EPA-650/2-75-018
FEBRUARY 1975
Environmental Protection Technology Series
ui S$i$S:iii::i:
O W$m
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4 ENVIRONMENTAL MONITORING
5. SOC1OECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
Tint, report liat, been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY scries. This scries describes rest-arch performed to
develop and demonstrate instrumentation, equipment and methodology
to repair or prevent environmental degradation Irom point and non-
point sources of pollution. This work provides the new or improved
technology required for the control and treatment of pollution sources
to meet environmental quality standards.
-------
EPA-650/2-75-018
DESIGN, DEVELOPMENT,
AND FIELD TEST
OF A DROPLET
MEASURING DEVICE
by
lU-ctnr Muducki, Myron Kaufman, and Danu-l E. Magnus
KLD Associates, Inc.
7 High Street
Huntmgton, New York 11743
Contract No. 68-02-1309
ROAP No. 21ADJ-081
Program Element No. 1AB012
EPA Project Officer: D. Bruce Harris
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL'PROTECTION AGENCY
WASHINGTON, D.C. 20460
February 1975
-------
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
11
-------
ABSTRACT
The design, development and field testing of a concept
to measure liquid droplets in the size range from lym to
600pm is described in the report. The measurement probe
is a platinum wire 5ym in diameter. With the probe
electrically heated to a predetermined temperature,
changes in the probe resistance are related to the size
of the impinging droplets in the gas flow. The electrical
signals from the probe are processed and used to classify
the droplets into six different size ranges or bins.
Two prototypes consisting of the probe and electronic
processing unit were constructed and tested in the field.
Actual droplet distributions and concentrations were
successfully measured.
iii
-------
CONTENTS
Abstract
List of Figures
List of Tables
Acknowledgments
Sections
I
II
III
IV
V
VI
VII
CONCLUSIONS AND RECOMMENDATIONS
INTRODUCTION
DESIRED CHARACTERISTICS OF
THE DROPLET MEASURING DEVICE
PRINCIPLE OF OPERATION AND
DESCRIPTION OF THE DC-1
LABORATORY AND ANALYTICAL
STUDIES
FIELD TESTING
REFERENCES
Page
111
v
vi
vii
1
3
21
41
47
iv
-------
FIGURES
No. Page
1 Principle of Operation of the Sensor
(Idealized) 10
2 Model DC-1 Droplet Counter Block Diagram 12
3 Photograph of Instrument and Probe 14
4 Temperature Calibration Curve 17
5 Air Velocity Calibration 18
6 Effect of Eccentric Collision on
Signal Output 23
7 Typical Sensor Electrical Output 26
8 Calibration Curve 29
9 Equipment Arrangement for Calibration 35
10 Calibration Circuit 37
11 Typical Droplets and Electrical Output 38
12 Demister Cross Section 42
-------
No.
I
TABLES
Typical Sequence of Measurements
Page
46
VI
-------
ACKNOWLEDGMENTS
The work reported herein was performed by KLD Associates/
Inc. under contract number 68-02-1309 from the Environ-
mental Protection Agency, Research Triangle Park, North
Carolina. Mr. Bruce Harris was the Project Officer for
the EPA and his support and suggestions are acknowledged.
The field testing of the instrumentation was performed
at the facilities of the Nassau County Sewage Treatment
Plant, Wantagh, New York. Mr. Frank Flood, Superintendent
of Operation and Maintenance in the Nassau County Depart-
ment of Public Works, made the necessary arrangements at
these facilities, and his support is gratefully acknowledged,
The cooperation of all individuals at the treatment plant
was outstanding and contributed greatly to the success of
the field testing.
vii
-------
SECTION I
CONCLUSIONS AND RECOMMENDATIONS
In this study, the concept of measuring droplet size and
concentration with a hot-wire sensor was investigated and
implemented. From the results of the study, the following
conclusions are made:
The concept of measuring droplets with a hot-
wire sensor is feasible over the size range
from lym to 60Gym. The measurement technique
has been demonstrated in the laboratory and
under field conditions.
A prototype design, consisting of the probe
and electronic processing unit, was implemen-
ted and used under field conditions. Actual
droplet distributions and concentration were
successfully measured and reported.
The prototype system is well suited to field
work since it is lightweight and easy to
operate. The probe itself is delicate and
must be handled carefully. However, the
probes, once installed, lasted for many hours
of measurement under the field environment.
Techniques were developed to discriminate the
signals generated by the droplets from the
noise due to turbulence. These techniques
proved very successful for the field environ-
ment encountered in this project.
Two prototype units were delivered to the Environmental
Protection Agency. These units are suitable for making
surveys of droplet distributions in scrubbers and mist
eliminators provided the following operating conditions
are not exceeded:
size: lym to 600ym
flow velocity: 3 m/sec
concentration: 500 droplets/cm3
temperature: 0°C to 100°C
-------
The prototype instruments are the first generation in the
development of the concept, and further improvements can
extend the operating range and usefulness of the device.
Some recommendations for further research and development
are:
The units should be used to perform a series
of measurements for numerous environmental
conditions encountered in different scrubbers
and mist eliminators. The data should be
carefully recorded and analyzed to establish
well defined performance characteristics for
the device.
The effects of turbulence and varying tempera-
ture conditions need more study in the
laboratory. Work, thus far, was limited in
this area. As the device is subject to new
operating conditions, data would become available
to help interpret measurements and to improve
calibration procedures.
The range of the device should be extended to
higher flow velocity and smaller droplet
sizes. Under a proper research program it
appears that measuring submicron droplets should
be possible.
The device is now powered using a 110 volt 60
c/s supply. The usefulness of the device can be
improved by making the unit battery powered.
The output from the electronic signal processor
is now a manual operation. For monitoring
droplet distribution over extended periods of
time in the field, the system should be
automated to provide hard copy printouts.
-------
SECTION II
INTRODUCTION
The measurement of small liquid droplets in the gas
stream of scrubbers and mist eliminators is a need in
pollution control technology. The capability to make
such measurements will lead to improved performance of
such devices and to a better understanding of the processes
involved. The work described in this report covers the
design, development and field testing of equipment to
satisfy this measurement need. Specifically, the
objective of the project is to produce a device capable
of measuring droplet size and concentration as it exists
in the gas stream in and around scrubbing systems.
Under this project, KLD Associates, Inc. developed a
droplet measuring device which uses a hot wire for the
sensor. As a droplet attaches to the wire, the local
cooling of the wire causes an electrical pulse which is
then analyzed for the droplet size. The instrument sorts
the sizes with respect to six ranges or bins, and the
number of droplets in each bin is accumulated to provide
a droplet distribution. The probe can be placed directly
in the gas stream and is connected to the electronics
system via a coaxial cable. The system is designated
the DC-1 Droplet Counter.
In this report, the work performed in the development of
the droplet measuring device is presented. The initial
effort involved a literature review to determine the
desired operating characteristics for the measuring device.
These specifications, which appear in Section III,
provided the basis for the design and laboratory work
presented in Section V. Extensive laboratory work was
required to better understand the droplet-wire attachment
mechanism over a wide range of droplet sizes. A special
optical test apparatus was developed to calibrate the
device, and, of course, a variety of droplet generating
devices were used to provide drops of different sizes,
concentrations and velocities.
-------
A most important phase of the project was the field testing
(Section VI) of the device in the demisters on the power
generation system for the Nassau County Sewage Treatment
Plant. These tests immediately demonstrated that most
droplets were less than lOym in diameter. The device was
then modified to measure droplets with diameters in the
range of lym to SOOym. In addition, the effect of flow
turbulence was noted and the device altered to discriminate
between turbulent noise and the actual droplet signatures.
The final design of the DC-1 incorporates these modifica-
tions and other secondary features to make it more suit-
able for field operation.
-------
SECTION III
DESIRED CHARACTERISTICS OF THE DROPLET MEASURING DEVICE
The initial effort under the research program was to
determine the specifications and operating conditions in
scrubbers and mist eliminators where the droplet measuring
device would be used. These specifications refer to
Range of droplet size
Concentration of droplets
Velocity of the gas
Temperature of the gas.
Such parameters are the basic information necessary for
the design and testing of the droplet measuring device
for the intended applications by the EPA.
The range of these parameters was initially determined by
a literature review on water droplets, scrubbers and mist
eliminators. A partial listing of the more relevant
documents is presented in References 1 through 18. In
addition, manufacturers and users of scrubbers, demisters
and gas cleaning equipment were contacted for information
on the operating parameters of their equipment. These
industrial contacts included visits to local sites and
telephone discussions with representatives who were outside
the New York City metropolitan region. The gathered
information was then analyzed and the specifications
established for the droplet measuring device. In the
following four sections, the results from this phase of
the study are presented. Also presented are some comments
on the specifications in view of the field measurements.
As the field work progressed, the characteristics of the
droplet measuring device had to be modified, since the
specification for the range of droplet size was not
adequate. By extending the range of the device to smaller
droplets, a better representation of the distribution in
the demister was possible.
RANGE OF DROPLET SIZE
The range of droplet size was defined using (1) empirical
data from the literature, (2) specifications from manu-
-------
facturers, and (3) the investigation of various criteria
setting limits to the attainable droplet size. The
relevant criteria are:
Thermodynamic minimum size
Aerodynamic maximum size
Evaporation minimum size.
The thermodynamic minimum size criteria postulates that
the amount of mechanical work required for droplet forma-
tion is inversely proportional to the droplet size. Data
on commercially available nozzles shows that the
efficiency in converting flow energy into droplet breakage
is of the order of 1% for sizes in the 10-micron range;
the efficiency increases for larger droplets. Based on
these results, the minimum droplet size can be predicted
for a given pressure drop across the nozzle. The aero-
dynamic maximum droplet size is estimated from the balance
between surface tension tending to keep the drop together
and the aerodynamic forces. The largest droplet size
corresponds to free fall in still air with diameters
reaching 8 mm to 9 mm. Practical conditions encountered
in scrubbers, with higher velocities and substantial
turbulence, reduces the maximum possible droplet diameter
down to one or two millimeters. The evaporation of
droplet produces a reduction in diameter in such a way
that the distribution becomes time dependent. The life
of small droplets depends on the gas temperature and
humidity being, in general, rather short of the order of
one second.
Using data on direct droplet measurement, as provided by
manufacturers and investigators, the drop size distribu-
tion for cone nozzles, rotating disk and Venturi scrubber
was analyzed. The result of this study shows that most of
the droplets would be within the range of 10pm to SOOnm.
This range was used for the initial development of the
device. However, the field measurements with the device
indicated that most of the droplets in the scrubber (in
the region of its demister) were less than lOym. As a
result of field work, the instrument was modified to
operate over the range of lym to about 600ym.
-------
CONCENTRATION OF DROPLETS
Values of droplet concentrations have been measured by
several investigators and computed for a variety of situ-
ations. The maximum concentration was found to vary over
several orders of magnitude. The following conclusions
may be reached from the available data:
Higher concentrations are achieved with small
droplet diameters; the limit is set by either
evaporation or coagulation due to Brownian
motion.
Monodispersions achieve larger concentration
than polydispersion.
Turbulence of the flow reduces substantially
the concentration, as compared to a quiescent
suspension.
Computed maximum concentrations are orders of
magnitude larger than measured values.
For the design of the instrument, the maximum droplet
concentration measured in fog was adopted, with an actual
value of 500 drops/cm3. Such extreme conditions were
observed in still air with droplets in the 10-micron
diameter range. This value of concentration is higher
than the data encountered in scrubbers where polydisper-
sions and a strong turbulence are present.
FLOW VELOCITY
This parameter varies within wide limits and two basic
types of scrubbers were considered; the spray scrubbers and
the Venturi scrubber. The first type is characterized by
a low velocity of the flow with substantial time for the
fluid-contaminant interaction.
Venturi scrubbers operate at high velocity; the gas flow
is totally or partially responsible for the generation of
small droplets. Flow conditions are changed downstream to
generate settling conditions.
-------
A survey of typical installations shows that gas velocities
between .6 and 2 m/sec (2 and 6 ft/sec) are encountered
in spray scrubbers and 40 to 100 m/sec on the venturi
throat. The droplet measuring device with a standard
probe is designed to operate with a flow velocity up to
3 m/sec (~10 ft/sec). At higher velocities, the flow may
become very turbulent and errors in the measured data can
result. To make measurements at high velocity, a special
probe needs to be developed.
TEMPERATURE OF THE GAS
The temperature of the gas can vary over a wide range in
scrubbers. For certain facilities, a portion of the
injected water is used to cool the gas and the proper
scrubber action takes place downstream where the gas
temperature has been substantially reduced. On the lower
side, freezing of the water sets a minimum limit to the
flow temperature. Hence, a temperature range between
20°C and 100°C is representative of conditions to be
encountered in typical scrubbers and mist eliminators.
This range was used in the development of the probe for
droplet measurements. All the field studies were made at
temperatures within this range.
-------
SECTION IV
PRINCIPLE OF OPERATION AND DESCRIPTION OF THE DC-1
The operation of the hot wire droplet sensor DC-1 is based
upon the localized heat transfer and cooling which is caused
by the droplet impinging on a small hot wire or surface.
The concept is schematically shown in Figure 1, where the
hot wire sensor and its longitudinal temperature distribu-
tion is shown (a) before and (b) after a water droplet
(cross-hatched circle) attaches. The electrical resistance
of the wire is a function of the wire temperature, which in
situation (a) is high and substantially uniform along the
wire. In situation (b) the portion of the wire covered by
the droplet is cooled to approximately the droplet tempera-
ture. A constant electrical current flows along the wire
creating a measurable voltage drop at the wire support.
The high voltage encountered before droplet attachment (a)
is reduced for (b) in direct proportion to the cooled
length of wire; i.e. the droplet diameter. The electrical
energy delivered to the wire evaporates the water, leaving
the sensor clean and ready for further interaction.
The above description of the operating principle is an
idealization and in actual practice many aspects influence
the electrical signal. For example, heat conduction in the
radial and longitudinal directions can have an important
effect on the electrical signal. These effects were exten-
sively investigated analytically and experimentally. The
analytical studies (described in section V) guided our
research in the selection of materials which are used in
the probe.
DESCRIPTION OF THE DC-1
The instrument consists of a box containing the circuitry
for analyzing the electrical pulses and a hot-wire probe
to be located in the region where the water droplets are
present. A single coaxial cable with lengths up to fifty
feet is used to connect the probe to the electronic circuit.
The probe was designed around a five micron platinum wire
for measuring droplets in the range from lym to 600ym. The
electronic circuit for analyzing the pulses from the probe
-------
Support
Support
Hot Wire
Wire Temp.
Hot-.
Ambient
Wire Longitudinal^
Dimension
(a) Temperature before Attathment
Wire Temp.
Hot-
ne
4
Ambient
Water Droplet
Wire Longitudinal
: Dimension
(b) Temperature after Attachment
Figure 1: Principle of Operation of the Sensor
(Idealized)
10
-------
is designed to classify the droplets into six subranges or
bins. The droplet size range in each bin can be modified
by changing a plug-in ladder network. For both the probe
and electronic circuit, particular attention was devoted
to:
Simplicity of the construction
Small size of the instrument
Operation by a non-skilled operator
Clear display of the measured parameters.
Electronic Subsystem
The front end and conditioner (FECA) of the instrument
(Figure 2) is a bridge able to operate in three modes:
temperature
velocity
droplet count*
For the temperature mode, the probe is fed with a low
current so as not to introduce appreciable electrical
heating; the wire assumes the temperature of the surround-
ing medium and the instrument may be used as a sensitive
thermometer. In the velocity mode, the probe current is
substantial and the probe wire is heated well above the
surrounding medium. Due to air movements relative to the
probe, a signal is generated in proportion to the fluid
velocity. In the count mode, a value of reference resis-
tance 1.5 larger than the corresponding cold probe
resistance is set. A feedback power amplifier supplies
sufficient current to normally balance the bridge. However,
fast fluctuation of the sensor temperature due to droplet
interaction is not compensated by this power amplifier and
the corresponding electrical signal appears at the bridge
output. This output signal is used in the remaining por-
tion of the electronic circuit to determine droplet size.
The signal from the FECA goes into an amplifier and after-
wards splits in two directions:
11
-------
Probe
Input
>-i
Front End
I Conditioner &
Anplifier
1C 38
Count
Tatp
Vel
\
nt*
p-: 0
Gain Ad].
Mode
S5
TP BB
1^
o Strobe I
iming & I
1C 13B /
$ u)
ef. Null
R 1 Meter
idge 1
Pk. S
^ & Hoi
1C
*1_
y Conp. #6 A ' N. EX-OR 6N
"V 1C 6A J 1 Y 1C 12A *
~T Ccnp. #5 ^\
*V 1C 6B )
y Conp. #4 \
V 1C 7A J
~r Com?. #3 -\
V ic 7B J r
~T Conp. #2 A L
*\ TP ftft y
^T Conp. «1 A
ample
d .
3
s
/
~~*^ EX-OR 5 \
Y 1C 12B j
*V EX-OR 4 ^
^
S.
7 1C 12C 7
V EX-OR 3X
7 1C 12D /
\ EX-OR 2 \
X 1C 13A /
~"\ EX-OR 1 N
y
I
-»
-
-»
-
AND
1C
14A
AND
1C
14B
m>
I.A.4C |
AND
1C
14D
AND
1C
15B
ANU
Strobe
Timing &
Logic
1C 4,5,13
1C 15,44,45
1C 58
S2
To
Timing
Logic
15D
Off
Power
-6 P. S.
.99 Counter
.999 Total
S3
Counter Select
17 0
Itot.Time Wlndow
123456
3 Decimal Digit LED |
Display I
1C 41,42", 43 J
Figure 2: Model DC-1 Droplet Counter Block
Diagram
-------
to a LIMITER, and
to PEAK SAMPLE AND HOLD circuit .
The LIMITER triggers the sampling when the electrical signal
has a fast front and exceeds the background noise level.
The droplet signal amplitude is determined at the PEAK
SAMPLE AND HOLD (PSH) circuit which provides a square pulse
with an amplitude directly proportional to the signal peak
and with a duration of two milliseconds.
The signal from the PSH goes to six COMPARATORS. Each com-
parator has a fixed D.C. reference established by a
RESISTANCE LADDER NETWORK, which determines the droplet
diameter intervals for each one of the six channels.
Several ladder networks are available and can be easily
changed since the network is a plug-in circuit. When the
peak detector signal is greater than the reference signal
at one comparator but less than the reference level at the
adjacent comparator, the pertinent logic provides a signal
output at this particular channel only. This output signal
increments a counter which is stored in a channel counter.
The TOTALS counter is activated whenever any one of the
channels is activated. Thus the sum of all the channels
counters equals the total count.
The counting process is started by the operator by pressing
a reset button which also activates a CLOCK with a .1 second
resolution. The counting process stops when either 99 or
999 droplets are counted in the TOTALS counter, or when
99.9 seconds has been reached, whichever occurs first. The
results of this counting function are stored until the re-
set button is pressed again.
At any time during the counting cycle, or at its completion,
the content of the counters may be displayed, one at a time.
This selection is performed with a thumb wheel switch which
controls the multiplexer interfacing between the selected
counter and the three decimal digit display.
All the instrument controls and display are mounted in the
front panel. The finished instrument is shown in Figure 3.
13
-------
FIGURE 3: PHOTOGRAPH OF INSTRUMENT AND PROBE
14
-------
Probe
The wire adopted for the sensor is platinum with a five
microns diameter (.0002 inches).
In its original form, the platinum wire is covered with a
thick (.1 mm; .004 inches) layer of silver. A piece of
this wire is soldered to the probe and the silver is
removed with nitric acid. During this etching process, the
sensor electrical resistance is controlled in order to
achieve four ohms resistance. The length of exposed plati-
num wire is approximately one millimeter.
The body of the probe is a metallic tube 6mm. (.25 inches)
in diameter and 300 mm (12 inches) in length. The assembled
probe is shown in Figure 3. The probe is connected to the
electronic subassembly with a coaxial cable.
USE OF THE INSTRUMENT
The setting of the instrument and interpretation of the
data requires the measurement of flow temperature, flow
velocity, counting of droplets and timing of the operation.
All these measurements are performed with this instrument
without requiring any additional components. A discussion
of each function follows.
Flow Temperature Measurement
The resistance of the wire sensor varies with its tempera-
ture, and this variation is used to monitor the flow
temperature. Both the electrical current flowing in the
sensor and the surrounding fluid determine the equilibrium
temperature of the sensor. With the function selector
switch in the TEMP position, the sensor operates on a very
low electrical current, thereby its joule heating is
negligible. The instrument acts as a sensitive thermometer
that is calibrated to the resistance required to achieve
equilibrium of the electrical bridge. The resistance is
measured with the potentiometer mounted on the front panel.
In general, a temperature versus resistance calibration
curve would be required for each probe. However, a simpler
procedure is to measure the sensor resistance at the known
ambient temperature and compute the sensor resistance at any
15
-------
other temperature using the temperature coefficient corres-
ponding to platinum, for which a = .0038 i .
To simplify the computation, a temperature calibration is
presented in Figure 4.
Flow Velocity Measurement
The flow velocity is measured by imposing a high electrical
current on the sensor, raising its temperature approximately
150° above the ambient. The reference resistance (potentio-
meter) required to balance the bridge provides an indica-
tion of the actual temperature of the probe. Thermal
equilibrium is reached when the electrical input power
balances the heat losses to the air surrounding the sensor;
such a heat transfer balance is a strong function of the
air velocity relative to the probe. A velocity calibration
curve is obtained by setting the function selector switch
to VEL and exposing the sensor to known flow velocities
while the corresponding electrical resistances are measured.
Such a calibration is valid for sensors of equal diameter
and material. For the five-microns-diameter platinum sen-
sor, the calibration performed in the laboratory is
presented in Figure 5. It should be noticed that changes
in sensor resistance are substantial at low flow velocities,
up to approximately one meter per second (3 ft per second).
For higher flow velocities small errors in the resistance
result in large errors in velocity. As a result of this
behavior, the resistance measured at zero velocity must be
precisely determined, to avoid unacceptable errors at high
velocity.
A satisfactory approach to measure the no-flow resistance
was developed using a shield which encloses the probe. The
shielded probe then is introduced into the flow and left in
place until thermal equilibrium is reached. The potentio-
meter is adjusted to null the bridge circuit and the value
of the reference resistance R, is recorded. The shield is
then removed and the probe is reinserted into the flow with
the sensor axis perpendicular to the flow. The potentio-
meter is readjusted to null the electrical bridge and the
value of R_ is recorded. The velocity is then determined
from the calibration curve shown in Figure 5.
16
-------
AT
°C
t
100
90
80
70
60
50
40
30
2Q
RCS = cold resistance in
stack
RCA = cold resistance in
ambient air
10
1 1.05 1.1 1.15 1.2 1.25 1.3 1.35
RCS
RCA
Figure 4: Temperature Calibration Curve
17
-------
FLOW VELOCITY
.125
nss ~
Hot Resistance in stack
(shielded)
Hot Resistance in stack
(unshielded)
.15 .175 .20
Figure 5: Air Velocity Calibration
18
-------
Droplet Counting and Sizing
As previously described, the flow temperature is measured
with the function selector switch on TEM. The reference
resistance, RCS, from the temperature measurement is
increased by a factor of 1.5 and this value set on the
potentiometer. The function selector switch is set to
COUNT and the bridge balance meter will automatically
balance itself because the feedback power amplifier is in
the circuit.
The reset button is pressed to clear all counters and the
interaction of droplets with the sensor may be observed
with the channel switch in position 7, corresponding to
TOTAL count. After the counting cycle is finished (after
99 or 999 droplets, as selected by the operator, or 99.9
seconds interval as dictated by the instrument clock), the
number of droplets in each one of the six channels plus the
time interval are recorded. For each channel the average
droplet diameter DI can be computed using the resistance
ladder network values.
The droplet concentration n. corresponding to the channels
is computed with the expression:
Nj
i = V7!7! (2D,. + d) Drop/cm3 (i = 1,6)
where
N. droplets counted in the i channel
V flow velocity (cm/sec)
t time interval (sec)
S. sensor length (cm)
D^ average droplet diameter for the i^n channel
d sensor wire diameter; 5x10 cm
19
-------
The sensor length may be measured under a low power micro-
scope or its value computed from the cold resistance value,
since a five micron platinum wire has a resistance of 39
ohms per cm. The sensors delivered with the instruments
are 1 cm long.
20
-------
SECTION V
LABORATORY AND ANALYTICAL STUDIES
As presented in Section IV, the principle of operation of
the DC-1 appears simple, but to achieve a good correlation
between liquid droplet sizes and the amplitude of the
electrical signal, several complicated effects required
careful investigation. An important aspect of this
project was the verification of underlying assumptions
behind the principle of operation. For example, the heat
transfer between the wire and droplet is much more compli-
cated than the idealized model of Section IV. Also, the
mechanism of droplet attachment bears an important re-
lationship to the performance of the device. These
types of investigation were necessary to optimize the
performance of the instrument and to better define its
range of operation. Some of these studies were analytical
while others were experimental.
A discussion of the laboratory apparatus to calibrate the
DC-1 is also included. An essential part of the laboratory
equipment is the means of generating droplets in a known
size range. Several techniques were required for this
project as presented in the latter part of this section.
The laboratory experiments of the sensor principle
involved three basic studies:
Droplet-wire attachment mechanism
Amplitude and shape of the electrical signal
Hot-wire versus hot-film sensors.
Each of these topics is described below.
Droplet and Wire Attachment Mechanism
Ideally, each droplet should surround the wire and establish
a close thermal contact over a well-defined portion of the
wire. An ample amount of theoretical and experimental
information is available in connection with the capture of
droplets by bodies of different geometry and materials.
The majority of this work deals with the performance of
filters. However, little, if any, information is available
on the droplet capture phenomenon when the wire is heated
21
-------
to a substantial temperature. Consequently, work was done
in the laboratory to gain knowledge of the attachment
mechanism under condition of heating.
The attachment of droplets to wires of different diameters
and temperatures was observed using a low magnification
microscope. A high speed electronic flash synchronized
with the electrical pulses from the probe permitted
photographing various stages of droplet attachment; and,
from these photographs, a thorough understanding of the
phenomenon was developed. Result of such work shows
that droplets with a diameter larger than twice the wire
diameter are centered with respect to the wire, and the
droplet shape remains spherical. No tendency to slide
along the wire was detected. During the evaporation of the
liquid, the shape remains approximately spherical. It was
observed that high speed droplets are sliced by the wire.
The thin water film left on the wire causes a thermal
effect similar to the actual droplet. Of course, the
evaporation time for this thin film is substantially
shorter than an attached droplet, but no appreciable
difference in the peak electrical signal is detected.
As computations predict, the experimentally observed
surface tension effects are stronger than inertial
effects for small droplets. Droplets in the millimeter
diameter range experienced an acceleration of a few "g"
when touching a 10-micron diameter wire. Accelerations
of the order of 1,000 "g" are estimated for droplets
approximately lOym in diameter. The effect of eccentric
collision between droplets and the wire is not a problem
for small droplets moving at slow speed. For instance, at
a velocity of 3 m/sec (10 ft/sec), the electrical signal
for droplets lOOym in diameter was unaffected by
eccentricity collisions. For the same velocity, the 1 mm
diameter droplets produced electrical signals which were
significantly affected by eccentric collisions.
In order to study this effect of eccentric collisions, a
stream of 300 micron drops was generated with a vibrating
rod. The sensor was mounted with its axis perpendicular
to the trajectory of the droplets and moved sidewise to
achieve a variety of eccentric collision conditions.
The test results from one set of experiments in the labor-
atory are shown in Figure 6. The results are for droplets
22
-------
Droplet diameter 300 microns
Wire diameter 3.8 microns
\sec
.
/ * ^3 m/sec
t
Relative
electric
signal
let
--- * -- *=-.
\
-.8
- .6
. .4
.2
150
100
50 0 50
x - microns
100
150
Figure 6: Effect of Eccentric Collision
on Signal Output
23
-------
SOOym in diameter and for velocities of 1 m/sec. and
3 m/sec. The electrical signal is normalized with respect
to the peak amplitude corresponding to a perfectly
centered collision. Each of the plotted points represents
the average value from several experiments. The small
scattering of the data is due to small variation in
droplet size and fluctuation in the trajectory. The
recorded data for the eccentricity x £. 150ym was repeatedly
checked and verified. These small signals are due to the
aerodynamic flow attached to the droplets and was readily
eliminated from the signal processing by selective
filtering.
These studies determine the operating limitations for the
measurement of large droplets. For large droplets (near
600ym), the flow velocity must be low (1 to 2 m/sec) to
allow the droplet to center itself. For smaller droplets
the flow velocity may be substantially increased without
affecting the attachment mechanism. The limitation arises
from the air turbulence which, at high flow velocity,
induces high frequency electrical noise masking the
droplet signal. Tests performed on low turbulence flow at
50 meters per second with droplets 20 microns in diameter
did not show appreciable errors as compared to the
results obtained for a flow velocity of 3 meters per second.
Amplitude and Shape of the Electrical Signal
For the purpose of discussing the electrical response
of the sensor to interacting water droplets and in order
to identify the relevant parameters, this discussion is
restricted to hot wire sensors, excluding hot films.
Assume a wire of length I and diameter, d. The resistance,
R. of such wire at a temperature, t, is related to its
resistance, RQ, at a reference temperature by
R = RO [1 + k (t - t0)]
where k is the resistance temperature coefficient.
A thermal equilibrium condition is established between
the electrical heating of the wire and the heat transfer
to the gas stream. When a water droplet of diameter, D,
24
-------
attaches to the wire, the thermal equilibrium condition of
the wire-droplet is altered. Except for very small sizes,
the water droplet thermal capacity is much larger than the
thermal capacity of the wetted portion of wire. Also the
thermal conductivity of both water and wire material is so
high that, for all practical purposes, the temperature of the
wetted portion of wire is made equal to the water temperature.
Then the resulting voltage pulse, 6V, measured at the wire
terminal as a result of the droplet attachment is
6V = - RO I k J (t - t0)
where I is the electrical current flowing in the wire.
The same electrical current heats the droplet, delivering
an electrical power
which increases the droplet temperature and causes
evaporation.
A typical electric signal obtained during the droplet-
wire interaction is shown in Figure 7. The initial
fast decay during the droplet attachment depends on the
attaching mechanism and is completed in approximately
ten microseconds.
After attachment, the droplet is heated, increasing its
temperature almost linearly as a function of time until
evaporation and shrinkage takes place. The duration of
the signal is a function of the droplet size and the power
dissipated per unit length of wire. As a rule, the time
required for a complete evaporation is proportional to the
square of the droplet diameter. By increasing the wire
temperature, the evaporation time is decreased.
Another important factor was observed during this study;
large droplets are in general sliced through by the wire
as a result of dominance of inertial forces over surface
tension forces. The water layer which is momentarily
left on the wire covers a length of wire equal to the
droplet diameter. As a result, the peak electrical
signal during the attachment portion is almost unaffected
while the decay after the peak is substantially shorter.
25
-------
Output
Voltage
Droplet
Heating at
"Constant
Diameter
to
a\
Droplet Evaporation at
'Constant Temperature^
Shrinking Diameter
Peak
Time
Figure 7: Typical Sensor Electrical Output
-------
For the conditions of operation selected for the DC-1,
the average signal duration is of the order of four milli-
seconds. This time interval does not necessarily
reduce the sampling rate to four milliseconds between
droplets. Other droplets interacting with other portions
of the wire generate the proper peak electrical signal,
which occurs in the initial 10 microseconds. If two
droplets are attached to a small portion of wire, the
electrical signal is less than the calibrated amplitude
for each drop. The final configuration of the instrument
is designed to measure up to 500 droplets per second.
The sensitivity of the probe (ratio of peak voltage, V,
to captured droplet diameter, D) is given by
X = 51 (n-1)
D a
where
R is the probe operating resistance (hot)
I is the current flow in the sensor
H is the length of the heated wire
n is the overheat ratio § =1.5
Typical values of these parameters for a five-micron
platinum wire are:
R = 4 ohms
I = .04 Amps
SL - 1 mm
For these values, the sensitivity of the probe becomes
X = 12 (microvolts per micron)
D
27
-------
For droplets with D <_ 2d (d is the wire diameter) , this
law does not apply. For such small drops the signal ampli-
tude is reduced because the thermal inertia of the wire
is important as compared to the thermal inertia of the
droplet, resulting only in a partial cooling of the
sensor. The change in behavior for the small drops is
clearly shown by the attached calibration curve
(Figure 8). In the region above 20ym, the actual probe
sensitivity is less than the theoretical value because
of the electronic circuit (filtering and feedback power
amplifier) which has been introduced in the implementation
of the final design.
Hot Wire Versus Hot Film Sensor
The laboratory work with a probe constructed from a small
diameter wire (5pm) is always susceptible to mechanical
breakage during handling and use. Consequently, several
concepts were considered to make the probe more rugged.
The use of a hot-film probe appeared most advantageous
and its operating characteristics were investigated
experimentally in the laboratory.
The film probe is very rugged and operated for a long
period of time without any mechanical failure. However,
the response characteristic of the film probe is
inferior to the wire probe. Specifically,
1. The film probe has insufficient sensitivity
for small droplets. The electrical response
of the sensor to small droplets decays rather
rapidly for D < 2d. For the wire, such
conditions are reached for droplets below ten
microns. For commonly available films, almost
the entire specified range of droplet diameter
(lym to SOOym) falls under these unfavorable
conditions.
2. There is a large contact surface between the
droplets and film. Consequently, large
droplets tend to attach to the film and can
only be removed by evaporation. The evapora-
tion can require up to 1/10 of a second during
which time a portion of the probe cannot be
28
-------
ro
2.8
2.6
2.4
2.2
co 2.0
i-H
g 1.8
I
+j 1.6
3 1.4
0 1.2
4-)
O
$ 1.0
Q>
Q
X 0.8
0)
04 0.6
0.4
0.2
Window
No.
1
2
3
4
5
6
/
I
f
Size Range of
Windows (um)
i L
#1
1 -1.5
1.5 -2.25
2.25-3.4
3.4 -5.0
5.0 -7.5
>7.5
X
X
0 10 20 30 4
adder Mo
#2
1-2
2-4
4-8
8-16
16 - 32
> 32
X
#3
1-3
3-9
9-27
27 - 81
81 - 243
> 243
\/
/
0 50 60 70 8
Droi
/
'
X
X
X
^s
s
X
^x^
^
^"
.
'
3 90 100' 110 120 130 140 150 160 170 180 190 20
alet Diameter - Microns
Figure 8: Calibration Curve
-------
used for droplet sensing. Therefore, the film
probe tends to have a low sampling rate.
Because of these disadvantages, the wire probe is
considered more suitable for measurements in scrubbers
and demisters.
ANALYTICAL STUDIES
The analytical studies performed under this project
were used to interpret experimental results and to assist
in the selection of the material for the sensor.
Several studies of the transient thermal response of the
probe were performed.
Initially, a study of the radial temperature distribution
was undertaken assuming a one-dimensional analysis with
various boundary conditions representing the interface
between the wire and the droplet.
Using tungsten as the material for the transducers, the
results showed that the radial temperature distribution
becomes approximately uniform in less than 10"6 seconds
after the droplet attaches. Consequently, the assumption
that the wire is at the droplet temperature appears
justified in the vicinity of the droplet.
Another analysis was undertaken to investigate the longi-
tudinal heat transfer along the wire and to establish
the validity of the idealized model described in
Section IV. This analysis includes the heat transfer to
the surrounding air. The differential equation represent-
ing the problem is
pcpTt = k Txx + Q - h (T - Ta)
where
Q is a source term for the electrical heating
h (T - Ta) is the heat transfer from the
wire to the air
The equation was analytically solved for a droplet of
30
-------
generic size at some position xo along the wire. With
the general analysis, it is possible to study the longi-
tudinal heat transfer from the instant of droplet
attachment until complete evaporation. However, the
laboratory measurements are restricted to a few milli-
seconds after droplet attachment, and consequently,
the analysis was further simplified to this short-
time interval. Then the solution becomes:
e*P - TT - * dT
where
Y - JL. r_L.l
pcp [sec J
n (x) = £_
4ct [sec]
= JS_
PCP liee
6 = T-T0
9
-------
With this model, the time and space distribution of the wire
temperature was determined for three candidate materials:
tungsten, platinum and nickel. The results and conclusions
from the analytic study are:
1. An equilibrium longitudinal temperature distribu-
tion is established in a few milliseconds. The
radial cooling of the wire is much faster than
the longitudinal cooling.
2. By eliminating electrical signals at frequencies
below two kHz most of the effects of the longitu-
dinal cooling on peak signal amplitude are removed.
3. Nickel has the most suitable properties (e.g., low
value of diffusivity and low value of density),
leading to a minimum longitudinal cooling.
However, nickel wire is not readily available in
small diameters except for nichrome alloy, which
has an unacceptable low temperature coefficient.
Tungsten wire exhibited the worst thermal behavior
of the three materials. Platinum wire is a good
compromise, since its thermal response was almost
equivalent to nickel.
Tests performed in the laboratory with both tungsten
and platinum wire corroborate the analytical study.
USEFUL LIFE OF SENSORS
A variety of tests were performed to determine the useful
operating life of the sensors. Several sensors with
promising performance characteristics were subjected to
continuous operation under a spray operating with tap water.
The wire temperature of operation was set for a convenient
signal amplitude (*150°C) while the wire resistance was
monitored. Observation under a medium power microscope was
used to detect accumulation of contaminants and mechanical
damage.
Specifically, the following sensors were used:
3. 8 microns diameter tungsten wire
5 microns diameter tungsten wire
32
-------
20 microns diameter tungsten wire
.5 mm diameter ceramic rod with a platinum
film and quartz coating
3 microns diameter platinum wire
5 microns diameter platinum wire
In all cases accumulations of deposits on the wire were
detected after one or two hours operation. After ten hours
of operation a change in the operating characteristics was
detected. The hot wire resistance measured was decreased,
indicating a stronger cooling due to the dirt accumulation.
Observation under a microscope showed that the accumulation
was larger and less uniform on tungsten than on the other
sensors.
Tungsten wires with diameters of 3.8 and 5 microns lasted,
on the average, six and ten hours respectively before
breakage. Overheat values of 1.4, 1.5 and 1.6 were used
and a slight tendency to reduce the sensor life was
correlated to the higher overheat. The reasons for the
mechanical failure are not clear; gradual reduction on the
wire cross-section due to corrosion does not seem to be
significant since no important changes in the cold wire
resistance were noticed prior to breakage.
Platinum wire (3ym in diameter) lasted, on the average,
twelve hours before breakage. Five microns platinum wire
seldom broke even after fifteen hours of operation with
substantial accumulation of dirt.
Film sensors did not break and showed accumulation of dirt
after twenty-five to thirty hours of continuous operation.
EXPERIMENTAL APPARATUS
The laboratory experiments with the droplet sensor required
two types of basic equipment:
1) Apparatus for studying the attachment mechanism
of the droplet to the wire and a means of cali-
brating the sensor
33
-------
2) Equipment -to generate droplets in the size range
of interest.
In the following two sections, the equipment for the labora-
tory experiments is described.
Equipment for Studying Droplet Attachment and Calibration
The function of the apparatus is to provide information on
the diameter of the water droplet interacting with the
sensor together with the corresponding electrical signal.
The resulting correlation between the amplitude of the
electrical signal and the droplet diameter is the data
necessary to calibrate the sensor.
The approach adopted for these experiments was optical,
where the interacting droplet is observed under a microscope
during its attachment to the hot wire. A photographic
camera records the image of both the wire and the droplet,
and a permanent record is obtained to study the attachment
mechanism and to provide the calibration data. Since the
interaction time is short, an electronic flash is used to
stop the motion at the instant of droplet contact, thereby
no appreciable shrinkage due to evaporation takes place and
accurate calibration data is obtained. The duration of the
flash is approximately one microsecond, which is compatible
with movement and shrinkage of the droplet in the diameter
range of interest. The electronic flash provides a bright
background for the hot wire; both wire and droplet appear
dark with sharp boundaries. Several other illumination
arrangements were studied in the laboratory for dealing with
droplets which have a diameter less than 10vim. Such small
droplets are difficult to locate along the wire which is
itself Sum in diameter. Figure 9 shows the arrangement of
the main components. The photographic camera is a Polaroid
operating with Type 107 film. Magnifications up to 100 were
used; the limits being set by the need to photograph the
entire wire length (~1 mm) into the 10 cm diagonal linear
dimension of the photograph. A standard Leitz compound
microscope was used with a x50 magnification in the objec-
tive lens. The microscope specimen holder was adapted for
the support of the probe, and advantage was taken of the
micrometric positioning set-up to properly position the
wire in the microscope field.
34
-------
Probe
m
Illumination
Control
Electronic
Flash
Figure 9: Equipment Arrangement for Calibration
-------
The calibration of a sensor requires an oscilloscope picture
of the electrical signal taken simultaneously with the drop-
let picture. Such synchronization is accomplished with the
circuit shown in Fig. 10. A controlled power supply feeds a
resistance bridge, one of whose arms is the hot wire. The
electrical signal resulting from the interaction of the
droplet with the hot wire is filtered to remove low frequen-
cies due to air turbulence and it is displayed on an oscillo-
scope. The same signal is amplified and fed to a trigger
circuit which controls the oscilloscope sweep. The start of
the sweep is also used to trigger the electronic flash.
Every effort was made to reduce to a minimum the delay
between droplet to wire contact and operation of the flash,
in order to prevent distortion and/or evaporation of the
droplet. This apparatus was used to obtain data from
hundreds of droplets which varied in size from 5ym to 600pm.
Three such droplet photographs are shown in Figure 11 along
with the corresponding electrical signatures from the sensor.
Droplet Generators
The generation of droplets over the total specified size
range (lym to 600ym)with low velocity and low turbulence
could not be achieved using a single technique. In the
course of the investigation, the following approaches were
used:
Saturated steam
Vibrating capillary
Micropump
Rotating disc
Berglund Liu generator
The saturated steam was generated with an electrical heater
in contact with a metallic tube which had one end in water.
The water was supplied to the heater by capillary action.
This technique was used to generate droplet sizes below
ten microns.
The vibrating capillary consisted of a metallic tube put
into vibration with an electromagnet. An audio frequency
generator was the source of power for the electromagnet
and both amplitude and frequency were adjusted to achieve
36
-------
Adjustable Power
Bridge and
Filter
Oscilloscope
i
Hot Wire Probe
Trigger
Electronic
Flash
Figure 10: Calibration Circuit
-------
46 y droplet
Figure 11: Typical Droplets and Electrical Output
38
-------
mechanical resonance. Water was fed to the tube through a
restriction. The size of the droplet is a function of the
maximum acceleration at the tip of the capillary tube where
the droplet forms. The minimum droplet diameter generated
by this apparatus was fifty microns using the highest
acceleration. Repeatable droplets with a diameter of
approximately one hundred microns are achieved without
difficulty. The trajectory and separation of these droplets
remains consistent for extended periods of time. Because of
the consistency of this device, it was used in the study of
the eccentric collision phenomenon reported under the section
entitled Laboratory Experiments and Results.
The micropump droplet generator uses pressure surges on a
metallic tube to eject small droplets out of a restriction.
Such pressure surges were created with a flattened portion
of a brass tube being impacted with an electromagnetic
hammer. The trajectory of the droplet could be controlled
but a range of droplet diameters was always present. Typi-
cal diameters were between twenty to one hundred microns.
The rotating disc is a well-known device for generating
droplets. A jet of water is aimed near the center of the
rotating disc and the water slowly flows to the edges of
the disc where it breaks into small droplets because of
centrifugal forces. The disc is attached to an electrical
motor whose speed is varied to achieve different distribu-
tions of droplet sizes. A chamber surrounding the disc
allows for the selection of a small portion of the droplets
being generated. Droplet diameters from thirty microns to
larger than one millimeter were obtained with this device
in the laboratory.
The Berglund Liu monodisperse droplet generator is a rather
sophisticated instrument which ejects droplets through small
orifices as a result of the pressure surge created with a
piezo electric crystal. The crystal is activated from an
oscillator. Three orifice diameters are available: five,
ten and twenty microns diameter; the corresponding water
droplet diameters' are 13, 23 and 45 microns. The droplet
diameter is repeatable and the droplets can be ejected at a
low speed. The device was used extensively in our studies.
However, operation of the generator involves a learning
period and keeping the device in operation is tedious and
time consuming.
39
-------
In all the experimental work, the droplet sizes were actually
measured using the photographic records rather than assuming
the generators produced a monodispersion.
40
-------
SECTION VI
FIELD TESTING
Several facilities using scrubbers and mist eliminators were
investigated for possible use during the evaluation of the
instrument. The final selection was a group of mist
eliminators which were attached to scrubbers used to clean
the exhaust of diesel engines operating with either standard
diesel fuel or methane. This facility is located at the
Nassau County Sewage Treatment Plant, Wantagh,' Long Island,
New York. Five such units are installed and normally only
one operates at a given time.
A cross section of a demister is shown in Figure 12. Every
effort was made to minimize interference with the normal
operation of the plant and to avoid modification of the
demisters. Since the demister is located on top of the
building, all measurements were taken by lowering the probe
down the open stack which was approximately 5 feet high.
The probe was mounted on a pole which was flexible enough to
pass the spinner blades and reach the region where the spray
nozzles are located. A jig was clamped to the stack and
used to position and hold the probe in the desired region
of the flow field.
A power line was installed to operate the instrument and
auxiliary equipment. In particular, an oscilloscope was
used during most of the tests to determine the level of the
turbulence signal and the response of the droplet measuring
instrument. A magnetic tape recorder was also used to
record pertinent signals which then were analyzed in the
laboratory.
Great importance was placed on the actual use of the
instrument under field conditions. The basic purpose of the
field studies was to obtain performance data which was then
used to modify and improve the device. The major observa-
tions and results from these studies are summarized below:
1} It was observed that the droplet size population
at the demister chamber extended below the
specified minimum size of 10 microns. Also the
flow velocity was somewhat higher than expected,
41
-------
Spinner
De-entrainment
Liquor
Inlet
t
Flow
r
Spray
4 1/2-
Figure 12: Demister Cross Section
42
-------
reaching three meters per second (ten feet per
second) . The flow turbulence was also very high.
2) A program of modifications of the instrument was
carried out to satisfy these stringent operating
conditions. Most of the effort was for improving
the electrical filter characteristic to reject
signals generated by turbulence and to increase
the sensitivity for small diameter droplets.
These modifications were sucessfully made, and the
instrument now can measure droplets as small as
at a flow velocity up to 3 m/sec.
3) When using the instrument, the operator must
avoid strong turbulence conditions. The DC-1
circuit is designed to discriminate the low fre-
quency and low signal levels of the turbulence
from the high frequency signals from droplets.
However, under extreme turbulent flow conditions,
counting errors can occur.
4) Some secondary problems were noted during the
tests. For example, strong electrical inter-
ference caused false counting during a phase of
the field testing. The device was modified by
introducing a proper filter of the power line.
All such secondary problems are corrected in
the final instruments.
5) The probe operated successfully for many hours
of testing without failure. If a probe did
malfunction, it was caused by mishandling.
6) There were particles in the flow, but their
presence did not affect the measurement nor did
they break the platinum sensor.
7) The DC-1 measuring device is considered
operational and ready for field usage.
In the following three sections, the temperature, velocity
and droplet field measurements are briefly described.
43
-------
TEMPERATURE OF THE DEMISTER FLOW
The temperature of the flow did not change substantially
over an extended period of time, fluctuating around ±2°C.
At the axis of symmetry of the scrubber, a temperature of
56°C is observed, while close to the walls 50°C is reached.
Below the spinner the temperature increases to 62°C. These
results were compared against measurements performed with a
mercury column thermometer. Both measurements showed dis-
crepancy of up to two degrees. In the laboratory, the
agreement between the instrument temperature and the cali-
brated thermometer temperature was close to a degree.
MEASUREMENT OF VELOCITY
The probe was enclosed in a wire cage to protect the sensor
from strong impacts against the wall. The no-flow conditions
were determined by wrapping the probe cage in aluminum foil
and inserting it into the demister. The probe was held in
place until equilibrium conditions were reached. With the
function switch on VEL (velocity), the corresponding value
of reference resistance (Rnss) required to balance the
bridge was measured and recorded.
The shield was removed and the probe located at various
points in the stack flow to measure and record the corres-
ponding values of Rnsv; i-e-» tne reference resistance
which balances the bridge. From these measurements the
velocity is determined for each location.
For these measurements, the sensor wire should be perpen-
dicular to the flow since the cooling effect varies with
the cosine of the angle between wire axis and velocity vec-
tor. To perform this alignment the probe itself is used,
modifying at each point its angular location for a maximum
meter reading.
The flow distribution in the demister is rather complicated,
with a strong rotary motion superimposed on the axial motion.
Fluctuations in the velocity are present so average values
were measured over a period of approximately ten minutes.
The velocity at one foot from the wall was between 3.0 and
3.1 meters per second. The velocity at the center of the
stack was between 2.8 and 2.9 meters per second. Due to the
44
-------
operation of the diesel engine the flow shows a pulsation
of the order of ten cycles per second.
MEASUREMENT OF DROPLETS
With the instrument as originally designed (minimum droplet
size ten microns), a low counting rate was observed. Es-
timates of the droplet concentration were performed based
on the opacity of the plume, and it was concluded that
most of the droplets were smaller than ten microns in
diameter. The instrument was modified to satisfy those
conditions.
For most of the field study, a particular resistance
ladder was used to obtain measurements down to one micron.
The bin sizes were:
Lower Limit Upper Limit D^
Bin (yim) (ym) (urn)
1 1.0 1.8 1.40
2 1.8 2.5 2.15
3 2.5 6.0 4.25
4 6.0 40.0 23.00
5 40.0 160.0 100.00
6 160.0
Large fluctuations in the concentration were observed so
that the average from several measurements extending over
periods of ten minutes was computed. The time required
to achieve a total count of 999 droplets varies between
20 and 600 seconds.
The measurements were performed along the axis of symmetry
of the stack and 6 inches from the wall. Starting from
the top of the stack, measurements were made every two
feet until the spinner was reached. Table 1 presents
the data obtained from a typical sequence of measurements.
45
-------
stack
position
top
center
top
side
21
center
21
side
41
center
4'
side
6'
center
61
side
6 1/2 '
center
6 1/2 '
side
time
interval
(sec)
130
105
140
131
122
135
93
83
87
68
BIN 1
Nl
505
380
550
620
590
670
580
630
540
545
"l
160
150
160
196
204
210
200
315
260
332
BIN 2
N2
290
251
285
310
205
267
312
252
246
271
"2
81
89
74
87
63
74
124
113
104
148
BIN 3
N3
170
208
175
215
63
140
133
182
153
190
°3
32
50
31
41
12.5
25
36
55
44
70
BIN 4
N4
0
1
0
2
3
22
12
53
5
90
%
0
.6
0
1.3
2
.6
.6
2.6
.5
.8
BIN 5
N5
0
0
0
0
0
0
0
0
7
25
BIN 6
N6
0
0
0
0
0
0
0
0
0
7
TABLE 1 - Typical sequence of measurements
N^ = count, ifck bin; n^ concentration, ith bin [droplet/cm ]
-------
SECTION VII
REFERENCES
1. Calvert, Seymour, et al, Wet Scrubber System Study,
Vol. 1, NTIS (1972).
2. Perry, J. H., "Chemical Engineer's Handbook," 4th
Edition, McGraw-Hill, New York (1963).
3. Pigford and Pyle, "Ind. Eng. Chem," 43, 1649 (1951).
4. Harris and Haun, "Chem. Eng. Progress," 6_2_ (4), 55
(1966).
5. Johnstone and Williams, "Ind. Eng. Chem," 31, 993
(1939).
6. Johnstone and Silcox, "Ind. Eng. Chem.," 39, 808 (1947).
7. Herring and Marshall, Am. Inst. Chem. Engrs. J., 1.,
200 (1955).
8. Sherwood and Pigford, "Absorption and Extraction,"
McGraw-Hill, New York (1952).
9. Byrd and Dewey, "Chem Eng. Progress," 53, 447 (1957}.
10. Darrah and Frega, "Plant Engineering," 26, 15 (1972).
11. Green and Lane, Particleate Clouds, Dusts, Smokes, and
Mists," E. & F.N. Spon Ltd, London( 1964).
12. Akbutt and Kropp, "Determination of the Average Size of
Droplets to Design a Scrubber with a Venturi Tube,"
UDC 621.182.94.541.182.
13. Gifford, "Chem. Eng. Progress," 6£ (6) (1973).
14. Myers, "Fog," Scientific American, Dec. 1968, Vol. 219,
Number 6.
47
-------
15. "Drop-Size Distributions from Large-scale Pressure
Nozzles," K.Y. Kim (Monsanto Company, St. Louis, Mo.)
and E. Saunders (Tycohafts, Inc., Waltham, Mass.).
Preprint 17C, Second joint AIChE-IIQPR meeting, Tampa,
Fla., May 19-22, 1968.
16. "Atomization and Spray Drying," W. R. Marshall, Jr.
Chem. Eng. Progress Monograph Series No. 2, Amer. Inst.
of Chem Engrs. (1954).
17. "On Sprays and SprayingA Survey of Spray Technology
for Research and Development Engineers," W. E. Ranz
(Eng. Research, Pennsylvania State University),
Bulletin No. 65 (1956)
18. "The Penn State Bibliography on Sprays," 2nd Ed.
Published by The Texas Company, Refining Dept.
(December, 1953).
48
-------
TECHNICAL REPORT DATA
(Please read Inanietiont on the reverse before completing)
1 REPORTED.
EPA-650/2-75-018
2.
3. RECIPIENT'S ACCESSION>NO.
4. TITLE AND SUBTITLE
Design, Development, and Field Test of a Droplet
Measuring Device
5. REPORT DATE
February 1975
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Hector Medecki, Myron Kaufman, and Daniel E.
Magnus
9. PERFORMING OR6ANIZATION NAME AND ADDRESS
KLD Associates, Inc.
7 High Street
Huntington, NY 11743
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADJ-081
11. CONTRACT/GRANT NO.
68-02-1309
12 SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COV
Final; 6/73 Through 6/74
VERED
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
The report describes the design, development, and field testing of a concept to
measure liquid droplets in the size range from 1 to 600 micrometers. The measure-
ment probe is a platinum wire 5 micrometers in diameter. With the probe electri-
cally heated to a predetermined temperature, changes in the probe resistance are
related to the size of the impinging droplets in the gas flow. The electrical signals
from the probe are processed and used to classify the droplets into six different
size ranges or bins. Two prototypes, consisting of the probe and electronic proces-
sing unit, were constructed and tested in the field. Actual droplet distributions and
concentrations were measured successfully.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Drops (Liquids)
Size Determination
Air Pollution Control
Stationary Sources
13B
07D
14B
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThuReport)
Unclassified
21. NO.OF PAGES
ft
20 SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
49
------- |