U.S. Environmental Protection Agency Industrial Environmental Research
Office of Research and Development Laboratory
Research Triangle Park, North Carolina 27711
EPA-600/7-78-011
January 1978
MEASUREMENT
OF HIGH-TEMPERATURE,
HIGH-PRESSURE PROCESSES:
Annual Report
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide'range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsemeht or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-011
January 1978
MEASUREMENT
OF HIGH-TEMPERATURE,
HIGH-PRESSURE PROCESSES:
Annual Report
by
Larry Cooper
Aerotherm Division/Acurex Corporation
485 Clyde Avenue
Mountain View, California 94042
Contract No. 68-02-2153
Program Element No. EHE623 and 624
EPA Project Officer: William B. Kuykendal
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY
APPENDICES . . .
A. Measurement of Velocity, Pressure and Temperature in HTHP A-i
Process Streams, V. Kulkarni and L. Cooper, Aerotherm
Report TN-76-34
B. Material Selection for Sampling in Coal Conversion Systems, B-i
J. Hull, Aerotherm Report TM-77-180
C. Tar Sampling in Coal Gasification Processes, W. Krill, C-i
Aerotherm Report TM-77-173
D. Field Testing of a Sampling System for High-Temperature/ D-i
High-Pressure Processes, W. Masters, Aerotherm Report
TM-77-177
E. IERL-CRB Sampling Manual for Level 1 Environment Assessment, E-i
W. Krill, Aerotherm Report TM-77-160
F. A Progress Review of Source Assessment of Sampling System p-i
Programs, D. Blake, Aerotherm Report TM-77-176
G. Conversion to International System of Units G-l
iii
-------
EXECUTIVE SUMMARY
Aerotherm Division of Acurex Corporation is conducting a program for
Measurements of High-Temperature/High-Pressure (HTHP) Processes for the
Process Measurements Branch of IERL under a 3-year term level-of-effort
contract. The planned level-of-effort is 4 man-years per year. This report
summarizes the activities of the first year of the program.
Under this contract, Aerotherm is evaluating and developing methods
for measuring process streams for particulates and harmful gaseous
components. These measurements assess the effluent cleanup requirements and
indicate techniques which minimize the generation of toxic substances that
require subsequent cleanup.
The tasks accomplished under this program were organized under three
categories:
• State-of-the-art review
• Research and development
• Support services
Reports have been prepared under each task summarizing the work performed
during the past year. Abstracts of these reports are given below; the
complete reports are included in the Appendix.
STATE-OF-THE-ART REVIEW
Measurement of Velocity, Pressure and Temperature in HTHP Process
Streams, V. Kulkarni and L. Cooper, Aerotherm Report TN-76-34
To determine particulate and gaseous concentrations in high-
temperature, high-pressure process streams, gas flow velocities, pressures,
and temperatures must be accurately measured. Various methods have been
reviewed to determine which methods of measurement are most suitable.
The pitot static probe accurately measures static and dynamic
pressures, and with the gas density can be used to determine velocity. The
probe has shortcomings (e.g., plugging by dust), although these may be
overcome by using purging techniques. The "S" type pitot static probe is
perhaps the best solution for the near future. This probe resists plugging
-------
much longer than ordinary probes. However, to use this probe, a calibration
factor dependent on the Reynolds number may be required to accurately
determine velocity.
Other devices are also available for measuring velocity. These
devices include flare gas probes, drag meters, and fluidic sensors. In
addition, the application of other types of equipment such as orifice and
venturi meters, as well as laser velocimeters is still being evaluated.
For measuring temperature, thermocouples are still the simplest and
most convenient devices available. The shielded thermocouple is likely to
give the best results in the immediate future. Other methods, such as
radiation pyrometers, molecular beam, and velocity of sound techniques still
have major drawbacks. More study including tests in HTHP environments would
be required to determine whether another method should replace the
thermocouple.
Material Selection for Sampling in Coal Conversion
Systems, J. Hull, Aerotherm Report TM-77-180
To investigate prospective materials for coal conversion system
sampling, Aerotherm made an extensive literature search and direct contacts
with persons active in material research. Aerotherm had access to the latest
information and data in this area, since one of our staff (the author of this
report) is a member of the Metal Properties Council and its Subcommittee on
Materials for the Gasification of Coal.
Sufficient laboratory and pilot plant data has been obtained to help
designers select materials that will survive for several hundred hours in the
environment found in coal conversion systems. For example, at 18000F, adding
at least 25-percent chromium establishes corrosion resistance in gasifiers
for 1000 hours. And while there are no safe alloys for extended use in high
sulfur gas above 1600°F, both Alloy 310 and Kanthal are acceptable below
16000F.
When selecting materials, the local environment surrounding each probe
is important, since temperature, pressure, gas composition, and flyash
characteristics often vary with probe location. In addition, surface finish
and other fabrication parameters such as heat treatment are also important to
material survival.
Tar Sampling in Coal Gasification Processes, W. Krill,
Aerotherm Report TM-77-173
Tars are a byproduct of many coal gasifiers. With decreases in
temperature, tars condense on surfaces often restricting the flow. Since
tars are chemically stable, it is hard to remove them from equipment. In
addition, some tars are highly carcinogenic and should not be released to the
environment. All of these factors make it difficult to extract particulate
and gaseous samples in the presense of tar.
-------
A comprehensive study was made to determine the best way to sample
under these conditions. Brian Macnick, a UCB researcher under ERDA contract,
"Physical Characteristics of Tars," was used as a consultant in this study.
The gasifier facilities at Grand Forks Energy Research Center, Riley-Stoker,
Blue Diamond Co., and Morgantown Energy Research Center were contacted to
relate their specific experience with tars.
While tars have widely varying condensation temperatures, they tend to
remain in the vapor state at temperatures above 160QOF. Thus, particulate
can be separated out if the temperature is high enough. In addition, higher
process pressures will result in fewer (if any) species condensing at a given
temperature. Of all the equipment used for tar extraction (filters,
cyclones, condensers, scrubbers, and electrostatic precipitators), the
electrostatic precipitator (ESP) is judged best. The electrostatic
precipitator collects tar particles of all sizes, while other collection
methods cannot collect the smaller particles. Also, the ESP handles a large
collected mass better than other methods.
To demonstrate the applicability of the electrostatic precipitator,
Aerotherm derived a conceptual design for a compact two-stage collector (6
inches in length, 7 inches in diameter), based on a Southern Research
Institute concept. The collector has been calculated to be 100 percent effi-
cient for all particles above 1 micron, 99.6 percent efficient at 1 micron,
and 99.2 percent efficient at 0.3 micron (tar particles can be as small as
0.1 micron).
Collecting solid particulate at stream temperatures avoids additional
condensation of tars. Therefore, a possible sampling scheme for a
gasification process begins by collecting solid particles at stream tempera-
tures with a particulate cyclone, followed by an appropriate reduction in
temperature and pressure using a condenser and then tar collection in an
electrostatic precipitator.
RESEARCH AND DEVELOPMENT
Field Testing of a Sampling System for High-Temperature/
High-Pressure Processes, W. Masters, Aerotherm Report
TM-77-177
The Aerotherm HTHP sampling system for measuring particulate
concentration and trace element content has recently been successfully demon-
strated in field tests at the Exxon Pressurized Fluidized Bed Combustor
(PFBC) Miniplant in Linden, N.J. The probe is of a hydraulically operated,
telescoping design for operation in a severe, high-pressure FBC environment.
The versatility of the probe was demonstrated by the fact that the
three sampling tests at Exxon used two types of particulate collectors: the
first test used a glass fiber thimble filter with large total mass capacity,
and the last two tests used a cascade impactor with seven stages for partic-
ulate sizing. Sample data obtained from the 136QOF, 9-atmosphere gas stream
demonstrated that the system can operate in an extreme PFBC environment. The
-------
only problem encountered was a malfunctioning impactor heater. This heater
was replaced in preparation for subsequent tests to be performed at Exxon.
The ability of the system to measure particulate loadings and size
distributions were fully demonstrated in these tests. Particulate size
distributions were obtained from the impactor runs, and most particulates
were between 1 to 10 micrometers. The chemical composition of collected
particulate was also analyzed using a dispersive X-ray fluorescence analyzer.
The results of these tests and the subsequent tests will be fully detailed in
a separate test report to be issued under EPA cover.
SUPPORT SERVICES
IERL-CRB Sampling Manual for Level, 1 Environmental Assessment,
W. Krill, Aerotherm Report TM-77-160
The IERL Combustion Research Branch plans to conduct a Level 1
environmental assessment at their combustion facilities at Research Triangle
Park, North Carolina. Drawings were prepared by Aerotherm showing facility
modifications which would allow a special sampling probe ("coke oven" probe)
to be used. These drawings will be used by IERL-CRB to modify hardware at
their facility. This manual includes instructions on test planning,
preparation of sampling systems, obtaining samples, recovery of samples from
sampling equipment, and data reduction.
A Progress Review of Source Assessment of Sampling System
Programs, D. Blake, Aerotherm Report TM-77-176
Several tasks for using and improving the Source Assessment Sampling
System (SASS) were undertaken in the past year. In one task, Aerotherm
supported two SASS test observations. The first observation, at an
experimental coal-fired boiler at K.V.B., Inc., established the need for
several minor modifications. In the second, Radian Corporation used the SASS
at the Comanche Power Plant of the Colorado Public Service Co. This test, at
an increased flowrate, established the need to incorporate two vacuum pumps
instead of one.
In another task, the performance of the three SASS cylcones was
experimentally evaluated at SoRI using "real" solid particles at 205°C.
Since the dye solution particles used at SoRI were not adequate for this
temperature, a different method of cyclone calibration was devised. In this
method, a test dust of known properties is used to generate a dust cloud
which is fed into the cyclone. The efficiency of the cyclone is measured
from the quantity and size distribution of dust collected in the cup and
passed by the cyclone to the filter.
Well-characterized silica dust was used in the first experiments, and
size distributions were measured using the X-ray sedograph at EPA/RTP.
Inconsistent results from these tests showed that the test dust degraded.
The easiest way to correct for this was to find a test dust that would not
erode. A spherical aluminum powder material was chosen to solve this
-------
problem, while making the data analysis easier because of the simple shape.
The possible problem introduced by the low melting point of aluminum (which
could cause particles to stick to walls) can be solved by lengthening the
heater to reduce wall temperatures.
A second series of tests has shown reasonable results, but no size
distribution analysis (needed to calculate cyclone efficiency curves) has
been performed yet.
-------
Appendix A
Acurex Corporation/Aerotherm Division
October 1976
Aerotherm Project 7237
MEASUREMENT OF VELOCITY,
PRESSURE AND TEMPERATURE IN
HTHP PROCESS STREAMS
V. Kulkarni
L. Cooper
AEROTHERM REPORT TN-76-34
Prepared for
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park
North Carolina 27711
Contract 68-02-2153
A-i
-------
T""L£ OF Di:.'T:!JS
Section -
1 INTRODUCTION A-l
2 GENERAL CONSIDERATIONS A-2
2.1 Operating Environment A-2
2.2 Accuracy A'3
2.3 Operating Characteristics A-6
3 VELOCITY AND PRESSURE MEASUREMENT A'11
3.1 General Considerations A-12
3.1.1 Compressibility of Gas A-12
3.1.2 Reynolds Number A-13
3.1.3 Turbulence A-l 5
3.1.4 Solid Particles in Suspension A-l5
3.1.5 Orientation of Probe and Flow Angularity Measurement A-16
3.2 Instruments for Velocity Measurement A-16
3.2.1 Pitot/Static Probe A-19
3.2.2 Disc/Static Heads A-21
3.2.3 "S" Type Pi tot Tube A-21
3.2.4 Other Special Probe Designs for Measurements in Dust Laden Gases A-21
3.2.5 Flare Gas Probe A-24
3.2.6 Drag Meter A-28
3.2.7 Fluidic Sensors A-30
3.2.8 Acoustic Velocimeter A-30
3.2.9 Orifice and Venturi Meters A-30
3.2.10 Laser Velocimeter A-32
4 TEMPERATURE MEASUREMENT A-34
4.1 Temperature Measurement in Gas Flow A-34
4.2 Temperature Measurements with Thermocouples A-35
4.2.1 Convection A-35
4.2.2 Radiation A-36
4.2.3 Conduction A-38
4.3 Other Temperature Measurement Devices A-38
4.3.1 Radiation Pyrometers A-38
4.3.2 Molecular Beam A-39
4.3.3 Velocity of Sound Method A-39
5 CONCLUSIONS A-40
APPENDIX A - BIBLIOGRAPHY A-41
APPENDIX B - COAL COMBUSTION PRODUCT ANALYSIS A-43
A-ii
-------
LIST OF ILLUSTRATIONS
Page
2-1 Examples of Anisokinetic Sampling A-5
2-2 Effect of Anisokinetic Sampling on the Measurement of Concentration A-7
2-3 Circular Cylinder in a Steady, Uniform Flow Field A-9
2-4 Velocity Disturbance with Distance, 6 = 90° A-10
3-1 Correction Factor k for Low Reynolds Numbers for a Prandtl Pitot Tube A-14
3-2 Influence of Pitot/Static Probe Orientation on Measured Pressure A-17
3-3 Flow Angularity Determination Using a Cone Flow Probe A-18
3-4 Typical Pitot/Static Probe A-20
3-5 Disk/Static Head A-22
3-6 "S" Type Pitot A-23
3-7 Special Probes for Measurement in Dust Laden Gases A-25
3-8 Purge Flow Pressure Sensors A-26
3-9 Flare Gas Probe A-27
3-10 Drag Meter A-29
3-11 Fluidic Sensor A-31
3-12 Schematic of the Interference Between Two Traveling Collimated Coherent Beams of
Light of Wavelength X Passing Through Each Other at an Angle 6 A-33
4-1 Relation Between Reynolds and Musselt Numbers for Flow Over a Thermocouple .... A-37
B-l Equilibrium Combustion Products from Typical Eastern Coal A-44
B-2 Ratio of Specific Heats of Combustion Products from Typical Eastern Coal A-46
B-3 Molecular Weight of Combustion Products from Typical Eastern Coal A-47
A-iii
-------
SECTION 1
INTRODUCTION
/
Accurate determination of gas flow velocities, pressures, and temperatures is essential for
measuring particulate and gaseous concentrations in high temperature, high pressure process streams.
Temperature measurement is required in velocity determination and along with pressure measurement
serves to characterize the gas flow being sampled. In this memorandum, the various aspects and methods
of velocity, pressure, and temperature measurements in a high temperature, high pressure, dust laden
environment are examined. This study has been limited to a survey of the various methods available.
The reader is referred to the bibliography (Appendix A) for further depth on the subject matter.
Some general considerations for the required measurements are presented in Section 2. Velocity,
pressure, and temperature measurement techniques are examined in Sections 3 and 4. Finally, conclusions
are given in Section 5.
A-l
-------
SECTION 2
GENERAL CONSIDERATIONS
In this section several aspects of the velocity and temperature determination in a high tempera-
ture, high pressure environment are discussed. In Section 2.1, the problems arising from operating
in a severe environment are presented. The need for accurate determination of velocity, pressure, and
temperature is discussed in Section 2.2. Finally, the operating characteristics of the instrumentation
are examined in Section 2.3.
2.1 OPERATING ENVIRONMENT
The operating environment for the high temperature, high pressure sampling systems of interest
are assumed to be as follows:
• Maximum operating temperature = 2000°F*
• Maximum operating pressure = 1500 psia
• Maximum velocity = 300 ft/sec
• Particulate size range = 0.2 to 7 microns
• Particulate loading = up to 0.2 grains/scf
t Highly corrosive atmosphere due to the presence of hLS, SO,, chlorine and flourine com-
pounds
• Highly erosive atmosphere due to high velocity particulate loading
• The typical pipe diameter where measurements' will be made is assumed to be 4 to 10 inches.
In addition, the stream conditions may be somewhat nonuniform or unsteady in some instances as a
result of large obstructions such as valves or bends in the flow. Under some operating conditions the
possibility exists that the instrumentation can become coated with materials such as slag and tars.
Tar condensation is most likely to occur at lower temepratures (800°F to 1200°F). Therefore, if the
environment is particularly "dirty", some means of keeping the measuring ports clear must be provided.
*
For the most part English units are shown which are customary to sampling technology See
Appendix B for appropriate conversion factors for metric units.
A-2
-------
At the high temperatures (2000°F) and pressures (up to 100 atmospheres), is is important to
maintain the structural integrity of the components since material strength is reduced at high tem-
peratures and loads increase in proportion to the pressure. The environment is dust laden and
highly corrosive and erosive. The presence of SO.,, flourine and chlorine compounds, alkali metal
compounds, and water at high temperatures, combine to produce an environment which vigorously attacks
most metals. The combination of high velocities (-300 ft/sec) and large dust loadings (-0.2 gr/scf)
can cause significant erosion (mechanical removal of material from surface) over a relatively short
period of time. The materials for the instrumentation should be chosen with these considerations in
mind. High strength stainless steels, inconels, and hastalloys are among those metals which seem
promising from the standpoint of survival. As mentioned previously, the ports in the instruments are
liable to be blocked in the "dirty" environment. Larger ports could be used, however these perturb
the the flow, significantly, and give less accurate results. Blockage by dust particles can be avoided
by periodic purging of the ports with a compressed gas. In this case, care must be taken to protect
the pressure transducers or manometers during blowing. After blowing, sufficient time must be allowed
for the system to achieve equilibrium before a measurement is taken. Analytical methods* have been
developed to predict the transient response characteristics of pressure instrumentation.
2.2 ACCURACY
The measurement of gas velocity, pressure, and temperature is required to set isokinetic sampling
rates and determine the thermodynamic state of the gas. The instrumentation should have sufficient
accuracy to provide valid and repeatable data. The measurements should be reasonably accurate over
the whole operating range. It is preferrable to avoid individual calibration of each probe at several
operating conditions. For example, standard pi tot tubes and thermocouples are attractive because
they are well characterized and there is no need for individual calibration.
To ensure collection of a representative sample of particles suspended in a moving gas stream,
it is necessary that the flow pattern upstream of the sampling probe be unaffected by the presence of
the probe. If the flowrate per unit area through the sampling probe is lower that the freestream flow-
rate per unit area, the streamlines ahead of the probe will diverge. This upstream influence is
typical of low speed (subsonic) flows. On the other hand, if the suction flow through the sampling
probe is set too high, the streamlines ahead of the probe will tend to converge. In either event,
*Aerotherm has recently programmed a method for evaluation of pressure line response times.
A-3
-------
particles contained in the flow will not follow the streamlines due to their inertia and, consequent1y»
sampling errors will occur. It has also been shown that errors in gas concentration measurements can
result from deviations from the isokinetic condition.*
These adverse effects can be avoided by setting the plane of the sampling orifice normal to
the direction of the flow and adjusting the flowrate of the sampler so that velocity into the probe
at the entrance plane is equal to that of the undisturbed stream. When this is done, sampling is said
to be isokinetic, and a representative sample enters the probe. Therefore, there is need for accurate
measurement of stream velocity in order to determine the proper suction rate through the sampling
probe.
As mentioned previously, errors associated with anisokinetic sampling are due to particle inertia.
If the sampling velocity is less than the stream velocity (Figure 2-1 (a)). Some particles originally
contained in the gas stream that pass around the probe are projected into the flow entering the probe,
causing the concentration to be overestimated. The opposite effect occurs (Figure 2-1 (b)), when the
sampling velocity exceeds the stream velocity.
When the sampling is not isokinetic, that is, the sampling velocity, V , differs from the
velocity, V of the gas stream, the measured concentration of the particles, C , defined as the ratio
of the weight of the particles to the volume of the gas drawn in, can be written as:
c* va
^-= a^+ (1 - a)
& S
Where C is the actual concentration and a is a parameter depending on the particle character
a
(shape, size, density), on the gas density and the pattern of the streamlines in the neighborhood of
the probe. If the particles are very small, a ->• 0, and C •*• C , regardless of the rate of suction,
which amounts to saying that very small particles follow the streamlines and behave like gas molecules.
On the other hand if the particles are very large, a -> 1
Multiplying this equation by _ s_, where S is the area of suction, gives:
SVa
See Reference 6.
A-4
-------
Particle
trajectory
cn
Streamline
Particle
trajectory
Streamline
Figure 2-1. Examples of anisokinetic sampling.
-------
CSV
CaSVa
s . ,
"
Where MS is the mass of the particles collected per unit time and Ma is the actual mass of the particles
in the stream.
Thus, in the case of very large particles, the mass collected does not depend upon the rate of
suction because large particles do not deviate from the original direction of the flow due to their
large inertia. The particulate concentration in this case is found by dividing this mass by the actual
volume flowrate. Figure 2-2 shows the effect of anisokinetic sampling on the measurement of particulate
concentration.
In the intermediate case, when the particles are neither too big nor too small, i.e., 0 < a < 1,
either isokinetic sampling should be ensured or the value of a should be approximated. An expression
for a is given by:
,
a = 1 -
LA
where: A = uV /g
u = Velocity of free fall of particles in an undisturbed flow
L = Length of purturbation upstream of the probe (function of probe geometry)
In the present application, the maximum particle size expected is 7 microns. For C./C =
S a
1.5 (i.e., 5 percent error in concentration measurement), V,/V = 1.3 (i.e., 30 percent error in
a S
isokinetic sampling condition can be tolerated). Thus, given the required accuracy limit on concen-
tration, the accuracy requirement on velocity, pressure, and temperature can be determined.
2.3 OPERATING CHARACTERISTICS
The instrumentation for velocity, pressure, and temperature measurement should be simple to use
and easy to maintain. The response time should be low, so the instrument can quickly react to changing
operating conditions. Traversing capability is important because of the need for obtaining data which
is representative of the duct cross section. The probe itself should not disturb the flow field sig-
nificantly. This problem is examined in more detail below:
In-situ probes disturb the velocity field of a surrounding flowstream. In HTHP applications
the low subsonic (virtually incompressible) character of the flowstream permits these disturbances to
A-6
-------
Very large particles
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
V /V
V s
Figure 2-2. Effect of anisokinetic sampling on the
measurement of concentration.
A-7
-------
propagate for some distance in all directions from the probe. Disturbance strength depends upon the
size of the probe; it also diminishes with increasing distance from the probe. Thus, it may be pos-
sible and practical in a given application to size the probe(s) so that interactions between the probe(s)
and the flowstream bounding walls or other members are negligible.
Conservative estimation of disturbance strength is possible with a simple two-dimensional model.
Figure 2-3 illustrates the configuration: a circular cylinder oriented parallel within a steady, uni-
form flow field. The fluid is assumed to be incompressible and inviscid.
Analysis of this flow model shows that the maximum disturgance to the surrounding flow occurs
at 6 = 90°. Along AA, the disturbance, is represented by the ratio of the local flow field velocity
(u) to the undisturbed flow field velocity (Um). The ratio $ = u/Uo, depends only upon the ratio (R)
of field position (r) to cylinder radius (a), RE r/a. Figure 2-4 demonstrates the inverse square
law effect of distance on the disturbance.
Probe sizing then involves distinct steps. First, determine the acceptable level of disturbance
at the bounding wall in terms of velocity ratio. Next, find the corresponding distance ratio from
Figure 2-4. Finally, calculate the probe diameter:
. h
d -?
where h is the distance between the flowstream bounding walls.
A-8
-------
Figure 2-3. Circular cylinder in a steady, uniform flow field.
-------
10
Figure 2-4. Velocity disturbance with
distance, 6 = 90°.
A-10
-------
SECTION 3
VELOCITY AND PRESSURE MEASUREMENT
Generally, the method used for the measurement of flow velocity in low speed flows is based on
the Bernoulli equation for incompressible flow. The relation is given by:
p + t> = constant
This equation expresses conservation of energy along a streamline in a perfect fluid (inviscid
and incompressible). The potential energy term (pgz) has been omitted in this equation because for
gaseous flows it is negligible compared to the other terms. Applying this relation at the stagnation
point of the measuring probe,
where: p = stagnation pressure
p = static pressure
V = freestream velocity
p = density of the fluid
Hence, the fluid velocity can be determined from the measurement of the differential pressure,
D - DI
|2(p0 - p)
v = / °
v V P
In most low speed flow situations temperature can be measured with sufficient accuracy by
means of an ordinary thermocouple and the fluid density can be determined from the measured tempera-
ture and static pressure through the equation of state. If it is assumed that a lack of isokinetic
sampling conditions can result in errors in gaseous concentration, then the determination of true gas
density will depend upon a correct setting of the sampling rate. However, the above equation indicates
that the density must be known apriori. Hence, the molecular weight, M, appearing in the equation of
A-ll
-------
state* must be determined either iteratively from the measured quantities of each constituent and then
adjusting the sampling rate, accordingly, or calculated by some means such as shown in Appendix B.
Also in high temperature applications the temperature sensor must be able to survive the environment
and, in addition, be free of errors produced by radiation effects. Several different instruments,
using the simultaneous measurement of pressure differential and temperature measurement to obtain
flow velocity are discussed in Section 3.2.
There are other methods of determining the flow velocity which do not require the measurement of
differential pressure and static temperature. Some of these methods for example, include use of laser
or sound beams. Others depend on determining the force exerted on a body of known drag characteristics.
These methods are also discussed in Section 3.2.
3.1 GENERAL CONSIDERATIONS
In this section certain additional aspects of the problem of velocity measurement using pressure
differential are discussed. First, we must consider under what conditions the flow may be treated as
incompressible.
3.1.1 Compressibility of Gas
Bernoulli's equation for compressible flow is:
V2 Y D
•j- + —1—r ^ = constant
Here, Y> is the ratio of specific heats** at constant pressure to that at constant volume.
Applying the equation between the freestream and stagnation conditions,
T + Y - 1 P = Y^nr pT
Assuming ideal gas behavior, this equation can be expressed as,
where: M = Mach number = - = -—= (ideal gas)
a
* PM
** j- r,
See Appendix B
A-12
-------
The Mach number expected (with maximum gas velocity = 300 ft/sec, and velocity of sound =
1100 ft/sec at room temperature)
Mma
- 300 _ n 07
-TTOO - °'27
room temperature
The error due to compressibility at M = 0.27 is less than 2 percent. At higher temperatures,
the Mach number will be even less, since the speed of sound is proportional to /T. Therefore, com-
pressibility effects are not important in the present application.
3.1.2 Reynolds Number
For low Reynolds numbers, viscous forces are no longer negligible in comparison to inertia
forces and the Bernoulli equation is no longer strictly applicable. In this case the measured pressure
differential can be corrected by dividing by a viscous correction factor, K. Figure 3-1 shows a
theoretical calculation as well as experimental values for the factor, K, for a Prandtl type pitot/
static tube. Several pitot/static tube designs exist; the so called Prandtl tube constituents one
particular design (See Reference 13).
Low Reynolds numbers are obtained at low velocities and high temperatures. Consider air at
2000°F (maximum temperature for the application under consideration) and flowing at 10 ft/sec under
1 atmosphere pressure. Also assume a pitot tube diameter of 0.1 inch.
n - -2- - 14'7 x 144 _ n m, Ibm
p ~ RT ~ 53.3 x 2460 ' °'016 7F
u = 338 x 10"7 JJ"" at ~2500°R
SGC ~ I t
. _ = 0.016 x 10 x 0.1 - .
'• Ke
12 x 338 x 10"7 '
This is about the lowest possible Reynolds number that can be expected since at lower tempera-
tures and higher pressures the Reynolds number will increase. Deviations in the correction factor, K,
from unity are not significant above this value of Reynolds number. If, however, Re drops below 20,
the factor K rises rapidly and the correction factor needs to be considered. It is highly desirable
to be operating above this critical Reynolds number at all times. It is, therefore, desirable to know
the critical Reynolds number at which viscous effects need no longer be considered. It should be
noted that for some other pressure probes the critical Reynolds number may be significantly higher and
under a range of operation they need to be corrected for viscous effects.
A-13
-------
1.6
1.5
1.4
o-fc. 1.3
<3t
I Q.
1.2
1.1
1.0
0.95
(Re) -
1000
5000
Figure 3-1. Correction factor k for low Reynolds numbers for a Prandtl pitot tube.
-------
3.1.3 Turbulence
The differential pressure measurements are affected by turbulent velocity components. These
turbulent components of velocity contribute to errors in the measurement of average static pressure
levels. Corrections for the effects of turbulence are difficult to make since the intensity of tur-
bulence is generally not known apriori and is difficult to measure. The effects of turbulence upon
measured values are generally small (<1 percent) and can probably be neglected for the present ap-
plication.'
3.1.4 Solid Particles in Suspension
The solid particles suspended in the gas stream may affect the measurement of velocity by direct
deposition on the probe thus blocking off the ports. As the pressure ports are blocked progressively
by particle deposition, a drift in the reading will occur. Also, the deposited particles may combust
under suitable conditions. Account also has to be taken of changes in the density of the fluid due to
the presence of the particles. These can be considered as two phase flow effects.
The static pressure is generally measured where the flow is relatively undisturbed from its
freestream condition. Under these circumstances the particles move at nearly the same velocity as
the gas flow and follow the streamlines. Each particle, provided it is not too large, behaves as a
fluid particle and the fluid density can be taken as the average density, i.e., total mass per unit
volume;
p = pgas + C
where C is the mass of particles per unit volume of gas under the measuring conditions. It is clear
that under the conditions of concern this effect is negligible since at the lowest gas density of
interest (p = 0.016 lb/ft3 at 1 atm and T = 2000°F and for C = 0.2 gr/ft3)the error in neglecting the
last term is less than 0.2 percent.
The effect of the particles on the total pressure measurement is more complicated since the total
pressure measurement is dependent upon the manner in which the particles are decelerated as they arrive
at the stagnation point. Influencing factors are: the characteristics of the particles (dimensions,
form, density, velocity, and concentration), of the gas (density, viscosity, and velocity) and of the
probe (dimensions, form, etc.). Two extreme cases can arise in which the particles are sufficiently
large so that they move independently of the streamlines, or alternatively they are sufficiently
small so that they behave as particles of fluid.
A-15
-------
In the case of large particles, there are two possibilities: (a) the particles will lose their
kinetic energy by collision or friction against the probe in which case the fluid will not recover the
energy and the density is to be taken as the density of the fluid itself and (b) the particles are
decelerated by the fluid, in this case the change in kinetic energy of the particles results in an
increase in the total pressure as given by the equation,
2C
that is,
The mixture is thus considered to have a density = p,,,,. + 2C.
yaS
When the particles are sufficiently small, they follow the streamlines and behave like elements
of fluid. There is no exchange of energy between the particles and the fluid. Kinetic energy is
transformed into potential energy according to the Bernoulli equation and the density is to be taken
the average density, i.e., p = p + C. However, in the present application, since C is less than
QaS
0.2 percent of the lowest gas density considered, the effect of particle concentration upon pressure
measurement can be neglected.
3.1.5 Orientation of Probe and Flow Angularity Measurement
Errors can arise when the probe is inadvertently placed at an angle to the direction of flow.
Figure 3-2 shows the typical errors that result when pressures are measured with a probe oriented
at some angle, 6, to the direction of the flow. For flow in a pipe, it is adequate to align the probe
along the centerline, hence, flow angularity is not considered a prime source of error in these appli-
cations.
Sometimes, it may be necessary to make a measurement in a region where the flow direction is
not known. In such a case the direction of flow may be of interest. For this type of application
a device known as a cone flow probe, which is intentionally sensitive to angle of attack, may be used.
Figure 3-3 shows such a probe. In practice, the angle of attack may be deduced by two methods. First,
by calibrating the static readings with angle of attack, or secondly, by moving the probe head, itera-
tively, until the static ports are nulled.
3.2 INSTRUMENTS FOR VELOCITY MEASUREMENT
In this section some of the instruments used for pressure and velocity measurement are described.
A-16
-------
-5 - •
-10 ••
-15 • -
-20 --
-25
Dynamic pressure
Static pressure
Total pressure
4-
12
6° pitot/static
24
Figure 3-2. Influence of pitot/static probe orientation on measured
pressure.
A-17
-------
0
a. Nonnulled position
(calibration map required)
Static ports
Total port
0
b. Nulled position
(probe iteratively positioned)
Figure 3-3. Flow angularity determination using a cone flow
probe.
A-18
-------
3.2.1 PI tot/Static Probe
Pitot/static probes can be used to obtain accurate velocity measurements. The typical probes
shown in Figure 3-4 has an impact port at the head to measure stagnation pressure and one or more
ports located on the side about 8 to 10 diameters downstream of the tip to measure static pressure.
Usually the static ports are manifolded to provide an average value. For low Mach number flow (<0.3)
the dynamic pressure is a direct measure of gas velocity and is simply the difference between stagna-
tion and static pressures. As mentioned previously, to obtain the velocity from the dynamic pressure
it is also necessary to know the gas density. This usually requires information about the gas composi-
tion, temperature, and pressure. Calculations of typical coal combustion products indicate that the
gas constant, R, does not differ greatly from that of air (see Appendix B).
Advantages of the pi tot tube are as follows:
• It is accurate
• Standard designs are available
• Calibration constants are well known
• Nominal flow field disturbance
t Relatively insensitive to flow angularity (especially with ellipsoidal nose design).
The disadvantages and possible remedies are:
t Small ports are likely to be clogged in dust laden environments — if feasible, back pres-
surization can be used to unclog the ports
• At low speeds, the pressure differential may be too small to measure with a standard manom-
eter - sensitive differential pressure transducers can be used
t At high pressure levels manometers may be impractical - use differential pressure trans-
ducers
• The "L" shape may make it difficult to insert in small diameter pipe - use hook shape
• For high velocity streams at high temperature and pressure, aerodynamic loads may lead
to structural problems —add reinforcement ribs (see Figure 3-4).
It seems likely that the shortcomings of the pi tot/static probe can be overcome, and therefore,
it is a viable candidate for pressure and velocity measurements.
A-19
-------
Static ports
Flow
-- --, "P - -T, . ' *"
^ r» ^x
tal port
>
\
J
\
\
N
^
1
1
3
2
<0
i
<
: Reinforcement
Rib
L-shaped
Hook shaped
Figure 3-4. Typical pitot/static probe.
A-20
-------
3.2.2 Disc/Static Heads
In dirty environments, the small holes in the static ports of the pitot tube are liable to be
blocked. It is, therefore, desirable to provide larger port openings. Under such conditions, a disc/
static head may be used with the usual form of pitot head as shown in Figure 3-5. Such a combination
may disturb the flow field considerably more than the standard pitot/static tube. Another disadvantage
is that the measurements are very sensitive to flow angularity that might be encountered near a pipe
bend or near a sudden pipe expansion or contraction. For these reasons, the disc/static head is probably
not appropriate for application to the present problem.
3.2.3 "S" Type Pitot Tube
The "S" type (stausscheibe) or reverse pitot tube (Figure 3-6) has gained popularity for dust
laden environments. The large port openings can delay plugging by dust for a considerable time period.
Also a suction effect on the downstream facing tube causes the pressure in the wake to be lower than
the stream static pressure. As a consequence, measured pressure differences for "S" type tubes are
greater in comparison to standard pitot devices thus increasing the sensitivity. However, it is
usually desirable to know static pressure for other reasons such as determining stream density.
The large tube diameter, plus the fact that two lengths of tubing are coupled back-to-back,
lead to more significant stream disturbances than the pitot/static probe. Hence, there is need for
correct calibration of the "S" type probe. It has been observed that the calibration factor for the
"S" type pitot tube depends on the overall probe configuration (i.e., sampling nozzle, thermocouple,
etc.). Therefore, the "S" type pitot tube must be mounted on the complete probe assembly for proper
calibration. The major disadvantage of the "S" probe is that the calibration factor may depend on
Reynolds number, whereas, the pitot/static probe is relatively insensitive to viscous effects as
mentioned previously. Therefore, the determination of Reynolds number becomes coupled to the measure-
ment of velocity. Also the calibration is somewhat affected by the amount of turbulence present in
the calibration stream compared to the test stream.
The "S" probe is also fairly sensitive to flow angularity. In spite of these negative aspects,
the fact that the probe is much less susceptable to plugging than the pitot/static, makes it an attrac-
tive choice for use in highly particle laden streams. Its use, thereby, avoids the use of unproven,
complex, purging systems.
3.2.4 Other Special Probe Designs for Measurements in Dust Laden Gases
Since an ordinary pitot tube can be easily blocked by particles, special probes may be used
for the measurements of velocity. When the particles are small and their concentration is low, blockage
A-21
-------
Total
Flow
\
Static
1—T
I I
JL_L
7
Figure 3-5. Disk/static head.
A-22
-------
Flow >•
Figure 3-6. "S" type pitot.
A-23
-------
of the total pressure port can be avoided by the method shown in Figure 3-7(a). A small tube facing
downstream is fitted into a large cavity at the stagnation point. The static pressure is not actually
measured, but a pseudo static (base) pressure is measured instead, by porting in an area protected
from the particles by a flange. Since this is a nonstandard probe design, the configuration must be
calibrated in a wind tunnel.
For large particles, the static pressure is measured with a cylindrical tube and instead of
stagnation pressure, pressure in the wake of the tube is measured (Figure 3-7(b)). This probe design
also has to be calibrated.
Another suggested means of preventing probe clogging is shown in Figure 3-8. Here, to avoid
interference effects, the static pressure and total pressure are measured on separate heads. Purge gas
is constantly pumped back through the port openings. The test stream pressure levels are not measured
directly, but instead are deduced from the purge gas flowrate. By maintaining subsonic flow conditions
in the purge flow, the local "back pressure" at the port opening acts as the control valve on the
purge flow. By separately measuring the purge gas flowrates outside of the test stream with some
device such as a venturi flowmeter, the test stream static and total pressure levels can be determined.
By using a conical design similar to the cone flow probe, angle of attack could also be determined.
The design of a system such as the one suggested above would be strongly tied to the particular
operating conditions and test configuration of interest. The supply lines and flow valves would have
to be designed such that internal flow "choking" (point where mass flow per unit area is maximized)
due to friction or area constriction would be avoided. Internal choking would obviate the fundamental
basis of operation, in that the control over purge gas flowrate would shift from the port opening to
the internal choke point. Therefore, the design and operation of such a system requires further
research and development.
3.2.5 Flare Gas Probe
This probe is similar to an "S" type pitot probe with two openings at the probe tip as shown in
Figure 3-9. These openings are connected by an internal tube. A portion of this tube is heated and
a thermoelectric sensor measures temperature gradients along the wall of the tube, external to the
flowstream. The sensor, in effect, acts as a flowmeter. Purge gas is injected into the tubing such
that a pneumatic bridge is formed. At zero line velocity, the bridge is balanced and purge gas exhausts
out of both tips equally. As flow across the tip occurs, a differential pressure is developed and the
bridge is unbalanced. Purge gas still exhausts from both openings, but now they are slightly unequal.
The thermoelectric sensors measure the shift in temperature gradients along the heated portion of the
A-24
-------
Pseudo-static pressure
Total pressure
Psuedo-static pressure
Total pressure
a. Small particles
Negative pressure in the wake
Kl CVVVNXSXVVVCVS.NX
Static pressure
b. Big particles
Figure 3-7. Special probes for measurement in dust
laden gases.
A-25
-------
Venturi meter
a. Total pressure probe (enlarged)
Static head probe
I
b. Static pressure probe (enlarged)
Supply
pressure
Figure 3-8. Purge flow pressure sensors.
A-26
-------
r
x— «w
f
1
V
i x
— ^y
/
/
/
i
\
*^
r
1
i
r^
f
^
- KL
:-1
rge gab
<~\
8H
' Bridge"X
/'"""
I
X) — i
v_-x
Sensors
>
[~
v
1
•>
1
1
1
1
1
1
1
1
1
n
f -~ U
( — -
/\ to- 1
>
1
/
M^"
^
r
r
I
/
i
j
^
^
1\
Figure 3-9. Flare gas probe.
A-27
-------
tube which are directly related to the main gas flow. Because the purge gas is continuously exhausting
into the flowing gas, corrosive of particulate laden gases are prevented from entering the probe.
Advantages of this probe are:
• It can be used in a dust laden environment
• Calibration is relatively constant
t The probe is most sensitive to low flowrates
t Good accuracy (±2 percent of full scale).
The disadvantages are:
• The output is nonlinear. The probe is most sensitive at low velocities. However, at high
velocities (>20 m/sec), the measurements have less accuracy
• The purge gas pressure must be higher than the line pressure. In high pressure environments
(-100 atm), this may be troublesome
t The probe must be isolated from vibrations
• Zero drift problems occur
• Initial setup and adjustment is complicated.
3.2.6 Drag Meter
The drag force on a target disc is measured and the velocity is calculated from knowledge of
the meter drag coefficient (Figure 3-10). Bonded strain gages, in a bridge circuit outside the fluid
stream, translate this force into an electrical signal.
Advantages are:
• Particulate loadings are acceptable (for heavy loading purge cleaning may be required)
• Good accuracy (+1 percent)-
Disadvantages are:
• The probe is sensitive to flow angularity
• The stream density must be well known
t Heavy particulate loading could cause errors due to impacts.
A-28
-------
K
Probe
bidirectional
!
Fiqure 3-10. Drag meter.
A-29
-------
3.2.7 Fluidic Sensors
Two types of fluidic sensors are generally used: crossflow and co-flow. The crossflow sensor
(Figure 3-11) employs a jet stream normal to the measured flow, two receiver ports and a device to
measure the differential pressure across the receiver ports. Sensor operation depends upon the jet
entraining and mixing with the surrounding fluid. This causes a jet deflection which is measured by the
receiver port. The differential pressure across the receiver ports is a measure of flow velocity.
The crossflow sensor is most accurate at low velocities (0.3 to 18 m/sec). At higher velocities (14
to 150 m/sec) co-flow sensors are used. Here the jet is in the same direction as the flow.
The advantages are:
t The probe can be used in particulate laden, corrosive environments
• Rugged construction suitable for field use
• Good accuracy (±3 percent)
• Good sensitivity at low flowrates.
The disadvantages are:
• Receiver ports may require purge to keep particulate matter from clogging ports
a Two sensors (cross- and co-flow) required to cover the required velocity range
• Calibration may be difficult since gas mixing may differ with test conditions.
3.2.8 Acoustic Velocimeter
Sound beams are transmitted both upstream and downstream of the flow. The speed of propagation
is the vector sum of the local sound velocity and the flow velocity. The parameter measured is the
time of transmission of a sound pulse or the change in phase of a continuous wave beam. The flow speed
is determined without an explicit determination of the sound speed. One disadvantage is that heavy
particulate loading can interfere with phase measurements. This technique is not yet commercially
available for the type of applications discussed.
3.2.9 Orifice and Venturi Meters
With this method, a nozzle or orifice plate is inserted across the pipe . This restricts the
area through which flow can pass and thus increases local velocity. A comparatively large drop in
static pressure is experienced by the flow, which can be used as a measure of flow velocity. These
devices are commonly used for measurement of process stream flowrates. They give average flow
velocity rather than velocity at a point (which is essential for isokinetic sampling). However,
A-30
-------
(
1
Measu
vel
oc
i\
\
S
Supply
^ i^^_i»
^*"™~~1
'A'ozzle flow
/ Jet
r
red
it)
P !
X
V.
t
Total head tubes
V -/
/
J-
/*
tf
^—
\
^
— ft/sec —
v J
^^ _^
^
y
|
Figure 3-11. Fluidlc sensor.
-------
if the stream is fully developed and well behaved, it may be possible to assume a fully developed
turbulent velocity profile. Then knowing the total flowrate one could compute the local velocity.
In the present case, the lowest pipe Reynolds number would be about 1600 (1 atm, 10 fps,
4 in. dia., 2000°F), The entrance length required to establish fully developed flow would be given
by Le = 0.035D, or about 18 ft. At the highest Reynolds number, Re = 4.7 x 10s (100 atm, 300 fps,
4 in. dia., 2000°F) the corresponding turbulent entrance length would be about 10 ft. It is,
therefore, conceivable that the required conditions could be obtained to make such a measurement.
Such a technique would require more detailed study and verification by test.
3.2.10 Laser Velocimeter
When two coherent beams of light intersect in space, stationary three-dimensional interference
fringe patterns are established. The fringe spacing, A, is a function of the wavelength, A, of light
and the angle, 6, at which the two beams intersect (Figure 3-12). The laser velocimeter utilizes the
wave interference properties of two coherent light beams.
Particulate laden combustion gases pass through the region of interference. A particle moving
through adjacent fringes of constant separation will scatter light periodically. This scattered
light is collected by a photodetector which produces a pulsating electric current whose frequency is
proportional to the particle's velocity.
The advantages of laser velocimeter are:
• High accuracy
t No perturbation of flow field
t No mechanical calibration errors.
The disadvantages are:
• High cost
• Elaborate setup required which may not be suitable for field use
• Problems of light transmission through heavily particle laden, radiating streams.
Although the laser velocimeter method is promising, considerable development problems must be
overcome before it can be considered to be suitable for field application. Programs designed to test
the feasibility of this method are underway at present. It is expected that within the next year
or two the desirability of using this approach will be well defined.
A-32
-------
CO
OJ
Note: Interference fringe
spacing, A, must be
greater than particle
size.
/- First traveling
/ wave of wavelength X
Interference
pattern
Second traveling wave
of wavelength X
Figure 3-12. Schematic of the interference between two traveling collimated coherent beams
of light of wavelength X passing through each other at an angle 6.
-------
SECTION 4
TEMPERATURE MEASUREMENT
As mentioned previously, temperature measurement is required in the determination of velocity.
It is also necessary to characterize the thermodynamic state of the gas flow being sampled. The various
aspects of temperature measurement in a gas stream are discussed in Section 4.1. Determinization of gas
temperature using thermocouples is discussed in Section 4.2. Some other methods of temperature measure-
ment such as pyrometry are also described.
4.1 TEMPERATURE MEASUREMENT IN GAS FLOW
Ideally, a temperature measuring probe should stagnate a moving gas stream and give measurement
of the stagnation temperature* of the gas. However, in reality, the temperature of the probe differs
from the stagnation temperature as a result of energy exchange within the boundary layer, so the tem-
perature measured by the probe, T , will be different from the stagnation temperature, T , of the gas
by the amount
\;2
o m
where: a is the recovery factor, with values for velocities up to Mach 3 given by,
a = 0.68 for probe perpendicular to the flow direction
a = 0.86 for probe along the flow direction
To establish the maximum error due to the use of recovery temperature in place of the total
temperature, assume V = 300 ft/sec, and the probe perpendicular to flow
.'. AT =(1-0.68) x.2x7?8 -2x0-24
= 2.4°F
Thus, error due to recovery factor is not significant for the present application.
The difference between the static and total temperature under the conditions of interest is negligible
V2 3002
Tt " Tt = 2JgC~ = 2 x 778 x 32.2 x 0.24 = 7'5°F
A-34
-------
The sensing probe and Its surreoundings exchange heat by convection, radiation, and conduction.
For radiant heat exchange, the probe will be "seeing" a temperature which is weighted towards the
temperature of the wall. The wall temperature may be significantly less than the gas temperature;
hence, there will be radiation heat loss from the probe to the walls. Also, the base of the probe may
be at a lower temperature which will cause conduction heat loss. These losses have to be balanced by
convective heat transfer from the gas to the probe and, hence, losses due to radiation and conduction
will increase the temperature difference between the gas and the probe. To minimize this error,
radiation and conduction losses should be minimized, while at the same time the convective heat transfer
coefficient to the probe should be maximized.
4.2 TEMPERATURE MEASUREMENTS WITH THERMOCOUPLES
The simplest and most convenient temperature measurement device is a thermocouple. The thermo-
couple must be in thermal equilibrium with the surrounding media being measured. This presents several
problems because the equilibrium temperature of the thermocouple junction inserted into a gaseous
stream is the temperature resulting from:
1. Heat transfer by convection between the thermocouple and the gases across the boundary layer.
2. Heat transfer by radiation between the thermocouple, the gases, suspended particles, and the
walls of the duct.
3. Heat conduction along thermocouple wires.
The true temperature of the gas can be determined by either minimizing the errors due to the above
factors or applying corrections to the measured values. The equations governing the heat transfer be-
tween the thermocouple and its surroundings are briefly discussed below.
4.2.1' Convection
The convective heat transfer, Q , to the thermocouple is given by,
c
Qc = hc (TQ - Tt)
where: h = convective heat transfer coefficient
T = temperature of the gas
T. = temperature of the thermocouple junction
The heat transfer coefficient, hc> is generally calculated from the knowledge of the Nusselt,
Prandtl, and Reynolds numbers. For combustion gases the Prandtl number is approximately 0.7 and the
relations between the Nusselt number (Nu = hcd/k) and the Reynolds number (Re = pVd/y) are as follows:
A-35
-------
Nil = 0.44 Re0'5 for a thermocouple perpendicular to the flow
Nu = 0.085 Re0-67" for a thermocouple parallel to the flow
k = gas conductivity
Figure 4-1 shows the relationship between Nusselt and Reynolds numbers. It is apparent that in order
to maximize convective heat transfer, the crossflow arrangement is preferable for Re < 15,000, whereas
the parallel flow arrangement gives higher convective heat transfer for Re > 15,000.
Aspirating the gases at high velocity past the thermocouple, and decreasing the diameter of the
thermocouple also increase the convective heat transfer coefficient according to:
hc a VP d'™
where: 0.5 < n < 0.7
0.3 < m < 0.5
Such a means of promoting convective heat transfer to the thermocouple will increase the accuracy of
the device by bringing the temperature of the thermocouple closer to that of the gas stream.
4.2.2 Radiation
The radiant heat exchange, QR) between the thermocouple at temperature, T,, and the surrounding
walls which are at temperature, T, is given by,
QR = ea (T* - T")
where: e = emissivity of the thermocouple
cr = Boltzmann constant
At high temperatures the radiative loss to the cool walls can be significant. Therefore, it
is necessary to minimize this radiant heat exchange. One way to do this is to reduce the emissivity,
e. Polished metal surfaces have low emissivity at low temperature, but the emissivity increases rapidly
with temperature. The emissivity also increases rapidly if the surface is altered by oxidation or
deposition of foreign materials.
Radiation sheaths made of refractory materials such as Al^O, possess, on the contrary, emissivities
that decrease with increasing temperature. Also, they are much less subject to chemical attack by the
gas mixture.
A-36
-------
TOO
60
20
10
6
10J
Thermocouple
perpendicular
to flow
Thermocouple
parallel to flow
Figure 4-1. Relation between Reynolds and
Nusselt numbers for flow over
a thermocouple.
A-37
-------
Another way to insure low radiant heat exchange is to create a local wall temperature, T, as
close as possible to the thermocouple temperature, TT. This can be achieved by surrounding the thermo-
couple by multiple radiation shields. In this configuration the emissivity of the probe and outside
of the radiation shield should be as low as possible, whereas the emissivity of the inside of the
shield should be as high as possible.
4.2.3 Conduction
Heat will be transferred from the tip of the thermocouple to its base by means of conduction
along the junction wires. If the base temperature is Tg, the difference, AT, between the measured tem-
perature and that of the gas is given by
TT - T
AT = -I 9.
L/4h/dk
where: L = length of thermocouple wire immersed in the flow
h = heat transfer coefficient between gas and thermocouple
d = diameter of the thermocouple wire
k = thermal conductivity of the thermocouple
It is likely that these conduction errors can be reduced, substantially, by adequate sheathing of the
wires in an insulating material.
In order to fully evaluate the error likely to be produced by radiation and conduction effects,
it is necessary to perform a heat balance on the thermocouple configuration of interest. The convective,
conductive, and radiative transport terms must be evaluated, simultaneously- The solution is necessarily
iterative owing to the transcendental nature of the equations.
4.3 OTHER TEMPERATURE MEASUREMENT DEVICES
4.3.1 Radiation Pyrometers
At high temperature, radiation pyrometers may be used for temperature measurement. Thare are
two classes of pyrometers: (1) total radiation pyrometers which measure total thermal radiation emitted
and (2) optical pyrometers which compare the brightness of the hot body with the brightness of a standard
radiation source. The most obvious problem with pyrometry in the present application is the optical
transparency of the gases. If the gas is transparent, the pyrometer will be seeing the walls and give
a measurement which is weighted towards the wall temperature rather than the gas temperature. On the
other hand, heavy particle loading will alter readings if the particles are not in thermal equilibrium
A-38
-------
with the gas. Other major drawbacks are a relatively high, lower temperature limit (~1500°F to 2000°F)
and poor spatial resolution.
4.3.2 Molecular Beam
The temperature of the gas can be determined directly from measurement of molecular motion. The
molecular beam sampling apparatus causes the gas to undergo a very rapid expansion to supersonic vel-
ocities. By intermittently chopping the molecular beam and measuring the time required for a chopped
portion of the beam to arrive at a detector site, the velocity of the molecules can be determined. The
velocity is a direct indication of the stagnation temperature of the gas. This technique has good
accuracy and time response. However, the probe size is relatively large and is susceptible to blocking
by parti cul ate matter.
4.3.3 Velocity of Sound Method
Measurement of the velocity of sound can indcate the temperature of the gas if its thermodynamic
properties are well known. Measurement of sound velocity results in determination of a value for
a = /yRT. This technique, however, has poor spatial resolution and requires a relatively large probe
size. It also assumes the gases behave as a perfect gas mixture.
A-39
-------
SECTION 5
CONCLUSIONS
Various aspects and methods of measuring velocity, pressure and temperature in a high temper-
ature, high pressure, dust laden environment were examined. These methods were examined on the
basis of accuracy, utility, and state of development. In the dusty environments encountered in the
present application some means need to be provided for keeping the sensing ports free of particulates,
tars, etc. In the near term, the "S" type pitot/static probe offers the simplest solution to the
problem of velocity measurement. Shielded thermocouples should provide adequate temperature
determination in the near term.
To improve the accuracy of the next generation of sampling systems, flow sensing devices
incorporating more refined means of cleaning pressure ports and thermocouples should be considered.
The transient purging of a pitot/static tube offers a logical solution to this problem. More
detailed analyses needs to be performed to determine the necessity for aspirating thermocouples.
Some of the newer developments in velocity, pressure, and temperature sensing were briefly
examined. Under their current state of development, it is difficult to assess their utility. At
present, some of these devices are being evaluated by Argonne Labs, Sandia Corporation and others
to determine their application towards HTHP measurements.
A-40
-------
APPENDIX A
BIBLIOGRAPHY
1. Hawksley, P. G., Badzioch, S., and Blackett, J. H., "Measurement of Solids in Flue Gases," The
British Coal Utilization Research Association, 1971.
2. Parker, G. J., "Some Factors Governing the Design of Probe for Sampling in Particle and Drop
Laden Streams," Atmospheric Environment, Volume 2, pp 477, 1968.
3. Sem, G. J., and Borgos, J. A., "Instrumentation for Measurement of Particulate Emissions from
Combustion Sources," Environmental Protection Technology Series, September 1973.
4. Heflinger, L. 0., Matthews, B. J., and Shelton, H., "Portable Laser Velocimeter for Stack Velocity
Measurements," TRW report 20852-6002-RO-OO, November 1972.
5. Cheremisinoff, P. N., "Industrial Pollution Control Measurement and Instrumentation," Technomic
Publishing Company, 1976.
6. Lengelle, G., Verdier, C., "Gas Sampling and Analysis in Combustion Phenomena," AGARDOGRAPH
No. 168, July 1973.
7. Pustinger, 0. V., Shaw, D. A., Sherman, P. L., and Snyder, A. D., "Instrumentation for Monitoring
Specific Particulate Substances in Stationary Source Emissions," Environmental Protection Technology
Series, EPA-R2-73-252, September 1973.
8. Brooks, E. F., Beder, E. C., Flegal, C. A., Luciani, D. J., and Williams, R., "Continuous Measure-
ment of Total Gas Flowrate from Stationary Sources," EPA-650-2-75-020, February 1975.
9. 0'Fallen, N. M., Beyerlein, R.A., Managan, W. W., and Karplus, H. B., "Monitoring Coal Energy
Processes," Industrial Research, June 1976.
10. Baas, P. B., and Mai, K., "Trends of Design in Gas Turbine Temperature Sensing Equipment," Tempera-
ture - Its Measurement and Control in Science and Industry, Reinhold Publishing Company, 1962.
11. Ladenburg, R. W., Lewis, B., Pease, R. N., and Taylor, H. S., "Physical Measurements in Gas Dynamics
and Combustion, Volume IX, Princeton Series on High Speed Aerodynamics and Jet Propulsion, Princeton
University Press, 1954.
12. Kesselring, J. C., Gorji, K. M., Young, W. S., Rodgers, W. E., and Knuth, E. L., "Flame Character-
ization Probes," EPA- 650/2-74-023, March 1974.
13. Ower, E., and Pankhurst, R. C., "The Measurement of Air Flow," Pergamon Press, 1969.
14. Gnyp, A. W., St. Pierre, C. C., Smith, D. S., Mozzon, D., and Steiner, J., "An Experimental Investi-
gation of the Effect of Pitot Tube Sampling Probe Configurations on the Magnitude of the S-type
Pitot Tube Coefficient for Commerically Available Source Sampling Probes," University of Windsor,
February 1975..
15. Brooks, E. F., and Williams, R. L., "Technical Manual for Process Stream Volumetric Flow Measure-
ment and Gas Sample Extraction Methodology," TRW 24916^6028-RU-OO, November 1975.
16. Stoller, H. M., Schrider, L. A., and Northrop, D. A., "The Status of Instrumentation and Process
Control Techniques for In-situ Goal Gasification," ISA-ASI 76224, pp 117-128, 1976.
17. Blann, D. R., "Measurement Methods at High Temperature and Pressure," Symposium on Particulate
Cpntrol in Energy Processes, May 1974.
A-41
-------
18. Beer, J. M., and Thring, M. W., "Measurements in Flames," Edward Arnold Publishers Ltd., 1972.
19. Mercer, T. T., "Aerosol Technology in Hazard Evaluation," Academic Press, Inc., 1973.
20. Vitols, V., "Theoretical Limits of Errors Due to Anisokinetic Sampling of Particulate Matter,"
Journal of the Air Pollution Control Association, p. 79, Feb. 1966.
A-42
-------
APPENDIX B
COAL COMBUSTION PRODUCT ANALYSIS
A series of combustion calculations typical of fluidized bed combustion (FBC) products was
performed. The purpose of the combustion calculations presented here are twofold. First, it was
desired to establish typical values of thermodynamic properties such as the ratio of specific heats
and molecular weight for evaluation of quantities required to perform flow calculations. Once the
molecular weight is determined, the gas constant, R, can be determined from R = 1545/molecular
wt[ft-lb/lb-°F].
The second purpose of these calculations was to provide some insight into the effect of
changes in pressure level and equivalence ratio upon the gaseous constituency. For instance, within the
sampling system, changes in pressure will occur owing to internal flow pressure drops. As a result,
the gas constituency will change as a result of these pressure changes. Physically, these changes
will occur at a finite rate in accordance with the laws of chemical kinetics. While no attempt
was made to model chemical kinetics, the equilibrium values shown, indicate the maximum change in
gaseous constituency between two end states.
The calculations were performed using the Aerotherm Chemical Equilibruim (ACE) Code. A fuel
typical of an eastern coal was used, while the equivalence ratio was varied from 0.5 to 2.0.
Figure B-l presents the mole fractions of some of the combustion products of interest. Other pro-
ducts have been ommited for clarity. Note that some species are fairly insensitive to changes in
pressure ranging from 1 to 100 atmospheres. Figure B-2 presents the ratio of specific heats
(necessary for compressible flow analysis) as a function of equivalence ratio and pressure level.
At equivalence ratios exceeding unity there is little effect of pressure level upon specific heat
ratio. Figure B-3 presents the corresponding values of molecular weight for the combustion products.
Over the range of equivalence ratios and pressures shown, the molecular weight remains between 27 to
30 Ib/lb-mole. Therefore, using an average value of about 28.5 would give reasonably accurate
results over an extended range of conditions since the error in velocity would be proportional to
the square root of the accepted error in molecular weight.
A-43
-------
T("K) P(atm) T(°K) P(atm) T(°K) P(atm)
10"
0 •
10"'
10"'
10"(1
10 " •
1087 - 100 1829 - 100 1397 - 100
1059 - 50 1828 - 50 1397 - 50 Adiab
1024 - 20 1027 - 20 1397 - 20 flame
997 - 10 1826 - 10 1397 - 10 tempe
914 - 1 1825 - 1 1397 - 1
A- 19807
• H2- 1 ' '
a H2-in
• H2-20
A u r-n .« Stoichiometric
• H9-bU
• H2-100
O
QH?S -1,10, 20, 50, 100
^ -N
• CII4 - ] I SO,
I rs so? - i, 10, 20
J 0 S02 - 50, 100
• m - 100 ° M2S " 10° m ' ] ' 10' 20' 50) 10°
• H\\3 - 50 o H2S - 50
O COS - 1, 10, 20, 50 ° S°3 " 10°
o SO^ - 50
» Hl\^ - 20 0 H9S - 20 •>
J ^ 0 S03 - 20
• NH3 - ° o S03 - 10
0 H9S - 10
• NH0 - 1 Mn i ° N09 - 10°
w 3 0 NO - 1 2
o SO-, - 1
o H9S - 1 J
« HCN - 100 o N02 - 50
• IICN - 50
.HCN -20 °N°- 10 °C1 - '
.HCN - 0 ° COS - ' „ N0 ,
o NO - 20 2
o Cl - 50
- o NO - 50 ri _ lnn
atic
rature
- 100
10
1 \J
1
10 ppm
0.5
1.0
1 .5
2.0
1 ppm
Air to fuel/fair to fuel/stoichiometric)
Figure B-l. Equilibrium combustion products from
typical eastern coal.
A-44
-------
T("K) P(atm)
1087 - 100
1059 - 50
1024 - 20
997 - 10
914 1
T(;K) P(atni
1P29 - 100 "
1828 - 50
1827 - 20
1826 - 10
182B - 1
T(JK) Pfatr"
1397 - 100
1397 - 50
1397 - 20
1397 - 10
1397 - 1
I
Adiabatic
flame
temperature
-Stoichiopietric
10°
N, - 1, 10,
i 20, 50,
100
Pressure (atm)
N, - 1, 10, 20,
c 50, 100
100%
i C02 - 1, 10, 20, 50, 100
V
£
\
-P=.
en
10"
CO,
•- 1
100, __10
5=18 >co
© CO, - 1, 10, 20,
*• 50, 100
IC02 - 1, 10, 20, 50, 100
10%
H20 - 100
H?0 - 50
HoO - 20
HjO - 10
§H.O - 1, 10, 20, 50, 100
CO - 1, 10, 20, 50, 100
|H20 - 1, 10, 20, 50, 100
1%
1.5
1.0 1.5 2.0
Air to fuel/(air to fuel/stoic)
Figure B-l. Concluded
-------
1 29
1.28'
1.27-
1.26-
1.25.
1.24-
"0" 1.23-
•r—
fO
<- 1.22-
ro
J= 1.21.
o
£ 1.20-
0
Ol
Q.
^ 1.19-
1.18.
1.17-
1.16-
1.15.
1.14-
1.13-
a
SI, 10, 20
50, 100
X 100
O 50
A 20
Q 10
O i
Pressure (atm)
1, 10, 20, 50, 100
(O
o
CD
0.5 1.0 1.5
Air to fuel/(air to fuel/stoic)
2.0
Figure B-2.
Ratio of specific heats of combustion products
from typical eastern coal.
A-46
-------
O(J
29
at
i
.a
1,28 -
'i
s_
as
MolecuL
27
O 1, 10, 20, 50, 100
f
X 100
O 50 £
ac
ll\
i**
A 20 7
a 10
0 i
O 1, 10, 20, 50, 100
0.5 1.0 1.5
Air to fuel/(air to fuel/stoic)
2.0
Figure B-3.
Molecular weight of combustion products
from typical eastern coal.
A-47
-------
Appendix B
Aerotherm Project 7237
MATERIAL SELECTION FOR
SAMPLING IN COAL
CONVERSION SYSTEMS
Jacques Hull
Acurex Corporation/Aerotherm Division
485 Clyde Avenue
Mountain View, California 94042
June 1977
AEROTHERM REPORT TM-77-180
Prepared for
EPA Project Officer — William Kuykendal
Industrial Environmental Research Laboratories
U.S. Environmental Protection Agency
Research Triangle Park
North Carolina 27711
Contract No. 68-02-2153
B-i
-------
TABLE OF CONTENTS
Section
1 INTRODUCTION B-1
2 BASIC METALLURGICAL TERMS B"3
3 GASIFIER CONDITIONS B"8
4 MECHANICAL PROPERTIES AT GASIFIER CONDITIONS B'21
5 CORROSION EFFECTS B"24
5.1 Basic Corrosion Mechanisms B"25
5.2 Gaseous Corrosion Reactions B~30
5.2.1 Hydrogen Embrittlement B"32
5.2.2 Oxidation B-32
5.2.3 Carburization B-34
5.2.4 Sulfidation B-35
5.2.5 Reactions With Other Gases B-38
5.2.6 Metal Surface Conditions (Fabrication Factors) B-38
5.2.7 Coal Gasification Gaseous Corrosion Data B-39
5.2.8 Refractories B-47
5.3 Liquid Corrosion Reactions B-48
5.3.1 Moderate Temperature Reactions B-48
5.3.2 Hot Corrosion B-50
5.3.3 Erosion-Corrosion B-50
6 EROSION B-52
7 GASIFIER FACILITIES PROBLEMS B-55
7.1 C0? Acceptor Problems B-55
7.2 Hygas Problems B-56
7.3 Synthane Problems B-58
7.4 Fluidized Bed Combustor Problems B-59
7.5 Miscellaneous Problems B-60
8 CONCLUSIONS -- SIGNIFICANCE TO SAMPLING PROBE MATERIALS B-62
BIBLIOGRAPHY B-65
B-ii
-------
LIST OF ILLUSTRATIONS
Figure Page
1 Consol C02 Acceptor process (Reference 11) B-ll
2 IGT Hygas process (Reference 11) B-12
3 Bureau of Mines Synthane process (Reference 11) B-13
4 Bituminous Coal Research - Bi-Gas process (Reference 11) B-14
5 Battelle-Union Carbide Agglomerated Ash process (Reference 11) B-15
6 Thermodynamic stability diagram for Cr-O-C (Reference 19) B-19
7 Thermodynamic stability diagram for metal-sulfur-oxygen (Reference 28) B-20
8a Magnification: 5000x B-33
8b Magnification: lOOOx B-33
8c Magnification: 500x B-33
LIST OF TABLES
Figure Page
1 Environmental Conditions in Various Coal Conversion Processes B-10
2 Inlet and Equilibrium Gas Composition of Phase I Corrosion Tests on
ERDA/MPC IITRI Program3 B-17
3 Basic Checical Composition of Selected Alloys3 B-27
4 Specifications Known to Cause Stress Corrosion Cracking When Present at low
Concentrations and as Impurities B-31
Corrosion Rates in Stainless Steels and Superalloys in Simulated Reducing
Coal Gas3 at 1500 F and 1000 PSI. Loss of Sound Metal in Mils per year0
5
per year"
(Based on 1000 hour data) B-41
6 Corrosion Rates in Stainless Steels and Superalloys in Simulated Reducing
Coal Gas3 at 1800°F and 1000 PSI. Loss of Sound Metal in Mils per year0
(Based on 1000 hour data) B-42
7 Corrosion Rates in Stainless Steels and Superalloys in Simulated Mildly
Oxidizing Coal Gas at 1850°F and 150 PSI. Loss of Sound Metal in Mils
per year (Based on 1000 hour data) B-43
8 Comparison of Corrosion of Selected Alloys in Laboratory and Pilot Plant
Exposures at 1850°Fa B-44
9 Comparison of Laboratory and Pilot Plant Corrosion Tests in Consol-Co^
Acceptor Regenerator Environment B-45
10 Acceptable Alloys for Service in Coal Gas Environments at 1800°Fa B-46
11 Comparison of Corrosion Rates of Steels in Liquid and Gas Quench Phases3 B-49
12 Failure Modes of Components in Coal Conversion Plants3 B-57
B-iii
-------
SECTION 1
INTRODUCTION
The currently available stainless steels and superalloys have been
designed to resist high temperature corrosion in highly oxidizing
atmospheres. Basically, these materials are alloyed to counteract high
temperature oxidation in air by relying on the selective oxidation of
chromium to form protective oxide scales on external surfaces at the
gas/metal interface.
The materials in coal conversion process equipment will be exposed at
high temperature to environments in which the oxygen activity is generally
secondary to the activity of other gaseous components, such as carbon and
sulfur in the form of CO, C02, CH4> S02, S03, H2S etc. Existing alloys have
not been designed to resist corrosion resulting from interaction of the
metallic components with these gases. Consequently, these materials are
susceptible to severe and sometimes catastrophic attack by sulfidation and/or
carburization. Protective compounds are not formed at the gas/metal
interface and unacceptable rates of corrosion can result.
The design of high temperature-high pressure (HTHP) probes for
sampling in coal conversion systems must rely heavily on experience gained
from the corrosion of materials in coal conversion pilot plants and in
laboratory investigations simulating coal conversion conditions. Because of
the existence of particulate matter such as char and flyash, erosive effects
B-l
-------
must also be considered as they tend to aggravate the corrosive attack.
In order to better understand the interrelation of the factors
influencing the selection of materials for HTHP sampling probes, this
discussion will deal with a series of topics in the following order: After
defining metallurgical terms, the environmental conditions existing in
various conversion systems are discussed. This leads to a discussion of the
mechanical properties of materials in these environments, and is followed by
a detailed discussion of the various corrosion modes, including gaseous
corrosion, liquid phase corrosion, hot corrosion and erosion/corrosion. The
direct effects of erosion are outlined, and this is followed by summaries of
investigations relating to a number of component failures in coal conversion
systems. Finally, the significance of this information to the selection of
materials for sampling probes is discussed, and a bibliography is provided.
B-2
-------
Alkali metals:
Aluminizing:
Annealing:
Anodic corrosion:
Austenite:
Brittleness:
Carbide:
Carburizing:
Caustic:
Cementite:
SECTION 2
BASIC METALLURGICAL TERMS
Principally sodium, potassium and lithium.
Forming an aluminum compound coating.
Heating to and holding at a suitable temperature and
then cooling at a suitable rate to produce a desired
microstructure.
The dissolution of a metal acting as an anode in contact
with a dissimilar metal acting as a cathode, as a result
of the formation of an electrolytic cell.
A solid solution of elements in face-centered cubic
iron. The 300 series stainless steel are austenitic.
The quality of a material that leads to crack
propagation without appreciable plastic deformation.
A compound of carbon with one or more metallic elements.
Introducing carbon into a solid ferrous alloy. In
stainless steels, this can result in chrome carbide
formation.
Strongly alkaline.
Iron carbide.
B-3
-------
Char:
Cold work:
Corrosion:
Crevice corrosion;
D i f f u s i on:
Ductility:
Erosion:
Eutectic:
Eutectic melting:
Exfoliation:
Ferrite:
Finish:
Coal combustion product containing unburned carbon.
Permanent strain produced by an external force.
The deterioration of a metal by chemical or
electrochemical reaction with its environment.
Corrosion caused by concentration of salts or gases in
crevices resulting in differential cell build-up
ultimately causing deep pitting.
Movement of atoms or molecules to new sites within a
material.
The ability of a material to deform plastically without
fracturing.
Destruction of materials by the abrasive action of
moving fluids usually accelerated by the presence of
solid particles in suspension.
An isothermally reversible reaction in which a liquid
solution is converted into two or more intimately mixed
solids on cooling.
Melting of localized micro-regions whose composition
corresponds to that of the eutectic.
A type of corrosion that progresses approximately
parallel to the outer surface causing layers of the
metal to be elevated by the formation of corrosion
products.
A solid solution of elements in body-centered cubic
iron. The 400 series stainless steels are ferritic.
Surface condition, quality or appearance of a metal.
B-4
-------
Fluxing:
Flyash:
Grain size:
Heat treatment;
Hydrogen
embrittlement:
Impurities:
Inclusions:
Intergranular
corrosion:
Martensite:
Microstructure
Modulus of
elasticity:
Nitriding:
Oxidation:
Removal of material from the surface by dissolving in a
molten flux.
A finely divided siliceous material formed during the
combustion of coal.
Average size of individual crystals in a polycrystalline
metal or alloy.
Heating and cooling a solid metal or alloy to obtain a
desired microstructure.
A condition of low ductility in metals resulting
from the absorption of hydrogen.
Elements or compounds whose presence in a material are
undesirable.
Nonmetallic materials in a solid metallic matrix.
Corrosion occurring preferentially at grain boundaries.
A metastable phase of steel formed by transformation
*.' ik. i,
from austenite on cooling at-a sufficiently rapid rate.
The structure of polished and etched metals as revealed
by a microscope.
A measure of the rigidity of the material.
Introducing nitrogen into a solid ferrous alloy by using
a nitrogenous material such as ammonia.
A reaction in which there is an increase in valence
resulting from a loss of electrons. Formation of an
oxide scale on a metal surface.
B-5
-------
Passivity:
Pear lite:
Phase:
Plastic deformation:
Polycrystalline:
Precipitation heat
treating:
Quenching:
Refractory:
Residual stress:
Scaling:
Sigma phase:
Slag: t
Solid solution:
Solution heat
treatment:
A condition 1n which a metal, because of an impervious
covering of oxide or other compound, has a potential
much more positive than where the metal 1s in an active
state.
A lamellar aggregate of ferrite and cementite.
A physically homogeneous and distinct portion of a
material system.
Deformation which remains permanent after removal of the
load that caused it.
Pertaining to a solid composed of many crystals.
Artificial aging in which a constituent precipitates
from a supersaturated solid solution.
Rapid cooling
A high melting point material.
Stress present in a body that is free of external forces
or thermal gradients.
Forming a thick layer of oxidation products on metals.
A hard brittle phase present in alloys containing
sufficient chromium.
A nonmetallic product resulting from the mutual
dissolution of flux and nonmetallic Impurities.
A single solid homogeneous crystalline phase containing
two or more chemical species.
Heating an alloy to a suitable temperature, holding
that temperature long enough to allow constituents
to enter into solid solution, and then cooling rapidly
to hold the constituents in solution.
B-6
-------
Spelling:
Stabilizing
treatment:
Stress-corrosion
cracking:
Stress-relieving:
Thermal fatigue:
Thermal shock:
Thermal stresses:
Toughness:
Welding stress:
The cracking and flaking of particles on a surface.
Heating austenite which contains stabilizing elements
to tie down its carbon by forming carbides with these
elements.
Failure by cracking under combined action of corrosion
and stress.
Heating to a suitable temperature to reduce residual
stresses, then cooling slowly to minimize development
of new residual stresses.
Fracture resulting from the presence of temperature
gradients which vary with time so as to produce cyclic
stresses.
The development of a steep temperature gradient and
accompanying high stresses.
Stresses in a material resulting from nonuniform
temperature distribution.
Ability of a material to absorb energy and deform
plastically before fracturing.
Residual stress caused by localized heating and cooling
during welding.
B-7
-------
SECTION 3
GASIFIER CONDITIONS
The term "gasifier" is used here to mean an assemblage of equipment
for the purpose of converting coal to gas. The range of conditions present
in gasifiers can vary widely (temperatures: room to beyond 3000°F, and
pressures: atmospheric to in excess of 1000 psi). Raw gas produced can
contain CO, C02, Hg, H20, CH4> NZ, H2$, can also contain NH3> HCN, phenols,
chlorides, organic acids, char, dolomite and flyash. The partial pressure of
hydrogen present can be high, leading to possible hydrogen embrittlement. The
dew point can be 300°F, leading to condensate attack, such as aqueous acid
corrosion and chloride stress-corrosion cracking. Finally, the hydrogen
sulfide content can be as much as 1.5 percent by volume, leading to sulfide
corrosion failure.
Table 1 presents the environmental conditions present in a number of
coal gasification processes. The moisture content can be as high as 50
percent* (Lurgi, Bi-gas), the hydrogen composition may reach 45 percent
(Consol-C02 Acceptor), the methane content is generally under 20 percent, but
*A11 percentages quoted herein are percent by volume unless otherwise
ctaf art
stated.
B-8
-------
may reach 73 percent (Hydrane), and carbon monoxide may be as high as 70
percent (Atgas).
Figures 1 through 5 represent schematics of five coal gasification
pilot plants: Consol C02 Acceptor, IGT Hygas, Bureau of Mines Synthane,
Bituminous Coal Research Bi-Gas, and the Battelie-Union Carbide Agglomerated
Ash.
The C02 Acceptor process utilizes lignite and calcined dolomite. A
slightly oxidizing condition is present in the regenerator, while reducing
conditions exist in the gasifier unit. Hygas conditions are reducing, with
a significant amount of h^S present. Synthane's f^S and moisture contents
are both high. In Bi-Gas, the moisture content is even higher, and the
temperature in the gasifier reaches 300QOF. The Agglomerated Ash pilot plant
has both a gasifier and a burner unit, with reducing and oxidizing
conditions, respectively. Two of these pilot plants have high dew points:
420°F in Hygas, and 492°F in Bi-Gas.
It is obvious that sampling conditions can vary widely from unit to unit, and
within a unit.
The gas composition varies with temperature and pressure. An example
of this is shown in Table 2. In the laboratory test program at the Illinois
Institute of Technology Research Laboratories (IITRI), the equilibrium gas
composition at 1000 psi varies considerably with temperature. For example,
the hydrogen content is 4 percent at 90QOF, 23 percent at 1500°F and 31
percent at 1800 F. Of course, this assumes equilibrium, which is not
necessarily attained in practice. Rate kinetics can be important in certain
instances. For instance, the C0/C02 equilibrium is rapidly attained in a
matter of seconds, while the CH4/H2 equilibrium is slow and takes several
minutes to achieve. Hot surfaces will accelerate reaction rates, and for
B-9
-------
TABLE 1. ENVIRONMENTAL CONDITIONS IN VARIOUS COAL CONVERSION PROCESSES
Coal
Gasification
Process3
1 Atgas
2 Bi-Gas
3 CO- Acceptor
4 Hydrane
5 Hygas
6 Koppers Totzek
7 Lurgi
8 Molten Salt
Coal Gasification
9 Morgantown
Producer
10 Synthane
11 U-Gas
12 Winkler
Raw Gas Analysis
(X)
H20
_._
48.0
17.1
—
24.4
9.6
50.2
22.6
10.0
37.1
12.0
23.1
(X)
H2
9.6
12.7
44.6
22.9
22.8
33.1
20.1
34.8
13.6
17.5
11.6
32.2
(X)
CH4
20.0
8.1
17.3
73.2
14.1
—
4.7
5.8
2.3
15.4
4.1
2.4
(X)
H2S
—
0.7
0.03
—
0.9
0.3
0.6
0.2
0.6
0.3
0.6
2500
ppm
(X)
CO
69.7
22.9
14.1
3.9
18.0
50.4
9.2
26.0
18.2
10.5
17.0
25.7
(X)
C02
—
7.3
5.5
—
18.5
5.6
14.7
10.3
7.6
18.2
8.8
15.8
(X)
N2
0.7
0.3
0.2
...
1.0
—
0.3
47.3
0.5
45.4
0.8
(X)
C2H6
—
...
0.37
—
0.5
...
0.5
—
0.5
___
Other
—
—
NH3/0.8
...
—
—
—
COS/0.1
—
...
COS/400
ppm
Exit Cond.
(°F)
Temp.
2600
1700
1500
1800
600
2750
700-1100
1700
1000
1800
1900
1500-1850
(psi)
Pres.
5
1000-1500
150-300
1000
1000
atm.
450
1200
300
500-1000
350
atm.
DO
I
Reference 23
-------
CONSOL C02 ACCEPTOR PROCESS
HEAT RECOVERY
03
I
OtVOlATIUZilt
IfN'F
IK KM
IMMTE
k
y
HEBMiRATOH
MOLE*
N2.
C02
CO
HjO
65
31
2
1.5
3MP8S
T
.JL^
[^
\
IV X-y
50
MOLE% H2 50 % HjS .017 %
CO 16
H^ 13
CH4 tO
C02 6
NH3 .5
N2
PH2
.25
143
.06
RAW CAS
TO PUHIFICATHM
AM HETHAMTMM
-a-
(TACK
UKHTOAS
rrtAm
Figure 1. Consol C02 Acceptor process (Reference 11).
-------
IGT HY6AS PROCESS
COAL
SLURRY
CXI
INi
STEAM
OXYGEN
fr
0
H
MOLE%
17
21
21
IS
25
CO
C02
CH4
H20
HyS
PH2
210
1.2
drapt-420'F
GASIFIER
H2 22 X
CO 19
C02 19
CH4 IS
H^ 23
MO LEX
NH4 as x
»•£ .89
HCN .02
PHj > 262
PHjS- 10.7
600 "f
12SO°F
1750 °F
1200 PSIG
1900°^
DRYING ZONE
LOW TEMPERATURE
REACTION ZONE
HIGH TEMPERATURE
REACTION ZONE
SYNTHESIS GAS
GENERATION
ZONE
1
MO LEX
H2 19 X
CO 17
C02 16
CH4 13
H^ 20
NH3 0.4
HCN 0.02
H2S 0.78
Oib 13
PH
232
4
WATER
QUENCH
RAW GAS TO
PURIFICATION
AND
METHANATION
w-
-^ CHAR REMOVAL
Figure 2. IGT Hygas process (Reference 11).
-------
BUREAU OF MINES SYNTHANE PROCESS
ca
co
UFTTUBE
DRIER
OXYGEN
STEAM
FEEOWATER
BOILER
o-
0-
RAW GAS TO
PURIFICATION
AND
UETHANATION
COAL
WATER
SLURRY
CHAR
Figure 3. Bureau of Mines Synthane process (Reference 11).
-------
BITUMINOUS COAL RESEARCH - BI-GAS PROCESS
-fri
6AS1FIER
03
I
STEAM
CIRCULATING
QUENCHWATER
RAW GAS TO
PURIFICATION
AND
METHANATION
COAL-WATER
SLURRY
SLAG REMOVAL
Figure 4. Bituminous Coal Research - Bi-Gas process (Reference 11).
-------
CO
CJ1
BATTELLE-UNION CARBIDE AGGLOMERATED ASH PROCESS
GASIFIER
H2 48.1
CO 26
C02 S
CH4 6
H20 14
MOLES
K NH3
«2
PH2
.002 %
0.3
0.5
• 49
' 84
X
"I
I
COAL
o-J
1800° F
100PSIG
L /
L
*
^
\
\
STEAM
ASH REMOVAL
RAW GAS TO PURIFICATION
AND METHANATION
Figure 5. Battelie-Union Carbide Agglomerated Ash process
(Reference 11).
-------
this reason hot beds are often used to speed up the shift to equilibrium.
Catalytic effects produced by the presence of such elements as nickel are
currently under investigation.
Particulates, such as char, ash and dolomite in the gas stream can
increase the carbon and sulfur activity potential. Thick deposits can also
prevent the formation of protective oxide scales. Ash components can react
with the surface scales and form condensed phases which can change the
corrosion process drastically.
The activity potential relationships for forming chrome carbides
versus chrome oxides are illustrated in Figure 6 at 1800°^ for both
atmospheric and 102 atmospheres pressure involving three hypothetical gas
mixtures at 10 percent and 30 percent CH4- The equilibria for iron and
nickel were not included for the sake of clarity, and because corrosion
protection of iron-nickel-chromium base alloys are derived from a Cr203"ricn
scale, and reactions which tie up chromium in other forms can detract from
the corrosion resistance of the alloy. The equilibrium diagram of Figure 6,
therefore, indicates the stable phases of chromium with respect to oxidizing
and carburizing potentials. The oxidizing/carburizing potentials of the gas
mixtures are represented in the diagram by the various symbols, and it is
clear that the third mixture will result in conditions conducive to carbon
deposition (carburizing) at atmospheric pressure, and to borderline
conditions at 102 atmospheres. Mixtures 1 and 2 will also be borderline at
high pressure, but conducive to protective chrome oxide formation at
atmospheric pressure. Of course, these equilibrium conditions will vary with
temperature.
The sulfide/oxide equilibrium for aluminum, chromium, iron, nickel and
cobalt are shown in Figure 7. It can be seen that in the presence of 0.1 to
B-16
-------
1 percent H?S in a hypothetical gas mixture, sulfides of iron and nickel
will form if the HUS content is above 0.1 percent and cobalt sulfide when
the HpS content exceeds 0.5 percent. (These gas compositions fall within
the sulfide phase fields of these elements, as can be seen by continuing
the horizontal lines delineating those phase fields.) Sulfides of chromium
and aluminum do not form, since gases with these particular H2S contents
fall wicnin the oxide stability region of these metals. It must be noted
that at this temperature the nickel and cobalt sulfide scales are liquid.
This fact has significant implications which will be discussed in the
section on erosion-corrosion.
TABLE 2. INLET AND EQUILIBRIUM GAS COMPOSITION OF
PHASE I CORROSION TESTS ON ERDA/MPC IITRI PROGRAM9
Gas Composition, Percent
Gasifier
Equilibrium
Component
H?
C§
CO?
CH4
NH3
H2S
H20
N2
Inletc
24
18
12
5
1
0-1.0
39-40
~
900°F
4
5
25
19
1
0-1.0
45-46
~
150QOF
23
11
19
9
1
0-1.0
36-37
—
180QOF
31
17
15
3
1
0-1.0
32-33
-
Regenerator
Inlet
1
1
30
-
_
0.1(S02)
0
68
Reference 9
3
At 1000 psi and indicated temperature
Inlet gas composition constant for all tests in gasifier gas
B-17
-------
In practice, several scale forming mechanisms compete with each
other, and the complex processes lead to multilayer scale formation whose
configuration is dependent on diffusion rates, nonequilibrium kinetics,
etc. Often there is a race between the formation of a protective oxide
scale and a damaging sulfide scale, and the outcome determines the corrosion
resistance lifetime of the alloy. Preoxidizing the alloy preconditions
it and improves the corrosion resistance lifetime of the alloy. The oxide
scales act as diffusion barriers which provide a prolonged incubation
period for such corrosion processes as the catastrophic formation of molten
metal sulfides.
B-18
-------
-2
I
-6
-8
(C
CrrC3
Cr23C6
Cr
Mutu.fi IO%CH«
M'XtUre lotmlK^otm
1 0 •
2 a •
3 A A
•IS}'
30%
latm|K
a
ffl
A
/
/
CH4
)2o1rr
O
a
^
T.
89
e
a
o
Cr203
Cr-O-C
1800 °F
—
—9
-4 •
MM
£
I
HO -
-12 _
-10
-36 -32 -28 -24 -20 -16 -12
Log Po2(atm)
-2
-8
Figure 6. Thermodynamic stability diagram for Cr-O-C (Reference 19)
B-19
-------
CO
I
8
-40
-30
log P0
-20
-10
0°
-5
-10
-15
1 I ' '
1800°F
980°C
1255°K
Solid
Liquid
A12S3
CoS
NiO
ff
10
I I
15
10
-5
Figure 7. Thermodynamic stability diagram for metal-sulfur-oxygen
(Reference 28).
-------
SECTION 4
MECHANICAL PROPERTIES AT GASIFIER CONDITIONS
At elevated temperatures, the atomic mobility is enhanced, and the
increased diffusion rates of the elements in the alloy lead to a loss in
stiffness, strength and compositional stability. Creep (strength) rather
than yield (strength) becomes the controlling mode of deformation, and
fracture is governed by stress-rupture rather than by ultimate strength
considerations. Thermal cycling can lead to fatigue failures, particularly
•*.
in conjunction with corrosion, since this leads to corrosion-fatigue
failures; the thermal cycling implies that a form of stress-cycling is
present, that is to say, there is a thermal effect on an existing stress
distribution. Of course, thermal cycling can also cause a change in the
compositional stability, in the sense that different metallurgical phases
are stable within given temperature ranges. Formation and dissolution
of phases (such as precipitates) occur at various reaction rates and influence
the mechanical properties since the alloy condition has changed. There
can be a severe loss in ductility, resulting in embrittlement of the alloy
and eventual failure. Low cycle fatigue is another form of failure which
may occur at elevated temperatures under high stress-amplitude levels;
B-21
-------
a stress of the same magnitude would result in nonsensitive amplitudes at
room temperature.
Ceramic materials have outstanding elevated temperature stability and
strength properties (tensile strength actually improves at elevated
temperatures) but are notoriously poor in their thermal fatigue properties
and crack very quickly when thermally cycled; their ductility is very low and
their notch-sensitivity very high, as they do not significantly deform
plastically at temperatures under 2000°F.
Superior mechanical properties at elevated temperatures are achieved
in metals by liberal addition of nickel. Among the metals having significant
elevated temperature resistance, ferritic alloys are generally limited to
service temperatures of lOOQOF, stainless steels to 1600°F, and high nickel
alloys to 20000F. Sigma-phase precipitate formation in high chromium alloys
containing a ferritic phase can lead to catastrophic embrittlement in service
in the temperature range 1000 to 1600°F. Chromium containing ferrite alloys
(such as 446 stainless steel) are susceptible to what is commonly called
885 F precipitation embrittlement (actual range 750 to 950°F). Both sigma-
phase and 885°F embrittlement can be avoided by redissolving the harmful
precipitates by heating to above 165QOF.
Pressure translates itself into bi-axial tension on vessel wall
components and into hydrostatic compression on internal parts. These
stresses must be added to the existing stress components (such as thermal
stress), resulting in pressure induced stress magnitudes which can approach
the yield strength of the material at elevated temperatures, causing a
significant amount of creep.
The effect of pressure on mechanical behavior appears to be minimal,
for the range of pressures under consideration. Basically, increasing the
B-22
-------
pressure is equivalent to lowering the temperature in solid state kinetics.
Reaction rates are slowed down somewhat, and hence, the alloy stabililty
is improved. Metallurgical phenomena which are significant under atmospheric
pressure may become unimportant at high pressures; the reverse may also
be true. Consequently, corrosive reactions, and hence mechanical properties
of partially corroded materials may be significantly different under high
pressure.
B-23
-------
SECTION 5
CORROSION EFFECTS
In order to understand the effects of corrosion on materials in coal
conversion processes, the basic corrosion mechanisms must be reviewed in
terms of the conditions existing in coal conversion equipment.
Many of the reactions involve gas/metal interactions which can lead to
such corrosion phenomena as hydrogen embrittlement, oxidation, carburization,
sulfidation, etc. In addition to the environmental factors, design and
fabrication characteristics of the alloys may strongly influence the
corrosion process.
Corrosion rates have been measured in laboratory tests and in pilot
plants, and the results give a significant indication of the anticipated
lifetime of various commercial alloys. The information is extended to
include refractory ceramics as these have a significant place in coal
conversion facilities, especially at the higher temperatures.
The liquid/metal interactions are also very significant. They may
occur in components exposed to moderate temperatures at which many gases
condense. They may also involve the highly significant hot corrosion
phenomenon, which can be devastating even when the damaging alkali species is
present only in minute quantities, such as parts per billion. And the
interactions may involve slagging phenomena, such as the erosion-corrosion
B-24
-------
mechanism which takes place when low melting sulfides form on the metal
surface.
Several of these phenomena can occur alternately, or even
simultaneously. Sometimes the corrosion damage is slow, extending over a
time frame of thousands of hours, and sometimes extremely rapid, occurring
over a matter of minutes. It is not uncommon for a 1/4-inch thick plate to
corrode through in less than 50 hours.
5.1 BASIC CORROSION MECHANISMS
Corrosion protection is imparted to stainless steels and superalloys
by the presence of chromium in excess of 12 percent, above which the alloys
resist attack by most mild corrodents, and do not rust in moist air. As the
chromium level reaches 18 percent, the alloys become resistant to more
aggresive corrodents, such as nitric acid. Pitting may occur in halide
solutions, general corrosion in hydrochloric and sulfuric acids, and
intergranular corrosion in sensitized stainless steels in contact with nitric
acid. As a rule, superalloys are more corrosion resistant than stainless
steels.
The chemical composition of a number of stainless steels and super-
alloys is tabulated in Table 3. These can be roughly classified as iron
base, cobalt base or nickel base alloys.
Corrosion may involve an electrochemical cell mechanism of anodic
attack. Consequently, the activity level of a surface can be reduced to a
less anodic (more cathodic) condition by passivation with nitric acid (6 to
15 percent solution of 70 percent HNOa for 1/2-hour). This strongly
oxidizing treatment builds up an oxide film on the surface. However,
oxygenating agents must be present and replenished constantly to maintain
passivity, otherwise, localized corrosion frequently occurs (as in crevices).
B-25
-------
For instance, hydrogen reduces passivity drastically, and may even destroy
it. It is also possible to overpassivate, which leads to increased
susceptibility to pitting corrosion.
Pitting corrosion results from a localized breakdown of the oxide
film, often under stagnant flow conditions, and even more often under unclean
conditions. Impurities can form small local galvanic cells which may act as
pit initiators.
Galvanic corrosion occurs when two dissimilar metals are joined. The
more noble metal will be protected by the less noble one which gets consumed.
This is a particularly severe problem when a massive noble metal is joined to
a much smaller less noble metal. A high degree of localized corrosion will
then ensue. One example of this occurs when a weld metal is less noble than
the parent metal.
Crevice corrosion occurs;in restricted areas where an occluded cell
can develop, such as at a metallic joint, under scale deposits, etc. Oxygen
becomes depleted inside the crevice creating an anode — while outside, the
presence of oxygen creates a cathode. Since oxygen cannot diffuse readily
into the crevice, the site remains active, and rapid corrosion is promoted by
the large cathode (i.e., the rest of the metal surface).
Environments containing moisture and halides can cause pitting and
crevice corrosion. This can occur when chlorides are present even in ppm
quantities. Such chlorides may be found in water and steam, in coal, and
even in insulation. One such example occurs by the thermal decomposition of
the phenolic resin binder in fiberglass insulation.
Severe general corrosion occurs when either hydrochloric or dilute
sulfuric acid is present in the absence of an oxidizer.
B-26
-------
TABLE 3. BASIC CHEMICAL COMPOSITION OF SELECTED ALLOYS9
Composition, percent by weight
Fe Ni Cr Co Al Mo
Other
304
316
310
446
In-600
In-601
In-800
In-793
In-671
Stellite 68
Haynes 188
Multimet N155
RA-333
HL-40
Armco-21-6-9
Inconel 625
Incoloy 825
Haste lloy X
Armco 22-13-5
Co-Cr-W No. 1
617
AL29-4-2
Thermal loy 63
HK-40
FSK-414
70
65
52
75
7
16
47
43
2
1
29
16
47
63
3
30
19
57
63
38
47
2
9
14
20
76
60
31
32
49
2
23
20
48
19
7
62
41
46
13
54
4
35
20
11
19
17
25
24
16
23
21
21
50
29
23
22
26
31
21
22
22
22
22
30
22
29
26
28
30
56
36
20
3
3
55
13
48
1
2
1
1
3
4
9
3
9
2
1 9
4
1
2 Mn
2 Mn
2 Mn, 1 Si
1 Mn
1 Mn
7 W, 1C,
15 W, 1 Mn
4 W, 1 Cb,
3 W, 2 Mn,
1 Si
8 Mn, 1 Si
4 Cb
2 Cu, 1 Ti
1 W, 1 Mn,
5 Mn
12 W, 3 C
2 Mn, 2 Si
7 W, 1 Mn,
1 Mn
1 Mn
1 Si
1 Si
1 Si
Reference 9
B-27
-------
Severe general intergranular attack can occur in noncorrosive environments if
grain boundary diffusion of readily soluble elements such as sulfur has
occured at high temperature during the fabrication process.
Sensitization in another problem encountered in alloys containing
chromium: the intergranular precipitation of chromium carbide in austenitic
stainless steels (such as 304 or 316) in the service temperature range 800 to
1500QF depletes the alloy of the chromium so necessary for protection from
corrosive attack. Sensitization can be avoided by reducing the carbon
content of the alloy (304L or 316L) or by stabilizing the alloy with strong
carbide forming elements such as titanium, tantalum or columbium (321, 347 or
348 stainless steels). Desensitization can be achieved by heating above
160QOF to redisolve the carbides; it is generally done at 1900 to 2100 F, and
must be followed by rapid cooling. It is important to remove residues of
cleaning solutions used for degreasing, etc., as these can cause severe
corrosion upon heating. It is also important that welding operations be
rapid, otherwise the metal adjacent to the weld may become sensitized. For
this reason, a desensitization post-weld treatment is recommended for
critical components.
Knifeline attack is also associated with Sensitization related
corrosion. Here, a very thin layer of metal immediately adjacent to the
fusion line of a weld is preferentially corroded in an intergranular manner.
Crater corrosion may show up at the stopping point of weld beads. End-
grain corrosion is similar to crater and knifeline attack. All three
usually occur only in strongly oxidizing environments. Knifeline attack
can be eliminated by reheating the weldments to 19500F and then quenching.
End grain attack can be avoided by not exposing the ends of bar stock
or tubing, or the parting lines of forgings to corrodents. If this cannot
B-28
-------
be done, the exposed ends should either be fused or covered with weld metal
deposit.
Chloride or caustic solution induced cracking can occur even when no .
stress is present, in which case nondirectional random cracking may occur
intergranularly. However, chloride-corrosion cracking is usually considered
to be one of the corrosion modes associated with stress-corrosion cracking
(SCC). Intergranular failure, when it occurs, consists of characteristic
branching normal to the stressed direction. The mechanism involves the
depletion of chromium next to the grain-boundaries. (Chromium carbide
precipitates in the grain boundary region enhance corrosion activity, but are
not necessary to the mechanism, since grain boundaries, by their nature are
more chemically active than the grains.) Failure occurs along the grain
boundaries. Or, failure can occur transgranularly, again with the cracks
forming normal to the stressed direction. The source of the stresses can be
from such service conditions as thermal gradients, thermal cycling, etc. or
from fabrication factors (such as bending and forming), surface finish
(residual grinding stress, etc.), stress raisers (notches), thermal
processing (welding), or fitting and assembly. Furthermore, prior cold
working reduces resistance to corrosion.
Chloride stress cracking usually occurs transgrariularly under stress
at temperatures above 160 F with only a few ppm of chlorides. Austenitic
stainless steels are particularly susceptible. Heat transfer intensifies
stress corrosion problems, probably by increasing the concentration of
chlorides at the metal surface. Nitrogen enhances stress-corrosion cracking
in the presence of magnesium chloride. Increasing the nickel content of the
alloy beyond 10 percent improves the resistance of the alloy to chloride
cracking.
B-29
-------
Caustic environments also contribute to stress corrosion cracking of
austenitic stainless steels, but only at temperatures near or above the
boiling point at ambient atmospheric pressure, and only with concentrated
caustic solutions. Again, failure can be either transgranular or
intergranular.
Intergranular stress corrosion cracking of austenitic stainless steels
also occurs in polythionic acids (H2OgS2-5), especially when the steels are
sensitized, but transgranular cracking of nonsensitized steels has been
observed. Proper care to prevent the entry of moisture during shutdown
prevents formation of polythionic acid.
Stress relieving of tensile stresses is helpful in preventing stress
corrosion cracking. The introduction of compressive residual stresses at
surfaces (such as by shot-peening or cold reduction) is also helpful. In the
case of chlorides, treating the water or steam helps prevent damaging
chloride concentrations. Sensitized 304 stainless steel fails readily at
180°F when stressed in water containing 70 ppm chlorides, but only 2 ppm
chlorides are required to crack desensitized stainless steel at 390op. While
100 ppm chlorides can crack sensitized stainless steel at room temperature,
only 2 ppm of fluorides are necessary to crack the same alloy at room
temperature. Specific ions known to cause stress-corrosion cracking are
listed in Table 4.
5.2 GASEOUS CORROSION REACTIONS
The gaseous equilibrium will determine the type of reaction which may
occur with each element present in the alloy under set pressure and
temperature conditions. The reaction may be oxidizing, carburizing or
sulfidizing. Other gaseous elements (Cl, F) may also react, and hydrogen can
react with stressed metal. Furthermore, the condition of the metal surface
B-30
-------
TABLE 4. SPECIFICATIONS
PRESENT AT LOW
Damaging Ion
Fluoride
Gaseous chlorine
Gaseous HC1 and HBr
Hal ides in aqueous
solutions
02 dissolved in liquid H20
Gaseous hydrogen
at ambient pressure
at high temperature
and pressure
Hydroxides (LiOH, NaOH,
KOH)
Gaseous H2o
Polythionic acids
H2S gas
MnS inclusions
S02 gas with moisture
Nitrogen oxides with
moisture
Ammonia in aqueous
solutions
Nitrates in aqueous
solutions
N, P, As, Sb, Bi impurities
in alloys
As, Sb, Bi in aqueous
solutions
Carbonates in aqueous
solutions
C0-C02-H20 9as
Lead in aqueous solutions
KNOWN TO CAUSE STRESS CORROSION CRACKING WHEN
CONCENTRATIONS AND AS IMPURITIES*
Susceptible Alloy
Sensitized austenitic stainless
steels
Low alloy-high strength steels
ibid
Austenitic stainless steels
Sensitized stainless steels
Low alloy-high strength steels
Low strength, medium strength
steels
Carbon steels, Fe-Ni-Cr alloys
(caustic cracking)
High strength aluminum alloys
Sensitized stainless steels and
Inco 600
High strength-low alloy steels
High strength steels (crack site
nucleation)
Copper alloys
Copper alloys
Copper alloys
Carbon steels
Stainless steels (with
Cl~ present)
High strength steels
(hydrogen cracking)
Carbon steels
Carbon steels
High nickel alloys
Temperature
room
room
room
hot
570°F
room
>390 F
>210°F
room
room
room
room
room
room
room
>210°F
room
room
2100F
Reference 24
B-31
-------
often controls the reaction with gases (or liquids), but refractories such as
ceramics react quite differently from metals. All these reactions will be
discussed in the following sections and corrosion data for alloys subjected
to coal conversion gas environments will be presented.
5.2.1 Hydrogen Embrittlement
Stress is necessary for the occurence of hydrogen embrittlement.
Atomic hydrogen enters high strength steels and forms bubbles at inclusions
(Figures 8a and 8b). The higher the stress, the more hydrogen remains in the
steel instead of diffusing through the steel. A decarburization process
follows with the cementite (FegC) in the steel decomposing to form methane
(CH/j) by combining with the hydrogen. The resulting fissures crack the steel
(Figure 8c). Austenitic steels are not generally subject to hydrogen
embrittlement; the occurence is favored in ferritic steels, since carbides
which can readily dissociate are present. The hazards of hydrogen
embrittlement can be avoided by making use of the API's Nelson Curves in
selecting steels for service in hydrogen environments.
5.2.2 Oxidation
Superior oxidation resistance is found in alloys containing a
significant amount of chromium, such as alloys In-601, In-617, Haynes
188, Multimet N155, 446 S.S., 310 S.S., RA-333, HL-40 and Thermalloy 63.
These types of alloys are basically oxidation resistant to 20000F. Above
that temperature, some dispersion strengthened nickel base alloys are
usable to 23000F, but their room temperature ductility is poor. Coatings
such as silicides, aluminides, etc. offer additional resistance to oxidation.
Of course, cooling the alloys by water-jacketing, or by other means to
the degree that chrome-moly steels and monel-clad or even plain carbon
steels become usable, may present a viable alternative. Ceramic insulation
B-32
-------
Figure 8a
Magnification: 5000x
Figure 8b
Magnification: lOOOx
Hydrogen bubbles at manganese sulfide inclusions in high strength steel
Figure 8c
Crack formation resulting from junction of the fissures formed in high
strength steel. Magnification: 500x.
B-33
-------
presents yet another choice. However, thermal cycling tends to crack the
ceramic, and the component must be designed so that hot spots do not develop
as a result of these fissures (flow paths must be periodically blocked). The
best oxidation resistant materials are oxide ceramics such as alumina or
zirconia.
Actually, most of the existing commercial alloys have been designed
for oxidation resistance. Thin, adherent protective oxide coatings are
formed (thick oxide films tend to spall off). However, there is a thermal
expansion differential between the metal and the oxide, and the mismatch is
greater for austenitic than ferritic steels, which can lead to cracking of
the protective oxide barrier. The thermal expansion mismatch is low in
superalloys: In-600, In-701, Hastelloy X and RA-333. Stainless steel
service limits are 1650°F for 304, 316, 321 and 347; 2000°F for 446, and
2100°F for 310.
5.2.3 Carburization
Hydrocarbon environments such as methane may cause carburization. It
may be avoided if a strong chromium oxide film has formed on a smooth
surface. Carburization occurs readily in low oxygen regions such as
crevices, surface flaws, etc., or inside intergranular corrosion fissures in
sensitized stainless steel, and in existing corrosion pits. It can be
counteracted by the use of alloys of high chromium content, or by the
addition of steam (oxidizing) or sulfur (sulfidizing).
Chromium carbides form during the carburization process, leading to a
loss of ductility and eventual brittle fracture. This carbide formation
depletes the chromium in the metal and leads to oxidation attack of the grain
boundaries. Both CO and C02 can contribute to carburization.
B-34
-------
Nickel can cause dissocation of methane at 1650°F, leading to carbon
deposition. The carbon acts as a reducing agent, and can combine with the
oxygen of the protective oxide, thereby permitting sulfides to form.
In atmospheres depleted of oxygen, pitting can result from
graphitization of the chromium depleted matrix, with iron-nickel and graphite
powders forming upon extended exposure.
Metal Dusting consists of catastrophic localized carburization
in the temperature range 900 to 200QOF which may cause severe pitting
corrosion. This can be avoided by adding sulfur or excess hydrogen or
CO, or by reducing the presence of hydrocarbons and methane. It can be
intensified by alternating reducing and oxidizing cycles.
5.2.4 Sulfidation
Sulfidation is a very significant problem in coal conversion. For
example, a steel containing 9 percent chromium and 1 percent molybdenum
subjected to a 6 percent H2S content in H? gas at 900°F and 2000 psig will
suffer a 128 mil loss of metal, while an In-800 alloy (20 percent chromium)
under identical conditions will lose less than 6 mils. At 1800°F in an
atmosphere containing 0.1 percent H^S, 316 stainless steel suffered a 480 mil
loss in 1000 hours, while an In-800 alloy lost only 35 mils under identical
conditions. Aluminizing the In-800 alloy further reduced the loss to 14
mils. Alloys that are most resistant in a reducing environment such as h^S
are not necessarily equally resistant to oxidizing environments such as S02-
Here, the 25 percent chromium containing alloys such as 310 stainless perform
best. At 185QOF in an atmosphere containing 0.13 percent S02» 31° stainless
lost only 13 mils in 1000 hours compared to In-800 alloy which lost 71 mils
under identical conditions.
B-35
-------
Chromium oxide scales are not generally protective in a reducing
environment. They allow diffusion of the metal ions outward and sulfur
inward. SuIfides form both above and below the oxide layer, and the sulfides
below destroy the oxide scale, as it will no longer adhere to a metallic
base. Thermal expansion differentials from thermal cycling speed up the
destruction process.
High sulfur contents accelerate nickel sulfide formation. Because the
nickel-nickel sulfide eutectic is molten at 1193°F, catastrophic attack will
occur. To avoid nickel sulfidation at 1800°F, the H2S/H2 ratio must be very
low (less than 1/300). This corresponds to no more than 0.1 percent H2$ at
1000 psi with a hydrogen content of 33 percent. Below this critical ratio,
sulfidation can only occur by H2S or S02 transport, and not by metal ion or
sulfur transport. Consequently, nickel sulfidation does not then occur at a
practical rate.
If enough oxygen is present in the gasification atmosphere, then a
protective chrome oxide scale is stable, and nickel sulfide does not form.
A moderately protective oxide scale will also form when sufficient
steam is present. With 310 stainless a 7.5 percent H20 content is required
with 1 percent H2S; or a 20 percent H20 content is required with 1.5 percent
H2$. In-800 has a higher nickel/chrome ratio and requires more water: 15
and 27.5 percent, respectively. Kanthal (Fe-Cr-Al) does not contain nickel
and is not easily attacked by high H2S, low H20 combinations at 15000F.
A high sulfur content (3 percent), such as is present in Illinois
No. 3 coal, limited the operation of a FBC operated at the Coal Research
Establishment in Great Britain to below 1600°F. This combustor operates with
10 to 20 percent air at low flow velocities of 3 ft/sec. While 16 alloys
B-36
-------
tested well at 1550°F for 2000 hours, all performed poorly at 1650°F. The
best performer at 1650°F was 310 stainless steel.
Catalyzers are known to have a significant effect on sulfidation. In
one experiment, the presence of calcium oxide in a sulfidizing atmosphere
accelerated the corrosion attack very significantly.
Because liquid metal sulfide formation occurs at 1193°F with nickel,
161 lop w-jth cobalt, and 1810°F with iron, there have been attempts to develop
new sulfidation resistant alloys: Kanthal (78Fe-16Cr-5Al) does not form
protective oxide scales, and molybdenum base alloys are effective only if
protected from oxidizing atmospheres.
The evidence indicates that there is no foolproof means to avoid
sulfidation in stainless steels and in superalloys at temperatures at which
sulfide slags will form with basic elements of the alloy. This is because
fully protective chome oxide scales do not develop. Of course, these scales
are most protective in highly oxidizing atmospheres, and least protective in
reducing atmospheres. However, cobalt sulfide and iron sulfide slags do not
form at 1500°F, and the formation of nickel sulfide slag is relatively slow
(if the H2$ content is below 0.5 percent) except in regions of high activity
such as at crevices and adjacent to or under char deposits or where
impurities acting as catalysts activate the reaction. This does not mean
that sulfidation will not occur, but only that severe slag-type sulfidation
may be avoided or delayed; the metal will gradually convert into corrosion
products. But if the temperature is raised to around 1800 F, both cobalt
sulfide slags and iron containing sulfide slags will form in reducing
atmospheres, and sulfidation will rapidly become catastrophic. In contrast,
sulfates rather than suIfides form in highly oxidizing atmospheres, and
B-37
-------
corrosion is not catastrophic. However, sulfidation can be rapid in
moderately oxidizing atmospheres of high sulfur content.
Aluminizing (coating the surface of the alloy with an aluminum
compound) will increase the incubation period for sulfidation to occur.
A pre-oxidation treatment helps form a protective oxide scale, and
thereby increases the sulfidation resistance of chromium containing alloys;
such pretreatment is practically mandatory.
5.2.5 Reactions With Other Gases
Little is known about the high temperature resistance of stainless
steels and superalloys to other gaseous products. Except as noted, methane,
ammonia, hydrogen, carbon dioxide and carbon monoxide do not produce
corrosive reactions with these alloys, but have a secondary role in reactions
involving sulfur, carbon and oxygen.
Gases such as chlorine and fluorine are highly damaging to metals at
temperatures in excess of 1000°F, but the ppm content threshold is not known.
It is suspected that even a few hundred ppm of these gases are strongly
damaging at 1500 F. Alloys with a high nickel content tend to be more
resistant to these gases. A high chromium content enhances surface
reactions.
5.2.6 Metal Surface Conditions (Fabrication Factors)
The state of the metallic surface is of vital importance for
developing corrosion resistance. This is clearly apparent when corrosion
resistance is expressed in terms of the existence of a diffusion barrier
consisting of a tough, adherent, protective oxide scale. It is necessary to
avoid regions of high chemical activity, such as are present in the vicinity
of char or bed materials, or at surfaces containing crevices, scratches,
holes, etc., where protective oxides are hard to grow. Not only will the
B-38
-------
carbon and sulfur activity be changed at such sites, but the equilibrium gas
composition can be affected locally.
Since cold work enhances diffusion, it is best to remove it by
annealing. Even stencil marks can be corrosion sites. Therefore, only a
full anneal is satisfactory, and further cold work (such as polishing, final
machining, etc.) must not be performed. Such an alloy has been desensitized
by the annealing treatment. Therefore, any welding must be performed prior
to annealing, or the alloy may become sensitized. Pickling and passivating
treatments are best avoided, otherwise grain boundary trenches can be opened
up by excessive acid pickling. It may be best to leave the mill scale
intact. If scale must be removed, only methods that do not create surface
defects should be used.
5.2.7 Coal Gasification Gaseous Corrosion Data
Considerable data have been accumulated since 1973 in the ITTRI-MPC-
ERDA program on screening alloys for coal gasification in the laboratory and
in pilot plants.
Alloys such as In-800 and In-671 have successfully performed for 7000
hours in the laboratory at 1800°F and 1000 psi in a simulated coal gas
containing 0.5 percent H2S. But both these alloys did poorly in the Consol-
CO? Acceptor Pilot Plant, and in erosion-corrosion tests; calcium oxide
is present in these environments and may be responsible for this result.
Coating the In-800 by aluminizing provided approximately 2000 hours protection,
after which its behavior was considerably poorer than that of the uncoated
alloy. The long term data indicate that extrapolation is unsafe in most
cases, and that the computed annual rate is actually too conservative.
(A sudden shift in the corrosion rate can occur at any time.) Only the
B-39
-------
uncoated In-800 alloy had a long term corrosion rate lower than the rate
extrapolated from 1000 hour data.
Tables 5, 6 and 7 present laboratory corrosion data for 11 alloys at
1500, 1800 and 1850°F, respectively. Four alloys have a projected corrosion
rate of less than 10 mils per year at 1500op for up to 1 percent H2$ in the
gas. These are 310 , 446, In-671 and In-800 aluminized. The situation is
quite different at 1800°F: few alloys have a corrosion rate of less than 20
mils per year, even without H2S present, in which only five alloys corrode at
less than 20 mils per year. When 0.1 percent H2$ is present, the number of
acceptable alloys is reduced to three (310 aluminized, In-800 aluminized, and
In-671). The number drops to two with 0.5 percent \\2$ (In-800 aluminized and
In-671) and to only one with 1 percent H2$ (In-800 aluminized). However,
when the atmosphere contains 0.13 percent S02, the situation changes
somewhat. Although three alloys (310, 310 aluminized, and In-800 aluminized)
again meet the acceptability test, alloy In-671 is replaced by 310 and the
310 alloy does better than the In-800 alloy. Consequently, 310 alloys are
recommended for oxidizing conditions and either In-671 or In-800 for reducing
conditions.
When comparison is made with pilot plant data, as in Tables 8 and 9,
these conclusions remain basically unchanged, although the corrosion rates
appear to be somewhat higher in a plant environment. Both In-671 and 310
stainless steel showed up well in the pilot plant oxidizing environment at
1500°F but not at 1700°F.
The corrosion-errosion data of Table 8 points out that a catastrophic
corrosion rate occurs when slagging takes place. The corrosion rate
increases by two orders of magnitude.
B-40
-------
TABLE 5. CORROSION RATES IN STAINLESS STEELS AND SUPERALLOYS
SIMULATED REDUCING COAL GASa AT 150QOF AND
1000 PS I. LOSS OF SOUND METAL IN MILS PER YEAR5
(BASED ON 1000 HOUR DATA)
IN
COMPOSITION
Alloy
304
316
310
310 aluminized
446
600
601
671
793
800
800 aluminized
0.1% H2S
16
21
4
13
2
19
25
4
13
24
4
0.5% H2S
46
36
3
11
3
>1000
>1000
2
21
17
5
1.0% H2S
39
41
4
11
8
»
657
6
15
15
10
Composition shown in Table 2,
Reference 9.
B-41
-------
TABLE 6. CORROSION RATES IN STAINLESS STEELS AND SUPERALLOYS IN
SIMULATED REDUCING COAL GASa AT 180QOF AND
1000 PSI. LOSS OF SOUND METAL IN MILS PER YEARb
(BASED ON 1000 HOUR DATA)
Alloy
304
316
310
310 aluminized
446
600
601
671
793
800
800 aluminized
0% H2S
219
115
17
17
194
38
16
19
36
24
13
COMPOSITION
0.1% H2S
549
480
159
16
84
68
26
16
36
35
14
0.5% H2S
584
365
34
43
25
>1000
62
19
123
164
20
1.0% H2S
>1000
>1000
164
21
20-254
>1000
26-75
17-28
>1000
67-876
12
Composition shown in Table 2.
Reference 9.
B-42
-------
TABLE 7. CORROSION RATES IN STAINLESS STEELS AND SUPERALLOYS IN
SIMULATED MILDLY OXIDIZING COAL GAS3 AT 1850°F
AND 150 PSI. LOSS OF SOUND METAL IN MILS PER YEARb.
(BASED ON 1000 HOUR DATA)
Alloy
304
316
310
310 aluminized
446
600
601
671
793
800
800 aluminized
Composition
0.13% SO?
313
891
13
13
>1000
885
51
25
79
71
17
Composition shown in Table 2.
Reference 9.
B-43
-------
TABLE 8. COMPARISON OF CORROSION OF SELECTED ALLOYS IN LABORATORY
AND PILOT PLANT EXPOSURES AT 185QOFa
Loss of Sound Metal in Mils Per Year
Alloy
310
310 aluminized
800
800 aluminized
671
446
Laboratory
Corrosion^
Test
13
13
71
17
25
1100
Laboratory
Corros i on-Eros i onc
Test
3000
3900
4700
3700
1200
1800
Pilot Plant
Consol-C02 Acceptor^
Regenerator Test
64
65
218
65
120
31
Reference 9.
'185QOF-1000 hrs-150 psi. Gas: 70% Ng, 27% C02, 3% CO, 0.13% S02
:18500p-50 hrs-14.6 psi. Gas: 33% H2, 18% CO, 14% C02, 2.5% CH4, 1%
0.5% H2S, 31% H20, 100 fps velocity (with
char loading)
^UOOOF-1800 hrs-150 psi. Gas: 70% N2, 27% CO, 3% CO, Trace H2S
B-44
-------
Additional data obtained on superalloys under the conditions of
Table 6, indicated that only the superalloys shown in Table 10 may be
acceptable in reducing environments at 1800°F.
TABLE 9. COMPARISON OF LABORATORY AND PILOT PLANT CORROSION TESTS
IN CONSOL-C02 ACCEPTOR REGENERATOR ENVIRONMENT3
Loss of Sound Metal in Mils Per
Alloy
304
310
21-6-9
22-13-5
446
800
793
825
Hastelloy X
671
310 aluminized
800 aluminized
Laboratory
Corrosion
Test
313
13
—
—
>1000
71
79
—
25
13
17
Pilot
Plant0
Test A
645
64
78
85
31
218
149
753
438
>1000
65
65
Pilot
Plantd
Test B
>1000
412
199
>1000
__
>1000
317
>1000
>1000
>1000
80
79
Year
Pilot
Plant6
Test C
38
__
32
>364
145
177
—
_ _
--
92
121
Pilot
Plantf
Test D
19
2
34
26
*m —
14
16
30
8
2
30
31
a Reference 9
b 185QOF-150 psi-1000 hours
c >17000F-150 psi-800 hours
d >17000F-150 psi-1600 hours
e >1700°F-150 psi-1127 hours
f 150QOF-150 psi-2390 hours
Gas: 70% N2, 27% C02, 3% CO, 0.13% S02
Gas: 70% HZ, 27% C02, 3% CO, Trace H2S
Gas: 70% N2, 27% C02, 3% CO, Trace H2S
Gas: 70% N2, 27% C02, 3% CO, Trace H2S
Gas: 48% H2, 23% H20, 12% CH4, 8.5% CO,
6% C02, 2.5% N2
B-45
-------
TABLE 10. ACCEPTABLE ALLOYS FOR SERVICE IN COAL GAS ENVIRONMENTS
AT 18000Fa
0.1% HgS Q.5% H2S 1.0%
Haynes 188 Haynes 188 Haynes 188
Haste Hoy X Haste Hoy X Haste Hoy X
Multimet N-155 Multimet N-155 Multimet N-155
Stellite 6B Stellite 6B Stellite 6B
Co-Cr-W No. 1 Co-Cr-W No. 1
AL29-4-2 AL29-4-2
800 aluminized 800 aluminized 800 aluminized
Inco 671 Inco 671
310 aluminized
Reference 9
Recent data at 1.5 percent H2S indicate that most alloys of Table 10
are unsatisfactory at high sulfur contents. Only Stellite 6B was fully
satisfactory, while Inco 671 and 310 stainless were marginally acceptable at
ISOOop .
The following conclusions can be drawn on the basis of laboratory
tests and limited field data:
1. The addition of at least 25 percent chromium is necessary to
obtain corrosion resistance in gasifier gas for 1000 hours
at 18000F.
2. A safe alloy for extended use in high sulfur coal gas at
temperatures above 1600°F does not exist.
3. Only aluminized 800 or 310 alloys are usable for 1000 hours above
1700°F in an S02 environment. Alloy 310 may be used
satisfactorily below 1600°F.
4. On the basis of limited data, it appears that Kanthal (Fe-Cr-Al)
is an acceptable alloy below 1600°F.
B-46
-------
5.2.8 Refractories
Refractories are used in diverse environments consisting of
temperatures up to 2000op in dry ash gasifiers, and in excess of 3000 of in
slagging reactors; pressures up to 1200 psi; gaseous environments which
include H20, H2, H?S (or S02, $03), CO, C02, Cfy and various hydrocarbons;
and erosive velocities ranging between 5 and 100 ft/sec.
Potential problems include:
1. Gaseous corrosion - attack by steam, H2, CO, C02> and alkali
vapors in decreasing order
2. Slag erosion and corrosion
3. Erosion and abrasion by high velocity particles
4. Thermomenchanical failures leading to hot-spots
5. Attack by acids such as HC1 and H2S04
6. Attack of iron impurities in the ceramic by CO below 18500F
7. Leaching out of silica by steam, and
8. Oxidation of the bond in silicon carbide bonded ceramics
Information to date indicates that high pressure steam is detrimental
to the strength of dense, high alumina (95 percent) concretes at temperatures
above 500op. On the other hand, 50 percent alumina concretes do not appear
to be affected. High temperature (1800°F) erosion of concrete is actually
less than the erosion occuring at room temperature. Additionally, 90 percent
phosphate bonded aluminas of high density have been found to be more stable
in steam/CO atmospheres than cement (calcium aluminate) bonded aluminas of
the same density. Superior slag erosion resistance is obtained with fused
cast spinel type refractories
B-47
-------
5.3 LIQUID CORROSION REACTIONS
Chlorides in ppm quantities can cause stress corrosion cracking only
in aqueous solutions. It appears, therefore, that liquid phase corrosion can
cause some alloys to fail even though they do not exhibit any corrosion
problems in gases at high temperatures.
Thus, at temperatures below the dew point, (acid) liquid phase
corrosion can be extremely serious. However, the gaseous environment may
still be controlling: For instance, chloride stress cracking has not been
observed in reducing environments.
Liquid phase corrosion in the form of slag occurs in both hot and
slagging corrosion. A severe case of slagging corrosion occurs in combined
erosion-corrosion phenomena. However, the most commonly observed liquid
phase corrosion occurs at moderate temperatures.
5.3.1 Moderate Temperature Reactions
Corrosion studies have also been conducted at coal gasification quench
phase conditions in the IITRI-MPC-ERDA program. It was found that at
moderate temperatures in laboratory experiments, corrosion was more severe in
liquid than in gas. The stress corrosion failure of 410 stainless at 462°F
and 1200 psi occurs in 50 hours in liquid and 150 hours in gas (50 percent
failures). Corrosion data also obtained at 4620F and 1200 psi in CO, C02>
H2$, NH3, H2, CH/j, N? gas and in HCN, phenol, chlorides, toluene and H20
liquid derived from saturating water with the gas at the quench conditions,
are described in Table 11.
B-48
-------
TABLE 11. COMPARISON OF CORROSION RATES OF STEELS IN LIQUID AND GAS
QUENCH PHASES*
(Mils per Year)
Alloy Aqueous Corrosion Gaseous Corrosion
Carbon Steel (Ferritic) 160 85
410 Ferritic Stainless Steel 300 150
304 Stainless Steel 9 6
316 Stainless Steel 6 5
a
Reference 9
It is evident that the austenitic stainless steels have a much lower
corrosion rate than the ferrite alloys.
Other evidence obtained in these corrosion studies indicate that
temperature is a key factor in these tests and that pitting occurs in the
heat-affected weld zone in both 304 stainless and carbon steels.
Low temperature exposures were also carried out in the pilot plants.
The annual aqueous corrosion rate in carbon steel is as high as 150 mils
compared to about 8.5 mils in 304 stainless and 6 mils in 316 stainless.
The removal of residual stresses by stress relieving often helps
reduce such corrosion problems as chloride stress cracking. Alloys containing
in excess of 20 percent nickel combat most forms of liquid phase corrosion
quite effectively. Specific acids such as sulfurous can be countered
by adding molybdenum to the alloy, while boiling nitric acid is resisted
by the addition of chromium. Getter ing agents, such as columbium, tantalum
and titanium are useful in neutralizing harmful alloying impurities.
A number of specific corrosion problems in the liquid phase are listed
in Table 4.
B-49
-------
The problem of condensates in coal conversion facilities is often
solved by a "wet wall" technique, such as cladding cabon steel or chrome-moly
steel with monel (a copper-nickel alloy) which effectively resists corrosion
at moderate temperatures.
5.3.2 Hot Corrosion
Chlorine acts as a carier for harmful low melting alkali metals such
as sodium and potassium, or for transition metals such as vanadium. The
result is slagging hot corrosion sulfate deposits at elevated temperatures,
especially in the 1000-1800°F range. Sodium impurities can be a serious
problem in quantities exceeding 40 parts per billion. Sulfates can be
gettered by adding hafnium, binding the sulfur and thus avoiding this form of
coal ash corrosion which can cause severe clogging of tubes, even in amounts
below 50 ppm. While vanadium can be gettered by magnesium salts, the
chlorine level can be reduced by preheating the coal.
Typical alkali contents of coal are of the order of several hundred
ppm, and the sulfur content of coal can exceed 3 percent. Sodium sulfate
forms a eutectic with potassium sulfate at 1526°F. Sodium chloride and
potassium chloride react with S02 to form molten sulfates and free oxide
ions, leading to a fluxing reaction on the metal surface, followed by
spalling which removes the protective oxide scale, thus catastrophically
destroying the metal. This is often called a "sticky ash" problem.
Cobalt-based alloys are less susceptible to hot corrosion than
nickel-base alloys.
5.3.3 Erosion-Corrosion
Erosion-corrosion test results from the IITRI-MPC-ERDA program
indicate that slagging erosion can virtually destroy 1/4 in. thick high
temperature alloys in 50 hours. Aluminide coatings provide a significant
B-50
-------
Degree of temporary protection, subject to an incubation period whose
duration has not as yet been measured.
Sliding erosion-corrosion is caused by the impacting gas and char
stream which is repeatedly peeling away the liquid phase from the metal
surface, thereby intermittently exposing unoxidized metal to corrosion.
Thus, the corrosion rate is increased even more than the slagging corrosion
in which the slag is not peeled away. Normally, without erosion, a corrosion
rate initially starts steeply linear with time, and rapidly levels off
parabolically as a protective oxide scale builds up. But when the scale is
prevented from building up by erosion, the rate remains steeply linear.
Thus, when such a diffusion barrier is removed, the parabolic rate reverts to
its former steep linear slope, and this causes catastrophic corrosion.
Another factor enhancing the corrosion rate is that of impacting
erosion increasing the true metal surface, thereby increasing the metal
reactivity.
One solution to erosion-corrosion is to hard face the metal with a
castable ceramic. While this will not prevent corrosion, it will protect the
oxide scale from erosive attack, and hence, the corrosion rate will be
increased by erosion. Such a barrier is equally effective in reducing
both hot corrosion and slagging corrosion, since small particle deposition
(such as flyash) which interacts with protective oxide barriers can no
longer take place on the metal surface. Of course, aluminizing a metal
provides this type of barrier protection, although the thickness (and
the lifetime) is significantly lower.
The erosion problem caused by char can be resolved either by adjusting
the carbon reaction to burn additional carbon, or by char removal.
B-51
-------
SECTION 6
EROSION
As stated previously, erosion itself increases the erosive wear by
increasing the true surface area, and hence, the mechanical abrasion.
It is known that the angle of attack in particulate impact is perhaps
the most important factor in erosive wear. The erosive rate is highest in
brittle materials with a 90o impact angle, while in ductile metals, glancing
angles of the order of 20° produce the higher erosion rates. This angle
decreases to about 10o with very ductile metals, or with metals coated with
thick soft corrosion products, or again with moderately ductile metals at
high temperature (temperature softens metals and reduces their elastic
moduli).
Other variables affecting erosive wear include particle velocity,
particle flux, surface hardness, and particle size. Erosion is approximately
proportional to the particle velocity to the two and one-half power for
ductile metals, and to the four and one-half power for brittle materials for
velocities up to 600 feet per second. Erosive wear appears to be
proportional to grain loading, and inversely proportional to the impacted
intrinsic hardness. (Hardening a steel does not improve the erosion
B-52
-------
characteristics.) Because of its spongy nature, char will not cause as much
erosive damage as flyash, since it will tend to crush on impact.
Particle size has a significant effect on erosive wear. For a given
gas velocity, particle velocity decreases with increasing particle size.
Particles, greater than 3 microns impact and cause significant erosive
damage. Smaller particles in the 0.3 to 1 micron range tend to deposit on
surfaces and cause slagging corrosion damage. Sufficiently small particles
tend to remain in the gas stream.
Erosive damage is essentially proportional to particle size in the 3
to 12 micron range. Particles of 20 micron size cause severe damage. The
smaller particles tend not to deviate much from the gas stream lines, they
most often contribute to low angle impact with long paths to impact, while
the larger particles tend to deviate substantially from the flow direction,
and consequently to impact at higher angles along a shorter path to impact.
Because ductile metals erode more rapidly at low glancing angles (20o
for flat surfaces, 40° for cylindrical surfaces), it is best to replace
conventional elbows with T-connections, as a 90o impact is far less severe
with metals. Furthermore, the char will cushion the impact by packing
one wing of the T. An alternative is to line or coat with a ceramic, since
while these materials are relatively sensitive to 90° impact, they are
minimally sensitive to low angle impact. Thus, alumina ceramics can be slip
cast over metals.
The most erosion resistant materials are (high density) ceramics,
carbides, and then cobalt base alloys in that order.
Relative wear tests at Battelle, Columbus indicate that in 2.5 hours,
Kennametal grade K703 with cobalt-chromium binder wore less than 7.6 microns
and boron carbide wore 9.6 microns, compared to 25.4 microns for alumina and
B-53
-------
31.7 microns for zirconium diboride. Silicon carbide erosion was in excess
of 100 microns.
An EPRI program at Battelle has demonstrated that at low flow
velocities of around 15 feet per second and with high char loadings, the
particles tend to stick to the metal, promoting slagging corrosion. Hence,
there should be no abrasive erosion at typical FBC velocities.
B-54
-------
SECTION 7
GASIFIER FACILITIES PROBLEMS
A significant number of failures have occurred in gasifier facilities,
including corrosive tube blockage, perforation, etc. A total of 374 failures
recorded to date are listed in Table 12. The largest number of failures are
attributed to erosion (73), then to design (46) or overheating (46) followed
by sulfidation (39) and chloride stress corrosion cracking (38). The
majority of failures occur in piping (170), then in valves (42), followed by
bellows (36).
Some of these failures and their causes are described in the following
sections.
7.1 C02 ACCEPTOR PROBLEMS
1. Preheater Coils
These coils consist of Inco 800 alloy tubing at 1600°F and 150 psi,
exposed to the regenerator gas. The tubing failed by "metal dusting".
This pitting resulted from carburization at low \\2® content and from
both carburization and sulfidation at high ^0 content. It was
initiated by local fracture of the protective scales, by stresses
generated possibly by thermal cycling. Possible solutions include the
development of more plastic protective scales and the addition to the
alloy of gettering elements, such as titanium, zirconium, hafnium,
columbium and tantalum, which combine strongly with the carbon and
B-55
-------
sulfur, thereby competing effectively with the carburization
and sulfidation of critical elements of the alloy.
2. Gasifier Probe
A gasifier probe made of Type 310 stainless steel and sensor tubes
made of Type 316 stainless steel failed as a result of "stress-
enhanced sulfur corrosion" at a temperature of 1600°F.*
3. Expansion Joint Failure
A 1-inch long crack occured at a convolution of a Type 304 stainless
steel bellows after 31 days in operation at 300op and 200 psi in the
recycle gas. Failure was the result of chloride stress cracking from
chlorides in the steam used to purge the line.
4. Furnace Tube Pitting
Both Inco 800 and Type 321 stainless steel tubes pitted severely at
temperatures above 85QOF, presumably from reaction with the zinc oxide
system used for sulfur removal. Up to 50 percent metal loss in the
321 steel resulted primarily from oxidation, and secondarily from
sulfidation, carburization and nitridation.
7.2 HY6AS PROBLEMS
1. Slurry Dryer Grid
An oxidation-sulfidation failure occurred in an Inco 800 grid located
in the fluidized bed at SOOop with the gas at 120QOF and 1000 psi
pressure; the gas contained 1 percent H2S, as well as CO, C02, H2, N2>
CH4 and
*Conoco Tech Service Report 116-74-1301
B-56
-------
TABLE 12. FAILURE MODES OF COMPONENTS IN COAL CONVERSION PLANTS'
Failure
Mode
Corrosion
C arborization
Metal Ousting
Oxidation
Pitting
Sulfidation
Creep
Design
Equipment
Malfunction
Overheating
Overstressing
Erosion
Fabrication
Fatigue
Quality Control
sec
SCC-C1
Unknown
TOTALS
Pressure
Vessels
3
1
0
2
2
1
2
4
0
3
2
2
1
23
Pumps
1
10
3
17
2
1
34
Piping
9
4
4
9
22
2
15
19
23
7
3
5
6
25
17
170
Valves
1
1
5
1
19
2
4
4
3
2
42
Bellows
4
2
4
8
3
2
1
1
5
6
36
Auxiliary
Process Equip
1
1
1
1
10
7
1
5
2
3
3
35
Thermowells
2
6
2
1
1
12
Thermocouples
1
2
6
2
2
2
2
1
1
3
22
Totals
16
7
8
15
39
4
46
46
1
73
13
9
15
10
38
34
374
co
en
Reference 7
T-355
-------
2. Thermocouple Sheath
Failure of a Type 316 stainless steel sheath required substitution by
Type 310 stainless steel. Additionally, an inert nitrogen purge of
the thermowell aided significantly in reducing this problem. Failure
of the 316 steel had occurred by sulfidation; a low melting
nickel sulfide eutectic had formed at temperatures below 1300 F.
However, the 310 stainless steel sheath eventually failed as well.
3. Pressure Taps and Thermocouple Sheaths
Several Inco 800 alloy components failed by sulfidation. Replacement
by 446 stainless steel proved superior to replacement by 310 stainless
steel (see 2, above).
4. Cyclone Dip-Leg Pipe
An Inco 800 alloy internal cyclone pipe failed by hot corrosion
combined with carburization and sulfidation. It was replaced by
Type 310 stainless steel. It had been exposed to 1900°F
at 20 psi for 2000 hours, of which 300 hours was with low steam
gas, and 1700 hours in a gas of composition 30 percent C02, 15
percent H2, 5 percent Ng, and 35 percent H£0- Tne presence of
molten slag (clinker) was detected on the dip-leg.
7.3 SYNTHANE PROBLEMS
1. Thermocouple Sheath
Two Inco 800 alloy sheaths failed by sulfide corrosion in the
distribution cone of the gasifier. After approximately 50 hours
service at 500°F in steam plus oxygen and some ash, failure resulted
from the presence of 0.25 percent sulfur during start-up at 1000°F,
thus sensitizing the Inco 800 alloy permitting sulfidation to occur in
B-58
-------
regions of residual stress (at bends). Type 310 stainless steel was
substituted for the Inco alloy, and the start-up fuel was changed to
sulfurless propane.
2. Stainless Steel Flange
A leaking crack was discovered in a Type 304 stainless steel weld neck
flange in the preheat line to the gasifier. The flange was exposed
for approximately 180 hours to 10 psi and llOOop flue gas and for 700
hours to steam at 600 psi and 75QOF (with 13 weight percent oxygen for
50 hours). While deionized water was used during operation, city
water containing 28 ppm chlorides was used during checkout pretesting.
Investigation revealed the presence of three long intergranular cracks
in the flange material, plus a state of internal stress due to heavy
cold working of the part during fabrication. It was deduced that the
cracks initiated in the heat-affected weld zone where the flange joins
the preheat piping, and that failure was due to intergranular stress
corrosion.
7.4 FLUIDIZED BED COMBUSTOR PROBLEMS
Reactions occur on fired metallic surfaces beneath deposits which
arise from condensation of components in the coal which are volatalized
by the flame, and from impaction of flyash entrained in the flue gases.
Salts responsible for accelerated corrosion are apparently molten at the
indicated metal temperature, and are capable of reacting with and removing
the protective scales. In the region of the flame, sulfur from the coal
can deposit and react to form low melting sulfides, resulting in very
rapid metal wastage. However, Inco 800 alloy tubes have been found to
show only superficial damage at temperatures around 1400 to 150QOF.
B-59
-------
7.5 MISCELLANEOUS PROBLEMS
1. Project Lignite Preheater Coil (U. of N. Dakota)
Three Type 316 stainless steel preheater coils used in a fluidized bed
slurry heater ruptured rapidly as a result of chloride stress
corrosion cracking; the make-up solvent provided the principal source
of chlorides ("200 ppm). They were replaced by Inco 800 alloy. Even
so, the failures occured in less than 30 hours each. A fourth coil
was annealed after fabrication to remove all residual stresses, and
its service life now exceeds 60 hours.
2. Project Gasoline Reactor Vessel (CONOCO)
A Type 316 stainless steel vessel was found to be subject to chloride
stress corrosion cracking resulting from use of a molten zinc chloride
catalyst.
3. MORGAS Butterfly Valve (ERDA-Morgantown)
Both solid Stellite 3 and Haste!loy-coated steel pressure valves
eroded in less than 30 hours. Only tungsten carbide and Inconel 900
fixed orifices were able to withstand the 1300op fine particulate
environment.
4. Low Btu Gasifier Gas Generator (Westinghouse)
The generator operates on a mixture of propane, C02> steam and air at
1800 to 2400op. The main burner components consisted of 304, 316 and
446 stainless steels and Inco 800 alloy. Carburization and oxidation
failures occured. The burner was then provided with internal cooling
to reduce temperatures to below 1000°F; a refractory flame shield was
B-60
-------
also provided. In addition, the burner material was changed to Inco
601, since it has superior resistance to cyclic carburization and
oxidation.
B-61
-------
SECTION 8
CONCLUSIONS — SIGNIFICANCE TO SAMPLING PROBE MATERIALS
There is no simple solution to coal conversion material problems. The
large number of documented material failures in gasifier facilities will
attest to that. Many of these recorded failures, such as in piping
components, have implications which are directly applicable to material
selection for sampling probes.
The actual gas composition in contact with the material may vary a
short distance away. The corrosion response of the material to a slight
shift in composition can be quite dramatic. In fact, each individual
sampling probe should be specifically designed for the actual gasifier
location which is being sampled: temperature, pressure and anticipated gas
composition are important considerations in the selection of sampling probe
materials, as are the flyash characteristics, including the original coal
composition. The alloys selected should, a,s far as possible, be chosen so
that the thermodynamically stable phases are oxides and not sulfides or
carbides.
The materials must of course be selected so that mechanical failure
does not occur under operating conditions, and must take into account the
effects of thermal cycling and of the phase stability of the alloying
constituents (avoidance of phase changes, precipitation, etc.) at operating
conditions and under transient heating and cooling as well.
B-62
-------
Of course, temperature ranges, within which specific phenomena enhance
corrosion for the particular alloys and operating conditions, must be avoided
or traversed quickly. Furthermore, the component should be purged of
condensable gases prior to cooling if possible, or at least flushed clean
immediately after use.
The discussion of basic corrosion mechanisms emphasizes the importance
of such factors as cleanliness, design considerations, fabrication parameters
and heat treating, on avoidance or minimizing of corrosion. For example,
designing to avoid char build-up, sharp edges, rough surfaces, etc..
Additionally, improper utilization or cleaning of the components can lead to
severe corrosion problems. Therefore, it is important to devise procedures
which will be applied from conceptual design to actual utilization. A pre-
oxidation treatment of the components is recommended to improve the
corrosion resistance.
Sufficient laboratory and pilot plant data have been obtained to
permit the designer to select materials compatible with the environments in
which they will be used, so that the equipment will survive for a reasonable
length of time. It must be noted that intermittent use shortens the lifetime
of the alloys compared to the life expectancy with uninterrupted use.
Furthermore, the 1 iftime of the material is predicted on the assumption that
design, fabrication and utilization "flaws" will be avoided. A single
scratch in a critical location can initiate catastrophic failure.
It appears that abrasion type erosion is not a problem for probes at
most sampling locations. Generally, the flow velocity is such that the
flyash will stick to the metal surface, which creates a severe problem of
slagging corrosion. Therefore, it is important to design to avoid char (or
B-63
-------
flyash) entrapment, in-so-far-as possible. If this is not possible, a
ceramic coating may be found to be a viable solution.
Of course, solutions to corrosion problems should not be left entirely
to material selection. The component and system design can also contribute
solutions: cooling to reduce temperatures, designing to reduce velocities
(wherever possible) and to reduce stresses in critical parts, etc. The fuel
may be modified by cleaning and treating it, and by removal or filtering of
harmful material before it can reach critical parts.
In conclusion, it may be said that probes for sampling at high
pressure and temperature can be built with existing commercial alloys, and,
provided that the proper design, fabrication and utilization precautions have
been taken, can provide several hundred hours of service prior to replacement
of critical components.
B-64
-------
BIBLIOGRAPHY
1. A. J. MacNab, "Design and Materials Requirement for Coal Gasification,"
Chem. Eng. Prog. 71 (11) 51-58 (Nov. 1975).
2. A. A. Foroulis and W. W. Smeltzer, Editors, "Metal-Slag-Gas Reactions
and Processes," The Electrochemical Society Inc., Princeton, NJ (1975)
Symposium Compilation of 78 papers.
3. J. P. Hirth and H. H. Johnson, "Hydrogen Problems in Energy Related
Technology," Corrosion, 32 (1) 3-26 (Jan. 1976).
4. K. Natesan, "Corrosion-Erosion Behavior of Materials in a Coal-
Gasification Environment," Corrosion, 32_ (9) 364-370 (Sept. 1976).
5. F. C. Schora and G. Patel, "Design Considerations for Low Btu Coal
Gasification," Industrial Heating, 43 (11) 15-21 (Nov. 1976).
6. M. S. Crowley: "Refractory Problems in Coal Gasification Reactors,"
Industrial Heating, 43 (11) 58-67 (Nov. 1976).
7. "Materials and Components in Fossil Energy Applications," ERDA Newsletters
No. 1 to 9 (6/75 to 4/77), J. Richard Schorr, Editor, Battelle;
Columbus, Ohio.
8. M. A. Howe, "Selection of Materials Used in Coal Gasification Plants,"
7th Synthetic Pipeline Gas Symposium, AGA (Nov. 1975).
9. A. 0. Schaefer, "A Program to Discover Materials Suitable For Service
Under Hostile Conditions Obtained in Equipment For The Gasification
of Coal and Other Solid Fuels," ERDA Reports FE-1784-12/15/18/21/24,
IITRI Program, Materials Properties Council, NY, NY (Annual 1975;
1st, 2nd and 3rd quarters 1976; Annual 1976).
10. R. A. Perkins, W. C. Coons and F. J. Radd, "Metal Dusting in Coal
Gasification Environments," Proc. Symp. Properties of High Temperature
Alloys, Electrochem. Soc. (Jan. 1977).
11. A. J. MacNab, "The Material Design Interface in Coal Conversion
Technology," Materials Problems and Research Opportunities in Coal
Conversion Vol. II, Corrosion Center, Ohio State U. (Apr. 1974).
12. A. W. Huff, J. P. Young, and L. K. Ives, "High Temperature Erosion in
Oxidizing and Reducing Atmospheres," NBS Reports ERDA £(49-18)1514 and
1536; OCR 14-32-0001-1514, EPA 65012-75-027-C.
13. "Materials for Coal Conversion Systems Design," ASM Conference,
Pittsburg, PA (April, 1976).
14. R. A. Perkins, "Sulfidation Resistant Alloys for Coal Gasification
Service," ERDA Quarterly Report FE-2299-6 (Dec. 1976).
B-65
-------
15. H. D. Bastow and G. C. Wood, "Multilayer Scale Formation During Sulfidation
of Nickel," Oxidation of Metals, 9 (6), 473-496 (Dec. 1975).
16. A. M. Hall, "Potential Materials Problems in Coal Gasification Systems,"
MFPG Coal Conversion Symposium, Battelle, Columbus, Ohio (April 1976).
17. "Materials for Coal Conversion Utilization," ERDA-EPRI-AGA Conference
on Materials for Coal Conversion and Utilization (Sept.-Oct. 1976).
18. W. R. Williams, "Lockhopper Valve Failures," ERDA-Morgantown, MFPG
Coal Conversion Symposium, Battelle, Columbus, Ohio (April 1976).
19. I. G. Wright, "Correlations of High Temperature Corrosion of Alloys
in Coal Conversion Environments," Battelle, Columbus Ohio, ERDA Contract
W-7405-Eng-92 (Sept. 1976).
20. S. Bhattachryya, F. Bock, A. MacNab, and T. B,. Cox, "Alloy Selection
for Coal Gasification Quench Systems," NACE Corrosion 77 Paper No.
51 (March 1977).
21. V. L. Hill and M. A. H. Howes, "Metallic Corrosion in Coal Gasification
Pilot Plants," NACE Corrosion 77 Paper No. 50 (March 1977).
22. Charles Spengler et al., "Possible Materials Problem Areas in Power
Generation Turbines Operating with Fuels or Gases Derived from Coal,"
NACE Corrosion 77 Paper No. 13 (March 1977).
23. "Clean Fuels from Coal Symposium II Papers," IGT, Chicago 111., p.
904 (June 23-27, 1975).
V
24 Metals Handbook, Vol. X, Failure Analysis and Prevention, A.S.M.,
Metals Park, Ohio, p. 210 (1975).
25. R. A. Perkins, "High Temperature Corrosion of Stainless Steels in
Coal Gasification Environments," NACE Corrosion 77 Paper No. 49 (March
1977).
26. L. H. Wolfe, "Laboratory Investigation of High Temperature Alloy
Failure Mechanisms," NACE Corrosion 77 Paper No. 12 (March 1977).
27. H. F. Wigton, "Corrosion of Superalloys, Inconels and Stainless Steels
by the Products from Fluidized-Bed Coal Combustion," NACE Corrosion
77 Paper No. 52 (March 1977).
28. N. Parikh et al., "Corrosion of Materials at Temperatures and Environments
Expected in Proposed Processes for the Gasification of Solid Fuels,"
IITRI-B8149-27, Summary Report for Metals Properties Council (Jan.
1975).
29. J. Hull et al., "Hydrogen Induced Delayed Failure of X60 High Strength
Steel Pipe," Summary Report No. 8679, SRI, Menlo Park, CA (Nov. 30,
1970).
B-66
-------
Appendix C
Acurex Corporatlon/Aerotherm Division
May 1977
Aerotherm Project 7237
TAR SAMPLING IN
COAL GASIFICATION PROCESSES
W. V. Krill
AEROTHERM REPORT TM-77-173
Prepared for
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park
North Carolina 27711
Contract 68-02-2153
C-1
-------
INTRODUCTION
With the advent of diminishing petroleum fuels as a natural resource
(i.e., natural gas and fuel oil), interest in coal as an alternate fuel
source is growing. Over 40 coal gasification processes from bench to full
scale are currently under development in this country. Other coal processes
producing solid and liquid fuels are also being developed. As a result, new
technologies in process stream handling and measurement are required if these
processes are to be demonstrated on a commercial scale.
High molecular weight hydrocarbons (tars) are a byproduct of many coal
gasifiers. These hydrocarbons which are entrained in the gas stream as
either vapors or mists are important to process development for several
reasons:
1. Entrained tars are difficult to transport over long distances (for
units producing pipeline gas), having a tendency to deposit on
pipe walls as stream temperature decreases or flow irregularities
are encountered
2. Tars may not be compatible with gas turbine fuel systems and
combustion chambers (for units producing electrical power),
causing tar buildup and flow restriction within these components
3. Some tars have been identified as being highly carcinogenic and
cannot be released to the environment
In addition, tars hamper sampling system operation which is normally required
for gas stream measurements by process developers.
Since coal gasification is largely in the developmental stage in the
United States, little attention has yet been paid to tar production.
Similarly, only limited tar quantity and physical property data is available.
Interest in tar production is increasing, however, as developers continue to
solve process operational problems and turn their attention toward producer
gas quality.
In response to the need for an understanding of the tar production
problem, the Aerotherm Division of Acurex Corporation has conducted a survey
of tar collection technology under its current contract "Measurements for
High-Temperature, High-Pressure Processes" for the Environmental Protection
Agency. Under this contract, a variety of measurement techniques and
equipment for use in high-temperature, high-pressure process streams have
C-li
-------
been investigated. The investigation of tar collection methods for sampling
systems is a natural extension of studies related to gasification sampling
problems.
This tar survey has focused on:
• Development of a thorough understanding of tar problems
• Compilation of a comprehensive list of tar physical properties
t Evaluation of existing detarring hardware for possible application
to sampling systems
• Formulation of tar collection design specifications and recommen-
dations for collector concepts for process sampling systems
Each of the above survey elements will be discussed in the following sections
of this document.
C- i i i
-------
TABLE OF CONTENTS
Section Page
1 TAR PRODUCTION VARIABLES C-l
2 TAR PROPERTIES C-3
2.1 Chemical Composition C-3
2.2 Condensation Temperature C-5
2.3 Chemical Stability C-5
2.4 Other Properties C-7
3 TAR COLLECTION CONCEPTS C-8
3.1 Electrostatic Precipitator C-9
3.2 Scrubber C-ll
3.3 Condenser C-ll
3.4 Cyclone C-14
3.5 Filter C-14
4 TAR COLLECTORS FOR SAMPLE SYSTEM APPLICATION C-l7
REFERENCES C-22
C-iv
-------
LIST OF ILLUSTRATIONS
Fjjjure Page
1 Tar yield as a function of distillation temperature .... C-6
2 Typical electrostatic precipitator C-10
3 Typical water-jet scrubber C-12
4 Typical quench chamber scrubber C-13
5 Typical shell-and-tube condenser C-15
6 Typical cyclone separator C-16
7 ESP collector for sampling system adaptation C-19
8 Gasification process sampling scheme C-21
LIST OF TABLES
Figure Page
1 Grand Forks Slagging Lurgi Tar Analysisa C-4
2 Typical Elemental Analysis of Gasifier Tars C-5
C-v
-------
SECTION 1
TAR PRODUCTION VARIABLES
An understanding of the severity of tar problems associated with
coal gasification processes has been gained by conducting a survey of process
developers. This survey has identified specific process problem areas
related to tar, summarized past experience with tar handling, and evaluated
present and future requirements for tar sampling.
Operational problems related to tars are quite diverse among process
developers. Tar properties, quantities produced, and the physical state are
largely dependent upon:
• Gasifier type and supporting cleanup devices
• The type of coal used in the process
t Gasifier temperature levels and temperature histories of the coal
and producer gas
• Other physical characteristics of the gasifier and support
equipment
As a result, each process developer is dealing with different tar production
parameters and tars of different physical properties. A given process may
produce a wide range of tar and oil types entrained within its producer gas
stream which will vary significantly with coal type, temperature, pressure,
gasification rate, etc.
There are nearly as many variations in gasifier types as there are
facilities in this country (40 or more). The units are largely distinguished
by such features as:
• Gasifier bed geometry and coal transport technique
• Oxygen or air firing
• Steam introduction
Since gasifier design features are usually quite unique to a given
unit, tar production is similarly a unique function of a given process. In
general, developers have not investigated the phenomena of tar production as
C-l
-------
related to the above process variables to any significant extent. The
uniqueness of gasifier units also makes it difficult to correlate information
among processes to further the understanding of tar problems. Some
preliminary work has been performed to demonstrate the effect of coal type on
tar production. Lowry (Reference 1) has stressed that composition is
strongly dependent upon coal type. More recently, Riley-Stoker Corporation
has suggested a 100-percent or more increase in tar production with
bituminous coals as compared to lignite (Reference 2). It also appears from
this study that tar yield may be a direct function of the percent of volatile
matter contained within a given type of coal.
The effects of coal type on tar production may be overshadowed by the
effects of reactor temperature and temperature history which have a strong
influence on the tar produced. Offtake temperature varies widely among
gasifier types and even among sizes of a given type. In general, commercial-
size plants will have higher offtake temperatures than their developmental
bench and pilot scale versions.
It has been suggested by several sources (Reference 2 for example)
that formation of heavy tars is favored by high temperatures and long residence times
of the gasifier products at those temperatures. Survey information compiled
by Aerotherm supports this hypothesis. Gasifiers with low offtake
temperatures, 300°F to 500°F (e.g., the slagging Lurgi at Grand Forks Energy
Research Center and the Riley-Morgan demonstration unit) reportedly produce
relatively small amounts of tar. Most of the heavy hydrocarbon output is
present as light oils. Many gasifiers operate with an offtake temperature in
the 800°F to 1400°F range and report severe tar problems related to cleanup
equipment use and sampling systems operation. Examples of these facilities
are: (1) 0)3 Acceptor, Rapid City, South Dakota; (2) G.E., Schenectedy, New
York (Lurgi); (3) Morgan town Energy Research Center, Morgantown, West
Virginia; and (4) Westinghouse Process Development Unit, Waltz Mill, Pennsyl-
vania.
Again, the quantity of tar production data related to process tempera-
ture is very limited. For many processes, in fact, potential tar problems
may exist that have yet to be identified. As the scale of gasification units
increases (and hence the offtake temperature increases in many instances),
greater attention will be paid to the related tar problems.
C-2
-------
SECTION 2
TAR PROPERTIES
A gasifier process survey in conjunction with a literature review was
conducted to identify current data available on the physical properties of
coal tars. This data is important to a fundamental understanding of tar
problems and to the evaluation and selection of tar collection equipment. The
properties that are discussed below include:
• Chemical composition
• Condensation temperature
• Chemical stability
t Other physical properties
• Health hazards
2.1 CHEMICAL COMPOSITION
The chemical composition of coal process tars is the one property
which makes all other properties difficult to quantify. Several thousand
molecular species have been estimated to be present in some isolated tar
samples which indicates that a wide range of associated physical properties
exist as well. Therefore, in most cases, only hydrocarbon groups and average
physical property data can be reported for a given sample.
An example of the hydrocarbon groups that have been reported are shown
in the data for the Grand Forks Energy Research Center slagging Lurgi
gasifier (Reference 3). A list of 24 groups is shown in Table 1. Note the
diversity of the identifed compounds. Typical analytical techniques for tars
have included gas chromatography and gas chromatography coupled with mass
spectrometry. Ultimate analyses have also been performed in several
instances (References 2 and 3), providing the typical results shown in Table 2.
C-3
-------
TABLE 1. GRAND FORKS SLAGGING LURGI TAR ANALYSIS^
Constituents Pet
Saturated hydrocarbon 10.3
Nonvolatile residue 7.7
Phenol/cresol/xylenol 22.4
Indanol 1.0
Dibenzofuran 3.78
Hydroxyanthracene 1.01
Indanes 0.22
Naphthol 3.93
Indenes 0.96
Pyridines 1.81
Quinclines 1.35
Naphthalene 4.76
Acenaphthene/biphenyl 0.37
Fluorene/acenaphthalene 1.76
Phenanthrene/anthracene 0.85
Dihydropyrene 0.81
Pyrene/fluoranthene 0.49
Chrysene 0.49
Benzenes 1.78
Indole 0.23
Carbazole 0.21
Benzocarbazole 0
Benzofuran 5.16
Benzonapthofuran 2.30
Performed by mass spectrometry
C-4
-------
TABLE 2. TYPICAL ELEMENTAL ANALYSIS OF GASIFIER TARS
Element Amount (by Mass)
C 85 %
H 8 %
N 1 %
0 5 %
S 0.5%
Ash 0.5%
2.2 CONDENSATION TEMPERATURE
Condensation temperature (or equivalently, distination temperature)
of tars is probably the most important physical property to process devel-
opers. It determines whether tars are entrained in the gas stream as vapors
or as liquid mist and thus their propensity for depositing internally on
process ducts and in cleanup equipment. Since condensation temperatures vary
widely among hydrocarbon groups (from room temperature for light oils to over
1500°F for tars) and even among species of a given group, tar samples
typically have a wide range of condensation temperature. Note that
condensation temperature is a function of pressure as well (Reference 4).
Some of the more common pure hydrocarbons (e.g., naphthalene) have been shown
to have condensation temperatures which are strongly pressure-dependent above
10 atmospheres. This is expected to be true for other hydrocarbon species as
well. Since a wide range of gasifier operating pressures has been identified
(atmospheric to 2500 psig), process pressure will play an important role in
tar collection. In general, higher process pressures will result in fewer
species condensing at a given temperature.
Condensation temperature data is normally obtained by fractional dis-
tillation of the tar sample. As an example of several techniques for vapor
pressure and distillation temperature measurements, see References 5 through
8. Reference 2 (see Figure 1) reports distillation temperatures from
approximately 300°F to 1600°F, illustrating the wide range usually obtained.
This data is taken from laboratory tests of References 9 and 10. The quan-
tity of tar which will condense at any given temperature is dependent upon
the distillation distribution. Distributions of the shape of Figure I are
typical.
The total tar sample typically represents as much as 10 percent
by weight of the coal feed. Process developers currently burn these tars
as product gas is flared or collect them by scrubbing (or other techniques)
for eventual dumping as landfill.
2.3 CHEMICAL STABILITY
Little fundamental work has been performed to indicate the chemical
stability of coal tars. Some experience does exist, however, to indicate
chemical changes which a tar sample can undergo during a period of time after
it is extracted from the process. Some samples have been reported to have a
large undisti1lable fraction, indicating that samples are subject to
C-5
-------
100
o
i
o
u
c
HI
u
k.
-C
D>
1
90
so
70
60
50
40
30
20
10
0
Lignite
.o
200
-O-
XX
Bituminous
400 600 80.0 1000
Distillation temperature, °F
1200
1400
1600
Figure 1. Tar yield as a function of distillation temperature.
-------
polymerization and oxidation reactions. Even light oil samples that are
initially collected floating upon a water surface seem to become inseparable
from the water or completely polymerized to the solid phase after a period of
time.
The importance of tar stability is twofold. First, the process
developer does not want to contend with a substance that is difficult to
remove from cyclones, scrubbers, etc. Second, the integrity of tar samples
taken for analysis may be difficult to maintain, providing data of only lim-
ited accuracy. Chemical stability also influences tar collection techniques,
and further problem definition is required for full development of sampling
systems capable of separating tars.
2.4 OTHER PROPERTIES
A number of other tar properties of interest have been investigated.
However, data is extremely limited. Lowry (Reference 1) lists data sources
and some data for:
• Specific gravity
• Temperature coefficient of specific gravity
• Heat capacity and heating value
• Heat of vaporization
t Thermal conductivity
• Surface tension
• Viscosity
t Dielectric properties
Much of this data is for coal tar products (e.g., from coke ovens) which are
known to have significantly different properties from gasification tars.
More recent gasifier studies (Reference 2) have reported some of the funda-
mental properties (specific gravity, viscosity), but data is generally
sparse.
One final property of gasifier tars should be emphasized. Many of the
hydrocarbon compounds that can exist in tars are known carcinogens,
particularly the aromatics and other benzene ring compounds (Reference 11).
It is especially important that all persons dealing with tar wastes and
samples avoid skin contact and breathing vapors. Also, safe procedures in
all aspects of tar handling require further study and definition.
C-7
-------
SECTION 3
TAR COLLECTION CONCEPTS
A variety of equipment has been used for extraction of tars from
process gas streams on the process level. Each of these concepts have
been reviewed in consideration of adapting the most appropriate technique
to a sampling system for collection of tar samples.
Very few of the tar collectors identified have been shown to per-
form satisfactorily on the process level. The unit must be capable of
performing several difficult functions in an efficient manner:
• Condensation of tar vapors to the mist phase (when required)
• Collection of all mist droplets in the stream regardless of
drop size and total mass loading
• Operation without internal fouling by the collected material
• Provision for simple access of collected material
Nearly all classical dust and mist collection techniques have been
applied to tar removal. These include:
• Electrostatic precipitators
• Scrubbers
• Condensers
• Cyclones
• Filters
The success or failure of each relates directly to the properties
of the tars themselves. An evaluation of each collection technique is
given below.
C-8
-------
3.1 ELECTROSTATIC PRECIPITATOR
The electrostatic precipitator (ESP) has historically been the
most widely applied technique for tar removal. Reference 9 shows
application to gas detarring as early as 1915 in the United States. Possibly
as many as 600 unit installations exist throughout the country, making the
detarring of fuel gases the second largest application of the ESP.
The basic ESP concept is illustrated in Figure 2. A high voltage
potential (typically 10 to 30 kV) exists at the discharge electrode providing
a potential difference between the electrode and precipitator shell. The
particles entering in the gas stream pick up charge from the field and then
migrate toward the shell as a result of Coulombic forces. Variation of the
unit's geometry, field potential, and field configuration provide different
design variables for specific applications.
The success of electrostatic precipitation for gas detarring has been
reported (see References 2 and 12 through 15 for example). The discussions
are generally qualitative in nature rather than quantitative, but the general
level of success is demonstrated. Only Reference 2 has attempted to provide
quantitative data in terms of both mass collected and ESP efficiency at vari-
ous tar mist particle size ranges.
In general, the wide range of tar particle size range and total mass
loading make collection difficult. Tars often exist in the gas stream as
both fog (particle diameters as small as 0.lu) and mist (particle diameters
from lu to 60u or greater). The ESP is one device which can be designed to
efficiently collect particles over this entire size range. The ability to
collect small particles is particularly significant as other collection
methods lack this ability. Very small particles tend to follow gas streamlines
and, therefore, devices which depend on condensation by impingement are
ineffective for fine mists. The unique ability of the ESP to collect
fine particles has been demonstrated in Reference 2. Other references
cited demonstrate the efficiency of electrostatic precipitation for both
tar and light oil collection.
In addition, the ESP can easily accommodate a large collected mass
over a long time interval. The collected tar mixture is usually of a low
enough viscosity that the collected material can be continually withdrawn
from the bottom of the unit (see Figure 2).
The possible explosion hazard associated with an electrical spark in
a combustible gas mixture caused initial apprehension in the gas industry
when electrostatic precipitation was introduced. Successful demonstrations
in the early 1900's, however, had shown that flammability limits for most
process gases were so low that the explosion risk was minimal. Therefore,
due to high efficiency and generally low maintenance, the ESP has become the
most popular means of gas tar removal on the process level.
C-9
-------
Insulator
Precipitalor
shell
Gas
entrance
Dust on
precipitator wall...
Collected dust
Figure 2. Typical electrostatic precipitator.
C-10
-------
3.2 SCRUBBER
A number of coal gasification facilities utilize scrubber units for
gas cleanup, usually intended for water, oil, and tar removal. A number of
scrubber types exist, but only the spray and water chamber scrubbers will be
discussed here.
A typical spray scrubber is shown in Figure 3. A high velocity,
atomized water stream is sprayed either axially or radially into the throat
of a venturi section. The tar laden gas stream also enters the venturi.
Particle collection is accomplished by impingement of the water droplets
upon the smaller tar particles, causing them to be entrained in the water
stream and removed at the bottom of the unit. Support equipment includes
the pump, a water supply, and waste water handling equipment.
The success of the venturi scrubber is related to tar particle size.
The impaction collection mechanism (of the water droplet on the tar particle)
is most efficient and can scrub smaller particles from the gas stream when
water droplets are very finely atomized and moving at high velocity. This
requires a large pressure drop at the atomizing nozzle and large pumping
requirements. The result is a very substantial increase in scrubber energy
input for particle removal down to ly in diameter. In general, collection of
submicron particulate by the spray scrubber is prohibited by the impaction
mechanism of collection. Since submicron tar particles (fog) exist in many
coal gasification processes, the venturi will allow some entrained tar to
remain in the process stream.
Quench chamber scrubbers force the gas to bubble through a contained
liquid water bath. The quench chamber is illustrated in Figure 4. Efficient
scrubbing of tars and oils from the gas stream relies on obtaining good
contact between the tar mist and the quench water. This is achieved by
breaking gas bubbles down to the smallest possible sizes.
Reference 2 has reported collection capabilities of the quench chamber
and compared them to those of an ESP unit. The quench chamber was found to
be less efficient in total collection, again limited by the submicron tar
particles existing in the producer gas. The chamber appears to be most ef-
fective in removing lower temperature condensate fractions (light oils) and
water, possibly due to the existence of larger particle sizes among these
fractions. In the study of Reference 2, it was concluded that scrubbing
techniques are significantly less effective then electrostatic precipitation
for tar removal.
In conclusion, scrubbing has been found to be effective in some tar
collection applications. However, the technique is ineffective for submicron
particulate collection which may result in significant tar carryover in some
process applications.
3.3 CONDENSER
Condensers differ from other tar collection devices (i-e., the ESP) in
that they are capable of condensing tar vapors to the liquid phase as well as
C-ll
-------
Pump
Spray
nozzle
Figure 3. Typical water-jet scrubber.
C-12
-------
Gas
inlet
Gas outlet
Gas dip tube
Quench water
Condensate
drain
Figure 4. Typical quench chamber scrubber.
C-13
-------
collect the tar mist. Other collector types operate effectively only if tars
are condensed to mist prior to entering the unit. In the condenser, the mist
collection is performed by exposing the gas stream to large, cooled surfaces,
allowing tar droplets to attach to the walls. Particle migration to the wall
is essentially unaided (contrasted to the Coulomb forces existing in the
ESP), and relatively large surface areas and small flow passages are required
for efficient collection.
Process survey and literature reviews have revealed several types of
condensers used for detarring of producer gas. The Riley-Stoker gasifier
process and a sampling system at the Grand Forks slagging Lurgi (References
2 and 3) have both utilized basic tube-and-shell heat exchange condensers for
tar collection (see Figure 5). In both instances, it was reported that light
oils were collected but not significant quantities of tar. As was noted for
scrubbers in Section 3.2, it appears as though the large drop size of light
oils facilitates collection by a number of devices. The smaller (often
submicron) particles of tar fog are more illusive and difficult to collect.
A rather novel condenser utilizing centrifugal impellers to aid the
tar collection was also identified by process survey. This unit has since
gone out of production, however, and little information is available as to
its functionality. In general, reports of the effectiveness of condensers
for tar collection are limited.
3.4 CYCLONE
The cyclone is a most common device for solid particulate removal in
many processes. A typical design is shown in Figure 6. Most gasification
facilities utilize the cyclone as the primary cleanup device. Hence, several
reports regarding cyclone collection of tars are in existence. Riley-Stoker
(Reference 2) has shown them to be relatively ineffective in collecting
entrained tar relative to other devices. It is interesting to note that
cyclones generally will not remove particulate that is smaller than 1 micron
in diameter. Carryover may result, therefore, if many fine particles exist
in the process stream.
3.5 FILTER
Various filtration techniques were also investigated for application
to tar collection. Ordinary fibrous filters are not a useful tar collection
technique as only a small collected mass would clog the filter and restrict
flow. Packed bed filters are one alternative as they are usually mounted
within a large chamber (to provide laminar gas flow) and can collect a
large volume of tar and other condensibles. Unfortunately, packed beds
are generally ineffective for particles of a few microns diameter or smaller.
A packed bed/solvent bath combination filter would increase small particle
collection efficiency but would require some additional development.
Therefore, filtration may be a viable technique for some process tar measure-
ments .
C-14
-------
Gas outlet
t
Gas inlet
o
Coolant
inlet
Coolant
outlet
Compensate drain
Figure 5. Typical shell-and-tube condenser.
-------
Receptacle
Body
Cylinder
Figure 6. Typical cyclone separator.
C-16
-------
SECTION 4
TAR COLLECTORS FOR SAMPLING SYSTEM APPLICATION
The survey of the physical properties of coal gasifier tars and of
existing and potential tar collector techniques has led to recommendations
for tar collection in process sampling systems. Since sampling system expe-
rience in tar collection is currently in its infancy, hardware development is
required. This development may be either a simple scaledown of existing full
scale process equipment or the development of new designs for specific
application to sampling systems.
To summarize the previously discussed survey findings, a list of
design specifications and hardware requirements can be formulated. All spec-
ifications fall into two basic categories:
1. Collector efficiency
2- Collector applicability to the high temperature and pressure
environment of a gasification system
Collection efficiency is a key requirement of sampling system hardware
as the objective of sampling is usually to obtain a quantitative measurement
of the process stream constituents. Any tar material which passes through
the collector, then, provides inaccuracies in the measurements obtained. It
is also important that the device be capable of interfacing with the severe
environment of the process stream. This requirement includes safe and
reliable operation over its lifetime.
The efficiency of the developed collector depends upon its ability to
collect tar mist particles over the entire size range present in the gas
stream. As has been noted, process experience suggests that fine particulate
(below 1 diameter) does exist in most gas streams and is particularly hard
to capture. The percentage of the total tar mass that is submicron in size
is largely unknown; thus, the collector should be required to have reasonably
high collection efficiency for small particles. Of the techniques discussed
in Section 3, only electrostatic precipitation and filtration have this
potential.
There are a number of other factors which will have an impact on
collection efficiency. In general, the longer the tar particles are exposed
to the collecting surfaces, the greater the number that will ultimately be
removed from the gas. Therefore, collector residence time may be an
C-17
-------
important design parameter. The designer must also consider the total
mass to be collected (in relationship to sampling time and tar loading).
This mass must be contained so as to not interfere with collector opera-
tion. All collectors discussed can be designed to meet these requirements
with the exception of some filter units which will clog under limited
operation and thus be unsuitable for sampling system application.
It almost goes without saying that a sampler component must be
compatible with the environment in which it will be used. It is pointed out
here, however, to illustrate requirements of the collector design. One must
as a minimum consider:
e Material requirements to withstand stream temperature and possible
corrosive and erosive properties
• Structural design to withstand pressure loadings at temperature
• Ease of sample extraction such that samples will not be contami-
nated
• Safety requirement of containing hot gases and possibly toxic tar
substances
• Geometry and size that are compatible with other system components
Probably all of the collector concepts of Section 3 could be designed
to incorporate these features. Collector selection, therefore, is mainly
dictated by collection efficiency requirements.
In consideration of tar properties, collection concepts, and collector
requirements, the electrostatic precipitator appears to be the most
appropriate for sampling system adaptation. Its general high efficiency,
compatibility with design requirements, and wealth of practical experience on
the process scale make it the optimum choice for further development. There
are applications where filters and/or condensers may also give adequate
performance. Cyclones and scrubbers do not appear adequate for tar
collection in any application. One should be fully aware of the limitations
and design problems associated with these units before development is begun.
Aerotherm has demonstrated the applicability of the electrostatic
precipitator by preliminary design calculations. A typical gasifier test
condition of 5000 Ibm of producer gas per hour with a density of 0.31 Ibm/ft3
and 0.03 Ibm of entrained tar per Ibm of gas was assumed for the design
specification. In addition, a sampling system operating at a rate of 1 acfm
was assumed. These assumptions result in 0.6 Ibm of tar that must be
collected by the sampling system every hour.
It was desirable to have the resulting ESP of compact design such that
it would be compatible with sampling systems. The result was a two-stage
collector based on a Southern Research Institute (SoRI) concept, shown in
Figure 7. The cylindrical design fits into a space of 7 inches in diameter
and 6 inches in length. A disc-shaped electrode charged to 10,000 volts
C-18
-------
Gas Inlet
Gas exit
Corona electrode
Collector electrode
Grounded
collection
surface
10 kV
-5 kV
Power leads
Figure 7. ESP collector for sampling system adaptation.
-------
promotes primary charging of incoming tar droplets. An annular secondary
field at 5000 volts provides a large surface area for particle collection.
The acceptability of the design is further demonstrated by its
computed collection efficiency. For all particle sizes above 1 micron, the
efficiency is virtually 100 percent. At 1 micron, efficiency falls to 99.6
percent and at 0.3 micron falls to 99.2 percent. Hence, the high collection
efficiency requirement of any tar collector is clearly demonstrated by the
ESP approach.
The sample flows by gravity to a cup at the bottom of the unit
for eventual removal: the majority of the collected tar fractions remain
fluid at the elevated operating temperature (~450°F). Once the sample
is condensed and collected, analysis should be performed as soon as possible
in order to avoid sample composition changes with time.
In order to decide which collector is best for a specific applica-
tion, an overall sampling strategy must be developed. This step is fol-
lowed by design of a sampling system concept and finally design of indi-
vidual components.
Most coal gasification developers are primarily interested in solid
particulate carryover from the gas producer and cleanup units. Therefore,
sampling systems are usually designed to provide appropriate solid collection
capabilities. This requires that solids be separated from tar mists and
collected independently. Usually, solid particulates are removed from the
sample stream first since any tar collection device would act efficiently on
solids as well. Solid particulate collection at stream temperature is desir-
able to avoid additional condensation of the tars. Thus, solid particulates
should be removed first while at or near stream temperature and pressure. For
tar removal, the stream must then be cooled (to allow condensation) and
passed into the collection device. The cleaned sample gas is then vented or
retained for further analysis. Figure 8 shows a diagram of this sampling
scheme.
This sampling scheme is one which can be envisioned for a multicompo-
nent producer gas stream. One might select particulate cyclones for solids
collection, a condenser for gas cooling, and an ESP for tar collection. The
exact configuration of the sampling system is necessarily process dependent.
The tar collection technique is but one system function requiring
consideration and development for future sampling systems.
C-20
-------
SAMPLE EXTRACTION FROM
PROCESS DUCT
SOLID PARTICULATE
COLLECTION
APPROPRIATE TEMPERATURE
AND PRESSURE REDUCTION
TAR
COLLECTION
SAMPLE GAS
VENTING OR ANALYSIS
Figure 8. Gasification process sampling scheme.
C-21
-------
REFERENCES
1. Lowry, H. H. (editor), "Chemistry of Coal Utilization -- Supplementary
Volume," John Wiley and Sons, Inc., 1963.
2. Lisaukas, Robert A., Johnson, Stephen A., and Earley, William P.,
"Control of Condensible Tar Vapors from a Fixed Bed Coal Gasification
Process," 4th Energy Resource Conference, Paper No. IMMR11-PD14-76, June
1976.
3. "Quarterly Technical Process Report, July-September 1976, Grand Forks
Energy Research Center," GFERC/QTR-76/5.
4. Prausnitz, J. M., "Molecular Thermodynamics of Fluid-Phase Equilibria,"
Prentice-Hall, Inc., New Jersey, 1969.
5. Sinke, G. C., "A Method for Measurement of Vapor Pressures of prganic
Compounds Below 0.1 Torr -- Naphthalene as a Reference Substance," J.
Chem. Thermodynamics, 6, 1974.
6. Smith, George, Winnick, Jack, Abrams, D. S., and Prausnitz, J. M.,
"Vapor Pressures of High-Boiling, Complex Hydrocarbons," The Canadian
Journal of Chemical Engineering, Vol. 54, August 1976.
7. McKay, J. F., Amend, P. J., Harnsberger, P. M., Cogswell, T. E., and
Latham, D. R., "Separation and Analysis of Petroleum Residues," Laramie
Energy Research Center, ERDA.
8. Green, J. E., Schmach, L. J., and Worman, J. C., "Simulated Distillation
by Gas Chromatography," Analytical Chemistry, Vol. 36, No. 8, 1964.
9. Rambush, N. E., "Modern Gas Producers," London, p. 102, 1923.
10. Rhodes, E. 0., "The Chemical Nature of Coal Tar," Chemistry of Coal
Utilization, Vol. 2, John Wiley and Sons, Inc., New York, p. 1290, 1945.
11. Sax, N. Irving, "Dangerous Properties of Industrial Materials," Van
Nostrand Reinhold Co., New York.
12. White, Harry J., "Fifty Years of Electrostatic Precipitation," Journal
of APCA, Vol. 7, No. 3, November 1957.
13. White, Alfred H., Rowley, R. B., and Wirth C. K., "Electrical Separation
of Tar from Coal Gas," Journal of Gas Lighting and Water Supply,
September 1914.
14. Cree, K. H., "Cottrell Electrical Precipitation as Applied to the
Manufactured Gas Industry," American Gas Journal, Vol. 162, March 1945.
15. Thompson, R. J. S. and Cosby, N. T., "Notes on the Removal of Particu-
late Matter and Vapour Phase Sulphur Compounds from Gases," The Gas
World — Coking, Vol. 148, August 2, 1958.
C-22
-------
Appendix D
Aerotherm Project 7237
FIELD TESTING OF A SAMPLING SYSTEM
FOR HIGH TEMPERATURE/HIGH
PRESSURE PROCESSES
William Masters
Acurex Corporation/Aerotherm Division
485 Clyde Avenue
Mountain View, California 94042
June 1977
AEROTHERM REPORT TM-77-177
Prepared for
EPA Project Officer — William Kuykendal
Industrial Environmental Research Laboratories
U.S. Environmental Protection Agency
Research Triangle Park
North Carolina 27711
Contract 68-02-2153
D-i
-------
TABLE OF CONTENTS
Page
1 INTRODUCTION ..................... D-l
2 SUMMARY OF RESULTS .................. D-2
3 EQUIPMENT DESCRIPTION ................. D-4
3.1 Sampling System ................. D-4
3.2 Exxon Mim'plant Test Facility .......... D-15
4 TEST DESCRIPTION ................... D-20
5 DATA ......................... D-23
D-ii
-------
LIST OF ILLUSTRATIONS
Figure Page
3-1 High-temperature, high-pressure sampling system .... D-7
3-2 System schematic D-8
3-3 Exploded view of HTHP probe D-9
3-4 Probe housing and duct interface valve D-10
3-5 Access valves D-ll
3-6 Dowtherm console D-13
3-7 Flow control oven and gas train D-14
3-8 Control consoles D-16
3-9 Pressurized fluidized bed coal combustor system .... D-17
3-10 HTHP probe assembly installed at Exxon miniplant .... D-19
5-1 Particle size distribution D-32
5-2 Impactor substrates D-34
5-3 Impactor substrates D-34
5-4 Impactor substrates D-35
5-5 Impactor substrates D-35
5-6 Impactor substrate Run 3, Stage 5 D-36
5-7 Impactor substrate Run 2, Stage 5 D-37
5-8 Particle photomicrographs Stage 1 D-38
5-9 Particle photomicrographs Stage 2 D-39
5-10 Particle photomicrographs Stage 4 D-40
5-11 Particle photomicrographs Stage 6 D-41
5-12 Particle chemical composition D-42
D- i i i
-------
LIST OF TABLES
Table Page
3-1 Sampling System Capabilities D~5
3-2 Utility Requirements D-18
5-1 Test Conditions D-24
5-2 Probe Instrumentation Readings D-25
5-3 Gas Train Instrument Readings D-26
5-4 Anisokinetic Correction Factors D-27
5-5 Structure Temperatures D-29
5-6 Particulate Content D-30
5-7 Particle Size Distribution D-33
D-iv
-------
SECTION 1
INTRODUCTION
This report describes field testing of an advanced sampling system
which can measure the participate concentration and trace element content
of high-temperature, high-pressure (HTHP) processes. Tests were held at a
pilot-scale coal combustion plant where pressurized fluidized bed combustion
is being developed. The field tests were successful in demonstrating sam-
pling system capabilities and producing sample data.
Acurex/Aerotherm has developed the HTHP sampling system for the
Industrial Environmental Research Laboratory of the U.S. Environmental Pro-
tection Agency. This work is part of a program Measurements of High-
Temperature, High-Pressure Processes (Contract 68-02-2153), intended to pro-
duce the new sampling technology needed for advanced coal conversion pro-
cesses. Fluidized bed combustion and coal gasification processes emit gases
containing large quantities of fine particles. These particles must be re-
moved to prevent damage to process equipment (mainly turbines) and to elim-
inate potential environmental pollution. Development of particulate removal
equipment is an important step toward making advanced coal conversion pro-
cesses practical. The sampling system described in this report is one of
the first tools available for measuring the collection efficiency of fine
particle removal devices operating in high-pressure, high-temperature en-
vironments.
D-l
-------
SECTION 2
SUMMARY OF RESULTS
The field demonstration of the high-temperature, high-pressure sam-
pling system was successfully completed. Three sampling runs were made:
one using a filter to collect particulate, and two using a cascade impactor.
Trace metals and trace organics sampling equipment was operated during the
filter run. The test produced the following data:
• Particulate concentration
• Particulate size distribution
• Moisture content
• Particulate chemical composition
• Particulate shape
• Duct gas temperature and pressure
• Access port and valve temperatures
The test also produced samples of trace organics collected on XAD-2 sorbent,
and trace elements collected in oxidizing impinger solutions. These samples
have not yet been analyzed.
For a more detailed description of the test data, see Section 5 of
this report.
The test series demonstrated the capability of the sampling system to
operate in the severe PFBC environment. Sample data was obtained from the
1360°F, 9-atmosphere gas stream. Generally, the system operated as designed:
obtaining access to the pressurized duct while the process was operating,
inserting the sampling probe, sampling the gas stream, and withdrawing the
sample. However, as might be expected in a first field test, a few hardware
D-2
-------
problems were found. Most of these were corrected before the sampling
tests, but one uncorrected problem, a malfunctioning impactor heater, gave
sample collection temperatures which were lower than desired. The heater
will be replaced before future use of the system.
During the test sequence, the sampling operations proceeded very
smoothly. The three sampling runs were completed within a 30-hour period
(20 working hours). The tests showed the versatility of the system, oper-
ating with two different types of particle collectors, with and without
trace element sampling equipment.
D-3
-------
SECTION 3
EQUIPMENT DESCRIPTION
The following discussion is divided into two parts: first, a de-
scription of the advanced sampling system itself; and second, a description
of the Exxon Mini pi ant PFBC facility where the sampler demonstration took
place.
3.1 SAMPLING SYSTEM
The sampling system described in this report samples participate,
trace organics and trace metal contaminants in high-pressure, high temperature
gas streams. The system represents an advancement in the state of the art de-
signed to sample new coal conversion processes. System capabilities are
summarized in Table 3-1.
The basic functions of the sampling system are to:
• Safely contain facility pressure
• Insert the sample probe into the process duct while the process is
operating
• Extract a representative sample
• Cool the sample to a temperature which is compatible with developed
particle collectors yet prevents condensation (~450°F)
• Collect and aerodynamically size particulates
t Collect trace organics and trace metals
• Monitor duct conditions and control sample flowrate to give accu-
rate isokinetic capture conditions
• Remove the sample probe and close off duct access so that collected
samples may be removed while the process remains pressurized
D-4
-------
TABLE 3-1. SAMPLING SYSTEM CAPABILITIES
Sample Environment
• Temperature
• Pressure
• Gas Constituency
CO
COz
NO
S02
HzO
NOX
H2S
COS
CS2
HCN
NH3
Organics
• Stream Velocity
t Partlculate Grain Loadings
• Partlculate Size Range
(for classification)
• Duct Size
Sampling System Configuration
Traverse Capability or Penetration
of Nozzle Into Duct or Vessel
Access Process Port Requirements
6as Constituency Analysis
1200'F - 1800°F
650°C - 1000'F
3-20 atmospheres
Concentrations subject to further
investigation, dependent on process
sampled
8-150 fps
0-15 gr/ft' (subject to further
consideration and actual process
characteristics)
0.2 — 10 microns (Notes:. Larger
particulates may be acceptable in
most cases of total mass deter-
mination or if classify, they
may be amenable to "scalping" ahead
of classification device)
Variable depending on probe; std.
is 8 inches I.D. minimum
Modular, so as to allow In-situ or
extractive sampling by cooled probe
Approximately 18" either in-situ
or extractive configuration (some
dependence on Internal configura-
tion of duct or vessejj&n be ex-
tended by relatively minor hardware
modification (longer probe, chamber
extension, spool piece, etc.)
Standard: 4' IPS minimum, 300 Ib
flange access through 4" IPS alloy
gate value (Note: Smaller ports
may be acceptable If special probe
assembly Is used)
Depends on specific process re-
quirements; will be better Identi-
fied during program based on
state-of-the-art review
D-5
-------
To perform these functions, the sampling system includes the follow-
ing subsystems:
• Sample probe assembly
• Dowtherm coolant system
t Hydraulics for probe traverse actuation
• Flow control oven
• Trace organics module
• Trace metal impinger train
• Control consoles
The sampling system is shown in Figures 3-1 and 3-2.
Probe Assembly
The probe assembly includes the sampling probe, probe housings and
duct access valves. The probe itself consists of the sample inlet nozzle,
cooler section, particulate collector and heated transport tube, as shown
in Figure 3-3. Two types of particulate collectors were used in the demon-
stration tests: one was a University of Washington Mark III cascade impac-
tor, with seven stages of particulate sizing; the other was a glass fiber
thimble filter with large total mass capacity, but no sizing capability.
The inlet nozzle diameter was 0.75 inch for all tests. The front portion
of the probe includes a sample conditioning cooler which reduces sample tem-
perature from process temperature to about 450°F. Electric heaters along
the rear portion of the probe maintain the sample temperature.
The sampling probe is mounted within a housing which contains pro-
cess pressure. The housing is a tube-within-a-tube assembly which telescopes
to traverse the probe into and across the process duct. Hydraulic cylinders
connecting the two housing tubes actuate probe motion. The housing assembly
is shown in Figure 3-4.
For the demonstration test series, the probe housing was connected to
the process duct by two 4-inch diameter gate valves in series. The 4-inch
valve bore provides sufficient clearance for the probe to be inserted through
the valves into the process stream. Two valves were used to give redundant
shut-off capability. The installed access valves are shown in Figure 3-5.
D-6
-------
C3
MICROSW1TCHES FOR
TRANSVERSE CONTROL
/-INNER TUBULAR HOUSING
To gas analysis equipment
and vent
HYDRAULIC SUPPLY SYSTEM
CONTROL CONSOLE (2)
Figure 3-1. High-temperature, high-pressure sampling system.
-------
I
00
1 Enclosure
I (valves & housings)
|
Motor
-j7*^— 1 driven
valve
Hand
valve
Sample
inlet
i ,
To vent
Impinger
train
Flow control
oven
Back
If-
— »•
Organics
modules
,
)
*/
Iff
0(
r
;
i
— A-
c /
/r?v\
;U I
I
/
f
regulator
. . ^
Heat
tracing
Figure 3-2. System schematic.
-------
CO
O
Transducers and
controls
Impactor
stacks
Heated
transport
tube
Probe tip
-Nozzles
Figure 3-3.
Exploded view of HTHP probe.
D-9
-------
co
o
Figure 3-4. Aerotherm HTHP sampling probe and duct interface valve.
-------
a
Figure 3-5. Access valves.
-------
Dowtherm Coolant System
The sample conditioning cooler in the probe assembly receives Dowtherm
coolant from the Dowtherm supply system shown in Figure 3-6. This system in-
cludes a pump, heater, surge tank, cooling coil and controls. During opera-
tion, the heater and heater controller maintain the Dowtherm temperature at
the desired sample conditioning temperature, which in this test series was
set at about 450°F.
Hydraulics
The hydraulic cylinders for probe positioning are supplied by a
portable hydraulic pump driven by an electric motor.
Flow Control Oven
After leaving the throttling valve at the probe exit, the sample
gas is brought to the flow control oven which contains the back pressure reg-
ulator, valves for diverting sample flow from the vent to the organic module
and impingers, and the orifice controlling the flowrate through these compo-
nents. The oven enclosure maintains all of the controls at 450°F, preventing
sample condensation. The flow control oven is shown in Figure 3-7.
Organics and Trace Metal Collectors
The gas collection equipment included in the sampling system consists
of an organics module and impinger train (see Figure 3-7). Both units are
identical to those used in Acurex/Aerotherm's commercially available Source
Assessment Stack Sampler. The organic module cools the sample gas to 70°F
and traps organic vapors in a porous polymer granular bed. The polymer used
in this test series is Rohm & Haas XAD-2 gas chromatographic packing material.
The impinger train uses three high-volume glass impingers to collect trace
metal vapors, followed by a silica gel dryer. For this test series the im-
pinger oxidizing reagents were as follows:
Impinger Solution
#1 6M H202
#2 0.2M (NH4)2S2Og + 0.02M AgN03
#3 0.2M (NH4)2S208 + 0.02M AgN03
#4 Silica Gel
D-12
-------
Figure 3-6. Dowtherm console.
D-13
-------
Flow
Control
Oven
rganic \ Impingers
odule x
Figure 3-7. Flow control oven and gas train.
-------
Controls
The sampling system includes instruments for measuring conditions in
the duct, gas sample, system heaters and coolant. The system also has con-
trols for the sample flowrate, traverse drive, heaters, coolant pump, and
purge gas. Most of the instrument readouts and controls are housed in the
two portable control consoles, shown in Figure 3-8.
3.2 EXXON MINIPLANT TEST FACILITY
This section describes the Exxon Miniplant facility itself, and the
deployment of the sampling system in the facility during the demonstration
tests.
The Miniplant is a pilot scale pressurized fluidized bed combustor
operated by the Exxon Research and Engineering Company in Linden, New Jersey.
The PFBC process is being developed as a more efficient and cleaner method
of burning coal. A sketch of the PFBC system is shown in Figure 3-9. Coal,
along with limestone or dolomite which act as SOp sorbents, is injected into
the bottom of the pressurized boiler. Coal is burned in the limestone bed
which is fluidized by the incoming combustion air. Sulphur dioxide formed
in the combustion process is removed by the limestone bed. Steam coils im-
mersed in the fluidized bed remove some of the heat of combustion and main-
tain the bed temperature in the range of 1500°F to 1700°F. Steam thus
generated operates a steam turbine. The desulphurized flue gas passes
through a particulate removal system and is then expanded across a gas tur-
bine. The particulate removal system must reduce the particulate loading
down to levels sufficiently low to protect the gas turbine and meet current
pollutant emission standards. The Miniplant facility does not presently
include a final gas cleanup device or turbines.
The Miniplant facility consists of the combustor tower and control
building. The combustor is a four-story structure, with platforms at each
level. Stairways connect the platforms. A crane on the top level is avail-
able for moving large equipment. The control building includes a laboratory
area.
D-15
-------
Figure 3-8. Control consoles.
-------
o
i
Coal and'
make-up
sorbent
Boiler
Figure 3-9. Pressurized fluidized bed coal combustor system.
-------
For the sampler demonstration tests, the sampling location was down-
stream of the secondary cyclone (participate removal device), as indicated
in Figure 3-9. At this location, there is a specially constructed duct
section with a sampling port. The sampling port has a 4-inch 300-pound pipe
flange which interfaces with the sampling system access valves. The duct
diameter at the sampling location is 10 inches.
The sampling location was physically located at the top of the com-
bustor tower. When installed, the probe assembly was horizontal, about 4
feet above the platform (see Figure 3-10). The coolant console and hydraulic
pump were also placed on the top platform, near the probe assembly. The con-
trol consoles and gas train equipment were set up one floor below, where a
partial enclosure gave some weather protection.
The route between the laboratory area and the sampling location in-
cluded four flights of stairs and about a 100-foot walk. The sample probe
assembly was hand-carried along this route before and after each sampling
run. Probe cleaning, assembly, disassembly and sample removal were all
done in the laboratory. The lab facility had & scale, oven, desiccator and
other equipment used in sampler preparation arid sample processing. Labware
and materials were supplied by Aerotherm.
The Exxon facility provided a number of utilities supporting the
sampler operation. Power connections, water, and nitrogen supplies are
summarized in Table 3-2. In addition, Exxon supplied technician support
during equipment setup and disassembly.
TABLE 3-2. UTILITY REQUIREMENTS
Electrical:
480 VAC, 3 phase, 40 amp
115 VAC, 15 amp
Water:
5 gpm, 50 psi
Purge Nitrogen:
(flow and pressure required depending on stream
conditions)
• For this test, about 2 scfm at 125 psi
D-18
1 line
6 lines
1 line
1 line
-------
Figure 3-10. HTHP probe assembly installed at Exxon miniplant.
-------
SECTION 4
TEST DESCRIPTION
This section describes assembly and operating procedures for the
sampling equipment and some of the significant events which occurred during
the test series. The narrative of events is divided into sections on: pre-
test activities, sampling runs, and post-test activities.
Procedures
Equipment setup and operation was done according to a formal pro-
cedure which defined proper installation of access valves and probe housing,
probe setup and assembly, system preparations for testing, test sequence,
shutdown and sample removal.
In several cases, we made decisions in the field to change predefined
procedures. For example, the exposure and limited space on the sampling
platform made impactor removal at the sampling location impractical. The
entire probe was carried to the lab for disassembly.
In precleaning the sampling equipment, the procedures in the IERL-
RTP Procedures Manual for Level 1 Environmental Assessment (EPA-600/2-76-160a)
were followed with one exception; the nitric acid passivation of some internal
surfaces of the probe, organic module and flow control oven was omitted be-
cause large acid containers were not available. Sample removal and post-test
cleaning also followed Level 1 procedures.
Pretest Activities
Pretest activities included planned unpacking, setup and checkout,
plus fixing several problems with the facility and sampling system. The
test preparations were completed between March 22 and March 30.
D-20
-------
Heavy equipment was installed with the help of Exxon personnel. For the
nitrogen purge gas, Exxon provided a connection to the facility nitrogen
supply. Exxon also assisted in making a support for the cantilevered
probe housing.
First Sampling Run
For the first test run, the sampling system was set up using the
thimble filter particulate collector and gas train equipment. Following
preheating, the duct access valves were opened and the sampling probe in-
serted into the duct stream. The sample flow control valve was opened
until the flow indicating orifice indicated a sample flow of 0.75 acfm at
nominal particulate collector conditions. When flow conditions were
established, the gas train flow control valve was opened, diverting total
sample flow through the organic module and impinger train. Sampling con-
tinued for 30 minutes. Instrument readings during the test run are listed
in Table 5-1. At the end of the test run, the motor driven sample flow
valve was left open and the sample flow was shut off using the manual ball
valve. The probe was then withdrawn and gate valves were closed. After
cool down, the probe assembly was removed. The probe and gas train were
taken to the lab area for sample recovery and cleaning.
Second Test Run
For the second test run, the cascade impactor was used for particu-
late collection. Since this was to be a very short test with a small
amount of gas sample collected, gas train equipment was not used. Based
on estimated particle concentration and impactor capacity, the maximum
sampling duration was estimated to be between 30 seconds and 1 minute. For
this test run, 30 seconds was chosen. To achieve proper sample flow as soon
as possible, the motor driven control valve was left at the same setting as
the earlier filter run, and on-off control accomplished with the manual ball
valve. As soon as the probe reached the in-stream position, the sample
flow was started. No attempt was made to adjust flow while sampling. After
30 seconds, sample flow was stopped with the ball valve, the probe was with-
drawn and access valves closed. After cooldown and probe removal, the
probe assembly was carried as carefully as possible down the combustor tower
D-21
-------
stairs to the lab. There, the impactor assembly was removed, disassembled
and inspected. The amount and patterns of the catch seemed to indicate
normal operation of the device (see Figures 5-2 through 5-5). However, on
one stage (Stage 4) the substrate shifted slightly, and on another (Stage 7)
some of the jets were plugged. After sample removal and cleaning, the probe
was ready to be set up for the third and final test run.
Third Test Run
The third test run also used the impactor for particulate collection
and omitted the gas train equipment. Based on the lightly loaded appearance
of the 30-second impactor catch from Run No. 2, we increased the duration
of this run to 1 minute. The flow control method for this run was identical
to Run No. 2. The manual valve was again used to start and stop sample flow,
with no attempt to adjust flowrate during sampling. Again the probe was
taken to the lab for disassembly and sample removal. The impactor substrates
were noticeably more heavily loaded than for the 30-second impactor run
(see Figures 5-6 and 5-7). With the completion of sample removal and
cleaning of the sampling equipment, the testing phase was finished.
Post-Test Activities
Following test completion, sampling system hardware was packed and
stored onsite at Exxon in preparation for a follow-on sampling program.
Test samples were brought back to Aerotherm, where some initial analysis
has been performed. The results of this evaluation are described in the
following section.
D-22
-------
SECTION 5
DATA
This section presents the detailed information collected during the
test series. The section is divided into descriptions of test conditions,
instrument readings, and sample properties.
Test Conditions
Plant generating conditions are listed in Table 5-1. The nominal
conditions were identical for all three sampling runs. The facility ran
steadily without interruption during the tests.
Instrument Readings
The sampling system includes a number of instruments which measure
duct, sample and equipment operating conditions. Readings from these in-
struments are presented in Tables 5-2, 5-3 and 5-4.
Table 5-2 gives readings from probe assembly instruments. Duct tem-
perature and pressure were somewhat lower than expected, at about 1360°F and
110 psig. Sample conditioning data showed that the sample gas was cooled
below the desired 450°F in passing through the unheated particle collection
device. In Run No. 1, with the filter collector, minimum temperature after
the impactor was 225°F. Temperatures remained above the dewpoint, however
(207°F for 120 psig, 6-percent water), so no water condensation occurred in
these tests. Correcting the impactor heater malfunction will eliminate the
low collection temperatures for future sampling.
The sample flowrate was maintained within the impactor operating
range throughout the test series (0.8 to 0.9 acfm at orifice inlet condi-
tions). This flowrate gave nozzle velocities which were considerably above
duct velocity (anisokinetic). However, for the high gas temperature and
D-23
-------
TABLE 5-1. TEST CONDITIONS
o
ro
50.4 50.5 50.5
Run: #1 #2 #3
Date
Time
Ambient temperature
Bed Conditions
Temperature
Pressure
Ca/Sulphur Ratio
Excess Air
Coal
Dolomite
Flowrate — scfm
Average Duct Velocity - ft/sec
3-31-77
3:30 p.m.
64°F
1650°F
9 atm
1.25
30%
Champion
Pfzizer
544
6.7
4-1-77
10:30 a.m.
70°F
1650°F
9 atm
1.25
30%
Champion
Pfzizer
546
6.3
4-1-77
2:40 p.m.
67°F
1650°F
9 atm
1.25
30%
Champion
Pfzizer
546
6.7
-------
TABLE 5-2. PROBE INSTRUMENTATION READINGS
Run No. 1
Insertion
Sample Flow
l
Shut-off
Run No. 2
Before Insert
Sample Flow
Run No. 3
Insertion
Sample Flow
Shut-off
Time
3:10 p.m.
3:22 p.m.
3:32 p.m.
3:35 p.m.
10:30 a.m.
2:58 p.m.
Elapsed
Time
(minutes)
0
0
0
0.1
5.0
10.0
15.0
20.0
25.0
30.0
-
0
0.5
0
0
0
0.33
0.67
0.83
1.0
Stack
Pressure
(psig)
(125)a
(125)a
no
109
no
no
109
no
no
no
(121)a
120
025)4
111
Stack
Gas
Temo.
(°F)
(170)
(190)
1360
1350
1350
1340
1320
1320
1320
1320
(195)
1360
190
1360
1360
1360
1360
1360
Dowtherm
Inlet
Temp.
(•F)
431
425
437
437
437
437
437
425
437
437
431
431
431
Dowtherm
Exit
Temp.
(°F)
420
407
435
435
435
435
435
407
435
435
420
420
420
Sample
Temp.
Impactor
Inlet
(°F)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Sample
Temp.
Orifice
Inlet
(°F)
322
332
332
352
373
395
410
421
441
459
332
225
327
327
Sample
Temp.
Transport
Tube Exit
(°F)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Sample
Flow-
rate k
(acfm)b
0
0
0
0.73
0.64
0.82
0.83
0.85
0.87
0.88
0
0.90
0
0
0
0.91
0.39
0.88
Impactor
Heater
Temp.
m
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Transport
Tube
Heater
Temp.
(°F)
450
452
450
450
450
450
445
460
460
460
445
445
445
o
-------
TABLE 5-3. GAS TRAIN INSTRUMENT READINGS
o
i
ro
CTl
Run No. 1
Start Flow
Stop Flow
Time
3:10 p.m.
3:22 p.m.
3:32 p.m.
Elapsed
Time
0
0
0
0
5.0
10.0
15.0
20.0
25.0
27.65
Flow-
rate
(acfm)
la
0
0
0
0.73
0.64
0.82
0.83
0.85
0.87
0.88
Transport
Line Temp.
(°F)
444
453
-
-
449
449
447
447
456
450
Flow Control
Oven Temp.
(°F)
449
-
-
-
396
425
458
454
446
444
Organic
Module
Temp.
(°F)
58
-
-
-
70
72
74
75
76
75
Impinger
Train
Temp.
(°F)
70
-
-
-
69
69
70
71
73
73
Ambient
Temp.
(°F)
65
64
65
67
67
65
66
aAt orifice conditions - see Table 4-2.
-------
TABLE 5-4. ANISOKINETIC CORRECTION FACTORS
o
ro
Run
1
2
3
Sample
Flowrate
(acfm)
0.85
0.90
0.90
Nozzle
Velocity
(ft/sec)
9.3
13.0
11.3
Estimated
Duct
Velocity
(ft/sec)
6.7
6.3
6.7
Velocity
Ratio
0.72
0.48
0.59
Particle Concentration3
Correction Factor
measured
Ctrue
0.99
0.98
0.99
Calculation per method in Handbook of Aerosols, TID-26608, 1976, Section 5.1-1 and
Figure 5-2.
-------
pressure, fine particles and low velocities involved, the variance from
isokinetic conditions has an insignificant effect on measured particulate
concentration. A comparison of duct and sampling velocities and the calcu-
lated correction factors for anisokinetic conditions is presented in
Table 5-4. As shown, the measured particulate concentrations are within 1
or 2 percent of isokinetic measurements.
Table 5-3 lists the instrument readings from the gas sampling equip-
ment used during Run No. 1. Gas sample flow was started shortly after
particulate sampling began, so the total elapsed time is less than shown in
Table 5-2. During gas sampling, all sample flow was diverted to the gas
train, so the flowrates given for the gas train are the same as those
through the particulate collector. The temperature readings show that all
gas train components were operating correctly.
During the test series, we measured the surface temperatures of the
access port, valves and probe housing. These readings are presented in
Table 4-5. The valve surface temperature remained, below 258°F at all times.
Accessible surfaces of the probe housing also remained cool, below 167°F.
Sample Properties
The tests produced data on particulate concentration, size distribu-
tion, appearance and chemical composition, ahd on moisture content. The
organic and trace metals samples have not yet been analyzed, so this report
contains no data on these constituents.
The measured particle concentrations are listed in Table 5-6. The
values of 0.43 to 0.64 grain/scf are reasonable compared to other measure-
ments made in unpressurized portions of the Exxon process. These have
ranged from 0.2 to 1.2 grain/scf. The 0.64 grain/scf value from the 30-
minute sample is the most accurate measurement from our tests. It comes
from the largest sample and best defined conditions.
The moisture content measured in Run 1 was 6.2 percent by volume.
This compares well with Exxon's preliminary estimate of 5.8 percent.
D-28
-------
TABLE 5-5. STRUCTURE TEMPERATURES
Pretest (11:00 a.m., 3-30-77)
Duct Wall
349°F
Nozzle Cylinder
Top 294°F
Side 280°F
Bottom 189°F
Valve - Duct Side
Top 258°F
Side 220°F
Bottom 195°F
Run No. 1 (3:30 p.m., 3-31-77)
Time
0
10 min
20 min
60 min
Valve -
Probe
Side
147
159
157
190
Inner
Probe
Housing
133
131
133
Outer
Probe
Housing
133
154
167
D-29
-------
TABLE 5-6. PARTICULATE CONTENT
Run #1 Run #2 Run #3
Date: 3-31-77 4-1-77 4-1-77
Time: 3:30 p.m. 10:30 a.m. 3:00 p.m.
Particle Catch:
(grams)
Filter 3.2515
Impactor - 0.0554 0.0892
Residue 1.8565 0.0334 0.0595
Total 5.108 0.0884 0.1497
Sample Volume: 122.5 3.13 5.03
(scf)
Particle Content: 0.64 0.43 0.46
(grain/scf)
Particle Content: 0.65 0.44 0.47
(gr/scf)
(Anisokinetic
Correction Applied)
D-30
-------
Particle size distribution information is presented in Figure 5-1
and Table 5-7. As shown, there is some difference in the results from the
two impactor runs. However, both show that most of the particulate falls
within the 1 to 10 micrometer range.
The impactor substrates are shown in Figures 5-2 through 5-7.
Generally, the patterns are typical of normal impactor operation. Stage 7,
however, shows evidence of several plugged jets. A comparison of Figures
5-6 and 5-7 show the differences in particulate loading for substrates from
Run 2 and Run 3. Run 2 substrates were lightly loaded, while those from
the longer duration Run 3 showed heavy, three-dimensional deposits.
The particulate sample from Run 2 were photographed using a scanning
electron microscope (see Figures 5-8 through 5-11). The particulate is ir-
regular in appearance, suggesting that it may be calcium sulphate crystals
from the dolomite bed and ash from low-temperature combustion. Some of
the photos show congealed masses of particles. The cause of this phenomena
could be any of the following: a property of the collected particulate,
condensation on the particulate or the conductive spray applied to the
sample for SEM photography.
The chemical composition of collected particulate was analyzed by
dispersive X-ray fluorescence analyzer. Spectra of X-ray emissions are
shown in Figure 5-12. The analysis shows detectable amounts of aluminum,
silicon, calcium, sulphur, iron, potassium, titanium and copper.
D-31
-------
100
3
6
Percentage smaller (by weight)
10 20 30 40 50 60 70 80 90
98
§
N
I/)
c
O
0
•I—
E
OJ
N
I/)
OJ
O
10
8
1.0
8
6
0.1
Figure 5-1. Particle size distribution,
D-32
-------
TABLE 5-7. PARTICLE SIZE DISTRIBUTION
OJ
00
Stage
1
2
3
4
5
' 6
7
Filter
D50
Microns
26.0
12.0
4.3
2.1
1.2
0.6
0.3
Run #2
Weight
Collected
0.0076
0.0080
0.0171
0.0139
0.0022
0.0036
0.0020
0.0010
0.0554
% Total
Weight
13.7
14.4
30.9
25.1
4.0
6.5
3.6
1.8
grams
%
Smaller
86.3
71.8
41.0
15.9
12.0
5.4
1.8
Run #3
Weight
Collected
0.0093
0.008
0.0221
0.0215
0.0135
0.0081
0.0039
0.0028
0.0892
% Total
Weight
10.4
9.0
24.8
24.1
15.1
9.1
4.4
3.1
grams
%
Smaller
89.6
80.6
55.0
31.7
16.6
7.5
3.1
-------
Figure 5-2. Impactor substrates.
Figure 5-3. Impactor substrates.
D-34
-------
Figure 5-4. Impactor substrates
Figure 5-5. Impactor substrates
-------
0
Figure 5-6. Impactor substrate Run 3, Stage 5.
-------
0
Cit
Figure 5-7. Impactor substrate Rune 2, Stage 5.
-------
in m
1000X
10 Microns
3000X
3 Microns
Figure 5-8. Particle photomicrographs Stage 1.
D-38
-------
1000X
10 Microns
3000X
3 Microns
Figure 5-9. Particle photomicrographs Stage 2.
D«
•~ 1
-------
1000X
10 microns
1 *
*. «*
Figure 5-10. Particle photomicrographs Stage 4.
D-40
-------
**
3000X
3 microns
10000X
1 micron
Figure 5-11. Particle photomicrographs Stage 6.
D-41
-------
Stage 1, Run 2
Stage 6, Run 2
Figure 5-12. Particle chemical composition.
D-42
-------
Appendix E
Acurex Corporation/Aerotherm Division
March 1977
Aerotherm Project 7237
IERL-CRB SAMPLING MANUAL FOR
LEVEL 1 ENVIRONMENTAL ASSESSMENT
AEROTHERM REPORT TM-77-160
Prepared for
EPA Project Officer: William B. Kuykendal
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Contract 68-02-2153
Task 03
E-i
-------
INTRODUCTION
This sampling procedures manual has been prepared for the Process Measurements Branch of the
Industrial Environmental Research Laboratory, EPA. The procedures that are described apply to
Level 1 environmental assessment of combustion facilities located at Research Triangle Park, North
Carolina.
Under the present assessment program, only combustion facilities are considered. It is an-
ticipated, however, that future needs will arise for extension of sampling equipment and analysis
techniques to noncombustion facilities as well. This manual deals with flue gas and influent fuel
stream sampling for specific combustion facilities. Sampling includes stack gases and vapors, stack
gases that contain particulate, and liquid fuels. Other Level 1 sampling requirements (fugitive
emissions, slurries, and solids) are not of concern here.
The approach to Level 1 environmental assessment sampling consists of five functions:
1. Test planning
2. Preparation of sampling systems
3. Obtaining the samples
4. Recovery of the samples from sampling equipment
5. Data reduction
Each of these functions is described in a separate section of this manual. At the end of Sections
1 to 4, an outline of the procedures to be used during that phase of the assessment is provided for
the operator's use.
Completion of the sampling procedures provides contained samples ready for analytical evalu-
ation. Analysis procedures to be performed in the IERL Level 1 EA laboratory are provided under a
separate manual prepared by A. D. Little, Inc. (Reference 1).
For reference, a list of the IERL-CRB combustion research facilities is given below.
• Aerotherm experimental furnace
E-ii
-------
i Residential furnaces (two)
t Avco and North American firetube boilers
• Caterpillar diesel engine
0 Dynamic Science Unit -model package boiler
• Solar gas turbine
Table 1 briefly describes each facility and its sampling conditions.
E-iii
-------
TABLE 1. IERL-CRB COMBUSTION RESEARCH FACILITIES
Facility
Description
Stack/Port Characteristics
Stack Gas Properties
1. Aerotherm Experimental Furnace
2. Residential Furnaces (2)
3. Avco and North American
firetube boilers
4. Caterpillar Diesel Engine
(D334)
5. Dynamic Science Unit
Model Package Boiler
6. Solar Gas Turbine
280,000 Btu/hr capacity
Fuels: oil, natural gas, methanol
Primary and secondary air systems
Exhaust gas recirculation system
Tube heat exchanger for cooling of
exhaust gases
Oil- or gas-fired home heaters.
Furnaces change periodically, at-
taching to permanent stacks
2 to 2.5 million Btu/hr
Fuels: #2 oil, natural gas
Steam recycled by tube heat
exchangers
Precombustion chamber design with
supercharge
Speed: 1,800 rpm
Used for power generation
Electrical power converted to heat
Located exterior to Wing G with
full weather housing
4 million Btu/hr with #2 or #6
oil or natural gas, Dowtherm heat
exchange system
Exhaust gas recirculation system
Dual impeller design for power
generation
Electric power converted to heat
for discharge
Located exterior to Wing G with
full weather housing
8" stack - has an elbow above
mezzanine level and runs
horizontally into a blower. Sample
ports located in insulated
horizontal section
6" galvanized stacks, uninsulated,
sampled 3' above mezzanine level
A damper is provided above the
sample location in each stack for
draft
14" stacks, insulated Avco unit
has sample port ~3' above mezzanine
level. North American unit has
offset in stack above mezzanine —
sample port located near mezzanine
level downstream of offset
12" stack, straight pipe attached
to horizontal engine muffler
Sample location ~3' above mezzanine
15" SST stack with surrounding pro-
tection cage
Sample locations are above EGR en-
trance into stack
16" stack, lower half included in
unit housing
Contains muffler
Sample locations are ~3' above
mezzanine
400°F to 500°F
Slight positive or negative
pressure
Velocity: 80 to 260 fpm
250°F to 600°F
Slight negative pressure
Velocity: 100 to 325 fpm
250°F to 600°F
Slight positive pressure
Small fluctuations may occur
due to flame instabilities
Velocity: 450 to 850 fpm
300° F to 900° F
Slight positive pressure
Velocity: 500 to 3,000 fpm
900°F to 1,500°F
Slight positive pressure
with some flame-induced
pulsations
Velocity: 650 to 1,000 fpm
500°F to 1,000°F
Slight positive pressure
Velocity: 6,000 to 14,000
fpm
-------
TABLE OF CONTENTS
Section page
1 SAMPLE PLANNING E-l
1.1 Personnel Training . E-l
1.2 Identification of Safety Requirements / E-2
1.3 Identification of Test Conditions and Data Requirements E-3
1.4 Sample Schedule and Manpower Requirements E-4
2 SAMPLE SYSTEM PREPARATION E-6
2.1 Source Assessment Sampling System E-6
2.2 Gas Grab Sampler E-12
2.3 Continuous Gas Analysis E-14
2.4 Liquid Fuel Sampling E-15
3 SAMPLING E-l 7
3.1 Sampling with the SASS Train E-17
3.2 Gaseous Grab Sampling E-24
3.3 Liquid Fuel Sampling E-27
4 SAMPLE RECOVERY E-30
4.1 SASS Train Breakdown E-30
4.2 Sample Extraction E-30
4.3 System Cleanup and Reassembly E-31
5 DATA REDUCTION E-33
REFERENCES E-34
APPENDIX A - SAMPLE DATA SHEETS E-35
APPENDIX B - LIST OF SAMPLING EQUIPMENT E-42
LIST OF ILLUSTRATIONS
Figure Page
1 Sample Planning Procedures E-5
2 Source Assessment Sampling System (SASS) E-7
3 Liquid Cooled Sampling Probe E-9
4 Probe Coolant System Schematic E-10
5 Grab Sample Probe E-l3
6 System Preparation Procedures E-l6
7 SASS Sample Port with Gate Valve E-20
8 Residential Furnace SASS Sample Port E-21
9 CRB Combustion Facility Stack Velocities E-22
10 Gas Grab Sample Port with Ball Value E-25
11 Residential Furnace Gas Grab Sample Port E-26
12 Sampling Procedures E-28
13 Sample Recovery E-32
E-v
-------
SECTION 1
SAMPLE PLANNING
Planning an environmental assessment must assure that a representative sample is extracted
from the process stream and provided to the EA laboratory for analysis. It is the intent of sample
planning to identify the procedures that are to be followed in obtaining specific emission data,
provide familiarity with sampling equipment and process operating conditions, and to identify poten-
tial sampling problems that may be encountered during the test. Coordination with laboratory per-
sonnel to identify sample container and analysis time requirements during the planning stage will
facilitate procedures during the sampling period.
1.1 PERSONNEL TRAINING
An important step in planning an environmental assessment is to familiarize all involved per-
sonnel with Level 1 sampling and analysis procedures. These procedures are documented in Reference 2.
For specific application of sampling procedures to combustion facilities at RTF, particular attention
should be directed to Chapters 1, 2, 3, and 5 which address overall Level 1 sampling strategy for
gaseous, particulate, and liquid phase sampling.
The Level 1 EA Procedures Manual also describes sampling equipment with which personnel must be
thoroughly familiarized prior to testing. Hardware of specific interest includes: (1) the Source As-
sessment Sampling System (SASS) for collection of particulate, trace elements, and organic compounds,
(2) a grab sampler system which collects gaseous phase emissions for subsequent analysis by gas chro-
matography, and (3) equipment required for collection of liquid fuel samples. Conceptual descrip-
tions of each type of sample hardware are presented. In addition, continuous gas analyzers may be
used for environmental assessment of specific gas species that are not routinely analyzed by gas chro-
matography in the Level 1 laboratory.
Specific hardware description for the SASS is presented in Reference 3. Assembly, setup, dis-
assembly, transport, and operation of the SASS must be totally familiar to all sampling personnel.
Since details of the gas grab and liquid fuel sample systems are not provided in other docu-
ments, a brief description of each is presented in Section 2 of this manual.
E-l
-------
1.2 IDENTIFICATION OF SAFETY REQUIREMENTS
Safety precautions to be exercised in the use of the Level 1 sampling systems are stated in
Section 3 of this manual. Potential hazards exist due to the high temperatures within and surround-
ing the combustion facility effluent stacks. Proper use of all sampling equipment at each of the
sampling sites will provide safe working conditions. Procedures for working in the high-bay area
should be followed explicitly.
NOTE
If the sampling operator feels that all safety requirements have not been satisfied
for all personnel involved, it is his obligation to discontinue all sampling proce-
dures until the unsafe condition is corrected.
1.3 IDENTIFICATION OF TEST CONDITIONS AND DATA REQUIREMENTS
The identification of test conditions can facilitate sampling procedures by providing the
operator with knowledge of probable stack conditions. The variables that should be identified by
the combustion facility operator for the sampling crew include:
• Fuel type and consumption rate
• Stoichiometry
• Changes in test configuration that may affect sampling operations
• Changes in operating conditions that may affect sampling results
• Period of operation
In many instances, stack conditions can be calculated or approximated from past test experi-
ence. Pretest knowledge of stack temperature and velocity are of particular value to the sampling
crew. Moisture content, gas composition, and particulate content of the combustion gases are also
of interest prior to testing as considerable preliminary sampling effort can be saved when these
variables do not have to be specifically measured.
Specific data requirements for a Level 1 environmental assessment are subject to EPA objec-
tives and formulated during the test planning stage. In general, data pertinent to the CRB combus-
tion facilities will include analyses for:
• Collected SASS samples
— Particulate
— Trace elements
— Organic compounds
E-2
-------
t Stack gases
- General inorganic combustion species (CO, C02, 02, NO , H20, etc.)
- Sulfur compounds
- Light hydrocarbon species (C1 to Cg hydrocarbons)
• Influent liquid fuel streams
The sampling operator should formulate test objectives and requirements with the facility manager
during test planning. He must also be familiar with the analysis techniques applicable to each
sample type, such that valid samples can be provided to the laboratory. Refer to Chapters 7 to 10
of Reference 2 for details of analysis techniques.
Sampling system data must be taken by the crew during operation of the systems. SASS data
requirements include periodic recording of the control module readouts:
• Pi tot tube pressure differential
• Orifice flow pressure differential
• Initial and final gas meter readings
• Sample elapsed time
• Stack temperature
• Gas cooler temperature
• Impinger outlet temperature
• Oven temperature
• Inlet and outlet gas meter temperatures
• Probe coolant temperatures
t Pump vacuum
Continuous and grab gas and liquid fuel sampling require records of:
• Sample time
• Gas concentrations from continuous analyzers
Sample data sheets are provided in Appendix A of this document for:
• Velocity and temperature traverse
• Stack gas moisture measurement
E-3
-------
• Stack gas composition (molecular weight) measurement
• Particulate sampling data
• Gas grab and liquid fuel data
• Continuous gas analysis
Similar sheets should be prepared for each of the functions required by the test plan. Final copies
(post-test) are provided to the facility manager, sampling operator, and laboratory personnel.
1.4 SAMPLE SCHEDULE AND MANPOWER REQUIREMENTS
Early identification of facility test periods allows planning of the environmental assessment
schedule. Since the SASS is the most time consuming of the sampling systems (~4 hours operation
per test point), its start time should be as early in the test period as possible to ensure adequate
sample time. Taking supplemental gas grab samples, liquid fuel samples, and operation of continuous
gas analyzers can then be performed periodically between SASS adjustments.
Two men are required for SASS setup. During operation, at least one person (and periodically
both) can be made available for gas and fuel sampling operations.
All steps identified above for sample planning are shown schematically in Figure 1.
E-4
-------
Personnel Training
Identify Safety
Requirements
Identify Test Conditions and
Data Requirements
Formulate Sampling Schedule,
Manpower Requirements
Familiarize with Level 1 sampling and
analysis procedures
Familiarize with applicable Level 1
sampling hardware
- SASS
- Gas grab sampler
- Liquid fuel sampling
— Continuous gas analyzers
t Working in high temperature areas
• Sampler/stack interface procedures
• System component operational
procedures
• Facility parameters
— Fuel and consumption rate
- Stoichiometry
— Facility configuration
- Operating period and conditions
• Prepare data sheets
• SASS assembly, setup, operation,
disassembly
• Gas grab sampling
• Liquid fuel sampling
t Continuous gas sampling
Figure 1. Sample planning procedures.
E-5
-------
SECTION 2
SAMPLE SYSTEM PREPARATION
The sampling systems that are used for Level 1 assessment of the CRB combustion facilities
and their sample types are:
• Aerotherm Source Assessment Sampling System (SASS) with special liquid cooled probe -
particulates, trace elements, organic compounds
• Gas grab sampler — stack gases
t Continuous gas analyzers - specific stack gas species
• Liquid containers — fuel sampling
Each sampling apparatus is described below. For reference, a complete list of all sampling
equipment to be assembled for the assessment is given in Appendix B.
2.1 SOURCE ASSESSMENT SAMPLING SYSTEM
A schematic of the Source Assessment Sampling System is shown in Figure 2. Main components
consist of a probe for extraction of gases from the stack, an oven with cyclones and filter for col-
lection of particulate matter, a gas cooler/organic adsorption module, impingers for trace element
collection, and gas transport and metering equipment. A single SASS has been provided for all
Wing G combustion facilities that interfaces with each stack. A special liquid-cooled probe inter-
faces with the standard SASS and is applicable to all combustion facilities in Wing G.
Before full setup of the SASS train at the sample location, the probe, cyclone and filter
holder, gas cooler, and impinger components should be prepared and partially assembled in the labo-
ratory. Two persons are required for these functions. Procedures for precleaning SASS components
are presented in Section 3.3 of Reference 2. Assembly of the cyclones, filter, probe, oven, and
impingers is outlined in Reference 3. Final assembly of the system at the sample location can be
completed after transport of assembled components from the laboratory.
A sampling cart to be used to transport sampling equipment to the sampling site is provided.
The cart can also be used during operation of the SASS, supporting the control module, gas cooler,
E-6
-------
Stack T.C.,
~N
Convection
r~ oven
Filter
Organic
adsorbent
cartridge
^S^s&i
Dry gas meter
Orifice meter
Centralized temperature
and pressure readout
Control module
— &±-
_r\_
i
Gas cooler
Imp/cooler
trace element
collector
10 cfm vacuum pumps
Impinger
T.C.
Figure 2. Source Assessment Sampling System (SASS).
-------
and impinger components. It is recommended that the control module and vacuum pumps not be trans-
ported to the laboratory on a routine basis, but hand-carried to the stack to be sampled. The
sample cart can be wheeled to the Wing G area, but components must be hand-carried to the mezzanine
level.
The gas sampling probe and cyclone oven are supported horizontally at the stack by a unirail
support system. Setup details for the unirail are detailed in Reference 3. The unirail should be
setup well in advance of facility startup to allow working around the stack in an ambient tempera-
ture environment. Once the unirail is in place (clamped securely to the sample port), the probe
is connected to the oven as shown in Reference 3, and the assembly mounted on the unirail support.
Assembly of the system is completed by connection of sample lines from the oven to the gas cooler,
the impingers to the sample pumps, and from pumps to the control module.
The sampling probe provided for the SASS train (a nonstandard component of the SASS,
shown in Figure 3 with support equipment) is temperature-conditioned by two independent cooling
fluids. A closed-loop oil circulation system is provided to condition and control the temperature
of the sample gases flowing through the probe liner. The outer fluid (water) maintains the probe
outer sheath temperature at a safe working level. The gas sample is to be conditioned to 400°F at
the entrance to the cyclone oven under all sampling circumstances. As a consequence, there are
sampling conditions where the oil system is either extracting heat from or providing heat to the
sample gases (dependent on stack temperature).
A temperature control unit supplied with the probe maintains the oil temperature at a set
value. This unit contains an electrically heated oil reservoir that supplies heat to the circulat-
ing oil. Heat may also be extracted from the oil by circulating through the adjacent finned tube
heat exchanger rather than the reservoir. A fan is provided on top of the heat exchanger to aid in
cooling. Switching between the heating and cooling loops of the unit is provided automatically by
a solenoid valve controlled by the dual setpoint temperature controller at the base of the unit.
A schematic of the probe system is shown in Figure 4.
Setup of the probe includes connection of both oil and water systems to their respective
supply and return connections on the probe body. Locate the oil pump in the probe vicinity to allow
traversing. To connect the probe hydraulic oil system, attach the supply and return metal braided
hoses to the two most-rearward tubes on the probe body by the quick-disconnects provided. The
oil supply line is permanently attached to the pump outlet (hand-valve side). The return line is
provided with disconnects at each end which are connected to the probe body at one end and the top
of the oil reservoir tank at the other end. This hose is provided with one female and one male end
E-8
-------
c
CO
Figure 3. Liquid cooled simpling probe.
-------
Reservoir
Water Cooled Sheath
il Conditioned M 11 i
Liner /
/
v \
V \
/ w t
f
s ^
\ \
'/
9'
1
9'
5
^_rv^
Heater
--i \ (
» U
— ^
4' *
Oil pump Oil conditioning module
25'
25'
'•B
Water pump
Cooler
Pressure
indication
115V/15 amp
60 cycle
Power switch
115V
60 cycle
Ice bath
Figure 4. Probe coolant system schematic.
-------
and can only be connected in one direction. Completion of the oil system is accomplished by con-
nection of the pump supply line to the bottom of the oil reservoir with the appropriate disconnect.
The water coolant system is connected by two garden hoses at the front-most fittings on the
probe body. Water circulation can be provided by either the water pump provided with the probe
system and an ice bath (see Figure 4) or a pressurized water system (city water or closed-loop).
If the ice bath option is employed for the water system, between 25 and 100 pounds of ice per hour
will be required, dependent upon the heat load.
CAUTION
Both water and oil should be supplied to the system prior to startup of circulation
pumps. Startup details are provided in Section 3.
The probe umbilical connection is somewhat simplified over that of the standard SASS probe.
As the umbilical provided with the system was constructed for the standard probe, it will contain
power connections that are not required for the liquid probe. The connections required are:
• Pi tot tube pneumatic lines (color-coded)
• Stack gas thermocouple
The umbilical connection is completed to the impinger, oven, and gas cooler thermocouples as de-
scribed in Reference 3.
Both electrical power and cooling water are required for operation of the SASS and probe.
Power requirements are one 115V, 15A circuit for each of the following units:
t Sample probe temperature control unit
• Oil pump
• Control module
• Gas cooler
• Sample vacuum pumps
In addition, a cube or crushed ice supply (-15 Ib/hr) is required for the impinger bath. Filling
of the water circulation system for the impinger ice bath requires several gallons and operates as
a closed-loop system after filling.
The final function to be performed prior to sampling with the SASS is to leak-check the
assembled train. The probe nozzle inlet is blocked off (a plug fitting is provided with the probe)
while 15 inches Hg vacuum is pulled on the system by the sample pumps as per the SASS operating
E-ll
-------
manual. If the noted flowrate is less than 0.05 cfm, the system is then ready for sampling. If
the leak rate is greater than 0.05 cfm, the source of the leak must be found and corrected.
Periodic calibration of the pitot tube, orifice meter, and dry gas meter SASS components is
required to maintain sampling accuracy. The following calibration program should be followed for
Level 1 environmental assessment. All calibration should be performed on schedule and logged in a
convenient location.
The accuracy of the stack flow velocity measurement is dependent upon knowing the coefficient
C of the S-type pitot tube. Calibrations are performed as follows:
• Prior to the first field test and at 1-year intervals - calibrate pitot tube by EPA
Method 2 as described in Reference 4
• Prior to and after each system use - visually inspect pitot tubes for signs of damage,
recalibrate if damage is noted
The orifice meter and dry gas meter provide accurate measurement of the sample gas flowrate
and total volume. Calibration procedures are detailed by EPA Method 2 and in Reference 3. The
Level 1 calibration schedule is as follows:
• Prior to the first field test and at 1-year intervals — calibrate as per EPA Method 2
• Prior to and after each system use — perform calibration check by comparing the volumes
measured by the dry gas meter and the orifice meter (flowrate multiplied by the run time).
Recalibrate both components if the measured volumes do not agree within ± 5 percent.
The above calibration check should be performed for at least two system flowrates. Those recommended
are 4.0 cfm and 8.0 cfm. An example of the check is as follows. The system flowrate is adjusted to
a steady-state condition at the approximate recommended flowrate. A timed interval is established
during which the orifice meter is read successively and the total gas meter volume is recorded.
The volume of gas passing through the orifice meter is the product of the run time and the flowrate
(corrected to calibration temperature conditions)-and is compared to the gas meter volume (also
corrected to calibration conditions).
2.2 GAS GRAB SAMPLER
Stack gases are to be grab-sampled from the stack periodically according to the test plan.
These samples are simply extracted from the stack into an evacuated flask and transported to the
laboratory for analysis.
E-12
-------
CO
Encased, evaluated
glass sample bottle
Figure 5. Grab sample probe.
-------
The grab sample system is shown in Figure 5. A stainless steel tube and ball valve are
fixed permanently to each facility stack. The sampling probe (stainless steel with a ceramic liner)
slides in through the fixed-tube port and is connected to the sample bottle by a flexible line.
The flexible line contains a pyrex glass wool plug to prevent particulate influx to the sample
bottle. Complete operating details are presented in Section 3.
Several functions are required in preparation of the grab sample system. Since the quality
of the sample is dependent upon maintaining the original composition of the gas from stack to
analyzer, the probe and connecting tube inner surfaces should be made inert prior to use. This
function is accomplished by allowing the probe inner tube surface to remain in contact with a 1:1
(v/v) aqueous nitric acid solution for a 1-hour period as performed for SASS train components.
Three successive cleaning stages follow with distilled water, isopropyl alcohol, and methylene
chloride (see Section 3.3.3.1 of Reference 2). The glass sample bottle, encased in its protective
cover and having been similarly cleaned and evacuated by laboratory personnel, is connected to the
flexible tubing and probe. The evacuation pressure and temperature should be labeled on each sample
bottle. The probe/bottle assembly, as well as additional sample bottles, are then transported to
the sample location.
2.3 CONTINUOUS GAS ANALYSIS
Certain combustion gas species generated in the Wing G facilities cannot be routinely ana-
lyzed by the grab sample/gas chromatography technique of the IERL Level 1 laboratory. These
species, which may include:
• Nitric oxides — NO, NO
X
• Carbon dioxide - C02
i Carbon monoxide - CO
• Oxygen - Op
as well as others of interest, can be analyzed by a continuous gas analysis system. The analyzers
require initial calibration with a standard calibration gas prior to providing continuous monitor-
ing of the stack gases. Setup of the system includes this calibration and provision of a sample
line from the analyzer bench to the appropriate stack. Alternatively, gases can be injected into
the analyzers from grab sample containers if a sample line to the stack is not available.
E-14
-------
2.4 LIQUID FUEL SAMPLING
The method of sampling liquid fuels in the Wing 6 facilities is extraction by tap and collec-
tion directly in a glass laboratory container. The container should be prepared according to
Level 1 procedures (cleaning, etc.) by laboratory personnel such that it is ready for transport to
the sample location.
Figure 6 shows the sample system preparation procedures described above.
E-15
-------
Source Assessment
Sampling System
Gas Grab Sampler
Continuous Gas Analysis
Liquid Fuel Sampling
• Collect all system components
• Perform periodic component calibrations (if required)
• Prepassivate and clean all sample contact surfaces for 1 hour
with 1:1 (v/v) nitric acid
- Probe liner and nozzles
- Cyclones and filter holder
— Heated sample line
— Gas cooler components
- Impinger bottles
t Subassemble cleaned components
• Transport components from laboratory to sample location
• Setup components at stack
- Unirail support
— Suspend oven and probe on unirail
— Place control module, sample pumps, probe heater, probe pump,
and probe ice bath (if used) at sample site
• Connect SASS component sample hoses
t Connect probe coolants
— Oil system
- Water system
• Connect probe umbilical
- Pi tot tubes
— Thermocouples
• Provide power, water connections
• Leak check assembled system
• Collect sample probe, tubing, sample bottles
• Prepassivate and clean probe and tubing
• Assemble probe and tubing
• Transport components to sample location
• Provide sample line, filters, and water removal device from sample
gases
• Calibrate analyzers with standard calibration gases
• Secure cleaned sample bottles from laboratory
• Transport to sample location
Figure 6. System preparation procedures.
E-16
-------
SECTION 3
SAMPLING
The sampling procedures for the SASS, grab sampler, and liquid tap sampling are discussed
in this section. Procedures for continuous gas analysis are specific to each analyzer and sampling
personnel are referred to the appropriate operator's manuals. Each Level 1 assessment will begin
and end with attention to the SASS train — its operation being the most time consuming of the sam-
pling procedures. Gas and liquid fuel samples can easily be taken after the SASS has achieved
steady operation.
The two persons that are required for SASS setup will also be required to perform the sam-
pling operations. One person can perform startup, operation, and shutdown procedures for the SASS.
The second person is then available to perform grab sampling and monitoring of continuous gas
.analyzers.
3.1 SAMPLING WITH THE SASS TRAIN
In order to reach a steady sampling operation with the SASS train, several functions are re-
quired:
• Attachment of the probe nozzle or pitot tubes (or both) to the probe
• Calculation of sampling points for velocity and temperature traverse by EPA Method 1
(Reference 4)
• Warmup of the probe coolant fluid to achieve proper sample conditioning i
• Insertion of the probe body into the stack
• Measurement of stack velocity and temperature by traversing
• Measurement of static pressure, ambient pressure,, moisture content, and gas composition
• Selection of the average velocity sampling point
• Selection of the proper nozzle size to achieve isokinetic sampling conditions
t Initiation of sample flow
E-17
-------
In general, stack velocity measurements will be required before actual sampling can begin.
The usual measurement method involves attachment of the pitot tubes to the SASS probe. The probe
is then inserted into the stack and a stack traverse made from which a velocity profile can be gen-
erated (one-dimensional). After calculation of the sample point of average stack velocity, the
proper nozzle size for isokinetic sampling can be selected.
Connection details for pitot and nozzle are given in Reference 3. Nozzle selection proce-
dures are also detailed.
The assessment test plan will require that a velocity traverse be performed when
• The average velocity sample point is not adequately known from previous testing
• Test conditions are significantly different from those of past assessments
The appropriate sample points across the duct are then calculated by EPA Method 1. The calculation
procedure for circular ducts is described in Reference 4.
Before insertion of the probe into the stack, probe cooling systems must be allowed to reach
steady state. Turn the probe power switch on and set the temperature controller to approximately
250°F. Allow the fluid to heat (-30 minutes). Similarly, the SASS oven must be heated to 400°F in
order to maintain a constant gas flowrate through the cyclone train.
Following heatup, turn the probe oil circulation pump on. The pump will begin laboring
hard due to the low temperature fluid in the lines and the probe. Soon after the pump has been
started, the motor should start to speed up and the pressure light on the module should extin-
guish (1 minute). Allow the probe liner to warm up (2 minutes).
Should the pump stall before warm fluid is passing through its line, adjust the relief valve
on the pump by turning the handle on the valve counter-clockwise until the motor can handle the
load. Once the pump is turning easily readjust the valve until the pressure light goes out. Now
allow some warm fluid to enter the cooler by turning the temperature control 50°F lower for approx-
imately 1 minute. The solenoid valve will chatter for a few seconds, and after the exchange of
fluid has taken place, settle down. Submerge the water pump in the ice bath (if used) and place
the ice bath in a convenient location. The water pump is rated at 24-foot 6-inch pressure shutoff
and 810 gph. Do not place the pump more than 12 inches above or below the probe before the lines
are filled with water.
Insure that the oil bath has reached the desired temperature, the pressure light is out, the
fan is running, the water is circulating, and the probe is ready for insertion.
E-18
-------
With the pitot tube and/or nozzle securely attached to the sample probe, and the probe fully
heated by the steps above, the probe can be inserted into the stack. Two sample port types are
provided on the Wing G facility stacks. The first, used on all facilities except the residential
furnaces, contains a gate valve for sealing of the combustion gases from the high-bay area. The
second, used only on the residential furnace stacks contains no valve but is a simple T section
through which the probe can be passed. Figures 7 and 8 show the two sample port configurations,
respectively.
Before inserting the probe into the gate valve port, the port cap shown in Figure 7 is
slipped over the end and onto the body of the probe. Verify that the port seal is in place on the
outlet edge of the port. The probe is first inserted through the seal and into the port to
Position 1 in Figure 7. The port cap is then screwed firmly to the port to hold the seal in place.
The port is now effectively sealed such that the gate valve can be opened without escape of stack
gases. With the gate valve in the full-open position, the probe is slid further into port Posi-
tion 2. The probe is generally left in this position (out of the gas stream) until sampling is to
begin. For sampling, the probe is slid further to the appropriate location (Position 3, Figure 7).
The above procedure is simply reversed for extraction of the probe.
CAUTION
The probe outside diameter must fit securely through the port seal before opening of
the gate valve. Care must be taken to avoid contact with hot stack and port surfaces.
Each residential stack port is covered with a cap which is removed just prior to probe in-
sertion. Since negative stack pressures are present, no gas will be emitted to the high bay. A
special port seal for the residential stack is supplied that is slipped over the probe (Figure 8)
before removal of the port cap. With the cap removed, the probe is inserted to a point where the
nozzle is just out of the gas stream. The seal can then be slid toward the front of the probe
and wedged between probe and port with moderate hand pressure. To insert the probe into the stream
for sampling, it is necessary to first loosen the seal (pull back along the probe body). The probe
can then be moved into the stack and the seal again wedged into the port.
Stack velocity and temperature measurements are performed as the first step in the sampling
procedure. Reference 3 provides details for calculation of stack velocity from SASS data. Figure 9
shows the approximate range of velocities to be encountered in each of the Wing 6 facilities.
In addition to stack velocity and temperature traverse data, selection of the proper sampling
nozzle is dependent upon the gas composition (molecular weight) and moisture content. The techniques
E-19
-------
ro
o
Stack
gas
n
val ve
f :•'
.r
Lil —
M
] Unirail
j support
r
Port cap
Sample probe
Pert _\
seal
res
4->
00
Figure 7. SASS sample port with gate valve.
-------
3" dia^
ro
V\\ \\Tti
Sample probe
SIiding.
seal
6" dia
L
30
•4
*
Stack gas
Figure 8. Residential furnace SASS sample port.
-------
m
i
i-o
ro
Facility
• Versatile furnace
• Residential furnace
• Firetube boilers
• Diesel engine
, DSU Boiler
• Gas turbine .
100
Nozzle sizes: Q Q Q
1-1/2 1-1/4 1
o o o o o
3/4 5/8 1/2 7/16 5/16
1,000
Velocity, f'pm
10,000
Figure 9. CRB combustion facility stack velocities.
-------
employed are EPA Methods 3 and 4, respectively, described in Reference 4. The equipment for per-
forming each of the EPA sampling methods is incorporated into the SASS, and the calculation pro-
cedures are outlined in Reference 3.
The measurements of moisture content and gas composition can be omitted from an assessment
if the process operating point is being duplicated and composition data is available from past
testing. When gas composition measurement is required, continuous gas analysis for CO, COp, and Op
is the recommended technique over the EPA standard Orsat analyzer.
Isokinetic sampling conditions are achieved by variation of the nozzle inlet area rather
than by changes in the SASS sample flowrate. Figure 9 indicates the nozzle sizes supplied with the
SASS plotted at their approximate isokinetic sampling velocities for the fixed system flowrate.
Therefore, a preliminary nozzle may be selected by use of the figure. Wherever multiple nozzles are
shown for a given facility, higher stack velocities dictate that the smaller of the listed nozzles
will be required. Note that only two of the Wing 6 facilities (the diesel engine and gas turbine)
require sample nozzles of 0.75 inch or smaller diameter. Since nozzles larger than 0.75-inch diam-
eter cannot be mounted on the probe simultaneously with the Ditot tubes, all other facilities will
require separate velocity measurements with use of the pi tot tubes only.
NOTE
Figure 9 is provided only as a guide to nozzle selection. The proper nozzle size for
isokinetic sampling must be selected by the calculations of stack velocity shown in
Reference 3.
The nozzles that are provided for the Wing G combustion facilities have the following diam-
eters: 5/16, 7/16, 1/2, 5/8, 3/4, 1, 1-1/4, and 1-1/2 inches. Two of the facilities, the versatile
and residential furnaces, require nozzles larger than 1-1/2 inches and hence, cannot be sampled
isokinetically with the nozzles provided. Larger nozzles are not compatible with the SASS probe and
these two facilities will always be sampled nonisokinetically using the 1-1/2-inch nozzle.
Following nozzle selection and mounting on the probe, the probe is reinserted through the
stack port and allowed to reach a steady-state temperature that provides the proper sample tempera-
ture (400°F). The probe nozzle should remain within the port until the probe body is fully heated.
When the operator is satisfied that the probe has reached steady state, the probe can be inserted
into the gas stream and sampling begun by steps listed in the SASS operating manual.
Level 1 environmental assessment specifies 30 m3 of sample by the SASS train which will re-
quire a 3- to 4-hour period. Periodic adjustments of the SASS are required during the sample
period, including:
E-23
-------
• Probe coolant temperature and flowrate adjustments to maintain cyclone collection temper-
ature (the objective is to maintain a sample temperature of 400°F at the entrance to the
cyclone oven)
• Possible filter change, should fine particulate cause clogging
• Maintenance of proper sample flowrate
• Monitoring of changes in stack temperature
t Recording of data at periodic intervals
See Reference 3 for more specific details regarding SASS train operation. Following comple-
tion of the sampling, the system is shutdown and extracted from the stack by pulling probe and oven
back to clear the valve, closing the gate valve, and final withdrawal from the port.
CAUTION
Avoid escape of hot gases through the port due to over-extraction of the probe prior
to closing of the gate valve.
3.2 GASEOUS GRAB SAMPLING
Gas grab samples are to be taken periodically as required by the test plan. As with the
SASS sampling system, grab sample port configuration is common to all Wing G facilities except the
residential furnaces. All other facilities are provided with a permanent sample probe tube and
valve.
The more common port with hand valve is constructed with an 0-ring seal at the port outlet.
Silicon grease should be applied to the probe's outer wall to lubricate this seal, taking care that
the grease does not come in contact with the probe inner surfaces. The probe is partially inserted
to seal the port, the hand valve opened, and the probe inserted until its stainless steel collar
butts up against the port opening.
The residential furnace ports contain no valve and simply require removal of the port cap
and insertion of the sample probe. The sample probe is of different construction than that used on
the other facilities. The two port and probe types are shown respectively in Figures 10 and 11.
The glass bottle (previously evacuated and attached to the sample probe) has a sample volume
of 3 liters. The sample is taken by opening of the bottle valve and allowing the stack gases to
be drawn into the sample bottle. Bottle and stack pressure equalize within a few minutes at which
time the bottle valve is placed in the closed position and the probe is removed from the stack.
E-24
-------
[\J
en
Flexible
coupling
0-ring
seal
Figure 10. Gas grab sample port with ball valve.
-------
ro
CTi
To sample
bottle
LZS
\ \ \ X X \ XI
XX XXX X XT
Flexible
coupling
Sample
tube
Xl
^
ilullite
1 iner
6" Furnace stack
Figure 11. Residential furnace gas grab sample port.
-------
CAUTION
The sample probe must remain in the outer housing during opening and closing of the
port valve to avoid escape of hot stack gases.
In order to take additional samples during the test period, the grab sample probe can be de-
tached from the evacuated bottle at the flexible line and a new bottle attached. The above sampling
procedure is then repeated. In order to maintain sample integrity, sample bottles should be taken
to the laboratory for analysis as soon as possible rather than waiting until completion of the
entire day's sampling.
3.3 LIQUID FUEL SAMPLING
The RTP Wing 6 facility area contains a fuel-oil system with taps at specific locations. In
order to sample the liquid fuel, these taps are opened as the fuel is flowing in the line, filling
a glass laboratory bottle. The tap is then closed off, completing the sampling procedure.
A schematic of all sampling procedures is shown in Figure 12.
E-27
-------
SASS Train
Gaseous Grab
Sampling
• Attach probe nozzle and/or pitot tubes
• Calculate stack traverse points
• Warmup probe coolant fluids
— Turn power switch on
— Set temperature controller and allow oil to heat
- Start oil pump
— Allow fluid to flow through cooling loop
— Start water flow circuit
— Verify proper operation
• Insert probe into stack
• Perform velocity and temperature traverse
• Measure appropriate stack variables
— Stack static pressure
— Ambient pressure
— Exhaust gas moisture content
— Gas composition
• Withdraw probe from gas stream
• Calculate average velocity sampling point
• Select proper nozzle for isokinetic sampling
0 Reinsert probe into stack
• Initiate sample flow
• Record data at periodic intervals
• Perform periodic SASS adjustments
— Probe coolant temperature and flowrate
— Filter change if required
— Maintain sample flowrate
• Shutdown sample gas flow at sample completion
t Withdraw probe from stack
t Shutdown probe coolant system
— Turn off temperature controller
- Turn off coolant pumps
— Turn power switch off
• Attach probe tubing to evacuated flask
• Insert probe into stack by appropriate procedure
• Open bottle valve to initiate sample flow
0 Allow stack and bottle pressures to equalize
0 Close bottle valve '
0 Withdraw probe from port and insure the port seal
0 Detach probe tubing from sample bottle
0 Label sample bottle for time taken
0 Transport contained sample to laboratory as soon as possible to avoid
changes in sample composition
Figure 12. Sampling procedures.
E-28
-------
Liquid Fuel
Sampling
• Verify flow through the fuel system
• Open tap at sample location
• Fill glass laboratory bottle
• Close sample tap
• Cap sample bottle
• Label bottle for time taken
Continuous Gas
Analysis
t Provide sample flow to analyzers
• Allow analyzers to stablize
t Record species concentrations at periodic intervals
Figure 12. Concluded
E-29
-------
SECTION 4
SAMPLE RECOVERY
Following completion of sampling operations, the sampling crew must transfer samples to lab-
oratory containers for subsequent analysis. The procedures include:
t Disassembly of sampling hardware
• Transport of equipment to the analysis laboratory
• Extraction of samples from the SASS train
• Equipment cleanup and reassembly
The most time-consuming sample recovery involves the SASS samples. The pertinent procedures
are discussed below.
4.1 SASS TRAIN BREAKDOWN
Following cooldown of probe and oven, the SASS is partially disassembled at the sample loca-
tion. Cap off the nozzle at the probe inlet and disconnect power, water, and probe coolant supplies
such that the probe and oven can be removed from the unirail support. Disconnect the sample hoses
from impingers and gas cooler and separate each component in the system. The probe, cyclones and
filter holder, impingers and case, gas cooler, and liquid fuel samples are then ready for return to
the laboratory. The SASS control module and pumps can be left in place. The mechanics of SASS
breakdown are more fully explained in the SASS operator's manual. The cart can be used to facili-
tate moving of the SASS components to the laboratory.
4.2 SAMPLE EXTRACTION
Sample material held within the SASS must be properly extracted and put into laboratory con-
tainers. The complex procedures, designed to avoid sample contamination, are outlined in Reference
2, Figures 12 to 14. All containers should be prepared by laboratory personnel. The particulate
catch includes not only that of the cyclone cups and filter, but also particles deposited within the
probe liner, nozzle, and cyclone interconnecting tubes. All particulate is to be washed from con-
tacted surfaces and placed in individual, weighed laboratory containers.
E-30
-------
The condensed water from the gas cooler is collected and bottled for future measurement and
analysis. The gas cooler unit also contains the organic adsorbent module that is emptied into a
laboratory container. Finally the four impinger bottles contain solutions that must be washed into
laboratory containers by the procedures of Reference 2.
Gas and liquid grab samples are collected directly into their storage containers. No addi-
tional preparation is required. All sample bottles should be labeled as to test numbers, sample
time, and contents as they are taken or prepared. The sample crew then stores all samples in the
area provided in the analysis laboratory.
4.3 SYSTEM CLEANUP AND REASSEMBLY
All SASS train components contacted by the sample gas require passivation by acid treatment
and cleaning to inert the surfaces prior to the next system use. The grab sample probe should be
similarly treated. If the sampling systems are to be used soon after the just-completed test, this
function may be performed before reassembly of the SASS. The final step in sample preparation is
reassembly of the cyclones/filter/oven and impingers and storage until the next Level 1 assessment.
Sample recovery procedures are outlined in Figure 13.
E-31
-------
Sampling System
Disassembly
Sample Extraction
Sample System
Storage
Partially disassemble SASS train at sample site. Isolate
- Probe
- Cyclone oven
- Gas cooler module
— Impinger bottles
Transport above SASS components to laboratory
Transport remaining gas and fuel sample bottles to
laboratory
Disassemble SASS components as required
Transfer particulate samples to individual containers
— Probe and nozzle catch
- Cyclones
— Filter and holder
- Tube connections
Collect gas cooler condensate in provided container
Transfer polymeric adsorbent material to provided
container
Wash reagents from impinger bottles to laboratory
containers
Label all containers for content, sample time, and test
number
Store sample containers in laboratory
— Gas grab bottles
- Liquid fuel bottles
- SASS collection containers
Passivate and clean SASS and grab sample equipment
Reassemble SASS components
- Cyclones, filter holder, and oven
— Gas cooler module
- Impinger bottles
Store above components in analysis laboratory
Figure 13. Sample recovery.
E-32
-------
SECTION 5
DATA REDUCTION
The responsibility for data reduction for the environmental assessment lies jointly with
sampling and analysis personnel. The sampling crew provides copies of all data sheets to the labo-
ratory, complete with final calculations of average stack velocity, moisture content, gas composi-
tion, particulate sample volume, and grab sample conditions. Following sample analysis, the labo-
ratory should provide concentration data for all sample components pertinent to the Level 1
requirements (Reference 2). The final step, converting sample data into process mass emission rates,
can be performed by either sampling or analysis personnel.
The calculated emission rates are the main objective of the environmental assessment. EPA
methods (Reference 4) are a good source of calculation details for total particulate weight and
gaseous species concentrations. These calculations as well as other specie and elemental mass data
are then easily converted to emission rates with the stack and sample system gas flowrates. All
emission rate data are finally reported to program, facility, sampling, and analysis personnel.
E-33
-------
REFERENCES
1. "IERL-CRB Analysis Manual for Level 1 Environmental Assessment," A. D. Little, Inc.,
February 1977.
2. Hamersma, J. W., Reynolds, S. L., and Maddalone, R. F., "IERL-RTP Procedures Manual: Level 1
Environmental Assessment," EPA-600/2-76-160a, June 1976.
3. "Operating and Service Manual - Aerotherm Source Assessment Sampling System," Acurex Corporation,
Aerotherm Division, April 1977.
4. U.S. Environmental Protection Agency, "Standards of Performance for New Stationary Sources,"
Federal Register, Vol. 36, No. 247, Part II, December 1971.
E-34
-------
APPENDIX A
SAMPLE DATA SHEETS
E-35
-------
Facility
Stack Diameter (in.)
Date
Time
Run
Operators
Static Pressure (in. WG)
Ambient Pressure (in. Hg)
Type S Pi tot Coefficient
Stack Pressure (in. Hg)
Molecular Weight (Ib/lb-mole)
B
wo
VELOCITY TRAVERSE DATA
Sampling
Point
Number
Velocity
Head
(in. WG)
ShP
Average
Stack Gas
Temp
(°F)
Average
Sampling Point
Velocity
(fps)
E-36
-------
Facility
Stack Diameter (in.)
Date
Time
Run
Operators
Static Pressure (in. WG)
Ambient Pressure (in. Hg)
Type S Pi tot Coefficient
Wet Bulb Temp (°F)
Dry Bulb Temp (°F)
Moisture (% vol.)
wo
STACK GAS MOISTURE (CONDENSATION)
Final
Initial
Liquid Collected
Condensate Collection
Bottle Weight
(gm)
Total, V^c
Clock
Time
Dry Gas
Meter
(ft3)
Total, Vm
Gas Flowrate
(cfh)
Average
Gas Meter
Temp
(°F)
Average, Tm
Velocity
Head
(in. WG)
Stack Gas
Temp
(°F)
Gas Cooler
Temp
(°F)
E-37
-------
Facility
Stack Diameter (in.
Date
Time
Run
Operators
Static Pressure (in. WG)
Ambient Pressure (in. Hg)
Type S Pi tot Coefficient
STACK GAS MOLECULAR WEIGHT
Clock
Time
*• •
Gas Flowrate
(cfh)
Velocity
Head
(in. WG)
Stack Gas
Temp
(°F)
co2
(% by Volume)
Average
o2
(% by Volume)
Average
CO
(% by Volume)
Average
E-38
-------
Facility
Stack Diameter (In.)
Date
Tine
Run
Operators
Static Pressure (in. WG)
Ambient Pressure (in. Hg)
Type S Pitot Coefficient
Stack Pressure (in. Hg)
Molecular Weight (Ib/lb-mole)
3wo
Meter Box Number
Orifice Meter Coefficient
Probe Length (ft)
Nozzle Diameter (in.)
PARTICULATE SAMPLING DATA
Sampling
Point
Number
^ ^
rinrt
Time
Stack Gas
Temp
Average
Probe
Temp
Gas
Temp
Impinger
Outlet Temp
Oven
Temp
Dry Gas t
Inlet
Average
Average
leter Temp
Outlet
Average
Velocity
Head
(in. WG)
Orifice
Meter
(in. WG)
Average
Gas Meter
Volume
(ft3)
Total
Pump
Vacuum
(in. Hg)
E-39
-------
Facility
Stack Diameter (in.)
Date
Time
Run
Operators
Stack Pressure
Ambient Pressure
GAS GRAB AND LIQUID FUEL SAMPLES
Clock
Time
Stack Gas
Temp
Gas Grab
Sample No.
Liquid Fuel
Sample No.
E-40
-------
Facility
Date
Time
Run
Operators
Stack Pressure
Ambient Pressure
Analyzer Calibrations
CO: ppm
C02: %
NO: ppm
NOV : ppm
A
CONTINUOUS GAS ANALYSIS
Clock
Time
Stack
Temp
(°F)
Analyzer
Gas Temp
(°F)
Gas Concentrations
°2
(%)
CO
(ppm)
co2
(*)
NO
(ppm)
N0x
(ppm)
E-41
-------
APPENDIX B
LIST OF SAMPLING EQUIPMENT
E-42
-------
I Source Assessment Sampling System*
1. Unirail support
2. Control module
3. Sample gas pumps
4. 25-foot sample hose (1/2 inch)
5. 4-foot pump-to-control module hoses
6. 25-foot umbilical
7. 3-cyclone assembly
8. 6-inch filter assembly
9. Cyclone oven
10. Organic module/gas cooler unit
11. 42-inch oven-to-gas cooler sample line
12. Impinger assembly and case
13. Water circulation pump
14. Sample probe
15. Probe temperature-control unit
16. Probe hydraulic oil pump assembly
17. Metal braided hydraulic hoses (3)
18. Garden hoses (2)
19. Water circulation pump
20. 8 probe nozzles, sizes 5/16 inch to 1-1/2 inches
21. S-type pi tot tubes
22. 142 mm glass fiber filters
23. Organic adsorbent material
24. Hydraulic fluid - Therminol "66" (Monsanto)
25. Crushed ice supply (impingers)
26. Sample containers
27. Vacuum grease
II Gas Grab Sampler
1. Sample probe and connecting tubing
2. Evacuated, encased sample bottles
3. Glass wool plugs
4. Silicon grease
III Liquid Fuel Sampling
1. Glass sample containers
*See Reference 2 for component descriptions and part numbers.
E-43
-------
Appendix F
Aerotherm Project 7237
A PROGRESS REVIEW OF
SOURCE ASSESSMENT SAMPLING
SYSTEM PROGRAMS
David Blake
Acurex Corporation/Aerotherm Division
485 Clyde Avenue
Mountain View, California 94042
June 15, 1977
AEROTHERM REPORT TM-77-176
Prepared for
Environmental Protection Agency
Research Triangle Park
North Carolina 27711
Contract 68-02-2153
F-i
-------
TABLE OF CONTENTS
Section Page
1 INTRODUCTION F-l
2 SASS TEST OBSERVATIONS F-l
3 TRAINING SESSIONS F-2
4 CYCLONE CALIBRATION F-3
5 COORDINATION WITH OTHER TLE CONTRACTORS F-6
6 SASS SUPPORT OPERATIONS F-ll
LIST OF ILLUSTRATIONS
Figure Page
1 Cyclone test apparatus ............... F-4
2a Si02 test dust ................... F-7
2b SiOp dust - small cyclone cup catch ........ F-7
2c Si02 test dust - small cyclone filter catch .... F-8
3a Aluminum test dust ................. F-9
3b Aluminum dust - small cyclone cup catch ...... F-9
3c Aluminum dust — large cyclone cup catch ...... F-10
F-ii
-------
1. INTRODUCTION
During the past year the TIE contract has supported several tasks
pertaining to the development, use, and improvement of the Source Assessment
Sampling System (SASS). The specific tasks involved were Task 07 (Cyclone
Calibration), Task 08 (SASS Support), and several Support Service Tasks of
less than 100 hours duration. Activities related to the SASS have been
conducted in five general areas:
• SASS test observations
t SASS training sessions
• Cyclone calibration
• SASS support operations
• Coordination with other TIE contractors.
Each of these activities will be discussed individually.
2. SASS TEST OBSERVATIONS
Two observations* of SASS tests were made by Aerotherm personnel in
the past year. In each case, the purpose of the observation was two-fold:
to advise the test crew on use of the train and to observe the train under
actual field conditions so that problems could be noted and corrected.
The first observation was of a SASS test of an experimental coal
fired boiler at KVB, Inc. The observation was very valuable in helping to
identify several problems associated with the use of the SASS train as it
was then configured. Changes subsequently made as a result of this obser-
vation included:
• Elimination of Swageloc connectors in the cyclone assembly, as
they are easily damaged and are inoperative when hot. Modified
pipe union fittings now provide good performance
• Provision for a Marmon clamp to secure the two parts of the
large cyclone, eliminating the screw fasteners
*The Plant Mitchell observation occurred in May 1977, which is after the
close of the reporting period (April 1976 -April 1977). It will be
covered in the next annual report.
F-l
-------
• Modification of the shape of the middle cyclone to make after-
test cleaning easier
t Increasing the size of the small cyclone cup
• Modification of the impinger bottles to reduce the possibility
of breakage
The other observation was of a SASS test of the Comanche Power Plant
of the Colorado Public Service Comapny, performed by Radian Corporation.
This test was the first one conducted at the increased flowrate of 4.0 scfm.
The results of this test indicated that a single vacuum pump was inadequate
for maintaining that flowrate. The SASS train's design was therefore modi-
fied to incorporate two pumps.
3. TRAINING SESSIONS
Twice during the past year training sessions were conducted for SASS
users by Aerotherm personnel. The first session was held at Battelle Columbus
Laboratories. The purpose was to train Battelle sampling crew members in the
setup, checkout, operation, and cleanup of the SASS train for Battelle's up-
coming test at the Exxon pressurized FBC. A secondary objective was to ex-
plain the design and use of the SASS to interested members of the Battelle
technical staff. Approximately eight field crew members were given 2 full
days of hands-on practice in the use of the trains. During the 2 days, the
SASS train was disassembled, reassembled, and leak-checked three times.
About 15 members of the Battelle technical staff attended a presentation on
the SASS design.
The other training session was conducted at EPA/RTP. The goals were
the same as at Battelle - to provide general information on the SASS design
to interested EPA staff members, and to allow EPA technicians to practice
assembling the SASS train under the direction of an Aerotherm engineer fa-
miliar with its use. A representative from Southern Research Institute was
also present. A 1-day session was held with about 15 people attending the
discussion of SASS design, and 5 people completing the practice sessions in
assembly, operation, and leak checking. A satisfactory leak check was
routinely achieved using inexperienced technicians under Aerotherm direction.
F-2
-------
4. CYCLONE CALIBRATION
«
The cyclone calibration task is intended to experimentally evaluate
the performance of the three SASS cyclones using "real" solid particles.
At the time the calibration task was begun, other calibration tests were
underway at Southern Research Institute. The SoRI method involves the use
of a vibrating orifice particle generator which produced precisely sized
particles of a dye solution. When dried by the carrying gas stream, a
monodisperse aerosol is produced. Collection efficiency versus particle
size curves that characterize the cyclone performance can be obtained by
feeding several different sizes of monodisperse dye particles into the
cyclone to be tested.
Several problems that existed with the SoRI method led to the de-
cision to try a completely different approach to cyclone calibration.
These problems were: unavailability of a dye material useable at 205°C,
and the extremely low particle mass concentrations used in the SoRI method.
It was felt that these low concentrations (several orders of magnitude lower
than field concentrations) might not adequately simulate wall-sticking and
particle-particle interaction effects.
The following describes the calibration method used by Aerotherm:
A dust cloud was generated from a test dust of known properties, then the
dust was fed into the cyclone and the cyclone efficiency measured from the
quantity and size distribution of the dust collected (in the cup) and passed
by the cyclone (in the filter). Figure 1 shows the experimental setup.
In the first series of experiments, a highly classified silica dust
material was used. The dust was classified by the Donaldson Company into
nine size fractions, starting with AC Fine test dust. This particular mate-
rial was chosen because of its availability, its wide use as a test dust in
the HVAC industry, and the familiarity of Donaldson Company with the behavior
of the material in their classifiers.
The approximate performance of each of the SASS cyclones was known
from previous SoRI tests. Accordingly, three to six of the classified dusts
were chosen for calibrating each cyclone, depending on the mean particle
size of the dust. The size distribution of the test dust and the dust col-
lected in the cyclone cup was measured using the x-ray sedograph at EPA/RTP.
F-3
-------
^THERMOCOUPLE
TEE
FILTER
CYCLONE
FILTER
Figure 1. Cyclone test apparatus.
-------
From the size distribution data, it should be possible to construct a
cyclone efficiency versus particle size curve for the particle size range
of the test dust. When this was attempted, it became apparent that the
experimental results were inconsistent and in some cases contradictory.
For several experiments, for example, the mass median size of the cyclone
cup catch was smaller than the feed material; the filter catch mass median
diameter was even smaller. This result is clearly impossible unless the
size distribution measurement method is faulty or unless the test dust is
changing its characteristics during the test.
There is some evidence that the latter explanation is the cause of
the unexpected test results.v Figures 2a, 2b, and 2c are scanning electron
micrographs of the feed, cyclone cup, and filter fractions, respectively,
from a calibration run with the small cyclone. The magnification is 3000X.
It is qualitatively apparent that the cyclone cup fraction is smaller than
the feed fraction, as indicated by the x-ray sedograph measurements. The
most interesting point, however, is the appearance of the particles. The
test dust particles (Figure 2a) are generally smooth and evidence cleavage
planes. The particles collected by the cyclone (Figure 2b), however, are
very rough and pitted, and seem to be rounded off. The filter fraction
largely consists of very small particles that are not evident in the test
dust. All of this seems to indicate that the test dust has been eroded and
reduced in average size somewhere in the calibration apparatus. As veloci-
ties in the dust cloud outlet tube and heater are deliberately kept high
(near sonic) to avoid reagglomeration of the dust, it is suspected that
particle-particle contact in this region is causing the erosion. The hard-
ness and frangibility of the test dust undoubtedly is also a major factor.
In order to overcome the problem of test dust degradation, either
the apparatus or the test dust will have to be changed. The easiest solu-
tion appears to be to obtain a test dust that is not subject to erosion.
After a search for such a dust, a spherical aluminum powder material was
discovered. Being malleable, no particle erosion is anticipated. Also,
the fact that the particles are spherical should make measurement of the
size distribution and subsequent data analysis easier. One possible prob-
lem is the low melting point of aluminum and associated possibility of
F-5
-------
sticking to the hot wall in the heater sections of the test apparatus.
If necessary the heater can be lengthened to reduce wall temperature
below the melting point.
Some calibration tests have been completed using the aluminum
test dust. Size distribution analysis has not been completed; however,
photomicrographs indicate reasonable results. Figures 3a, 3b and 3c show
the test dust, the small cyclone cup catch, and the large cyclone cup catch.
The test dust is primarily spherical with a relatively narrow distribution
of particulate sizes. Since the small cyclone is collecting almost all of
the test dust, the small cyclone cup catches (Figure 3b) look very much
like the test dust (Figure 3a). The large cyclone cup catch, however,
clearly shows preferential collection of the larger particles, as it should.
Completion of the size distribution analysis will allow calculation of
cyclone efficiency curves.
5. COORDINATION WITH OTHER TIE CONTRACTORS
Two trips to EPA/RTP were made during the past year to attend the
TLE contractor's meetings. The purpose of these meetings was a general
review of programs and progress. In addition, numerous contacts with the
other contractors occurred on a less formal basis. Some of these con-
tracts were with
• Southern Research Institute - numerous discussions regarding-
cyclone calibration tasks at SoRI and Aerotherm. Also, dis-
cussions and planning for the SoRI solvent-refined coal SASS
test
• Arthur D. Little, Inc. - many discussions about the performance
of various organic sorbents and the applicability of Viton seals
to the SASS trains
• Research Triangle Institute - the main contact was a meeting at
Aerotherm where the field performance of the SASS train was
discussed
• TRW, Inc. — numerous conversations about use of the SASS train
on the incinerator ship Vulcanus, performance of the KVB boiler
test, and corrosion problems in the SASS trains.
F-6
-------
Si02 test dust
Figure 2(a).
SiO? dust - small cyclone cup catch.
Figure 2(b).
F-7
-------
SiCL test dust - small cyclone filter catch
Figure 2(c).
F-8
-------
Aluminum test dust
Figure 3(a).
Aluminum dust - small cyclone cup catch
Figure 3(b).
F-9
-------
Aluminum dust - large cyclone cup catch
Figure 3(c).
F-1D
-------
SASS SUPPORT OPERATIONS
Task 08 of the TLE contract is intended to support a number of
ctivities concerning development of the SASS train. The major activities
re listed here
t An operating manual for the SASS train was prepared. The
manual provides complete, detailed instructions for setting
up, checking-out, running, and trouble shooting the train.
• An oversize filter holder was designed and constructed to
reduce the number of filter changes during a SASS test. The
oversized filter has 2.3 times the area of a standard filter.
For most sources, one filter should last an entire 5-hour run even
when testing ahead of the control device (1-3 gr/scf).
t A spare parts inventory has been provided for government
owned SASS trains.
• A list of desirable SASS train modifications has been assembled,
in response to comments by field crews. Some 22 system
modifications are now under consideration.
• The two partial SASS trains used for sampling the incinerator
ship, Vulcanus, have been upgraded to full SASS status.
• Six complete sets of SASS train blueprints have been submitted
to the Project Officer.
• A comprehensive final report, detailing the SASS development
project, is now underway.
• A corrosion/welding consultant has been retained to examine
and report on ways of minimizing corrosion in the SASS train by
proper choice of materials and construction techniques.
F-ll
-------
APPENDIX G
CONVERSION TO INTERNATIONAL SYSTEM OF UNITS
To convert
acfm
atmospheres
°F
°R
ft3
ft/sec, fps
gpm
grain
grain/scf
inches, in
Ibm
microns, urn, y
pound
psia
to
meter3/minute, m3/min
newton/meter2, newt/m2
Celsius, °C
Kelvin, °K
meter3, m3
meter/second, m/s
meter3/minute, m3/min
kilogram, kg
kilograms/meter3, kg/m3
meter, m
kilogram, kg
meter, m
newton, newt.
newton/meter2, newt/m2
multiply by
2.8317 x 10-2
1.0133 x 105
°C = 5/9(°F - 32)
°K = 5/9 °R
2.8317 x 10-2
3.048 x 10"1
4.5461 x ID'1
6.4780 x 10~5
2.2877 x 10-3
2.54 x TO'2
4.5359 x 10-1
1.00 x 10-6
4.4482
6.8948 x 103
G-l
-------
TECHNICAL REPORT DATA
(Please rcail Instructions on the reverse before completing}
. REPORT NO.
EPA-600/7-78-011
3. RECIPIENT'S ACCESSION-NO.
I. TITLE AND SUBTITLE
Measurement of High-temperature, High-pressure
Processes: Annual Report
5. REPORT DATE
January 1978
6. PERFORMING ORGANIZATION CODE
'. AUTHOR(S)
Larry Cooper
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Aerotherm Division/Acurex Corporation
485 Clyde Avenue
Mountain View, California 94042
10. PROGRAM ELEMENT NO.
EHE623 and 624
11. CONTRACT/GRANT NO.
68-02-2153
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Annual; 4/76-7/77
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES T£RL-RTP project officer is William B. Kuykendal, Mail Drop 62,
919/541-2557.
16. ABSTRACT Tne repOrj- reviews the first year's efforts under a planned 3-year program
to develop measurement techniques for high-temperature, high-pressure (HTP) pro-
cesses. Several related topics are discussed in detail, including: (1) the design,
development, and successful demonstration of a system for sampling particulate from
a pressurized fluidized-bed combustor operating at 740 C and 9 atmospheres; (2) a
review of existing measurement methods to determine the best available techniques
for measuring gas flow velocities, pressures, and temperatures in HTP process
streams; (3) a review of various HTP coal conversion processes and recommendations
for material selection for sampling them; and (4) recommendations for sampling tars
in a coal gasification process.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Pollution Fluidized-bed
Measurement Processing
Sampling Gas Flow
Industrial Processes Velocity
High Temperature Tests Pressure
High Pressure Tests Temperature
Coal Gasification pn«t Tars
b.lDENTIFIERS/OPEN ENDED TERMS
Pollution Control
Stationary Sources
Particulate
c. COSATI Held/Group
13B
14B
13H
07A
20D
18. DISTRIBUTION STATEMEN1
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
!1. NO. OF PAGES
254
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
G-2
------- |