United States Industrial Environmental Research EPA-600/7-79-180
Environmental Protection Laboratory August 1979
Agency Research Triangle Park NC 27711
Development of
Superior Entrainment
Separators
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-600/7-79-180
August 1979
Development of Superior
Entrainment Separators
by
Seymour Calvert and Harry F. Barbarika
Air Pollution Technology, Inc.
4901 Morena Boulevard, Suite 402
San Diego, California 92117
Contract No. 68-02-2184
Program Element No. EHE624
EPA Project Officer: Leslie E. Sparks
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
-------
ABSTRACT
An experimental and theoretical program was carried out to
develop an improved design for entrainment separators for scrubbers
The problems of separation efficiency, suspended solids deposition
and plugging of the entrainment separator were of primary concern.
A pilot scale entrainment separator (E.S.) coupled to a
scrubber and designed to handle a nominal gas flow rate of 1.4
m3/s (3,000 ACFM) was designed, built and tested. Vertical discon-
tinuous, zigzag baffles was the E.S. design selected after a re-
view of -.both theory and practical experience with slurry scrubbers.
The effect of E.S. performance on particulate emissions of a typi-
cal fossil-fueled boiler was evaluated.
The experimental program included measurements of entrainment
size distribution and loading, entrainment collection efficiency,
solids deposition character and rate, and E.S. washing efficiency.
Results were compared with available models and new criteria for
effective washing were developed.
111
-------
TABLE OF CONTENTS
ABSTRACT iii
FIGURES vii
TABLES xi
ABBREVIATIONS AND SYMBOLS xiii
ACKNOWLEDGMENT xvi
SECTIONS
1. SUMMARY AND CONCLUSIONS 1
Summary 1
Conclusions 5
2. INTRODUCTION 11
Application to Slurry Scrubbers 12
Background 13
Other Programs 13
Conclusions from EPA Studies 14
Objectives 14
3. ENTRAINMENT SEPARATOR DESIGN FOR PILOT PLANT 17
Design Approach 17
Background Information 17
Emission Limit 19
Entrainment Separator Types 20
Performance of Selected Entrainment Separators. ... 21
Scrubber/E.S. Performance Balance 21
Final Selection and Performance Prediction 25
Pilot Plant 32
Scrubber/Entrainment Separator Shell 34
Entrainment Separator Design 37
Scrubber Operating Conditions 37
Drop Disengagement in Vertical Duct 41
Particle Collection Efficiency of Turning Vanes . . .41
IV
-------
TABLE OF CONTENTS, continued
Gravity Settling in the Horizontal Duct 43
Overall Penetration Through Duct 46
4. ENTRAINMENT CHARACTERIZATION 49
Drop Sizing Techniques 50
Range of Drop Sizes Expected 50
Review of Drop Sizing Methods 50
Sizing Methods Chosen. 51
Entrainment Data 54
E.S. Drainage Data 54
Sphere Drainage Data 56
Concentration from C.I. Measurements 56
Auxiliary Sizing Techniques 58
Entrainment Drop Size Distributions 58
Types of Size Distributions 59
Cascade Impactor Data 60
Size from E.S. Penetration 62
Size Distributions from Sphere Collector Data 72
Expected "a " of Large Drops .76
o
Sphere Data Analysis 77
Nozzle Manufacturers' Drop Size Data 79
Conclusions 81
5. ENTRAINMENT SEPARATOR PERFORMANCE 91
E.S. Configurations 91
Performance Measurements 92
Comparison with Theoretical Model 95
Pressure Loss 95
Reentrainment Characteristics 97
E.S. Outlet Solids Emission Levels 98
6. SOLIDS DEPOSITION 100
Introduction 100
Slurry 100
Procedures 100
-------
TABLE OF CONTENTS, continued
Measurements 101
Results 104
Comparison with Solids Deposition Model 106
Conclusions 110
Operating Experience 110
7. WASHING SYSTEM 114
Introduction 114
Needs and Requirements 114
Requirements/Constraints 115
Experimental Design 117
Results 117
Washing Flux Calculation 117
Minimum Flux Required 119
Comparison with Prediction 121
Design Criteria 121
Typical Design 121
8. FUTURE RESEARCH RECOMMENDATIONS 124
Grade Efficiency 124
Reentrainment Limit 125
Solids Deposition Model 125
Demonstration Program 126
REFERENCES 131
APPENDIX "A" 134
VI
-------
FIGURES
Number Page
1 Performance cut diameter for several types of
entrainment separators 22
2 Cumulative solids concentration from a mobile bed
scrubber for two levels of entrainment concentration. 24
3 Top view diagram of example zigzag baffles 27
4 Predicted grade efficiency of 3 rows of zigzag
baffles 29
5 Predicted grade efficiency of 6 rows of zigzag
baffles 29
6 Predicted grade efficiency of zigzag baffles at 4.6
m/s superficial gas velocity 30
7 Predicted overall penetration for 6 rows of baffles . 31
8 Measured overall collection efficiency of vertical
zigzag baffles 33
9 Reentrainment limits for 6 rows of vertical zigzag
baffles (data by Calvert, et al., 1974) 33
10 Schematic diagram of scrubber/entrainment separator
system 35
11 Scrubber/E.S. system 36
12 Scrubber/entrainment separator system ........ 38
13 Gravity settling velocity of water drops in air
(Fuchs, 1964) 42
14 Theoretical centrifugal deposition velocity of water
drops in air 44
15 Theoretical grade efficiency of turning vanes .... 44
16 Theoretical grade efficiency of horizontal duct ... 48
VII
-------
FIGURES (continued)
Number Page
17 Theoretical overall penetration through turning
vanes and horizontal duct of log-normally distri-
buted drops ..................... 4°
18 Photograph of glass cascade impactor ........ 52
19 Body collector drop sizing system .......... 57
20 Theoretical overall penetration through zigzag
baffles of drops with cubic distribution ...... 64
21 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the mobile bed (mode 1)
scrubber ...................... 66
22 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the mobile bed (mode 2)
scrubber ...................... 66
23 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the mobile bed (mode 3)
scrubber ...................... 67
24 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the mobile bed (mode 4)
scrubber ...................... 67
25 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the mobile bed (mode 5)
scrubber ...................... 68
26 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the sieve high scrubber. . . . 68
27 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the sieve low scrubber .... 69
28 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the spray 5 scrubber ..... 69
29 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the spray 7 scrubber . . „ . . 70
30 Measured entrainment concentration vs. drop diameter
at the E.So inlet for the spray 4 scrubber . „ . . . 70
31 Measured entrainment concentration vs. drop diameter
at the E.S. inlet for the spray 6 scrubber ..... 71
viii
-------
FIGURES (continued)
Number Page
32 Theoretical overall efficiency of sphere collectors
on log-normally distributed drops 75
33 Theoretical overall efficiency of sphere collectors
on drops with the cubic distribution ?5
34 Calculated outlet drop size distribution of turning
vane/horizontal run '8
35 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for mobile
bed (mode 1) scrubber 85
36 Composite drop size distribution at entrainment
separator inlet and at scrubber outlet, for mobile
bed (mode 2) scrubber 85
37 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for mobile
bed (mode 3) scrubber. . 86
38 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for mobile
bed (mode 4) scrubber 86
39 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for mobile
bed (mode 5) scrubber 87
40 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for the
sieve high scrubber 87
41 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for sieve
low scrubber 88
42 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for spray 5
scrubber 88
43 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for spray 7
scrubber 89
IX
-------
FIGURES (continued)
Number Page
44 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for spray
4 scrubber ..................... 89
45 Composite drop size distributions at entrainment
separator inlet and at scrubber outlet, for spray
6 scrubber ..................... 90
46 Comparison of E.S. size distributions of mobile
bed, mode 1 vs. data of Calvert, et al. (1977). . . 90
47 Measured overall penetration correlated to cumula-
tive mass fraction of drops smaller than 20 pmA . . 94
48 Measured deposition thickness rate for spray
scrubber operation ................. 105
49 Measured deposition thickness rate for sieve plate
scrubber operation ................. 105
50 Comparison of measured and predicted deposition
rate for spray scrubber run sets 2 and 3 ...... 109
51 Comparison of measured and predicted rates for the
sieve plate scrubber operation ........... 109
52 Slurry flux for maximum deposition rate
53 Washing flux vs. percent of baffle surface cleaned
after 10 minutes of u~ = 2.3 m/s ..........
54 Washing flux vs. percent of baffle surface cleaned
after 10 minutes of UG = 3.4 m/s .......... 118
55 Proposed program schedule ............. 130
-------
TABLES
Number Page
1 Summary of Experimental Conditions .......... 4
2 Mobile Bed Scrubber Operating Conditions ....... 39
3 Single Sieve Plate Scrubber Operating Conditions. . . 39
4 Spray Scrubber Operating Conditions ......... 40
5 Superficial Velocity in Entrainment Separator .... 55
6 Average Size Distributions Based on Cascade Impactor
Data at E.S. Inlet .................. 61
7 Measured Overall Penetration Through E.S ....... 63
8 Size Distributions Derived from E.S. Penetration Data 65
9 Overall Efficiency of Sphere Collectors ....... 73
10 Size Distributions at E.S. Inlet Derived from Sphere
Collector Data .................... 80
11 Summary of Performance of the Entrainment Separator
Based on Cascade Impactor Sampling .......... 93
12 E.S. Outlet Solids Emissions Based on Drop
Penetration ..................... "
13 Calculated Slurry Flux on Rows 1 Through 6 ...... 108
14 Calculated Penetrations of Wash Spray Drops ..... 12°
15 Predicted Total Washing Rates for Typical E.S. for
Ten-Minute Operation ................. 123
16 Program Cost Estimate ................ 129
XI
-------
TABLES (continued)
Number Page
Appendix
A-l Entrainment Separator Inlet Data Mobile Bed
Scrubber Operation 135
A-2 Entrainment Separator Inlet Data Sieve Plate
Scrubber Operation 138
A-3 Entrainment Separator Inlet Data Spray
Scrubber Operation 140
A-4 Entrainment Separator Outlet Data Mobile Bed
Scrubber Operation 142
A-5 Entrainment Separator Outlet Data Sieve Plate
Scrubber Operation 144
A-6 Entrainment Separator Outlet Data Spray
Scrubber Operation 145
XII
-------
LIST OF ABBREVIATIONS AND SYMBOLS
Latin
a - constant in sphere collection efficiency equation
a - centrifugal acceleration, m/s2
AD - deposition area, cm2 or m2
b - section width per baffle, m
C1 - Cunningham slip correction factor, dimensionless
CD - drag coefficient, dimensionless
C - drop concentration hitting row n, ml/m3
CT - total drop concentration, ml/m3
d, - drop diameter, ym, (or ymA, cm, m)
dj - geometric mass mean drop diameter, ym
d(j5Q - drop diameter collected with 501 efficiency, ym
cL,,,, " cubic distribution maximum drop diameter, ym
m 3.x.
d - cubic distribution mass median drop diameter, ym
d - particle diameter, ym (or ymA, cm, m)
P L*
d - particle aerodynamic diameter, ymA = ym (g/cm3)^
pa
d - geometric mass mean particle diameter, ym
d - mass median particle diameter, ym
f - distribution density function, dimensionless
f , - cubic density function
£TJ, - log-normal density function
f - pressure loss coefficient, dimensionless
f - volume fraction
f - deposition thitkness rate correction factor
F , - cubic distribution function
F - spray nozzle distribution function
noz r
g - acceleration of gravity, m/s2 or cm/s2
h - height, cm or m
k - factor in solids deposition equation
K - particle impaction parameter, dimensionless
Xlll
-------
R<
T
u
u
u/
D
Dl
LIST OF ABBREVIATIONS AND SYMBOLS (continued)
Latin
L - length, cm or m
L - critical length, cm or m
n - number of rows of baffles
n, - number of baffles in a row
NR - Reynolds number, dimensionless
Pt - penetration, fraction
PT - overall penetration, fraction
QG - gas volumetric flow rate, m3/s
QL - liquid volumetric flow rate, m3/s or £/s
r - radius of curvature, cm or m
v_.
solid deposition rate, mg/cm2-s
temperature, °C or °K
deposition velocity, m/s or cm/s
deposition velocity on first row, m/s or cm/s
superficial gas velocity, m/s or cm/s
actual gas velocity between baffles of row i, m/s or cm/s
terminal gravity settling velocity, m/s or cm/s
br
ut - turbulent eddy velocity, m/s or cm/s
w - baffle or duct width, cm or m
W - solids weight fraction in slurry
x - log-normal distribution parameter
Greek
6 - film thickness, ym
AP - pressure loss or drop, Pa or cm W.C.
e - porosity or void fraction
n - efficiency, fraction
r\ - overall efficiency, fraction
9 - baffle angle, degrees or radians
yG - gas viscosity, kg/m-s or g/cm-s
HT - liquid viscosity, kg/m-s or g/cm-s
VG - gas kinematic viscosity, m2/s or cm2/s
p^ - drop density, kg/m3 or g/cm3
uGi
u^.
xiv
-------
LIST OF ABBREVIATIONS AND SYMBOLS (continued)
Greek
PG - gas density, kg/m3 or g/cm3
PT - liquid density, kg/m3 or g/cm3
LJ
p - particle density, kg/m3 or g/cm3
p - solid cake density, kg/m3 or g/cm3
a - geometric standard deviation of diameter
T - deposition thickness rate, mm/hour
T' - modified deposition thickness rate, mm/hour
$ - slurry flux, mg/cm2-s
*™ov ' maximum slurry flux, mg/cm2-s
HI 3.X
* - slurry flux on row n, mg/cm2-s
$ - washing flux, mg/cm2-s
W
Abbreviations
ACFM - actual cubic feet per minute
BTU - British thermal unit
C.I. - cascade impactor
E.S. - entrainment separator
FGD - flue gas desulfurization
GPM - gallons per minute
i - liter
N - normal conditions : 20°C, 101.325 kPa
NSPS - new source performance standards
psig - pounds force per square inch, gauge
SCF - standard cubic foot : 20°C, 101.325 kPa
W.C. - length of water column
\,
ymA - microns, aerodynamic, ym (g/cm3)
xv
-------
ACKNOWLEDGEMENT
A.P.T., Inc. wishes to express its appreciation for excellent
technical coordination and for very helpful assistance in support
of our technical effort to Dr. Leslie E. Sparks, EPA Project
Officer.
xvi
-------
SECTION 1
SUMMARY AND CONCLUSIONS
SUMMARY
Entrainment separation (or mist elimination) has been a
major operating problem in scrubber systems, especially those
in which dissolved and suspended solids are present in the
scrubber liquid, as in lime/limestone scrubbers. The program
reported here was initiated by the Particulate Technology Branch
of IERL/RTP, EPA in order to develop a superior entrainment
separator for FGD lime/limestone scrubbers.
Entrainment separator (E.S.) requirements and performance
have been studied in a sequence of research and development pro-
grams under several EPA contracts. The results of this work and
information from the literature were used as the basis for an
experimental program to develop a superior ES for lime/limestone
scrubbers. The problems of ES efficiency, suspended solids
deposition, and washing were of major concern. The effect of
chemical reactions and the resultant precipitation of solids
were not included among the parameters studied. It has been
shown that proper control of FGD "system chemistry" can essen-
tially eliminate the formation of tightly cemented scale which
cannot be removed by washing. Consequently, the present study
was restricted to the use of suspended solids.
Pilot Plant
A pilot-scale scrubber and ES system was built to handle a
nominal gas flow rate of 1.4 m3/s (3,000 ACFM). Several scrub-
bers, typical of those which might be encountered in FGD systems,
were used so that the entrainment characteristics would realis^
tically represent what might be encountered. A single stage
mobile bed ("ping pong ball"), 1-sieve plate, and spray scrubber
-------
were used in this program.
Design criteria for the superior E.S. system were estab-
lished by considering the effect of the E.S. on particulate
emissions from a typical fossil fuel boiler. Estimates of the
uncontrblled emissions characteristics and the probable new
source performance standards were used to predict the overall
efficiency required. Outlet particle concentration depends on
both the scrubber and the E.S. acting in series. Consequently,
the scrubber and E.S. performances were balanced to define the
overall system which would achieve the necessary efficiency with
the least power consumption.
The basic theoretical and experimental work on major types
of E.S. was carried out on EPA contracts and was presented by
Calvert et al. (1974). This research covered the types of E.S.
available, drop collection efficiency, gas handling capacity,
pressure drop, reentrainment, and solids deposition. Reports
of operational experience at the EPA alkali scrubbing facility
at the Shawnee Power Station, Paducah, Kentucky detailed many
problems with E.S. when used with lime/limestone slurry scrubbers
Consideration of the overall system, discussed above, led
to the conclusion that the E.S. should have an efficiency char-
acterized by a cut diameter (drop collected at 50% efficiency)
of about 15 to 20 ym , corresponding to a pressure drop of
about 2 cm W.C. A zigzag-type E.S. with vertical or sloping
blades and horizontal gas flow was selected for this service.
In order to reduce the liquid loading on the E.S. and thereby
to increase the allowable gas velocity before reentrainment
and to decrease the amount of solids deposition, some "pre-
separation" features were employed as fallow:
1. Sufficient "free board" was used above the scrubbing
zone to permit large drops to settle out or reach
the scrubber walls and flow down. A vertical dis-
tance of 137 cm was used. Drop diameters with
settling velocities corresponding to the upward air
velocities used were 700 to 950 pm for the mobile
-------
bed, 300 to 400 ym for the sieve plate, and 600 ym
for the spray scrubber.
2. A 90° turn from vertical to horizontal gas flow was
fitted with curved turning vanes. Large drops were
collected on the turning vanes and flowed back to
the scrubbing liquid tank. The turning vanes had a
predicted drop cut diameter of 65 ym.
3. A horizontal flow section of sufficient length
(235 cm) for substantial drop sedimentation was used
between the end of the 90° turn and the upstream face
of the E.S. The predicted cut diameter of the hori-
zontal run was 180 ym and the total for freeboard,
bend, and horizontal run was 55 ym.
Experimental Program
The experimental program included measurements of entrain-
ment size distribution and concentration, drop collection effi-
ciency, solids deposition character and rate, and E.S. washing
efficiency. Flow rates and pressure drops over various seg-
ments of the apparatus were also measured. The scrubber had
a 76 cm square cross section in which different types of
internal structure could be placed to provide the mobile bed,
sieve plate, and spray scrubber entrainment emissions.
The E.So was built in row modules to allow testing with
1 to 6 rows of baffles. Both vertical and forward sloping
baffles were tested in 3-row and 6-row configurations. The
experimental conditions that were studied are listed in Table 1.
A washing system consisting of sprays of recycled slurry
placed upstream of the E.S. was used. The wash period, fre-
quency, drop size, and liquid flow rate were varied to deter-
mine the minimum requirements for washing.
Evaluation and Modeling
Experimental results were evaluated to determine how
well the theoretical models predicted performance. The drop
collection efficiency model agreed generally with the data.
-------
TABLE 1. SUMMARY OF EXPERIMENTAL CONDITIONS
Scrubber
Mobile Bed - 1
Mobile Bed - 2
Mobile Bed - 3
Mobile Bed - 4
Mobile Bed - 5
Sieve Plate-low
Sieve Plate-high
Spray - 5
Spray - 7
Spray - 4
Spray - 6
vv
Vm3
6.5
6.0
4.5
7.4
8.4
11.3-13.4
8-9.5
1.9
2.6
1.8
2.7
E.S. Velocity,
m/s
6.9
6.9
6.9
6.1
5.3
2.3
3.4
4.6
4.6
4.6
4.6
Scrubber Velocity,
m/s
3.7
3.7
3.7
3.3
2.8
1.2
1.8
2.44
2.44
2.44
2.44
Type of Runs
Blank (B), 3 vert.
rows (3V), 6V,
3 sloping (3S), 6S
Drop size and "Pt"
B, 3V, 6V, 3S, 6S
Drop size, "Pt", and
solids deposition
Same as sieve plate
-------
The inlet drop size distribution and concentration varied
greatly over the cross section of the duct and because of
this and the radical difference between the inlet and outlet
drop size distributions, the accuracy of the model could not
be evaluated with much precision. Reentrainment was not
observed for the 6-row E.S. configurations, as would be ex-
pected from previous studies. The measured pressure drops
agreed with those from previous studies and a mathematical
model was developed.
Locations where heavy deposition of solids occurred were
predicted with the solids deposition mathematical model,
although the deposition rate could not be. Some improvements
in the model were made and it is accurate enough to predict
potentially troublesome conditions and locations.
The required washing rate was determined experimentally
and empirical criteria for satisfactory washing were developed.
The continuous flow solids deposition model was not rigorously
applicable to the experimental case in which washing was inter-
mittent, but it could be applied to predict a minimum washing
flux. General trends can be predicted but not precise flow
rates.
CONCLUSIONS
The experience and information gained in the course of
this program have led to the formulation of a number of con-
clusions. Some of these relate to the research objectives,
some to the methodology, and some to problems requiring
study. For clarity the conclusions are grouped under topical
headings below.
Performance Requirements
1. An E.S. drop cut diameter of 15 to 20 ym generally is
adequate for lime/limestone scrubbing to meet the EPA NSPS of
13 mg particulates/MJ. (For water drops >5 ym diameter, physical
and aerodynamic size are practically equal).
-------
2. Pressure drop, or power requirement, for the combina-
tion of scrubber and E.S. can be minimized with an E.S. pres-
sure drop of 2 to 3 cm W.C.
Pre-Separation
1. Pre-separation of entrained drops upstream of the E.S.
can significantly reduce the liquid/gas ratio and, thereby, can
increase the gas velocity for re-entrainment. Ample freeboard
between the scrubber and the next element, turning vanes, and
a horizontal run were used in series for pre-separation and
gave a drop cut diameter of 55 ym.
2. Turning vanes were the most effective pre-separator
and had a predicted drop cut diameter of 65 ym.
Primary Collection Efficiency
1. The vertical, discontinuous zigzag baffle E.S. (Z.Z.E.S.)
used with horizontal gas flow in this study had a drop cut dia-
meter of about 20 ym and removed drops larger than 20 ym physical
diameter over the range of flow rates investigated.
2. A new mathematical model developed in this study for
predicting the primary collection efficiency of zigzag baffles
is basically the same as our previous model but easier to com-
prehend. The effects of drop deposition area and other design
parameters are directly apparent in the new model.
3. A Z.Z.E.S. with baffles sloped from vertical and with
horizontal gas flow had essentailly the same primary collection
efficiency as the vertical baffles.
Reentrainment
1. Both the vertical and sloping baffle, 4-row Z.Z.E.S.
had adequate reentrainment characteristics. They did not show
drop reentrainment at the limits of pilot plant flow rates
(i.e., 8.4 m/s air velocity and 2.4 1/min3 slurry/air ratio).
2. While four rows of baffles have high primary collec-
tion efficiency, they do not have adequate reentrainment char-
acteristics to handle high entrained liquid flow rates. The
4-row Z.Z.EoSo exhibited reentrainment at 6.9 m/s air velocity
-------
and 3 1/m3 slurry/air ratio.
Solids Deposition
1. Solids deposition is inevitable at some point in the
E.S. where the slurry flux to the baffle decreases to a rate
which is not "self washing." The predicted minimum rate of
washing was approximately 10~k cm3/cm2-s and the flux to cause
solids deposition on the top 10 cm of a vertical baffle was
approximately 2 x 10~5 cm3/cm2-s.
2. The mathematical model for solids deposition on verti-
cal surfaces which was developed previously is adequate for
estimating the conditions under which deposition will occur.
3. No solids deposition was observed on the backs of
baffles, as had been in previous experiments on a single baffle.
The gas flow pattern in a multiple baffle Z.Z.E.S. appears to
prevent the wake eddy which brought small drops to the back
surface of a single baffle.
Washing
1. Solids deposited without cementation (i.e., due to
suspended, not dissolved, solids can be removed from baffles by
means of intermittent washing with coarse sprays of slurry on
the upstream face of the E.S.
2. Effective washing with slurry requires that the slurry
flux reaching the last baffle be higher than the minimum
self washing rate.
3. Experimental results confirmed the utility of the
mathematical model for solids deposition in estimating minimum
washing rates.
Drop Sizing Methods
1. A cascade impactor method using a salt tracer provided
good drop size distribution data for diameters between 1 and
20 unio
2. The overall penetration and outlet size distribution
data from the E.S. were correlated to log-normal and cubic
-------
inlet size distributions using a mathematical model for E.S.
grade penetration. The correlations were very good in a few
cases and moderately good in most cases.
3. A new method was developed for sizing drops by means
of spherical and cylindrical collection bodies. The volumes
of drops collected by the sphere collectors were correlated
to log-normal and cubic inlet size distributions using a
mathematical model for sphere collection efficiency.
4. The hot wire drop counter did not perform successfully
under the conditions of these tests. The fragility of the 5 ym
diameter wires and the use at high gas velocities were the prob-
able causes of the poor performance.
Entrainment Size Distribution
The size distributions of the entrainment produced from
the three types of scrubbers are represented graphically as the
average of the data based on cascade impactor, sphere, and
E.S. sampling at the E.S. inlet.
Comparison with Other Data
The only entrainment size distribution data for scrubbers
and conditions similar to those used here that could be found
were those of Calvert, Yung and Sparks (1977). The present
data show a drop concentration about double that of the 1977
measurements for drops smaller than 20 ym diameter.
Future Research Recommendations
The major goals of this research were met, however, a few
problems remain to be more fully resolved. These include a
more precise determination of the entrainment separator grade
efficiency and reentrainment limits and more experimental work
to improve the solids deposition model.
-------
The superior E.S. system should undergo a demonstration
phase to prove that the design can maintain high performance
and not be degraded by a slurry in an actual practical appli-
cation. Demonstration at a coal-fired utility boiler that
uses a lime or limestone flue gas desulfurization scrubber
would be desired. The design gas flow rate should be an
order of magnitude greater than the present; i.e., at least
14 actual m3/s (30,000 ACFM).
Report Guide
The following report is devoted mainly to describing the
experimental program, its results, our analysis, and conclusions.
While only one type of E.S. was used, the program yielded some
results which have general applicability. To assist the reader
in locating the kind of information he needs, an overview of
the major subject categories of the report are given below.
Section 2 presents a brief summary of the background work
in the field and the objectives of this program.
Section 3 gives the design basis and rationale for the
pilot plant used in this program. The performance characteris-
tics required of the E.S. were established upon consideration
of the overall emission limits and the characteristics of FGD
scrubbers.
Methods for predicting the primary collection efficiency
and re-entrainment conditions for a given zigzag baffle design
are given in "Final Selection and Performance Prediction, page 25.
The design rationale and details of the experimental pilot
plant are given in "Pilot Plant," page 32. Pre-separation effects
on drop size distribution are predicted in this section.
Section 4 deals with the measurement of drop size and
concentration. Several methods were used and the results are
compared for a number of experimental conditions.
Section 5 presents experimental results on E.S. performance
and compares them with theoretical predictions. Penetration,
pressure drop, reentrainment, and outlet solids emissions are
discussed.
-------
Section 6 covers the experiments and analysis of experi-
mental results on solids deposition. Comparison of experimen-
tal results and a mathematical model illustrates the use and
limitations of the model.
Section 7 describes the E.S. washing system, experiments,
results, and analysis of data.
Section 8 gives the recommendations for future research.
10
-------
SECTION 2
INTRODUCTION
Entrainment separators, or mist eliminators, are widely
used to prevent the carry-over of liquid drops from scrubbers.
The liquid drops are the result of the thorough and vigorous
liquid-gas contacting in the scrubber which atomizes and entrains
some of the scrubbing liquor. Undesirable particulate emissions
as well as numerous operational problems may occur if this
entrained liquid is not removed from the gas stream.
The liquid entrainment will generally contain both suspended
and dissolved solids. The suspended solids may be the particles
collected by the scrubber, substances carried in the scrubbing
liquid, or products of chemical reaction occurring within the
scrubber. Similarly, the dissolved solids may come from the
impurities in the gas, reagents carried in the scrubber liquid,
or products of reaction.
The residue remaining after the emitted entrainment drops
have dried is a particulate emission into the atmosphere. The
composition of such residues can be quite different from that
of the particles entering the scrubber, especially where reactive
solutions or slurries are used for gas scrubbing. Thus, a
consequence of excessive entrainment carry-over can be that the
particulate concentration in the effluent gas within a certain
size range could be higher than that in the gas entering the
scrubber.
Many operational problems can arise if the entrainment
level is excessive. The drops can be responsible for corrosion,
erosion, and ultimately, mechanical failure of fan blades or
housing. Pools of liquid and/or residual solids can also deposit
in ductwork, on heat-exchange surfaces, and in the stack, causing
11
-------
corrosion or plugging. In the area immediately surrounding
the point of emission harmful and unsightly "rain-out" of liquid
drops can occur.
Because of the emissions and operational problems that arise
from excessive levels of entrainment, the entrainment separator
is an important component of the air pollution control system.
In many applications it is no less important than the scrubber
itself since its inefficiency can allow reemission of particles
collected and generated in the scrubber. And, should the
entrainment separator fail to remove enough corrosive, erosive,
or solids-laden drops, operational problems will occur to
seriously affect the availability of the scrubber.
APPLICATION TO SLURRY SCRUBBERS
Many scrubbers which are designed for particle collection
use fresh or fairly clean recirculated water as the scrubbing
liquid. The solids in the form of collected particulates and
impurities would, in this case, usually constitute less than
1$ of the liquid mass, depending on the particle loading and
the efficiency of the cleaning of the recycled liquid. Thus,
the design of the entrainment separator (E.S.) need be mainly
concerned with collecting liquid drops. However, if a considerable
percentage of suspended solids is present in the scrubbing
liquid, one must pay attention to the possibility of solids
deposition and accumulation.
Slurries are used primarily when a gaseous component of the
effluent stream is to be scrubbed. Chemically reactive slurry
solids are used to increase the mass transfer rate and capacity
of the liquid. One of the most widely used slurry scrubbers is
the lime/limestone system for controlling both particulates and
sulfur oxides from coal-fired utility boilers. In these scrubbers
the suspended solids concentrations are often in the range of
10 to 20% by mass so that the potential for plugging and
other operational problems, as well as high particulate emissions
is great.
12
-------
BACKGROUND
Because of the increasing usage of lime/limestone scrubbing
for flue gas desulfurization (FGD), the U.S. Environmental
Protection Agency (E.P.A.) has carried out several research and
development programs aimed at developing our capability and
engineering knowledge in this subject.
In the "Scrubber Handbook" prepared under E.P.A. contract
by Calvert et al. (1972), a subsection was devoted to entrain-
ment separators. It was noted that entrained drops can be
removed from the gas stream by a variety of devices similar
in principle to particle collecting scrubbers. The devices for
which some theoretical or experimental collection information
was available were gravity settlers, knitted wire mesh, rods
arranged in staggered rows, and cyclones.
Because the available design information was so scarce
the E.P.A. contracted with Air Pollution Technology, Inc. (A.P.T.)
in 1973 to conduct extensive theoretical and experimental studies
of entrainment separators. The study included an evaluation of
current technology, experimental studies in a pilot-scale system,
and development of theoretical performance models.
Zigzag baffle, knitted mesh, tube bank, packed bed, and
cyclone devices were tested. Preliminary studies of the mechanisms
for and effects of solids deposition were also made. Two reports
were issued for this contract- Calvert et al. (1974), and
Calvert et al. (1975). In addition, the results have been
reported in the literature by Calvert et al. (1974b), Calvert
et al. (1977), Calvert (1978a), and Calvert (1978b).
Other Programs
Since 1972 E.P.A. has been operating a prototype FGD test
facility at the Tennessee Valley Authority's Shawnee power
plant near Paducah, Kentucky. A number of reports have given
details of the operational problems; e.g. Epstein (1975) and
Head (1976).
13
-------
Mist eliminators have also been listed as a major problem
area for lime/limestone FGD systems by the Electric Power
Research Institute (EPRI) - see Nannen and Yeager (1976). One
result of EPRI's efforts to investigate entrainment separator
problems has been a study on mist eliminator design by Conkle
et al. (1976).
Conclusion of EPA Studies
Several conclusions were drawn by Calvert et al. (1975) con-
cerning primary collection efficiency, capacity, nature of re-
entrainment, pressure drop, and solids deposition. And, as a
result of the studies, it was thought that "Entrainment sepa-
rator design or specifications by means of rational methods is
possible to a useful degree." However, "Several important areas
require further study before the state of knowledge will be
adequate for the reasonably thorough and accurate design of an
entrainment separator. Some of these are:
1. Reentrainment mechanism and loading for separators
under various operating conditions,
2. Entrainment loading and drop size distribution from
various scrubbers under different operating conditions.
3, Solid depositions and factors affecting the deposition
rate.
4„ Effective separator washing method and flow rate of
washing liquid,"
The program that is reported here was undertaken to design
and test a superior E.S. for lime/limestone scrubbers and to
develop better information for the design of E.S.s. The latter
three areas mentioned above were specifically studied in detail.
Fulfilling these needs as well as helping to solve the persis-
tent operational problems with slurry scrubber FGD entrain-
ment separators were goals of this program.
OBJECTIVES
The objectives of the program included the following:
14
-------
Design and Build Pilot Plant
Design and build a pilot scale (1,000 to 3,000 ACFM)
entrainment separator, coupled with a scrubber.
Experimental Study
Conduct a pilot scale experimental program using the pilot
scale entrainment separator to:
1. Determine locations in the entrainment separator where
liquid and entrained solids tend to concentrate.
2. Develop washing techniques to effectively clean the
entrainment separator, especially those locations
where liquid and solids concentrate. In developing the
washing techniques the contractor shall keep in mind
that the amount of fresh water available for washing
field entrainment separators is limited. Thus, the
washing techniques must make use of recirculated scrubber
liquid and a minimum of fresh water.
3. Determine the primary droplet collection efficiency,
reentrainment properties, and pressure drop of the
entrainment separator.
4. Investigate potential operation and maintenance
problems such as plugging, corrosion, etc.
Develop Design Criteria
Develop design equations including scale-up criteria
which can be used to design larger units.
Determine Emission Effects
Determine the effects of entrained solids (dissolved or
suspended) that escape the entrainment separator on particulate
emissions from a scrubber system.
Modify Design Models
Previous design methods for entrainment separators should
be modified to incorporate the results of this research.
15
-------
Recommend Test Program
Recommend a test program to demonstrate superior entrain-
ment separators at large pilot scale (up to 30,000 ACFM of
gas) for example at EPA's scrubber test facility at Shawnee.
Include estimates of the cost and time required for the
demonstration.
16
-------
SECTION 3
ENTRAINMENT SEPARATOR DESIGN FOR PILOT PLANT
DESIGN APPROACH
The approach to the design of the pilot plant entrainment
separator (E.S.) was the same as should be followed in designing
any E.S., namely:
1. Obtain background information on the gas stream, parti-
culate emissions, and entrainment size and concentration,
2. Determine the required emission limit.
3. Identify suitable types of E.S.
4. Estimate the entrainment separator performance.
5. Determine the compatible combinations of scrubbers
and E.S.
6. Final selection and performance prediction.
Background Information
The E.S. design process for a new installation will involve
a balance between scrubber and E.S. performance for an optimum
design. Thus, particle as well as entrainment information is
required. When the E.S. is to be used on an existing scrubber
whose performance is fixed, the E.S. efficiency required will
also be fixed. Particle information of interest includes:
1. Particle concentration
2. Particle size distribution
3. Particle density
4. Nature, shape, volatility, etc. of the particles
5. Gas temperature, pressure, and composition
17
-------
To design a suitable E.S. or to select one and predict its
collection efficiency, one needs information on the liquid and
gas phases, as listed below.
A. Liquid Phase
1. Entrainment drop size distribution
2. Entrainment loading
3. Suspended and dissolved solids
4. Densities
5. Vapor pressure
6. Nature of the entrainment; i.e., is it sticky,
corrosive, oily, etc.?
B. Gas Phase
1. Temperature
2. Pressure
3. Composition
4. Flow rate
5. Allowable pressure drop
Assumed Properties -
The particle and entrainment properties assumed for this
study were based on a typical coal-fired utiltiy boiler appli-
cation as described by Epstein (1975) for the EPA alkali scrub-
bing test facility at TVA, Shawnee Station.
Particle concentration: 8.6 g/Nm3*
Particle size distribution: d = 35 ymA** and a = 2.3
Scrubber type: Mobile bed (because of severe entrainment
problems experienced with this type)
*'N' refers to "normal" gas conditions, i.e. 20°C and 1 atm.
**tymA1 (micrometers, aerodynamic) are the units of drop or
particle aerodynamic impaction diameter (see Galeski, 1977),
defined as: ,
dpa [ymA] = dp[ym] x (C' pp)*
where d = Stokes or physical diameter, ym
C' = Cunningham slip correction factor
p = particle density, g/cm3
18
-------
Entrainment drop size distribution: As described by Cal-
vert, et al. (1977)
Entrainment loading: As described by Calvert, et al. (1977).
Suspended and dissolved solids: Total up to 25% by mass;
primarily CaCOs. No attempt was made to match the
dissolved and suspended sulfates and sulfites, or the
pH of FGD slurries. This study concentrated on perfor-
mance with suspended solids deposition without chemical
precipitation effects. Work at EPA's Shawnee station
has shown that scaling resulting from sulfate/sulfite
precipitation and other causes can be effectively con-
trolled through the system chemistry - see Williams (1978).
Nature of entrainment: Abrasive
Gas conditions in E.S.: Saturated with H20 at about 50°C,
pressure = 1 atm.
Emission Limit
In order to specify the E.S. performance required one must
first determine the overall emission limit for the scrubber/E.S.
air pollution control system. The E.S. for this study was designed
for a pilot plant which was to simulate a lime/limestone scrubber
system used for flue gas desulfurization and particulate removal.
Although it might be ideal to remove all the entrainment,
economics dictates some limit, and it is necessary that a tolerable
level of removal efficiency be determined. The first consideration
should be that of complying with legislated emission standards.
In the limestone FGD case the applicable standard would be that
for particulate emissions.
The standards for existing fossil-fueled boilers vary from
state to state and the EPA was in the process of promulgating a
new source performance standard (NSPS) for new facilities. At the
time of this study the NSPS for total particulates from fossil-
fueled boilers is 13 mg/MJ heat output of fuel (0.03 lbm/106
BTU). In terms of concentration of particulates this is about
34 mg/m3 at 20°C, 1 atm (0.015 grain/SCF).
19
-------
Entrainment Separator Types
Entrainment separators for scrubbers fall into the following
categories:
1. Cyclone
2. Gravity settler
3. Wire mesh
4. Packed bed
5. Discontinuous baffles
6. Tube bank
7. Continuous wave form
The properties and limitations of each category of E.S. have
been studied by Calvert, et al. (1975) and descriptions of commer-
cially available designs can be found in Mcllvaine (1978). In
general the wire mesh, packed bed, and waveforms with "hooks" and
other projections are not suitable for applications using slurry
as they are easily filled by solids buildup.
The gravity settler is usually not feasible because the
efficiency is low and the space required is large. Of the re-
maining three types, the cyclone and the baffle type are the
most widely used on slurry scrubbers. Baffle types are often
preferred because of their large clearance dimensions, ease of
washing, relatively small size, and moderate cost.
An open, or discontinuous, baffle type E.S. has been used at
EPA's Shawnee test facility (Epstein et al.,1976), while continuous
sharp-cornered and waveform chevrons have been used extensively
in other lime/limestone facilities and pilot plants. Weir,
et al. (1976) have successfully used the waveform chevrons at
the Mojave Generating Station of Southern California Edison Com-
pany. The sharp-cornered chevrons have been used in pilot plant
studies at EPA by Hollinden et al. (1976, 1977) and at Ontario
Hydro by Sekhar (1977). Conkle et al. (1976) in a broad survey
or lime/limestone FGD systems have concluded that "the chevron
or baffle-type mist eliminator is used nearly universally in the
United States."
20
-------
Based on its general usage and on our extensive experience
with the design (Calvert, et al., 1975) we selected the discon-
tinuous zigzag baffle type of entrainment separator for the pilot
plant study.
Performance of Selected Entrainment Separators
The approximate performance of various types of entrain-
ment separators is shown in Figure 1. The performance parameter
used in the figure is the aerodynamic cut diameter in units of
ymA, which is the drop diameter that will be collected with 50%
efficiency. The collection efficiency of drops larger than the
cut diameter is usually very high so it can be approximated by
assuming that 100% of them are collected.
For our system we arrived at a design pressure drop for the
E.S. of about 2 cm W.C. This pressure drop corresponds to a cut
diameter ranging from about 15 ymA for the least efficient type
of E.S. in Figure 1 to about 2 ymA for the wire mesh. To be con-
servative, we assumed that 100% of drops larger than 20 ymA are
collected for AP = 2 cm W.C. in a zigzag baffle E.S. The impli-
cations of this E.S. performance on the required scrubber per-
formance are discussed below.
Scrubber/E.S. Performance Balance
The system performance for a mobile bed scrubber and baffle-
type E.S. was calculated for representative conditions. The
required overall efficiency is assumed to be 99.7% and the
assumed entrainment parameters have been listed previously.
Most of the particle penetration will be allocated to the
scrubber, as discussed below.
In order to predict the E.S. efficiency required we had to
assume the entrainment drop size, concentration, and solids con-
tent. The drop size and concentration were taken from Calvert,
et al. (1977) for QL/QG = 7.8 1/m3 and yfi * 4 m/s in a small,
3-stage mobile bed scrubber. A conservatively high slurry
concentration of 25 wt% solids was assumed, although this is
almost double the usual concentration for lime/limestone
scrubbers.
21
-------
100
to
H
W
U
g
u.
K
PJ
10
0.01
1 B
I I I I I
I I I I I I I I
A. Baffles, 6-row, 30°
B. Baffles, 6-row, 45°
C. Tube bank, 6-row, 1 cm spacing
D. Tube bank, 6-row, 0.3 cm spacing
E. Packing, 2.5 cm
F. Mesh, 0.028 cm wires
G. Cyclone, 2 m diameter
i i i i i i
I j LI i ii
I I I 1 I I 1 I
0.1
1.0
10 50
AP, Pressure Drop, cm W.C.
Figure 1. Performance cut diameter for several types of entrainment separators.
-------
Figure 2 is a plot of cumulative mass concentration of
solids emission from mobile bed scrubbers as a function of
dry particle residue diameters and original drop diameters.
The solid curve labeled "1977 data" shows that the cumulative
mass concentration of solids in drops 20 ymA diameter and
smaller was 4 mg/Nm3. Consequently, an E.S. with a 20 ymA cut
diameter would emit about 4 mg/Nm3 of solids carried by the drops,
According to Epstein (1975), a mobile bed scrubber with a
21 cm W.C. pressure drop would have a penetration of 0.0035 and
would emit 30 mg/Nm3 (for the representative conditions), so
the total emission of ash and entrainment solids would be 34
mg/Nm3. The solid curves on Figure 2 labeled "Ash" and "Total,
1977 + 21 cm W.C." show the predicted emissions as a function
of dry particle diameter.
As will be seen later, the amount of entrainment measured
in this study was about twice that in the 1977 research. The
dashed curves in Figure 2 shows the cumulative solid mass
emitted as entrainment, the ash emission for a 27 cm. W.C.
mobile bed scrubber pressure drop, and the total dry particle
emission. A relatively small increase, 6 cm W.C., of scrubber
pressure drop would permit attainment of the 34 mg/Nm3 limit
with the high assumed value of 25 wt% solids in the entrainment.
At the usual lime/limestone slurry concentration of around
15%, the mass emission due to entrainment would be 4.8 mg/Nm3
rather than the 8 mg/Nm3 for 25% slurry. Consequently, even
with the benefit of hindsight, the design objective of 15 to
20 ymA drop cut diameter for the E.S. seems reasonable.
The sensitivity of overall pressure drop to E»S. pressure
drop can also be estimated from Figures 1 and 2, as given below.
A. Increasing efficiency by having an E.S. absolute cut
at 10 ymA instead of 20 ymA would result in an increase in
pressure drop of the E. . from 2 cm W.C. to about 20 cm W.C.
for the baffle-type E.S. The emissions would be reduced by
only about 0.6-1.8 mg/Nm3 or 2-61 for an 18 cm W.C. increase
in APo
23
-------
40
CO
E
o
u
CO
a
o
LO
LU
u
30
20
10
0.1
I 111
TOTAL, 1977 DATA, AP = 21 cm W.C.
TOTAL.PRESENT DATA
(34 mg/Nm3)
SCRUBBER AP = 21 cm W.C.
ASH;
SCRUBBER"AP = 27 cm w.c.
PRESENT STUDY
1977 DATA
0.5 1.0
10
50 100
DROP OR PARTICLE DIAMETER, d, or d , ymA
d pa
Figure 2. Cumulative solids concentration from a mobile bed scrubber for
two levels of entrainment concentration.
-------
B. Increasing efficiency by raising scrubber pressure
drop by only 6 cm W.C. from 21 to 27 cm W.C. results in the
scrubber being about 99.8% efficient and reduce particulate
loading by 4 mg/Nm3 , or 13%.
Thus, increasing E.S. pressure drop has very little effect
on overall efficiency, compared to increasing the scrubber pres-
sure drop. This conclusion generally applies, since the scrubber
is a much more efficient particle collection device than the
entrainment separator.
Final Selection and Performance Prediction
The discontinuous zigzag baffle design was selected for
the pilot study. This E.S. type could provide the needed
efficiency and had been applied to lime/limestone scrubbers. It
was also considered that this design would be less susceptible
to solids buildup than others and could easily be washed.
The configuration of the baffles is shown in Figure 3.
The approximate performance cut diameter of this configuration
was shown as curve "A" in Figure 1. In the following a more
detailed performance prediction will be presented.
Primary Collection Efficiency -
The penetration of drops through zigzag baffles, based
on perfect turbulent mixing, is:
- exp-
where A~ = deposition surface area, cm2
UD = deposition velocity, cm/s
QG = gas flow rate, cm3/s
The deposition surface area is the upstream surface area of the
baffles :
AD = w h nb n (2)
25
-------
where w = baffle width, cm
h = baffle height, cm
n, = number of baffles in a row
b
n = number of rows
The drop deposition velocity is the terminal centrifugal velocity,
where the centrifugal force acting on the drop is due to the
average radius of curvature through which the gas turns in
passing through the baffles (see Calvert et al., 1974).
- • (s %¥)'"
where p, = drop density, g/cm3
PG = gas density, g/cm3
d, = drop diameter, cm
CD = drop drag coefficient
a = centrifugal acceleration, cm/s2
2 u2 sin 0
a = ^—.—-—
w cos^ 0
where u~ = superficial gas velocity, cm/s
0 = angle through which gas must turn, degrees or radians
When the drop Reynolds number is less than 0.1 Stokes flow may
be assumed, so that:
CD = ^e NRe < °-1 ^
p u d,
where NR = — (6)
i\c \JL f~t
b
and p., = gas viscosity, g/cm-s
A more general drag coefficient for spheres, due to Dickinson
and Marshall (1968) is,
r = n ?? +
LD - u.zz + jj-
Re
°-6 \
'15 NRe)
26
-------
Combining equations (1), (2), (3) and (4)
Pt = exp
where
UD1 [s sin (0/2) pd dd
w
(9/2)
]
J
i/ 2
(8)
(9a)
1/2
Up r8 sin 0 Pd dd 1
UG [3 w cos 3 0 CDPGJ
(9b)
It is assumed that 0 is the angle the baffle makes with the
flow and that except for the first row the gas turns through an
angle of 2 0. And,
0~
(10)
nb = 6
G —
n
For the dimensions shown in Figure 3:
w = 7.6 cm
b = 8.5 cm
0 = 30 degrees
7.6 cm
t
x
>
7.6 cm
51 cm
i
2.5 cm
T
Figure 3. Top view diagram of example zigzag baffles.
27
-------
yr = 1.8 x 10" g/cm-s
We assume typical values for the air and drop properties:
PG = 1.2 x 10"3 g/cm3
Pd = 1.05 g/cm3
'G
The penetrations calculated from these equations for varying
drop diameter and superficial gas velocity are shown graphically
in Figures 4,5, § 6. The plots show fractional efficiency, which is
one minus penetration. The first two show the effect of
varying gas velocity and the third shows the efficiency as a
function of drop diameter for each row of baffles at a set gas
velocity. Coordinates for these plots are logarithm of drop dia-
meter and the Weibull cumulative distribution function. Weibull
coordinates have been used because efficiency plots as a straight
line function of drop diameter in the low Reynolds number (Stokes
flow) regime.
Overall penetration can be predicted by integrating equation
(8) over the size distribution of'drops. Size distributions,
which will be discussed in Section 4, may take a number of forms.
Typical size distributions are the log-normal and the cubic.
Figure 7 presents the overall penetration based on numerical in-
tegration of equation (8) for two log-normal distributions
(a = 1.5 and 2.0) and the cubic distribution.
o
Reentrainment Prediction -
Drop collection in an E.S. is governed by the combination
of the primary collection of drops and the reentrainment of col-
lected liquid. The liquid and gas flow capacities of an entrain-
ment separator are limited by the occurrence of reentrainment
at high velocities. Experimental data on collection efficiency
for various separators have been summarized by Calvert (1978), as
follows.
Figure 8 shows collection efficiency vs. horizontal air velocit
for vertical zigzag baffles. For horizontal gas flows, Figure
9 shows the superficial (empty cross-section) gas velocity in
the separator as a function of the inlet liquid/gas ratio, with
reentrainment rate as a prameter. The indicated zones are
approximate, and above the shaded area some slight reentrainment
was noted.
28
-------
0.9
2 5 10 20
DROP DIAMETER, ym (physical)
Figure 4. Predicted grade efficiency of 3 rows of
zigzag baffles.
0.9
0.5
0.1
0.01
0.001
I III I 1 I
5 10 20
DROP DIAMETER, ym (physical)
50
Figure 5. Predicted grade efficiency of 6 rows of baffles.
-------
UJ
u
K-l
u.
u.
LU
0.01 -
0.001
10 100
DROP DIAMETER, ym (physical)
1,000
Figure 6. Predicted grade efficiency of zigzag baffles at 4.6 m/s superficial
gas velocity.
-------
1.0
CO
o
o
2
<4-t
tu
z
o.
_]
a
0.1
0.01
0.001
u = 4.6 m/s
u
LOG-NORMAL
a = 1.5
g
CUBIC
DISTRIBUTION
1.0
10 20
50 100
ddg °r dnun'
200
500
o
tH
M-l
o
I—I
i
UJ
UJ
a.
Si
uu
o
la.
0.1
0.01
0.001
- LOG-NORMAL
a = 1.5
g
CUBIC
DISTRIBUTION
u = 4.6 m/s
LOG-NORMAL
2 .
10 20 50 100 200
ddg °r dnun' ym
500
Figure 7. Predicted overall penetration for 6 rows of baffles.
-------
From Figure 8 it appears that about 6 m/s is the limiting
superficial gas velocity. This limit is somewhat dependent
on entrainment loading and may be raised to about 8 m/s at
low loadings, based on data shown in Figure 9. A conservative
velocity, allowing for a relatively high entrainment loading,
would be 4.6 m/s (15 ft/s), which was selected as the design
velocity.
Pre-Separator -
It follows from the relationship between the reentrainment
gas velocity and the liquid/gas ratio that higher gas capacity
could be obtained if the liquid/gas ratio were reduced. This
is the reason for using various means for removing some entrain-
ment upstream from the final E.S. Relatively simple provisions
for "pre-separation" can effect significant improvement in the
E.S. capacity.
One should note, however, that the use of a pre-separator
will not reduce the collection efficiency capability of the final
E.S. For example, if a 15 ymA cut diameter is required in order
to meet the emission limit, the use of a pre-separator with per-
haps, a 50 to 100 ymA cut diameter will not alleviate that re-
quirement.
Pressure Drop Prediction -
Calvert, et al. (1975) present an equation for predicting
pressure drop based on the drag of inclined plates. For a
velocity of 4.6 m/s the predicted pressure drop was about 1.1
cm W.C. for 6 rows of baffles. Their experimental measurement
was slightly higher, about 1.5 cm W.C.
PILOT PLANT
The pilot plant was designed to be representative of an
industrial scrubber/entrainment separator system in that it was
large enough and of such configuration that its performance
could be "scaled-up" reliably. The design gas flow rate was a
minimum of 1.42 m3/s (3,000 CFM) and could be 50% higher, depen-
ding on the system pressure drop. Windows and special access and
32
-------
£-
Ui
0.5
0.4
0.3
0.2
0.1
0
Bell & Strauss (1973)
2 chevrons
Houghton 6 Radford (1939)
6 rows, 9 = 30°
Calvert, et al. (1974)
6 rows, 9 = 30°
d, = 380 urn, o =1.5
dg g
0123456 8
u , SUPERFICIAL HORIZONTAL VELOCITY, m/s
Figure 8. Measured overall collection efficiency
of vertical zigzag baffles.
1000
S
i—*
e
en
O
t-H
E-
_
Z
O
100
50
40
30
20
10
•o
Reentrainment <1%
Reentrainment in part of duct
Primary efficiency <100%
1°0% efficiency
I
I
I
I
I
_L
0 246
UG> SUPERFICIAL HORIZONTAL GAS VELOCITY, m/s
Figure 9. Reentrainment limits for 6 rows of
vertical zigzag baffles. (Data
by Calvert, et al., 1974}.
-------
sampling conveniences were included in the system. The system
is shown schematically in Figure 10 and pictorially in Figure 11
Location of the entrainment separator in a horizontal duct
was an important design feature. Previous work reported by
Calvert, et al. (1977) had demonstrated that a configuration
with the gas flowing horizontal and the drainage surfaces in
vertical planes was least subject to reentrainment. The E.S.
designed for horizontal gas flow was therefore able to handle
the highest entrainment loading and gas velocity for its size.
Pre-Separation -
The scrubber/entrainment separator system was designed with
enough vertical height between the scrubber and the entrainment
separator so that substantial disengagement of large drops could
occur. The loading on the entrainment separator is consequently
greatly decreased without using any energy.
Another pre-separator stage was easily incorporated in the
form of the turning vanes in the elbow transition between the
vertical scrubber and the horizontal entrainment separator duct.
The turning vanes collected a considerable amount of the large
size entrainment which would otherwise settle on the horizontal
duct or overload the E.S.
Finally, the horizontal run between the turning vanes and
the E.S. enabled additional entrainment to settle out of the
gas stream before the E.S. As discussed previously, the purpose
of designing the system was to minimize the loading on the
E.S. and thereby increase the allowable gas velocity.
Scrubber/Entrainment Separator Shell
The scrubber and entrainment separator shells were made of
molded fiberglass, 4.8 mm (0.19 in) thick. The vertical
scrubber section stood in an open polypropylene tank, from
which the scrubbing liquor was recirculated. The entrainment
separator section was in a horizontal run, accessible from a
platform which was 2.44 m (8 ft.) above the floor. The air was
moved by a forced draft centrifugal blower. The system had
34
-------
OUTLET
ENTRAINMENT SEPARATOR
SECTION
Cn
SCRUBBER
MODULE
1
i
1 L
1 L
1
I
IkrJ
*
I
'
— - — ^
\> \< 1 '
LIQUID DRAIN LINE
SCRUBBER LIQUID
LINE
FORCED-
SCRUBBER SUMP TANK
Figure 10. Schematic diagram of scrubber/entrainment separator system.
-------
Figure 11. Scrubber/E.S . system,
-------
numerous windows and sampling ports located at strategic points.
Basic dimensions are shown in Figure 12.
The scrubber section was square, with a cross-sectional
area of 0.581 m2 (3.33 ft2). It was 1 m high and flanged at
both ends to allow easy removal. The scrubber section was 1.4 m
below the horizontal section to allow room for disengagement
of large drops generated in the scrubber.-
The E.S. section was rectangular, 61 cm high by 51 cm wide
(24 in x 20 in). There is a 0.13 m deep recess below the entrain-
ment separator section to collect liquid drainage and to run internal
piping. The floor was thin gauge stainless steel sheet with
cross-wise slots to accommodate drainage from the E.S. baffles.
Entrainment Separator Design
The entrainment separator was a discontinuous zigzag baffle
type, containing up to six rows in two configurations. One
configuration was a vertical orientation and the other was a
forward sloping orientation of 30° from the vertical. The
baffles were made of type 304 stainless steel and constructed
in row modules so that any number of rows from one to six could
be used. A plan view of the baffles was shown in Figure 3.
The lower front edge of the first row was located 2.35 m hori-
zontally downstream from the transition section. Vertical
height was 61 cm for all baffles.
Scrubber Operating Conditions
Three types of scrubbers were used to generate entrainment:
mobile bed, sieve plate, and spray. A single stage mobile bed
scrubber was operated in five modes as described in Table 2.
The bed packing consisted of 3.8 cm (1.5 in) diameter hollow
polypropylene spheres. A single sieve plate was operated at
two conditions, as shown in Table 3 . Two types of spray
nozzles were used in the spray scrubbers, as described in
Table 4 to achieve different drop size distributions.
37
-------
INLET
PORTHOLE
61
137
oo
132
76
1
UtJL
I I
ENTRAINMENT SEPARATOR
235
184
SCRUBBER
TANK
BLOWER
OUTLET
PORTHOLE
TO
+
OUTLET
Figure 12. Scrubber/entrainment separator system
(dimensions in cm).
-------
TABLE 2. MOBILE BED SCRUBBER OPERATING CONDITIONS
Mode
1
2
3
4
5
(
m3/s
2.12
2.12
2.12
1.89
1.65
CFM
4,500
4,500
4,500
4,000
3,500
UG
m/s
3.66
3.66
3.66
3.25
2.84
£/m3
6.54
5.94
4.46
7.35
8.40
AP
cm W.C.
3.9
3.6
3.1
3.7
3.4
Bed static height - 20-23 cm; bed packing- polypropylene
spheres 3.8 cm dia., each weighing ~4.7 g; bed support •
0.088 cm dia. galvanized wire screen openings -1.3 cm x
2.5 cm; located 1.37 m below horizontal section.
TABLE 3. SINGLE SIEVE PLATE SCRUBBER OPERATING CONDITIONS
Stainless Steel - 0.16 cm thick (0.0625 inches)
Hole Diameter - 0.476 cm (3/16 inch)
Column Cross Section - 0.581 m2 (6.25 ft2)
Downcomer Cross Section - 0.021 m2 (0.22 ft2)
Plate Area Exposed to Flow - 0.386 m2 (4.15 ft2)
Open Area - 0.033 m2 (0.36 ft2)
Weir Height - 5.1 cm (2 inches)
Located 1.47 m below horizontal section
1. Low AP Operation
Gas flow rate = 0.71 m3/s (1,500 CFM)
Water flow rate = 8-9.5 i/s (125-150 GPM)
Pressure drop ~ 11 cm W.C.
2. High AP Operation
Gas flow rate = 1.04 m3/s (2,200 CFM)
Water flow rate = 8-9.5 £/s (125-150 GPM)
Pressure drop ~ 23 cm W.C.
39
-------
TABLE 4. SPRAY SCRUBBER OPERATING CONDITIONS
Cocurrent operation, nozzles located 2.13 m below horizontal
section
Scrubber cross-sectional area = 0.581 m2 (6.25 ft)
Gas flow rate = 1.42 m3/s (3,000 CFM)
Scrubber superficial gas velocity = 2.44 m/s
Nozzle type: Spraying Systems Whirljet
Model no.:
Pressure:
Flow rate:
Drop sauter
mean dia:
1/2-B-50
207 kPa (30 psig)
5 nozzles - 2.68 A/s
7 nozzles - 3.75 H/s
1,000 ym
Bete Spiral
3/4-ST-24-FCN
55 kPa (8 psig)
4 nozzles - 2.52 i/s
6 nozzles - 3.79 £/s
500 ym
40
-------
Drop Disengagement in Vertical Duct
In a vertical gas flow a drop with a terminal settling
velocity greater than the gas velocity will not be emitted
from the scrubber, provided enough disengagement space above
the scrubber is provided. The terminal settling velocity of a
drop or particle is the velocity at which the gravity and fluid
drag forces balance. The disengagement space is needed to allow
for the higher velocities given to some drops at the top surface
of the sieve plate, mobile bed, or spray.
For an air-water drop system at 20°C and 1 atm, the
terminal settling velocity of various diameter drops is given
in Figure 13. One should note that in systems which are not
saturated with water vapor (100% relative humidity) the drops
will decrease in size due to evaporation.
Superficial gas velocities for the three scrubber types and
the drop diameters corresponding to these terminal settling
velocities are given below.
Scrubber Gas Velocity, m/s Settling Drop Diameter,ym
Mobile Bed 2.8 to 3.7 700 to 950
Sieve Plate 1.2 to 1.8 300 to 400
Spray 2.44 600
Particle Collection Efficiency of Turning Vanes
The turning vanes cause drop separation due to centrifugal
deposition. The four vanes are quarter-sections of 15 cm radius
duct arranged diagonally across the turn. The centrifugal de-
position velocity is computed by balancing the centrifugal force
acting on the drop and the drag.
(113
The deposition area of the four vanes is
AD = 2 x TTX 0.15 x 0.76 = 0.73 m
2
41
-------
1.0
u
i— i
.j
UJ
0,1
0.01
10
T = 20°C
i i i I i i i I
I i i i i 1 11
i i i
100 1,000
dd, DROP DIAMETER, ym
10,000
Figure 13. Gravity settling velocity of water drops in air,
(Fuchs, 1964).
42
-------
Penetration through the turning vanes can be analyzed like
that through a horizontal duct by substituting centrifugal
deposition velocity for gravity settling velocity. Figure 14
is a plot of the centrifugal deposition velocity vs. drop diam-
eter for Ug = 4.6 m/s and r = 0.15 m. The penetration equa-
tions for the three drop Reynolds number regimes (described later)
are:
un
Pt = l - — d >100 ym (12)
d
Pt = exp |- Y7M I dd <10° ym
UD = 4.47 x 10"" dj dd < 30 ym
Equation (12) predicts that no drops larger than 88 ym
will penetrate the turning vanes, which makes it unnecessary
for this analysis. The efficiency of the turning vanes is
plotted in Figure 15 for u~ = 4.6 m/s. This figures shows
that very few drops larger than 200 ym should penetrate the
turning vanes.
Gravity Settling in the Horizontal Duct
The deposition due to gravity settling of water drops in
a horizontal rectangular duct was analyzed. The assumed condi
tions and dimensions were:
vn * 0.15 cm2/s (kinematic viscosity)
(j
h = 0.61 m (duct height)
w = 0.508 m (duct width)
L = 2.35 m (duct length)
The Reynolds number of the duct is then
for Up = 4.6 m/s
b
NRe = 3.69 x 10" uG(m/s)
ND = 170,000
Ke
43
-------
0.01
100
dd, DROP DIAMETER,
1,000
Figure 14. Theoretical centrifugal deposition
velocity of water drops in air.
0.999
0.01 ;
0.001
10
20 30 40 SO 100
d, DROP IHAMFTRR, um
son
Figure 15. Theoretical grade efficiency of
turning vanes.
-------
The turbulent eddy velocity (also known as the friction
velocity) is approximately,
ute = u
where UG = mean gas velocity , 4.6 m/s
f = friction factor
At Re = 170,000, f = .016 for smooth pipes. Thus,
ute z 0.21 m/s
It is assumed that drops with terminal settling velocities,
u., greater than u will not be affected by the turbulent eddies
Thus, for the present case drops larger than 100 ym diameter
(see Figure 13) are not affected by the turbulence. Drops this
size deposit as if in laminar flow.
Laminar Analysis for Drops Larger than 100 ym-
The penetration is
u A u L
Pt = 1 - -£-^ = 1 - ,-1— (16)
UG
which in our case reduces to
ut L
- 1- - (17)
for 0% penetration, the minimum or critical length needed is
Lcr - %?i (18)
For 100 ym diameter drops, L = 11 m.
For the laminar deposition region, at L = 2.35 m
Pt = 1 ~ T77 (19)
where ut is found from Figure 13. A plot of predicted efficiency
(1-Pt) for laminar settling of large drops is given in Figure 16.
45
-------
Turbulent Analysis for Drops Smaller than 100 ym -
In turbulent flow the drops are uniformly mixed over the
cross section of the duct, except near the walls. Since the
vertical eddy velocity approaches zero at the bottom wall of
the duct drops which enter this region will be deposited. The
concentration of drops entering the wall region is proportional
to the concentration in the free stream so the deposition is
exponential:
Pt
/-ut M /-ut w
= exp \ ~$~) = exp \ w~TrTr
(20)
For drops smaller than 50 ym the terminal settling velocity
can be found assuming Stokes law (viscous) resistance.
(Pd-PG) g^d
dd<50 ym : ut- 18 (22)
b
where g = 980 cm/s2
yG = 1.84 x 10"1* g/cm-s
so u = 3.10 x 10~5 d|, m/s (d, in ym)
For d,<50 ym, u must be found from Figure 11. The
penetration at L = 2.35 m, h = 0.61 m, and Up = 4.6 m/s is
Pt = exp(l77 ' (23)
which is plotted on Figure 16, as efficiency, which is 1 - Pt.
Overall Penetration Through Duct
The overall penetration through the turning vanes and
horizontal duct is the penetration integrated over the size
distribution function:
46
-------
Ft = / Pt f(x)cbc (24)
where for a log-normal inlet distribution,
* • fa,,"" <25>
£(x) = exp (-x2/2) (26)
V2u
Pt = Pt (turn vanes) x Pt (horizontal duct)
Figure 17 is a plot of overall penetration for various values
of geometric mass mean drop diameter for a standard deviation,
a = 2.0, using the grade penetrations from Figures 15 and 16.
At higher gas velocities the turning vanes will be more effi-
cient and the horizontal duct will be less efficient, so that
Figure 17 is approximately correct over a moderate range of velo-
cities about 4.6 m/s. It also shows that if the scrubber produces
a log-normal distribution of drops (a = 2) less than 1% by mass
o
will reach the E.S. if their geometric mass mean diameter is
over 550 ym.
47
-------
oo
0.999
0.99
0.9
0.5
0.1
0.01
0.001
I r T I \II I I
10
STOKES
LAW
LAMINAR
SETTLING-
TURBULENT
SETTLING
u = 4.6 m/s
L = 2.35 ra
J I
20 30 4050
100
d DKOI' niAMETER, \im
500
Figure 16. Theoretical grade efficiency of
horizontal duct.
0.5
0.4
0.3
0.2
0.1
0.01
u = 4.6 ra/s
L = 2.35 m
r = 15 cm
c
J
I I I I I I I i
I
10 :0 30 40 50 100 i.ooo
d, , DROP GEOMETRIC MASS MEAN DIAMETER, um
Figure 17. Theoretical overall penetration through turning
vanes and horizontal duct of log-normally
distributed drops.
-------
SECTION 4
ENTRAINMENT CHARACTERIZATION
The measurement of the drop size distribution and concen-
tration of the entrainment entering the entrainment separator
(E.S.) was an important part of this study. Drop size and con-
centration had to be measured at the E.S. inlet and outlet in
order to determine the total and grade efficiency of the E.S.
While the efficiency of the pre-separation stages could be
estimated, the entrainment leaving the scrubber modules was not
defined. Some data were available for the mobile bed (Calvert,
et al., 1974) but substantially none for the other modules.
Most of the drops ejected from the scrubbers are very large and
subject to sedimentation, which is greatly dependent on gas
flow patterns. Consequently, the nature of the entrainment is
highly dependent on the configuration and operation of each
individual scrubber and would be difficult to generalize.
The approach taken in this study was that the important
thing to know about entrainment is its cumulative mass distri-
bution in the size range which penetrates the pre-separator.
Application of good engineering design will result in a system
which keeps a massive concentration of large drops from reaching
the E.S. The difficult problem is to design for a small cut
diameter with reasonable pressure drop and freedom from plugging.
In this chapter the available methods and the methods
selected for measuring entrained drop sizes are described. Then
the data are presented and analyzed. The drop size distributions
measured by several methods are discussed.
49
-------
DROP SIZING TECHNIQUES
Range of Drop Sizes Expected
Most sizing techniques depend on the size of the drops to
be measured. The most important range of drop sizes is the
range where the E.S. efficiency is changing most rapidly; i.e.,
about the diameter collected with 50% efficiency. Based on
Figures 4 and 5, presented in Section 3, the drop diameters
expected to be collected with 50% efficiency range between
20 ym and 50 ym. Thus, the sizing techniques must be able to
measure drops with diameters between about 5 ym and 100 ym in
order to allow determination of efficiency as a function of
drop diameter.
Review of Drop Sizing Methods
Methods for drop size analysis have been developed to meet
the requirements of many applications in science, industry, and
environmental monitoring. Spray and mist drop size characteri-
zation is required for various applications, such as spray
nozzles, cooling tower drift eliminators, gas-atomized scrub-
bers, fuel injection, fog formation and dissipation, and scrub-
ber and chemical process mist eliminators or entrainment
separators(E.S.).
Several authors have recently reviewed the state-of-the-
art of drop size analysis. Davies (1976), Carter (1970), and
Roffman and Van Vleck (1974) present excellent discussions.
Chan and Golay (1977) also survey measurement techniques in
their report. Davies classified the methods into six categories
1. Photographic Imaging
2. Collection and Deposition
3. Momentum Transfer
4. Hot-wire Anemometry
5. Electrical Mobility
6. Optical
50
-------
Total or overall concentration of drops in the gas stream is
usually determined by isokinetic sampling with a chemical or
radioactive tracer in the liquid; e.g., the method used by
Johnson and Statnick (1974).
Sizing Methods Chosen
Hot-wire anemometry and a collection and deposition method
were initially selected for use, based on equipment availability
and previous experience. A KLD Associates model DC-1 Droplet
Counter was available and appeared suitable,, As described by
Medecki, et al. (1975) it consisted of a 5 ym diameter platinum
hot-wire probe sensor and associated electronics to analyze the
electrical pulses generated when drops attach to the hot wire.
The unit has the following specifications:
Drop size range: 1 ym to 600 ym
Flow velocity: 3 m/s, maximum
Concentration: 500 drops/cm3, maximum
Temperature: 0°C to 100°C
As discussed later, the KLD device was found not to be practical
for this application and yielded no useful data.
The first collection and deposition method used in this
research involved a specially designed cascade impactor. In
the course of the program we developed a drop sizing method
which utilized single body collectors (spheres, cylinders).
These methods are discussed below.
Cascade Impactor
The A.P.T. designed cascade impactor (C.I.) consisted of a
series of stages of a horizontal flow through single round jet
impinging on a flat plate. The nozzle and plate were made of
stainless steel and the housing glass, similar in shape to
a test tube. Figure 18 is a photograph of the A.P.T. drop
sizing C.I. The ratio of jet diameter to jet-to-plate distance
was such that the impaction parameter corresponding to 501
collection efficiency, K , was approximately 0.2. For the
stages with particle cut diameters about 2 ym and smaller,
this value was verified by calibration with monodisperse
51
-------
Figure 18. Photograph of glass cascade impactor.
52
-------
polystyrene latex spheres.
The C.I. was located in the gas stream and isokinetic sam-
pling was used. A glass fiber filter was used after the final
impaction stage. While measurable amounts of water were often
collected in the first stage, the use of a reasonable sampling
time did not cause accumulation of enough water in the other
stages.
As in previous studies, we used a salt tracer in the liquid,
Sodium chloride was dissolved in the scrubbing water, and the
NaCl collected on each C.I. stage and on the final filter was
detected using a chloride specific ion electrode method. The
measured amount of NaCl was then related to the volume of water
drops on each stage since the NaCl concentration in the scrubber
was also measured. In addition to its tracer function, the salt
reduced the water vapor pressure of the drops to inhibit evapo-
ration and size change before deposition in the C.I.
Single Body Collectors
An impactor for sizing drops larger than 20 ym diameter
would be very difficult to design and use. For one thing, it
would be so large that it would interfere with the performance
of the entrainment separator. As an alternative a system using
collection by inertial impaction on single bodies was developed.
Spheres and cylinders of different diameters were placed in the
entrainment stream. Drops which impacted on the bodies would
flow down the wet surface and into a collection flask by way
of a thin supporting rod.
Collection efficiency correlations from the literature
were used to relate the integrated efficiency of collection
to the drop size distribution. The total drop concentration
in the stream was measured by the cascade impactor.
While we used both spheres and cylinders, most of the
data were taken with spheres. The spheres seemed to provide
slightly more reproducible results than the cyliners.
53
-------
ENTRAINMENT DATA
The measured entrainment loadings from the three types
of scrubbers used are compiled in Tables A-l to A-3 in Appen-
dix "A". The scrubber operating conditions were given in
Tables 2-4 and the superficial gas velocities in the entrain-
ment separator (E.S.) are presented in Table 5.
No measurements taken with the KLD Drop Counter are shown
because operational difficulties rendered the data useless.
The 5 ym probe wires were apparently too fragile for this appli-
cation. Also, the velocity range was exceeded in all but two
scrubber configurations. The data will be discussed in detail
in the following paragraphs.
E.S. Drainage Data
The first column of data, after the run number in Tables
A1-A3, presents the measured water drop concentrations (ml/DNm3)
based on collection by the E.S. These measurements are based
on the water flow rate in the two downstream E.S. drain lines
during steady-state operation. (See Figure 10). The upstream
drain line, separated from the other drains by a dam, carried
away water which settled in the horizontal duct and was forced
to flow (creep) along the floor by the drag of the air flow.
All the drain lines were sealed so that no gas could flow
through them.
The concentrations based on E.S. drawings given in Tables
A-l through A-3 must be considered approximate because they
were based on the assumption that the E.S. was 1001 efficient.
Creep of liquid into the E.S. along the duct walls could not
be prevented. The data shown in parentheses qualitatively show
the effect of the creeping liquid since no E.S. baffles were in
the duct during these runs. Some settling of drops in this 53
cm long space during these empty-duct runs mayaccur. Thus
concentrations based on the E.S. drainage are higher than they
should be.
54
-------
TABLE 5 . SUPERFICIAL VELOCITY IN
ENTRAINMENT SEPARATOR
Scrubber Configuration
Mobile Bed
Mode 1
Mode 2
Mode 3
Mode 4
Mode 5
Sieve Plate
Low AP
High AP
Spray
E.S. Superficial Velocity
u~, m/s
6.9
6.9
6.9
6.1
5.3
2.3
3.4
4.6
55
-------
Sphere Drainage Data
Data are presented in Tables Al - A3 for drop collection by
spheres of two diameters, 7.6 cm and 3.8 cm. The spheres were
solid hardwood whose porous surface caused the liquid to spread
in sheets rather than rivulets. The spheres were held in the
gas stream on a thin wood rod which was held centered inside a
length of stainless steel tubing which was connected to a flask.
The AoP.To body collector drop sizing system is shown schema-
tically in Figure 19.
The drops impacting on the spheres formed a sheet which
flowed by gravity down the wood supporting rod (which was
exposed to the gas stream for less than 0.5 cm) into the tube
and then into a flask. The collection system was sealed to pre-
vent gas flow into the drain tube. The rate of liquid flow into
the collection flasks was measured after the spheres had become
thoroughly wetted.
Drop concentrations were calculated by dividing the liquid
flow rate from the sphere by the gas velocity and the sphere
cross-sectional area. These concentrations were generally
higher than those based on C.I. sampling through a sampling
nozzle. The reason is not known, as discussed below.
The relatively large horizontal surface of the spheres
offers a possibility for collecting large drops which settle
from the gas or drip from the ceiling of the duct. However,
computation of the probable drop sedimentation rate based on
the drop size measurements indicated that only a small
increase in concentration could be accounted for. Sedimenta-
tion might increase the measured concentration by 0.1 ml/DNm3
out of a total on the order of 2,0 ml/DNm3.
It is possible that the Col. might "under-sample" large
drops by some mechanism such as interception by the nozzle
inlet, followed by shattering and/or deflection from the inlet.
Concentration from C.I. Measurements
The concentrations shown for the C.I. method are based on
56
-------
AIR
FLOW
WOOD SPHERE
SUPPORT ROD
TUBE
BOTTOM OF DUCT
FLASK
Figure 19. Body collector drop sizing system.
57
-------
on analysis of the amount of tracer NaCl collected in the
nozzle, cascade impactor (C.I.) and final filter. The sampling
nozzle was located 6.5 cm above the center of the duct, about
45 cm upstream of the E.S. (see Figure 12). All the stages
of the C.I. were inside the duct.
Because the sampling is done upstream of the E.S. the size
distribution and total loading that the E.S. encounters may be
slightly different than that measured by the C.I. The location
of the sampling nozzle was chosen so that it would sample drops
which would enter the center of the E.S. Measured concentrations
of drops 20 ym diameter and smaller will not be affected signifi-
cantly by the sampling location.
AUXILIARY SIZING TECHNIQUES
In addition to making direct measurements one can estimate
the maximum drop sizes to reach the entrainment separator. In
the vertical section of the scrubber one can estimate that all
drops with a settling velocity greater than the upward gas
velocity will not reach the E.S. Settling in the rectangular
horizontal duct upstream of the E.S. is also predictable.
Centrifugal deposition on the turning vanes in the transi-
tion elbow between the vertical tower and the horizontal duct
can be predicted with good accuracy because the geometry is
simple. When the spray scrubber is operating, the drop size
distribution in the scrubber can be estimated by extrapolation
of published data,, The maximum drop size reaching the E.S.
inlet would be approximately the diameter corresponding to the
overall mass fraction of the spray at the E.S. inlet.
ENTRAINMENT DROP SIZE DISTRIBUTIONS
As discussed previously, it is important to know the drop
size distribution in the crucial range where collection effi-
ciency changes the most with drop size. Several techniques
were used to measure drop size over different size intervals
58
-------
within this range. It is necessary to reconcile all of these
discontinuous measurements to give a continuous distribution
of drop sizes.
The sizes larger than 20 um were computed from data on
drop collection on spheres and collection in the entrainment
separator. Comparisons were also made with the nozzle manu-
facturer's drop size data, theoretical fallout due to gravity
in sections between the spray scrubber and the entrainment
separator, and theoretical collection by the turning vanes.
Types of Size Distributions
In order to describe the size distribution of drops with a
few parameters it is necessary to relate the data to mathe-
matical distribution functions. For liquid atomization pro-
cesses the log-normal and cubic functions seemed suited to
describe the size distribution.
Log-Normal Distribution
The cumulative mass distribution for a log-normal distribu-
tion of drop diameters is
In d
where d
and
where
drop diameter , ym
log-normal cumulative distribution, fraction
f™(ln d'} * ^H
o
o
-(
lnd
2 In2 a
g
= log-normal density function
, • geometric mass mean diameter, urn
geometric standard deviation
(28)
59
-------
These last two parameters fully describe the log-normal distri-
bution.
Cubic Distribution
Plots of the drop size distributions published by spray
nozzle manufacturers often fit a cubic distribution:
(29)
where F , = cubic cumulative distribution, fraction
d = maximum drop diameter, ym
IHclX
The cubic density function is,
f . = — L- d/ (30)
1
.
cub 13
max
The mass median diameter or diameter at which F , = 0.5 is,
d = 0.794 d (31)
mm max
Thus, only one parameter, either "dmm" or dmax" fully describes
the cubic size distribution. The "spread" of the distribution
has been fixed by the cubic exponent.
Cascade Impactor Data
Data for the concentration of drops on each stage have
already been presented. The use of these measurements and the
stage cut diameters yields a limited size distribution. Since
the maximum stage cut diameter was 20 ymA, only the lower end of
the size distribution can be determined. The averaged data are
summarized in Table 6. The total size distribution requires
correlation with data from the sphere collectors and other means
60
-------
TABLE 6. AVERAGE SIZE DISTRIBUTIONS BASED ON CASCADE
IMPACTOR DATA AT E.S. INLET
Scrubber Type
Total Loading
ml/DNm3
Cumulative Mass Fraction
of drops smaller than indicated
diameter (ymA)
Mobile Bed
Mode 1
Mode 2
Mode 3
Mode 4
Mode 5
Sieve Plate
Low AP
High AP
Spray
5-B50
7-B50
4-ST24
6-ST24
2.44
1.54
1.26
1.19
0.85
0.086
0.14
9.15
7.32
3.22
3.76
8
0.008
0.026
0.016
0.013
0.080
20
0.36
0.17
11
0.005
0.007
0.011
0.009
5
0.007
0.023
0.014
0.014
8
0.29
0.14
8^
0.003
0.006
0.009
0.007
1.5
0.005
0.008
0.008
0.010
0.041
1.5
0.21
0.10
1.6
0.002
0.002
0.006
0.005
61
-------
Size from E.S. Penetration
The efficiency characteristics of the entrainment separator
itself can be used to compute the inlet drop size distribution.
The overall penetrations through 3 and 6 rows were determined
by the cascade impactor/sampling train upstream and downstream
of the E.S. Table 7 presents the data.
For zigzag baffles, penetration as a function of drop diameter
for drops less than 100 ym can be expressed as:
Pt = exp
- In 2
(32)
where d, is the drop diameter collected with 50% efficiency.
The overall penetration is the penetration integrated over
the drop size distribution. The two types of size distributions
were used to match with the penetration data. The log-normal dis-
tribution requires numerical integration, which Calvert, et al.
(1972) performed and presented in their Figure 5.3.1-4. The cubic
distribution has a mathematical solution:
PT =
1.5
mm
erf / 1.05
1.3
mm
s o,
mm d 5 o
f^YTl
c Ap
1.1
/d \2 "
-l.l[_mm_)
*dds/-
(33)
where "erf" is the error function. This equation is plotted on
Figure 20.
Results
The derived size distribution based on penetration through
the E.S. are presented in Table 8. These derived size distribu-
tions and the data from the cascade impactor sampling are plotted
on Figures 21 through 31. These figures also show the results of
the sphere collector data which will be discussed next. In general
these figures show a more uniform drop diameter for the larger size
62
-------
TABLE 7. MEASURED OVERALL PENETRATION THROUGH E.S.
Scrubber
Configuration
Mobile Bed, Mode 1
2
3
4
5
Sieve Plate, Low AP
High AP
Spray - 5 B-50
7 B-50
4 ST-24
6 ST-24
Pt, Overall Penetration (fraction)
3 Rows
6 Rows
0.0082
0.015
0.31
0.14
0.029
0.014
0.014
0.030
0.0082
0.0084
0.013
0.025
0.055
0.26
0.20
0.0049
0.089
0.013
0.011
63
-------
o
z
o
w
z
tu
Cu
UJ
0.001
0.01 -
d /d
nun
Figure 20. Theoretical overall penetration through zigzag
baffles of drops with cubic distribution.
64
-------
TABLE 8. SIZE DISTRIBUTIONS DERIVED FROM
E.S. PENETRATION DATA
Scrubber
Configuration
Mobile Bed 1
2
3
4
5
Sieve Plant - High AP
Low AP
Spray 5
7
4
6
ddso» pm*
3 Row
31 .
31
31
33
35
44
54
38
38
38
38
6 Rows
21
21
21
21
24
30
37
26
26
26
26
Log-Normal
d _ , UIH
dg
110
90
80
95
65
50
85
120
125
120
175
Dist.
a **
g
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5***
1.5
1,5
2.0***
Cubic
Dist.
nun
135
110
95
110
65
55
85
145
150
140
125
* See Figures 4 and 5
** Assumed
*** Actual best fit values
65
-------
1,000
100
10
LOG-NORMAL
CUBIC
E.S. DATA
SPHERE DATA
C.I. DATA
0.01
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNrn3
10
Figure 21. Measured entrainment concentration vs. drop diameter at the
E.S. inlet for the mobile bed (mode 1) scrubber.
1,000
100
10
E.S. DATA
LOG-NORMAL
CUBIC
SPHERE DATA
DATA
10
CUMULATIVE MASS CONCENTRATION, ml/DNm3
Figure 22.
Measured entrainment concentration vs. drop diameter at the
E.S. inlet for the mobile bed (mode 2) scrubber.
-------
l.OOC
100
10
LUB1C
SPIIF.RE DATA
UX;-NORMAI.
1,000
100
10
U.S. DATA
CUBIC
LOG-NORMAL
SPIII kl DATA
C.I. DATA
(1.01
0.1
CUMUUTIVE MASS CONCENTRATION, nl/TOm
0.1 1
CUMIII.ATIVI: MASS CONCKXIRATION,
10
Figure 23. Measured entrainnent concentration vs. drop diameter at the U.S. inlet for the
•obile bed (aodc 3) scrubber.
Figure 24. Measured entrainnent concentrution vs. drop diameter at the
I:.S. inlet for the mobile bed (mode 4) scrubber.
-------
1,000
100
10
LOG-NORMAl
SPHERE DATA
E.S. DATA
C.I. DATA
0.01
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNm3
10
Figure 25. Measured entrainment concentration vs. drop diameter at the
E.S. inlet for the mobile bed (mode 5) scrubber.
1,000
100
10
CUBIC
LOG-NORMAl.
E.S. DATA
C.I. DATA
I I i I
I I
0.01 0.1 1 10
CUMULATIVE MASS CONCENTRATION. ml/DNm3
Figure 26. Measured entrainment concentration vs. drop diameter at the
E.S. inlet for the sieve high scrubber.
-------
1,000
100
10
LOG-NORHAL . CUBIC
E.S. DATA
C.I. DATA
i i i i 11
1,000
0.001
0.1 1
OJHULATIVE MASS CONCENTRATION, •1/DN»'
10
0.001
0.01 0.1
CUMULATIVE MASS CONCENTRATION, ml/DNm1
Figure 27. Measured entrainrent concentration vs. drop diameter at the
E.S. inlet for the sieve low scrubber.
Figure 28. Measured entrainnent concentration vs. drop diameter at the E.S. inlet for the
spray 5 scrubber.
-------
1,000
100
IS
s
s
10
0
E.S. DATA
CUBIC
LOG-NORMAL
C.I. DATA
0.001
0.01 0.1 ( 1
CUMULATIVE MASS CONCENTRATION, ml/DNm]
10
Figure 29. Measured entrainment concentration vs. drop diameter at the E.S. inlet for the
spray 7 scrubber.
1,000
100
10
E.S. DATA
CUBIC
LOG-NORMAL
C.I. DAT,
0.001
0.01 0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNm3
10
Figure 30. Measured entrainment concentration vs. drop diameter at the E.S. inlet for the
spray 4 scrubber.
70
-------
1,000
• I I I III I II
100
a.
§
o
10
LOG-NORMAL
C.I. DAT
CUBIC
I 1 I I I I I I I I A t I I I I I
0.001 0.01 0.1 1 10
CUMULATIVE MASS CONCENTRATION, ml/DNm3
Figure 31. Measured entrainment concentration vs. drop diameter at the E.S. inlet for the
spray 6 scrubber.
-------
indicative of the sharp depletion of the large drops due to
gravity settling and deposition in the turning vanes. On the
other hand, the C.I. data for the smaller sizes show a steep
slope, indicating the minor contribution to the total drop
concentrations of drops of less than 10 ym diameter. The C.I.
distributions for all types of scrubbers and conditions do .not
vary a great deal, which shows a relative independence of the
size distribution of drops smaller than 20 urn diameter to the
atomization mechanism within the scrubber.
Size Distributions from Sphere Collector Data
Inertial collection efficiency of drops in a gas stream
on solid spheres depends on a number of factors, indluding the
drop diameter. If all the factors except drop diameter are
known, then the experimentally measured collection efficiency,
when compared with empirical correlations, should yield infor-
mation about the drop diameter. Complications arise when a
distribution of drop sizes exists, requiring either a large
number of experimental conditions (collector diameter, gas
velocity, etc.) or an assumed type of size distribution.
In the experiments the amount of liquid collected by
the sphere was divided by the volume of gas swept by the sphere
(based on the cross-sectional area) to determine the collected
drop concentration. The collection efficiency was then deter-
mined by dividing the collected concentration by the total
concentration measured at the same location by another means.
This other means was usually the cascade impactor. The mea-
sured sphere collection efficiencies are presented in Table 9.
These are overall efficiencies since a distribution of sizes
impacted on the sphere collectors.
The efficiency of collection by spheres has been empir-
ically correlated by Calvert (1970):
L
+ » C34)
72
-------
TABLE 9 . OVERALL EFFICIENCY OF SPHERE COLLECTORS
Scrubber Configuration
Mobile Bed, Mode 1
Mode 2
Mode 3
Mode 4
Mode 5
n, Overall Collection Efficiency, fraction
3.8 cm Sphere 7.6 cm Sphere
1.16
1.83
2.05
1.53
1.20
.96
1.40
1.67
1.30
.93
Sieve Plate, Low AP
High AP
Spray, 5 B-50
7 B-50
4 ST-24
6 ST-24
0.94
0.94
1.09
0.90
0.68
0.69
0.80
0.56
* The entrainment loading for the sieve plate runs was too light
to collect a measurable amount of liquid with the spheres.
73
-------
where K =
V
C'
-, drop impaction parameter
Because the drops collected are not of uniform diameter, the
efficiency of the sphere collector is the collection efficiency
at each drop diameter integrated over the drop size distribution;
n - / n
(dd) f(dd) ddd
(35)
where
drop size distribution density function
overall efficiency, fraction
The log-normal distribution required numerical integration
of equation (35) . The results are shown in Figure 32 for 3
different values of the geometric standard deviation (a ) .
o
For the cubic distribution, equation (36) could be inte-
grated analytically, with the result:
n =
3.675
4.392
tan
(36)
where the arctangent function is in radians and
2.7
a =
(37)
and d, was defined on Figure 32 . The equation is plotted on
Cl 5 o
Figure 33.
Correlation of the collection data in Table 9 with these
log-normal and cubic distributions would normally involve the
following procedure:
74
-------
Figure 32. Theoretical overall efficiency of sphere
collectors on log-normally distributed drops.
0.01
10
d /d,
mm d s o
Figure 33. Theoretical overall efficiency of sphere
collectors on drops with the cuhic
d i st r ihut i on.
-------
1. Select a "d " or the cubic distribution.
2. Find the "d^ " or "dmm" for which the two data points
Cn" vs. d, /d(j50) for each d^5fl lie on the theoretical
line.
Since only two data points were available a unique "o~ , d, "
combination could usually not be determined for the log-normal
distribution. However, the standard deviation was expected to be
small (~1.5) so that the data could be used to determine "d, ".
A rationale for this expectation is given below.
Expected "a " of Large Drops
Interpretation of size distribution data from the various
collection methods would be greatly aided by any additional
knowledge of the true distribution. One bit of knowledge avail-
able is the standard deviation of large drop diameters.
Few drops larger than 200 ym are expected to penetrate the
turning vanes and horizontal run. Predictions of the collec-
tion efficiencies of these collectors, made in Section 3,
lead to this expectation. Because of the lack of large drops
the drop size region immediately below about 200 ym will have
a fairly uniform distribution. If we assume a log-normal form
of the distribution the expected geometric standard deviation
(a ) for the distribution of diameters between 50 ym and 200 ym
o
would be around 1.5. In the following discussion this hypothe-
sis is supported.
The cumulative mass concentration is the product of the
total mass concentration and the cumulative mass distribution
function, CTF(dd). The penetration of a drop of diameter, d,,
through a collection device, such as a horizontal run or turn-
ing vanes, is the ratio of the derivatives of the outlet and
inlet cumulative mass concentrations:
d [CTn Fn (d,)]
Pt (dd) = . T° ° d (38)
a [CT1 F. (d,)]
where 'o'refers to outlet
'i'refers to inlet
76
-------
Since the derivatives of the cumulative mass loading function
(F) are the density functions (f) and the ratio of the total
concentrations is the overall penetration, the outlet density
function is:
£ » — £. (39)
0 Ft x
Finally, the outlet cumulative mass distribution is the integral
o
In Section 3 the penetration (Pt) through the turning
vanes and the horizontal run were predicted for an E.S. velo-
city of 4.6 m/s. Also, the overall penetration (Pt) through
the two collectors was predicted, assuming that the drops up-
stream of the turning vanes had a log-normal distribution with
a geometric standard deviation of diameters of 2.0. We need
only assume a geometric mass mean inlet drop diameter (dj )
to use these predictions to calculate the distribution at the
eml of the horizontal run.
Figure 34 shows the calculated size distribution using the
two above equations for an assumed d, of 500 ym at the scrubber
which is a reasonably expected drop diameter from the spray
scrubber. The distribution in the region of diameters greater
than 50 ym is approximately log-normal with d, =125 and a =1.6.
Below ym the size distribution changes, with the standard de-
viation in diameters increasing. Similar results occur for
other types of scrubbers. Thus, in subsequent discussions we
will assume where necessary, that the large drops entering the
E.S. have a a =1.5 which is very close to that calculated above.
o
Sphere Data Analysis
Correlations of the sphere collection data in Table 9.
with the theoretical graphs could not be made directly because
the measured efficiencies were usually above 100%. As dis-
cussed previously, gravity settling of drops on the spheres
contributed somewhat to the higher measured efficiencies.
77
-------
oo
300
200
e 100
of.
OJ
06
Q
50
30
20
10
u., (E.S.) = 4.6 m/s
\3
dd (INLET) = 500 urn
a (INLET) = 2
o
I I I
0.001 0.1
10
30
50 70
90 95 98
F , CUMULATIVE % BY MASS
Figure 34. Calculated outlet drop size distribution of turning vane/horizontal run.
-------
However, this contribution was estimated to be only about 5%.
Another contributor to the higher measured collection
efficiencies relative to the cascade impactor sampling nozzle
could be the higher probability of the larger spheres collec-
ting a few large reentrained drops that drip from the duct
ceiling or are sheared from the side walls. Also, the varia-
tion in size distribution in the air stream from top to
bottom of the spheres was greater than that across the sampling
nozzle. Thus, the spheres may have encountered a slightly
larger size distribution than the sampling nozzle.
The sphere data are useful if the ratio of the two mea-
sured efficiencies are used so that most discrepancies between
actual and ideal collection mechanisms can be neglected.
Correlation to the assumed size distributions was done as
follows:
1. Assume the smaller, more efficient, sphere has an
efficiency predicted by the theory.
2. Assume either a "o"e" or "oV1 or the cubic distribu-
tion, find the "ddgM or "dmm" for which the ratio of the mea-
sured efficiencies matches the theoretical efficiency ratio
for the two "dj50's" °^ t*ie two sPneres-
The results of this correlation procedure are given in
Table 10. A standard deviation for the log-normal distribution
of 1.5 was assumed, for reasons discussed previously. The size
distributions are plotted on Figures 21 through 31 for the
three collection devices. For the mobile bed the E.S. and
sphere collection data agree closely. The sphere data indi-
cate a smaller size distribution than do the E.S. data for the
spray scrubber runs.
Nozzle Manufacturers' Drop Size Data
Another way to check our measured drop size distribution
from the spray scrubber is to use the nozzle manufacturer's
data to describe nozzle output and to estimate the change in
drop size distribution as the entrainment travels to the E.S,
79
-------
TABLE 10.SIZE DISTRIBUTIONS AT E.S. INLET DERIVED FROM SPHERE COLLECTOR DATA
Scrubber
Configuration
Mobile Bed 1
2
3
4
5
Spray 5
7
4
6
d, urn
dso»
3.8 cm
Sphere
39
39
39
42
45
48
48
48
48
dd50, ^
7.6 cm
Sphere
56
56
56
59
63
68
68
68
68
Log-Normal Dist.
dd , ym
90
80
85
100
65
55
55
55
40
a *
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
Cubic Dist.
dmm' »m
100
70
95
110
80
65
70
70
50
* Assumed
80
-------
Spraying Systems Company provides some drop size distri-
bution data for their nozzles. The Whirljet size 50, which
we used, has a cumulative mass fraction that is a function of
the cube of the drop diameter at the lower end (less than 700
ym diameter) at 150 psi (1034 kPa). Assuming that the size
distribution is similar at the 30 psi (207 kPa) pressure we
used, the equation for the cumulative loading fraction at
thw lower end would be: ,
in ym) (41)
This equation is an extrapolation for diameters smaller
than 440 ym.
Based on the spray concentration at the nozzle and the
measured entrainment concentration entering the E.S. a maximum
expected drop diameter, and a mass median diameter of drop
entering the E.S. can be calculated. For the "Spray-5" scrub-
ber configuration the liquid flow rate through the spray noz-
zles was 2.68 i/s and the gas flow rate was 1.42 m3/s. Thus,
the total initial spray drop concentration was 1.89 £/m3. The
measured concentration at the E.S. inlet was 9.15 ml/m3, so
that Fnoz = 9.15 x 10"3/1.89 =0.00484. Using the above equa-
tion the drop diameter corresponding to this "F " is 249 ym.
no
-------
drop size measurements, it is important to point out what has
been demonstrated and what improvement remains to be done.
Entrainment characteristics were well enough established to
give a useful basis for E.S. design or specification.
Sizing Methods
1. The cascade impactor method using the salt tracer
provided good drop size distribution data for diameters between
1 and 20 ym.
2. The overall penetration and outlet size distribution
data from the E.S. were correlated to log-normal and cubic
inlet size distributions using a mathematical model for E.S.
grade penetration. The correlations were very good in a few
cases and moderately good in most cases.
3. The volumes of drops collected by the sphere collectors
were correlated to log-normal and cubic inlet size distributions
using a mathematical model for sphere collection efficiency.
After compensating for the greater than theoretical efficiency
of the spheres, the data correlated to the size distributions
adequately. Further investigations of sphere collection effi-
ciency under more controlled conditions would be needed to
completely define the mechanisms involved.
4. The hot wire drop counter did not perform successfully
under the conditions of these tests. The fragility of the 5 ym
diameter wires and the use at high gas velocities were the prob-
able causes of the poor performance.
Entrainment Size Distribution
The size distributions of the entrainment produced from
the three types of scrubbers are summarized in Figure 35 through
45. Drop size was measured at the E.S. inlet not at the scrubber
"outlet", so it was necessary to compute the probable distribu-
tion at the scrubber outlet. The solid lines represent the
average of the data based on cascade impactor, sphere, and E.S
sampling at the E.S. inlet. The dashed lines represent the
82
-------
size distribution directly above the scrubbers, before the
turning vanes and the horizontal run.
Scrubber outlet size distribution was calculated from
the E.S. inlet curve and the theoretical penetrations through
the turning vanes and the horizontal run. Because the E.S.
inlet data end at a drop diameter of about 150 urn, the scrub-
ber size distribution is shown to end there also. The actual
scrubber outlet drop size distribution could include a high
concentration of large drops not indicated in these figures.
Good scrubber E.S. design practice would prevent these large
drops from entering the E.S.
in these figures.
Effectiveness of the Pre-Separators
The turning vanes and horizontal section were very effec-
tive at reducing the loading on the E.S. when the entrainment
contained a large percentage of large (> 150 ym) drops. This
effectiveness can be seen in Figures 35 through 45 where the
differences in the two size distributions are due to the pre-
separators. As is seen in the case of the sieve plate scrub-
bers, the pre-separators were not so effective for drops smal-
ler than about 150 ym.
Comparison with Other Data
The only entrainment size distribution data for scrubbers
and conditions similar to those used here that could be found
were those of Calvert, et al. (1977), cited earlier in Section
3. They measured the entrainment size distribution above a
mobile bed scrubber that had a superficial gas velocity of 3.7
m/s and a liquid to gas ratio of 7.8 fc/m3, which was very simi-
lar to our Mode 1 conditions.
They sampled in a vertical duct using a University of Wash-
ington cascade impactor with filter paper substrates and a pre-
cutter to remove large drops. The substrates were analyzed for
NaCl used as a tracer. A comparison of the 1977 data with the
present data is given in Figure 46.
83
-------
The present data show a drop concentration about double that
of the 1977 data. The reasons for the difference are not clear,
although the mobile beds were of different size and configuration
and there were dissimilarities in sampling devices and methods.
We have noted that gas flow distribution is extremely important to
the mode of operation of a mobile bed scrubber, especially if a
single stage is used.
84
-------
OO
tn
1,000
1,000
100
10
0.01
AVERAGE MEASURED
AT E.S.
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNm3
10
Figure 35. Composite drop size distributions at entrainment separator
inlet and at scrubber outlet, for mobile bed (mode 1) scrubber.
100
I
10
AVERAGE MEASURED
AT E.S..
V
:ALCULATED AT
SCRUBBER
o.oi
o.i i
CUMULATIVE MASS CONCENTRATION, ml/UNn1
10
Figure 36. Composite drop size distribution at entrainoent separator
inlet and at scrubber outlet, for mobile bed (node 2) scrubber.
-------
1,000
100
oo
P
I
3
10
AVERAGE MEASURED
AT E.S.
LCULATED AT
SCRUBBER
1,000
100
10
AVERAGE MEASURED
E.S.
.LCU LATE D AT
SCRUBBER
0.01
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNm3
10
Figure 37. Composite drop size distributions at entrainment separator
inlet and at scrubber outlet, for mobile bed (mode 3) scrubber.
0.01
Figure 38.
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNmJ
10
Composite drop size distributions at entrainment separator
inlet and at scrubber outlet, for mobile bed (node 4) scrubber.
-------
1,000
100
I
ct
I
00
10
AVERAGE MEASURED.
E.S.
:ALCULATED AT
SCRUBBER
1,000
100
10
AVERAGE MEASURED
AT E.S.
CALCULATED AT
SCRUBBER
0.01
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNm3
10
0.01
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNm!
10
Figure 39. Composite drop size distributions at entrainment separator
inlet and at scrubber outlet, for mobile bed (mode S) scrubber.
Figure 40. Composite drop size distributions at entrainnent separator
inlet and at scrubber outlet, for the sieve high scrubber.
-------
1,000
1,000
100
a.
12
1
o
10
AVERAGE MEASURE
E.S.
CALCULATED AT
SCRUBBER
0.01
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNm3
10
Figure 41. Composite drop size distributions at entrainment separator
inlet and at scrubber outlet, for sieve low scrubber.
100
10
AVERAGE MEASURED
E.S.
'CALCULATED AT
SCRUBBER
0.01
0.1 1 10
CUMULATIVE MASS CONCENTRATION, ml/DNm5
100
Figure 42. Composite drop size distributions at entrainment separator inlet and at scrubber
outlet, for spray 5 scrubber.
88
-------
1,000
100
1
o
a.
i
10
AVERAGE MEASURED
AT E.S..
r
CALCULATED AT
SCRUBBER
0.01
0.1 1 10
CUMULATIVE MASS CONCENTRATION, ml/DNm3
100
Figure 43. Composite drop size distributions at entrainment separator inlet and at scrubber
outlet, for spray 7 scrubber.
1,000
100
10
AVERAGE MEASURED
AT E.S.^
CALCULATED AT
SCRUBBER
0.01
0.1 1
CUMULATIVE MASS CONCENTRATION, ml/DNm3
10
100
Figure 44. Composite drop size distributions at entrainment separator inlet and at scrubber
outlet, for spray 4 scrubber.
89
-------
VO
o
1,000
100
10
AVERAGE MEASURED
AT
CALCULATED AT
SCRUBBER
1JO
10
0.1
1977 DATA
\
RESENT DATA
0.01
0.1 1 10
CUMULATIVE MASS CONCENTRATION, ml/DNm3
100
Figure 45. Composite drop size distributions at entrainment separator inlet and at scrubber
outlet, for spray 6 scrubber.
0.001 0.01 0.1
CUMULATIVE MASS CONCENTRATION, ml/DNmJ
Figure 46. Comparison of E.S. size distributions of
mobile bed, mode 1 vs. data of Calvert, et al.
(1977).
-------
SECTION 5
ENTRAINMENT SEPARATOR PERFORMANCE
The performance of the entrainment separator (E.S.) on
entrainment from the three scrubbers was measured, and compared
with the model of Calvert, et al. (1974). The efficiency deter-
minations were based on the drop loading and size distribution
at the E.S. inlet, as discussed in Section 4, and outlet drop
loading and size distribution as measured with cascade impactors.
E.S. CONFIGURATIONS
The zigzag baffle E.S. depicted in Figure 3 was built in row
modules. This construction allowed testing of any number of
rows from 1 to 6 at any one time. In this program only 2 con-
figurations were run: 3 rows and 6 rows of baffles.
One other design parameter was varied: the slope of the
baffles. In addition to the vertical baffle configuration, a
series of tests were run with forward (i.e., upstream) sloping
baffles. The design was similar to the vertical design except
that all the baffles were sloped forward at the top, 30° from
the vertical.
This design should allow the liquid film, which has a hori-
zontal velocity component due to the drag of the gas, to have
a better chance flowing the full length of the baffle before
reaching the aft edge. Should much liquid flow to the aft edge
the film thickness there would increase greatly and the liquid
would be susceptible to reentrainment. Thus, it was expected
that the sloped configuration would be more efficient than the
vertical. The sloped baffles were tested in 3 and 6 rows con-
figurations .
91
-------
PERFORMANCE MEASUREMENTS
The outlet drop loadings and size distributions were
measured with the special cascade impactor (C.I.) using the
chloride ion tracer method described in Section 4. The mea-
surements are presented in Tables A4-A6 of Appendix "A". Over-
all penetrations were calculated based on inlet and outlet C.I.
measurements and are presented in Table 11. Overall penetrations
for vertical and sloped baffles are not shown separately because
there was no significant variation.
Grade Penetrations
Determination of the grade penetration was very difficult
because of uncertainties in several measurements. The inlet
and outlet size distributions were so different that the outlet
size distributions usually had only a few percent of the drops
larger than 20 ymA. In the case of the sieve plate scrubber,
grade penetrations determination accuracy was limited because
the drop concentration was so low.
Figure 47 presents a plot of the overall penetration data
versus the cumulative mass fraction of inlet drops smaller than
15-20 ymA diameter. It can be seen that the penetration is sub-
stantially equal to the fraction of inlet drops smaller than
20 ym diameter. This is clear evidence that the E.S. cut dia-
meter is about 20 ym, as is should be.
The high penetration of the drops through the 3 row E.S.
during spray scrubber operation is an indication that reentrain-
ment was occurring. The drop concentration from the spray
scrubber was relatively high and this probably caused more re-
entrainment. The low penetration for the mobile bed operation
suggests that the cut diameter was smaller than about 20 ymA,
which is in keeping with the high gas velocity in the E.S. during
the mobile bed runs.
92
-------
TABLE 11. SUMMARY OF PERFORMANCE OF THE ENTRAINMENT
SEPARATOR BASED ON CASCADE IMPACTOR SAMPLING
Scrubber
Configuration
Mobile Bed - 1
Mobile Bed - 2
Mobile Bed - 3
Mobile Bed - 4
Mobile Bed - 5
Sieve Plate - low
Sieve Plate - high
Spray - 5
Spray - 7
Spray - 4
Spray - 6
E.S.
Velocity
m/s
6.9
6.9
6.9
6.1
5.3
2.3
3.4
4.6
4.6
4.6
4.6
Overal
Penetrat
%
3 -Row
0.82
1.5
31.0
14.0
2.9
1.4
1.4
3.0
1
ion
6 -Row
0.82
0.84
1.3
2.5
5.5
26.0
20.0
0.49
0.89
1.3
1.1
93
-------
H
0.1
m
Cu
i
w 0.01
o
0.001
I I I I I I I I I I I
SCRUBBER TYPE E.S. CONFIGURATION
D Mobile Bed
O Sieve Plate
Spray
Mobile Bed
Sieve Plate
A Spray
0.001 0.01 0.1 1
CUMULATIVE MASS FRACTION OF INLET DROPS SMALLER THAN 20 ymA
Figure 47. Measured overall penetration correlated to cumulative
mass fraction of drops smaller than 20 ymA.
94
-------
COMPARISON WITH THEORETICAL MODEL
The theoretical efficiency of zigzag baffles was predicted
in Section 3, based on the model developed by Calvert, et al.
(1974) . Theoretical grade efficiencies were presented in
Figures 4, 5, and 6 for the geometry and gas velocities used in
the pilot E.S.
For 3 rows the predicted cut diameters (50% efficiency)
were between 30 and 50 ym while for 6 rows they ranged from 21
to 36 ym. Predictions of collection efficiency for 15 ym dia-
meter drops ranged from 0.05 to 0.3 which agrees with the experi-
mental results that most of the drops smaller than 15 ym diameter
penetrated the E.S.
When compared to the entrainment size distributions given
in Figures 35 to 45, these results show the expected trends of
lower penetrations for higher velocity and larger drop sizes.
The low penetrations during the high velocity mobile bed tests
indicate that reentrainment was not occurring. The relatively
high penetrations for the sieve scrubber operation are due to
a combination of the small drop size from the scrubber and the
low gas velocity in the E.S.
The data agree with the model and do not indicate any errors
in the model. The accuracy of the grade penetration curves was
limited, so the model could not be tested with precision. The
model is useful for design purposes, where ignorance of the
entrainment drop size distribution usually imposes a severe
limitation on one's ability to predict performance.
PRESSURE LOSS
The pressure loss in the E.S. was measured for superficial
gas velocities between 2.2 and 7.0 m/s for 3 and 6 rows of baf-
fles. The data correlate to the following equations:
for 6 rows, AP = 5.88 UG* , Pa (42)
for 3 rows, AP = 2.65 UG* , Pa (43)
95
-------
where AP = pressure loss, Pa
UG = superficial gas velocity, m/s
These data agree with those of Calvert, et al. (1974). The
data may be correlated to a general pressure loss equation as
follows. Assume the equation has the form,
n
£P IpG (uGi')2 (44)
where f = pressure loss coefficient, dimensionless
PG • gas density, kg/m 3
uGi'= actual velocity between baffles of row i, m/s
n = number of rows of baffles
for our geometry,
(45)
then, AP - fp 7 PG u> f, 15u + cos^oO (47)
The average value of f based on the data, was f =1.2.
Conversion of Units to cm W.C.
It is often convenient to express pressure loss in cm of
water column. The conversion factor is: 1 cm W.C. = 98.06 Pa,
Thus equation (44) would become:
n
AP (cm W.C.) = 0.0102 f P(, (U(Ji ') (48)
96
-------
REENTRAINMENT CHARACTERISTICS
A few runs were made to determine the onset of reentrain-
ment for the 6 rows of vertical baffles. The determination
of onset of reentrainment was made visually. With all 6 rows
in place reentrainment was not observed even at the limit of
system performance: volume flow rate = 2.6 m3/s, superficial
gas velocity, UG = 8.4 m/s, and spray scrubber QL/QG, •
2.4 i/m3.
When 2 rows were removed (4 baffles in place) reentrain-
ment was observed beginning at a volume flow rate of 2.12 m3/s,
UG = 6.9 m/s and spray scrubber, QL/QG, =2.97 A/m3. Heavier
reentrainment was observed at higher velocities.
The onset of reentrainment occurred simultaneously with
the flow of the liquid from the front baffle surface to the
rear surface. The flow of the liquid around the rear edge
and onto the rear surface was caused by the combination of the
large liquid film thickness due to heavy loading and the trans-
verse drag of the high velocity gas on the film. Apparently
as the liquid flowed around the baffle edge, some of it was
sheared off and consequently reentrained into the gas flow.
The fact that reentrainment will occur at lower velocities
when a fewer number of rows of baffles are used indicates that
the final rows are collecting reentrained drops from the for-
ward baffles. In many installations where the entrainment
loading is high, extra rows are needed for this reason and
not to increase the primary efficiency. If the entrainment
drop sizes are relatively large, there is little advantage to
more than about 3 rows for a system such as our pilot E.S.
unless the loading is heavy enough to cause reentrainment
in the forward rows at the operating velocity.
97
-------
E.S. OUTLET SOLIDS EMISSION LEVELS
The penetration data of this section of the report and
the inlet size distribution data of the previous section can be
used to estimate E.S. solids emissions from the configurations
we tested. Slurry concentrations of 15 wt. % and 25 wt. % were
used to compute solids emissions over the range from the usual
level to a convservatively high one.
Table 12 presents the solids emission concentrations for
the E.S. used in this program. Emissions from the mobile bed
E.S. would range from 4 to 14 mg/DNm3 for 25% slurry and from 2.3
to 8.4 mg/DNm3 for 15% slurry. The highest emission occurred for
mode 5, for which the liquid/gas ratio in the scrubber was the
highest studied.
In the section on design (Section 3) we assumed that the
E.S. operating on entrainment from a scrubber similar to our
Mobile Bed, Mode 1, would emit 4 mg solids/DNm3 while we measured
an emission level corresponding to a 6 mg/DNm3 solids. Since
the entrainment emission rate from the scrubber in the present
study was twice that assumed for the specification of design
efficiency, while the E.S. emission rate is only 1.5 times the
design rate, it is obvious that the E.S. is more efficient than
specified. Predicted solids emissions are within an acceptable
limit for the mobile bed and sieve plate scrubbers, even for 25%
slurry. Spray scrubber emissions are higher, due to the high E.S,
inlet liquid loadings, but could be tolerated at 15% slurry con-
centration. The use of a more efficient pre-separator would be
advisable for the spray scrubber in order to reduce entrainment
emission and ease the required scrubber efficiency.
98
-------
TABLE 12. E.S. OUTLET SOLIDS EMISSIONS BASED ON DROP PENETRATION
10
Scrubber
Configuration
Mobile Bed
Mode 1
Mode 2
Mode 3
Mode 4
Mode 5
Sieve Plate
Low
High
Spray - 5
Spray - 7
Spray - 4
Spray - 6
E.S., AP
cm W.C.*
2.9
2.9
2.9
2.2
1.7
0.3
0.7
1.3
1.3
1.3
1.3
Inlet Liquid
Drop Cone. ,
ml/DNm3
2.44
1.54
1.26
1.19
0.85
0.086
0.14
9.15
7.32
3.22
3.76
Pt for
6 rows
%
0.82
0.84
1.3
2.5
5.5
26.0
20.0
0.49
0.89
1.3
1.1
Outlet Solids
Cone, in Gas
@25% solids
6.0
3.9
4.9
8.9
14.0
6.7
8.4
13.5
19.5
12.6
12.4
mg/DNm
(§15! solids
3.6
2.3
2.9
5.3
8.4
4.0
5.0
8.1
11.7
7.6
7.4
*1 cm W.C. = 98.06 Pa
-------
SECTION 6
SOLIDS DEPOSITION
INTRODUCTION
The characteristics of solids deposition in the pilot scrub-
ber/E.S. were determined during operation with a limestone slurry
scrubbing liquid. The deposition rate on the baffle surfaces
was measured for a series of run conditions and E.S. configu-
rations. Comparison with predictions was made and attempts
were made to modify the model to more accurately predict the
deposition rate.
SLURRY
The slurry consisted of commercially available fine lime-
stone (CaCOs) powder, nominally 325 mesh size. When sized with
a Coulter counter the mass median particle diameter was found
to be 12 ym. The particle density of CaC03 is 2.7 g/cm3.
PROCEDURES
The character of the solids deposition on the zigzag
baffle surfaces was determined by operating the pilot system
in a normal manner with slurry. For each condition the scrub-
ber was operated approximately 7 hours per day for at least 7
days. The slurry was well stirred before starting each day
and samples were taken twice a day for concentration.
At the end of a run set the entrainment separator (E.S.)
baffles were removed, inspected, and photographed. The deposits
were non-uniform in most cases, but average thickness could be
estimated. These average deposit thicknesses were used to
determine a rate of deposition, or a thickness buildup rate
for each run set.
100
-------
MEASUREMENTS
A total of 6 series of runs were made, each lasting a
number of days. The run conditions and observations were as
follows:
Run Set 1
Scrubber nozzles: 6 Bete ST24FCN
Scrubber nozzle pressure: 69 kPa gauge (10 psig)
Duration: 6 days; 38 hours, total
Average slurry concentration: 2.7% by mass
Entrainment separator configuration: 6 rows of vertical
baffles
E.S. superficial gas velocity: 4.6 m/s
Because of inadequate mixing, the slurry solids concentra-
tion was only 2.7% by mass while the goal was 10%. The first 3
rows had no solids deposition and the last 3 had slight uni-
form deposition. The fifth row had the thickest deposit, which
was still less than 1 mm. The back surfaces of all 6 rows
showed no deposition.
Run Set 2
Scrubber nozzles: 5 Spraying System Company 1/2B50
Scrubber nozzle pressure: 207 kPa, gauge (30 psig)
Duration: 9 days; 48.5 hours, total
Average slurry concentration: 6.4% by mass
E.S. configuration: 6 rows of vertical baffles
E.S. superficial gas velocity: 4.6 m/s
The same trends were noticed as for the previous run set
but the deposition was heavier. The first 3 rows had no deposi'
tion except at the top, leading edge of the first row. The
last 3 rows had uniform deposition over the front surface with
row No. 5 having the heaviest deposition, reaching a thickness
of about 3 mm. No deposition occurred on the back surface of
the baffles.
101
-------
Run Set 3
Scrubber nozzles: 5 Spraying System Company 1/2B50
Scrubber nozzle pressure: 207 kPa, gauge (30 psig)
Duration: 8 days; 54 hours, total
Average slurry concentration: 6.4% by mass
E.S. configuration: 6 rows of vertical baffles
E.S. superficial gas velocity: 4.6 m/s
The solids deposition characteristics were:
1. The first 3 rows were clean, front and back, except the
first row had slight buildup on the leading edge for 5 cm from
the top.
2. Row 5 had heaviest deposition - about 3 mm thick over
the first third of the front surface.
3. Row 6 had about 1 mm of deposition, uniformly distri-
buted over the front surface.
4. The back surfaces of rows 4, 5, and 6 had slight
deposition.
Run Set 4
Scrubber nozzles: 4 Bete ST24FCN
Scrubber nozzle pressure : 69 kPa, gauge (10 ppig)
Duration: 13 days; 87.4 hours, total
Average slurry concentration: 5.3% by mass
E.S. configuration: 6 rows of vertical baffles
E.S. superficial gas velocity: 4.6 m/s
The solids deposition characteristics were:
1. Row 1 - Front has thick deposit on the top 12 cm. At
the front edge the thickness was the maximum, 0.6 cm. The lower
49 cm of the front surface and the rear surface were clean
except for a slight (<0.1 cm) deposit on the trailing edge.
2. Row 2 - Pattern similar to Row 1 but less in extent.
The rear surface was generally clean.
3. Row 3 - Front surfaces clean except for the baffle
next to the wall. This baffle had about 0.5 cm deposit over
the lower 40 cm of surface. The top of the baffles also had
deposits up to 1 cm thick on the front edge. The rear sur-
face had a very thin even deposit («0.1 cm).
102
-------
4. Row 4 - The lower 5 cm of the front surface was clean
and there was an even deposit about 0.2 cm thick on the upper
56 cm of the baffle. The rear surface had a thin (<0.1 cm),
even deposit.
5. Row 5 - Front surface had no deposit on lower 5 cm
nor at the very top. Thickness varied from top to bottom with
a maximum of 0.3 cm deposit 40 cm from the top. The rear sur-
face had an even, thin (<0.1 cm) deposit.
6. Row 6 - Front surface had a thin deposit (-0.1 cm
average) with the maximum at 53 cm from top. The lower 8 cm
were clean. The rear surface had a very slight deposit.
Run Set 5
Scrubber nozzles: 4 Bete ST24FCN
Scrubber nozzle pressure: 69 kPa, gauge (10 psig)
Duration: 8 days; 52.9 hours, total
Average slurry concentration: 3.3% by mass
E.S. configuration: 6 rows of baffles, sloped 30° forward
of the vertical
E.S. superficial gas velocity: 4.6 m/s
The solids deposition characteristics were:
Row 1 - Front surface was clean except for the top 5 cm
which had lumps on the front edge. The baffle next to the
wall had deposits over most of the front surface about 0.1 mm
thick. The rear surfaces were clean.
Row 2 - Same as row 1 except the baffle next to the wall
was clean.
Row 3 - Same as row 1 except the baffle next to the wall
had a deposit thickness of about 0.5 mm.
Row 4 - The upper third of the front surfaces had deposits
0.1 to 0.5 mm thick; the lower 40 cm was clean. The baffle next
to the wall had a 0.1 mm thick, uniform deposit on the front
surface. The rear surfaces were clean.
Row 5 - The baffle next to the wall was heaviest, but all
the baffles had 0.1 to 0,5 mm thick deposit on the upper 40 cm
of the front surface. The lower third was almost clean. The
rear surfaces were clean.
103
-------
Row 6 - All the baffles had a thin (0.1 mm) deposit on the
upper 45 cm of the front surface, except for the baffle next to
the wall. Its entire front surface was coated. The back sur-
faces were clean.
Run Set 6
Scrubber: Sieve plate § 1.0 m3/s, QL/QG =9.1 A/m3
Duration: 7 days; 47.8 hours, total
Average slurry concentration: 4.5% by mass
E.S. configuration: 6 rows of vertical baffles
E.S. superficial gas velocity: 3.4 m/s
The solids deposition characteristics were:
Row 1 - The upper 50 cm had 1 mm thick deposits in patches,
not uniform. Channels or clean vertical spaces were present.
The rear surface was clean.
Row 2 - Thick (1-2 mm) deposits occurred in spots. The
thickest region was 30 cm from the top and the thinnest was at
55-61 cm from the top. The rear surface was clean.
Row 3 - The front surfaces had a thin, 0.1 mm deposit.
Slight carry-over around the leading edge to the rear surface
was present, resulting in a thin coating near the leading edge.
Row 4 - A slight (<0.1 mm) deposit occurred on the front
surfaces. The rear surfaces were clean.
Row 5 - A very thin, even deposit coated the front sur-
faces. There was nothing on the rear surfaces.
Row 6 - Same as row 5.
RESULTS
Figures 48 and 49- present the average thickness rates for
spray scrubber and sieve plate scrubber operation, respectively.
The thickness rate has been normalized for a 101 slurry for
comparison purposes. Curves have been drawn for the average
of each row that had a measurable deposit.
The first 3 rows of the E.S. during spray scrubber operation
had no deposit except over a few centimeters at the top, leading
edge. There was no measurable deposit on the rear surfaces of
104
-------
0.1
3
o
e
0.05 _
y
P
5 Numbers refer
to baffle row
20 30 40 50
h. VERTICAL DISTANCE FROM TOP, cm
l-'ipurc 48. Measured deposition thickness rate for spray
scrubber operation.
0.1
3
O
c/)
c,=
o
0.05
Row 2
1 I
Run Set 6
J 1 1 1 '''I
10 20 30 40 50
h, VERTICAL DISTANCE FROM TOP, cm
60
Figure 49. Measured deposition thickness rate for
sieve plate scrubber operation.
105
-------
the baffles of any row. Based on the thickness rates shown in
the figures, it would take at least 100 hours of operation
before the deposits became thick enough to affect the pres-
sure drop
COMPARISON WITH SOLIDS DEPOSITION MODEL
Using data from a small single baffle, Calvert, et al (1974)
derived an empirical equation for the solids deposition rate.
We have modified the model slightly to account for slurry
particle size and to account for the deposition occurring
while the dry baffle surface is being wetted. The equation
for deposition rate is,
Rs = W $ exp [-k ( d^m- and when a < d :
— pmm pmm
Rs " W * (51)
106
-------
The thickness rate is related to the deposition rate by
36 R
s
(52)
ps
where T - thickness rate, mm/hour
p = solid deposit density, g/cm3
At the start of each daily run the deposit on the baffle
surface is dry. Before any slurry will run off the surface it
must be wetted. To account for the solids deposition occurring
during this initial wetting of the surface, the following
equation has been derived:
T' = T(l+fT)m (53)
where T' = corrected thickness rate, mm/hour
m = number of times the dry surface is wetted
and fT - V (54)
with e ~ porosity of surface deposit, fraction
f = volume fraction of slurry particles in the slurry,
fraction
The slurry flux, $, was not measured directly. It was
calculated based on the entrainment loading and size distri-
buiton and the penetration through the various rows of the
E.S. The penetration model was essentially that derived by
Calvert et al. (1974). The calculated slurry fluxes for the
operating conditions of the spray and sieve plate scrubbers
are given in Table 13.
The model is compared with the data in Figures 50 and 51
for the 5-nozzle spray scrubber amd the sieve plate scrubber,
respectively. The agreement is fairly good for row 6 of the
107
-------
TABLE 13. CALCULATED SLURRY FLUX ON ROWS 1 THROUGH 6*
Scrubber
Spray 5
Spray 4
Sieve-High
"G
in E.S.
m/s
4.6
4.6
3.4
Total
Loading
ml/m3
9.0
3.2
0.14
\
pmA
160
130
100
a
1.6
1.5
1.5
*»
(mg/cm
2.4
0.71
0.017
•i
2-s)
1.3
0.49
0.016
0
0
0
3
(mg/cm2
.31
.14
.0066
t
-s)
0.097
0.049
0.0031
<3
0.
0.
0.
5
fag/
037
019
0015
oo
* Method described with equations (56) - (58) in Section 7
0.016
0.0086
0.0008
-------
0.1
cc
a
3
ROW S
0.05 -
X
s-
MKASIWF.I)
— PREDICTED
I
10 20 SO 40 50
h, VERTICAL DISTANCE PROM TOP, cm
Figure 50. Comparison of measured and predicted
deposition rate for spray scrubber
run sets 2 and 3.
0.1
3
O
ac
a:
3
to
O
u.
01
f-
in
tu
0.05
i I I i r
\
1 I 1
MEASURED
PREDICTED
ROW 2
JL
10 20 30 40 50
h, VERTICAL DISTANCE FROM TOP, cm
60
Figure 51. Comparison of measured and predicted
rates for the sieve plate scrubber
operation.
109
-------
spray scrubber run and row 3 of the sieve scrubber run where
the deposit was the most uniform. A significant underesti-
mate is evident for row 5 of the spray scrubber operation
and row 2 of the sieve plate operation. It appears that
"end effects" or the effects of the ceiling and the end con-
struction of the baffles has caused a downward shift of the
location of heaviest deposit. Another real effect causing
this shift is gravity which reduces the entrainment loading
somewhat on the top few centimeters of the baffles. Also, the
higher than predicted deposit thickness could be caused by
periods of unsaturated operation.
Should the water drops be evaporating within or upstream
of the E.S., the liquid film thickness on the baffle would
decrease below that predicted by equation (50)« A thinner
liquid film would cause a higher deposition rate. The
humidity was monitored closely and on days with low ambient
humidity an extra fog nozzle was used upstream of the scrubber
to keep the air as nearly saturated as possible.
CONCLUSIONS
Because of the non-ideal flow pattern at the top of the
E.S., the non-uniform concentration of drops due to gravity
and some periods of unsaturated operation, the model under-
predicted the solids deposition rate. It was accurate to
within an order of magnitude and indicated correctly the
rows of the E.S. which should have significant solids deposi-
tion. The present model is useful for design purposes but
it should be developed further.
OPERATING EXPERIENCE
The main problems encountered due to use of the slurry
were pump seal wear and difficulty in mixing of the slurry
in the main sump tank. A ceramic mechanical seal was used
during the salt solution runs but it quickly wore out during
slurry operation. A packing seal was then installed. The
seal naturally dripped to enhance lubrication; a fresh water
110
-------
barrier was not used. The seal drippings were pumped back into
the scrubber sump with a small venturi pump attached to one of
the mixing pumps. Two small pumps were used to circulate the
slurry at various locations within the main sump tank to pro-
vide extra mixing.
The main pump was a standard design cast iron close-
coupled centrifugal pump. During the salt solution runs a
bronze, closed impellor was used. However, it became eroded
noticeably after a few days of operation on the slurry. A
cast iron,open impellor was substituted and no further prob-
lems were encountered.
Unsaturated Operation of E.S.
An unsaturated condition was noticed at the outlet of the
entrainment separator on a day when the relative humidity was
about 50%. Evaporation of drops in the entrainment separator
would reduce the amount of liquid flowing on the baffle sur-
face. This was undesirable since solids deposition rate
increases as the slurry flux and liquid film thickness decrease,
To try to ensure saturation a fog nozzle was placed
between the blower and the scrubber inlet. Thereafter the
humidity was closely monitored. The differences between inlet
and outlet wet bulb temperatures were kept within 0.5°C. In
addition to using the fog nozzle on dry days the E.S. baffle
surfaces were closely watched for wetness. Test runs were
not begun (blower started) until they were visibly wet from
the convection of mist from the fog nozzle and spray scrubber
nozzleSo
Solids Deposition Outside the E.S.
Solids deposition was significant in only one location
other than the E.S. in the pilot system. This was the hori-
zontal section of duct upstream of the E.S. which experienced
buildup on its floor. It was flushed periodically with scrub-
ber liquid (slurry). A sloping floor and a shorter horizontal
section would alleviate this deposition.
Ill
-------
A potential solids buildup location was the turning vane
section at the top of the vertical scrubber duct. However,
the slurry flux was large enough to keep the vanes self-washed
and no significant solids buildup occurred.
Design Considerations -
In general the designer of an entrainment separator system
for use on a slurry scrubber should consider the potential for
solids buildup throughout the system. For horizontal surfaces
solids deposition will occur wherever the liquid flow is slow
enough to allow slurry particles to settle out and slow enough
not to scour the deposited solids. These horizontal surfaces
include any ducting in which entrainment drops may settle out
by gravity and horizontal parts of the E.S. which carry away
the entrainment collected on the baffle surfaces.
Solids will deposit on vertical surfaces, such as the
E.S. baffles, when the liquid flow on the surface is too slow
to be self-washing (by scouring, etc.). Prime examples of
vertical surfaces that have potential for solids buildup are
horizontal ducting bends or turns. As in the E.S. a change
in gas flow direction will cause some of the slurry drops to
impact on the solid surface, creating a potential for solids
buildup.
Both horizontal and vertical surfaces will experience
rapid solids deposition if the carrier gas is allowed to
become unsaturated. If the gas is not saturated the entrain-
ment will evaporate resulting in either a higher solids
concentration in the drops or even dry particles instead of
drops. Besides causing a faster deposition rate, evaporation
of the entrainment will also reduce the efficiency of the
entrainment separator due to the smaller drop size of the
entrainment.
The solution to unwanted solids buildup is washing.
Spray nozzles should be located so as to wash or flush any
surface that has a high potential for solids buildup.
112
-------
Recirculated slurry can usually be used, although if fresh,
make-up water must be added to the system the wash system
would be an excellent location.
113
-------
SECTION 7
WASHING SYSTEM
INTRODUCTION
In this section the need for and requirements of a washing
system to remove deposited solids from entrainment separator sur-
faces will be described. A series of experiments will be discussed
to verify the performance of a candidate washing system. Finally,
criteria for washing systems of any scale will be developed.
NEEDS AND REQUIREMENTS
There exists a strong possibility that an entrainment
separator designer will be unable to avoid solids deposition,
either because the entrainment loading will be below that which
is self-washing or it will be above the loading which causes an
insignificant deposition.
In the previous section on solids deposition it was demon-
strated experimentally that a high slurry flux was self-washing
and will not permit solids to build up. Also in the case of
very light entrainment the deposition rate was insignificant.
Between these limits the deposition rate is significant and
unavoidable. It would appear that solids deposition could be
eliminated by operating at high enough entrainment loading to
keep the baffles self-washing. However, at these high loadings
the E.S. will probably not have the required efficiency due to
either low primary efficiency or reentrainment. Also, solids
will always deposit at the top of the leading edges of the first
few rows because the film thickness is essentially zero at those
locations. In some E.S. installations this buildup could slip
down, because of vibrations, etc., and cause significant plug-
ging problems.
114
-------
The location of the maximum solids buildup can be approxi-
mated by differentiating the deposition rate equation with res-
pect to the slurry flux, *, setting the result equal to zero, and
solving for *.__... The location on a certain row a certain dis-
IHcLX
tance from the top of the baffle is,
h-1:000!1'8 („„.„ "•' V ess)
3 yT »D I 0.0433 + 0.7067 $D J
L K \ K /
where $D = slurry flux for which the deposition rate is a
K
maximum, mg/cm2-s
and the other terms were defined in Section 6. If the vertical
distance of film travel, h, calculated from equation (55) is
between 1 cm and the baffle height then that row of baffles
will have the maximum deposition rate at that height. If the
"h" calculated from equation (55) is much less than 1 cm the
slurry flux is too great for significant solids deposition. And,
if the "h" calculated is much greater than the baffle height
then the deposition rate is insignificant, because the slurry
flux is so low. Equation (55) is plotted on Figure 52 for
typical conditions.
For the typical conditions given and a baffle height of
61 cm, Figure 52 shows that the rows for which the slurry flux,
$ was between 0.005 and 0.06 mg/cm2-s would have significant
solids buildup. These rows would be the ones requiring washing.
Requirements/Constraints
The basic requirements of the washing system is that it
wash the baffle surfaces in a short time using a minimum of
liquid and pumping power. If spray nozzles are used, they must
be arranged to provide coverage of at least the upper half of
the baffles. Recycled slurry should be used unless the system
requires fresh make-up water, which could be conveniently intro-
duced in the wash system.
115
-------
ioo r
s
*
a,
e
1
a,
8
H
(-1
O
u
10
0.1
EQUATION (55)
r PL = 1.05 g/cm3
y, = 0.01 g/cm-<
g = 980 cm/s2
0.001 0.01 0.1
*R, SLURRY FLUX FOR MAX. DEPOSITION RATE, mg/cm2-s
Figure 52. Slurry flux for maximum deposition rate.
116
-------
EXPERIMENTAL DESIGN
The wash system for experimentation consisted of several
small spray nozzles on a manifold located about 23 cm upstream
of the E.S. The nozzles were Spraying Systems Company Model
1/4-B-3 of type 316 stainless steel. They were operated at
210 and 350 kPa gauge pressure (30-50 psig) using slurry and the
drops produced had mass median diameters of 400 and 360 ym,
respectively. The rate of washing was controlled by the number
of nozzles and the pressure.
The washing system was tested at the end of solids deposi-
tion runs and using artificially coated baffles. An inter-
mittent mode was found to be satisfactory with the wash sys-
tem being run for only 10 minutes during the cycle.
RESULTS
The results of the wash system tests are shown in Figures
53 and 54 where the washing flux on each row of baffles is
plotted against the percent of the surface cleaned. Since
the "percent cleaned" may be imprecise due to the non-uniformity
of the deposit, these figures should be used primarily to deter-
mine the washing flux required for total cleaning. The washing
flux was calculated from the liquid flow rate, the spray drop
size distribution, the gas velocity, and the E.S. geometry.
Washing Flux Calculation
The flux of washing slurry, S> , hitting each baffle depends
w
on the amount of slurry penetrating the baffles of the upstream
rows.
Equation (8) is used for calculating the penetration of
zigzag baffles:
Pt = exp
w
(56)
where the terms were defined in Section 3.
117
-------
10.0
1.0
E
.C
00 -
0.1
0.01
-------
The deposition velocities, UDI , and UD, are evaluated using
the general drag coefficient relation, equation (7) . The
overall penetration, Ft, is equation (8) integrated over the
size distribution.
The loading hitting each row, n, can then be found from:
Cn * CT C^n-l * *•!> (57)
where n = row number
C = loading hitting row n, ml/m3
C™ = total loading, ml/m3
Pt0 = 1.0
and the loading hitting each row is related to the flux on
the surface by
PT C ur
»n ' -^^ (58)
where p, = liquid density, g/cm3
Up = superficial gas velocity, m/s
$n = slurry flux on row n, mg/cm*-s
The calculated penetrations are given in Table 14 for the 4 condi-
tions under which the wash system was run. The loading, CT, was
calculated from the liquid flow rate through the nozzle, account'
ing for wall losses.
Minimum Flux Required
For the superficial gas velocities of 2.3 and 3.4 m/s
the minimum slurry flux required to clean the baffles was 0.9
and 0.6 mg/cm2-s, respectively. The difference due to velocity
is probably the result of more reentrainment from upstream rows
at the higher velocity. The flux calculation is conservative
as it assumes no reentrainment.
119
-------
TABLE 14 . CALCULATED PENETRATIONS OF WASH SPRAY DROPS
Overall Penetration, Pt , fraction
UG, m/s
Row d, , ym
" '«'
1
2
3
4
5
6
2.3
360
2.2
0.27
0.069
0.029
0.016
0.0096
0.0064
2.3
400
2.2
0.24
0.057
0.023
0.012
0.0072
0.0047
3.4
360
2.2
0.24
0.053
0.020
0.0102
0.0059
0.0038
3.4
400
2.2
0.21
0.043
0.016
0.0077
0.0043
0.0027
120
-------
Comparison with Prediction
The solids deposition model is not directly applicable
for predicting required washing fluxes. However, from Figure
52 one would predict that the washing flux should be greater
than 0.1 mg/cm2-s. This value was also the flux, above which
no deposition occurred in the solids deposition experiments.
Since this is the flux required to continuously wash the baf-
fles it is logical that a somewhat higher flux would be re-
quired for short duraction intermittent washing of accumulated
deposits. Thus, the experimentally determined washing fluxes
do not conflict with the deposition model but they could not
be predicted with any precision by the model.
DESIGN CRITERIA
The variable that determines the washing effectiveness is
the washing flux on the baffles. It is a function of the
total wash spray flux into the E.S. and the penetration of
the spray drops through the preceding rows of baffles. Since
the penetration is a function of the E.S. geometry, gas velocity,
and the wash spray drop size distribution, the calculation of
the flux is complicated.
Typical Design
In order to estimate the amount of washing required in a
typical E.S. installation, we will assume the following typical
conditions:
E.S. geometry- as we used (Figure 3)
E.S. superficial gas velocity = 4.6 m/s (15 ft/s)
Spray drop size distribution:
ddg = 400 um
cg = 2.2
Washing duration = 10 minutes
Required washing flux, $wagh = 0.6 mg/cm2-s
121
-------
The calculated overall penetrations for each row are:
Row, n 1 2 3 45 6
PTn 0.161 0.0301 0.0104 0.0048 0.0026 0.0016
The total washing spray flux required upstream of the first row
to wash row n is then,
, mg/cm2-s (59)
i-l "n
which is given in Table 15 in terms of both spray flux (liquid
flow rate per unit E.S, cross section) and Qj/Q^. Rows 4, 5,
and 6 would probably require less washing than shown since
there would be considerable reeentrainment from preceding
rows at these high liquid loadings. The amount of washing
could be reduced by decreasing the gas velocity or lowering
the spray drop size at the cost of E.S. efficiency or increased
spray nozzle pressure, respectively. Washing rates could be
significantly reduced by separating a 6-row E.S. into two
3-row modules, also at the cost of E.S. efficiency.
122
-------
TABLE 15. PREDICTED TOTAL WASHING RATES FOR TYPICAL
E.S. FOR TEN-MINUTE OPERATION
(UG =4.6 m/s)
Row
Wash Spray Flux
Jl/m2-s
GPM/ft2
Wash Spray QL/QG
£/m3
gal/1,000 ft3
1
0.0072
0.011
0.0016
0.012
2
0.046
0.068
0.010
0.075
3
0.30
0.44
0.065
0.49
456
1.1 2.7 6.0
1.6 4.0 8.8
0.24 0.59 1.30
1.8 4.4 9.8
123
-------
SECTION 8
FUTURE RESEARCH RECOMMENDATIONS
The major goals of this research were met, however, addi-
tional work on some problems would be useful. Since the major
goals of the project were attained, the superior E.S. design
should be used in a demonstration phase. This phase would prove
that the design can maintain high performance and not be degraded
by a slurry in an actual practical application. Demonstration
at a coal-fired utility boiler that uses a lime or limestone
flue gas desulfurization scrubber would be desired. The design
gas flow rate should be an order of magnitude greater than the
present; i.e., at least 14 actual m3/s (30,000 ACFM).
GRADE EFFICIENCY
The nature of E.S. operation is such that precise measure-
ment of the drop size distribution and concentration is extremely
difficult. Drop size and concentration vary greatly in three
dimensions, especially at the upstream side of the E.S., where
large drops abound. One would have to make a tedious traverse
across the inlet face of the E.S. along a plane close enough to
the face so that drop sedimentation has a tolerably small effect.
The lack of a good drop sizing method for the 20 to 50 ym
diameter range is another deterrent to good drop size analysis.
Consequently, one does not have a good measurement of the average
drop size and concentration going into the E.S. The outlet drops
are better defined but the net result is that measurement of
drop collection efficiency as a function of drop diameter is
subject to considerable error if done for polydisperse entrain-
ment from a scrubber.
We have done some exploratory experiments which showed that
the use of monodisperse liquid drops is a promising way to
124
-------
measure E.S. collection efficiency. Drops in the broad range of
10 to 100 ym diameter are relatively easy to generate with such
devices as the spinning disk atomizer.
Sampling and analysis based on the use of a tracer compound
in the drops or optical methods can provide the data needed to
compute drop collection efficiency. Problems of drop distribu-
tion across the E.S. face and along the axis of the gas flow can
be readily resolved.
REENTRAINMENT LIMIT
Extension of our knowledge of the reentrainment limit of
the zigzag E.S. was not a part of the present study but it would
be useful to do. Several E.S. configurations which have been
developed in recent years are claimed to have high limiting gas
velocities and these should be investigated in addition to the
zigzags.
Because these newer configurations are complex and generally
involve the use of hooks or small depressed passages for liquid
flow, they are probably subject to plugging by solids. The influ-
ence of solids deposition on the performance of such E.S. con-
figurations should be determined experimentally.
SOLIDS DEPOSITION MODEL
The only available model for solids deposition was not satis-
factory for anything more than predicting general trends. It
was based on a simplified theoretical treatment which yielded
a semi-empirical equation. Constants in the equation were
evaluated with data taken in a limited experimental program.
In that experiment a single impaction baffle was sprayed with
slurry from an air-atomizing nozzle.
Observations made in the present study showed that air flow
patterns and drop deposition on the single element are not the
same as on one baffle among a zigzag assembly. It is clear that
a more realistic and extensive experiment should be performed
to obtain good data on solids deposition rate.
125
-------
A special experiment is required because the rate of solids
deposition in a simulated scrubber system such as used in the
present study is so low that the experimental run requires at
least several days of operation. An accelerated deposition
experiment, perhaps with automated, continuously operated
equipment, is needed.
The recommeded program should investigate important var-
iables, such as E.S. Type, surface cake type, surface roughness
and wettability, and slurry particle size. From the data and
applicable theory, engineering models should be developed to
predict deposition rate more accurately.
DEMONSTRATION PROGRAM
The advantages and performance of the type of entrainment
separator described in this report would be demonstrated at a
large scale pilot facility. The facility should be a large coal-
fired boiler that has a lime or limestone wet scrubber to con-
trol sulfur dioxide and particulate emissions. The scale of the
demonstration should be at least an order of magnitude greater
than the small pilot facility used in the program described in
this report; i.e., 14 m3/s (30,000 ACFM).
Program Objectives
The program objectives would be to demonstrate that the
E.S. performs with the desired efficiency and operates econom-
ically without being adversely affected by solids deposition.
Overall economics of this E.S. relative to other types used on
similar applications would be estimated.
Program Plan
The general program would be conducted as follows:
1. Site Selection:
A utility that has a pilot scale wet scrubber operating
on flue gas from a large coal-fired boiler is the desired
facility. An ideal site would be the EPA Alkali Scrubbing Test
Facility at the TVA Shawnee Power Station, Paducah, Kentucky.
This facility has 3 pilot scrubbers, each capable of treating
126
-------
up to 16.5 m3/s (35,000 ACFM) of flue gas at 150°C. The scrub-
ber outlet gas glow rates are about 80% of the inlet because of
the pressure and temperature drops through the scrubber. Two
of the pilot scrubbers are currently in operation: a venturi/
spray tower and a mobile bed. Both have horizontal zigzag baffle
entrainment separators located at the top of vertical scrubbing
towers.
2. Source Evaluation:
The drop size distribution, entrainment loading, and liquid
and gas properties and flow rates would be determined by a source
test. The efficiency of the present E.S. (if any) would be mea-
sured to permit performance and cost comparison. Overall des-
cription of the scrubber/E.S. system would be compiled to deter-
mine available structures and equipment that could be adapted
for a new E.S. and to facilitate evaluation of the impact of the
new E.S. on overall system performance and economics.
3. Preliminary Design:
The basic E.S. design will be the same as the vertical zig-
zag baffles described in this report. Several design variables
must, however, be considered:
a. Height of the vertical disengagement space above the
scrubber.
b. Incorporation of a pre-separator, such as turning vanes
in the vertical-to-horizontal elbow duct transition.
c. Length and slope of horizontal duct between the scrub-
ber tower and the E.S.
d. Cost/performance trade-off to decide pressure drop,
superficial gas velocity, scrubber cross sectional,
and number of rows of baffles,
e. E.S. baffle length.
f. E.S. drainage and slurry recirculation system.
g. Washing system design - slurry or make-up water.
h. Materials.
i. Provisions for performance testing.
127
-------
j. Instrumentation.
k. Test plan.
4. Engineering Design:
This phase would produce engineering fabrication, construc-
tion and installation, drawings and procedures based on the
design decided upon in the preliminary design phase. Detailed
costs would be determined and prospective subcontractors and
equipment vendors would be evaluated.
5. Installation:
Procurement of equipment, selection of subcontractors and
construction and installation of the E.S. system would occur
in this phase.
6. Startup:
During this phase the various problems that accompany the
operation of new equipment and systems would be solved. Inspec-
tion and maintenance procedures would be set up. Special instru-
mentation required for performance testing would be installed
and checked out.
7. Performance Testing:
As in the source evaluation phase the upstream and down-
stream entrainment and gas properties and conditions would be
measured to allow determination of overall and grade efficiencies.
The following variables would be evaluated:
a. Number of rows in E.S.
b. Superficial gas velocity and pressure drop.
c. Washing rate, frequency, and liquid.
d. Slurry concentration, particle size, and chemistry.
e. Materials - plastic versus stainless steel.
8. Evaluation:
This phase would involve the analysis and evaluation of the
data taken during the testing phase. The effects of the several
variables on the E.S. performance would be determined. The
economic benefits of the E.S. system would also be evaluated.
128
-------
Program Schedule
The demonstration program would be performed over a 2-year
period and involve 6 months of performance testing. A proposed
schedule is presented in Figure 55.
Program Costs
The program would cost approximately $400,000 and the break-
down is presented in Table 16.
TABLE 16 . PROGRAM COST ESTIMATE
Task
1
2
3
4
5
6
7
8
9
Labor
$ 20,000
30,000
50,000
50,000
30,000
10,000
90,000
40,000
30,000
$350,000
Other
$ 1,000
5,000
30,000
10,000
2,000
$48,000
129
-------
Figure 55. Proposed program schedule.
PROGRAM SCHEDULE
TITLE
Task
No.
1
2
3
4
5
6
7
8
9
• Entrainment Separator Demonstration | |
Task Description
Site selection
Source evaluation
Preliminary design
Engineering design
Installation
Startup
Performance testing
Evaluation
Report
2 4
—
MONTHS AFTER START
6 8 10 12 14 16 18 20 22 24
^^v
—
^ Scheduled Start or Completion ^ Actual Start or Completion
-------
REFERENCES
Bell, C.G. and W. Strauss. Effectiveness of Vertical Mist
Eliminators in a Cross Flow Scrubber. J, Air Pollution
Control Assoc. iZ.3:967-969, 1973.
Calvert, S. Venturi and Other Atomizing Scrubbers; Efficiency
and Pressure Drop. American Inst. of Chemical Engineers J.
16^:392-396, 1970.
Calvert, S. State-of-Art Survey of Mist Elimination in the USA.
In: Second US/USSR Symposium on Particulate Control. EPA
600/7-78-037, 1978. pp 180-193.
Calvert, S. Guidelines for Selecting Mist Eliminators. Chemical
Engineering, pp 109-112, Feb. 27, 1978.
Calvert, S., J. Goldshmid, D. Leith, and D. Mehta. Scrubber
Handbook. PB 213-016, 1972.
Calvert, S. , I.L. Jashnani, and S. Yung. Entrainment Separators
for Scrubbers. J. of Air Pollution Control Assoc. 24:971-975,
1974. ~
Calvert, S. , S. Yung, H.F. Barbarika, and L.E. Sparks. Entrain-
ment Separators for Scrubbers. In: Second EPA Fine Particle
Scrubbers Symposium. EPA 600/2-77-193. pp 75-95, 1977.
Calvert, S., S. Yung, and J. Leung. Entrainment Separators
for Scrubbers - Initial Report. EPA 650/2-74-119a. PB 241-189
1974. '
Calvert, S. , S. Yung, and J. Leung. Entrainment Separators
for Scrubbers - Final Report. EPA 650/2-74-119b. PB 248-050
1975. '
Calvert, S., S. Yung, and L.E. Sparks. Liquid Entrainment
from a Mobile Bed Scrubber. J. Air Pollution Control Assoc
2^:768-770, 1977.
Carter, W.J. A Literature Survey and Analysis of Mists 1-10 u
Size and Moisture Separators, MSAR 70-10, January 21 1970.
131
-------
Chan, J.K. and M.W. Golay. Comparative Evaluation of Cooling
Tower Drift Eliminator Performance. MIT-EL 77-004. PB 272-366,
June 1977.
Chan, J. and M.W. Golay. Comparative Performance of Current
Design Evaporative Cooling Tower Drift Eliminators. Atmos-
pheric Environment. 11^:775-781, 1977.
Chaput, L.S. Federal Standards of Performance for New Stationary
Sources of Air Pollution. J. of Air Pollution Control Assoc.
2^:1055-1060, 1976.
Conkle, H.N., H.S. Rosenberg, and S.T. Dinovo. Guidelines for
the Design of Mist Eliminators for Lime/Limestone Scrubbing
Systems. EPRI FP 327, December 1976.
Davies, R. A Review of the Methods for the Particle Size Analysis
of Droplets, Sprays, and Mists. Dechema Monogr. 79, n 1589-1615,
pt B, pp 115-134, 1976.
Dickinson, D.R. and W.R. Marshall, AIChE Journal. 14:541-552,
1968.
Epstein, M. EPA Alkali Scrubbing Test Facility: Summary of
Testing through October 1974. EPA 650/2-75-047. NTIS PB
244-901, 1975.
Epstein, M. H.N. Head, S.C. Wang, and D.A. Burbank. Results
of Mist Elimination and Alkali Utilization Testing at the
EPA Alkali Scrubbing Test Facility. In: Proceedings: Symposium
on Flue Gas Desulfurization New Orleans, March 1976, Volume I,
EPA 600/2-76-136a, 1976. pp 145-204.
Fuchs, N.A. The Mechanics of Aerosols. Pergamon Press, New
York, 1964.
Galeski, J.B. Particle Size Definition for Particulate Data
Analysis. EPA 600/7-77-129, 1977.
Gieseke, J.A. and R.I. Mitchell. Size Measurement of Collected
Drops. J. of Chemical and Engineering Data. l£:350-353, 1965.
Head, H.N. EPA Alkali Scrubbing Test Facility: Advanced Program,
Second Progress Report. EPA 600/7-76-008. NTIS PB 258-783, 1976.
Head, H.N. EPA Alkali Scrubbing Test Facility: Advanced Program,
Third Progress Report. EPA 600/7-77-105, 1977.
Hollinden, G.A., R.F. Robards, N.D. Moore, T.M. Kelso, and
R.M. Cole. TVA's 1-MW Pilot Plant: Vertical Duct Mist Elimi-
nation Testing - Progress Report, PRS-14. EPA 600/7-76-021,
October 1976.
132
-------
Hollinden, G., R. Robards, N. Moore, T. Kelso, and R. Cole.
TVA's 1-MW Pilot Plant: Final Report on High Velocity Scrub-
bing and Vertical Duct Mist Elimination, PRS-19. EPA 600/7-
77-019. NTIS PB 269-850, March 1977.
Houghton, J.G. and W.H. Radford. Trans. Am. Inst. of Chemical
Engineers. 3^:427, 1939.
Johnson, L.D. and R.M. Statnick. Measurement of Entrained Liquid
Levels in Effluent Gases from Scrubber Demisters. EPA 650/2-74-
050. NTIS PB 233-739, June 1974.
Medecki, H., M. Kaufman, and D.E. Magnus. Design, Development,
and Field Test of a Droplet Measuring Device. EPA 650/2-75-018,
1975.
Mcllvaine Company. The Mcllvaine Scrubber Manual. Northbrook,
Illinois, 1978.
Nannen, L.W. and K.E. Yeager. Status of the EPRI Flue Gas
Desulfurization Development Program. In: Proceedings: Symposium
on Flue Gas Desulfurization, New Orleans, March 1976, Vol. I.
EPA 600/2-76-136a, 1976. pp 102-113.
Roffman, A. and L.D. Van Vleck. The State-of-the-Art of Mea-
suring and Predicting Cooling Tower Drift and Its Deposition.
^4:855. J. Air Pollution Control Assoc. 1974.
Sekhar, N. Demister Design for Limestone Slurry Scrubber,
Paper No. 77-17.3. Presented at 70th Annual Meeting of APCA,
Toronto, Ontario, Canada, June 20-24, 1977.
Weir, A., L.T. Papay, D.G. Jones, J.M. Johnson, and W.C. Martin.
Results of the 170 MW Test Modules Program. Mohave Generating
Station, Southern California Edison Company. In: Proceedings:
Symposium on Flue Gas Desulfurization, New Orleans, March 1976
Volume I. EPA 600/2-76-136a, 1976. pp 325-353.
Williams, J.E. Mist Eliminator Testing at the Shawnee Prototype
Lime/Limestone Test Facility. In: Second US/USSR Symposium on
Particulate Control. EPA 600/7-78-037, 1978. pp 194-214.
Yao, S.C. and V.E. Schrock. Aerodynamic Design of Cooling Tower
Drift Eliminators, Transactions of the ASME: Journal of Engi-
neering for Power, 450-456. October 1976.
133
-------
APPENDIX "A"
ENTRAINMENT SEPARATOR INLET AND OUTLET
ENTRAINMENT CONCENTRATION AND SIZE DISTRIBUTION DATA
134
-------
TABLE A-l.
ENTRAINMENT SEPARATOR INLET DATA
MOBILE BED SCRUBBER OPERATION
Concentration (ml/DNm3) based on 4 collection methods
C/l
Run
No.
E.S.
Drainage
7.6 cm dia. 3.8 cm dia.
Sphere Sphere
Drainage Drainage
Sample
Cum.
Total
Train w/Cascade Impactor
Cone. Less Than (ymA)
14 8 5 1.5
Scrubber Mode 1
13-3
13-4
14-4
13-15
13-16
13-26
13-27
13-5
13-6
14-5
13-19
13-29
13-24
13-25
Average
13-11
13-12
Average
(1
(1
Cl
3
3
3
3
4,
4,
3.
4.
4.
3.
4.
4.
4.
.65)
.45)
.61)
.55
.61
.17
.36
.07
,39
,87
45
45
88
65
49
57
2
2
2
2
2
2
2
2
2
2
2
2
2
2
.37
.05
.29
.41
.41
.38
.28
.89
.23
.47
.34
.40
.22
.10
2.35
2.13
2.19
2.16
2
2
2
3
3
2
2
2
2
2
2
2
2
2
2
Scrubber
2
.69
.47
.82
.33
.30
.88
.91
.64
.69
.89
.61
.95
.69
.75
.83
Mode 2
.75
2.89
2.82
2
1
2
2
2
1
2
2
2
3
3
2
2
1
2
1
1
.08
.86
.83
.68
.51
.86
.40
.03
.08
.28
.36
.56
.98
.67
.44
.47
.60
1.54
.034
.049
.010 .006
.016
.018
.012 .008
.003 .002
.022
.049
.028 .012
.036
.010
.006 .004
.004 .004
.019
.035
.046
.040
.032
.029
.001
.012
.012
.001
.020
.040
.035
.010
.001
.001
.016
.028
.043
.036
.011
.011
.007
.007
.007
.015
.032
.005
.012
.014
.011
.012
-------
TABLE A-l. (continued)
Run
No.
13-13
13-14
Average
13-17
13-18
14-1
14-2
13-1
13-2
13-7
13-8
14-6
13-21
13-22
13-23
14-3
Average
E.S.
Drainage
4.
4.
4.
2.
3.
3.
2.
(5.
(4.
(4.
3.
3.
4.
4.
4.
4.
3.
13
13
13
75
48
04
93
29)
97)
67)
87
62
34
05
34
27
67
7.6 cm dia. 3.8 cm dia.
Sphere Sphere
Drainage Drainage
2.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
03
84
94
47
57
56
64
59
43
33
54
62
54
67
64
61
55
Scrubber
2
2
2
Scrubber
1
1
1
1
1
1
1
2
2
1
2
1
2
1
Mode 3
.34
.41
.38
Mode 4
.15
.98
.85
.93
.85
.46
.46
.24
.00
.54
.06
.84
.24
.82
Sample Train w/Cascade Impactor
Cum. Cone. Less Than (ymA)
Total 14 8 5 1.5
1
1
1
1
0
1
1
1
0
0
1
1
1
1
1
.955
.57
.26
.47
.15
.637
.31
.24
.27
.817
.720
.94
.20
-
.42
.06
.19
.024
.016
.020
.017
.016
.005 .003
.006 .004
.033
.011
.028
.031
.010 .005
.037
-
.008
.023 .012
.016
.023
.013
.018
.012
.011
.029
.010
.020
.025
.036
-
.007
.001
.017
.007
.012
.010
.007
.009
.020
.008
.010
.018
-
.012
-------
TABLE A-l. (continued)
Run
No.
E.S.
Drainage
7
. 6 cm dia.
Sphere
Drainage
3
.8 cm dia.
Sphere
Drainage
Sampl
Cum.
Total
e Train
Cone.
14
w/Cascade
Less Than
8
Impactor
(umA)
J J. •
5
13-9
13-10
14-7
Average
3.61
3.32
3.17
3.37
0.81
0.77
0.79
0.79
Scrubber Mode 5
0.81
0.77
0.79
0.79
0.99
1.05
1.01
1.02
0.344
1.36
0.852
0.852
.005
056
145
003
068
.046 .029
.132 .040
.089 .035
*Note: Figures in ( ) not included in averages.
-------
TABLE A-2. ENTRAINMENT SEPARATOR INLET DATA
SIEVE PLATE SCRUBBER OPERATION
Concentration (ml/DNm3) based on drainage from
entr. sep. and sample train w/cascade im-
pactor collection
Run E.S. Sample Train w/C.I.
No Drainage Cum. Cone. Less Than (ymA)
Total 20 11 7
Low Pressure Drop Operation
14-19 (-451) .081 .023 .018 .016
15-3 (-291)
14-17 .516 .092 .043 .041 .036
14-21 .773 .089 .028 .025 .020
14-10 - .126 .039 .030 .018
14-11 .387 .074 .027 .022
14-12 .451 .085 .014 .013 .009
14-13 .773 .092 .028 .027 .018
15-6 .507 .048 .016 .013
14-23 .773 .094 .033 .028 .020
14-24 .773 .120 .034 .029 .022
14-26 .677
14-27 .516 .087 .031 .025 .014
14-28 .644 .040 .030 .021 .017
14-29 .516 .056 .035 .027 .019
15-1 .581 .131 (.122) (.115) (.108)
15-2 .387 .068 .033 .022 .014
Average .591 .086 .031 .025 .018
138
-------
TABLE A-2. (continued)
Run
No
14-18
15-4
14-16
14-20
14-8
14-9
14-14
14-15
15-5
14-22
14-25
Average
E.S.
Drainage
Cum.
Total
Sample
Cone.
20
Train w/C.I.
Less Than (vimA)
10 6 1.8
High Pressure Drop Operation
(.483)
(.327)
.878
1.23
1.06
1.32
1.23
1.06
1.14
1.23
1.32
1.16
.208
.197
.133
.167
.108
.136
.101
.038
.106
.197
.139
.024
.016
.020
.013
.020
.026
.018
.032
.047
.024
.022
.015
.018
.009
.018
.024
.017
.016
.030
.041
.021
.018
.009
.011
.013
.013
.012
.014
.016
.018
.014
.001
.011
.006
Note: Figures in parentheses ( ) not used in average calculation,
139
-------
TABLE A-3. ENTRAINMENT SEPARATOR INLET DATA
SPRAY SCRUBBER OPERATION
Concentration (ml/DNm3) based on 4 collection methods
Run
No.
15-21
22
17
23
16A
16B
26
Average
15-19
20
18
24
12
13
14
15
25
Average
E.S.
Drainage
(5.04)
(4.44)
8.80
9.23
10.8
10.4
11.0
10.0
(4.45)
(4.14)
7.18
8.10
8.17
8.38
8.31
8.66
9.65
8.35
7.6 cm dia. 3.8 cm dia. Sample
Sphere Sphere Cum.
Drainage Drainage Total
5 ea.
-
-
6.22
6.24
6.19
6.24
6.32
6.24
7 ea.
-
-
5.52
5.13
4.09
4.97
5.31
5.55
4.58
5.02
1/2-B-50 Spray
-
-
8.77
8.14
9.03
9.03
8.61
8.72
1/2-B-50 Spray
-
-
7.59
6.91
5.62
7.23
7.23
7.33
6.17
6.87
No z z 1 e s
8.01
8.19
9.28
8.74
9.17
11.9
8.77
9.15
Nozzles
7.17
10.8
7.05
6.40
5.10
6.56
6.18
9.84
6.76
7.32
Train w/Cascade Impactor
Cone. Less Than (vunA)
17 8 5 1.5
.060
.058
.040
.026
.048
.028
.043
.053
.061
.028
.088
.047
.033
.052
.039
U71)
.036
.026
.018
.033
.020
.029
(.132) (.121)
.041 .017
.047
.020
.055 .044
(.134) (.112)
.075
.032
.022
.042 .031
.024
.020
.014
.010
.008
.012
.015
.015
.012
.012
.018
.012
.014
-------
TABLE A-3. (continued)
Run
No.
16-3
16-4
16-5
15-29
16-8
15-28
Average
16-1
16-2
16-6
15-30
16-7
15-27
Average
E.S.
Drainage
(2.99)
(3.47)
6.12
6.31
8.17
7.75
7.09
(3.18)
(2.69)
6.12
6.29
7.46
7.46
6.83
7.6 cm dia. 3.8 cm dia
Sphere Sphere
Drainage Drainage
4 ea.
-
2.51
2.75
2.74
2.36
2.59
6 ea.
-
-
1.74
2.20
2.13
2.33
2.10
Sample
Cum.
Total
Train w/Cascade Impactor
Cone. Less Than (ymA)
17 8 5 1.5
ST-24-FCN Spray Nozzles
3.12
-
2.70
4.14
3.77
3.45
3.52
ST-24-FCN Spray
-
-
3.66
3.56
3.03
3.25
3.38
2.03
3.21
3.64
3.64
3.66
3.22
Nozzles
4.04
4.82
2.83
3.65
2.57
4.62
3.76
.069
.022
.016
.024
.038
.034
.038
.032
.030
.021
.048
.034
.033 (.030)
.060
.017
.012
.018
.030
.028
.029
.031 .027
.028
.024
.015
.040
.028 .027
.043
.012
.009
.010
.017
.018
.019
.019
.015
.009
.034
.019
-------
TABLE A-4. ENTRAINMENT SEPARATOR OUTLET DATA
MOBILE BED SCRUBBER OPERATION
Concentration (ml/DNm3) based on sample train
w/cascade impactor collection
Run No.
13-3
13-4
14-4
13-15
13-16
13-17
13-18
13-26
13-27
14-1
14-2
13-5
13-6
14-5
13-11
13-12
13-13
13-14
13-1
13-2
13-7
13-8
14-6
13-9
13-10
14-7
Scrubber Cum. Cone. Less Than (ymA)
Mode Total 8 5 1.5
No Entrainment Separator
1
1
1
1
1
4
4
1
1
4
4
1
1
1
2
2
3
3
4
4
4
4
4
5
5
5
.607
.302
.333
3 Rows of Vertical
.031
.022
.017
.015
3 Rows of Slanted
.015
.011
.025
.016
6 Rows of Vertical
.018
.056
.032
.013
.012
.016
.016
.010
.011
.042
.061
.031
.085
.036
.019
(Blank
.021
.021
.021
Baffles
.021
.016
.012
.014
Baffles
.012
.008
.019
.011
Baffles
.014
.042
.018
.010
.010
.014
.015
.009
.010
.032
.055
.022
.068
.032
.015
Run)
.013
.020
.019
.019
.014
.012
.014
.012
.008
.017
.010
.013
.036
.016
.009
.010
.014
.010
.007
.009
.028
.048
.020
.059
.027
.013
.008
.016
.013
.015
.013
.010
.012
.011
.007
.013
.008
.010
.028
.012
.006
.008
.012
.010
.007
.006
.021
.038
.015
.040
.023
.010
142
-------
TABLE A-4. (continued)
Scrubber Cum. Cone. Less Than (ymA)
Run No. Mode Total 8 5 1.5
6 Rows of Slanted Baffles
13-19 1 .012 .009 .008 .007
13-20 1 .004 .004 .004 .004
13-24 1 .012 .012 .010 .008
13-25 1 .016 .016 .010 .009
13-21 4 .034 .030 .029 .025
13-22 4 .010 .010 .010 .010
13-23 4 .014 .012 .012 .011
14-3 4 .034 .028 .024 .018
143
-------
TABLE A-5. ENTRAINMENT SEPARATOR OUTLET DATA
SIEVE PLATE SCRUBBER OPERATION
Concentration (ml/DNm3) based on sample train with cascade
impactor collection
Run No,
14-19
14-18
14-17
14-16
14-21
14-20
14-10
14-11
14-12
14-13
15-6
Average
14-8
14-9
14-14
14-15
15-5
Average
14-23
14-24
Average
14-22
14-25
Average
Scrubber
AP
Cum Cone. Less Than
Total 8 5
(yrnA)
No Entrainment Separator (Blank Run)
Low .051 .017 .016
High .102 .017 .015
3 Vertical Rows of Baffles
Low .037 .032 .028
High .019 .012 .011
3 Slanted Rows of Baffles
Low .019 .013 .011
High .022 .014 .013
6 Vertical Rows of Baffles
Low .039 .036 .026
Low .012 .009 .009
Low .010 .008 .007
Low .022 .016 .015
Low .015 .013 .012
Low .020 .016 .014
High .078 .073 .019
High .013 .011 .009
High .010 .009 .008
High .015 .012 .012
High .022 .020 .019
High .028 .025 .013
6 Slanted Rows of Baffles
Low .027 .024 .021
Low .031 .027 .026
Low .029 .026 .024
High .019 .016 .015
High .038 .030 .023
High .029 .023 .019
1.6
.012
.009
.025
.006
.007
.008
.021
.006
.006
.015
.009
.011
.014
.005
.005
.008
.016
.010
.014
.020
.017
.008
.013
.011
144
-------
TABLE A-6. ENTRAINMENT SEPARATOR OUTLET DATA
SPRAY SCRUBBER OPERATION
Concentration (ml/DNm3) based on sample train w/cascade impactor
Run
No
Spray
Config.1
Cum. Cone.
Total 17
Less Than
8
(ymA)
5
1.5
No Entrainment Separator (Blank Run)
15-21
15-22
15-19
15-20
16-3
16-4
16-1
16-2
15-17
15-18
16-5
16-6
15-23
15-24
15-29
15-30
15-16A
15-16B
15-12
15-13
15-14
15-15
16-8
16-7
5
5
7
7
4
4
6
6
5
7
4
6
5
7
4
6
5
5
7
7
7
7
4
6
2.73
3.88 .165
4.43 .061
1.95
.590 .024
1.30
.914
.467 .023
3 Vertical Rows
.355
.128
.059
.109
3 Slanted Rows
.198
.069
.036
.075
6 Vertical Rows
.043
.046
.042
.107
.041
.064
.051
.030
.057
.132
.042
.061
.017
.033
.030
.019
.299
.061
.019
.018
.095
.031
.019
.039
.019
.031
.027
.073
.026
.039
.040
.016
041
048
.027
.026
.288
.051
.016
.015
.084
.026
.016
.034
.015
.027
.021
.065
.023
.032
.035
.015
.021
.079
.022
.013
.015
.269
.037
.012
.012
.071
.020
.013
.028
.009
.017
.011
.046
.014
.022
.031
.012
145
-------
TABLE A-6. (continued)
Run
No
Spray
Config.1
15-26
15-25
15-28
15-27
5
7
4
6
Cum. Cone.
Total 17
Less Than (ymA)
8 5 1.5
1 Spray Configurations:
6 Slanted Rows
5
7
4
6
.051
.053
.045
.045
- 5 ea.
- 7 ea.
- 4 ea.
- 6 ea.
.033 .028
.036 .031
.034 .031
.031 .027
1/2-B-50 spray nozzles
1/2-B-50 spray nozzles
ST-24-FCN spray nozzles
ST-24-FCN spray nozzles
.022
.024
.024
.022
146
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/7-79-180
2.
RECIPIENT'S ACCESSION- NO.
TITLE AND SUBTITLE
Development of Superior Entrainment Separators
, BEPOHT DATE
August 1979
PERFORMING ORGANIZATION CODE
AUTHOR(S)
eymour Calvert and Harry F. Barbarika
B. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
EHE624
. PERFORMING ORGANIZATION NAME AND ADDRESS
Air Pollution Technology, Inc.
4901 Morena Boulevard, Suite 402
San Diego, California 92117
11. CONTRACT/GRANT NO.
68-02-2184
2. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PEI
Final; 10/76 - 3/79
PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/13
5. SUPPLEMENTARY NOTES IERL-RTP project officer is Leslie E. Sparks, Mail Drop 61,
919/541-2925.
ABSTRACT The report describes an experimental and theoretical program carried out
to develop an improved design for entrainment separators for scrubbers. The pro-
blems of separation efficiency, suspended solids deposition, and plugging of the en-
trainment separator (ES) were of primary concern. A pilot scale ES, coupled to a
scrubber and designed to handle a nominal gas flow rate of 1.4 cu m/s (3000 acfm),
was designed, built, and tested. Vertical discontinuous zigzag baffles was the ES
design selected after a review of both theory and practical experience with slurry
scrubbers. The effect of ES performance on particulate emissions of a typical fossil-
fueled boiler was evaluated. The experimental program included measurements of
entrainment size distribution and loading, entrainment collection efficiency, solids
deposition character and rate, and ES washing efficiency. Results were compared
with available models and new criteria for effective washing were developed.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
Pollution Control
Stationary Sources
Entrainment Separators
Mist Eliminators
Particulate
c. COSATI Field/Group
Pollution
Separators
Washing
Scrubbers
Dust
Flue Gases
Fossil Fuels
I3F
131,07A
13H
11G
2 IB
21D
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (THis Report)
Unclassified
21. NO. OF PAGES
163
20. SECURITY CLASS (This page I
Unclassified
22. PRICE
EPA Form 2220-1 (»-7J)
147
------- |