United States Industrial Environmental Research EPA-600/7-79-196
Environmental Protection Laboratory August 1979
Agency Research Triangle Park NC 27711
Assessment of
Instrumentation for
Monitoring Coal Flowrate
and Composition
Interagency
Energy/Environment
R&D 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 endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-196
August 1979
Assessment of Instrumentation for
Monitoring Coal Flowrate and Composition
by
E. F. Brooks and C. W. Clendening
TRW Systems and Energy
One Space Park
Redondo Beach, California 90278
Contract No. 68-02-2165 (Task 9) and 68-02-2613 (Task 2)
Program Element No. INE624
EPA Project Officer: Frank E. Briden
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
11
-------
ABSTRACT
The report gives results of an assessment of instrumentation for
the measurement of coal flowrate (either as a dry solid or in a coal/water
slurry) and composition. Also investigated was the appropriateness of
EPA/IERL-RTP involvement in the development or evaluation of such devices.
Findings for flow measurement hardware were that dry coal flow can be
easily and accurately measured using weigh belt devices, and that the mass
flow of coal in a coal/water slurry stream can be measured using a flow-
meter (electromagnetic flowmeters are preferred) and a nuclear density
gage. The most promising analysis concept under development is fast neutron
activation, with delivery of a sulfur and ash meter anticipated by the end
of 1979. Other techniques, such as X-ray fluorescence, work on only a very
small coal sample. It is recommended that further EPA investigation deal
with system, rather than component, evaluations.
This report was submitted under Task 2 of Contract No. 68-02-2613,
by TRW Systems and Energy, under the sponsorship of the Industrial Environ-
mental Research Laboratory - RTP, of the Environmental Protection Agency.
Work was completed and a draft report submitted in February, 1978. A re-
port was issued under TRW cover in August, 1978. Only very minor updating
has occurred between February, 1978, and publication of the present report.
iii
-------
TABLE OF CONTENTS
Page
Abstract i i i
List of Figures v
List of Tables vii
Acknowledgements ix
Sections
I Introduction 1
II Principles of Instrument Operation 3
2.1 Coal Analysis Instruments 5
2.2 Flow Measurement Instruments 35
III Hardware Status 66
3.1 Coal Analysis Instruments 66
3.2 Flow Measurement Instruments 76
IV Ten Year Projection of Instrument Development
and Availability 104
V Summary and Conclusions 107
VI References 109
VII Glossary 113
VIII Appendix - Procedure for Weigh Belt Calibration
and Use 117
iv
-------
FIGURES
No. Page
1 Compton Effect Geometry 15
2 Mass Attenuation Coefficient for Photons in Air 16
3 Thermal Neutron Capture Gamma Ray Analysis Experimental
Configuration 20
4 Gamma Ray Spectrum of Coal Measured with Cf 252 Source
and 6 by 7 Inch Nal Detector 21
5 Inhomogeneity Effect Upon Gamma Ray Transmission 28
6 Principle of Neutron Moisture Meter 30
7 Term Scheme, Showing the Origin of the X-Ray Spectrum
of the Elements 31
8 X-Ray Fluorescence Configuration 33
9a Weigh Belt Installation for Long Belt 37
9b Weigh Belt Installation for Short Belt 37
10 Typical Weighbridge Unit 38
11 Schematic of Loss-In-Weight System 43
12 Illustration of "Solids Turbine" Concept 44
13 Schematic of Screw Feeder 46
14 Schematic of Electromagnetic Flowmeter, Taken from
Reference 25 48
15 Illustration of Magnetic Flowmeter Error Due to Non-Full
Pipe 50
16 Ultrasonic Flowmeter Schematic 52
17 Schematic of NMR Flowmeter, Taken from Reference 33 55
18 Schematic of Vortex Shedding Meter 58
19 Schematic of Target Meter 60
20 Schematic of Venturi Flowmeter 61
-------
FIGURES (Cont.)
No. Page
21 Accuracy of Coal Mass Flow Measurement 1n a Coal/Water
Slurry for Various Accuracies of Slurry Density, Coal
Density, and Volumetric Flow Rate as a Function of
Coal Concentration in the Slurry 64
22 Conceptual Design of Continuous Nuclear Analyzer of Coal
(CONAC) System Including a Nuclear Scale (Ref. 39) 68
23 Detail of Overrange Mechanism for Target Meter 90
24 Ultrasonic Density Gage 95
25 Slipstream Concept for Ultrasonic Density Gage 97
26 Typical Flowmeter/Density Gage Combination, Taken from
Texas Nuclear Catalogue 98
vi
-------
TABLES
No. Page
1 ASTM Designation D388 - Classification of Coals by Rank 7
2 Typical Analysis of Coal of the United States 8
3 Typical Ash Contents 9
4 Classification of Neutrons by Energy Level 17
5 Neutron Scattering Properties of Nuclei 18
6 Analytic Sensitivities for Thermal Neutron Capture 23
7 Z/A Values 25
8 Coal Elements Analyzed By CONAC System 69
9 Density Gage Data 71
10 Moisture Meter Data 72
11 LECO IR-33 Sulfur Analyzer 73
12 Perkin-Elmer Model 240 Elemental Analyzer 73
13 X-Ray Analyzer Characterization 75
14 Portable Elemental Analyzer for Ash in Coal 75
15 Partial List of Weigh Belt Vendors (from Ref. 48) 77
16 Typical Weigh Belt Specifications 78
17 Typical Weigh Belt Hardware Costs 78
18 Magnetic Flowmeter Vendors 81
19 Typical Magnetic Flowmeter Specifications 83
20 Common Magnetic Flowmeter Liner and Electrode Materials 83
21 Typical Magnetic Flowmeter Dimensions and Costs 84
22 Ultrasonic Flowmeter Data 88
23 Target Meter Data 91
vn
-------
TABLES (Cent.)
No. Page
24 Vortex Meter Data 93
25 Venturi Meter Data 94
26 Vendor Addresses 99
-------
ACKNOWLEDGEMENTS
The authors wish to thank the numerous vendors who supplied catalogue
information, and assistance over the telephone. In addition, we sincerely
appreciate the help given by Mr. Martin Gruber, Jr. of the Southern Weighing
and Inspection Bureau, Dr. Robert Stewart of the Morgantown Energy Research
Center, who spent a full day showing us the MERC facility and explaining
instrument operation, and Mr. James Rudd, who gave us an instructive tour
of the Homer City, Pennsylvania, power plant.
ix
-------
-------
I. INTRODUCTION
The purpose of this report is to present current technology for
measurement of the following parameters:
Flowrate of coal as a dry solid
Flowrate of coal in a coal/water slurry
Coal composition
Implicit in some techniques for slurry flow measurement is a supplementary
measurement of slurry density. Composition measurements may apply to a
single constituent, such as sulfur, or to a more complete elemental analysis.
The reason for this technology assessment is to keep the Environmental
Protection Agency abreast of instrumentation developments in the above areas.
There are many potential applications for environmental quality improvement,
especially in the areas of process control. At the source of coal combustion,
continuous real time or quasi-real time (instrument response times of
1-5 minutes) coal flow rate and composition data could be used effectively
to optimize both the combustion process itself and the performance of
pollution control devices. This same knowledge would also be useful at coal
cleaning facilities, and directly in the mine for such activities as prelim-
inary sorting of high and low sulfur coals. The possible growth in the
near future of long distance coal conveyance via slurry pipelines also
creates an interest in slurry instrumentation.
The basic finding during the investigation has been that adequate
standard instrumentation is available for the flow measurements, but
continuous analysis equipment is still largely in the development stage,
with the first production unit scheduled for delivery at the end of 1979.
The presently preferred technique for dry coal flow measurement is the
weigh belt, using either a mechanical or electronic load sensor. Due to
their simplicity, nuclear weight scales using a gamma ray technique are
also becoming popular for conveyor belt measurements. Magnetic flowmeters
-------
are most common for slurry flow measurement, and are acceptable when
properly used. They must be used in conjunction with a density meter to
determine the percentage of coal in the slurry. Several new ultrasonic
flowmeters also appear capable of good accuracy in dense slurry flow
measurement. Fast neutron activation techniques are the most promising
for coal analysis instrumentation, but, as mentioned above, are still
largely in the development stage.
The report gives the technical background (physics of operation)
of the methods considered applicable. This is followed by discussion of
the development status of each of the techniques. Since not all tech-
niques have reached the production stage, estimates of instrument
development and availability over the next ten years are given, along
with potential instrument uses. This last section, as originally written
in early 1978, was updated in mid-1979, primarily to note progress by
SAI on nuclear coal analysis instrumentation. As an appendix, procedures
for weigh belt calibration and use formulated by the Southern Weighing and
Inspection Bureau are presented.
There has long been a strong need for flow measurement instruments
in the coal industry for accounting/sales purposes, and this resulted 1n
the development of high accuracy weigh belt techniques. Magnetic flow-
meter development arose from a more general need for a non-intrusive flow
measurement technique. System accuracy for the slurry flow measurement is
not, and probably never will be, as good as that for conveyor belt measure-
ments. Finally, instrumentation for continuous coal analysis is needed to
satisfy more recent demands for improved environmental quality and efficient
energy usage, and lags in development compared to the flow instruments,
primarily because the analysis measurements are inherently more difficult
to make. The combined flow and analysis measurements will be able to be
used to determine elemental mass flow rates, which are required for opti-
mum process control.
-------
II. PRINCIPLES OF INSTRUMENT OPERATION
This report is focused on flow and composition measurements of coal
streams, either dry or entrained in water. For the types of instruments under
consideration, flowrates are measured as volumetric flow or mass flow, and
composition measurements are performed in terms of mass fractions. Flowrate
data is primarily of interest for process control and also directly for
pollution monitoring (e.g., sulfur content). The combination of flowrate and
composition can yield constituent mass flow rates (e.g., carbon, ash, sulfur)
which may be of greater use for accurate process control than either the total
flow or the composition measurements. A brief mathematical base, from a system
viewpoint, is presented below to provide a context for the variety of measure-
ment concepts which follow. The following single equations describe the
measurements analyzed in detail in this report:
1. Dry Solid Mass Flow:
I
(1)
where
m-r = total mass flowrate, kg/s
m = mass present in sensing section at time of measurement, kg
a = length of sensing section, m
9. = speed of conveyor belt, m/s
2. Slurry Mass Flow
where
u
psl
A
m-,
Jsl
u A
(2)
= average slurry velocity, m/s
3
= local slurry density, kg/m
f
= local pipe cross-sectional area, m'
3. Constituent Mass Fraction
Vm
(3)
-------
where
V
m
mass fraction of constituent i, dimensionless
sensed mass of constituent i within the effective
sampling volume during the total sampling time, kg
effective sampling volume (must be established by
o
in-place calibration), m
density of material in sampling volume, kg/m
4. Constituent Mass Flow
IT!. = Ply X-j
(4)
where
m. = mass flow rate of constituent i, kg/s
5. Slurry Density
V
(5)
where
o
VT = locally defined control volume, m
3
V = volume of solids in control volume, m
s o
p = average density of solids in control volume, kq/m
3
V = volume of liquid in control volume, m
3
p,, = average density of liquid in control volume, kg/m
These equations are set up to reflect the way in which the measurements are
actually performed. Equation (4) shows the simple but key point that both
mass flow and composition measurements are required to obtain constituent mass
flow rates. Equation (5) is used by itself, with p , as the actual measured
parameter, to determine percent solid and percent liquid in slurry streams,
and then in conjunction with equation (2) to determine solid and liquid mass
flow rates.
The detailed presentations in the remainder of this section deal with the
physics of the instruments used to measure the parameters in Equations (l)-(g\
Since the coal analysis techniques are not yet fully developed, more detail -js
presented for them in order to give an idea of their ultimate capabilities and
limitations.
-------
2.1 COAL ANALYSIS INSTRUMENTS (Ref. 1-3)
Any analysis of instrumentation designed to measure coal must take
account of its extremely inhomogeneous nature with regard to composition
and physical form. For this reason before embarking upon a discussion
of the instrumental technologies, the range in magnitude of applicable
quantities and their definitions are summarized in this introductory
section.
In most cases the definitions employed will be those of the ASTM.
The terms of greatest interest are defined as follows in ASTM Designation
D121-72:
Standard Definitions of Terms Relating to Coal and Coke -
As Specified in ASTM Designation D121-72
" proximate analysis - in the case of coal and coke, the determination,
by prescribed methods, of moisture, volatile matter, fixed carbon (by dif-
ference), and ash.
ultimate analysis in the case of coal and coke, the determination
of carbon and hydrogen in the material, as found in the gaseous products
of its complete combustion, the determination of sulfur, nitrogen, and ash
in the material as a whole, and the estimation of oxygen by difference.
NOTE 1 -The determination of phosphorus is not by definition
a part of the ultimate analysis of coal or coke, but
may be specified when desired.
NOTE 2 - When the analysis is made on an undried sample, part
of the hydrogen and oxygen as determined is present
in the free moisture accompanying the coal. Therefore,
in comparing coals on the basis of their ultimate
analysis, it is advisable always to state the analysis
on both the "as-received" and "dry" bases.
NOTE 3 Inasmuch as some coals contain mineral carbonates, and
practically all contain clay or shale containing com-
bined water, a part of the carbon, hydrogen, and oxygen
found in the products of combustion may arise from these
mineral compounds.
moisture -essentially water, quantitatively determined by definite
prescribed methods which may vary according to the nature of the material.
NOTE 1 -Such methods may not determine all of the water present.
NOTE 2 - In the case of coal and coke the methods employed shall be
those prescribed in ASTM Methods D 271, Laboratory Sampling
and Analysis of Coal and Coke.
5
-------
equilibrium moisture of coal - the moisture content retained at
equilibrium in an atmosphere over a saturated solution of potassium
sulfate at 30 C, and 96 to 97 percent relative humidity. When the
sample, before such equilibration, contains total moisture at or
above the equilibrium moisture may be considered as equivalent to
inherent or bed moisture; and any excess may be considered as extraneous
moisture.
ash inorganic residue remaining after ignition of combustible
substances, determined by definite prescribed methods.
NOTE 1 -Ash may not be identical, in composition or quantity,
with the inorganic substances present in the material
before ignition.
NOTE 2 In the case of coal and coke, the methods employed
shall be those prescribed in Methods D 271.
volatile matter - those products exclusive of moisture, given off
by a material as gas or vapor, determined by definite prescribed methods
which may vary according to the nature of the material.
NOTE - In the case of coal and coke, the methods employed shall
be those prescribed in Methods D 271.
fixed carbon - in the case of coal, coke, and bituminous materials:,
the solid residue other than ash, obtained by destructive distillation,
determined by definite prescribed methods. "
The ASTM utilizes these definitions plus calorific value to classify
coals by rank. Table 1 summarizes the system. The substantial variations
in composition associated with the various ranks are typified by the example
in Table 2. The reported ranges are by no means all-inclusive but only
representative of a typical bed within the listed state. Of particular
importance are the wide variations in moisture, ash and sulfur content.
Moisture content is extremely variable because it depends both upon
the mine and upon the coal treatment after mining. For example, air
drying can reduce the moisture content. On the other hand exposure to
rain or snow can raise it. The situation is further complicated by the
-------
Table 1. ASTM DESIGNATION D 388 CLASSIFICATION OF COALS BY RANK
Class
I. Anthracite
II. Bimuninous
III. Subbitunn'nous
IV. Lignitic
Group
1. Meta-anthracite
2. Anthracite
3. Semi anthracite
1. Low volatile bitumi-
nous coal
2. Medium volatile
bituminous coal
3. High volatile A bi-
tuminous coal
4. High volatile B
bituminous coal
5. High volatile C
bituminous coal
1. Subbituminous A coal
2. Subbituminous B coal
3. Subbituminous C coal
1. Liynite
2. Lignite
Fixed Carbon
Limits, percent
(Dry, Mineral-
Matter-Free Basis)
Equal or
Greater
Than
98
92
86
78
69
Less
Than
98
92
86
78
69
Volatile Matter
Limits, percent
(Dry, Mineral-
Matter-Free Basis)
Greater
Than
2
8
14
22
31
Equal or
Less
Than
2
8
14
22
31
Calorific Value Limits
Btu per pound (Moist.
Mineral Matter-
Free Basis)
Equal or
Greater
Than
14,000
13,000
C 11,500
I 10,500
10,500
9,500
8,300
6,300
Less
Than
14,000
13,000
11,500
11,500
10,500
9,500
8,300
6,300
-------
Table 2. TYPICAL ANALYSIS OF COAL OF THE UNITED STATES
State
Pennsylvania
Arkansas
West Virginia
Pennsylvania
Kentucky
Ohio
Illinois
Wyomi ng
Utah
Washington
North Dakota
County
Lackawanna
Usual range
Johnson
Usual range
McDowell
Usual range
Clearfield
Usual range
Harlan
Usual range
Perry
Usual range
Christian
Usual range
Carbon
Usual range
Morgan
Stevens
Billings
Rank
Anthracite
Semi anthracite
Low-volatile
bituminous
Medium-volatile
bituminous
High-volatile A
High-volatile B
High-volatile C
High-volatile C or
subbituminous A-
subbituminous B
Subbituminous B
Subbituminous C
Lignitic
Proximate analysis, percent by weight
Moisture
4.3
Dry coal
2.5
1-3
2.7
1-5
2.7
1-4
3
2-4
7.8
7-10
13.5
12-14
12.8
9-17
17.9
16.4
38.0
Volatile
Matter
5.6
4-7
12
11-14
17.6
1 5-22
21.6
21-26
36.9
37-40
36
37-42
35.6
40-42
37.1
39-45
27.7
29.8
25.8
Fixed
Carbon
80.6
80-86
76.1
75-82
73.7
72-74
66.9
64-70
56.6
57-60
47.3
49-53
39.3
44-47
43.5
47-56
36.7
34.8
29.4
Ash
9.5
9-14
9.4
6-12
6
4-9
8.8
7-13
3.5
3-6
8.9
7-11
11.6
12-15
6.6
4-10
17.7
19
6.8
Sulfur
0.6
0.5-0.9
2.3
1.3-3.8
0.7
0.5-0.9
1.8
1-3.3
0.7
0.5-1.2
2.0
1-3.5
4.2
4.4-5.5
0.5
0.3-1.3
0.5
2.0
0.6
00
-------
fact that moisture may deliberately be added to the coal either to
prevent spontaneous combustion or as a by-product of coal cleaning
operations. Such additions are not uniformly distributed throughout
the coal but tend to be more concentrated near the surface.
The variations in coal ash are due to similar complexities. The
ash arises from several sources. Because coal is extremely inhomogeneous
there are bulk contaminants distributed throughout the coal even within
the seam. Furthermore, inevitably contaminants such as clay, shale and
other rocks from the floor and roof of the seam are included along with
the coal. These types of contaminants are the ones typically removed
by coal cleaning methods. In addition, inorganic materials which are
inherently chemically contained in the coal contribute to the ash content.
Typical ranges for the ash content associated with the sum of both of these
contributions are shown in Table 3.
Table 3. TYPICAL ASH CONTENTS
Si02
A1203
Fe203
CaO
MgO
Ti02
Na20 + K02
S03
20-60
10-35
5-35
1-20
0.3-4
0.5-2.5
1-4
0.1-12
It is apparent from the above discussion that there are substantial
variations in the composition of coal obtained from different sites.
A far more critical problem from a measurement standpoint is that large
variations occur within even relatively small volumes of coal. For this
reason extremely detailed sampling procedures have been developed and
are documented by organizations such as ASTM. The basic objective of
-------
these techniques is to obtain a small representative sample from a
large volume of material. Typically this is achieved by taking a
large number of moderate size samples. These samples are then ground
to a reduced size and thoroughly mixed. New samples are obtained from
the resultant mixture and the process iterated until a suitably uniform
mixture is obtained. Details of sampling equipment, procedures and
accuracy can be obtained in volume 26 of the ASTM Standards and will not
be repeated here. From a process control standpoint the most significant
feature of these procedures is that they are quite time consuming. In
practice typically hours are required and even under the most optimistic
of conditions it is hard to imagine.sampling times less than several
minutes. Some of the nuclear measurement techniques are particularly
attractive for process control because they probe large sample values
and do not require such sampling approaches.
2.1.1 Background Information on Nuclear Interaction (Ref. 4-6)
In order to develop a basis for understanding the nuclear diagnostic
technologies a brief introduction to the pertinent interactions is presented
in the following sections. First, the concept of a cross-section is reviewed
and the distinction between attenuation and absorption is made. Next,
photon-matter interactions are discussed. Finally, neutron-matter interac-
tions are considered.
2.1.1.1 The Cross Section Concept
The concept most often utilized to quantify the probability of nuclear
reactions is the cross-section and is typically denoted by the symbol a with
an appropriate subscript. A cross section can be visualized as an area
presented by scatterers to an incident beam and is obtained from the ratio
# incident particles scattered per second
# incident particles per unit area per second
Consider for example, a slab containing scatterers which can be thought
of as hard spheres of radius R which absorb an incident particle if 1t strikes
them. The geometrical cross section of each sphere is obtained from
the expression
A = irR2 cm2/sphere
10
-------
Assume that the slab has an area A, a thickness t and contains N scatterers
per unit volume. The number of scatterers per unit area is therefore:
A~cm " on
If the slab is thin enough that the scatterers do not shadow each other
(i.e., Nta « 1), then the probability of an incident point particle striking
a scatterer is
P = (Nt) a (6)
Usually single scattering events are not investigated but rather a number of
particles are incident on the scattering material. For instance, if n inci-
3
dent particles per cm moving with a velocity V are incident on the target,
2
the flux S = nV particles/cm and the anticipated number of scattering events
p
per cm of target is equal to the expression
(nV) (Nta) (7)
from this viewpoint an alternative expression for the cross section is
collisions per square centimeter per second
a (particle flux) (scatterers/square centimeter)
In practice this expression is the one which is most frequently utilized.
The above simple case can also be used to derive the well-known exponen-
tial attenuation behavior. Consider the penetration of an incident flux S(o)
to a depth X into the slab. Evidently from equation (7):
dS(X) = -S(X) (No- dX)
dx =
the obvious solution is the function
S(X) = S(o) e'NaX (8)
The quantity Na is referred to as a linear attenuation coefficient per cm
11
-------
of travel and is frequently utilized instead of the cross section to charac-
terize a target material.
The cross section concept is much more general than the above discussion
indicates. It should be appreciated that the effective area concept is just
a visualization aid and that in fact nothing in the system actually need have
this particular geometrical area. Furthermore, although the interaction of
a particular incident particle is indeed either an all or nothing process,
the event need not be an absorption. In a more general sense any type of
interaction process can be said to possess a cross section. A complete
specification of a scattering event requires a delineation of the incident
particles, the target, and the products after scattering. These products
can include particles not even present before the interaction. The complete
specification also requires a detailed knowledge of the quantum states of
all involved particles.
As an illustration of the subtleties involved, consider the concepts
of attenuation, scattering and absorption of photons. The linear attenuation
coefficient y is equal to the total cross section per scatterer (i.e., the sum
of all processes) multiplied by the number of scatterers per cm3. It includes
both elastic processes in which the photon is merely deflected without giving
up energy and processes in which the photon energy is actually absorbed by
the scattering medium. The elastic processes are referred to as elastic
scattering and are characterized by a scattering coefficient y . The absorp..
tion processes are characterized by an absorption coefficient y . The atten-
uation coefficient incorporates both processes:
y = ys +ya (9)
It should be appreciated that if desired, both y0 and y= can be further
s a
broken down into various subclasses.
2.1.1.2 Interaction of Gamma Radiation With Matter
Gamma rays are photons which accompany nuclear transitions. The gamma-
ray matter interaction is characterized by the fact that each y-ray is
removed in an all-or-nothing manner from the incident beam. Several competing
12
-------
alternative reactions can occur. The principle possibilities are an interaction
with the entire atom, an interaction with a single electron, or with the atomic
nucleons.
In each type of interaction the number of photons removed is proportional
to the thickness traversed and is evidently a solution of the differential
equation .,
&=-yl (10)
where
I = incident intensity
u = linear attenuation coefficient
X = depth into scatterer
For the simple homogeneous monoenergetic case the solution to the equation
is the well known exponential attenuation I = IQe~y . This simple viewpoint
expresses the basic concept embodied in the study of nuclear interactions.
The study of y interactions focusses upon the determination of y as a function
of y ray energy and material and upon the products of the incident gamma rays
that are interacted with.
In applications relevant to the diagnostic utilization of gamma rays sev-
eral types of process are significant and dictate the selection of suitable gamma
rays. At low energies the predominant interaction process is the photoelectric
effect. The precise energy at which this statement is correct is highly depen-
dent upon the atomic number (Z) of the scattering atom but roughly speaking
these effects dominate at energies of order 1 Kev-500 Kev. In the photo-
electric effect the incident photon is absorbed by a bound electron and one
electron is emitted usually from the K or L shell. Energy conservation yields
hvi = Be + KE (11)
where
hv- = incident gamma ray energy
Be = binding energy of electron prior to ejection
KE = electron kinetic energy
The excited atom promptly emits a characteristic X-ray or an Auger electron
as it relaxes to a lower level. This type of process is quite sensitive to
the energy of the gamma ray and to the specific atomic scatterer involved and
13
-------
is the basis of X-ray fluorescence studies as discussed further below.
As the gamma ray energy increases to become much larger than the binding
energies eventually the precise bound-state of the electron becomes of minor
importance and the probability of interacting with single electrons becomes
the dominant interaction. The scattering is as if the electrons were free
and at rest. This is known as the Compton Effect and dominates at energies
of ^ 0.5-1.5 Mev. The situation is illustrated in Figure 1 and energy and
momentum conservation yield the equations
5" (1 " cos*)
(M)
The important points to note are that the Compton shift in wavelength in any
particular direction is independent of the incident photon energy but that
the energy shift increases strongly as hvQ increases. Solving for v1
yields:
' >
v " 1 + o(l - cost))) (14)
hvn
where a = V
"V
It is apparent that if this type of scattering is averaged over all angles, a
continuous distribution of energies is observed. In Y- ray-scintillation spectre*
scopy this continuous spectra is referred to as the Compton edge. There 1s
however a maximum possible shift namely the value obtained when =180°. This
minimum value of hv' is called the backscatter peak energy and determines the
extent of the Compton edge as measured from a total energy peak. The most Impor-
tant feature of Compton scattering from an application standpoint is that it
ignores the nucleons and is primarily an electron counter.
If the gamma energy is sufficiently high additional processes become
possible. The one of greatest interest for diagnostic purposes is pair
production. In the field of a charged particle a photon may be totally absorb*
and a positron-electron pair emitted. A minimum eneergy of 1.02 Mev is requi^
for this process to occur. Pair production begins to be noticeable at about
1 Mev and dominates over Compton scattering at F«\*5-10 Mev. Because the proces-
14
-------
Figure 1. Compton effect geometry
15
-------
usually involves the atom nucleii's electric field the cross section is depen-
dent on the particular nucleii and typically increases as Z2. At still higher
energies additional processes are possible but are not of interest to current
discussion.
As stated earlier the three important interactions dominate attenuatl
By way of illustration of the magnitudes Figure 2 shows their relative
magnitudes in cm2/g Air.
10
MeV
Figure 2. Mass attenuation coefficients for photons in air.
Note that the total attenuation is almost entirely
due to Compton effects in the intermediate
energy range.
This figure illustrates all the comments made about relative importance of
the processes. It also shows the relative magnitudes of the absorption.
In particular note the high cross sections at low energy as this will later be
important in the realization that X-ray fluorescence only probes surface
layers. On the other hand for photons > 1 Mev substantial penetration thru
thick material layers makes it possible to probe into macroscopic volumes
of coal.
16
-------
2.1.1.3 Neutron Matter Interactions
The neutron-matter interaction is quite different from the y-ray
matter one and thus requires a separate discussion. Like the y-ray matter
interaction it is convenient to divide the interaction types into inelastic
and elastic cases. The neutrons themselves are classified by kinetic
energy. Those with energies greater than 0.5 Mev are considered fast and
those less than 0.5 Mev are considered slow. A more detailed classifica-
tion scheme is also sometimes used as follows:
Table 4. CLASSIFICATION OF NEUTRONS BY ENERGY LEVEL
% 0.25 ev
^ 0.2 ev
^ 0.4 ev
% 0.6 ev
1-10 ev
10-300 ev
500 ev-0.5 Mev
> 0.5 Mev
> 20 Mev
Thermal
Epithermal
Cadmium
Epicadmium
Slow
Resonance
Middle
Fast
Ultrafast
The simplest type of interaction is just an elastic collision of a
neutron with an atomic nucleus. The problem is easily treated by clas-
sical mechanics and the results basically indicate that the most effective
energy transfer occurs when a neutron strikes a nucleus with a mass
approaching that of the neutron. More quantitatively it is possible to
calculate the average value of the quantity obtained from the expression:
- ln
(Eo/El)
(15)
where
= logarithmic decrement of energy
E = initial neutron energy
E, a energy after on collision
It can be shown that
(16)
where
A = atomic mass of the target
17
-------
For a neutron to slow from an energy of E^ to an energy Ef requires
[In (E./Ef)]/£ collisions. Table 5 illustrates some typical values.
This table shows that elastic neutron scattering is strongly dominated
by low atomic number elements. Notice that if the scattering is evalued
on a per mass basis rather than per atom, the low A values are even more
dominant. As will be discussed later, this behavior can be utilized to
detect moisture via the hydrogen induced scattering and is also a very
important consideration in selecting shielding materials.
Table 5. NEUTRON SCATTERING PROPERTIES OF NUCLEI
Element
Hydrogen
Helium
Carbon
Sulfur
Iron
A
1
4
12
32
56
e
1.0
0.425
0.158
0.061
0.035
Number of collisions
from 10 Mev to 0.025 ev
19.8
46.6
125.0
325.0
566.0
In addition to the elastic (n,n) type collisions which have been
described above several types of important inelastic collision processes
are possible. The most important possibilities of interaction with a
nucleus are as follows:
1. The nucleus is excited to a higher energy level and returns to
its ground state by the emission of one or more photons.
2. The neutron is captured and forms a compound nucleus. In some
cases this new nucleus is in an excited state and decays to the
ground state via the emission of y-rays. This process is
referred to as radiative capture and is represented by the
symbol (n.y).
3. The neutron is captured and another elementary particle is emitted
18
-------
4. Additional processes such as fragmentation and fission of the com-
pound nucleus may also occur.
The relative probabilities of those processes are highly dependent
upon both the nucleus involved and upon the energy of the neutron. In
the current work the neutrons will typically have a fairly low energy
and the atomic weight of the target nucleii will be low. Under those
conditions the (n,y) processes are usually the most important. Further-
more, the cross-sections for such reactions are typically largest for the
slowest neutrons and vary approximately inversely with the neutron velocity.
Thus, those neutrons that have been slowed to near Boltzman room tempera-
ture velocities (thermalized) dominate the inelastic reactions.
2.1.2 Activation Analysis Instruments (Ref. 7-12)
Inelastic neutron scattering can be effectively utilized for elemental
analysis. Prompt gamma radiation associated with thermal neutron capture
is especially useful. Figure 3 shows how the technique is employed. A
suitable neutron source such as Californium 252 is utilized to provide a
suitable neutron flux. These neutrons are thermalized and diffuse through
a substantial probe volume of material. Occasionally a neutron interacts
with a nucleus via an (n,y) reaction. The energy of the promptly emitted
gamma ray is characteristic of the energy level structure of the
nucleus and can be used to infer the element's presence. Scattered gamma
rays are detected by energy sensitive scintillation crystals, usually Tl
activated Nal. The resultant signals are analyzed to determine the gamma
ray energy spectrum and the energy distributions utilized to infer the
elemental composition.
Figure 4 shows a typical spectrum obtained by Hall et al. To under-
stand this spectrum it is first necessary to appreciate how a scintillation
spectrometer works. Basically it consists of a scintillator material, a
photomultiplier, and a pulse height analyzer. The function of the scintil-
lator is to convert the gamma ray energy to visible radiation. This visible
radiation is then converted to an electrical pulse and amplified by the
photomultiplier. Each gamma ray produces an electrical pulse with a height
ideally linearly related to its energy. The pulse height analyzer sorts
pulses with respect to their amplitude and thus forms a histogram of the
energy distribution as shown in Figure 4.
19
-------
PULSE HEIGHT ANALYZER
NEUTRON
SOURCE
r>
SAMPLE
PHOTOMULTIPLER
SCINTILLATION
CRYSTAL
Figure 3. Thermal neutron capture gamma ray analysis
experimental configuration
20
-------
ro
Tenfold
scale *
increase]]
CHANNEL NUMBER
Figure 4. Gamma ray spectrum of coal measured with Cf-252 source and 6- by
7-inch Nal detector
-------
Even monochromatic y-rays unfortunately produce complicated scintil-
lation spectra. In addition to a total absorption peak, they produce both
Compton continuums and for energies above 1 Mev pair production peaks.
Furthermore, because positrons or electrons can escape rather than being
detected, escape peaks downshifted by multiples of 0.51 Mev may also appear.
Finally, because intrinsic efficiencies are not perfect even the total
absorption peak has a finite width. Typical values of resolution are
highly dependent on crystal size and approach 10* for large crystals.
Here resolution is defined-as the full width of a peak at half maximum
energy divided by the energy value of the peak measured for the 0.66 Mev
y-ray of Cs 137.
For real spectra the situation is even more difficult because many
different peaks each with their own complex spectra may be present.
Ambiguities resulting from the overlap of different spectra are referred
to as interference.
Figure 4 illustrates many of these features. Peak P3 corresponds
to the 5.4 Mev gamma ray from sulfur. Gamma rays from sulfur also contr1»
bute to P2 which is the first escape peak at 4.9 Mev. For the large crystal
employed the second escape peak at 4.4 Mev is quite minor. Unfortunately
4.9 Mev gamma rays from Si also contribute to P^ thus interfering with the
sulfur measurement. They also contribute to P, via their first escape
peak. Finally, iron and other elements contribute further interfering
signals at 6.0 Mev and its first escape peak.
Martin and Hall have carefully modeled the elements present in a
typical coal sample and concluded that these interferences could be com-
pensated for by assuming that the percentage sulfur could be obtained from
the expression:
%S = K (B-A/2) (17)
where
%S = percent sulfur
K = constant
B = area under inner two peaks
A = area under all four peaks
They tested this approach both experimentally and theoretically and
22
-------
found it to be fundamentally sound.
It is also possible to infer other elements by activation analysis.
For instance the 2.23 Mev peak shown in Figure 4 is due to hydrogen and
as stated earlier, peaks due to Si and Fe are also readily observed.
Table 6 lists some analytical sensitivities for thermal neutron capture
gamma rays as tabulated by D. Duffey et al. Sensitivity is expressed as
S = Io/A
(18)
where
S = sensitivity
I = number of gamma rays per 100 neutrons captured
a = neutron absorption cross section in Barns
A = atomic mass
It must be appreciated that it is not practical to resolve all those lines
but the possibility of obtaining further information is clearly present.
Table 6. ANALYTIC SENSITIVITIES FOR THERMAL
NEUTRON CAPTURE
El ement
C
N
Na
Mg
Al
Si
S
Ca
Fe
Energy
(Mev)
4.95
6.32
5.27
6.40
3.98
3.59
3.92
7.72
6.38
4.93
5.42
4.87
4.42
7.65
7.63
6.02
5.92
Sensitivity
(la/A)
0.019
0.089
0.136
0.592
0.501
0.402
0.10
0.175
0.072
0.402
0.678
0.132
0.116
1.04
1.27
0.379
0.389
23
-------
2.1.3 Gamma Ray Density Gages "(Ref. 13-15)
For a homogeneous material of thickness L and attenuation coefficient y
(cm } the simple expressions discussed in Section 2.1.1.2 imply that if a collimatct
monochromatic gamma ray flux I is incident upon the material then the transmitted
flux is obtained from the expression:
I = I0 e-yL (19)
Thus, if I and I are measured the quantity yL can be calculated:
yL = In ]- (20),
o
This approach is the basis of virtually all gamma ray density gauges.
Even for homogeneous materials and mono-energetic gamma rays there are limitatio;
to the above analysis. These equations assume that only unscattered gamma rays
are measured. In practice this idealization is not achieved. To even approach this
situation would require an energy spectrum analysis of the detected radiation and
a computer analysis of the peak area to subtract scattered radiation. Furthermore,
real gamma ray sources are only approximately collimated so that angular variations
may also influence the results. These complications are normally reduced by doing
in situ calibrations so that to first order they are included in the system normal-
ization.
Although the above discussion concerned only homogeneous materials, it is
possible to extend the results to inhomogeneous ones. In particular if the energy
of the gamma ray source is judiciously selected it can be shown that the product
vL is a direct measure of the amount of mass interposed between the incident and
transmitted flux. As previously discussed for gamma ray energies of order 0.5 -
1.5 Mev and low atomic weight elements the attenuation is dominated by Compton
scattering. This means that the product yL is directly proportional to the number
of electrons in the material volume. The following equations describe the situation:
Ne = ANoZ (21)
24
-------
where N = number of electrons per unit volume
p = average density
Z = average atomic number
A = average atomic weight
N = Avagadro's number
v = f(E) N (22)
where
p = attenuation coefficient
f(E)= energy dependent, material independent scattering function
= KE = const for a particular energy
The mass contained in a unit area slab of thickness L is obtained from
m = PL (23)
The ratio of the quantity yL to this mass can be obtained by substitution into
the above equations and is equal to
f(E) N Z
R = A° (24)
Thus the only material dependence in this ratio arises from the Z/A term. Table 7
shows that with the notable exception of H Z/A is virtually independent of material
for low atomic number constituents. Thus the product pL is a direct measurement
of the interposed mass independently of its composition.
Table 7. Z/A VALUES
Atom
H
C
0
S.
i
S
Z
1
6
8
14
16
A
1.008
12.01
16.00
28.09
32.06
Z/A
0.992
.500
.500
.498
.499
25
-------
The mass fractions associated with H atoms may induce some errors into
density measurements but these usually only amount to a few percent and to the
extent that hydrogen fractions are constant can easily be absorbed into the
calibration. The only case where significant problems may arise is the measure-
ment of low density solids in a water slurry. In this case the effects can be
substantial.
The trade-offs associated with designing a gamma ray density measurement
system are best appreciated by performing standard error analysis.
Let S = the signal obtained with no interposed material in the
probed volume
S, = the signal transmitted through the sample
then g
yL = In ^- (25)
Differentiating yields:
dSQ dSL
d(wL) = 3 - 5 (26)
0
for independent measurements of SQ and SL the fractional uncertainty in pL 1s
obtained by standard techniques:
oS,
where the a quantities denote standard deviations for the subscripted quantities
On this basis it appears advantageous to make yL as large as possible. Howevo
the transmitted signal decreases exponentially with yL thus yL can't be too
large or the statistical uncertainty in SL will be large. This trade-off is
integrally related to the -y source size and the thickness of the material
measured. A numerical example will be discussed in the applications section.
26
-------
The utilization of the technique to assess coal flow also may be influenced
by spatial inhomogeneity of the material. It can be shown that if a material
contains inhomogeneously distributed regions of low density across the path
of the gamma rays that the non-linearity of the technique will systematically
lead to underpredictions of mass attenuation. The example of Figure 5 illus-
trates the situation. Case 1 consists of a simple homogeneous uniform slab
with thickness L. The transmitted beam intensity is evidently I e"ML. Case 2
consists of a scatterer with the same total volume and frontal area as in case 1
but with two thicknesses. Half of the sample is L-X thick and half is L+X thick.
The transmitted radiation is therefore:
!o -u(L-X) . jo -u(L+X) = I e'uLcosh UX (28)
2~e 2
Cosh nX is always greater than 1 and approaches eyXfor large values of yX.
Thus the case 2 situation would lead to an underestimate of the amount of
attenuating material present. In the extreme case when pX » 1 the fractional
error approaches X/L. It is clearly necessary to minimize such effects for very
inhomogeneous products such as coal by making the mixture a's uniform as possible.
This can frequently be achieved by either integrating over a large area or over
a long period of time. Even when these techniques are successful it is still
desirable to perform a calibration with the coal itself to normalize out such
systematic errors.
2.1.4 Neutron Moisture Measurements (Ref- 16-19)
Fast neutrons are readily produced with continuous energy distributions
in the range of 1 to 6 Mev by utilizing radioactive sources such as 210 Po-Be
or 252 Cf. If such neutrons are incident upon a medium, a portion of the
neutrons are scattered and slowed down to thermal energies through a series
of elastic collisions with the target nucleii. As described in Section 2 ,
this process is strongly dominated by light nucleii particularly hydrogen.
Hydrogen not only is more effective at absorbing energy but also has a high
collision cross-section. For instance, representative cross-sections for
24 2
C and H atoms are 4 and 20 X 10 cm , respectively. Furthermore, according
27
-------
Ioe
-Ml
CASE 1
t t t
_L
f
I. e coshyX
L+X
I
t f t
CASE 2
Figure 5. Inhomogeneity effects upon gamma ray transmission
28
-------
to Table 7 hydrogen is 6 times more effective than C at transferring energy
per collision. Thus, a hydrogen atom is approximately 20/4 x 6 = 30 times
more effective at slowing fast neutrons than a carbon atom. On a weight
basis the ratio is even more striking and becomes 360.
This strong preferential scattering by hydrogen can be utilized to infer
hydrogen content of materials by employing detectors sensitive only to low
energy neutrons. The approach is to direct a beam of high energy neutrons
into the material of interest as shown in Figure 6 . Typical penetration
depths of order 0.5 m are achieved. Thermal neutron detectors mounted nearby
detect those neutrons which have been thermalized and nearly backscattered.
The primary source of such neutrons is thermalization by hydrogen atoms and
the measured signals are proportional to the amount of H present.
In some cases it is thus possible to infer moisture content once the
hydrogen content is known. Unfortunately, this is not always the case
because this method does not distinguish the chemical state of the hydrogen.
Thus in the case of coal it is necessary to independently determine the
hydrogen content of the dry coal before a moisture determination can be made.
Under favorable circumstances when only one type of coal is utilized,this
may be practical but in cases where different types are involved it may prove
impossible to separate the intrinsic and moisture contributions.
2J.5 X-Rav Fluorescence (Ref. 20-24)
If a target material is irradiated with high energy X-rays, it emits
fluorescent radiation with an energy spectrum consisting of discrete lines.
The discrete spectrum energies are characteristic of the energy level
structure of the scattering atoms and allow a determination of the elements
present in the target. Figure 7 shows the general classification scheme
utilized to describe the transitions responsible for a given line. For
example, when a K electron is knocked out of an atom, another electron can
replace it giving rise to characteristic K series emission lines. Most
X-ray fluoresence applications are based upon K or L lines.
In general X-ray fluorescence becomes an easier technique to apply
as the atomic number of the atom of interest increases because atoms with
larger atomic numbers produce characteristic X-rays of higher frequency.
It was found empirically by Moseley that the characteristic X-ray frequen-
29
-------
FAST
NEUTRON
SOURCE
SLOW
NEUTRON
DETECTOR
Figure 6. Principle of neutron moisture meter
30
-------
M
L
, 1
^CL fro
I Na
^_
1111
Figure 7. Term scheme, showing the origin of the X-ray
spectra of the elements.
31
-------
cies are well fitted by the function:
v = K(Z-a)2 (29)
where v is the characteristic frequency, Z is the atomic number, and K
and a are constants common to all elements but different for the various
lines ie. K ,, K^ etc. At low photon energies the cross section for a
given atom varies as (v2)'7/2. Thus roughly speaking the penetration
depth of a characteristic X-ray into an arbitrary material varies as Z7.
Evidently low Z characteristic X-rays are much more readily absorbed
than high Z ones. There is another experimental difficulty associated
with the excitation of the light elements. It is called fluorescent yield
The problem is that for low atomic number elements, K emission radiation
may knock out an L electron rather than leaving the atom. This process 1s
called the Auger effect and becomes increasingly important for low atomic
number elements. For example, between Z = 30 and 15 the observed yield
drops from approximately 45% to 5% due to this effect alone. Both this
effect and the very low penetration depths make it impractical to utilize
X-ray fluorescence for low atomic number elements.
Figure 8 shows schematically how X-ray fluorescence is used to
analyze materials for the presence of high Z elements. The first require-
ment of the system is an excitation source. In large analytical instruments
this may consist of an electron beam or an X-ray tube. In small portable
devices isotopic X-ray sources have proved to be more convenient. As this
source bombards the target a portion of the target atoms emit characteris-
tic X-rays. Part of the fluorescent radiation is received by the detection
system and some sort of energy separation scheme employed. In analytical
instruments which have very high sensitivity, typically a crystal system
utilizing Bragg scattering is employed and a detector only processes those
X-rays dispersed at the appropriate angle. Somewhat simpler but less sensi-
tive systems have successfully utilized material filters instead. The
detector utilized is usually a simple proportional or ionization counter
which produces a discrete count for each X-ray detected.
For purposes of quantitative analysis it is necessary to correlate
the measured intensity of each line with the percentage composition of the
specimen. This signal is related to a number of features of the measure-
ment system response and geometry. It is however possible to control for
32
-------
EXCITATION
SOURCE
RGET MATERIAL
ENERGY SELECTION
SYSTEM
DETECTOR
Figure 8. X-ray fluorescence configuration
33
-------
these factors so that the limiting feature is the matrix effect which
varies with the sample composition. Matrix effects include both absorption
and enhancement effects. The absorption effect is caused by the fact that
both characteristic and primary X-rays must pass through increasing amounts
of matrix material as penetration goes deeper into the sample. Enhancement
is a further complication that occurs because characteristic fluorescent
lines of some matrix elements may excite the characteristic spectra of the
desired element. These effects are highly complex and dependent upon both
chemical composition and physical form of the matrix. These effects make
careful corrections for matrix composition necessary to obtain useful
results. The attenuation problem is sufficiently acute that X-ray flucres
cence techniques for atomic numbers less than 10 are almost never used and
for atomic numbers between 10 and 20 it is necessary to eliminate air
attenuation effects by utilizing either a vacuum path or purging with
Helium.
34
-------
2.2 FLOW MEASUREMENT INSTRUMENTS
2.2.1 Dry Coal Flow
Of the types of measurements considered in this report, this is the one
which can be performed with the least difficulty, lowest cost, and greatest
accuracy. In any total instrumentation system where the final output is mass
flow rate of a coal constituent, such as carbon, ash, or sulfur, if the total
mass flow measurement is that of a dry (i.e., non-slurry) coal stream, the total
mass flow measurement should have a negligible impact on system error in a pro-
perly designed and operated system.
Measurement of dry coal flow rate does not differ in any important aspects
from measurement of a large variety of other solids, The widespread need for
solid flow measurement for everything from potatoes to iron ore has resulted
in very accurate and reliable instrumentation. The most widely used measurement
techniques can be categorized as static (i.e., batch) or dynamic, with the latter
category subdivided into vertical motion techniques and horizontal motion
techniques. Weight scales are means by which the static measurements are taken.
Since weighing is the means by which raw materials have been bought and sold
for thousands of years, there has been a corresponding great interest in accurate
weighing techniques. In the context of our current interest, we need to note
the following:
Quantities of coal at least up to that contained in a typical
railroad car can be weighed with very high accuracy (^ .1% or
better).
As is discussed below, static weighing of a given amount of material
is the preferred means by which a calibration of a dynamic (continuous)
system is obtained.
Flow measurements using static weighing techniques can be done on a batch, or
semi-continuous, basis, such as having individual coal-carrying railroad cars,
or trucks, momentarily come to rest on a static scale, or by having coal fed
into a hopper, stop the flow, take a weight measurement, dump the coal, and
then resume the feed. The development of instrumentation which can be used
directly on conveyor belts has eliminated the need for batch weighing measure-
ments other than for calibration purposes. Presented below is a background
section on conveyor belt instrumentation (here termed weigh belt instrumentation),
35
-------
followed by a short section on techniques such as loss-in-weight, which apply
to vertical flow systems.
2.2.1.1 Weigh Belt Techniques
The term "weigh belt", "belt scale", and "conveyor scale" are all used
to describe the type of instrument shown schematically in its two most common
configurations in Figures 9a and 9b. These instruments are used in conjunction
with a conveyor belt of some sort which moves the material of interest between
two points. The configuration in Figure 9a is generally used in conjunction
with long conveyor belts whose primary purpose is conveyance, as from a coal
stock pile to a power plant hundreds of meters away. The configuration in
Figure 9bis normally used at the outlet of a hopper, so that the coal falls
vertically onto the belt, and then the coal is conveyed only a short distance
(^ 3 meters typical) before being unloaded. As far as the instrumentation systen
is concerned, the only significant difference between these two types of
applications is the length of the belt. All common weigh belt instruments for
either application have three features in common:
1. A sensing section of the belt, which is a specified length along
which the contents being conveyed are continuously weighed, providing
the 7- term in Equation (1).
K*
2. A device to measure linear belt speed through the sensing section,
providing the a term in Equation (1).
3. An integrator to combine the two measurements to produce m,., the
total mass conveyed since a specified starting time.
There are instruments in which these elements are done wholly by mechanical
means. Most currently marketed instruments use a combination of mechanical
and electronic devices, and a nuclear radiation technique is becoming popular
for the weight measurement term.
The basic mechanical or electro-mechanical weighing technique is shown
in Figure 10. The basic elements in the assembly are a continuous conveyor belt
supported periodically along its length by idlers (cylindrical rollers) which
are attached to a metal frame, a belt drive mechanism, one or more idlers or
idler assemblies attached to a floating weigh bridge, a belt speed sensor
either in contact with the belt itself or with one of the idlers, and a
36
-------
COAL
FEED
IDLER
WEIGH BRIDGE
IDLERS
BELT
BELT
SPEED
SENSOR
LOAD CELL
(FIXED TO
FRAME)
WEIGH BRIDGE
(FLOATING)
DISCHARGE
GRAVITY
TAKE-UP
00
Figure 9a. Weigh belt installation for long belt
COMPONENTS
SAME AS
ABOVE
Figure 9b. Weigh belt installation for short belt
-------
RAMSEY MODEL 10-4 WEIGHBRIDGE ASSEMBLY
(4 IDLER SUSPENSION SHOWN, ALSO AVAILABLE IN 2 AND 3 IDLER SUSPENSIONS)
00
51R75
WEIGH IDLERS
WEIGH BRIDGES
Factory pre-assembled, unitized construction
Only eight mounting bolts required for fast
and easy installation in field.
Full-floating, rigid mechanical tubing
weigh platform--maintains alignment
permanently.
Four, super precision strain gauge
load cells (with special temperature
compensation) applied in tension
less than .001" deflection on weigh
PAMSEYlHGIKtllHMB COMMMV
-------
weighing device which senses the instantaneous.weight of the floating weight
bridge and the length of belt and its contents in the sensing region. In a
strictly mechanical system, the weight of the "live" components are transmitted
to a mechanical integrator via lever arms, pully and/or gears. Load
cells are the most commonly used weighing device presently sold. All manu-
facturers contacted who deal in electro-mechanical mechanisms use one of a
variety of types of load cell, sometimes in conjunction with a lever arm system,
to sense weight. Belt linear velocity is typically determined in a digital
manner by measuring the rotational speed of an idler which contacts the belt
or that of a precision wheel which is held in contact with the belt by a spring
loaded lever arm. The measurement is digital in that full rotations of a
wheel, or partial rotations based on sensing teeth regularly spaced around the
wheel circumference for higher resolution, are sensed as a function of time.
This rotation rate measurement is transformed into a belt speed measurement by
multiplying the rotational rates of the wheel by the wheel radius.
It is appropriate in terms of report organization to present the physics
of nuclear weighing devices along with that of other nuclear radiation techniques,
This is done in Section 2.1. For the purposes of this section, it is appropriate
to simply state that in a nuclear radiation weighing device, a nuclear radiation
source sends out gamma rays, a percentage of which are absorbed by any matter
placed between the source and the detector. The gamma ray attenuation is
proportional to the amount of mass placed between the source and the detector.
This mass normally consists of a very short (compared to the electro-mechanical
instruments) length of conveyor belt and the material on it, and, usually, a
part of the fixed framework below the belt. In principle, then, a nuclear
scale acts just the same as an electro-mechanical one, in that it measures
m/i, and it must deal with the same type of tare, i.e., the belt and some
portion of fixed structure. Here are two major differences between the two
approaches: The obvious one is that the weight sensing approaches are quite
different; the other is that the nuclear scale has no mechanical contact with
the conveying system, whereas an electro-mechanical device must be mechanically
integrated and in physical contact with the conveying system. The latter may
influence instrument selection in retrofit or unusual applications.
39
-------
The aspect of weigh belt use which it is appropriate to treat in this
section is calibration. Both electro-mechanical and nuclear scales have belt
and framework weight to be dealt with as tare in addition to the weight
of the coal itself. A partial list of possible causes of calibration shifts
is:
Change in belt tension
Material sticking to belt and being weighed more than once
Belt erosion
Change in density and/or positional distribution on the
belt of material being conveyed
Change in material loading
Change in composition of material being conveyed (applies to
nuclear scale)
Accumulation of coal dust on sensitive portions of mechanism
Sensor drift as a function of ambient atmospheric conditions
and time.
Several of these are rather easily eliminated through proper equipment
design, installation, and operation (belt tension, material sticking, distrl
bution and loading, dust accumulation, atmospheric condition changes). Others
must be dealt with by periodic recalibration.
To achieve optimum accuracy, we need to be concerned about two aspects
of system design and operation; material loading on the belt and method of
calibration. In each case, there will usually be cost/accuracy trade-offs
to be considered. The feed rate of material, ITU, as noted in Equation (1)
is the product of the material loading on the belt, m/£, and the belt speed
SL. Either or both can be varied to provide a desired total flow. The belt
speed measurement accuracy is essentially independent of the belt speed itself
i.e., it is a "percent reading" type accuracy. By contrast, it is normally
the case that load cells and nuclear scales tend to behave as "percent full
scale" accuracy instruments, so that significant changes in belt loading v*in
produce larger random errors. In addition, changes in material loading will
also introduce non-linearities in the calibration curve of the weighing dev-l
40
-------
(output vs m/£), which could be eliminated only through meticulous calibration
over a wide loading range. Consequently, optimum accuracy will be obtained
when the material loading is held constant at a specified level and belt speed
is used as the primary variable to obtain desired variations in tru. The other
benefit of operating in this mode is that the system time response is shorter
for a variable speed belt system, since the delivery rate changes concurrently
with changes in belt speed. In a constant belt speed system, there is an
inherent delivery delay time of ijl, where £ is the length of the belt
between the loading and unloading points.
Three methods of system calibration are in general use at present:
Static calibration, dynamic calibration with chains, and dynamic calibration
with material being conveyed. It must be emphasized that only the last method
constitutes a true system calibration. In this method, a known quantity of
material, an established by static weighing either before or after conveyance
along the belt, is conveyed past the instrument. This is a true in-situ
calibration, and is the most time consuming and therefore most expensive method
of calibration. A less rigorous method of calibrating electro-mechanical
systems is to use a calibration chain of known m/a affixed to the belt. The
objection to this common technique is that the chain will not load the belt
in quite the same manner as the actual material, producing systematic errors.
In the static calibration method which is becoming popular because it is quick
and inexpensive, a known dead weight is attached to the load cell in an electro-
mechanical system, or a metal plate of known attenuation properties is placed
between the source and detector in a nuclear scale system. This is not a
system calibration - it is merely a single point load sensor check. Its
primary value would be as a quick sensor check between material calibrations.
It should not be used in lieu of material calibrations.
The current status of weigh belt instrumentation, including manufacturers,
type of equipment, accuracy and cost data, is presented in Section 3.2.1.
2 2.1-2 Vertical Flow Techniques and Miscellaneous Techniques
The low cost, accuracy, and reliability of weigh belt techniques have
resulted in their "cornering the market" for high flow rate dry granular
material bulk flow measurement. Some alternate techniques are mentioned
41
-------
below. These are usually employed for "hard to handle" materials (coal is
an "easy to handle" material) or are used in situations where a conveyor
' belt is not practical for some reason.
Conveyor belts must operate at an angle close enough to the horizontal
that the material will not slide. Under some circumstances, it may be necessary
or desirable to have a measurement system through which the material passes
vertically. Two such techniques are loss-in-weight systems and what will be
referred to here as a solids turbine.
A loss-in-weight system is shown schematically in Figure 11. The has*
components are a hopper, load cell, outlet metering valve or gate, and materl 1
feed line. Principles of operation are simple. With the outlet closed,
material is fed into the hopper until a desired total weight is reached, fh
material feed is then stopped, and the outlet is opened to achieve a desired
flow rate. The flow rate is determined by the change in the load cell readi
per unit time. The load cell is often connected via a servo loop to the out
let controlling mechanism so that a desired flow can be maintained. When th
load cell reaches a specified minimum reading, the outlet is closed and the
procedure repeated. This approach involves more and bulkier equipment than
a weigh belt for the same level of flow, and it is clearly a batch process
rather than continuous. Due to the relatively wide range which the load cell
must cover, the accuracy will generally not be as good as for a weigh belt
installation using a comparable load cell. Best accuracy for a loss-in-wei
system is about 1% of reading.
The solid turbine is shown schematically in Figure 12, and is directl
analogous to a turbine motor used for fluid flow. While a paddle wheet is
shown in the schematic for simplicity, actual devices use multiple helical
blades, so that the turbine axis is vertical. The kinetic energy of the
falling solid material impinging on the rotor blades causes the rotor to ti
at a rate proportional to the solid mass flow. This is a viable technique f
continuous flow measurement, but has not achieved the accuracies of weiqh
belt systems.
42
-------
//
RIGID
SUPPORT
FILL
HOPPER
GRANULA
MATERIAL
CONTROL
VALVE
OR
GATE
PROCEDURE:
1. HOPPER IS FILLED
(OUTLET CLOSED)
2. CONTROL VALVE IS
OPENED TO DESIRED
POSITION AND LOAD
CELL OUTPUT VS.
TIME GIVES MASS
FLOW RATE
3. CONTROL VALVE IS
CLOSED AND
HOPPER REFILLED
DISCHARGE
Figure 11. Schematic of Loss-In-Weight system
43
-------
SOLIDS FEED
FALLING SOLIDS
TURN TURBINE AT
A RATE PROPORTIONAL
TO MASS FLOW
DISCHARGE
Figure 12. Illustration of "Solids Turbine" concept
44
-------
One final technique worthy of mention, but applicable only for
rather low flow rates, is the screw feeder, shown in Figure 13. It is
similar in concept to positive displacement metering pumps for fluids.
Material from the hopper falls onto the screw, is conveyed along an annual
passage, and then falls off the screw at the exit. For a given screw, feed
rate is proportional to the screw rotation rate. A good feeder flowrate
accuracy is about 5%.
2.2.2 Coal/Mater Slurry Instruments
Coal/water slurry flow measurement is inherently more difficult than
dry coal flow measurement, and will rarely be as accurate. The measurement
is also complicated by the fact that there are inherently two parameters of
interest: The total flow rate (both coal and water) and the percent coal in
the slurry. This means in normal practice that two separate instruments are
required - one for flow and one for coal content.
The general approach which has been taken to slurry flow measurement
has not actually been an approach but rather an avoidance - this type of measure-
ment is performed only when necessary. Slurry flow instruments all have one
feature in common - they were originally developed for liquid flow application
where the solids content in the stream is negligible. It is also fair to say
that any flow device which is acceptable for cool/water slurry flow measure-
ment would work more accurately and reliably if the flowing medium were
water with only trace amounts of solids or non at all. Thus the coal in the
slurry is, as far as the instrument is concerned, a contaminant, which tends
to make its presence known by accumulating in undesirable places, acting as
an abrasive, and/or obstructing signal paths.
For purposes of this report, instruments are being separated into two
categories: Obstructionless (with respect to the slurry flow, the instrument
behaves as an equivalent length of straight pipe) and intrusive (one or more
instrument components are physically inserted into the flow stream). This
distinction is being made from the standpoint of applicability - experience
has shown obstruction!ess flow meters to be much more suitable for slurry
flows, although some manufacturers claim that their intrusive devices are
45
-------
DRIVE
UNIT
HOPPER
SCREW
DISCHARGE
Figure 13. Schematic of screw feeder
46
-------
acceptable for the present application. Both instrument categories are
discussed below, followed by a discussion of a system consisting of a
flow meter and density gage.
2.2.2.1 Obstruction!ess Flowmeter
2.2 2.1.1 Magnetic Flowmeter
As manufacturers are quick to point out, the magnetic flowmeter has
been around, at least in concept, since the 1830's when Faraday attempted
(unsuccessfully) a demonstration in the Thames River. Reference 25 notes
that the first electromagnetic flowmeter patent was issued in 1915.
The magnetic flowmeter is shown schematically in Figure 14. Basic de-
scriptions of its workings are presented in References 25-29. The operation
principle is Faraday's Law of Induction, which may be stated as (Ref.26):
"... motion at right angles between a conductor and a
magnetic field will develop a voltage in the conductor".
The device itself is rather simple, although not inexpensive, consisting of a
magnetic field across the pipe, and a pair of electrodes. The field lines,
direction of flow, and the line between the electrodes form a mutually
orthogonal system. The electrodes are typically mounted flush with the pipe
wall, or in some cases are slightly recessed from the wall. The material being
conveyed must have a minimum electrical conductivity for the meter to be able
to operate (i.e., generate a voltage between the electrodes). Ordinary tap
water has sufficient conductivity, so there is no problem with coal/water
slurries in that regard. Performance of the instrument is described by the
following equation:
E = KBdu" (30)
where
E = emf produced between the electrodes, volts
K = calibration factor, dimensionless
p
B = magnetic flux density, volfs/m
d = distance between electrodes (normally pipe I.D), m
u~ = slurry average axial velocity, m/s
47
-------
120V
60Hz
FIELD COIL (1 OF 2)
J*T J !
Amtf
LIQUID VELOCITY, yf
ELECTRODE (1 OF 2)
= CALIBRATION CONSTANT
FOR THE METER
B = MAGNETIC FLUX
GENERATED BY THE
FIELD COILS
Figure 14. Schematic of electromagnetic flowmeter,
taken from Reference 25.
48
-------
The desired outputs are volumetric flow and/or mass flow, given by
2
where
j
ty = slurry volumetric flow, m /s
o
A = pipe cross-sectional area, m
and . -n-pfi
*T = psi v = psi m '32>
Equation (30) is very illustrative of the workings of this type of meter.
The form of Equation (30) makes it clear that the magnetic flowmeter senses
velocity - not volumetric flow or mass flow. This means that a separate
density determination is required to obtain mass flow, as in Equation (32).
Another not quite obvious but very important point is that the meter output
will be essentially the same for constant average slurry velocity whether
or not the pipe is full. This is illustrated in Figure 15. As long as there
is a conductive path through the slurry between the electrodes, the flow-
meter cannot distinguish between a full and a partly full pipe. This imposes
the requirement on the system that the pipe always be full.
In a given setup, the calibration factor K in Equation (30) is
established by in-place calibration (usually by dumping the slurry into a
volumetric tank). The calibration factor is not subject to change due to
changes in slurry properties such as conductivity (as long as the minimum
conductivity requirement is met), nor is it subject to changes in the
coal /water ratio or coal composition. It is, however, subject to changes
in flow profile. Equation (30) was derived from the case of an axisymmetric
flow profile. Thus the flow as it goes through the meter may be uniform,
partially developed (as long as it is axisymmetric) or fully developed
(either laminar or turbulent) without causing a calibration shift. The
following quote from Reference 27 describes meter performance in asymmetric
f 1 ows :
"... it is easy to show that a velocity variation along the electrode
diameter, but with constant velocity along each vertical line
(parallel to the flux lines), will produce significant error. Such
49
-------
PIPE FULL
u" = 2.0 m/s
b. PIPE PARTIALLY FULL
IT = 2.0 m/s FOR ACTUAL
SLURRY
METER READING * 2.0 m/s METER READING = 2.0 m/s
RESULT: CORRECT VOLUMETRIC RESULT: INCORRECT (HIGH)
VOLUMETRIC FLOW
SENSED
FLOW SENSED
Figure 15. Illustration of magnetic flowmeter error due to
non-full pipe
50
-------
a profile is farfetched. ... The author has been unable to produce
errors discernible in a background scatter of ± 0.5 percent by any
piping configuration, including a throttling gate valve 4 [pipe
diameters] upstream from the electrode plane."
One source of velocity-related systematic error which we have not found
treated in the literature is the error due to a net velocity differential
between the water and the coal in the slurry. Fortunately, with respect
to this error source (but unfortunate for density determination), the
density difference between coal and water is small (^ 30%), so the velocity
difference would tend to be small except for large coal particles. This
differential will be a maximum for vertical flow. Vertical flow upward
through the meter is often desired to insure that the pipe is full. If
there is a significant velocity differential between the coal and the water,
significant being 1% or more, the meter will register the average velocity
of what is passing through it, but will give no indication of the coal/water
velocity differential. The differential would have to be detected by com-
paring the density of the material in a volumetric tank to that in the flow
line. If the latter is higher, it is an indication that the mean water
velocity is higher than the mean coal velocity. This potential error
source, then, must be handled through the calibration process.
Achievable accuracies for magnetic flowmeters are dependent on the
electronics package. Typical manufacturer's specifications are ± 1% of
reading over a 10 to 1 flow range or ± 0.5% of full scale, according to
the customer's desires. As is shown below, this does not translate into
1% accuracy on the coal flowrate that accuracy will not be as good due
to density measurement errors.
2 2.2.1.2 Ultrasonic Flowmeter (Reference 30-31)
The most basic type of ultrasonic flowmeter would be as shown sche-
matically in Figure 16. It consists of a source of high frequency pressure
pulses and a detector on either side of a pipe, typically at a 45° angle
with respect to the pipe axis. In the configuration shown in Figure 16,
the transit time of a discrete pulse is given as
a
r " (c + v ) cose (33)
51
-------
TRANSMITTER
PIPE
PULSE TRAIN
RECEIVER
FLOW INCREASES PULSE TRAIN SPEED IN A LINEAR MANNER,
AND IS SENSED BY DIFFERENCE IN PULSE TRANSIT TIME
FOR FLOWING AND STATIC CONDITIONS.
Figure 16. Ultrasonic flowmeter schematic
52
-------
where
i = acoustic path length, m
t = transit time, seconds
c = speed of sound in the slurry, m/s
v = mean slurry axial velocity along the line between the source
m and detector, m/s
9 = angle between the pipe axis and the line between the source
and detector, degrees
It will always be the case in coal/water slurries that c » v . Also,
the mean velocity vm along what is effectively a diameter, is not the
same as the average velocity in the pipe, U. These two parameters are
related in a known way for the case of fully developed pipe flow by a
function which is dependent only on the pipe Reynolds number:
" = f (%) vm (34)
where
f = functional relationship, dimensionless
Re = pipe Reynolds number
There are no devices on the market which use hardware corresponding to the
technique described by Equation (33). Changes in the sound speed, c, due
to concentration, composition, or temperature are typically of the same
order as the slurry velocity, which would lead to very poor accuracy.
Consequently, the normal hardware approach is to use two source/detector
pairs, with one signal propagating upstream and one propagating downstream.
Sinusoidal pressure pulse trains are then propagated in each direction,
and the instrument output becomes a frequency difference which is directly
proportional to v and is independent of c. This leaves us with three
primary error sources: the frequency difference, the path length, and the
relationship between v and u~. The last two sources should be minimized
by in-place calibration, while the first is dependent upon the basic
quality of the instrument itself and the slurry properties.
Like the magnetic flowmeter, the ultrasonic flowmeter is inherently a
velocity measuring device - it does not sense volumetric flow or mass flow.
Unlike the magnetic flowmeter, which integrates over the entire pipe cross-
section, the ultrasonic flowmeter integrates along a line, so that an
53
-------
in-place calibration is required in many situations. Traditionally, ultra-
sonic flowmeters have been limited by the solids concentration in a stream
the beam is attenuated too much by the solid particles. Another traditional
problem has been buildup of material on the sensor surfaces. The latter
problem has been largely eliminated through development of "clamp-on" instru-
ments which are totally non-intrinsive and require no facility modifications.
Some of the more recently developed instruments have also demonstrated capabil-
ities to make measurements in streams with 50% or more solids content, which
is typical for coal slurries. These recent developments have made ultrasonic
flowmeters much more viable instruments for slurry flow measurement.
2.2.2.1.3 Nuclear Magnetic Resonance Meter (Ref. 26,33)
A brief description will be given here of this instrument. More
detail can be found in Reference 9. The device is currently being offered
only for one inch diameter lines.
This technique requires that a major stream constituent be an element
(in either elemental or molecular form) with a nucleus having a magnetic
moment. Hydrogen is such an element, so the concept will apply to coal/
water slurries. The meter is shown schematically in Figure 17, adapted
from Reference 9. The apparatus consists of two magnets, a modulator,
resonator, detector, and electronics circuitry. The first magnet (usually
permanent) creates a nuclear magnetization in the hydrogen atoms, producing
a spin alignment. The slurry then passes into the field of the second
magnet, where a modulated signal is induced. This causes precession of
the nuclear magnetic dipoles in the flow. The receiver then picks up a
signal which is different in phase from the modulating signal. By virtue
of a control look, the phase difference is held at a specified constant
value by adjusting the frequency of the modulating signal. This ultimately
results in sensing of the flow velocity as
G = I w (35)
where
i = distance between centers of modulating and receiving coils, m
= phase difference between modulator and receiver signal, radians
w = modulator frequency, s"1
54
-------
en
Counter
readout
Detector
Resonator
1
se
rator
Frequency
controller
i
Modulator 1
HoM Magnet-S
~-P* Fluid flow
HQD Magnet-S
ttt&KUWWfiS,
SEE TEXT FOR EXPLANATION
Figure 17. Schematic of NMR flowmeter, taken from Reference 33
-------
Thus the output of the device is proportional to the average flow velocity,
so that the NMR meter is identical in this respect to the magnetic flow-
meter. One additional comment should perhaps be made explicit here - the
term "nuclear" in nuclear magnetic resonance is used because it is the
atomic nucleus which is magnetized as part of the sensing technique;
however, no nuclear reactions or radioctive substances are part of the
concept.
The main problem with commercialization of NMR has apparently been
cost - its accuracy capabilities appear to be on a level with those of
electromagnetic flowmeters. There appear to be no limitations on solid
content, as there are for the ultrasonic flowmeters. In summary, the
concept is valid, but at present there is hardware available only for very
small (1 inch diameter or less) line sizes.
2.2.2.2 Intrusive Flowmeters
Most liquid flow measurement techniques are of an intrusive nature.
These include orifices, nozzles, Venturis, turbine, target meters, vortex
shedding meter, etc. Any of these can be adapted with varying degrees of
success to coal/water slurry flow measurements. The extent to which intru-
sive techniques are tried is often a function of two things: the amount
of pressure put on an instrumentation engineer to cover up with something
that works, usually on a very small hardware budget, and/or the enthusiasm
and salesmanship of an instrument manufacturer's marketing group. Our
feeling is that the intrusive flowmeter most likely to succeed in coal/water
slurry applications are the vortex shedding meter, target meter, and venturi
meter. Abrasive degradation would be of concern for all three; in addition
there is concern about clogging of the pressure lines in the venturi meter
The only major reason we can see for selecting any of these rather than an
electromagnetic flowmeter for service in a dense coal/water slurry line
would be a lower purchase price. Initial hardware savings may be offset,
however, by higher maintenance costs for the intrusive device.
2.2.2.2.1 Vortex Shedding Meter
The vortex shedding meter is a relative newcomer in the flow measure-
ment field, but has reached a high state of development in a short period
of time. It is based on a simple, reliable fluid mechanics phenomenon.
56
-------
When a cylinder with a cross sectional height to width ratio not much
oreater than unity (e.g., circular or square cross section) is aligned
normal to a flowing stream, vortices will be shed periodically from either
side of the cylinder as indicated in Figure 18, so long as the Reynolds
number based on the cylinder height is high enough. For sufficiently high
Reynolds number, the shedding fequency becomes indicpendent of the Reynolds
number and the following relation is attained:
f = k % (36)
where
f = shedding frequency, s"
k = numerical constant, dimensionless
u = fluid velocity approaching cylinder, m/s
d = cylinder height, m
The shedding produces substantial pressure fluctuations, as was evidenced
by the complete collapse of a large suspension bridge due to wind induced
vortex shedding from the structure. In process stream flowmeters, the
shedding body spans the pipe, except in very large pipes where a smaller
point-type sensor may be used. The solid blockage imposed by the shedding
body is substantial, about 40%, to minimize nonlinearities due to changes
in the velocity profile with pipe Reynolds number, and to insure that
vortices of only one frequency are shed at any one time (too great a change
in velocity along the length of the shedding body would result in multiple
"cells" of vortices at different shedding frequencies).
In recent years, a number of novel ways of detecting the shedding
frequency have appeared. The ones most applicable to slurry flows are
thermal and solid state pressure detection methods which involve no moving
parts or flow through small passages.
Typical accuracy capability is ± 0.5% of reading above a known mini-
mum flow for a given pipe size. Erosion of the sensing mechanism at the
flow interface and, to a lesser extent, erosion of the shedding body
itself, are the main concerns in slurry applications. Also, there would
be a lower maximum allowable particle size for the vortex shedding meter
than for a non-intrusive device.
57
-------
PIPE
SHED VORTEX
SHEDDING
BODY
VORTICES ARE SHED ALTERNATELY FROM EACH SIDE OF THE
SHEDDING BODY. PRESSURE FLUCTUATIONS CAUSED BY THE
SHEDDING CAN BE SENSED BY FORCE MEASUREMENT, THERMAL,
OR ACOUSTIC MEANS. SHEDDING FREQUENCY IS A FUNCTION
OF FLOW VELOCITY.
Figure 18. Schematic of vortex shedding meter
58
-------
2.2.2.2.2 Target Meter
The target meter, shown in Figure 19, is a force measuring device.
A drag body is placed in the atream at the end of a lever arm. The sensor
output is a torque usually monitored by means of a strain gage bridge.
The operable relation is
T = £D = ilk CD \ p IP2 A (37)
where
T = sensed torque, N-m
a = leverl arm length, m
D = drag force on body and exposed arm, N
k = calibration factor (function of blockage), dimensionless
CD = body drag coefficient, dimensionless
p = stream density, kg/m3
A = drag body frontal area, m2
It may be noted that, unlike the other sensors thus far considered, the
target meter is not a velocity measurement device - its dependence on
density puts it more in the mass flow measurement category.
The drag meter is typically capable of 1% of reading accuracy if
a large enough drag body is used. Its only major deficiency in the
present application is erosion of the drag body with time.
2.2.2.2.3 Venturi Meter
This well known differential pressure device is shown in Figure 20.
It operates on the following basis:
AP = Pe - Pe = k \ P "*
where
Ap = sensed differential pressure, pascals
p = static pressure at exit (or entry) plane, pascals
p = static pressure at throat, pascals
L
k = calibration factor, dimensionless
The venturi meter is preferred over the orifice meter or nozzle because
it does not have recirculating regions. Accuracies to 1%
59
-------
PIPE
STRAIN GAGE
BRIDGE
LEVER ARM
TARGET DISC
TARGET DISC IS DEFLECTED BY ITS DRAG IN FLOW STREAM.
DRAG FORCE BECOMES A TORQUE VIA THE LEVER ARM, AND TORQUE
IS SENSED BY STRAIN GAGE BRIDGE. TORQUE IS A FUNCTION
OF MASS FLOW IN THE PIPE.
Figure 19. Schematic of target meter
60
-------
PIPE
HIGH
PRESSURE
LINE
LOW PRESSURE
LINE
DIFFERENTIAL PRESSURE
GAGE
VENTURI GEOMETRY PRODUCES A DIFFERENTIAL PRESSURE PROPORTIONAL
TO THE SQUARE OF THE AVERAGE FLOW VELOCITY
Figure 20. Schematic of venturi flowmeter
61
-------
of reading or better can be attained for a sufficient investment in the
pressure transducer. Clogging of the pressure taps and, to a lesser
extent, erosion of the venturi itself, are the major problems to be dealt
with in adapting a venturi meter to coal/water slurry applications.
2.2.2.3 Slurry Flowmeter/Density Meter Systems
As noted above, the most applicable slurry flowmeters must be used
in conjunction with a density meter to determine the coal/water ratio
and coal and water mass flow rates. This section presents a brief error
analysis to show what coal flowrate accuracies can be achieved as a
function of meter accuracy and coal content in the slurry. The results
of the analysis show that even though good accuracy can be obtained by
the meter, the system accuracy will not be nearly as good.
For the case where the average coal velocity in the slurry is equal
to the average water velocity in the slurry, the coal mass flow rate can
be derived from Equation (2) and expressed as follows:
psl " pw
"c = "P" PC " A (39>
where
m = coal flow rate, kg/s
F = average flow velocity, m/s
A = pipe cross-sectional area, m2
Psl = slurry density, kg/m3
p = water density, kg/m
w
,3
p = coal density, kg/m3
Standard error analysis techniques (Reference 34) were used to derive an
error equation for mc of the following form (assuming no error in the
density of water):
V
mc
(40)
-------
where
an= standard deviation of parameter n,
in same units as parameter n.
Equation (40) presents the system error, am , as a function of the indi-
vidual error sources. Furthermore, normalizing the error terms in the
form a /n allows for easy computation in terms of percent errors rather
than absolute ones. A main point in this analysis is that the terms
sl and 1 (especially the former) are greater than unity for
Psl"pH20 PsTpH20
coal/water slurries. This amplifies slurry density and coal density
measurement errors. It is reasonable in terms of existing technology
to say that we can achieve
aa2 aA2 *
+ -7 = .01
u A
It is reasonable to take
PH 0 = 1.000 g/cm3
p £ 1.35 g/cm3
C
1 - psl - pc
These numbers were substituted into Equation (40), using a range of values
for p -i, and a- /m was calculated for
s c
- = .001, .01, .02, .05
psl
Results shown in Figure 21 as a function of percent coal by volume in the
slurry. As the coal content drops to zero, the error becomes infinite
psl
once the term - becomes infinite. Figure 21 illustrates that when
5 psl"pH20
63
-------
ACCURACY OF COAL MASS FLOW RATE, %
0
20 40 60
% COAL BY VOLUME IN SLURRY
80
100
Figure 21. Accuracy of coal mass flow measurement in a coal/
water slurry for various accuracies of slurry
density, coal density, and volumetric flow rate
as a function of coal concentration in the slurry
64
-------
individual parameters are measured to a given accuracy, say 1%, in this
system, the system accuracy will be much poorer than 1%. It is clear
that for this particular process measurement, a system error analysis
should be rigorously carried out to determine the error in the real
parameter of interest, which is the coal flow rate measurement. Such an
analysis may have a strong impact on the instruments selected and on
operating conditions such as slurry density.
65
-------
III. HARDWARE STATUS
The previous section dealt with the physics of instrument operation.
This section is concerned with the current state of development of the
instruments. This ranges from readily available equipment offered 1n
slightly varying forms by a number of(manufacturers, such as weigh belt
devices and electromagnetic flowmeters, to instruments such as sulfur or
ash monitors which will not be commercially available for several years.
The information in this section was obtained from the technical literature
vendors, users, development organizations, and regulating agencies. In
the case of commonly available devices such as weigh belt Instruments, we
did not do an exhaustive vendor survey, but stopped after reaching the con-
clusion that high accuracy equipment can be obtained from a number of vendors
Where applicable, a list of vendors for a specific type of instrument 1s
given. Such lists are acknowledged not to be all inclusive, and do not
constitute recommendations for a listed vendor's product, since evaluations
needed to make such a recommendation are outside the scope of the task.
Vendor data should be Interpreted as an indication of instrument avallabllit
and general accuracy capabilities. A list of vendors mentioned 1n this
section and their addresses is presented at the end of the section.
3.1 COAL ANALYSIS INSTRUMENTS
3.1.1 Neutron Activation (Ref. 35-40, 10-12)
The most promising elemental analysis technique for process control
is thermal neutron capture gamma ray analysis. Because the neutrons are
thermal it is possible to probe a substantial volume. Furthermore, be-
cause the emitted radiation consists of fairly energetic gamma rays, the
emitted signal also can propagate through substantial thicknesses of coal
Thus neutron activation allows one to sample a large volume of material
simultaneously.
The basic work in this area was performed by Stewart and Hall at the
Morgantown Energy Research Center, and was directed toward development of
a sulfur meter capable of handling a 7 to 10 ton per hour slipstream.
The pioneering efforts at MERC involved two relatively recent state
of the art advances. The heart of the technique is a high flux neutron
66
-------
source. The availability of man-made Californium 252 which fissions
spontaneously makes it practical to achieve high neutron fluxes with
relative safety. Furthermore, state of the art electronics developed at
MERC make it possible to analyze over 300,000 gamma rays per second thus
obtaining usable statistics in time constants of order 2 minutes. The
work at Morgantown was mothballed in June, 1978, due to manpower shortages
and a reordering of technical priorities (as opposed to any problems with
the instrument). Good results had been obtained up to the time work was
stopped.
Similar instruments are presently being developed by commercial
contractors. One is a sulfur and ash meter being developed for the Homer
City coal cleaning plant by MDH Industries, Inc. The MDH monitor has tar-
get accuracies of .05% for sulfur and .25% ash. As in the MERC device, a
californium source is used for neutron production. The unit is to be
delivered at the end of 1979, and will process approximately 100 TPH of
cleaned coal. Unit costs for instruments in the near future will be about
$150,000 to $200,000, and expected costs for diverting a slipstream to the
instrument from a given process line, and then returning it, are $100,000
to $500,000. The device will weigh approximately two to three tons, with
the bulk of that being shielding fo the radioactive source. MDH is accept-
ing orders for the device at this time.
The most advanced system under current development is the continuous
Nuclear Analyzer of Coal (CONAC) system devised by Science Applications,
Inc. (SAI), in Palo Alto, California. As shown schematically in Figure 22,
CONAC is a complete coal handling and analysis system consisting of a belt
for flow control, a nuclear weigh scale for total mass measurement, and a
neutron activation device for composition measurement. The detection
scheme allows for measurement of all the elements listed in Table 8. Oxy-
gen is determined by the difference between total mass as measured by the
nuclear scale and the total mass of the other designated elements. Since
the meter cannot distinguish chemical forms, a separate measurement of
moisture is used. Electromagnetic or nuclear magnetic resonance approaches
are the present leading candidates for the moisture measurement. The
CONAC approach as it is being presently pursued seems to work best with
about a 30 cm thickness of coal on the belt. This will allow for monitoring
67
-------
1. Feed Surge Bin
2.A
00
Figure 22. Conceptual design of Continuous Nuclear Analyzer of Coal (CONAC)
system Including a nuclear scale (Ref. 39)
-------
Table 8. COAL ELEMENTS ANALYZED BY CONAC SYSTEM
AND THEIR CONCENTRATIONS IN TYPICAL COAL
SAMPLES (Reference 39).
Element
Carbon
Hydrogen
Sulfur
Nitrogen
Silicon
Aluminum
Iron
Titanium
Sodium
Potassium
Calcium
Chlorine
Concentration, %
Pittsburgh #8 Coal
72.0
5.3
2.9
1.4
1.8
1.3
1.3
.030
.026
.11
.096
.079
Wyoming Coal
58.5
5.5
.40
.90
1.0
.68
.24
.030
.050
.018
1.2
.01
Of an entire coal stream in most applications, rather than just a slip-
stream. This is an important feature in and of itself, since it elimi-
nates inaccuracies involved in obtaining a representative coal sample
for analysis. Most of the work at SAI to date has been sponsored by the
Electric Power Research Institute (EPRI). A "full-fledged" system is
scheduled to be installed for EPRI at an as yet undetermined site in
1981. Under a separate program, a sulfur and hydrogen monitor is to be
delivered to Detroit Edison in 1980.
69
-------
3.1.2 Gamma Ray Density Gages (Ref. 13-15)
The use of gamma ray density gages is already standard practice in
a variety of industries. Their most straightforward application is just
as level sensors used to determine when a container is full or empty.
More sophisticated applications include sensors used to determine bulk
density variations and slurry sensors used to measure slurry composition.
The precise design employed is usually strongly application dependent.
Typically either Cs 137 or Co 60 radioactive isotopes are used to produce
the required gamma rays. Because nuclear density gages depend upon exponen-
tial attenuation, the size of the probed volume is critical. For example
for coal and 1.25 Mev gamma rays the attenuation coefficient is roughly
0.05 cm" . Thus if a 6 foot volume is to be probed the attenuation factor
is equal to:
e6 x 12 x 2.54 x 0.05 = fi9 = 8 x 1Q3
Evidently because the thickness appears in the exponential a factor of two
difference can be highly significant. Consider that in the above example
doubling the thickness decreases the transmitter signal by a factor of
o
8 x 10. Notice that merely increasing the source size is not necessarily
a viable approach to recover this loss because of the increased health
hazard.
Available commercial units incorporate a number of desirable features
Among the more prominent are source decay compensation, temperature com-
pensation and linearization circuitry. These features automatically cor-
rect for such effects as decay of the radioactive isotope source and the
intrinsic non-linearity of the exponential attenuation.
Table 9 summarizes some specifications for a representative device.
It should be appreciated that the best way to calibrate such a device
is in the actual application. By doing this systematic absorptions caused
by pipe walls or belt thicknesses can be readily compensated for. Further-
more, irregular volume distribution effects and the presence of hydrogen
atoms can also be partially compensated for in this manner.
70
-------
Table 9. DENSITY GAGE DATA
Size (I.D.):
Nominal accuracy:
Ambient temperature
range, °C:
Response time:
Nominal cost,
incl. indicator:
Vendors contacted:
2.5 cm to > 46 cm
±.5% to ± 1% of range, where
a nominal range might be
specific gravity 1.1 to 1.3
-30 to +60
15-240 seconds
$3600 for 5 cm-46 cm I.D.
Texas Nuclear Division of
Ramsey Engineering;
Kay-Ray
71
-------
3.1.3 Neutron Moisture Meter (Ref. 16-19)
Neutron moisture determination in coal has been demonstrated by work
at the Bureau of Mines and commercial instruments are now available from
several vendors. It should be noted however as observed earlier that so-
called moisture meters are in fact indicators of hydrogen content so that
the moisture determination is only as good as a priori knowledge of the
hydrogen content of the coal.
A typical system is described in Table 10. This system probes for a
depth of approximately 50 cm into the coke being measured for moisture
content. The required high energy neutrons are produced by an Americium/
Beryllium source and detected by ionization chambers. A steel backed
ceramic waveplate supplied by the vendor is fitted as an integral part
of the coke hopper. In order to determine water on a percentage basis
the vendor also recommends the inclusion of a density gage to normalize
out bulk density variations.
Table 10. MOISTURE METER DATA
Size:
Nominal Accuracy:
Range:
Vendor Contacted:
Probes 50 cm radius hemisphere
±0.5$ moisture when density
corrections are available
0-20% moisture
Kay-Ray
72
-------
3.1.4 Chemical Elemental Analysis Techniques (Ref. 1)
All commercial analysis techniques suffer from the necessity to obtain
representative samples. For this reason our discussion of these approaches
w-jll be quite brief because they do not appear promixing in process control
applications. Details of wet chemistry measurements are outside the scope
of the current discussion and hence are omitted. The interested reader can
find detailed discussions in Volume 26 of the ASTM Standards.
Commercial instruments which automatically perform most of the required
chemistry are available. Tables 11 and 12 illustrate the characteristics
of two representative instruments. The Perkin-Elmer Model 240 performs an
elemental analysis by first combusting the sample and then utilizing thermal
conductivity detectors in conjunction with different types of traps which
selectively absorb gases. Note in particular the very small sample size
and the slow response time of the instrument. The LECO sulfur analyzer also
employs combustion, but infers sulfur via IR absorption of SOp. Note its
more favorable sample size and response time. However, it still requires a
representative sample and thus is of limited potential utility for on-line
control.
Table 11. LECO IR-33 SULFUR ANALYZER
Sample Size
Range
Accuracy
Response Time
Vendor Contact
Cost
0.2-0.7 g
0.01 to 3% S
± 3% S content
2 minutes
LECO Corporation
$17,000
Table 12. PERKIN-ELMER MODEL 240 ELEMENTAL ANALYZER
Sample Size: 1-3 mg
Range: C, H. N simultaneously + conversion to S
Precision: Better than 0.3% for C, H, N or 0
Response Time: 13-20 minutes
Vendor Contact: Perkin-Elmer
Cost: $12-20,000, dependent upon options
73
-------
3.1.5 X-Ray Fluorescence (Ref. 20-24)
Both large scale and portable X-ray fluorescence spectrometers are
available commercially. The primary limitations of these devices are as
follows:
1. Insensitivity to low atomic number elements.
2. Small sampling volume due to attenuation by absorption.
3. Sensitivity to sample matrix composition.
Inspite of these problems X-ray fluorescence devices are currently being
used in process control applications in favorable cases such as the cement
industry. Furthermore, at least one company merchandises a unit especial!
designed to measure coal ash content.
Table 13 illustrates the features of a large spectrometer system.
To achieve the stated high degree of accuracy it is necessary to have a
highly homogeneous, finely ground sample. It is also necessary to provide
some sort of matrix correction approach to eliminate systematic problems
Both of these features are probably incompatible with coal process control
applications unless the coal is prepared in a highly homogeneous manner.
Such preparation would almost certainly take too much time to serve as
on-line diagnostic.
Table 14 describes a much simpler system marketed by Texas Nuclear
Corporation. This device utilizes a combination measurement of backscatte
and fluorescence radiation to infer ash content. Although the backscatter
radiation can penetrate a depth of order 2 cm, the fluorescence radiation
has a penetration depth less than 1 mm. Thus if a matrix composition cor-
rection is necessary very fine powers are required, particularly for Ca
and Ti. The device is designed for use only with pulverized coal and the
quoted specifications assume that the unit is precalibrated by a sample
of the type of coal to be measured. The accuracy of the measurement 1s
considerably reduced if different types of coal are analyzed. If on the
other hand uniform types of coal are analyzed, then accuracies as good as
± 0.2% ash are possible.
74
-------
Table 13. X-RAY ANALYZER CHARACTERIZATION
Analyzing Range: F to U
Nominal Accuracy: better than 1% with suitable standard
Sample Requirements: 1-3/4" dia. by 1" thick maximum
Response Time: typically 40 sec for 8 elements
Approximate Cost: $100,000
Vendor Contact: Applied Research Laboratory, Phillips
Diano Corporation
Table 14. PORTABLE ELEMENTAL ANALYZER
FOR ASH IN COAL
Nominal Accuracy*: + 10% relative
Range: 5 to 25% ash
Response Time: 60 seconds
Sample Requirements: 100 mesh
Vendor Contact: Texas Nuclear Division of
Ramsey Engineering
*Specific to precalibration of a particular coal type
75
-------
3.2 FLOW MEASUREMENT INSTRUMENTS
3.2.1 Dry Coal Flow
3.2.1.1 Weigh Belt Techniques
A partial list of suppliers of weigh belt equipment is given in Table 15
The size of the list indicates the easy availability of this type of equip-
ment, as well as its widespread use. A sampling of literature from manufac-
turers in Table 15was used to generate the specifications shown in Table 16
These specifications are typical of several manufacturers, although not all
vendors cover the full range of belt widths. Belt scales tend to fall into
one of three categories:
t High accuracy (± 0.1% to ± 0.25% F.S. accuracy)
electromechanical scales
t Medium accuracy (± 0.5% to ± 1%) electromechanical
scales
Nuclear scales (± 1%)
Typical hardware costs for these three types of systems are shown in Table 17
along with a relative ranking of operating and maintenance costs. The main
differences between the high and medium accuracy electromechanical scales
are number and accuracy of load cells (up to four high accuracy cells for
the high accuracy systems vs. one for the medium accuracy system) and
number of idler rollers (usually four for the high accuracy system vs. one
or two for the medium accuracy system). The nuclear scale, of course, is
a totally different concept. While the two electromechanical systems are
almost identical in terms of interfaces with the conveyor belt system
the nuclear scale has no physical interface except for the relatively
insignificant belt speed sensor. It will generally be the case that a
nuclear scale will be easiest to use, and, due to its simplicity, will tend
to be preferred as long as its accuracy is compatible with system require-
ments. While a 1% accuracy would be acceptable for most process control
requirements, it is typically not acceptable in true commerce applications
where the scale is metering the coal for sales accounting. In these
applications, ± 0.25% accuracy is a typical requirement.
76
-------
Table 15. PARTIAL LIST OF WEIGH BELT VENDORS
(FROM REF. 50)
Type
General
Vendor
Aggregates Equipment Inc.
ASEA Inc.
*Auto Weigh Inc.
Cardinal Scale Mfg. Co.
Fairbanks Weighing Div., Colt Industries
Fairfield Engineering Co.
Howe Richardson Scale Co.
*Inflo Resometric Scale Inc.
Jeffrey Mfg. Div., Dresser Industries Inc.
KHD Industrieanlagen AG, Humboldt Wedag
Kilo-Wate Inc.
K-Tron Corp.
Lively Mfg. & Equipment Co.
*Merrick Scale Manufacturing Co., Inc.
* Ramsey Engineering Co.
Revere Corp. of America, Sub. of Neptune Intl. Corp.
Rexnord Inc.
Rexnord Inc., Process Machinery Div.
*Thayer Scale Hyer Industries
Thurman Scale Co., Div. Thurman Mfg. Co.
Webb Jarvis B., Co.
Wilson, R. M., Co.
Nuclear
*Kay-Ray Inc.
*0hmart Corp.
*Ramsey Engineering Co.
literature received and evaluated.
See Table 26 for addresses
77
-------
Table 16. TYPICAL WEIGH BELT SPECIFICATIONS
Belt Width
Turndown*
Accuracy, % reading
% F.S.
Ambient Temperature
Range, °C
Mechanical or Electromechanical
High Accuracy
Medium Accuracy
Nuclear
30 cm - 300 cm
± .1 - 2.5
> 3/1
+ .5 - 1
± 1
<_ 0 to >_ 40
Table 17. TYPICAL WEIGH BELT HARDWARE COSTS
Type
Cost for 1000 ton/hour capacity unit
Electromechanical
High Accuracy
Medium Accuracy
Nuclear
$10,000
$ 5,000
$ 6,000
Installation Costs:
High Accuracy
Medium Accuracy
Nuclear
Highest
Next Highest
Lowest
78
-------
Weigh scales are sensors which are often used in control systems.
A detailed discussion of control systems is beyond the scope of this report,
but a few comments are in order here. The basic function of the system
will be to provide specified flow rates of coal as a function of time.
As noted in Section 2.2, there are two basic ways of adjusting the flow
rate in a conveyor belt system -belt speed and belt loading. As is
emphasized in the Appendix, optimum weigh scale accuracy is achieved for
a constant belt loading, preferably about 80% of capacity. Also, system
time response is minimized, especially for long belts, when the belt speed
is the primary variable. Thus a system with variable belt speed and con-
stant loading is preferable from several standpoints. A variety of control
systems is discussed in References 26, 41, and 42. Both single belt systems
and multiple belt blending systems are discussed. General discussions of
belt scales and comments on their usage in specific situations are given
in References 43-49.
The widespread use of belt scales for commerce has (surprise!) led to
widespread regulatory agencies' monitoring of their use. We have been in
contact with the Southern Weighing and Inspection Bureau, an acknowledged
expert agency in this area, regarding procedures for calibration and use
of weighing scales. These procedures are presented in annotated form in
the Appendix . The procedures show that, in comparison with other process
nstruments of comparable accuracy, weigh scales require little attention
nd calibration when properly used.
The conclusions on weigh belt devices, in the context of this report,
are as follows:
Weigh scales are simple, reliable, accurate, and not notably
expensive.
Nuclear scales are easiest to install and use, and will probably
be least expensive over a period of years. They are not how-
ever, as accurate as the best electromechanical scales.
In a system involving analysis intruments to obtain constituent
(e.g., carbon, sulfur, ash) flow rates, the total flow error
from a weigh scale will have a negligible effect on system
error.
79
-------
3.2.1.2 Miscellaneous Techniques
Due to the wide applicability and availability of weigh belt devices
we did not spend a great deal of time investigating availability of other
dry coal flow devices. Two devices are briefly noted here, and both are
metering devices with sufficient stability and repeatability to serve as
measuring devices as well.
The first is a line of screw feeders offered by Vibra Screw Incorporated
These devices operate as discussed in Section 2.2.1.2. Capacities range from
about 0.05 to about 2,000 tons per hour. These devices have claimed accura-
cies of l%-2% of reading. On another project, we have recently calibrated
a small veeder manufactured by K-Tron, and found a repeatability of about
± 5% of reading. At first glance, the K-Tron feeder tested appeared to
have a design which would lead to similar, if not better, performance than
the Vibra Screw feeders. It is quite possible that increased physical size
may lead to higher accuracy. These feeders are designed to attach directly
to the outlet of a storage hopper. To the best of our knowledge, K-Tron
feeders are only manufactured for relatively low feed rates, and cannot
cover the range Vibra Screw does.
The second device is the Merrick Scale Manufacturing Company "Flow Sta "
feeder, which also operates at the outlet of a hopper. Maximum flow rate 1
about 25 tons per hour. The device may be described as a "metering turbi "
feeder (as opposed to a passive sensing turbine) in which a rotating blade
causes material to be removed from the bottom of the bin.
Both devices are volumetric feeders which would require recalibratio
in the event of coal density or particle size distribution changes. Nelth
would be as accurate as a good weigh belt system.
3.2.2 Coal/Water Slurries
3.2.2.1 Obstruction!ess Flowmeters
3.2.2.1.1 Magnetic Flowmeter
A list of magnetic flowmeter manufacturers is given in Table 18. Th
magnetic flowmeter is not as widely used in the coal industry as are wei
80
-------
Table 18. MAGNETIC FLOW METER VENDORS
*Brooks Instrument
Fielding Crossman & Assoc., Ltd.
*Fischer & Porter Co.
Foxboro Co.
Hoke Controls Ltd.
*Honeywell Int.
Kent Cambridge Ltd.
Ontor Ltd.
Taylor Instrument
*Warren Automatic Tool Co.
*literature received and evaluated
see Table 26 for addresses
81
-------
belt devices, but demand for the instruments has led to a rather sophisticated
state of development. Typical instrument specifications are listed in Table
19. These do not vary a great deal from vendor to vendor. Materials of con-
struction are listed in Table 20, and can be supplied by a majority of vendors.
For typical coal/water applications, meter dimensions are given in Table 21
along with unit costs which basically vary linearly with pipe size for pipes
over 20 cm (8 in.) diameter.
A typical magnetic flowmeter system consists of three components - the
sensing unit, a signal converter, and a recorder and/or controller. It Is
generally not a problem to locate the converter up to hundreds of meters from
the sensing units. A typical magnetic flowmeter for coal/water slurry appij.
cations would have the following characteristics:
Pipe size: £46 cm (18 in.) diameter
Maximum sensor dimensions: 60-cm long x 75 cm x 75 cm
Minimum sensor dimensions (up to 2.5-cm diameter pipe):
35-cm long x 25 cm x 35 cm
Materials (see Table 20)
Body: 304 Stainless Steel
Wetted surfaces: Liner - Rubber or Teflon
Electrodes - 316 Stainless Steel or a
more exotic material in
unusual applications
t Maximum flow velocity: 9 m/s (30 ft/s)
Preferred maximum velocity to minimize abrasion: 2 m/s (6 ft/s)
Calibration: Sensor unit and signal converter are calibrated
independently. Converter calibration devices are available
The sensor must be calibrated with actual flow, preferably
using typical slurry composition dumping into volumetric
tank of adequate accuracy.
Achievable accuracy: ±0.5% full scale or +_ 1% of reading over
10-to-l range.
Power usage: 30-200 watts depending on size.
t Optional ultrasonic electrode cleaner to avoid buildup of sludge
or grease on electrodes.
82
-------
Table 19. TYPICAL MAGNETIC FLOWMETER SPECIFICATIONS
Size (I.D.)
Accuracy
% Reading
% F.S.
Power Required
Max Temperature, °C
Max Velocity, m/s
Min Velocity, m/s
.25 cm - 120 cm
+_ 1, over 10/1 range
± -5
~ 30 watts for .25 cm I.D.
~ 100 watts for 15 cm I.D.
^ 200 watts for 46 cm I.D.
65 (polyurethane lines)
150 (teflon lines)
9
.6
Table 20. COMMON MAGNETIC FLOWMETER LINER AND
ELECTRODE MATERIALS
Liner
Electrode
Teflon
Polyurethane
Ceramic
Fiberglas
Neoprene
Misc. rubber compounds
Stainless Steel, 304 or 316
Hastelloy
Platinum
Tantalurn
Titanium
83
-------
Table 21. TYPICAL MAGNETIC FLOWMETER DIMENSIONS
AND COSTS
Pipe
cm
5.1
10.2
20.3
46
Size
in
2
4
8
18
Dimensions
L,cm
39
42
44
67
W,cm
30
42
42
68
H,cm
34
46
46
72
Post*
$3500
$3800
$4500
$12,000
*includes meter, converter, and recorder
84
-------
In some literature reports, such as Reference 51, performance criticisms
are lodged against the magnetic flowmeter. As the same reference shows,
the bulk of these ojbections (e.g. "the pipe is not always full"; "the
meter body is poorly grounded") are a result of poor communication between
vendor and user or inattention to detail on the part of the user. It is our
conclusion from reviewing both the literature and vendor data that problems
which would be applicable to coal/water slurries have been properly addressed
and overcome. A potential user should be able to purchase an adequate
device, as long as stated accuracies are acceptable, from one of a number of
different vendors. One item to keep in mind when purchasing a magnetic
flowmeter is that there is a preferred flow velocity range for the device,
which may result in selection of a meter with a different I.D. than that of
the pipe to which it will be attached. If a smaller size can be selected,
a cost savings will be realized.
For coal/water slurry applications, it will be the case that use of a
magnetic flowmeter will probably result in lower operating/maintenance
costs than will the use of any other flow device of comparable accuracy,
due to the relatively noncritical wetted pieces and the obstructionless
nature of the device. It is the approach which we most strongly recommend
for slurry flow measurements. It must, of course, always be recognized that
the magnetic flowmeter inherently measures an average flow velocity of the
material passing through it, so additional diagnostic instrumentation,
usually in the form of a density gage, will be required to determine the mass
flow of coal in a coal/water slurry.
3.2.2.1.2 Ultrasonic Flowmeters
The "traditional" ultrasonic flowmeter shown in Figure 16 works best
in a homogeneous, particle free stream. When the solid loading becomes
high, the particles scatter and attenuate the pressure signals. As the
loading increases, eventually a point is reached such that the signal is
lost and the device will not work. A representative of the Badger Meter
Manufacturing Company related during a telephone conversation that the
upper limit they have found on solids content for their instrument is 25%.
This is well below a typical coal slurry solids content of about 48%.
85
-------
The other important limitation on traditional ultrasonic flowmeters 1s
that they are line averaging devices which take readings along a pipe diameter.
This average velocity must be related analytically, or preferably, empirically
to the average flow velocity. If the relationships between the two changes
as it almost always will, as a function of flowrate and/or solids content
then the output will be nonlinear with flowrate and extensive clibration
will be required. This is in contrast to the more linear magnetic flowmeter
which would generally require only a single point calibration.
Presently available ultrasonic flowmeters are of two types, as distin-
guished by the transducer/fluid interface. The more "traditional" instru-
ments have the transmitters and receivers in direct contact with the fluid
while the "clamp-on" devices, so-called because they can be clamped onto
an existing pipe, send and receive signals through the pipe wall. To
obtain parallel faces between the transmitter and receiver, recesses in
the pipe wall are used (always in the case of wetted transducers and some-
times in the case of clamp-on transducers) Clamp-on transducers are
especially attractive, as long as accuracy can be maintained, since they
require no modifications to the flow line for installation or removal.
They also eliminate the problem of coating of the sensors which often
occurs in the case of wetted sensors.
Two relatively new clamp-on flowmeters appear capable of handling the
heavy solids loading found in coal slurries. One is a system which trans-
mits a signal in one direction across the pipe, and is manufactured by
Controlotron Corporation. In appearance it is similar to the device offered
by Badger, but has been found to be acceptable in slurry lines with up to
at least 50% solids content. The other is a device manufactured by
Polysonics, and makes use of a single transducer clamped to one side of
a pipe. In a coal slurry application, the transmitted signal would be
reflected off the coal particles and back to the sensor. The particle
velocity is then determined by a Doppler shift. This instrument has also
been found acceptable for use in dense slurries. It will work best when
the flow field is uniform with the particles moving at the fluid velocity.
Since these conditions are not present in real life, in-place calibration
as a function of flow rate, particle size, and solids loading should be
performed for optimum accuracy.
86
-------
Nominal manufacturer specifications are shown in Table 22, along with
representative costs. It is usually the case that clamp-on devices will be
preferred, since they are less expensive and require minimum modifications
to an existing piping system. Clamp-on transducers perform best on a steel
pipe, which is what would commonly be used in coal/water slurry systems.
For larger pipe sizes, 30 cm (12 in.) or greater, the ultrasonic flowmeter
Wi11 tend to be less expensive than the magnetic flowmeter in terms of pur-
chase price and installation cost. The newer ultrasonic flowmeters have not
been available long enough to permit long term comparisons with magnetic
flowmeters. They may eventually prove to be at least as acceptable as mag-
netic flowmeters.
3.2.2.1.3 Nuclear Magnetic Resonance Meter
The Badger Meter Manufacturing Company holds the patent on the Nuclear
Magnetic Resonances flow meter. At present, the largest pipe size for which
they offer a device is 2.5 cm (1 in.). There are at present hardware diffi-
culties relating to the size of magnet coils which make larger sizes imprac-
tical. The listed accuracy of the device is 1% of reading, and a typical
system with accessories costs about $5500. Due to the size limitation,
this device in its current state of development is not applicable to coal/
water slurry measurements.
3 2.2.2 Intrusive Flowmeters
The survey-type articles reviewed (References 26, 29, 51 and 52) are
unanimous in not recommending intrusive flowmeters for this type of appli-
cation. The devices considered below are generally less expensive to pur-
chase than magnetic flowmeters, but it is likely that the purchase price
savings would be more than offset by maintenance costs. In addition, there
is greater probability of data loss due to instrument failure as well as
the threat of possible line blockage or erratic flow due to the presence
of the meter in the stream. We recommend that intrusive devices by considered
for coal/water slurries only in the event that obstruct!'onless devices such
as the magnetic flowmeter are clearly unacceptable for some reason. Presented
below are discussions of various devices in order of our estimation of their
probable success.
87
-------
Table 22. ULTRASONIC FLOWMETER DATA
Size (I.D.)
Accuracy
% Reading
Temperature Range,
°C
Nominal system cost
Vendors contacted
2.5 cm - 50 cm
± 1 above 30 cm/s
(±.1 for special order equipment)*
'v 0 to 50
$2000-$6000 for 15 cm - 50 cm pipes
Badger Meter Mfg. Co.
Controlotron Corp.
Panametrics
Polysonics
*Panametrics offers custom systems in this range
but has not worked with coal/water slurries.
See Table 26 for vendor addresses.
-------
3.2.2.2.1 Target Meter
The blockage of this type of device is primarily in the center of the
pipe (Figure 19) rather than at the walls, so accumulation problems would
be minimized. There are three main potential problems with use of target
flowmeters in coal/water slurries in addition to the problem of possible
pipe clogging. These are:
0 Loss of accuracy due to abrasion of the drag body.
9 Mechanical deformation due to large particle impact.
0 Clogging of flowmeter components such as the overrange mechanism.
Abrasion will change the drag characteristics, as would any deformation.
The drag body would need periodic replacement, which will show up as a labor
cost and a minor hardware cost. A typical overrange mechanism is shown in
Figure 23, and is a device to prevent damage to the strain gage bridge/lever
assembly. Coal particles could lodge in this mechanism and render the instru-
ment useless. Removing the overrange mechanism leaves the instrument more
susceptible to major damage.
Nominal cost and specification data are listed in Table 23. Target
meters work well in clean liquid streams and in clean or dirty gas streams.
Our own evaluation in a gas stream on another EPA program (Reference 51)
confirms a 1% accuracy capability. There are no truly fundamental reasons
whv the target meter cannot be used in slurry lines, but they have yet to
be recommended, especially for dense slurries, by unbiased sources.
3 2-2.2.2 Vortex Shedding Meter
Potential problems with this approach are possible accumulation of coal
near the relatively large shedding body, leading to inaccuracies, and
abrasion of the shedding body, in addition to possible pipe clogging. As
. the case of the target meter, the intrusive body would require periodic
replacement.
89
-------
OVERRANGE
MECHANISM
LEVER ARM
MECHANISM PREVENTS
EXCESSIVE LEVER ARM
DEFLECTION. COAL
PARTICLES MAY LODGE IN
ANNULUS OF MECHANISM
ANNULUS
Figure 23. Detail of overrange mechanism for target meter
90
-------
Table 23. TARGET METER DATA
Size (I.D.)
Accuracy
% Reading
Power required
Max Temperature, °C
Max Flow Velocity, m/s
Min Flow Velocity, m/s
Approximate system cost,
including indicator
Vendors contacted
1.9 cm to > 50 cm
± .5 to 2 over 10/1 range
of order 10 watts
177
400 (optional)
x, 4.5
* .1
-v $2700 for 30 cm dia
Ramapo Instrument Co.
Engineering Measurement Co.
See Table 26 for vendor addresses.
91
-------
Costs and specifications are shown in Table 24. In their current state
of development, vortex shedding meters are reliable and accurate in clean
flows, but there is insufficient data available to determine their capabilities
in dense slurries.
3.2.2.2.3 Venturi Meters
Some manufacturers of venturi meters do not recommend their use for
anything worse than "reasonably clean liquid" applications, while others
are more adventurous and feel they can handle slurry flows. Venturi meters
have a better probability of success than orifice plates or nozzles because
they do not cause recirculating regions to be set up. The overriding
concern about use of a venturi meter, of course, will be clogging of the
relatively small pressure taps. This would require use of an automatic
purge system to maintain clear lines. The duty cycle of the purge system
would likely have to be high (20% or more) so that truly continuous measure-
ment would not be obtained. Some users have obtained good results with an
added purge system.
Costs and specifications for typical venturi meter installations are
shown in Table 25. A purge system would be an additional cost item, and
would be the key to a successful system. Since most manufacturers do not
generally recommend venturi meters for this type of application, great care
must be taken if this approach is selected.
3.2.2.3 Density Gages
Two types of density gages are in common use in slurry flow systems -
nuclear (gamma ray) gages, as discussed in Section 3.1, and ultrasonic gages.
An ultrasonic gage is shown schematically in Figure 24. Like the ultra-
sonic flowmeter, it has a transmitter and receiver of high frequency pulses.
Received signal strength is the measured variable, and is inversely propor-
tional to the solids content in the stream. Eventually a point is reached
where signal attenuation is too great and no usable signal is received.
This upper limit is a function of both solids content and pipe diameter.
92
-------
Table 24. VORTEX METER DATA
Size (I.D.)
Accuracy, % Reading
Max Temperature, °C
Min. Flow Requirement
Approximate System
Cost, including
Indicator
Vendor Contacted
2.5 cm to 274 cm
+ .25 to + 1 (10/1 range)
120
Pipe Reynolds number >_ 10,000
$4700 (5 cm I.D.)
$5700 (10 cm I.D.)
$6600 (20 cm I.D.)
$13,000 (46 cm I.D.)
Neptune/Eastech
See Table 26 for vendor addresses.
93
-------
Table 25. VENTURI METER DATA
Size (I.D.)
Accuracy of Ap
Measurement,
% F.S.
Approximate system
Cost, including
Indicator but not
Purge System
Vendors Contacted
.6 cm to 150 cm
+ .25
$2000 (10 cm I.D.)
$2200 (20 cm I.D.)
$3200 (46 cm I.D.)
Fischer & Porter
Flow-Dyne Engineering
See Table 26 for vendor addresses.
94
-------
PIPE
TRANSMITTER
FLOW .
PULSE TRAIN
RECEIVER
STRENGTH OF RECEIVED SIGNAL FOR PURE LIQUID ESTABLISHES
REFERENCE. SIGNAL IS ATTENUATED BY SOLID PARTICLES
IN FLOWING (OR STATIC) STREAM. CALIBRATION ESTABLISHES
RELATION BETWEEN % SOLIDS AND SIGNAL STRENGTH.
Figure 24. Ultrasonic density gage
95
-------
For the device manufactured by National Sonics, the product of solids
content in percent (by volume) and pipe I.D. in inches must be less than
or equal to 60. This means that the dense coal/water slurries typically
encountered would limit pipe size to 2.5-5 cm (1-2 inches), meaning that
the device would be applicable primarily in situations as depicted in
Figure 25, where a small diameter slipstream is taken from the main line
and the sensor mounted on the small slipstream line. Such a configuration
would have to be carefully designed to avoid obtaining a non-representative
sample in the slipstream. Except in the case of very dilute slurries or
very small lines, the ultrasonic density gage will not be applicable to
coal/water slurries.
3.2.2.4 Slurry Flowmeters/Density Gage Systems
A typical system involving use of a density gage and flowmeter is
shown in Figure 26, taken from a Texas Nuclear catalogue. The example
shown would be for a case where the desired parameters are mass flow rate
and stream density. For the case of coal mass flow as an output, a micro-
processor or other logic device would be needed to compute coal mass flow
as in Equation (39). In light of discussion above of various instruments
it is reasonable to expect a slurry volumetric flow accuracy of ± 1% of
reading, and a slurry density measurement accuracy of ± .1% to ± .2% of
reading. If we arbitrarily assume that the accuracy of the coal density
measurement is the same as the slurry density measurement, and that the
coal content of the slurry is 20% to 60% by volume, then the accuracy of
the coal mass flow measurement will be between about 1.8% and 4%. This
is a significantly lower accuracy than would tend to be inferred from the
individual component accuracies, and clearly shows why flow rate measure-
ments of a dry coal flow is the preferred technique when possible.
96
-------
LARGE PIPE
ULTRASONIC DENSITY GAGE REQUIRES SMALL PIPE FOR
DENSE SLURRY. SLIPSTREAM APPROACH PROVIDES SMALL
DIAMETER, REASONABLY REPRESENTATIVE SLIPSTREAM
Figure 25. Slipstream concept for ultrasonic density gage
97
-------
DENSITY
GAUGE
vo
00
FLOW
METER
TEXAS NUCLEAR
MASS FLOW
TRANSMITTER
FLOW
TRANSMITTER
DENSITY
RECORDER/
CONTROLLER
EXTERNAL
TOTALIZER
(TONS)
MASS FLOW
RATE
RECORDER
(TPH)
Figure 26. Typical flowmeter/density gage combination, taken
from Texas Nuclear catalogue
-------
Table 26. VENDOR ADDRESSES
Weigh Belts - General
Aggregates Equipment Inc.
9 Horseshoe Rd.
Leola, Pa 17540
ASEA Inc.
4 New King St.
White Plains, NY 10604
Auto Weigh Inc.
P.O. Box 4017
1439 N. Emerald Ave.
Modesto, Cal 95352
Cardinal Scale Mfg. Co.
203 E. Daugherty
Webb City, Mo 64870
Fairbanks Weighing Div., Colt Industries
711 E. St. Johnsbury Rd.
St. Johnsbury, Vt 05819
Fairfield Engineering Co.
324 Barnhart St.
Marion, Ohio 43302
Howe Richardson Scale Co.
680 Van Houten Ave.
Clifton, NJ 07015
Inflo Resometric Scale Inc.
2324 University Ave.
St. Paul, Minn 55114
Jeffrey Mfg. Div., Dresser Industries, Inc.
912 No. Fourth St.
Columbus, Ohio 43216
KHD Industrieanlagen AG, Humboldt Wedag
Weirsbergstrasse, D5
Koeln 91, Fed. Rup. of Germany
Kilo-Wate Inc.
Box 798
Georgetown, Tex 78626
99
-------
Weigh Belts - General (Cont.)
K-Tron Corp.
P.O. Box 548
Glassboro, NJ 08028
Lively Mfg. and Equipment Co.
P.O. Box 338
Glen White, W. Va 25849
Merrick Scale Manufacturing Co., Inc.
180-192 Autumn Street
Passaic, N.J. 07055
Ramsey Engineering Co.
1853 W. County Rd. C
St. Paul, Minn 55113
Revere Corp. of America, Sub. of Neptune Intl. Corp.
North Colony Rd.
Wallingford, Conn 06492
Rexnord Inc.
P.O. Box 2022
Milwaukee, Wis 53201
Rexnord Inc., Process Machinery Div.
P.O. Box 383
Milwaukee, Wis 53201
Thayer Scale Hyer Industries
Rt. 139
Pembroke, Mass 02359
Thurman Scale Co., Div Thurman Mfg. Co.
1939 Refugee Rd.
Columbus, Ohio 43215
Jervis B. Webb Co.
9000 Alpine Ave.
Detroit, Mich 48204
R. M. Wilson Co.
Box 6274
Wheeling, W. Va 26003
100
-------
Weigh Belts - Nuclear
Kay-Ray Inc.
516 W. Campus Dr.
Arlington Heights, 111 60004
Ohmart Corp.
4241 Allendorf Drive
P.O. Box 9026
Cincinnati, Ohio 45209
Ramsey Engineering Co.
1853 W. County Rd. C
St. Paul, Minn 55113
Magnetic Flowmeters
Brooks Instrument Div., Emerson Electric Co.
Hatfield, Pa 19440
Fielding Crossman and Assoc., Ltd.
232 Yorkland Blvd.
Willowdale, Ont. Canada
Fischer and Porter Co.
War-minster, Pa 18974
Foxboro Co.
Foxboro, Mass 02035
Hoke Controls Ltd.
2240 Speers Road
Oakville, Ont., Canada
Honeywell Int.
1100 Virginia Drive
Fort Washington, Pa 19034
Kent Cambridge Ltd.
80 Doncaster Ave.
Thornhill, Ont., Canada
Ontor Ltd.
12 Leswyn Rd.
Toronto, Ont., Canada
Taylor Instrument, Process Control Div., Sybron Corp.
95 Ames St.
Rochester, NY 14601
Warren Automatic Tool Co.
P.O. Box 18345
Houston, Tex 77023
101
-------
Ultrasonic Flowmeters
Badger Meter Mfg. Co., Precision Product Div.
6116 E. 15th St.
Tulsa, Okla 74115
Controlotron Corporation
155 Plant Avenue
Hauppauge, L.I., New York 11787
National Sonics
250 Marcus Boulevard
Hauppauge, New York 11787
Polysonics
3230 Mercer Street
Houston, Texas 77027
Target Meters
Engineering Measurements Company, Inc.
P.O. Box 346
1840 N. 55th St.
Boulder, Colo 80302
Ramapo Instrument Co., Inc.
2 Mars Court
P.O. Box 428
Montville, NJ 07045
Vortex Meters
Neptune/Eastech
308 Talmadge Rd.
Edison, NJ 08817
Panametrics
220-T Cresent St.
Waltham, Mass 02154
Venturi Meters
Fischer and Porter Co.
Warminster, Pa 18974
Flow-Dyne Engineering, Inc.
P.O. Box 9034
Fort Worth, Tex 76107
102
-------
Density Gages
National Sonics
250 Marcus Blvd.
Hauppauge, NY 11787
Texas Nuclear
9101 Research Rd.
P.O. Box 9267
Austin, Tex 78757
(See also Kay-Ray Inc., listed in Weigh Belts-Nuclear section.)
Coal Analysis Instruments
Applied Research Laboratories
P.O. Box 129
Sunland, CA 91040
DIANO Corporation
9 Commonwealth Ave.
Woburn, Mass. 01801
LECO Corporation
3000 Lakeview Ave.
St. Joseph, Mich. 49085
MHD Industries, Inc.
426 West Duarte Road
Monrovia, CA 91016
Perkin-Elmer Corporation, Instrument Div.
Main Avenue
Norwalk, Conn. 06852
Philips Electronic Instruments, Inc.
85 McKee Drive
Mahwah, NJ 07430
Science Applications, Inc.
5 Palo Alto Square
Suite 200
Palo Alto, CA 94304
Watkins Johnson
3333 Hileview Ave.
Palo Alto, CA 94304
(See also Kay-Ray Inc., listed in Weigh Belts-Nuclear section.)
103
-------
IV. TEN YEAR PROJECTION OF INSTRUMENT
DEVELOPMENT AND AVAILABILITY
This section consists of a list of predictions for instrumentation
improvements over the next decade. The reason for presenting these predic-
tions is that we expect that coal analysis instruments, specifically of
the neutron activation type, will become commercially available during this
time period, and will potentially have a large effect on overall instrumen-
tation and control systems.
4.1 WEIGH BELT INSTRUMENTS
Electromechanical weigh belt devices have pretty well reached their
limits of accuracy, reliability, and simplicity. No significant future
improvements are considered likely. Nuclear weigh scales are cost com-
petitive with the "medium" accuracy weigh belt, but cannot meet the
accuracies required by some regulatory agencies. While the nuclear scale
is insensitive to belt/weighbridge interface factors which can lead to
errors in electromechanical systems, it is more sensitive to coal composi-
tion variations. It appears, then, that electromechanical scales are
ultimately limited by systematic errors primarily related to the belt,
while the nuclear scale is limited by random errors due to coal property
variations. Two basic paths are available for nuclear scale development:
improvement of accuracy to become competitive with high accuracy electro-
mechanical scales, and lowering of cost to stay financially more attractive
than medium accuracy electromechanical scales. There is insufficient
information at present to determine if the high accuracy goal can be
achieved.
4.2 SLURRY INSTRUMENTS
At present, there is not a great demand for coal/water slurry instru-
mentation. If slurry pipelines become more common (the critical issue
being right-of-way conflicts with railroads), demand may increase. Like
weigh belt devices, the electromagnetic flowmeter has pretty well reached
a maximum state of development. Recent improvements have been in the
104
-------
materials area, to decrease maintenance, and in installation details to
try to make operation as foolproof as possible. A more widespread need
for coal/water slurry flowmeters may provide incentive to manufacturers
of devices other than magnetic flowmeters (particularly intrusive devices)
to improve their instruments and/or adequately demonstrate that they will
work well in dense slurries. In large pipes, where the cost of magnetic
flowmeters is high, ultrasonic flowmeters would be significantly less
expensive. It is likely that nuclear density gages will maintain their
superiority in dense slurries, while ultrasonic density gages will be
more accurate in very dilute slurries. New developments in slurry
instrumentation would be spurred by increased demand, or by success in
a similar application.
4.3 COAL ANALYSIS INSTRUMENTS
The problems associated with obtaining a representative coal sample
(typically a few grams from hundreds or thousands of tons) for instruments
using techniques such as X-ray fluorescence are very difficult and expen-
sive to overcome, when they can be overcome at all. That is a major reason
why techniques such as thermal neutron activation, which can look at a
major portion, or even all, of the coal stream, are so attractive and hold
the greatest promise for future development. The concept has been adequately
demonstrated. MDH and SAI are currently working on deliverable hardware
systems. The early units will concentrate on sulfur, hydrogen, and ash.
further development should lead to a system capable of monitoring all
important coal constituents. As the eighties progress, the technology
should expand from one of a kind devices with rather limited capabilities
to general availability of complete monitoring systems, which, like the
CONAC system being developed at SAI, will include total mass measurement
and flow control in addition to the basic analysis instrumentation.
4.4 INTEGRATED INSTRUMENTATION SYSTEMS
The impact of increasing high accuracy instrument availability and the
proliferation of low cost microprocessors is offering the possibility of a
revolution in process instrumentation and control. For example, continuous
knowledge of the flowrate and heating value of coal being fed to a furnace
105
-------
would allow for the possibility of continuous control of combustion param-
eters such as temperature and air/fuel ratio to optimize the combustion
process. Monitoring of sulfur content of the incoming coal would allow
for optimum control of a clean-up device such as a wet scrubber. It 1s
being shown routinely that proper instrumentation pays for itself, and
more, very rapidly through increased energy efficiency. With the antici-
pated availability of good analysis devices, the prospects for better
energy utilization and optimum environmental quality control become even
better.
106
-------
V. SUMMARY AND CONCLUSIONS
5.1 PRINCIPLES OF INSTRUMENT OPERATION
Relevant techniques were presented in detail to give a general under-
standing and to give an idea of fundamental limitations. For example, a
difference in ultimate utility between neutron activation devices and
fluorescence radiation devices is shown by the fact that the former can
operate with a penetration depth of order one meter, while the latter is
limited to penetration depth of order one millimeter. This has strong
implications on instrument configuration and utility.
In contrast to the rather sophisticated physics of coal analysis
equipment, operation principles of most flow measurement devices are
fairly simple. This simplicity has meant that flow measurement devices
have been more easily developed, and have been in wide use for many years.
The material presented makes it clear that there are many techniques
available for the measurements of interest. Availability of instrumenta-
tion, however, is another matter, especially in the case of coal analysis
devices.
5.2 HARDWARE STATUS
Instruments available now for coal analysis, with the exception of
the neutron moisture meter, are capable of analyzing only small samples,
which means that a sophisticated sampling system is required. The mois-
ture meter itself is actually a hydrogen indicator, which means that the
coal hydrogen content must be known independently to determine the
jnoisture content. It has been the observation of MERC personnel that
the hydrogen content in coal is in fact constant for a given coal type.
Although the MERC program has been halted, instrument development is pro-
ceeding at MDH and SAI, with delivery of the first MDH sulfur and ash
monitor expected by the end of 1979. Neutron activation devices are of
particular interest because they hold the promise of sampling entire coal
streams, thus eliminating the major problem of obtaining a representative
sample.
107
-------
Weigh belt devices, either electromechanical or nuclear, and electro-
magnetic flowmeters (typically coupled with nuclear density gages) are the
most used devices for dry coal flow measurement and coal/water slurry flow
measurement, respectively, and will be acceptable for most situations.
Weigh belt technology is especially highly developed, in part due to
regulatory considerations. Intrusive devices may be applicable to slurry
flow measurements, but this has not been clearly demonstrated as yet.
Recent developments in ultrasonic flowmeters have made them much more
applicable to slurry flow measurement.
5.3 EPA/IERL AND COAL MONITORING INSTRUMENTS
Weigh belt devices are sufficiently advanced, and their use in many
situations sufficiently monitored by various regulatory agencies, that thev
should not need EPA attention. Similarly, electromagnetic flowmeters are
a very standard item and do not appear to need attention, in terms of field
evaluations for example, except in unusual circumstances. As far as other
slurry flow measurement devices are concerned, we feel that a reasonable
approach is to let vendors demonstrate that their devices are acceptable
for (dense) coal/water slurries.
At such time as devices like the neutron sulfur meter become com-
merically available, IERL will likely wish to be involved in collaborative
testing or other evaluation of the devices. This, however, will not be
taking place for at least a couple of years.
Perhaps the most reasonable area for IERL investigation in the near
future is instrumentation/control systems which make use of the kinds of
instruments discussed in this report. The thrust of the investigation
would be to determine how to make the best uses of available and soon-to-
be-available instrumentation to control the operation of sources such as
coal-fired utilities to minimize pollutant emissions through direct com-
bustion control and/or optimizing operation of devices such as wet scrub-
bers. The greatest ultimate value of the instruments discussed herein 1s
not just as monitoring sensors, but as components in an active control
system, either on sources of direct coal combustion, or in coal cleaning
plants.
108
-------
VI. REFERENCES
1. Annual Book of ASTM Standards. Part 26 Gaseous Fuels; Coal and Coke.
American Society for Testing and Materials, 1975.
2 Fuels and Fuel Technology. Volume 1, Wilfred Francis, Pergamon Press
Ltd., 1965
3. Coal Coke and Coal Chemicals. P. Wilson and J. Wells, McGraw-Hill,
4. Nuclear Physics, I. Kaplan, Addi son-Wesley Publishing Co., Inc., 1962
5. American Institute of Physics Handbook, D. Gray Editor, McGraw-Hill
Book Company, 1972.
6. Nuclei and Particles. E. Segre, W. A. Benjamin Inc., 1965.
7. Activation Analysis Handbook, R. Koch, Academic Press, 1960.
8. Activation Analysis Handbook, Miloslav, Rakovic, CRC Press, 1970.
9 Instrumentation in Applied Nuclear Chemistry. J. Krugers, Plenum,
___ _
10. Neutron Activation Analysis. Chemical Analysis Volume 34, De Soete,
Wiley Interscience, 1972.
11. "Analytical Sensitivities and Energies of Thermal Neutron-Capture
Gamma Rays," D. Duffey, A. Elkady and F. Senftle, Nuclear Instruments
and Methods 80, 1970, pp. 149-171.
12. "Nuclear Meter for Monitoring the Sulfur Content of Coal Streams,"
R. Stewart, A. Hall, J. Martin, W. Farrior and A. Poston, Bureau of
Mines Advancing Energy Utilization Program Technical Progress Report
74, Jan. 1974.
13 "Gamma-Ray Absorption Coefficients," C. Davisson and R. Evans,
Reviews of Modern Physics. Vol. 24, No. 2, 1952, pp. 79-107.
14. "Take a look at nuclear gauges," B. Burton, Instrumentation and
Control Systems, December 1976, pp. 41-48.
-15 Nuclear Gages for Monitoring the Coal bed of Commercial -Scale Pres-
surized Gas Producers," G. Friggens and A. Hall, U.S. Department of
the Interior Report of Investigations 7793, 1973.
15. "Continuous Monitoring of Coal by a Neutron Moisture Meter," A. Hall,
J. Konchesky and R. Stewart, U.S. Department of the Interior Report
of Investigations 7807.
109
-------
REFERENCES (cont.)
17. "Plant Tests of a Neutron Moisture Meter for Coal," R. Stewart,
A. Hall, J. Konchesky, Mining Congress Journal, November 1970
pp. 44-48.
18. "Testing for Moisture Content in Foods by Neutron Gaging," S. Helf
Practical Applications of Neutron Radiography and Gaging. American*
Society for Testing and Materials, 1976, pp. 277-291.
19. "Measurement of Consistency of Pulpwood-Water Slurry Based on
Neutron Slowing-Down and Diffusion," J. Hewitt and V. Slobodian,
Applications of Neutron Radiography and Gaging, American Society
for Testing and Materials, 1976, pp. 292-302. *
20.
X-Ray Spectrochemical Analysis. L. Birks, Interscience Publishers Tnr
1959. c"
21. "X-Ray Spectrochemical Analysis of Materials Cement and Dental Alloys "
B. Bean and B. Mulligan, Application of Advanced and Nuclear Phvsir<- '
to Testing Materials, American Society for Testing and Material's!
22. Physical Chemistry, E. Moelwyn-Hughes, Pergamon Press, 1961,
pp. 218-222.
23. Applications of Low Energy X and y Rays. Editor, R. Ziegler.
24. "Trace Elements in Whole Coal Determined by X-Ray Fluorescence "
J. Kuhn, Norelco Reporter. Vol. 20, No. 3, 1973, pp. 7-10. '
25. "Electromagnetic Flowmetering . . . basics, products, and Apnlicatinnc «
Instrumentation Technology. March 1974, pp. 29-36. «nons,
26. Instrument Engineers Handbook, Volume 1, Process Measurement Bel a R
Liptak, Editor, Chilton Book Company, 1969.
27. "Electromagnetic Flowmeter Primary Elements," V. P. Head, Transaction*
of the ASME - Journal of Basic Engineering. December 1959, pp. 660-665"
28. "An Investigation of Electromagnetic Flowmeters," H. G. El rod and
R. R. Fouse, Transactions of the ASME. May, 1952, pp. 589-594.
29. "Selecting the Right Flowmeter, Part I: The Six Favorites," D J
Lomas, Instrumentation Technology. Vol. 24, No. 5, May 1977, pp. 55.53
30. "Ultrasonic Flowmeter Basics," J. L. McShane, Instrumentation
July 1971, pp. 44-48. '
31. "Clamp-on Ultrasonic Flowmeters . . . Limitations and Remedies,"
L. C. Lynnworth, Instrumentation Technology. September, 1975, pp> 37.44
32. "How the Coal Slurry Pipeline in Arizona Is Working," Environmental
Science and Technology, Vol. 10, No. 12, November 1976, pp. 1086-
TOST!
110
-------
REFERENCES (cont.)
33. "NMR Applied to Flow Measurement," W. K. Genthe, et al, Instrumentation
Technology, November 1968, pp. 53-58.
34 The Analysis of Physical Measurements. E. Pugh and G. Winslow,
Addison-Wesley, 1966.
35. "Monitoring the Sulfur Content of Coal Streams by Thermal-Neutron-
Capture Gamma-Ray Analysis," J. Martin and A. Hall, Morgantown Energy
Research Center Report RI-76/4. July 1976.
36 "Fast Pulse Processing for Gamma-Ray Spectroscopy," G. Fasching,
G. Patton, U.S. Bureau of Mines Report of Investigations 7909, 1974.
37 "On-Stream Measurement of Sulfur in Coal, Californium 252 Application
Note," Californium-252 Information Center, Savannah River Laboratory,
1973.
38. "Nuclear Measurement of Carbon in Fly Ash," R. Stewart, W. Farrior,
U.S. Bureau of Mines Report of Investigations.
*Q "Nuclear Assay of Coal" (Volume 1-11), T. Gozani, et al, EPRI FP-989,
January, 1979.
"Advanced Techniques and Instrumentation for Real Time On-Line and
Laboratory Analysis of Coal," T. Gozani, et al, presented at 1979 DOE-
ANL Symposium on Instrumentation and Control for Fossil Energy Processes
August 20-29, 1979, Denver, Colorado.
41 . "Feeders and Conveyor Scales as Control Elements," W. Harris, ISA FCE
736461, 1973 (obtainable from Ramsey Engineering Co.).
.2 "Control Systems for Belt Feeders," L. D. McEvoy, Instrumentation
Technology, February 1968, pp. 41-45.
43. "Belt Scales - White Elephants or Invaluable Measuring Instruments,"
. " Bruce McCarthy, Australian Mining. October 1976, pp. 26-27.
44. "Planning and Specifying a Conveyor Scale System," C. Marinelli,
Plant Engineering, April 29, 1976, pp. 237-239.
"A User's View of Conveyor Weigh Scales," R. J. Bailey, Canadian
Mining Journal. June, 1975, pp. 24-25.
"Costs of Viscosity, Weight, Analytical Instruments," Bela G. Liptak,
Chemical Engineering. Sept. 21, 1970, pp. 175-179.
"Adjustable-Speed Feeder Driver," A. C. Lordi, Instrumentation
Technology. January 1969, pp. 46-50.
Ill
-------
REFERENCES (cont.)
48. "Weigh Belt Scales - Nuclear Approach," G. Bauer, Canadian Mining
Journal, June, 1975, pp. 25-26.
49. "Belt Scale Design Considerations," Transactions of the Society of
Mining Engineers, AIME, Vol. 252, December, 1972, pp. 425-432.
50. "Coal Age Buying Directory," Coal Age, September, 1976, pp. 240-286.
51. "Where We Stand In Slurry Flow Measurements," G. M. Behrend,
Transactions of the Canadian Institute of Mining, Vol. 76, 1973
pp. 166-171.
52. "Selecting the Right Flowmeter, Part II: Comparing Candidates,"
D. J. Lomas, Instrumentation Technology, Vol. 24, No. 6, June 1977
pp. 71-77. *
53. "Continuous Measurement of Total Gas Flowrate from Stationary
Sources," E. F. Brooks, et al, EPA-650/2-75-020, February, 1975.
112
-------
VII. GLOSSARY
SYMBOL
A
A
B
B
Be
c
c
CD
D
d
d
E
Eo'El
f
f
h
hv
I
I
K,k
Ka,K3,
KE
L
USAGE
area
atomic mass
magnetic flux density
area
binding energy
speed of sound
speed of light
drag coefficient
drag
pipe diameter
cylinder height
electromotive force
energy
functional relationship
frequency
Planck's constant
DIMENSIONS
m2
amu
volt-s/m2
m2
erg
m/s
m/s
N
N
m
volt
erg
_i
s
erg-s
erg
photon energy
gamma rays/100 neutrons
intensity photon/cm2
constant or calibration factor
energies erg
kinetic energy erg
thickness cm
113
-------
GLOSSARY (cont.)
SYMBOL
a
a
*B
m
mT
m.
m
m.
mo
N
Ne
No
n
pe
Pt
Ap
R
R
ReP
S
S
VS1
T
t
USAGE
length
belt speed
belt length
mass
total mass flowrate
coal flow rate
sensed mass of constituent i
mass flow rate of
constituent i
electron mass
number density
electron number density
Avagadro's number
particle density
entry static pressure
throat static pressure
differential pressure
radius
ratio
pipe Reynolds number
analytical sensitivity
flux
gamma ray signals
torque
time
DIMENSIONS
m
m/s
m
kg
kg/s
kg/s
kg
kg/s
9
scatterers/cm3
cm'3
particles/cm3
pascals
pascals
pascals
cm
#Y-rays-barn
loO neutrons
particle/cm2
N-m
s
114
-------
GLOSSARY (cent.)
SYMBOL
t
u
u
V
V.
m
X
Z
6
X
y
v
p
p,
m
USAGE
thickness
fluid velocity
average velocity
velocity
effective sampling volume
solids volume
control volume
liquid volume
volumetric flow
mean velocity along pipe
diameter
depth
atomic number
angle between instrument
and pipe axes
wavelength
absorption coefficient
frequency
logarithmic decrement of
energy
density
coal density
liquid density
material density
solid density
DIMENSIONS
cm
m/s
m/s
m/s
m3
m3
m3
m3
m3/s
m/s
cm
cm
cm"1
hertz
kg/m3
kg/m3
kg/m3
kg/m3
kg/m3
115
-------
GLOSSARY (cont.)
SYMBOL USAGE DIMENSIONS
p , slurry density kg/m3
p density of water kg/m3
w
a cross section cm2 or
barn = lO
a constant
a standard deviation of
parameter n
angle
4> phase difference
X.j mass fraction of constituent i
oj modulator frequency s"
116
-------
APPENDIX
PROCEDURE FOR WEIGH BELT
CALIBRATION AND USE
Of the types of instrument systems considered in this report, only
weigh belt systems are sufficiently advanced and identifiable as systems
to warrant recommendations of operating procedures. Continuous coal anal-
ysis instrumentation is not sufficiently developed, and there are at
present too few applications for coal/water slurries.
Our research has led to the conclusion that the Southern Weighing
and Inspection Bureau in Atlanta, Georgia has promulgated weigh belt
urocedures which provide maximum accuracy with a minimum of calibration
cost to the user. The Bureau has supplied us with their standard procedures,
After a quick review of the received procedures, we decided that the
accuracy requirements could probably be loosened a bit without adversely
-jmpacting overall system requirements in a system which included (less
accurate) coal analysis instrumentation. We then proceeded to look for
oossible shortcuts which could save calibration time while maintaining a
flow accuracy of about 1%. After a closer inspection and conversations
yiith the Bureau, we came to the conclusion that the procedures already were
designed to minimize calibration work, and that their requirements should
Ot be relaxed. This also indicates that the required accuracy levels can
he achieved by available hardware systems.
Presented below are verbatim copies of the information supplied by the
Bureau. Not all items, such as requirements to supply data to the Bureau,
re applicable to the present case. We have annotated on an item-by-item
basis* as felt necessary, by using an encircled number to the left of the
Bureau's item number, as in
Qj 2. All inspection requirements . . .
notes are presented by number after the Bureau's last report to avoid
The notation primarily deals with two topics - items which are
f a purely regulatory nature and directly involve the Bureau, and items
117
-------
would not apply if a nuclear weight detection instrument were used. Three
documents are presented: "Requirements for the Approval of Belt Conveyor
Scales," "An Explanation of Rules for the Inspection, Testing, and Cali-
bration of Belt Conveyor Scales," and "Chain Test". They are presented in
chronological order, and there is some redundancy. The first two reports
are broad in scope and offer a wealth of user data on installation and use
as well as calibration, and the second report offers a complete list of
definitions of terms associated with the instruments. The third document
deals solely with chain calibration checks for mechanical and electro-
mechanical scales.
All readers of this report must keep in mind that the annotation of
the Bureau's documents is our own, and does not reflect Southern Weighing
and Inspection Bureau policy.
118
-------
jcument 1
Page
SOUTHERN WEIGHING AND INSPECTION BUREAU
REQUIREMENTS
FOR THE APPROVAL OF
BELT CONVEYOR SCALES
CIRCULAR 9585-5
jSSUED: ISSUED BY:
JUNE1, 1971 C. E. Pike. Manager
Southern Weighing and Inspection Bureau
Suite 306 - Transportation Building
151 Hlis Street. N. E.
Atlanta, Georgia 30303
119
-------
CONTENTS
Belt Scale Requirements
Tolerance Values
Additional Requirements
Vanders Obligation
Tests - Types Defined
4-5 Tests - Procedures
120
-------
)cument
Page
BUREAU
REQUIREMENTS FOR BELT SCALES
1. The system must be designed and rated so that material flowing over the
scale will remain within 50% to 98%of scale capacity and must be adequate and constant.
2. All conveyors on which a belt scale is located must be rigid in design and
so constructed that it is free from vibration and is not subject to stress that would cause
any deflection and shall not exceed 1,000 feet in length from center of head pulley to
center of tail pulley.
3. All conveyors on which a belt scale is located must be equipped with a gravity
take-up that is of adequate weight so as not to allow slippage of the belt on the drive pulley.
With such weight properly weighted, the belt must remain in contact, with all troughing
weigh bridge idlers or sensing elements, at all times.
4. The conveyor on which the scale is located must be free from interference
from all other operations.
5. The system must be so designed that the complete contents of each individual
car may be guaranteed to pass over the scale.
6. Sufficient impact idlers must be provided in the conveyor under each infeed
so as not to cause a deflection of the belt during the time material is being introduced.
7. Incline of the scale conveyor belt must not be in excess of 18 to 20 degrees.
8. For maximum reliability, mechanical scales should be equipped with a multi-
idler suspension.
9. Adequate simulated load testing equipment, to be applied on the belt over the
weight sensing element, approximating 80% of the rated capacity of the weighing system,
is required. A suitable storage area must be provided so that such equipment does not
rust or deteriorate from abrasive or coating action. Note: See Standard Procedure for
Test of Belt Scales. (Rige 5 - No. 17)
10. A rate of flow indicator must be installed and correctly calibrated. The indi-
cator must also be equipped with a twenty-four hour disc or strip chart to serve as a per-
manent record, along with a high and low cut-off alarm system. This is to help prevent
under or overloading of the scale. The alarm should be operational at not lower than 35%
or greater than 98%of scale capacity. The type of alarm used (i.e. audio or visual) must
be determined by the merits of each individual application. Such system will not be out-
fitted with a switch to disconnect either the chart or alarm system. Any exceptions to
this rule will be at the discretion of the Bureau.
11. A fast-count totalizer must be installed and must register in units of 1/10 of
a ton, or as determined by this Bureau.
12. A ticket or tape printer (determined by each individual application) and
printing to the equivalent numerals as indicated in No. 11, must be installed. It is
essential that the printer mounting will guarantee vibration free operation.
13. Scale and instrumentation cabinets must be purged with clean dry air.
14. Where scales and component parts are subject to a rapid or extreme change
in temperature, heaters must be installed as suggested by manufacturer.
121
-------
Document 1
PageA
-2-
©
15. Scales located near large bodies of water, operations that require use of
quantities of water, or any area of high humidity, must have heaters installed in scales
and/or instrumentation as suggested by manufacturer.
16. Scales and especially instrumentation should be protected from the elements
by weather-proof structures. Control rooms that house instruments must be air purged
with air conditioning, fans with filters, heaters, etc. so that a dust free area of greater
density must be created on the inside than is outside. Such system is required to preservl
all electronic instruments.
17. Wind screens must be erected around the entire weighing element.
18. Adequate access to scale must be provided. This includes a walk-way, step,
etc. so that servicing can be accomplished with ease.
19. Sufficient guards must be erected around "live" scale parts located near
walk-ways so that persons will not touch such parts or deposit equipment in their vicinity
Such application will be at the discretion of this Bureau.
20. In some cases, "live" parts of scales should be painted in contrasting colors
so as to warn persons against touching them. The application of this Rule will be at the
discretion of this Bureau.
21. Hydraulic machinery, large motors, or any equipment which will cause ex-
cessive vibration or noise is prohibited from being placed in, or affixed to, the control
room.
22. All infeed gates, feeders, etc. must be positive in action and so designed
that material will flow freely through them when opened or placed in operation. They
must also be positive in their closing action so that leakage does not occur.
23. Scale housing and instrumentation must be securely locked.
TOLERANCE VALUES
1. Material Test Tolerance
All belt scales must be material tested and adjusted to within 0.25% with
repeatability. The spread between plus or minus figures shall not exceed 0.25%.
2. Maintenance Tolerance
All belt scales must maintain a 0.5% tolerance between material tests. The
spread between plus or minus figures shall not exceed 0.5%.
3. Zero Tolerance
All belt scales must maintain a zero balance of 0.01 % for 10 minutes before
use, and after sufficient warm-up of the belt.
4. Repeatability Tolerance
All belt conveyor scales using simulated load testing equipment must be tested
for a repeatability of not greater than 0.1% for at least 5 consecutive tests. The spread
between plus or minus figures shall not exceed 0.1%.
122
-------
Document 1
Page
-3-
ADDITIONAL REQUIREMENTS
1. A spare parts kit, as suggested by manufacturer, must be kept on hand at all
times, with replacement of any parts used as soon as possible.
Note: This requirement is not intended as a means to heavily stock the scale owner with
unnecessary and expensive parts, but is intended to be of an assistance. Such Kit is to
contain only parts that can be easily replaced or installed by the owner, and not an elec-
trical or scale technician.
2. A service contract, preferably with the scale manufacturer's service repre-
sentative, must be evidenced. Such contract must provide for at least one yearly over-
haul of scale.
3. At least two persons or more on location must be thoroughly familiar with the
scale and its operation.
4. Control over entire system must be exercised by the scale operator to insure
adequate loading of the weighing device.
5. Unauthorized personnel should be discouraged from loitering in the scale and
instrumentation areas.
6. Where test chains are used, adequate cables, bridles, hooks, etc. must be
provided to hold the chains in place on the belt as recommended by the scale manufacturer.
VENDER'S OBLIGATION
IT SHALL BE THE DIRECT OBLIGATION OF THE PERSON OR FIRM SUPPLYING
A RELTM3ONVEYOR SCALE TO THE USER TO CONTACT THE SOUTHERN WEIGHING
"AND INSPECTION BUREAU AS FAR IN ADVANCE OF THE TESTING DATE AS POSSIBLE.
YN QRDHR^O^EXPgpiTE SUCH TESTS. IT IS SUGGESTED THAT THE BUREAU'S GENERAL
nflPICEBlTCONTA'CTED AS SOON AS DEFINITE PLANS HAVE BEEN MADE FOR THE
TMSTATJjgfiON OFTHE SCALE. FURTHER. AFTER INSTALLATION. THE PERSON OR
PTRM SUPgYINGTHjE BELT-CONVEYOR SCALE MUST ATTEST (IN WRITING) TO THE
FACT THAT ALL THE AFOREMENTIONED RULES. REGULATIONS. AND CONDITIONS
pfAVE~BEEf? METrAfrD THAT THE SCALE IS WEIGHING WITHIN TOLERANCE. ONLY
foHKNTHirtJSER jjgg BEEN SO INFORMED AND A COPY OF THE VENDER'S LETTER ~
HAS BEhg'FURNISHED' TOJTHE SOUTHERN WEIGHING AND INSPECTION BUREAU SHALL
THE BUREAUjjEWflFIED THAT THE WEIGHING SYSTEM IS READY FOR TESTING.
TYPES OF TESTS DEFINED
Material Testing.
1. All belt conveyor scales must be material tested before acceptance. Three to
five such tests are required, with 10 to 15 car (or in some cases heavy truck) loads per
test. If the weighing system is designed for individual car weights, the weight of each in-
dividual car must be in tolerance. Results must be furnished this Bureau.
Maintenance Testing
T. Sll belToonvey°r scales must be maintained within a 0.5% tolerance between
material tests, by using a known weight load test. Such tests must be conducted on a
weekly basis, preferably as soon after actual use of the belt as possible. Results of
each test must be furnished this Bureau.
123
-------
-4-
Repeatability Test
3 . All belt conveyor scales must be tested for repeatability each time a
tenance Test is made and before Material Tests are run.
STANDARD PROCEDURE FOR TESTS OF
BELT SCALES
1. Scale and entire installation will be checked for conformity with the above
listed rules and regulations. The scale and all component parts must be completely
installed and in good operating condition before any tests will be made.
2. The belt length, belt speed, and weight of chains per foot must be accurately
measured so as to determine tons per foot of belt travel during chain tests.
it up.
3. The belt must be run at least twenty minutes before test begins in order to warn
At temperatures below 40 degrees, the warm up time should be of longer duration.
4. Following warm-up, belt must be run for at least ten minutes to determine zero
of scale. Daring this time, the zero reading should be constant.
5. After establishing that the scale holds its zero, ten circuits of the belt will be
run with the simulated load testing equipment in place on belt. In some instances, where
the belt is long, only three circuits of the belt will be required.
If test chains are used, they shall be held securely in place on the belt as the
scale manufacturer recommends.
If tests utilizing other than the above methods are contemplated, the approval of
the Bureau must be secured.
Note: When other than test chains are used, only idlers of the highest quality and
requiring lubrication daily, must be used. Such idlers must be installed on the weighing
element and for at least five idler spaces before, and for at least five idler spaces after
the weighing element. Further, each week correct idler level must be determined with a
string level. If any one, or more, of the above named idlers are out of level, the scale
must not be used until correction is made. Idlers with worn bearings must be replaced.
6. Five more simulated load tests must then be run, with results of each test
reading within 0.1% of the first test outlined in No. 5. This will establish repeatability.
7. If the scale has repeatability, fifteen loads will then be weighed over the scale,
at 50% to 98%of capacity. A load is construed as a rail carload or truck load of high
capacity. (Note: fa instances where procurement of rail cars proves difficult, ten loads
per test may be substituted.) This is a calibration test and may have to be made several
times in order to bring the weighing device within the acceptable tolerance of 0.25%.
8. After the scale has been properly calibrated, three separate material tests
all within 0.25% with material being loaded at 50% to 98%of capacity, must then be made.
Each test will consist of fifteen loads. (Here again ten loads per test may be substituted
in the event rail cars are difficult to obtain.)
9. Following three successful material tests, five simulated load tests must
again be made in order to establish a new calibration factor.
124
-------
Document 1
Page
-5-
10. On successful completion of all tests, the scale will again be checked for zero.
At mis time, it will also be established that a service contract has been consummated with
a qualified belt scale service organization, preferably the scale's manufacturer.
11. Material scale tests should be made once every 90 to 120 days, or as directed
by the Bureau.
^6^ 12. The material acceptance tolerance of 0.25%must be maintained for thirty
days, and at any time the scale is overhauled or material tested by an authorized scale
man.
13. A maintenance tolerance of 0.5% must be maintained by the user at all times
after the scale is accepted. This may be assured by conducting tests at least once a week,
as outlined in No. 9, using the newly established calibration factor.
14. The rail carrier should be contacted at least two weeks before material tests
are made to insure that sufficient cars are available.
15. Copies of weekly tests must be sent to the General Office of this Bureau. Also,
copies of the results of the scalemen's tests must be sent to the General Office of this
Bureau every time they are made.
16. Proof of weight for accepted and certified belt scales will be in the form of a
printed weight ticket. The rate of flow chart must be dated corresponding with cars in
loading or unloading sequence. The correct rate of flow over a belt scale is essential
for good weights.
All of the records outlined herein must be preserved together so that they may be
audited Under terms of the weight agreement.
17. fa instances where material testing can be satisfactorily performed weekly,
Simulated Load Tests are not required. Copies of weekly material tests, however, must
be submitted to the General Office of this Bureau.
125
-------
Document 2
Page 1
AN EXPLANATION OF
RULES FOR THE INSPECTION, TESTING, AND CALIBRATION
OF
BELT CONVEYOR SCALES
*******
Southern Weighing and Inspection Bureau
Suite 306 - Transportation Building
151 Ellis Street, N. E.
Atlanta, Georgia 30303
C. E. Pike, Minager
July 16, 1973
126
-------
INDEX
fage
Belt Conveyor Scale Rules and Regulations Appendix
Definition of Terms 8-10
Inspection of Weighing Facilities 1-4
Nfeterial Test 5
Other Steps to be Taken 7
Simulated Load Test - Electronic 7
Simulated Load Test - Nfcchanical 6
Vander's Obligation 4
127
-------
Document 2
Page 3
BELT CONVEYOR SCALES
There are four basic types of continuous weighers which are known as belt or
conveyor scales. These are: Ivfechanical, Electronic, Atomic, and Rieumatic.
Basically, all types except atomic operate on the principle of material passing
over a weighing element on a belt conveyor. By integrating the speed of the belt with
the scale indication, tons (or pounds) per an hour are produced. Atomic scales, on
the other hand, are actually measuring devices. They operate by placing a radio
active mass over the belt and a form of radiation counter under the belt. Usually a
foot (or more) of pre-weighed material is placed on the belt under the radio active
mass and the radiation counter counts the number of rays not absorbed by the material
This in turn is translated or "calibrated" into tons (or pounds). As with the other
types of scales, the speed of the belt is integrated with the weight indication to produce
tons (or pounds) per hour.
INSPECTIONS
Rate of Flow: By experience it has been found that belt conveyor scales only
produce reliable weights when the material flow is constant between 50% and 98% of
scale capacity. Valley (pits), slugs, and other interruptions in an otherwise even
flow can produce erratic weights, as can long tailing or dribbles from the feeders.
Stability of Scale: One of the basic laws of weighing machines is "plumb and
level"and belt conveyor scales are no exception. Although the conveyor belt might
be on an incline, the basic weighing system is plumb and level. Therefore, not only
should the conveyor be stable but adequate bracing should be maintained under the
scale to prevent vibration or movement.
Tension: One of the greatest forces that will affect belt conveyor scales is
tension. A scale belt cannot be too tight and by the same token the belt must not be
too loose. The very best way to insure even tension is with the use of a gravity
take-up. This is a weight suspended on a roller from the underside (return side)
of the belt. The weight must be free in its movements and not bind on guide rails
when they are used. Another type of gravity take-up usually used when limited space
is a problem is to mount the rear pulley of the belt on a movable carriage. A bridle
and cable are then affixed to the rear of the carriage, with a weight being affixed to
the other end of the cable. With the use of pulleys the weight can be hung from a wall
or other convenient place. In both cases only enough weight is used so that the belt
will not slip on the drive pulley while under load. If the belt sags between idlers
the idler spacing must be changed. Under no circumstances should excessive belt
sag - under load - be countered by increasing the gravity take-up-weight. This would
create too much tension and could cause massive errors.
Conveyors: All conveyors used in connection with belt conveyor scales should
be rigid in design and should be used for weighing only. Additionally, all material
must pass over the scale. Trippers, etc. must not be installed in the belt. Further
the slope of the conveyor should not exceed 18 to 20 degrees; otherwise, the material'
would tend to roll back down the belt causing some material to be weighed more than
once. Flited conveyor belts (belts with ridges molded onto working surface) should
not be used as calibration is extremely difficult.
128
-------
Testing Equipment: Almost all belt conveyor scales are installed with testing
equipment. Atomic scales employ a calibrated plate which absorbs a given (or known)
amount of radiation. Other types of scales are calibrated with anything from test
weights, logging chains, etc. to test chains. The only practical method found thus far
has been with test chains. ID reality, a test chain is not a chain but a series of rollers
connected together with links and calibrated to the nearest thousandth of a ton (or pound)
per foot. These so-called test chains provide the nearest simulation to a loaded belt
and reproduce not only weight but attempt to form the belt configuration under load.
Use of test weights, logging chains, etc. has not proved satisfactory and for this
reason are unacceptable.
Totalizer: All belt conveyor scales are equipped with a master totalizer which
shows the total tons (or pounds) weighed. Weight indication for test purposes should
only be taken from the master totalizer. Some belt conveyor scales are also outfitted
with a percentage rate of flow meter as part of the master totalizer instrumentation,
while others supply this unit as an accessory.
Alarms: Since almost all belt conveyor scale instrumentation is located in the
same area as other controls (control room), the operator's attention to the rate of
flow chart cannot be constant. Therefore a high and low cut-off alarm system must
be installed to attract the operator's attention. The low alarm should be operational
from 0 to approximately 35% of scale capacity; and the high alarm should operate at
approximately 98%of scale capacity. This is to alert the operator that the scale is
loaded to a point below its capacity to register weights correctly or when the scale is
at the point of being overloaded. It should be borne in mind that the type of alarm to
be used (audio or visual) depends on the noise level and presence of other warning
lights in the control room. Certainly, the type of alarm used must be such that it will
be immediately noticed by the operator. Additionally, it is advised that the alarm
system not be equipped with a cut-off switch or time delay device.
Instrumentation: Control rooms or areas should also receive close attention.
Since usual proof of weight is in the form of a printed weight ticket, the scale should
be equipped with a printer. Additionally, since it has been proved that belt conveyor
scales only produce accurate weights when loaded between 50% and 98% (to 100%) of
capacity, a rate-of-flow disc or strip recorder is also necessary. The printed ticket
and the recorded rate-of-flow chart then comprise proof of weight for belt conveyor
scales.
Air Rirging and Heaters: Dust and moisture are two problems that always affect
instrumentation; therefore, the control area is required to be air purged with air
conditioning and heating equipment (as the season demands). This is to create an area
of greater density inside the instrumentation area than is outside. By these means
dust is held to a minimum. By the same token, scales having cabinets must also be
air purged and equipped with thermostatically controlled strip heaters. (One or more
light bulbs placed in cabinets are not sufficient.)
Wind Screens: All conveyors not located in buildings on which a belt conveyor
scale is used must be outfitted with wind screens. Wind screens are merely walls (of
metal, wood, etc.) placed around the weight sensitive area (20 feet before and 20 feet
after the weight sensitive element, including the weight sensing element). Such
screening must also include a roof and doors at least at one end. The purpose is to
keep the effects of wind to a minimum. Since the scale and testing equipment should
also be protected from the elements, the wind screens serve a dual purpose.
129
-------
ive Scale farts: All "live" parts of the belt conveyor scale near walkways
or where they can be reached should be outfitted with guards so that such parts can't
be accidentally touched. Such parts should also be painted in a bright contrasting
color so that everyone will immediately recognize them for what they are. Additionally,
easy access must be made to all scale parts so that the scale can be observed while in
operation and servicing can be accomplished with minimum difficulty.
Location of Nfachinery: Large motors, machinery, or hydraulic equipment
should not be affixed to the scale conveyor or control house, as excessive vibration
and/or harmonics could cause malfunctions in the rate-of-flow chart, printer,
integrator, etc.
Infeeds: Infeeds to conveyor belts should receive special attention, as they could
become troublesome. Feeders, gates, etc. must be positive in action and must not
leak. Long dribbles before weighing commences or long tailings after the gates close
can only produce inaccurate weights. Additionally, the conveyor belt under the infeeds
must not sag excessively under impact of material being introduced. If this situation
exists, only the placement of more idlers (or impact idlers) in the affected area will
cure the problem.
Scale Housing: All belt conveyor scales utilizing cabinets over or beside the belt
must be securely locked to prevent tampering from unauthorized persons. Cabinets
located elsewhere should also be locked for the same reason. Keys should be in the
possession of the top official at the site, or only one other person he so designates.
Idlers and Rilleys: The sensitive area of weighing on this type of scale is
approximately twenty feet before and twenty feet after the weight sensing element, fa
this area, idler spacing must be exact and alignment must be true. An out-of-round
idler will introduce errors which cannot be factored out of the scale. Restrictions
must be placed on the use of training or trueing idlers, as they can introduce adverse
tension into the belt. Wing pulley use should be discouraged as too much vibration to
be compensated for is usually present.
Conveyor Belting and Splices: All splices should be of the vulcanized type-
however, one good metal splice can be used in most cases. The reason for this'pro-
vision is because most splices (other than vulcanized) tend to create heavy spots in
the belt, making calibration difficult. It should also be borne in mind that the higher
rated scales usually utilize heavy multi-ply belting. Since belting must be as flexible
as possible to achieve good weights, it is necessary to run the belt for a minimum of
twenty minutes empty before use.
Conveyor Length: Conveyors on which a belt conveyor scale is located must not
exceed 1,000 ft. in length, from center of head pulley to center of tail pulley. By the
same token, conveyors under 50 feet should receive the explicit approval of the scale
manufacturer before attempting to make tests. Long conveyors tend to make the zero
calibration very difficult, if not impossible. Short conveyors, unless specifically
designed for the job, can cause uneven tension, rapid idler bearing wear, etc.
Chutes and Surge Bins: The discharge chute and/or surge bin should be carefullv
inspected. Chutes that spew material out over the sides of cars are unacceptable. This
condition can usually be cured by placing metal guides in the chute trough so that the
stream is concentrated. On large operations (that load at better than 1500 tons per
hour), a surge bin must be placed at the discharge end of the belt. This is a collectuur
bin equipped with a gate and loading chute. With this equipment the loading can be
130
-------
form with no spills between cars. The best "rule of thumb" to determine size of
the bin is to: estimate the rate of flow per minute; times the minutes required to
move the rail cars 20 feet; plus 10%. Usually bin capacity is rated on approximately
three minutes running time. Therefore, a system rated at 3000 TPHwill be equipped
with a bin of 150 ton capacity.
Discharge: The discharge end of the conveyor belt should also receive close
attention"! Some types of material tend to cling to the belt or leave a residue. As
this material or residue builds up, the belt becomes heavier, thereby affecting the
zero calibration of the scale. Also, when this condition exists, it is usually found
that when the material or residue forms a thick enough layer, parts of it will "sluff"
off, thereby creating an erratic load situation on the belt, not to mention uneven
tension. ft is strongly suggested that a gravity type wiper be installed at the end of
the belt on the return side (underside),
Sinks, Pits, or Dumping Areas: When belt conveyor scales are used to weigh
inbound commodities, special attention must be given the device into which the
material is dumped from rail cars. The sink or pit must be so designed that the
complete contents of a rail car enter that area only and are not spilled outside the
intended sink or pit. Grating must contain holes of sufficient size to guarantee the
passage of the complete contents of the car. Sides of the pit must also slope
sufficiently so that all material flows to the feeder gate. When practical, the un-
loading area must be enclosed to prevent wind-blown loss and detrimental effects
from the forces of nature.
Scale Location: Location of the scale on the conveyor belt is a problem for
the scale supplier and system owner; however, it is well to know where a scale will
produce the best weights. As a "rule-of-thumb", a belt conveyor scale will work
better as near to the in-feed as possible, ft should be remembered that this will be
at least twenty feet, unless the scale company has made specific provisions to move
the weight sensing element closer. Also, a scale must not be placed in a concave or
convex curve of the conveyor. This is to insure complete contact of the belt at all
times with the weight sensing element and to prevent excessive tension in the weighing
area.
(7) VENDER'S OBLIGATION
IT SHALL BE THE DIRECT OBLIGATION OF THE PERSON OR FIRM SUPPLYING
A BELT^CONVEYOR SCALE TO THE USER TO CONTACT THE SOUTHERN WEIGHING
AND INSPECTION BUREAU AS FAR IN ADVANCE OF THE TESTING DATE AS POSSIBLE.
IN ORDER TO EXPEDITE SUCH TESTS, IT IS SUGGESTED THAT THE BUREAU'S
GENERAL OFFICE BE CONTACTED AS SOON AS DEFINITE PLANS HAVE BEEN~MADE
FOR THE INSTALLATION OF THE SCALE. FURTHER, AFTER INSTALLATION^
THE PERSON OR FIRM SUPPLYING THE BELT-CONVEYOR SCALE MUST ATTEST
(IN WRITING) TO THE FACT THAT ALL THE AFOREMENTIONED RULES, REGU-
LATIONS. AND CONDITIONS HAVE BEEN MET, AND THAT THE SCALE IS WEIGHING
WITHIN TOLERANCE. ONLY WHEN THE USER HAS BEEN SO INFORMED AND"A~
COPY OF THE VENDER'S LETTER HAS BEEN FURNISHED TO THE SOUTHERN
WEIGHING AND INSPECTION BUREAU SHALL THE BUREAU BE NOTIFIED THAT THE
WEIGHING SYSTEM IS READY FOR TESTING.
The purpose of this requirement is threefold: to protect the buyer; to protect the
seller; and to protect the carrier. The scale supplier may refuse a letter stating that
the belt conveyor scale system is correct and in good weighing condition, until all con-
ditions are such that good weights can be achieved. On the other hand the buyer is
guaranteed that the system will produce good weights. Lastly, the carrier or its repre-
sentative does not have unnecessary expenditures in weighing cars, traveling and living
expenses of its representatives, and wasted time in testing when the system is not
cabable of producing reliable weights.
131 4.
-------
Document 2
Page 7
Testing Belt Conveyor Scales
There are two types of tests used to calibrate belt conveyor scales: simulated
load tests; and material tests. Of the two, the best way (and the only acceptable method)
is the material test.
Ivteterial Test: After finding that all conditions comply with rules for belt conveyor
scales a material test is made. (This test consists of moving previously weighed - or
material to be weighed - over the belt conveyor scale. Extreme care must be taken to
see that all material is weighed both on the check weigh scale and on the belt conveyor
scale. Then the two weights are compared, the difference figured, and the error in
percentage computed.)
1. The check-weigh scale (track scale, truck scale, etc.) is checked to determine
that it is in good condition and has been calibrated.
2. The scale conveyor belt is run for twenty minutes or more (at below 42 degrees
the belt should be run for 60 minutes).
3. A short meeting should be held with all personnel at the conveyor scale site to
explain what is to happen.
4. After running the belt empty (warming up the belt), a reading is taken from
the master totalizer.
(ll) 5. The belt is run for approximately 10 minutes and the reading is again taken
It should not vary more than one increment of the scale. If the reading varies more
the zero must be adjusted. This process is repeated until an acceptable zero condition
is achieved.
6. After taking the master totalizer reading, material is introduced onto the
scale belt and the rate of flow should be carefully observed to rise to better than 50%
of rated capacity. The ideal operating and weighing range is 77 to 85 percent of rated
capacity. (As a rule of thumb, if the time the scale operates after the infeed is opened
and closed at under 50% capacity is no more than 10% of the running time, acceptable
weighing conditions are present.)
7. After the weighing has been completed, the belt should be running and empty
(The belt should not be stopped.)
8. The reading is then taken from the master totalizer again. The "start"
figure is subtracted from "stop" figure, which shows tons (or pounds) weighed. This
figure is compared with the printer. The printer may show as much as two of the
smallest increments difference, which is permissible.
9. Since three to five tests are required, the other tests are then made, follow-
ing the same procedure as above.
10. When all tests have been completed and the material has been weighed on the
check weigh scale, the percentage error is computed. The check weigh scale must
always be considered the "K" factor (or constant) with the belt conveyor scale being
calibrated to it. (Note: In all cases, the cars or trucks must be both light weighed and
gross weighed on the same scale. Weights taken from two different scales should not
be used, and in every test the cars or trucks must be light weighed. Also, weighings
must be made to the smallest increment of the weight indicator.)
132
-------
Document 2
Page 8
Simulated Load Test of Mechanical Belt Scales: The scale and all component
parts must be thoroughly inspected to determine that compliance with the rules has
been met.
A simulated load test is used as a reference only and should not be considered
as a calibration test. The following are the steps to be taken for a simulated load test
of a mechanical belt conveyor scale.
1. Msasure the exact length of the conveyor belt on the belt center to the nearest
1/10 foot.
2. Find the stamped weight per foot on the test chains (usually at or near the end
of the chain on a brass plate). When more than one chain is used, add the weights per
foot.
3. Run the conveyor for at least twenty minutes. (Below 42 degrees the conveyor
should be run for 60 minutes.)
4. Take the reading from the master totalizer and run belt for ten minutes.
5. Take the reading again and compare. If the reading varies by more than one
increment, the scale should be adjusted for "zero". This process should continue until
the final adjustment proves the scale to be holding zero.
6. Stop the conveyor and apply test chains, the center being over the center of
the weight sensing element. Bridles made of cable or heavy rope should be affixed to
the ends of the chains to securely hold them in place on the belt.
7. The conveyor should be run several circuits to center the chains on the belt.
Adjust the bridle if necessary.
8. Select one idler at the scale and mark it as the reference point.
9. Kferk the side of the belt and run three complete revolutions and stop the belt
when the mark exactly crosses the center of the marked idler. Nfark the belt again
exactly opposite the reference point.
10. Read the master totalizer to at least 3 decimal places.
11. Run the conveyor (10 circuits for a short belt and 3 circuits for a long belt).
Stop the conveyor when the first mark passes the center of the reference point idler.
12. Read the master totalizer.
13. Nferk the belt exactly in the center of the reference point idler.
14. Nfeasure the over-run or under-run from the mark in the latter part of (9).
15. The number of circuits of the belt times the number of feet in one circuit
(length of belt) plus or minus the over-run or under-run gives the length the belt
traveled. Miltiply this by the weight per foot of test chains (see Item 2). The result-
ing figure is total weight. Convert the resulting figure to tons (or pounds), if necessary,
by dividing the scale registration unit. Compare the results with the results of
subtracting (10) from (12).
16. Run this same test the required number of times (at least three tests or at
least five if excessive errors are shown).
133
-------
Document 2
Page 9
Electronic scales should be tested as follows:
Since electronic scales are equipped with a switch which can disengage the inte-
grator, measuring over-runs or under-runs from the belt is not necessary. The scale
operator merely runs the belt and engages the integrator as a mark on the belt passes
a reference point. When the belt has run the desired number of circuits, the switch is
again activated to disengage the integrator as the mark on the belt passes the reference
point. Therefore, Items 1 through 13 should be followed as should Item 16.
Other Suggested Steps to be Taken:
1. Since a belt conveyor scale must have repeatability (ability to repeat the same
indication when the same test load is applied successively) it is wise to conduct at least
three simulated load tests before the initial material test is made. If such tests prove
the scale does not have repeatability, further testing would be impractical as the scale
would not be capable of weighing correctly.
2. If a belt conveyor scale has been calibrated to test chains before material
testing, and it is necessary to further adjust the scale to the material tests, it then
proves the test chains are not in calibration with the weighing system. It would be
necessary to then conduct at least five simulated load tests following the material tests
These chain tests would produce an error. The errors should then be averaged to
produce a calibration factor. If this factor is plus a certain percentage, then such a
factor should be subtracted from future chain tests. By the same token,' if the factor
shows minus a certain percentage then such a factor should be added to future chain
tests.
3. Because material tests are considered the only satisfactory means to calibrate
a belt conveyor scale, simulated load tests are only considered as a reference point
In order to show that the belt conveyor scale is holding its calibration, weekly simulated
load tests should be made and reports of such tests should be furnished to the person
responsible for supervision of the scale. By this means, monitoring of the weighing
system is possible on a weekly basis without incurring the expense of railroad weighing.
(7J 4. When weekly tests show that the belt conveyor scale is not out of calibration
more than 0.5%, material tests consisting of ten to fifteen cars (or trucks) may be
made only once every six months. When the material being handled leaves a residue
on the belt, such as wood chips, the material tests should be made approximately once
every 90 days, not to exceed once every 120 days.
5. It is advisable whenever practical to have the check-weigh scale tested and
inspected before material tests are made.
6. Before final approval is given to any belt conveyor system, the user should
provide evidence that at least two inspections will be made each year by a qualified
scale service man. Of these visits, at least one should include a complete overhaul
134
-------
Document 2
Page_K)_
Definition of Terms Used in connection with Belt Conveyor Scales
Air Rirge: The act of creating an atmosphere greater on the inside than outside, using
clean dry air, to discourage dust and other foreign matter.
Belt Conveyor: An endless moving belt for transporting material from place to place. *
Belt Conveyor Scale; (Belt Scale); (Conveyor Scale): A device installed on a belt
conveyor to measure the weight of bulk material being conveyed. *
Bend Rilley: A roller placed on the return side (underside) of a conveyor belt to turn
its direction or to measure speed of belt.
Cable Conveyor: A belt conveyor utilizing cable or rope as the supporting member for
the conveyor on which the idlers are mounted.
Calibrated Hate: A suitable metal plate, provided by the scale manufacturer, deter-
mined to have the same effect on a nuclear scale as a specified load or bulk material
on the belt conveyor. A calibrated plate is the equivalent of a test chain or test
weights used with other types of belt conveyor scales. *
Chain: See Test Chain.
Chart Recorder: A device used with a belt conveyor scale which records the rate-of-
flow of bulk material over the scale at any given time. A recorded chart together
with a record of weight constitutes proof of weight.
Concave Curve: A change in the angle of inclination of a belt conveyor where the center
of the curve is above the conveyor. *
Convex Curve: A change in the angle of inclination of a belt conveyor where the center
of the curve is below the conveyor. *
Conveyor Stringers: Support members for the conveyor on which the idlers are
mounted.
Counter (Remote): A numerical display in a location remote from the scale showing
the tons (or pounds)of material that have been conveyed over the scale.
Disc Chart: See Chart Recorder.
Drive: The device or apparatus used to transmit energy from a motor to move the
conveyor belt.
Feeders: See Infeed.
Gate: See Infeed.
Gravity Type Wipers: A scraping or wiping device used to clean residue from the belt
on the return side (underside)of the belt in the vicinity of the head pulley. The
wiper is affixed to one end of an arm which has a weight hanging from the other end.
The weight is such that the wiper is held against the belt.
Head Rilley: The pulley at the discharge end of the belt conveyor. The power drive to
drive the belt is generally applied to the head pulley. *
135
-------
iler Space: The center-to-center distance between idler rollers measured parallel
to the belt.*
Idler Frame: The frame or device which holds the idler rollers, affixed to the conveyor
stringers.
Mler or Idler Rollers: Freely turning cylinders mounted on a frame to support the
conveyor belt! For a flat belt the idlers may consist of one or more horizontal
cylinders transverse to the direction of belt travel. For a troughed belt, the idler
will consist of one or more horizontal cylinders and one or more cylinders at an
angle to the horizontal to lift the sides of the belt to form a trough. *
Ihfeed: The gate, short belt, vibrator feeder, stroker feeder, etc. that deposits
material on the belt conveyor to be weighed.
Integrator: The "heart" of the belt conveyor scale. A device which integrates the
speed of the belt with the weight of the material to produce tons (or pounds) per
hour. The integrator may be electronic or mechanical and may be one of numerous
patented designs.
Loading Rnnt: The location at which material to be conveyed is applied to the conveyor.
\fester Totalizer: See Totalizer.
Nuclear Type (non-contract) Scale: A device consisting of a source of nuclear radiation
and a detector for that radiation. Absorption of radiation determines the mass of
the material passing between the source and the detector. *
Printer: A device used to imprint on tickets, tape, or other papers, the tons (or pounds)
of material that have passed over the scale in a given time. (Such as per car or per
train weights.)
Rilley: A cylindrical roller over which the belt passes to change direction, such as
the head pulley or tail pulley or bend pulley.
Rate of Speed Detector: A device usually operated in conjunction with electronic load
cell belt conveyor scales which transmits the speed of the belt to the Integrator.
Several types are in use, the most popular being either a small generator whose
voltage output is in direct relationship to the speed of the belt; or a slotted disc
mounted between a photo electric cell and light source which converts each pulse
of light to an electronic signal to the integrator.
Rated Capacity: That value representing the weight that can be delivered by the device
in one hour. *
Registration: The unit of weight in which the scale is calibrated, such as 5,000 Ibs.,
tons, long tons, metric tons, etc.
Rope Conveyor: See Cable Conveyor.
Simulated Weight Test: A test using artificial means of loading the scale to determine
the performance of a belt conveyor scale. *
Skirting: Stationary side boards or sections of belt conveyor attached to the conveyor
support frame or other stationary support to prevent the bulk material from falling
off the side of the belt. * Usually used at infeeds.
136
-------
^
Stringer: See Conveyor Stringer.
Strip Chart: See Chart Recorder.
Strip Heater: A thermostatically controlled heating strip (usually of the Calrod type)
used to heat the scale or component parts sufficiently to prevent condensation.
Swedged Rilley: A drive pulley with a vulcanized rubber coating molded in a high
friction pattern to prevent slippage of the belt under load.
Swivel Idler: An idler frame pivoted in its center to the idler stringers so that it may
change position. Tracking idlers are mounted beside the troughing idlers so that if
the belt rides off center the pressure on the tracking idler forces the swivel idler to
turn, thus realigning the belt on the conveying idlers.
Tail Rilley: Hie pulley at the opposite end of the conveyor from the head pulley. *
Take-up: A device to assure sufficient tension in a conveyor belt that the belt will be
positively driven by the drive pulley. A gravity take-up consists of a horizontal
pulley free to move in either the vertical or horizontal direction with the dead weights
applied to the pulley shaft to provide the tension required. A jack or screw type take-
up consists of a device that must be manually adjusted to move the tail pulley to
increase or decrease tension. A hydraulic take-up consists of hudraulic cylinders
mounted on either side of both ends of the tail pulley. When properly activated, the
hydraulic cylinders will move the tail pulley to increase or decrease tension on the
belt.
Take-up Guides: Glides or tracks on either side of the counter weight to prevent
vertical movement of that weight.
Test Chain: A device consisting of a series of rollers or wheels linked together in such
a manner as to assure uniformity of weight and freedom of motion to reduce wear,
with consequent loss of weight to a minimum. *
Totalizer: A device used with a belt conveyor scale to indicate the weight of material
which has been conveyed over the scale. The master weight totalizer is the primary
indicating element of the belt conveyor scale. An auxiliary vernier counter used for
scale calibration is not (necessarily) part of the master weight totalizer. Auxiliary
remote totalizers may be provided. The totalizer shows the accumulated weight and
may be non-resettable or may be reset to zero to measure a definite amount of
material conveyed.
Tracking Idlers: Usually small cylinders vertically mounted on shafts affixed to a
swivel idler frame. The purpose is to allow the side of the belt to rub against the
tracking idler, forcing the swivel idler frame to turn, thus realigning the belt.
Training Idlers: Idlers of special design or mounting intended to shift the belt sideways
on the conveyor to assure the belt is centered on the conveying idlers. *
Tripper: A device for unloading a belt conveyor at a point between the loading point and
idle head pulley. *
Wing Rilley. A pulley usually used as the tail pulley, made of widely spaced metal bars
in order to set up a vibration to shake loose material off the underside (return side)
of the belt. The use of such a pulley is definitely not recommended unless the con-
veyor stringers under the scale are thoroughly braced with their own support.
Wiper: See Gravity Type Wiper.
10.
(*) Taken from MBS H-44.
137
-------
Document 3
Page 1
File 958S-5-H
CHAIN TEST
(1) Run belt for 20 minutes when temperature is above 42°; 1 hour when temperature is
between 32° and 42° ; for as long as possible more than 1 hour when temperature is
below freezing.
(2) Take scale reading, and 10 minutes later take reading again. If scale is on "zero",
reading will be the same, ff not, adjust "zero" and repeat above procedure until no
movement of read out is detected.
(3) Look at plate on chains and determine weight per foot.
(4) Measure belt to nearest inch or 1/10 foot.
(5) Hace chains on belt and tie down.
(6) Select one idler as a reference point and mark three long stripes on side of belt
opposite idler. Start belt running.
(7) Multiply length of belt times chain weight per foot and divide by 2000 to determine
tons per 1 revolution of belt. This is the K(l) factor.
(8) Divide K(l) by belt length to determine tons per foot. (If an inch rule is used, divide
above result by 12 to determine weight per inch.)
(9) Multiply K(l) by number of circuits of belt to be run to determine K factor.
(10) When the last of the three stripes on the edge of the belt pass the reference idler,
shut off the belt. Mark a spot on the belt opposite die reference idler and further
mark the direct center in a contrasting color, i.e., yellow or white mark with a
pencil line in the direct center.
(11) vVrite down the scale reading.
(12) Run the belt the required number of circuits and on the last circuit, when die last
of the three stripes passess the reference idler shut off the belt.
(13) If the previous mark does not perfectly align itself with the center of the reference
idler, again mark the edge of die belt and locate a contrasting line directly in die
center. Measure the distance between die marks. K an overrun, it is a "plus"; if
an under run, it is a "minus". Multiply die measurement by die K factor for feet or
inches, as die case may be. This produces tons over/underrun.
(14) Read scale and subtract (11) from it.
(15) H overrun was noted, subtract tonnage from reading; if underrun, add tonnage to it.
(16) Compare results with K factor. If K factor is die larger of die two figures, die
resulting error is "minus"; if die K factor is die smaller of die two figures, die
resulting error is "plus".
(17) Determine error percent by dividing K factor into error tonnage. (Be sure to move
decimal point 2 places to right in computations.)
138
-------
Document 3
Pagej.
-2-
(18) Repeat steps (11) through (17) at least two more times to determine error repeatability.
When error has been determined, adjust scale span as per directions from manufacturer.
Note: Each time step (13) is repeated, remove old centerline mark on over/underrun so
as not to confuse future measurements.
(19) K adjustment to span was made, repeat steps (11) through (18) at least three times to
determine adjustment was correct and scale is in correct calibration.
/2Q) Hie best possible time to run a chain test is immediately after material handling has
taken place so that belt is in proper flexible condition.
« Georgia,
17. 1975.
139
-------
(7)
ANNOTATIONAL COMMENTS
Requirements for reporting to the Bureau or for adhering to the
Bureau's regulations apply only to devices falling under Bureau
jurisdiction. The appropriateness of a given data requirement or
regulation must be determined for each situation not governed by
the Bureau or the regulatory organization.
Items under this note, such as for idlers and location of scale,
will generally not apply to systems using a nuclear scale. The
manufacturer should be consulted if applicability is unclear.
If a nuclear scale is used, the only contact requirement is for
the belt speed indicator, which has a negligible impact on belt
tension requirements.
"Car" refers to railroad car - for other coal sources, read as
"The system must be so designed that all calibration material may
be guaranteed to pass over the scale."
Incline is to prevent badslide of material and applies to nuclear
scale instruments as well as mechanical and electromechanical ones.
Nuclear scale manufacturers' specifications indicate that the accuracy
level may not be achievable for that device. If manufacturer's speci-
fications are considered acceptable for a given application, then the
scale is acceptable upon verification that it meets those specifica-
tions (in the absence, of course, of more stringent requirements from
a regulatory body such as the Bureau).
The requirement for 10 to 15 car loads corresponds to'vZOOOtons of
coal, and is usually used in calibration of conveyors with approxi-
mately 8000 tons per hour capacity, which corresponds to a nominal
calibration run time of 15 minutes. This quantity of coal would be
inappropriate for very small capacity units. As a rule of thumb,
it may be considered acceptable to use a quantity of coal which would
result in a run time of at least 15 minutes at 80% capacity and at
least three complete circuits of the belt.
140
-------
The nuclear scales, a (static) plate takes the place of test chains.
The belt should be running, empty, when the plate is being used to
account for effects of the belt itself.
Some manufacturers offer a static test weight for calibration purposes.
This can be used to accurately check the load cell itself, but will
not indicate systematic errors due to changes in belt tension and other
factors related to the conveyor belt/weighbridge interface in mechani-
cal and electromechanical systems.
Belt length concerns show a need for good communication between user
and vendor so that appropriate hardware is purchased. Very short
(1-3 m) conveyors are usually purchased as a packaged system, often
including the hopper which feeds the belt.
"One increment of the scale" requirement is listed as .01% in the
first report.
Afterthought: Nowhere in manufacturers' data which we have examined is
there an indication of need for the care and attention to detail
indicated in these reports. Many of the small details are not
expensive or time-consuming, and several need to be performed only
once. Simple operation descriptions by sometimes overzealous
marketing people should not lead users to the conclusion that
installation and operation details as described in these reports
may be glossed over or dismissed.
141
-------
REVIEW COMMENTS
In his review of the draft report, Mr. Martin Gruber of the Southern
Weighing and Inspection Bureau returned the following comments (presented
verbatim):
1. - Page 121. No. 10, Footnote 1. Of primary importance to anyone
who depends on results from belt conveyor scales is an accurate weight.
A rate-of-flow chart allows complete supervision of the instrument without
the necessity of having an employee making constant observation.
2. - Page 122. No. 23, Footnote 1. It has been our experience that
the mere act of barring access to a belt scale adjustment mechanism by the
simple process of locking cabinet doors has resulted in more accurate scales
and less frequent overhauls. Locking access doors allows the key holders to
have complete control over unauthorized tampering.
142
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO. 2.
EPA-600/7-79-196
4, TITLE AND SUBTITLE
Assessment of Instrumentation for Monitoring Coal
Flowrate and Composition
7 AUTHOR(S)
E . F . Brooks and C . W . Clendening
9 PERFORMING ORGANIZATION NAME AND ADDRESS
TRW Systems and Energy
One Space Park
Redondo Beach, California 90278
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION- NO.
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
INE624
11. CONTRACT/GRANT NO.
68-02-2165 (Task 9) and
68-02-2613 (Task 2)
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 7/76 - 7/78
14. SPONSORING AGENCY CODE
EPA/600/13
,5. SUPPLEMENTARY NOTES ffiRL-RTP project officer is Frank E. Briden, Mail Drop 62.
919/541-2557.
i6. ABSTRACT
repOrt. gives results of an assessment of instrumentation for the mea-
surement of coal flowrate (either as a dry solid or in a coal/water slurry) and com-
position. Also investigated was the appropriateness of EPA/TERL-RTP involvement
in the development or evaluation of such devices. Findings for flow measurement
hardware were that dry coal flow can be easily and accurately measured using weigh
belt devices , and that the mass flow of coal in a coal/water slurry stream can be
measured using a flowmeter (electromagnetic flowmeters are preferred) and a
nuclear density gage. The most promising analysis concept under development is
fast neutron activation, with delivery of a sulfur and ash meter anticipated by the end
of 1979. Other techniques , such as X-ray fluorescence, work on only a very small
coal sample. It is recommended that further EPA investigation deal with system,
rather than component, evaluations.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
:. cos AT I Field/Group
Pollution
Assessments
Instruments
Monitors
Coal
Flow Measurement
Measurement
Compositions
Electromagnetis m
Flowmeters
Fast Neutrons
X Ray Fluorescence
Pollution Control
Stationary Sources
Weigh Belts
Nuclear Density Gages
13B
14B
20C
20H
08G,21D 20F
^"DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
153
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA
Form 2220-1 (9-73)
143
------- |