublished report, 1974.

-------
                                            79
                  UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

                      Research  Triangle  Park, North Carolina  27711

SUBJECT:  spurious  Acid Mist  Results  Caused  by Peroxides in     DATE:  January 22, 1976
          Isopropvl  Alcohol Solutions  Used in EPA Test Method (M-8)
FROM:     Dr.  josepTT.  Knoll,  QAB/EMSL  (MD#77)
TO.
Mr. Roger T. Shigerhara, EMB/ESED (MD#19)
                An  evaluation  study  of EPA  Test Method  (M-8) for the Determination
          of Sulfuric  Acid  Mist  and  Sulfur  Dioxide Emissions from Stationary Sources
          has been  carried  out in  the  Quality Assurance Branch.  One result of this
          study has been  the finding that peroxide impurities  in the isopropyl alcohol
          used for  acid mist collection can convert  sulfur dioxide to sulfuric acid
          and result in erroneously  high acid mist values.  The quantities of sulfur
          dioxide collected as sulfuric acid were of the order of from  ten to twenty
          five percent of the  EPA  compliance standard.  It was independent of the
          quantity  or  concentration  of sulfur dioxide that had passed through the
          system and only dependent  on the  quantity  of  peroxide, traces of which may
          occasionally be found  in reagent  grade isopropyl alcohol.

                The following  test is  tentatively proposed for detecting peroxides in
          isopropyl alcohol:

                    Shake 10 ml  of isopropyl alcohol with 10 ml of freshly
                    prepared  10% potassium  iodide solution.  Prepare a  blank
                    by similarly treating 10 ml of distilled water.  After
                    one minute,  read the absorbance  at  352 nm. If absorbance
                    exceeds 0.1, reject alcohol for  use.

                Peroxides may  be removed from isopropyl alcohol by redistilling or
          by passage through  a column  of activated alumina.  However, it is possible
          to obtain reagent grade  isopropyl alcohol  with suitably low peroxide levels
          from commercial sources, so  that  rejection of contaminated lots may be a
          more efficient  procedure.

          cc:  M. R. Midgett
 EPA Form 1320-6 (Rev. 6-72)

-------
                                           80
                  UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                    Office of  Air  Quality  Planning and Standards
                    Research Triangle  Park,  North Carolina 27711
SUBJECT:  Determination of  Isopropanol  Loss During Method 8     DATE:
            Simulation Tests

FROM:    Peter R. Westlin, Test  Support Section
         Emission Measurement  Branch  (MD 19)

TO:      Roger T. Shigehara, Chief, Test Support Section
         Emission Measurement  Branch  (MD 19)
                                                                      JUN  2 5 197&
              In answer to questions regarding potential  loss  of isopropanol(IPA)
         through evaporation and a subsequent error in moisture determination
         when using Method 8, a laboratory program was undertaken at the  IRL
         during June 17 and 18.  A Method 8 sampling train was set up without
         the glass filter between impinger 1, the IPA impinger, and number 2, the
         first hydrogen peroxide, \\2Q2» impinger.  The third impinger in  the
         train was also a \\2^2 impinger, while the fourth was  left dry.   The
         fifth and last impinger contained silica gel.  A standard Method 5 meter
         box was used to draw and measure the volume of the sample.

              Two test runs were completed.  For the first run, 200 milliliters(ml)
         of IPA was placed in the first impinger, 100 ml  of ^Og in each  of the
         next two, the fourth impinger was left dry, and about 300 grams(g) of
         silica gel were placed in the last impinger.  About 1420 liters(l) (50 ft3)
         of room air were drawn through the train at a flow rate of about 70 liters
         per minute(lpm) (0.8 cfm).

              The results showed a loss of 40 ml of solution in the first impinger,
         a gain of 15 ml in the second, a 12 ml gain in the third, a negligible
         gain in the dry impinger, and 20.5 g gain in the silica gel.  The net
         change across the train was 7.5 ml (assuming the mass gained on  the silica
         gel was water).  A specific gravity determination showed that  the original
         IPA solution had been prepared incorrectly and was 67% IPA rather than the
         specified 80% IPA.  The solution remaining in the IPA impinger after the
         test run was shown to be 52% IPA.  The 15% loss corrected for  total
         volume change represented a loss of 51 ml of IPA.

              For the second run, 100 ml of IPA solution was placed in  the first
         impinger and the rest of the sampling train was the same as for  run 1.
         The IPA solution was prepared as specified in the Federal Register
         (December 23, 1971) and a specific gravity check of concentration showed
         the solution to be 73%  IPA.
                                      <3
              After some 1500 1  (54 ft ) of sample were drawn through the train,
         the first IPA impinger showed a 50 ml volume loss, the second  measured a
         75 ml gain, the third showed an 18 ml gain, the third showed a 2 ml gain,
         and the silica gel mass total increased 11 g.  The net volume  change across
EPA Form 1320-6 (Rev. 6-72)

-------
                                   81

the train was a 6 ml increase.  The final IPA solution concentration
was 27.4% representing a loss of about 59 ml of IPA.

     These test results indicate that measurement of volume gain in the
impingers of the Method 8 for the purposes of calculating sample mois-
ture content is not impaired by any loss of IPA through evaporation.
In neither test run was there a net loss of volume from the sample trains.
The net gain was approximately equivalent to 0.6% moisture or about 25%
relative humidity.

     A notable secondary finding of this short study was the great change
in IPA solution concentration during a test run.  Approximately 1  ml of
IPA was removed from first impinger per 30 1 of sample gas for each test
run.  Initial volume of IPA solution or IPA concentration appear to
have little effect on this ratio.  Some IPA may have been evaporated and
condensed farther down the train.  More IPA was probably carried through
as a mist and collected later.

     This IPA loss may be significant if the concentration of IPA gets
too low to effectively inhibit oxidation of SO* during Method 8 sampling.
Joe Knoll was not aware of this potential problem and could not tell me
what a lower effective limit of IPA may be.  He agreed that it could be a
significant problem not only in the possible interference from S02 oxida-
tion, but also by meeting the titration end point analysis.

     I suggest further work in this problem area be undertaken.  Such a
project may be suitable for one of the co-op students in the next several
months.

-------
          COMPARISON OF EMISSION RESULTS FROM IN-STACK FILTER

                  SAMPLING AND EPA METHOD 5 SAMPLING

                  Peter R. Westlin and Robert L.  Ajax
                               Abstract

      A  series  of replicate emission  tests  using  in-stack and out-of-stack
sampling  trains  were conducted at each of  four fossil-fuel-fired power
generation  stations.   The sampling train used for measuring in-stack
particulate included a probe nozzle  and an  in-stack glass fiber mat
filter, followed by  a heated probe extension and an out-of-stack filter
The  Environmental  Protection Agency  Method  5 particulate sampling system
was  used  as the  out-of-stack sampling  train.  The two sampling trains were
operated  simultaneously at approximately the same point in the stack gas
streams with no  traversing.

      The  particulate  catch from each sampling train was analyzed for
particulate mass,  sulfate content, organic  content, and acidity.  For
the  in-stack train,  the results are  reported for both the in-stack catch
(the  particulate obtained from the nozzle and the in-stack filter), and
the  total catch  (the  in-stack particulate,  plus the particulate washed
from  the  probe extension  and the out-of-stack filter).

      The  tests at  two coal-fired units  with electrostatic precipitators
and an oil-fired unit with no control  device resulted in the out-of-stack
train catch exceeding the in-stack catch, in each case.  The difference
varied with the  sulfur content of the  fuel   and ranged from 10 mg/dscm at
the unit  firing  0.3%  sulfur  oil,  to  112.6 mg/dscm at the unit firing 3%
sulfur coal.  The  measured sulfate did  not, however, fully account for
this  difference.

     Opposite results  were obtained at  a second oil-fired unit with a
wet limestone scrubber.   At  this  unit,  which was burning 2.5% sulfur
fuel, the in-stack catch  was  significantly greater than the out-of-stack
train catch (421.3 mg/dscm versus 217.6 mg/dscm respectively).   This
difference  was the apparent  result of a reaction occurring on the wet
in-stack  filter.
    Emission  Measurement Branch,  ESED,  OAQPS, EPA, RTP, NC
    Presented at  the  annual  APCA  Meeting, June 1975
                                     82

-------
                                                                  75-19.1

Introduction

     During the summer of  1973  the  Emission Measurement Branch of the
Environmental  Protection Agency (EPA)  undertook a project in which particu-
late emissions were sampled with  in-stack filter and out-of-stack filter
sampling trains.

     The purpose of the project was to obtain and compare particulate
emission sampling  results using sampling trains in which  all  components
except  filter  location were identical.  The equipment used for measuring
the in-stack particulate, as shown in Figure 1, consisted of  a probe nozzle
and an  in-stack filter, followed by a probe extension, and an out-of-stack
filter.  The EPA Method 5 sampling train,1  shown in Figure 2, was used to
measure particulate out-of-stack at 120°C.   The two trains were operated
simultaneously side-by-side at approximately the same point in the stack
gas streams at each of four fossil-fuel-burning power generating stations.
Two plants were coal-fired with electrostatic precipitators and two plants
were oil-fired, one using a wet limestone scrubber, and the other having  no
supplementary  emission control.  The sulfur content of the fuels ranged
from 0.29  to  3.3 percent.

     The particulate  catch from each sampling system was analyzed for
particulate mass,  sulfate  content, organic content, and acidity.  The
results are reported  for the in-stack catch (the particulate obtained from
the nozzle and the in-stack filter), the total in-stack (the in-stack par-
ticulate plus  the  particulate  washed from the probe extension downstream
and the out-of-stack  filter),  and  the EPA Method 5 catch (the particulate
from the nozzle, the  probe, and  the out-of-stack filter).  The impinger catch
results are not reported  in this paper as the dry or front half results were
of concern in this project.

Methods

     A special dual-probe  sampling box was constructed to  house the  two
sampling  trains and to allow for simultaneous operation of both systems.
Two equal-length sampling  probes were  employed side-by-side with the  probe
tips approximately 10 centimeters  apart.  No  provisions were made for tra-
versing of the stack  cross-section as  only relative  concentrations were
 desired.   Although only one point  was  sampled,  isokinetic  conditions  were
maintained.   A pi tot tube was  attached  to  the EPA  Method  5 sampling  probe
 to permit velocity head measurements,  and  adjustments  in  the  sampling rate
 of each train were made every  five minutes  during  sampling to maintain iso-
 kinetic sampling  conditions.   Other  measurements  recorded  at  regular inter-
 vals  included stack temperature, dry gas volume,  meter vacuum,  gas meter
 temperatures, orifice pressure drop, and sample  box  temperature.  Sample  box
 temperature in the enclosure  housing the box filters was  carefully monitored
 and maintained at or above 120°C.   The sample box temperature was measured
 with a thermocouple located in the downstream half of one of the  box filters.

      The  in-stack sampling train was composed of a button-hook sampling
 nozzle; a 5.7 cm  diameter glass-fiber mat  filter and an  in-stack  filter
 nolder; a heated, glass-lined probe; a second 7.6 cm diameter glass-fiber

                                     83

-------
                                IIKTACK    FILTER HOLDS
CO
                                                                                AIR-TIGHT PUMP
                                               Figure 1.  In-slack paniculate sampling tram.
                                                                                                                         en

-------
00
en
                           VACUUM
                  MAIN       LINE
r	1     O     VALVE
                                                  n  L_iH>
-------
                                                                    75-19.1
mat filter  in  a  heated  sample  box; and condensation impingers.   This arrange-
ment along  with  the  pump  and metering equipment is displayed in Figure 1.
The EPA Method 5 train  (Figure 2) was described in the December 23, 1971
Federal Register.'

     Cleanup procedures were as prescribed for Method 5 in the Federal
Register except  that the  nozzle tip and trie filter holder upstream of the
in-stack filter  were cleaned with acetone and stored separately from the
rest of the probe.   Tiie probes were rinsed and brushed with acetone, and
the catch was  saved  for analyses.  The out-of-stack filters and the in-stack
filters were stored  in  glass petri dishes.  The filter holders were rinsed
with acetone as  were the  impingers after the water condensate was saved.

     Each sample was carefully analyzed for particulate mass, sulfate content,
organic content, and acidity,  except for one test where only mass and sulfate
were analyzed.   The  samples were divided into aliquots in order to obtain all
the necessary  information.  The acetone solutions were divided into three
aliquots:   the first was  used  to determine mass of particulate, the second
was titrated for acidity  and for sulfate as S04, and the last was extracted
for organic materials.  The filters were first weighed for particulate mass
and then divided in  half:  one half used to determine organic materials, and
the other half analyzed for acidity and sulfate contents.

     Particulate mass was determined gravimetrically after proper dessication.
Sulfate (S0|)  content was determined using the thorin titration technique.
Ether-chloroform extraction was used to establish the organic content of each
sample and  an  acid-base titration was employed to determine the acid content
of each sample._ The analytical results were expressed in milligrams (mg) for
particulate, $04, and organics; and in milliequivalents (meq) of h^SC^ for
acidity.  These  analytical results were then converted into concentration
units—milligrams per dry standard cubic meter (mg/dscm)--for statistical
analyses and reporting.   The sulfate catch was assumed to exist as sulfuric
acid, and the  concentration of sulfate was expressed as mg/dscm of H?SO/i +
2H20.2                                                              *  *

     The statistical  significance of differences between the various data
sets was determined  by  the t-test.  For the purposes of this report, a  0.05
percent probability  level was  set as the minimum of acceptance or rejection
of the hypothesis.

Results

     Tables 1  through 4 show emission concentrations as determined by the
emission tests at the four power plant facilities.  Comparisons are made
between the EPA  Method  5  concentrations and the in-stack concentrations using
data obtained  during simultaneous, single point sampling.  The pollutant
emission data  and the oxygen data supplied by plant personnel were also used
to estimate the  emission  rate  from grams per standard cubic meter to grams
per million calories.  The oxygen measurements were made at sampling points
other than  the particulate sampling points and as a result the emission rates
are only approximate values.

                                     86

-------
                                                                 75-19.1

                                 Facility A

     Table 1  shows  the  data  obtained  at facility A, a coal burning power
plant employing an  electrostatic  precipitator.  The coal being fired had a
reported sulfur content of 3.3 percent.  Only  particulate mass and sulfate
concentrations  were determined for this facility.  The average EPA Method 5
particulate emission concentration was  129.8 gm/dscm corresponding to an
approximate emission rate of 0.49 grams per million calories  (g/10° cal).J
As shown in Table 5, the average  particulate concentration determined from
the EPA Method 5 train  was significantly greater than the particulate catch
for the in-stack sampling train,  129.8  versus  17.2 mg/dscm.   Adding the probe
wash and the filter catch downstream  of the  in-stack filter  to the in-stack
catch produced a total  particulate concentration of 124.6 mg/dscm which was
not significantly different  from the  Method  5  dry particulate concentrations.

     Similarly, the sulfate  found in  the EPA Method 5 train,  76.3 mg/dscm,
was significantly greater than the 4.7  mg/dscm in-stack  catch.   The total
catch of the in-stack train  indicated a sulfate concentration of 60.6 mg/dscm
which was not significantly  different than  the 76.3 mg/dscm  EPA  Method  5
sulfate catch.

                             Facility B

     The emission concentration data  obtained  at facility  B,  an  oil-fired
power plant with no control  devices,  are shown in Table  2.   Sulfur content
of the oil was reported to be 0.29 percent.   The average EPA Method  5  particu-
late emission concentration, 38.8 mg/dscm,  was significantly greater  than  the
in-stack particulate concentration of 30.1  mg/dscm.   Compared with facility A,
however, the actual magnitude of this difference  is  small:   8.7  vs 112.4  mg/dscm.
No significant difference was found between the particulate  concentration  found
in the total in-stack dry train, 42.0 mg/dscm, and  that found in the  EPA
Method 5 equipment.  The EPA Method 5 particulate  concentration  corresponded
to an approximate emission rate of 0.05 g/lQo cal  for facility  B.

     The difference between EPA Method 5 sulfate concentration,  13.6 mg/dscm,
and  the  in-stack sulfate concentration, 8.4 mg/dscm, was small  and was not
statistically  significant.  Similarly, the total  in-stack train sulfate con-
centration of  15.7 mg/dscm was not significantly different from the EPA
Method  5 catch,  13.6 mg/dscm.

      The organic matter concentration  of the  EPA Method 5 catch was 12.8
mg/dscm, a  level significantly greater than the 7.9 mg/dscm  found in the
 in-stack filter assembly, but significantly less than the 17.6 mg/dscm
 captured in the total  in-stack dry sampling train.  The acidity  concentration
 of the EPA Method  5 sampling  train,  18.9 mg/dscm of H?S04, was  significantly
 greater than both  the  in-stack concentration, 3.4 mg/dscm, and  the total
 in-stack sampling  train concentration, 10.7 mg/dscm.
                                     87

-------
                                                                      75-19.1

                             Facility C

     Facility C was a coal-fired boiler using coal  reported at 0.85 percent
sulfur and controlling emissions with an electrostatic  precipitator.  Neither
the difference between the dry particulate concentration  of the EPA Method 5
train, 226.2 mg/dscm, and the particulate concentration of the in-stack train,
207.9 mg/dscm, nor the difference between the EPA Method  5 particulate con-
centration and the total in-stack train concentration of  223.4 mg/dscm was
statistically significant.   It is, however, noteworthy  that the difference
of 18.1 mg/dscm between the  EPA Method 5 train catch and  the  in-stack catch,
and the corresponding values of 112.6 and  8.7 mg/dscm  fj"nj»t fa"!^?. A
and B  respectively, each show a consistent relationship to  the fuel sulfur
cSntent-0.85, 3.3  and 0.3% sulfur  for facilities C, A,  and  J respectively.
This relationship is shown graphically in Figure 3.  This is  in spite of  the
fact that  the average EPA Method  5 particulate concentration  at facility  C  is
equivalent to an approximate emission  rate of 0.34 g/10* cal  which differs
from facility B by a factor  of  7.

     Sulfate emissions  for  facility  C  found  using  the  EPA Method  5 train
averaged  5.7 mg/dscm, a level  significantly  greater  than th*^?"!™110"
determined from  the  catch of the  in-stack sampling train, 2.9 mg/dscm.   The
sulfate  concentration of the EPA  Method  5 catch was  not  significantly  different
from  the  sulfate  concentration, 4.7  mg/dscm, of  the  total in-stack dry
sampling  train  catch.   The  average  EPA Method 5  organic  concentration  for
facility  C,  12.4 mg/dscm, was  not significantly  different from the in-stack
organ  c  concentration  of 13.5 mg/dscm.   Neither  was  It different  from the
total  in-stack train concentration  of 16.4  mg/dscm  The acidity  concentra-
tion  (H?S04) of the EPA Method 5  train was  small,  3.2  mg/dscm, for facility
but was2significantly greater than  the acidity concentration found by the
 in-stack sampling train, 2.0 mg/dscm.   When the back catch was added tthe
 in-stack concentration, the resulting total  i n-stack acidity """^ration
was 4.0 mg/dscm.  This number was significantly greater  tnan the  acidity
 concentration obtained by the EPA Method 5 sampling assembly.  Note that the
 actual magnitudes of the components-sulfates, organics. a;J acidity--are
 relatively small and are less than  about 5 percent of  the  total  particulate
 mass  for  both the EPA Method 5 catch and the in-stack  catch.

                              Facility D

       Facility D was an oil-fired steam generating station  using  oil with a
 sulfur content of 2.45 percent.  The  plant  employed a limestone  scruboer as
 the emission control system.   No reheat device was  present in  the gas  stream
 prior lo  the sampling  location.  This,  along with apparent problems  in the
 dem?ster  resulted  in'an exhaust  gas stream  which was  supersaturated with
 moisture    The  mist caused  some  problems in sampling  and may be  the source of
 tte  anomalies  in  tSe  comparison  results  that follow  The particu a e con-
 centration  found by the EPA Method  5 system was 217 6 mg/dscm, significant^y
 less  than the  particulate  catch  of  the in-stack filter, 421.3 mg/dscm.  The
 EPA  MethSd 5epaPrticulate concentration corresponded to  an approximate nass
 emission rate of 0.31  g/105 cal  for facility D.   The  total in-stack train
  particulate concentration was 727.2 mg/dscm.
C
                                      88

-------
                                                           75-19.1
                    SULFUR CONTENT !N F'JEL.%


Figure 3.  EPA method 5 particulate minus in-stack particulate
           versus sulfur content in fuel.

                              89

-------
                                                                       75-19.1

     The results of the comparison of sulfate concentrations  showed different
relationships than for the participate concentrations.   The sulfate concen-
tration (H?S04 + 2H20) in the EPA Method 5 train averaged  119.2 mg/dscm, a
value significantly greater than the 18.4 mg/dscm found in the in-stack fil-
ter train.  Further, the EPA Method 5 value was greater than  the  total in-stack
train sulfate concentration of 55.5 mg/dscm.  The acidity  concentration of
the EPA Method 5 train, 36.6 mg/dscm, was also significantly  greater  than
the concentration found in the in-stack filter train,  10.2 mg/dscm, although
no significant difference was observed between the acidity concentration found
in the total in-stack sampling train, 31.7 mg/dscm, and the EPA Method 5
train.  On the other hand, comparisons of the organic  catch of the  two samp-
ling trains resulted in relationships similar to the particulate  concentration
comparisons.  The average organic concentration of the EPA Method 5 train,
60.7 mg/dscm, was significantly less than the organic  concentration of the
in-stack filter train, 120.9 mg/dscm, for facility D.   The same  relationship
was true for the total in-stack train concentration of 137.9  mg/dscm.

                       Evaluation of Results

     Various combinations of the different portions of the total  particulate
catch of each of the sampling trains were studied in order to determine  the
source of the differences between measured concentrations.  One  combination
studied was designed to determine if sulfates as ^804 + 2^0 make up the
difference between  in-stack particulate catch and the EPA Method  5 particulate
catch.3  To do this, the EPA Method 5 particulate concentration was compared
with the sum of the EPA Method 5 sulfate catch plus the in-stack  non-sulfate
eaten.  If these new, concentrations were  found  not  to be significantly dif-
ferent, then the difference between the in-stack particulate catch and the
EPA Method 5 particulate catch could be attributed  to the sulfate caught in
the EPA Method 5 sampling train.  A similar  analysis was done to determine
if the difference between the  in-stack sampling  train catch and the EPA
Method 5  catch could be condensible organic  matter  for those tests in which
organic data were available.

      For  facility A,  the coal-burning  power  plant with 3.3 percent sulfur
coal,  the comparison of the  sulfate  test  showed  that a significantly greater
amount of material  was caught  in  the  EPA  Method  5  train than could be ac-
counted for  by  the  sulfate  as  H2S04 +  2H20  in  the  EPA  Method 5 catch.  In
this  case,  the  difference between  in-stack  filterable material and EPA
Method 5  catch was  apparently  not  all  sulfate  matter.

      The  data  from  facility  B, an  oil-fired generator with 0.29  percent
sulfur fuel,  showed that  no  significant  difference could  be  found between
tne  in-stack catch  plus  the EPA Method 5 sulfate catch  and the EPA Method  5
dry  particulate  catch.   Thus the sulfate found in  the  EPA Method 5 train
could have  accounted  for  the difference  between the in-stack dry particulate
 concentration and the EPA Method 5 concentration.   A  similar comparison
 using the organic catch instead of the sulfate showed, however,  that the
                                     90

-------
                                                                  75-19.1

difference between the in-stack catch and the EPA Method 5 catch could also
have been the organic matter found in the EPA Method 5 dry particulate catch
which could also have accounted for the difference.   This indicates that the
variation in the data had as great an influence on the statistical  comparison
results as did the sulfate or organic catch of the EPA Method 5 sampling train.

     The coal-burning power station, facility C, firing 0.85 percent sulfur
fuel produced a comparison of emissions results similar to that of facility B.
That is, either the sulfate or the organic matter found in the EPA Method 5
particulate catch could have accounted for the differences between the EPA
Method 5 particulate concentrations and the in-stack particulate concentrations.
This is as expected since there was no significant difference between in-stack
particulate collections and EPA Method 5 collection for this site.

     Test results of the emissions from site D do not fall into the pattern
set previously by the other three test sites.  Comparisons of concentrations
using the sulfate data or the organic data produced no significant results,
as might be expected.  The particulate concentrations from the in-stack
filter were significantly greater than the dry particulate concentration
from the EPA Method 5 train and could not be accounted for with either the
sulfate catch or the organic catch.

Conclusions

     The in-stack sampling train does not produce results equivalent to the
EPA Method 5 sampling train results at all power plant sites.  At two power
plants where samples were collected in dry stack gases, the in-stack filter
tended to collect less material  than the EPA Method 5 sampling train.  There
was no significant difference between the particulate catch of the  two trains
at a third power plant with dry stack gas and low sulfur fuel.   At  another
site where stack gases were supersaturated with water following a wet scrubber,
the in-stack filter collected considerably more particulate than the EPA
Method 5 train.

     The magnitude of the differences in the material  collected by  the in-
stack filter and the EPA Method 5 train was much greater for the high
sulfur fuel  power stations than for the low sulfur fuel power plants and
showed a consistent relationship to the fuel  sulfur content.   The differences
in the amounts were, however, neither directly attributable to the  sulfates
found in the EPA Method 5 catch nor to organic matter.   Particulate matter
collected outside the stack,  downstream of the in-stack filter made up the
difference between the in-stack catch and the EPA Method 5 catch, but no
definite conclusion as to what this material  was and why it passed  the in-
stack filter,  can be drawn from this study.

     As  for the  cause of the high in-stack filter catch compared to the EPA
Method 5 catch in wet stack gases, chemical  reaction between  the mineral
scrubbing medium and the sulfur oxides in the gas  stream may  be occurring.
These reactions  may occur in  the stack gas streams,  or the sulfur oxides may
react with the minerals  and the moisture on the wet  filter surface  of the
                                     91

-------
                                                                  75-19.1

in-stack filter.  These salts would not be collected on the EPA  Method 5
filter as this filter is heated above the dew-point of water and is relatively
dry.


     Further study in the area of in-stack filters in wet gas streams should
answer tnese questions.  Other types of sampling methods may be  found more
appropriate under these conditions.
 References


 ft- IU1? 4°T-prote  "^9* of Sulfur Dioxide on Particulate  Test  Results.
 Fnnnf communication with Chief of Combustion and Incineration Section,
 Environmental Protection Agency, Research Triangle Park, N.  C.  August 23,



 3.   Shigehara, R  T. R. M. Neulicht, and W.  S.  Smith.   A Method for Calculat
 ing Power Plant Emission Rates.  Stack Sampling News.   1 (1), July 1973.

 4.   Hemeon, W.C.L., and A. W.  Black.  Stack  Dust Sampling:   In-stack Filter
 or  EPA Train.   J.  Air Pol. Control  Assoc.  22 (7):   516-518, July 1972.
                                   92

-------
                                                                    75-19.1
                                  TABLE 1
                 SUMMARY OF EMISSION CONCENTRATIONS  FOUND
              DURING SIMULTANEOUS IN-STACK AND EPA METHOD 5
               EMISSION TESTS AT A COAL-FIRED POWER PLANT3
                                (mg/dscm)
            Particulate concentration
Run
1
2
3
4
5
6
EPAD
93.9
155.0
156.2
65.7
175.3
133.7
In-stack
10.1
20.5
30.6
22.7
8.4
10.7
Total c
97.5
190.8
133.1
127.5
76.3
122.2
EPAb Di
43.8
92.7
80.9
33.0
114.6
93.0
                      H2S04 + 2H20 concentration
                      EPAb Dry   In-stack   Total0
                                   1.7         5.0
                                   2.2        26.2
                                   3.6       108.0
                                  14.3        87.2
                                   2.5        52.7
                                   3.9        84.3
Average   129.8
17.2
124.6
76.3
4.7
60.6
  Sulfur content of coal  =3.3 percent,  average  stack  temperature =  139°C.
  Based on catch of EPA Method 5 sampling  train.
  Based on sum of catches of in-stack  filter and probe and dry filter of
  in-stack filter.
                                    93

-------
                                                          TABLE 2
                  SUMMARY OF EMISSION CONCENTRATIONS FOUND DURING SIMULTANEOUS IN-STACK AND EPA  METHOD  5
                                       EMISSION TESTS AT AN OIL-FIRED POWER PLANT3
                                                         (mg/dscm)
        Participate Concentration
    + 2H20 Concentration
                       c
Run
1
' 2
3
4
5
6
7
8
9
10
11
12
EPAb dry
44.8
51.0
43.2
42.5
43.6
26.8
25.5
35.1
49.9
48.6
28.1
27.1
In-stack
26.0
24.7
27.2
25.5
28.2
28.6
22.5
23.7
50.9
45.7
27.8
29.9
Total0
44.7
35.8
50.9
42.6
47.4
34.1
37.1

55.0

34.2
37.8
EPAb dry
30.9
33.8
4.5
3.9
5.8
13.2
4.5
3.4
20.1
18.3
12.0
12.5
In-stack
0.6
0.7
0.9
0.6
0.7
13.9
13.3
13.5
22.6
21.2
11.9
0.7
Tota
12.6
7.9
14.8
9.5
13.6
24.0
22.1
20.1
24.4
23.4
13.2
2.8
                                                Organic Concentration
                                               EPAb dry In-stack Total0
Average  38.8
30.1
42.0
13.6
8.4    15.7
19.5
 7.4
20.7
19.1
16.2
10.1
 5.6
15.0
17.3
12.1
 5.0
 6.1

12.8
                                                                              4.4
                                                                              6.4
                                                                              6.4
                                                                              9.5
                                                                              9.5
                                                                              6.5
                                                                              7.3
                                                                              4.7
                                                                             10.3
                                                                             14.7
                                                                              4.2
                                                                             10.9
                                                                 17.7
                                                                 17.0
                                                                 23.8
                                                                 26.5
                                                                 24.3
                                                                 10.3
                                                                 18.5

                                                                 14.2

                                                                  7.4
                                                                 16.9
                                                                Acidity(H^SO.Concentration
                                                                  EPAb dry In-stack  Total0
7.9   17.6
28.2
35.4
27.9
30.2
24.9
12.7
19.4
34.8
 3.4
 2.3
 3.0
 4.3

18.9
                                                                              4.3
                                                                              3.8
                                                                              3.2
                                                                              4.1
                                                                              9.2
                                                                              3.6
                                                                              3.8
                                                                              3.3
                                                                              1.2
                                                                              1.0
                                                                              1.5
                                                                              2.0
                                                                         17.5
                                                                         11.9
                                                                         17.0
                                                                         11.5
                                                                         20.1
                                                                         18.4
                                                                          9.8
                                                                          8.8
                                                                          2.9
                                                                          2.0
                                                                          5.4
                                                                          3.1
3.4   10.7
    Sulfur content of oil  =  0.29  percent, average  stack temperature = 168°C.
    Based on catch of EPA  Method  5  sampling  train.
    Based on sum of catches  of  in-stack  filter and probe and dry filter downstream of the in-stack filter.

-------
                                                          TABLE 3
                  SUMMARY OF EMISSION CONCENTRATIONS FOUND DURING SIMULTANEOUS IN-STACK AND EPA METHOD 5
                                        EMISSION TESTS AT A COAL-FIRED POWER PLANT3
                                                          (mg/dscm)
        Particulate Concentration
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
EPAb dry
205.1
159.3
133.0
265.6
147.4
310.4
265.6
279.5
216.9
204.6
275.6
220.1
207.2
In-stack
161.0
170.5
53.6
227.4
131.2
316.9
262.9
321.8
201.1
213.2
235.5
222.8
184.6
Total0
167.6
180.8
170.3
233.8
136.4
320.4
267.0
326.3
206.0
218.6
245.1
235.8
195.8
EPAb
3.3
3.9
5.7
6.9
5.1
5.5
8.0
5.7
4.6
5.7
9.9
5.7
4.2
Average 226.2     207.9     223.4
H2S04 + 2H20 Concentration
     dry  In-stack  Total0
5.7
            2.6
            1.8
            6.8
            2.6
            1.9
            2.8
            3.1
            1.5
            3.1
            4.8
            3.0
            1.9
            1.8

            2.9
4.8
2.1
9.3
4.4
2.2
4.4
5.1
3.3
5.2
6.6
5.9
4.1
3.6

4.7
                            Organic Concentration    Acidity(H?SOJConcentration
                              K                   *+       L«   ^   »            _
                           EPA1
   dry  In-stack Total0
12.7     32.0   34.4
12.7     12.4   16.6
 9.5      7.8   18.8
17.1      8.0   10.5
13.6      8.3   10.4
12.5     10.8   12.3
17.0     13.2   15.1
10.7     23.4   25.0
16.1     17.6   19.7
 7.4     15.2   17.6
14.3     12.5   13.8
 7.2     11.0   13.2
10.7      2.9    4.5
                                                                                              EPA  dry In-stack Total
12.4
13.5   16.4
    Sulfur content of coal  =  0.85  percent,  average  stack  temperature =  199°C.
2.5
1.8
1.8
2.9
3.3
2.5
2.4
3.9
3.3
4.7
6.6
3.5
2.6

3.2
1.8
1.7
1.8
2.0
2.9
1.4
2.4
2.4
2.4
2.4
1.8
1.8
1.8

2.0
    Based on catch  of EPA Method  5  sampling  train.
    Based on sum of catches  of in-stack  filter and probe and dry filter downstream of the in-stack filter.
3.6
3.4
3.7
3.9
5.9
2.9
4.0
4.1
4.0
4.0
4.6
3.5
3.7

4.0
VO
01

-------
                                                          TABLE 4
                   .SUMMARY OF EMISSION CONCENTRATIONS FOUND DURING SIMULTANEOUS IN-STACK AND EPA METHOD 5
                                        EMISSION TESTS AT AN OIL-FIRED POWER PLANT3
                                                          (mg/dscm)
Particulate Concentration
Run
1
2
3
4
5
6
7
8
9
10
11
12
Averai
EPAb dry
245.3
101.9
170.9
68.6
203.3
180.6
419.8
171.4
211.3
432.0
279.8
126.9
ge 217.6
In-stack
52.4
314.4
425.7
390.7
482.5
483.8
476.7
525.6
425.7
632.8
644.9
200.7
421.3
Total0
167.8
825.6
868.0
603.1
921.0
618.7
905.7
713.5
788.8
1042.8
930.8
340.7
727.2
H9SOA + 2H90 Concentration
£ *f £•
EPAb dry
102.6
88.0
173.4
53.8
178.8
78.5
110.8
112.6
107.8
127.5
159.5
136.8
119.2
In-stack
7.3
9.0
21.5
8.0
27.2
20.3
16.5
18.0
19.3
33.0
30.0
11.3
18.4
Total0
166.0
25.4
36.2
179.0
30.5
24.9
19.2
21.8
40.4
54.4
48.7
19.3
55.5
Organic
EPAb dry
62.2
37.4
58.8
34.3
70.4
40.0
78.5
63.7
79.9
68.7
65.2
69.1
60.7
Concentration Acidity (FLSOJConcentration
In-stack
6.4
74.5
102.6
124.8
141.6
132.1
209.1
143.5
117.6
162.5
142.6
93.7
120.9
total0
74.9
88.9
111.0
133.2
153.9
139.6
233.7
158.1
128.2
171.2
156.9
101.7
137.6
EPAb dry
30.3
27.5
46.1
21.8
43.6
27.4
53.3
42.1
47.3
17.1
41.9
40.3
36.6
In-stack
11.1
7.8
11.2
8.1
14.3
11.6
12.3
8.7
9.6
14.5
5.5
8.2
10.2
Total0
55.3
30.2
30.2
57.1
33.5
21.1
30.7
19.2
32.7
34.7
20.8
12.4
31.7
a  Sulfur content of oil = 2.45 percent, average stack temperature = 60"C.
b  Based on catch of EPA Method 5 sampling train.
c  Based on sum of catches of in-stack filter and  probe and dry filter downstream of in-stack filter.
10
en

-------
                                                         TABLE 5
                                RESULTS OF t-TEST COMPARISONS OF CONCENTRATION DETERMINED
                          FROM IN-STACK AND EPA METHOD 5 SAMPLING AT FOSSIL FUEL POWER PLANTS
Facility
   A
   Participate Mass
Method 5a > In-stackb
Method 5  = Total0
Sulfate (H2S04 + 2H2
 Method 5 > In-stack
 Method 5 = Total
      Organic
  Acidity(H2S04)
                Method 5
                Method 5
          > In-stack
          = Total
 Method 5 = In-stack
 Method 5 = Total
Method 5 > In-stack
Method 5 < Total
Method 5 > In-stack
Method 5 > Total
                Method 5
                Method 5
          = In-stack
          = Total
                Method 5  < In-stack
                Method 5  < Total
 Method 5 > In-stack
 Method 5 = Total

 Method 5 > In-stack
 Method 5 > Total
Method 5 = In-stack
Method 5 = Total

Method 5 < In-stack
Method 5 < Total
Method 5 > In-stack
Method 5 < Total

Method 5 > In-stack
Method 5 - Total
   a  Based on participate catch of EPA Method 5 dry sampling train.
   b  Based on participate catch of in-stack filter.
   c  Based on sum of participate catch of in-stack filter and probe and filter downstream of in-stack filter.
VO

-------
               EPA METHOD 5 SAMPLE TRAIN CLEAN-UP PROCEDURES
                              Clyde E. Riley*
Introduction
     In the performance of participate source emission tests, an
important procedure affecting the accuracy is sample recovery.  Accurate
results are not possible unless proper procedures are conscientiously
applied in recovering and quantitatively transferring particulate matter
from the sample train to the storage  container.  Often, however, these
procedures receive only minimum attention.   Well-trained and highly
experienced technical staff are normally employed to design and oversee
the performance of a test and the writing of a  test report while, in
contrast, the least experienced personnel are often assigned sole
responsibility with limited guidance  for the recovery of sample from the
train—a task which includes a high potential for producing significant
errors.

     The accuracy of sample recovery  procedures are, of course, not only
dependent on  the ohysical  transfer of sample from the train to the storage
containers; the procedures  also  involve  the  selection of proper equipment,
use of  proper materials,  application  of  proper  cleaning, handling, and
shipping techniques, and  an overall awareness of the importance of each
phase of the  sample handling  procedure.  The following guidelines describe
procedures which are employed by  the  Emission Measurement Branch to assist
in minimizing sources  of  error  in  EPA Method 5  sample train cleanup.  These
are presented here, both  to  call  attention  to the degree of detail which must
be considered in sample recovery,  and to make the procedures available to
others engaged in source sampling.  These guidelines do not include techniques
* Emission Measurement Branch, ESED,  OAOPS,  EPA, RTP, NC
  Published in Stack Sampling News  3(1): 4-7, July 1975
                                      98

-------
for analysis of the impinger catch or any specific procedures other
than those necessary for analyzing the sample for mass only.  Also, it
should be noted that these procedures are not regulatory requirements;
rather, they are procedures                    to be used by contractors
employed by the Emission Measurement Branch.  Although these procedures
reflect the collective experience gained by EMB in the conduct of several
hundred source tests, we recognize that other source sampling groups may use
different cleanup techniques.  It is hoped,  therefore, that this publication
will provide the impetus to others to publish such alternate or improved
techniques.
Pretest Preparation
     1.  Brushes and sample  recovery support items shall be properly
cleaned and enclosed in dust-free packaging before being used in the
sample recovery operations.  This includes the sample containers as well
as  the sample collector glassware.
     2.  Sample containers to  be used for the liquid samples shall be
Type I, chemically-resistant,  borosilicate narrow-mouth glass bottles
(500 mis. or 1000 mis. size).  Screw-cap closures with Teflon  rubber-
backed liners shall be used  on all such sample containers.  Use of any
other type liquid sample container, closure, or liner shall be verified
acceptable prior to use.
     3.  Glass or plastic petri dishes shall be used to contain the filter
samples, unless otherwise specified by EPA.
     4.  Pre-weighed indicating silica yel shall be acceptable only if  the
containers are completely full and tightly sealed.
  Trade  Name
                                       99

-------
                        2
     5.  Only fresh ACS  reagent  grade  chemicals shall be used for sample
cleanup and recovery.
     6.  All reagents and samples shall  be  stored in sealed, non-contaminating
containers.  This includes  acetone which shall  be purchased and stored in
glass containers.  Only acetone with  blank  values less than 0.001 shall be
acceptable for sample recovery operations.
     7.  If water is required for cleanup of  the probe and filter assembly,
it shall be distilled and stored  in non-contaminating containers.
Sample Recovery
     1.  Proper sample recovery procedure begins as soon as the probe is
removed from the stack at  the  completion of the sampling period.  When the
probe can be safely handled, wipe off all external particulate matter near
the tip of the probe nozzle and place a cap over the nozzle tip.  Do not
cap off the probe tip tightly  while  the sampling train is cooling as this
will create a vacuum in  the filter holder, thus drawing water from the
impingers into the filter  holder.
     2.  Before moving the sample train to the cleanup site, remove the probe
from the sample train, inspect for condensed water, wipe off the silicone
grease, if used, and cap the  open end of  the probe.  Be careful not to lose
any condensate.  Wipe off  the silicone grease from the filter inlet where
the probe was fastened and cap it loosely.  Remove the umbilical cord from
the last impinger and cap  the impinger opening.   If a flexible  line is used
between eicher  the  first impinger or condenser and the filter holder, dis-
connect the  line at the  filter holder and  drain any condensed liquid into the
 2
  American Chemical  Society
                                        100

-------
impingers or condenser and remove the line from the impinger.   After
wiping off the silicone grease, cap off the filter holder outlet and
impinger inlet and the flexible line, if used.  Either ground  glass
stoppers or their EPA approved equivalent may be used to close these
openings.
     3.  Transfer the probe and filter-impinger assembly to the cleanup
area.  Exercise care in moving the collection train from the test site to
the sample cleanup area to avoid the loss of collected sample or the gain of
extraneous particulate matter.  This area shall be clean and protected from
the wind to minimize the chances of contaminating or losing portions of
the sample.
     4.  Prior to sample cleanup and during disassembly, an inspection shall
be made  of the individual components of the sample collector.  This
inspection should reveal whether or not the sample collector was functioning
properly.  Also by observing  the quantity of  sample, it can be estimated  if
a sufficient  amount of matter has  been collected  for proper analysis.
Record any  items  that could possibly affect the  results  (e.g., cracked
or broken glassware, water  in the  filter  holder,  unexpected residue, spent
silica gel).   State whether or not the sample is  still  valid  and give  basis.
      5.  A  consistent  procedure shall  be  used for the  sample  collector
disassembly  and cleanup.   The following order is  recommended:
          General
               a.   The sample  containers  shall be  tightly  capped after  the
 sample recovery operation.   The closure  caps  shall be sealed  to thp "arrow-
 mouth containers  with shrink bands, plastic  tape, or their equivalent.
                                     101

-------
 The glass petri dishes shall be sealed around their circumference with
 large rubber bands and secured with plastic tape or its  equivalent.
              b.  All samples including blanks shall be assigned  individual
 identification numbers by using pre-numbered EPA sample  identification labels.
 Where more than one container is needed to contain a given  sample, each
 additional container shall  be assigned the same basic identification number.
 All  such  multiple containers shall  be further marked to  indicate  the total
 number of containers used for that  sample and which container of  the series
 each  represents (examples 1  of 3, 2 of 3, etc.).
             c.   After the  recovery operation,  the volume of all  liquid
 samples including rinses  shall  be documented either by using graduated sample
 bottles and  recording  the sample volume on the  recovery  sheet or  by permanently
marking the  sample container and/or label  to indicate the liquid  level.   By
doing  this,  the  laboratory will  be  able to determine whether or not sample
 leakage occurred  during transport.
             d.   A 200 ml blank  reagent sample  shall be collected for each
lot of rinse reagents  used.   Representative  blank  samples of the acetone  or
other solvents, distilled water  (if used), and  preweighed filters (quantity
three) shall be collected during the test  program.  The acetone and water
samples shall be  analyzed to determine  the amount of contamination attributed
to the sample reagents.
                                     102

-------
         Filter - Remove the filter holder and inspect the  filter mat
for punctures or tears before removing and placing it in  an identified
glass or plastic petri dish container.  Use a pair of parallel  tweezers
and/or clean disposable surgical  gloves to handle the filter.   If it is
necessary to fold the filter, do so such that the particulate  cake is
inside the fold.  Quantitatively remove any particulate matter and/or
filter media which may adhere to the filter holder or support  by carefully
using a dry nylon bristle brush,  rubber policeman, or a sharp-edged blade.
Place this matter into the same container as the filter.   Seal  the container
as described in the General Section.
         Probe - It is recommended that two people be used  to  clean the
probe to minimize altering the sample.  The probe cleanup and  disassembly
shall be conducted in the following order.  Making sure that dust on the
outside of the probe or other exterior surfaces does not  enter into the
sample, quantitatively transfer the particulate matter and  condensete from
the probe nozzle, probe fitting, probe liner, and front half of the filter
holder to container No. 2.  Rinse these components with acetone, distilled
water (if required), or other appropriate rinsing solvents  that have been approved
                                     103

-------
by EPA.  In all cases,  collect  a representative  blank of the rinse
solvents.  Specific  steps  are as follows:
              a.  Carefully  remove  the  probe  nozzle and clean the inside
surface by triple rinsing  with  acetone  from a glass wash bottle and brushing
twice with a precleaned nylon brush.  Continue brushing until the acetone
rinse shows no visible  particles, after which perform a final rinse of the
inside surface with  acetone.
              b.  Brush and  rinse with  acetone the inside parts of the
probe fitting in a similar way, i.e., until the  rinse shows no visible
particles remaining.
              c.  Rinse the  probe liner with  acetone by tilting and
squirting acetone into  its upper end, while rotating the liner in a 360°
manner so that all inside  surfaces will be rinsed.  Let the acetone drain
from the lower end into the  sample container.  A second acetone rinse shall
be performed with the aid  of a  probe brush.   Position the liner as before
and squirt acetone into the  upper end while pushing the brush through the
entire length of the liner using a twisting action.  Repeat the brushing and
rinsing operation (minimum two  times) until no particulate matter remains
in the probe liner upon visual  inspection.  With stainless steel  or other
metal liners, brush  and rinse in the above prescribed manner at least six
times; metal liners  have small  crevices in which particulate matter can be
entrapped.  Upon completion  of  the brushing and  rinsing operation, rinse the
brush with acetone and  perform  a final  acetone rinse on the liner.  Collect
these rinsings in the same sample container as before.
                                    104

-------
               d.  After ensuring that the filter holder has been
wipped clean of silicone grease, clean the inside of the front half of
the filter holder by double brushing with a nylon bristle brush while
rinsing with acetone or brush and rinse until all visible particulate is
removed.  Make a final rinse of the brush and inside surface of the front
half of the filter holder.  Again these rinsings are placed in the No. 2
sample container.  (Note:  Do not rinse or brush the fritted-glass support.)
         Silica Gel  - Record the color and condition of the indicating
silica gel in the last impinger and determine if it is completely saturated.
Weigh the used silica gel to the nearest 0.5 gm and determine the amount
of moisture collected.  The silica gel shall  be transferred to a shipping
container or discarded if contaminated.
         Impinger Catch - If analysis of the  impinger catch is not required,
discard the liquid after measuring and recording the volume or weight.
                                   105

-------
TECHNICAL REPORT DATA
'/J'i'tjc reau liKlnicnru'i i»» ill
1 REPCPT NO |2
£PA-450/2-78-042d |
4 T'T._E AND SUBTITLE
Stack Sampling Technical Information: A Col
of Monographs and Papers Volume IV
7 AUTHOR(S)
Roger T. Shigehara (Editor)
9 PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Environmental Protection Agency
i -i iffji be for.; co' '


lection



Emission Standards and Engineering Division
Emission Measurement Branch
Research Triangle Park, NC 27711
12 SPONSORING AGENCY NAME AND ADDRESS
Same as above.




15 SUPPLEMENTARY NOTES
16 ABSTRACT

"Stack Sampling Technical Information" is a four-volume
and papers which have been compiled by the
. ii-if}

3 WtC'PicNT S ACCESSION NO
5 REPOFtf^DATc

October 1978
6 PERFORMING ORGANIZATION CODE
8 PERFORMING ORGANIZATION REPCRT NO


10 PROGRAM ELEMENT NO


11 CONTRACT/GRANT NO




13 TYPE OF REPORT AND PERIOD COVERED
14 SPONSORING AGENCY CODE



collection of monographs
Emission Measurement Branch, ESED, OAQPS.
The information specifically relate to current EPA test
methods and compliance
test procedures. The data presented in some of these documents have
served as
the basis for a number of revisions made in the EPA Reference Methods 1 through 8.
Several of the documents are also useful in determining
procedures .

acceptable al

ternative

17 KEY WORDS AND DOCUMENT ANALYSIS
J DESCRIPTORS
Gas Sampling
Filtered Particle Sampling
Gas Analysis
1R 3 ;-RiB'_|T,r>N ST/l'LMC WT
Unlimited

b IDENTIFIERS/OPEN ENDED TERMS
Stack Sampl
ing
19 SECURITY CLASS !lln\ Report i
Unclassified


20 SECURITY CLASS iTIni pugei
Unclassified
c COSATl 1 icId/Croup
14B
14D
21 NO OF PAGES
118

22 PRICE

£Pi For-, 72?"  ! iR- «  4-~
                                                      '106

-------