&EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 2771 1
EPA-600 9 82 005d
July 1982
Research and Development
Third Symposium on the
Transfer and
Utilization of Particulate
Control Technology:
Volume IV. Atypica
Applications
-------
EPA-600/9-82-005d
THIRD SYMPOSIUM ON THE
TRANSFER AND UTILIZATION OF
PARTICULATE CONTROL TECHNOLOGY
VOLUME IV. ATYPICAL APPLICATIONS
Compiled by:
F.P. Venditti, J.A. Armstrong, and M. Durham
Denver Research Institute
P.O. Box 10127
Denver, Colorado 80208
Grant Number: R805725
Project Officer
Dale L. Harmon
Office of Environmental Engineering and Technology
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
Prepared for:
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental
Research Laboratory-Research Triangle Park, North Carolina, Office of
Research and Development, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
ii
-------
ABSTRACT
The papers in these four volumes of Proceedings were presented at the
Third Symposium on the Transfer and Utilization of Particulate Control
Technology held in Orlando, Florida during 9 March through 13 March 1981,
sponsored by the Particulate Technology Branch of the Industrial Environ-
mental Research Laboratory of the Environmental Protection Agency and
coordinated by the Denver Research Institute of the University of Denver.
The purpose of the symposium was to bring together researchers,
manufacturers, users, government agencies, educators and students to
discuss new technology and to provide an effective means for the transfer
of this technology out of the laboratories and into the hands of the users.
The three major categories of control technologies — electrostatic
precipitators, scrubbers, and fabric filters — were the major concern of the
symposium. These technologies were discussed from the perspectives of
economics; new technical advancements in science and engineering; and
applications. Several papers dealt with combinations of devices and
technologies, leading to a concept of using a systems approach to partic-
ulate control rather than device control. Additional topic areas included
novel control devices, high temperature/high pressure applications, fugitive
emissions, and measurement techniques.
These proceedings are divided into four volumes, each volume contain-
ing a set of related session topics to provide easy access to a unified
technology area.
ill
-------
VOLUME IV
Page
VOLUME I. COAL FIRED BOILERS—CONTENTS x
VOLUME II. ELECTROSTATIC PRECIPITATORS—CONTENTS ... xiv
VOLUME III. PARTICULATE CONTROL DEVICES—CONTENTS . . . xviii
Section A - Advanced Energy Applications
HIGH TEMPERATURE PARTICLE COLLECTION WITH
A.P.T. EPxP DRY SCRUBBER I
S. Yung, T. Lee, R.C. Patterson, S. Calvert and D.C. Drehmel
PARTICLE COLLECTION IN CYCLONES AT HIGH TEMPERATURE
AND HIGH PRESSURE 2
R. Parker, R. Jain, S. Calvert, D.C. Drehmel and J. Abbott
OPERATING RESULTS OF ELECTROSTATIC PRECIPITATORS
AT HIGH TEMPERATURE AND HIGH PRESSURES 3
P.L. Feldman and K.S. Kumar
CONTROL OF PARTICULATES IN PROCESS AREA 12, SOLVENT
REFINED COAL PROCESS 15
W.H. Wilks, P.O. Wilkinson and J.A. Schlosberg
NON-PLUGGING RETAINING STRUCTURE FOR GRANULAR
BED FILTER FOR HTHP APPLICATION 26
A.M. Presser and J.C. Alexander
PARTICULATE EMISSIONS CONTROL FROM A COAL-FIRED
OPEN-CYCLE MAGNETOHYDRODYNAMICS/STEAM POWER PLANT ... 36
H.H. Wang and T.E. Dowdy
REAL TIME COARSE PARTICLE MASS MEASUREMENTS IN
A HIGH TEMPERATURE AND PRESSURE COAL GASIFIER
PROCESS TREATMENT 46
J. Wegrzyn, J. Saunders and W. Marlow
THE DESIGN, ENGINEERING, AND STARTUP OF A VENTURI
SCRUBBER SYSTEM ON AN OIL SHALE OFF-GAS INCINERATOR ... 55
P.A. Czuchra and J.S. Sterrett
FLUIDIZED-BED COMBUSTION HOT FLUE GAS CLEANUP
PERSPECTIVE ON CYCLONES AND OTHER DEVICES 63
R.F. Henry and W.F. Podolski
V
-------
VOLUME IV CONTENTS (cont.)
Page
PRESSURIZED AND NON-PRESSURIZED ACOUSTIC
AGGLOMERATORS FOR HOT-GAS CLEANUP APPLICATIONS 73
K.H. Chou and D.T. Shaw
ALKALIS AND THEIR CONTRIBUTIONS TO CORONA CURRENT
AT HIGH TEMPERATURE AND HIGH PRESSURE 74
R.W.L. Snaddon
HOT GAS CLEANUP IN PRESSURIZED FLUIDIZED
BED COMBUSTION 83
L.N. Rubow and M.G. Klett
VENTURI SCRUBBING FOR CONTROL OF PARTICULATE
EMISSIONS FROM OIL SHALE RETORTING 95
G.M. Rinaldi and R.C. Thurnau
OVERVIEW OF THE DEPARTMENT OF ENERGY'S PRESSURIZED
FLUIDIZED-BED COMBUSTOR CLEANUP TECHNOLOGY PROGRAM . . .105
W.E. Moore
THE CYCLOCENTRIFUGE™—AN ADVANCED GAS/SOLIDS
SEPARATOR FOR COAL CONVERSION PROCESSES 116
P.R. Albrecht, J.T. McCabe and W. Fedarko
Section B - Fugitive Emissions
DEMONSTRATION OF THE USE OF CHARGED FOG IN
CONTROLLING FUGITIVE DUST FROM LARGE-SCALE
INDUSTRIAL SOURCES 125
E.T. Brookman, R.C. McCrillis and D.C. Drehmel
THE CONTROL OF FUGITIVE EMISSIONS USING WINDSCREENS . . .135
D. Carnes and D.C. Drehmel
THE INFLUENCE OF AGGREGATE PILE SHAPE AND
ORIENTATION ON PARTICULATE FUGITIVE EMISSIONS 145
D. Martin
SPRAY CHARGING AND TRAPPING SCRUBBER FOR
FUGITIVE PARTICLE EMISSION CONTROL 155
S. Yung, S. Calvert and D.C. Drehmel
IMPROVED STREET SWEEPER FOR CONTROLLING URBAN
INHALABLE PARTICULATE MATTER 156
S. Calvert, H. Brattin, S. Bhutra, R. Parker and D.C. Drehmel
vi
-------
VOLUME IV CONTENTS (cont.)
Page
A WIND TUNNEL FOR DUST ENTRAINMENT STUDIES 168
A.S. Viner, M.B. Ranade, E.J. Shaughnessy, D.C. Drehmel
and B.E. Daniels
TECHNIQUES AND EQUIPMENT FOR MEASURING INHALABLE
PARTICULATE FUGITIVE EMISSIONS 179
H.J. Kolnsberg
BALLOON SAMPLING TO CHARACTERIZE PARTICLE
EMISSIONS FROM FUGITIVE SOURCES 188
J.A. Armstrong and D.C. Drehmel
AN ELECTROSTATICALLY CHARGED FOG GENERATOR FOR
THE CONTROL OF INHALABLE PARTICLES 200
C.V. Mathai, L.A. Rathbun and D.C. Drehmel
RELATIVE EFFECTIVENESS OF CHEMICAL ADDITIVES
AND WIND SCREENS FOR FUGITIVE DUST CONTROL 210
D.C. Drehmel and B.E. Daniel
PARTICULATE IMPACT COMPARISON BETWEEN CONTROLLED
STACK EMISSIONS FOR A 2000 MW ELECTRICAL GENERATING
STATION 222
H.E. Hesketh and F.L. Cross
OPERATING EXPERIENCE AND THE TECHNIQUES IN THE
CONTROL OF COAL DUST EMISSIONS FROM LARGE
STORAGE PILE AT NANTIC9KE TGS 232
N. Krishnamurthy, W. Whitman and Y.V. Nguyen
Section C - Opacity
MODELING SMOKE PLUME OPACITY FROM PARTICULATE
CONTROL EQUIPMENT 242
D.S. Ensor, P.A. Lawless, S.J. Cowen
TETHERED BALLOON PLUME SAMPLING OF A PORTLAND
CEMENT PLANT 252
J.A. Armstrong, P.A. Russell, M.N. Plooster
THE RELATIONSHIP OF FLY ASH LIGHT ABSORPTION TO
SMOKE PLUME OPACITY 264
S.J. Cowen, D.S. Ensor
vii
-------
VOLUME IV CONTENTS (cont.)
Section D - Measurements
A SPECIAL METHOD FOR THE ANALYSIS OF
SULFURIC ACID MISTS 275
P. Urone, R.B. Mitchell, J.E. Rusnak, R.A. Lucas and
J.F. Griffiths
A MICROCOMPUTER-BASED CASCADE-IMPACTOR
DATA-REDUCTION SYSTEM 285
M. Durham, S. Tegtmeyer, K. Wasmundt and L.E. Sparks
DEVELOPMENT OF A SAMPLING TRAIN FOR STACK
MEASUREMENT OF INHALABLE PARTICULATE 297
A.D. Williamson, W.B. Smith
INHALABLE PARTICULATE MATTER SAMPLING
PROGRAM FOR IRON AND STEEL: AN OVERVIEW
PROGRESS REPORT 306
R.C. McCrillis
DEVELOPMENT OF IP EMISSION FACTORS 317
D.L. Harmon
INHALABLE PARTICULATE EMISSION FACTOR PROGRAM
PURPOSE AND DEVELOPMENT 326
F.M. Noonan and J.H. Southerland
INHALABLE PARTICULATE EMISSION FACTORS FOR BLAST
FURNACE CASTHOUSES IN THE IRON AND STEEL INDUSTRY . . . .335
P.D. Spawn, S. Piper and S. Gronberg
INHALABLE PARTICULATE EMISSIONS FROM VEHICLES
TRAVELING ON PAVED ROADS 344
R. Bohn
QUALITY ASSURANCE FOR PARTICLE-SIZING MEASUREMENTS . . .353
C.E. Tatsch
PARTICULATE EMISSIONS CHARACTERIZATION FOR
OIL-FIRED BOILERS 363
D. Mormile, S. Hersh, B.F. Piper and M. McElroy
A CONTINUOUS REAL-TIME PARTICULATE MASS MONITOR
FOR STACK EMISSION APPLICATIONS 373
J.C.F. Wang, H. Patashnick and G. Rupprecht
viii
-------
VOLUME IV CONTENTS (cont.)
Page
Section E - Mobile Sources
STUDIES OF PARTICULATE REMOVAL FROM DIESEL EXHAUSTS
WITH ELECTROSTATIC AND ELECTROSTATICALLY-
AUGMENTED TECHNIQUES 383
J.L. DuBard, M.G. Faulkner, J.R. McDonald, D.C. Drehmel
and J.H. Abbott
STUDIES OF PARTICULATE REMOVAL FROM DIESEL EXHAUSTS
WITH MECHANICAL TECHNIQUES 395
M.G. Faulkner, J.L. DuBard, J.R. McDonald, D.C. Drehmel
and J.H. Abbott
UPDATE ON STATUS OF CONNECTICUT'S CONTROL PROGRAM
FOR TRANSPORTATION-RELATED PARTICULATE EMISSIONS . . . .406
H.L. Chamberlain and J.H. Gastler
AUTHOR INDEX 413
ix
-------
VOLUME I
COAL FIRED BOILERS
Section A - Fabric Filters
COAL PROPERTIES AND FLY ASH FILTERABILITY 1
R. Dennis, J.A. Dirgo and L.S. Hovis
PULSE-JET FILTRATION WITH ELECTRICALLY
CHARGED FLYASH 11
R.P. Donovan, L.S. Hovis, G.H. Ramsey and J.H. Abbott
ELECTRICALLY CHARGED FLYASH EXPERIMENTS IN A
LABORATORY SHAKER BAGHOUSE 23
L.S. Hovis, J.H. Abbott, R.P. Donovan and C.A. Pareja
ELECTROSTATIC AUGMENTATION OF FABRIC FILTRATION .... 35
D.W. VanOsdell, G.P. Greiner, G.E.R. Lamb and L.S. Hovis
FABRIC WEAR STUDIES AT HARRINGTON STATION 45
R. Chambers, K. Ladd, S. Kunka and D. Harmon
SPS PILOT BAGHOUSE OPERATION 55
K. Ladd, W. Hooks, S. Kunka and D. Harmon
REVIEW OF SPS INVESTIGATION OF HARRINGTON STATION
UNIT 2 FABRIC FILTER SYSTEM 65
K. Ladd, S. Kunka
A SUMMARY OF PERFORMANCE TESTING OF THE APITRON
ELECTROSTATICALLY AUGMENTED FABRIC FILTER 75
D. Helfritch and L. Kirsten
FABRIC FILTER OPERATING EXPERIENCE FROM SEVERAL
MAJOR UTILITY UNITS 82
O.F. Fortune, R.L. Miller and E.A. Samuel
EVALUATION OF THE 25 MW KRAMER STATION BAGHOUSE:
TRACE ELEMENT EMISSION CONTROL 94
M.W. McElroy and R.C. Carr
CHARACTERIZATION OF A 10 MW FABRIC FILTER
PILOT PLANT 96
W.B. Smith, K.M. Gushing and R.C. Carr
SPECIFYING A FABRIC FILTER SYSTEM 107
R.L. Ostop and D.A. Single
-------
VOLUME I CONTENTS (cont.)
Page
EVALUATION OF THE 25 MW KRAMER STATION BAGHOUSE:
OPERATIONAL FACTORS IN PARTICULATE MATTER
EMISSION CONTROL 118
R.C. Carr and M.W. McElroy
PULSE-JET TYPE FABRIC FILTER EXPERIENCE AT AIR TO
CLOTH RATIOS OF 5 TO 1 ON A BOILER FIRING PULVERIZED
COAL 120
G.L. Pearson
SELECTION AND OPERATION OF BAGHOUSES AT R.D. NIXON
STATION, UNIT #1 129
R.C. Hyde, J. Arello and D.J. Huber
POTENTIAL FOR IMPROVEMENT IN BAGHOUSE DESIGN 138
R.M. Jensen
REVIEW OF OPERATING AND MAINTENANCE EXPERIENCES WITH
HIGH TEMPERATURE FILTER MEDIA ON COAL-FIRED BOILERS . . .148
L.K. Crippen
Section B - Electrostatic Precipitators
PILOT DEMONSTRATION OF THE PRECHARGER-COLLECTOR
SYSTEM 157
P. Vann Bush, Duane H. Pontius
REMEDIAL TREATMENTS FOR DETERIORATED HOT SIDE
PRECIPITATOR PERFORMANCE 165
R.E. Bickelhaupt
EVALUATION OF THE UNITED McGILL ELECTROSTATIC
PRECIPITATOR 176
D.S. Ensor, P.A. Lawless, A.S. Damle
PREDICTING THE EFFECT OF PROPRIETARY CONDITIONING
AGENTS ON FLY ASH RESISTIVITY 185
R.J. Jaworowski and J.J. Lavin
SO3 CONDITIONING TO ENABLE ELECTROSTATIC
PRECIPITATORS TO MEET DESIGN EFFICIENCIES 197
J.J. Ferrigan, III
ENHANCED PRECIPITATOR COLLECTION EFFICIENCIES
THROUGH RESISTIVITY MODIFICATION 206
D.F. Mahoney
xi
-------
VOLUME I CONTENTS (cont.)
Page
DEVELOPMENT OF A NEW SULFUR TYPE ASH CONDITIONING . . . .216
R.H. Gaunt
OPERATING EXPERIENCE WITH FLUE GAS CONDITIONING
SYSTEMS AT COMMONWEALTH EDISON COMPANY 226
L.L. Weyers and R.E. Cook
THE APPLICATION OF A TUBULAR WET ELECTROSTATIC
PRECIPITATOR FOR FINE PARTICULATE CONTROL AND
DEMISTING IN AN INTEGRATED FLY ASH AND SO, REMOVAL
SYSTEM ON COAL-FIRED BOILERS 236
E. Bakke and H.P. Willett
FIELD EVALUATIONS OF AMMONIUM SULFATE CONDITIONING
FOR IMPROVEMENT OF COLD SIDE ELECTROSTATIC PRECIPITATOR
PERFORMANCE 237
E.G. Landham, Jr., G.H. Marchant, Jr., J.P. Gooch and
R.F. Altman
EVALUATION OF PERFORMANCE ENHANCEMENT OBTAINED
WITH PULSE ENERGIZATION SYSTEMS ON A HOT SIDE
ELECTROSTATIC PRECIPITATOR 253
W. Piulle, L.E. Sparks, G.H. Marchant, Jr. and J.P. Gooch
A NEW MICROCOMPUTER AND STRATEGY FOR THE CONTROL
OF ELECTROSTATIC PRECIPITATORS 265
K.J. McLean, T.S. Ng, Z. Herceg and Z. Rana
ASSESSMENT OF THE COMMERCIAL POTENTIAL FOR THE HIGH
INTENSITY IONIZER IN THE ELECTRIC UTILITY INDUSTRY . . . .272
J.S. Lagarias, J.R. McDonald and D.V. Giovanni
APPLICATION OF ENERGY CONSERVING PULSE ENERGIZATION
FOR PRECIPITATORS—PRACTICAL AND ECONOMIC ASPECTS . . . .291
H. H. Petersen and P. Lausen
Section C - Dry SO9 Scrubbers
SO, REMOVAL BY DRY INJECTION AND SPRAY ABSORPTION
TECHNIQUES 303
E.L. Parsons, Jr., V. Boscak, T.G. Brna and R.L. Ostop
DRY SCRUBBING SO2 AND PARTICULATE CONTROL 313
N.J. Stevens, G.B. Manavizadeh, G.W. Taylor and M.J. Widico
FIBER AND FABRIC ASPECTS FOR SO2 DRY SCRUBBING
BAGHOUSE SYSTEMS 323
L. Bergmann
xii
-------
VOLUME I CONTENTS (cont.)
Page
TWO-STAGE DRY FLUE GAS CLEANING USING CALCIUM
ALKALIS 333
B.C. Gehri, D.F. Dustin and SJ. Stachura
CONTROL OF SULFUR DIOXIDE, CHLORINE, AND TRACE
ELEMENT EMISSIONS FROM COAL-FIRED BOILERS BY FABRIC
FILTRATION 341
RJ. Demski, J.T. Yeh and J.I. Joubert
Section D - Scrubbers
FLYASH COLLECTION USING A VENTURI SCRUBBER—MINNESOTA
POWER'S COMMERCIAL OPERATING EXPERIENCE 352
C.A. Johnson
AUTHOR INDEX 361
xiii
-------
VOLUME II
ELECTROSTATIC PRECIPITATORS
Section A - Fundamentals
Paqe
MATHEMATICAL MODELING OF IONIC
CONDUCTION IN FLY ASH LAYERS 1
R.B. Mosley, J.R. McDonald and L.E. Sparks
MEASUREMENTS OF ELECTRICAL PROPERTIES
OF FLY ASH LAYERS 13
R.B. Mosley, P.R. Cavanaugh, J.R. McDonald and L.E. Sparks
LASER DOPPLER ANEMOMETER MEASUREMENTS OF PARTICLE
VELOCITY IN A LABORATORY PRECIPITATOR 25
P.A. Lawless, A.S. Damle, A.S. Viner, E.J. Shaughnessy and
L.E. Sparks
PROGRESS IN MODELING BACK CORONA 35
P.A. Lawless
A COMPUTER MODEL FOR ESP PERFORMANCE 44
P.A. Lawless, J.W. Dunn and L.E. Sparks
MEASUREMENT AND INTERPRETATION OF CURRENT
DENSITY DISTRIBUTION AND CHARGE/MASS DATA 54
M. Durham, G. Rinard, D. Rugg and L.E. Sparks
THE RELATIONSHIP BETWEEN GAS STREAM TURBULENCE
AND COLLECTION EFFICIENCY IN A LAB-SCALED
ELECTROSTATIC PRECIPITATOR 66
B.E. Pyle, J.R. McDonald, W.B. Smith
PARTICLE DEPOSITION PROFILES AND REENTRAINMENT
IN A WIRE-PLATE ELECTROSTATIC PRECIPITATOR 76
E. Arce-Medina and R.M. Felder
PARTICLE TRANSPORT IN THE EHD FIELD 87
T. Yamamoto
SURFACE REENTRAINMENT OF COLLECTED FLY ASH IN
ELECTROSTATIC PRECIPITATORS 97
M. Mitchner, M.J. Fisher, D.S. Gere, R.N. Leach and S.A. Self
ELECTROMECHANICS OF PRECIPITATED ASH LAYERS 109
G.B. Moslehi and S.A. Self
xiv
-------
VOLUME II CONTENTS (cont.)
Page
EXPERIMENTAL MEASUREMENTS OF THE EFFECT OF
TURBULENT DIFFUSION ON PRECIPITATOR EFFICIENCY 120
G.L. Leonard, M. Mitchner and S.A. Self
CAN REENTRAINMENT BE EXPLAINED USING A NEW
PRECIPITATOR FORMULA? 130
S. Maartmann
A LABORATORY FURNACE FOR THE PRODUCTION OF
SYNTHETIC FLY ASH FROM SMALL COAL SAMPLES 141
K.M. Sullivan
COMPUTER SIMULATION OF THE WIDE PLATE
SPACING EFFECT 149
E.A. Samuel
SIMULTANEOUS MEASUREMENTS OF AERODYNAMIC SIZE
AND ELECTRIC CHARGE OF AEROSOL PARTICLES IN REAL
TIME ON A SINGLE PARTICLE BASIS 160
M.K. Mazumder, R.G. Renninger, T.H. Chang,
R.W. Raible, W.G. Hood, R.E. Ware and R.A. Sims
APPLICATION OF LASER DOPPLER INSTRUMENTATION TO
PARTICLE TRANSPORT MEASUREMENTS IN AN ELECTROSTATIC
PRECIPITATOR 169
M.K. Mazumder, W.T. Clark III, R.E. Ware, P.C. McLeod,
W.G. Hood, J.E. Straub and S. Wanchoo
THE APPLICATION OF MEASUREMENTS OF AEROSOL
CHARGE ACQUISITION BY BIPOLAR IONS TO THE PROBLEM
OF BACK CORONA 179
R.A. Fjeld, R.O. Gauntt, G.J. Laughlin and A.R. McFarland
IDENTIFICATION OF BACK DISCHARGE SEVERITY 189
S. Masuda and Y. Nonogaki
Section B - Operations and Maintenance
MODELING OF ELECTROSTATIC PRECIPITATORS WITH RESPECT
TO RAPPING REENTRAINMENT AND OUTLET OPACITY 199
M.G. Faulkner, W.E. Farthing, J.R. McDonald and L.E. Sparks
NEW PRECIPITATOR TECHNOLOGY FOR PARTICULATE
CONTROL 208
J.R. Zarfoss
XV
-------
VOLUME II CONTENTS (cont.)
Page
AN APPLICATION SUMMARY OF HIGH ENERGY SONIC
CLEANING APPLIED TO ELECTROSTATIC PRECIPITATORS 218
M.J. Berlant
THE IMPACT OF INTELLIGENT PRECIPITATOR CONTROLS 230
N.Z. Shilling, R.O. Reese and J.A. Fackler
AN ENERGY MANAGEMENT SYSTEM FOR
ELECTROSTATIC PRECIPITATORS 242
R.R. Crynack and M.P. Downey
RELATIONSHIP BETWEEN ELECTROSTATIC PRECIPITATOR
PERFORMANCE AND RECORDKEEPING PRACTICES 252
S.P. Schliesser
AN OPERATION AND MAINTENANCE PROGRAM FOR
A PHOSPHATE ROCK ELECTROSTATIC PRECIPITATOR 262
D.B. Rimberg
Section C - Advanced Design
ELECTROSTATIC PRECIPITATOR PERFORMANCE
WITH PULSE EXCITATION 273
D. Rugg, M. Durham, G. Rinard and L.E. Sparks
DEVELOPMENT OF A CHARGING DEVICE FOR HIGH-RESISTIVITY
DUST USING HEATED AND COOLED ELECTRODES 283
G. Rinard, M. Durham, D. Rugg and L.E. Sparks
THE EVALUATION OF NOVEL ELECTROSTATIC PRECIPITATOR
SYSTEMS USING A TRANSPORTABLE PROTOTYPE 295
G. Rinard, M. Durham, D. Rugg, J. Armstrong,
L.E. Sparks and J.H. Abbott
ANALYSIS OF THE ELECTRICAL AND CHARGING
CHARACTERISTICS OF A THREE ELECTRODE PRECHARGER . . . .304
K.J. McLean
PARTICLE CHARGING IN AN ELECTROSTATIC
PRECIPITATOR BY PULSE AND DC VOLTAGES 314
L.E. Sparks, G.H. Ramsey, R.E. Valentine and J.H. Abbott
PARTICLE COLLECTION IN A TWO STAGE ELECTROSTATIC
PRECIPITATOR WITH VARIOUS COLLECTOR STAGES 326
L.E. Sparks, G.H. Ramsey, R.E. Valentine and J.H. Abbott
HIGH INTENSITY IONIZER DEVELOPMENT 334
M.H. Anderson, J.R. McDonald, J.P. Gooch and D.V. Giovanni
xvi
-------
VOLUME III
PARTICULATE CONTROL DEVICES
Section A - Scrubbers
Page
THE CALVERT SCRUBBER 1
S. Calvert, R.G. Patterson and S. Yung
FLUX FORCE/CONDENSATION SCRUBBER SYSTEM
FOR COLLECTION OF FINE PARTICULATE EMISSIONS
FROM AN IRON MELTING CUPOLA 10
S. Calvert and D.L. Harmon
DEMONSTRATION OF HIGH-INTENSITY-IONIZER-ENHANCED
VENTURI SCRUBBER ON A MAGNESIUM RECOVERY
FURNACE FUME EMISSIONS 21
A. Prem, M.T. Kearns and D.L. Harmon
A NEW ENTRY IN THE HIGH EFFICIENCY SCRUBBER FIELD .... 33
L.C. Hardison and F. Ekman
PERFORMANCE OF PARTICULATE SCRUBBERS AS
INFLUENCED BY GAS-LIQUID CONTACTOR DESIGN
AND BY DUST FLOCCULATION 43
K.T. Semrau and R.J. Lunn
INVESTIGATION OF VENTURI SCRUBBER EFFICIENCY
AND PRESSURE DROP 51
R. Parker, T. Le and S. Calvert
SCRUBBER TECHNOLOGY AND THE INTERACTION OF
A UNIQUE STRUCTURE AS MIST ELIMINATOR 60
G.C. Pedersen
NOVEL ANNULAR VENTURI SCRUBBER DESIGN REDUCES
WASTE DISCHARGE PROBLEMS 71
H.P. Beutner
CONSIDERATION OF THE PERTINENT DESIGN AND
OPERATING CHARACTERISTICS ESSENTIAL FOR
OPTIMIZATION OF VENTURI SCRUBBER PERFORMANCE 80
H.S. Oglesby
APPLICATION OF SCRUBBERS FOR PARTICULATE
CONTROL OF INDUSTRIAL BOILERS 90
M. Borenstein
xviii
-------
VOLUME II CONTENTS (cont.)
Page
DEMONSTRATION OF AIR POLLUTION SYSTEMS HIGH
INTENSITY IONIZER/ELECTROSTATIC PRECIPITATOR ON .
AN OIL-FIRED BOILER 349
G.A. Raemhild, A. Prem and F. Weisz
PRIMARY AND SECONDARY IONIZATION IN AN
ELECTRON BEAM PRECIPITATOR SYSTEM 358
W.C. Finney, L.C. Thanh, J.S. Clements and R.H. Davis
INFLUENCE ON PARTICLE CHARGING OF ELECTRICAL
PARAMETERS AT DC AND PULSE VOLTAGES 370
H.J. Joergensen, J.T. Kristiansen and P. Lausen
BOXER-CHARGER MARK III AND ITS
APPLICATION IN ESP'S 380
S. Masuda, H. Nakatani and A. Mizuno
THE PERFORMANCE OF AN EXPERIMENTAL
PRECIPITATOR WITH AN ALL-PLATE ZONE 390
J. Dalmon
THE PHYSICS OF PULSE ENERGIZATION OF
ELECTROSTATIC PRECIPITATORS 404
L. Menegozzi and P.L. Feldman
ADVANCED ELECTRODE DESIGN FOR
ELECTROSTATIC PRECIPITATORS 405
S. Bernstein, K. Ushimaru and E.W. Geller
Section D - Industrial Applications
PROBLEMS IN APPLYING AN ELECTROSTATIC
PRECIPITATOR TO A SALVAGE FUEL-FIRED BOILER 415
C.R. Thompson
THE APPLICATION OF ELECTROSTATIC PRECIPITATORS
TO BOILERS FIRING MULTIPLE FUELS 425
R. L. Bump
AUTHOR INDEX 435
xvii
-------
VOLUME III CONTENTS (cont.)
Page
APPLICATION OF HIGH ENERGY VENTURI SCRUBBERS
TO SEWAGE INCINERATION 102
F.X. Reardon
AN INCINERATOR SCRUBBER THAT WORKS:
A CASE STUDY Ill
C. Menoher
EVALUATION OF ENTRAINED LIQUOR CONTRIBUTION TO
TOTAL MASS EMISSIONS DOWNSTREAM OF A WET SCRUBBER . . .119
W. David Balfour, L.O. Edwards and HJ. Williamson
Section B - Fabric Filters
A DUAL-BEAM BACKSCATTER BETA-PARTICLE GAUGE
FOR MEASURING THE DUST CAKE THICKNESS ON OPERATING
BAG FILTERS INDEPENDENT OF POSITION 128
R.P. Gardner, R.P. Donovan and L.S. Hovis
DIAGNOSING FILTER FABRIC CAPABILITIES WITH LIGHT
SCATTERING AND NUCLEI DETECTING INSTRUMENTATION . . . .140
R. Dennis, D.V. Bubenick and L.S. Hovis
ACID DEWPOINT CORROSION IN PARTICULATE
CONTROL EQUIPMENT 150
T.E. Mappes, R.D. Terns and K.E. Foster
SECOND GENERATION OF EMISSIONS CONTROL
SYSTEM FOR COKE OVENS 160
J.D. Patton
EFFECTS OF FLYASH SIZE DISTRIBUTION ON THE
PERFORMANCE OF A FIBERGLASS FILTER 171
W.F. Frazier and W.T. Davis
FUNDAMENTAL STUDY OF A FABRIC FILTER
WITH A CORONA PRECHARGER 181
K. linoya and Y. Mori
ECONOMIC EVALUATION FACTORS IN BID
EVALUATIONS—A SENSITIVITY ANALYSIS 193
J.G. Musgrove and J.E. Shellabarger
FLY ASH RE-ENTRAINMENT IN A BAGHOUSE—
WHAT DOES IT COST? 201
J.G. Musgrove
xix
-------
VOLUME III CONTENTS (cont.)
Page
WHY PERFORM MODEL STUDY OF FABRIC FILTER
COLLECTOR? 211
W.T. Langan, N.Z. Shilling, W.A. Van Kleunen and O.F. Fortune
EXPERIENCES OF A SMALL INSULATION MANUFACTURER
IN MAINTAINING COMPLIANCE WITH AIR POLLUTION
CONTROL REGULATIONS 221
R.L. Hawks
ADVANCED FABRIC FILTER TECHNOLOGY FOR
DIFFICULT PARTICULATE EMISSIONS 228
H.P. Beutner
DEVELOPMENT OF GUIDELINES FOR OPTIMUM BAGHOUSE
FLUID DYNAMIC SYSTEM DESIGN 238
D. Eskinazi, G.B. Gilbert and R.C. Carr
THEORETICAL ASPECTS OF PRESSURE DROP REDUCTION
IN A FABRIC FILTER WITH CHARGED PARTICLES 250
T. Chiang, E.A. Samuel and K.E. Wolpert
EXPERIMENTAL CORRELATION OF DUST CAKE POROSITY,
AIR-TO-CLOTH RATIO AND PARTICLE-SIZE DISTRIBUTIONS . . . .261
T. Chiang and R.L. Ostop
MODEL FOR DUST PENETRATION THROUGH A
PULSE-JET FABRIC FILTER 270
D. Leith and MJ. Ellenbecker
PERFORMANCES OF DUST LOADED AIR FILTERS 280
C. Kanaoka, H. Emi and M. Ohta
ELECTROSTATICALLY ENHANCED FABRIC
FILTRATION OF PARTICULATES 290
T. Ariman and S.T. McComas
A STAGGERED ARRAY MODEL OF A FIBROUS FILTER
WITH ELECTRICAL ENHANCEMENT 301
F. Henry and T. Ariman
Section C - Granular Beds
AEROSOL FILTRATION BY A COCURRENT MOVING
GRANULAR BED: PENETRATION THEORY 311
T.W. Kalinowski and D. Leith
FUNDAMENTAL EXPERIMENTS ON A GRANULAR BED FILTER 321
K. linoya and Y. Mori
XX
-------
VOLUME III CONTENTS (cont.)
Page
DRY DUST COLLECTION OF BLAST FURNACE
EXHAUST GAS BY MOVING GRANULAR BED FILTER 332
A. Wakabayashi, T. Sugawara and S. Watanabe
Section D - Novel Devices
IRON AND STEEL AIR POLLUTION CONTROL
USING MAGNETIC SEPARATION 341
D.C. Drehmel, C.E. Ball and C.H. Gooding
TECHNICAL AND ECONOMIC EVALUATION OF TWO
NOVEL PARTICULATE CONTROL DEVICES 353
R.R. Boericke, J.T. Kuo and K.R. Murphy
TM
THE ELECTROSCRUBBER111 FILTER—APPLICATIONS
AND PARTICULATE COLLECTION PERFORMANCE 363
D. Parquet
HIGH EFFICIENCY PARTICULATE REMOVAL WITH
SINTERED METAL FILTERS 373
B.E. Kir stein, W.J. Paplawsky, D.T. Pence and T.G. Hedahl
APPLICATION OF ELECTROSTATIC TECHNIQUES TO
THE REMOVAL OF DUST AND FUME FROM THE
INDUSTRIAL ENVIRONMENT 382
S.A. Hoenig
k
THE DRY VENTURI 393
A.J. Teller and D.R.J. Roy
FIBER BED FILTER SYSTEM CONTROL OF
WELDING PARTICULATES 398
J.A. Bamberger and W.K. Winegardner
THE USE OF GLASS CAPILLARY FILTERS TO
CLASSIFY ACTINOLITE FIBERS 406
J.W. Gentry, T.C. Chen, S.W. Lin and P.Y. Yu
ULTRA-HIGH EFFICIENCY FILTRATION SYSTEMS
(AIR RECIRCULATION) 417
R.W. Potokar
THE WET WALL ELECTROSTATIC PRECIPITATOR 428
J. Starke, J. Kautz and K-R. Hegemann
xxi
-------
VOLUME III CONTENTS (cont.)
Page
Section E - Mechanical Collectors
TROUBLESHOOTING MULTIPLE CYCLONES ON
FUEL-OIL-FIRED BOILERS 438
F. Crowson and R.L. Gibbs
COLLECTION EFFICIENCIES OF CYCLONE SEPARATORS 449
P.W. Dietz
ELECTROSTATICALLY AUGMENTED COLLECTION
IN VORTICAL FLOWS 459
P.W. Dietz
HIGH PERFORMANCE CYCLONE DEVELOPMENT 468
W.G. Giles
AUTHOR INDEX 481
xxii
-------
HIGH TEMPERATURE PARTICLE COLLECTION
WITH A. P. T. EPxP DRY SCRUBBER
By: S. Yung, T. Lee, R.C. Patterson and S. Calvert
Air Pollution Technology, Inc.
4901 Morena Blvd., Suite 402
San Diego, CA 92117
B.C. Drehmel
Particulate Technology Branch
Industrial Environmental Research Branch
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
The A. P. T. EPxP dry Scrubber is a novel device for controlling fine
particle emissions at high temperatures and pressures. It uses relatively
large particles as collection centers for the fine particles in the gas
stream. Fine particles and collector granules are contacted in a venturi
type contactor and fine particles are collected by the granules through
the mechanisms of inertial impaction, diffusion and electrostatic deposi-
tion. For maximum efficiency the particles are pre-charged and the
collectors are polarized.
3 3
Bench scale (0.5 Am /min) and pilot scale (4.8 Am /min) experiments
have been run at temperatures from 20 C to 820 C to determine the perform-
ance characteristics of the system. This paper presents the system design
and experimental results for the bench scale and pilot scale tests.
NOTE: Please contact the authors for further information regarding
this paper.
-------
PARTICLE COLLECTION IN CYCLONES AT
HIGH TEMPERATURE AND HIGH PRESSURE
By: R. Parker, R. Jain, and S. Calvert
Air Pollution Technology, Incorporated
4901 Morena Boulevard, Suite 402
San Diego, CA 92117
D. Drehmel and J. Abbott
Particulate Technology Branch
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
An experimental study of cyclone efficiency and pressure drop was
conducted at temperatures up to 700°C and pressures up to 25 atm. The
cyclone efficiency was found to decrease at high temperature and increase
at high pressure for a constant inlet velocity. Available theoretical
models could not predict the observed effects of high temperature and high
pressure on collection efficiency. Pressure-drop models predict the
effects of temperature and pressure fairly well. Collection-efficiency
data correlated well against Reynolds number and the square root of Stokes'
number. This correlation accurately accounted for the effects of both
temperature and pressure. These data are for a 2-ln. diameter cyclone at
relatively low velocities (<5 m/s). Data for a 6-in. cyclone of similar
configuration and operating at 635°C, 700 kPa, and 36 m/s also agreed well
with this correlation.
NOTE: Published in Environmental Science & Technology. 15-451-58
April 1981.
-------
OPERATING RESULTS OF ELECTROSTATIC PRECIPITATORS
AT HIGH TEMPERATURE AND HIGH PRESSURES
By: Paul L. Feldman and K. Sampath Kumar
Research-Cottrell, Inc.
P. 0. Box 1500
Somerville, New Jersey 08876
ABSTRACT
This paper presents Research-Cottrell's experience with
pilot electrostatic precipitators (ESPs) under extreme conditions
of temperature and pressure. Case studies of five pilot systems
are provided which describe operations at temperatures upto
1700°F and pressures upto 850 psig. This discussion brings out
the lessons learned from these pilot ESP's and points out the
future development needs in selection of materials, mechanical
design and process understanding. Some case studies have a
direct bearing on the design and operation of ESP's for hot gas
cleanup in pressurized fluidized bed combustion systems.
INTRODUCTION
The investigation of electrostatic precipitator for applica-
tions in unusual temperature and pressure environment has been
underway for several decades. During mid 1920's research began
on the application of high pressure ESP's for high pressure coal
gas pipelines to collect oil and tar impurities from the gases.
Elaborate research work was later pursued by Wintermute(1' in the
30's, Cooperman(2) in the 50's, Shale(3), Brown(^), and
Robinson'-*' in the 60's. Bush, Feldman and Robinson' continued
the research work on corona stability in the 70's covering a
wide range of temperature and pressure.
Because of the strong foundation of research established by
workers during earlier decades, opportunities surfaced during
the 60's for pilot scale experiments on the performance charac-
teristics of the high pressure and high temperature(HTHP) pre-
cipitators.
CASE STUDIES
First of the five pilot programs to be discussed is the
Union Carbide pilot plant. Table 1 is a concise compilation of
all the important aspects of these five pilot plants.
-------
TABLE 1. PILOT UTUP ELECTROSTATIC 1-RECIPITATOR EXPERIENCE
Item
Dates Tested
Location
Fuel
System
Dust: Type
Mean 0 CM)
T <°K)
P (ATMl
* Pipes Tested
Pipe Diameter (cm)
Pipe Length (m)
** Gas Velocity (m/s)
Gas Flow (kG/hr)
SCA (ftVlOOO acfm)
Union Carbide
Oleflns
1962-1964
Institute, W. VA
Coal
PFB Burner
Ash
2.6
870-1000
3-8
19
15.2
1.83
.2-1.2
S70-820
200-1200
Bureau of Mines
1963-1968
Morgantown, W. VA
Natural Gas
H.P. Gas Burner (lab)
Flyash - Reinjected
30
1000-1100
4.5-6.5
16
15.2
1.83
.3-1.1
2600
•v.230
Combustion Power Co.
1966-1968
Montebello, CA
Methanol
S.U.E. Burner
Alumlna/Flyash In-
jected
6-10
1150-1200
4.5-11
1
20.3
4.57
2.9
980
157
Texaco Co.
1961-1963
Montebello, CA
Heavy Oil (*840
12° APT)
Gaslfler
Carbon
0.007 to 0.01
400-465
21-25
5, 3, 1
10.2, 15.2, 20.3
1.83
.61-1.1
490
180-300
Texas
Gaa
Transmission Co.
1965-1967
Lake Cormorant, MI
Pipeline
Pipeline
Lube - Oil
0.7
310-316 *
53-56
Plates: 7
7.6 (Plat*
3.05 x .6
3-9
163,000
96
Ducts
spacing)
high
Discharge Electrode:
Type Hires
Discharge Electrodei
Diameter (mm) 2.1, 3.4
Typical Collection
Efficiency (%) 98.8
Cleaning System)
HI Rapper
LT Rapper
Hires
2.1
91-96
Rapper
Rapper
Twisted Hires
3 x 2.8
80-82
Rapper
Rapper
Twisted Hires, Rigid
Electrodes
Hires
One 3 x 1.8 Twisted, .25
Five 3 x 1.8 Twisted,
6 Vane Paddlewheel
98.8
Hater Flush
99
Spray Systems
-------
Union Carbide Olefins Co.
Union Carbide, in 1961, operated a pilot pressurized
fluidized bed combustor (PFBC) in Institute, W. Virginia. A
HTHP precipitator was supplied in early 1963 with the purpose of
removing particulates at 1300°F and 100 psig, prior to the gas
turbine cycle.
The test program consisted of over one hundred tests in a
six month period. Current-voltage characteristics were obtained
for various temperature and pressure conditions, which are des-
cribed in Figure 1. It could be seen from this illustration that
operating pressure had an very important bearing on the ESP oper-
ation at Union Carbide. Even though the temperature was about
1300°F, the operating pressure of 70 psig shifted the i-v curve
to the right, making possible the application high electric
field strengths, often exceeding 20 kV/in. Figure 2 illustrates
the effect of operating field strengths on collection efficiency
of ESP. Higher operating field strengths resulted in higher col-
lection efficiency, conforming to the dust collection phenomena
of ESP's in high sulfur coal applications.
The ESP consistently achieved collection efficiencies of
98-99%. Flyash particulates were rather fine with a mass median
diameter of 2.6 microns. Precipitator gas velocities were gen-
erally above 4 ft/second. No difference in particulate collec-
tion efficiency was reported for two wire diameters that were
tested. There were no operating problems encountered with this
unit. Both the collecting tubes and discharge electrodes were
kept clean by the cleaning system arrangement.
Texaco
In early 1960 Texaco operated a pilot synthesis gas genera-
tor at Montebello, California. Oxidation of heavy fuel oil had
geneated fine carbon material which was limiting the life of
shift converter catalysts downstream. Research-Cottrell was
asked to design a system that would reduce the concentration of
these submicron (0.007 to 0.01 ym) carbon particles from 400 ppm
down to 5 ppm or less. Operating temperatures were typically
350°F but the operating pressure was high, about 350 psig.
Research-Cottrell designed a wet wall precipitator to pre-
vent both carbon reentrainment and build up on electrodes. Two
rigid discharge electrode designs are reported here, one with a
six vane "paddlewheel" structure and another with a rigid cen-
tral mast with five twisted wires mounted around it.
Seventy short runs and one extended test run of one hundred
hours were examined. The eight inch collecting pipe had maximum
electrical stability with the wet wall opeation. Current-voltage
-------
.6
*"!. .3
d=1.0
0 PSIG, 70°F
50 PSIG, 1125°F
70 PSIG 1290°F
10 15 20 25 30
E (kV/ln)
Figure 1 CURRENT-VOLTAGE CURVES
TAKEN IN PROTOTYPE
HTHP PPTR PILOT
(Union Carbide Olefins Co.)
99.9
99.8
O 99.5
111
O 99
EL
111 98
O
§ «
111
d go
80
50
10 15 20 25
AVERAGE FIELD STRENGTH, kV/ln
Figure 2 PERFORMANCE of a HTHP
WIRE-PIPE PPTR FOLLOWING an
OPERATING PFBC at 1300°F
(Union Carbide Olefins Co.)
-------
characteristics with the eight inch tube are shown with 5-wire
stiff discharge electrode on Figure 3, while those with the
paddlewheel design are shown in Figure 4. It can be seem from
these figures that good electrical stability could be maintained
by these electrodes, resulting in high operating voltages (and
hence field strengths).
Both the short test runs and the extended operation showed
that particulate loading after the ESP could be consistently
maintained below the 5 ppm level, resulting in particulate col-
lection efficiency near 99%. Above results were possible only
with negative polarity of the applied voltage. When positive
polarity runs were attempted, particulate loading increased to
19 ppm.
Bureau of Mines
During the mid 1960's C. C. Shale extended his HTHP precipi-
tator investigations to pilot scale. Flue gas was generated by
burning natural gas and flyash particle were reinjected to simu-
late the dust burden. Research-Cottrell supplied the precipita-
tor to operate upto 1500°F and 80 psig. Precipitator consisted
of sixteen six inch diameter pipes with 0.083" discharge elec-
trodes. Both pipes and wires were cleaned by pneumatic rapping
systems similar to the Union Carbide system.
Shale encountered several opeating problems with this pilot
unit. Tube sheet, which held all the collecting tubes, was dis-
torted due to thermal expansion problems, and the horizontal
rapping system added to the tubes' further dislocation from the
vertical axis. Misalignment had reduced the clearance between
discharge and collector electrode by an inch, resulting in re-
duced applied voltages. The high voltage support insulator had
overheating problems, which further limited the application of
high voltage. In spite of these adverse conditions, the ESP
performed reasonably well when the polarity of applied voltage
was negat ive.
Collection efficiency under the operating condition of
1470°F and 80 psig was 95% with negative polarity. Operating
voltage was 38 kV at a precipitator current of 180 mA. However,
the collection efficiency with positive polarity was 75%.
Though the operating voltage was high, at 55 kV, the precipita-
tor current was only about a fifth of that with negative polar-
ity at 40 mA.
It is felt that the high current densities in the precipita-
tor can introduce sufficient adhesive forces on collecting elec-
trodes overcoming the reentrainment problems associated with
high conductivity of particulates at the high operating temper-
atures. Positive polarity inspite of higher voltage, may be
-------
00
1.60
1.40
UI 1.20
-UI
£°
gO
<=»-
S=o
3UI
O-l
0.80
0.60
CC
O
O 0.40
CC
£
Q.
0.20
5-WIRE STIFF DISCHARGE
ELECTRODE
(8' Diam. Pipe 6' Long)
• = ATMOSPHERIC AIR
• = NATURAL GAS 310 PSIG 90°F
X= SYNTHESIS GAS
360 PSIG 250-280'F
(Texaco)
20 40 60 80
PPTR VOLTAGE - kV
100
avg
Figure 3 CURRENT-VOLTAGE
CHARACTERISTICS
120
1.60
1.40
O
111 1.20
O-J
gg 0.80
0.60
O 0.40
CC
£
O.
0.20
6-VANE PADDLEWHEEL
DISCHARGE ELECTRODE
(8' Diam. Pipe 6 Long)
• = ATMOSPHERIC AIR
• = NATURAL GAS 350 PSIG
A = SYNTHETIC GAS
355 PSIG 375'F
(Texaco)
40 60 80
PPTR VOLTAGE - kV
avg
Figure 4 CURRENT-VOLTAGE
CHARACTERISTICS
100
120
-------
suffering from particulate reentrainment problems associated
with low current densities.
Figure 5 shows the current voltage relationship with nega-
tive polarity at various relative densities, at a temperature of
1470°F. Again, we see the shift in these curves to the right as
the relative density (or pressure) goes up. Runaway currents
are possible at high temperatures if the operating relative
density is too low. Shale noted in his work that negative
field required a relative density of at least 1.25 for corona
stability.
Combustion Power Company
In the mid 60's Combustion Power Company was investigating
pressurized incineration of municipal solid waste for U.S. De-
partment of H.E.W. Particulate collection was to be accomp-
lished by an electrostatic precipitator at 1600-1700°F and 50-
150 psig.
Figure 6 shows the typical i-v data at two levels of gas
density for positive and negative polarity. Testing showed that
both negative and positive polarity were stable at 1700°F as
long as pressure was greater than 3.5 atmospheres. Also, the
positive polarity showed higher sparkover voltage than negative.
However,as noted 'in the earlier work, negative corona re-
sulted in much higher collection efficiencies than positive.
Figure 7 correlates electric field strength (E) to collection
efficiency (ri) for both negative and positive polarity. Good
correlation with E^ and In (l-r|) is seen, which is typical of
conventional precipitators.
Alkali (KC1) addition to flue gas was tested to determine
if high alkali levels would result in excessive unstable cur-
rents at the high temperatures. Figure 8 describes the effect
of KC1 addition at 0, 40 and 400 ppm at 1700°F. The addition
of alklai increased the current between 60 to 90% at 40 ppm and
several times at 400 ppm. Corona instability is indicated at
the high ppm levels of alkalis. The real effect of alkali pres-
sure in PFBC effluents on the ESP remains to be seen during the
ensuing pilot plant work with DOE.
One of the problems noted in this work was the discharge
electrode oscillation at high temperatures. Oscillation spans
of 2-4" had limited the electrical operation to 12 kV/in. Much
higher field strengths would have been possible in the absence
of this phenomenon. This adverse phenomenon may partially ex-
plain the low efficiency of 87% attained during the tests. In
addition to low field strengths (and hence low precipitator
current), high gas velocities of 9.5 ft/sec during the tests
9
-------
.7
.5
*"H .3
.2
10
15
20
25
30
E (kV/in)
Figure 5
HTHP PILOT PPTR CURRENT-VOLTAGE
DATA TAKEN by C. SHALE at MORGANTOWN
COAL RESEARCH CENTER, US BUREAU off MINES
Figure 6
4=2.31
2.27
E (kV/in)
CURRENT-VOLTAGE DATA FOR
T m 1650-1720 »F, P - 50-140 PSIA
(Combustion Power Co.)
-------
20
40
o
UJ
o
u.
u.
Ul
a.
a.
60
80
90
CORONA
STARTING FIELD
POSITIVE
POLARITY
NEGATIVE
POLARITY
T = 1172°K
P = 8 atm
1.0
0.8
0.6
,7 r
.5
0.4
0.2
0)
w
o
O.
Q.
—— 40 ppm kCL
400 ppm kCL
10
15
20
25
E (kV/in)
20
40
60
80
100
120
0.1
140
E2 - PPTR FIELD STRENGTH (kV/ln)2
Figure 7 PPTR COLLECTION EFFICIENCY ON
ALUMINA TEST DUST AS A FUNCTION
OF FIELD STRENGTH
(Combustion Power Co.)
Figure 8 EFFECTS of ALKALI ADDITION
AT1700°F, 115PSIA
(Combustion Power Co.)
-------
could also have contributed to a reentrainment problem.
Texas Gas Transmission Co.
Natural gas pipelines had faced the problem of submicron
oil mist from oil lubricated compressors, thereby limiting the
pipeline transmission efficiency. In 1965 Research-Cottrell
worked with Texas Gas Transmission Co. to develop a precipitator
for removing the oil mist and increasing pipeline efficiency.
The precipitator operated for one year in steady state oper-
ation and was a successful demonstration. Collection efficien-
cies were above 99%. Inlet grain loadings were 0.03 grains/
acf with a particle size of 0.7 micron. The pipeline efficiency
improved from 93.6% to 99% with the use of ESP's.
FUTURE DEVELOPMENT NEEDS
Although the case studies of five pilot precipitators pro-
vide data base for design with more confidence, there is a de-
finite need for further extended testing on the demonstration
scale to resolve several key issues.
Ash Characteristics
Ash characteristics such as stickiness, ease of dislodge-
ment are important issues affecting precipitator performance
and these can only be evaluated under actual operation condi-
t ions.
Reentrainment
It is important to characterize the parameters that can exa-
cerbate the reentrainment of ash back into the gas stream, thus.
lowering the overall collection efficiency. Factors such as
gas velocity, applied electric field strengths and rapping char-
acteristics need to be evaluated to ensure high efficiency oper-
ation .
Electrode Alignment
Good alignment between discharge electrodes and collecting
tubes is of paramount importance to maintain high precipitator
efficiency. Mechanical designs should account for thermal ex-
pansion effects and rapping stress effects to overcome the mis-
alignment potential. Further, only an on-site demonstration
can confirm the efficacy of such designs.
Discharge Electrode and Collector Tube Cleaning
Though rappers have been used in the past work, there is a
12
-------
necessity to minimize impacting requirements on component mem-
bers at high temperature to improve hardware reliability. Use
of sonic energy to dislodge collected particulates has been
successfully demonstrated in several applications. Sonic energy
systems do not impose severe stresses on structural members.
Therefore, optional combination of cleaning systems is essen-
tial to improve component reliability.
Tube Shape
Collector tubes using cylindrical pipes have inherent defi-
ciency in packing more collector area for a given volume. This
leeds to larger pressure vessels and hence are less cost effec-
tive. Designs should consider "hexagonal" type collector tubes
with a honeycomb arrangement. If this design proves out to be
successful, significant cost savings can be achieved by im-
proving packaging efficiency of precipitator components.
Power Supply and Bushings
Optimum operation may require use of high voltage power
supplies that are at present beyond the state-of-the-art tech-
nology. Data for such applications depend on the temperature,
pressure condition of the process. Further, temperature con-
trolled insulators are necessary to overcome heating problems
at high temperature.
Materials of Construction
High temperature applications beyond 1500 F need careful
choice of materials that have both oxidation resistance and high
yeild strengths under typical ESP operation.
Economics
Cost analysis performed by Bush, et al have shown that
high temperature high pressure precipitators can be cost effec-
tive for PFBC conditions, assuming reasonable sizing and design
parameters. Demonstration scale tests are needed to generate
data for commercial offerings.
CONCLUSIONS
Though the past pilot plants at extremes of temperature and
pressure have provided a data base for design and sizing of
ESP's, several development questions need to be answered. Ex^
tended testing on "demonstrator" pilot plants are needed prior
to reliable product offerings. Research-Cottrell's ongoing
contract with DOE to demonstrate HTHP precipitator performance
on a 2 MW PFBC system for hot gas cleanup is a step in the
right direction. It is felt that emerging energy technologies
13
-------
will revive the interest in ESP use for extreme temperature
and pressure applications.
ENDNOTES
1. H. Wintermute, "Report on Voltage and Current Tests at High
Pressures", Research Corporation, April (1931).
2. P. Cooperman, "The Effects of Gas Temperature, Pressure and
Composition on the Electrical Characteristics of the Corona
Discharge", Research-Cottrell, September (1953).
3. C. C. Shale, "High Temperature, High Pressure Electrostatic
Precipitation", APCA Paper No. 66-125, 1966.
4. R. F. Brown, "An Experimental High Temperature High Pressure
Electrostatic Precipitator Module Design and Evaluation",
Research-Cottrell, (1969).
5. M. Robinson, R. F. Brown, F. W. Schmitz, "Pilot Plant Tests
of an Electrostatic Precipitator on High Pressure Synthesis
Gas (Texaco Process)", Research-Cottrell, Inc., April (1963),
6. J. R. Bush, P. L. Feldman, M. Robinson, "Development of a
High Temperature, High Pressure Electrostatic Precipitator",
EPA-600/7-77-132, November (1977).
7. J. R. Bush, F. L. Blum, P. L. Feldman, "Comparative Economic
Analyses of Selected Particulate Control Systems for Ad-
vanced Combined Cycle Power Plants", Second EPA Symposium
on Transfer and Utilization Control Technology, July, 1979.
14
-------
CONTROL OF PARTICULATES IN PROCESS AREA 12,
SOLVENT REFINED COAL PROCESS
By: W. H. Wilks, P.D. Wilkinson
Catalytic, Inc.
1500 Market Street
Philadelphia, Pa. 19102
J.A. Schlosberg
International Coal Refining Company
P.O. Box 2752
Allentown, Pa. 18001
ABSTRACT
The Solvent Refined Coal Process is one of the major undertakings in the
Synthetic Fuels Program of the U.S. Department of Energy (DOE). Internationa]
Coal Refining Company (ICRC), a partnership of Air Products and Chemicals
and Wheelabrator-Frye, is now in the design stage of a 6000 TPD demonstration
plant. There are many facets of the process where proven technology is
not yet available, which is part of the reason for a relativly large demons-
tration plant. The environmental atmospheric emission problems on the
other hand, were investigated and solved with established control technology.
Catalytic, Inc. has the responsibility for the design, engineering and
special equipment procurement in one process area. This paper describes
the environmental approach used by Catalytic, Inc. to control particulate
emissions in that area. The contents include: a description of the SRC-1
process, identification of the emission sources, definition of the problem
area, application of basic control principles, and equipment selection
and design based on accepted and proven approaches.
INTRODUCTION
The Solvent Refined Coal (SRC-1) Process, one of the most advanced direct
coal-liquefaction processes available, has attracted considerable national
attention as a partial answer to the continuing energy crisis. The SRC-1
Process will utilize America's abundant supply of high-sulfur bituminous
coal to produce a varied selection of reliable and environmentally acceptable
products. These include low-sulfur ash-free boiler fuels, metallurgical
and anode grade coke; and with further refinement distillate fuels as
well as petrochemical and gasoline feed stocks. Air Products and Chemicals,
Inc. and Wheelabrator-Frye Inc. both bring much research, development,
and engineering experience in coal-liquefaction technology to the Demonstra-
tion Plant Program. Catalytic, Inc. (a subsidiary of Air Products) has
engineered, built, and now operates a six ton-per-day SRC Pilot Plant
at Wilsonville, Ala., while Rust Engineering (A subsidiary of Wheelabrator-
Frye) designed built and operates the DOE Pilot Plant at Fort Lewis, WA.
During the initial phase of the current project (referred to as Phase 0),
Catalytic has used this experience in the conceptual design and environmental
appraisal in its assigned process area. In addition, Catalytic has evaluated
process options and identified critical technology areas requiring additional
15
-------
data. Alternatives and refinements are being evaluated in Phase 1 to
improve overall process and environmental systems. The ultimate purposes
of the 6000 TPD Demonstration Plant are to establish the feasibility of
large-scale production economics, to prove the applicability and reliability
of commercial mechanical equipment, and to define environmental problems
and evaluate the applied control technology. Ultimately, this will serve
as the basis for expanding to a 30,000 TPD commercial plant.
DESCRIPTION OF THE SRC-1 PROCESS
The SRC-1 process begins with mixing of dry pulverized coal with a process-
derived solvent in the Coal Slurry Drums. (See Figure 1.) The resulting
slurry is pumped to reaction pressure, a hot hydrogen-rich recycle gas
is added before the mixture is heated to reaction temperatures in the
Coal Slurry Heaters. Additional hot hydrogen-rich recycle gas is added
at the exit of these fired heaters, and the mixture flows through the
first coal dissolver. Cold hydrogen-rich recycle gas is added as a quench
at the exit of the first dissolver before entry into the second dissolver
for further hydrogenation and desulfurization, producing a slurry of low-
sulfur SRC (Solvent Refined Coal) and mineral ash, plus liquid and gaseous
by-products.
Vapors are separated from the liquids and the SRC/mineral ash slurry in
a High-Pressure (HP) Separator (operating at reaction conditions) and
in the Medium-Pressure (MP) and Low-Pressure (LP) Flash Drums, in which
the pressure is progressively lowered resulting in further vapor/liquid
separation. Vapors containing excess unreacted hydrogen are partially
condensed in the Hydrogen Recovery Section, medium-pressure hydrogen-rich
sour gas is separated, compressed, and combined with high pressure sour
gas and forwarded to a gas-treatment unit for acid-gas removal before
being recycled to the coal dissolvers. Hydrocarbon liquids recovered
from the Hydrogen Recovery Section and vapors from the LP Flash Drum are
sent to the Solvent Column for fractionation into medium and heavy oils.
Over-head distillates from this column are partially condensed, yielding
a medium oil product and a low-pressure noneondensable off gas which is
compressed and forwarded to gas treatment for use as process fuel gas.
The bottoms product is the process solvent or heavy oil; part is sent
to the Process Solvent Drum for recycle to the Coal Slurry Drums, and
the rest removed as heavy oil product.
The SRC/Mineral ash slurry exiting the LP Flash Drum is charged to a Vacuum
Column for further recovery of process solvent. Non-condensable overhead
vapors are collected in a vent system and incinerated in the Coal Slurry
Heaters. The bottoms product (SRC, mineral ash, and unconverted carbon)
is sent to the Kerr-McGee Critical Solvent Deashing Unit (CSDU) where
solid mineral ash and unconverted carbon are separated from the SRC using
a proprietary process. Ash concentrate from the CSDU is conveyed to a
gasifier for hydrogen production and to convert the ash to an environ-
mentally acceptable slag. Molten ash-free SRC from the CSDU is then divided
into three approximately equal streams. One-third is delivered to an
Expanded-Bed Hydrocracker for further hydrogenation to liquid distillates,
one-third is sent to a Coker-Calciner for coke production, and the remainder
16
-------
is cooled and solidified on the SRC Solidifiers (sufficient capacity is
provided to solidify 100 percent of the Molten SRC produced). The low-sulfur,
ash-free solid SRC product is then conveyed to storage.
BASIS FOR SELECTION OF CONTROL SYSTEMS
Government Requirements
The entire facility is considered a "major source" that will be constructed
in an environmental area designated as attainment. It is therefore subject
to EPA's PSD (Prevention of Significant Deterioration of Air Quality)
regulations, which have been promulgated in 40 CFR 52.21. This requires
the application of BACT (Best Available Control Technology).
This requirement for the application of BACT means that the emissions
limitations (including a visible-emissions standard) must be based on
the maximum degree of reduction achievable for each pollutant subject
to control under the Clean Air Act. The maximum degree of reduction achiev-
able for each pollutant emitted is determined by the Administrator on
a case-by case basis. This determination takes into account energy, environ-
mental, and economic impacts and costs. Beside the conventional air-pollu-
tion-control approaches, this effort involves application of in-plant
pollution-abatement measures in the production processes and of other
control methods, including fuel cleaning or treatment, and innovative
combustion techniques.
The application of BACT cannot result in emissions of any pollutant that
would violate any applicable standard under the New Source Performance
Standards (40 CFR 60), National Emission Standards for Hazardous Air Pollut-
ants (40 CFR 61), or an approved State Implementation Plan.
Process Restrictions
Some information required for the Kentucky Air Permit applications was
not available, either because of process decisions that were yet to be
made, detailed engineering that was not yet completed, or data to be supplied
by a vendor who had yet to be selected. If permit application preparation
were delayed until all such needed information was available, construction
could be delayed a year or more. We therefore addressed all relevant
air pollution control and air quality concerns in the most thorough manner
possible. Kentucky had agreed to accept preliminary information based
on best engineering judgement as to the outcome of process option studies
and detailed design. This was contingent upon the willingness to make
conservative judgements with respect to potential air quality impacts
and to update the preliminary information as necessary. This approach
is typical of environmental input at the onset of a project, where the
environmental engineer works closely with process engineers and other
disciplines. It has resulted in a mutual understanding of each other's
responsibilities and problems and has established a consensus for selecting
the best solutions, whether through process change and/or control-equipment
application.
17
-------
IDENTIFICATION OF PROBLEMS AND DEVELOPMENT
OF SELECTED CONTROL SYSTEMS
Process Heaters
The process area (Refer to Figure 1) will have four indirect-fired types
of heat exchangers, which are governed by Kentucky Air Pollution Control
Regulation 401 KAR 59:015, Section 4:
1. The coal-slurry heaters - to heat the coal slurry to the reaction
temperature required in the dissolvers.
2. The CSD heater - to heat the proprietary solvent and the SRC
mixture.
3. The vacuum heater - used only on startup, to heat the SRC solvent
and ash slurry and to raise the vacuum column to designated
normal operating temperature.
4. The hot-oil heater - to raise heat-transfer fluid to operating
temperatures.
The total heat level for all heaters is 522 MM BTU/Hr. During normal
operation the heaters will be fired with process-derived fuel gas, which
is equivalent to natural gas. During startup the heater will be fired
with process-derived fuel oil, equivalent to No. 6 fuel oil. Regular
operating schedule will be 24 hours per day, 330 days per year, based
on one scheduled 35-day shutdown.
Proposed Controls
The control of particulate emissions is a function of the design and operation
of the combustor, which will be designed for low excess air combustion
with commercially-proven burners. No attempt will be made to operate
at off-stoichiometric combustion conditions; thus any unburned hydrocarbons
due to incomplete combustion will be insignificant.
The residence time in the heaters will be 1-2 seconds at 2200 F, plus
an additional second at 1500 F. This time/temperature combination will
achieve 100 percent combustion of hydrocarbons, with a corresponding reduction
in particulates approaching zero emission. A continuous monitoring instrument
will be used to measure unburned hydrocarbon in each stack. If the emission
levels approach design limitations, excess air rates will be changed to
insure complete combustion.
The opacity regulation for individual indirect heat exchangers with heat
input capacity less than 250 MM/BTU/Hr is that a maximum of forty (40)
percent opacity shall be permissible for not more than six (6) consecu-
tive minutes in any sixty (60) consecutive minutes during cleaning of
the fire box or blowing soot. This regulation will be met by determining
the frequency of soot blowing during actual operation of each heater.
18
-------
Expected Control Efficiency
The design and operation of the heaters, as confirmed by the monitoring,
will result in emissions levels for particulate well below the value stated
in the Kentucky regulation, which is 0.10 Ib/MM BTU for a source (all
heaters) having a total heat input greater than 250 mm BTU/Hr. The values
submitted on the Kentucky Air Pollution Control Permits Applications were
from EPA AP-42 "Compilation of Air Pollutant Emission Factors", which
are 0.0095 Ib/MM BTU for gas-fired and 0.056 Ib/MM BTU for oil-fired heaters,
Based on vendor information, the actual emissions should be less than
these factors. The schedule of soot blowing will insure compliance with
the opacity regulation.
BACT Rationale
The operational controls adopted for these heaters will reduce the emissions
of particulate (including opacity) well below the allowable for applicable
State regulations and also meet the BACT control requirements. Additional
end-of-pipe controls for further reduction are not warranted.
Emissions from Process Operations
The following two vent streams emitting particulate are governed by Kentucky
Air Pollution Control Regulation 401 KAR 59:010 for New Process Operations:
1. The emissions produced when pulverized coal and process solvent
are charged to the coal slurry drums
2. The emissions produced from exhausting the solidifier hoods
during the solidification of SRC in the SRC solidifiers.
Proposed Control System
Proposed Control System 1 is shown schematically in Figure 2 and described
in the following paragraphs.
Emissions from the Coal Slurry Drums (water vapor, nitrogen, argon, process
solvent vapors, and coal particles) will be treated for particulate removal
in a venturi scrubbing system consisting of coal slurry vent ejectors,
vent separators, and vent condensers.
Recovered solvent/water will be reprocessed and the solvent reused. Coal
Slurry Vent Condenser emissions, which will consist of small amounts of
particulate and solvent vapor in nitrogen and argon, will be sent through
Coal-Slurry Vent Blowers to the SRC H.C. Vent Vapor Seal Pot.
Emissions from the SRC Solidifier hoods through the Solidifier Fume Blowers
(SRC particulate and hydrocarbon vapors) will be sent to the Solidification
Vent Vapor Seal Pot. The emissions from points 1 & 2 (see Figure 2) will
be combined and ultimately incinerated in the Coal Slurry Heaters, along
with fuel gas normally used to fire these heaters.
19
-------
When the Coal-Slurry Heaters are shut down, the vapor stream from the
Solidification Vent Vapor Seal Pot will be diverted to the Vent Waste
Incinerator because the SRC solidifier has to operate after the rest of
the process is down in order to process residual molten SRC material in
the process vessels and in the Molten SRC Tank. The solidifier will operate
at a reduced rate under these conditions to keep the incinerator size
to a minimum. The incinerator is expected to operate approximately 1-2
days per shutdown, and fired with process fuel oil.
When the entire process is shut down, the stream from the SRC Vent Vapor
Seal Pot will be diverted to the Flare Header. The vapor stream at this
time will contain only small quantities of emissions, resulting from the
breathing of several vessels.
The residence time in the coal slurry heaters will be 1-2 seconds at 2200 F,
plus an additional second at 1500 F.
The residence time in the Vent Waste Incinerator will be 2 seconds at
1000 C (1832°F), which are the operating criteria proposed by EPA in Section
250.45-1 of Hazardous Waste Guidelines and Regulations dated December 1978.
These criteria were suggested by the Kentucky Division of Air Pollution
Control as acceptable without further qualification. Other operating
conditions can be submitted when updating the permits, as long as there
is sufficient technical backup and documentation.
Expected Control Efficiency
The vent-ejector type of scrubber, operating with liquid at the appropri-
ate pressure (minimum of 100 psig) and liquid/gas ratio (40 gpm/1000 acfm),
will remove a minimum of 94 weight percent of particles larger than 2 microns.
In the condenser/accumulator, the coolant temperature will be such that
96.8 percent of the total vapors will be condensed and reprocessed. Also,
during the condensing unit operation, the remaining particulate will act
as a nucleus for formation of condensed droplets. This will result in
removal of at least 80 percent of the remaining particulate.
In the heater operation, there will be more than enough time and temperature
for 100 percent combustion of process solvent, waste vapors, and combustible
particulate (coal), and also for 100 percent conversion of the H_S to
SO . 2
x
The time and temperature conditions selected for the Vent Waste Incinerator
will assure 100 percent combustion of process solvent, waste vapor, and
combustible particulate.
BACT Rationale
The vapors from the Coal-Slurry Drums will contain a considerable amount
of water, resulting in high humidity; the particulates will be sticky
and tend to agglomerate in a solid mass. A wet scrubber, therefore, appears
to be the most appropriate device for removal of the particulates in this
20
-------
gas stream. Solvent will be used as the scrubbing liquid because it has
a natural affinity for the coal dust (aiding in the removal of particulate).
The slurry blowdown is returned directly to the Coal Slurry Drums.
No control system that provides a greater degree of particulate removal
is applicable to this process. The physical characteristics of the particu-
late and of the vapors emitted by the Coal Slurry Drums preclude the use
of fabric filters and electrostatic precipitators. The hydrocarbon boiling
point range is such that the condenser/accumulator is highly efficient
(96.84%) in removing total vapors. Incineration of the remaining solvent
vapors and coal dust (along with waste vapors) in a process heater that
operates continually will provide for emission control with a high degree
of efficiency and reliability. Other feasible alternatives such as a
separate thermal combustor or carbon-bed adsorber would not provide any
greater degree of control or reliability; neither would be cost effective,
as each would require additional capital expenditures, additional energy
usage, and manpower for operation and maintenance.
Use of an incinerator as a standby thermal combustor on process heater
shutdown will provide a safe, efficient and reliable means of combusting
the exhaust from the solidifier hoods, which will consist primarily of
air. The use of process fuel will make this unit energy self-sufficient.
A wet scrubber system is not feasible for removing hydrocarbon vapors.
The solubility of the organic vapors in water is very low, thereby resulting
in poor removal efficiencies and creating a water pollution problem.
A carbon bed adsorber although it is highly efficient may present problems
in the disposal of the spent carbon in a sacrificial bed or the condensed
vapors from a regenerative bed.
Decoking Operation
The emissions produced during tube decoking of process heaters are also
governed by Kentucky Air Pollution Control Regulation 401 KAR 59:010 for
New Process Operations.
s
The process heaters will require tube decoking at periodic intervals,
at least once per year. This operation will consist of introducing steam
and air into the tubes while operating the heaters at approximately 10
percent of normal heat capacity. The mechanics of decoking are as follows:
1. Shrinking and cracking coke loose by heating the tubes from
the outside while steam is introduced from the inside. This
dislodges the coke, and the steam flow carries the particles
through the tubes. This is called spalling.
2. Reaction of hot coke with steam, to produce CO, CO., and H-.
3. Reaction of coke and oxygen in air, to produce CO and C0_.
-------
Each decoking operation will require approximately 48 hours per pass.
Some heaters will have multiple passes, and in order to reduce the total
decoking time, several passes can be done in parallel.
Emissions of "coke" particles, CO, and CO- from the decoking operation
will vary from a maximum at the initial stages and taper off to zero at
the end of the cycle. The end of the cycle will be determined by measuring
the CO and CO. levels; a zero reading will indicate completion.
Proposed Control System
The control system for the decoking operation (System 2) is shown schematic-
ally in Figure 3 and described below.
The emissions from the tube decoking of the Coal-Slurry Heaters will be
sent to the Slurry-Heater Decoking Drum, those from the Hot-Oil Heater
to the Hot-Oil-Heater Decoking Drum, and those from the CSD Heater and
Vacuum Column Heater to the CSD/Vacuum-Heater Decoking Drum. The drums
will be cyclonic separators capable of removing more than 90 percent of
the particulate (designated as "Coke"); water will be fed to the top of
the drum to enhance collection of the particulate. The resulting vapor
streams from the decoking drums (essentially water vapor, air, CO, C0_,
H? and "Coke" particles) will be sent to the Vent Waste Incinerator for
combustion of the CO and "Coke" to CO,,. The operating criteria for the
Vent Waste Incinerator for this application will tentatively be the same
as those for Proposed Control System 1.
Expected Control Efficiency
The particulate from the decoking operation will be generated mostly during
the spalling part of the cycle. The particle size can vary from a fraction
of an inch to the sub-micron range. The fines will be produced by the
abrasion of the coke particles along the inside of the tubes due to the
high velocity gas flow. Because there is no documented information on
the expected size distribution of particulate emission from decoking oper-
ations, the efficiency of the decoking drums has to be stated in terms
of typical theoretical efficiencies. Estimated grade efficiencies for
Cyclone separators are 65 percent at 20 microns, 83 percent at 60 microns,
and 96 percent at 60 microns. Since less than 1 percent by weight of
the coke particles will be less than 20 microns, an average efficiency
of greater than 90 percent is estimated. Since this is not acceptable
by Kentucky as BACT, the discharge of the decoking drums will be sent
to the Vent Waste Incinerator. The operation of the incinerator in terms
of residence time and temperature will assure 100 percent conversion of
CO and "coke" to C02.
BACT Rationale
The vapors from the decoking drums will contain CO and particulate plus
a high percentage of water vapor. A cyclonic separator is the most suitable
22
-------
device for this application, because of the high humidity and the size
of the particles more that 99 percent of which will be larger than 20
microns in diameter. Overall efficiency will be greater than 90 percent.
A bag filter would not be appropriate because the high humidity would
cause condensation on the bags and result in plugging of the unit. A
wet-type scrubber, which may be slightly better in collection efficiency,
would require more equipment (ejector venturi, separator, tanks, liquid-
recirculation system, instrumentation, etc.), resulting in higher capital
costs, energy usage and operating labor; and would create a water-pollution
problem. Neither device would remove any of the CO.
The Vent Waste Incinerator is the logical backup control unit for disposing
of the CO and the particulate carryover from the decoking drums. The
equipment has already been justified for use in the previously described
Control System 1, thereby making it even more cost effective. It would
be 100 percent efficient in converting the contaminants. (The use of
process fuel will produce negligible amounts of criteria pollutants.)
Fugitive Particulate Emission
Any areas in the process where there is a possibility of fugitive particulate
emissions will be equipped with a hood or similar device to capture the
fugitive particulates, which will then be exhausted to a control system.
Therefore, no additional "Fugitive" control measures for particulate are
required.
Summary
This paper is an example of what can be accomplished in a complex source
(requiring a PSD Permit Application) when there is cooperation and under-
standing between industry and the regulatory agencies.
Ground rules were initially set by the agencies to allow for a lack of
well-defined process design information and final equipment selection
by accepting good engineering judgement with the option to make necessary
revisions when more data becomes available. This paper demonstrates what
can be accomplished when the environmental discipline is actively involved
in the initial stages of a project in an integrated effort with other
engineering disciplines to mutually define and solve environmental problems.
23
-------
FIGURE 1
SIMPLIFIED SRC-1 PROCESS FLOW DIAGRAM
-------
F \ G, O
POC
PCOC&SS
25
-------
NON-PLUGGING RETAINING STRUCTURE FOR GRANULAR BED FILTER
FOR HTHP APPLICATION
By: A. M. Presser, J. C. Alexander
EFB, Inc.
Woburn, MA 01801
ABSTRACT
Electrically augmented granular bed filters, modeled after commercial
Electrified Filter Bed (EFB) technology, are being considered as HTHP partic-
ulate control devices. High efficiency filtration in a moving granular bed is
achieved through the combined action of inertial impaction and electrical
image forces. Plugging of granule-retaining structures by buildups of sticky
dust, a fundamental problem encountered in HTHP gas cleanup is prevented by
employing large inlet louvers backed by a shallow "fast face" region. In this
region, granules are moved at a significantly faster rate than in the remain-
ing bulk of the bed, sweeping away dust accumulations.
An 84-hour test of an EFB collecting Burgess #10 pigment dispersed in
0.38 NnH/sec (800 scfm) of room temperature air at concentrations up to
4 gm/Nm3 showed no filter plugging. Dust buildups on superfluous metal struc-
tures, which might lead to problems in longer duration tests, were shown to be
eliminated in another set of similar tests
INTRODUCTION
The success of power generation systems utilizing direct fired turbines
with fluidized bed coal combustors may ride on the ability to prevent erosion
of gas turbine elements by hot particulate matter in the combustion gas stream.
Granular bed filters have been the focus of much investigation because of
their inherent resistance to degradation in hostile environments. However,
filter plugging problems caused by the sticky nature of flyash at HTHP condi-
tions have slowed development of these filters. The tests reported here
demonstrate a non-plugging granular bed filter design.
The plugging problems which must be surmounted can be illustrated by
previous work. The DUCON filter is a granular bed in which filtration is
effected by a cake of particulate which is allowed to build on the front face
of the granule bed. Periodically, the bed must be blown back to remove the
cake and keep the pressure drop within reasonable limits. In tests by
Exxon (1), the front cake built to a level where it could not be blown back
or, in other tests, removed by fluidizing the bed. This is an example of
front face plugging, where dust actually blinds the incident face of the
granular ged filter.
Combustion Power Co- (2) has tested a moving granular bed where, by
utilizing relatively large granules, filtration occurs in the interstices of
the bed. Here, however, the dust loaded the bed void regions to such an
extent that bed motion was inhibited, leading to unacceptable buildups. This
is termed bed freezing.
26
-------
Problem Approach
General Electric Company and EFB, Inc. are currently investigating a novel
approach to the HTHP cleaning problem. In the Electro-granular Bed (EGB) the
combined action ot inertial impaction and electrical image force provide for
dust filtration in the interstices of a bed composed of large granules. The
EGB is a direct extension of commercially available Electrified Filter Bed
(EFB) air pollution control eqiupment developed and marketed by EFB, Inc.
Several patented features of this technology are directly applicable to the
prevention of plugging problems. These features are described now with refer-
ence to Figure 1, a schematic of an EFB pilot system also used for part of the
current tests. The schematic represents a cross-sectional view of a circular
cylindrical unit in which two filter stages are vertically arranged. Inlet
dust laden gas passes first through the lower stage. It is on this incident
filter face, the outside of the lower tube, that the severest problems are
expected because much of the filtration occurs in this first filter. The
partially cleaned gas then travels upward in the hollow center where a corona
discharge emitted from the centrally hung high voltage electrode electrically
charges the dust. Finally, the gas exits through the upper filter stage in
which the granules are electrically polarized by an applied electric field in
order to collect the charged dust effeciently. In the EGB, the collection
force due to the applied field is replaced with the force due to the electri-
cal image of the charged particles.
On a continuous basis, the granules move downward through the beds with
granule speed being controlled by the screw feeder at the bottom. The granule-
dust mixture is separated, and the cleaned granules are transported to the top
of the unit by an exterior pneumatic conveying system.
In order to achieve long term, non-plugging performance, a retaining
structure for the granules has been developed which incorporates large
inclined slats of metal which hold granules in the bed by action of their
angle of repose, while providing a large open area where the dust laden gas
essentially impinges on a free surface of gravel. The continuous motion of
the gravel breaks up bridges of dust between granules and carries away any dust
buildups. This feature prevents the so-called front face plugging.
As dust deposits in the bulk of the bed, it accumulates most heavily in
the front regions of the bed. Excessive dust buildup in any region can
cause bed freezing. The EFB system incorporates a "fast face" region defined
by the inlet side louvers and a coarse mesh screen located about 1/4 of the
bed overall depth away from the front louvers. Granules in this region are
moved downward at a faster rate than in the remaining bed, thus removing the
greater accumulations of dust faster.
The standard EFB design needed to be modified to prevent accumulation of
dust on exposed portions of the bed inlet side louvers. This problem was
identified in the course of these tests and is illustrated in Figure 2. The
proposed solution, reducing louver area to the minimum necessary to retain
the granules, proved effective in a second set of tests performed in a special-
ly prepared model unit.
27
-------
Test Program Description
As part of the research program at EFB, Inc., cold flow experiments were
performed to provide data that could be used to project HTHP performance of
the EGB. The desired objectives of the test program were to demonstrate the
non-plugging design during a relatively long duration at nominal dust loading,
and to simulate an operational upset in which gravel motion is slowed and
observe the recovery capabilities of the design.
The 800 cfm pilot unit was run at room temperature and pressure and
Burgess #10 dust was dispersed as the sticky dust simulating HTHP flyash.
Details of the pilot unit design are shown in Table 1. Pressure drop across
each of the beds was monitored as a means of detecting plugging problems which
might develop. Also, visual observations were useful in spotting areas of
particulate buildup on support structures in the EGB.
The 84-hour (3y day) test was preceded by one and ten hour duration tests
to shake down the test equipment and procedures and to indicate problems which
might occur. Table 2 shows the relevant parameters for the 84-hour test.
During the run, the pressure drop across each bed was continuously monitored,
and inlet and outlet dust loadings were periodically monitored. To simulate
the image force that is expected to be present at HTHP conditions, an electric
field of 2x10^ v/m was imposed in the second stage. A field on the outlet bed
has no effect on the performance of the anti-plugging features. At the end
of 100 hours, the unit was visually inspected.
For the first 30 hours, the dust source was freshly dispersed Burgess #10
pigment. From 30 hours on, the dust source was a combination of freshly
dispersed pigment and pigment cleaned from the gravel and carried by the
exhaust pneumatic transport air.
A smaller filter with "reduced area louvers" was employed in another test
to demonstrate the resistence of this louver design to buildups on exposed
louver surfaces. This test was run at very high inlet dust loadings to
accentuate the problem. Only visual observation was necessary to verify
performance. Table 2 shows the test parameters.
Test Results
Test results from the 84-hour test are shown as the evolution of pressure
drop with time (Figure 3) and photographs of the louver regions after the test
(Figure k-1). Initially, pressure drop was relatively constant, with small
variations caused by changes in the dust feed rate and air volume flow through
the bed. After 42 hours, the granule downward motion was slowed, simulating
an operational upset. Pressure drop rose in the inlet bed due to the
increased loading of dust in that bed. At 52 hours, full bed motion was
restored and by 60 hours the inlet bed pressure drop had recovered to its
former level.
Figures 4 and 5 illustrate the state of the inlet face at the end of the
test. Both are views of the louvers from above, but Figure 4 is a tangential
view while Figure 5 is looking straight pn. One observes that there is a heavy
28
-------
growth of dust on the louvers, but open regions are visible and individual
granules are discernable with no bridges between granules. That is, the fast
face motion apparently destroys any bridges before they can act to freeze the
bed.
Figure 7 shows a level view of the louvers. The growth of dust struc-
tures, as described in Figure 2, is evident. To confirm the theory that this
growth would only occur on louver area that was not swept by granules, the
Reduced Area Louver Test was run.
The test filter of Figure 2d had shortened louvers so that there were no
extraneous metal surfaces where dust could accumulate. This filter was sub-
jected to a loading of dust of roughly 13 times the loading used in the 84-
hour test. Table 3 shows the parameters used in this test. The increased
dust loading accentuated deposition and as a result the test time was consid-
erably shortened. To carry off the large amount of dust, the bed granules
were also removed at an increased rate over normal.
Figure 8 shows the modified louvers before and after the test. During
the first two hours, dust accumulated on the louvers and built up a rounded
edge on the head of the louvers. After that, a steady state buildup appeared
to be reached so the test was terminated at the end of the third hour.
Figure 8b shows the louvers with some deposition, but none of the vertical
dust structures which were observed in the 84 hour test. Deposition appeared
limited to the 1/8" of metal at the top of the louver which was unswept by
granules. This provided too small a base for large dust structures to grow.
Conclusions
The EGB design, with large inlet louvers backed by a fast face region,
modified by eliminating exposed louver surfaces has proven effective in
eliminating all types of plugging problems. Long term tests with Burgess #10
pigment to simulate a "sticky" dust typical of HTHP flyash were successful in
maintenance of low filter pressure drop and in avoiding dust buildups on meta]
structures. As a result of these tests, work is progressing to demonstrate
this non-plugging design for filtration of flyash from high temperature
(1550°F) fluidized bed coal combustor effluent gas.
Acknowledgement
This work was supported by the Department of Energy under Contract
DE-SC21-79ET15490 to General Electric Company. The work was performed by
EFB, Inc. under subsontract to General Electric.
End Notes
1. Hoke, R. C. and Gregory, M. W. Evaluation of a Granular Bed Filter for
Particulate Control in Fluidized Bed Combustion. (Presented at EPA/DOE
Symposium on High Temperature, High Pressure Particulate Control,
Washington, D.C. Sept. 20-21, 1977.
2. Wade, G. L. Performance and Modeling of Moving Granular Bed Filters.
29
-------
(Presented at EPA/DOE Symposium on High Temperature, High Pressure
Particulate Control, Washington, B.C. Sept. 20-21, 1977).
TABLE I. EGB DESIGN AND OPERATING PARAMETERS
Bed Depth
Bed Outer Diameter
Bed Inner Diameter
Active Bed Face Area
Fast Face Depth
Media Diameter
Gas Face Velocity
Lift Pipe Air Velocity
Charger
Louver Vertical Spacing
Louver Angle
Ratio Fast Face Velocity:
Bulk Velocity
Media Feed Rate
10 cm (4 in.)
60 cm (24 in.)
40 cm (24 in.)
1.1 m2(12.6 ft2)
2.5 cm (1 in.)
2 mm
0.3 m/s (60 ft/min)
40 m/s (120 ft/s) .j^
Sharpened square rod (-jin.)
10 cm (4 in.)
60° from horizontal
300 pounds/hour
TABLE 2. 84 HOUR TEST PARAMETERS
Injected Dust
Granular Bed Media
Inlet Gas Flow
Fast Face Media Velocity
Charger Voltage
Bed Applied Field
Media Feed Rate
Time into test
18 hours
38
72
Burgess No.10 pigment
NJ #2 Sand (2-3 mm. diameter)
800 acfm
5.8 ft/hr
45 KV
2xl05 v/m
300 pounds/hr
Inlet dust loading
4.0 grams/m3
2.6
1.8
TABLE 3. MODIFIED LOUVER TEST PARAMETERS
Bed Depth
Active Bed Face Area
Fast Face Depth
Media Diameter
Face Velocity
Louver Spacing
Louver Angle
Average Dust Loading
Test Duration
10 cm (4 in.)
0.14 m2 (1.5 ft2)
2.5 cm (1 in.)
4 mm
0.3 m/s (60 ft/min)
10 cm (4 in.)
60° from horizontal
20 g/m3
3 hours
30
-------
Dust/gravel
Separation
Chamber
Pneumatic
Transport
Line
Inlet
Sampling
Dust
Feeder
Screw
Feeder
Transport
Air
Figure 1. 800 cfm Pilot Unit Schematic
31
-------
(a)
(b)
(c)
(d)
Figure 2. Deposition Problem and Solution
(a) Impaction on louver tips creates small blockages
which accentuate impaction and dropout.
(b) Dropout behind blockage build a thick base for
dust growths.
(c) The thick base provides support for more growth
of these blockages.
(d) Proposed solution to this problem is staggered louvers
so there is little bare metal for structures to build on.
32
-------
Pressure
Drop
(in. w.g.)
Lower bed
upper bed
10
20
30
40
50 60
70
80
90
Time From Test Start
(hours)
Figure 3. Bed Pressure Drop During The 84-hour Test.
33
-------
Figure 4
Figure 5
Figure 6
Figures 4-7.
Figure 7
Photographs of the Inlet Louvers of 800 cfm pilot unit
after 84-hour test.
34
-------
(b)
Photographs of 14" x 18" Reduced Area Louvers Model
(a) Before Test.
(b) After 3-Hour Test.
35
-------
PARTICULATE EMISSIONS CONTROL FROM A COAL-FIRED OPEN-CYCLE
MAGNETOHYDRODYNAMICS/STEAM POWER PLANT*
By: Hsuan-Hsien Wang and Thomas E. Dowdy
The University of Tennessee Space Institute
Energy Conversion Division
Tullahoma, TN 37388
ABSTRACT
A coal-fired open-cycle magnetohydrodynamics (MHD)/steam power plant
will generate participates that differ significantly from those produced in
conventional coal combustion. Potassium carbonate is added to the extremely
hot combustion gases (approximately 3000K) to increase gas electrical con-
ductivity and to capture the sulfur as potassium sulfate. This material,
along with coal ash constituents that have been melted and/or vaporized and
recondensed, form the particulates that must be collected. The size distri-
bution of these particles, along with chemical composition and dust loading
must be experimentally verified. A test facility designed to fire 0.9 Kg/s
of coal is under construction at the University of Tennessee Space Institute
(UTSI) near Tullahoma, Tennessee. In order to specify and design the most
appropriate particulate collection system for the Coal Fired Flow Facility
(CFFF), a review has been made of available information concerning MHD pro-
cess particulates and high efficiency particle collection devices. High
energy wet scrubbers, electrostatic precipitators and fabric filters, with
and without agglomerating pretreatment, were considered.
INTRODUCTION
Magnetohydrodynamics (MHD) electric power generation is based on the
direct conversion of thermal energy to electric energy by passing a high
temperature (approximately 3000K), high velocity (approximately 1000 m/s)
combustion gas through a magnetic field. The gas is made electrically
conducting by adding a seed, typically potassium. The principle is similar
to that of a conventional turbine generator system, the difference being
that the rotating conductors of a turbo generator are replaced by a par-
tially ionized combustion gas which interacts with the magnetic field. The
interaction of the high velocity conducting fluid with the intense trans-
*This work was sponsored by the Department of Energy under Contract No.
DE-AC02-79ET10815.
36
-------
verse magnetic field induces an electric field within the fluid. Electrodes
convey the electricity to an inverter where the direct current power pro-
duced is transformed to alternating current which can be transmitted
directly through an electric power grid. The gases leaving the MHD channel
pass through a steam bottoming plant which produces additional power by
means of a conventional steam turbine generator.
The University of Tennessee Space Institute under Department of Energy
(DOE) sponsorship is currently developing and constructing a Coal-Fired Flow
Facility as part of the overall DOE effort to advance coal-fired MHD
electric power generation. The conceptual arrangement of the CFFF is shown
in Figure 1. The CFFF flow train consists of twelve components including
vitiation heater, coal-fired combustor, nozzle, diagnostic channel or MHD
generator channel, power conditioner, diffuser, radiant slagging furnace,
secondary combustor, superheater, air heater and particulate collector. The
design goal for all components in the CFFF, is the demonstration of pre-
dicted system characteristics required of an MHD system (e.g. life, perfor-
mance, reliability, failure mode, etc.) so that the data obtained can be
used as the technological data base for future scale-up of a coal-fired
open-cycle MHD/steam power plant.
The vitiation heater reacts mixtures of oxygen and nitrogen with No. 2
fuel oil to produce a gas at an elevated temperature having the required N/0
ratio for the coal-fired combustor. In the coal-fired combustor vitiated
oxidizer reacts with pulverized coal and its carrier gas. The coal is mixed
with a seed material, potassium carbonate and/or potassium sulfate, and
burned substoichiometrically at high temperatures to produce an ionized gas
suitable for MHD power generation. The nozzle accelerates the gas from the
combustor and serves as a geometric transition between the coal-fired com-
bustor and the channels. The diagnostic channel is a Hall device which
measures plasma average electrical conductivity. The generator channel con-
verts some of the thermal energy released from the coal to electrical
energy. The power conditioner is designed to be compatible with the
generator/diagnostic channel operating characteristics so as to operate
during both steady state and transient periods. The diffuser recovers the
static pressure from the generator/diagnostic channel(s) to a value near
atmospheric pressure and reduces the velocity of the plasma to values accep-
table for the downstream components. These seven components make up the MHD
topping cycle.(1)
The radiant slagging furnace rejects the major amount of entering slag
from the gas stream and removes the slag to a transfer point while retaining
the seed in a gaseous state. The radiant slagging furnace absorbs heat from
the gas stream, flashes heated water to steam, vents the steam to
atmosphere, and further reduces the velocity of the gas stream. The radiant
slagging furnace has sufficient gas residence time which, along with proper
temperature reduction, promotes the decomposition of NOX. The secondary
combustor mixes air, either at ambient temperature or preheated by the air
heater, with the gas stream to complete combustion of the gases. Excess air
is provided to complete combustion while maintaining NOX levels below
Environmental Protection Agency (EPA) utility limits. The superheater simu-
37
-------
lates the operating conditions in a large scale superheater by duplicating
temperatures of the bulk gas and metal temperature of the tubes. The air
heater provides heated air to the secondary combustor. The particulate
collector ensures that particulate emissions do not exceed EPA limits and
demonstrates adequate particulate collection performance under MHD operating
conditions. The above five components belong to the steam bottoming
cycle. (1 )
The design for the air heater and superheater, which are components
upstream of the particulate collector in the CFFF, have not been completed.
In addition, there is limited prior experience with the seed/slag chemistry,
slag/seed separation, slag/gas separation, strong magnetic field
(approximately 3 Tesla), high combustion temperature (approximately 3000K),
flow pattern and heat transfer characteristics of the MHD process. Thus,
information about the inlet particle-laden gas stream of the particulate
collection system of the CFFF is incomplete. This information must be
experimentally verified. This is a major objective of the CFFF Low Mass
Flow (LMF) test train in the future.
The MHD particulates are significantly different from particulates
generated from the conventional coal -fired power plant. These particulates
must be collected efficiently so that the coal -fired open-cycle MHD/steam
power plant will be environmentally as well as economically acceptable. The
purpose of this paper is to discuss available information concerning MHD
particulates and to evaluate the feasibility of applying commercially
available fine particulate collection technology to the CFFF.
MHD PARTICULATES
Several factors unique to a coal -fired open-cycle MHD/steam power plant
influence the generation of particulates. The coal -fired combustor operates
at temperatures sufficiently high (approximately 3000K) that a major portion
of the mineral content of the coal will be vaporized. This slag will begin
to condense and solidify in the cooler downstream components to form small
particulates. The greatest fraction of these particles will be removed in
the radiant slagging furnace. These slag particulates will have a different
chemical composition as compared with the slag generated in conventional
coal -fired power plants because of different gas characteristics and slag
condensation as well as solidification environments.
The seed material, potassium carbonate (1^03), is added to the coal-
fired combustor to increase the gas electrical conductivity and to capture
the sulfur present in the coal as potassium sulfate (KgSC^). This material
solidifies at temperatures around 1342K in the cooler downstream components
primarily in the air heater and superheater to form predominately submicron
particulates.
There is some experimental evidence obtained in small-scale facilities
which indicates that the strong magnetic field (approximately 2 Tesla) in
MHD generators may influence the formation of particulates. Apparently this
results in the formation of a larger number concentration of submicron
38
-------
particulates than is obtained without the field.(2) The magnetic field may
suppress turbulence and prevent nucleation of new particulates until high
supersaturation ratios are reached. In addition, when thermal energy is
extracted from the MHD channel, the gas cools more rapidly as it travels
through the channel, which directly alters the particulate size
distribution.(2) Consequently, both of these factors will contribute to
decrease the time available for agglomerating smaller particulates into
larger ones.
The entrained coal slag particulates may serve as condensation nuclei
for the seed material when the combustion gas is cooled below the dew point
of K2S04 of approximately 1943K. Tempelmeyer, et al compared Scanning
Electron Microscope (SEM) micrographs of fly ash from a conventional power
plant with that obtained from an MHD system.(3) They observed a coating of
K2S04 on the surface of the MHD particulates, illustrating that the
entrained coal slag particulates serve as nuclei upon which K2S04 vapor con-
denses.
The condensation of seed will cause the particulates to grow into the
size range of 0.3 to 1.0 microns, depending on the number of coal slag par-
ticulates and the amount of I<2S04.(4) The MHD particulate size is projected
in the submicron range. Im, et al estimated that the mean size of MHD par-
ticulates is in the range of 0.2 to 0.5 microns.(4) Some USSR data from
the U-02 facility shows a mass mean diameter of 0.2 microns.(5) It should
be noted that this facility fires clean fuel and injects ash. Researchers
at UTSI have measured and counted the particulate matter taken from the bag
filter of their Energy Conversion Facility (ECF) which has no magnet on the
test leg. The result of this mesured MHD particulate size distribution is
shown in Figure 2. In this figure, the estimated particulate size distribu-
tion of MHD particulates and the measured particulate size distribution
typical of fly ash from a conventional coal-fired power plant(6) is also
presented.
The MHD particulates size is projected to be one to two orders of magni-
tude smaller than fly ash particulates generated from conventional coal-
fired power plants. In addition, the chemical composition of MHD
particulates is quite different from the conventional fly ash particulates.
Furthermore, the electrical resistivity of particulates is strongly depen-
dent on their chemical composition and the characteristics of their carrier
gas stream. At present, it is not known how the high concentration of
K2S04 will affect the MHD particulate electrical resistivity. Researchers
at UTSI have measured the electrical resistivity of the MHD particulate
taken from their ECF. It is ranged from 3.8 to 4.3 x 1010 ohm-cm. Based on
this information, the inlet gas stream characteristics to the particulate
collector of the CFFF has been estimated so that evaluation of commercially
available fine particulate collectors is possible. The estimated charac-
teristics of the gas stream are shown as follows:
Pressure: 0.88 atm
Temperature: 41 IK
Total mass flow rate: 5.88 Kg/s
39
-------
Volumetric flow rate of gas:
Dust loading:
Gas composition:
02
N2
C02
H20
NOX
so¥
7.13 m3/s
5.75 g/m3 (average)
41.63 g/m3 (maximum)
Volume Fraction
0.04530
0.52179
0.29026
0.14186
0.00069
0.00010
1.00000
3.8 to 4.3 x
2.66 g/cm3
See Figure 2
1010 ohm-cm
Particulate electrical resistivity:
Particulate density:
Particulate size distribution:
Particulate chemical composition:
Weight Fraction
K2S04 0.939
K2S03 0.004
K2S 0.001
K2C03 0.050
Fly ash 0.006
T7TJOD"
PARTICULATE COLLECTORS
High-energy wet scrubbers, electrostatic precipitators and fabric
filters are commercially available fine particulate collectors. There are
also many agglomerating methods which can be used as a gas preconditioner to
alter the size distribution of particulates so that it is easier to collect
them in the major collector. At present, sonic agglomeration seems the most
appropriate compared with space-charge agglomeration, magnetic agglomera-
tion, turbulent agglomeration, thermal agglomeration, and radiation agglo-
meration based on the theoretical and experimental background, although
there is much uncertainty in sonic field generation with respect to wave
form, power consumption and the coagulation rate.(7)
Typical features of fabric filters (FF), electrostatic precipitators
(ESP), wet scrubbers (WS) and sonic agglomerators (SA) are shown in Table 1.
Wet scrubbers are considered because of the high water solubility of K2S04.
However, high energy consumption, insufficient grade mass collection effi-
ciency over the submicron range, scaling and fouling, and low operational
reliability ruled them out as a major particulate collector for the CFFF.
Electrostatic precipitators are considered due to their wide use in
industry. However, insufficient grade mass collection efficiency over the
submicron range and the considerable difficulties achieving a 10 percent
opacity may force them to connect with a preconditioner to meet the par-
ticulate emission standards. Fabric filters are of interest because of
their high grade mass collection efficiency over the submicron range. In
addition, they are the only type of particulate collector that can perform
40
-------
adequately under cyclic loading. Furthermore, the expected promulgation of
a fine particulate emission standard of 0.02 lb/106 Btu during the 1980's
and the uncertainties as well as variabilities of coal source which the
utility industry may face in the future, make fabric filters more attractive
for collecting MHD particulates.
Researchers in Electric Power Research Institute (EPRI) have conducted
research on the economics of fabric filters versus electrostatic precipita-
tors for particulate collection.(8) They concluded that the trend toward
increasingly stricter particulate emission standards has raised the cost for
electrostatic precipitators such that fabric filters have become cost com-
petitive. (8) The relative economic feasibility of a well designed fabric
filter and sonic/electrostatic precipitator for the CFFF have not been done.
At present, we believe the fabric filter approach may be the less
expensive.
Although we presently favor the fabric filter, it is not without
problems. The MHD particulates are fine, hygroscopic and sticky. They have
the potential to blind a fabric filter. The performance of the fabric
filter in collecting MHD particulates must be experimentally verified.
The performance requirements of the particulate collection system for
the CFFF is shown as follows:
Durability: 104 hours
Particulate mass collection efficiency: 99 percent
Exit temperature of gas: 41 IK
Exit mass flow rate of gas from the collector: 5.88 Kg/s
Particulate emission standard: 0.03 lb/106 Btu (i.e. 13 ng/J)
Maximum allowable exit mass flow rate of particulates: 0.36 g/s
Maximum allowable particulate concentration: 0.038 grains/SCF
Most of the MHD particulates are removed in the radiant slagging fur-
nace. Only a small mass fraction of particulates entrained in the gas
stream go to the particulate collection system of the CFFF. In order to
meet the seed recovery and particulate emission goal of the CFFF, the choice
of particulate collection system is strongly dependent on the performance of
slag rejection of the radiant slagging furnace.
Babcock and Wilcox under DOE contract will provide a heat recovery/seed
recovery (HRSR) system to the CFFF to be tested by UTSI. The particulate
collector of the B&W HRSR is an electrostatic precipitator. We currently
plan to test both the B&W's ESP and a slip stream fabric filter.
SUMMARY
The fabric filter approach for the seed recovery and particulate
emissions control from MHD systems may be the system of choice. However,
there is uncertainty about the slag rejection capability which may force us
a change in the performance requirements of the particulate collection
system of the CFFF. Furthermore, there is potential of blinding the fabric
41
-------
filter by the fine, hygroscopic and sticky MHD participates. Since a recom-
mendation has been made to utilize the electrostatic precipitator from
Babcock & Mil cox heat recovery/seed recovery contract for the CFFF, we have
also elected to test a slip stream fabric filter.
ACKNOWLEDGEMENTS
The efforts of many individuals in the Energy Conversion Division of the
University of Tennessee Space Institute were required to perform the work
which is described in this paper.
ENDNOTES
1. Strom, S. S., M. H. Scott, and R. C. Attig. Overall Test Plan for the
Low Mass Flow Test Train at the Coal-Fired Flow Facility. The
University of Tennessee Space Institute, Tullahoma, Tennessee, September
1980.
2. Tempelmeyer, K. E., et al. Investigation of Factors Influencing
Potassium Seed Recovery in Direct Coal-Fired Generator Systems.
Proceedings of the 16th Symposium on Engineering Aspects of MHD,
University of Pittsburgh, April 1977.
3. Tempelmeyer, K. E., et al. Recent Experimental Studies of the
Interaction of Potassium Seed with Coal Slag in a Direct Coal-Fired MHD
Generator. Proceedings of the 15th Symposium on Engineering Aspects of
MHD, University of Pennsylvania, May 1976.
4. Im, K. H., J. Patten, T. Johnson, and K. Tempelmeyer. Condensation and
Deposition of Seed in the MHD Bottoming Plant. Proceedings of the 18th
Symposium on Engineering Aspects of MHD, Butte, Montana, June 1979.
5. Selby, R. C. Department of Energy, Chicago Operations and Regional
Office to J. Epstein. Letter communication, March 28, 1980.
6. Cheng, R. J. Characteristics of Particulates from Power Plants.
Journal of the Air Pollution Control Association, Vol. 26, No. 8, August
1976. p. 787.
7. Wang, H. H. An Evaluation of Various Particle Agglomeration Methods for
Fine Particle Control in Low Mass Flow Coal-Fired Flow Facility System,
Internal Report. The University of Tennessee Space Institute,
Tullahoma, Tennessee, August 1980.
8. Severson, S. D., R. W. Scheck, K. S. Campbell, and F. A. Homey. The
Economics of Fabric Filters and Precipitators. Chemical Engineering
Progress, January 1980. p. 68.
42
-------
TABLE 1. FEATURES OF FINE PARTICULATE COLLECTORS
Features
Specific
Energy
Consumption
Pressure Drop
Grade Mass
Efficiency for
Submicron
Particulate
Maximum
Operating
Temperature
Effluent
Opacity
Effluent
Particulate
Loading
Electrostatic
Fabric Filters (FF) Precipitators (ESP)
Wet Scrubbers (WS)
1 to 3 hp/103 acfm 1 to 1.5 hp/103 acfm 10 to 20 hp/103 acfm
0.5 to 10 in. w.g.
> 99%
560K
Q%
< 0.02 gr/SCF
0.5 to 1.5 in. w.g.
95% to 99% for hot-
side ESP
90% to 95% for cold-
side ESP
700K
10% to 20%
< 0.04 gr/SCF
60 to 100 in. w.g.
90% to 99%
10% to 20%
< 0.07 gr/SCF
Sonic
Agglomerators (SA)
1.5 to 46 hp/103 acfm
for traveling wave SA
0.5 to 2 hp/103 acfm
for standing-wave SA
(only used as a
precollector)
-------
RADIANT
SLAGGING
FURNACE
AIR HEATER
CHANNEL
NOZZLE
POWER CONDITIONER
SECONDARY
OMUUSTO
ELECTROSTATIC
PRECIPiTATOR
SUPERHEATER
DOWNSTREAM
COAL FIRED
COMBUSTOR
VITIATION
HEATER
FIGURE 1
Conceptual Arrangement of MI-ID System Including Steam Bottoming Plant
-------
FIGURE 2. Particulate Size Distribution of the MHD Particulates
and the Participates Generated from the Conventional Coal-Fired
Power Plant.
i i i i i i I i i i i i i i 11 i i i i i i i i I i i
i i i i i
o>
N
-------
REAL TIME COARSE PARTICLE MASS MEASUREMENTS IN A HIGH TEMPERATURE
AND PRESSURE COAL GASIFIER PROCESS TREATMENT
By: J. Wegrzyn, J. Saunders, and W. Marlow
Brookhaven National Laboratory
Department of Energy & Environment
Upton, New York 11973
ABSTRACT
A sampling system has been designed and is under construction at
Brookhaven National Laboratory that employs a probe appropriate for direct
extracted sampling of erosive range particulate matter from a coal gasifier
outlet or a high pressure fluidized bed combustor. The sampling train is
scheduled to be tested at the Morgantown Energy Technology Center on their 42
inch coal gasifier unit. This sampling system consists of four modules: 1) a
null balance extractive probe with injection through a porous lined tube to
minimize wall loss, 2) a stem type virtual impactor to separate coarse from
fine particles, 3) a filter tape collector, and 4) a Beta gauge total mass
detector. The key design feature of this system is a stem type virtual
impactor which separates at ambient gas stream conditions the coarse particles
from the sampling stream so that at upon filtration no condensible vapors,
fine particles or reactive gases pass through the filter tape. This system
should provide coarse particle mass flux data with a time resolution of thirty
seconds or better.
INTRODUCTION
Erosion by particles greater than 2 microns of turbines located in post
combustion gas streams from coal fired combustors is currently one of the
major obstacles to the increased utilization of coal for direct power
generation either with or without gasification. Various sampling methods
exist and others are under development for the measurement of these erosive
particles, but there is currently no accepted approach for the measurement of
post combustion stream, particulate mass content to assert the effectiveness
of the cleanup devices. Measurement of the particulate mass flux in the
erosive size range, as it exists under the appropriate high temperature
pressure conditions and in the presence of high partial pressures of
condensible vapors, is needed to provide the necessary information on the
effectiveness of the cleanup devices. The effort at Brookhaven National
Laboratory is towards the construction of a general purpose sampling train to
determine the effectiveness of the various air cleaning devices in either a
coal combustor or gasifier. The validity of aspiration sampling is well
established for both laboratory^ and stack applications (EPA Method-5).
However, care must be taken when employing these techniques to the problems of
sampling a high temperature and pressure environment where there is a presence
of large amounts of condensible vapors. Since condensation, evaporation, and
additional chemical reactions within the sampling train can alter the
46
-------
particulate size, density, and composition. The objectives of this work are
to 1) maintain sample integrity by perserving the representativeness of the
particulate sample 2) produce readily interpretable data by measuring mass
'directly 3) minimize and measure any errors associated with extractive
sampling.
The following sections of this paper details some of the critical design
features of the Brookhaven National Laboratory sampling train and further
discusses some of the major problems associated with aspiration sampling. A
brief theoretical discussion is also being presented to facilitate in
discussing some of the inherent problems in aspiration sampling.
Figure 1 illustrates the distortion in measurement from sampling at a
higher then free stream velocity. Particles whose stopping distance X p are
larger than the length of the disturbance of stream L will penetrate past the
probe, while these with A p
-------
T7
V0 PROCESS
STREAM
Figure 1 - Particle Trajectories for Sampling (V > VQ) °
48
-------
Figure 2 - Sampling Train Module Components
49
-------
DESCRIPTION
Sampling requirements of particle integrity, artifacts of measurements,
and validity of signal for high temperature, high pressure aerosols have to be
addressed if meaningful particulate measurements are to be made on a
combustion or gasifier process stream. These problems as applicable to the
Brookhaven National Laboratory probe have been addressed by W. Marlovr1' and
will not be reviewed here.
The sampling train as it is being built at Brookhaven National Laboratory
is constructed of sections of schedule 40 304 stainless steel pipe, that are
connected by either 1 inch or 2 inch ASME rated 300 psi flanges. The hinge
cap closures on the tape reels and particle deposition chamber also have a 300
psi rating. Thermocouples and pressure transducers are placed along the
sampling train and cooling coils to monitor the system. If any of these
signals exceeds the safety limit the probe is programmed to go to a safety
mode by closing the main valves and purging the system with dry nitrogen. The
hinge cap closures are also equiped with safety locks which prohibit opening
if that part of the train is under high pressure. The electronics, flow
meters, secondary valves, and data logger are all included in a heated
cabinet, not shown in this drawing, which protects them from the environment.
Figure 2 is a box diagram of the sampling train. The train has been
designed in module form for two reasons. First, a module design permits
tailoring the train for the problems that may exist at the specific sampling
site. Secondly, any deficient module can be easily replaced without major
redesigning of the whole probe.
An ideal sampling head consists of a flow sensor mounted near the
entrance of thin wall knife edge gas skimmer. The hostile environment of the
process stream eliminates the use of most flow sensors to set the isokinetic
condition and only null balance static pressure taps have been employed
successfully. This system balances the static pressure inside and outside of
the sampling head. However, the thick walls required to mount the static taps
result in a flow disturbance that causes a pressure difference between the two
taps even under isokinetic condition. Brookhaven National Laboratory is
currently studying this error as well as exploring alternative isokinetic
sensors.
More than 75% of the particulate matter can be lost in the walls of a
poorly designed sampling train.3 To overcome this problem, work is being done
at Brookhaven National Laboratory on the use of a porous lined sampling tube.
Initial cold tests of preventing wall deposition in a porous tube by the
injection through the porous tube of a secondary gas stream have been
encouraging. These tests however were performed on the straight section of
tubing and further tests are planned to assert whether this technique will
also prevent particle deposition in a right angle bend. The major
disadvantages of using porous lined sampling trains are flow control and flow
measurement.
50
-------
Figure 3 shows a stem type virtual impactor. The scaling of this device
follows the theoritical restraints as put forth by V. A. Marple^ for the
overall design of a virtual impactor, and its concentric flow pattern is
similiar to a virtual impactor built and tested by the H. Masuda6. The
separator as seen in Figure 3 is built within a 300 psia s.s. orifice flange.
This unique design has the capability of separating at high temperatures and
pressures the coarse particles from the process stream, as nitrogen gas is
leaked through the conical tip of the stem. This secondary flow of gas helps
to divide the process stream into concentric flows, and this flow is
accelerated to about a third of a Mach by the converging nozzle which is
directed at a knife edge skimmer. The inner most part of this flow, that is
the nitrogen which came from the stem tip and the coarse particles whose
inertia carried them into this region, is then separated off into the lower
section while the remaining stream consisting of condensible vapors, reactive
gases, and fine particles, is forced to reverse its direction and is removed
by the tee connection located above the separator. Once the coarse particles
are removed from the process stream they can be cooled down and collected for
analysis.
The coarse particles are filtered from the gas in the deposition chamber
by a continuously fed filter tape. This tape runs from the supply reel to the
take up reel and consists of a one inch wide filter material of glass
microfibers supported by a glass cloth. The length of the tape is 250 feet
and it has a temperature limitation of 1200°F. The hinge pipe cap closure
permits easy access and maintenance of the filtering region. Non-metallic
seals are used, however, which necessitates selectively cooling the seals to
temperatures below 500°F.
Figure 4 shows construction of the beta gauge. Beta gauges have long
been recognized as a preferred solution to the, problem of automated mass
measurement of particulate matter in flue gases.1 The theory and operation of
Beta gauges for particulate mass measurement is given in a paper by P.
Lilienfeld.8 The BNL gauge consists of a New England Nuclear carbon-14 beta
source, a Nuclear Enterprize NE-102 plastic scintillator which converts the
beta particles into a light pulse and a EMI model 9824B photo multiplier tube
which detects the light pulses. An analysis of the response data for the
NE-102 scintillator for .1 to 1 MeV Beta particles has been given by R. L.
Craun.^ The attenuation of the light pulses is directly proportional to the
amount of mass that is present between the source and the scintillator. A
reading of the particulate mass on the tape is achieved by recording the
attenuation of the beta particles and subtracting out the attenuation due to
the mass of the filter tape and the surrounding gases. The combination of a
beta sensitive scintillator and a PM tube is used instead of the conventional
Geiger counter because of its superior frequency response. This permits
higher counting rates which results in lower statistical counting error.
51
-------
Figure 3 - Virtual Stem Type Impactor
52
-------
H.V.
TAPE
SIGNAL
PM TUBE
C-14
TAPE
NE-102
Figure 4 - Beta Gauge Mass Detector
53
-------
CONCLUSION
Brookhaven National Laboratory is constructing a sampling train for the
measurement of the total particulate mass in a high temperature and pressure
coal stream. The probe attemps to maintain integrity of the sample and has
high mass loading applications. The train is designed in module form so that
it can be tailored to the particular needs of the process stream. The
accuracy of the system and its components, including sampling, transport, and
detection have to be determined. Accurate measurement of the aspiration
velocity and setting of the isokinetic conditions appears to be a problem.
REFERENCES
1. Balyaev, S. P. and L. M. Levin. J. Aerosol Sci., Vol. 5, p.325 (1974).
2. Hawksley, P. G. W., S. Badzioch, and J. H. Blackett. Measurement of
Solids in FLue Gases. The Institute of Fuel London, (1977).
3. Bajura, R. A., et al. Morgantown Energy Technology Cneter, Project
DE-AT21-78MC07087/TO-36.
4. Marlow, W. H. and N. Abuaf. The Proceedings of the 1978 Symposium on
Instrumentatin and Control for Fossil Demonstration Plants, Newport Beach,
CA, June 19-20, 1978. ANL-78-62
5. Marple, V. A. and C. M. Chien. Env. Sci. Tech. Vol.14, No. 8, p. 976,
August 1980.
6. Masuda, H. D. Hoehrainer, and W. Stober. J. Aerosol Sci., Vol 10, p.275
(1979).
7. Sem, G. J., et al. "State of the Art: 1971 Instrumentation for
Measurement of Particulate Emissions from Combustion Sources", EPA
Contract No. CPA-70-23 (1971).
8. Lilienfeld, P. Staub-Reinhalt Luft, Vol. 35, No. 12, p.458 (1975).
9. Craun, R. L. and D. L. Smith. Nuclear Instruments and Methods, Vol. 80,
p. 239 (1970).
ACKNOWLEDGEMENT
The authors are pleased to acknowledge the support of this work by the
Department of Energy under Contract No. DE-AC02-76CH00016.
54
-------
THE DESIGN, ENGINEERING, AND STARTUP OF A VENTURI SCRUBBER SYSTEM
ON AN OIL SHALE OFF-GAS INCINERATOR
By: P. A. Czuchra
FMC Corporation
Air Quality Control Operation
1800 FMC Drive West
Itasca, Illinois 60143
J. S. Sterrett
Rio Blanco Oil Shale Company
Rio Blanco, Colorado
Figure 1 - Rio Blanco Job Site
55
-------
ABSTRACT
This paper describes the design, engineering and startup of a venturi
scrubber removing particulate matter and sulfur dioxide from an oil shale
retort off-gas incinerator. Since oil shale recovery is a new technology,
a scarcity of data exists concerning conditions to be expected going to
the scrubber. Therefore, certain assumptions concerning design conditions
had to be made. The assumptions made to develop the design criteria for
the scrubber system will be discussed along with the preliminary results
concerning the accuracy of these assumptions.
TEXT
Background Information
Rio Blanco Oil Shale Company is a general partnership consisting of
Standard Oil Company of Indiana and Gulf Oil Corporation. Together they
are testing and developing a process in Rio Blanco County, Colorado,
which will extract oil shale (kerogen) from the shale.
The process currently being tested is a modified in-situ process.
In this process, the shale is fragmented by use of explosive charges,
and a portion of the fragmented shale is mined out to create voids which
enhance flow through the retort. The retort is ignited and heat from
partial combustion is used to evaporate the kerogen from the rock. The
oil then flows out the bottom of the retort and is pumped to the surface.
The off-gas from this process is sent to an incinerator where the gaseous
sulfur compounds formed in the process are converted to sulfur dioxide
This is shown in figures 2 and 3.
(so2).
MIS RETORT SCHEMATIC
Air & Steam
150°C (300°F)
Shale Combustion
750'-970°C
(1400'-1800°F)
Shale Retorting
370°-480°C
(700*-900°F)
Figure 2
Shale Cooling
150--750°C
(300°.1400°F)
Shale Preheating
480° 750°C
(900°-1400°F)
Oil Condensing
70°-370*C
(160°-.700°F)
Oil & Gas
70°C
<160°F)
56
-------
Ventilation
exhaust
Air
intake
Figure 3 - Modified In-Situ Process Pictorial
The gas leaves the incinerator and is subjected to a series of
water sprays prior to entering the scrubber. The water sprays cool the
gas from incinerator fire box temperature to approximately 260°C (500°F).
The gas from the incinerator also contains particulate matter and hence
would be out of compliance with air quality regulations on both S0~ and
particulate emissions. For this reason, it was decided to select a
venturi scrubber to simultaneously collect the SCL and particulate
matter which is generated. A venturi scrubber is a flexible device. It
can yield higher particulate removal efficiencies by simply increasing
the pressure drop.
Equipment Description
In March of 1978, FMC Corporation received a contract to design and
furnish a scrubber system for Rio Blanco. The heart of the system is an
FMC Dual Throat Venturi Scrubber. The scrubber, as shown in figure 4,
incorporates a rectangular housing with a wedge-shaped insert. The insert
moves up and down to change the cross-sectional area of the venturi throat.
This allows the venturi to maintain a constant pressure drop with varying
gas flow rates or to have the ability to change the pressure drop at constant
gas flows. The geometry of rectangular housing and wedge-shaped insert
allows a true venturi configuration of a constant angle inlet, a throat
of constant cross-sectional area, and a constant angle of divergence in the
expander section to be maintained throughout the turndown range of the
scrubber.
57
-------
SPRAY NOZZLES
TEFLON INSERT GUIDES .
INSERT
FLOODED ELBOW
Figure 4 - FMC Dual Throat Venturi
The insert is raised or lowered by a hydraulic cylinder at the bottom
of the unit. Teflon guides mounted on the sides of the housing prevent the
insert from fluttering in the gas stream while providing a smooth surface
for easy movement of the insert.
Liquid is introduced to the scrubber by large-diameter, low-pressure-
drop spray nozzles. The nozzles provide good water coverage to the throat
while minimizing the chances of pluggage. The nozzles are recessed out of
direct contact with the gas stream, thus diminishing the possibility of
material buildup from a wet/dry interface. The venturi is approximately
6.7 meters (22 feet) high. The coarse separation of the water droplets from
the gas stream is accomplished by a cyclonic entrainment separator. Gas
58
-------
from the venturi is directed into the separator tangentially. The water
droplets impact into the side of the separator and drain down into an
integral recirculation tank. Additional gas/liquid separation is accomp-
lished by a chevron mist eliminator located near the top of the separator.
The separator is approximately 14.6 meters (48 feet) high and 6.7 meters
(22 feet) in diameter. The venturi is constructed of Incoloy 825 and the
separator and mist eliminator of 316L stainless steel. The venturi and
separator are shown together in figure 5.
Figure 5 - FMC Dual Throat Venturi and Separator at Rio Blanco
Due to the potential abrasiveness of the particulate being collected
and the corrosive nature of the recirculation liquor, rubber-lined recir-
culation pumps were chosen for the system. Also due to the conditions
described above, FMC recommended the use of either abrasion resistant
FRP or rubber-lined cast iron for the recirculation piping. It was decided
to install the rubber-lined pipe. The neutralization tank was constructed
of corrosion-resistant FRP. All piping and tanks associated with the recir-
culation system are well insulated to prevent any possibility of freezing
the scrubber and soda ash solutions. This was necessary due to local
weather conditions.
The soda ash storage tank is a unique feature of the system. Soda
ash is pneumatically conveyed into the tank through a device called a
Slur-0-Lyzer®. The device contains a series of spray nozzles which wet the
soda ash as it is blown in. Some of the wetted soda ash forms a monohydrate
crystal bed at the bottom of the tank while the rest goes into a saturated
solution. Solution is circulated through this bed and maintains a constant
concentration of approximately 30% dissolved solids. As the solution is
withdrawn from the tank, it is replaced with fresh water. More soda ash
59
-------
then dissolves from the crystal bed to maintain the 30 percent dissolved
solids concentration. When the bed is reduced to a certain depth, more
dry soda ash is added to the tank. The monohydrate crystals can exist only
at temperatures above 110°F; otherwise a different, less soluble crystal
forms. Therefore, it is important that the tank be kept above this tem-
perature. The tank has steam plate coils to maintain the temperature above
110°F. The Slur-0-Lyzer® and tank stand approximately 9.2 meters (30 feet)
high and 6.1 meters (20 feet) in diameter.
Process Description
The main mechanism for particle collection in the FMC Dual Throat
Venturi Scrubber is inertial impaction. In this phenomenon gas is accel-
erated in the venturi section, and the particulate matter accelerates along
with it. Water droplets are injected from the spray nozzles, and the parti-
culate matter impacts into these droplets. The water droplets containing
the particulate matter are then removed from the gas stream in the separator
in the manner previously described . The water collected in the integral
recirculation tank is recirculated back up into the liquid introduction
system of the venturi. A bleed stream is taken from the recirculation system
at a rate which maintains either the suspended solids concentration at
approximately 0.5% or the dissolved salt concentration at about 10%, whichever
is closest to its limit. Since the abrasion characteristics of the
particulate being collected were unknown, the 0.5% limit was considered
to be a conservative compromise between water conservation and equipment
life. This bleed stream is then neutralized and sent to a series of
ponds where the solids are expected to settle out. Fresh water is added
to the integral recirculation tank to replenish the water lost from the
bleed stream and evaporation in the venturi. It is also intended that
some water from the ponds will be used as make-up water.
Sulfur dioxide absorption is accomplished in the venturi throat using
a sodium based alkaline solution. A schematic of this process is shown in
Figure 6. Sulfur dioxide is absorbed in. a solution containing sodium
sulfite (Na2S03), sodium bisulfite (NaHSO,,), and sodium sulfate (Na-SO,).
As S02 is absorbed, the sodium sulfite is converted to sodium bisulfite
and a percentage is oxidized to sodium sulfate. These reactions are shown
below:
S02 + Na2S03 + H20 — > 2NaHS03 (1)
2Na2S03 + 02 — > 2Na2S04 (2)
The bleed stream taken from the recirculation system for the particulate
removal also prevents the solution from becoming saturated with dissolved
solids. To replenish the sodium lost in this bleed stream, soda ash is
added from the soda ash storage tank at a rate which maintains the solution
at the design pH of about 6.5. The Na2C03 reacts with NaHSO- forming
Na.SO, as shown below:
Na2C03 + 2NaHS03 — > 2Na2S03 + C02+ + H20 (3)
The bleed stream is sent to a neutralization tank where soda ash is added
to convert any remaining NaHSO- to Na«SO~.
60
-------
-*»TO EXHAUST STACK
Na2C03-
DUAL THROAT
VENTURI
AND
SEPARATOR
INCINERATOR OFF-GAS
Na2S03
Na2S04
NaHS03
H20—>2NaHS03
Na2C03+2NaHS03 -»2NaS03+C02
I—Na2C03
Na2S03
Na2S04
NEUTRALIZATION
TANK
^7~
o t=±i
TO DISPOSAI
PONDS
Na2C03+2NaHS03 —»2Na2S03-»-C02
Figure 6 - Sodium Based S0_ Absorption Process
Design of the System
There was a significant challange in predicting the design conditions
for this application, since this particular oil shale extraction process had
never been done before. Lab studies had indicated subsequent to the design
and construction of the unit that SO- inlet concentrations might be about 10
times the original design concentration. Immediate plans by Rio Blanco,
however, are to run the scrubber at only about 25% of original flow design.
It is our belief that the scrubber will still remove sulfur under these
conditions within compliance of Rio Blanco's PSD permit (90% removal, 250
ppm SO- in the outlet, and 63 kilograms per hour (138 pounds per hour) SO-
emission).
If Rio Blanco were to operate additional test retorts with this
equipment at higher flow rates and at high inlet concentrations of SO-,
modifications would probably have to be made to the equipment. A preliminary
study by FMC recommended additional spray nozzles and revisions to the soda
ash feed system, as minimum requirements to incrementally increase the
capacity of the scrubber. Addition of spray nozzles to the separator would
increase the SO- collection efficiency of the system. Also, addition of a
dry soda ash storage silo was recommended to increase the storage capacity
of the soda ash system. The soda ash from these silos would be pneumatically
61
-------
transferred into the Slur-0-Lyzer® as required. Lastly, the scrubber blow-
down ponds would also have to be expanded if higher flow rates are planned.
Determining the proper scrubber pressure drop for particulate collection
also posed a design problem. It was believed that the bench scale testing
would not be effective for determining the particle size distribution coming
into the scrubber. Therefore, FMC decided to use a particle size distribu-
tion from previous analogous applications to determine the proper operating
pressure drop. This pressure drop was found to be 114 centimeters (45 inches)
water gauge for an efficiency of 97.8%.
System Construction and Startup
The fabrication and construction of the system went smoothly. Rio
Blanco was able to erect the system with no need for construction supervision
from FMC. A three day training session was conducted by FMC to familiarize
Rio Blanco personnel with the operation of the system.
The system was started up on October 13, 1980, and after some initial
fine tuning was ready to be placed on line later that same day. The FMC
startup engineer was on the job site for approximately one week during the
startup period.
Initial Operation and Test Results
During Retort "0", the scrubber was operated in a manner to minimize
operability problems, i.e. dilute soda ash feed and dilute scrubber/blowdown
liquor. Once the monohydrate crystal bed is formed in the Slur-0-Lyzer® it
must be continuously heated to maintain the monohydrate structure. Because
of the short amount of time the system would be operated for burn "0", the
soda ash storage tank was operated without the monohydrate bed. Retort
1, on the other hand, will need to be operated at design conditions with
a monohydrate crystal bed soda ash feed system and a concentrated scrubber/
blowdown liquor. Retort 1 is currently scheduled to begin in April 1981.
During Retort "0", the scrubber was operated at only about 15% of its
rated capacity of 6,160 m /minute (220,000 acfm). Temperature into the
scrubber has been approximately 230°C (450°F). The scrubber was initially
operated at a 102 centimeter (40inches) water gauge pressure drop across
the throat, but was later reduced after it was found that particulates in
the blowdown were well under 0.1 wt%. Cursory results also showed that
changing the pressure drop produced no measurable effect on outlet S0?
emission or perceived opacity. Outlet S02 concentrations have averaged
approximately 100 parts per million based on an in-stack SO- analyzer.
Scrubber efficiency has been approximated at well over 90%. Although
no particulate testing was done during the short retort operation, the FMC
start-up engineer (who is an untrained stack observer) reported that the
stack was clear after dissipation of the steam plume.
62
-------
FLUIDIZED-BED COMBUSTION HOT FLUE GAS CLEANUP PERSPECTIVE
ON CYCLONES AND OTHER DEVICES
By: R.F. Henry, W.F- Podolski
Chemical Engineering Division
Argonne National Laboratory
9700 South Cass Avenue
Argonne, IL 60439
ABSTRACT
Pressurized fluidized-bed combustion combined cycle generation of
electricity promises appreciably higher efficiencies and less impact upon the
environment than conventional boilers with flue gas desulfurization.
Implementation of the "combined cycle" requires that flue gas to be expanded
in the gas turbine be cleaned of particulates to tolerable levels. This
cleanup problem is currently the focus of research and development into
several types of improved hot gas cleaning equipment, but to date the best
performance has been obtained with the more conventional type cyclones.
These performance data have been compared to theory, and predictions for
performance for several cyclones in series show that required turbine
tolerance levels should be attainable with reasonably sized equipment.
Gas cleanup downstream from the turbine might still be required with some
coal/sorbent combinations to comply with EPA-NSPS.
INTRODUCTION
Increased utilization of coal as an alternative to hydrocarbon fuels
would reduce the United States dependance on foreign oil. However, use of
higher-sulfur-content coals requires provisions for control of potential
emissions. This may be accomplished by adding a flue-gas scrubbing system to
a conventional coal boiler, by processing of the coal (cleaning), by con-
verting to a synthetic fuel and removing pollutants from the intermediate
fuel before burning, or by burning in a bed of sulfur sorbent. Of these
alternatives, the latter is of interest in this paper.
Fluidized bed combustion (FBC), at either atmospheric or elevated
pressure has potential for meeting environmental regulations concerning
sulfur and nitrogen oxides, and it has the potential for generating elec-
tricity at increased efficiency, particularly if carried out in the pres-
surized mode with combined gas and steam turbine generators.
Implementation of the combined cycle concept depends on interfacing of
the gas turbine with the pressurized, fluidized-bed combustor. Even though
the PFBC has low NOX and S02 emissions, the "dirty" nature of coal (ash and
alkali), coupled with the highly turbulent nature of contact in the PFBC and
the characteristics of sulfur sorbents leads to an effluent with a consid-
erable loading of particles as well as alkali metal concentrations in excess
of current limits for oil- and gas-fired turbines. Removal of these
particles hot and at elevated pressure is necessary for best efficiency of
63
-------
the gas turbine. Adsorption of the alkali from the hot flue gas and develop-
ment of resistant cladding or coatings for turbine blades are being consid-
ered as alternatives for corrosion protection.
BACKGROUND
Allowable levels of particulate matter and alkali compounds for econom-
ically-long gas turbine lifetimes are not well defined for PFBC service, but
estimates for particulate limitations have been made based on calculations,
burner-rig tests and several 1000-hour tests of turbine parts in PFBC system
effluent. Extrapolation to reasonable lifetimes has yielded relations as
shown in Figure 1. The Westinghouse curves (1) represent a calculation cor-
responding to an allowable loading and size distribution based on gas turbine
inlet specifications and an alternate allowable loading with finer size
distribution which would have equivalent erosive properties. The General
Electric curves represent extrapolation of some early data (2) obtained at
Morgantown and G.E. from coal-fired turbine experiments (the nearly vertical
curve) and more recent estimates of allowable particulate matter based on
results (3) obtained in PFBC experiments at Exxon and CURL.
Next we look at combustor effluents. Loading and particle size distri-
bution curves are shown in Figure 2 for a typical effluent from the CURL
combustor during the 1000-hour test (4) and from data analyzed by G.E. in
characterizing the efflux for the PFB/CFCC development program.(5)
If these two groups of data, the PFB effluent and the estimate of the
allowable particulate concentration at the turbine inlet are placed on the
same plot as in Figure 3, the degree of performance required for the hot gas
cleanup system is placed in perspective. Removal of greater than 90 per cent
of the total particulate matter loading will surely be required—in fact,
removal of greater than 99 per cent may well be required for all particles
larger than 10 micrometers (perhaps even for particles larger than 5 micro-
meters) .
Particulate control is of importance in the PFBC concept for another
reason—compliance with the EPA-NSPS.(6) As shown in Figure 3, turbines may
well tolerate a higher total loading - as long as it consists mainly of finer
particles - than will be permissible. This leads to the following "trade-
off" which must be made: Hot gas cleanup, at pressure, may be carried out
for turbine protection only, with post-turbine, lower temperature and pres-
sure cleanup to meet emission requirements; or the cleanup required to meet
NSPS may be carried out before the turbine. In either case, high removal
efficiencies of fine particles will be required, but in the former case, at
lower temperature and pressure. Cleanup at the lower temperature and pres-
sure is considered to be possible with existing equipment—electrostatic
precipitator (ESP) or baghouse technology, for example.
Investigations in PFBC pilot-scale facilities have routinely employed
cyclones for at least one or more stages of particulate removal while occa-
sionally testing other devices with potential for performance improvements.
To date, no test of a device other than a cyclone has been completely
64
-------
satisfactory at HTHP conditions experienced in the PFBC pilot-scale facili-
ties.
Three different PFBC development units have completed extensive testing
of turbine parts: Exxon in the Miniplant (7), CURL in the PFBC unit at
Leatherhead, U.K. (8), and Curtiss-Wright in the PFB/Small Gas Turbine rig
(9). At each of these locations, 1000-hour tests have been completed for the
specific purpose of exposure of turbine components (stationary cascade of
turbine blades in the former two tests and a rotating turbine in the latter)
to actual PFBC effluent after some degree of cleaning. Concurrent evalu-
ations of the hot gas cleanup systems used were also made during these tests.
A brief summary of these tests is shown in Table 1. Note first that
there are several significant differences: The Exxon and CURL PFBC's are
steam-cooled, while Curtiss-Wright's is air-cooled; Exxon and Curtiss-Wright
have first-stage cyclones which recycle material to the fluidized bed while
CURL does not; the Exxon cleanup system has 2 units (in addition to the
recycle cyclone), while CURL and Curtiss-Wright have 3 (in addition to the
recycle cyclone at Curtiss-Wright); temperatures at CURL were somewhat lower
than at Exxon and Curtiss-Wright. Conclusions from these tests are:
Exxon - Some damage to turbine parts (higher temperature,
large particles)
CURL - Essentially no damage
Curtiss-Wright - Essentially no damage
In each test there were turbine parts of several different alloys and in
each case one or more materials tolerated the exposure with minimal damage.
These results are encouraging and do justify the prediction that a large
facility could be operated, but probably at less than 1600°F turbine inlet
temperature (perhaps less than 1500°F) with hot gas cleanup comparable to
that demonstrated. Improved performance from the hot gas cleanup system
would allow higher turbine inlet temperatures and increased cycle efficiency.
There is also concern that scale-up of hot gas cleanup equipment will
cause degradation of performance, since in these three tests, cyclone-type
units were used, and the effect of size on separation has been decumented.
Calculations were made to estimate performance of a cleanup system comprising
three cyclones in series, based on existing theory and results from the
recent PFBC tests described above.
PERFORMANCE CALCULATIONS
Choice of a predictive relation for these calculations was made after
review of the literature, including earlier reviews of cyclone and hot gas
cleanup work by Stern (10) in 1955, Jackson (11) in 1963, the American
Petroleum Institute (12) in 1975, Razgaitis (13) in 1977, and many others.
Many approaches to characterization of cyclone performance have been
65
-------
suggested - with various levels of empiricism. In general, agreement exists
on the use of operating conditions as part of a performance prediction
model - usually as an impaction number grouping. There is some agreement on
use of a geometric factor, ranging from a simple trend based on the ratio of
a single dimension to a geometric number to account for all dimensional para-
meters which is used as part of the overall equation for prediction.
The approach selected for use in this analysis was developed by Leith
and Licht.(14) This approach employs the concepts of backmixing, residence
time, the tangential velocity profile and natural length phenomenon exhibited
by vortex flows. It does not account for any effects of agglomeration or
loading—both of which contribute to some conservatism as the general effects
of both are to cause higher collection efficiencies than predicted. Probably
most importantly, the Leith and Licht model does not depend on scaling of
results from a similar unit with the same dust, but allows for prediction of
performance a_ priori.
Comparison of calculated dso's using the method developed by Leith and
Licht was made for Exxon tertiary cyclone performance data (7) and CURL
primary and secondary cyclone performance data.(8) As shown in Table 2 good
agreement was obtained. Performance estimates for a larger hot gas cleanup
system comprising three cyclones in series were calculated.
Operating conditions were assumed to be 1600°F and 10 atmospheres pres-
sure. All cyclones were assumed to have the geometry of the Exxon tertiary
unit.(7) Primary and secondary units would have 100 feet per second inlet
velocity and be 6 feet in diameter, while the tertiary units would have 150
feet per second inlet velocity and be 3 feet in diameter. This information
allows calculation of fractional efficiencies with the equations of Leith and
Licht, and when used with a typical loading and particle size distribution
for combustor effluent from a test period during the 1000-hour series at
CURL (8), stage-wise calculation of cyclone outlet loadings is possible.
Results of these calculations, the cyclone outlet loadings and particle
sizes, are shown in Figure 4, along with the combustor effluent and turbine
tolerance estimates. This hypothetical hot gas cleanup system is seen to be
capable of resulting in a turbine inlet (3° cyclone outlet) loading and size
distribution very close to those estimated as tolerable for gas turbines.
This cleanup system (of 6-foot and 3-foot diameter cyclones) would
provide overall removal efficiency of about 98 per cent, with a pressure drop
of about 6 pounds per square inch—based on the correlation of Shepherd and
Lapple.(14) The system, as envisioned for cleanup of the hot gas in a PFBC
module of 100 MW capacity (25 MW gas turbine in a steam-cooled cycle), would
comprise 3 parallel units each in the primary and secondary stages and 8
parallel units in the tertiary cleanup stage.
These calculations indicate a PFBC system could be built and operated
with a cyclone type hot gas cleanup system for gas turbine protection.
Demonstration of this technology for a long-term and at a large-scale is the
--'- of PFBC development being conducted by the International Energy Agency at
66
aim
-------
Grimethorpe (16) , Curtiss-Wright in their DOE-sponsored pilot plant program
(9), and the team of American Electric Power Service Corporation/STAL-Laval
Turbin AB/Deutsche Babcock at their component test facility in Sweden.(17)
Results from turbine materials testing indicate that at higher tempera-
tures, which are desired for better gas turbine efficiency, lower particulate
levels and/or smaller sizes will be required for acceptable turbine life-
time. (18) This is due to the complex nature of the corrosion, erosion and
fouling mechanisms and their dependance on temperature.
Hot gas cleanup by means of cyclones appears capable of providing ade-
quate turbine protection in first generation PFBC plants, but incentives are
clear for operation at higher temperatures. Considerable value accrues to
alternative, more efficient hot gas cleanup which would allow higher turbine
inlet temperatures and increased cycle efficiency.
There are several classes of devices which have inherently higher par-
ticulate capture capabilities (19) - such as electrostatic precipitators
(ESP's), scrubbers and several types of filters, but each of them is much
more complex than a simple cyclone and has yet to be demonstrated in PFBC
service. Some development work and testing has been carried out on ceramic
fabric and granular bed filters, as well as preliminary testing of HTHP ESP
and an electrostatically enhanced cyclone. Further development of these and
several other types of possible "second generation" equipment has been sup-
ported by USDOE and USEPA. Some of the current projects are shown in Table
3.(20) Germany and England have active programs for the development of PFBC
and Germany has several HTHP gas cleaning processes under development.(21)
Ultimately, development of PFBC systems will be directed by economic
decisions. PFBC's are known to be capable of meeting current emission
requirements for S02 and NOx, and for given coal feed, PFBC's are smaller
than conventional boilers. Hot gas cleanup cost and performance versus
turbine lifetime and various other trade-offs must be made as firm design,
performance and cost data are generated in development, demonstration and
first-generation units. Table 4 lists some areas of concern which might well
govern these economic decisions.
In conclusion, PFBC appears to be a viable alternative for generation of
electricity from high-sulfur coal in an environmentally acceptable manner.
Hot gas cleanup and its impact on turbine lifetime is the area of most
technical and economic uncertainty. At conditions envisioned for first-
generation plants, cyclones should provide adequate turbine protection.
Future plants will probably be designed for higher turbine inlet temperatures
and lower particulate loadings which may be satisfied by the new generation
of hot gas cleanup equipment now being developed.
ENDNOTES
1. Keairns, D.L. et al. Fluidized Bed Combustion Process Evaluation. EPA-
650/2-75-027-c, September 1975.
67
-------
2. General Electric, CFCC Development Program. Advanced Clean-up Hardware
Performance Guidelines for Commercial Plant. FE-2357-37, September 1977 .
3. Boericke, R.R. et al. Assessment of Gas Turbine Erosion by PFB
Combustion Products. The Proceedings of the Sixth International
Conference on Fluidized Bed Combustion, April 1980, CONF-800428 Vol. 2,
p. 724.
4. Roberts, A.G. et al. Fluidised Bed Combustion. 1000 Hour Test
Programme in a Pressurised Fluidised Bed Combustion Facility. Quarterly
Report January - March 1980. FE-3121-14, May 1980.
5. General Electric, PFB/CFCC Development Program. Gas Cleanup Performance
Requirement Report. FE-2357-66, May 1980.
6. U.S. EPA, New Stationary Sources Performance Standards. Federal
Register ^4(113) 33580, 1979.
7. Hoke, R.C. et al. Miniplant and Bench Studies of Pressurized Fluidized-
Bed Coal Combustion: Final Report EPA-600/7-80-013, January 1980.
8. Roberts, A.G. et al. Fluidised Bed Combustion. 1000 Hour Test
Programme in a Pressurised Fluidised Bed Combustion Facility. Vol.
I - IV. FE-3121-15 (a - d), June 1980.
9. Curtiss-Wright, Engineer, Design, Construct, Test and Evaluate a
Pressurized Fluidized Bed Pilot Plant Using High Sulfur Coal for
Production of Electric Power. PFB/SGT Technology Unit Extended Test
Report. FE-1726-51A, April 1980.
10. Stern, A.C. et al. Cyclone Dust Collectors. American Petroleum
Institute, February 1955.
11. Jackson, R. Mechanical Equipment for Removing Grit and Dust from
Gases. Published by BCURA, Leatherhead, England, 1963.
12. American Petroleum Institute, Cyclone Separators. API Publication 931
Ch. 11, May 1975.
13. Razgaitis, R. An Analysis of the High-Temperature Particulate
Collection Problem. Argonne National Laboratory Report, ANL-77-14,
October 1977.
14. Leith, D. and Licht, W. The Collection Efficiency of Cyclone Type
Particle Collectors - A New Theoretical Approach. AIChE Symposium
Series 68(126) 196, 1972.
15. Shepherd, C.B. and Lapple, C.E. Flow Patterns and Pressure Drop in
Cyclone Dust Collectors. IEC 32(9) 1246, 1940.
68
-------
16. Carls, E.L. et al. The IEA Grimethorpe Pressurized Fluidized Bed
Combustion Experimental Facility. The Proceedings of the Sixth
International Conference on Fluidized Bed Combustion, CONF 800428
Vol. 2, p. 225, August 1980.
17. American Electric Power, Pressurized Fluidized Bed Combustion Program,
Summer 1980.
18. McCarron, R.L. and Grey, D.A., Materials Problems in Fluidized-Bed
Combustion Systems. EPRI CS-1469, August 1980.
19. Spaite, P.W. and Burckle, J.O. Selection, Evaluation and Application
of Control Devices. Air Pollution 3rd Ed., Vol. IV, Ch. 2, Academic
Press, New York, 1977.
20. U.S. DOE High Temperature, High Pressure Particulate and Alkali
Control in Coal Combustion Process Streams—Contractors' Meeting,
Morgantown, West Virginia, February 1981.
21. Proceedings of International Conference on Gas Cleaning at High
Temperatures and High Pressures. Julich, Federal Republic of Germany
(VDI-Berichte 363), May 1980.
69
-------
10000
E 1000
z
o
°
§
100
10
GE 77
GE80
Wspec
I
W equiv
I 10 100
PARTICLE SIZE, pm
Figure 1
Estimated Turbine Tolerances
10000
e iooo —
o
g
>
S
100 —
I 10
PARTICLE SIZE,
100
Figure 2
PFBC Effluent Data
I 10
PARTICLE SIZE ,
100
10000 —
1000
100
10
T
I 10 100
PARTICLE SIZE, Mm
Figure 3 """"" ""' "m Figure 4
Effluent, Turbine Tolerance, EPA NSPS Calculated Cyclone Capabilities
70
-------
TABLE 1
TEST FACILITY SUMMARIES
Exxon Miniplant - Steam Cycle Simulation
Recycle and 2 Cleanup Cyclones
20-110 ppm with d5Q = 1-3 ym
but 3-5%>10 ym
1500-1590°F at Cascades
Slight Effects
CURL - Steam Cycle - 2 Streams
1-3 cyclones 2-3 cyclones
1° van Tongeren 1° chamber
2° van Tongeren 2° hybrid volute entry
3° Stairmand 3° Stairmand
105 ± 35 ppm 235 ± 40 ppm
(155 ± 40) 2 cyclones
1425°F 1350 - 1475°F
d_., both 1-2 ym
cascades - no damage
Curtiss-Wright - Air Cycle
Recycle, Dyna-therm, Ducon, Aerodyne Cyclones
250 ppm — Dilution to 80 ppm
1600°F - d5Q = 1-2 ym
Rover IS/60 Turbine OK
TABLE 2
PREDICTED AND ACTUAL CYCLONE PERFORMANCE
_ , m Predicted d,-,, Actual d
Cyclone Test ym 50 ym 50
CURL 1° van Tongeren 2.5 2.7
CURL 2° van Tongeren 1.93 1.0
CURL 3° Simple 1.12 0.88
Exxon 3° of Scaled CURL 1° 1.26 1.1
Exxon 3° of Scaled CURL 2° 1.36 1.5
71
-------
TABLE 3
HTHP GAS CLEANUP PROJECTS
TABLE 4
TRADE-OFF'S, INTERACTIONS, ECONOMIC CONCERNS
US DOE PRDA
I General Electric - Electrostatic Granular
Westinghouse - Ceramic Crossflow Filter
Exxon - Magnetic Granular Bed
Air Pollution Technology - Dry Plate Scrubber
SUNY Buffalo - Accoustic Agglomeration
II General Electric - Electro Cyclone
Westinghouse - "Ducon" Granular Bed
Acurex - Ceramic Bag Filter
Research Cottrell - HTHP ESP
US DOE Non-PRDA
General Electric - Glass Slag Scrubber
M.I.T. - Electro Fluid Bed Filter
Combustion Power - Granular Bed Filter
Penn. State - Accoustic Agglomeration
Denver Res. Inst. - HTHP ESP Test Facility
Linhardt Assoc. - Wedge Separator
Mechanical Technology - Cyclone Centrifuge
Argonne - Alkali Adsorption
Foreign
German, England, Sweden, Canada
South Africa, Australia, Japan
Turbine Tolerance - Particulate
VS
EPA NSPS Limits
(Final Stage Cleanup HTHP or Post-Turbine)
Turbine Lifetime/Tolerance Relations
Erosion Temperature
Corrosion VS Loading
Deposition Alkali
Erosion/Corrosion Interaction
Fluxing Protective Oxides
Exposing Grain Boundaries
Metal Attack (Sulfidation)
Economic Implications of Each (All) of Above:
Conventional Cleanup - Post-Turbine - Turbine
Loadings Higher/Lifetime Shorter (Temperature Lower)
VS
Advanced Method of Cleanup HTHP — Turbine
Loadings Lower/Lifetime Longer (Temperature Higher)
-------
PRESSURIZED AND NON-PRESSURIZED ACOUSTIC
AGGLOMERATORS FOR HOT-GAS CLEANUP APPLICATIONS
By: K.H. Chou and D.T. Shaw
Laboratory for Power and Environmental Studies
State University of New York at Buffalo
4232 Ridge Lea Road
Amherat, NY 14226
ABSTRACT
The application of an acoustic agglomerator (AA) as a dust-particle
preconditioner has been recently investigated in our laboratory. Several
experiments have been conducted with the following AA parameter values:
AA tube diameter: 1" to 18"; flow rate: up to 200 SCFM; pressure:
1 to 7 atm; temperature: room to 500°C; sound pressure level: up to
170 db; sound frequency: 0.4 to 20 kHz; acoustic generator types: siren,
whistles and EM speakers.
The major advances may be summarized as follows: 1) AA works effec-
tively when acoustic-induced turbulence is initiated at about 160 db.
This threshold value is experimentally determined and is shown to be inde-
pendent of the tube size. 2) The acoustic turbulent agglomeration rate is
relatively independent of the sonic frequency. To avoid the mechanical
vibration and the intensive sonic attenuation loss along the AA tube, the
optium frequency is in the range of 0.8 to 1.2 kHz. 3) The immediate
application of AA technology is limited by the lack of experience in large-
scale sound generator design and operation. Our preliminary results show
that large air-jet choppers driven by dusty air and pneumatically control-
led oscillation metallic-diaphragm generators are two possibilities with
the least R&D cost.
NOTE; Please contact the authors for further information regarding
this paper.
73
-------
ALKALIS AND THEIR CONTRIBUTIONS TO CORONA CURRENT
AT HIGH TEMPERATURE AND HIGH PRESSURE
By: Robert W.L. Snaddon
General Electric Company
Corporate Research & Development
Schenectady, New York
ABSTRACT
The corona discharge is being investigated as a means of providing the charging neces-
sary for the electrostatic augmentation of particle removal from high temperature, high pressure
(HTHP) gas streams. In coal fired, combined cycle power generation, the HTHP gases contain
alkali in the vapor phase. Because of their low ionization potentials, it is feared that alkali me-
tals could have a detrimental effect on the operation of corona chargers in these gases. In this
paper the effects of adding alkali to a HTHP gas stream are compared with a chemical analysis
of the test gases. It is proposed that examination of the gas phase chemistry offers a useful
way of assessing the potential alkali problem. For the limited results reported here, tentative
conclusions are put forward regarding the nature and origins of the observed increases in
current.
Introduction
With the current emphasis on the utilization of coal as a primary energy source, considerable
effort is now being delpoyed to develop effective devices for controlling particle concentrations in the
high temperature, high pressure (HTHP) gas streams encountered in the combined cycle, Pressurized
Fluidized Bed Combustor (PFBC) method of energy conversion. In addition to Electrostatic Precipita-
tors, the Electrocyclone (1) and the Electrostatic Granular Bed Filter (2) are two new electrostatically
augmented collection devices presently under investigation. Like the precipitator, these devices rely
on a corona discharge to produce the particle charging necessary for the enhanced performance.
Since it is understood that the charge acquired by a particle in a corona charger is inherently limited
by the intensity of the electric field (3), a good understanding of corona discharges in the PFBC
environment is required if a proper appreciation and optimization of the above techniques is to be
realized.
The early work by Shale (4) and more recent studies by Feldman et al. (5) have demonstrated
that, on their own, the combined operating temperatures and pressures envisaged for the PFBC (i.e.,
900 to 1000 °C and 7 to 10 atm) do not preclude the generation of stable corona discharges. How-
ever, in the real PFBC environment a further potential problem is posed by the presence of alkalis,
where the vapor phase concentrations are expected to amount to a few ppm. Cooperman (6) has
pointed out that the low ionization energies of alkali metals could mean that corona currents and
operating voltages in PFBC type environments might be dictated by small concentrations of these
species. Nevertheless, little attention has been paid to these effects since the work reported by Brown
and Walker (7); presumably, because these authors observed only small changes in current when
alkali was added to their system.
In recent studies of the PFBC gas chemistry (8), it has been shown that the alkalis in the vapor
phase constitute a variety of compounds. The equilibrium concentrations of these compounds as well
as the concentrations of atomic sodium and potassium and their ions are controlled by multiple reac-
tions related to the composition of the coal and the conditions under which it is combusted. Even
though tests have demonstrated that high fields can be established in a real PFBC environment (9), a
wide variety of coals and combustion conditions are to be expected in PFBCs and one must anticipate
different effects arising from different levels of thermally ionized species. Also, some caution should
74
-------
be exercised when drawing conclusions from tests conducted in simulated environments. All the
relevant chemical species are not usually present and added alkalis might form stable compounds
rather than thermally ionized species.
For the above reasons it is suggested that a proper evaluation of the alkali problem cannot be
made without consideration of the gas phase chemistry. In this paper preliminary equilibrium calcula-
tions are presented for a simple set of experimental conditions where alkali salt has been added to a
simulated HTHP gas stream. An attempt is made to identify the nature of the observed increase in
current and the mechanisms controlling its magnitude. The relevance of extrapolating these findings
to real PFBC conditions is also discussed.
Experimental Equipment
The HTHP test facility used in this study is shown schematically in Figure 1. In this system it
was possible to carry out experiments at temperatures and pressures up to 950 °C and 10 atm. The
high temperature gas stream was obtained by combusting carbon monoxide and air in a swirl-cup
burner. The system pressure was controlled by a back-pressure regulator located downstream of the
burner and test section.
The burner and test section were located in a vertically mounted stainless steel pressure vessel.
The 50.8 mm diameter gas flow channel was formed with a 12.7 mm thick castable ceramic and a
further 12.7 mm of ceramic fibre thermal insulation was employed between this and the pressure
vessel wall.
Temperatures and pressures were measured with thermocouples and simple Bourdon gauges.
The fuel and air flowrates were measured using Fischer and Porter flowmeters and the quantity of
water in the test gas was determined by measuring the relative humidity of the air leaving the dryer.
The alkali salt used in this study was sodium chloride. This was injected into the burner air line
in the form of a fine powder (3.9/urn mean diameter) using a small fluidized bed. For reasons
described later in the text, the NaCI was introduced to the system in pulses. This was achieved with a
solenoid valve located upstream of the fluidized bed.
The details of the electrode system are shown in Figure 2. Observations at ambient conditions
indicated that the curvature on the supporting body and the spherical termination were sufficient to
ensure that, at high voltages, the corona discharge was confined almost entirely to the 76.2 mm long,
3.18 mm diameter wire electrode. These components as well as the high voltage lead-through were
manufactured from stainless steel.
At the test temperatures even high quality ceramic insulators tend to become electrically con-
ducting. Therefore, to minimize the leakage current it was necessary to cool the beryllium oxide and
Macor" insulators with air.
The high voltage was delivered by a Hipotronics 875-13, 0- ± 75 kV, stabilized DC unit. The
voltage was monitored with the power supply meter. At the test temperatures the entire inner lining of
the pressure vessel was electrically conducting. Consequently, it was not possible to employ the
usual technique of measuring the corona current between the collecting electrode and ground. Rather,
the currents had to be measured on the delivery side of the system. This was done with a Kiethley
616 digital electrometer.
Experimental Procedure
The procedures adopted during testing were simple. Before and after HTHP testing the electrode
alignment was checked by measuring the current voltage characteristics at ambient conditions. Suffi-
cient time (usually four hours) was allowed for the system to reach thermal equilibrium prior to making
measurements at any given HTHP condition. Current voltage characteristics were recorded at regular
intervals to monitor drift in background or leakage conduction and to ensure that there were no major
changes in the corona threshold and breakdown voltages.
Because the leakage current was found to change significantly over the duration of a test, it was
necessary to inject the alkali in pulses. Observation of the current transients during these pulses then
75
-------
allowed the real changes due to the presence of the alkali to be identified. The output of fluidized bed
injector was calibrated in independent tests using absolute filters and gravimetric analysis.
Experimental Results
The experimental results reported in this paper were all obtained at the same test conditions; i.e..
900 °C and 7 atm. The fuel, air and water vapor flows to the system under these conditions are listed
in Table 1.
TABLE 1. GAS FLOWS TO THE SYSTEM AT 900 °C AND 7 ATM
Gas Flowrate (gs~1)
CO Fuel 0.798
Burner Air 3.641
Cooling Air 0.493
H20 10.3 x 10~3
Measurements were made at four different alkali injection rates. The concentrations, the injection
rates and the transport air and water vapor flowrates through the fluidized bed are listed in Table 2.
TABLE 2. ALKALI, TRANSPORT AIR AND WATER VAPOR FLOWRATES
AT 4 TEST CONCENTRATIONS
NaCI NaCI Air H2O
Cone, (ppm) Flowrate fcgs~1) Flowrate (gs~1) Flowrate (mgs~1)
53 275 0.261 0.65
169 900 0.386 0.96
367 2000 0.510 1.28
717 4000 0.640 1.60
The corona threshold and the breakdown voltages without alkali present were found to lie around
7 kV and 12 kV for positive corona and 13 kV and 17 kV for negative corona. With the present equip-
ment it was difficult to establish exact breakdown or sparkover voltages. As breakdown was
approached the currents increased rapidly, often activating the current limiting trip on the power sup-
ply before any sign of the sparkover could be detected.
Figure 3 shows typical negative current voltage characteristics recorded at the test conditions.
The two curves exemplify the changes in leakage current which occurred during the duration of a test;
the two curves recorded approximately an hour apart.
Figure 4 shows a typical current transient recorded when alkali was injected into the system.
The initial spike resulted from very high alkali concentrations which were caused by the accumulation
of NaCI in the transport line between pulses. Once this was carried away the current dropped to a
steady level corresponding to steady injection from the bed.
The measured increases in current at added alkali concentrations up to 700 ppm were small,
Figure 5 showing these increases for a given constant applied voltage. Tests were also performed at
constant alkali concentrations and different applied voltages. Figure 6 gives the results of one such
test 76
-------
Gas Phase Chemistry
In order to examine the composition of the test gas and to ascertain the numbers of thermally
ionized species which could be expected under these conditions, gas phase equilibrium calculations
were performed for the test conditions listed in Tables 1 and 2. A proprietary computer program
which performs a system free energy minimization using fundamental thermodynamic properties of the
elements in the system was employed in these calculations. The computed equilibria considered 27
species resulting from the known elements in the system and the most likely combinations of these.
Examination of the chemical activities of potential liquid phase species indicated that at the
three highest concentrations (i.e., 169, 367 and 717 ppm NaCI) sodium carbonate (Na2CO3) fume
would be formed. Since only gas phase reactions were considered, these observations imply that the
calculated equilibrium concentrations are only valid for the 53 ppm NaCI case. The calculated equili-
brium concentrations of the four major uncharged and the four major charged species in this case are
listed in Table 3.
TABLE 3. CALCULATED EQUILIBRIUM CONCENTRATIONS OF MAJOR
SPECIES AT 53 ppm ADDED NaCI
Uncharged Concentration Charged Concentration
Species (Part. Press, at 7 atm) Species (Part. Press, at 7 atm)
N2
CO2
02
H20
4.954
1.328
6.891 x 1(T1
2.848 x 10"2
Na+
cr
N02~
co2~
4.906 x 1(T13
4.786 x 10~13
1.201 x 10~14
4.646 x 10"17
Discussion and Conclusions
For the reasons stated in the previous section, comparison of the experimental measurements
with information deduced from the chemical analysis must be confined to the 53 ppm added NaCI
case. Although this restricts the relevant data available for analysis, certain tentative conclusions are
possible and these are discussed below.
In terms of the charging capability of a corona, it is of interest to establish (i) whether the
increased current is unipolar or bipolar and (ii) what mechanism controls the production of this
current. In a corona device the current which performs the charging originates in the ionization zone
and is unipolar. If the increased current results from equal numbers of positive and negative ions in
the charging zone it is bipolar. Bipolar current does not aid the charging process and in the limit of
high concentrations of these bipolar ions, effective charging can be inhibited.
The above two issues are in essence related. To explore these, four limiting cases of the rate of
production of thermal (bipolar) ions and the rate of their removal by the field are considered. These
are:
a. No rate controlling processes; i.e., both the reaction kinetics and the ion removal by the field are
very fast and the reactions giving rise to the thermal ions are driven to completion within the
residence time of the gas in the charging field.
b. Reaction kinetics very slow and rate of removal of ions by the field very fast. In this case the
measured increases in current would correspond to the rate at which the equilibrium concentra-
tion of ions is delivered to the charger by the gas flow.
c. Reaction kinetics very fast and the current controlled by the rate at which ions can be removed by
the field. Provided the total concentrations of the elemental constituents of the gas are not altered
significantly by the removal of ions, the increased current in this case would be given by
-------
A/- neb / EwoWv<, A (D
where n is the concentration (nrT3) of positive and negative ions, e is the electronic charge (1.6 x
1CT19 C), b is the mobility (m2V~1s~1) of the ions, Ew is the electric field (VrtT1) at the electrode
surface and Aw is the area (m2) of the electrode surface.
d. The rate of ion removal by the field is fast and the increased current is controlled by the reaction
kinetics and the production of thermal ions.
From simple calculations it is clear that Cases a and b are not tenable. In Case a all the
sodium would be ionized and for 53 ppm NaCI this would result in a current of approximately 1 A.
For Case b the equilibrium concentrations of ions would only yield a current of some 10~9 A and
this current would not depend on the strength of the applied field.
Since the measured increase in current is very small compared with 1 A, it is feasible to
consider Case c. For the given electrode system, evaluation of the integral in equation 1 is diffi-
cult. Instead, this was deduced by invoking Gauss' law
*o$E-d£- q C (2)
and measuring the capacitance of the electrode arrangement (without the HV lead-through). The
capacitance was found to be 6 pF. For the ion concentrations in Table 3 and a mobility
~ 1(T4nn2V~1s~1 this yields
A/ = 0 x V A (3)
where
ft ~ 1(r10A\r1
and V is the applied voltage.
Comparing this with the data in Figure 6, it is apparent that although the linear dependence
on voltage is evident, the predicted increase in current is somewhat lower than the measured
values. However, given the relative crudity of the experiment and some uncertainty in the value of
ion mobility, better agreement would perhaps have been fortuitous. Good values for the ion mobil-
ities at these conditions are not readily available and extrapolation from ambient measurements
using the T^p~1 temperature pressure dependence (10) was used. Since the interaction of large
ions with surrounding gas molecules is expected to be smaller at the HTHP conditions, the
estimated value of b might be low. Also, field distortions resulting from the high voltage lead
through aren't included in the estimated A/ and it is difficult to assess what effect these might
have.
Since the predicted increases in current for Case d would lie between those for Cases b and
c, the above analysis and arguments strongly suggest that the observed increases in current were
due to bipolar ions and that the current was field limited. This conclusion is enhanced by the fact
that if the current was unipolar the observed increases would likely have been non-linear with vol-
tage and would likely have only appeared above the corona threshold.
Although the alkali concentrations in the PFBC environment are expected to be much lower
than those considered here, it might well be premature to assume that the increased current due
to alkali will be small or bipolar. The gas chemistry in the PFBC is entirely different and one must
expect varying concentrations of thermal ions depending on coal composition and combustion
conditions. Also, one can not ignore the possibility of the presence of low ionization energy or low
work function materials in sufficient concentrations which could significantly alter corona ioniza-
tion or electrode processes. Chemical analysis of the sort presented here appears to offer a
cheap expedient means of assessing these possibilities.
78
-------
tift—i
COMPRESSOR DRYER
•*—»-^ FILTER
BURNER
FLUIDIZED BED INJECTOR
Figure 1. SCHEMATIC LAYOUT OF HTHP TEST FACILITY
133mm
50.8mm
GROUNDED
ELECTRODE
INSULATION
r~ STAIN LESS
STEEL
MACOR"
BERYLLIUM OXIDE
COOLING AIR
Figure 2. HIGH VOLTAGE ELECTRODE ASSEMBLY
79
-------
0.5
0.4
LU
cc
0.3
o
s
DC
<
uu
0.2
0.1
0
I ' I ' I ' I ' I ' I ' I ' I ' I
_ LEAKAGE CURRENT '
24 6 8 10 12 14 16 18 20
APPLIEDVOLTAGE(kV)
Figure 3. NEGATIVE CORONA CHARACTERISTICS AT 900°C AND 7 ATM
I I
05 '
04
< :- '
T' n ^
2
II | ;
DC n?*4**1
-, U.«-
o
0.1 —
Ort .
.U
II
II
• »
|
\
^^»
0
•MM
10 20
TIME(s)
i i i i i
TEMPERATURE = 900 °C
PRESSURE = 7 ATM.
APPLIED VOLTAGE = 14kV
NaCIConc. = 368 ppm
Figure 4. CURRENT TRANSIENT FOR 20s INJECTION PULSE
80
-------
0.1
LU
DC
DC
O 0.01
Z
LU
CO
DC
O
Z
0.001
10
A + 13kV
• - 13kV
100
1000
NaCI CONCENTRATION (PPM)
Figure 5. INCREASE IN CURRENT AT CONSTANT APPLIED VOLTAGE
QC
O
LU
CO
DC
O
Z
15
10
I
"024 6 8 10 12 14 16 18 20
APPLIED NEGATIVE VOLTAGE (kV)
Figure 6. INCREASE IN CURRENT AT 53PPM NaCI
81
-------
Acknowledgements
The author would like to thank Dr. K.L. Luthra of General Electric Corporate Research and
Development for performing the equilibrium calculations presented in this paper and for the many
helpful discussions regarding this work.
This work was performed by Corporate Research and Development under NYS ERDA contract
No. 237-ET/FUC-80 and DOE contract No. DE-AC01-80ET17091 to the General Electric Energy Sys-
tems Programs Department.
References
1. Boericke, R.R.. Kuo. J.T., Dietz, P.W. and Giles, W.B., "Electrocyclone for High Temperature High
Pressure Dust Removal." Presented at A1AA 19th Aerospace Sciences Meeting. St. Louis. MO.,
January 1981.
2. Kallio. G. and Dietz. P.W., "Image Charge Collection of Fine Particles in Granular Beds," To be
presented at the 7th TDFM Convention, Oxford. England, June 1981.
3. Melcher, J.R. and Sachar, K.S., "Charged Droplet Scrubbing of Submicron Particulate." Environ-
mental Protection Technology Series, EPA-650/2-74-075, 1974.
4. Shale, C.C., "The Physical Phenomena Underlying the Negative and Positive Coronas in Air at High
Temperatures and Pressures," IEEE National Convention Record, Pt. 7, 1965.
5. Feldman, P., Bush, J. and Robinson, M. "High Temperature. High Pressure Electrostatic Precipita-
tion" Presented at EPA Symposium on the Transfer and Utilization of Particulate Control Technol-
ogy, Denver, CO., 1977.
6. Cooperman. P. "Spontaneous lonization of Gases at High Temperature," IEEE Transactions on
Industrial Applications, vol. IA-10, No. 4, p. 520. 1974.
7. Brown, R.F. and Walker, A.B. "Feasibility Demonstration of Electrostatic Precipitation at 1700°F,"
JAPCA, vol. 21. no. 10. p. 617, 1971.
8. Spacil. H.S. and Luthra, K.L., "Thermochemistry of PFB Coal Combustion/Gas Turbine Combined
Cycle," General Electric Corporate Research and Development Report no. 80-CRD-238, 1980.
9. Dietz, P.W. and Kallio, G., "Experiments on Electrostatic Charging of Dust in the PFB Combustor
Environment," Presented at 6th International Fluidized Bed Conference, Atlanta, GA. 1980.
10. Nasser, E., "Fundamentals of Gaseous lonization and Plasma Electronics," John Wiley and Sons
Inc.. 1971.
82
-------
HOT GAS CLEANUP IN PRESSURIZED FLUIDIZED BED COMBUSTION
L.N. Rubow
M.G. Klett
Gilbert Associates, Inc.
P.O. Box 1498
Reading, PA 19603
ABSTRACT
Pressurized fluidized-bed (PFB) combustion for electric power
generation provides a direct combustion process for coal and low-grade fuels
with the potential for improved thermal conversion efficiency, reduced costs
and acceptable environmental impacts. For the successful operation of
combined cycle PFB power plants, particulate removal from the combustion
off-gases at operating temperatures and pressures is critical. The gas
clean-up system must be capable of reducing the particulate loading and
alkali metal concentration in the combustion off-gas to levels compatible
with gas turbine operating conditions and environmental standards. With the
rapid development of PFB combustors, hot gas clean-up threatens to become a
bottleneck in the development of a combined cycle power generation system.
This paper summarizes the status of hot gas clean-up development for
PFBC. Included is information on approaches taken for hot gas clean-up, an
outline of present development programs, and a review of other special
problems associated with high temperature and pressure gas clean-up.
Special attention is given to recent PFBC tests with cyclones in series,
with encouraging results. Projections of performance have been made for
commercial sized equipment which demonstrate that cyclones may not provide
sufficient clean-up. The effect of clean-up system requirements on the PFBC
cycle have also been analyzed and are considered.
INTRODUCTION
The U.S. Department of Energy is sponsoring development of
fluidized-bed combustion (FBC) technology as a promising alternative for
utilization of our abundant domestic coal resources in an environmentally
acceptable manner.
Fluidized beds as utility boilers have several advantages over
conventional pulverized coal fired boilers, including an increase in heat
transfer rate, low combustion temperatures and NO levels, toleration of a
wide range of coal properties, and S02 removal. By pressurizing the
combustor, additional advantages are realized, including higher heat
transfer and volumetric heat release rates, (resulting in a smaller volume),
fewer coal feed points, lower NO emissions, and improved SO capture. A
pressurized combustor can be applied to a combined cycle pTant to take
advantage of its inherent high efficiencies.
83
-------
In 1969, testing began in Leatherhead, England to pressurize the
fluidized bed in hopes of realizing these greater potential advantages over
conventional coal fired boilers. These tests were successful, and, as a
result, programs aimed at developing this technology were initiated in the
United States by DOE, EPA, and others.
Pressurized Steam — Cooled Combuslor
Pressurized Air-Cooled Combuslor
Steam Turbine
Ash Disposal
Waste
Heat
Boiler
FIGURE 1
PFBC CONCEPTS UNDER DEVELOPMENT
Ash Disposal
Figure 1 shows the two primary PFBC concepts being developed. The
fluidized bed effluent in both concepts contains a high concentration of
particulates which must be removed prior to expansion in the turbine. Since
no device is available commercially which will remove these particulates at
the required temperature, hot gas particulate removal is a key technical
issue which may be an obstacle to near term commercialization.
Another potential problem pertaining to hot gas cleanup is that of
alkali metal removal, since measured alkali metal emissions are one to two
orders of magnitude greater than those produced by the combustion of
residual oils in large industrial turbines. Suppressing or filtering the
alkalis and/or developing turbine materials which will resist corrosion and
sulfidation are under development. Deposition of particulate matter on
turbine parts affects the performance of the turbine and may also contribute
to the corrosion process.
EMISSION CHARACTERISTICS
The hot gas cleanup system in PFBC should be capable of reducing the
particulate loading in the hot (1500-1700F) combustion off gas to levels
compatible with gas turbine operating conditions and environmental
84
-------
standards. Alkali metal species in the combustion gas may also be too high
for adequate turbine life and a filtering device may be necessary.
Particulate Loadings and Size Distribution
Although cyclones have not been successful in reducing particulate
levels to the EPA New Source Performance Standard (NSPS) requirement
(equivalent to about 0.02 gr/scf at 20% excess air), they will still play an
important part in PFBC cleanup since an advanced cleanup device (i.e.,
baghouse, precipitator) will likely be preceded by two or three cyclones.
Table 1 presents a summary of the data from various technology rigs after
cleaning the combustion gases with two or three cyclones. As indicated by
this data, the best that has been achieved using "small diameter" cyclones
is a particulate carryover of 0.04-0.07 gr/scf with a median particle size
of about 1.3 microns. Since this performance level is expected to degrade
significantly when large commercial scale cyclones are used, conventional
cyclone separators alone appear inadequate for reducing exhaust loadings
sufficiently to meet NSPS. They may also be unable to permit long term gas
turbine operation, and a tertiary device will be needed in order to produce
a viable system to meet turbine requirements.
TABLE 1. PARTICULATE LOADINGS FROM PFB CYCLONES
Average Loadings Median Particle Size
Technology Rig gr/scf ym
CURL 0.04 1.3
ANL 0.52
CPC 0.42 3-4
Exxon (Miniplant) 0.04 2
C-W 0.07 1.3
Chemical Characteristics
PFBC particulates from the primary and secondary cyclones and exhaust
gas captures have been chemically analyzed by various investigators to
determine carbon, sulfur, and calcium contents and, in some cases, ash
constituents. Preliminary experiments on the fate of 19 trace elements in a
pressurized Jfluidized-becL combustor have been reported by Argonne National
Laboratory ' and Exxon . The alkali content of the effluent stream
(primarily Na and K) is a matter of concern because of the potential for
fouling and hot corrosion of the gas turbine blading. The alkali present in
the form of solids is subject to removal by the various hot gas particulate
cleanup devices. However, alkali present in the vapor phase will not be
significantly reduced by particulate filtration devices. The experimental
data on alkali metal emissions in the vapor phase from tests conducted at
NCB/CURL (Leatherhead, England) indicate that about 0.1 to 5 ppm sodium and
potassium are liberated into the gas phase from a FBC at about 6 atm
pressure and temperatures of 1450°F-1750°F. ' '
85
-------
Cleanup Requirements
The particulate removal system should be capable of reducing the
particulate loadings in the combustion off-gas to levels compatible with
environmental standards and with gas turbine requirements. Specific
potential problems related to the gas turbine include blade erosion, hot
alkali corrosion of turbine parts, and gas turbine fouling.
Environmental
The EPA particulate emission standard for new coal-fired steam
generators is presently 0.03 pounds of particulates per million Btu of fuel.
For a PFBC operating in the 20-30% excess air range (steam-cooled and
air-cooled respectively), this is equivalent to an exhaust emission of less
than 0.02 gr/scf. To date, there is no specific restraint on particle size
distribution.
Erosion
Particulate tolerance for gas turbines has been estimated by turbine
manufacturers to be as low as 0.0002 gr/scf, based on liquid fuel data.
Manufacturers of expander turbines (low temperature) specify a 0.07 gr/scf
loading for their large cat-cracker installations. Consideration of recent
data by General Electric has resulted in their proposed requirement of not
more than 0.001 gr/scf greater than 10 |jm and not more than 0.005 gr/scf
greater than 5 Mm-
Corrosion and Fouling
Gas turbine components exposed to the hot combustion gases are made of
materials that form oxide scales to protect themselves from oxidation. In
the presence of alkali metal compounds, which react with the sulfur oxides
and chlorides in the combustion gas, liquid films of sulfate and sulfate-
chloride mixtures can be deposited on the turbine hardware. These melts
must be prevented because they can initiate hot corrosion and can lead to
substantial deposit formation (fouling). Actual experimental results to
date indicate that gas temperature in terms of both level and consistency,
has a critical bearing on the extent of hot gas corrosion.
Fouling is the accumulation of unwanted deposits on the hot turbine
components due to ash constituents in the fueJL In tests conducted by the
Solar Division of International Harvester Co., the fouling characteristics
of particulate from the CPC combustor were investigated. Solar concluded
that the CPC flyash had a marked tendency to form dense deposits on
impaction with hot turbomachinery surfaces.
HOT GAS CLEANUP DEVELOPMENT STATUS
Several approaches have been proposed and are being developed for
particulate cleanup of high temperature and high pressure gases. The
86
-------
processes which have been most actively investigated include inertial
devices, granular bed filters of various designs, electrostatic
precipitators, fabric filters, metal or ceramic filters, and various hybrid
designs. Except for conventional cyclones, these methods must be classified
as developmental hot gas cleanup techniques and require considerable
developmental effort before they can be reliably used. While some methods
and processes are more advanced than others, questions of ultimate
efficiency, reliability and equipment life are still unanswered and must be
confirmed by both rig and pilot scale experiments.
Technology Rig Results
Technology rigs at Leatherhead England (CURL), Combustion Power Co.
(CPC), Argonne National Labs (ANL), Exxon Research Labs (ERE) NASA-Lewis,
and more recently, Curtiss-Wright (C-W) Corporation have employed various
hot gas cleanup devices.
Cyclones, granular bed filters and baghouse devices have been tested at
these technology rigs. Of these devices, only cyclones have proven operable
in the system;,however, outlet particulate loadings are above the NSPS limit
of 0.03 lb/10 BTU (0.015 gr/scf at 20% excess air). Recent results have
shown cyclone "trains" to reduce particulate emissions to 0.04-0.07 gr/scf,
with mean and maximum particle sizes of 1.3 and 8 |Jm respectively. Although
granular bed filters and baghouses have produced encouraging results, their
operability has not been proven. Following is a brief summary of three
recent rig programs.
Work at the Coal Utilization Research Labs CURL, Leatherhead, England
has provided the basis for many PFBC designs, but until recently has not
been involved in HTHP cleanup programs. The majority of the testing at CURL
has been done with two 10 inch diameter cyclones. An Aerodyne cyclone,
modified by GE was then tested. As part of the 10 x 100 hour test sequence
sponsored by DOE, Van Tongeren and Stairmand cyclones were tested, producing
an outlet loading of 0.04 gr/scf.
From 1976 to 1979, Exxon Research Engineering operated a 12 inch
diameter PFBC, conducting a number of investigations of combustion, sorbent
regeneration, and hot gas cleanup. The initial cleanup configuration
consisted of two cyclones. Some modification to the cyclones resulted in an
outlet loading of 1 gr/scf, 4 |jm mean particle size, considerably above EPA
limits. Subsequent testing with a Ducon granular bed filter showed
promising results of 0.05 gr/scf, but never achieved "steady state"
operation due to cleaning cycle difficulties. The addition of a third
cyclone, in addition to a recycle and second stage cyclone, resulted in
loadings of 0.02 - 0.05 gr/scf with approximately 2 [im mean particle size.
Slipstream testing on a bag filter was encouraging, but an "integrated"
cleaning cycle was not demonstrated. This program is currently inactive.
C-W has operated their rig for over 2000 hours, but have tested only
cyclone configurations. Testing began with a recycle cyclone followed by an
Aerodyne 2-stage cyclone with dirty secondary flow. Subsequent
87
-------
modifications included the change to secondary flow followed by the addition
of two cyclones between the primary cyclone and the Aerodyne. The Ducon
cyclone which was installed has performed satisfactorily but the Dynatherm,
a somewhat unconventional swirl design, has failed to perform. After
modifications to the cyclones, an outlet loading of 0.07 gr/scf with a
1.3 Mm mean particle size was achieved. The present configuration has
succeeded in supporting a 1000 hour turbine run with minimal erosion damage.
Developmental Programs
DOE and EPA have sponsored most of the development work in hot gas
cleanup for PFB combustion. The Office of Energy, Minerals and Industry of
the EPA has sponsored fluidized bed combustion for several years. The
objectives of EPA were to identify potential environmental problem areas and
to develop environmental controls, while the fluidized bed combustion
technology is still under development. As part of this program, EPA has
sponsored .several projects in particulate control at high temperature and
pressure. However, present EPA programs in PFBC are primarily in the
area of environmental assessment and not in the development of control
equipment.
In the past, DOE's work in hot gas cleanup has been in support of
specific technology rig and pilot plant programs. There was not a single
unified program to develop gas cleaning equipment. In addition to their
developmental work for specific technologies DOE has now begun a unified hot
gas cleanup developmental program through a multi-contract PRDA (Program
Research and Development Announcement). The DOE programs now comprise most
of the on going work in hot gas cleanup for PFB combustion.
Research work to date has been primarily concerned with determining
process feasibility and in developing lab scale equipment. Correlations of
performance to cleanup has received little attention. High temperature and
pressure sampling equipment is now becoming available which may allow better
correllation of results with data. More effort is needed to develop a basic
understanding of high temperature and pressure collection mechanisms.
DOE Sponsored Projects
o Program Research and Development Announcement
DOE issued a Program Research and Development Announcement (PRDA) to
initiate the development of new concepts and also to test rig scale
equipment. The program participants will address not only particulate
removal, but also the removal of alkali and other trace metals.
The solicitation is divided into two categories:
Category I -- Concept Definition, Laboratory and Bench Scale Testing,
and Evaluation (estimated duration - two and one-half
years)
88
-------
Category II — Process Design Testing and Verification (estimated
duration - three and one-half years)
Five contracts were awarded in Category I. Firms selected by DOE to
develop hot gas cleanup techniques include:
1. Exxon Research and Engineering Co. - Magnetic stabilized-bed filter.
2. General Electric Co. - Electrostatic granular bed filter.
3. Westinghouse Electric Corp. - Ceramic cross flow filters.
4. University of Buffalo - Acoustic agglomeration of particles.
5. Air Pollution Tech. - Dry plate scrubber.
Four firms are contracted to produce rig scale devices in Category II.
These devices are scheduled to be tested at either the C-W Technology Rig or
the Westinghouse Waltz Mill facility. Firms/concepts selected by DOE are
the following:
1. Acurex Corporation - Ceramic Bag Filter
2. General Electric - Electrocyclone
3. Cottrell Environmental Sciences - Electrostatic Precipitator
4. Westinghouse - Granular Bed Filter
o Granular Bed Filter (Combustion Power Company). - The Combustion Power
Company, under EPA and ERDA sponsorship, designed a moving bed granular
filter for use with their CPU-400 pilot plant. A moving bed filter was
constructed in November 1975 but during the first system heat up to
1300°F, the inner panel of the filter buckled due to excessive stress.
After an ERDA review, a more gradual and lower risk development program
was initiated. A series of performance evaluations were carried out
under low pressure conditions on hot and cold models. A 1000 hour hot
flow performance test was completed in 1979-
Electric Power Research Institute
EPRI sponsored work at Westinghouse has been centered on baghouse
development, and recent rig tests have been encouraging.
EFFECT OF CLEANUP SYSTEM DESIGN ON PLANT PERFORMANCE
Minimization of temperature and pressure losses in hot gas cleanup
systems is desirable and is also an important consideration in the design
and selection of equipment. Cyclones are typically low AP devices in the
10-30 inch of water range. In the PFBC application, with inlet pressures of
100 to 150 psia, cyclone system AP's have risen to the 4 to 6 psi range.
89
-------
This is due to the direct relationship of AP to velocity, thus relating to a
AP/P ratio, and also to the high design velocities chosen for high
efficiency PFBC cyclones.
Figure 2 illustrates the effect of particulate cleanup system AP on
plant efficiency for both an air cooled and steam cooled PFBC cycle. Also
shown is the approximate range of APs expected for various advanced cleanup
devices. The advanced cleanup systems under development are being designed
not only for collection performance superior to cyclones but also for
relatively low AP, thus improving plant performance. Figure 3 shows the
effect of particulate cleanup system temperature loss on plant efficiency
for the air cooled and steam cooled PFBC cycles. Since the temperature loss
occurs only with the combustion gas, the air and steam cooled PFBC concepts
experience nearly the same effect. The trend with cyclones will be to
improve the performance with smaller diameter units which will increase
surface heat losses. Advanced devices will be enclosed in large vessels,
thus minimizing heat losses.
1.5 -I
1.0 —
0.5-
0 •
I.O-i
if 0.5 -
0 5 10
HOT GAS CLEAN UP SYSTEM PRESSURE LOSS~-%6P/p
FIGURE 2
EFFECT OF HOT-GAS CLEANUP SYSTEM
PRESSURE LOSS ON PLANT EFFICIENCY
i i i i i
0 10 20 30 40 50
HOT GAS CLEAN UP SYSTEM TEMPERATURE LOSS °F
FIGURE 3
EFFECT OF HOT-GAS CLEANUP SYSTEM
TEMPERATURE LOSS ON PLANT EFFICIENCY
With these relationships established, the designer of PFBC systems must
choose a device which will not only provide acceptable collection
efficiency, but will result in the minimum AP and heat loss penalty to the
system.
ALTERNATIVES TO HOT GAS CLEANUP
As an alternative to the present PFBC concepts, two additional
approaches were investigated (PFBC Expander Cycle and a Recuperative Heat
90
-------
Exchanger Cycle) which, while reducing the cycle efficiency somewhat provide
a viable means of gas cleanup.
PFBC Expander Cycle
Increased awareness of the necessity for energy conservation has
provided incentive for the development of rugged turbines to recover energy
from many processes requiring pressurization. Among these process are fluid
catalytic cracking, nitric acid and ethylene oxide production, and recently,
steel production. While many of these processes impose severe conditions on
the expander none duplicate those anticipated in PFBC operation. However,
it has been demonstrated that particulate loadings much heavier than those
specified by leading industrial turbine manufacturers can be tolerated in
"heavy duty" expanders designed for several years of continuous operation.
Particulate limits have been set by many turbine manufacturers to 0.0002
gr/scf. Expanders, however, can withstand loadings of .07 gr/scf at a
lower temperature. High temperature turbines are limited to less than 0.01
gr/scf to achieve 50,000 hours lifetime. Whatever the ultimate
requirement, a turbine built to withstand a relatively heavy particulate
loading is an attractive option for PFBC.
A study to examine the cycle efficiencies of PFBC systems employing an
expander has recently been conducted by Gilbert Associates. Two basic
systems were investigated: 1) An air-heater PFBC cycle and 2) a steam
generating PFBC. The air heater cycle was derived from an EPRI study
conducted by Curtiss-Wright. A compressor driven by a low temperature
"expander" turbine (1150°F) supplies combustion air to the PFBC (1650°F).
The combustion air temperature is reduced by an economizer, which is part of
the waste heat steam system. This is a modification of the reference system
which increases the cycle efficiency from 31.7 to 35.2%.
The, steam cycle was based on the GE CFCC (Coal Fired Combined Cycle)
design. As in the air heater cycles, combustion air is supplied by an
expander driven compressor and clean-up is provided by a cyclone train.
Supercritical steam (3500/1000/1000) is produced within bed heat exchanger
tubes and from turbine exhaust heat. The power produced by the turbine is
small relative to the steam turbine power production, and therefore, the
system is not sensitive to the low temperature required by the expander.
The resulting cycle efficiency is 38.3%, very close to the 40% with a "high"
temperature turbine. A change to more conservative subcritical steam cycle
results in an efficiency of 37.4%, still an attractive system.
It appears that utilization of an expander in PFBC is a viable option
and will result in a system which is both cost and performance competitive
with conventional technology.
Cold Gas Cleanup With a Recuperative Heat Exchanger
The use of low temperature conventional cleanup in a pressurized
fluidized bed combustor system requires the use of recuperative heat
exchanger in the gas cycle to maintain a reasonable turbine inlet
91
-------
temperature. The high pressure combustion gases are cooled by a
recuperative heat exchanger, cleaned of particulates, and then reheated by
the recuperator before exhausting to the turbine (Figure 4). A high
efficiency venturi scrubber was selected as the particulate cleaning device.
The hot gas from the pressurized fluidized bed combustor passes through
two stages of cyclone type air cleaners where the entrained medium and
coarse dust is separated and removed from the gas stream. The gas then
enters the recuperative heat exchangers. In the counterflow heat exchanger,
the hot gas in the tubes is cooled as it transfers heat to the clean gas
which is flowing on the shell side. The clean gas from the venturi scrubber
flows back to the shell side of the recuperative heat exchanger where it is
reheated and flows to the gas turbine inlet.
The basic recuperative heat exchanger is presently restricted to about
1000°F due to the limiting factors of materials, thermal expansion and
welding procedures. This study was based on a split or two unit heat
exchanger which permits approximately two thirds of the required heat
transfer area to be fabricated in lower priced material, while the remaining
third must be Haynes 25, 310SS, or equivalent.
FIGURE 4
COMPARISON OF PFBC HOT AND COLD GAS CLEANUP SYSTEMS
A study was carried out to determine the performance penalty which
accompanies reduced temperature particulate cleanup. The hot and cold gas
cleanup alternatives were compared for both the air cooled and steam cooled
combustor concepts. Two hot gas cleanup cycles were evaluated for the air
cooled system, one with a granular bed filter and the second using only
cyclones. Cold gas cleanup results in an efficiency loss of 4 to 4.5
92
-------
percentage points. The reduced performance combined with the added cost of
the heat exchanger makes this PFBC option relatively less attractive.
CONCLUSIONS
Pressurized fluidized bed combustors will require high temperature
particulate cleanup for turbine protection. While significant work has been
accomplished in the area of hot gas cleanup, progress generally has been
slow and encouraged mainly by Government funding. This has been true for
the most part because these systems present difficult technical problems
which are costly to solve. Manufacturers have not seen assurance of a
potential market that would justify development of these systems. With the
rapid development of fluidized bed combustion and gasification, hot gas
cleanup has threatened to become a bottleneck in the development of a
competitive, coal-fired combined cycle power generation system.
The DOE hot gas cleanup program is very important since output from
that program may be requisite to PFBC commercialization. Several promising
devices are being developed and these devices appear to improve PFBC cycle
performance. PFBC cycles using an expander or a recuperative heat exchanger
were assessed as an alternative approach. The expander cycle looked very
promising whereas the recuperative heat exchanger cycle was less attractive.
Questions of expander availability, development and costs must be answered
to determine if this option is practical.
REFERENCES
1. Vogel, G.J., et al. Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion and Regeneration of
Sulfur-Containing Additives. EPA-650/2-74-104. September 1974.
2. Vogel, G.J., et al. Annual Report on a Development Program on
Pressurized Fluidized-Bed Combustion. ANL/ES-CEN-1011. July 1975.
3. Hoke, R.C., et al. Studies of the Pressurized Fluidized-Bed Coal
Combustion Process. EPA-600/7-76-011. September 1976.
4. National Research and Development Corporation. Pressurized
Fluidized-Bed Combustion. Office of Coal Research. R&D Report No. 85
Interim #1. July 1974.
5. National Research Development Corporation. Pressurized Fluidized-Bed
Combustion, Complete Report of Test Run No. 5. BCURA Ltd. US ERDA.
September 1975.
6. National Research Development Corporation. Pressurized Fluidized-Bed
Combustion, Report No. 43, Annual Report - July 1976 to June 1977. US
ERDA. September 1977.
93
-------
7. J. P. Meger and M. J. Edwards, "Survey of Industrial Coal Combustion
Equipment Capabilities: High Temperaure High Pressure Gas
Purification," Oak Ridge National Laboratories, TM-6072, June 1978.
8. General Electric Energy Systems Programs Development. CFCC Development
Program, Cleanup Equipment Performance Specification of Commercial
Plant, Task 4.1.2. ERDA Contract No. EX-76-C-01-2357. August 1977.
9. Corrosion and Erosion Evaluation of Turbine Materials in an Environment
Simulating the CPU-400 Combustor Operating On Coal, Final Report. US
ERDA FE/1536-3. April 1977.
10. Roberts, A. G., et. al, Fluidized Bed Combustion - 1000 Hour Test
Program in a PFBC Facility, Final Report, (FE-3121-15), December 1980.
11. Exxon Research and Engineering, "Miniplant Studies of Pressurized
Fluidized Bed Coal Combustion: Third Annual Report", April 1978,
EPA-600/7-78-069.
12. Henschel, D.B. "The EPA Fluidized Bed Combustion Program - An Update."
Proceedings of the Fifth International Conference on FBC. December
12-14, 1977. Mitre Corp. for U.S. DOE, December 1978.
13. "Exploratory Research Development, Testing and Evaluation of System or
Devices for Hot Gas Cleanup." Program Research and Development
Announcement No. (PRDA RA01-79ET15055) , Issuance Date April 4, 1979.
14. General Electric, CFCC Development Program, Taks 4.1.2 Report, March
1978, U.S. DOE FE-2357-37.
15. R. Schaeffer, "Cost Estimate of Fluidized Bed Combuster Air Heater for
Gas Turbine," EPRI Report FP-995, Technical Planning Study TPS 76-664,
February 1979.
16. General Electric, CFCC Development Program, Commercial Plant Design
Definition, (FE-2357-28), March 1978.
94
-------
VENTURI SCRUBBING FOR CONTROL OF PARTICULATE EMISSIONS
FROM OIL SHALE RETORTING
By: Gerald M. Rinaldi
Monsanto Research Corporation
1515 Nicholas Road
Dayton, Ohio 45407
Robert C. Thurnau
U.S. Environmental Protection Agency
5555 Ridge Avenue
Cincinnati, Ohio 45213
ABSTRACT
During a week-long field study in September 1980, a mobile pilot-scale venturi
scrubber was tested for control of particulate emissions from Laramie Energy
Technology Center's 150-ton simulated in-situ oil shale retort. The retort
off-gas flow of 545 acfm, discharged from a heat exchanger at a temperature of
136°F and saturated with water, was scrubbed at liquid-to-gas ratios of 11 to
18 gpm/103 acfm. Sampling and analysis of the scrubber inlet and outlet gases
were conducted to determine particulate removal. Outlet particulate concen-
trations were consistently reduced to 35 mg/m3, even though inlet loadings
varied from 125 to 387 mg/m3 and 50 weight percent of the particles were less
than four microns in diameter. Particulate control efficiencies up to 94
percent were achieved, although no correlation to liquid-to-gas ratio was
apparent. Simultaneous control of ammonia emissions, at efficiencies up to
75 percent, was also observed.
INTRODUCTION
Oil shale has been recognized as a potentially substantial energy resource in
the United States for more than 100 years. Recently, increasing dependence
on foreign oil supplies and rapidly escalating oil prices have provided new
incentive for shale oil recovery from deposits in Colorado, Utah, and Wyoming.
At least four domestic firms (Colony Development Operation, Paraho Development
Corporation, Superior Oil Company, and Union Oil Company) have developed sur-
face retorting processes, in which oil shale is mined and crushed prior to
thermal processing in aboveground facilities. In-situ or modified in-situ
processes, in which the shale bed is hydraulically or explosively fractured
and retorting is carried out underground, are now being developed by Equity
Oil Company, Geokinetics, Inc., Occidental Oil Shale, Inc., and Rio Blanco
Oil Shale Company.
Despite the benefits of oil shale retorting as an alternative energy source,
byproduct gases containing a complex mixture of criteria pollutants (particu-
late matter, hydrocarbons, carbon monoxide, and oxides of nitrogen and sulfur)
and non-criteria pollutants (e.g., ammonia and hydrogen sulfide) could have an
adverse impact on the pristine air quality of the Rocky Mountain region if
released uncontrolled. Therefore, pollution control methods capable of ade-
quately reducing environmentally harmful discharges must be available to assure
95
-------
the emerging oil shale industry's compliance with future standards. Toward
this end, the U.S. Environmental Protection Agency (EPA) has contracted with
Monsanto Research Corporation (MRC) to characterize point-source air pollution
due to surface and in-situ oil shale retorting, focusing on particulate emis-
sions, and to evaluate available particulate control methods. This paper pre-
sents results of a field test in which a mobile venturi scrubber was used to
control particulate emissions from a simulated in-situ oil shale retort oper-
ated by the Department of Energy's Laramie Energy Technology Center (LETC) in
southeastern Wyoming.
DESCRIPTION OF LETC RETORTING FACILITY
In 1969, a batch-type retort with the capacity to process 150 tons of shale
was constructed by the U.S. Bureau of Mines at a site north of the Laramie
Energy Technology Center in Laramie, Wyoming (1). Figure 1 is a flow diagram
of LETC's 150-ton retort and its auxiliary equipment (1). Combustion of oil
shale is initiated via a natural gas burner mounted on the retort lid. Air
is forced downward through the shale bed, simulating the vertical modified
in-situ technologies being developed by Occidental Oil Shale, Inc., and Rio
Blanco Oil Shale Company. Shale oil and byproduct water drain to a collection
tank located below the retort. The retort off-gas, containing residual oil
mist and water, passes through four packed towers (not two as originally
designed and indicated in Figure 1) and then through a water-cooled heat
exchanger in an attempt to remove this entrained material. After line pres-
sure is increased by a positive-displacement blower, the gas stream is split
to allow for recycle. The excess is vented to a waste-gas stack equipped with
a natural gas burner to oxidize combustible components before release to the
atmosphere. Retorting continues until the bottom grate temperature increases
to approximately 500°F, by which time shale oil production has already stopped
and the oxygen content of the off-gases has begun to rise.
Relevant details of the operation of LETC's 150-ton oil shale retort during
Run 19, when the pilot scrubber test was conducted, are as follows. The shale
retorted was medium-grade material, containing about 28 gallons of oil per ton
of rock, taken from the DOE mine at Anvil Points, Colorado. A mixture of air
and steam was injected to the retort at feed rates of 13,250 dry standard
cubic feet per hour and 121.5 pounds per hour, respectively. Ignition and
air input began at approximately 10:45 AM MDT, 8 September 1980; retorting
continued until 12:30 AM MDT, 15 September 1980. Total oil recovery for the
5.5-day burn was 1,990 gallons, equivalent to a production rate of 360 gallons,
or 8.6 barrels, per day. For comparison, commercial oil shale processing
facilities will produce as much as 50,000 barrels per day, continuously, for
20 years or longer (2).
CONTROL DEVICE SELECTION AND OPERATION
Choice of Wet Scrubbing
During the first phase of EPA Contract No. 68-03-2784, MRC evaluated electro-
static precipitation, fabric filtration, and wet scrubbing relative to their
potential ability to control particulate emissions from in-situ and surface
96
-------
retorting (2). The utility of electrostatic precipitation in the application
of interest is questionable because of the lack of information on particle
resistivity and because of the presence of flammable or explosive gases that
represent a potential safety hazard. The formation of a porous dust cake
necessary for high particulate removal efficiencies in fabric filtration
equipment may be precluded by shale oil mist and possibly by condensed mois-
ture in retorting off-gases. Unlike electrostatic precipitation, there is
much less fire or explosion hazard associated with wet scrubbing of retort
off-gases containing combustible species. Scrubbers also are preferable to
baghouses because the former are not only effective for control of liquid or
solid particles in high-temperature, moisture-laden gas streams, but also
suitable for treatment of gas streams with fluctuating characteristics, for
example, flow rate, temperature, and composition. Wet scrubbing also has the
advantage that simultaneous particulate removal and absorption of gaseous
pollutants can be accomplished using an appropriate liquid medium. Primarily
based on these considerations, MRC recommended wet scrubbers for pilot-scale
testing to study the control of particulate emissions from in-situ and sur-
face oil shale retorting (2).
Operation of EPA's Mobile Scrubber at LETC
For the pilot-scale particulate control test at Laramie, MRC used a venturi-
cyclone scrubber housed in a standard freight trailer and developed by EPA as
a mobile research unit. A schematic diagram of the mobile scrubber is given
in Figure 2. Prior testing of the mobile wet scrubber, first by MRC and then
by Acurex Corporation, led to some design and hardware modifications not
indicated in the figure. A single Roots blower replaced the original two
Paxton blowers because of the latter's inability to handle residual particu-
late matter and entrained water droplets passing the demister. The original
demister was replaced by an epoxy-coated, reinforced plywood compartment
containing a series of zigzag baffles to collected entrained droplets.
Of the three interchangeable venturi throats provided with the mobile scrub-
bers, the "medium" throat was used at Laramie. This throat has a diameter of
2.36 inches, a length of 12 inches, and two radially opposed liquid feed
nozzles two inches below the throat entrance. Piping throughout the scrubber
trailer, and that used for external hookup to the retort off-gas line, was
six-inch-diameter stainless steel. Because the gas stream treated at Laramie
was saturated with moisture and its temperature relatively low, neither the
presaturator shown in Figure 2 nor four banks of band heaters along the scrub-
ber inlet line were needed.
LETC's existing ductwork was breached just upstream of the inlet to the ther-
mal incinerator (see Figure 1), venting the entire gas flow to the mobile
scrubber. In order to reduce any hazards of environmental contamination or
worker exposure associated with emissions of hydrocarbons and other gaseous
pollutants, the scrubber outlet gases were piped to the incinerator. The
retort off-gases treated had an average temperature of 136°F, an average
moisture content of 21 percent by volume, and an average flow rate of 545
acfm, or 314 dscfm.
97
-------
During the field testing at LETC's 150-ton retort, MRC varied the scrubbing
liquid feed flow, and thus the liquid-to-gas (L/G) ratio, over a range typical
of commercial installations in order to observe the effect of this parameter.
Table 1 describes the various operating conditions employed, in chronological
order, specifying the volumetric feed rate and origin of scrubbing liquid,
L/G ratios, and durations. The relationship between liquid feed rate, or L/G
ratio, and the pressure drop across the venturi scrubber and cyclone is quanti-
fied in Table 2. Up to 5 gpm of well water was supplied by LETC, and the
recycle consisted of scrubber discharge treated in a mobile stream stripping
unit operated by Denver Research Institute. Steam stripping was intended to
remove dissolved gases, such as ammonia, carbon dioxide, and hydrogen sulfide,
from the recycle stream. Suspended solids were removed by a traveling-grate
deep-bed filter mounted in the scrubber trailer. The temperature of the com-
bined scrubber feed varied from 55° to 110°F and the pH from 6 to 9, the
higher values reflecting the effect of heating during steam stripping and
the uptake of basic dissolved gases, respectively.
TABLE 1. SCRUBBER OPERATING CONDITIONS
Operation
code
8A
10A
10B
6A
6A
IOC
8B
Scrubbing liquid
Well water
3
5
0
0
0
5
4
feed rates, gpm
Recycle
5
5
10
6
6
5
4
L/G ratio,
gpm/103 acfm
15
18
18
11
11
18
15
Duration,
hr : min
2:25
14:10
2:00
24:50
6:50
24:00
5:45
Numerals in codes indicate total scrubbing liquid feed rates, and
letters indicate different well water/recycle ratios.
Calculated using average gas flow rate of 545 acfm.
TABLE 2. EFFECT OF LIQUID-TO-GAS RATIO ON PRESSURE DROP
Total scrubbing liquid Liquid-to-gas ratio, Pressure drop,
feed rate, gpm gpm/103 acfma in. H20^
5 9 22
6 11 26
7 13 28
8 15 32
9 17 35
10 18 38
o
Calculated using average gas flow rate of 545 acfm.
Measured from venturi inlet to cyclone outlet.
98
-------
SAMPLING AND ANALYSIS METHODS
The principal objective of gas and liquid sampling and analysis during the
venturi scrubbing test at LETC's 150-ton oil shale retort was to determine
the particulate control efficiencies achievable. Points sampled included,
for the gas phase, the scrubber inlet and outlet and, for the liquid phase,
the scrubber feed and discharge water. Equipment and procedures used to
determine particulate loading were essentially those described in EPA Meth-
od 5 as outlined in the Federal Register (3). The back half of the Method 5
sampling train was modified by including an XAD-2 resin trap (to collect
organic vapors) between the filter and the impingers and by using an oxida-
tive reagent solution in the impingers (to collect volatile compounds of
trace elements).
Because of the small duct diameter (6"), single-point sampling was used to
determine the particulate emission rate, with the probe tip placed at the
point of average gas velocity, as determined by a detailed preliminary veloc-
ity traverse (EPA Method 1). Particulate mass, as specified by the Federal
Register method, is reported as the sum of that collected on the filter and
that rinsed from the probe and connecting glassware upstream of the filter.
In addition, MRC measured the amount of "back half" residue rinsed from the
downstream half of the filter holder and the glassware down to the XAD-2 resin
trap.
Particle size distribution was measured using Andersen Mark III cascade im-
pactors, including the preimpactor for separation of coarse particles (4).
Glass fiber collection substrates were used because the oily nature of the
particulate might otherwise cause problems of carryover from stage to stage.
The isokinetic impactor sampling rate with a 1/8-inch-diameter nozzle was
approximately 0.2 acfm, and sampling durations were 10 and 20 minutes for the
scrubber inlet and outlet, respectively. Due to the control efficiency of the
pilot scrubber, the outlet sampling time was too brief to collect a measurable
amount of sample; therefore, only the inlet particle size distribution is
reported.
RESULTS AND DISCUSSION
Particle Size Distribution
MRC's measurements of the size distribution of particles emitted by LETC's
150-ton retort represent the first such determination for that specific
facility and one of only very few for the oil shale industry in general. An
average distribution, based on three measurements, of particle size in the
retort off-gases, that is, the scrubber inlet, is listed in Table 3. Two
significant conclusions can be drawn from this data. First, more than half
the particulates, by weight, have a diameter less than five microns, with
approximately 10 percent less than one micron in diameter. Also, the size
distribution appears to be bi-modal, with the fractions larger than 20 microns
(^35 percent) and between one and two microns (^30 percent) predominating.
Visual inspection of the impactor collection substrates indicated the presence
of both straw-colored oily material and a black, possibly inorganic dust,
perhaps due to attrition of the shale in the retort.
99
-------
TABLE 3. SIZE DISTRIBUTION OF PARTICLES IN RETORT OFF-GAS
Weight percent
Stage
Size range,
microns
Preimpactor
+ Stage 0
1
2
3
4
5
6
7
>19
13.6
9.1
6.3
4.0
2.0
1.3
0.9
.5
- 19.5
- 13.6
- 9-1
- 6.3
- 4.0
- 2.0
- 1.3
In size Cumulative less
range than size range
36.8
0.6
4.4
1.3
7.0
11.0
28.4
5.0
63.2
62.6
58.2
56.9
49.9
38.9
10.5
5.5
Backup
filter
<0
.9
5.5
0
Averages calculated from results of three separate
samples.
Particulate Loading and Control Efficiency
Table 4 presents the results of seven Method 5 measurements of particulate
loading in LETC's retort off-gases and in the venturi scrubber outlet for
three different scrubber operating conditions. These calculated values are
based only on the mass collected from the filter and washes of the front-half
glassware in the sampling train, thus making this data equivalent to others
taken according to the Federal Register method. By comparison, the mass of
condensible organic "particulate" matter rinsed from the glassware between
the filter and the resin trap (back half) ranged from 2.7 to 5.5 times the
amounts in Table 4 for the scrubber inlet and 4.8 to 21 times as much for the
scrubber outlet. Therefore, LETC's in-situ oil shale retorting off-gases con-
tained a very substantial quantity of potential "particulate" emissions that
condense between 250°F and about 70°F, the approximate temperature of the
resin trap. The larger ratios of back-half to front-half "particulate" in the
scrubber outlet merely indicate, as expected, that venturi scrubbing is not
suitable for removal of condensible organic compounds.
Several qualitative conclusions can be derived from the particulate emissions
data in Table 4. The overall average concentration of 214 mg/dncm in LETC's
retort off-gases falls within the range of 20 to 750 established from available
estimates and data MRC gathered during Phase I of this EPA contract (2) . The
inlet particulate concentration was however highly variable, ranging from 125
to 387 mg/dncm, with no apparent correlation to the progress of the retort
burn. By extrapolating the overall average particulate emission rate of 113
grams per hour for LETC's 8.6-barrel-per-day facility, the predicted uncon-
trolled emissions from a 50,000-barrel-per-day commercial oil shale plant
are approximately 5,200 metric tons per year, assuming a 90 percent stream
factor.
100
-------
TABLE 4. CONTROL OF PARTICULATE EMISSIONS
FROM OIL SHALE RETORTINGa
Operation
code°
6A
-Averages-
SB
IOC
-Averages-
Concentration, mg/dncm
Inlet
245
275
152
224
144
125
173
387
228
Outlet
37
32
32
34
38
46
35
24
35
Mass flow
Inlet
128
145
80
118
78
69
96
198
121
rate, g/hr
Outlet
17
14
14
15
20
23
18
jl
17
Percent ,
control
87
90
82
87
74
67
82
94
86
Based on mass collected from front-half train wash and filter.
See Table 1 for explanation.
Milligrams per dry normal cubic meter (20°C, 760 mm Hg).
100 x (inlet flow rate - outlet flow rate) * inlet flow rate.
Regarding the efficiency of venturi scrubbing for control of particulate emis-
sions from oil shale retorting, the most significant conclusion from Table 4
is the ability to consistently achieve an average outlet concentration of
35 mg/dncm despite the three-fold variation of inlet concentrations. An unex-
pected result is that increasing the liquid-to-gas ratio from 11 gpm/103 acfm
to 18 gpm/103 acfm did not improve the particulate removal efficiency, as
would be predicted by theory, despite the increased pressure drop. Future
experiments should consider determining the minimum liquid-to-gas ratio that
yields similar particulate control in order to reduce water consumption.
Alternatively, determination of the extent to which the liquid-to-gas ratio
must be increased beyond 18 gpm/103 acfm in order to achieve control effi-
ciencies greater than 90 percent would also be of interest, particularly in
light of the following. Hesketh (5) has developed an empirical relationship,
based on data from a variety of venturi scrubbers, to predict control effi-
ciency for particles less than five microns in diameter. Using this model
under the conditions of the mobile scrubber test at Laramie, the liquid-to-gas
ratios used - 11, 15, and 18 gpm/103 acfm - should have given control effi-
ciencies of 93.8, 95.5, and 96.5 percent, respectively. The corresponding
actual removals achieved were only 87, 74, and 86 percent, considerably less
than predicted. One possible explanation for this performance as applied to
particulate emissions from oil shale retorting may be the inability to control
the relatively large fraction of sub-micron particles.
Additional characterization of the emissions from LETC's 150-ton in-situ oil
shale retort indicated the effect that venturi scrubbing had on air pollut-
ants other than particulate matter. As is typical of this source type, retort
101
-------
off-gas concentrations of ammonia, carbon monoxide, hydrocarbons, and hydrogen
sulfide were all substantial. The average concentration of ammonia in the
scrubber inlet was 3,000 mg/dncm. On a mass flow basis, 50 to 75 percent
control of ammonia emissions was achieved, probably due simply to its high
solubility in water because no special measures were taken to obtain such
performance.
ACKNOWLEDGMENTS
This research was funded by the U.S. Environmental Protection Agency (EPA)
under Contract No. 68-03-2784. The contents of this publication do not neces-
sarily reflect the views or policies of the EPA, nor does mention of organiza-
tions imply endorsement by the U.S. Government. Both MRC and EPA would like
to thank Andy Long and Jas Lotwala of the Department of Energy's Laramie
Energy Technology Center for their invaluable assistance during preparation
for and implementation of this program. The quantity and quality of data
generated during the brief field test are due in large part to the efforts of
MRC's Chuck Duncan, sampling team leader.
ENDNOTES
1. Harak, A. E., L. Docktor, A. Long, and H. W. Johns. Oil Shale Retorting
in a 150-Ton Batch-Type Pilot Plant. U.S. Bureau of Mines, Laramie,
Wyoming, 1974. 35 pp.
2. Rinaldi, G. M., D. L. Zanders, and G. D. Rawlings. Air Pollution Investi-
gations of Oil Shale Retorting: In-situ and Surface; Phase I: Review of
Retorting Processes and Selection of Technology for Control of Particulate
Emissions. Contract No. 68-03-2784, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1980. 63 pp.
3. Determination of Particulate Emissions from Stationary Sources. Federal
Register, 42:41776-41782, August 1977.
4. Operating Manual for Andersen 2000 Inc. Mark II and Mark III Particle
Sizing Stack Samplers, Revision A. Andersen 2000 Inc., Atlanta, Georgia,
1977. 50 pp.
5. Hesketh, H. E. Fine Particle Collection Efficiency Related to Pressure
Drop, Scrubbant and Particle Properties, and Contact Mechanism. Journal
of the Air Pollution Control Association, 24:939-942, October 1974.
102
-------
Noturol
gas
burner
o
OJ
Grote-
1
1
V
s
s
s
s
J 150-
^
i
| FRC[j PRCM
Aif 4 K m r\ •
* VQ Q) {J-
FM
•ton
ri
Blower j
!
P
\
^
s
V
\
\
s
s
y
\
\
v
s
^~\
*nnfl
^
S
y
\
s
s
s
s
s
y
\
y
s
s
\
fimrf
Pocked
\towers
\
\
'•
Wofer-cooled
heat exchanger
r n v I
0
stock
0
-t*3-
LC
Water and oil
-cxi-
^ 1^— Noturol gos
burner
Legend
To From
EPA/MRC
Scrubber
LC-Lood cell
FM-Flov meter
FR-Flov recorder
FRC-Flow recorder controller
PRO —Pressure recorder controller
RD —Rupture disk
Figure 1. Gas and liquid flows in LETC's 150-ton retorting equipment (1).
-------
SCRUBBED
GAS EXIT
STACK GAS
INLET
FLOW
NOZZLE
SUPPLY i *
WATER v-11
^y^/m^y^y/^y^^^y/^0/
«UPPLYTANK
PRESATURATOR
^
QSCRUBBER PUMP
PRESATURATOR PUMP
OVERFLOW
Legend
n BALL VALVE * THROTTLE VALVE • AIR
[XI BUTTERFLY VALVE >Li SOLENOID VALVE —WATER
^ BLn-rtRn-Y'vALVE ^ CHECK VALVE P PROCESS INSTRUMENTATION STATION
Figure 2 - Schematic diagram of EPA mobile scrubber.
104
-------
OVERVIEW OF THE DEPARTMENT OF ENERGY'S
PRESSURIZED FLUIDIZED-BED COMBUSTOR
CLEANUP TECHNOLOGY PROGRAM
By: William E. Moore
Office of Coal Utilization
Fossil Energy
U. S. Department of Energy
Washington, D. C. 20545
ABSTRACT
This paper provides a review of the Department of Energy's (DOE) Fossil Energy,
Office of Coal Utilization's Pressurized Fluidized-Bed Combustor (PFBC) Clean-
up Technology Program. The PFBC cleanup program has two goals:
• Removing contaminants from the gas stream to meet turbine tolerance
for erosion/corrosion and deposition.
• Removing pollutants from the gas stream to meet the New Source
Performance Standards (NSPS).
The program involves three development areas: particulate control, alkali
control and particulate-alkali measurement. The overall program schedule is
presented along with a summary status. Finally, the key technical issues are
outlined which must be solved for the cleanup equipment to meet contaminant
removal goals. To achieve overall program goals will require the coordinated
effort of both the public and private sector. This paper seeks to promote
this coordination by presenting DOE's PFBC cleanup program, and it is hoped
that this program description will improve communication and the end result
will be a well coordinated program leading to development of an acceptable
cleanup train for ensuring early commercialization of PFBC combined cycle
power systems.
INTRODUCTION
The Department of Energy's (DOE) coal-fired pressurized fluidized-bed
combustor (PFBC) cleanup technology development activities are conducted as a
part of Fossil Energy's Advanced Environmental Control Technology Program.
This PFBC cleanup work is part of the overall gas stream turbine (Figure 1)
cleanup effort which also includes cleanup of fuel gas from gasifiers for
combined cycle power systems.
105
-------
This PFBC gas stream cleanup program includes projects that were transferred
from the prior Energy Research and Development Administration as well as new
projects that were awarded under DOE through a competitive procurement with a
Program Research and Development Announcement (PRDA). This PRDA pursued two
stages of development: Category I - "Laboratory Bench Scale Testing and Eval-
uation" and Category II - "Subpilot Scale Design, Testing and Verification."
These projects are to develop and demonstrate hardware-oriented prototype
systems or devices for cleanup of the PFBC gas stream at system temperature
and pressure for use with open cycle turbines in the utility or industrial
sector.
The PFBC cleanup program has two goals:
• Removing contaminants from the gas stream to meet turbine tolerance
for erosion/corrosion and deposition.
• Removing pollutants from the gas stream to meet New Source Performance
Standards (NSPS).
PFBC systems have demonstrated their capability to meet NSPS with respect to
NO and SO emissions leaving particulates as the remaining contaminant of
concern. ¥o meet NSPS, particulates have to be removed to levels equal to or
less than 0.03 lbs/10 BTU. Hopefully, the selected cleanup system will meet
both the NSPS goal as well as the turbine tolerance goal thereby improving the
economics by eliminating the requirement for a downstream particulate removal
device.
To meet the NSPS, the particulate removal system must effectively remove
particles down to one micron in average diameter. (Reference is made to
Figure III — Depiction of Difficulty with NSPS.) The barrier filtration
projects presently underway such as the ceramic felt bag and the granular bed
filters are expected to filter the sub-micron particles as well as the par-
ticles over two microns sufficiently to meet the NSPS. The electrostatic
precipitators removal efficiency under PFBC conditions has to be demonstrated
while the enhanced inertial devices such as the electrocyclone are not ex-
pected to meet the required mass loading level for NSPS. However, it is
exepcted that future commercial selection will depend upon cost, maintainabil-
ity and pressure drop which have to be determined and verified under high
temperature, high pressure PFB conditions.
The greatest obstacle to PFB combustor-power conversion system development is
cleaning the gas stream contaminants to levels that are acceptable for commer-
cial utility turbine operation. The off-gas causes erosion, corrosion and
deposition problems in the open cycle gas turbine. Corrosion in PFBC turbine
operation is primarily caused by the presence of alkali sulfate compounds in
the liquid and vaporous state which deposit as liquids on the blade causing
sodium/potassium sulfate corrosion. Erosion results from the impact of the
high velocity particulate matter while deposition occurs through a buildup
of fine particles on the turbine surfaces. Gas turbines have historically
106
-------
been primarily operated with relatively clean petroleum based fuels and there
is only a very limited data base for durability predictions with coal-derived
contaminant laden gases. There is no concensus within the turbine industry
as to what erosion tolerance the PFB combustor turbines should have other than
the safety of adopting the clean fuel specification. There are considerable
differences between the design philosophies for aircraft derivative turbines
and gas expanders as used in petroleum refineries. The cleanliness goals
DOE's program is working toward are 0.02 PPM for alkali (adopted from the
clean fuel turbine specifications) and 0.001 grains per standard cubic foot
for particles greater than five microns. Since the PFB gas turbine operates
solely on sensible heat from the gas stream, it is essential that cleanup
processes be operated at or near bed temperature to achieve acceptable
efficiency.
DOE's PFBC cleanup program is concerned with three development areas:
• Particulate Removal
• Alakli Removal
• Instrumentation
Particulate Removal
For particulate removal, the only state-of-the-art device available for com-
mercial application with the PFB combustor system is the cyclone. Based on
both 1000 hour tests on the subpilot plants located at CURL Leatherhead,
England and Curtiss-Wright, Woodridge, New Jersey, it is very doubtful as to
whether the conventional cyclone when scaled-up for commercial PFB use can
meet acceptable turbine particulate limits (See Figure IV) due to the reduced
performance of large size cyclones as predicted by cyclone performance scaling
laws. The cyclone train utilized in both these tests consisted of small, under
20-inch diameter cyclones, and achieved an overall particulate removal effi-
ciency of approximately 97 percent for the Curtiss-Wright cyclone train and
over 98 percent for the CURL train. The average total dust loading that
passed through both cyclone trains was 0.04 to 0.06 grains/scf with a mean
particle size of one and one-half to two microns. The emission levels
achieved were approximately 0.2 lbs/106 BTU thus necessitating the use of
downstream particulate removal to meet the 0.03 lbs/106 BTU NSPS level. (For
the PFB combustor approximately 99.9 percent removal efficiency is required to
meet NSPS.) Cyclone efficiency suffers considerably when scaled to commercial
sizes and the use of multi-clones to circumvent scaleup problems has proven
difficult since these multi-clones are very prone to plugging.
In view of the aforementioned conventional cyclone problems, DOE is investi-
gating a number of advanced particulate removal concepts including enhanced
inertial devices, electrostatic precipitators, fabric filter, dry plate
scrubber, acoustic device, granular bed filters, molten glass scrubber and
ceramic cross-flow filter.
The following projects that are being pursued are listed below by principle of
operation and by contractor.
107
-------
• Inertial Devices
- General Electric - Electrocyclone (Subpilot Tests at Curtiss-Wright
S.G.T.)
- Mechanical Technology, Inc. - Cyclocentrifuge
- General Electric - Glass Slag Particle Separator
* Filters
- Granular Bed Filters
- Westinghouse - Fixed Granular Multi-Bed Filter (Subpilot Tests
on Simulated PFBC)
Combustion Power Company - Moving Granular Bed Filter
- Exxon - Magnetically Stabilized Granular Bed
General Electric - Electrostatic Granular Bed
- Air Pollution Technology - Dry Plate Scrubber
Ceramic Filter
- Westinghouse - Ceramic Cross-Flow Filter
- Bag Filter
Accurex - Ceramic Felt Bag (Subpilot Tests at Curtiss-Wright
S.G.T.)
• Electrostatic Precipitator
- Research-Cottrell - Tubular ESP (Subpilot Tests at Curtiss-Wright
S.G.T.)
- Denver Research Institute - ESP Test Facility
• Acoustically Enhanced
- University of Buffalo - Acoustic Agglomerator as Preconditioner
Coupled to Cyclone
The above particulate removal projects include both laboratory scale and
subpilot scale projects. The laboratory project work is to define and verify
the concept in sufficient detail to determine overall system feasibility. The
subpilot scale projects are to establish and verify the cleanup process feasi-
bility including scale-up parameters at the subpilot scale level.
The laboratory scale projects are at various stages of completion and some of
the accomplishments to date are:
• The operation of an acoustic agglomerator contrary to previous theory
is independent of acoustic frequency.
• A two stage electrostatically enhanced GBF has proven feasible at
laboratory scale atmospheric conditions and has illustrated the
significant collection efficiency through electrostatics for PFB
applications.
108
-------
• Electrostatically charging a dry plate scrubber at atmospheric
conditions has demonstrated a substantial increase in collection
efficiency.
• Through an interagency agreement with the Environmental Protection
Agency, DOE has completed and now has available a flexible HTHP
electrostatic precipitator test facility.
These laboratory scale projects are presently preparing for hot pressurized
testing from which the overall feasibility of specific particulate removal
concepts for future PFB applications can be determined.
The subpilot scale projects involving the electrocyclone, the fixed GBF, ESP
aiid ceramic felt bag have progressed to the stage of final design with PFBC
testing scheduled to start the latter part of 1981. Thus, there are no results
to discuss at this time. Three of the four devices as indicated will be tested
in the small gas turbine facility located at Curtiss-Wright in Woodridge,
New Jersey, which is being supplied as Government Furnished Equipment by DOE
(Figure V). The fixed granular bed filter pursued by Westinghouse is sched-
uled for testing at their Waltz Mills PFBC simulator.
The future work for PFBC particulate removal will focus on the completion of
bench scale and subpilot testing (See Figure VI). The most promising system
will be scaled-up to pilot size (10 to 20 MWe) for pilot plant test verifica-
tion in the 1984-1985 time period. A possible pilot plant test site is the
Curtiss-Wright pilot plant presently under construction at Woodridge,
New Jersey (See Figure II).
Alkali Removal
Alkali vapor removal to the partial PPM levels required for reliable turbine
operation is extremely difficult. One of the basic problems is that the
presence of chlorine and other contaminants limit the alkali vapor pressure
in equilibrium with the alumino - silicates. Thus, the alkali vapor level
still remaining with in-bed gettering at 1700°F appears to be an order of
magnitude above the desired 0.02 PPM turbine tolerance level. Measurements
downstream with positive filters indicate that the gas leaving the PFB com-
bustor contains 10 to 20 PPM of alkali vapor consisting of both sodium and
potassium compounds. The combustor gas also contains sulfates, chlorine,
vanadium and lead presenting a somewhat different gas chemistry problem than
that with petroleum based fuel containing alkali. Our present program goal
of 0.02 PPM of alkali has been adopted from experience with petroleum fuel.
Due to the complexity of alkali vapor removal to the desired low concentration
levels for reliable turbine operation, the following process options for
alkali removal are being pursued.
• Fixed sorbent beds utilizing activated bauxite or diatomaceous earth
for alkali absorption on solid surfaces.
• Addition of alumino-silicates to the combustion bed as chemical
gettering agents for equilibration of the gas phase alkali with the
alumino-silicate to form solid compounds.
109
-------
• Absorbent smoke ^ubmicron sized sorbents) to increase the effective
sorbent surface to alkali gas ratio for increased reaction rates.
• Surface ionization wherein alkali vapors are ionized by a platinum
catalyst surface and then the alkali ions are directed to an electric
field from which they are focused for collection on a collector elec-
trode surface.
Along with these alkali control options, modeling and experimental studies
are being conducted to improve the understanding of the PFB combustion stream
gas composition. The characterization of the alkali formation processes are
being studied through the application of chemiluminescent and laser floures-
cent techniques.
The following projects are underway in the alkali removal area:
Alkali Removal Concepts
• Combustion Power Company - In-bed Chemical Gettering with Alumino-
Silicate
• General Electric - Sorbent Smoke and Surface Ionization
• Argonne National Laboratory - Fixed Sorbent Bed
• Westinghouse - Fixed Sorbent Bed
• APT - Fixed Sorbent Bed
The following alkali characterization/formation and modeling projects are
underway:
• National Bureau of Standards
- Data Base and Experimental Studies of Gas Phase SCL Interactions
- Alkali Release from Fly Ash, Slag and Glass
• Georgia Institute of Technology
Spectroscopy to determine alkali gas species.
• Aerodyne
- Alkali Release from Model Organic Materials and Computer Modeling
of Gas Stream Chemistry
• MIT
Thermodynamic Measurements of Slag Properties
• University of California
110
-------
- Laser Studies of Alkali Species in Flames
• Morgantown Energy Technology Center
- Alkali Surface Enrichment and Release from Minerals
The primary results to date from these ongoing alkali projects are:
• Laboratory tests using sorbent beds at high temperature, high pressure
conditions using a simulated gas stream have measured removal of up to
99.9 percent of the alkali metal compounds present. (It should be
noted that the gas stream used was in the absence of sulfur, chlorine,
vanadium and particulate dust contaminants.)
• Preliminary assessment of the alumino-silicate smoke approach has
demonstrated that a monolayer coverage can be obtained on the smoke
absorbent surface at alkali concentrations low enough to prevent tur-
bine corrosion. Therefore, there is optimism that this increased sur-
face area approach can economically achieve the desired low levels of
alkali vapor.
• Combustion stream modeling to determine the condensed alkali species
has indicated through calculations that the major percentage of the
alkali content in the stream would exist in the solid form as alkali
alumino-silicate compounds.
In conjunction with the alkali removal tasks, the following options are being
considered:
• Blade replacement as periodic maintenance.
• Operation with cooled blades.
• Operation at lower turbine inlet temperatures.
• The development of improved corrosion resistant blade coatings or a
combination of these approaches in order that practical utility
operation is achieved.
Instrumentation Efforts
To date an unresolved key bottleneck in the development of particulate and
alkali control devices is the lack of reliable accurate instruments for
measurement of the subject contaminants at the concentrations of interest.
Instruments are needed that can accurately measure and determine in-situ
particulate loadings and size distributions as well as alkali concentrations
to the desired concentration levels. State-of-the-art impactor devices are
not calibrated to the desired temperatures for accurate measurement. Con-
sidering these instrument problems, several programs are underway to develop
equipment to isokinetically sample PFB effluent to determine the particle
loading and distributions. In addition, the direct readout instruments such
as laser interferometer analyzers are being developed. Such noncontact
111
-------
instrument projects have been pursued under the PFB combustion program with
Leeds and Northrup and Spectron but a significant amount of baseline cali-
bration testing has to be completed to establish instrument confidence.
On the recent 1000 hour CURL test, the "standard" particulate measuring
technique was the isokinetic fixed sampler probe followed by a small high
efficiency cyclone with a total backup filter. Size distributions were
determined with a coulter counter. DOE for the forthcoming subpilot tests of
the prototype scale particulate cleanup devices at Curtiss-Wright is developing
an improved version of this high efficiency cyclone collection concept which
will provide real-time particle-mass measurements. Real-time measurements
will be possible with tapered element oscillating microbalances (TEOM) (one
per cyclone) coupled with a multistaged cyclone train of progressively small-
er diameter. The multi-cyclone portion of the instrument was tested on the
1000 hour CURL test. For the first time, the sampling/sizing classification
will be done at PFB temperature and pressure with real-time mass measurements
per each size classification.
For alkali measurement, the atomic emission spectrometer appears to be a
promising alkali detector and will be used for the PRDA tests. This instru-
ment was recently tested on DOE's Morgantown Energy Technology Center's
gasifier and performed satisfactorily. Due to the difficulty of obtaining
accurate alkali particulate measurements, the PFB combustor cleanup program
is pursuing the following instrument related programs. A significant amount
of PFB combustor testing will be required before these instruments can be
reliably used for future demonstrations and commercial applications.
Instrumentation Projects
• Ames Laboratory - Development of alkali monitor (Atomic Emission
Spectrometer)
• Brookhaven National Laboratory - Direct Sampling of Particulate Mass
Loadings
• Sandia Laboratories - Multi-staged Cyclone Train Coupled to a
Tapered Element Oscillating Microbalance
• Spectron - In-Situ Determination of Particle Sizing (laser
Interferometer)
• Inhalation Toxicology Research Institute - High Temperature, High
Pressure Impactor Measurement
Key Technical Issues
DOE's PFBC cleanup program has a substantial challenge to meet in achieving
its contaminant removal goals for alkali and particulates. To achieve these
removal goals accurate instruments have to be developed in parallel to verify
the results. Some of the key technical issues that have to be confronted and
resolved include:
112
-------
• Measurement of alkali vapors and participates (particle-mass concen-
trations) accurately to very low concentrations is necessary for
adequate contaminant control. Obtaining meaningful measurements will
require instruments that provide real-time in-situ measurements.
• Instruments have to provide for uniform baseline measurements in order
to make direct efficiency comparisons of one cleanup concept versus
another.
t The PFB combustor gas stream presents a sticky environment which may
prevent the barrier devices that will meet NSPS from operating with-
out plugging. Also, the electrostatics may be detrimentally effected.
• Particulate control devices that are capable of meeting NSPS must be
extremely efficient while maintaining an acceptable overall system
pressure drop.
• The amount of alkali metal compounds that appear in the vapor state
versus the percent that is tied up in a solid compound is to
some degree still unknown. System tradeoffs regarding alkali metal
compound removal versus improved turbine materials remain to be
performed.
• Can turbine cladding and coating materials be developed that are
resistant to alkali corrosion at the full (1700°F) temperature
capability of the PFB system?
• Tolerance to erosion/corrosion of gas turbine alloys and coatings in
a PFBC gas environment have not been obtained for exposure durations
of interest to utility operators.
• Can aluminio-silicate gettering tie up a sufficient amount of the
alkali to be considered as a means for alkali control?
• Can cyclones be realistically scaled up to meet the turbine erosion
limits or can banks of appropriate sized cyclones especially for the
first generation PFBC power system (considering valve and lock hopper
problems) be developed with acceptable performance and economics on
large commercial PFB utility systems?
• The use of actual PFB combustors for the basic and applied research
associated with cleanup development is very expensive. Therefore, a
better understanding of the overall PFB combustor gas stream chemistry
is needed to allow for closer simulation of the gas stream so that
less expensive testing of prototype devices is possible. The impor-
tance of realistic simulation is especially true for understanding
the alkali problem and developing an effective absorbent since chlorine,
sulfur, lead, and vanadium will drastically effect the physical as
well as the chemical adsorption/absorption rates.
113
-------
Fig. 1 Advanced Environmental Control
Technology Program Structure
s
IOP
PROCESS
STEAM
—
WASTE
MCA1
DOII1B
£=
STACK
•» WATCH
COMPRESSOR
EXPANDERS ALTERNATOR
Fig. 2 Simplified Flow Diagram
of C.W. Pilot Plant
TYPICAL
COMBUSTOR
BED EFFLUX
INEFFICIENT REMOVAL
OF 5 MICRON OR SMALLER
PARTICLES PREVENTS
MEETING NSPS
.03 LB/10" BTU
POSITIVE
FILTERS
NSPS
.0001
5 10 100
PARTICLE SIZE-MICRONS
1000
CURL Combuster
Bed Efflux
4 Stages
30" Cyclones
(Projected Scale Up)
Projected Turbine
C.W. 1000 hr SGT
3 Stages 18"
lOyna, Ducon, Aerodyn
w/ Clean Air Sheath)
11 X
CURL 1000 hrLUj
STAL LAVAL! ,4
IVT-20",15",ST-13"I
0001
5 10 100
PARTICLE SIZE MICRONS
1000
Fig. 3 PFBC Challenge to Meet NSPS
Fig. 4 "State of Art" Cleanup
114 W/Cyclones
-------
SC3T/PPB TECNMOI_CK3Y UNIT
Fig. 5
DOE'S PFB CLEANUP TECHNOLOGY
PROGRAM
CY 79 CY 80 CY 81
TASK
TURBINE CLEANUP
PFBC SYSTEMS
(LAB & SUBPILOT
TESTS)
ISORBENT TESTS
£) MODELING)
(ALKALI &
PARTICULATE
MEASUREMENT)
FY79
CAT I A
'-
FY80
CAT 1
•„
WARDS
LAB T
LAB 1
FY81
AWARDS
STING
ESTING
CY82 CY83 CY 84 CY 85 CY 86 | CY 87
FY82
FY83
CAT 1 TESTS
COMPLETED
7
r
*
\
\ SUBPILOT
\ TESTS
) UVO
^T DEC
FY 84
FY85
PILOT PLANT
TESTS
SION FOR 1
V O
ILOT
/ PLANT CLEANUP TRAIN
/ W/SELECTED INSTRUMENTS
/
FY86
FY87
INITIATE
DEMO PLANT
A
Fig. 6
115
-------
TM
THE CYCLOCENTRIFUGE
AN ADVANCED GAS/SOLIDS SEPARATOR FOR COAL CONVERSION PROCESSES
By: P.R. Albrecht, J.T. McCabe
Mechanical Technology Incorporated
Latham, New York
William Fedarko
U.S. Deparment of Energy
ABSTRACT
Various coal conversion processes under development for the emerging United States Energy Program
have focused renewed attention on the need for efficient gas cleanup systems. One approach being funded
by the U.S. Department of Energy (DOE) is the development of a high-velocity particulate separator called
a Cyclocentrifuge. The separation principle is based on centrifugal force generated by a bladed rotor
located in a cyclone-shaped vessel. A 1000-acfm laboratory model tested at one atmosphere and 100°F is
described. Test results include velocity profiles and particulate separation efficiency measurements.
The design of a high-temperature unit is also discussed and plans to test this unit using 1000°F, 250-psi
synthesis gas from a coal gasifier are outlined.
INTRODUCTION
This paper describes the continuing development of a centrifugal device for separating fine particu-
late matter from a gas. As the United States coal program gains momentum, the status of particulate
control technology will become increasingly more important, not just because maintaining air quality is
generally desired, but also because particulate control is essential to the economic viability of certain
coal conversion processes. In one process, for example, hot, pressurized, coal-produced, fuel gas is
used to fire the gas turbine combustor in a combined-cycle power system. For reasonable gas turbine
blade life, this application requires a degree of particulate separation that is substantially greater
than can be achieved by an ordinary cyclone. The present alternative to mechanical separation for coal-
derived fuel gas is to scrub the gas with an aqueous solution and accept the energy loss associated with
cooling the gas.
Although there are other areas where high-efficiency separation of fine particulate matter can make
a beneficial contribution towards accelerating the commercial viability of some coal conversion pro-
cesses, the cleanup of low-Btu fuel gas for a gas turbine application was designated as the immediate
target for this program.
The work described herein follows a feasibility study [1] in which the role of gas centrifuge tech-
nology in coal conversion processes was examined. This study concluded that particulate separation was an
area in which some of the concepts from centrifuge technology could be applied to improve the performance
of existing equipment. As part of the study, a design approach following centrifuge concepts was estab-
lished and, to show feasibility, preliminary design calculations and drawings were made for a separation
device based on a combined-cycle electric power application. The selected application required stringent
particulate removal from 125,000 scfm of coal-produced, low-Btu gas at 1000°F and 250 psia. In this
study, the projected separation efficiency was estimated to be about 85% for a 1-micron ash particle.
This percentage was a significant improvement over existing cyclones. The separation device was named a
Cyclocentrifuge because it consisted of a cyclone containing a centrifuge. Basically, the concept was
founded upon the following critical assumptions:
• The normal vortex in a cyclone could be accelerated by appropriate aerodynamic design of a bladed
rotor to produce a stable, high-strength vortex.
• The separation potential of the high-strength vortex would not be nullified by secondary mixing
effects.
• The high-strength vortex flow would provide high efficient separation of particles in the 1-
micron range.
• The rotating assembly could be designed and manufactured to provide high on-line availability
while operating in a hot, pressurized coal-produced gas stream.
• The required rotor power would be acceptably low and could be extracted from the cleaned process
gas efficiently, using a reaction turbine fitted to the rotor.
Phase II of the program, which is the subject of this paper, provided engineering data on all of the
above items, except those relating to hot, pressurized testing. The data were based on a program scope
that included:
116
-------
• The design and construction of 1100-scfm engineering prototype Cyclocentrifuge for cold-flow
testing
• The design and construction of a laboratory test loop including a measurement system for per-
formance evaluation
• A test program focusing on measurements of velocity distribution and separation efficiency.
Phase III of the program, which is now in progress, will provide further information on Cyclocentri-
fuge efficiency at ambient conditions as well as on a unit to be tested at the DOE Morgantown Energy
Technology Center (METC) facility. The ongoing Phase III program is intended to expand the data produced
to date to include manufacturing considerations for high-temperature service, on-line particulate moni-
toring, parametric cold-flow laboratory tests, and tests using the full output stream from the METC
gasifier.
GENERAL DESCRIPTION
Figure 1 shows the cross section of the Cyclocentrifuge. The particulate-laden gas enters a cyclone-
shaped vessel containing a hollow, high-speed, bladed rotor. The rotor acts to uniformly accelerate the
vortex flow causing the centrifugal force on the particles to be increased significantly compared to a
conventional cyclone. The centrifugal force drives the particles to the outside wall where they flow
downward with the boundary layer into a collection hopper. A small percentage of the process gas is used
as continuous blowdown flow to help prevent dust retainment in the hopper. The remainder of the cleaned
process gas passes through a reaction turbine mounted in the hollow rotor. This turbine serves a twofold
purpose; it drives the bladed rotor and it removes the swirl from the gas. Swirl removal recovers energy
and helps minimize particle deposition on the inside of the rotor. The cleaned gas passes through the
hollow rotor, then exits the Cyclocentrifuge.
Variable Speed Drive
(For Test Arrangement
Only; Not Shown)
-.-- *---^ Upper Support
^Struts
- Buffered
Labyrinth Seal
i Gas Inlet
LABORATORY MODEL DESIGN
The laboratory model was designed to the
following specifications:
• Inlet Flow
• Air Temperature
• Rotor Speed
• Pressure
500 - 1300 acfm
(14.2 - 36.8 m3/min)
45 - 100"F
(7.2 - 37.7°C)
5,000 - 15,000 rpm
(523.6 - 1570.8 rad/sec)
9-14 psia
(3136.1 - 4878.4 N/m2)
As shown in Figure 1, the rotor is an assembly
consisting of the swirl augmentation rotor, drive
turbine, and bearings.
The swirl augmentation rotor is a subassembly
consisting of the bladed centrifuge shell, upper
and lower shell support struts, and three nose
cones. The bladed centrifuge shell has eleven blade
rows with 16 blades in each row. Each blade has a
double circular arc profile and is 0.923 in. (2.39
cm) high with a maximum thickness of 0.1 in. (0.25
cm) and a chord length of 0.577 in. (1.47 cm).
Each blade row is oriented with respect to the
rotor axis to achieve a smooth, uniform increase in
the gas tangential velocity. The cord of each
blade in the first blade row makes an angle of 70°
with respect to the "rotor axis. The angle in each
successive blade row decreases by 7" making the
cords of the blades in the last row parallel to the
rotor axis.
The centrifuge shell has an outer diameter of
6 in. (15.24 cm), a shell thickness of 0.15 in.
(0.38 cm), and is 15.375 in. (39.05 cm) long.
The total stress in the shell at 15,000 rpm is
about 25,000 psi (8.7 x 10 x 10 N/m2) which comes
from two principle sources. Hoop stress contributes
about 65% of the total stress while the remainder
is bending stress caused by the blades. The maxi-
mum stress occurs at the outer surface where the
blade is attached to the shell.
Gas Discharge
Upper
Support
Bearing
Upper Nose Cone
Lower Nose Cone
Bearing
Lower Bearing
Support
Stationary
Nose Cone
Vortex Core'
Envelope
Centrifuge Rotor
Swirl Augmentation
Blade Cascade
ii Lower Support
Struts
Drive Turbine
Cyclone Shell
Cone
Dust Collection
Hopper
"Dust
Fig. 1 Cross Section of Laboratory Cyclocentrifuge
117
-------
The upper and lower support struts connect the stub shafts to the bladed shell via eight spokes.
The spokes are designed for high axial and radial stiffness and are shaped to minimize flow disturbance.
The lower support strut is electron beam welded to the rotor shell. The upper support strut is bolted to
permit access to the rotor interior. This strut also contains a buffered labyrinth seal to prevent the
dirty inlet gas from leaking into the cleaned gas discharge plenum.
The bearings are 25-mm, grease-packed, shielded, deep-grooved ball bearings axially preloaded to
give a radial stiffness of between 2 x 105 and 5 x 105 Ib/in. (1.8 to 4.4 x 10& N/m). This bearing
stiffness coupled with the structural stiffness of the bearing support places the first critical speed
207. above the design speed. Figure 2 shows the critical speed map for this design.
Rotor drive power is provided primarily by a
reaction turbine that extracts power from the
process gas. Thus, the Cyclocentrifuge rotor does
not require an external drive which would involve
additional bearings and pressure seals.
"• Maximum Operating Speed
i 2
Range of Bearing
Stiffness
Design
Flexible Rotor Supported
by Flexible Bearing
1st Critical
Speed
I I I I i i i II
I i I I 11
I I i i I i i
"•tO4 2 5 »10J
Bearing Stiffness (Ib/ln.)
Fig. 2 Critical Speed Map - Flexible Rotor Assumed
For speed control during testing a variable
speed drive motor is connected to the rotor through
a quill shaft.
The stationary assembly of the Cyclocentrifuge
is composed of the cyclone vessel, dust collection
hopper, discharge gas plenum, and bearing supports.
The overall height of the cyclone shell is 7 ft 9
in. (236.2 cm) with a maximum diameter of 16.25 in.
(41.3 cm).
The cyclone vessel consists of two parts: a
cylindrical upper part referred to as the inlet
casing and a lower part which has a cylindrical and
conical section, known as the cyclone. The inlet
casing houses the rotating assembly and includes
the tangentially arranged inlet gas duct.
The bearing supports are located on both ends
dynamically designed to minimize flow restriction
while maintaining the high structural rigidity
necessary for meeting the r'otor critical speed
criteria. The bottom support includes two types of
struts. The outboard struts are angled to achieve
zero lift for the average downward velocity outside
the stiffening ring while the inboard struts are
angled to achieve the same condition for the upward
velocity inside the stiffening ring.
The upper support serves to separate the
discharge plenum from the inlet casing. The struts
in the center of the upper support are not angled
because the discharge flow vector from the rotor is
parallel to the rotor axis.
The dust collection hopper is located at the
bottom of the cyclone and is continuously purged
with 2 to 3% of the inlet flow.
HIGH-TEMPERATURE, HIGH-PRESSURE DESIGN
The design shown schematically in Figure 3 is
for the maximum operating conditions of the METC
gasifier. These conditions are as follows:
e Volume Flow 450 acfm (12.74 m3/min.)
• Pressure 350 psig (1.2 x 10 N/m2)
• Gas Temperature 1000°F (537.7°C)
• Rotor Speed 9,000 to 13,000 rpm
(942.5 to 1,361.4 rad/sec)
The basic design principles are the same as
the laboratory model, so only the differences be-
tween the two designs will be discussed here.
of the inlet casing. The support struts are aero-
n
Buffered
Labyrinth Seal
Upper Support
Strut
Swirl Augmentation
Rotor
Lower Support/'
Strut
Bearing Water
Return
Ceramic Liner
Insulation
Reaction Drive
Turbine
Bearing Water
Supply
Journal Bearing
Fig. 3
Schematic of Cyclocentrifuge Designed
for the METC Gasifier
118
-------
The bearings have been changed from grease-packed ball bearings to water-lubricated hydrostatic
bearings. The rotor is still supported between two journal bearings to maintain a critical speed that is
greater than the design speed. The thrust bearing has been located at the top of the rotor to promote
rotor stability. The number of blades rows has been reduced to compensate for a longer bearing span.
The blades have been coated with a wear-resistant coating.
A ceramic liner has been added to the cyclone shell to reduce wear, and insulation has been added to
maintain the gas temperature above the condensation point of the tars contained in the gas stream.
The high-temperature Cyclocentrifuge is currently being manufactured, and testing is scheduled to
begin in late 1981.
LABORATORY TESTS
Aerodynamic Test Method
The aerodynamic tests were made by measuring velocity profiles at the axial stations located along
the Cyclocentrifuge shell, as shown in Figure A, Each station was fitted with a traversing mechanism,
carrying a pitot tube. A two-degree-of-freedom traversing mechanism was used and carried a three-hole,
wedge-shaped, two-dimensional pitot probe. This probe permitted measurements of the total and static
pressures as well as yaw angle.
Aerodynamic Test Results
Although the aerodynamic tests were conducted to map the general velocity profiles within the Cyclo-
centrifuge, the effects of flow and rotor speed on tangential velocity were of particular interest be-
cause this velocity component is a measure of the separation force. To measure this relationship, the
inlet flow was varied between 500 and 1,000 acfm, and the rotor speed between 5,000 and 15,000 rpm for a
range of flows. During the testing, the breaking characteristics of the motor drive did not allow for
testing of all combinations. Most notably, the low rotor speed, high-flow combinations could not be
achieved.
Station 1
£ 400
"I-300
i*?
Ill200
10 o"lOO
12 o
•
. Data Not Taksn
-
.
• i i i i i
400
II200
-"
100
983 acrm
15,000 rpm
4.0 4.5 5.0 5.5 6.0 6.5 7.0
Radius (In.)
4.0 4.5 5.0 5.5 6.0 6.5 7.0
Radlu* (In.)
a. 40°
,1^300
J> ?
_« Jzoo
Ha«
oi 100
i
" a
635 acfm
15,000 rpm
^*"^^""'™™^™*^""^ta^l""l"^"™^«^
.
1 1 1 1 1 1
4.0 4.5 5.0 5.5 6.0 6.5 7.0
Radlu* (In.)
737 acfm
15,000 rpm
400
-,300
| 200
100
0
651 acfm
15,000 rpm
-
-
1 1 1 1 I.I
4.0 4.5 S.O S.S 6.0 6.5 7.0
Radlu* (In.)
4.0 4.S 5.0 5.5 6.0 6.5 7.0
Radius (In.)
1.0 4.5 5.0 5.9. «.0 6.5 7.0
Radlu* (In.)
Fig. 4 Schematic of Measurement
Station Locations
Fig. 5 Tangential Velocity Profiles Adjacent to the
Swirl Augmentation Rotor for Various Flow Rates
119
-------
Figure 5 shows the velocity profiles at the stations in the neighborhood of the swirl augmentation
rotor for various flow rates. From this set of curves it can be seen that the tangential velocity in-
creases as tne flow passes axially through the centrifuge blades, and that a flow turndown ratio of 0.25
does not cause a substantial reduction in the maximum velocity at the end of the rotor.
A velocity map (Figure 6) taken near design point condition shows that the maximum tangential velo-
city of 350 ft/sec at Station 5, which is near the last blade row of the swirl rotor, is exceeded at
Station 7 in the cone, and is almost equalled at the bottom of the cone at Station 9. Thus, the high
centrifugal force that was generated by the swirl augmentation rotor acts to produce and maintain par-
ticle separation throughout the length of the Cyclocentrifuge. The plots on Figure 6 showing axial
velocity indicate the boundary of the rising core of cleaned gas. At Station 7, the radius of the core
is 3 in., which is equal to the radius of the hollow rotor that acts as the exit pipe. During the aero-
dynamic tests, the boundary of the core was studied and found to be a stable surface at design condi-
tions. At Stations 7 and 9, the points of maximum tangential velocity, which correspond to the maximun
separation force are inside the rising clean gas core. Considering the particle residence time, the
magnitude of the force field, and the stability of the flow, a high percentage of particles ranging
downward to a diameter of approximately 0.5 micron should reach, and be held in, the cyclone boundary
layer.
Sh.ll
Flow
Upward Flow
100
50
'
' -50
-100
; t
1 2 ,
" vX^
r IDof
L. Cyclone
x— • p Shell
/3 4 5 6 7
Radius (In.)
J Downward
Row
Axial Velocity
(ft/sec)
I S o S 1
Upward Flow
. ,, IDof
T L- Cyclone
1 JT.-
,{,,,. 11
\
/234S
Radius (In.)
(Downward
Flow
6 7
1234567
Radius (In.)
Fig. 6 Gas Tangential and Axial Velocity Profiles at Design Flow
The pressure drop through the Cyclocentrifuge is shown in Figure 7 as a function of rotor speed for
different flow rates. The dashed lines in Figure 7 indicate the calculated pressure drop using the
method described in Reference 1, with measured variables and an empirical correction factor for the
= 4
in
0.
i3
Pressure Di
o -* w
— Experimental Measurements
____ Theoretical Predictions
Using Empirical Coefficient
Row - 500 acfm
, , -T ~" i i i i
24 6 8 10 12 14 16
Rotor Speed (thousands of rpm)
3 -
2 -
1 .
r a 4
jiL
Flow - 800 acfm a 3
__^ |,
~~~ | 1
i i i i i i | | i Q
Flow = 1000 acfm ^^*f
" Predicted Pressure
Drop without Empirical
Correction
1 1 1 1 1 1 L 1
Rotor Speed (thousands of rpm)
Fig. 7 Pressure Drop Versus Speed
8 10 12 14 16
Rotor Speed (thousands of rpm)
120
-------
Cyclocentrifuge. The point shown in the bottom curve of this figure is the calculated design point value
for pressure drop without empirical correction. Note that about half the pressure drop would occur in a
conventional cyclone operating at the same flow conditions. The other half of the pressure drop is taken
across the rotor drive turbine.
Dust Tests
Methylene Blue Method
The dust tests were divided into two sections. The particulate generation and measurement method-
ology for the first section were selected on the basis of: (1) ease of generating a controlled mono-
dispersed particle size distribution, and (2) rapid, accurate data reduction.
Several well-established techniques exist which will produce an aerosol of repeatable size distribu-
tion and predictable concentration. The Collison nubulizer/methylene blue dye aerosol generation system
is one of the best established of these techniques and years of experience with this system have been
used to formulate British Standard BS-2577 for its use [2]. A detailed description of the design of a
Collison nebulizer is given by May [3],
The Collison nebulizer is an atomizer which is used to generate relatively uniform water droplets
from a reservoir which contains dissolved methylene blue dye. The droplets from the nebulizer are mixed
with a stream of moisture-free nitrogen gas in a drying column where the water is evaporated, leaving a
methylene blue dust suspended in the gas. This dye is then injected into the test loop, and isokineti-
cally fed impactors are inserted upstream and downstream of the Cyclocentrifuge to sample the aerosol.
The collected particles are then washed from the various stages of the impactor with a small quantity of
water, and the strength of the blue dye solution measured photometrically. A concentration of one part
dye in 10 parts water will produce sufficient coloration to be detected, allowing efficiencies to be
determined by comparing the upstream and downstream mass of dye from equivalent-size isolation stages of
the collector.
The total amount of aerosol injected during this section was 6.6 mg/hr. Preliminary tests with a
cascade impactor indicated that well over 3/4 of the mass was in the aerodynamic size range below 3
microns in diameter, and no particles larger than 6 microns in diameter were detected. Similar results
have been obtained during subsequent test runs. In general, 90% of the injected dye is within ± 0.5
micron of a prescribed value that depends primarily on the percent concentration of methylene blue in the
water solution.
The filter holders, or May cascade impactor, are mounted at the exit of the ball valve, and a frac-
tion of the inlet flow is sampled during each test to determine either the mass concentration when the
filter is used or the size distribution of dye particles when the impactor is used. During preliminary
testing, inlet flow was sampled for periods of 3, 6, 12, 24, 12, 6 and 3 minutes to determine the stabil-
ity of the system. These tests indicate that 6 minutes of sampling is sufficient to accurately determine
the rate of injected dye.
At the completion of the intial test sequence, a 0.4-micron-pore filter was substituted for the
usual paper filter at stage five of the impactor. The Collisons were operated at their normal injection
rate, and a two-minute exposure of the impactor stage was completed at design flow. The cascade impactor
stages were then removed and, rather than washed for color analysis, they were examined microscopically
using a Leitz Ultra Pak with total magnification of 800X. Transmitted light, with slight crossed polars
in conjuntion with supplementary incident light, showed the methylene blue particles as easily identi-
fied, slightly flattened, shiny spheres. A Porton eyepiece was used to determine the equivalent diameter
of particles larger than 0.4 micron in diameter by traditional counting techniques. The optical proce-
dure was used to verify the calibration of the more rapidly executed photometric analysis.
A similar cascade impactor sampling system is located in the cylindrical duct downstream of the
centrifuge. A three-probe integrated sampling system is used to verify the absence of stratification or
other similar phenomena. Because of the low particle loading downstream of the Cyclocentrifuge, a mini-
mum of 30 minutes of sampling time is required to obtain a measurable concentration of dye. The mass of
dye measured on each stage of the inpactor at inlet and discharge is compared, and the efficiency of the
Cyclocentrifuge is then determined for each size class.
Methylene Blue Test Results
The dust tests were conducted at the design point flow while varying the rotor speeds and blowdown
flow. The rotor speeds at which tests were conducted were 10,000, 12,500, and 15,000 rpm. The blowdown
flows were approximately 2 and 3% of the main stream flow.
The tests at each combination of parameters were repeated ten times, and the average of the ten runs
was used as the datum point for each combination. The results of the individual tests were generally
found to be repeatable within 10% of average.
121
-------
Figure 3 shows the measured grade-efficiency curve for the Cyclocentrifuge operating at design poin
conditions with a 3% blowdown flow, and using methylene blue dust suspended in air. The vertical line
of the curve represent the standard distribution ot the ten test runs. The dashed line shown on thi:
curve is the theoretical grade efficiency for this dust calculated by the method described in Refereno
1.
10C
J80
> 60
I 4°
g 20
0
Theoretical
Prediction
V,
Test Data
Rotor Speed = 15,000 rpm
Stream Flow = 1,000 acfm
Blowdown Flow = 3%
Inlet Pressure - 12.85 psla
Inlet Temperature - 75° F
Particle Matter: Methylene Blue Oust,
s.g. - 1
SUMMARY OF SEPARATION EFFICIENCY
FOR PARTICLES 1/2, 1, AND 2 MICRONS IN DIAMETER
12345
Particle Diameter (microns)
6
Fig. 8 Grade-Efficiency Curve for Design Point
Conditions
Rotor
Speed
rpm
10,000
12,500
15,000
2% Blowdown Flow
Separation Efficiency
(Percent)
1/2
micron
22
25
25
1
micron
44
47
47
2
micron
80
80
80
3% Blowdown Flow
Separation Efficiency
(Percent)
1/2
micron
32
38
54
1
micron
55
60
73
7
micron
82
85
92
Microscopic examination of the impactor plates revealed that some of the larger particles (>4
microns) tended to bounce through the impactor and get trapped in the final filter. Attempts to correct
this situation by coating the collection plates of the impactor led to erroneous readings on the photo-
meter. The microscopic analysis of the collection plates indicated that some of the particles that should
have been intercepted by the 0.5-micron plate had passed through to the final filter. Thus, the photo-
metric analysis of this plate indicated a higher separation efficiency than actually existed. These
points are shown on the 0.5-micron line of the separation efficiency curves given on Figure 9. However,
the plotted curves statistically correct this situation using the microscopic data. The data given on
these curves generally show an increase in particle separation efficiency as the rotor speed and blowdown
flow increase. The above table shows this trend clearly.
The effect of rotor speed and blowdown flow on the overall efficiency, and on the efficiency for all
particles less than L.4 microns, is shown in Figure 10. The upper curve shows a general increase in
overall efficiency with increasing rotor speed and blowdown flow. The lower curve, however, shows a
different trend. In this curve, when 3% of the process gas flow is used for continuous blowdown, the
100
e°°
,60
.|40L
| / Row * 1,000 acfm
*" I Rotor Speed = 15,000 rpm
/ Slowdown Row - 3°/«
123456
Particle Size (microns)
100
80
I6"
5 20
0
--*—
;5 /
/ Row = 1,000 acfm
. / Rotor Speed - 12,500 rpm
/ Blowdown Flow = 3%
123456
Particle Size (microns)
iu 20
" I / Row = •
/ Row = 1,000 acfm
. Rotor Speed = 10,000 rpm
/ Blowdown Flow * 3%
123456
Particle Size (mlcrona)
100
- 80
| 60
|40
S20
0
x—
/
• * /* Row = 1,000 acfm
/ Rotor Speed = 15,000 rpm
- / Blowdown Row - 2%
/
1 2 3 4 5 6
Particle Size (microns)
100
— ao
? 60
u
•§ 40
=
20
0
---*—
//'
'I ^
1 / Row = 1,000 acfm
_ / Rotor Speed = 12,500 rpm
/ Blowdown Row - 2%
/
7
12345
Particle Size (mlcrom)
100
~ 80
d
S-60
•§ 40
S20
0
-*— — —
/
. J / Row * 1,000 acfm
/ Rotor Speed * 10,000 rpm
— i Slowdown Flow ~ Vn
1 23456
Particle Size (microns)
Fig. 9 The Effects of Rotor Speed and Blowdown Flow Rate on Grade Efficiency
122
-------
100
£80
I-
I40
5 20
0
3% Slowdown Flow
2% Slowdown Flow
9 10 11 12 13 14 15 16
Rotor Speed (thousands of rpm)
100
££ 80
o-o
clency on Par
tan 1.4 micror
S S
|^ 20
I
°.
-
3% Blowdown Row
i -"
» * —
2% Blowdown Flow
-
i i i i
13 14 15 1
Rotor Speed (thousand, of rpm)
Fig. 10 Rotor Speed Versus Efficiency for 2% and 3% Slowdown
efficiency increases the rotor speed up to the speed test limit. But when the blowdown flow is decreased
to 2%, the efficiency shows a maximum value at about 13,500 rpm. The decrease in efficiency after 13,500
rpm can probably be attributed to increased particle reentrainment in the blowdown hopper with increased
vortex strength.
Figure 11 shows the calculated grade-efficiency curves for high-pressure, high-temperature gasi-
fication conditions based on extrapolating the cold-flow test results. The theoretical performance
curves for the Cyclocentrifuge and a high-efficiency cyclone are the curves published in the Phase I
Final Report [1]. The curve labeled "Based on Cyclocentrifuge Cold-Flow Test Results" is the solid line
curve from Figure 8 corrected for the conditions listed. The theoretical curve and the performance curve
based on the cold-flow test results are in close agreement, and are substantially better than a theoret-
ical high-efficiency cyclone, based on Reference 4.
Fly Ash Tests
The second set of dust tests used fly ash recovered from an electrostatic precipator used by
Rochester Gas and Electric on one of their coal—fired boilers. The fly ash was injected into the loop
from a fluidized bed at a rate of about 1.5 grains per actual cubic foot (3.5 grams/nr') or about 3500
ppm. The inlet and discharge were sampled isokinetically and the sample was collected on filters for
subsequent analysis. The samples were weighed for total mass, and centrifugal sedimentation techniques
were used for size analysis.
Fly Ash Test Results
Testing with fly ash is still in progress and grade efficiency data is still subject to some refine-
ments. However, Figure 12 shows average grade efficiency for several test runs at 13,000 rpm rotor speed
(approximately 85% design speed). Also shown on this figure is the methylene blue results for similar
conditions. Further testing at design ^peed is in progress.
99.97 -
100
90
80
_ 70
^60
8 50
I 40
* 30
20
:
0
/
»
r
/
/
/
nT
if'
Nj/
/
hec
s*^
/
ret
cal
,'
Pra
^^
die
--
Ion
--
for
"
Cy
elm
•' High Efficiency Cyclone*
Based on Cyclocentrifuge
"Cold Flow Test Results
•Co
rrec
:sn
rlfu
!•'"
i
.—
-'-
—
• —
—
""
Pressure = 250 psla
Temperature = 1000° F
Viscosity = 1.82 « 10^ f)^c
Particle Density - 100 ^-
ad
or vitcoalty and Particle Density
1 2 3 45 6 7 8 9 10 11 12 13 1
Particle Diameter (microns
t 15 16 17 18 19 2
10'12 0,3 0.4 0.6 0.8 1.0 2.0
Particle Size (microns)
Fig. 11 Grade Efficiency Curves for Gasification
Conditions
Fig. 12 Grade Efficiency for Fly Ash and
Methylene Blue Tests at 13,000 rpm
123
-------
SUMMARY
Laboratory tests were made that show the vortex flow in a conventional cyclone can be accelerated by
an aerodynamically designed rotor to a stable tangential velocity that is three to five times greater
than a high through-put conventional cyclone. The tests also showed the separation efficiency was in-
creased to the level that would be expected from the increased centrifugal force and that the penalty for
this efficiency increase was increased pressure drop. Design analysis completed to date indicates a
durable mechanical design for the planned 1000°F tests is feasible.
ENDNOTES
1. McCabe, J.T. Centrifuges for Coal Conversion and Related Processes. Phase I, Feasibility Study.
Mechanical Technology Incorporated, DOE Report FE-2428-10, Contract No EF 76-C-01-2428.
2. British Standard For Methylene Blue Tests For Respirator Canisters. British Standards Institution,
B52577; 1955.
3. May K.R. The Collision Nebulizer: Description, Performance and Application. Aerosol Science,
Vol. 4, 1973.
4. Kotch, H. , and Licht W. New Design Approach Boosts Cyclone Efficiency. Chemical Engineering, Nov
7, 1977.
124
-------
DEMONSTRATION OF THE USE OF CHARGED FOG IN CONTROLLING
FUGITIVE DUST FROM LARGE-SCALE INDUSTRIAL SOURCES
By: Edward T. Brookman
Project Engineer
TRC - Environmental Consultants, Inc.
Wethersfield, CT 06109
R.C. McCrillis, B.C. Drehmel
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
A unique device for the control of particulate emissions works on the
principle that most industrial pollutants acquire an electrostatic charge as
they are dispersed into the air. If this charged airborne material is ex-
posed to an oppositely charged' water fog, the charges act to enhance the
contact between the particulates and the fog droplets, resulting in rapid
agglomeration and particle fallout. A device that generates charged fog has
now been substantially developed and is being offered commercially by The
Ritten Corporation.*
TRC-Environmental Consultants, Inc. has been contracted by EPA/IERL-RTP
to test The Ritten Corporation's Fogger IV on several large-scale fugitive
dust sources. This paper discusses the initial test at a sand and gravel
operation and then describes current and future test sites. Preliminary
test results are presented in terms of percent reduction in total suspended
particulates (TSP). Changes in fogger effectiveness due to variations in
operational parameters are discussed. Initial tests indicate overall fogger
efficiencies of approximately 70 percent.
INTRODUCTION
A spray of fine water droplets is a well-known means of airborne dust
removal. Unfortunately, water sprays are not very efficient in removing
dust from ambient air. One means of improving the efficiency of water
sprays is by applying a charge to the spray that is opposite in polarity to
that of the dust to be suppressed. Most industrial pollutants and naturally
occurring fugitive dusts acquire an electrostatic charge as they are dis-
persed into the air. Exposing this charged airborne material to an oppo-
sitely charged water spray enhances contact between the particulates and the
water droplets. After contact is made, the wetted particulates agglomerate
rapidly and fall out of the atmosphere.
*The Ritten Corporation, 40 Rittenhouse Place, Ardmore, PA 19003
125
-------
Charged sprays can be improved further by atomizing the water droplets
to produce a fog. The fineness of the particles enhances their charge
carrying capabilities in the spray. Furthermore, Hoenig (1) has demon-
strated that the greatest effectiveness is obtained when the size of the
water droplets is similar to that of the dust particles to be controlled.
Lastly, less water is required when fog is used, thus reducing operating
costs.
A device capable of producing this fine spray and applying a charge to
it is known as a charged fogger. The charged fogger is intended primarily
for fugitive dust sources that cannot reasonably be controlled using conven-
tional means such as hooding. Such sources include materials handling oper-
ations (transfer points and conveyors), truck and railroad car loading and
unloading, front end loaders, ship loading, grain silos, and mining opera-
tions. The charged fog concept has been applied to operations as small as a
hand grinder and as large as a quarry.
Although the charged fog concept has been widely applied to industrial
souces of fugitive dust, little data is available on fogger control effi-
ciency. To obtain such data, the Industrial Environmental Research Labora-
tory of the Environmental Protection Agency at Research Triangle Park, North
Carolina EPA/IERL-RTP contracted TRC-Environmental Consultants, Inc. to con-
duct a full scale demonstration of a charged fogger on several industrial
fugitive emission sources. In particular, EPA/IERL-RTP was interested in
testing the largest fogger, "Fogger IV," manufactured by the Ritten Corpora-
tion, on several sources in the iron and steel and sand and gravel
industries.
Sources considered for testing included materials transfer, conveying,
grinding, crushing, and truck and railroad car loading and unloading. Re-
quirements for a test site included isolability from other dust sources,
availability of necessary utilities, relative difficulty of control by other
methods, representativeness to the general industry, relatively continuous
operation, and fairly heavy dust production to facilitate sampling.
Following numerous visits to iron and steel and sand and gravel sites,
several suitable sources were selected for field testing the charged fogger.
The source chosen for the first test was the primary rock crusher operation
at a sand and gravel site in Connecticut.
DESCRIPTIONS OF SITE AND TEST EQUIPMENT
Figure 1 is a plot plan of the primary crusher operation showing the
locations and dimensions of the various structures. The site and test
equipment are described in the following subsections.
Test Site
The inital fogger test site was a primary rock crusher. Approximately
100 dump trucks per day, each carrying loads of approximately 45 Mg (50
tons) of quarry rock (basically basalt) mixed with dirt, back up to the
crushing pit to unload. Unloading times vary from 30 to 60 seconds,
126
-------
J
12m
PRIMARY
CRUSHING J_
PIT
BREAKER ARM
PRIMARY
CRUSHER
15m
COMPUTER
CONTROL
BUILDING
7m
CEMENT
BLOCKS
PAVED AREA
EDGE OF
EMBANKMENT
12m
CRUSHER
CONTROL
SHED
CRANE FOR
DISLODGING
JAMS AND
REMOVING
OVERSIZE
MATERIAL
Figure 1. Primary crusher plot plan
127
-------
depending on conditions in the pit. The pit itself is roughly 8 meters long
and 6 meters wide. The crushing is done by a Superior 4265 gyrotory rock
crusher. There is a two story computer control building north of the
crushing pit, a control shed to the east, and a large paved area to the
south. All approach roads and areas around the buildings and pit are paved
and kept reasonably clean through frequent sweepings and waterings.
Fugitive dust emissions result from the dumping and crushing operations.
The truck unloading is the primary source of dust with the major portion
coming from dust boil-up at the rear of the pit. There is also dust at the
rear of the truck during the dump. The crushing procedure itself also pro-
duces dust, but to a much lesser degree than the unloading process.
Charged Foggers
Two identical foggers were designed especially for TRC and EPA by the
Kitten Corporation. Ritten's standard Fogger III was modified and upgraded
to allow for variations of parameters. The final configuration, "Fogger IV,"
is shown schematically in Figure 2.
In the generation of the charged fog by the Fogger IV, water is atomized
by compressed air and ejected from a nozzle. As the fog leaves the nozzle,
it passes through an induction ring where either a positive or negative
charge, depending on the nature of the dust, is applied to the spray. A
flow of air around the nozzle, provided by a centaxial fan, projects the fog
toward the dust source. A control panel, on the back of the fogger, allows
for fogger operation and parameter variability.
The requirements for and capabilities of the operational parameters are:
n
• Air supply to nozzle - A compressed air supply of 5.6-8.8 kg/cm
(80-125 psi) is required. For the tests the air was supplied by a 2
hp compressor. The air flow through the nozzle is variable from 0
to 11.3 m3/hr (0 to 400 scfh). ' '
• Water flow - The water supply to the fogger should be around 3.5
kg/cm (50 psi) which is typical "shop" water pressure. The water
flow through the nozzle is variable from 0 to 151 1/hr (0 to 40 gph).
• Power - The foggers require a power supply of 230 volts, single
phase, 60 Hz. Current requirements do not exceed 35 amps.
• Centaxial fan - The fan, driven by a 5 hp explosion-proof motor,
operates at a maximum of 79 m/min (2800 scfm). The maximum
output air velocity is approximately 3048 m/min (10,000 fpm). The
fan flow rate is variable from 0 to 100 percent of capacity.
• Charge per droplet - Assuming an average droplet size of approxi-
mately 60 Vim, the average number of elementary charges per droplet
was calculated to be approximately 8 x IQ1* for 75 1/hr (20 gph)
water flow. This can also be expressed as a charge/mass ratio of
0.11 uC/g.
128
-------
BELT GUARD
^JUNCTION BOX (MOTOR)
lOsI
22.9 cm
t-1—5HP MOTOR
15.2 cm INDUCTION RING
AIR ATOMIZING NOZZLE
NOSECONE
15 A\ WATER
cm-f
V
BELT DRIVEN CENTAXIAL FAN
-48.3 cm
I-WEATHER-PROOF CONTROL
PANEL ENCLOSURE
CONTROL PANEL
\v* y—CONTROL BOX
AIR AND WATER INPUT Jfr
CONNECTION PORTS"
230 VAC RECEPTACLE
230 VAC MAIN
DISCONNECT SWITCH
CONTROL CABINET—
*-LIFTING EYE FOR SKID JACK
Figure 2. Schematic of The Ritten Corporation's Fogger IV.
-------
• Flow spectra - Two different flow nozzles were used for the tests,
both manufactured by Delavan.* One nozzle produced a conical spray
of droplets estimated to be in the 50-70 ym size range; the other
had a heavier flow capacity and produced a conical spray of droplets
estimated to be in the 60-80 ym size range. A third nozzle, which
produces a flat spray, was not available for the tests, but may be
used later.
The two foggers were tested at various locations around the pit to
determine the arrangement for optimum dust control. Placement was also de-
pendent on wind direction. The most effective positions seemed to be at the
rear corners of the pit with one fogger aimed at the dust boil-up and the
other, at the rear of the truck.
Sampling Equipment
The equipment used for particulate measurements included seven high-
volume samplers (hi-vols) and a wind recording system. The hi-vols were
manufactured by Misco Scientific** and had automatic flow control, enabling
the mass flow rate to be held constant regardless of filter loading, atmos-
pheric conditions, and line voltage changes. Two of the hi-vols were fitted
with Andersen Model 7000 Size Selective Inlets (SSIs) which are designed to
remove all particulates larger than 15 ym from the sampled air before fil-
tering the remaining particulates onto a standard hi-vol filter. Two other
hi-vols were fitted with Sierra Instruments Series 230 four—stage cascade
impactors (CIs). By using the SSIs and CIs, the charged fogger efficiency
could be examined for various particle size ranges.
The wind velocity and direction measurements were recorded using a
Climatronics Mark III wind system. Wind speed is measured with a three-cup
anemometer coupled to a light chopper. The chopper output is converted to
DC voltage and recorded on a chart. The wind direction is measured by a
wind vane coupled to a precision low-torque potentiometer. The wiper vol-
tage of the potentiometer is recorded on another chart.
The hi-vols were at various locations and in various combinations around
the pit, depending on wind direction. The locations were selected to pro-
vide variation in downwind distance.
TEST PROGRAM AND PROCEDURE
The test program consisted of 32 runs during 6 days of testing. Tests
were run with no charge, with fog only, with uncharged fog, with positive
fog, and with negative fog. Water flow, air flow, and fan speed were only
varied slightly since conditions at the crusher prevented extensive para-
meter variations. Water was provided by a tank with a small pump which
limited nozzle flow to approximately 80 1/hr. Fan speed was reduced to 80
percent of capacity to help reduce excessive dust reentrainment in the pit.
* Delavan Mfg. Co., 811 4th St., West Des Moines, IA 50265
** Misco Scientific, 1825 Eastshore Highway, Berkeley, CA 94710
130
-------
The sampling procedure was essentially the same for each test. Upon
arrival at the test site, the wind recording system was set up and the wind
direction determined. The hi-vols were then positioned in a sampling array
downwind of the crushing pit. The foggers were positioned to control the
dust cloud while not spraying directly into the samplers. Once the equip-
ment was positioned, the pre-weighed hi-vols were placed in the samplers.
The samplers were turned on simultaneously just prior to the first truck
dump of a predetermined sequence of trucks (typically eight trucks provided
sufficient material for sampling purposes). For the runs with the foggers
in operation, the foggers were also turned on at this time and adjusted to
the predetermined fogger operational parameter conditions. After the last
truck of the sequence had dumped into the pit and crushing was completed,
the samplers and foggers were all stopped and the filters removed. At the
end of the day, all of the filters were returned to TRC's chemistry labora-
tory where they were subsequently desiccated and weighed.
PRELIMINARY RESULTS AND DISCUSSION
The majority of the test runs at the primary crusher, numbers 7-31, were
completed before the final filter weights were available from the chemistry
laboratory. Examination of the data revealed several important factors. In
almost all cases, the TSP levels, as measured by the various samplers,
showed increases above the uncontrolled levels when uncharged fog was ap-
plied to the crushing operation. Although this result was unexpected, fur-
ther analysis revealed the problem: the fans in the foggers which create
the airflow that projects the fog toward the dust source are so powerful
that they were actually creating an artificial wind effect. The uncontrolled
dust plume was subject only to the ambient wind; whereas, the controlled
plume was being "directed" radically by the fogger air jets. This discovery
produced the need for a final series of tests wherein the uncontrolled base-
line TSP levels were recorded with the fans on with no water added.
Another concern that developed was with the intermittent nature of the
truck dumps. In some cases, Tight trucks would unload within 20 minutes; at
other times it would take 30 or 40 minutes. The data was therefore reduced
on a per-truck basis since the unloading and crushing times, the times when
the vast majority of the dust is produced, were essentially the same for all
dumps. The data were also adjusted slightly to account for deviations of
the actual sampler flow rates from the design flow of 1.1 m /min (40 cfm).
While the data from runs 7-31 did not reveal information regarding over-
all fogger efficiency, they did provide insight into the increase in effi-
ciency due to charging the fog. This efficiency could be further examined
with regards to particle size, distance from the pit, and charge polarity.
Not all of the runs produced usable data: the fog impinged on the samplers
during some tests.
The data from runs 32-39 were used to determine fogger efficiency with
respect to uncharged fog. The spacing of the samplers also allowed for the
examination of efficiency versus distance from the pit. This information
was then combined with the data from runs 7-31 to calculate overall fogger
efficiency.
131
-------
Attempts were made to obtain information on visibility improvement using
EPA Method 9 (visual determination of opacity). It was found that the opac-
ity of the fog was similar to that of the uncontrolled dust plume so that no
real visibility improvement was noted.
Figure 3 presents the preliminary test results from the initial charged
fogger tests at the primary crusher site. The left side of the figure shows
the percent reduction in TSP levels when an uncharged water fog was used to
control the fugitive dust. The right side of the figure shows the addi-
tional percent reduction in TSP levels when a charge was applied to the
fog. The data from this figure reveal several important results, as dis-
cussed in the following subsections.
Uncharged Fog Efficiency
Based on the limited amount of data for fan-only versus uncharged fog,
it appears that a water spray alone is approximately 30-40 percent efficient
in reducing fugitive dust levels from the primary crusher. It also appears
that this efficiency is independent of particle size and distance from the
pit.
Efficiency Increase Due to Charging of Fog
Applying a charge to the water spray reduced the fugitive dust levels
40-70 percent over those recorded using uncharged fog. There appears to be
a trend of increasing reduction with increasing distance from the pit. This
apparent phenomenon is not explainable at this time, but it may have some-
thing to do with agglomeration and particle fallout. This possible distance
factor will be examined further in future tests.
Figure 3 also shows that TSP reduction due to charging is essentially
the same regardless of the polarity of the charge applied to the spray.
This indicates that the dust cloud contains a mixture of particles, some
negative and some positive. This is consistent with the findings of other
researchers, namely Hoenig (1) and Kunkel (2).
TSP reduction appears to be the same for the respirable size range
(<_ 15 ym), as measured with the hi-vols with size selective inlets, as
for the size range sampled with the standard hi-vol (<_ 30 ym). It was
hoped that using cascade impactors would provide additional information on
efficiency versus particle size, but the results proved unusable. Almost
all of the material collected by the hi-vols fitted with the impactors was
collected on the back-up filter, indicating severe particle bounce between
the impactor stages. Perhaps tests at sources with finer dust will yield
more useful information.
Overall Fogger Efficiency
By combining the results presented in Figure 3, it is possible to calcu-
late an overall collection efficiency for the charged foggers. Based on the
preliminary data, the use of charged fog can reduce the fugitive dust levels
132
-------
IUU
0
gao
Z m
O CJ
flee
o 5
§060
•"• WH
UJZ
CC3
z2
HI
O ^" 4.Q
oc ~
ui z
O.O
z
u.
20
^^
_ 1 I 1 L
O STANDARD HI-VOL
A SIZE SELECTIVE INLET
O _
o
^^^
Q
— —
lilt
100
o
o
UL 8O
Q
HI
Z°
1§
••* ^*^
I-I
00 6O
QO)
I-S
Z |T_
U 40
OQ
ecu
UJO
Q.OC
Z
Z 20
n
~ O STANDARD* HI-VOL ' '~
n CASCADE IMPACTOR -
u TOTAL LOADING
A SIZE SELECTIVE INLET
OPEN SYMBOLS: NEG. FOG
SOLID SYMBOLS: POS. FOG
D
0
i •
o 8 •
O A
A. A
__ o
LJ
~» ..
A
i i i i
25 5 75
DISTANCE FROM PIT (m)
1O
25 5 7.5
DISTANCE FROM PIT (m)
10
Figure 3. Percent reduction in TSP levels due to fogging of primary crusher.
-------
that result from the primary rock crushing operation approximately 65-75
percent. It is felt that this reduction could be even greater through the
use of additional foggers, wind baffles to reduce turbulence, and increased
water flow.
FUTURE WORK
The foggers are currently being tested on two different sources in U.S.
Steel plants near Pittsburgh, Pennsylvania. One source is a sinter plant
pug mill which mixes water with baghouse dust for dust suppression before
recycling the material back through the plant. Even with water addition,
there is a significant amount of fugitive dust around the source. The se-
cond source is the hot fume that results in a cast house from the filling of
a ladle car with molten iron from a cast. Following the tests at U.S.
Steel, a coke screening operation at Stelco in Hamilton, Ontario, will be
tested. At this site, coke is transferred onto a shaker screen where it is
sorted by size. The shaking results in copious amounts of dust.
The steel company tests mentioned above will most likely be completed by
this summer. Following completion of all field tests, a final report will
be prepared for the EPA presenting the results.
ENDNOTES
1. Hoenig, S.A. Use of Electrostatically Charged Fog for Control of Fugi-
tive Dust Emissions. The University of Arizona, Tucson, AZ.
EPA-600/7-77-131 (NTIS PB 276-645). November 1977.
2. Kunkel, W.B. The Static Electrification of Dust Particles on Dispersion
into a Cloud. Journal of Applied Physics. Volume 21. August 1950.
pp. 820-832.
134
-------
THE CONTROL OF FUGITIVE EMISSIONS USING WINDSCREENS
By: David Games
Project Scientist
TRC - Environmental Consultants, Inc.
Wethersfield, CT 06109
Dennis C. Drehmel
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
TRC-Environmental Consultants, Inc., was contracted by EPA/IERL-RTP to
perform a project that would provide information on the velocity
distributions downwind of windscreens. This was accomplished via field
testing of several windscreen sections erected on smooth terrain. This
paper presents the results of these tests as well as a discussion of
windscreen theory in general. A theoretical case study of the potential
reduction of emissions from a coal storage pile using a windscreen as a
control measure is also presented. The results of this case study indicate
80 percent fugitive emission reduction.
INTRODUCTION
Despite extensive controls on traditional point sources, the National
Ambient Air Quality Standard (NAAQS) for total suspended particulate (TSP)
is still being exceeded in many areas of the country. Among the reasons for
these excessive TSP levels is the fugitive dust that is generated from
non-traditional sources such as unpaved parking lots, construction sites,
exposed plant areas, and st<^-age piles. To reduce the impact of these
sources on ambient air quality, EPA/IERL-RTP has undertaken a program of
area source control technology research and development. One of the control
methods under investigation is a windscreen.
Wind-generated fugitive dust is the result of wind pressure against
surface particles overcoming the force of gravity on the particles and the
force of adhesion between the particles. For each size and type of dust
particle, there is some "threshold" velocity below which no wind erosion
will occur. Windscreens or wind barriers function as do other surface
barriers in providing wind erosion control; i.e., they take up or deflect a
sufficient amount of the wind force to lower the wind velocities to leeward
below the threshold required for initiation of soil movement.
THEORY
A windscreen exerts a drag force on the incident wind field, causing a
net loss of momentum and a sheltering effect in the lee of the screen.
135
-------
Airflow characteristics within the sheltered area vary with:
1. windscreen height and width;
2. windscreen permeability;
3. size of the elements or openings in the windscreen;
4. terrain roughness; and
5. turbulence intensity in the incident wind field and in the airflow
downwind of the screen.
The design and siting of a wind fence for a specific application should
reflect the optimum combination of the above. Typical field data presented
by Raine and Stevenson (1) indicate that an effective windscreen should give
a mean velocity reduction of 50 percent up to 10 fence heights downstream
and a maximum reduction of 70 to 80 percent 1 to 5 fence heights downstream
(measurements taken at 0.5 fence heights above ground).
Windscreen Height and Width
The downwind extent of sheltering is typically reported in terms of the
number of equivalent fence heights. Figure 1 depicts isotachs for a 4.9
meter high windscreen (50 percent porosity) that is 122 meters long. The
data was collected by Radkey and MacCready (2) for an incident wind of 17
m/sec. Using this case as an example, wind velocity is reduced by 50
percent or greater for a distance of 55 meters downwind. Doubling the
height of the fence will double the distance downwind for the 50 percent
isotach.
The fence width determines the lateral extent of the sheltered area.
The downwind isotachs are essentially parallel to the windscreen, with the
exception of edge effects which reduce the effectiveness of the windscreen
locally. Edge effects consist of displacement of the adjacent flow and the
development of turbulent eddies.
Windscreen Permeability
Air that passes through the windscreen elements is referred to as "bleed
flow." Air that is displaced upward and passes over the fence is referred
to as "displacement flow." As the permeability of a windscreen is reduced,
the "bleed flow" through the windscreen decreases and the drag force
increases. Also, there is a greater upward deflection of the incident
wind. As the permeability approaches zero, a region of large-scale
separated flow in the lee of the windscreen develops, turbulence in the
separation zone intensifies, and stagnation occurs downstream in the "bleed
flow." The lower the permeability of the wind break, the closer behind it
lies the stagnation point. The mean flow pattern in the vicinity of the
stagnation point is poorly defined. With a solid fence, there is a
well-defined recirculating eddy in the separation zone. A less permeable
windscreen may cause a greater reduction in mean wind speed; however, the
136
-------
(0
QC
UJ
H
Ul
Z
g
uj
z
9
6
3
0
100%
80%^
70% 7
-10H
VERTICAL SECTION
ALONG < OF FENCE.
100%
10H 20H 30H 40H
DISTANCE IN FENCE HEIGHTS
50H 60H
60
30
(0
cc
UJ
I-
III
I-
z
UJ
X
UJ
flC
u-30
-60
110%
GROUND PLAN
WIND READINGS 0.4M
ABOVE THE GROUND-)
5?% 70% 80% 90%
60%
100%
110%
-10H
50H 60H
0 10H 20H 30H 40H
DISTANCE IN FENCE HEIGHTS
NOTE: VELOCITIES ARE % OF UPSTREAM VELOCITY AT THE SAME HEIGHT
Figure 1. Wind velocity pattern above a mown field during a 17 m/sec wind
blowing at right angles to a 4.9 m high wood fence 122 m long of 50%
porosity.
137
-------
greater turbulence in its wake may make it less effective for overall wind
protection than a more permeable one.
With a more permeable fence and consequently higher "bleed flow"
velocity, there is less shear in the flow at the fence top and thus weaker
streamwise momentum transfer from the "displacement flow" back into the
sheltered region. As a result, there is less reduction in mean wind speed
and lower turbulence intensity in the near wake, but a slower recovery to
the upstream condition in the far wake. Above a geometric permeability of
35 to 50 percent, a zone of stagnated flow and large scale eddying is no
longer evident in the leeward flow.
Size of the Elements in the Windscreen
The high shear initiated at the fence top is the dominent source of
turbulent velocity fluctuations in the leeward flow, even with the most
permeable fence. However, the size of the elements in the windscreen
significantly affects the turbulent structure in the "bleed flow." For a
windscreen of a given geometric permeability, the turbulence intensity and
scale at a given location downstream increases with the size of the elements
in the barrier; i.e., with the coarseness of the permeability.
Terrain Roughness
The smoother the terrain on which a windscreen of given height and
permeability is erected, the greater will be the reduction in mean velocity
in the leeward flow. It has been found that with increasing upstream
terrain roughness and more turbulent approach flow, the smaller is the zone
of reduced wind velocity downstream of the wind break. Also, experimental
data from field and wind tunnel tests show that drag increases with an
increase in the fence height to the effective terrain roughness ratio.
Turbulence Intensity of the Incident Wind Field
In fully aerodynamically rough, neutrally stable flow, windscreen
leeward flow patterns should be independent of approach flow velocity.
However, decreasing the turbulence intensity in the approach flow increases
the zone of reduced velocity downwind of the screen.
Turbulence Intensity Downwind of a Windscreen
At any downstream point the root mean square velocity fluctuation, u1,
varies much less with windscreen permeability than does mean velocity, U.
This indicates that higher values of u'/U behind denser fences are due more
to a large reduction in U than to higher u1 levels. Therefore, regions of
high turbulent intensity tend to be regions of low mean velocity. Also, the
high shear initiated at the fence top is the dominent source of turbulent
velocity fluctuations, even with the most permeable fence. For a given
approach flow, u'/U tends to be slightly higher with less permeable fences,
apparently due to stronger shear in the flow at the fence top.
138
-------
TESTING PROGRAM
TRC completed tests on two commercially available polyester windscreens
provided by Julius Koch, USA, Inc.* The windscreens had permeabilities of
65 percent and 50 percent, respectively. Testing was completed at the state
airport in Quonset, Rhode Island. The windscreens were erected on a flat
open field which was located adjacent to Narragansett Bay. Each fence was
1.8 meters high, with the 65 percent permeable fence 55 meters wide, and the
50 percent permeable fence, 59 meters wide.
The objective of the field testing was to document the effectiveness of
a commercially available windscreen. Measurements were also made of wind
acceleration on the slope of a hill located on the Quonset airport grounds.
Measurements were made using either Teledyne or Climatronics wind
anemometers and wind direction sensors. Data were recorded on Soltec strip
chart recorders.
Windscreen Results
Results for the 65 percent permeable windscreen are presented in Figures
2 and 3. For an average incident wind (Uo) of 3.0 m/sec, the observed
wind velocity reduction was 70 percent in the adjacent area downwind of the
windscreen. A reduction of 40 percent was observed as much as 14 fence
heights (H) downwind. A second test was completed for an average incident
wind of 5.2 m/sec. The observed reduction in wind speed adjacent to the
windscreen was only 40 percent. For an elevation of 1/2H, the downwind
distance experiencing a 40 percent reduction extends to approximately 6H for
U0 = 3.0 m/sec, and 20H for Uo = 5.2 m/sec. With increasing incident
wind speed, the area experiencing a 40 percent reduction is reduced
significantly. The decrease in fence efficiency is attributed to an
increase in the turbulence intensity within the incident wind field.
Fluctuations of the instantaneous velocity in the incident flow field was 25
percent greater for the Uo = 5.2 m/sec case.
For U0 = 3.0 m/sec, the data collected at 0.5H, 1H, and 2H downwind
shows that there is a loss of momentum in the "bleed flow" and development
of a zone of stagnation. However, the few data points collected for the
U0 = 5.2 m/sec case exhibit no apparent zone of stagnation. The increased
mixing behind the windscreen and possibly the very coarse nature of the
elements in the windscreen contribute to the elimination of the area of
stagnation.
Data for the 50 percent permeable fence (Figure 4) were collected for an
incident wind of 4.2 m/sec. The downwind area adjacent to the windscreen
shows a reduction in wind velocity of 60 percent, which is comparable to the
efficiency of the 65 percent permeable fence for an incident wind of 3.0
m/sec. However, there is no apparent zone of stagnation separating the
"bleed flow" from the "displacement flow." Also, recovery of the incident
flow strength downwind occurred relatively fast.
* Julius Koch, USA, Inc., 387 Church Street, New Bedford, MA' 02741
139
-------
W3.0
cc
H 2.7
UJ
5 2.4
h-
O
• •i
I 1-8
•h*
J 1.5
O
P 1.2
CC
£0.9
0.6
0.3
n
• FENCE HEIGHT= 1.8 METERS
• AVERAGE INCIDENT WIND = 3.0 M/SEC
• READINGS
ARE % OF AMBIENT
• 65% PERMEABLE FENCE
• TEST DATE: 7/1/80
_
.8 1.0
je/£— *•-*•» ^ .6
'XXX ^*x X
.4.4.4 vxx
x '*<+
•4
-XXX %x
.4.3.2
• X X X
.4.3.3
-XXX
.4.3.3
i i i i
.6
X
50% ISOTACH
^^^*-^
K«*^
^x
• i i
70%
"**"^t**»^'
^**^ "**^^^
^^^ ^x
X \ XX
.6 x j
\ _jo%
^x^
^
1 1 1 1 I
6 9 12 15 18 21 24 27 30
DOWNWIND DISTANCE (METERS)
33 36
Figure 2. Measured velocity reductions due to a 65% permeable windscreen
with an average incident wind speed of 3.0 m/sec.
-------
(0
oc
111
o
Hi
z
-I
<
o
B
3.0
2.7
2.4
2.1
1.8
1.5
1.2
0.9
0.6
0.3
m
1^% ^K ^K
.0 .9 .9
X X X^<*
S
/
M
tf
i H^fl-tl^
^^^^^^
•• ^^^fc
.6.6 .6
-xxx
-
-
"
,
•
•
•
•
•
**** X X
.9 .8
.8 """""^
^rf M'
*"""*^«^ .8
\ -7 ^^^
V * ^
%x -7
\
\
^
\
\
, , , 1
FENCE HEIGHT =1.8 METERS
t
AVERAGE INCIDENT WIND = 5.2 M/SEC
READINGS ARE % OF AMBIENT
65% PERMEABLE FENCE
TEST DATE: 6/6/80
*r*"**»«^^ x
9X^ g
*i
XNX
fc .9 .8 X
*««»^x x
^^
». *w ^^ ^8O%
^^ C. S
^ X
X%x^- 70% \
^ X
V \
60% \ \
\ \
\ \
%
i i i i i i i
6 9 12 15 18 21 24 27 30 33
DOWNWIND DISTANCE (METERS)
36
Figure 3. Measured velocity reductions due to a 65% permeable windscreen
with an average incident wind speed of 5.2 m/sec.
-------
2.4
2.1
H
U!
S 1.5
*•••*
I1'2
us
1 0.9
O
0.6
ur
>0.3
.4 .4 .4
"XXX
• FENCE HEIGHTS 1.8 METERS
• AVERAGE INCIDENT WfND = 4.2 M/SEC
• READINGS ARE % OF AMBIENT
• 50% PERMEABLE FENCE
* TEST DATE : 8/15/8O
.9
ISOTACH
6 9 12 15 18 21 24
DOWNWIND DISTANCE (METERS)
27
30
Figure 4. Measured velocity reductions due to a 50% permeable windscreen
with an average incident wind speed of 4.2 m/sec.
-------
The test results suggest that, in practical application, fence
efficiency in the region where "displacement flow" predominates is less than
predicted. The lower efficiencies observed are attributed to the higher
turbulence intensity observed for higher incident wind velocities and
shearing at the fence top. Both of these effects extend the penetration of
the turbulent eddies into the sheltered zone which, in turn, transfers
momentum and increases mixing.
APPLICATION OF WINDSCREEN - CASE STUDY
The effectiveness of a windscreen to control fugitive emissions from a
coal storage pile was appraised using the equations developed by Blackwood
and Wachter (3). Solutions were obtained for a 54,000 metric ton coal pile
with an 8 meter high windscreen. The surface area of the coal pile was
determined from the following relation:
S = Tc/PbH
For this equation, S is the surface area of the pile; Tc the weight of
the coal stored; PJ-, the bulk density (800 kg/m3) ; and H the pile height
(7.6 m). For the study case, S equals 8900 m2.
The incident wind velocity distribution under consideration is presented
in Table 1. Average velocities were multiplied by a factor of 1.5 to
account for wind acceleration along the upwind slope of the pile. The
particulate emission rate was estimated from the following equation:
= 336 u3 Pg S
(P-E)2
For this equation, Q is the entrainment/emission rate; u is the wind speed;
and (P-E) is the precipitation-evaporation index.
The total estimated emiosions from the coal pile using no control
measures were 26.3 metric tons. The total estimated emissions using the 8
meter high windscreen were only 5.4 metric tons. Using the windscreen
offers the potential of reducing fugitive emissions by 80 percent.
TABLE 1. VELOCITY DISTRIBUTION FOR THE COAL STORAGE PILE
Velocity
Interval
(m/sec)
0
2.2
4.5
6.7
8.9
>]
- 2.2
- 4.5
- 6.7
- 8.9
- 13.4
L3.4
Average
Velocity
(m/sec)
1.3
3.6
5.8
8.0
11.2
15.6
Percent Time
of
Occurrence
19
38
19
10
10
4
Average Velocity
x 1.5*
(m/sec)
2.2
5.4
8.9
12.1
17.0
23.7
Multiplier of 1.5 applied to account for wind acceleration along the
upwind slope of the storage pile.
143
-------
ENDNOTES
1. Raine, J.K. and D.C. Stevenson. Wind Protection by Model Fences in a
Simulated Atmospheric Boundary Layer. Journal of Industrial
Aerodynamics Vol. 2, 1977, pp. 159-180.
2. Radkey, R.L. and P.B. MacCready. A Study of the Use of Porous Wind
Fences to Reduce Particulate Emissions at a Coal-Fired Generating
Station. Prepared by Aerovironment, Inc. February 1979.
3. Blackwood, T.R., and R. A. Wachter. Source Assessment: Coal Storage
Piles. EPA-600/2-78-004k (NTIS PB 284 297). May 1978.
144
-------
THE INFLUENCE OF AGGREGATE PILE SHAPE AND ORIENTATION
ON PARTICULATE FUGITIVE EMISSIONS
By: Dennis Martin
Principal Engineer
TRC - Environmental Consultants, Inc.
Wethers field, CT 06109
ABSTRACT
While it has long been recognized that pile shape and orientation have
some influence on the amount of particulate matter emitted, quantitative
data have been lacking. This is due primarily to the inherent difficulties
of any field program which would attempt to establish such a relationship.
To correct this lack, studies were conducted in a wind tunnel to determine
quantitatively how pile shape influences emissions. It was found that the
slope of the pile played an important part with respect to the acceleration
of the wind up the front of the pile. Also, it was shown that emissions
from the top of the pile were dependent upon the orientation of the pile
with respect to the wind. The implications of this study are that signifi-
cant reductions in fugitive emissions from a pile can be accomplished simply
by changing its shape slightly and reorienting it with respect to the wind.
Estimates of the potential reduction are given.
Also, it is shown that use of water sprays, chemical stabilizers, and
wind screens can be made more cost effective by selectively applying them to
the high emission areas of the piles.
INTRODUCTION
The fugitive particulate emissions from active and inactive storage
piles can be a significant portion of the total emissions from a facility
such as a coal port, power plant, steel mill, or coal mine. As with most
other fugitive emission sources, however, past efforts to precisely quantify
them with respect to those parameters which affect their rate of emission
have met with little success. It has been only recently that factors such
as silt and moisture content have been taken into account in predictive
equations. However, large inconsistencies are still evident in the work
done by the various researchers .
As part of a research project for the Particulate Technology Branch of
EPA's Industrial Environmental Research Laboratory at Research Triangle
Park, North Carolina, TRC - Environmental Consultants, Inc. (TRC) investi-
gated several control strategies applicable to various fugitive areas
145
-------
sources. It was during this work that the aforementioned inconsistencies
were noted. Special interest was taken in the wide variation with which
researchers related air velocity to emissions. Since it had been previously
observed that pile contour and alignment probably had some impact upon the
magnitude of the emissions, it was decided that a more detailed study of
this effect was now in order.
A better understanding of how fugitive emissions are generated is now
necessary since the quantification of fugitive sources is part of many pre-
vention of significant deterioration applications and offset type situa-
tions. With various factors available in the literature for any given oper-
ation, the potential for unnecessary confrontation exists between EPA or
state enforcement officials and industry representatives. It is not unusual
for a situation to occur where enforcement will choose an emission factor
available for an operation that is two orders of magnitude higher than that
picked by industry. Through a better understanding of how emissions are
generated, more accurate factors can be obtained through a planned testing
program and such conflicts can be avoided in the future.
EMISSION FACTORS FOR STORAGE
Table 1 lists some of the more frequently referenced emission factors
for wind erosion from aggregate storage piles. Two of the factors do not
include wind speed directly in the equation. Another factor relates emis-
sions to the cube of the velocity while the final one does not include a
velocity term at all. Investigations performed in wind tunnels and in the
field have related particulate fugitive emissions from piles to the air
speed raised to the 2.1, 2.7, 3, and 9.0 powers (Reference 4).
The above variations indicate that those factors which influence the
emissions from piles are complex in nature and in interaction and are prob-
ably not accurately characterized by the parameters used in the equations.
Partially, this is due to the use of surrogates acting in place of what
might be the preferred factor. For example, it is obvious that high surface
moisture content of dust would act in a number of ways to inhibit emissions.
The surface moisture content of a material varies with its physical proper-
ties and those meteorological conditions such as rainfall, temperature, and
incident radiation levels particular to a given area. Rather than measure
the surface moisture contents of the dust during testing programs and relat-
ing the result to emissions, however, Thornthwaite's precipitation-
evaporation index Ls commonly used, as can be seen in Table 1. While there
are good and compelling reasons to use this index, it must be realized that
its use could impart considerable inaccuracy to the resulting factor. A
second reason for such variations is that the velocity was measured in dif-
ferent locations around the pile and at different heights during the differ-
ent programs. This, of course, could lead to a scatter in the results. A
third reason for variation in emission factors is that the sampling is
usually performed by the upwind-downwind technique. Due to the inherent
imprecision of this method, short-term testing programs could result in
factors which are only appropriate to the sites tested being developed.
146
-------
TABLE 1. WIND EROSION EQUATIONS
Factor
0.11
iooy
Parameter Definition
E " emission factor, Ib/ton placed
in storage
PE = Thornthwaite's index
Reference
Comments
Test performed at sand and gravel plants.
Upwind-downwind test method used.
0.05
(l.5JV235Jll5A90/
E = emission factor, Ib/ton put
through storage
s » silt content, %
d » No. of dry days per year
f = % of time wind speed exceeds 12
mph at 1 foot above ground
D - duration of storage, days
Does not include emissions caused by pile
maintenance and traffic. Iron pellet and
coal piles were sampled.
E = emissions of suspended particulate,
Ib/acre/year
e * soil erodibility
s = silt content, %
f * % of time wind speed exceeds 12
mph at 1 foot above ground
PE • Thornthwaite's index
Portable wind tunnel used to modify original
soil loss equation
iit / \3/~ \ 2 /c\°-
336 (u) (pb)' (S)
(PE)2
Q s emissions, mg/sec
u « wind speed, m/sec
Pl, = bulk density, g/cm*
S * surface area, m2
PE » Thornthwaite's index
Limited testing performed at a coal pile
-------
Having reviewed the available emission factors along with experimentally
derived relationships between emissions and velocity, it remains unclear
exactly what role wind velocity plays in the particulate emissions from
storage piles.
PARTICLE TRANSPORT
To properly understand why pile shape and orientation could have an
effect upon emissions caused by wind erosion from storage piles, it is first
necessary to understand those forces which act upon the particles causing
them to become airborne. Basically, there are four forces which act upon a
particle immersed in a moving fluid. The first force is caused by the
impact of the fluid upon the particle and is called the impact pressure or
form drag for a particle lying on the ground. The second force acts as a
negative pressure on the leeward side of the particle and is termed the vis-
cosity pressure. This pressure varies with the velocity of the fluid as
well as its density and its coefficient of viscosity. This pressure is also
termed the skin friction drag. Together, these two forces are the total
drag exerted on the particle and are symbolized by the term Fc in Figure
1. The third force acting upon the particle is the lift force (Lc) caused
by the Bernoulli effect which states that whenever the fluid velocity is
increased, as at the top of a particle, the pressure transverse to the
direction of the fluid motion is reduced. Since the air speed increases
travelling over the top of the particle, the decrease in pressure at the top
of the particle causes a lifting force. This force acts through the center
of gravity as is shown in Figure 1. The final force acting on the particle
is that of gravity.
To get the particle airborne, particle movement must first be initiated.
From Figure 1, the threshold drag acting on a particle would be:
Fc = (0.52 g D3p' - Lc) tan ^ (1)
where:
Fc = threshold drag
g = acceleration due to gravity
D = particle diameter
p' = immersed density of the particle
L- = lift force
= angle of repose of the particle with respect to the average
drag level of the fluid
It has been experimentally determined that Lc = 0.75 Fc for any size
ot spherical particle. Also, since the drag and lift forces per unit of
horizontal area acting on the topmost particle are larger than drag and lift
for the entire bed, the equation is modified to include a factor n which
expresses this ratio. Another factor, T, is introduced to express the ratio
of maximum to mean lift and drag on the particle. This factor is important
since it has been observed that the longitudinal turbulence intensity is
about 33 percent near the bed. If a normal distribution is assumed for
velocity fluctuations, then instantaneous velocities would be twice the mean
while instantaneous drag values would be four times the mean (Reference 6).
148
-------
WIND
DIRECTION
/— HEIGHT k
MEAN BED
LEVEL, Z0
Figure 1. Forces of lift, drag and gravity acting on a soil grain
in a windstream at the threshold of movement of the grain. (Ref. 5)
WIND
DIRECTION"
AVERAGE GROUND SURFACE
RESULTANT
Figure 2. Pattern of approximate pressure differences. (Ref. 5)
149
-------
Combining all of these factors and dividing by the area of the particle
(0.7854D*) gives:
_
_ 0.66 g D p1 n tan
(1 + 0.75 tan 0) T
where Tc is the mean threshold drag per unit horizontal area of the
whole bed of material. Since the wind induced surface shear stress is:
T = p U£ (3)
where :
p = density of air
U* = friction velocity
= _ U _
2.5 In (z/z0)
U = mean wind velocity at height z
zo = height where U = 0
it can be seen therefore that the force necessary to initiate particle move-
ment is directly related to the air velocity squared.
Having initiated movement, the forces acting on the particle change as
the particle moves upwards. The drag force increases with height as does
the velocity while the lift force due to the Bernoulli effect decreases
rapidly. This is shown in Figure 2. It is expected that at sites such as a
coal pile thermal drafts would help to raise the particles still higher into
the air.
In summary, therefore, a relationship between the square of the velocity
and emissions is indicated by theory.
RESULTS OF STUDIES
Since it has been shown that the air velocity is the primary force
governing individual particle transport, the question remains why such a
relationship is not more evident in the available emission factors. Seeking
to show that pile shape and orientation have an effect upon the local velo-
city field around a storage pile, two studies were undertaken. The first
study was performed in a wind tunnel and its purpose was to determine if the
velocity at the top of a model storage pile was affected by the angle of the
slope leading to it. A second part of this study was to determine if emis-
sions from the model varied according to its orientation. The wind tunnel
and measurement methods used were described previously (Reference 7).
The second study was a field program performed to measure the wind flow
over a hill. The purpose of this study was to determine if the velocity
increases noted in the wind tunnel study would also be observed.
150
-------
Wind Tunnel Study
Figure 3 shows the results of the effect of the slope of the leading
edge upon the air velocity at the top of the storage pile. In the ratio
Vp/V0, V0 is the velocity taken at the same height as the top of the
pile but 6 inches in front of it. Little or no change in the value of
Vp/V0 with increasing wind speed was noted for very shallow (10°) or
steep (90°) leading edge slopes, but marked increases were observed for
the intermediate (30°, 45°, and 60°) slopes.
To determine the effect of pile orientation upon emissions, 50g of coal
dust were sprinkled on the top of the model at each orientation, and the
tunnel free stream velocity was set to 20 mph. Although various sizes of
coal dust were used, all of the results were within about 20 percent. When
the pile was placed length-wise in the tunnel, emissions averaged 0.008
g/ft . When placed width-wise, the emissions were about 0.02 g/ft3. It
was obvious from observation that the leading edge of the pile was where
most of the erosion took place. By placing the pile width-wise in the air
stream, the area of the high erosion zone more than doubled.
Wind Flow Over a Hill
The incident wind speed (V]^) was monitored approximately 15m to one
side of the hill, at an elevation of 1m. For various locations on the hill,
wind velocity measurements (¥2) were also obtained at 1m above the surface
(Figure 4). The data recorded at each monitoring location are presented in
Table 2. In addition, average V2/V^ ratios are noted in Figure 4.
As the incident wind approached the hill, V^/V^ was observed to
decrease at the foot of the hill. On the upwind slope of the hill, V2/Vi
values increased, with a maximum of 1.6 recorded at the hill crest.
values then decreased to the lee of the hill crest.
TABLE 2. MEASURED WIND ACCELERATION ALONG THE SLOPE OF A HILL
Location, X,
M from Crest
4.90
3.25
3.25
3.25
1.38
0
0
0
0
0
-1.71
-3.20
Incident
Wind Speed (Vj.)
8.1
6,5
7.0
6.2
6.5
7.0
8.1
6.2
5.5
6.7
5.5
6.7
Measured
Wind Speed (V2>
6.6
6.9
6.5
7.8
8.5
9.6
11.6
10.4
8.6
11.6
8.1
8.7
V2/Vi
0.8
1.1
0.9
1.2
1.3
1.4
1.4
1.7
1.6
1.7
1.5
1.3
Average
V2/Vi
0.8
1.1
1.3
1.6
1.5
1.3
151
-------
1.5
1.4
1.3
1.2
1.1
1.0
V0 = 6.3 mph
i i
10 30 45 60 90
SLOPE, DEGREES
0
a
o
>
^
a
>
1.9
1.4
1.3
1.2
1.1
1.0
1.5
1.4
1.3
1.2
1.1
1 n
i i i i
v'
V0= 11.1 mph
i i i i
10 30 45 60 9(
SLOPE, DEGREES
i i i i
; / " '^
S
VQ =15.6 mph
i i i i
10 30 45 60
SLOPE, DEGREES
90
Figure 3. Relationship between velocity-at top of pile (Vp) divided
by upstream velocity at same height (Vo) vs. angle of slope of
leading edge.
152
-------
Ln
U)
v,/v.=:i.e
V2/V,81.5 JL
V2/V,:s1.1
V2/Vf =0.8
—T
-3.20
———f——
-1.71 0 1.38
X, METERS-*
PREFERENCE SENSOR WAS LOCATED IN A
FLAT AREA 15 M TO ONE SiDE OF THE HSLL.
3.25
4.9 5.6
v2/v,=si:o
REFERENCE
SENSOR*
Figure 4. Location of sensors for monitoring wind velocities over a hill.
-------
CONCLUSIONS
From theoretical consideration it was shown that the square of the velo-
city is related to the initiation of particle movement and subsequent trans-
port. From the results of our study it was shown that the orientation of
the pile could affect the emissions by as much as a factor of 2.5. Also, it
was observed that the air will accelerate up the sides of these piles making
velocities at the top 20 to 50 percent higher than anticipated. These
results could explain why emission factors developed from field studies do
not correlate well with V^ since this relationship is masked by the
effects of pile shape and orientation. A further result of this study is
the possibility that the existing emission factors could be modified to
account for the orientation effect. If it is assumed that the field studies
which developed the data for these factors were performed with the samplers
placed parallel to the length of the pile then the factors re:>rv-.sent a maxi-
mum emission rate potential. If the prevailing wind direction is perpendic-
ular to the length of the pile then the emission factor should be able to be
reduced by about 60 percent.
Control techniques applicable to storage piles can be better applied if
the effect of shape and orientation are taken into account. Windscreens
placed on top of the pile would control the acceleration effect and reduce
or eliminate the zone of high erosion. Watering or chemical stabilization
should be concentrated on the edge of the pile most affected by the wind.
The slope of the pile in the prevailing wind direction should be limited to
less than 10°.
REFERENCES
1. U.S. Environmental Protection Agency. Compilation of Air Pollutant
Emission Factors. Third Edition. AP-42 (NTIS PB 275525). August 1977.
2. Cowherd, C., Bohn, R., and T. Cuscino. Iron and Steel Plant Open Source
Fugitive Emission Evaluation. Midwest Research Institute. EPA-600/2-
79-103 (NTIS PB 299385). May 1979.
3. Cowherd, C., Cuscino, T., and D. Gillette. Development of Emission
Factors for Wind Erosion of Aggregate Storage Piles. APCA Paper No.
79-11.1 Presented at t-'ne 72nd Annual Meeting of the Air Pollution Con-
trol Association. June 1979.
4. Blackwood, T.R. and R.A. Wachter. Source Assessment: Coal Storage
Piles. Monsanto Research Corporation. EPA-600/2-78-004k May 1978.
5. Chepil, W.S. and N. P. Woodruff. The Physics of Wind Erosion and Its
Control. Advances in Agronomy, Vol. 15, pp 211-302. 1963.
6. Lyles, L. and R. K. Kraus. Threshold Velocities and Initial Particle
Motion as Influenced by Air Turbulence. Transactions of the ASAE, Vol.
14, pp 563-566. 1971.
7. Martin, D. and D. Drehmel. Control Methods for Fugitive Area Sources.
In: Proceedings, Fourth Symposium on Fugitive Emissions: Measurement
and Control, New Orleans, LA, May 1980. TRC - Environmental Consul-
tants, Inc. EPA 600/9-80-041. pp 402-415, December 1980.
154
-------
SPRAY CHARGING AND TRAPPING SCRUBBER
FOR FUGITIVE PARTICLE MISSION CONTROL
By: Shui-Chow Yung and Seymour Calvert
Air Pollution Technology, Inc.
4901 Morena Blvd., Suite 402
San Diego, CA 92117
Dennis C. Drehmel
Particulate Technology Branch
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
The control of fugitive process emissions (FPE) with Spray Charging
and Trapping (SCAT) scrubber was evaluated both theoretically and experi-
mentally. The SCAT uses air curtain and/or jets to contain, convey, and
divert the FPE into a charged spray scrubber.
Experiments were performed on an 8000 cfm bench-scale spray scrubber
to verify the theory and feasibility of collecting fugitive particles with
charged water spray. The effects of charge levels on drops and particles,
nozzle type, drop size, gas velocity, and liquid/gas ratio on collection
efficiency were determined experimentally. The results of the experiments
and the comparison between theory and data are presented.
An air curtain was developed for conveying the FPE to the spray
scrubber, deflecting the crosswind, and containing hot buoyant plume.
The design and air flow field for the air curtain are presented.
NOTE; Published in the Journal of the Air Pollution Control
Association, 30:1208-11, November 1980.
155
-------
IMPROVED STREET SWEEPER FOR CONTROLLING URBAN INHALABLE PARTICULATE MATTER
By: S. Calvert, H. Brattin, S. Bhutra, R. Parker
Air Pollution Technology, Inc.
4901 Morena Blvd., Suite 402
San Diego, CA 92117
D. Drehmel
Particulate Technology Branch
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
Dust emissions from paved roads are a major source of urban inhalable
particulate matter. A.P.T. is conducting an experimental program to develop
design modifications which can be used to improve the ability of municipal
street sweepers to remove inhalable dust particles from the street.
A commercial regenerative air sweeper has been purchased and modified.
Major modifications include a charged spray scrubber for fine particle col-
lection and a gutter broom hood to help contain redispersed dust particles.
Design information and preliminary test data are presented.
INTRODUCTION
Dust emissions from paved roads are a major source of urban air pol-
lution. Draftz (1978) estimated that 40 to 70% of the total suspended par-
ticulate matter in many urban areas comes from dust particles redispersed
by road traffic. Pitt (1978) estimated that city streets can contribute
from 5 to 50 yg/m3 of particulate matter to the urban air.
The existing types of street sweepers include broom, vacuum, regenera-
tive vacuum, and water flushing sweepers. These generally have low-to-moder-
ate efficiency for removing inhalable particulate matter (IPM) from streets
and test data scatter widely. Buchwald and Schrag (1967) sampled the air
inside the driver's cab and determined that there is a serious exposure to
respirable dust. Closing the windows and the use of a cyclone separator for
dust collection had small effect on respirable dust concentration in the cab.
It is clear that urban street dirt can cause air pollution, water pollution
from rain water runoff, and an occupational hazard for the sweeper operator.
156
-------
When evaluated with regard to IPM removal, each of the conventional types of
sweepers has a significant deficiency, as indicated below.
Type of Sweeper Deficiency
Broom Disperses IPM while sweeping.
Vacuum Either emits IPM or must clean a
large volume of vent air. Gut-
ter broom disperses IPM into air.
Regenerative vacuum Disperses IPM while sweeping.
Water flush Moves IPM to gutter and causes
water pollution.
OBJECTIVE
Under E.P.A. contract, A.P.T. has conducted a research and development
program with the primary objective of developing practical means for improv-
ing the efficiency of a suitable street sweeper for the removal of IPM from
urban streets.
A.P.T. evaluated the problem and proposed to develop a street sweeper
which would use a regenerative air flow system, gutter broom hooding, and
a low-volume scrubber on a vent air stream. The use of a vent air stream
would permit a positive inward flow of air from the sweeping area into the
sweeper body rather than allowing this dusty air to be dispersed. The appli-
cability of the SCAT (Spray Charging and Trapping) system was to be investi-
gated.
The SCAT system uses combinations of air curtaining, hooding, and spray
scrubbing (with or without electrostatic augmentation) to capture and retain
particles for subsequent disposal. Specific circumstances dictate the SCAT
features which are used in each case.
Additional objectives of the program were to:
1. Build a modified street sweeper and demonstrate its capabili-
ties for urban street sweeping.
2. Develop and use methods for evaluating sweeping efficiency.
APPROACH
The general approach of the program was consistent with the premise
that a suitable scrubber could be designed if one knew what air flow rate
and particle collection capability were required. In other words, much more
was known about scrubber design than about street sweeping in terms of IPM
157
-------
control parameters. Consequently, the main effort was placed on determining:
1. Background information on street sweeping.
2. The street sweeper best suited for study.
3. A tentative set of design criteria to use for designing the
experimental street sweeper, including:
a. Air flow rates required to control dust emissions.
b. The total air flow rate which has to be scrubbed.
c. Particle size distribution and particle concentration
in the uncontrolled effluent air.
d. Utilities limitations.
4. Design concepts for the modifications needed for IPM control.
5. An experimental procedure for determining the efficiency of
removing IPM from urban streets.
A limited effort was expended on charged spray scrubbing and scrubber
design. Some experimental work was done to confirm earlier experiments and
to evaluate concepts for atomizers and drop chargers which could produce
smaller charged drops than used in previous work. The further development
of a mathematical model for particle collection by charged spray also re-
ceived some effort.
RESULTS
A literature search and interviews of qualified persons were conducted
to determine:
1. The nature of dirt on paved streets, including information
on the sources of particulate matter, the methods of deposi-
tion, and various methods of removal.
2. The prevalent types of street sweepers available, their cost,
and the potential for improving them.
3. Methods for sampling and analysis of the IPM on the street
and in the air.
There is abundant literature on street dirt and its contribution to air
and water pollution. There is, however, very little on the subject of street
sweeper efficiency as a function of dust particle size. Brookman and Martin
(1979) present an extensive literature review and recommendations for re-
search areas. Some of the works which provide useful quantitative informa-
tion are those of Axetell and Zell (1977), Cowherd et al. (1977), Draftz
(1978), Pitt (1978), Sartor and Boyd (1972), and Sehmel (1973).
158
-------
At the time the research began, broom sweepers were the type most com-
monly used. However, the trend appeared to be toward the use of regenerative
vacuum sweepers, so this was the type selected for our use.
as:
Street sweeping effectiveness has been determined in several ways, such
1. Ambient air was sampled upwind and downwind from the street
to determine the contribution of dust from the Street (e.g.,
Cowherd et al., 1977, and Pitt, 1978).
2. Street dirt loading was measured by vacuum-cleaned sample
areas before and after sweeping (e.g., Cowherd et al., 1977).
3. Street dirt loading was measured by a combination of sweeping
and water-flushing sample areas before and after street sweep-
ing (e.g., Sartor and Boyd, 1972).
Preliminary Design Criteria
Observation of street sweeper operation clearly showed that gutter
brooms (GB) and leakage from the "pickup hood" were the major sources of
IPM emissions from the regenerative vacuum sweeper (RVS). In order to design
the vent air scrubber and the GB hoods, the required air flow rate was esti-
mated from industrial ventilation practice to be about 28 to 56 m3/min (i to
2 x 106 ftVmin) vented through the scrubber.
The sweeper has a blower which the manufacturer rated at 280 to 340 m3/
min (10 to 12 x 10s ft3/min). Our tests showed that it can move about 50%
of the rated capacity. This proved to be adequate for venting about 30 m3/
min while providing effective dust pickup in the hoods.
The SCAT scrubber uses preformed spray drops to collect the fine par-
ticles, and an entrainment separator to collect the drops and large parti-
les. The particles may be charged in an ionizing section before reaching the
spray section, and the spray may be charged with induction electrodes near
the nozzle water outlet. -
The scrubber needs about 0.9 £/m3 (3 gal. water/1,000 ft3 air handled).
For convenience, this translates to a 0.76 m3 (200 gal.) capacity to allow
once-through operation, and a reasonable time between refills. An existing
0.12 m3 (30 gal.) tank plus an additional 0.64 m3 (170 gal.) auxiliary will
provide the required water. The scrubber comes equipped with a water pump
which can provide 0.9 £/min (3 gpm) at 3,450 kPa (500 psi). For the first
experimental sweeper, it was assumed that the existing water pump would suf-
fice and that once-through water use would be acceptable.
159
-------
Experimental Apparatus
The basic sweeper was analyzed for air flow and dirt conveyance. Ex-
tra ducting was needed to transport the dirt from the RVS recirculating air-
flow system to the scrubber. Hoods were designed to contain the dust gen-
erated by the gutter brooms, and ducting was designed to carry this dirt to
an appropriate place in the hopper.
It was necessary to develop a hood arrangement that allowed normal
broom operation and also reduced fugitive broom emissions to an acceptable
level. An abbreviated hood was designed that allowed brushing on the curb
side. This hood operates with a high velocity suction airflow similar to
a vacuum cleaner, and also uses the interaction of the rotating broom and
street dirt to capture dust. The captured dust is conveyed to the main hop-
per with piping and flexible hose. A damper valve in the pipe section con-
trols the airflow from the hood to the hopper.
Several street dust samples were taken from the plumes dispersed by
broom and vacuum type sweepers. The representative particle size distri-
bution had a geometric mean diameter, dpg, of 4 ymA and geometric standard
deviation, ag, of 2. Experiments on charged spray scrubbing were conducted
to verify ana extend the results of previous studies.
Predictions of scrubbing efficiency based on the preliminary informa-
tion on street dust size indicated that a minimum of 90% efficiency could be
obtained within the constraints of the available power, space, and water.
The scrubber and accessories were designed conservatively so that
enough space was allowed for additional airflow. Other main elements, such
as the entrainment separator, were also designed so that their capacity could
be increased if the need arose.
The sampling system provides for the measurement of airflow from the
gutter brooms to the scrubber inlet. Particle size and concentration were
measured with cascade impactors. Sample ports were provided for the scrub-
ber inlet and outlet at locations at least 8 duct diameters from upstream
transitions or bends. Sample trains of the type used for EPS Method 5 were
used. Scrubber and GB airflow rates were measured with Venturi meters and
were controlled by dampers.
The positive displacement water pump was calibrated in terms of aux-
iliary engine speed. Additional utilities (such as the electrical power
supplies for sampling pumps and scrubber charging, water tank and transfer
pump, sampling personnel station, and platform for sampling trains) were
placed on a trailer pulled by the street sweeper.
Street Dust Sampling
Initial full scale operation of the sweeper and auxiliary equipment in
San Diego was used to observe the modified components in use, and to provide
more data on the nature of street dirt. Later, as the apparatus was refined,
additional samples were gathered on other streets, and eventually more were
160
-------
gathered in heavy industrial areas in Los Angeles.
Twelve sampling runs were made, and results are summarized in Table 1.
As can be seen, the concentration and particle size distribution vary greatly
from location to location. Sampling results agree with visual observations
in that dirtier streets resulted in higher particle concentrations and larger
particles.
The data in Figure 1 show the effect of vent rate on mass concentration
at the scrubber inlet. These data were taken from the same street site; only
the vent rate varied. Run 18/7 had a higher vent rate and a higher mass con-
centration than Run 18/11.
Street samples collected so far indicate a difference in characteristic
size dirt when neighborhoods are considered. The major distinction noticed
between neighborhoods is the amount of traffic; street dirt is reduced in
size by continuous traffic.
Vent Air Rate
During the initial sampling runs it was noted that "puffs" of dust
escaped from the main pickup hood, suspended dust was continuously gener-
ated by the G.B.'s, and the G.B. water sprays were ineffective for reducing
fugitive emissions. When the pickup head travels on uneven street surfaces,
dust puffs emerge from the pickup head.
These dust puffs are eliminated by opening the vent valve. This de-
creases the pressure on the discharge side of the blower and increases the
suction in the hopper. As a result, the average pressure in the pickup
head is reduced and air flows into the hood through any openings between the
rubber seals and the pavement.
To minimize the scrubber size and water consumption, the air flow to
be vented through the scrubber should be kept at a minimum. However, the
air flow should not be so low that dust puffs occur around the gutter broom
and the pickup head.
The minimum air flow needed to be vented through the scrubber to pre-
vent the occurrence of dust puffs was determined when the gutter broom hoods
were turned either ON or OFF. When ON, the air flow in each hood was main-
tained at 0.17 m3/s (350 acfm), which was the minimum required flow for sat-
isfactory hood operation.
It was found that the minimum air flow to be vented through the scrub-
ber to prevent dust puffs with the gutter broom hoods OFF was about 0.33 nr/s
(700 acfm), and about 0.38 m3/s (800 acfm) with the gutter brooms ON.
Efficiency Measurement Method
To determine the sweeping efficiency of the sweeper, a method is needed
which measures the amount of dust that can be dispersed into the ambient air
161
-------
by the mechanisms actually occurring on streets. The mechanisms of street
dust removal are (Brookman and Martin, 1979):
1. Reentrainment (by air currents around moving vehicles).
2. Wind erosion (similar to 1, but due to natural air currents).
3. Displacement (similar to 1, re-deposition near street).
4. Rainfall runoff.
5. Street cleaning.
The first three mechanisms result in airborne dispersal of street dirt
so the sampling method should measure the amount of IPM which can be dis-
persed by these.
Preliminary experiments were performed to measure street dust density
with a brush-type vacuum cleaner, as performed by Dahir and Meyer (1974).
The dust on the street was first loosened with a brush, and then taken up
and filtered by the vacuum cleaner. The dust density (mass/unit street
area) was determined from filter weight gain.
This method was found to have several deficiencies. The vacuum cleanei
can remove dust which is deposited deeply in cracks and is not normally re-
entrained. Further, the amount of dust vacuumed increases with each pass
of the vacuum nozzle; therefore, there is no logical end point for sampling.
Consequently, the above method was abandoned and a new method devel-
oped. The new method is a vacuum cleaner and a modified pickup nozzle.
The pickup nozzle is shown in Figure 2. The nozzle is shaped to create a
uniform 97 km/hr air flow field at the street surface. The dust dislodged
by the airflow is sucked into the vacuum and filtered by the vacuum bag.
This method should simulate the dust reentrainment mechanisms.
A detailed definition of reentrainment would require a complex model
of airflow around an automobile. Assuming that the maximum air velocity
caused by the passage of an automobile will not exceed its velocity, and
that 97 km/hr (70 mph) is a reasonable maximum automobile velocity, the re-
sultant maximum reentrainment velocity would be 97 km/hr.
Wind speeds are generally much less than 97 km/hr (27 m/s), so this
velocity is conservatively high to represent the influence of wind erosion.
The displacement mechanism is essentially similar to that for reen-
trainment, and due to the same vehicular motion. The definition differs in
that the particles stay in suspension only long enough to reach the area
immediately adjacent to the street. An air velocity of 97 km/hr would ade-
quately represent this vehicle-induced mechanism.
Efficiency Data
The jet nozzle street dust sampling method was used; representative
results are shown in Figure 3, a plot of street dust density before and
after sweeping. Street dust is concentrated in the gutter area. The street
sweeper sweeping efficiency can be calculated from the results presented in
162
-------
Figure 3. The sweeping efficiency ranges from 75% to 97% for measurements
to date.
CONCLUSIONS
Results to date show that the application of SCAT to the control of
fugitive road dust emissions is feasible. The regenerative vacuum sweeper
(RVS) is ideally suited to the SCAT technique, as it allows a discharge
stream of air to be cleaned and recycled into the atmosphere. This feature
reduces the size and power requirements necessary for control of fugitive
emissions of IPM.
Gutter broom dust emissions can be substantially improved with an ad-
vanced, interactive gutter broom hood. Since most of the dirt is in the
gutter, this improvement is very significant. Additional power require-
ments for the SCAT system are minimal. Existing standard equipment pumps,
blowers, and auxiliary power are sufficient to do the job.
END NOTES
1. Axetell, D., and T. Zell. "Control of Reentrained Dust from Paved
Streets". U. S. Environmental Protection Agency, Washington, DC, NTIS.
PB 280325, July 1977.
2. Brookman, E. T., and D. K. Martin. "Assessment of Methods for Control
of Fugitive Emissions from Paved Roads". EPA-600/7-79-239 (NTIS PB80-
139330), November 1979.
3. Buchwald, H., and K. R. Schrag. "Dust Exposure in Mechanical Street
Sweepers". J. of A.I.H.I., 28, pp. 485-487, 1967.
4. Cowherd, C., et al. "Quantification of Dust Entrainment from Paved
Roadways". U. S. Environmental Protection Agency, Washington, DC, EPA-
450/3-77-027, NTIS PB 272613, July 1977.
5. Dahir, S. H. and W. E. Meyer. "Bituminous Pavement Polishing". The
Pennsylvania State University, University Park, PA. November 1974.
6. Draftz, R. "The Impact of Fugitive Dust Emissions on TSP Concentra-
tions in Urban Areas". In Third Symposium on Fugitive Emissions:
Measurement and Control, U. S. Environmental Protection Agency, Re-
search Triangle Park, NC. EPA-600/7-79-182 (NTIS PB 80-130891), August
1979.
7. Pitt, R. "Demonstration of Non-Point Pollution Abatement Through
Improved Street Cleaning Practices". Prepared for the City of San
Jose, Public Works Department, San Jose, CA. August 1978.
163
-------
Sartor, J. D. and G. B. Boyd. "Water Pollution Aspects of Street Sur-
face Contaminants". U. S. Environmental Protection Agency, Washing-
ton, DC. EPA-R2-72-081, November 1972.
Sehmel, G. A. "Particle Resuspension from an Asphalt Road". Atmos-
pheric Environment, pp. 291-309, March 1973.
TABLE 1. SUMMARY OF SAMPLING RESULTS FROM LOS ANGELES AREA
Traffic
Total
Scrubber Mass Mass Geometric
Flow Concen- Median Standard
Rate tration Diameter Deviation Street
Industry
Refinery
Bridge
Steel
plant
Refinery
Coke
plant
Coke
plant.
Steel
plant
Steel
plant
Steel
plant
Warehouse
and steel
plant
Grain
mill
Conditions
Medi uin-
heavy
Very
heavy
Light-
medium
Heavy
Heavy
Heavy
Light-
medi urn
Medi urn
Medi um-
heavy
Medi urn
Heavy
acfm
710
710
1,130
1,130
1,130
1,130
710
710
710
710
1,130
mg./DNm3
3,190
770
3,503
684
132
193
386
265
362
134
556
ymA
10.0
5.2
26.0
65.0
42.0
100.0
7.7
17.5
13.1
15.4
170.0
ag
4.5
3.5
15.3
8.1
7.1
8.7
7.0
4.9
5.0
4.3
9.4
Loading
Heavy
Heavy
Heavy
Medi urn
Light-
medium
Light-
medium
Heavy
Medium
Medi urn
Medi urn
Medium
heavy
Warehouse Medium
1,130
2,186 47.0
5.4
Heavy
164
-------
UJ
Q_
[HIGH
AIRFLOW
0.53 m3/s
10 20 30 40 50 100 200 300 500
CUMULATIVE MASS CONCENTRATION, mg/DNm3
1000 2000 300
Figure 1. Effect of vent air flow on scrubber inlet dust concentration,
165
-------
1 TO VACUUM
^COLLECTION
UNIFORM
AIRFLOW
FIELD
STREET
SURFACE
Figure 2, Uniform airflow vacuum nozzle.
166
-------
Ll.]
Ci
JUU
200
100
50
An
30
20
10
9
8
7
6
5
4
3
?
AT-
2?E
.
1
\
\
l\
CURB
i
— T h
K i . M
ij):_i :
\
1
V
\
A
i :~^\
: RIGHT
r PI
B0
^— •
f
1 ; ! '
| !
•
_r_i
—
^
i
"I
,
1
. ' i
i i . !
! '
1 1 1
' | 1
1 i ,
tfEFOR!
i~^
PATH
a
1-
E=
IAIN
100D
— 1 — •
— • — r
V^
-^
V
PIC
PA1
DEEP
AT ST
JUNCT
^^
^
— i
r4
,
!
: SWEE
1 1 LE
^G°
KUP
H >
i
1
1 i
i
, i j
r— — =
3ING E
FT
Bo PAT
^v
3sa
i 3-3
: [ V ,
^L
H:
3
I
-X-i
\
i i
GROOVE
REET/GUTTER
ION
R SWEEPING
I ' 1
1
; i
1 M ' -
3
i
'
— ni IM
— r
dl
|
-l\
MO
vuii nu.
8/12
'ENTING
1.34 m3/s
0 PARKING
,
1 | ' 1
LVLiL
iC
:/
<.
j —
1
5
V
; i I
-/
-
A
=
N
i — i — • — • — i—
I
i 1 1
III
I'll
1 , , —
0
4
1 2 3
DISTANCE FROM CURB, m
Figure 3. Street dust distribution before and after sweeping,
167
-------
A WIND TUNNEL FOR DUST ENTRAINMENT STUDIES
By: A.S. Viner, M.B. Ranade
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
E.J. Shaughnessy
Department of Mechanical Engineering
Duke University
Durham, North Carolina 27706
D.C. Drehmel, B.E. Daniels
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ABSTRACT
A wind tunnel facility has been fabricated at the Industrial Environmental
Research Laboratory (IERL) of the Environmental Protection Agency (EPA) at
Research Triangle Park, NC, to assess potential emissions from different dust
types and source configurations, and to evaluate control techniques to reduce
emissions.
The aerodynamics of the wind tunnel is designed to study the effects of
velocity profiles and turbulence on dust entrainment. A sampling protocol was
developed to measure emission rates from several model sources. The aerody-
namics of the wind tunnel, experimental procedures, and prelimianry experimen-
tal data on dust entrainment are described in this paper.
INTRODUCTION
Dust entrainment from open sources first gained major scientific attention
when wind erosion caused the dust bowls of the 1930's in the Great Plains of
North America. Entrained dust from soils, storage piles, and waste sites has
been recognized as a significant source of ambient particulates (1). Control
of these fugitive dust emissions is desirable to prevent loss of resources
(such as coal from storage piles) and to prevent the spread of potentially
toxic dusts from waste dumps.
The EPA is responsible for developing control options for minimizing the
impact from these emissions. Evaluations of these control options on a field
scale are expensive and the measurements are subject to variations in atmos-
pheric conditions. A reduced-scale experimental facility such as a wind tun-
nel is desirable to allow controlled study of several variables that influence
erosion in the open atmosphere. This approach is of special value in screen-
ing control options.
168
-------
A wind-tunnel facility was considered to be ideal for such investigations.
Gillette (2) described a wind tunnel used to study threshold velocities for
erosion. His tunnel was a portable, open-floored unit in which soil formed
the floor. Soo, Perez, and Rezakhany (3) also used a wind tunnel to study the
use of wind barriers for reducing entrainment from coal-dust piles.
Description of the Wind Tunnel
Review of past work on particle entrainment by wind indicated that the
dust entrainment test facility should be characterized by a long, square test
section and should offer maximum access and visibility throughout its length.
The flow near the floor over the dust sample should have a thick, turbulent
boundary layer with a logarithmic velocity profile to simulate the form of the
atmospheric boundary layer.
Options for design were somewhat limited because the test section would be
a modification to the existing aerodymanic test facility which could be best
described as an environmentally controlled, closed-loop wind tunnel. Because
of the size of the fan already in place, the maximum available airflow was
680 m3/min. With a fixed flowrate, a trade-off had to be made between the tun-
nel cross-sectional area and the wind velocity. A large cross-section was
desirable because it would eliminate the boundary-layer effects of the side
walls and ceiling. However, a conflicting concern was that a sufficient veloc-
ity be available to entrain, say, coal dust with an entrainment-reducing
coating applied.
Based on these requirements, a new test section with a 60 cm x 60 cm cross-
section was designed and constructed. This cross-sectional area allows a maxi-
mum velocity of 30 m/sec and is of sufficient size to minimize ceiling and wall
effects. The whole assembly, shown in Figure 1, is a bolt-on replacement for
the original wind-tunnel test section. The structure is 9.8 m long and is
constructed of aluminum, plexiglass, and plywood. The primary elements are
aluminum channels and angle beams which ensure straightness and rigidity. The
walls of the wind tunnel are virtually all plexiglass for maximum visibility.
The plexiglass sheets are mounted on aluminum frames to form swinging doors
which provide access to all points in the test section. Each door is 1.2 m
long and there are eight doors to a side. The plexiglass ceiling is suspended
from cross supports so that the height can be adjusted by adjusting the ceiling
height at different points down the length of the tunnel. A larger floor-to-
ceiling height may be provided at the tunnel exit than at the tunnel inlet.
This divergence permits a uniform static pressure while slightly decreasing the
velocity. Keeping the static pressure slightly below atmospheric eliminates
sealing difficulties and ensures that dust entrained in the gas stream stays in
the tunnel.
Several ways of generating a thick, turbulent boundary layer near the
floor were considered. In short wind tunnels, complex vortex generators are
required to get the necessary boundary layer. In this tunnel, however, the
length was sufficient to develop the logarithmic profile using roughness ele-
ments. Two different types of roughness elements were tried; the first were
lengths of 2.5 cm aluminum angle and the second were 1.25 cm, quarter-round
wooden strips. The roughness elements were attached to the floor and placed
169
-------
across the width of the tunnel. The first element was placed at the inlet, the
second one 30 cm downstream, and so on until the last one located 4.6 m down-
stream of the inlet. The dust sample tray is 45 cm downstream of the last rough
ness element.
Velocity profiles were measured with a pitot tube and a hot-film anemometer
to verify the aerodynamic design. The velocities were measured vertically from
the floor, just upstream of the test section along the vertical centerline. The
profile of mean velocities as measured at different nominal velocities is shown
in Figure 2. The indicated flow settings correspond to the nominal flow speed
in ft/sec. The nearly 30 cm thick boundary layer resulting from the 2.5 cm
roughness elements is evident. Although not shown in these measurements, the
ceiling boundary layer is less than 5 cm thick. Hot-film anemometer studies
indicate that the flow is laminar between 30 and 55 cm from the floor. Hori-
zontal velocity profiles, the profiles between the walls and parallel to the
floor, are shown in Figure 3. The mean velocity is uniform to within 5 percent
over the central 45 cm of the tunnel, and virtually constant over the central
30 cm.
The atmospheric boundary layer has a velocity profile which can be repre-
sented by:
U = ^ £n (Y/Y0) (1)
where U is the velocity at height Y,
U* is the friction velocity,
k is Von Karman's constant (k=0.4), and
Y0 is the roughness length.
The vertical profiles as measured with a hot-film anemometer were compared to
Equation 1. Figure 4 shows the vertical velocity profile resulting from 2.5 cm
roughness elements at a nominal velocity of 23 m/sec. Figure 5 shows similar
data for the 1.27 cm elements. A best-fit line through the data points indi-
cates that U* = 3.5 m/sec and Y0 = 1.7 cm for the 2.5 cm elements. For the
1.27 cm elements, U* = 1.5 m/sec and Y0 = 0.33 cm. From the definition of
friction velocity (U* = /T0/S), the shear stress can be calculated for known
U* and 6. The 2.5 cm surface roughening elements produce a shear stress of
14.4 N/m2 and the 1.27 cm elements produce a value of 2.7 N/m2. These values
are typical of values encountered in the atmosphere.
The turbulence intensity was measured as a function of height from the
floor. The results of these measurements are shown in Figure 6. The turbu-
lence is greatest near the floor and rapidly drops to zero at the centerline
of the tunnel.
Experimental Setup and Procedure
A test procedure was developed to measure dust emissions and evaluate con-
trol options. The advantage of reduced-scale tests is that they offer the
ability to determine the sensitivity of dust entrainment to individual para-
170
-------
meters. The disadvantage of reduced-scale tests is that the relationship
between data from wind-tunnel tests and data from actual performance tests in
the open atmosphere is uncertain at best.
The most important parameter with regard to particle entrainment is the
shear stress at the surface of the dust sample. As noted above, the shear
stress at the surface is related to the shape of the velocity profile. There-
fore, a change in the shape of the velocity profile would result in a change
of the shear stress. For example, when wind flows over a pile, eddies intro-
duced in the flow affect the shear stress. The inclusion of secondary flows
would complicate the comparison of results from tests on different control
options.
A dust tray was designed which could be inlaid in the floor of the test
section as seen in Figure 7. When the tray is filled, the top surface of the
dust is flush with the floor of the tunnel and the velocity profile is not
appreciably altered. The tray has an area of 900 cm2 and is 2.5 cm deep. It
is equally spaced between the walls of the tunnel and 46 cm downstream of the
last surface roughening element. At this location, the tray is not affected
by the wall or ceiling boundary layers. The tray can be lifted out of the
tunnel either to be filled with dust or to be weighed on a triple beam balance.
To measure the emissions from the dust sample, two sets of sample ports
were drilled. There are two ports in each set, one in the floor and one in
the wall of the tunnel. The first set of ports is located 2.44 m downstream
of the dust tray and the second set is located 4.88 m downstream of the tray.
A probe connected to either a Climet optical particle counter or a filter may
be inserted into a. port to determine the particle size distribution or mass.
A similar set of ports located 15 cm upstream of the sample tray is used to
measure the velocity profile.
Each test of entrainment involved filling the sample tray with dust and
smoothing the surface. Then the tray was placed in the tunnel and exposed to
a known velocity profile. Three methods to measure the rate of dust emission
from the sample were investigated. For Methods 1 and 2, a probe was inserted
in one of the ports located downstream of the tray. In Method 1, the probe
was used to collect mass samples. In Method 2, the probe was used to collect
samples for size-distribution analysis with an optical particle counter. A
large number of particles were blown off very soon after the beginning of the
test; after this initial phase, hardly any dust was entrained. These emission-
measurement methods are sensitive to the transient structure of the sample
surface; consequently, the results obtained with these methods were virtually
impossible to reproduce. Method 3 was much more direct. By measuring the pre-
test and post-test weights of the sample tray, it was possible to calculate
the amount of mass entrained per unit area per unit time averaged over the
duration of the test. A drawback of pre-test and post-test weighings is the
requirement that the dust tray be removed from the tunnel for each data point
to be obtained. Consequently there is the possibility of errors due to han-
dling.
171
-------
Preliminary Results
Preliminary results on dust entrainment which were obtained using this
facility are shown in Figure 8. A variety of dust types and sizes were used
for these verification studies including flyash, flyash mixed with glass beads,
and coal dust. The threshold velocities shown in Figure 8 compare well with
the results of Gillette (2) who reported threshold friction velocities on the
order of 50 cm/sec for grains with 1 mm diameter. The most important feature
to note in Figure 8 is the higher velocities, and consequently higher shear
stresses, required to entrain the smaller particles. These results agree with
the observations of Chepil and Woodruff (4) who note, "Loose particles smaller
than 0.01 mm (10 jum), if not mixed with coarser particles and if placed in a
bed that is thoroughly smoothed, are not moved even by an exceedingly strong
wind."
To entrain a small particle, the inertial and adhesive forces acting on
it must be overcome by the shear stress of the wind. The shear stress in the
laminar sub-layer which is present at the surface is considerably less than the
shear stress in the turbulent layer surface. For this reason, a particle which
is too small to protrude above the laminar layer is less likely to be entrained
than a larger particle.
Summary
In summary, we were able to design and construct a wind-tunnel section for
modification of the EPA aerodynamic test facility. The design offers access to
all points of the tunnel and visibility throughout its full length. The tun-
nel is long enough for surface roughening elements to generate turbulent bounda-
ry layers with a wide range of friction velocities.
Preliminary tests indicate that dust emissions can be measured in the tun-
nel and the shear stresses generated are sufficient to cause entrainment of
particles of various sizes. Detailed tests of the effectiveness of emission
control options are being studied.
References
1. Cuscino, T.A., C. Cowherd, and R. Bohn. Fugitive Emission Control of Open
Dust Sources. In: Proceedings: Symposium on Iron and Steel Pollution
Abatement Technology for 1980, EPA-600/9-81-017, pp.71-84, March 1981.
2. Gillette, D. Tests with a Portable Wind Tunnel for Determining Wind
Erosion Threshold Velocities. Atmos. Env. V. 12, pp.2309-2313, 1978.
3. Soo, S.L., J.C. Perez, and S. Rezakhany. Wind Velocity Distribution Over
Storage Piles and Use of Barriers. In: Proceedings: Symposium on Iron
and Steel Pollution Abatement Technology for 1980, EPA-600/9-81-017,
pp.85-106, March 1981.
4. Chepil, W.S. and N.P. Woodruff. The Physics of Wind Erosion and its Con-
trol. Advances in Agronomy, V. 115, pp.211-302, 1963.
172
-------
Figure la. The wind tunnel.
Figure Ib. Profile forming roughness elements,
173
-------
a 3o-
fe25.
t—
z
CD
E 20-
x
15-
10-
5 -
Nominal Velocity (m/s)
B =
C =
12 15 18 21
VELOCITY (m/s)
24
27
30
33
--60
-55
-50
-45
-40 =
N-t
-35 §
o
-n
30 <«
o
Ms |
•20 g"
•15
. 10
. 5
0
Figure 2. Vertical mean velocity profile at tunnel centerline.
0
5
ioH
15
25-
25
3
20-
15-
10-
5-
0
AIR
12
—I 1
15 18
VELOCITY (m/s)
21
—T
24
—r
27
T~
30
~T
33
Figure 3. Horizontal mean velocity profile at tunnel centerline.
174
-------
100
B 10 .
V " 1.7 cm
u" "3.5m/s
r° ' 14.4 N/m2
12 18
VELOCITY (m/s)
24
30
Figure 4. Boundary layer profile at 23 m/s, 2.5 cm roughness elements.
175
-------
100
10
y* • 0.33 cm
u* » 1.5 m/s
TO - 2.7 N/m2
6 12 18
VELOCITY (m/s)
24 30
Figure 5. Boundary layer profile at 12 m/s, 1.27 cm roughness elements.
176
-------
TURBULENCE INTENSITY (%)
25 H
20 H
10-
10
15
20
25
30
35
40
9 12
VELOCITY (m/s)
15 18
Figure 6. Boundary layer profile at 18 m/s, 1.27 cm roughness elements.
Figure 7. Dust tray in the wind tunnel.
177
-------
5 •
4 -
I
12 15 18
VELOCITY (m/s)
21 24
27
30
Figure 8. Preliminary results of entrainment tests.
178
-------
TECHNIQUES AND EQUIPMENT FOR MEASURING INHALABLE
PARTICULATE FUGITIVE EMISSIONS
By: Henry J. Kolnsberg
TRC - Environmental Consultants, Inc.
Wethersfield, CT 06109
ABSTRACT
The United States Environmental Protection Agency has initiated an
extensive program to measure inhalable particulate matter (IP) emissions
from a large number of industrial sources to obtain data for the development
of IP emission factors. About half of the effort has been reserved for
fugitive emissions measurements at a variety of industrial sites.
To respond to the special demands of sampling in the limited IP size
range, the recognized sampling methods have been modified or restricted and
a number of specialized sampling devices have been developed. It is
expected that the modified sampling methods and special devices will provide
the bases for future standard sampling procedures once an IP standard has
been adopted.
This paper describes the sampling techniques and modifications being
employed in the program. The IP sampling devices are also described, along
with the results achieved in their initial applications.
INTRODUCTION
One result of the Environmental Protection Agency's current review of
the criteria and standards for the regulation of particulate matter emis-
sions is a proposed revision of the existing standard for total suspended
particulate matter (TSP). The proposed revision will either replace the TSP
standard with one based on the fraction of particulate matter in the inhal-
able size range (currently defined as 15 micrometers and smaller in aero-
dynamic diameter) or modify the TSP standard to include inhalable particu-
late matter (IP) limitations for specific areas or sources.
To obtain background data for the proposed IP standard, the Agency has
initiated a program to measure IP emissions from sources in a variety of
industries, prioritized in consideration of their estimated impact. The
measurements will be used to update emission factors for industrial proces-
ses and operations as published in EPA's Compilation of Air Pollutant Emis-
sion Factors, commonly referred to by its report number, AP-42. The meas-
urement program, which is divided about equally between point and fugitive
sources, will include simultaneous determinations of TSP and IP from each
179
-------
source; and will thus provide bases for comparison and guidance in the pro-
posed revision of the standard.
The measurement of the fugitive IP emissions presents some rather unique
problems. Point source IP emissions can, in most instances, be quantified
quite readily using the standard techniques and equipment currently employed
for point source TSP - or relatively simple modifications of such techniques
and equipment. There are no such "standard" measurement methods for fugi-
tive emissions, however, and most of the equipment employed in the recog-
nized techniques is designed for sampling TSP. The starting point for the
fugitive IP measurement portion of the program then had to be the definition
and specification of the techniques and equipment to be employed. This
paper summarizes the measurement techniques and sampling equipment for IP
fugitive emissions currently being used in the program and described in an
as yet unpublished draft version sampling protocol prepared by TRC - Envi-
ronmental Consultants, Inc. as the initial phase of the program effort.
MEASUREMENT TECHNIQUES
There are four basic measurement techniques for airborne fugitive emis-
sions generally recognized by the air pollution community as capable of pro-
viding acceptable data. Each is used to determine the generation rate of
emissions from a source as the product of a measured concentration and the
volume flow rate of the air carrying the emissions past the measurement
location. While most of the measurements made until very recently have been
designed to quantify emission rates of TSP, the four basic techniques may be
employed for IP measurements with only minor modifications to program
designs and sampling procedures and the utilization of some specifically
designed size-selective sampling equipment.
The most accurate fugitive emissions measurements can usually be made at
the source before the emissions begin to diffuse into the atmosphere. The
quasi-stack sampling technique provides a means for the capture of the emis-
sions in a small volume of transport air and for the control of the emis-
sions flow from the source to the sampling location. The next most accurate
measurements can be obtained after the emissions have diffused into a
larger, enclosed volume of air, such as that inside a plant structure. The
roof monitor sampling technique is used in such instances. The least accur-
ate measurements are those made after the emissions have diffused extensive-
ly into the ambient air at significant distances from the source, where the
upwind downwind sampling technique may be employed. The fourth technique,
exposure profiling, is a specialized form of downwind sampling restricted to
a specific category of sources. Its accuracy is comparable to roof monitor
sampling.
The IP sampling program prioritizes the application of the techniques in
their order of accuracy—quasi-stack, roof monitor and upwind/downwind sam-
pling. Exposure profiling receives higher priority for applicable sources
than upwind/downwind sampling.
The basic techniques their applicabilities and limitations, and the
modifications or special sampling equipment required for IP measurements are
180
-------
described in the following subsections. Descriptions of the IP sampling
equipment are provided under Sampling Equipment, below.
Quasi-Stack Sampling
In this technique the fugitive emissions are captured at their source in
a temporarily installed hood or enclosure and transported by the capturing
air to an exhaust duct or stack of regular cross-sectional area. The emis-
sions are measured in this ductwork using standard stack sampling procedures
such as EPA Method 5.
The precision and accuracy limits of the quasi-stack method are the best
of the noted sampling techniques and are also the best defined. The accur-
acy of the quasi-stack method is only slightly less than that of a normal
stack test in that fewer points are usually sampled and that a constant bias
may be introduced by a failure to capture all of the emissions from the
source being tested. Care in the design of the capturing system can reduce
the latter error to a negligible amount. A tolerance in the isokinetic sam-
pling range of +20 percent, rather than the +10 percent allowed in stack
sampling, is usually permitted.
Quasi-stack sampling is necessarily limited to sources that can be iso-
lated from other sources and effectively enclosed or hooded to capture their
emissions. Careful consideration must be made in the design of the enclo-
sure or hood to ensure total capture of the emissions and to provide a vol-
ume of emission-transporting air sufficient to carry the emissions intact to
the sampling equipment. The hooded enclosure design should not interfere
with normal plant operations, and the capturing air flow across the process
should not be so large as to alter the nature of the process or affect the
amount or character of the emissions.
No major modifications to the quasi stack sampling method procedures are
required for IP measurement. A small, dual- cyclone probe to provide for
particle size fractionation has been designed for use with the EPA Method 5
or 17 trains for in-stack sampling. The probe is ideally suited for IP
quasi-stack sampling and is described under Sampling Equipment, below.
Standard cascade impactors may be used with the addition of a cyclone to
provide the 15 micrometer cut.
Roof Monitor Sampling
This method is used to measure the emissions generated by sources locat-
ed within a building or similar structure as they are transmitted into the
atmosphere through a roof monitor or other opening. The total emission rate
for all sources within the structure is determined as the product of the
emissions concentration measured in the air at the opening and the air flow
rate through the opening. Most roof monitor sampling programs require the
collection of samples and the measurement of air velocities simultaneously
at a number of points in the plane of the opening to ensure that
representative average values of concentration and flow rate are obtained.
181
-------
Roof monitor sampling is most effectively employed for larger sources
located within structures with only a few openings. It requires sampling
and measurement devices capable of making accurate determinations of
relatively small masses of emissions and very low air velocities. It may be
utilized for the characterization of specific sources within an enclosure if
operating schedules include or permit the arrangement of emissions genera-
tion by only that specific source.
The accuracy and precision of this technique vary with characteristics
of the sources and are definable only in general terms for each source test-
ed. In general, the roof monitor method is not as accurate as the quasi-
stack method since a significant portion of the emissions may escape through
other openings before reaching the measurement point and since a much higher
degree of dilution with transport air occurs before measurement.
No procedural modifications to the roof monitor method are required for
IP measurements. The method is probably more effective for particles in the
smaller IP size range since a smaller proportion of such particles will be
lost to fallout resulting from relatively low entrainment velocities and
large vertical transport distances inherent in roof monitor arrangements.
Size-selective devices, such as a horizontal elutriator to reject particles
larger than 15 micrometers and cascade impactors to provide size-fractionat-
ed samples, are employed with high volume (hi-vol) filter samplers to obtain
concentration measurements. Transport air velocities for calculating volume
flow rates are determined by traversing anemometers across the monitor
opening^
Upwind/Downwind Sampling
This method is utilized in the measurement of emissions after they have
entered the ambient atmosphere from open or area sources, or from enclosed
sources not amenable to quasi stack or roof monitor sampling^ The emission
rate for such sources is determined by measuring the concentration of the
emissions in the ambient air downwind of the source, subtracting the portion
of the concentration attributable to other sources and that measured as
background upwind of the source, and using the thus-determined concentration
in proven diffusion equations or mathematical models to back-calculate the
source's rate of emission. Measurements of other contributing parameters,
such as wind speed and direction during the emission sampling, location of
samplers relative to the source, and atmospheric and topographic conditions,
are also required.
The upwind/downwind method is the least accurate of the methods de-
scribed, since only a very small portion of the emissions are captured for
measurement after an extreme degree of dilution in the transporting ambient
air. Careful sampling system design is required to approach the accuracy of
the equations or models used in calculating the emission rate.
Upwind/downwind sampling is probably the most universally applicable of
the fugitive emissions measurement techniques, since it is not usually
limited by source location or geometry.
182
-------
Procedure modifications required for IP sampling are limited to allow-
ances for longer sample collection periods to permit the collection of suf-
ficient particulate material mass for analysis, required by the rejection of
particles larger than 15 micrometers. Longer sampling period requirements
can sometimes be relieved by locating samplers closer to the emissions
source where the concentration is higher. A size-selective inlet for stand-
ard hi-vol samplers has been developed and is already commercially
available. It is described under Sampling Equipment, below. More detailed
particle size data may be obtained by adding a cascade impactor.
Wind speed and direction measurements are made throughout the sampling
period using standard meteorological devices with recorders.
Exposure Profiling
The exposure profiling technique is employed in the sampling of line or
moving point sources and relatively small virtual point or area sources
where the samplers can be located in the ambient atmosphere within a few
meters downwind.
Samples are taken at a number of points within the emission plume at
near-isokinetic sampling velocities, selected to approximate the measured
local mean wind speed. The plume boundaries are determined by spatial
extrapolation of the measured exposures and the source strength is deter-
mined by integration of the measured exposures over the cross-sectional area
of the plume.
The exposure profiling method is not as accurate as the quasi stack
method since only portions of the total emissions and transport air are sam-
pled or measured. Its use of near-isokinetic sampling and provisions for
correcting to isokinetic conditions in the exposure calculations tend to
minimize errors, making the method about as accurate as the roof monitor
method.
IP exposure profiling requires only that the larger particles be reject-
ed from the sample while maintaining the near-isokineticity of the sampling.
This is readily accomplished by the use of either a horizontal elutriator or
a cascade impactor in the inlet to a hi-vol type sampler. Sampling
velocities are matched to transporting air velocities by an inlet nozzle of
suitable dimensions with the constant-flow rate samplers.
SAMPLING EQUIPMENT
The modifications to adapt the basic fugitive emissions measurement
techniques for IP measurements all involve the addition of size-selective
devices for the rejection of most of the particulate matter larger than 15
micrometers from the samples before they are collected for gravimetric anal-
ysis. Each of the devices makes some provision for the determination of the
mass of large particles included in the IP catch, usually as a function of
the mass of larger particles collected by the device. The devices include
the dual cyclone train for quasi-stack measurement; a horizontal elutriator
for roof monitor and exposure profiling measurements; and a size selective
183
-------
inlet for hi-vol samplers for upwind-downwind measurements. Each of these
devices has been developed specifically for the IP measurement program and
is being used by the program's contractors. Possible modifications to elim-
inate excess particle bounce and carry-over in existing cascade impactors
designed to be used with hi-vol samplers are being investigated by the
Process Measurements Branch of EPA's Industrial Environmental Research
Laboratory at Research Triangle Park, North Carolina.
Each of the devices is described in some detail in the following sub-
sections, along with some of the procedural or operational techniques
involved in its application to IP measurements.
Dual Cyclone Train
The dual cyclone train was designed and fabricated by the Southern
Research Institute (SoRl) of Birmingham, Alabama, as their IP Cyclone Sam-
pler specifically for the IP measurement program. It consists of two
cyclones and a backup filter arranged in series, and can be connected
directly to an EPA Method 15 or Method 17 sampling train. The initial
cyclone has a D$Q of 15 micrometers at the design flow rate of 23 liters
(0.8 ft^) per minute at 150°C (300°F). The second cyclone has a 050
of 2.5 micrometers. The backup filter may be either a 2.5 inch disc or a
Method 17 thimble type.
In operation, the selection of one of a set of 11 inlet nozzles of dif-
ferent diameters permits sampling at near-isokinetic conditions over a wide
range of gas velocities at the design flow rate. The initial cyclone
separates the portion of particles larger than 15 micrometers from the
sample and passes the IP fraction. The second cyclone then separates the
fraction between 15 and 2.5 micrometers and passes the fine fraction to the
filter. The overall diameter of the IP Cyclone Sampler with any of the
inlet nozzles is just under 6 inches so that two 6-inch diameter sampling
ports are required for the recommended four-sampling-point array for round
or rectangular ducts.
Horizontal Elutriator
A horizontal elutriator consists of a stack of horizontal plates, spaced
so that each pair of plates forms a settling chamber. As particles are
transmitted through the chambers at a velocity determined by their carrier-
gas volumetric flow rate, the larger particles are deposited on the chamber
bottom as the result of gravitational acceleration, while smaller particles
are carried through. The elutriator designed by SoRI for the IP program has
a collection efficiency of 50% for 15 micrometer particles when used as a
directional inlet for a standard 8 x 10 hi-vol sampler operating at 40 CFM
(1.13 m-Ymin). The elutriator, since it is a passive device, requires no
special calibration but must be kept level during sampling to ensure proper
operation. A modified design has been developed that will provide 15
micrometer particle separation at 20 or 40 CFM sampling rates and allow the
use of impaction type samplers in series with the elutriator and hi-vol
sampler.
184
-------
Size-Selective Inlet
Designed for .non-directional sampling applications as a precollector of
15 micrometer particles with a standard hi-vol sampler, the size-selective
inlet (SSI) is commercially available from at least two manufacturers. The
SSI draws its sample through a circumferential inlet slot that eliminates
almost all wind speed and direction effects, and transmits it through a
buffer chamber into an itnpaction chamber where all particles larger than 15
micrometers are collected. Smaller particles are transmitted through the
impactor chamber's vertical vent tubes to the hi-vol filter for collection.
The SSI thus provides a measure of the IP concentration as a single mass/
sampled-air-volume ratio in the same manner as hi-vol sampling for TSP.
Cascade Impactors
At least two manufacturers currently market multi-stage impaction
devices for use with standard hi-vol samplers. Theoretically capable of
separating the particulate matter into as many as six different size ranges
within the IP category, most of the impactors have been found to provide cut
points different from their design or calibrated values due to excessive
bouncing of larger particles and the resultant carry-over to subsequent
stages. Experiments by the EPA and manufacturers have indicated that the
bounce problem can be largely eliminated by operating at reduced sampling
rates, but this also results in a shift in the SSI cut point to greater than
15 micrometers. The Process Measurements Branch of EPA's Industrial
Environmental Research Laboratory at Research Triangle Park is currently
conducting tests to evaluate impactors modified to eliminate the bounce
problem while operating at the standard hi-vol sampling rate of 40 CFM (1.13
m^/min.). Such modified impactors will provide particulate size
distribution information in addition to the IP fraction concentration when
used in conjunction with an SSI on a standard hi-vol sampler.
APPLICATIONS AND RESULTS
To date, only the SSI's and cascade impactors designed for use with
hi-vol filter sampler have been used in the sampling of fugitive emissions.
Three SSI/impactor/hi-vol samplers were included in an array of 25 samplers
used to determine the fugitive emissions generation rate of a large, active
utility coal storage pile in a TRC project to evaluate the
cost-effectiveness of the upwind/downwind technique. The combined samplers
were installed at locations side-by-side with standard TSP hi-vols and
operated simultaneously with them. The combined samplers were operated
through five successful test runs at 20 CFM (0.57 m3/min.) after a trial
run at 40 CFM (1.13 m^/min.) indicated so much bounce and carry-over that
almost all of the sample was deposited on the final hi-vol filter.
Results of the five sampling runs are summarized in Table 1 as average
values of the IP sampled at three distances downwind of the emissions
source. The increasing percentages of particles in the smallest size range
displayed as the distance from the source increases is indicative of effec-
tive operation of the combined samplers, reflecting the settling of larger
particles in the diffusing cloud of emissions.
185
-------
Current plans to increase the effort on the IP measurement program will
provide additional applications data to either verify the effectiveness of
the procedures and equipment designated in the IP sampling protocol or indi-
cate changes to make them more effective.
186
-------
TABLE 1. INHALABLE PARTICULATE MASS IN COAL PILE FUGITIVE EMISSIONS
SAMPLER
LOCATION
Meters
downwind
16
to
PERCENT
10.2
to
IP SAMPLE
4.2
to
IN SIZE RANGE
,2.1
to
of source 10.2ym 4.2ym 2. lym 1.3ym 1.3-Oym
210 25.4 22.0 6.4 6.5 ,. 39.6
470 15.1 18.8 3.6 6.4 56.0
600 17.6 21.2 1.2 2.5 57.4
187
-------
BALLOON SAMPLING TO CHARACTERIZE PARTICLE
EMISSIONS FROM FUGITIVE SOURCES
by
James A. Armstrong
Denver Research Institute
University of Denver
Denver, Colorado 80208
and
Dennis C. Drehmel
U. S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
ABSTRACT
A balloon sampling method being developed to characterize fugitive
particle emissions from line, point, and area sources is presented. The
method consists of using a series of tethered balloon sampling systems
positioned at selected distances downwind from the source being sampled.
Three lightweight wind directional samplers will be flown simultaneously at
different heights by each balloon system with a fourth sampler located at
ground level at each launch site. The samplers will utilize horizontal
elutriation to obtain a 15 urn aerodynamic diameter particle cut and will be
capable of sequential filtration to provide a second cut to separate the
fine and coarse size fractions of the inhalable particles sampled. The
multiple balloon sampling technique will be coupled with the exposure
profiling method to determine improved source emission factors. In addi-
tion, the balloon network will be used to characterize the downwind trans-
port of fugitive emissions.
INTRODUCTION
It is now widely recognized that fugitive particle emissions contribute
significantly to the measured levels of total suspended particulates (TSP)
found in both urban and rural environments. These emissions are believed
to be responsible for the failure of many areas to attain the national
primary air quality standards for TSP. Major open sources of fugitive
emissions include unpaved roads, paved roads, surface mines, storage piles,
construction activities, agricultural tilling, wind erosion of harvested
cropland, and forest and brush fires, including both wildfires and prescribed
burns(l). Other sources include iron and steel foundries, smelters, and
refineries(2,3,4).
Of the fugitive sources listed above, open sources (except forest and
brush fires) emit particles primarily in a coarse size range of greater
than approximately 2.5 ym in equivalent aerodynamic diameter. These par-
ticles, classified as mechanically generated, are believed to be the most
significant sources of fugitive emissions contributing to measured ambient
TSP levels. Fugitive particle emissions produced by combustion processes,
188
-------
such as from forest and brush fires and from furnace operations at foundries
and smelters, are typically in a fine particle size range less than 2.5 ym.
The chemistry of the coarse versus fine particles also differs: coarse
particles are usually more basic; and fine particles, more acidic(5).
The traditional method of monitoring airborne particles from fugitive
sources has been to use ground-based high-volume (hi-vol) samplers. The
samplers are usually at considerable distances both upwind and downwind
from the source being sampled. Dispersion modeling is normally used to
interpolate the source strength from the hi-vol data. Considerable error
can be incurred when using this technique because of the often limited
number and location of sampling sites and due to the current simplicity of
the atmospheric dispersion models(6). Complications are further increased
when the source is not well defined.
Samplers mounted on towers have been used successfully to characterize
fugitive dust emissions from paved and unpaved roads employing an "exposure
profiling" technique which is based on the isokinetic profiling concept
used for conventional source testing(7). Particulate emissions are measured
directly downwind of the source being sampled by placing a grid of samplers
in the effective cross-section of the plume. Emission factors are subse-
quently determined by a mass-balance calculation. The exposure profiling
technique eliminates the need to use dispersion models to estimate emission
rates from fugitive sources. A two dimensional vertical array of samplers
must be used when sampling a point or area source with this technique while
line sources (e.g., haul roads) require a one dimensional vertical grid of
samplers. Major limitations of the exposure profiling technique are the
practical height of the sampling towers and their lack of mobility.
A tethered balloon sampling system has been developed which overcomes
these limitations so that the exposure profiling technique can be extended
to a larger number of fugitive sources. This sampling system basically
consists of a tethered balloon, three wind vane samplers, and a ground-
based sampler. The wind vane samplers attach to the balloon and its
tetherline at different sampling heights. The balloon is positioned at
selected distances downwind from a source being sampled. The ground-based
sampler is located near the balloon launch site to monitor dust concentra-
tion at ground level. This sampling method has been successfully employed
to investigate the vertical extent of fugitive dust emissions from various
operations at a surface coal mine(8). A program has been started to extend
this method by using a series of four balloon systems equipped with newly
designed inhalable particulate (IP) samplers capable of sequential filtration.
The multiple balloon sampling method will be coupled with the exposure
profiling technique to determine improved emission factors from line,
point, and area sources.
SYSTEM DESIGN
The sampler design parameters are presented. This is followed by a
detailed discussion of the IP wind vane samplers and the tethered balloon
systems.
189
-------
Sampler Design Parameters
EPA is in the process of establishing a new ambient air quality size
specific particulate standard(S). Particles classified as "inhalable" form
the basis of the proposed standard. Inhalable particles are currently
defined as those having an aerodynamic diameter of 15 ym or less. The 15
ym diameter constitutes a conservative upper limit for the size of inhaled
particulate matter that can affect the human lower respiratory system.
Consequently, the new wind vane particle samplers are being designed for
50% capture of 15 ym aerodynamic diameter particles by means of horizontal
elutriation. Inhalable particles can further be subdivided into groups of
coarse and fine fractions with a cut point of approximately 2.5 ym. The
fine particles are primarily responsible for air quality visibility degrada-
tion and are primarily produced by combustion processes. The coarse-
fraction inhalable particles are those between 2.5 and 15 ym and are nor-
mally classified as mechanically generated. Most wind generated dust, for
example, would consist of coarse size particles. Depending on the type of
source being sampled, it may sometimes be desirable to be able to separate
inhalable particles into coarse and fine fractions. For example, fugitive
particulate emissions from an iron foundry are expected to be a mix of fine
and coarse particles. By being able to quantify the fractions in terms of
amount and chemical composition, the actual emission sources should be able
to be identified. When desired, the balloon samplers will employ sequential
filtration to separate inhalable particles into coarse and fine fractions.
The samplers are also being designed such that their performance will
be independent of wind speed and direction. Sampler weight will allow
three units to be flown by each balloon. A fourth unit will be stationed
at the winch site of each balloon system to monitor ground-level dust
concentrations.
IP Wind Vane Samplers
A schematic of the sampler type being developed for this program is
shown in Figure 1. The vertical tube of the sampler attaches to the
horizontal tube at a point where the wind vane balances the forward tube
extension such that the extension is always horizontal. The tubes are
fiberglass; the wind vane consists of aluminized Mylar sheets stretched
over a wire loop. A particle sampling assembly, consisting of a fiberglass
entrance cone, a horizontal metal tube elutriator, and a polystyrene
cassette capable of holding one or two filters, attaches to the front end
of the horizontal tube. The wind vane configuration ensures that the
entrance cone always points into the wind during sampling. A housing
containing a diaphragm pump, battery pack, and voltage regulator is attached
to the lower end of the vertical tube. The weight of these components
ensures that the sampler maintains the orientation shown in Figure 1. The
pump voltage will be regulated so that a constant flow rate of 2 £/min
through the sampling assembly will be maintained. The pump can operate for
-10 hours before the lithium battery pack needs replacement. Nylon lines
equipped with swivels and quick disconnect clamps will be used to attach
the samplers to the balloons or tetherlines. A prototype IP wind vane
sampler, weighing -320 g, is nearing completion and will be laboratory
190
-------
tested before mass production of additional units commences. The sections
that follow discuss horizontal elutriation, sequential filtration, entrance
cone considerations, and the planned laboratory evaluation of the IP wind
vane sampler.
Horizontal Elutriation
Several types of inertial mechanisms, such as centrifugal force and
impaction, were considered for separating particles having aerodynamic
diameters larger and smaller than 15 ym. Due to weight and pressure drop
limitations of the pumps capable of being used for the IP wind vane samplers,
however, it was decided to use horizontal tube elutriators.
The governing equation for predicting the collection efficiency of
particles from a laminar airstream passing through a circular horizontal
tube is (9):
Eff. - ^[2e/l-ez' * - e1' Vl-ez;a + arcsin e1/3] (1)
where e = (3V L/8aV), V is the particle settling velocity, L is the tube
length, a is the tube radius, and V is the average air velocity. The
settling velocity of spherical particles in air is:
where C is the slip correction factor, g'is the gravitational constant, p
is the particle density, d is the particle diameter, and y is the viscosity
of air.
For a 15 ym diameter aerodynamic particle, p = 1 g/cm3 and C-l. From
Equation 2, V =0.67 cm/s. Now by specifying the average air velocity
through a tube of a given radius and by setting Eff. = 0.5, the tube length
can be calculated using Equati
-------
fine particles. Nuclepore filters are excellent for this application since
their low blank weight and low hygroscopicity facilitate gravimetric and
elemental analyses. By controlling the face velocity of air passing through
the filters, it is possible to obtain a good size cut between coarse and
fine particles. For the 8.0 ym Nuclepore filter, a face velocity of 2.8
cm/s results in 50% capture efficiency for 3.1 ym aerodynamic diameter
particles. If the face velocity is increased to 7 cm/s, the cut point
shifts to 2.8 ym while a face velocity of 15 cm/s results in a cut point of
2.2 ym(ll). A thin coating of Apiezon grease is used on the 8.0 ym filter
to prevent particle bounce.
For this program, 25 mm diameter Nuclepore filters having the above
pore sizes will be used. Sequential filter cassettes are commercially
available for this filter size. (The filter cassette in Figure 2 shows
only the base section of the stacked filter assembly. This configuration
would be used when sequential filtration is not employed.) The actual
filtration area of the filters is 2.4 cm2 so for an air flow rate of 2
£/min, the face velocity would be 14 cm/s resulting in a cut point of ~2.3
ym. The outer cassette cap will be epoxied to the back of the elutriator
housing tube.
Entrance Cone Considerations
The diverging entrance cone of the particle sampling assembly will
have a total included angle of 30° or less to ensure that separation from
the walls will not occur as the air is decelerated prior to entering the
elutriator. The cone will slip over the front of the elutriator housing
tube as shown in Figure 2.
An error analysis, which considered the ratio of sampled particle
concentration to true concentration, was conducted to determine the effects
of anisokinetic sampling(12) . The analysis showed that one cone having a
set inlet diameter was insufficient for the anticipated range of wind
speeds (2 to 10 m/s) which will be encountered during balloon sampling. In
order to reduce the potential sampling error to less than 20% for 15 ym
aerodynamic diameter particles, three interchangeable cones having differ-
ent inlet diameters will be used. The decision of which cone to employ
will be dictated by the average wind speed encountered at the start of a
sampling period. Table 1 shows the calculated sampling error for three
different cone inlets. Since the sampling error is proportional to the
particle diameter squared, the error will be much less for particles <15
ym.
The samplers' wind vanes should keep the entrance cones aligned to
±10° of the wind direction. Since the maximum sampling error is equal to
1 minus the cosine of the angle, the maximum error for misalignment will be
Laboratory Evaluation of the IP Wind Vane Sampler
The particle collection performance of the prototype IP wind vane
sampler will be evaluated in the laboratory prior to field use. A
192
-------
closed-loop experimental electrostatic precipitator will be used as a wind
tunnel for this evaluation. The test section, where the wind vane will be
suspended, will be the precipitator collector section. The collector
plates and corona wire frame assemblies will be removed from the precipi-
tator prior to these tests. The length of the collector section is 2.1 m
while its height and width are 1.5 and 1.8 m, respectively. Diffuser
screens in the precipitator inlet cones ensure a uniform average airflow
velocity of 1.5 m/s through the collector section. Fly ash, injected
upstream of the precipitator, will be used for the sampler performance
evaluation. The mass loading and particle size distribution of the sus-
pended fly ash will be measured in the precipitator inlet and outlet
sampling ports by means of a cascade impactor. Results will be used to
determine the aspiration efficiency of the sampler, the effectiveness and
accuracy of the horizontal elutriator, and the performance of the sequen-
tial filter assembly.
TABLE 1
CALCULATED SAMPLING ERRORS FOR THREE CONE INLETS
Wind Speed Inlet Dia. Stokes No. for Velocity Ratio Sampling Error
d=15 ym
(m/s) (cm) (%)
2 0.45 0.31 1.0 0
3 0.45 0.46 0.67 -16
4 0.29 0.95 1.25 +18
5 0.29 1.2 1.0 0
6 0.29 1.4 0.83 -14
7 0.23 2.1 1.14 +12
8 0.23 2.4 1.0 0
9 0.23 2.7 0.89 -10
10 0.23 3.0 0.80 -18
The sequential filter collection efficiency will be evaluated by
weighing the dust catch of both the 8.0 and 0.3 ym Nuclepore filters by
means of a Cahn electrobalance. Scanning electron microscopy will also be
employed to evaluate the size cut performances of the elutriator and the
8.0 pm filter by determining the size distributions and concentrations of
particles collected on each filter.
Tethered Balloon Systems
Blimp-shaped, red polyethylene balloons, inflated with 3.25 m of
helium, will be used to carry aloft the IP wind vane samplers. The balloons
attach to portable winches by means of lightweight tetherlines having a
breaking strength of 535 newtons. The winches, which weigh 27 kg, contain
forward/reverse variable speed motors and rechargeable, sealed lead/acid
battery packs.
This size balloon has a lifting capacity of 1.2 kg to altitudes of 800
to 1,000 m in winds of 6 to 8 m/s. Balloon flight performance is still
193
-------
reasonable in winds of 10 to 12 m/s but the maximum altitude attainable is
somewhat reduced. The balloons are designed to survive in winds up to 20
m/s. Four balloon systems are being developed for this program.
FIELD PROGRAMS
A demonstration program, monitoring emissions from an unpaved road,
will be conducted first in order to evaluate and establish sampling param-
eters and analysis techniques. This will be followed by a field program at
a site where chemical surface treatments for unpaved roads are being eval-
uated. Later field programs will characterize fugitive emissions from
point and area sources.
Demonstration Program
An initial demonstration field program is planned to evaluate the
balloon sampling method using dust emissions from an unpaved road. Since
this is a line source, only a one dimensional analysis need be applied.
The present site being considered for this demonstration is a lightly
travelled unpaved county road located along the north perimeter of the DRI
Cherry Creek Field Site (CCFS) in Englewood, Colorado. This well maintained
gravel road runs in an east-west direction in an open area which is bordered
to the north by a Colorado state recreation area. Winds are normally from
either the north or the south so that downwind sampling sites would be
either on DRI or State of Colorado property. Permission to conduct sampling
tests on state property is currently being requested. The CCFS is located
-16 km southeast of the University of Denver. This location is ideal since
inflated balloons can be stored indoors at the CCFS overnight or during
inclement weather.
The objectives of the demonstration program are to evaluate the
performance and reliability of the balloon systems during actual field
operations, to determine appropriate sampling heights at selected downwind
sampling distances, and to establish sampling times versus road traffic
necessary to collect sufficient dust samples for the exposure profiling
analysis.
Parameters controlling the emission levels from unpaved roads include:
the silt and moisture contents of the surface road material, the average
weight and number of wheels of vehicles passing over the road, the average
vehicle speed, and the local wind speed(14). For this demonstration, as
many of these parameters as possible will be held constant. The maximum
speed limit for this road is 65 km/hr. During non-rush hour periods the
average traffic on the road is ~2 to 4 automobiles or light duty trucks per
hour. It is planned that during balloon sampling, a DRI truck will be
continuously driven at 65 km/hr across a 200 m test section of this road
(20 to 30 passes/hr). Dust emissions due to other vehicles should have a
minor impact on the test results. Time lapse movies of the road test
section will be taken to record all traffic during sampling.
All demonstration tests will be conducted under dry road conditions.
At least 2 days of dry weather will precede any test. Samples of the road
194
-------
surface material will be analyzed for silt content and type of soil. Field
operations will be conducted over periods in which the wind is relatively
steady and predominately from either the south or the north. A meteoro-
logical station will be located at the test site to record wind speed,
direction, and temperature during all tests.
The sampling network for the demonstration is shown in Figure 3.
Separation distances of 25 m between the road and first balloon and between
the three balloons will be evaluated first, followed by other separation
distances. Tests over a 2 to 3 week period will be conducted where the IP
wind vane samplers will be operated both with single and sequential filters.
Emphasis will be placed on determining appropriate sampling heights and
times for selected "source to balloon" separation distances.
It is believed that three to four people will be required to operate
the balloon systems. This demonstration will allow the field personnel
requirements to be established.
All dust samples collected by the IP wind vane samplers during the
demonstration tests will be analyzed gravimetrically. Background par-
ticulate matter collected by the upstream samplers will be subtracted from
the downwind sampler catches. Scanning electron microscopy will be used to
analyze selected samples from both single and sequential filters to deter-
mine particle size distributions and number concentrations. The mass
emission rates of inhalable particles per unit length of the unpaved road
will be calculated by the exposure profiling method for each of the down-
wind balloon systems. Comparisons of the calculated mass emission rates
will be made between the three downwind balloon systems as well as with
published results for unpaved roads.
Treated Unpaved Road Study
Following the demonstration program, balloon sampling will be employed
to monitor fugitive emissions from an unpaved road which has a number of
test sections treated with different chemical stabilizers. A long term
sampling program has been started at the test road located at the Fort
Carson Army installation, south of Colorado Springs, Colorado. An exten-
sive ground-based network of hi-vol and dichotomous samplers is being used
to monitor emissions from the road test sections. A direct comparison
between the ground-based and balloon sampling results at this site will be
possible.
Later Field Studies
The balloon sampling method will be used to monitor and characterize
fugitive emissions from an open area source, such as a coal pile that is
actively worked or a construction site, during the second program year.
The area source will require two dimensional sampling and analysis.
Efforts will be made to select a field test site which is isolated from
other particle emission sources. The sampling network which will be
employed for the field program is shown in Figure 4. This first test of an
area source can again be considered a demonstration program in that considerable
195
-------
effort will be made to define the two-dimensional envelope of the resultant
particle plume for selected downwind locations and to establish appropriate
sampling times for the collection of adequate particle mass loadings for
gravimetric analysis. The total mass emission rate of inhalable particles
from the area source will be calculated and compared to past investigations.
Again, all sampling operations will be documented using time lapse photography.
Future field studies will be conducted to sample particle emissions
from both paved and unpaved roads and both process and open sources. It is
anticipated that fugitive particle emissions will be monitored and analyzed
at sites such as copper and lead smelters, iron and steel foundries, and
aluminum mills. The sequential filtration capability of the IP wind vane
samplers will be extremely useful in identifying the fractions of coarse
versus fine particles from such sites. This will aid in the identification
of the specific sources of the emissions. All collected samples will be
gravimetrically analyzed. In addition, selected filter samples from process
sources could be analyzed by x-ray fluorescence to determine major, minor,
and trace elements, and by scanning electron microscopy/energy dispersive
x-ray spectrometry to determine particle size distributions, concentrations,
and compositions. Transmission electron microscopy/selected area electron
diffraction may also be employed to determine the species of individual
particles.
Further studies at open pit mining sites are also envisioned as well
as investigations at oil shale extraction and process operations, which are
just now being developed.
REFERENCES
1. Evans, J.S., D.W. Cooper, M. Quinn, and M. Schneider. Setting Prior-
ities for the Control of Particulate Emissions from Open Sources.
Symposium on the Transfer and Utilization of Particulate Control
Technology, Vol. 4. EPA-600/7-79-044d (NTIS PB 295229), February
1979, pp. 85-103.
2. Scott, W.D., and C.E. Bates. Measurement of Iron Foundary Fugitive
Emissions. Symposium on Fugitive Emissions: Measurement and Control.
EPA-600/2-76-246 (NTIS PB 261955), September 1976, pp. 211-237.
3. Lindstrom, R.N., and S.E. Sundberg. Fugitive Particulate Emission
Rate and Characterization for Electric Arc Steelmaking Furnaces.
(Presented at the 72nd Annual Meeting of the Air Pollution Control
Association, Cincinnati, Ohio. 79-39.3, 1979.)
4. Kalika, P.W., R.E. Kenson, and P.T. Bartlett. Development of Procedures
for the Measurement of Fugitive Emissions. EPA-600/2-76-284 (NTIS PB
263992), December 1976.
5. Miller, F.J., D.E. Gardner, J.A. Graham, R.E. Lee, W.E. Williams, and
J.E. Bachmann. Size Considerations for Establishing a Standard for
Inhalable Particles. JAPCA 29:610-615, June 1979.
196
-------
6. Zeller, K.F., D.G. Fox, and W.E. Marlatt. Estimating Dust Production
from Surface Mining. Third Symposium on Fugitive Emissions: Measure-
ment and Control. EPA-600/7-79-182 (NTIS PB 80-130891), August 1979,
pp. 103-117.
7. Cowherd, C., C.M. Maxwell, and D.W. Nelson. Quantification of Dust
Entrainment from Paved Roads. EPA-450/3-77-072 (NTIS PB 269944), July
1977.
8. Armstrong, J.A., P.A. Russell, and D.C. Drehmel. Particle Production
from Surface Mining - Part 1: Vertical Measurements. Proceedings:
Fourth Symposium on Fugitive Emissions: Measurement and Control.
EPA-600/9-80-041 (NTIS PB 81-174393), December 1980, pp. 37-63.
9. Gushing, K.M. Development of Horizontal Elutriators for Sampling
Inhalable Particulate Fugitive Emissions. Proceedings: Fourth
Symposium on Fugitive Emissions: Measurement and Control. EPA-600/9-
80-041 (NTIS PB 81-174393), December 1980, pp. 208-242.
10. Cahill, T.A., L.L. Ashbaugh, J.B. Barone, R.A. Eldred, P.J. Feeney,
R.G. Flocchini, G. Goodart, D.J. Shadoan, and G.W. Wolfe. Analysis of
Respirable Fractions in Atmospheric Particulates via Sequential
Filtration. JAPCA 27:675-678, July 1977.
11. Barone, John B. Air Quality Group, Crocker Nuclear Laboratory,
University of California, Davis, California. Personal communication,
June 1980.
12. Belyaev, S.P., and L.M. Levin. Techniques for Collection of Representa-
tive Aerosol Samples. J. Aerosol Sci. 5:325-338, 1974.
13. Durham, M. Isokinetic Sampling of Aerosols from Tangential Flow
Streams. Ph.D. Dissertation, University of Florida, Gainesville,
Florida, 1978.
14. Cowherd, C. Jr., K. Axetell, Jr., C.M. Guenther, and G. Jutze.
Development of Emission Factors for Fugitive Dust Sources. EPA-450/3-
74-037, 1974.
197
-------
LITHIUM
BATTERIES
DIAPHRAGM PUMP
SWITCH
VOLTAGE REGULATOR
Figure 1. Schematic of the IP wind vane sampler.
ENTRANCE CONE'
/J A
WuTRIATOR TUBES
FILTER
SECTION A-A
Figure 2. Isometric and sectional views of the
particle sampling assembly.
198
-------
LAYOUT OF TETHERED BALLOON SAMPLING SYSTEM FOR PROFILING THE PLUME
FROM A LINE SOURCE AND FOR DETERMINING THE DOWNWIND TRANSPORT OF PARTICLES
Figure 3.
LAYOUT OF TETHERED BALLOON SAMPLING SYSTEM FOR PROFILING THE PLUME
FROM AN AREA SOURCE
Figure 4.
199
-------
AN ELECTROSTATICALLY CHARGED FOG GENERATOR
FOR THE CONTROL OF INHALABLE PARTICLES
By:
C.V. Mathai
L.A. Rathbun
AeroVironment Inc.
145 Vista Avenue
Pasadena, California 91107
and
D.C. Drehmel
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ABSTRACT
Devices using electrostatically charged water droplets for the control of
inhalable particulate matter from fugitive emission sources have recently been
introduced commercially (Foggers). This paper introduces a new spinning cup fog
thrower (SCFT) which uses a different method for droplet charging and overcomes
certain drawbacks of commercial Foggers.
The new SCFT is a modified Ray Oil Burner. Centrifugal forces and a high
velocity air stream generate fine water droplets from water flowing into an
atomizing cup. These droplets are charged by directly connecting a high voltage
power supply to the inflowing water and electrically isolating the waterfiuntil
it becomes fog. The typical charge-to-mass ratio recorded is 1.2 x 10 C/g,
with a median mass diameter of 200 ym. The fog covers an area of approximately
20 sq. ft. and the water flow rate can be varied from 8 to 68 Iph.
The SCFT, which requires less than 1 kW power, is being field-tested to
evaluate its potential to control inhalable particles from certain fugitive
emission sources where conventional methods of control cannot be applied
effectively.
200
-------
INTRODUCTION
Although significant progress has been made in the past in controlling
emissions from conventional industrial sources, air pollution regulatory
agencies are becoming increasingly concerned with the non-attainment of ambient
total suspended particulate matter standards, due mainly to fugitive emissions
from various sources. In the past, fugitive emissions have been of concern only
from a nuisance standpoint, increasing soiling potential and decreasing atmos-
pheric visibility only in the immediate vicinity of the source. It has also
been recognized that inhalable particles (<15ym), fine particles (<2.5pm)
in particular, have a dominant role in causing hazards to human health (1) and
degradation of visibility. The fine particles can remain in the atmosphere for
extended periods and can be transported over long distances.
Indiscriminate spraying of large amounts of water has been the most common
control method used in mines and other fugitive emission sources and it has been
only marginally successful (2). In a mine atmosphere, 30-40% is the accepted
value for the dust control efficiency of water sprays (3). However, even this
method has problems with clogged water spray nozzles and the need for large
quantities of water.
It has recently been shown that most industrial pollutants and naturally
occurring fugitive dusts acquire electrostatic charges as they are dispersed
into the air (4). Hoenig (4) has also shown that the charge on the particles of
fugitive emission is usually negative. Therefore, if positively charged water
droplets are sprayed on the dust to be suppressed, the particles which collide
with the charged droplets will agglomerate rapidly and settle out of the atmos-
phere. The collection efficiency of water sprays can thus be improved signifi-
cantly if well-charged water droplets (fog) of approximately 20 to 200 pro
diameter can be generated. Commercially available instruments to generate
charged fog, called Foggers are manufactured by the Ritten Corporation,
Ardmore, Pennsylvania. These Foggers use either a standard pressure-type or
twin-fluid nozzle to atomize the water, thus necessitating a substantial supply
of high pressure water or air for proper atomization. These types of nozzles
have a tendency to clog if the water supply contains a high concentration of
suspended solids (5).
Under EPA sponsorship, AeroVironment Inc., in consultation with
S.A. Hoenig of the University of Arizona, developed an instrument called the
spinning cup fog thrower (SCFT). This instrument was tested in a wind tunnel
facility during 1979 and these test results were reported by Kinsey (5). Since
then the SCFT has been completely redesigned and is now capable of generating
well-charged water droplets which guarantees a greatly increased collection
efficiency. This paper reports the details of this modified SCFT and the field-
testing program currently under way.
THEORETICAL BACKGROUND
A mathematical relationship describing the collection efficiency, E, of
suspended particles by water droplets is given by (6, 7)
201
-------
F - 1 PVD
t - 1 - exp
where Q-, is the volumetric water flow rate, Q is the volumetric gas flow rate,
L is the characteristic length for the total capture process, D is the mass mean
droplet diameter, and TJ is the single droplet collision efficiency.
The single droplet collision efficiency depends on the droplet diameter
and its lifetime in the air. It is determined by several mechanisms of interac-
tion between the droplet and the particles which collide with it. These
mechanisms include direct interception, inertia! impaction, Brownian diffusion,
and electric, diffusiophoretic, and thermophoretic forces (8-11). The dominant
mechanism for large particles (>lym radius) is impaction and interception and
for small particles it is Brownian diffusion. To obtain best collection
efficiency, the droplets must be small enough to provide both an adequate spray
rate per volume of gas treated and sufficient contact time, yet large enough not
to evaporate too quickly. The lifetime of a water droplet depends on the
temperature and relative humidity of the medium into which it is introduced.
Single droplet collection efficiency (which is assumed to be equal to the
collision efficiency) has been investigated by many (8-11). Figures 1A, B, and
C show the collection efficiency plotted as a function of particle radius with
droplet radius, relative humidity, and electric charge on the droplet as varying
parameters, respectively (8). From Figure 1A, it may be noticed that the
collection efficiency of larger particles is greater for larger droplets while
the collection efficiency of smaller particles is greater for smaller droplets.
The minimum in E occurs near a particle radius of ~lym (due to the ineffec-
tiveness of different interaction mechanisms) and, for droplets less than
~55 ym radius, E becomes infinitely low giving rise to the so called
"Greenfield gap." Figure IB shows that the collection efficiency increases
with decreasing relative humidity for smaller particles and E is independent of
relative humidity for larger particles. Figure 1C, the most interesting one to
us, shows that the introduction of electric charge on the aerosol particles and
water droplets completely eliminates the minimum in E and substantially
increases the collection efficiency depending on the amount of charge. Figure 2
shows a comparison of calculated and experimentally measured values of the
single particle collection efficiency as a function of droplet radius for
aerosol particles of 0.25ym at about 20% relative humidity and there is good
agreement between them.
The collection efficiency over a continuous distribution of aerosol
particle sizes by water droplets having a spectrum of radii may be quite dif-
ferent from that shown above for the case of a single droplet. Very little data
are currently available from an operating system. Therefore, our field tests of
the spinning cup fog thrower are even more significant.
THE MODIFIED SPINNING CUP FOG THROWER
As mentioned earlier, Kinsey (5) presented the results of wind tunnel tests
of an earlier version of the SCFT. The poor collection efficiency of that
202
-------
instrument has been attributed to: (1) the induction ring method of charging
the droplets gave a charge-to-mass ratio of only 1 x 10~-"-r c/g; (2) non-
uniformity of the charge on the droplets -- the outer edges of the spray pattern
exhibited a slight negative charge whereas the droplets in the center of the
pattern exhibited a positive charge; and (3) only a small fraction of the
inflowing water was converted to droplets suitable for the capture and removal
of dust while the large fraction ended up as large satellite drops which simply
hit the tunnel walls. Because of these problems, we have developed a completely
different instrument.
The new spinning cup fog thrower is a modified Ray Oil Burner (type AG)
which consists of only one movable part, a hollow steel shaft upon which are
mounted the atomizing cup, the fan, and the motor rotor. Figure 3 is a
schematic representation of the new SCFT. The modifications to the oil burner
include replacement of the fuel tube, air cone, and the spinning cup with non-
conducting materials. The water inlet tube is attached to the spinning cup and
its rear end is connected to the water supply using a rotating seal. We thus
have a completely closed water system which is kept at a high potential of 15 kV
by direct connection to the positive terminals of a d.c. high voltage supply
(12).
Water from the water tube falls into the 3,600 rpm spinning cup whose
inside is fabricated to a gradual smooth taper. Because of the centrifugal
forces, the water becomes a thin film and moves forward into a high velocity air
stream where it breaks up into fine droplets. These droplets are projected
forward by the air stream from the fan and an air butterfly is used to set the
airflow speed, thus controlling the spray pattern. Figures 4A and B show
typical fog patterns obtained with the modified SCFT. The spray patterns cover
approximately 20 sq. ft. and the water flow rate was about 8 Iph when water was
flowing under gravity alone. With the addition of an optional small pump we
were able to increase the water input to 68 Iph. The total power requirement of
this instrument is less than 1 kW. The whole instrument is mounted on a
portable platform for easy transport to a remote location. With the addition of
a small generator, the SCFT can be operated where commercial electric power is
not available.
The droplets were charged by directly connecting the positive terminal of a
15 kV d.c. power supply to the inflowing water. The charge-to-mass ratio was
measured using a sample train designed at AeroVironment. This consisted of a
stainless steel probe tip mounted on a glass midget impinger. The probe was
connected to copper wool packing placed inside the midget impinger and also to
an electrometer. The impinger was immersed in a grounded Dewar flask containing
dry ice. A sample of droplets was isokinetically collected inside the impinger
where the droplets transfer their charge to the copper wool. The current
produced in the copper wool was measured with the electrometer and the mass of
water collected inside the impinger was obtained gravimetrically. Then the
charge-to-mass ratio was calculated. This procedure was repeated several times
and the typical value obtained for a 200 ym droplet (see later) was
1.2 x 10"6 C/g.
203
-------
The size distribution of the water droplets was measured using a cloud
optical array probe for droplets in the range of 30 to 300 urn (diameter) and a
precipitation optical array probe for droplets in the range of 125 to 1,875 urn.
These measurements gave a concentration median droplet diameter of 100 ym and a
mass median droplet diameter of ~200 ym. These determinations were supple-
mented by microscope measurements of droplets collected on greased glass
slides, with a correction factor of 1.26 applied to account for the change in
shape of droplets on slides. This method gave a value of~90ym for concentra-
tion median droplet diameter.
To determine the particle control efficiency of the SOFT, a field test
program has been designed. The field tests are currently under way. Fog from
the SCFT will be applied to the sample volume and the SCFT's dust control
efficiency will be determined under various experimental conditions. It will be
tested with no fog, with uncharged fog, and with charged fog, for a total of 104
test scenarios as shown in Figure 5. The variable parameters for these
scenarios are wind conditions, relative humidity, water flow rate into the SCFT,
fog pattern, and nature and amount of charge on the droplets.
A sample train has been assembled to extract a representative sample
isokinetically. This sample train consists of an elutriator to cut off
particles over 15 ym and a five-stage cascade impactor. The samples collected
on the filters will be analyzed gravimetrically to determine both the total mass
concentration and particle size distribution for each test scenario. From these
measurements, we can calculate the control efficiency of the SCFT under various
experimental conditions.
CONCLUSIONS
The new spinning cup fog thrower is a Ray Oil Burner modified to generate
electrically charged water droplets for the control of inhalable particles. We
were successful in developing a well charged fog which should ensure excellent
particle control efficiency. The total power requirement of the SCFT is under
1 kW; therefore, with the addition of a small generator, it can be operated from
a moving vehicle such as a road sweeper or a front-end loader.
This modified spinning cup fog thrower is being field tested to determine
its particle control efficiency as a function of various field conditions and
water and power requirements. These results will be analyzed to evaluate the
effectiveness of the spinning cup fog thrower in controlling dust from certain
types of sources where other conventional methods can not be applied effectively
and economically.
ACKNOWLEDGEMENTS
The authors are grateful to S.A. Hoenig, J.S. Kinsey, and C. Lyons for
their contributions to the Fogger project at AeroVironment.
204
-------
REFERENCES
1. Cowherd, C. The Technical Basis for a Size-Specific Participate Standard.
J. of Air Poll. Cont. Assn. 30: 971, 1980.
2. Emmerling, J.E., and R.J. Seibel. Dust Suppression with Water Sprays
During Continuous Coal Mining Operations. Bureau of Mines Report of
Investigations, RI 8064, 1975.
3. Courtney, W.G., and L. Cheng. Control of Respirable Dust by Improved Water
Sprays. Bureau of Mines Information Circular, 1C 8753, 92, 1976.
4. Hoenig, S.A. Use of Electrostatically Charged Fog for Control of Fugitive
Dust Emissions. University of Arizona. EPA-600/7-77-131, (NTIS PB
276645), November 1977.
5. Kinsey, O.S., C.E. Lyons, S.A. Hoenig, and D.C. Drehmel. A New Concept for
the Control of Urban Inhalable Particulate by the Use of Electrostatically
Charged Fog. In Proceedings: Fourth Symposium on Fugitive Emissions,
Measurement and Control, New Orleans, La, May 1980 (OTRC -- Environmental
Consultants, Inc., EPA-600/9-80-041, pp. 388-402, December 1980.
6. Calvert, S., J. Goldschmid, D. Leith, and D. Metha. Wet Scrubber System
Study. Volume I, Scrubber Handbook EPA-R2-72-118a, (NTIS PB 213016),
August 1972.
7. Cheng, L. Collection of Airborne Dust by Water Sprays. Ind. Eng. Chem.
Process Develop. 12: 221, 1973.
8. Wang, P.K., S.N. Grover, and H.R. Pruppacher. On the Effect of Electric
Charges on the. Scavenging of Aerosol Particles by Clouds and Small Rain
Drops. J. of Atmospheric Sciences. 35: 1735, 1978.
9. Young, S.C., S. Calvert, and D.C. Drehmel. Fugitive Dust Control Using
Charged Water Spray. (Presented at the 72nd Annual Meeting of the Air
Pollution Control Association, Cincinnati, Ohio, 1979.)
10. Wang, P.K., and H.R. Pruppacher. The Effect of an External Electric Field
on the Scavenging of Aerosol Particles by Cloud Drops and Small Rain Drops.
J. of Colloid and Int. Science. 75: 286, 1980.
11. Pilat, M.J. Collection of Aerosol Particles by Electrostatic Droplet
Spray Scrubbers. J. of Air Poll. Cont. Assn. 25: 176, 1975.
12. Mathai, C.V., and L. Rathbun. New Concepts of Control Applied to Urban
Inhalable Particulates. Phase I Preliminary Report Submitted to U.S. EPA
Contract No. 68-02-3145 EPA, IERL-RTP, December 1980.
205
-------
IU
i
o
o
3
"3.
I
90
c
10-' -
10
0.01
0.1 1.0
Particle Radius (um)
10.0
u 10
UJ
1
3 10
-2
I
o
1? 10
-3
10'
20 esu/on .<•
0.01 O.i 1.0
Particle Radius (un)
10.0
10"
FIGURE 1. Calculated single droplet collection
efficiency (8) in air at 10 C and
900 mb as a function of particle
radius: A) for various droplet radii
and at 75% R.H.; B) for a 72 ym
droplet and various R.H.; C) for a
102 ym droplet with various charges
and 75% R.H.
0.01
0.1 1.0
Particle Radius (um)
10.0
10
-1
10
-2
10'
-3
10
.11
10
30 50 100 200 300 500 1000 3000
Droplet Radius
FIGURE 2. Comparison of cal-
culated (smooth line)
and measured (shaded)
single droplet collec-
tion efficiency (8) as a
function of droplet
radius for uncharged
aerosol particles of
0.25 ym radius in air
(22 C, 1000 mb, and
-20% R.H.)
206
-------
I
Air Fan
Non-Conductive
Air Cone
NJ
O
Water Barrier
Non-Conductive
Spinning Cup
Water
Flowmeter
Hollow Shaft
Motor
\
Non-Conductive
Water Tub
Spinning Shaft
Isolated
Water
Supply
Ground
Water Tube
Alignment
Plate
DC Power
Supply
FIGURE 3. Schematic diagram of the SCFT.
-------
FIGURES 4A and B. Typical spray patterns of the SCFT.
208
-------
VARIABLES
N>
O
VO
VINO
#lt
Medium 3-7 mpB
RELATIVE HUMBWTY
Uw «W*
SPRAY PATTERN
(AIR FLOW SETTING)
WATER FLOW RATE
LOW » 3011*
HIGH
CHARGE ON FOG
A - 10 KV, 2 mA
B - 10 KV, 3 nA
C - 13 KV, 2 M C - 15 KV, mA
0-1} KV, 3 mA
POSITIVE
NEGATIVE
CRAMfl I«>IAI. IM RUNS • EX IRA IIMTOMf R(H.lEI>fi«>N5
FIGURE 5. Field-test program of the SCFT.
-------
RELATIVE EFFECTIVENESS OF CHEMICAL ADDITIVES AND WIND SCREENS
FOR FUGITIVE DUST CONTROL
By: Dennis C. Drehmel, Bobby E. Daniel
U.S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina
Introduction
It has been estimated that in 1976 particulate emissions from open
sources were 580 million tons as compared to emissions from ducted
sources of 15 million tons. Even if most of the particulate emissions
from open sources are large and will settle out quickly, it is evident that
open sources will make a substantial impact on ambient air quality.
Consequently, the Environmental Protection Agency is investigating control
methods for open sources including road carpet for unpaved roads, improved
street sweeping for paved roads, charged sprays for industrial fugitives,
and wind screens for storage piles. In conjunction with those studies,
the control of dust from unpaved roads and storage piles with water and
chemical additives must be determined.
The purpose of this work is to evaluate the cost effectiveness of
chemical additives and wind screens as opposed to other control methods.
Before this can be accomplished, it is necessary to evaluate the wide range
of additive properties and effects. The results reported in this paper
are from wind tunnel* experiments planned to determine the extent of
diversity of chemical additives. Future work on the effects of aging
and weathering on additives will be needed to allow comparisons between
additives and wind screens for control of dust from storage piles.
Comparisons for control of dust from unpaved roads will require tests of
vehicular wear in addition to aging and weathering tests. Thus, the current
study should be seen as beginning a series of studies to evaluate best
practice for fugitive dust control.
Experimental
All tests were conducted in a wind tunnel with a 2 ft square
cross section. The active section of the tunnel was 32 ft long. The
first 16 ft was used for establishing the desired velocity profile.
Next was the 2 ft square dust holder with a 1 ft square recess below
the level of the floor of the tunnel. Past the dust holder was a 14
ft long disengagement section. Besides the active section of the
210
-------
tunnel Is an air return loop. In the air return section is a baghouse
to clean recycled air to the active section of the tunnel. Design of
the active section was intended to produce a turbulent boundary layer
similar to that found for wind passing over open ground or a storage
pile. Details of this design are given by Shaughnessey et al.
The dust holder located in the active section was a removable
section of the tunnel floor. The dust holder was weighed, filled with
coal dust, treated, reweighed, and mounted in the tunnel. The dust
holder was always filled loosely with dust, even with the top, so
that the top of the dust would be even with the floor of the tunnel.
Excess dust was pushed horizontally and brushed off. The treated coal
layer would dry overnight before testing. The wind tunnel would be
operated at several velocities beginning at the lowest and advancing
monotonically. After maintenance of a velocity for 3 min, the tunnel
would be shut down to remove and reweigh the dust holder. With the
dust holder put back, the next velocity was tested. The process was
repeated until the maximum test velocity was reached or until the emission
rate exceeded 200 gr/min- During some runs, isokinetic sampling
after the disengagement section was compared to dust loss results.
For testing the effects of chemical additives, two types of coal
dusts were used to determine the importance of dust particle size. The
mass median diameters of these two coals were different by a factor of
3. For testing wind screens, the test dust was road silt with a mass
median diameter of 148 ym. Ten types of chemical additives
were,tested as listed in Table 1. These additives were organic
salts, oils, acrylics, synthetic polymers, or copolymers. While the
recommended total application rates (water + additive) ranged from
0.016 to 8 gal./100 sq ft, the additive rate would vary more widely
because of recommended dilution factors ranging from 1 (no dilution)
to 6000. For these tests, the additives were diluted with water
within the recommended range and applied with either a spray bottle
or a spray gun. The application rate was within the recommended range
and verified by the volume applied and the weight gain of the coal
layer in the holder.
Results
The emission rate from road dust is shown in Figure 1 for several
cases using wind screens. With no wind screen, the loss of dust began
at 24 ft/sec and accelerated with increasing velocity. With a 2 in. solid
barrier in front of the road dust, incipient reentrainment was in front
of the road dust, slightly delayed but the loss of dust accelerated
faster than before so that the free stream case and the solid barrier
case were very similar in reentrainment profiles. With 1 in. high and 2 in.
high mesh wind screens of 50 percent porosity in front of the road dust, the
threshold velocity for reentrainment was delayed and the reentrainment
profiles were significantly different. A 2 in. high fence of 50 percent
porosity was also tested; it gave similar results to the 2 in. high mesh
screen. Examination of the velocity profiles before and after the 2 in.
211
-------
Table 1. TYPES OF CHEMICAL ADDITIVES TESTED
NS
Additive
Arquad
2C-75
Coal Dyne
Coherex
CPB-12
Gantrez An-119
Lignosulfonate
Oil & Water
SP-301
Pentron DC-3
Polyco 2151
Chemical
dimethyldicocoamonium
chloride
petroleum resin in water
acrylic latex
polymeric anhydride
liqnosulfonate
mineral oil
synthetic polymer
acrylic in water
synthetic copolymer
Additive
Supplier
Armak
Aquadyne
Witco
Wen Don
G.A.F.
Wen Don
Wen Don
Johnson & March
Apollo
Borden
Range of
Cost
$/gallon
1.66-2.16
4.25-4.75
1.60-2.55
3.50-4.75
4.41
2.00
2.00
2.8 - 3.0
4.00
2.95
Recommended Total
Application Rate
gallons/ 100 sq ft
0.016
0.4 - 8
0.7 - 4
0.7 - 6
1.1
1
0.4 - 3
Recommended
Range of
Dilution Factors
30
6000
4-5
10 - 20
1-4
1
1
4-10
-------
600
500
-400
~
s»
«
^300
c
»-
*200
100
I I I
WIND SCREEN TESTS
ROAD DUST 12X12 TRAY
— BEFORE SEC 3
T
T
O NO SCREENS
G 2 in. SOLID BREAK
A 1 in. MESH • FRONT
O2 in. FENCE -FRONT
v 2 in. MESH - FRONT
I
10
20
30
70
80
90
100
40 50 60
VELOCITY, ft/sec
Figure 1. FUGITIVE DUST EMISSIONS (WIND SCREEN TESTS)
-------
to
2
UJ
u
Ul
V)
8
7
6
5
4
3
2
1
I
.2 in. MESH. SO percent OPEN
CENTERLINEOFTRAY
02 8 14
PITOT PROBE POSITION, in.
Figure 2. VELOCITY PROFILE AROUND A WIND SCREEN
-------
600
500
400
5
300
200
100
I I T
CHEMICAL ADDITIVE STUDY
COHEREX 15 percent (BY WT)
12X12 TRAY BEFORE SEC 3
~CO ALDUS!-16 MESH
EF = ft/sec AT 75 gr/min
T
T
ONOADD. EF = 30.5
D 80ml EF = 43.4
A110 ml EF = 50.7
O140 ml EF = 68.6
I
I
I
10
20
30
Figure 3,
70
80
90
100
40 50 60
VELOCITY, ft/sec
FUGITIVE DUST EMISSIONS (CHEMICAL ADDITIVE STUDY)
-------
high mesh screen provides an explanation of these results. As shown in
Figure 2, the upstream velocity profile and the profile at 14 in. down from
the screen are almost the same. At 2 in. down from the screen, the air
velocity is nearly zero up to a height of 1.5 in. At 8 in. down from the
screen, the air velocity is still near zero for the first 0.5 in. above the
test dust-- Hence the 2 in. high screen is providing a wind shadow of 8 to
14 in. which should protect the 12 in. long section of test dust. If the
1 in. high screen provided a wind shadow at the same ratio, it would protect
only the first half of the test dust section from reentrainment. Typical
results of treatment with chemical additives are shown in Figure 3.
With or without additive, the entrainment rate rises rapidly with wind
velocity. Increasing the amount of additive increases the threshold
velocity and the placement of the profile. For each combination of coal
type, additive type, and application method, there is a set of curves as
shown in Figure 3.
To reduce a set of curves, several measures of the velocity-entrain-
ment profile were studied. First, the data was fit with either a second or
third order polynomial and solved for the threshold velocity. Second,
the lowest velocity causing 75 gr/min reentrainment was recorded.
Third, the data was replotted on log-log paper to extrapolate back to
the velocity giving 1 gr/min reentrainment. In general the curve
fitting for threshold velocity and the back extrapolation on log-log
paper gave nearly the same result: the lowest velocity at 75 gr/min
was usually 50 percent higher. The last number, the lowest velocity causing
75 gr/min entrainment, was named the entrainment velocity (or EV) to
distinguish it from the threshold velocity. As an example of the relation-
ship between the entrainment velocity and the threshold velocity, the EV
of the coarse coal dust was 28 ft/sec (19 mph) and the threshold velocity
was 19 ft/sec (11.5 mph). Because the velocity-entrainment profiles often
varied greatly in slope (for example, see Figure 1), it was felt that a
measure of both the intercept and the rate of rise was needed to character-
ize performance. Hence, the entrainment velocity was used primarily for
comparison. Note that entrainment rate varies inversely with entrainment
velocity.
1) The addition of water without additive had
little effect on entrainment velocity.
2) Compacting the coal with 5 Ib/sq ft decreased
the entrainment rate by a factor of 4.
3) The coarser coal dust had a 50 percent lower
entrainment velocity.
Q
The last result agrees with previous work. As noted by Chepil, for sizes
smaller than the most erodible (about 100 pm for soil), erodibility
decreases with particle size.
216
-------
Results for chemical treatment are shown in Table 2. Costs vary
over four orders of magnitude with Coal Dyne (the cheapest by virtue of a
high dilution factor) and SP-301 (the most expensive). All but Coal Dyne
appear to be effective at the manufacturer's recommended rate. In
comparing the entrainment factor at a standard cost of $300/acre, the
most cost effective additives are Lignosulfonate, Oil and Water,
Coal Dyne, and Polyco 2151. Of this group, Polyco 2151 is also effective
at a lower cost of $68/acre.
Discussion of Results
A large number of aggnts to prevent wind erosion and air pollution
have been previously reported. ' ' It is evident from these papers
that the effectiveness of a chemical additive to water for fugitive dust
control will vary dramatically according to the extent of dilution,
application rate, and method of application. Moreover, the optimum
conditions for one additive is often unlike that for another. The
problem is that for each additive, changes in dilution, application
rate, and method of application have different advantages and disadvantages.
The greatest cost effectiveness for a given additive may only be determined
after the effects of all parameters are understood. Consequently, results
as shown in Table 2 must be taken only within the context of the parameters
varied during this study.
Method of application was varied only by changing spray technique.
The effects of multiple applications and multiple applications with
mixing of the top layer were not tested. The two application techniques
tested were application with a spray bottle and with a spray gun. One
of the additives, Gantrez, could not be applied with a spray bottle.
Water modified with this additive would not atomize with the spray
bottle application. For some additives there was no apparent difference
in the effectiveness using different sprays; these were CPB-12, Pentron,
and Coal Dyne. For Coherex, the spray bottle application was more
effective. For Arquad, Lignosulfonate, Oil and Water, and SP-301 the
spray gun application was more effective.
The effect of dilution appeared to be discontinuous in relation
to application,rate. If the dilution was too low in combination
with a low application rate, poor coverage would result. The poor
coverage was evidenced by a non-uniform crust which would have crusted
areas in patches, lumps, or "ropes." Between the crusted areas would be
uncovered areas ranging from cracks to patches as large as a dime.
The effects of application rate on different additives can be
categorized in three groups, depending on the general shape of the curve
relating effectiveness to the rate. In all cases but SP-301, this curve
is either linear or tending to level off. For SP-301 the curve rose
sharply. The three groups are those in which:
217
-------
Table 2. RESULTS OF CHEMICAL TREATMENT
Chemical
Additive**
Coal Dyne
Coherex
£ CPB-12
Lignosulfonate
Oil & Water
Pentron DC3
Polyco 2151
SP-301
Cost at Entrainment Velocity*
Recommended Rate At Manufacturer's At a Cost of
$/acre Recommended Rate $300/acre
0.15
470
354
460
968
667
68
1223
19
55
35
61
61
44
48
55
755
44
32
61
60
30
50
21
*Lowest wind velocity in mph causing 75 gr/min reentrainment.
**Manufacturer's recommended rates not available for Arquad 2C-75 and
Gantrez An-119.
-------
a) Increased rate caused a proportionate rise in effective-
ness: Lignosulfonate, Coherex, CPB-12, Polyco, Coal Dyne.
b) Increased rate caused little or no rise in effectiveness:
Arquad, Pentron, Gantrez.
c) Increased rate caused a rapid rise in effectiveness:
SP-301 only.
It was expected that effectiveness would be proportional to
additive rate and this response was the most common. The response of
Arquad, Pentron, and Gantrez was not expected. This tendency to produce
less effect with more additive may also be viewed as producing more effect
at lower application rates than expected. Either interpretation carries the
warning that merely using more additive may not solve a problem and
that the most cost effective additive rate may not be assumed
but should be found by experiment. Lastly, the response of SP-301 may
be understood in the content of its chemical nature. As a concentrated
latex, SP-301 was observed to form a leathery layer on top of the coal.
At increasing additive rates, this layer becomes more uniform with fewer
imperfections to the point where the coal covered absolutely. Just
before that point, the imperfections in the crust layer are important
sites for disturbing the air flow and for weakening the strength of the
crust. With a small increase in additive, these imperfections are
eliminated with a concomitant dramatic improvement in the effectiveness
of control.
Conclusions
This study helped determine that:
1. Wind screens and fences are effective for prevention of
reentrainment and hence for fugitive dust control.
2. A 50 percent porosity screen or fence will cast an
effective wind shadow of 4 to 7 times its height.
3. A solid barrier provides no protection against fugitive
dust reentrainment. A 50 percent porosity barrier provides protection in
that the free stream wind velocity can be doubled before encountering
emissions equivalent to those with no barrier.
4. All chemical additives studied significantly increased the
threshold velocity and entrainment velocities and hence are effective
for fugitive dust control.
5. All chemical additives except Coal-Dyne were effective at
the recommended dilution and application rate. Coal Dyne could be made
cost effective by lowering the dilution rate to a more typical value (from
6000:1 to 4:1).
219
-------
6. The effectiveness of additives was strongly dependent on
application rate, dilution ratio, and method of application.
7. Too high a dilution ratio may render the additive ineffect-
ive and be coupled with run-off problems; too low a dilution ratio may
give poor coverage.
8. Increasing the total application rate does not affect the
performance of all additives in the same way. Although a proportionate
improvement is most common, some additives will not benefit from increasing
application.
Table for Conversion to Metric Units
To convert to multiply by
3
acfm am /hr 1.70
in. cm 2.54
gal. liter 3.79
mile km 1.61
mph km/hr 1.61
gr g 0.065
acre hectares 0.405
ft m 0.305
sq ft sq m 0.093
220
-------
References
1. Evans, J. S. and Cooper, D. W., An Inventory of Particulate
Emissions from Open Sources, JAPCA, 30 (12) December 1980.
2. Monitoring and Data Analysis Division, U.S. EPA, National
Air Quality and Emissions Trends Report, 1976, EPA-45Q/1-77-002,
December 1977.
3. Levene, B. and Drehmel, D. C., Civil Engineering Fabrics Applied
to Fugitive Dust Control Problems. In Proceedings: Fourth Symposium
on Fugitive Emissions, Measurement and Control, New Orleans, LA,
May 1980, EPA-600/9-80-041 (NTIS No. PB 81-174393), August 1980.
4. Calvert, S. et al., Improved Street Sweepers for Controlling Urban
Inhalable Particulate Matter. Presented at Third Symposium on the Transfer
and Utilization of Particulate Control Technology, March 9-12, 1981,
Orlando, Florida
5. Yung, S., Calvert, S., and Drehmel, D. C., Spray Charging and
Trapping Scrubber for Fugitive Particle Emission Control, JAPCA, 30 (11),
November 1980.
6. Carnes, D. H. and Drehmel, D. C., The Control of Fugitive Emis*
sions Using Wind Screens. Presented at Third Symposium on the Transfer and
Utilization of Particulate Control Technology, March 9-12, 1981, Orlando,
Florida.
7. Shaughnessey, E. J. et al., A Wind Tunnel for Dust Entrainment
Studies. Presented at Third Symposium on the Transfer and Utilization of
Particulate Control Technology, March 9-12, 1981, Orlando, Florida.
8. Chepil, W. S. and Woodruff, N. P., Journal of Advances in Agronomy,
Vol. 15, pp 211-302, 1963.
9. Ambrust, D. V. and Dickerson, J. D., Temporary Wind Erosion
Control, Journal of Soil and Water Conservation, July-August 1971, pp
154-7.
10. Canessa, W., Chemical Retardants Control Fugitive Dust Problems,
Pollution Engineering, 9 (7) pp 24-26, July 1977.
11. Cross, F. L., Control of Fugitive Dust from Bulk Loading
Facilities, Pollution Engineering, 12 (3) pp 52-53, March 1980.
221
-------
PARTICULATE IMPACT COMPARISON BETWEEN CONTROLLED STACK EMISSIONS
FOR A 2000 MM ELECTRICAL GENERATING STATION
By
Dr. Howard E. Hesketh, P.E.
College of Engineering
Southern Illinois University
Carbondale, Illinois 62901
Frank L. Cross, P.E.
Cross/Tessitore and Associates, P.A.
1611 E. Hillcrest St.
Orlando, Florida 32803
ABSTRACT
The impact on air quality is evaluated in regard to the National Ambient
Air Quality Standards (NAAQS) and PSD increments for a coal burning 2000 mw
electrical generating station. The hypothetical western station discussed
is in mountainous terrain, with elevation of up to 1800 m (6000 ft.) within
35 km (20 miles). The comparison is based on the use of NSPS control for
the boilers and BACT control for the fugitive emissions. The "valley" model
is used for the impact evaluation.
Combustion emissions are estimated using two different stack heights.
The flue gas emission control system consists of cold side electrostatic
precipatators followed by limestone flue gas scrubbers. Fugitive emissions
include estimates of controlled emissions from the coal and limestone piles,
conveyors, stackers, cooling towers, dumpers, and waste ponds plus vehicular
traffic in the various work areas. The importance of fugitive emission
control relative to stack emissions in meeting PSD and air quality require-
ments is shown.
INTRODUCTION
It is difficult to comprehend the magnitude and significance of parti-
culate emissions from large emission sources. In order to compare the effects
of elevated controlled stack emissions and low level controlled fugitive
emissions this study assumes a hypothetical 2000 mw coal fired utility located
in mountainous western terrain.
Both the National Ambient Air Quality Standards (NAAQS) and the new
Prevention of Significant Deterioration (PSD) regulations are applicable to
utilities and mean that many millions of dollars in control system costs must
be expended. In addition, the costs will differ by millions of dollars
dependent upon the actual impacts and the specific alternate control techno-
logies deemed necessary to meet these needs. The effects are reflected in
all of capital, operating and energy costs.
The PSD particulate increment restrictions which exist for Class I,
J II areas respectively in yg/m3 are 5,
mean and 10, 37, and 75 for the 24 hr maximum
II, and II areas respectively in jjg/m3 are 5, 19. and 37, for geometric annual
222
-------
Stack Emissions and Height
NSPS dictate that the allowable participate emissions should not exceed
(0.03 lb/106 btu).(2j At full load this is equivalent to approximately
1.93 x 1010 btu/hr heat input or 73.4g/sec (582 Ib/hr) of particulate
emissions. This requires a control efficiency of about 99.6% which can
be obtained by ESPs, scrubbers or fabric filters. Filters would be most
efficient, however they do not remove the required S02 without incorporation
of alkali spray dry facilities which are currently in a demonstration phase
of development. ESPs can achieve this level of particulate control but a
large SCA is required and ESPs do not remove S02 (spray dry systems can use
ESPs for the solids removal, but this technology has not been accepted).
The new type FGD scrubbing systems, which incorporate gas conditioning, have
been shown to achieve 99.6% particulate removal with simultaneous S02 control
when operated with good mist elimination.!3) For emission comparisons, this
study assumes that ESPs are used in parallel with limestone FGD wet scrubbers
and that the scrubbers have good mist eliminators. No plume reheat is
assumed.
Stack emission plume height is dependent upon stack height, stack
diameter, type of control system used, exit vel . and temp., and meteorological
conditions. This is to say that wet scrubbing emissions would be at adiabatic
saturation temperatures of about 328°K (130°F) unless plume reheat is used.
For purposes of the comparisons presented in this paper, it is assumed that
the plumes are unheated and to evaluate the results of reheat, two stack
heights are used. These heights are 114.4m (350 ft.) and 152.5m (500 ft.).
For this facility, four separate stacks are assumed, one for each boiler,
each having an exit velocity of 25.2 m/s (82.6 ft/sec) at full load. These
data are summarized in Table I for each of the four stacks.
TABLE 1. EMISSION INVENTORY AND STACK DATA
FOR FOUR STACKS
Comparison
Parameter
Stack Height, m
Stack Diameter, m
Exit Velocity, m/sec
Exit Temperature, °K
Particulate Emissions, g/s
Fugitive Emissions
Because of the theoretical nature of this report, certain assumptions
on the general arrangement of the power station and the types of equipment
for coal handling had to be assumed. Figure 1 is a sketch of the general
arrangement of the plant indicating the size and location of coal piles,
223
-------
conveyors, cooling towers, and train unloading facilities. Table II includes
a list of the control systems considered for the fugitive sources of emissions,
and the estimated emission from these sources. A 90 day coal and chemical
storage pile quantity is assumed. A 1.1 limestone supplied to S02 removed
stoichiometry is assumed.
It is assumed that two 100 car unit trains could be dumped each day
at full load operation of the utility. This would amount to a transfer of
20,000 tons coal per day. Actual transfer would occur over a 6 to 8 hr
period of time. The values shown in the table are shown as if unloading were
continuous so values are lower than actual during unloading periods.
Dispersion Modeling
Models and Wind Data
Mountainous terrain is assumed for the location of the power station,
and therefore the EPA Valley Model is used to estimate downwind ground level
concentrations (GLC). The meteorological data used in the Valley Model
calculations for the 3 hr winds is presented as a frequency distribution
in Table III. These data are restrictive, but are reported in the mountainous
regions of the western United States. EPA "default" data are used for the
24 hr meteorology. A 10 meter plume height is used in these calculations for
all ground level emissions sources that are actually less than 10 m high.
The fugitive emission sources were considered to originate at 4 locations
throughout the area as shown in Figure 1, and appropriate coordinates for
these are used as inputs to the Valley Model.
All models are subject to certain imperfections and deficiencies. Yet
certain models do consider factors that at least attempt to correct for
conditions not considered in other models. It is recognized that Williams
and others are in the process of developing high terrain models which may be
more accurate, yet at this time the EPA accepts the Valley Model as being
most applicable to mountainous terrain modeling. Data basis are being
obtained for attempted validation of the EPA Valley and CRSTER (for conven-
tional terrain) models and for other comparable models such as the "Texas"
and "Small Hill" models. It is not the intent of this paper to attempt to
validate the Valley Model nor is it desired to extend the dispersion predic-
tions to the 1 ug/m3 level over hundreds of miles as has been done by others.
The Valley Model is used here as it is the only high terrain model accepted
by the EPA without validation(7).
Terrain
A schematic of the downwind terrain showing relative distances and
elevations is given in Figure 2. This figure is not to scale and shows only
the specific locations corresponding to the wind directions used in Table
III. Obviously, other wind directions and terrain conditions could be used.
224
-------
FIGURE 1
POSSIBLE PLANT ARRANGEMENT
(Simplified Flow Diagram)
[ABLE II
INVENTORY OF FUGITIVE EMISSIONS
w Modeling
T Sites
II I 111! •
Rail Car Dumper
Conveyor
Stacker
Coal Pile (and Reclaimer)p
-------
Figure 2. Panoramic Sketch of Generatiny
Station and Local Topography
TABLE III
METEOROLOGY 3 HOUR WIND MATRIX
CASE
NO.
1 .
2
3
4
Dt'SCJUP'i'JON
Low wind spued
Unstable
Iliyh wind upobcJ
Un;; table
.
Low wind upccd
Stable
•ludium. wind upd
J t a b 1 c
WIND
SPEKD
' in/ sec
1.0
2.0
'i.'i
n . 9
U . '1
10.1 •
0 . 7
1.0-
1.0
2.6
2,'j •
2. 0 '
WIND
DIRECTION
20G
192
203
1U3
179
103
101
179
107
109
IOC
102
STAUILITV
CLASS
1
1
1
4
•1
4
c
C
(j
** c
^
^
Personal files
Case <\ is most restrictiVL-; these data are average for this case
and are used for their GLC effects study
226
-------
Fiyurc 3A - North Receptors
24 Hour Downwind Participate GLC's Due to North Wind :nd EPA Star
Meteorological Data for Receptor Locations as
noted from 2000 MM Utility
Figure 3B - NHE Receptors
24 Hour Downwind Particulate GLC's Due to North Wind and EPA Star
Meteorological Data for Receptor Locations as
noted from 2000 MW Utility
o 700
^ COO
£ t SCO
'Z .£ 400
"Z^ 300
<* £ 200
;;.•? 100
"'d 0
N)
-J
> i-
•r- >
700
600
500
400
300
200
100
0
-100
o --,
w o>
u. ^
-J *J
.o C
^ O
.35 -
30 -
25'' -
20
10
Fugitive
Stack v.
cm
o E
O —.
f~ o*
O 2.
OJ
N
5 10 15 20
Distance Dov.nward fro:n Utility, i'J-l
KEY.:
Fugitive —
Stack — -
10 15 20
Distance Downward from Utility, KM
-------
Figure 3C - Nt Receptors
2< Hour Downwind Particulate GLC's O'ji to North Wind and EPA Star
f'.eteorolosical Data for Receptor Locations as
noted from 2000 ;•'.* Utility
Figure 30 - NNW Receptors
24 Hour Downwind Particulate GLC's Due to Nor:h Wind i.nd EPA Star
Meteorological Data for Receptor Locations as
noted .frora 2000 MW Utility
N5
N3
00
700
600
500
400
300
200
100
0
-100
.
2 700
, o i* 600
/ •- 5 500
/ .2 £ 400
r / g .300
/ 2 i1 200
-100
-
-
^
^s
_
KEY:
•o
en
=3 E
o — -
Cl •
•M 0
-------
Figure 3£ - NW Receptors
21 Hour Downwind Participate GLC's Due to North Wind and EPA Star
Meteorological Data for Receptor Locations as
noted from 2000 MW Utility
> v.
•r-
£ X
HI
"
700
COO
500
400
300
200
100
0
-100
35
3U
25
20
15
10
i:uyi ti vi:
Stack
10 1G 20
Downward from Utility. KM
Impact Analysis Results
Most data show only GLC's for stack emissions. These data show GLC's
for the fugitive and the elevated stack emissions. The complete data are
given in the figures for all receptors for 24 hour wind persistance averaging
times for the most restrictive meteorological conditions. A screening series
of model runs were made to establish the most restrictive meteorological
data-which turned out to be Case No. 4 as shown in Table III for the particu-
lar wind directions and terrain chosen for this study. An actual wind speed
of 2.C in/s and stability Class 6 are used.
Figure 3 shows these data for GLC's due to ground level and elevated
emissions. The total emission increases at the specific ground receptors
would be the sum of the two individual offects noted in Figure 3. A profile
of receptor heights is superimposed on each part of Figure 3. Note that
only the 114.4 in elevated source height impact is shown as the impact from
the 152.5 in stack are very similar.
229
-------
Figure 3C shows that for relatively smooth terrain the downwind GLC's
of fugitive dust decrease exponentially with distance as would be expected.
This figure further shows that effects of controlled elevated emissions
are much less than the effects of the fugitive emissions for all distances
shown (up to 25km). A comparison of Figures 3A-E show all GLC's decrease
with deviation from source wind direction. Figure 3A indicates that a large
elevation receptor would block further significant transport of fugitive
emissions whereas Figure 3B suggests that a lesser receptor elevation could
result in higher fugitive downwind GLC's at greater distances.
GLC's from elevated emissions are more significant at remote receptors.
Figure 3/\ shows the effects of both distance and plume height. This figure
shows source emissions could be more significant than fugitive emissions,
under these circumstances, only when high elevations exist downwind.
It is apparent from this type of comparisons that coal burning facilities
must not overlook fugitive emissions when considering their air impact
analyses.
Summa ry
Comparisons of downwind particulate GLC's from a hypothetical 2000 mw
coal burning utility show that "controlled" fugitive emissions can have a
much greater impact on air quality than controlled elevated stack particulate
emissions. The data from this study are summarized in Figures 3A-E, and
show that Class I PSD increments would De exceeded because of fugitive
emissions and depending upon the terrain, distance, and meteorological
conditions, Class II increments may also be exceeded.
In addition, the following comparisons are shown relative to controlled
fugitive dust and controlled stack particulate emissions from a large coal
fired utility on downwind particulate GLC air quality:
- fugitive dust can be much more significant
- fugitive dust GLC decreases with downwind distance for relatively
smooth terrain
- fugitive dust GLC decreases as source wind direction deviates
from source-receptor direction
- a large elevation receptor (>200 m above source) may prevent
further downwind transport of fugitive dust
- substantial downwind elevations of <200 in may result in higher
fugitive GLC's at greater distances
- GLC's from elevated emission become more significant at remote
receptors
- GLC's from elevated emissions may dominate when downwind receptor
elevations are "very" high.
230
-------
Bibliography
(1) Prevention of Significant Deterioration, Clean Air
Aramendments of 1977, PL-95-95.
(2) U.S. Federal New Source Performance Standards, Subpart
Da, for New Electric Utility Steam Generating Units
started after Sept. 18, 1978, promulgated 6/11/79, CFR
44 FR 33580.
(3) Grimm, C., et al, "The Colestrip Flue Gas Cleaning System",
CEP, Vol. 85, No. 5, pp. 51-57, Feb., 1978.
(4) Evaluation of Fugitive Dust Emission from Mining, IERL,
EPA, Contract 68-02-1321, April, 1976.
(5) Potential Site, Study for a 1500MW Coal Fired Generating
Station near Creston, Washington. Prepared for the Energy
Facility Site Evaluation Council, Olympia, Washington,
prepared by URS Company, (1979).
(6) Currier, E.L., and Neal, B.D., "Fugitive Emissions from
Coal Fired Power Plants", paper NO. 79-11.4, proceedings
72nd Annual APCA meeting, Cincinnati, Ohio, June 24-29,
1979.
(7) "Winds of Change Reshape Air Pollution Modeling", Chemical
Engineering, Vol. 87, No. 24, pp. 35-37, December 1, 1980.
231
-------
NA59-07175
OPERATING EXPERIENCE AND THE TECHNIQUES
IN THE CONTROL OF COAL DUST EMISSIONS
FROM LARGE STORAGE PILE AT NANTICOKE TGS
By: N. Krishnamurthy
Ontario Hydro Corporation
700 University Avenue
Toronto, Ontario
Canada, M5G 1X6
W. Whitman
Nanticoke Thermal Generating Station
Nanticoke, Ontario
Canada
Y.V. Nguyen
Ontario Hydro Research Division
800 Kipling Avenue
Toronto, Ontario
Canada, M8Z 5S4
ABSTRACT
The large coal storage and handling operation at Nanticoke Thermal
Generating Station has caused significant coal dust control problems.
Investigations were made to identify the principal sources of airborne coal
dust. The dust emission rates and particle size distribution from different
sources were measured and dispersion calculated. An estimate of the impact
of these emissions on the area surrounding the station were made. Physical
model studies were carried out in an open channel water flume to determine the
optimum shape of the coal pile and derive optimum techniques in operation.
Several coal dust control agents to reduce the dusty character of the coal
were tested and evaluated. The application of these results to the
Nanticoke TGS coal pile management and experience are discussed.
INTRODUCTION
Nanticoke TGS is a 4 000 MW coal-fired generating station located on the
north shore of Lake Erie. Although this station has not operated at its full
capacity to date, ultimately it will receive a total of over 7 000 000 Mg of
eastern US and western Canadian coals per year and store 4 500 000 Mg on the
stockpiles for the winter peak period when the lake is closed to shipping =
During the shipping season, on average, a 30 000 Mg shipload is received each
day. Material handling of this coal is accomplished with conveyers, 2
stacker-reclaimers, 1 radial stacker, 10 tractor-scrapers, and other mobile
equipment.
232
-------
Since 1976, Ontario Hydro has received a number of complaints from the
cottagers in the northeast of the plant that coal dust is being blown from the
plant coal piles onto their properties. The dustfall rate in the cottage area
was in the range of 2.5 Mg/(km2.mo) to 5.9 Mg/(km2.mo)/l/ which is below the
Ontario ambient air quality criterion of 7 Mg/(km2.mo)/2/ but coal dust is a
nuisance problem and the Ontario Environmental Protection Act prohibits adding
any contaminent to the environment which can cause impairment to the quality
of the natural environment. The Ontario Ministry of the Environment expressed
concern about this problem to Ontario Hydro and accordingly Ontario Hydro
formed a task force to study the problem and recommend solutions. This paper
outlines the findings on the sources and control of coal dust emissions. The
control technology is not complex, but for such a large materials handling
operation, an organized, integrated approach was required. The approach taken
was to identify the principal sources of coal dust emissions, estimate the
impact of emissions on the surrounding environment, and then modify the coal
handling operation to minimize the dust generation.
COAL HANDLING OPERATION AND SOURCES OF COAL DUST
Coal Handling Operation
Figure 1 shows schematically the coal handling operations at Nanticoke in
1977. Coal was unloaded at the dock hopper from a self-unloading ship at a
maximum rate of 5500 Mg/h. This operation was dusty under high wind
conditions.
From the dock hopper coal was conveyed to the transfer house in two
covered conveyors and then either to the plant or to two stacker-reclaimers
for stockpiling. Each stacker-reclaimer is rail-mounted and forms two long
and narrow temporary surge piles, one on each side of the rail track. This
stockpiling operation generated dust especially under high wind conditions.
From the surge piles, coal was moved to the main storage pile by
tractor-scrapers to provide the winter coal requirement for the plant. The
tractor-scrapers weigh 100 Mg when loaded and pulverize the coal on the haul
roads to a fine dust which can be picked up in the wake of moving vehicles.
Haul roads were identified as sources of airborne dust. After the coal was
unloaded from the tractor-scrapers onto the main coal pile, the coal surface
was smoothed out and compacted by a rubber-tired bulldozer and sprayed with
water using a water wagon. This reduced the amount of airborne dust but under
high wind conditions and with dry coal, dust emissions could still be seen
from the top surface of the coal pile as well as from the sides and at the
leading edges. In the winter, when the lake is closed to shipping, coal was
reclaimed from the main storage pile by the tractor-scrapers running down the
slope of the pile. This operation generated significant amounts of dust due
to the wind turbulence at the reclaiming area.
Figure 2 shows a recent modification to the coal pile handling operation
to accommodate western Canadian coal. This coal contains a much larger
proportion of fine particles than the eastern US coal and worse problems were
anticipated with this coal. This coal is received at a separate dock and
transferred to a temporary surge pile created by a radial stacker. This
233
-------
stacker is equipped with a water spray system which forms a curtain around the
falling coal. The dock hopper is fitted with a dust hood and rubber curtains
at the opening. These facilities control and contain dust effectively.
Airborne Dust and Impact of Emissions
Airborne dust sampling was carried out at the locations identified above
viz stacker and reclaimer, haul roads, and on top surface and sides of the
piles. The results of the dust concentrations at these locations are shown in
Table 1. The geometric mass mean diameters of the samples collected were:
(a) Stacker-reclaimers surge piles - 18 iim-26 pm
(b) Haul roads with vehicle travelling - 12 um-30 jim
(c) Top and sides of coal pile due to turbulent wind conditions
12 um-35 um
These particles were small enough to be dispersed to the cottage area and
thus had to be controlled.
The impact of these dust emissions on the area surrounding the station
was assessed using a downwash dispersion model/2/. The maximum coal dust
deposition rate is a function of wind speed and distance downwind from the
coal pile, and it was estimated that the maximum deposition at the nearest
cottage was between 3.8 Mg/(km «mo) and 5.2 Mg/(km «mo) which is similar to
the monthly dustfall measurements made by Ontario Ministry of the Environment,
2.5 Mg/(km .mo) to 5.6 Mg/(km «mo)/l/« This agreement was probably fortuitous
but it did indicate that dust from the coal pile could be a nuisance problem
for the cottage area and furthermore, the dispersion model showed that dust
can be dispersed to as far as 3 km downwind from the plant. This indicated
strongly that there was a need to control coal dust from the Nanticoke coal
pile operation.
PHYSICAL MODELLING STUDY
As one step in the control program, Ontario Hydro hired a consultant to
examine the effect of wind turbulence on the coal pile by constructing a
physical model, 720 times reduction of the coal pile and placing it in an open
channel water flume to simulate the effect of wind turbulence on the coal
pile.
This modelling study allowed local wind flows to be visualized and
permitted modifications to be introduced without difficulty into the coal pile
model to reduce the wind blowing coal dust. The transport of coal particles
was simulated by the introduction of silica sand into the model area. Areas
of scouring or quiescence could be easily identified visually and problem
areas could be examined by the introduction of a coloured dye at the problem
spot.
The above study made the following recommendations for reducing the wind
blown dust around the coal pile.
234
-------
(a) The coal pile surface should be smooth and erosion should be controlled.
(b) The coal pile should have gentle slopes, as low as 1:7 or should be
treated with a dust suppressant*
(c) Semi-permeable fences with 50 percent open area should be installed on
and around the coal pile to reduce the wind velocity and turbulence.
(d) Operation of equipment which creates dust should be restricted to wind
protected areas.
Practical applications of these recommendations were considered when
modifying the coal pile operation. The pile building techniques were changed
to allow compaction of the coal on the sides and sealing the surface with
waste oil to give a smooth surface and prevent wind and water erosion.
Due to space limitations, the gentle sloping coal pile with a gradient of
1:7 was not possible.
The semi-permeable fences with 50 percent open area, would need to be at
least 4 m high and portable. The fence would be too costly to construct to
withstand the expected wind load and not practical for regular operation.
Thus, the use of these fences has been postponed pending further studies of
other coal dust control techniques.
The major application to the coal pile of the results of the modelling
study was a complete revision of the coal reclaiming technique so that the
operation of equipment that creates dust can be restricted to wind protected
areas. In previous years, coal was reclaimed from a bowl surrounding the
reclaim hopper by driving the tractor-scrapers down the slope. Modelling
studies showed that there was considerable turbulence in the bowl and that
dust generated by the scraper movement would not settle out. In 1977/1978,
the coal pile was redesigned so that the pile itself acted as a protective
berm allowing drifting coal to settle.
COAL DUST SUPPRESSANT EVALUATION
A wide range of coal dust suppressant agents were studied in both the
Research Division laboratory and on experimental coal piles at Nanticoke.
Weight loss data and the capability of the suppressant agent to retain its
dust binding properties for longer periods were used as performance
indicators.
A bitumastic emulsion, and waste oil were selected for trials on the
Nanticoke coal pile. The bitumastic emulsion was used to seal the sides of
the coal pile and the initial characteristics of the treated surface were most
satisfactory. But after about two months, weathering began to deteriorate the
quality of the surface crust and large sheets of the treated surface were
slipping down the pile sides exposing an unsealed surface. Waste oil was used
to seal the haul roads on the coal pile, at an application rate of
approximately 1 L/m^. The resulting coal surface had no crust, but the small
coal particles were well retained on the coal surface. The effectiveness of
the oil depended on how soon the tractor-scrapers were allowed to travel on
235
-------
these haul roads after they were oiled, and the extent of coal spillage from
tractor-scrapers. At Nanticoke the haul roads need to be oiled every two
weeks to maintain their dust-free character.
COAL DUST CONTROL PROGRAM AT NANTICOKE
Following the identification of the sources of airborne coal dust at the
coal pile, a coal dust control program was implemented which took into account
the results of the modelling and coal dust suppressant studies as far as
practically possible. This program involves modifications to the coal
handling operation at the dock hoppers, stacker-reclaimers, and at the main
coal pile areas.
At the dock hopper for receiving the eastern US coals, the distance that
coal falls through the air has been minimized by lowering the height of the
ship boom as the ship unloads. The ships have installed sprays, hoods, and
side screens around the end of the boom. At the dock hopper for receiving the
western Canadian coal, the hopper is equipped with two rotating sector
screened hoods to provide a totally protected environment when unloading coals
from the ship boom. Any airborne dust particles in the hood can be collected
by a baghouse filter.
At the stacker-reclaimer surge piles, dust generation has been reduced by
lowering the free fall distance of the coal. A water spraying to agglomerate
any remaining dust from this falling coal is under consideration. Water
spraying has also been applied to the radial stacker for the western Canadian
coal and is found to be effective for dust control.
On the main eastern US and western Canadian coal piles, the wind
turbulence along the pile sides has been reduced by a revised method of
compacting coal and using waste oil to retain dust on the surface. The dust
from the haul roads on the coal piles is being controlled by using waste oil
and the dust from the top surface of the pile is being controlled by using two
water wagons. The water wagons are used whenever there is any coal movement
on the coal pile in the summer as well as in the winter.
The coal reclaiming technique from the main coal piles has been
completely revised, as mentioned earlier, to permit the coal to be reclaimed
from the lee side (Figure 2) of the pile so that the pile itself acts as a
protective berm for any airborne coal dust to settle out.
The effectiveness of this control program has been monitored by a
particulate monitoring network which has been in operation since 1977.
COAL DUST MONITORING NETWORK
The monitoring network includes nine monitoring sites located between
0.2 km and 4.2 km from the coal pile. All sites monitor dustfall, the one
site about 1 km to the northeast of the coal pile, has additional monitoring
equipment consisting of a standard hi-vol, an isokinetic hi-vol sampler and
wind velocity and direction instrumentation/4/.
236
-------
The dustfall data show no significant change following implementation of
the coal dust control program but the coal fraction (by volume) in dust has
decreased since 1978. The total carbon fraction are in the range of
10 percent to 30 percent of the total dustfall sample.
The hi-vol data between July 1977 and April 1979 show four exceedances of
the 24 h Ontario air quality criterion of 120 ug/m3; however, in only one case
was the free carbon content of the dust sample high (40 percent); the other
three samples contained less than 10 percent free carbon. The hi-vol sampling
data were of limited value in characterizing the impact of coal pile
operations in the cottage area because the sampler has a low collection
efficiency at high wind speeds. Only at consistently low wind speed
(< 10 km/h) is a representative sample of airborne particulate likely to be
collected and these are not the conditions likely to result in significant
fugitive emissions from the coal pile.
The isokinetic hi-vol sampling data are the only data considered to be of
value in characterizing the impact of coal pile operations in the cottage
area. This sampler, developed by Ontario Hydro, Research Division, consists
of a cyclone collector ahead of the standard hi-vol sampler/5/. This sampler
collects dust samples isokinetically over a wind speed range of 15 km/h to
45 km/h whenever wind is from the direction of the coal pile. The sampler was
in operation after the implementation of the coal dust control program and
therefore, the sampling data cannot be used to assess the effectiveness of the
control program. However, the data point out that high dust concentrations
(> 150 ug/m3) occurred infrequently, were of short duration (< 10 hours) and
the highest frequency of occurrences was in the winter months. Even during
these high dust episodes our very limited data show that the free carbon
contents of the dust were about 12 percent.
Since the implementation of the coal dust control program, the number of
complaints of coal dust in the cottage area to the northeast of the coal pile
have reduced. Ontario Hydro is continuing to operate this monitoring network
to examine the effect of significantly greater usage of the fine western
Canadian coals on particulate levels in the cottage area and the monitoring
data obtained to date will provide a basis for comparison of the relative dust
levels.
CONCLUSIONS
1. The main sources of airborne coal dust are the operation of the
tractor-scrapers on the haul roads and the wind blown dust from the top
surface and sides of the coal piles.
2. Dispersion model studies indicate that if coal dust emissions are not
controlled, dust can be dispersed as far as 3 km downwind from the coal
piles.
237
-------
3. The windblown dust from the sides of the coal pile has been reduced by a
revised method of compacting coal and covering it with waste oil. The
dust from the top surface of the pile has been controlled by spraying
water from two water wagons and the dust from the haul roads controlled
by spraying waste oil on the roads.
4. The coal reclaiming operation has been completely redesigned and coal is
now being reclaimed from the leeward side of the pile so that the coal
pile itself acts as a protective berm to allow any airborne coal dust to
settle out.
5. The particulate monitoring network is a valuable tool to assist in
designing coal dust control strategy, but conventional dust sampling
equipment such as hi-vol and COM tape samplers are of very limited value.
ACKNOWLEDGEMENTS
The authors acknowledge the contributions of the members of the task
force on coal dust control, including Messrs. G.R. Taber, D.W. Hotchkiss,
T.B. Walls, M.J. Northfield, J.R. Ryan, M.R. Booth, R.W. Glass, and others.
Modelling studies were performed by Morrison, Hershfield, Theakston, and Rowan
Limited.
Permission for publication of this paper has been given by Ontario
Hydro. The opinions stated are those of the authors and do not necessarily
reflect official Ontario Hydro policy.
REFERENCES
1. Nguyen, Y.V. Airborne Coal Dust at Nanticoke GS. In: Summer Operating
Conditions. Ontario Hydro Research Division Report, 77-467-K, 1977.
2. The Environmental Protection Act, Ontario Regulation 872/74, 1971.
3. Nguyen, Y.V. Evaluation of Coal Dust Control Agents. Ontario Hydro
Research Division Report, 78-70-K, 1978.
4. Glass, R.W. Nanticoke TGS Coal Dust Monitoring. Ontario Hydro Research
Division Report, 79-556-K, 1979.
5. Glass, R.W. An Isokinetic Hi-Vol Air Sampler. Ontario Hydro Research
Division Report, 78-79-K, 1978.
6. Booth, M.R. Report of Task Force on Coal Dust Control. Thermal
Generation Division Report, NA59-07112.
7. Booth, M.R., Whitman, W.N., and Krishnamurthy, N. Coal Dust Management
at Nanticoke TGS. Report TG-07112. This paper was presented at the
Canadian Electrical Association, Thermal and Nuclear Power Section
Meeting, in Calgary, October 15-17, 1979.
238
-------
TABLE 1. SUMMARY OF COAL DUST CONCENTRATIONS
AT VARIOUS SAMPLING LOCATIONS DURING
HANDLING OF US AND WC COAL
Date
Aug 3,77
12,77
Aug 4,77
12,77
Aug 4,77
4,77
12,77
12,77
12,77
12,77
Aug 3,77
3,77
Aug 2,77
3,77
Location
1
2
3
4
7
8
9
10
11
12
14
15
5
6
Type
of
Coal
US
WC
US
WC
US
US
US
US
US
US
US
US
us
us
Dust
Concentration
mg/m3
1.2
7.3
1.3
18.9
75.3
97.2
146.7
171.4
194.4
185.8
7.7
0.9
0.9
0.3
Comment on Location
Near the stacker pile,
during stacking.
Near the north stacker
coal pile in a breeze.
In the wake of a van
travelling on the haul
roads and ramps.
On the side of the main
coal pile in a breeze.
On top of the main coal
pile.
Note;
Airborne dust samples were collected 1.5 m above the ground level near
the sources identified above by drawing air through a 0.8 cm diameter
nozzle onto a millipore filter paper, 0.45 jim pore size. Isokinetlc
samples were obtained by measuring the wind speed at about 15 cm above
the sampling nozzle with a portable anemometer and the sampling rate
adjusted to match the wind speed.
Legend
mo - Month
// - Reference Number
239
-------
tacker/
Reclaimer No 1
NANTICOKE TGS
ORIGINAL COAL HANDLING LAYOUT
DOCK HOPPER TO:
1 - POWERHOUSE BUNKERS DIRECTLY
• 2 -STACKER/RECLAIMER PILES
• 3 - S/R PILES TO MAIN STORAGE PILE
BY MOBILE EQUIPMENT
240
-------
stacker
• reclaimer
S/R1 /
5400 Mg/hr
conveyor
FIGURE 2
PRESENT COAL HANDLING & BLENDING
FACILITIES LAYOUT
prevailing
wind
direction
241
-------
MODELING SMOKE PLUME OPACITY
FROM PARTICULATE CONTROL EQUIPMENT
By: David S. Ensor and Phil A. Lawless
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, North Carolina 27709
Stanton J. Cowen
Atmospheric Research Group
464 West Woodbury Road
Altadena, California 91001
ABSTRACT
The ability to predict plume opacity from control equipment is important
in meeting air pollution regulations. This paper describes the applications
of a model to electrostatic precipitators and the implications to the mass/
opacity relationship for variation in inlet size distribution.
INTRODUCTION
Objective
The objective of this research was to develop convenient and accurate
methods to permit the design of control devices to meet opacity limitations.
Included are methods implemented on computation equipment such as program-
mable calculators, personal interactive microcomputers, and large computers.
Background
The opacity determined by observation of the smoke plumes emitted from
stationary sources has long been used as a means of regulation in the United
States. The Clean Air Act currently specifies that, if a source meets the
mass emission standards, a special regulation can be developed for that stack.
Thus, an understanding of the significant causes of opaque plumes is impor-
tant for the application of the regulation. Also, it has been observed that
identically designed plants may yield plumes with differing opacities. A
systematic investigation of the causes of excessive opacity requires the de-
velopment of scientific tools. The current research traces its history to
the work of Ensor.(1) Ensor assumed that the particle size distributions
were lognormal and the particles were spherical. Results of extensive com-
puter calculations were reported in graphs to allow an estimation of opacity.
242
-------
A different approach was adopted by Cowen et al.(2) They wrote general
opacity prediction programs for programmable calculators, which allowed com-
putations without the restriction of a lognormal particle size distribution.
Work is in progress to apply these programs to microcomputers. (3)
COMPUTATION OF OPACITY
The present discussion is limited to in-stack opacity as measured by a
transmissometer. Even with this restriction, the light scattering equations
that form the basis of opacity are complex and require modern computational
equipment .
Light Scattering
Opacity is defined by the light transmitted through the smoke by:
0 - (1 - I/Io) 100 (1)
where Op is percent opacity, I0 is incident light intensity, and I is trans-
mitted light intensity.
The light transmittance is related to the aerosol or particulate concen-
tration by:
I/Io - exp (- bext L) (2)
where bext is the extinction coefficient (m2/m3) and L is the light path-
length, the stack diameter in this case.
For computational work and data analysis, it is useful to express the ex-
tinction coefficient in terms of mass concentration as :
S
b - — W (3)
ext p
where Sv is specific volume extinction coefficient, m2/cm3; W is mass concen-
tration, g/m3; and p is particle specific gravity, g/cm3. The Sv is the re-
ciprocal of the "K" used previously by Ensor.(l) The Sv is a function of
only the particle size distribution and refractive index. The Sv is computed
by:
s „ 3/D2 Q(D.A.m) N(D) dD (4)
v 2/D3 N(D) dD
where D is particle diameter, ym; N(D) is particle size distribution by num-
ber; Q is light scattering efficiency, the effective light scattering area
divided by the projected particle area; X is the wavelength of light, usually
0.55 ym for visible light; and m is the particle refractive index (a complex
number), the real part of which is the ratio of the speed of light in media
divided into the speed in vacuum and the imaginary part of which describes
the absorption or conversion of light to heat. The Q is computed by applying
243
-------
the Mie equations; the solution involves a series of spherical Bessel's func-
tions. The details of the computation are described by Ensor(l) and Cowen
et al.(2) and will not be covered here.
The magnitude of the specific extinction coefficient is a measure of
opacity expected per unit particle volume. The relative nature of the
parameter is very useful in analysis of opacity problems. The extinction
coefficient is a function of concentration as well as size distribution and
refractive index. Thus, the extinction coefficient is not as useful as the
specific extinction coefficient for opacity prediction.
Control Device Models
Each type of air pollution control equipment acts on the inlet aerosol
in its own way. The reader is referred to Mosley et al.(4) and Lawless et
al.(5) for detailed descriptions of the electrostatic precipitator models.
ELECTROSTATIC PRECIPITATOR/OPACITY MODELING
Combustion Aerosol
The aerosol from the combustion of coal is composed of two distinct pop-
ulations of particles: The fly ash mode in the 5 to 50 ym particle diameter
range, formed from the gaseous phase and composed of fused ash; and the ac-
cumulation mode in the 0.08 to 0.2 ym diameter range, composed largely of
condensed materials. Typically, the accumulation mode contains less than
2 percent of the total mass. The physical basis of aerosol formation from
coal combustion has been extensively reviewed by Flagen and Friedlander.(6)
Experimental results from field measurements of the aerosol size distribution
were reported by Markowski et al.(7)
Ensor(8) reported a relationship between the geometric mean diameter of
the fly ash mode of the distribution and the ash content of the coal. The
distribution appears to be a stronger function of the ash concentration than
was predicted by the Flagen-Friedlander theory.
Size Distribution Case Studies
The model fly ash distributions were compiled from the literature rely-
ing heavily on Gooch and Marchant(9), and Ensor et al.(10,ll) These size
distributions are idealized but are a good representation of expected size
distribution for coals of low, medium, and high ash content. The model fly
ash size distributions used in this study are summarized in Table 1.
Computational Results
The electrostatic precipitator conditions are summarized in Table 2. The
precipitator parameters were constant for computation of all three particle
size distribution cases.
244
-------
Specific Extinction Coefficient Distributions
The behavior of the specific extinction coefficient as a function of par-
ticle diameter provides insight into how an electrostatic precipitator affects
the opacity-to-mass concentration relationship. The convention first used by
Ensor et al.(12) for atmospheric aerosol is a plot of the Abext/AlnD as a
function of diameter on semilogarithmic paper. If ASV is plotted, the area
under the curve is a portrayal of the total specific extinction coefficient.
In Figure 1 the incremental curves are presented for the medium ash coal model
size distribution. The specific extinction distribution is shown for the in-
let and the outlet conditions. For each model aerosol, the extinction distri-
butions have the same behavior. As the aerosol passes through the precipitator,
the relative importance of the submicron particles becomes less for opacity,
while the - 6-ym particles dominate opacity. This observation is contrary to
the popular conception that submicron particles determine the opacity and large
particles determine the mass concentration. The explanation is the dominance
of the rapping re-entrainment for both mass concentration and opacity. The
precipitator is effectively acting as a. source of aerosol in the - 6-ym par-
ticle diameter region.
Specific Extinction Coefficient as a Function of Specific Collection Area
In Figure 2 the specific extinction coefficient is shown for the inlet
and after each section as a function of cumulative specific collection area
(plate area/gas volume flow rate). The aerosol distribution is transformed
to approximately the same shape in each case. The emissions gradually assume
the properties of a rapping puff as the primary aerosol is removed.
The significance of the rapping puff in determining the outlet opacity
characteristics requires an examination of rapping. Extensive data on the
rapping size distribution were reported by Gooch and Marchant(9) for six
sites and by Ensor et al. (10,11) for rapping distribution at single sites.
The striking conclusion is that the rapping puffs have quite similar size
distributions for a wide range of equipment designs and test sites. The range
of the rapping puff distributions and corresponding specific extinction co-
efficients are shown in Table 3.
The expected specific extinction coefficient (and extremes) computed
from the rapping distribution is 0.85 ± 0.3 cm2/m3 and this variation can ex-
plain much of the reported variability of opacity-to-mass correlations.
The in-stack opacity as a function of outlet mass concentration is shown
in Figure 3. Two features of the curves are apparent: the test cases produce
similar opacity-to-mass curves, a consequence of the similar outlet Sv values;
and the curves, although computed with much different inlet conditions, are
consistent with the federal regulations.
The present analysis indicates that, for high efficiency electrostatic
precipitators, the outlet optical properties will generally be dominated by
the properties of the rapping puff. There are three reasons why this would
not occur: 1) with low efficiency precipitators, the properties of the inlet
distribution dominate the outlet; 2) precipitation problems occur such as back
245
-------
corona, excessive sneakage, and sparking; and 3) the second mode of the size
distribution makes an optically significant contribution to the emissions.
The effects of low efficiency precipitators on mass-to-opacity relationships
generally have not been considered in studies such as those reported by
Brennan et al.(13) In fact, one popular way of varying opacity and concen-
trations to develop correlations is to shut off sections of the precipitator.
This could be an explanation of why correlations for an individual stack may
be excellent, but correlations of one stack to another may be poor.
The second consideration of pathological precipitator operation is less
likely on high efficiency installations. However, one interesting well-
documented case was reported by Ensor et al. (11) for a large cold-side pre-
cipitator operating in extreme back corona. The submicron particles were
collected with abnormally low efficiency in the precipitator. The specific
extinction coefficient was about twice as large as calculated for a rapping
puff. (The rapping puff data are consistent with those measured at other
installations.)
The importance of the accumulation mode (0.1 ym) on in-stack opacity is
worthy of discussion. The fly ash cases were computed without consideration
of the accumulation mode. The specific extinction coefficients for the con-
densed metal oxides and soot are shown in Table 4 with estimates of the con-
tribution to the emission. Soot is very effective in screening light as
indicated by the magnitude of the specific extinction coefficient. For this
reason, the optical properties of soot and amorphous carbon are an active
topic of research.(14)
For widely separated distributions as reported here, the total specific
extinction coefficient is a sum of the mass fraction times the specific ex-
tinction coefficient for each component. As shown in Table 5, the contri-
bution of the case of the accumulation mode is small and much less than
15 percent for the soot case.
CONCLUSION
The in-stack opacity/mass concentration relationships have been ex-
plored for electrostatic precipitators. The opacity for high efficiency
precipitators will be determined by the rapping puffs. The relationship of
opacity-to-mass concentration for these precipitators may be quite similar
because rapping puffs have similar particle size distributions. Low effi-
ciency precipitators will retain the specific extinction coefficients of the
inlet ash. Nonideal conditions such as back corona may alter the specific
extinction coefficient by reducing the efficiency of submicron particles.
The combustion aerosol accumulation mode is believed to have an insig-
nificant effect on opacity from high efficiency electrostatic precipitators
unless soot is present.
246
-------
ACKNOWLEDGEMENTS
This research was supported by U. S. Environmental Protection Agency
Grant No. R-806718010 to the Atmospheric Research Group, U. S. Environmental
Protection Agency Cooperative Agreement No. R-805897-03 with the Research
Triangle Institute, and Research Triangle Institute program development
funds.
ENDNOTES
1. Ensor, D. S. Smoke Plume Opacity Related to the Properties of Air Pol-
lutant Aerosols. Ph.D. Dissertation, University of Washington, 1972.
2. Cowen, S. J., D. S. Ensor, and L. E. Sparks. TI-59 Programmable Calcu-
lator Programs for In-Stack Opacity, Venturi Scrubbers, and Electro-
static Precipitators. U. S. Environmental Protection Agency.
EPA-600/8-80-024 (NTIS PB80-193147), May 1980.
3. Cowen, S. J., D. S. Ensor, and L. E. Sparks. In-Stack Opacity Computer
Programs. Report in preparation under U. S. Environmental Protection
Agency Cooperative Agreement No. R-806718010.
4. Mosley, R. B., M. H. Anderson, and J. R. McDonald. A Mathematical Model
of Electrostatic Precipitation (Revision 2). U. S. Environmental Pro-
tection Agency. EPA-600/7-80-034 (NTIS PB80-190994), February 1980.
5. Lawless, P. A., J. W. Dunn, and L. E. Sparks. A Computer Model for ESP
Performance. (Presented at the EPA Third Symposium on the Transfer and
Utilization of Particulate Control Technology, Orlando, Florida,
March 9-12, 1981.)
6. Flagan, R. C. and S. K. Friedlander. Particle Formation in Pulverized
Coal Combustion—A Review. In: Recent Developments in Aerosol Science,
Shaw, D. T. (ed.). New York, John Wiley, 1978.
7. Markowski, G. R., D. S. Ensor, R. G. Hooper, and R. C. Carr. A Sub-
micron Aerosol Mode in Flue Gas from a Pulverized Coal Utility Boiler.
EST. 14:1400-1402, 1980.
8. Ensor, D. S. Aerosol Emissions Due to the Combustion of Coal. (Presented
at the Symposium on Plumes and Visibility: Measurements and Model Com-
ponents, Grand Canyon, Arizona, November 10-14, 1980.) Submitted to
Atmospheric Environment.
9. Gooch, J. P. and G. H. Marchant. Electrostatic Precipitator Rapping Re-
Entrainment and Computer Model Studies. Electric Power Research Insti-
tute Report No. EPRI FP-792, 1978.
10. Ensor, D. S., P. A. Lawless, and L. E. Sparks. Evaluation of the United
McGill Electrostatic Precipitator. (Presented at the EPA Third Sym-
posium on the Transfer and Utilization of Particulate Control
Technology, Orlando, Florida, March 9-12, 1981.)
247
-------
11. Ensor, D. S., S. J. Cowen, R. G. Hooper, G. R. Markowski, R. Scheck,
J. D. Farrell, and J. Burnham. Evaluation of the George Neal No. 3
Electrostatic Precipitator. Electric Power Research Institute Report
No. EPRI FP-1145, August 1979.
12. Ensor, D. S., R. J. Charlson, N. C. Ahlquist, K. T. Whitby, R. B.
Husar, and B. Y. H. Liu. Multiwavelength Nephelometer Measurements in
Los Angeles Smog Aerosol: I. Comparison of Calculated and Measured
Light Scattering. In: Aerosols and Atmospheric Chemistry,
Hidy, G. M. (ed.). New York, Academic Press, 1972.
13. Brennan, R. J. , R. Dennis, and D. R. Roeck. Review of Concurrent Mass
Emission arid Opacity Measurements for Coal-Burning Utility and Indus-
trial Boilers. U. S. Environmental Protection Agency.
EPA-600/7-80-062 (NT1S PB80-187420), March 1980.
14. Cowen, S. J., D. S. Ensor, and L. E. Sparks. The Relationship of Fly
Ash Absorption to Smoke Plume Opacity. (Presented at the EPA Third
Symposium on the Transfer and Utilization of Particulate Control Tech-
nology, Orlando, Florida, March 9-12, 1981.)
O Inlet
V Outlet
Geometric Mean Diameter • 10 fim
Geometric Standard Deviation - 4
Refractive Indax 1.525-0.06 i
Specific Gravity 2.27 a/cm3
Wavelength of Light » 0.65 ion
RTI/EPA Electrottatic Precipitator Modal
01
Particle Diameter jum
Figure 1. Incremental specific extinction coefficient as a function
of particle diameter at the inlet and outlet of an electro-
static precipitator.
248
-------
1.4-
1.3-
« 1.2-
o
N- 1.1-
E
I i.o-
I 0.9-
^ 0.8 ,
o
I 0.7-
•c
i2 0.6
0
I 0.5-
> 0.4-
o
|0.3-
W 0.2-
0.1
0
2gr/ft3
O 6XAsbinCo.lDB-4,,m
D 10%AihinCoalDB-10»,m
A 20% Aril In Coal Dg - 60 urn a, . 5 8 gr/ft3
18 3 e/lt)3
•*— Ripping Puff Dfl - 6 ^m og . 2S
RTI/EPA MODEL
RwittlvHv 2x1010 a-cm
Ripping Lou 10X
Section
0 10 20 30 40 50
Specific Collection Area m2/m3/sec
Figure 2. Change in specific volume extinction coefficient
as a function of precipitator collection area.
100.0
10.0
1.0-
0.1
O&XAihinCoilDa-4jun o,-3
D 10\A.hinCoalDB-10Mm aR-4 B.lSj/m3
A 20% Arii In Coal D(-SO Mm 0|-B 18.3 «/m3
Spfeif Ic Gravity 2.27 g/cm3
T.mperatur. 433° K
Pressure 1.0 Atm
Computed with RTI/EPA Model
0.001
0.001
Figure 3.
0.01 O10 1.0 10
Particulate Mass Concentration g/m3 Actual Conditions
100
0.01 0.1 1.0
Approximate Mass Emissions lb/106 Btu at 3% 02
10
Too
In-stack opacity as a function of outlet mass
concentration for the example size distribution.
249
-------
TABLE 1. MODEL AEROSOL SIZE DISTRIBUTIONS FOR THE FLY ASH MODE*
Case
Very low
Average
High ash
ash
ash
coal, 5%
content coal, 10%
content coal, 20%
4.
9.
18.
Concentration
6
2
3
g/m3
g/m3
g/m3
(2
(4
(8
gr/ft3)
gr/ft3)
gr/ft3)
Dg
(Vim)
4
10
50
sv
0g (cm2/m2)
3
4
5
1.
0.
0.
447
784
231
* Particle specific gravity =2.27 g/cm3; refractive index = 1.525-0.05 i;
wavelength of light =0.55 pm.
TABLE 2. SUMMARY OF ESP PARAMETERS
Wire-to-plate spacing 0.127 m
Wire radius 2.54 x 10~10 m
Area per section 1477 m2
Sections 4
Gas flow rate 117 m3/sec
Temperature 433° K
Pressure 1.0 atm
Length of section 9.76 m
Resistivity 2 x 1010 ft-cm
Applied voltage up to 55 kV
(depending on section and operating conditions)
Current density up to 4 x 10~" A/.m2
(depending on section and operating conditions)
Sneakage/section 10%
Rapping re-entrainment/section 10%
[An average cold-side puff size distribution from
Gooch and Marchant(9) was used.]
Sigma (standard deviation of gas velocity) 0.34
250
-------
TABLE 3. RANGE OF SPECIFIC VOLUMETRIC EXTINCTION
COEFFICIENTS EXPECTED FOR RAPPING PUFFS
Diameter
(ym)
5.0
* 6.0
7.0
°g
1.5
2.5
1.5
2.5
1.5
2.5
Specific volumetric
extinction coefficient
0.784
1.17
0.637
0.896
0.534
0.761
* Generally believed to be the most representative,
TABLE 4. PROPERTIES OF ACCUMULATION MODE*
Maximum contribution by mass (%)
Geometric mean diameter (ym)
Geometric standard deviation
Specific gravity (g/cm3)
Refractive index
Condensed
material
2.0
0.1
1.2
2.6
1.50
Soot
10
0.08
2.0
2.0
1.96-0.66 i
Specific volume extinction coefficient
(cm2/m3) 2.86 11.15
* Wavelength of light is 0.55 ym.
TABLE 5. CONTRIBUTION OF THE ACCUMULATION MODE TO
SPECIFIC VOLUME EXTINCTION COEFFICIENT
Case
Low ash-soot
Medium ash-soot
High ash-soot
Low ash-condensed mode
Medium ash-condensed mode
High ash-condensed mo.de
Sv
(cm2 /m3)
0.961
0.894
0.833
0.884
0.899
0.748
Percent
increase
9.9
9.8
15.5
1.4
0.5
3.8
251
-------
TETHERED BALLOON PLUME SAMPLING OF A
PORTLAND CEMENT PLANT
By
James A. Armstrong
Philip A. Russell
Myron N. Plooster
Denver Research Institute
University of Denver
Denver, Colorado 80208
ABSTRACT
A remote-controlled tethered balloon sampling system was used to collect
aerosols from a Portland cement plant which was experiencing a persistent
plume opacity problem. The persistent plume was always associated with the
prior formation of a water condensate plume which at times would be detached
from the stacks of the plant's electrostatic precipi tator. When the ambient
air temperature was high enough that the water condensate plume did not form,
the plume was clear.
Aerosol samples were collected from both persistent and clear plumes.
The plume sampling was conducted concurrently with in-stack gas sampling.
The plume samples were analyzed by scanning and transmission electron micro-
scopy to determine the concentration and composition of the collected
aerosols. The field program is discussed as well as analytical results and
a proposed hypothesis of the persistent plume formation.
INTRODUCTION
The existence of a high-opacity persistent plume from a Portland cement
kiln located in Glens Falls, New York, is dependent upon whether or not a
water condensate plume is first formed when exhaust gases exit the plant's
electrostatic precipi tator. When the ambient air temperature is sufficiently
cold, quenching of the stack gases leads to the formation of a water conden-
sate plume. This plume prevails for a few tens of meters until it evaporates
due_to mixing with drier ambient air. From this point on, a very dense,
bluish-white Plume_Persists for kilometers downwind of the plant. Depending
upon the ambient air temperature, the water condensate plume often is de-
tached from the precipitator stacks. The persistent plume opacity is observed
eS ''
hatth n the ambient t«»Perature is warm enough
that the_water condensate plume does not form at all, a clear plume from
the precipitator stacks results.
252
-------
The water condensate plume and the persistent plume occur consistently
during cold winter periods but are not observed during the summer. In the
spring and fall, when the morning ambient air temperature is sufficiently
cold, the plume formations occur until the air temperature rises above
an appropriate level.
An electrostatic precipitator is used to control particulate emissions
from the exhaust gases of the cement kiln. The kiln is fired using No. 6
fuel oil. The plant utilizes the "dry" process in which crushed raw
materials are fed into the kiln in their dry state. The precipitator, how-
ever, was designed for controlling emissions from a "wet" process kiln in
which crushed raw materials are mixed with water and fed in a slurry to the
kiln. The plant utilized the wet process when the precipitator was first
installed, but soon after changed over to the dry process. In lieu of
replacing the precipitator, a water injection system was installed downstream
of the kiln and upstream of the precipitator to condition and cool the
exhaust gases to the precipitator.
Each of the two stacks of the precipitator is equipped with a trans-
missiometer to continuously monitor in-stack opacity levels. Under normal
plant operations these levels are less than 10%. The production of the per-
sistent plume with an opacity often as high as 85% occurs outside of the
stacks, however, which indicates that additional control strategies are in
order to eliminate the problem. Before such strategies can be instituted,
however, definitive knowledge concerning the formation process and composi-
tion of the persistent plume need to be established. A number of past
in-stack sampling programs had been conducted at this plant to obtain this
knowledge but the results were inconclusive(1,2).
During April 1979, the Denver Research Institute (DRI) conducted a
field study using a tethered balloon sampling system to collect particles
directly from both persistent and clear plumes of this plant, in order to
determine the nature of particles causing the visibility problem. The DRI
field study was sponsored by the Particulate Technology Branch of the
Industrial Environmental Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina. The plume sampling was con-
ducted concurrently with in-stack gas sampling by a Northrop Services field
team sponsored by the Gaseous Emissions Research Section of the Environmental
Sciences Research Laboratory, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. In-stack measurements of h^SOi,, S02, sulfur
salt ions, NH3, and HC1 were made by the EPA/ESRL sponsored team(3).
TETHERED BALLOON SAMPLING SYSTEM
A detailed description of the tethered balloon sampling system used
during the investigation is presented elsewhere(A). The system basically con-
sists of a remote controlled particle sampler which is carried aloft by a
balloon tethered to a portable winch. The sampler, which weighs 950 g,
contains eight 37 mm diameter filters in cassette holders which are connected
through a motor-driven rotary valve to a piston-type suction pump. The
filter being sampled and the sampling time are controlled via a ground-based
253
-------
transmitter and an airborne receiver/servo system. An airborne verification
transmitter and ground-based receiver monitor the sampler for correct opera-
tion. Collection substrates for this program included 0.4 pm Nuclepore
filters and prefired glass fiber filters. The sampling flow rates were 3-9
£/min and 4.1 Vmin for the Nuclepore and glass fiber filters, respectively.
This sampler also contains a heated-surface semiconductor gas detector, the
output of which is coupled to the verification transmitter, so that continuous
relative information of the sampler position in a plume envelope is provided
to the ground-based receiver.
A blimp-shaped, red polyethylene balloon, which is inflated with 3-25 m3
of helium, is used as the airborne platform for the particle sampler. The
balloon is attached to a battery-powered winch by means of a lightweight
tetherline having a breaking strength of 535 newtons.
FIELD OPERATIONS
The sampling operations took place on April 24 and 25, 1979- Weak sur-
face radiation inversions occurred in the early morning hours of both days,
during which time light winds of 0.5 to 1.5 m/sec were from the north. The
topography of the area where the cement plant is situated slopes down to the
Hudson River, which is located a few kilometers south of the plant. As the
ambient air temperature increased from ~8°C to ~13°C on both days, the wind
remained light but shifted to a southerly direction. This occurred by 0800
hours EST on April 24 and by 0900 hours on April 25. The persistent plume
existed and was sampled during the time that the winds were from the north.
The plant electrostatic precipitator is oriented in a north-south direc-
tion with the two stacks on the west side of the unit. The stacks are 1.83
m in diameter and are separated by approximately 8.7 m. The north stack was
sampled by the EPA/ESRL sponsored field team. With the winds from the north,
the individual plumes of the two stacks merged soon after the exhaust gases
exited the stacks.
Limited plume sampling was accomplished on April 24. The sampling oper-
ations were started at 0607 hours from the open field to the south of the
precipitator. The visible persistent plume was sampled at a location just
downwind of the zone where the condensed water plume evaporated. The plant
was experiencing operational problems, and by 0628 hours shut down due to a
blockage in the cement kiln. The plant was out of service until 0300 hours
on April 25. The samples collected on April 24 were not analyzed.
Sampling of the persistent plume at a location just downwind of the
condensate plume was started at 0516 hours on April 25. The cement plant
operated continuously during the entire plume sampling period which ended
at 1015 hours. During this time, the kiln feed rate varied between 1,630
and 1,660 kg/min, the fuel feed rate varied between 85.9 and 87.1 £/min,
and the opacity of the north and south precipitator stacks varied between
8 and 3% and 3 and 41, respectively. The kiln exhaust gas temperature to the
spray tower was 288°C and the precipitator inlet temperature was 111°C
during the sampling period.
254
-------
Balloon sampling operations were documented by photographic slides
and time-lapse movies. A 16 mm movie camera, equipped with a zoom lens and
an intervalometer, was positioned at a site ~100 m west of the precipitator.
The movies proved to be invaluable in verifying the vertical position of the
balloon in the persistent plume; in determining the residence time of a parcel
of gas passing through the condensate zone; in clearly showing the phenomenon
of the condensate plume becoming more detached from the stacks and the persis-
tent plume becoming less intense or opaque as a function of time as the
ambient air temperature increased; in determining the horizontal and vertical
sampling distances from the precipitator stacks; and, finally, in confirming
that the balloon was indeed in the clear plume during the later morning
sampling operations. This was evidenced by observing the flight behavior
of the balloon in the clear but turbulent plume.
The ambient air temperature was 5°C at the start of plume sampling
operations on April 25. At this time the water condensate plume formed
within the first 3 m and evaporated at -45 m downwind (south) of the stacks.
From 0516 to 0651 hours, four samples of the persistent plume aerosols were
collected by the balloon system with sampling times ranging between 5 and 20
minutes. The estimated average horizontal and vertical sampling distances
from the top of the precipitator stacks were 55 and 35 m, respectively, for
these samples. By 0700 hours, the ambient air temperature had increased to
8°C, causing the condensate plume to form and evaporate at estimated downwind
distances from the stacks of 10 and 25 m, respectively. At this time the
wind was beginning to shift to the northeast. Between 0756 and 0823 hours,
one 20 minute and one 5 minute sample of the less intense persistent plume
were collected. For these samples the balloon was located at average hori-
zontal and vertical distances from the stacks of 30 and 35 m, respectively.
By 0900 hours the ambient temperature was 13°C, the wind had shifted to
the south, and clear plume conditions existed. Two 30 minute samples of
the clear plume aerosols were taken between 0903 and 1015 hours. The
estimated average horizontal and vertical sampling distances from the
stacks were 25 and 50 m, respectively.
LABORATORY ANALYSIS
Examination of materials collected on the Nuclepore and glass fiber
substrates was first made using a scanning electron microscope (SEM) equipped
with an energy dispersive X-ray spectrometer (EDXS) capable of detecting
elements with the atomic number of sodium or greater. All substrates were
precoated with approximately 20 nm of carbon by vacuum evaporation prior to
SEM examination. Particle size parameters, concentration, morphology, and
elemental composition were estimated using the information provided by this
examination. The discussion that follows concerns the selected Nuclepore
substrates containing aerosols collected in the persistent plume over a 20
minute period starting at 0523 hours on April 25 (Sample 25~2) and in the
completely clear plume over a 30 minute sampling period starting at 0903
hours on the same date (Sample 25-9).
These selected samples were further examined using a transmission elec-
tron microscope (TEM). Selected area electron diffraction (SAED) patterns
255
-------
were obtained from crystalline materials. Substrates to be examined by TEM
were coated with a thin layer (^ 10 nm) of carbon by vacuum evaporation,
transferred to a TEM grid, wicked with chloroform to dissolve the Nuclepore
polycarbonate filter, and covered with another thin layer of carbon by vacuum
evaporation. Particles collected on Samples 25~2 and 25-9 were also sub-
jected to treatment with NH3 vapor and heat (100°C for 30 minutes and 150°C
for k2 hours) with subsequent TEM-SAED examination.
SEM-EDXS Results
In Sample 25-2 the dominant particles by number had submicrometer dia-
meters and were observed to be of a liquid nature. SEM micrographs of this
sample are presented in Figures 1 and 2. Figure 1 shows these particles when
examined by SEM at a low tilt angle to the normal electron beam axis; Figure
2 illustrates the droplet or liquid morphological characteristics of these
particles when examined at a high tilt angle.
An examination of the individual liquid particles by EDXS showed
that they were rich in sulfur. These particles were observed to be stable in
all but the most intense SEM beam, where some decomposition was observed.
The liquid morphology of these sulfur-rich particles strongly suggested
that HaSOij was the dominant component. HaSOi, possesses an extremely low
vapor pressure and will exist as a liquid in the vacuum necessary for SEM
analysis(5,6).
Mineral particles, with diameters ranging from 0.2 to 15 ym, were
observed in Sample 25-2 but at concentrations much lower than the liquid
particles. The composition of the mineral particles, which were sometimes
alone but usually in agglomerates, included quartz, clays, feldspars, and
calci tes.
The characteristic morphology of the most numerous particles on Sample
25-9 is illustrated in Figures 3 and k. These particles were again sub-
micrometer in size. They, however, were roughly spherical and did not "wet"
or spread on the collection media surface. Figure 3 is a low tilt angle
view of these particles while Figure k is taken at a higher tilt angle. The
round submicrometer particles were rich in sulfur and were stable in an in-
tense electron beam. As in Sample 25-2, mineral particles in a size range
of 0.2 to 15 ym in diameter were observed at concentrations much lower than
the sulfur-rich particles.
It was assumed that the liquid particles of Sample 25-2 existed as
spheres in the persistent plume. In order to determine the equivalent spheri-
cal diameters of the liquid particles on the Nuclepore substrate, the volumes
of these droplets as observed on high tilt SEM micrographs were estimated
assuming they had the shape of spherical segments. The diameters of spheres
having equal volumes to the spherical segments were then calculated. The
equivalent spherical diameters of a significant number of liquid droplets
were used to plot a cumulative number distribution on a log-probability
graph. The resulting geometric median diameter by count, d , for the
liquid particles of Sample 25-2 was 0.32 ym with a geometri^standard
256
-------
Figure 1. SEM micrograph. Sample 25-2
at -25° tilt angle to normal beam axis,
K3
Ul
Figure 3. SEM micrograph, Sample 25-9
at -20° tilt angle to normal beam axis,
Figure 2. SEM micrograph, Sample 25-2
at ~60° tilt angle to normal beam axis.
Figure 4. SEM micrograph, Sample 25-9
at -55° tilt angle to normal beam axis.
-------
deviation, a , of 1.69. /The number concentration of these particles was
calculated tS be 1.46x10 particles/^ of sampled air.
The cumulative number distribution of the sulfur-rich particles from
Sample 25-9 was also plotted on a log-probability graph. For this sample,
d = 0.35 ynvand a =1.4. The particle number concentration was calculated
tocbe 1.l8x105 part9cles/£.
The size distributions of the two samples were unexpectedly similar and
the persistent plume sulfur-rich particle concentration was only one order of
magnitude larger than that of the clear plume. Based upon these facts, the
difference in opacity between these plumes would not be expected. A possible
explanation for the high opacity of the persistent plume is discussed later.
TEM-SAED Results
TEM-SAED examination of Sample 25~2 particles indicated that many of the
original particles on the Nuclepore substrates became connected during the
transfer process to TEM grids to form "bridges" of material. Some particles
were relatively unchanged and were observed to still be oblate when examined
at high tilt angles. While particle morphology was altered by the chloroform
dissolution of the polycarbonate substrate and the addition of the second car-
bon coating, the composition appeared unchanged. The material in the par-
ticles was heat labile in a more intense electron beam. No residues were
left after the material evaporated and no SAED patterns were observed. This
is shown in Figures 5 and 6.
No major changes were observed for Sample 25~9 particles in comparing
SEM and TEM micrographs. These particles under TEM examination appeared to
have thick shells surrounding less electron-dense matrices. They were not
heat labile. The particles were also not crystalline since they produced
no SAED patterns. Electron density appeared homogeneous; i.e., there were no
obvious inclusions of different material. A TEM micrograph of one of these
particles is shown in Figure 7.
After discovering the liquid nature of the sulfur-rich persistent plume
particles, portions of both the persistent and clear plume samples were ex-
posed to ammonia gas in order to determine if the particles would transform
into crystals. If the particles contained HaSOit, it would be expected that
exposure to NH3 would cause ammonium salts of sulfur oxides to form. TEM
grid-mounted samples were either exposed to vapors from NHi»OH in a petri
dish or had a gentle jet of NH3 gas directed onto their surfaces. Both
aerosol samples reacted with NH3 to form stable crystals. Depending if
the particle samples were precoated with carbon or not and/or if the samples
had been subjected to a heat treatment before being exposed to NH3, the
resulting crystals possessed different structural characteristics (micro-
crystalline structure, often showing preferred orientation, versus thin
rod-like crystals) and produced different SAED patterns. Rigorous crystal
indexing was not conducted on the crystalline particles generated by NH3
exposure but it was established that similar diffraction patterns could have
been produced by crystals of ammonium sulfate [ (NH^aSOj , ammonium hydrogen
258
-------
Figure 5. TEM micrograph, Sample
25-2, chloroform filter dissolution
only, low electron beam current.
Figure 6. TEM micrograph, Sample 25-2,
after exposure to higher beam current.
Figure 7. TEM micrograph, Sample 2.5-9.
259
-------
sulfate [NHifHSOj, ammonium sulfamate [NHi»S03 NH2], ammonium dithionate
sulfamic acid [NH3S03], and ammonium hydrogen sulfate sulfamate
OaNHa]. A more detailed discussion concerning the generation
of the stable crystals from the persistent and clear plume aerosol samples
is presented elsewhere(7)•
SULFUR MASS BALANCE
To make any estimate of the mechanism of formation of the particles in
the persistent plume it was necessary to compare the quantity of matter in
the particles with the quantity of the "raw material" in the stack gases.
Since sulfur was the main identifiable element in the particles sampled from
the persistent plume, and the S02 concentration in the stack gases was
measured, a sulfur mass balance computation was carried out.
The effluent from the electrostatic precipitator was released into the
atmosphere through two 1.832m diameter stacks. The total stack cross-sec-
tional area was thus 5.25 m . The S02 concentration in the stack gas was
-200 ppm as determined by the EPA/ESRL sponsored in-stack measurements (3)•
The mean diameter of the persistent plume at the sampling point was
determined from several frames of the time lapse movie taken at the time of
sampling. The balloon itself provided a length scale for plume measurements.
The balloon is k m long; the diameter of the plume, from the movie, was
about 7-5 times the length of the balloon, or 30 m. Assuming that the plume
was cylindrical, its cross-sectional area was then ~700 m . The dilution
ratio (plume area/stack area) was thus 700/5.25 = 133- The average concen-
tration of S02 at the sampling point was then 200 ppm/133 = 1-5 ppm.
The filter sampler was in the persistent plume for 20 minutes. Since the
sampling rate was 3-9 £/min, the total volume of air processed for this filter
sample was 78 £.
To make an estimate of the mass of the particulate matter collected on
the filter sample, it was necessary to make some assumptions as to its
chemical composition and density. For the purposes of this calculation, it
was assumed that the particles were pure H2SOit (molecular weight 98, density
1.834 g/cm ). (This means that this calculation probably gave an upper
limit for the quantity of sulfur present in the particles in the plume.)
Integration of the particle size distribution gave a total mass of 2.67 x
10 1T g. This particle size distribution was obtained by counting and sizing
the particles on a filter area of 1.1 x 10~5 cm . The total filter area was
8.55 cm . The mass loading of particles per unit volume of air was thus:
2.67x10-11gx8l55_cmi=2.65 x 10^ f-
1.1 x 10 ' cm 78 £
The volume of S02 gas needed to generate this massjJoading was easily
found using the ideal gas law; the result was 6.k x 10 £ S02/£ air or a
volume concentration of 0.064 ppm. This was about k% of the quantity of S02
estimated to be present in the plume, a relatively small fraction.
260
-------
A similar calculation was carried out for the particles observed on
Sample 25-9, taken in the clear plume later in the day. The mass loading
of particulate matter on this filter sample was smaller by almost exactly a
factor of 10. However, it is not possible to conclude from this that the
relative ratio of sulfur in the particles to sulfur available in gaseous form
was also smaller by a factor of 10, for two reasons: 1) the physical charac-
teristics of the particles in Sample 25~9 were different in the electron
microscope studies, suggesting a somewhat different chemical composition (i.e.,
less like H2S04); and 2) because the plume was invisible, it was not possible
to estimate the plume size from the photographic records, so that the dilution
ratio at the sampling point is not known.
POSSIBLE MECHANISMS OF AEROSOL FORMATION
The combined results of the in-stack and plume sampling investigations
have generated nearly as many questions as answers. Nevertheless, a plausible
case can be made for the hypothesis that the particles in the persistent
plume were sulfuric acid most likely containing some dissolved ammonium sul-
fate and/or ammonium hydrogen sulfate, created by the aqueous oxidation of
SOa in the presence of ammonia. In addition to an average S02 concentration
of 200 ppm, the stack gases contained about 20% water vapor and 170 ppm of
NHs(3). They also were known to contain a low concentration of mineral par-
ticles from the cement-making process, and very probably also contain trace
quantities of heavy metals from the fuel oil. NHs is known to catalyze the
aqueous oxidation of S02 to HaSOi,; more precisely, it acts as a buffer in the
solution to limit the hydrogen ion concentration and thus keep the concentra-
tion of sulfite ions (reportedly the most readily oxidizable species) high(8).
Some heavy metal ions (Mn, V, etc.) which are commonly found in fuel oils,
are also known to catalyze the sulfite-to-sulf5te oxidation reaction(9).
The laboratory investigation of the persistent plume particles by elec-
tron microscopy also was consistent with the above hypothesis, although not
conclusive. Morphologically, the particles appeared to be droplets; they
behaved in the electron beam in much the same way as HaSOi*; and they reacted
with NHs vapor to give a crystalline deposit whose behavior in the electron
beam was similar to that of ammonium sulfate. (Incidentally, if this hypo-
thesis is even only partially correct, the particle size distribution in the
plume could be quite different from the "size" distribution" on the fiIters.
At the relative humidities present in the plume, HaSOi, droplets could contain
a considerable quantity of water, which would evaporate when the particles
were placed in a vacuum environment for electron microscopy. The persistent
plume itself was bluish-white in color, which is consistent with scattering
from a median particle size about twice that estimated from the SEM micro-
graphs; humidification of an HaSOi,. aerosol distribution reported above could
easily enlarge the median particle size to this extent.)
The time-lapse movies suggest a residence time of 20 to 30 seconds for
gases and particles in the condensate plume, which seems very short. On the
other hand, the high concentrations of both SOa and, especially NHa, plus the
almost certain presence of heavy metal ions from the fuel oil, and the fact
that the above mass balance calculation showed that only a very few percent
261
-------
of the available S02 needed to react to produce the particles observed, make
the situation at least plausible. This is also supported by calculations made
by the EPA/ESRL sponsored research team who performed the in-stack gas
measurements at the time of the plume sampling. The calculations indicated
that, in the presence of an aqueous environment containing the measured level
of NH3, all of the S02 present in the stack gases could have been oxidized
to sulfate in ~15 seconds(3).
The particles in the clear plume are at present somewhat of an enigma.
They were high in sulfur, appeared to be solid, and were much more stable to
high temperatures and electron bombardment than the persistent plume aerosol.
Yet they too reacted with NH3 to give crystals similar to those formed from
the persistent plume aerosol. Moreover, particles with these properties
were not found in the persistent plume aerosol, which suggests that they did
not survive the water condensation/evaporation cycle which gave rise to the
persistent plume. A more detailed study of these "strange" particles could
possibly shed some light on the overall chemistry of the plume from this
cement plant.
CONCLUSIONS
Tethered balloon sampling of the persistent plume from the Glens Falls
Portland cement plant revealed the plume to be dominated by submicrometer,
sulfur-rich liquid particles. These particles reacted with NH3 gas to form
stable crystals which were similar to (NHOaSOi^, NHijHSOi,, and other ammonium
salts of sulfur oxide. It is proposed that these liquid particles consisted
of HaSOij most likely containing some dissolved (NHiJaSOi,. and/or NHitHSOit,
created by the aqueous oxidation of S02 in the presence of NH3.
Sampling of the clear plume from this plant revealed the presence of
submicrometer, sulfur-rich spherical particles which were more stable to
electron bombardment than the persistent plume particles. The clear plume
particles also reacted with NH3 to form stable crystals which were similar to
ammonium salts of sulfur oxide.
More refined collection and analysis procedures based upon the findings
of this investigation should lead to a more complete definition of the
nature and formation processes of the persistent and clear plume particles
from this Portland cement plant.
REFERENCES
1. Kawahata, M.. Study on Rotary Kiln Exhaust Gas Emissions in Relation
to Blue Haze Formation. Environment/One Corp. Report to Glens Falls
Portland Cement Company, November 1974.
2. Aggarwal, V., B. Pine, D. Daoust, H. Potter, B. Gould, and T. Clark.
New York State Stack Test Report, Glens Falls Cement Company, Glens
Falls, N.Y. New York State Department of Environmental Conservation
Report, September 1975.
262
-------
3. Del linger, B., G. Grotecloss, C.R. Fortune, J. L. Cheney, and J. B.
Homolya. Sulfur Dioxide Oxidation and Plume Formation at Cement Kilns.
Environ. Sci. Technol. Ik: 1244-12^9, October 1980.
4. Armstrong, J.A., P. A. Russell, I.E. Sparks, and D.C. Drehmel. Tethered
Balloon Sampling Systems for Monitoring Air Pollution. Submitted to
JAPCA for publication, August 1980.
5. Lodge, J.P. Jr. Identification of Aerosols. In: Advance in Geophysics,
Vol. IX, Landsberg, H.E. and J. Van Migham (eds.). New York, Academic
Press, 1962. p. 97-139.
6. Gras, J.L. and G.P. Ayers. On Sizing Impacted Sulfuric Acid Aerosol
Particles. J. Appl. Meteor. 18: 63A-638, May 1979.
7. Armstrong, J.A., P. A. Russell, M.N. Plooster, and I.E. Sparks. Balloon
Borne Particulate Sampling of a Portland Cement Plant Plume. In pre-
paration.
8. Hegg, D.A.. Atmospheric Sciences Department, University of Washington,
Seattle, Washington. Personal communication, June 1980.
9. Hegg, D.A. and P.V. Hobbs. Oxidation of SOz in Aqueous Systems with
Particular Reference to the Atmosphere. Atmos. Environ. 12: 241-253,
February 1978.
263
-------
THE RELATIONSHIP OF FLY ASH LIGHT ABSORPTION TO SMOKE PLUME OPACITY
By: Stanton J. Cowen
Atmospheric Research Group
Altadena, California 91001
David S. Ensor
Research Triangle Institute
Research Triangle Park, North Carolina 27709
ABSTRACT
Measuring the fly ash light absorption for coal-fired boilers with the
Integrating Plate Method (IPM) is discussed. Measurement of the optical pro-
perties of fly ash may be useful for stack opacity modelling and identifying
the relative contribution of particle light absorption to downwind visibility
impairment. The IPM technique is defined as comparing the light absorption
through a clean filter to one with a single layer of aerosol by integrating
the scattered light so only absorption is measured. Since the light absorp-
tion is a strong function of particle size, careful sizing is required for
accurate measurement. Preliminary calibration and fly ash data are reported.
INTRODUCTION AND SCOPE
Smoke stack opacity regulations have become an important regulatory
issue, and compliance can be difficult for stationary sources. The ability
to predict opacity depends on the quality of particle size and refractive
index data of the emission aerosol. A careful examination of the complex
index of refraction is necessary to improve the accuracy of opacity predic-
tion.
This paper discusses a preliminary measurement of the optical properties
of fly ash from pulverized-coal-fired boilers as part of work to determine
the relative importance of light absorption compared to light scattering.
Such work is needed, since the optical absorption of fly ash has been mea-
sured by Volz (1973) and Nolan (1977) under very limited conditions. The
large variability in the composition of fly ash, especially carbon content,
prompted further investigation of optical properties of this aerosol.
264
-------
This paper is, an initial, not comprehensive, examination of light ab-
sorption by fly ash from different types of coals and combustion conditions.
This information will help opacity modelers to more accurately predict stack
opacity.
EXPERIMENTAL PROCEDURE
This study includes the experimental measurement of fly ash absorption
and the theoretical modeling of smoke plume opacity. The experimental tech-
nique is an application of the Integrating Plate Method (IPM), as developed
by Lin et al. (1973), to fly ash. Preliminary measurements were undertaken
to determine the importance of fly ash light absorption. Absorption measure-
ments and analysis of the carbon content of the ash (assumed to be "graph-
itic") were performed to determine the correlation between absorption and
carbon content. This "graphitic" assumption is valid, because little "or-
ganic" carbon is thought to be present in the fly ash. The measurement
strategy has two assumptions regarding the nature of fly ash and its optical
properties:
First, samples were collected from electrostatic precipitator hoppers
downstream of a coal-fired boiler and stored in hard plastic containers for
future analyses. It is assumed that this sample is representative of the ash
emitted from the stack and that storage of the sample does not change its
optical properties.
Second, the absorption technique was used to examine submicron fly ash
deposited on a filter. The assumptions are that the submicron ash has optical
characteristics similar to other size particles and that fly ash light absorp-
tion on a filter can be related to its behavior while suspended in a gaseous
medium. Work is needed to assess the size dependency of carbon in fly ash
to certify that absorption properties of submicron ash are representative for
opacity calculations.
Lin et al. (1973) developed the Integrating Plate Method to measure the
absorption coefficient of ambient aerosol. Weis et al. (1978) further tested
and calibrated this technique with monodisperse aerosols of methylene blue.
A main goal of this study is to examine the application of the IPM procedure
to source aerosols such as fly ash.
IPM is a simple direct measurement of aerosol light absorption based on
comparing the transparency of a clean nuclepore filter with one containing
a thin layer of particles. A Hitachi Model 60 double-beam spectrophotometer
with a specially designed filter holder was used for the comparison. Opal
glass behind each filter integrates the transmitted light as a function of
angle so that only absorption is recorded. Back scattering probably can be
neglected, because the reflectivity of the filter is unchanged by the fly ash
particle; i.e., the real refractive indices of particles and filters are simi-
lar. The double-beam automatically gives the ratio of the aerosol absorption
to the blank filter. The optical arrangement for the beam measuring the
sample is shown in Figure 1.
265
-------
MIRROR
MIRROR
LIGHT SOURCE
TUNGSTEN LAMP
FIL'
APER
r
^>
FER
rURE
NUCLEPORE
1
1
FILTER WITH SAMPLE
t
PHOTOMETER
DETECTOR
OPAL GLASS
FILTER
Figure 1. Schematic diagram of the optical system incorporating
the integrating plate method.
The real part of the index of refraction is measured with oil immersion
techniques. The ash sample is spread over a microscope slide so that individ-
ual particles can be identified. The particle is then immersed in an oil of
known refractive index. The Becke line test shows the particle index of re-
fraction to be lower or higher than the immersion oil.
Converting the specific absorption coefficient to an imaginary refractive
index with an aerosol of spherical ammonium sulfate particles containing meth-
ylene blue dye is discussed at the end of Section 3, Data Analysis.
Bulk ash samples obtained from electrostatic precipitation required re-
suspending before measurement. A common sandblast gun was used as a simple
technique for deagglomerating fly ash by producing a high pressure aerosol
stream.
A single layer of submicron aerosol is required to easily evaluate the
light absorption effects. The absorption for submicron aerosol is simply pro-
portional to the product of the imaginary refractive index and volume of par-
ticles as described by Weis et al. (1978). The resuspended fly ash was first
drawn through an impactor to remove large particles and then through a Model
3071 Thermo System Incorporated Electrostatic Classifier to obtain a near
monodisperse aerosol with a particle diameter of approximately 0.8 jum.
Transmission Electron Microscope photographs (Figure 2) show the coagulation
of very small ash speres (particle diameter ~ 0.07 jLtm) on the surface of
larger 0.8 m particles. The quantity of material collected on each filter
was approximately 50 to 100 micrograms as measured by Cahn 25 balance. A dry
gas meter downstream of the filter measured gas volume passing through the
266
-------
Figure 2. Transmission electron microscopy photograph of
submicron fly ash particle at a magnification
of 44.350X. The length of the line is approxi-
mately 1 micron.
filter. The gas volume was used with the filter area to compute an effective
optical path length, which was used to calculate the absorption coefficient.
West Paine Laboratories performed a carbon analysis of the ash samples by
standard methods. Approximately 1.0 gram was accurately weighed and placed in
a tube furnace. The furnace was heated to 800°C until the samples were com-
pletely ashed. The carbon dioxide given off was trapped in a CC>2 free NaOH
solution. The carbon dioxide was then measured by a Beckman Total Organic
carbon analyzer.
DATA ANALYSIS
Data reduction for the Integrating Plate Method uses Mie theory to de-
termine absorption effects. It is assumed that the filter does not interfere
with the light absorption of the particles. Nuclepore membrane filters were
used, because the particles are on the surface of the filter and not embedded
as they would be in a fiberglass filter.
267
-------
The transmission of light through a volume containing an aerosol is de-
scribed by the Beer-Lambert Law:
Light transmittance = I/IO = exp (-bextL)
where I is the transmitted light, I0 is the incident light, bext is the light
extinction coefficient, and L is the illumination path length. The light ex-
tinction coefficient is the sum of the scattering coefficient (bscat) and the
absorption coefficient (babs)«
The ratio of intensity (I1) of light transmitted through the opal glass
and a filter deposited with aerosol to the intensity (Io') transmitted
through an opal glass and blank filter can be used to calculate the absorption
coefficient:
I'/V - exp (-babsL)
Ensor and Pilat (1971) have shown that the aerosol mass concentration is re-
lated to optical transmittance through a modified form of the Beer-Lambert
Law:
I _ exp
M L Sve
where M is the particle mass concentration, Sve is the ratio of the light ex-
tinction coefficient to the specific particulate volume (m^/cnH), (reciprocal
of "K" in the original paper), is the average particle specific gravity,
and L is the optical path length. Optical absorption can similarly be formu-
lated by
M L Sva
exp
where Sva is the ratio of the light absorption coefficient to the specific
particulate volume. The optical path length is determined by the length of
the volume of sampled air of cross-sectional area equal to the filter.
The specific absorption coefficient (Sva) is converted to an imaginary
refractive index (n") by calibrating the instrument with a laboratory aerosol
of similar particle shape and size as fly ash with known absorption character-
teristics. This was accomplished with an aerosol of spherical ammonium sul-
fate particles containing various amounts of methylene blue dye to obtain var-
ious degrees of light absorption. The calibration Sva is then proportional to
the product of imaginary refractive index (n") of methylene blue and its
weight percent in the ammonium sulfate aerosol, a nonabsorbing material. The
imaginary refractive index of the ash sample is calculated in the following
manner: The specific absorption coefficient (Sva) of the ash sample is
plotted on the methylene blue ammonium sulfate calibration curve (Figure 3).
The fraction of methylene blue corresponding to this plotted point multiplied
268
-------
0.0 1.0 2.0 3.0 4.0 5.0
% METHYLENE BLUE IN
AMMONIUM SULFATE AEROSOL
6.0
Figure 3. The absorption per unit volume (Sva) as a function
of methylene blue content.
by the imaginary refractive index of methylene blue (in this case, 0.1) equals
the imaginary refractive index of the ash sample. There is very little depen-
dence on the real part of the refractive index because the Integrating Plate
Method diminishes scattering effects by comparing a blank filter to one with
a layer of aerosol where the real refractive indices of aei^bsol and filter are
similar. The computed imaginary refractive index may then be used to calcu-
late its effect on stack opacity with the opacity model described earlier.
RESULTS
The absorption measurement of the calibration methylene blue/ammonium
sulfate aerosol are presented in Figure 3. Samples weighing from 100 to 500
micrograms were generated at the 1st International Workshop on Light Absorp-
tion by Aerosols at Colorado State University Cloud Simulation and Aerosol
Laboratory. The correlation coefficient of the line with the least squares
fit to the absorption data is 0.96. (One is a perfect correlation.) The
linear relationship is predicted by Mie theory for this submicron absorption
269
-------
aerosol. The particle size was determined by scanning electron microscopy
and an Electrical Aerosol Size Analyzer.
This experimental technique has been calibrated with monodisperse pure
methylene blue 0.8 to 2 pirn particles with absolute error of ten percent as
reported by Weis et al. (1978). Our experiments with pure methylene blue
particles with a mass mean diameter of 1.5 jam and 200 microgram filter load-
ings produced absorption readings approximately twenty-five percent higher
than predicted by Mie theory. The high absorption readings may indicate in-
terference from multiple scattering effects for high mass loadings (greater
than 100 micrograms on a 47 mm filter). The difficulty of separating scatter-
ing from absorption effects, especially for low absorption aerosols, precludes
a simple determination of fly ash absorption.
The interference effects are indicated in Figure 3 by the apparent ab-
sorption for the nonabsorbing pure ammonium sulfate samples. The high corre-
lation for the absorbing aerosols indicates a systematic offset related to
light scattering of the particles on the filter.
The coal, combustion conditions, and refractive indices for the fly ash
samples are listed in Table 1. These characteristics are probably not as im-
portant in determining light absorption as the quantity of soot emissions.
These carbonaceous emissions are more important for determining fly ash ab-
sorption than the mineral content of the ash.
Results of five fly ash samples analyzed for absorption and carbon con-
tent are shown in Figure 4. The absorption measurements were performed on
submicron ash sample with a mass mean diameter of approximately 0.8 jum as
indicated by Transmission Electron Microscopy. Although the trend indicates
increasing absorption with increasing carbon (graphitic) content, as expected,
the limited number of samples precludes statistical analysis. The scatter may
be the result of small differences in the particle size distributions and the
nonuniform distribution of carbon in the submicron ash compared to the hopper
ash.
The absorption measurements of the ash samples produce results similar to
the absorption calibration measurements of methylene blue ammonium sulfate
aerosol. The similarity in particle size and the similar low absorption of
these aerosols may explain their similar behavior. Again interference effects
like those in the calibration curve are noted by the back extrapolation of the
curve in Figure 4 to 0 percent carbon. Also the linear results are present in
both calibration and fly ash curves as predicted by Mie theory.
The effect of imaginary refractive index on opacity was estimated by com-
puting the ratio of opacity from pure soot to the opacity from transparent
material (Table 2). The calculated ratio of the specific aerosol volume to
the extinction coefficient reported by Pilat and Ensor (1971) was used to com-
pute the effect of light absorbing material on opacity less than 20 percent.
The opacity equation can be approximated for opacity less than 20 percent by
taking Taylor's expansion of an exponential as follows:
270
-------
Table 1
DESCRIPTION OF COAL AND COMBUSTION
CONDITIONS FOR FLY ASH SAMPLES
N)
Power Plant.
Power Plant A
Power Plant. B
Power Plant. C
Power Plant D
COAL PROPERTIES
Heating Type
Sub-bituminous,
pul ver ized,
western
Sub-bituminous,
pulverized,
western
Bituminous,
pulverized,
eastern
Bituminous,
pulverized,
western
Heating Valve
(BTU/lb)
10,000 to
11,000
10,500
12,000
12,000
Ash Content.
(X)
3 to 7
19
20
10
COMBUSTION CONDITIONS
Boiler Type
Front-fired
Reheat, natural -
circulation
steam generator
Controlled circu-
lation steam
generator
Unit Rating
(MM)
40
314
540
320
Refractive
Index*
1.565 -
0.0055i
1.57 -
0.0012i
to 0.0025
1.5 -
0.0043i
1.61 -
0.00041
*Real part of the refractive index determined by oil immersion techniques. Imaginary part of the refractive
index determined by Integrating Plate Method. Methylene blue has an imaginary refractive index equal to
0.1 at 0.56 urn wavelength reported by J. Gillespie (1980).
-------
3.0 -
2.5
2.0
1.5
1.0
0.5
e
PLANT C
,' PLANT B
1 i i I i I I L L
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
% CARBON
1.8 2.0
Figure 4. Absorption per unit volume (Sva) as a function
of carbon content.
Opacity =
Opacity
Opacity fts
Sve M L
1 - exp
1 - 1 -
Sve M L
x 100
13 ve M L
x 100 for opacity <20%
x 100
Opacity (soot) Sve (soot)
Opacity (transparent) ~
Sve (transparent)
272
-------
Table 2.
THE EFFECT OF ABSORBING MATERIAL
ON OPACITY AS A FUNCTION OF PARTICLE SIZE
\xjgw
g ^^
1
2
3
4
5
0.1
6.5
3.38
3.09
3.11
3.33
0.5
1.08
0.95
1.29
1.56
1.75
1.0
0.895
0.929
1.0
1.32
1.47
5
0.977
0.996
0.974
1.02
1.12
10
1.0
1.0
0.985
0.99
1.0
If aerosols are compared with similar particle mass concentration (M),
path length (L), and particle specific gravity (P), the ratio of their opaci-
ties is then simply the ratio of their specific extinction coefficients (Sve).
Table 2 is a calculation ratio of Sve for transparent aerosol with refractive
index equal to 1.5. These results show the absorbing refractive index is most
significant for the monodisperse aerosol with a mass mean radius equal to 0.1.
SUMMARY AND CONCLUSIONS
Applying the Integrating Plate Method to a source aerosol such as fly ash
requires careful treatment, especially in the area of particle sizing. Scan-
ning or transmission electron microscopy may be required for certain measure-
ment of the actual deposited particle size. Coagulation, especially for
larger particle mass concentrations, can change the effective particle size on
the filter. The application of this experimental technique to low absorbers
can create problems with filter interferences and scattering differences be-
tween the filter substrate and the particulates. The measured absorption may
be an overestimation of actual absorption because of these interference prob-
lems. More work is required to quantify these interference effects and to de-
velop a sound theoretical basis for this quantification. Fortunately, these
interferences seem to be consistent for a given particle size and absorptive
range, thus allowing an empirical correction.
The usefulness of this technique for fly ash absorption requires that the
optical properties of submicron ash represent the bulk sample emitted from the
stack, especially for modeling purposes. Additional research is required to
compare the carbon content and other absorbing species of the bulk ash to the
submicron ash concentrations. Careful work may require direct source sampling
rather than resuspension of bulk samples.
273
-------
This preliminary work indicates that imaginary refractive index is re-
lated to carbon content in a manner similar to that reported for atmospheric
aerosol by Weis et al. (1978) and Salder (1979). However, the number of sam-
ples needs to be expanded to include the range of carbon content expected in
fly ash (up to 20 percent or more). Also in cases of good combustion with
low carbon carryover from the boiler, matrix elements such as calcium, titan-
ium, and iron are believed to affect imaginary refractive index.
ACKNOWLEDGMENTS
We wish to thank G. Woffinden, G. Markowski and E. Hindman for technical
advice. This research was supported by the Environmental Protection Agency
under grant number R806718010 under the direction of L. E. Sparks.
ENDNOTES
Berstrom, R. W. (1972): Predictions of the spectral absorption and extinction
coefficients of an urban air pollution aerosol model. Atmos. Environ.,
6^ 247-258.
Lin, C., M. Baker and R. J. Charlson (1973): Absorption Coefficient of atmo-
spheric aerosol: A method for measurement. Appl. Optics, 12, 1356-1363.
Ensor, D. S., and M. J. Pilat (1971): Calculation of smoke plume opacity from
particulate air pollutant properties. J. Air Poll. Contr. Assn., 21, 496.
Gillespie, J., Personal Communication (1980).
Nolan, J. L. (1977): Measurement of light absorbing aerosols from combustion
sources. M. S. Thesis, University of Washington.
Pilat, M. J., and D. S. Ensor (1971): Comparison between the light extinction
aerosol mass concentration relationship of atmospheric and air pollutant
emission aerosols. Atmos. Environ., 5, 209-215.
Sadler, M. S. (1979): A comparative study of light attenuation, filter media,
and total carbon content. M. S. Thesis, University of Washington.
Twitty, J. T., and J. A. Weinmen (1971): Radiative properties of carbonaceous
aerosols. J. Appl. Meteor., 10, 725-731.
Volz, F. E. (1973): Infrared optical constants of ammonium sulfate, Sahara
dust, volcanic pumice, and fly ash. Appl. Optics, ^L2, 564-568.
Weis, R. E., A. P. Waggoner, R. J. Charlson, D. L. Thorsell, J. S. Hall and
L. A. Riley (1978): Studies of the optical, physical, and chemical
properties of light absorbing aerosols. Presented at Conference on
Carbonaceous Particles in the Atmosphere, Berkley, CA. 20-22 March.
274
-------
A SPECIFIC METHOD FOR THE ANALYSIS OF SULFURIC ACID MISTS
By: P. Urone, R.B. Mitchell, J.E. Rusnak, R.A. Lucas, J.F. Griffiths
Environmental Engineering Sciences Department
University of Florida, Gainesville, FL 32611
ABSTRACT
A number of organic dye precursors were investigated for their use
as specific reagents for the detection and measurement of sulfuric acid
mists in air. The underlying principle of the method was to have the
sulfuric acid react with the precursor to form a sulfonated product with
a specific color that could not be caused by other acids or pollutants
in air. At least five organic dyes were found to react with sulfuric
acid to form a uniquely colored sulfonation product. In all cases,
reversal of the reaction was a serious problem. It was found that
heating of the filter holder while sampling with a precursor coated
glass fiber filter fixed the sulfonated product. Fifty yg H-SO, aerosol
are readily observed by a color change on the filter. Basic theory,
methodology, sensitivity, and reproducibility are discussed.
INTRODUCTION
The sampling and analysis of sulfur trioxide and sulfuric acid in
emission sources continues to be a source of difficulty. At present a
number of methods are used to determine sulfuric acid mist in either
emission sources or ambient air. These methods can generally be divided
into two basic operations: 1) the sampling or collection of the aerosol;
and 2) the analysis or measurement of the collected aerosol (1).
The major methods of coll°ction of sulfuric acid aerosols are:
absorption into isopropyl alcohol or alkaline solutions, controlled and
uncontrolled condensation, impaction, and filtration.
The major methods of measurement and their chemical basis for the
analysis of H-SO, at present are:
sodium hydroxide titration (1); barium ion titration (2);
chloranilate (1); turbidimetric (1); flame photoluminescent
detector (1,5); ammonium vanadate (3); bromphenol blue (4);
coulometry (5) and conductimetry (1,7).
The major difficulties encountered by the presently available
methods are due generally to the sampling methods with their inherent
interferences from copollutants and sampling efficiency. Recently
Jaworowski and Mack evaluated the performance of EPA's Method 8, the
controlled Condensation Method, and the Dew Point Measurement Method
(7).
275
-------
SULFONATION
A specific analytical technique for the analysis of sulfuric acid
is possible due to the sulfonating ability of sulfuric acid with aromatic
and heterocyclic compounds. Sulfonation by definition is the process of
replacement of a hydrogen atom in an organic compound with the sulfonic
group (SO,,H ). The replaced hydrogen atom and the hydroxyl ion (OH )
from sulfuric acid form a water molecule (8, Fig. 1).
There is no clear cut guideline to follow when considering reaction
temperatures, concentrations, percent product recovery and reaction
rates when the sulfonation of species not generally dealt with in industrial
applications are studied. However, sulfonation of rings which are
substituted with electron withdrawing groups (-C1 , -Br , -SO^H, -
CCLH, -N02) proceeds with difficulty since the ring_is already electron
poor. Electron donating groups (-NH2, _OH, -OR, -0 ) attached to the
ring aid sulfonation and are ortho and para directing.
Kinetic and mechanistic studies of the sulfonation of aromatic
compounds have led to the conclusion that it is an S reaction with
monomeric S0_ as the effective reacting species, not only with SO.,
itself, but also when sulfuric acid or oleum are used. It has been
proposed that in sulfuric acid an acid solvate of SO. is the active
species. This is considered possible, but less likely than free S0«
(8). J
Aromatic compounds can be sulfonated with concentrated H_SO,, but
as the concentration of the water increases during reaction, the rate of
sulfonation steadily decreases, the reaction rate being inversely propor-
tional to the square of the water concentration. The reaction ceases
when the acid concentration reaches a level characteristic of each
compound. Sulfonation differs from most other types of electrophillic
substitution in that the process is readily reversible under mild conditions
(9). Desulfonation of some aromatic compounds proceeds rapidly and in
good yield by simple dilution in aqueous medium. Therefore, to carry
sulfonation reaction to completion, the removal of water formed from the
reaction is essential. However, the hydrolysis of sulfonated aromatics
will not occur if the acid strength is kept high.
EXPERIMENTAL
Five indicator dye precursors were studied because their reactions
with sulfuric acid produced a sulfonated product having a distinctive
and intense color. The compounds were: acid orange A, phenolphthalein,
thymolphthalein, methylene blue, and o-cresolphthalein. The reactions
of all gave satisfactory color changes with intensities useful for both
source and ambient air concentrations. Problems with reversibility of
the sulfonated products were encountered in all cases making it desirable
to develop a technique for preventing the reversal of the reaction.
276
-------
0-cresolphthalein was selected as one of the more promising of the
precursors. It was easily available in relatively pure form as a dry
slightly yellow powder. It reacted with sulfuric acid to give an intense
red-orange color. It did not react with concentrated solutions of
hydrochloric or phosphoric acids. Tests were designed: 1) to study the
optimal method for dissolving and coating the o-cresolphthalein on a
glass fiber filter; 2) to use the coated filter to sample a highly
dilute sulfuric acid aerosol under conditions that would promote and fix
the reaction of the sulfuric acid aerosol particles with the o-cresolpht-
halein; and 3) to selectively extract the sulfonated product from the
filter for spectrophotometric study.
Materials and Apparatus
The acids used were HCl, HN03> H^SO^, and ^PO,. They were all
concentrated and laboratory grade, or better. A 0.045 M H?SO, solution
was used in generating the acid aerosol to be sampled.
The solvents used were acetonitrile, dimethyl sulfoxide, and methanol.
They were selected from solubility tests with ortho-cresolphthalein
powder. They were all laboratory grade, or better.
The standard solution for the calibration curve was prepared by
mixing 0.03531 gm. of ortho-cresolphthalein powder with 5 ml. of 18 M
H_SO, in a 100 ml volumetric flask. It was gently heated until all the
OCP had gone into the solution. The flask was immersed in an ice water
bath and 95 ml. of 12 M HCl was added, at first drop by drop until the
violent reactions ceased. Then the dilution proceeded normally.
The aerosol generating and sampling train system is shown in Figure
2. The Dwyer type rotameters and the critical orifice, Type # 22-F,
were calibrated using a precision wet test meter. The aerosol was
generated by a DeVilbis nebulizer Model 40. The manufacturer has rated
the aerosols generated to have a mean diameter of 4.5 microns with a
geometric standard deviation of 1.8 at 5 psi. Aerosol flowrates and
concentrations were measured using distilled water and various molarities
of tLSO.. Heat checks were made before and after each run.
2 4
Purified compressed air was used for aerosol generation and a
Thomas vacuum pump, Model/Serial #907CA18-2, for sampling. The filter
holder was an in-line stainless steel Gelman Type 303. The plastic
inlet and outlet nipples were replaced with metal nipples to withstand
temperatures not to exceed 110°C. The inlet dispersing screen was
removed to avoid any loss of aerosol.
The type of filter used to collect the sulfuric acid aerosol was
the Gelman Type A/E glass fiber filter. These filters are devoid of any
organic binders and have a minimum retention efficiency of 99.7% for 0.3
Urn particles as measured by the dioctyl phthalate penetration test.
Barton and McAdie developed a method of treatment of the glass fiber
filters by soaking in 20% sulfuric acid followed by gentle boiling for
277
-------
10 minutes, rinsing with distilled water, 80% isopropanol, and finally
acetone (10-13). The filters are then dried for 15 minutes at 105 C and
stored in a desiccator (14). This method of filter treatment was used
in this study.
The heating unit was made from a copper pipe coupler covered with
alternate layers of asbestos and nichrome wire. The voltage to the
heater was controlled by a Powerstat. Two wooden caps were cut to fit
the inlet and outlet ends and to hold a thermometer. A four-liter Pyrex
bottle was used to receive and mix the aerosol. All excess gases were
exhausted to a laboratory fume hood.
Each filter was weighed and placed for approximately 2 minutes in
various solutions of ortho-cresolphtalein powder (OCP) in acetonitrile
(OCP-A), dimethyl-sulfoxide (OCP-D) or methanol (OCP-M). They were
drained of any excess solution by touching the edge of the beaker,
placed on Petri dishes or watch glasses, and allowed to air dry. Then
the filters were either placed directly in a desiccator or after oven
drying at 105 C for approximately 15 minutes.
Sulfuric Acid Aerosol Sampling
A pre-weighed and coated filter was placed in the filter holder,
the filter holder encased in the heater unit, and the Powerstat switched
on and set to a predetermined positon that would yield the desired
sampling temperature of 105 C.
Approximately 5 ml. of 0.045 M H?SO, was placed in the nebulizer.
Just before achieving the desired sampling temperature the aerosol
generating system was turned on with the compressed air valve. The
desired flowrate of 2.89 1pm, used in all of the experiments, was
regulated with the calibrated rotameter. This compressed air flowrate
produced an average sulfuric acid aerosol flowrate of 0.178 ml/min. ,
giving a concentration of 0.272 mg H SO. per liter of undiluted aerosol.
Dilution of the aerosol with one cubic root air per minute produced a
concentration of 26.0 yg HLSO./l air. The aerosol was sprayed into the
mixing chamber, diluted when desired, and vented to the hood. When
assured that the sulfuric acid aerosol was evenly and well dispersed,
the sampling pump was turned on. The sample flowrate was controlled by
the critical orifice at 1.913 1pm. Sampling periods were timed such
that 48-520 yg of sulfuric acid was collected on the sampling surface.
Immediately following the sampling run, room air was pulled through the
filter for 1 minute. Then the filter was kept in the heater unit for an
additional 10 minutes at 105 C.
The filters showed a light to a strong red-orange color of o-
cresolsulfonphthalein, tests were performed on them. The tests were for
(1) color stability by exposing a filter to either room air, oven drying
at 105 C, dessication, or combinations of each; (2) solvent extraction
by various organic and inorganic solvents, and (3) spectrophotometry of
the solutions.
278
-------
The same sampling procedure was used with a 12 M HC1 solution with
no dilution air added. One-minute samples were taken at temperatures in
the filter holder of 50 C and 100°C. The 50°C samples showed some
coloration but the 100 C samples showed none. Solutions of OCP in
concentrated HC1 and H PO^ gave no color. Concentrated nitric acid gave
a yellow color.
RESULTS
It was found that the prescribed method gave a stable and very
sensitive red-orange color of o-cresolsulfonphthalein on the filter.
The color was specific for sulfuric acid, and was easily discernable in
the range of 50 to 500 micrograms of sulfuric acid. This would cover
the range of one to 10 parts per million for a 10 liter (5 minute) stack
sample. Higher or lower concentrations can be measured with shorter or
longer sampling times. Field tests could be quickly and easily conducted
by comparison of the color obtained with those on a series of standards.
The color is not, at present, easily separated from the filter and
the organic precursor. Extraction of the filters with concentrated
hydrochloric acid can give semi-quantitative results when the extracted
color is compared to that of a blank run. Extraction with water causes
reversal, but titration of the ensuing sulfate is easily possible.
Recent work shows that phosphoric acid dissolves the o-cresolphthalein
without dissolving the sulfonated product. Such a separation would be
most useful for exact laboratory measurements.
The method has not as yet been field tested or tested for the
possible interference of volatile bisulfate salts. The latter should
easily be removed at higher temperatures by filtration if their interference
is significant.
ACKNOWLEDGMENTS
This work was supported by grants from the Environmental Protection
Agency and the Smelter Environmental Research Association.
279
-------
REFERENCES
1. Urone, P., "Source and Ambient Air Analysis of S(L and H2SO.:
State of the Art Report," Health Lab. Sci., 11;246 (July 1974).
2. Barton, S.C., and McAdie, H.G., "An Automated Instrument for
the Specific Determination of Ambient H-SO, Aerosol," Presented
before the Divison of Water, Air and Waste Chemsitry, ACS, New
York (1972).
3. Baviha, C.J., and Shinkararenka, L.S., "Determination of Sulfuric
Acid in Air by a Vanadate Method," Neftepererab Neftehhim, 9;40-41,
(1971).
4. Barrett, W.J., Miller, H.C., Smith, J.E., and Guin, C.H. ,
"Development of a Portable Device to Collect Sulfuric Acid
Aerosol," Interim Report from EPA (February 1977).
5. Scaringelli, F.P., and Rehme, K.A., "Determination of Atmospheric
Concentrations of Sulfuric Acid Aerosol by Spectrophotometry,
Coulometry and Flame Photometry," Anal. Chem., 41:707-713 (June
1969).
6. Thomas, R.L., Dharmarajan, V., Lundquist, G.L., and West, P.W.,
"Measurement of Sulfuric Acid Aerosol, Sulfur Trioxide, and
Total Sulfate Content of Ambient Air," Anal. Chem., 48:639-642
(April 1976).
7. Jaworowski, R.J., Mack, S.S., "Evaluation of Methods of Measurement
of SO /H SO, in Flue Gas," J. Air Pollu. Control Assoc., 29:43-46
(Jan. 1979).
8. Gilbert, E.E., Sulfonation and Related Reactions, Interscience
Publishers, New York (1965).
9- Butler, G.B., and Berlin, K.D., Fundamentals of Organic Chemistry
Theory and Application, Ronald Press Company, New York (1972).
10. Gelman Instrument Company Catalog (1975).
11. Scaringelli, F.P., and Rehme, K.A., Anal. Chem., 41:707 (June
1969).
12. Byers, R.L., and Davis, J.W., "Sulfur Dioxide Adsorption and
Desorption on Various Filter Media," Air Poll. Control Assoc.,
20:236 (April 1970). ~~
13. Pate, J.B., Ammons, B.E., Swanson, G.A., and Lodge, J.P., Jr.,
"Nitrite Interference in Spectrophotometric Determination of
Atmospheric Sulfur Dioxide," Anal. Chem., 37:942 (1965).
14. Barton, S.C., and McAdie, H.G., "Preparation of Glass Fiber
Filters for Sulfuric Acid Aerosol Collection," Env. Sci. Tech.,
4:769-770 (September 1970). 280 ~
-------
SULFONATION REACTION
I. 2H2 S04 *=* H30* + HS04 + S03(Equilibrium)
2. -H S0 —* (Slow)
3. ($) +HS04 —* (o) +H2S04(Fast)
S03 S03H
4. (o) + H30+ *==* (o) + H20(Equilibrium)
DESULFONATION REACTION
S03H
Figure 1. The Sulfonation Reaction
281
-------
To Hood
Filter Holder
and Heater /
/
Compressed
Air
Atomizer
4-Liter Pyrex
Bottle
Powerstat
Figure 2. Experimental Apparatus
-------
2.0
1.5
o
c
o
- 1.0-
0.5-
0.2-
o-Cresolphthalein
S03H
A=6.9l/xg/ml
•=l3.8/xg/m!
400 440 480 520 560 GOO
Wavelength, nm
Figure 3. o-Cresolsulfonphtalein Adsorption Curves
283
-------
14-
o-Cresolphthalein- S03H
0 0.2 0.4 0.6 0.8 1.0 12 1.4
Absorbance,5IOnm
Figure 40
Standard Curve for the H2S04 - o-Cresolphthalein Reaction Product
284
-------
A MICROCOMPUTER-BASED CASCADE-IMPACTOR DATA-REDUCTION SYSTEM
By: Michael Durham
Scott Tegtmeyer
Ken Wasmundt
Denver Research Institute
University of Denver
P.O. Box 10127
Denver, Colorado 80210
Leslie E. Sparks
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
ABSTRACT
This paper describes a cascade-impactor data-reduction system incorpora-
ting the Radio Shack TRS-80* computer. The system provides the computational
facilities required to reduce raw impactor data to a particle size distribu-
tion. The program first calculates the 50% cut-points for each stage and
then determines the cumulative mass distribution. A linear regression
analysis is then applied to this distribution in log-normal space to deter-
mine the log-normal distribution parameters. A spline-fit routine is used to
mathematically describe the cumulative distribution and to generate the
differential distribution. A mathematical function is used to extrapolate
the data to the maximum particle diameter in order to determine the amount of
mass below the Inhalable Particulate Matter (IP) 15 pm cut. Additional
programs provide the capability to statistically combine similar runs and
to calculate the efficiency of a control device as a function of particle
size by comparing impactor runs from the inlet and outlet. The system
provides hard copy of input and results in a combination of tables, line
printer graphics, and/or plotter graphics.
INTRODUCTION
Cascade impactors have become the standard instrument for measuring size
distributions of particulate matter entrained in gas streams. However, since
impactors measure the aerodynamic characteristics rather than particle size,
elaborate data reduction procedures are required to obtain particle size data
from an impactor run. This is due to the fact that the aerodynamic charac-
teristics of particles are dependent upon interactive functions of such
variables as: particle size and density; gas composition, density,
viscosity, temperature, pressure, and mean free path; and impactor flow rate,
number and diameter of jets, and stage-calibration constant. Even using the
simplest data-reduction procedures, these variables must be taken into
account.
*TRS-80 is a trademark of Tandy Inc. 235
-------
Another problem that exists is that, since the cut points on each stage
are a function of sampling conditions, even if the same impactor is used, it
is unlikely that the cut-diameters will be the same for two runs. This means
that, in order to statistically combine similar runs or to compare data from
different conditions, it is necessary to interpolate between data points.
Because of the extensive nature of the data reduction required for
cascade-impactor runs, a computer-based system was developed not only to
reduce the time required in making these tedious calculations, but also to
provide a measure of data-quality assurance. The system was designed for a
microcomputer to eliminate the necessity of a full-sized computer. Besides a
savings in cost, the microcomputer is physically small and portable and thus
can be transported to the source of the data and provide quick feedback to
the field-test program.
TRS-80 CASCADE-IMPACTOR DATA-REDUCTION SYSTEM
The TRS-80 cascade-impactor data-reduction system is designed to func-
tion on a Tandy Model I TRS-80 microcomputer system equipped with 48k bytes
of central memory (RAM), two or more diskette drives, and a line printer. In
addition, some of the programs have options for plotting data on either a
Houston Instruments DP-1 plotter or an HP 7225A plotter using the TRS-80
serial interface.
The system is intended to provide the computational facilities required
to reduce raw cascade-impactor data (stage weight gains, etc.) to a particle
size distribution at a consistent set of particle sizes, and then to combine
multiple impactor runs and compare the results of the inlet and outlet
samples to calculate the collection efficiency of an air-pollution control
device as a function of particle size. The system allows the user to select
whether the data are reduced according to physical diameter, classical aero-
dynamic diameter, or impaction aerodynamic diameter depending on how the data
are to be used. Each program, as appropriate for the data forms, provides
hard copy in a combination of tables, line printer graphics, and/or plotter
graphics.
The programs use a "menu" format for data entry in which the operator is
presented with a list of variables and their values. The user may review and
change any parameter up until the time that he initiates the calculations.
Figure 1 is a flow chart depicting the general organization of the system
showing the interactions between the programs (boxes) and the files produced
by the programs (circles).
GO
The "GO" program provides a central directory for the other programs in
the system. It provides a list of the other programs, and can be used to
give a brief description of each, or to initiate the execution of the other
programs. All of the other programs automatically return to this program
unless the return is disabled by the system manager.
286
-------
Figure 1. Impactor Data-Reduction System Flow Chart
DEFIMP
In order to make the system versatile enough to handle any type of
impactor, the "DEFIMP" or "Define Impactor" program was written. This
program creates a permanent file for an impactor which includes all the
information that will be required to reduce data from the particular
impactor. Table I shows the type of data required for each impactor. The
operator first provides a name for the impactor which will be used in the
IMPACTOR program to identify the correct impactor file. A brief description
of the impactor is also included to help the operator distinguish between
similar impactors. The shape of the jets (circular or rectangular) and the
number of stages are defined by the operator. Then for each stage the number
of jets, a calibration constant (square root of PSI), and the jet diameter
for circular jets or the slot length and width for rectangular jets are
input. The calibration constant, PSI, is defined as:
where
PSI =
D =
D 2
P
(D
18MD.
Particle diameter (cm),
Cunningham slip factor,
Particle density (g/cm3),
287
-------
V = Jet velocity (cm/sec),
(j = Gas viscosity (poise), and
D. = Jet diameter (cm).
•J
PSI is a dimensionless value and can be obtained empirically
(Harris, 1977) or theoretically (Marple, 1970; Ranz and Wong, 1958).
Empirical values for several of the commercial impactors are presented in
Gushing et al., (1976).
A pressure drop calibration constant (K) is input from the equation:
DP = K P Q2 (2)
where DP = Impactor pressure drop (in. Hg),
P = Density of the gas stream (g/cm3), and
Q = Actual flow rate (ft3/min).
If the value for K in not known, zero is entered, and the measured value is
then input after each test.
After all the data are input, the operator initiates the calculation of
stage constants. The stage constants are purely a function of the geometry,
and for given sampling conditions will be proportional to the cut diameter of
the stage. The stage constants are compared in order to determine if there
will be any overlapping of stage cut-points. When this occurs the masses of
the overlapping stages will be combined and assigned the smaller of the two
cut-points. Information concerning stage overlapping as well as geometric
and calibration data are then stored in a file for later retrieval by the
IMPACTOR program. The data are also output on the line printer in the form
of Table I. DEFIMP can be run for as many impactors as desired. Since it
has been found that the stage calibration constants (PSI) vary for different
collection substrates (Balfour, 1977) by changing the name, the same
impactor can be defined for different substrates.
DEFTEST
This program is used to provide the specification for each test per-
formed by the data-collection facility. A "test" is a series of impactor
runs made on a single source while that source is operating with a fixed
nominal set of operating parameters (e.g., flow rate, boiler condition).
For each test DEFTEST creates a file of documentation of the source, control
device, and test conditions. This eliminates the necessity to repeat the
test documentation for each impactor run. Table II is an example of the
output of this file. Besides the documentation, the operator also selects
which diameter definition will be used in the calculations for this particu-
lar test. The particle density is also input but is only used in the
calculations if the physical, or Stokes, diameter is used.
288
-------
MRI165
DESCRIPTION:
CIRCLE
7
8860.000
TABLE I. OUTPUT OF DEFIMP PROGRAM
YES HARD COPY
YES DISK UPDATE: FILENAME MRI165/IMP
IMPACTOR NAME
DRI IMPACTOR CALIBRATION
SHAPE OF JETS (CIRCLE/RECTANGLE)
NUMBER OF STAGES
K IN 'DP=K*P*Q*Q' (0 IF UNKNOWN)
STAGE #
1
2
3
4
5
6
7
MASS ON STAGES
# OF JE1
8
12
24
24
24
24
12
1 AND 2
rs SQR ROOT (PSI;
0.1100
0.2700
0.3400
0.3200
0.3200
0 . 3400
0.3600
WILL BE ASSIGNED THE
) JET DIAMETER
0.8700
0.4770
0.1980
0.1190
0.0838
0.0539
0.0542
CUTPOINT OF STAGE
STAGE CONSTANT
0.2325
0.2838
0.1352
0.0593
0.0350
0.0192
0.0145
1
DEFTEST OUTPUT
*****
YES
YES
29
2.430
SOURCE DESC.:
LOCATION:
CONTROL DEV.:
TEST PURPOSE:
REMARKS:
PHYSICAL
TABLE II.
DEFTEST VER-3.0
HARD COPY
DISK UPDATE OF FILENAME: DEFT30/DAT
TEST NUMBER
PARTICLE DENSITY (G/M[3)
230 MW COAL FIRED BOILER
ASPEN COLORADO
APC ELECTROFORCE ESP 375 SCA
DETERMINE FRACT. EFFICIENCY DURING 803 CONDIT.
IMPACTOR PROGRAM DEMONSTRATION
PARTICLE DIAMETER TYPE (I, C, P)
289
-------
ORSAT, METH4, AND METH5
ORSAT, METH4, and METH5 are programs to calculate the results of tests
for molecular weight, moisture content, and mass concentration using EPA
Reference Methods 3, 4 and 5 (Federal Register 1971 and 1977), respectively.
The values of the molecular weight and moisture content are stored in files
and used by the IMPACTOR program. The METH5 data are not required by any
other programs, and so the program creates no files but outputs results to
the line printer.
IMPACTOR
This is the main program in the system. It is used to perform all of
the calculations required to reduce the data from an impactor run to the
particle distribution at a specified set of diameters. The program uses the
contents of files created in the DEFTEST, DEFIMP, METH4, and ORSAT programs
along with the operator input, and produces optional disk files, CRT
graphics, and listings.
The input menu for this program is on three pages. The first page
contains descriptive data, which are also used to gather data from the test-
desciptor file (after the test number is specified), and from the impactor
descriptor file (after the impactor name is given). The run number is used
to form the output file name and therefore must be unique for each run in a
test series. The second page of the input is used to describe the sampling
conditions. The third page is used to enter the mass gain for each stage and
the final filter of the impactor. The number of stages is derived from the
impactor description file.
After reviewing and correcting any incorrect data, the operator
initiates the calculations. The FLOW subroutine calculates the actual flow
rate through the impactor, the flow rate corrected to standard conditions,
and the percent isokinetic sampling rate. The JET subroutine then calculates
the jet velocity and pressure drop for each stage of the impactor. The
pressure drop for each stage is equal to the sum of the pressure drops across
all stages preceding and including the stage being considered. Since only
the total drop across the entire impactor is known (measured or calculated),
it is necessary to estimate the fraction of the pressure drop occuring at
each stage. Because the gas flow through the impactor is accelerated and
decelerated at each stage, it is reasonable to assume that the pressure drop
is due to a loss of velocity pressure, and therefore the pressure drop at
each stage is proportioned according to the square of the jet velocity.
The VISCOSITY subroutine calculates the viscosity of the gas stream as a
function of the temperature and gas constitution. The viscosities of each
component of the gas stream are first calculated as a function of temperature
and then are combined into the viscosity of the total gas stream.
The CUT subroutine then calculates the 50% cut-point for each stage
(based on the selected diameter definition) and the cumulative frequency of
particles collected up to each stage. An iterative loop is used to determine
290
-------
the Cunningham correction factor. During the CUT routine, the stage number
is displayed to keep the operator confident that the program's progress is
satisfactory. Finally, the cumulative mass frequencies are computed for each
stage, and the results are displayed for the operator to confirm. The data
are then ready for the spline fit of the data.
Spline-Fit Subroutine
The Spline-Fit subroutine is used to extrapolate between the data points
to determine values of the distribution at a given set of diameters. The
program uses a spline routine described by Lawless (1978) which first trans-
forms the data to log-normal space and then fits the data with a natural
cubic spline curve. This is different from the routine used by Johnson
et al. (1978) in which a series of parabolas was used to fit the data in
log-log space.
In this program the impactor data are first transformed to log-normal
space; that is, the logarithm of the cut-point diameters is used and the
normal transform of the cumulative probabilities is used. In other words a
curve is fit to the data as plotted on log-probability paper. In this
manner, the spline will fit a true log-normal distribution exactly, and the
inherent errors in fitting near-log-normal distributions are substantially
reduced. Fitting the curve in log-normal space in no way forces the condi-
tions of log-normality on the data.
The natural cubic-spline-curve-fitting method is used to fit the trans-
formed data. The properties of such splines are well defined mathematically;
these splines are the smoothest curves that can be passed through all the
data points, and they are capable of fitting non-log-normal distributions
with small error.
The third-order natural spline that is used in the program consists of a
series of cubic polynomials, joined at the data points with continuous first
and second derivatives and with the second derivatives at the first and last
data points equal to zero. Outside the data range, the cubic polynomial is
reduced to a linear polynomial, so that the natural spline has some utility
in situations calling for extrapolation. Another interesting characteristic
of this spline is its insensitivity to variations at a single point. Changes
at a single point in the value of the function being fitted change the value
of the spline function only at that point and only affect the first and
second derivatives of its nearest neighbors. Thus, an error in the original
data does not propagate throughout the entire fitted region, as can happen
with the interpolating polynomial.
Once the cumulative distribution is fit and values are determined for
the specific diameters, the differential distribution (deltaM/delta(logD)) is
determined by calculating the values of the first derivative at the same
diameters. The data are then transformed back to linear space.
In addition to the spline fit, a linear-regression analysis is performed
on the data in log-normal space. This is done to obtain values of mass-
median diameter and geometric-standard deviation. These log-normal charac-
teristics are calculated so that the data can be used in the computer models
-------
for control devices. The correlation coefficient is also calculated to give
an estimate of how close the data fit a log-normal distribution.
Extrapolation of Data for Inhalable Particulate Information
Since the largest cut-point diameter of most commercial impactors is on
the order of 10 pm aerodynamic diameter, it is necessary to extrapolate
beyond the last data point in order to provide useful data for the inhalable-
particulate program (i.e., cumulative mass less than 15 (Jm aerodynamic
diameter). To provide a degree of control on the extrapolation, the
inhalable-particulate program committee decided that the mathematical routine
should be based on the following steps:
1. Obtain an estimate of the maximum particle diameter, d
I
2. Set the cumulative mass fraction less than d equal to 1.
3. Set the slope of the cumulative- size distribution curve at d
equal to zero.
4. Draw a smooth curve between d and the last impactor data point.
max r
5. Match the slope of the extrapolated curve with the slope of the
spline-fit curve at the last impactor data point.
To meet these criteria, the extrapolation was given the general form of
a hyperbola plus a quadratic. In log-normal space the cumulative distribu-
tion approaches 1 at infinity. Therefore, steps 2 and 3 required that the
curve approach infinity at d and bend over at infinity so that the slope
was zero. The hyperbola was incorporated to satisfy these requirements. To
provide a smooth transition from the hyperbola to the last data point the
quadratic was added. With the combined spline and extrapolated curve, it is
then possible to calculate the amount of mass less than 15, 10, 5, and
2.5 (Jm.
Presentation of Data from the IMPACTOR Program.
Following the calculations, the results of the spline fit are displayed
on the screen. The operator can select to store the data on diskette for
later use in the GRAPH, STAT, and EFFIC programs. The program then displays
a graph of the cumulative distribution on the CRT. The display consists of
the points calculated by the spline fit routine, plus the actual points
calculated at the stage cut-points which blink on the screen. The differen-
tial distribution is then displayed showing the points calculated in the
spline-fit routine. Finally the operation is given an option of line-printer
output of the results. Table III shows an example of output from the
IMPACTOR program which consists of three pages of data. The first page is
documentation and input data. The second page is calculated results
including sampling rate, percent isokinetic, cut-points, and spline fit
results. The third page gives the results of the extrapolation for
inhalable-particulate matter and the results of the linear regression to
determine the log-normal-size distribution parameters.
292
-------
TABLE III. LINE PRINTER OUTPUT OF THE RESULTS FROM THE IMPACTOR PROGRAM.
IMPACTOR VER-3.0
CO
DATE OF TEST
TIME OF TEST
LOCATION OF TEST
TEST NUMBER
SOURCE DESCRIPTION:
SOURCE LOCATION:
CONTROL DEVICE:
TEST PURPOSE:
TEST REMARKS:
PART. DIAM. IS
TEST TYPE
RUN NUMBER
RUN REMARKS:
IHPACTOR TYPE
WATER VAPOR
CARBON DIOXIDE
OXYGEN 2
PARTICLE DENSITY
GAS METER VOLUME
IHPACTOR DELTA P
ORIFICE DELTA P
STACK PRESS. (BELOW ATMOS
BAROMETRIC PRESS.
STACK TEMPERATURE
METER TEMPERATURE
IMPACTOR TEMPERATURE
SAMPLE TIME
AVERAGE VELOCITY OF GAS
GAS METER PRESSURE
NOZZLE DIAMETER
MAXIMUM AERODYN. DIAMETER
MASS GAIN OF STAGE 1
MASS GAIN ON STAGE 2
MASS GAIN OF STAGE 3
MASS GAIN ON STAGE 4
MASS GAIN OF STAGE 5
MASS GAIN OS STAGE 6
MASS GAIN OF STAGE 7
MASS GAIN ON FINAL FILTER
DATE OF TEST 1/6/81
1/6/81 TIME OF TEST 13:30
13:30 LOCATION OF TEST PORT 3A
PORT 3A TEST NUMBER 29 - IHPACTOR PROGRAM DEMONSTRATION
29 RUN NUMBER 1
230 MW COAL FIRED BOILER ACTUAL FLOW RATE (STACK CONDITIONS) 0.461 CFH
ASPEN, COLORADO FLOW 8*™ (STANDARD CONDITIONS) 0.244 CFM
APC E&ECTROFORCE ESP 375 SCA PERCENT ISOKINETIC SAMPLING 103.57%
DETERMINE FRACT. EFFICIENCY DURING S03 CONDIT.
IHPACTOR PROGRAM DEMONSTRATION
ClASS. AERO VISCOSITY OF GAS STREAM 0.0002171 GRAMS/CM-SEC
OUTLET STAGE CCF DP(CLASS. AERO) DP(IMP. AERO) CUM FRACTION
1 - FILE NAME: T29R1/OT
GREASED SUBSTRATES
MR 1 165
DRI IMPACTOR CALIBRATION
8.88% (FROM 'HETH4')
4.00% CARBON MONOXIDE 2.00%
0.00% NITROGEN 73.00%
1.00 GRAMS/CM
10.200 CUBIC FEET
1.70 INCHES HG (CALCULATED)
3.60 INCHES H20
.) 14.00 INCHES H20
24.75 INCHES HG
302 DEGREES F
78 DEGREES F
260 DEGREES F
30.00 MINUTES
240.00 FEET/MINUTE
2.85 INCHES HG
0.60 INCHES
60.00 MICRONS
12.20 MG
13.80 MG
14.46 MG
13-13 HG
6.23 HG
1.81 MG
0.72 HG
0.43 MG
1 1.029
2 1 , 024
3 1.051
4 1 . 120
5 1.212
6 1.442
7 1.725
NOTE: THE MASS ON
OUTPUT ON S
TOTAL MASS PER DRY NO
CLASS. AERO.
PARTICLE SIZE
(MICRONS)
0.20
0.25
0.40
0.50
0.75
1.00
1.50
2.00
2.50
4.00
5.00
7.50
10.00
15.00
20.00
25.00
40.00
50.00
STAGES 1 AND 2
AGE 1, FOR SPL
IMAL CUBIC METE
CUMFR
(STD. BEV.)
-3.3870
-3.1214
-2.5620
-2.3191
- -9951
- .8472
- .5301
- .1978
- .9636
- .6012
- .4321
- .0332
.3247
.9977
.5722
.1738
5.2478
12.1548
9 . 344 9 . 480
11.433 11.569
5.378 5. 13
2.283 2. 16
1.296 1. 27
0.646 0. 76
0.434 0. 70
WILL BE COMBINED AND ASSIGNED
NE FITTING ANALYSIS.
1 332.1070 MG/CUBIC METER
CUHFR CUM. MASS
(PER CENT} MC/DRY fi.
0. 0.1180
0. 0.3002
0. 1.7313
1. 3.3895
2. 7.6472
3. 10.7483
6. 20.9196
11. 38.3563
16- 55.6704
27. 90.9565
33. 110.5450
48. 161.6490
62. 208.3240
84. 279.2380
94. 312.8610
98.5 327.1680
100-0 332.1070
10D.O 332.1070
0.8057
0.5859
0.3555
0.1464
0.0471
0.0183
0.0068
TO THE
DM/DLOGD
CU- METER)
1 . 2862
3.0528
14.9665
22.8697
25.5045
32 . 5045
112.0330
188.5060
194.3580
200.0610
250.5770
370.7900
506-7230
363.4890
222.7530
101.7230
0.0049
0.0000
RESULTS (CONTIUNED)
TIME OF TEST 13:30
LOCATION OF TEST PORT 3A
TEST NUMBER 29
RUN NUMBER 1
INHALABLE PARTICULATE MATTER
CUM. MASS LESS THAN 2.5 MICRON: 55.67 MG/DRY NORMAL CU. METER (16.8%)
CUM. MASS LESS THAN 5.0 MICRON: 110-55 MG/DRY NORMAL CU. METER (33-3%)
CUM. MASS LESS THAN 10.0 MICRON: 208.32 MG/DRY NORMAL CU. METER (62-7%)
CUM. MASS LESS THAN 15.0 MICRON: 279.24 HG/DRY NORMAL CU- METER (84.1%)
NOTE: DIAMETERS FOR INHALABLE PARTICULATE ARE CLASSICAL AERODYNAMIC DIAMETERS.
LOG-NORMAL SIZE DISTRIBUTION PARAMETERS
LEAST SQUARES. LINE: Y=1.77-1.98X
MASS GEOMETRIC MEAN DIAMETER 7.846
GEOMETRIC STANDARD DEVIATION 3.192
CORRELATION COEFFICIENT 0.995
-------
GRAPH
This program is used to graph the output of the IMPACTOR program using
either the Radio Shack line printer, a Houston Instruments DP-1 plotter, or a
Hewlett Packard 7225A plotter. The program first plots the cumulative
distribution in log-normal space for diameters from 0.3 to 30 pm. The
differential distribution is then plotted in log-normal space from 0.3 to
30 pm. Figures 2 and 3 are examples of plots from the HP plotter.
STAT
This program is used to combine several impactor runs into a single file
for later processing by the EFFIC program. This program first checks to see
if the data being combined were calculated based on the identical diameter
definition. The program then proceeds to calculate the mean and the con-
fidence intervals for each of the diameters in the impactor file. The
statistical calculations are performed twice, with points which are widely
separated from the data ("outliers") eliminated from the second set calcula-
tions. The standard deviation is then converted to a confidence interval
range using an abbreviated table of the t-test for significance at the 90%
level. Von Lehmden and Nelson (1976) describes the criteria for outliers.
EFFIC
This program is used to compute the particulate collection efficiency of
particulate control equipment by comparison of simultaneous tests at the
inlet and outlet of the control device. The calculation can be done for
single impactor runs or multiple runs combined by the STAT program as long as
all data are based upon the same diameter definition. Calculations are made
for both total particulate mass and as a function of particle size, and the
results are expressed both in terms of fractional penetration and percent
efficiency. The results are displayed on the screen of the TRS-80, and are
optionally available as line-printer listings, line-printer plots, and/or
plots from one of the plotters. Figure 4 shows an example of the plot
produced by the HP plotter.
DIR/BAS
This program is used to examine and modify the directories IMP/NAM and
IMPDATA/DIR. The first of these is a list of the file names and descriptions
of the impactors in the system, and the second is a list of the files pro-
duced by the IMPACTOR program. One file is produced by IMPACTOR each time it
is run. The program can be used to provide a detailed listing of the names
and remarks for the files in the IMPDATA/DIR file, or it can be used to
delete file names and the associated files from either of these files.
CONCLUSIONS
The TRS-80 cascade-impactor data-reduction system provides an inex-
pensive computerized system to reduce raw cascade-impactor data to size
294
-------
OUTLET TEST # Z9 RUN - 1
DATE & TIME i 1/B/B1 13i 3Z
LOCATION i PORT 3A - GREASED SUBSTRATES
13 IB 30
CLASS. AERO. DIAMETER, MICROMETERS
OUTLET TEST # 29 RUN - 1
DATE & TIME i l/B/81 13:30
LOCATION sPORT 3A - GREASED SUBSTRATES
200
1
CLASS. AERO. DIAMETER, MICROMETERS
Figure 2. Cumulative Distribution* Figure 3. Differential Distribution*
INLET FILENAME! T30R2/IT
OUTLET FILENAMEi T30R2/OT
h 99.9
r 99.99
.3
13 10 30
PHYSICAL DIAMETER - MICROMETERS
Figure 4. Fractional Efficiency Plot*
* Actual plots on 8-1/2 by 11 inch paper.
295
-------
distributions, reducing the time required for the necessary calculations and
providing a measure of quality control for the data processing. Besides
standardizing the listing of input data and results, the system provides
report quality plots of the results. The hardware for the system is
available through Radio Shack and the software is currently available through
the Denver Research Institute and will soon to be available through NTIS.
REFERENCES
Balfour, W.D., (1977): "Use and Limitations of In-Stack Impactors," Masters
Thesis, University of Florida, Gainesville, Florida.
Gushing, K.M., G.E. Lacey, J.D. McCain, and W.B. Smith, (1976): Particulate
Sizing Techniques For Control Device Evaluation: Cascade Impactor
Calibrations. EPA-600/2-76-280 (NTIS PB 262849), October.
Federal Register, (1971): "Standards of Performance for New Stationary
Sources," Vol. 36, No. 247, December 23.
Federal Register, (1977): "Standards of Performance for New Stationary
Sources: Revision to Reference Methods 1-8," Vol. 42, No. 160,
August 18.
Harris, D.B., (1977): "Procedures for Cascade Impactor Calibration and
Operation in Process Streams," EPA-600/2-77-004 (NTIS PB 263623),
January.
Johnson, J.W., G.I. Clinard, L.G. Felix, and J.D. McCain, (1978): A
Computer-Based Cascade Impactor Data Reduction System. EPA-600/
7-78-042 (NTIS PB 285433), March.
Lawless, P.A., (1978): "Analysis of Cascade Impactor Data for Calculating
Particle Penetration," EPA-600/7-78-189 (NTIS PB 288649), September.
Marple, V.A., (1970): A Fundamental Study of Inertial Impactors. Ph.D Thesis,
Mechanical Engineering Department, University of Minnesota, Minneapolis,
Minnesota 55455.
Ranz, W.D., and J.B. Wong, (1958): Impaction of Dust and Smoke Particles,
Ind. and Eng. Chem., 50, No. 4, April.
Von Lehmden, D.J., and C. Nelson, (1976): "Quality Assurance Handbook for
Air Pollution Measurement Systems," EPA-600/9-76-005 (NTIS PB 254658),
March.
296
-------
DEVELOPMENT OF A SAMPLING TRAIN FOR STACK
MEASUREMENT OF INHALABLE PARTICULATE
By: A.D. Williamson and W.B. Smith
Southern Research Institute
2000 Ninth Avenue South
Birmingham, AL 35255
ABSTRACT
A new system, consisting of two cyclones operated in situ followed by a
diluter operated outside the process stream, has been developed to measure the
emission of inhalable particles from stationary pollution sources. Collec-
tion efficiency of 50% for 15 ym particles was achieved in the initial cyclone
at flow rates of 11, 20, and 23 Ji/min, respectively, at temperatures of 23°C,
93°C, and 150°C. At each condition the collection efficiency of the second
cyclone was found to be 50% for particles of 2.5 + 0.6 ym diameter. From the
second cyclone the fine particles pass through a heated probe into the dilu-
tion device. The temperature and relative humidity of the dilution air are
adjustable, and dilution ratios from 10:1 to 40:1 are possible, with standard
operation at a dilution of 25:1. Provision is made to sample the resulting
"plume" in the diluter with absolute filters, cascade impactors, electrical
aerosol analyzers, optical particle counters, or diffusion batteries.
INTRODUCTION
In reassessing the ambient air standards for suspended particulate,
health effects professionals have suggested that particles having aerodynamic
diameters of 15 pm or less are most likely to be deposited in the human res-
piratory system.1 These particles, called inhalable particulate (IP), are
considered to represent the primary health hazard for airborne pollution.
In order to estimate the impact of sources of IP upon the environment,
the U.S. Environmental Protection Agency is sponsoring a large sampling pro-
gram to measure emission factors of IP from source categories of significant
interest. The objectives of this program include measurement of total mass
smaller than 15 ym, the condensable fraction of total emissions, mass in dis-
crete size fractions within the IP range, and continuous size distribution up
to 15 ym or up to 80 ym. Lower priority goals include chemical and biologi-
cal characterization of the sized particulate. As part of the program to
measure emission factors, standardized methods are being developed to sample
stack, ambient, and fugitive aerosols. The methods that are currently con-
sidered to be appropriate for stack sampling include cyclones, impactors with
appropriate precollectors, and dilution for recovery of condensables.
297
-------
The minimum requirements of the instack sampler are that the IP be sepa-
rated into two fractions: one fraction with aerodynamic diameters lying be-
tween 2.5 and 15 ym, and the other, fine-particle, fraction with diameters
smaller than 2.5 ym. The condensable fraction of the aerosol is included with
the fine particles. The aerosol fractions larger and smaller than 15 ym must
be separated using an inertial device having an efficiency (Eff. vs. dia.)
curve bounded by two cumulative log-normal curves with geometric standard de-
viations, ag, of 1.0 and 1.7 and median diameters (Dso) of 13 and 17 ym aero-
dynamic. It has been shown that the measured IP concentration, for a wide
range of typical aerosols, is rather sensitive to the value of the median
diameter but much less sensitive to the ag of the sampler's collection effi-
ciency curve.z'3
This paper contains a brief description of the systems that have been
developed for sampling in process streams (instack).. The systems consist of
two cyclones, or a cyclone and impactor in series, followed by a dilution
device.
TECHNICAL DISCUSSION
Three cyclones were developed for use in the IP sampling program: two
for use in a dual series-cyclone sampler, and the other for use as an IP pre-
cutter with cascade impactors. Their dimensions are given in Figure 1.
Each of the cyclones was calibrated at three temperatures using monodis-
perse dye particles generated with a vibrating-orifice aerosol generator.
The aerosol was passed through a heated tube into the test cyclone and a fil-
ter which were inside an oven. For the larger cyclones, particles of 15 ym
diameter were generated and sampled over a range of flowrates until the flow-
rate for 50% efficiency was found at each temperature. Tests were also made
with the cyclone axes in different orientations to determine if particle
settling was important. The data from these experiments are summarized in
Figures 2 and 3. After the proper flowrates were determined for the IP cy-
clone, the small cyclone (Dso ^ 2.5 ym) was calibrated at the same flowrates
and temperatures to quantify its performance. Data from these tests are
shown in Figure 4.
The data from calibration tests of the IP cyclone and the IP precutter
cyclone are shown plotted as flowrate (for Dso = 15 ym) vs. gas viscosity in
Figure 5. The curves fitted to the data will allow the operators to inter-
polate between the calibration points in choosing the flowrate corresponding
to a Dso of 15 ym at different temperatures.
In order to meet the IP program goal of obtaining measurements of con-
densable matter, a dilution system has been designed and prototypes tested
under laboratory and field conditions. The design objective of this Stack
Dilution Sampling System is to simulate the plume/ambient mixing process as
closely as possible within the size constraints of a field sampling train.
The dilution system is illustrated schematically in Figure 6 and its specifi-
cations are summarized in Table 1. The principal component of the diluter is
a cylinder in which flue gas is mixed with filtered air and the resulting
298
-------
DIMENSIONS (CENTIMETERS)
CYCLONE
SRI-X
SRI-m
SRI-IX
D
6.14
3.11
5.12
Din
1.83
0.75
1.53
De
2.17
0.83
1.81
B
2.92
0.76
2.43
H
8.47
4.91
7.06
h
2.82
1.40
2.35
Z
5.65
3.51
4.71
S
2.40
1.08
2.00
"CUD
2.635
2.22
2.26
Dcup
6.14
3.10
5.12
Figure 1. Dimensions of cyclones used in sampling for
inhalable particles.
299
-------
100
o
a
40
20
•
/
I
I
f
1
/23°C
/ r
930C / /
f 4
/ r
4*1
•
if
i (
i
T
,150°C
4 6 8 10 20 40
FLOW RATE, liters/min 4101-78
Figure 2. Determination of flow rate at three temperatures for D5o = 15 ym
in Cyclone IX.
100
O 40
20
0 VERTICAL, OUTLET UP
& VERTICAL. OUTLET DOWN
• A >
ORI2
ONT
AL
i
/
/
/23°C
/
T,
//
//
/ '/
* ]/
I
1
/
4
x'
7
>•
150°C
4 6 8 10 20 40
FLOW RATE, liters/m n 4181.7S
Figure 3. Determination of flow rate at three temperatures for DSQ = 15
in Cyclone X.
300
-------
456 8 10
Figure 4.
AERODYNAMIC PARTICLE DIAMETER, micronuurs
4181-87
Collection efficiency versus aerodynamic particle diameter for
smaller IP cyclone at 22°C and 11.3 £/min (P), 93°C and 19.8 £/min
(O), and 150°C and 22.7 £/min (A). '
30
25
20
I
§ 15
I
CYCLONE X
Q - 10S log i» -225
CYCLONE IX
= 69.5lo8 1 -150
150
200 250 300
VISCOSITY (ij), micropoise
400
4181-121
Figure 5. Gas flow rate versus viscosity at DSO
cyclones IX and X.
301
= 15 ym aerodynamic for IP
-------
EXHAUST BLOWER
HI-VOL IMPACTOR
FILTER ASSEMBLY
GAS FLOW
TO ULTRAFINE
PARTICLE SIZING
SYSTEM (OPTIONAL)
DILUTION AIR
BLOWER
ICE BATH
TO HEATERS, BLOWERS
TEMPERATURE SENSORS
TO ORIFICE
PRESSURE TAPS
Figure 6.
FLOW, PRESSURE
MAIN CONTROL MONITORS
Diagram of Stack Dilution Sampling System.
4U1-J16B
100
10
•8
1=
I
0.1
a STACK CONCENTRATION
• DILUTED IN SDSS
I I I I I I I ll I
0.01
0.1
PARTICLE DIAMETER, urn
1.0
4181*331
Figure 7.
Submicron size distribution in stack gas from a coal-fired boiler
before and after dilution in Stack Dilution Sampling System. Note
evidence of sulfurie acid condensation mode.
302
-------
TABLE 1. SPECIFICATIONS FOR DILUTION SAMPLING SYSTEM
GEOMETRIC
• Active length of dilution chamber: 122 cm
• Diameter of dilution chamber: 21.3 cm
• Diameter of sample inlet tube: 4.27 cm
• Active dilution volume: 43600 cm3
FLOW
• Sample flow (determined by IP cyclone train) : Ml Jl/min
• Sample velocity: ^27 cm/sec at 150°C
• Dilution air flow: 425 £/min
• Dilution air velocity: 20 cm/sec
• Dilution ratio: ^25:1 (up to 40:1
possible)
• Residence time: 6.2 sec
GAS CONDITIONS
• Sample gas: T <250°C; particles >2.5 ym removed by cyclones
• Dilution air: T = 21.1°C; relative humidity 24%; filtered ambient air
SAMPLE COLLECTION
• Particulate collected on glass fiber filter
• Optional impactor gives cuts at 0.5, 1.0, 2.0, 4.0 jam
• Optional extraction of diluted stream for sizing by optical counter,
electrical mobility analyzer, CN counter, etc.
aerosol-laden mixture analyzed. The sample gas is introduced through a tube
oriented along the axis of the cylinder and the dilution air is introduced as
an annular sheath around the sample. As the sample is cooled by dilution,
condensable vapors form particles under conditions close to those which occur
in actual plumes. The system is designed for a "standard" dilution air at
21°C, 24% relative humidity, mixing time of 6.2 seconds, and a dilution ratio
of 25:1. All of these parameters can be adjusted to some extent. The diluter
can be adapted to a variety of sampling devices to measure the condensed and
diluted aerosol. The total flow can be directed into a cascade impactor or
filter, or portions of the diluted stream can be extracted, further diluted
if necessary, and sampled using electrical aerosol analyzers, optical particle
counters, or diffusion batteries with condensation nuclei counters.
The Stack Dilution Sampling System has been subjected to a variety of
laboratory and field tests. Using a transparent prototype, the formation and
mixing of a condensation plume were studied. For sample gas at typical stack
temperatures, thorough mixing was found to occur before the diluted stream
reached the top of the dilution chamber.
After laboratory characterization, the Stack Dilution Sampling System
was tested at the EPA facilities in Research Triangle Park, North Carolina.
The SDSS was one of a battery of devices used to sample the effluent of a
domestic oil-fired heater. Other sampling devices included SASS trains, gas
analyzers, and an automotive dilution tunnel. Analysis of extracts from the
303
-------
SASS solvent module demonstrated that the majority of the emissions were light
(TCO fraction) organic compounds, whose vapor pressure was too high to be cap-
tured in any of the particulate sampling devices including the dilution sys-
tems. However, both the SDSS and the automotive dilution tunnel captured
about 25% more mass than the SASS trains, indicating that condensation did
significantly increase the particulate loading under the conditions similar
to those at the flue exit.
A more dramatic confirmation of the operating behavior of the system was
obtained in a test at a power plant fueled by high-sulfur coal. The dilution
sampling system was used to sample the emissions of a 50 MW boiler downstream
of the electrostatic precipitator. Since independent measurements recorded
SO3 levels of 7 to 8 ppm in the same duct during the sampling periods, it was
anticipated that a measurable amount of sulfuric acid would be detected in the
SDSS. Since the duct did not have the 6 in. (15 cm) ports required for the
IP dual cyclone train, only the smaller cyclone was used on the probe of the
dilution sampling system. In addition, a Teflon membrane hi-vol filter was
used in order to ensure low sulfate background and minimize artifact sulfuric
acid formation due to the high concentrations of SOa in the stack gas. Aside
from these changes the dilution sampler was operated in a standard manner. All
data from this field test have not been analyzed at this time; however, pre-
liminary results are summarized in Table 2. The mass train and impactor
measurements are consistent with a stack particulate loading of 36 mg/m3 ac-
tual gas volume during the approximate time period of the SDSS test, with ap-
proximately 2 mg/m3 in the fine fraction as defined by the cut of stage 4 of
the U.W. Mark V impactors used. In contrast, the fine particulate plus con-
densable mass as measured by the SDSS was about 26 mg/m3. Although some dif-
ferences are to be expected due to the fact that the impactors and mass
trains were run at different times in different positions in the duct, the
order of magnitude increase in mass is clearly beyond experimental uncertain-
ty. To obtain further evidence that sulfuric acid condensation was occurring
in the SDSS, a portion of the diluted stream was extracted into the Southern
Research Institute Ultrafine Particle Sizing System for size analysis of the
particulate matter smaller than 1.0 ym. As shown in Figure 7, the size dis-
tribution of particulate after dilution was dominated by a mode centered at
about 0.08 Urn. In contrast, the size distribution in the stack peaks at lar-
ger diameters with an integrated mass roughly 5% that observed in the diluted
sample. The data in Figure 7 indicate that a sulfuric acid mist is formed by
condensation in the Stack Dilution Sampling System and, presumably, in the
plume of the power plant under test.
The Stack Dilution Sampling System will be tested at a variety of sta-
tionary emission sources in the course of the IP emission factor measurements
program. We anticipate that this program will provide extensive field vali-
dation of all the devices described in this paper. In addition, independent
laboratory and field characterization of each device is continuing.
304
-------
TABLE 2.
COMPARATIVE MASS LOADINGS (IN MG/ACM) IN TEST OF
UTILITY BOILER BURNING HIGH SULFUR COAL - PRELIMINARY
DATA
Sample
Fine (<2 yim)
Coarse (>2 ym)
Total
Method 17
Impactor
Impactor
SDSS
SDSS
(2/4/81)
(2/3/81)
(2/4/81)
(2/3/81)
(2/4/81)
2.0
2.2
25
27
21
25
35.7
23
27
ENDNOTES
1.
2.
3.
Miller, F.J., D.E. Gardner, J.A. Graham, and R.E. Lee, Jr.
ations for Establishing a Standard for Inhalable Particles.
Pollut. Contr. Assoc., 29(6): 610 (1979).
Size Consider-
J. Air
Kapadia, A., D.Y.H. Pui, and K.T. Whitby. Modeling of Impactor Data and
Calculation of IP Concentration. Final report on subcontract to Southern
Research Institute, EPA Contract No. 68-02-3118, November 1979.
Johnson, J.W-, B.E. Pyle, and W.B. Smith. Extending Precision in a Com-
puter-Based Cascade Impactor Data Reduction System. (Presented at the
Second Advances in Particle Sampling and Measurement Symposium, Daytona
Beach, Florida, October 1979, EPA-600/9-80-004, January 1980).
305
-------
INHALABLE PARTICULATE MATTER SAMPLING PROGRAM FOR IRON AND STEEL
AN OVERVIEW PROGRESS REPORT
By: R.C. McCrillis
Metallurgical Processes Branch
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park
North Carolina 27711
ABSTRACT
EPA's Office of Research and Development has entered into a major pro-
gram to develop inhalable particulate matter (.IP) emission factors, where IP
is defined as airborne particles of <15ym aerodynamic equivalent diameter.
The Metallurgical Processes Branch of EPA's IERL-RTP is responsible for the
iron and steel industry segment of this program. This paper presents a
summary of efforts to date. Every effort is being made to mesh the IP
requirements with those of other EPA sampling programs, thus reducing
overall cost to EPA and minimizing inconvenience to the host plants.
A preliminary literature review indicated the existence of particle
size data for several of the major iron and steel sources. However, none of
these data were obtained using current measurement technology: most of it
does not cover the full IP size range and, in many cases, there is
insufficient documentation to determine test procedures followed and to
adequately define the process operation during the tests. A thorough
compilation and review of existing data, currently underway, will be
completed in June 1981. This will augment the current field test program
which is directed toward those sources with the combination of high priority
and low existing data quality. To date, processes tested are basic oxygen
furnace (EOF) charging and tapping, hot metal desulfurization, blast furnace
cast house (building evacuation approach), and BOF main stack (limited
combustion system after scrubber). Open sources tested are paved and unpaved
roads and coal storage pile wind erosion and maintenance, all both with and
without controls, and an uncontrolled open area. Discussions are now
underway with several plants to test other high priority sources. If funds
allow, duplicate tests will be undertaken at another plant of at least the
highest priority sources.
306
-------
INTRODUCTION
The U.S. Environmental Protection Agency is required, under the amended
Clean Air Act of 1977, to review the scientific basis for the total suspended
particulate ambient air quality standard. Major consideration is being
given to a size-specific particulate standard focusing on inhalable parti-
culate (IP) matter, defined as airborne particles of <_15 pm aerodynamic
equivalent diameter.d' EPA has initiated an extensive program to compile
and review existing data and, where necessary, conduct field sampling
programs from which reliable emission factors for IP can then be determined.
EPA's Office of Research and Development is responsible for developing
these IP emission factors. A major part of this effort is directed toward
the steel industry. This paper discusses briefly the review of existing
iron and steel source particle size data and the field sampling program
being undertaken. The rationale and approach being followed to select
test sites are discussed and results obtained to date are summarized.
Conceptual plans for the remainder of the program are presented.
DISCUSSION
Source Selection
At the outset of the iron and steel sampling program, the decision
was made to proceed with field test site selection voluntarily, rather
than through the application of Clean Air Act, Art. 114. Industry contact
was initiated through the American Iron and Steel Institute (AISI) who
established an ad hoc coordinating committee. Meetings with this committee
were held to present an overview of the whole program and, following
resolution of outstanding issues, to review sources selected for testing.
The EPA/AISI cooperative effort has thus far resulted in the mounting
of an extensive field sampling program at Armco, Inc.'s Middletown Works.
Discussions will soon be initiated with several other companies; several
field test programs should be getting underway soon. The source selection
priority list, shown in Table 1, was developed based on estimated controlled
particulate emissions from each source on a nationwide basis. This priori-
tization represents an average of emission factors developed under separate
efforts: one represented factors from specific short term emission tests
(2); the other, presented values which might be termed typical for long
term operation (•*). it is only fair to say that, at best, this procedure
is only qualitative, but does nevertheless provide a rational approach to
source selection.
Review of Existing Data
The source selection priority list is based on total particulate data
due to the paucity of particle size data. Only six data sets are currently
contained in EPA's Fine Particle Emissions Information System (FPEIS)^.
These data sets, consisting of three open hearth furnace stack tests, two
electric arc furnace tests, and coke oven pushing shed tests, are summarized
in Figures 1-3, respectively. These data are judged to be good, although
307
-------
consideration must be given to when they were obtained (1974-77) and the
advancements made in particle size sampling technology since then. These
three sources should be tested again but not before other equally high
priority sources with no existing data are tested.
A thorough review of both published and unpublished literature is
currently underway in a concerted effort to ferret out all existing particle
size data. A first cut at the review has produced 30 unpublished test reports
containing particle size data. It is felt that there may be a significant
amount of additional data still to be located. It is apparent that some of
these data are of sufficient quality to warrant delaying new tests until
other sources with little or no data are tested. None of these old data were
obtained using the current measurement technology embodying a straight
nozzle into a 15 ym cyclone precutter. Thus, if funding permits, all
significant sources will be tested under this program.
In addition to the selection procedure coordinated with AISI, every
effort has been made to combine measurement of IP with other EPA sampling
programs. Not only does this serve to reduce EPA expenditures, it also
reduces inconvenience to the host companies. During the initial stages of
the program these "piggy back" projects constituted the greatest area of
activity due primarily to the fact that initial groundwork had already been
laid by the Agency, making it possible to mount the field effort- relatively
quickly.
Results To Date
Kaiser Steel Corporation
The first two sources tested under the iron and steel IP program were
the hot metal desulfurization (HMDS) and EOF charging and tapping emission
control systems recently installed at Kaiser Steel Corporation's Fontana,
CA, plant. These tests, performed in coordination with EPA's Office of
Enforcement through the Region V office, consisted of total particulate by
EPA Method 5 and particle size before and after the control device which
was, in both cases, a baghouse. EPA's contractor was Acurex Corporation.
Particle size before control was measured with the EPA two-cyclone IP
train developed for EPA by Southern Research Institute. Due to the low
particulate concentration, baghouse outlet particle size in both cases was
measured with an Andersen Mark III impactor fitted with the 15 ym cut point
cyclone precutter.
Results of the HMDS tests were reported in detail at an earlier EPA-
sponsored symposium^-''. In summary, the data showed an average un-
controlled total particulate emission factor of 1.01 kg/Mg of hot metal
desulfurized (range: 0.23 to 1.82 kg/Mg) with an average IP fraction of 25
percent. Nonsimultaneous measurements of the baghouse discharge showed an
average controlled total particulate emission factor of 0.0045 kg/Mg of hot
metal desulfurized (range: 0.0025 to 0.0055 kg/Mg) with an average IP
fraction of 62 percent. Thus the baghouse collection efficiency averaged
99.6 percent for total particulate and 98.9 percent for the IP fraction.
308
-------
EOF fugitive emissions from hot metal charging and tapping of finished
steel were measured separately. Charging measurements did not include the
addition of scrap. Results of the tests will be reported later in this
session*-"' .
Armco, Inc., Middletown
The extensive testing program undertaken for EPA by Midwest Research
Institute at Armco, Inc.'s Middletown Plant encompassed both open dust
sources and ducted process sources; emphasis was on the former. Speci-
fically, emissions from paved and unpaved roads were measured before and
after the initiation of emission reduction procedures. For paved roads, the
emission reduction consisted of flushing with water and/or sweeping —
vacuuming at regular intervals. Berms of paved roads were treated with
Coherex . The unpaved roads were treated either with Coherex® or water to
suppress emissions. Tests of the Coherex® treated road were conducted on
the second and third days after application of the suppressant. Additional
tests would be required to determine the long term control efficiency decay.
Preliminary results of the paved and unpaved road tests are currently under
review; final results should be available in May.
In addition to the road tests, measurements were made of windblown
emissions from the coal pile and emissions arising as a result of coal pile
maintenance activities. These data, still being analyzed, should also be
available in May.
All of these open source measurements were made using the MRI source
exposure profile technique. Road surface silt and moisture content were
sampled to allow correlation with the measured emission rates. Numerous
samples were taken from other inplant roads to develop an idea of the
representativeness of the sampled road segments and also to allow a more
realistic extrapolation of the test data to the whole plant.
Ducted process emissions measured at Middletown were from the EOF main
stack'9). Although the IP protocol calls for measurements before and after
the control device, the nature of the limited combustion EOF operation at
Middletown precluded measurements before the scrubber. Measurements after
the scrubber included total particulate and particle size as per the
protocol. Results are shown in Table 2. The results are presented for two
production rates, normal and intermediate. The ratios of IP to total
particulate are 69 percent and 57 percent, respectively, for the controlled
emission.
Dominion Foundries and Steel Company (DOFASCO)
DOFASCO first installed cast house emissions controls several years
ago. Their No. 1 blast furnace cast house control system, installed in
1975, was tested by EPA^10^ in 1976. This system employs the total building
evacuation concept. Although particle size measurement of uncontrolled
emissions was attempted, the data were not reliable due to particle bounce
(no precutter was used). Using present techniques, which employ the 15 ym
cyclone precutter ahead of the cascade impact-or, has solved this problem for
309
-------
the ongoing IP program. The combined control system for cast houses No. 2
and 3 was started up in November 1978. These cast houses are evacuated by
a common fan and baghouse sized to control emissions in either one by
isolation valves. Cast house No. 4 is currently being used to evaluate
concepts for local emission control.
The IP emission tests were run during the week of November 10, 1980, on
the combined control system serving cast houses No. 2 and 3; however,
measurements were made only when furnace No. 3 was casting. Emission tests
followed the protocol for ducted sources. Measurements were made for EPA by
GCA/Technology Division in the duct upstream of the baghouse; the open
monitor discharge on the baghouse was not tested due to the perceived
difficulty in obtaining meaningful results. A detailed discussion of the
results is the subject of the next
Bethlehem Steel Corporation, Sparrows Point
GCA/Technology Division is planning to conduct emission tests for EPA
at the new "L" furnace cast house at Sparrows Point this month. This is a
large modern furnace employing close-fitting hoods and covers over the
trough, iron runners, and spouts, a practice pioneered in Japan. Emissions
are ducted to a large baghouse. Emissions, following the ducted source
protocol, will be measured in the duct upstream of the baghouse. Since this
baghouse also controls emissions from numerous other fugitive sources, no
attempt will be made to sample the open monitor discharge. Installation of
scaffolding and ports was completed in February, several months later than
planned, causing the delay in testing from the original November date.
Results will be available in draft form within 3 months following completion
of field tests.
Future Tests
EPA is continually reviewing the overall status of the field testing
program in light of the source priority list and tests already completed or
firmly planned^12). Future tests will continue to address the highest
priority sources first. Testing of high priority sources suspected of
containing a significant fraction of condensible emissions (e.g., sinter
plant windboxes, electric arc furnaces, and coke ovens) will be initiated
in a few months when the condensible emission sampling equipment becomes
available. In the meantime, noncondensible sources (e.g., Q/BOP and full
combustion EOF main stack, material stock piles, and coke pushing) will
continue to be tested as rapidly as possible.
Once the field tests are completed, all data, new and old, will be
pulled together in one source category document giving emission factors
versus particle size for all major sources in the iron and steel industry.
In additon to the source test data, emission factors will be summarized in
formats appropriate for AP-42'13'. The current schedule calls for testing
to be completed in September 1981 and the source category report published
in January 1982.
310
-------
TABLE 1. IRON AND STEEL SOURCE PRIORITY RANKING FOR IP STUDY,
CONTROLLED EMISSIONS
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Process
Coke quenching
Blast furnace cast house
BOF stack
Material stockpiles
Roadway travel
Coke combustion stack
BOF charge and tap
Coke pushing
Sinter, misc. fugitives
Sinter windbox
EAF charge, tap, slag
Coal preparation
OH stack
Coke door leaks
EAF stack
Sinter discharge end
Blast furnace top
Teeming
Ore Screening
BOF misc. fugitives
Coke topside leaks
Industry total
particulate
emissions, Mg/yr
34,500
22,700
20,000
16,300
16,300
16,300
14,500
8,900
8,700
8,200
7,600
7,400
7,300
7,100
6,600
5,700
3,700
3,700
3,300
2,200
2,100
(Continued)
311
-------
TABLE 1. IRON AND STEEL SOURCE PRIORITY RANKING FOR IP STUDY,
CONTROLLED EMISSIONS (Continued)
Rank
Process
Industry total
particulate
emissions, Mg/yr
22
23
24
25
26
27
28
29
30
31
32
Reheat furnaces
Blast furnace combustion
OH roof monitor
Coal charging
Open area
Machine scarfing
BOF HMT
OH misc. fugitives
Soaking pits
EAF misc. fugitives
OH - HMT
2,000
2,000
2,000
1,800
1,100
670
650
640
570
540
190
TABLE 2. SUMMARY OF LIMITED COMBUSTION BOF MAIN STACK PARTICULATE EMISSION FACTORS(9)
Date Run Nos.
7/11/80 PSD-lb -
through PSD-6
7/14/80
7/14/80 PSD-7C
and PSD-8
Cumulative emission factor, kg/Mg (Ib/ton) steel3
<2.5 u">
0.010
(0.020)
0.007
(0.013)
<10 urn
0.010
(0.021)
0.007
(0.014)
<15 urn
0.011
(0.022)
0.008
(0.015)
Total
0.016
(0.031)
0.014
(0.027)
Low carbon steel
produced, Mg (tons)
196
(216)
152
(168)
a Steel produced.
b Results are the average of the first three heats and are considered to represent emissions
during normal production rates.
c Results are for the last heat tested and represent emissions for intermediate production
rates.
312
-------
Co
o
K
U
Q.
U
2
3
O
99.990
99.950
t».90
tt.ao
99.30
99
98
95 i
90
"
ro
60
so
40
30
20
10
5
Z
I
0.3
0.2
O.lS
O.I
0.0
10
AVCMACE
A Ot LANCIM6
CNAM«IN6
i i i i i i i i i I I J I i I I I I I
IU
o
ec
tu
a.
iu
3
o
99.990
99.950
99.90
99.80
99.30
••
98
93
90
: so
701
60
so
.
40
30
20
10
3
2
I
0.9
0.2
0.15
O.I
10° 10 • I02
PARTICLE DIAMETER, micrometers
i (5)
0.0
COVERALL PROCESS AVERACE
O MELT
D TAP-MELT
'"i
'
Figure 1. Average size distribution - open hearth furnace
emissions, uncontrolled.
10' IOW 10'
PARTICLE DIAMETER, micrometers
Figure 2.'^' Average size distribution - Marathon
LeTourneau Electric arc facility.
10*
-------
99.990
99.950
99.90
99.80
99.50
99
98
95
90
£ 80
g 70
5 60
0. 50
Ul 40
1 30
3 20
3
D «0
U
5
2
1
0.5
0.2
OJ5
O.I
0.0
• O Run One
- D Run Two
- X Run Three
.
5r
El/ (D
J>
: / :
JSr^
i.jfec'
K 0^^
/ Q
/
y4
/
-
D
-
-
-
•
-
- .
iii tiiiiil i i iiiiiil i > iiit||
10"' 10° 10 ' 1C
PARTICLE DIAMETER, micrometers
Figure 3.^ •' Average particle size distribution—coke oven pushing.
314
-------
ENDNOTES
References
1. Miller, F.J. et al. "Size Consideration for Establishing a Standard for
Inhalable Particles," J. Air Pollu. Contr. Assoc., 29(6): 610 (1979).
2. Cuscino, T. A., Particulate Emission Factors Applicable to the Iron and
Steel Industry, EPA-450/4-79-028, Midwest Research Institute, August
1979.
3. Barber, W.C., Particulate Emissions from Iron and Steel Mills, EPA/OAQPS
Internal Memorandum, dated November 6, 1978.
4. Reider, J. P., and R. F. Hegarty, Fine Particle Emissions Information
System: Annual Report (1979), EPA-600/7-80-092 (NTIS PB 80-195753),
Midwest Research Institute, May 1980.
5. Fitzgerald, J. , D. Montanaro, and E. Reicker, Development of Size-
Specific Emission Factor (Draft), EPA Contract 68-02-3157 TD 3, GCA/-
Technology Division, October 1980.
6. Gronberg, S., Test Program Summary for Characterization of Inhalable
Particulate Matter Emissions (Draft), EPA Contract 68-02-3157 TD 5,
GCA/Technology Division, September 1980.
7. Steiner, J., and D. Bodnaruk, "Particulate and SC>2 Emission Factors for
Hot Metal Desulfurization," Symposium on Iron and Steel Pollution
Abatement Technology for 1980 (November 1980, Philadelpia, PA).
8. Steiner, J., "Inhalable Particulate Emission Factors for BOF Furnaces in
the Iron and Steel Industry," Third Symposium on the Transfer and
Utilization of Particulate Control Technology (March 1981, Orlando, FL) .
9. Inhalable Particulate Emission Characterization Report: Armco Steel's
No. 16 Basic Oxygen Furnace, Middletown, OH (Draft), Prepared by PEDCo
Environmental, Inc. under subcontract, EPA contract 68-02-3158 TD 6,
Midwest Research Institute.
10. May, W. P., Blast Furnace Cast House Emission Control Technology
Assessment, EPA-600/2-77-231 (NTIS PB 276999), Betz Environmental En-
gineers, November 1977.
11. "Inhalable Particulate Emission Factors for Blast Furnace Casthouses in
the Iron and Steel Industry," Third Symposium on the Transfer and
Utilization of Particulate Control Technology (March 1981, Orlando, FL).
12. Preparation of Overall Iron and Steel Industry Inhalable Particulate
Test Plan (Draft), EPA Contract 68-02-2687 TD 13, GCA/Technology
Division, December 1980.
315
-------
13. Compilation of Air Pollutant Emissions Factors: Third Edition, No. AP-
42 (NTIS PB 275525), US Environmental Protection Agency, August 1977.
ACKNOWLEDGEMENTS
The three sampling contractors for EPA's IP program are GCA/Technology
Division, Midwest Research Institute, and Acurex Corporation. The AISI and
its member companies (in particular Armco, Inc., Bethlehem Steel Cor-
poration, Dominion Foundries and Steel, Ltd., and Kaiser Steel Corporation)
have been most helpful.
316
-------
DEVELOPMENT OF IP EMISSION FACTORS
BY: D. L. Harmon
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
ABSTRACT
In response to the Clean Air Act Amendments of 1977, EPA is consider-
ing an inhalable particulate (IP) ambient air standard. It will be
necessary to have emission factors for IPs to provide for implementation
of the standard. Steps have been taken to develop IP sampling techniques,
extrapolate existing data on particulate characteristics to the IP size
range, and select sampling strategies for major sources based on national
and regional impacts within budgetary constraints. Three year contracts
were awarded to three contractors in September 1979 to conduct plant
surveys and characterize sources for IPs. A priority list of major
sources has been developed and testing is underway to provide data for
IP emission factors for the priority sources.
INTRODUCTION
EPA is considering an inhalable particulate (IP) ambient air standard.
Particulate emission factors, as compiled in AP-42 (1) ("Compilation of
Air Pollutant Emission Factors"), estimate the emission of total suspended
particulates from uncontrolled sources. It will be necessary to have
similar emission factors for IP to provide for implementation of an IP
standard. Early in 1978, a task force on Particulate Emission Character-
ization was formed to develop a program for determining emission factors
based on cut-off size for IPs from both controlled and uncontrolled
sources.
In September 1978, a draft plan (2) was completed by EPA's Office
of Research and Development (ORD) for a program to obtain the IP data as
specified by a priority listing developed by EPA's Office of Air Quality
Planning and Standards (OAQPS). Since the amount of money required for
this program was obviously not available, both ORD and OAQPS decided
that a second draft was needed which would present potential strategies
for partial characterization of the priority list based on the amount of
resources available. The second draft (3) was produced and is the basis
for the ongoing IP emission factor development program.
During the time the IP emission factor program was being developed,
it became apparent that the measurements problem for obtaining IP source
data was much more difficult than originally believed. It was determined
that a sizeable effort would have to be devoted to this area before any
field measurements were actually undertaken. To this end a meeting was
sponsored by EPA's Industrial Environmental Research Laboratory and
Environmental Sciences Research Laboratory, both of Research Triangle
317
-------
Park (IERL-RTP and ESRL), bringing together many of the nation's measurements
experts to develop a program to provide the necessary measurement and
sampling techniques and expertise.
Three steps were deemed necessary to achieve the objectives of the
program proposed to develop IP emission factors:
1. Develop sampling techniques to allow assess-
ment of the particulate characteristics based on
the inhalable potential.
2. Extrapolate existing data on particulate
characteristics to the inhalable size range where
possible.
3. Select sampling strategies for major sources
based on national and regional impacts within
budgetary constraints.
DISCUSSION
The IP Emissions Factor Task Force Executive Committee has been
meeting about once a month since February 1979 to direct the work for
development of IP emission factors. The committee has members from the
following EPA organizations:
Office of Air Quality Planning and Standards (OAQPS)
Industrial Environmental Research Laboratory-RTF (IERL-RTP)
Industrial Environmental Research Laboratory-Cincinnati (lERL-Cin)
Environmental Sciences Research Laboratory (ESRL)
Division of Stationary Source Enforcement (DSSE)
An early task for the committee was to establish a list of priority
sources to be tested for IP emissions. Since the range of sources
emitting IP emissions was large and the time and resources available
were limited, it was necessary to make use of existing data to the
maximum extent practical and test those sources which would provide the
highest practical level of return. The priority sources identified
early in the program are listed in Table I. The only change made in the
priority list since it was developed has been to add Iron Foundries as a
source to be tested. Also, a recommendation to remove primary non-
ferrous zinc smelters from the priority list is being considered. At
the time the original priority list was developed and funding levels
were established, funds were included for lower priority sources to be
added when identified. Delays in testing and increased testing costs
have eliminated any testing beyond that now planned unless additional
funds are made available for this program.
At the beginning of the IP emission factor program, IP's were
defined as those having aerodynamic diameters less than 15 ym. In
addition to the IP size cut it was also desired to collect data for a
318
-------
possible future secondary fine particulate standard. Fine particles
were defined tentatively as less than 2.5 ym in diameter.
A large data bank of particle size data on various industrial
sources is in existence. The most useful data was taken with impactors
which have an upper stage cut-off of 10 ym or less. For this data to be
used to develop IP emission factors it was necessary to extrapolate from
existing data to 15 ym. An extrapolation procedure has been developed
and data from the Fine Particle Emissions Information System (FPEIS), a
data bank developed for EPA containing much of the existing particle
size data, is now being extrapolated to 15 ym. This data should be
adequate to provide IP emission factors for some industrial sources
without additional testing.
The measurement experts assembled to recommend IP test methods
decided that a 2-stage cyclone set would be the best device to use for
point source sampling. Such a system was developed for EPA by Southern
Research Institute to provide 15 ym and 2.5 ym size cuts. Before fabrica-
tion of these cyclone sets was complete, however, questions began to
arise as to selection of a 15 ym upper limit for IPs. In the spring of
1980 it was determined that the definition o-f IP might be revised to
some cut point less than 15 ym so that the cyclones could no longer be
used. At this time, the approved point source sampling method became an
impactor with a 15 ym cyclone precutter. Data is reported as a continuous
plot of emission factor vs particle size from 15 ym down. When the IP
cut point is finalized the appropriate emission factor can be read from
the graph. An example curve is shown in Figure 1.
It is necessary to obtain IP fugitive emission data on industrial
sources as well as point source data. Four test methods have been
approved for sampling IP fugitive emissions depending on the character-
istics of the site to be sampled:
Quasi-Stack Sampling
Roof Monitor Sampling
Upwind/Downwind Sampling
Exposure Profiling
Where cooling and dilution of ducted gas streams result in the
formation of condensed particles, it is necessary to measure these
condensibles. One of the early tasks in the IP program was to develop a
condensibles sampler. A prototype device has been designed, fabricated,
and laboratory tested by Southern Research Institute for EPA. This
device is currently being field tested and should be available for
testing by early spring.
Draft copies of protocols for these test methods have been developed
for those doing the IP testing. (4)(5)(6)(7)
319
-------
In the spring of 1979 a competitive procurement was issued for
three IP characterization contractors to conduct plant surveys and
source assessments for IPs and to support OAQPS in the development and
implementation of an IP standard. The continental U.S. was divided into
three separate geographical areas for contract award purposes:
Area A Area B Area C
EPA Regions EPA Regions EPA Regions
1, 2, 3, and 4 5, 6, and 7 8, 9, and 10
Separate proposals were required for each area an offerer wished to be
considered for. The level of effort for each area was estimated to be
43,500 direct technical labor hours with a period of performance of 36
months.
Contracts were awarded in September 1979 to three contractors, each
with a major subcontractor:
Area A Area B Area C
Contractor GCA Corp. Midwest Res. Inst. Acurex Corp.
Subcontractor TRC Env-, Inc. PEDCo Env., Inc. TRC Env., Inc.
The scope of work for these contracts required the contractor to
conduct on-site plant inspections or surveys for the purpose of defining
and evaluating the particulate pollution problems and for determination
of the fugitive sources. After approval of a test plan the contractor
was to make all necessary studies as directed. Testing would vary from
site to site but could include:
(1) Ducted Particulate Matter - When required, grainloading measure-
ments by mass and particle size and particulate samples for chemical
analysis would be obtained with the approved sampling train. Whenever
particulate control equipment was present, measurements would usually be
taken simultaneously up- and downstream of the control device. Generally,
four independent measurements at each sampling location would be required
for particle size measurements.
(2) Fugitive Emissions - Particulate emissions from both fugitive
process sources and fugitive dust sources would be determined using the
measurement methods approved by EPA. In the case of fugitive emissions,
sampling would include a full cycle of the process during representative
operation. In the case of fugitive dust emission, measurements would be
made under varying wind, moisture, and/or traffic conditions.
(3) Condensible Particulate Matter - Where cooling and dilution of
ducted gas streams resulted in formation of condensed particles, mass
measurements of such material may be required in addition to conventional
mass and particle size measurements of primary particulate.
320
-------
(4) Hazardous Particulate Matter - Where significant emissions of
hazardous materials were part of the IP, it may be required to include
determination of the composition of the particles. This determination
would be for inorganic materials such as lead, arsenic, beryllium, and
mercury.
When the IP characterization contracts were initiated, it was
planned to use EPA personnel from IERL-RTP and lERL-Cin who were familiar
with the industries to be tested to make test site selections. This did
not work out in most cases because many of the EPA personnel did not
have adequate time to devote to test site selection. It was necessary
to direct the IP characterization contractors to develop priority lists
of various industries and recommend specific sites for testing. Early
in the program, it was decided to try to work on a cooperative basis
with industrial organizations to gain access to test sites rather than
use Clean Air Act Section 114 letters to gain access. Most of the
industrial organizations contacted have been willing to work on this
project on a cooperative basis but working out such agreements has been
time consuming.
It was originally planned that much of the testing would be done
early in the contract period but this has not been possible. In addition
to the delays in test site selection and gaining access to test sites,
significant delays have resulted from the need to develop IP sampling
equipment including a condensibles sampler and size selective inlets and
elutriators for use in the fugitive measurements program.
Table II shows the current target dates for completing the high
priority source tests and the source category reports giving IP emission
factors. This testing is scheduled to be completed in 1981. If additional
funds are made available in FY-82 then testing will be extended to
include other sources.
As each source test is completed an individual test report is
written. When all testing is completed for a given source category, such
as Iron and Steel, then one of the three IP characterization contractors
will be given the task of preparing a final report providing emission
factors for that source category. This contractor will combine test
results from all contractors, extrapolate valid pre-existing data,
develop source category IP emission factors, and prepare a report which
will be used for AP-42 input.
Budgetary constraints limit the number of source categories and
number of sources within a category to be tested. Where possible, tests
for IP emissions are combined with other EPA field tests to stretch the
limited funds. Table III lists the IP tests which have been completed
to date. Table IV lists the tests which are in process or scheduled.
Several of these tests were "piggybacked" onto other EPA sponsored
tests.
321
-------
Most of the Paved Road tests for IP emissions have been completed.
Testing on unpaved roads will begin this spring. Midwest Research
Institute has developed predictive emission factor equations for traffic-
entrained dust from paved and unpaved roads based on field tests of a
variety of uncontrolled roads. Sufficient data will be collected from a
small number of field tests of industrial paved and unpaved roads to
verify or revise the emission factor equations. During most field tests
samples of silt from road surfaces will be taken and other data will be
collected so the emission factor equations can be used to develop industrial
road emission factors for various industries.
Tests at three different steel mills have been completed on different
sources and tests at a fourth plant are scheduled. An overall test plan
for the remainder of the Iron and Steel Industry tests has been submitted
to EPA by one of the characterization contractors. Tests at a lime
plant with one kiln controlled by a fabric filter and one kiln controlled
by an ESP have been completed. Tests at a wet cement plant are scheduled.
One secondary lead plant has been tested and site surveys to select
remaining secondary lead plants for testing are scheduled. One primary
copper plant has been tested and site surveys are scheduled to select
sites for other primary nonferrous smelter tests. Tests of two ferroalloy
plants are scheduled. Test sites for the Pulp and Paper testing have
been selected and testing will start as soon as the condensibles sampling
system is ready.
The categories not included in Tables III and IV which list completed
or scheduled tests are Iron Foundries, Asphaltic Concrete, Combustion,
and Incineration. Technical Directives have been issued to IP character-
ization contractors to prepare overall test plans for Iron Foundries and
Asphaltic Concrete. A report giving IP emission factors for combustion
sources based on existing data extrapolated to 15 urn is due in March.
Any field tests required to fill gaps in the existing data base for
combustion sources will be scheduled after this report is received. A
Technical Directive will be issued to one of the IP contractors in March
to prepare an overall test plan for Incineration. With this directive,
work will be initiated for all sources included on the priority list.
REFERENCES
1. Compilation of Air Pollutant Emission Factors, Third
Edition and Supplements, AP-42, U. S. Environmental
Protection Agency, Research Triangle Park, N. C.
August 1977.
2. Kashdan, E., Emission Factors for Inhalable Particulate
Matter: Draft Proposal Plan, U. S. Environmental
Protection Agency, Research Triangle Park, N. C., 1978.
322
-------
3. Abbott, J. H. , D. C. Drehmel, E. Kashdan, and M. D. Ranade
Emission Factors for Inhalable Particulate Matter:
Draft Proposal Plan, U. S. Environmental Protection Agency,
Research Triangle Park, N. C., 1979.
4. Harris, D. B., Procedures for Cascade Impactor
Calibration and Operation in Process Streams-Revised
1979, 2nd Draft, U.S. Environmental Protection Agency,
Research Triangle Park, N. C., 1979.
5. Protocol for the Measurement of Inhalable Particulate
Fugitive Emissions from Stationary Industrial Sources,
Draft, U. S. Environmental Protection Agency, Research
Triangle Park, N. C., March 1980.
6. Wilson, R. R., W. B. Smith, Procedures Manual for
Inhalable Particulate Sampler Operation, Draft,
U. S. Environmental Protection Agency, Research
Triangle Park, N. C., November 1979.
7. Williamson, A. D., Procedures Manual for Operation of the
Dilution Stack Sampling System, Draft, U. S.
Environmental Protection Agency, Research Triangle Park, N.
October 1980.
TABLE I. IP EMISSION FACTOR PRIORITY SOURCES
1. Paved and Unpaved Roads
2. Iron and Steel
3. Secondary Lead Smelters
4. Portland Cement
5. Lime Manufacture
6. Asphaltic Concrete
7. Ferroalloy
8. Primary Nonferrous
a. Copper
b. Lead
c. Zinc
d. Aluminum
9. Kraft Pulp Mills
10. Combustion (Coal/Oil)
a. Utility
b. Industrial
c. Commercial/Residential
11. Incineration
323
-------
TABLE II. IP EMISSION FACTOR PROGRAM TESTING SCHEDULE
Industry
1. Paved and Unpaved Roads
2. Industrial Roads
3. Iron and Steel
4. Iron Foundries
5. Secondary Lead Smelters
6. Portland Cement
7. Lime Manufacture
8. Asphaltic Concrete
9. Ferroalloy
10. Primary Nonferrous
a. Copper
b. Lead
c. Aluminum
11. Kraft Pulp Mills
12. Combustion (Coal/Oil)
a. Utility
b. Industrial
c. Commercial/Residential
13. Incineration
Complete Testing
8/81
5/81
9/81
11/81
9/81
5/81
5/81
8/81
6/81
9/81
8/81
12/81
12/81
Complete Source
Category Report
9/81
7/81
1/82
3/82
12/81
7/81
7/81
12/81
10/81
12/81
12/81
4/82
4/82
TABLE III. COMPLETED IP EMISSION TESTS
Source Category
Paved and Unpaved Roads
Iron and Steel
Cement and Lime
Primary and Secondary Nonferrous
Sources Tested
Paved Roads
EOF
Hot Metal Desulfurization
Paved Roads
Coal Storage Pile
Cast House
Lime Plant
Kiln-ESP
Kiln-Fabric Filter
Material Transfer
Product Loading
Secondary Lead-various ducted
and fugitive sources
Primary Copper-matte tap,
slag tap, and idle ladle
324
-------
TABLE IV. IP EMISSION TESTS IN PROCESS OR SCHEDULED
4.0
2.0
Source Category
Cement and Lime
Ferroalloy
Iron and Steel
Pulp and Paper
Sources Being Tested
Wet Cement Plant
Kiln-ESP
Product Grinder
Raw Material Transfer
Clinker Transfer
Product Loading
Mill Building
Silicon Metal Furnace
FeMn Furnace
Cast House
DCE Furnace
Non-DCE Furnace
Smelt Dissolve
Lime Kiln
f 1.0
of 0.8
o
u
< 0.6
0.4
LLI
I1J
0.2
0.1
1 Ib/ton = 0.5 kg/metric ton
0.1
0.2
0.4 0.6 0.8 1.0 210
PARTICLE DIAMETER, pm
8 10
20
Figure 1. Emission factor for controlled emissions from hot metal desulfurization plant based
on one test.
325
-------
INHALABLE PARTICULATE EMISSION FACTOR PROGRAM PURPOSE AND DEVELOPMENT
By: Frank M. Noonan and James H. Southerland
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Monitoring and Data Analysis Division
Air Management Technology Branch (MD-14)
Research Triangle Park, NC 27711
ABSTRACT
The Environmental Protection Agency is in the process of reviewing the
technical criteria and data bases to determine whether the establishment of
a particle size based National Ambient Air Quality Standard (NAAQS) for
particulate matter is warranted. Upon adoption of such a size specific
particulate standard, the Clean Air Act requires the States to develop and
submit revisions to their State Implementation Plans to define how they will
carry out a program to attain and maintain such a standard. Such revisions
would necessitate the collection and use of information related to size
selective particulate emissions from existing and future sources. Thus a
need exists to initiate development of an emission factor data base to allow
meeting such objectives.
A review of the basis and objectives of EPA's Inhalable Particulate
(IP) program currently underway to obtain emission factors is presented.
Due to the limitations of resources and the large number of sources to be
characterized, a prioritization of source categories had to be developed for
the emissions testing program. The criteria for selection of and prioriti-
zation of source categories for particle size characterization are presented.
INTRODUCTION
The basis of the IP emission factor program is to provide data to
support and meet the legal requirements of the Clean Air Act as amended in
1977.1 The 1977 Amendments require the Administrator to complete reviews of
all existing National Ambient Air Quality Standards (NAAQS) and the air
quality criteria upon which they are based before the end of 1980 and at
five year intervals thereafter, and to revise NAAQS as appropriate. The air
quality criteria documents contain the latest scientific knowledge useful in
indicating the kind and extent of all identifiable effects on public health
or welfare. The law provides that the primary NAAQS must be based on health
effects criteria and provide a reasonable margin of safety to ensure protec-
tion of the public health. Secondary NAAQS define levels of air quality
which the Administrator judges necessary to protect the public welfare from
any known or anticipated adverse effects of a pollutant.
326
-------
In late 1977, EPA's Office of Research and Development (ORD) initiated
review of the existing criteria documents. To date, ozone, carbon monoxide,
hydrocarbons, lead, and nitrogen oxides have been completed, and sulfur
oxides, particulate matter draft documents are in the process of review.
The major concern regarding the current particulate standard is with
the human health effects associated with size specific particulate exposure.
In addition to total suspended particulates the agency is examining the
effects of several size fractions including IP matter, <_ 15 micrometer (ym)
aerodynamic equivalent diameter and fine particulate £ 2.5 ym aerodynamic
equivalent diameter and fractions between these size cut points.2 The IP
fraction is based on particulate matter which can deposit in the conducting
airways and gas exchange areas of the human respiratory system during mouth
breathing. A fine fraction <_ 2.5 ym is based upon chemical composition and
the bimodal size distribution of airborne particles and the predominant
penetration of particles <_ 2.5 ym into the gas exchange region of the respi-
ratory tract.3
Program Development
In order to provide for attainment of these standards, States are
charged with the responsibility of developing and implementing emission
control programs for individual sources through the mechanism of State
Implementation Plans (SIP) which are subject to EPA approval. A stepwise
development of the SIP process is briefly presented in Figure 1. The legal
time requirement for the States to submit a plan to attain and maintain a
NAAQS after promulgation by the agency is nine months. Upon receipt of the
SIP, the EPA has four months to approve or reject the plan.
Should a size specific particulate standard be established, it will be
necessary for States to develop and submit revisions to their SIP to demon-
strate attainment and maintenance of the standard. These revisions will
entail the development and utilization of emission inventories, air quality
simulation models, and air quality baseline data for purposes of control
strategy development. Emission inventories of the stationary source IP
fraction of emissions will be needed for controlled as well as uncontrolled
sources, similar to those presently inventoried for the existing particulate
standard. Emission limitations and control measures must be adopted by the
States to ensure attainment and maintenance of air quality standards through
rules and regulations legally enforceable by the State agency. Additionally,
the SIP revisions must set forth legally enforceable procedures for review
of new sources and modification of existing sources based upon control
requirements anticipated necessary to meet the NAAQS.
Therefore, formulation of a size specific particulate control strategy
if needed begins with development of a baseline emission inventory of existing
particulate emission sources (including both ducted and nonducted fugitive
emissions) within a study area. Projected size specific emission inventories
will also be required for five year forecast periods, taking into account
source growth, more stringent controls, etc. The baseline inventory would
be input information to appropriate air quality simulation models or
327
-------
Va)ldace/Calibrate_
Loop
Develop Uuta Base
Air Quality Monitoring
Meteorological
Emission Inventory
Predicted
vs.
Measured
Air Quality
Control Strategy
Development
Account for
Emission
Growth
A
/ \
Yes XttainX
t I£S / M4An
00
Yes
Public Hearing
Loop
Control Strategy
Translated to
Regulations
Hearing
Review and
Record Public
Hearings
Amend and
Modify Regulations
and Plan
No
FiRuru 1. STATE IMPLEMENTATION PLAN DEVELOPMENT PROCESS
-------
techniques which would be calibrated against observed ambient size specific
particle concentrations. The calibrated models would then be used in conjunc-
tion with baseline air quality data and projected emissions from the sources
to determine the source specific (or source category) emission reductions
needed to attain and maintain the particulate standard.
It is too costly for States to conduct exhaustive source by source
monitoring to determine size specific particulate emission rates. Thus
emission or estimation factors to be utilized in estimating these emissions
are required. EPA has historically provided a compendium of emission factors
that can be used to estimate source emissions; AP-42,^ "Compilation of Air
Pollutant Emission Factors." This document provides emission data for
typical, uncontrolled, existing sources. Data are presented for most major
source categories, and are listed for each significant process within the
source category. Individual source category sections of AP-42 are periodi-
cally updated as new information become available. The IP testing program
underway is directed toward producing size specific emission factors for
many major particulate source categories, which will be published in AP-42,
or in a similar format for use by the States in development of SIP revisions.
Resource constraints prevent this program from addressing all possible source
categories at the present time. Therefore, some data gaps will remain.
Early in the IP program, a search for reliable particulate mass fraction
data for uncontrolled and controlled process emissions was conducted.
Computerized data bases that contain mass emission data and particle size
distribution data such as the Source Test Data System (SOTDAT), Regional Air
Pollution Study (RAPS), and the Fine Particle Emission Information System
(FPEIS) were searched. Also, support document or New Source Performance
Standards, other published data, such as Handbook of Emissions, Effluents,
and Control Practices for Stationary Particulate Pollution Sources,5 and
other related EPA documents were reviewed for applicable particle size
source related data. Particle size distribution data found in the existing
data bases were considerable for some source categories, (e.g., coal fired
power plants and some metallurgical foundries) but rather limited to non-
existant for many expected significant emission sources. Some of the
existing data bases were incomplete or not useful because the limited data
for sources of interest lacked detailed information pertaining to sampling
characteristics, particle sizes tested and reported (data was typically
confined to the 0.2 to 2.0 ym range), and source operating conditions. Much
of the data provided only particle size emissions information after controls.
The available particle size data was generally confined to ducted sources,
and obtained mostly by cascade impactors designed for particle size classi-
fications, well below the tentative 15 ym value for IP. Data on process
fugitive emissions, and on open dust sources were sparse. Moreover, little
data exists regarding condensible particulate which may be a major contribu-
tion to ambient fine particulate (<_ 2.5ym) levels.
Due to the lack of sufficient technical data to develop IP emission
factors, Office of Air Quality Planning and Standards (OAQPS) requested
ORD's Industrial and Environmental Research Laboratory (IERL) to support
development of a program which would meet these objectives. A list of
329
-------
industrial source categories was compiled, which OAQPS viewed as potentially
significant source categories based on the following criteria:
"Source particulate emissions characteristically found by existing
measurements or judgements were estimated to have substantial portions
less than 15 ym.
"Source particulate emissions in the National Emission Data Systems
(NEDS) were reported to have high mass emissions with implied national
and regional impact.
"Sources where particle size data and characterization is virtually
nonexistant.
"Fugitive emissions from processes are usually associated with the
source category.
"Source categories where condensible emissions, were expected and
thus no appropriate data base, could be anticipated, (e.g., metal fume,
sulfate, etc.)
Efforts were also made to identify and separate sources which character-
istically produce particulate emissions totally greater than 15 ym or totally
less than 15 ym since currently available emission factors compiled in AP-42
would potentially be sufficient in such cases, assuming final selection of a
15 ym upper limit.
The priority list of source categories recommended to IERL for inclusion
in the test program included Iron and Steel, Primary and Secondary Nonferrous
Smelters, and selected sources such as cement manufacturing, mineral (ore)
extraction and beneficiation, etc. The list did not recommend early sampling
of combustion sources because it was believed that available particle size
data might be sufficient for development of an IP emission factor should
sound data extrapolation methods become available. Also it was believed
that combustion sources are largely being addressed by existing control
programs. Further, a time delay before testing could be initiated was
anticipated due to the necessity for development of a suitable condensible
particulate test measurement method and development of methods to obtain a
15 ym upper limit, since available particle sizing techniques were not then
applicable above about 10 ym.
During the program development, the question of balancing program
resources and anticipated source emission characterization testing require-
ments was considered. A first estimate of costs to perform the work neces-
sary for developing IP emission factors for the complete initial priority
list of sources submitted by OAQPS was in excess of resources available to
ORD or OAQPS for this purpose. Additionally, time considerations for develop-
ment of IP emission testing equipment and methods along with strategies for
sampling listed source categories led to revisions in the priority list.
OAQPS and IERL jointly reviewed the budget constraints and selected a final
priority list of source categories, presented in Table 1, which were expected
to have the greatest impact on the SIP process. This priority list was
developed with the expectation that, from time to time, revision of source
330
-------
priorities would be made as the IP testing program progresses and data
become available from various source assessment projects.
The IP emission factor priority list emphasizes fugitive emissions
(process and dust) since the quantity and size distribution of these emis-
sions are uncertain and are expected to be a significant factor in the
implementation of IP standards. For some sources, total particulate "mass"
emission data are available but particle size data are not. Efforts on
these sources would routinely be expected to require particle size charac-
terization only. This is not an ideal situation, but is generally conceded
to be a reasonable approach. Thus, further cost adjustments had to be made.
When the final plan was being firmed up, it was decided that emission factors
would be developed not only for the 15 ym cut point but also for 2.5 ym as
this information would be needed should a "fine" particulate standard be
developed in the future and size distribution data between 2.5 and 15 ym be
required. Provision for collection and storage of particulate samples for
future chemical analyses and characterization was made. There was also a
consensus on the desirability to conduct a number of separate emission tests
rather than conducting single tests because multiple data points would
provide more confidence in the emission factors. This is especially true
since emission rates can vary by a factor of 2-3 among similar sources
within the same source category.
TABLE 1. PRIORITY LIST OF SOURCE CATEGORIES3
Paved and Unpaved Roads
Iron and Steel
Portland Cement
Lime Plants
Asphaltic Concrete
Ferroalloy
Secondary Nonferrous Lead
Primary Nonferrous
Copper
Lead
Zinc
Aluminium
Kraft Pulp Mills ,:; :
Combustion (Coal-Oil)
Utility :
Industrial
Commercial and Residential
Incineration
Lower Priority Sources
Priority list not in absolute order of data needs or source category's
potential impact.
331
-------
To ensure that the testing phase would yield desired objectives,
requirements for plant surveys and source characterization reports to IERL
project officers would be made available prior to testing so that the rela-
tionship of the plant to "typical" existing plants could be assessed. Plant
survey reports and preliminary assessments include process descriptions, pro-
duction and operating data, current operating practices, emission estimates,
number and kinds of fugitive sources and plant representativeness of the
industry. Source category reports include review of available emission data
published in literature, EPA data files (e.g., FPEIS, NEDS, etc.), and AP-42
emission factors.
Source characterization requires using sampling protocols developed by
EPA for both ducted and fugitive emissions. Where control equipment is
present, IP measurements shall be taken simultaneously up stream and down-
stream so that uncontrolled and controlled emission factors are developed.
All pertinent process equipment operating parameters required to support the
emission factor development are required as part of the test report. Gener-
ally, four independent measurements at each sampling location are required.
In the case of fugitive process emissions, the sampling must include a full
cycle of the process during representative operation, with process equipment
parameters recorded. Fugitive dust emission measurements shall be made
under varying wind, moisture, and other (e.g., traffic) characterizations as
appropriate.
Program Progress
From mid 1978 to date, implementation of the IP program has gone forward
at the direction of the IP Emission Factor Task Force Executive Committee,
which is composed of members of Office of Air Quality Planning and Standards,
Industrial Environmental Research Laboratory, Environmental Sciences Research
Laboratory, and Division of Stationary Source Enforcement. The purpose of
this committee has been to direct the overall program, keep its members
informed of how the program is proceeding and, to decide on issues related
to the IP program such as source testing schedules and review of lists of
tasks necessary to finalize IP emission factors. Several major steps have
been successfully realized since the beginning of the program. EPA and many
of the nation's particulate measurement experts have developed necessary
sampling techniques for measurement of IP particulate matter. During the
intervening time, formulation of particle size extrapolation techniques,
which can be applied to existing data bases such as the FPEIS to estimate IP
emissions have been developed. Measurement protocols for sampling IP have
become available and are presently in use in source characterization tests
today.
Performance of the IP testing program as specified in the established
protocols will result in reports and will be summarized in a format that
presents the emission factors over the entire particulate size distribution
ranges up to 15 ym for each measured process in the source category. The
necessity for developing the emission factors over the entire range is due
to the possibility that some cut points other than 15 ym (or 2.5 ym) may be
332
-------
adopted as IP standards. The final step will be the compilation and publi-
cation of IP emission factors for each source category in a format that will
be appropriate to supplement AP-42 along with a background documentation for
the emission factors. Table 2, lists major elements of the IP emission
factor program development.
TABLE 2. INHALABLE PARTICIPATE EMISSION FACTOR PROGRAM DEVELOPMENT
0 Priority List of Source Categories
Development of IP Measurement Equipment and Sampling Protocols
0 Source Site Selection and Evaluation Reports
0 Specific Plant Survey with Recommended Plans
0 Testing Program Implemented
0 Data Analysis and Development of IP Emission Factors for Processes
Tested
0 Emission Factors for Source Category Compiled and Formated for AP-42
0 IP Emission Factor Guidance for SIP Preparation When Standard Is
Promulgated
This paper has focused on the requirements and development of the IP
emission factor program, and provides a brief rationale for prioritization
of resources available for this characterization effort. The ultimate
purpose of the IP emission factor program is to provide assistance to the
States who will need guidance in the preparation of their emission inven-
tories and SIPs, should a size specific National Ambient Air Quality Standard
be adopted by the agency. As a result of this program the Agency will
provide States guidance information through size specific emission factors
for a wide range (though not exhaustive) of source categories that may be
impacted by a size specific particulate matter standard.
References
1. The Clean Air Act as amended August 1977 (42 U.S.C. 7414).
2. Staff Paper Outline for Particulate Matter, Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency, Research
Triangle Park, NC, July 31, 1980.
3. F. J. Miller, et al., "Size Considerations for Establishing a Standard
for Inhalable Particles", Journal of the Air Pollution Control
Association, 29(6); 610-615, 1979.
4. Compilation of Air Pollutant Emission Factors, Third Edition and
Supplements, AP-42, U.S. Environmental Protection Agency, Research
Triangle Park, NC, August 1977.
333
-------
5. A. E. Vandergrift, et al., Handbooks of Emissions, Effluents and
Control Practices for Stationary Particulate Pollution Sources,
National Air Pollution Control Administration, U. S. Department of
Health, Education and Welfare, Cincinnati, OH, November 1970.
334
-------
INHALABLE PARTICULATE EMISSION FACTORS FOR BLAST FURNACE
CASTHOUSES IN THE IRON AND STEEL INDUSTRY
Peter D. Spawn
Stephen Gronberg
Stephen Piper
GCA/Technology Division
213 Burlington Road
Bedford, MA 01730
ABSTRACT
Total mass emissions and the inhalable particulate fraction were measured
in the baghouse inlet ductwork at the No. 3 blast furnace casthouse emission
control device at DOFASCO in Hamilton, Ontario. The tests were sponsored by
EPA's Industrial Environmental Research Laboratory as part of the inhalable
particulate (IP) measurement program. The uncontrolled mass emission rate,
measured in the inlet ductwork, averaged 0.39 Ib particulate per ton of hot
metal cast for four Method 5 test runs. Results of 12 Andersen impactor runs,
conducted concurrently with the mass tests, found 0.24 Ib/ton of inhalable
particulate (less than 15 ym aerodynamic diameter). These measurements did
not include drilling and plugging emissions. Less than 100 percent capture
of the total uncontrolled emissions by the control system was noted. Some
casting emissions escaped capture and were not measured in ductwork.
BACKGROUND
Under the Clean Air Act Amendments of 1977, the U.S. EPA is required to
review the scientific basis for the total suspended particulate ambient air
quality standard. A size-specific ambient standard based on inhalable par-
ticulate matter (IPM), defined as airborne particles £ 15pm aerodynamic equiv-
alent diameter, is being considered.
EPA's Office of Research and Development (ORD) is responsible for devel-
oping IPM emission factors for possible use by states in revising State Im-
plementation Plans if an ambient standard based on particle size is promul-
gated. A major portion of ORD's IPM program is directed towards establishing
size-specific emission factors for major iron and steel making processes.
Based on a survey of best available emission factors, the blast furnace
casting operation is considered the second largest particulate emission source.
in the integrated iron and steel industry, generating about 22,000 metric tons
of total particulate per year, industry wide. Casting emissions are not con-
trolled at most U.S. blast furnaces, and are discharged to the atmosphere
through open roof monitors. Consequently, accurate measurement of total mass
emissions and particle size distributions has historically been unfeasible or
335
-------
quite difficult.
The first casthouse control system on a basic iron-producing furnace in
North America entered service at Dominion Foundries and Steel, Ltd. (DOFASCO)
in Hamilton, Ontario 1975. A blast furnace that produced ferromanganese, op-
erated by Bethlehem Steel, was fitted with controls in the early 1970s; the
furnace was retired later in the decade.
By the Fall of 1980, three new and seven retrofit casthouse emission
control systems had been installed in North America, and a number of prototype
systems were under evaluation by steel companies and EPA. By 1981, the industry
had made committments to install controls on an additional 58 furnaces in the
U.S.
As casthouse control technology began to develop in North America, it
became possible to accurately measure emission rates in the control device
ductwork. When the inhalable particulate program was initiated in 1979,
DOFASCO operated the only continuous-service control systems in North America,
and was a logical choice for measuring uncontrolled mass and inhalable par-
ticulate emissions.
PLANT DESCRIPTION
Process Description
DOFASCO operates four blast furnaces at the Hamilton, Ontario plant to
supply molten (pig) iron to the two EOF steelmaking shops. A typical blast
furnace and casthouse is shown in Figure 1. The furnace is continuously
charged at the top with iron ore pellets, coke, and limestone. Preheated
blast air blown through the furnace supports combustion of the coke, and
smelting of iron from the ore. Molten iron accumulates in the hearth (furnace
bottom). After several hours, a taphole in the furnace base is drilled open,
molten iron flows from the furnace through refractory-lined runners in the
casthouse floor, and into waiting railcars.
The No. 3 furnace is rated at 2150 tons of hot metal per day. It has a
working volume of 32,346 ft^ and a hearth diameter of 22 feet, 3 inches.
Entering service in 1956, its most recent reline at the time of testing was
in 1972. Furnace charge composition, operating parameters and hot metal
chemistry data are summarized later in this paper.
Casthouse and Emission Control Description
Figure 2 shows the runner layout at the No. 3 casthouse. Casting emis-
sions result from the hot metal flow through the brick-lined runners and the
interface of the hot metal with the ambient air inside the casthouse. Emis-
sions consist of micron-sized particles of iron oxide and graphitic carbon.
If not captured, casting emissions discharge to the atmosphere through open
roof monitors.
Casting emissions at the No. 3 blast furnace are controlled by a total
casthouse evacuation system which entered service in November 1978. A down-
336
-------
stream fan supplies an exhaust flowrate of about 480,000 acfm to a roof-level
intake plenum, withdrawing casting fumes into a 10 ft. duct leading to the
baghouse. Efforts were made to seal potential leakage paths that would allow
casting emissions to bypass the intake plenum. The total enclosed casthouse
volume is 365,000 ftz, and the 480,000 acfm exhaust rate measured in these
tests provides 1.3 casthouse air changes per minute.
The baghouse serving No. 3 furnace also controls No. 2 furnace. Dampers
in the ductwork are activated so only one casthouse is controlled at any one
time. The tests were conducted when No. 3 was controlled, and no exhaust was
supplied to No. 2 furnace.
The baghouse is a positive pressure unit, with an air-to-cloth ratio of
3.3:1 (at 400,000 acfm). The 1,360 polyester bags measure 11-3/4 inches in
diameter and 30 feet, 7 inches in length. The exhaust fan is designed for
400,000 acfm at a static pressure of 16 in. W.G. using a 2,500 hp, 900 rpm
direct drive motor. When not casting, the fan idles with dampers 90 percent
closed.
Shortly after these tests, the No. 3 blast furnace was taken out of ser-
vice for a reline. DOFASCO was planning to convert the total evacuation sys-
tem to a local hood system. The No. 2 casthouse, which shares the baghouse
with the No. 3 furnace, was also being converted to local hoods. Since local
hoods require less exhaust flowrate (^200,000 acfm), the existing 400,000 acfm
baghouse will be able to control both furnaces simultaneously.
SAMPLING EQUIPMENT
Total Particulate Sampling
EPA Methods 1 through 5 were used to measure total mass emissions. RAC
Stacksampler™ sampling trains were used in conformance with EPA Method 5
specifications. Each train consisted of a stainless steel, heat-traced probe
with a stainless steel button hook nozzle, thermocouple, pitot tubes, and a
Method 3 sampling probe attached as per EPA Method 5. Following the probe was
a glass fiber filter/cyclone bypass assembly installed in the hot box. A
Reeves Angel type 900AF 4-inch diameter filter was used in a filter compartment
maintained at 248°F + 24UF. The gas sample exits the hot box into impingers
where condensible materials are removed. The first, third, and fourth
impingers were modified Greenburg-Smith designs; the second was a standard
Greenburg-Smith. The first two impingers each contained 200 ml of distilled
deionized water, the third was empty, and the fourth contained silica gel.
Following the impingers, the sample gas flow rate and volume were measured at
the control box containing the pump, a calibrated orifice, and dry gas meter.
An integrated multipoint gas sample was collected in a tedlar bas as part
of each Method 5 test run and analyzed with an Orsat analyzer. The samples
were obtained and analyzed in accordance with EPA Method 3 - Gas Analysis for
Carbon Dioxide, Oxygen, Excess Air, and Dry Molecular Weight.
337
-------
Particle Size Distribution: Andersen Impactors
Particle size in the 0.8 to 15 micron range was determined using Andersen
Mark III cascade impactors equipped with a 15 micron cyclone inlet (precutter),
or button-hook nozzle. This in-stack fractionating sampler provides eight size
range measurements, and a cutpoint of greater than 15 microns. The substrates
and backup filters were Reeve Angel type 934AH glass fiber filters, and were
preconditioned for sulfate adsorption.
SAMPLING PROCEDURES
The test site was located in the horizontal inlet ductwork leading to the
baghouse, 11 diameters downstream and 2.6 diameters upstream of any flow dis-
turbance. The horizontal duct was traversed from ports located on the side
and the bottom. A bottom port was used instead of a top port; top port sam-
pling was impossible due to interference with an ore yard crane. A preliminary
velocity traverse determined the following:
• Absence of cyclonic flow conditions
• Average flow and temperature points for particle size measurements
• Proper sized nozzles needed to perform isokinetic sampling
Four mass tests were performed, each requiring two casts per run. Con-
current with the mass tests, a total of 12 Andersen runs were conducted, each
requiring about 15 minutes sample time to obtain a minimum of 50 mg sample,
per impactor. Leak checks were performed before and after each sampling run,
and after changing or disconnecting any train component.
The signal to begin sampling was provided by radio communication from
the process engineer stationed inside the casthouse. The inside observer
carefully documented the origin and magnitude of all emissions generated with-
in the casthouse. An outside observer recorded the opacity of emissions that
escaped capture and were emitted to the atmosphere.
TEST RESULTS
Mass and Visible Emissions Data
Mass emissions data for the four Method 5 test runs are summarized in
Table 1. Each test run sampled two casts, excluding the initial drilling and
final plugging operations. The average uncontrolled mass emission rate for
the four test runs during eight casts was 118 Ib/hr, or 0.39 Ib/ton of hot
metal cast.
Table 2 elaborates on the mass sampling data, showing the gas volume sam-
pled, particulate mass actually collected, and grain loading in the inlet duct-
work. The back-half analyses were low, ranging from 0.4 to 4.6 Ib/hr.
The Method 9 opacity observations indicated that less than 100 percent
capture was achieved by the TE system. The average opacity of emissions es-
caping the No. 3 casthouse during these tests was about eight percent (Method 9)
338
-------
Although it is not possible to quantitatively determine system capture effi-
ciency, the GCA engineers on-site felt that the uncontrolled emission measure-
ments may be 5 to 15 percent lower than the total emission.
DOFASCO has conducted long-term weight measurements of the baghouse hopper
catch and converted these data to emission factors. Unlike the GCA emission
factors reported herein, the baghouse catch data includes captured drilling
and plugging emissions, but excludes the small amount of baghouse emissions.
Review of the baghouse catch data, and comparison with the GCA Method 5 tests
of only four casts suggests that the Method 5 tests represent individual casts
and may not be representative of long-term average casthouse emissions at
DOFASCO.
Particle Size Distribution Results
Figure 3 shows the average particle size distribution for 12 Andersen im-
pactor runs conducted during 8 casts. Aerodynamic diameter is shown on the
horizontal axis, and the cumulative percentages are shown on the left. The
right-hand scale shows the uncontrolled emission factor for each particle cut
size.
Visual inspection of the Andersen substrates showed that particles less
than about 2 ym were red in color, typical of iron oxide fume. Particles
greater than about 2 urn were black, probably graphitic carbon.
Table 3 provides a numerical summary of the particle size data. Inhalable
particulate, less than 15 ym aerodynamic diameter, averaged 62 percent of the
total mass emissions.
Process Parameters During Testing
To facilitate comparison of the DOFASCO tests to other furnaces Table 4
shows the average of furnace process data during testing. Table 5 summarizes
hot metal characteristics. The sulfur content during these tests was approxi-
mately 2-3 times greater than normally encountered in domestic furnaces since
DOFASCO uses external desulfurization for sulfur control.
Casthouse emissions tend to increase as hot metal sulfur content increases
However, the sulfur levels during the DOFASCO tests varied over a narrow range,
and no conclusions could be drawn regarding the influence of sulfur on measured
emission rates. Assessment of other process parameters relative to the emis-
sions data found no statistically significant relationships. This may be due,
in part, to the limited amount of test data, i.e., only four mass emission test
runs.
CONCLUDING COMMENTS
The following comments summarize the results of the DOFASCO No. 3 cast-
house tests for uncontrolled emission rates.
• Total mass emissions averaged 0.39 Ib/ton of hot metal, excluding
drilling, plugging, and clean-up operations.
339
-------
• Inhalable particulate fraction, i.e.; < 15ytn, averaged 0.24 Ib/ton, or
62 percent of the total mass emission.
• Capture efficiency varied somewhat between casts, and overall, some-
what less than 100 percent capture was generally achieved. This tended
to bias the emission rates reported herein slightly low.
• The data showed no discernable impact of hot metal temperature sulfur,
or other process parameters on mass emissions or particle size distrib-
utions.
The Final Report to EPA contains additional information describing this pro-
gram.
ACKNOWLEDGEMENT
This study was partially funded by the U.S. EPA under lERL's inhalable
particulate program (Contract No. 68-02-2687, Technical Directive No. Oil and
Contract No. 68-02-3157, Technical Directive No. 002), with Mr. Robert C.
McCrillis of IERL serving as Task Manager. EPA's Division of Stationary Source
Enforcement provided some of the funding under EPA Contract No. 68-01-4143,
Technical Service Area 1, Task No. 96C with Mr. Thomas J. Maslany serving as
Task Manager. Mr. Richard Craig, also of EPA, served as Task Coordinator.
A high level of cooperation was received from DOFASCO during the field
tests, and planning efforts, especially from Al Kruzins, Murray Greenfield and
Warren Rombough. The contents of this paper do not necessarily reflect the
views or policies of the U.S. EPA, nor does mention of trade names, commercial
products, or organizations imply endorsement by the U.S. Government.
340
-------
TABLE 1. MASS EMISSION DATA-FRONT-HALF
TEST
NUMBER
1
2
3
4
Average
TONS
IRON
CAST
500
488
496
458
485
TOTAL MASS
LB/HR
86
108
134
131
118
EMISSIONS
LB/TON
OF IRON
0.27
0.39
0.46
0.45
0.39
TABLE 2. TOTAL MASS PARTICULATE TEST DATA-FRONT-HALF
TEST
NUMBER
1
2
3
4
GAS
VOLUME
SAMPLED
(dscf)
106
114
118
102
TOTAL
PART.
COLLECTED
(mg)
144
202
249
227
PART.
CONC.
(gr/dscf)
0.021
0.027
0.032
0.034
PART.
EMISSION
RATE,
(lb/hr)
86
108
134
131
NOTE: BACK-HALF CATCH RANGED FROM 0.4 - 4.6 LB/HR
FOR FOUR TESTS
TABLE 3. COMPARISON OF TOTAL MASS AND IP (<15 ym) FRACTIONS
TEST
NUMBER
1
2
3
4
AVERAGE
IP MASS
EMISSIONS,
LB/TON IRON
0.14
0.29
0.28
0.30
0.24
TOTAL MASS
EMISSIONS,
LB/TON IRON
0.27
0.39
0.46
0.45
0.39
% IP IN
TOTAL MASS
EMISSION
51
74
61
65
62
341
-------
TABLE 4. FURNACE OPERATION DURING FOUR TEST DAYS
COKE RATE, LB/NTHM
OIL RATE, LB/NTHM
CARBON, LB/NTHM -
FLUX, LB/NTHM -
°F -
°F -
BLAST TEMP.,
TOP TEMP.,
BLAST PRESSURE, PSI
TOP PRESSURE, PSI
0 IN BLAST, % -
AVERAGE WIND, CFM -
- 961
- 149
1014
282
1457
423
- 24.0
- 5.1
21.9
89,500
TABLE 5. HOT METAL CHEMISTRY DURING FOUR TEST DAYS
RUN
NUMBER
1
2
3
4
AVERAGE
TONS
CAST
500
488
496
458
485
RUNNER
TEMP.
OF
2725
2735
2710
2720
2720
7*
/o
.095
.090
.113
.089
.097
Si
%
0.75
0.69
0.53
0.55
0.63
Mn
%
0.85
0.79
0.77
0.85
0.82
Percent by weight.
ORE
COKE
FLUX
OFP-GAS TO
GAS CLEANING PLANT
SLAG
RUNNER
PREHEATED
BLAST AIR
IRON RUNNER
IRON LADLE
Figure 1. Typical blast furnace plant
layout (side-view).
342
-------
Figure 2. Runner Layout, No. 3 Casthouse.
OVERALL EMISSION RATE - 0.59 ""' ^ARTICULATE
Ion iron cost
77.77V
99.950
99.90
99.80
99.50
99.
98.
95.
90.
80.
70.
60.
! 50.
40.
30.
20.
• 10.
: 5.
2.
1.
0.5
0.2
0.15
O.I
n n
' I = AVERAGE OF
12 IMPACJOR RUNS W/ 9OV. C I
.
•
'
•
-
-
-
X-
-
-
r
^r_J
/-
.
r
-
-
-
-
-
, i , i , , ,
0.582
O.STI
0.591
0.312
0.275 '
0.254
0.199
0.196
0.117
0.078 ~
0.059 i
0.019 2
i
0.008 u
10
10"
10'
10'
PARTICLE AERODYNAMIC DIAMETER,micrometers
RHO - i.OO
Figure 3. Average size distribution of casting fumes at DOFASCO
No. 3 Blast Furnace Casthouse, Hamilton, Ontario.
343
-------
INHALABLE PARTICULATE EMISSIONS FROM
VEHICLES TRAVELING ON PAVED ROADS
By: Russel Bohn
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
ABSTRACT
This paper presents the results of a field sampling program designed
to test the quantity of dust emissions from vehicles traveling on paved
urban roadways. The sampling protocol used in this program focused on the
exposure profiling technique. Emission factors were determined for inhal-
able particulate, i.e., particles less than 15 micrometers in aerodynamic
diameter. Testing was performed in the winter and spring seasons of the
year in Kansas City and St. Louis, Missouri, respectively.
Samples of the dust found on the road surfaces were collected to de-
termine total loading and silt content. The relationship of measured emis-
sion factors to road surface silt loadings and vehicle characteristics was
investigated.
INTRODUCTION
Traffic entrained particulate from vehicles traveling on paved roadways
has been identified as a major cause of nonattainment of air quality stan-
dards for total suspended particulates (TSP) in urban areas. Therefore,
the quantification of this source is necessary to the development of effec-
tive strategies for the attainment and maintenance of the TSP standards,
as well as for the potential standard for inhalable particulate, i.e., par-
ticles less than 15 micrometers in aerodynamic diameter.
The following paper summarizes the methods used to quantify and char-
acterize the nature of the traffic entrained dust emissions. Preliminary
results are presented showing the variability in the measured emission fac-
tors and road surface dust loadings.
METHODS
Sampling Sites
Roadway sites which were selected for this study were located in TSP
nonattainment areas in the Kansas City and St. Louis metropolitan areas.
In Kansas City, three general site locations were chosen. They were
344
-------
7th Street and Osage, an industrial park area; Volker Boulevard and Rockhill
Road, an urban arterial; and 4th Street in Tonganoxie, Kansas, a small rural
town. These roadways were sampled in February and March of 1980.
In St. Louis, four general site locations were chosen. Interstate 44,
a high speed expressway; Kingshighway at Penrose Park, an urban arterial;
24th and Madison, a commercial arterial; and Benton Road, an urban arterial.
The last two sites were in Granite City, Illinois. These roadways were sam-
pled in May of 1980.
Emissions Sampling Equipment
The Exposure Profiling Technique was used as the protocol for the source
sampling. Previous roadway sampling using this technique and an explanation
of the sampling methods are reported elsewhere (1,2,3). Downwind of the
roadway, in addition to the vertically distributed exposure profiling sam-
plers, size selective inlets (SSIs) for the high volume air sampler were
located at two heights in the dust plume. These SSIs were fitted with
slotted high volume cascade impactors to measure particle size distribution.
The substrates of these impactors were greased to mitigate particle bounce.
Upwind of the roadway, the SSIs were placed at two heights to determine
vertical profile IP concentrations. In addition to the previously mentioned
sampling equipment, one standard high volume air sampler and one dichotqmous
sampler were placed upwind and downwind of the roadway. These instruments
were used to measure TSP and elemental concentrations, respectively. Fig-
ure 1 presents the general sampling equipment deployment for the roadway
testing.
Road Surface Dust Sampling Equipment
Samples of the dust found on the road surface were collected by using
a portable hand-held vacuum cleaner. A typical length of 5 feet was sampled
for every roadway travel lane. All travel lanes of the roadway were sampled
in the vicinity of the air sampling equipment. An exception was the high
speed expressway travel lanes in St. Louis. Due to safety constraints,
only one travel lane could be hand vacuumed during the study.
The collected total road particulate was weighed and then sieved to de-
termine physical particle size and silt content. Samples of,the silt were
further analyzed for elemental composition.
345
-------
RESULTS AND DISCUSSION
Particulate Emissions
Inhalable particulate emission factors for the 19 roadway tests are
presented in Table 1. The emission factor units are pounds of particulate
per vehicle mile traveled. The information is tentative and subject to
change before the final report for this study is submitted to U.S. EPA in
the spring of 1981.
Concerning the ratio of IP to TSP particulate concentrations, the av-
erage ratio of downwind IP/TSP was 0.47 with a standard deviation of 0.13.
The upwind IP/TSP particulate concentration ratio was 0.57 with a standard
deviation of 0.15. The above values are reported for a 2-meter sampling
height. The data indicates that background TSP contains a higher fraction
of IP than the TSP found 5 to 10 meters from the roadway edge.
Street Loadings
For each of the 19 source tests, a sample of the road surface dust in
the immediate area of the sampling site was collected. The sample consisted
of a representative strip across all the travel lanes. The total particu-
late sample was subsequently sieved to determine silt content. Table 2
presents the 19 total surface loadings, silt contents, and silt loadings.
The loading values are based on mass of material per area.
Vehicle Characteristics
Throughout the source testing, the amount and type of vehicular traffic
was monitored. Vehicle counts by type (cars, light trucks, and heavy trucks)
were tabulated. Vehicle speed was monitored with a radar gun. Vehicle
weights were estimated by vehicle type. Table 3 presents the monitored
vehicle characteristics for the 19 tests.
Emission Factors and Adjustment Parameters
A previously developed predictive emission factor equation for paved
roads was developed which accounted for the major adjustment parameters
which control the extent of vehicular entrained TSP emissions. Figure 2
presents the equation (3). An objective of this study is to develop an
equation of a similar form for inhalable particulate. Multiple regression
analysis will be used to establish the dependence of the emission factors
with the source adjustment parameters of road surface silt loading and the
vehicle characteristics of speed and weight. The equation will be devel-
oped in a predictive form, i.e., by quantifying the values of the adjustment
346
-------
parameters for a road segment, ,a source specific emission factor can be
estimated to aid emission inventory development.
SUMMARY
In this study, inhalable particulate emission factors were determined
for traffic entrained paved roadway emissions. A total of 19 source tests
were performed in the Kansas City and St. Louis metropolitan areas using
the Exposure Profiling Sampling Protocol. In addition to the collected
airborne particulate, samples of the dust loading on the paved road surface
were collected and characterized. Measured inhalable particulate emission
factors varied from 0.00026 Ib/VMT for an expressway to 0.036 Ib/VMT for an
uncurbed road adjacent to an unpaved parking area in a small rural town.
Measured silt.loadings found on the paved roads ranged from 0.022 g/sq m
for the expressway to 2.6 g/sq m for the road in the rural town. Typical
average emission factors and silt loadings for an urban arterial would be
summarized as 0.0042 Ib/VMT and 0.47 g/sq m, respectively.
ACKNOWLEDGMENT
This work was supported by the U.S. Environmental Protection Agency's
Industrial Environmental Research Laboratory under EPA Contract No. 68-02-
2814, Work Assignment No. 32. Dr. Dennis C. Drehmel, Particulate Technol-
ogy Branch, was the requester of this work.
ENDNOTES
1. Cowherd, C. , Jr., C. M. Maxwell,,-and D. W. Nelson. Quantification of
Dust Ent;rainment from Paved Roadways. EPA-450/3-77-027. U.S. Envi-
ronmental Protection Agency. July 1977.
2. Bohn, R., T. Cuscino, and C. Cowherd. Fugitive Emissions from Inte-
grated Iron and Steel Plants. EPA-600-2-78-050. U.S. Environmental
Protection Agency. March 1978.
3. Cowherd, C., Jr., R. Bohn, and T. Cuscino. Iron and Steel Plant Open
Source Fugitive Emission Evaluation. EPA-600/2-79-103. U.S. Environ-
mental Protection Agency. May 1979.
347
-------
TABLE 1. INHALABLE PARTICULATE EMISSION FACTORS
Test No.
Location
Kansas City
M-l
M-2
M-3
M-4
M-5
M-6
M-7
M-8
M-9
St. Louis
M-10
M-ll
M-12
M-13
M-14
M-15
M-16
M-17
M-18
M-19
7th St.
7th St.
7th St.
Volker Blvd.
Volker Blvd.
Rockhill Rd.
Volker Blvd.
Tonganoxie , Ks .
7th St.
1-44
1-44
1-44
Kingshighway
Kingshighway
Kingshighway
1-44
24th St. -Granite
24th St. -Granite
City
City
Benton Rd. -Granite City
IP Emission Factor
tlb/VMT)
0.013
0.0036
0.0072
0.00070
0.0020
0.0037
0.011
0.036
0.0078
0.00073
0.00080
0.00026
0.0011
0.0017
0.0037
0.00066
0.00080
0.0016
0.0010
348
-------
TABLE 2. PAVED ROAD SURFACE DUST LOADINGS
Total Surface Loading Silt Content Silt Loading
Test No- (g/sq m)
M-l
M-2
M-3
M-4
M-5
M-6
M-7
M-8
M-9
M-10
M-ll
M-l 2
M-13
M-14
M-15
M-16
M-l 7
M-18
M-19
4.3
4.2
4.2
4.9
4.7
3.3
2.7
17.6
2.5
0.047
0.047
0.047
0-77
0.58
0.58
0.047
13.9
10-1
10.5
_ \"* f
10.7
6.2
3.5
18.8
21.4
21.7
22.7
14.5
12.2
46.0
46.0
46.0
13.7
8.1
8.1
46.0
5.7
7.1
8.6
\&f "M "*/
0.46
0.26
0.15
0.43
1.0
0.72
0.61
2.6
0.31
0.022
0.022
0.022
0.11
0.047
0.047
0.022
0.79
0.72
0.90
349
-------
TABLE 3. VEHICLE CHARACTERISTICS
Test No.
M-l
M-2
M-3
M-4
M-5
M-6
M-7
M-8
M-9
M-10
M-ll
M-12
M-13
M-14
M-15
M-16
M-17
M-18
M-19
Cars
2,100
1,800
1,900
2,800
2,400
3,800
3,100
1,500
2,100
10,050
9,280
17,400
5,000
3,800
3,900
13,200
3,390
3,670
5,500
Number of Vehicle
Light Trucks
330
120
120
50
60
40
50
30
230
450
470
1,200
40
30
30
1,010
0
0
300
Passes
Heavy Trucks
320
120
200
0
0
0
20
0
150
1,090
1,160
1,140
150
110
110
1,220
0
0
0
Speed
(mph)
30
30
30
35
35
30
35
20
30
55
55
55
35
35
35
55
30
30
30
Weight
(tons)
5.6
3.8
4.5
2.1
2.2
2.1
2.3
2.2
4.1
4.5
4.8
3.8
2.7
2.7
2.7
4.3
2.0
2.0
2.4
350
-------
2m A
1m
3m
u> •:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:
m :•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•
4m!
2m
j:;:v':j:£:£:\-.£^y':::x':x':;:x'x; Roadway
. 4m
LEGEND:
A Standard Hi Vol
D Profiler Head
£^> Dichotomous Sampler
Q Hi Vol with SSI
Q Hi Vol with SSI/lmpactor
O Hi Vol with Cyclone/lmpactor
G Greased Filters
Figure 1. Sampling equipment deployment,
-------
OPEN DUST SOURCE: Vehicular Traffic on Paved Roads
QA RATING: B for Normal Urban Traffic
C for Industrial Plant Traffic
EF =
0.026 if—]
\ n/
[-M/-M
\10/\ 280 1
/W \°'7
("2~"7J kg/veh-km
veh-ml
Determined by profiling of
emissions from traffic (mostly
light-duty) on arterial road-
ways with values for s and L
assumed.
Determined by profiling of emissions
from industrial plant traffic yielding
higher than predicted emissions,
presumably due to resuspension of
dust from vehicle underbodies and
from unpaved road shoulders.
Assumed by analogy
to experimentally
determined factor
for unoaved roads.
Determined by profiling of emissions from
light-duty vehicles on roadway which was
artificially loaded with known quantities
of gravel fines and pulverized topsoil.
metric non-metric
EF = suspended particulate emissions kg/veh-km Ib/veh-mi
I = industrial road augmentation factor (see text) -
n = number of traffic lanes
s = silt content of road surface material % %
L = surface dust loading on traveled portion of road kg/km Ib/mi
W = average vehicle weight tonnes tons
Figure 2. Predictive TSP emission factor equation for vehicular
traffic on paved roads.
352
-------
QUALITY ASSURANCE FOR PARTICLE-SIZING MEASUREMENTS
By: C. E. Tatsch
Systems and Measurements Division
Research Triangle Institute
Research Triangle Park, NC 27709
ABSTRACT
Quality assurance for particle-sizing measurements comprises three
components: planning, implementation, and appraisal. The first two compo-
nents relate to activities designed to ensure that data generated will
satisfy project requirements. The latter component concerns assessment and
documentation of the data quality. Suitable resources for routine field
calibration of entire particle-sizing systems are not readily available.
Only certain subsystem parameters, such as flow rates, may be checked in the
hostile environments generally associated with such tests. Thoughtful
execution of a well-designed test plan is thus crucial to the production of
data of the required quality.
Several constraints are important in the design of a test plan that
will ensure the collection of high quality data. Qualified personnel must
be available who are experienced with equipment-specific details, aware of
anticipated process fluctuations, and capable of evaluating the consequences
of changing the test plan to accommodate unforeseen circumstances. The
capabilities of available equipment (including adequate spare parts, tools,
etc.) together with the characteristics of the field facility must also be
considered. Support of a comprehensive, consistent data quality program
must include consideration of additional factors such as the necessity for
laboratory aerosol calibration and adequate resources for assembly, opera-
tion, and disassembly of the particle-sizing measurements system(s). There
must be the capability for timely (preliminary) data reduction in order to
provide the necessary feedback for execution of an adaptable test plan.
Appraisal of data quality consists of qualitative and quantitative
checks on operations that are critical to measurement data quality. Qualita-
tive checks include reviewing the adequacy of, and adherence to, written
operating procedures, consistent with the particular project data require-
ments and site. Quantitative checks may include checks of various subsystem
components, such as flow rate, nozzle quality, weighing checks, and collocat-
ed measurements. At present, emphasis must be on the comparability of test
data, but there is increasing potential for improved precision and accuracy
of particle-sizing data.
INTRODUCTION
Particle-sizing measurements of all types are fundamentally scientific
observations, whether they are made in the laboratory or in the field. They
have an initially well-defined end use, such as characterization of the
performance of a scrubber or validation of a model of a precipitator. But
in research programs the goal frequently shifts before the end of the proj-
353
-------
ect, testing conditions are not what was desired, and the data are used by
someone with quite different motives than were expected. Data analysis may
occur long after data collection, and many things which were obvious earlier
in the project are forgotten. Discrepancies become more obvious as data are
available for extensive review and analysis, which is increasingly prevalent
in our data-base-oriented society. Unless we are willing to accept the
validity of nonrepeatable data, a conscious effort must be made to ensure
and document the quality of data, whether in the laboratory or in the field.
Adding weight to this is EPA's recent commitment to a mandatory Agency-
wide Quality Assurance program. Following issuance of memos from the Agency
Administrator in May and June 1979 requiring QA for "...all environmentally-
related measurements which are funded by EPA or which generate data mandated
by EPA," the Quality Assurance Management Staff (QAMS) was created in the
Office of Research and Development. The QA program being developed by QAMS
is motivated by EPA's need to have data of documented, adequate quality.
The program is being applied from "the top down" through QA programs for
Laboratories and Regional Offices to the project level. QA requirements for
contracts were developed late last year and will be part of all new EPA
contracts involving environmentally related measurements. QAMS is establish-
ing consistent, specific QA guidelines for routine measurements. Although
particle-sizing measurements do not fall into the routine category, evalua-
tion for QA aspects of particle-sizing measurement is closely related to the
basic scientific issues of planning for, and documenting, project data
quality.
The five basic data quality criteria being scrutinized by QAMS are
shown in Figure 1. Accuracy and precision are defined in familiar terms.
Completeness, although frequently applied to automated monitoring networks,
also applies to manual methods.
Two key concepts in the Quality Assurance program are Quality Control
and Quality Assurance. Quality control activities are those quality-related
activities conducted internally by the testing organization. Quality assur-
ance activities are those quality-related activities conducted externally by
the testing organization. Many of the quality-related activities are simi-
lar. The distinguishing feature is the organization responsible: personnel
of the testing organization may perform quality control functions, but it
requires personnel from an outside organization to perform quality assurance
functions. Because of this objectivity, the QA program is a tool by which
management may ensure and assess the quality of the measurement data.
Management commitment to data quality assurance must be demonstrated to
all personnel of the organization to ensure the success of the program.
Responsibility must be delegated for checking on data quality, but all
personnel, beginning with management, must realize and accept their personal
responsibility for producing high quality data.
Figure 2 depicts QC and QA activities at various stages in the life of
a project. The project team is responsible for all test preparation and
execution. The project QC personnel frequently evaluate all aspects of the
354
-------
project. QA personnel normally interact infrequently with both the project
team and the project QC personnel, with followups only if data quality
problems are discovered.
The goal of the project QA program is to ensure and assess that project
data are of adequate quality. If the majority of the QA effort is aimed at
ensuring project data quality, the job of assessing and documenting it will
be a straightforward and inexpensive one. For this reason, a proper, cost-
effective QA program is integral to a well-designed measurement program and
provides a tool for ensuring and evaluating project data quality before and
during data production activities, as well as at the end of the project.
PLANNING
Figure 3 outlines the problems faced in designing any measurement
program: the real focus is not on the sample collection or analysis, but on
the data analysis, which is the result of these activities. Obtaining
useful data as a final result is the motivation for the measurement effort;
if any of the links between the "real world" and the data interpretation
fail, the entire effort may be wasted. The principle of continual, logical
cross-checking and critiquing of the entire data collection effort should be
applied—not just by project management, but by each team member, at each
phase of a project.
The data quality question central to the entire sample collection and
analysis effort is how accurately the accepted data reflect the real world.
Conscious consideration of this criterion throughout the project will enhance
development and implementation of an acceptable quality assurance program.
Figure 4 is an outline of quality control planning applicable to field
tests. The level of acceptable data quality will affect the planning of
many operations and can be helpful in avoiding later misunderstandings with
the customer. Therefore, it should be established very early in the life of
a project. Tested procedures for routinely performed operations should be
available and should be followed by the test crew.
There are three major, and obvious, considerations of any sampling and
analysis effort, regardless of the particle-sizing method being used.
Despite the fact that they are obvious, we have encountered so many omissions
of them in our audits that it seems worthwhile to review them here. It is
useful to recall that all measurement programs suffer from the 'weak-link1
phenomenon: omission of certain key considerations can easily invalidate
excellent efforts in every other area of a project.
The measurement system component, which is the limiting factor in the
quality of data that can be obtained, is the team personnel. The best test
conditions and the best and most elaborate equipment cannot substitute for
personnel who are trained, experienced, and motivated to obtain high quality
data. Often there are too few of these people available, and the test crew
is made up of inexperienced, although well-intentioned, individuals. This
would be fine if no problems occurred during the test; but problems are the
355
-------
rule, rather than the exception. Training and experience are traditionally
considered to be key personnel qualifications. However, it has been demon-
strated repeatedly that workers also must be motivated to high quality work.
The equipment available to the test crew must be known to produce as
high a quality of data as possible. In some disciplines there are Standard
Reference Materials available that can be used to verify a total sampling-
measurement system. However, only selected component checks can be made for
particle-sizing equipment. That is all the more reason to check and document
the performance of every possible component of the measurement system.
Laboratory calibration of instrumentation before and after field use is
highly desirable and should be omitted only if cost is clearly prohibitive.
Careful documentation of calibration procedures and comparison of trends may
be useful. For example, if the calibration constant for an orifice suddenly
changes by 20 percent, there may be corrosion problems. Similarly, a sudden
shift in the calibration constant of a dry gas meter may indicate internal
problems. This information can then be used to develop a rational mainte-
nance program. Anticipating and preventing equipment failure improves the
completeness of the data, which is an aspect of the quality assurance program
being evaluated by QAMS.
The third important general aspect of a QA program for particle-sizing
is timely and valid data reduction. The reduced data and inferences made
from them are the "product" received and used by your customer. As the
sophistication of your data processing increases, so does the potential for
undetected errors. It is especially important to remember that no computer
program can ever be proven correct; it only takes one counterexample to
demonstrate a weakness. One of the most effective means of validating a
program is to have a variety of users with a variety of data use a program.
Also, it is now feasible to compute quite respectable results in the field.
This may be done using one of the first-generation hand-held computers, a
desktop microcomputer, or a portable terminal and telephone link to the home
laboratory computer. In any case, provision should be made for physically
and mathematically appropriate data reduction in as short a time as possible
for the final report. Only in this way can a test crew make informed deci-
sions during the progress of a test. Since the purpose of the intermediate
calculations is different from the final calculations, their content and
format need to be tailored to this purpose. For example, in the operation
of a dilution system for conditioning stack gas, a program for calculating
dilution flow, given the various orifice constants and pressure drops in a
multiple dilution system, would be helpful. The program could be designed
to allow changing (and entering) only the new flow value, with all the
previous data retained in continuous memory. The dilution rate results
would be available within seconds, thus allowing a trained operator to fine-
tune his adjustments in a knowledgeable manner.
Other common sense aspects of a measurement program, which may actually
be considered as part of the QA program, include allowing adequate time for
test crew planning, travel, setup, shakedown, operation, and critiquing of
test results as they are recorded. Provision for adequate utilities and
workspace (which are part of good operating practice) is also part of the
356
-------
data quality program. Considering the many possible sources of error, it is
a good investment of time to map out and document a flow chart of activities,
including reasonable alternatives, as part of the test plan. Suitable forms
for data collection should be included. Although development of these forms
and flow charts is time-consuming, they serve as good tools for reviewing
operations and pinpointing potential data quality problems. They also
provide materials for organizing subsequent tests at a fraction of the
effort yet with no relaxation of the quality planning. This approach to
methodical data collection will require some adjustment for you and your
crew members. At first, there may be a tendency not to complete data forms
"until later"—which seemingly never arrives. Perhaps this early development
stage of the QA program is the best time to demonstrate your commitment to
the QA management system by reviewing and critiquing some of the incoming
worksheets. This is also a good time to include each of the crew members in
the quality planning. Especially as new systems and forms are developed,
implementation of feedback concerning the ease of use, items to be added or
deleted, and format changes will clearly demonstrate that management is
committed to making the new system work. The people who use these tools
generally have useful suggestions for their improvement that may improve
your data quality.
Figure 5 shows an example of an impactor test plan that includes quality
assurance considerations. This is essentially a one-sheet directory that •
serves as a pointer to related test information. It covers not only the
aspects of data quality that factor into calculations, but the additional
operating aspects of a test that impinge on data quality and that are easy
to overlook. This sheet points to additional test-related documentation
that may be useful in substantiating the test. For example, recordkeeping
may be adequate with one notebook per test crew member; at least the hazards
of recording data on available scraps of paper can be avoided. The matter
of reweighing frequency can be evaluated and provided for before the test.
Transportation of the substrates, both from the preweighing to the impactor
assembly and from impactor disassembly to the postweighing area, can be
evaluated and a decision made on whether to do the weighings onsite or at
the home laboratory.
Because each of these decisions involves different resource require-
ments, preliminary cost-benefit types of information should be used; for
example, if substrates are to be weighed at the home laboratory, a set of
preweighed loaded substrates should be transported to the test site and
reweighed upon their return to determine typical effects of transport on the
substrate weights. The effects observed must then be compared with the
desired data quality before a decision can be made to exclude a portable
balance from the test equipment. Similarly, provision for near real-time
data processing should be evaluated. This provides the only secure means of
evaluating the progress of a test in the field. A distinct time lag occurs
after an impactor run until the substrates with samples can be conditioned
and reweighed; however, provision should be made for data analysis very
shortly following the postweighing. This may be performed using hand-held
calculators, desktop computers, or large mainframe computers, and one of
several available programs. However, if this is done, the program should be
357
-------
verified to give the correct results with a minimum of user effort. The
latter criterion will be well received by the users and is an important
factor in ensuring the quality of the results—the fewer user inputs to a
valid program, the fewer chances there are for mistakes.
One approach to data analysis that we have been developing with the
Fine Particle Emissions Information System is to have all data reduction
performed on EPA's computer at Research Triangle Park, with all data entry
in response to a prompt from the computer. This permits the software program
to be well-documented and provides users with comparable and accurately
reduced data. It is available in a uniform manner to the person doing the
preweighings, to the test crew who have recorded the impactor run parameters,
and to the person doing the postrun weighing. Various validation checks can
be built in to assist in evaluating the quality of the raw data. It is also
designed to permit significant changes and extensions as better algorithms
are developed for processing fine particle data. The output is designed to
provide useful summary information to the test crew during a test, as well
as report-ready raw and reduced data on completion of the test. And all
data are entered only once and retained by the system for subsequent use,
which should enhance the quality of the calculated results.
IMPLEMENTATION
Following the extensive planning activities that have been outlined, it
is important that the plans actually be carried out. Insofar as these are
developed and agreed on by management, the test leader, and test crew, much
of the implementation effort of quality assurance will be simplified. These
plans, implemented by those qualified, motivated crew members, who are
cross-checking and supporting each other in collecting good data and respond-
ing to changing test conditions, will reliably lead to high quality data.
The only other activity which is needed to complete the QA program is project
appraisal.
APPRAISAL
Appraisal of the project for its quality aspects is the final part of
the QA effort. Appraisal can be performed by project personnel—and then it
is discussed as part of the project quality control. If it is performed by
personnel who are not organizationally involved in the project, it is known
as a quality assurance activity, and will most likely be requested by your
project officer as part of the routine Agency-wide QA program. The types of
activities will remain similar, and so I will briefly outline the approach
you may take to quality control, or internal QA, appraisals.
Appraisal activity necessarily requires a degree of objectivity on the
part of the person assigned to this job. On a large project, it may require
the full-time effort of a crew member. On smaller ones, explicit assignment
of data quality responsibilities to one person is sufficient. In any case,
the purpose of this assignment is to have clear delegation of responsibility
for quality control and adequate time for review and assessment of project
data quality without conflict from other commitments. The basic tool to be
used is project documentation and its comparison with actual practice,
assuming, of course, that there has been adequate documentation.
358
-------
Review of completed data forms, observation of the performance of
specific tests, comparison of instrument readings against available stand-
ards, and other checks that may be developed are all part of the reasonable
routine for an internal (or external) auditor. Especially as your quality
system is being established, the auditor can provide useful information on
the fine-tuning of the system to your own data quality needs. As the inter-
nal audit program is implemented and documented, the credibility of the data
that you generate will be enhanced.
SUMMARY
In summary, aspects of Quality Assurance that can be applied to specific
programs, and that are consistent with EPA's Agency-wide Quality Assurance
program as applied to particle-sizing measurements, have been described. In
particular, use of feedback mechanisms for collecting, processing and evalu-
ating data were mentioned as an effective tool by which qualified, motivated
personnel can evaluate and improve data quality.
359
-------
Accuracy:
Precision:
Representa-
tiveness:
Completeness:
Comparability:
The degree of agreement of a measurement (X)
with an accepted reference or true value (T)
usually expressed as the difference (X-T).
A measure of mutual agreement among
individual measurements of the same property,
usually under prescribed similar conditions.
A numerical statement of how well a sample or
group of samples or the data derived therefrom
represent the actual parameter variations at the
sampling point, and how well that sampling
point represents the actual parameter variations
under study.
The amount of valid data obtained from a
measurement system compared to the amount
that was expected to be obtained under correct
normal conditions.
The measure of how well data may be
compared within a study and with data from
other studies.
Figure 1. Basic quality criteria.
Planning
Implementation
Data
Analysis
PERFORMING
ORGANIZATION
Project Team
Project QC
QA Project QA
ORGANIZATION
" '• I.I.I
,,,,,, j, 4 ; ,, i. " Project
,. ,, ,, ,, | « v " Report
\ ^
i
Project
T >, QA
Report
Figure 2. Relationships of project QA and QC.
360
-------
Real
Sample Sample Data
— Extraction "Analysis *• Analysis ^" INTERPRETATION
The purpose of SAMPLE extraction and analysis is to obtain DATA for analysis.
Figure 3. Summary of sample/data flow.
QUALITY CONTROL
• Specify data quality required
• Establish and evaluate adequacy of
procedures
• Make sure procedures are followed
• Employ trained and motivated personnel
• Ensure that equipment and supplies are
adequate, calibrated and properly maintained
• Perform independent check of experiment
• Use controls and blanks
• Provide for field data reduction
• Initiate corrective action based on field data
reduction
• Establish chain of custody of data and
samples
Figure 4. Outline of field test Quality Control planning.
361
-------
• Pitot tube
• Analytical balance
• Reweighing checks
• Stage loadings appropriate?
GENERAL
• Procedures available to test crew
• Data forms—velocity traverse, gas composition, impactor
operation
* Provision for range-finding runs
• Recordkeeping
CALIBRATION
• Impactor
• Dry gas meter and orifice
WEIGHING (pre/post)
• Filter conditioning
• Transport provisions
• Comparison of samples with
blanks and substrates
ASSEMBLY
• Leak check (train assembled)
Nozzle/flow rate selection
OPERATION
• Constant flow rate
• Blank—appropriate tinning,
location (using filtered
stack gas, etc.)
DISASSEMBLY
• Filter handling/storage
• Transport to weighing?
DATA ANALYSIS
• Model matches sampling
train
• Minimize user effort
• Impactor at desired
temperature
• Isokinetic operation
• Orientation: absolute and
with respect to flow
• Probe/impactor brushdown
• Check for re-entrainment
• Real-time capability/use
Figure 5. Outline of cascade impactor test plan.
362
-------
PARTICULATE EMISSIONS CHARACTERIZATION FOR OIL-FIRED BOILERS
By: Dominick Mormile
Consolidated Edison Company of NY Inc.
4 Irving1 Place
New York, NY 10003
Stuart Hersh and Bruce F. Piper
KVB Incorporated - A Research-Cottrell Company
175 Clearbrook Road.
Elmsford, NY 10523
Michael McElroy
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94303'
ABSTRACT
The size distribution and:composition of particulate matter emissions
from three oil-fired boilers, representative of utility usage, were determ-
ined for nominal operation under full-load operating conditions. The select-
ed boilers were a 360 MW tangentially fired boiler, a 346 MW face-fired boiler
and a 150,000 Ib/hr steam flow steam sendout boiler. In addition, assessments
of the variability of these particulate characteristics with changes in excess
air level and atomization quality were obtained as an indication of the fluct-
uations which could be anticipated as a result of operating practices. Siz-
ing characteristics were established with a low pressure, cascade impactor
system (University of Washington), an electric aerosol analyzer (which clas-
sifies size according to electric mobility) and electron microscopy tech-
niques with automatic counting. Particle composition analyses utilized
atomic absorption on bulk samples and electron dispersive X-ray (EDX), Auger
Microprobe and electron spectroscopy (ESCA) on individual particles.
INTRODUCTION
Despite the move towards increased coal utilization by the utility in-
dustry, many regions of the country rely on existing oil-fired power plants
as their principal energy source and are likely to continue to do so for many
years. Historically, oil-fired plants have been viewed as clean sources of
energy compared to conventional coal-fired plants. However, oil-fired plants
may represent an increasingly significant source of utility particulate mat-
ter emissions as advanced high efficiency particulate controls for coal-fired
units, continue to be developed and enter into service. Furthermore, since
oil-fired units are situated predominately in metropolitan areas with large
populations, the potential impact of particulate matter emissions from these
sources is of special concern.
In this context, a detailed understanding of the physical and chemical
characteristics of particulate matter emissions from oil-fired boilers is
important for a number of reasons. First, there is growing concern about the
363
-------
particle size dependent nature of possible health effects from particulate
matter emissions as evidenced by regulatory interest in fine "inhalable" par-
ticulates in the respirable size range (<10 micrometer diameter). Secondly,
to formulate emission control strategies, the size distribution of particu-
late matter is of critical importance due to the size dependent collection
performance of various control options. Third, from the standpoint of re-
ducing plume opacity and visibility, the size and composition of particulate
matter emissions play a key role due to the highly size dependent nature of
particle light scattering and plume formation mechanisms. Detailed particu-
late matter emission data from oil-fired boilers, similar to the data de-
veloped for coal, are virtually nonexistent and has resulted in a major gap
in the utility's emission data base.
Purpose and Scope
The objective of this project was to quantify the size and chemical com-
position of particulate matter emissions from representative oil-fired util-
ity boilers. Specifically, the size distribution and composition of emis-
sions from three boilers operated by the Consolidated Edison Company of New
York were measured for normal (baseline) operation under full load conditions.
In addition, measurements were performed under conditions of low furnace ex-
cess air and deteriorated fuel atomization to provide an indication of fluc-
tuations in particulate matter which could be anticipated as a result of
daily boiler operating practices. Polycyclic organic material (POM), S0_,
CO and NO emissions were also measured to provide a more complete character-
ization of the boiler emissions.
This project was jointly funded by the Consolidated Edison Company and
EPRI. The test program was conducted by personnel of the East Coast Division
of KVB Inc. This is the first phase of a continuing research effort at Con-
solidated Edison to characterize particulate emissions. Future phases will
consist of obtaining effects of fuel additives, off nominal operation such
as low NOx and ultimately to correlate particulate characteristics and
other flue gas properties with stack opacity.
Program Description
Test Boilers
The three test boilers, a 360 MW tangentially fired unit (.Ravenswood 1Q),
a 345 MW face-fired unit (Arthur Kill 20) and a 150,000 Ib/hr steam sendout
package boiler (East River 114) were selected because of the distinct dif-
ferences in their combustion designs. The objective of this selection was
to determine if major differences in the particulate size characteristics
could be attributed to specific boiler types. The excess air and atomization
quality tests were conducted only on the tangentially fired unit. All the
boilers were equipped with steam atomized oil burners.
The fuel burned in all of the boilers tested was a low sulfur (maximum
0.3 percent) relatively light No. 6 oil currently used throughout the Con-
solidated Edison sytem. While different fuels might be expected to influ-
ence the overall level of particulate emission and composition, the particu-
364
-------
late size characteristics presented in this report may be generic to con-
ventional oil-fired utility systems due to the suspected commonality of the
particle formation mechanisms involved.
Measurement Methods
Three complementary measurement techniques were used to measure parti-
culate matter size distributions: 1) in-situ low pressure cascade impactor
(particle segregation via inertial separation), 2} electric aerosol analyzer
(based on charged particle mobility in an electric field} and 3) electron
microscopy with automated particle sizing utilizing samples collected with
a miniature electrostatic precipitator. For the latter two approaches, a
prototype dilution-probe sampling system, specifically designed by KVB, was
used. Total mass loading was measured using EPA method 5. Particulate
matter chemical composition was determined for bulk samples and individual
particles by X-ray analysis, scanning Auger mircoprobe (with argon ion
milling) and electron spectroscopy (ESCA).
Field Test Procedure
At each boiler test condition a minimum of two replicate samples were
obtained with each of the particulate measurement techniques to demonstrate
repeatability. The actual test procedure consisted of two phases. The first
phase verified the desired boiler operating condition through inspection of
the burner front and flue gas measurements of 02, CO and NO. The second
phase consisted of performing the various particulate measurements as the
boiler conditions were held constant.
Baseline operating conditions at the three boilers were generally re-
peatable from day to day as evidenced by the consistency of the emission re-
sults. However, the reduced excess air and deteriorated fuel atomization
test conditions were not as repeatable due to the unavoidably subjective
nature of specifying these conditions. The reduced excess air tests were
conducted as close as practical to the boiler smoke threshold. The poor atom-
ization condition was established by decreasing the atomizing steam supply
pressure to the burners and adversely adjusting the fuel oil temperature,
thereby producing large oil droplets.
Results and Discussion
Influence of Boiler Operating Variables
Typical particulate mass size distributions at Ravenswood Unit 10 under
baseline, low excess air and poor fuel atomization conditions are compared
in Figure 1-1. The data are presented in a so-called "differential" plot
where the differential emission rate is plotted versus the log of the par-
ticle diameter. For purposes of interpretation this format can be viewed
as an emission rate histogram where the area under the curve is equal to
the total emission rate in units of lb/106 Btu.
A striking feature of the data is the concentration of fine, submicron
size mass emissions in a narrow "peak" at approximately 0.2 micrometer dia-
365
-------
meter. For baseline conditions, virtually all the particulate emissions are
generated in this size range. This baseline distribution is in great con-
trast with coal fired boilers where the mass mean particle diameter typically
ranges between 10 and 20 micrometers. It is also important to note that the
submicron peak is at a size region which can contribute significantly to
plume opacity due to the light scattering characteristics of these particle
sizes.
A typical photomicrograph of baseline particulate matter in Figure 1-2
shows spherical shaped particles with a relative size uniformity consistent
with the size distribution results. Chemical analysis indicates mostly sul-
fate compounds with smaller quantities of carbon and oil ash mineral con-
stituents.
Poor Fuel Atomization: Deteriorated fuel atomization resulted in the
generation of a substantial mass of large particulates (.greater than 10
micrometer diameter) as indicated in Figure 1-1. The fine submicron distri-
bution and composition was essentially unaltered compared to baseline. Sig-
nificantly, no changes in stack opacity were noted despite an up to a 10-
fold increase in total particulate emissions. Photomicrographs of the large
particles as typified by Figure 1-3 indicate irregularly shaped, porous par-
ticles with sponge-like appearance. Surface analysis of individual large par-
ticles indicated mostly carbon, but with an abundance of sulfur and oxygen
(sulfate) and lesser quantities of trace elements. However, particle depth
analysis illustrated in Figure 1-4 revealed that the dominant portion of the
sulfate was contained in a surface layer while the bulk of the particles con-
sisted mainly of carbon.
These results indicate that reducing the fuel atomization quality pro-
duces large particulates which appear to be derived from the carbonization
of large fuel droplets which neither completely vaporize nor burn out in the
furance. Thus, the presence of large porous carbonaceous particulates in
the emissions from an oil-fired boiler may indicate the existence of an
atomization deficiency or mismatch between the furnace burnout characteris-
tics and the type of atomizer used. Measurements of the type made in this
study might be utilized as a diagnostic procedure in situations where exces-
sive particulate emissions are encountered.
Reduced Excess Air: Referring again to Figure 1-1, the low excess air
test conditions produced a particle size distribution very similar to base-
line, but the generation of a small quantity of large particles was apparent.
Chemical analysis showed a measurable increase in carbon compared to base-
line measurements. Nevertheless, it appears that the particulate matter is
fundamentally similar in size to that produced under baseline conditions and
that carbon soot generated during low excess air operations does not neces-
sarily produce larger size particles. This conclusion may not be applicable
to extreme low excess air conditions (smoking) where large carbonaceous
particles might be expected. Unlike poor fuel atomization, increases in-
stack opacity were noted during reduced excess air conditions.
Summary: A summary of the overall particulate matter emission rates and
composition for the three Ravenswood Unit 10 test conditions is shown in
366
-------
Figure 1-5. Compared to baseline, increased carbon accounts for the slight
increase in emissions under low excess air conditions. Higher emission rates
with poor fuel atomization result from increases in both carbon and sulfate.
It is speculated that the large surface area associated with the atomization-
derived carbonaceus particles provide sites for deposition of additional sul-
fate compounds.
Comparison of Boiler Designs
The particulate emissions size distribution for the three test boilers
under full load baseline operation are shown in Figure 1-6. These results
indicate that both the face-fired and steam sendout boilers exhibit a hi-
modal distribution with a large size particle peak similar, but lesser in
magnitude, to that obtained from the tangentially fired boiler under reduced
atomization quality conditions. Photomicrographs of the large particles from
both the face-fired and steam sendout boilers show a sponge-like appearance
typical of the atomization derived particulates from the tangentially fired
unit. Chemical composition of these particulates were also similar to those
of the reduced atomization tests. Thus, both of these units appear to have
deficient atomization and/or furnace design characteristics less conducive to
carbon burnout. For the steam sendout unit this conclusion is further rein-
forced by visual flame observations which indicate the presence of numerous
large burning droplets throughout the furnace.
The baseline particulate mass emission rates and composition for three
boilers are summarized in Figure 1-7. Compared to federal New Source Perform-
ance Standards of 0.03 lb/106 Btu, the emission rates are generally low, re-
flecting the low ash, low sulfur properties of the fuel. The differences
among the three units can be attributed mainly to differences in particulate
carbon content. Reduction in carbon emissions from the face-fired and steam
sendout boilers is presently being pursued.
Polycyclic Organic Matter
POM emissions were measured at Ravenswood Unit 10 during the three pre-
viously discussed test conditions: baseline, low excess air and poor atomi-
zation. Although POM analyses must he interpreted with caution due to sampl-
ing and analytical uncertainties, emission rates were small, varying from
.0000013 to .000013lb/106 Btu. No definitive correlation with boiler operat-
ing condition was established. Up to 15 individual POM species were detected.
CONCLUSIONS
Based on the results of this program, the particulate matter emissions
from oil-fired boilers (at least firing fuels similar to those tested) can
be expected to exhibit 1) relatively narrow submicron size distributions
with a mass mean diameter in the 0.1 to 1.0 micrometer range and 2) a potent-
ial for the addition of a large particle distribution, depending upon fuel
atomization quality and boiler design characteristics which affect particle
burnout. The large particles generated as a result of these latter affects
have a distinctly different structure and composition than the submicron
367
-------
size particulates. The large particles are porous, roughly spherical and are
composed principally of carbon with a surface layer of sulfates. This is in
contrast to the submicron particles which appear to be generally spherical
solids of sulfate, carbon and trace amounts of fuel metals.
These test results also indicate the value of using particulate matter
size distribution measurements as a diagnostic indicator of potentially re-
solvable emissions problems. However, substantial work remains to be per-
formed to further differentiate between operational and design'generated
emissions, to quantify fuel and additive influences and to integrate parti-
culate, flue gas and ambient properties into an opacity model.
368
-------
FIGURE 1-1 PARTICLE SIZE MASS DISTRIBUTIONS
Ravenswood Unit 10 - Comparing Baseline, Low Excess Air and
Deteriorated Fuel Atomization Quality
/ A A
/ (11 3. 27) (243. 28)
Baseline 0
Low Excess Air
Atomization Quality
(Low Pressure
Cascade Impactor)
Particle Diameter ~ Microns
FIGURE 1-2 TYPICAL SEM PHOTOMICROGRAPH
Of Baseline Particulate Matter Illustrating
Relatively Uniform Submicron Size Particles
(Ravenswood Unit 10)
369
-------
FIGURE 1-3
FIGURE 1-4
TYPICAL SEM PHOTOMICROGRAPH
Of Large Carbonaceous Particles Resulting From
Poor Fuel Atomization
(Ravenswood Unit 10)
1100 microns!
CHEMICAL DEPTH ANALYSIS
Of Large Atomization Derived Particle
c
s
0
Na
,\_,_
........ .-»,••-
Symbol Element Mult. Factor
Na Sodium 10
O Oxygen 10
S Sulfur 5
C Carbon 1
- — <*• Time
\\
\xv
••^A^^jy^otts,**^^
I I I
Surface
,O1 02 .03
Distance from Surface (microns)
370
-------
FIGURE 1-5 COMPARISON OF PARTICULATE MATTER
Emission Levels and Composition Under Baseline,
Low Excess Air And Deteriorated Fuel Atomization Conditions
(Ravenswood Unit 10)
i .05
i
EPA Method 5
• Carbon
D SO.
• Remaining
D Not Analysed
for Composition
Cascade Impactor
m
Baseline Operat
U
Atomizalion Quality Series
FIGURE 1-6 COMPARISON OF PARTICULATE SIZE DISTRIBUTION
Normal Baseline Operation For Three Oil Fired Utility Boilers
I.016H
I" S .012 -
i!
£ f .008
<
x
• Ravenswood 10
- Arthur Kill 20
- East River 114
(Low Pressure
Cascade Impactor}
Particle Diameter ~ Microns
371
-------
FIGURE 1-7
COMPARISON OF BASELINE PARTICULATE MATTER
Emission Levels and Composition
For Three Oil-Fired Boilers
EPA Method 5
@ Carbon
El SO.
• Remaining
D Not Analysed
for Composition
Cascade Impactor
u-> en
Ravenswood Unit 10 Easl River Uml 114 Arthur Kin Unit 20
372
-------
A CONTINUOUS REAL-TIME FARTICULATE MASS
MONITOR FOR STACK EMISSION APPLICATIONS*t
By: James C. F. Wang
Combustion Research Division
Sandia National Laboratories, Livermore, CA
Harvey Patashnick and Georg Rupprecht
Rupprecht and Patashnick Company
Englewood, CO
ABSTRACT
The mass loading and size distribution of particulate emissions from
industrial stacks are critical parameters which the EPA uses for air pollution
monitoring and control. Commercially available instruments for detection of
particulates are either optical types, which do not measure particle mass
directly, or sampling types which do not provide real-time measurements. Last
year we demonstrated the feasibility of using the tapered-element oscillating
microbalance (TEOM) for real-time particle mass measurements at room tempera-
tures. Currently, we are developing a new TEOM for high temperature applica-
tions (up to 300 C). Other features of this new real-time mass monitor
include periodic backflush for long-duration continuous operations and an
interface to a multi-staged cyclone-train for real-time aerodynamic particle^
size distribution measurements. Fractional particulate mass loadings, at 50%
cut size of 1, 3 and 10 ym, are obtained using the modified cyclones in the
Source Assessment Sampling System (SASS). Test data on the performance of
this new TEOM/cyclone system are reported.
INTRODUCTION
The mass loading and size distribution of particulate emissions from
industrial stacks are the critical parameters which the EPA uses for air
pollution monitoring and control. Presently, commercial particle detectors
are either optical devices which do not measure particle mass directly, or
physical sampling devices which do not provide real-time measurements.
Furthermore, the accuracy of these measurements suffers from many inherent
shortcomings in the instruments when they are adapted from the laboratory
bench to an industrial stack. For example, the optical properties of the
particulate matter are required for optical techniques, but in a stack
exhaust, these properties are generally not known a priori and many also
^Presented at the Third Symposium on the Transfer and Utilization of
Particulate Control Technology, Orlando, Florida, March 9-12, 1981.
tThis work supported by the Department of Energy, the Office of Fossil Energy.
373
-------
change in time. Maintaining a clean window for optical access is also a very
difficult task.
Physical sampling techniques, on the other hand, involve sample with-
drawal by a probe, with consequent dilution or cooling of the sample, or both.
Also, the particle size in the sample may change because of agglomeration and
condensation of gaseous species during dilution or cooling. Furthermore, the
particle analysis is usually performed off-line and may be biased or dubious
because of difficulties associated with redispersion of the particles in the
analyzer.
Recently, at Sandia National Laboratories we have demonstrated the
feasibility of using the tapered-element oscillating microbalance (TEOM) for
real-time particle mass measurements at room temperature.(1) This mass
monitor measures the actual weight of the particles collected from a sampled
gas stream in real. time. Thus, measurement biases of the type encountered in
conventional physical sampling techniques are minimized.
To further improve the measurement accuracy of the physical sampling
technique, we are developing a new TEOM for high-temperature applications (up
to 300°C) in which the sampled gas can be kept at the stack gas temperatures
during measurements. Furthermore, the new TEOM is designed to be interfaced
to a multistaged cyclone train (one TEOM per cyclone) for real-time aero-
dynamic particle size fraction measurements. This new TEOM/cyclone system has
been designed, laboratory tests on various components have been conducted, and
preliminary results on these tests are reported in this paper.
NEW TEOM/CYCLONE SYSTEM
The new TEOM detector operates on the same principle as that of the room
temperature device reported previously.(1) Figure 1 shows the schematic of
the TEOM detector. Briefly, the tapered fiber element is clamped rigidly at
the wide end with high-temperature cement. A dust collection cup is mounted
rigidly on top of the narrow end of the tapered element, which is set in
oscillation between a set of electrically charged plates. The oscillation of
the fiber element is monitored by an LED and a photo-transistor at 90° to the
direction of the oscillation. Because of the temperatures encountered, the
LED and photo-transistor are placed in a box outside the detector assembly.
Two high-temperature optical fibers are used to transmit the LED signal and
to carry the oscillation signal to the photo-transistor, respectively. The
oscillation signal from the photo-transistor is then amplified and fed to the
conductive surface of the TEOM fiber. This feedback arrangement maintains the
fiber oscillation, whose natural frequency will change in relation to the mass
deposited on the dust collection cup. The sensitivity and frequency can be
chosen at will by proper dimensioning of the oscillating tapered element.
Three cyclones used to aerodynamically separate particles in the sampled
stream were originally part of a Source Assessment Sampling System from Acurex
Corp. They are modified here to interface with TEOM detectors for real-time
particle fraction measurements. The 50 percent cut sizes of these cyclones at
room temperature and atmospheric pressure are 1, 3, and 10 pm . With
modifications to accommodate the TEOM detectors and changes in operating
374
-------
conditions, the 50 percent cut sizes and collection efficiency curves of these
cyclones are expected to change.(2) A calibration test for the cyclones with
modifications for high-temperature and high-pressure conditions is planned.
Figure 2 shows the interface between the 3 ym cyclone and the TEOM
detector. A funnel is placed under the dust outlet of the cyclone to guide
the dust onto the dust collection cup. Our flow visualization tests indicate
that a gap of less than 6 mm between the bottom of the funnel and the cup is
desirable to ensure complete dust collection on the cup. A high-speed air
jet is designed to blow the dust out through the side holes periodically to
clean the TEOM detector. Currently, we are testing this concept for self-
cleaning operation and have not finalized the design yet.
Figure 3 shows the schematic of the total assembly of the TEOM/cyclone
system. An oven housing the cyclone train and TEOM detectors is capable of
keeping the internal temperature at levels up to 230C. A secondary sampling
line in the exhaust of the third cyclone is designed to be connected to a
fourth TEOM detector, which will be operated with a positive suction filter
attached to the tapered element to measure particles less than 1 ym. However,
this secondary sampling line is currently not implemented.
Because the cyclone cut size depends on the sampling velocity, in our
preliminary tests we decided to sample the gas at a fixed flow rate for con-
sistent particle size-fraction measurements. This introduces a nonisokinetic
sampling error which must be accounted for in the data interpretation.
PRELIMINARY TEST RESULTS
In laboratory tests of the TEOM/cyclone system, conducted on the atmos-
pheric combustor exhaust simulator (ACES) (3) at Sandia, fly ash from the
Exxon pressurized Fluidized Bed was injected into the ACES by a cyclone-type
fluidized bed dust feeder. The weight loss of the dust feeder was measured
in real time by an electronic scale and was related directly to the dust
loading in the sampled stream. The rate of dust injection was controlled by
the fluidizing air velocity in the dust feeder and could be changed within a
few seconds. With this feeder assembly, the dust loading density in the test
section of the ACES can be varied between 0 and 15 g/nr*. The fly ash used
was classified at a mean diameter of 5 ym and a standard deviation of 1.5 ym.
An original SASS train sampling system from the Acurex Corporation was
used, including sampling probe, cyclones, and oven. Only the 3 ym cyclone
was modified to have a TEOM detector coupled at its dust outlet. Regular dust
collectors were used for the 10 and 1 ym cyclones. The sampling flow was
established with a vacuum pump and measured with a mass flow meter. Because
of the size distribution of the fly ash, most of the particles were collected
by the 3 ym cyclone. The accumulated weight of the dust collected at the 3 ym
cyclone was measured in real time by the TEOM detector and compared with the
weight loss of the dust feeder.
The sampling system was tested in three temperatures, 20, 130 and 190 C.
During each test, the dust feeding rate was changed to test the time response
of the TEOM detector. At the end of each test, the dust collection cup of the
375
-------
TEOM detector was removed and weighed to compare the total dust weight with
that from the TEOM frequency shift measurement. Close agreement of the total
dust weight with that measured by the TEOM detector was obtained in every test;
the discrepancy between the two was always less than 2 percent. However, the
calibration factor which converts the frequency shift measured by the TEOM
detector to the dust weight must be obtained in the actual measurement envir-
onment to achieve this accuracy.
Figure 4 shows typical real-time measurements of the dust collected at
the 3 ym cyclone compared with the weight loss of the dust feeder at room
temperature. At the end of the test, the total weight loss from the feeder
and the total weight gain from the dust collected on the TEOM were compared,
and the ratio of these two final weights yielded a constant factor used to
scale both measurements for comparison throughout each test. It is note-
worthy that this scale factor changes for tests performed at different condi-
tions and/or arrangements. Moreover, this scale factor, SF, is proportional
to (1) the ratio of the total flow rate, QT, in the test section and the
sampling flow rate, Qg, and (2) wall losses of sampled particles in the
sampling probe and cyclones, WL-
QT WL
s, = ^a^>
where WQ = weight gain on the TEOM dust collection cup. If one has fixed
total and sampling flow rates, the scale factor provides a direct measurement
of wall losses of the sampling system.
Figure 4 shows close agreement between the dust collected at the TEOM and
the weight loss in the dust feeder as a function of time. The same slope
changes were obtained from the TEOM output at the instant the dust feed rate
changed. Similar results were obtained at 130 and 190 C, as shown in
Figures 5 and 6, respectively. The closeness of the agreement shown in these
figures indicates that the TEOM/cyclone arrangement works properly and is
capable of providing real-time particle size fraction measurements at condi-
tions which simulate power plant exhaust stack environments.
Because only one scale factor was used in each test, the results shown in
Figures 4 through 6 indicate that the ratio of dust collected on the TEOM cup
to the wall loss of particles in the sampling probe and cyclones is constant
throughout each test. If particles come off from walls of the sampling probe
and cyclones during operation, the TEOM detector should pick up these
particles and the weight gain should deviate from the weight loss curve of
the dust feeder. Similarly, if the particle wall loss is not related by a
constant factor to the particles collected, the TEOM output will not closely
follow the weight loss curve of the feeder. However, there may be an incu-
bation time required for the particle deposition on the wall to reach a
saturation or equilibrium state. We did not test the TEOM/cyclone system
long enough to observe these conditions.
Figure 7 shows the effect of temperature on particle wall loss, which is
calculated from Eq. (1) based on the measured flow rates, Qx and Qg, and the
TEOM weight gain, Wc. In the present sampling system, it is evident that the
376
-------
particle wall loss is reduced as the temperature of the sampled gas is
increased. This result can be attributed to the fact that the fly ash used
becomes less sticky at higher temperatures. It is also an important reason
to insist that stack sampling be performed at the characteristic temperature
of the stack gas for minimizing measurement errors.
CONCLUSIONS
The new TEOM detector has been demonstrated to measure particle mass in
real time at stack gas temperatures (up to 200 C). The combination TEOM and
cyclone also provides aerodynamic particle size-fraction measurements. The
importance of sampling at stack gas temperatures and in real time has also
been clearly shown. A complete real-time TEOM/cyclone particle size-fraction
monitor is being assembled and is scheduled to be tested in the near future.
ACKNOWLEDGEMENTS
The authors wish to thank E. Porter for his help in performing the
experiments and M. A. Libkind for many helpful discussions. This work is
sponsored by the Office of Fossil Energy, U.S. Department of Energy, and is
a part of the Sandia National Laboratories Diagnostics Assessment for
Advanced Power Systems Program.
. ' ENDNOTES
1. Wang, J.C.F., H. Patashnick, and G. Rupprecht, "A New Real-Time Isokinetic
Dust Mass Monitoring System," APCA Journal. Vol. 30, No. 9, September 1980,
pp 1018-1021.
2. Parker, P., R. Jain, S. Calvert, D. Drehmel, and A. Abbott, "Particle
Collection in Cyclones at High Temperature and Pressure," submitted to
Environmental Science and Technology, October 15, 1980.
3. Wang, J.C.F., "Cotnbustor Exhaust Simulation Facilities," Sandia Combustion
Research Program Annual Report 1979, UC-96, Sandia National Laboratories,
Livermore, CA, pp 15-16.
377
-------
FUNNEL
u>
•vj
oo
CYCLONE
DUST COLLECTION
CUP
FIELD
PLATES
OPTICAL
FIBER
TAPERED ELEMENT
CONDUCTIVE
PATH TO FIBER
///////
SIDE VIEW
TOP VIEW
DATA
(i PROCESSING r
Figure 1. Schematic of the High Temperature TEOM Detector
-------
Cyclone
Clean-Out Jet
Dust Collection Cup
TEOM Fiber
Clean-out
Vent
Figure 2. Schematic of the Interface Between Cyclone and TEOM
379
-------
OJ
oo
o
DATA LOGGER/CONTROLLER
Temperature
Pressure
Mass
Alarms
L
TEOM Clean-Out
Controls
TEOM Controls
Flowmeter
Flowmeter
Sample Outlets
Figure 3. Schematic of the Real-Time Particle Mass Monitor
-------
250
1-0
200 -
iy 111 150
>u-
100
IS
TEMP. @20°C
DUST FEEDER
TEOM
10
15 20
TIME (MIN.)
25
30
100 _
90 _|
80 "5
70 uJ LU
60 £
50
40
30
20
10
0
Figure 4.. Real-Time Test Results of the TEOM/Cyclone
System at 20°C
T= 130"C
— DUST FEEDER
• TEOM
10 15
TIME (MIN.)
Figure 5. Real-Time Test Results of the TEOM/Cyclone
System at 130°C
381
-------
"S
3 70
S 60
§ 50
o
LL
O
<& 40
I 30
£ 20
§
2 10
T=190°C
DUST FEEDER
TEOM
i i i i I i i i i I i i i i I i i i i I i i i
5 10 15 20
120
110
100
90
80
70
60
50
40
30
20
10
25
o -
TIME (MIN.)
Figure 6. Real-Time Test Results of the TEOM/Cyclone
System at 190°C
15
10
0
50
100
150
200
250
Temperature (°C)
Figure 7. Effect of Sampling Temperature on Particle Wall Loss
382
-------
STUDIES OF PARTICULATE REMOVAL FROM DIESEL
EXHAUSTS WITH ELECTROSTATIC AND ELECTRO-
STATICALLY-AUGMENTED TECHNIQUES
By: James L. DuBard, M. Greg Faulkner, Jack R. McDonald
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35255
Dennis C. Drehmel, James H. Abbott
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
ABSTRACT
Laboratory experiments on a 5.7 liter GM diesel truck have established
the concept of removing particulate emissions from diesel exhaust by a com-
bination of agglomeration and trapping devices. A two-stage electrostatic
precipitator (ESP) is used to agglomerate the primary particulate matter,
resulting in an order-of-magnitude increase in mass median diameter. The
agglomerated particulate is characterized. Aerosol sampling data are pre-
sented for the variation in particle size distribution and the efficiency
of trapping the agglomerated particulate in cyclones, fiber filters, and a
granular bed filter. Overall mass removal efficiencies greater than 80%
have been achieved with an ESP/granular bed filter system for a duty distance
greater than 800 km (500 miles) at a constant highway speed of 88 km/hr
(55 mph). Methods of cleaning the devices and removing collected particulate
are discussed.
INTRODUCTION
This paper will discuss the application of electrostatic gas cleaning
technology to the aftertreatment of vehicular diesel exhaust. Electrostatic
precipitation has the possible advantage of high-efficiency collection of
submicron diesel particulate matter with low engine backpressure and the
expenditure of low electrical power. On the other hand, the low resistivity
or the diesel particulate matter leads to the possible difficulties of sur-
face current leakage through sooty deposits on insulators and poor particle
fdhesion £ the coflectionplates. Electrostatic gas cleaning devices were
tested on the exhaust from a 5.7 liter GM diesel pickup truck mounted on a
chassis dynamometer.
ELECTROSTATIC PRECIPITATOR
An electrostatic precipitator (ESP) was developed and tested at Southern
Research Institute in Birmingham, Alabama(1). The ESP is a cylindrical two-
staee device designed to obtain maximum collecting plate area in a small
stage device design desiened to be periodically wet-flushed in a ver-
Hrrss -vsAss »
383
-------
TO SPRAY
NOZZLES
INSULATOR
CABLE TO HIGH
VOLTAGE SUPPLY
SPRAY NOZZLES
CHARGING
SECTION
COLLECTING
SECTION
0.30 m LONG
(12 in.)
CABLE TO HIGH
VOLTAGE SUPPLY
i '
=8
( B
C E
t 3
C
0.20 n
ID (S
/
/
/
/
/
/
/
(/////// /
rt -
i
i '
.-— —
> \
a
) :
i DIA.
in. y
/
^
/
/
/
/ / / ijj
o:::
STAR-SHAPED
ELECTRODES
(DETAIL)
SUPPORT SPIDER FOR
GROUNDED CYLINDERS
SUPPORT SPIDER FOR
HIGH VOLTAGE
CYLINDERS
INSULATOR
"- SUPPORT SPIDER
OUTLET
REMOVALBE
FIBER FILTER
STORAGE TANK
PUMP
4181-299
Figure 1. Schematic diagram of the electrostatic precipitator.
384
-------
gives^an average linear gas flow down through the ESP of 3.2 m/sec. For the
charging section, five star-shaped corona electrodes are mounted on a vertical
high-voltage rod chosen for mechanical strength. The corona electrodes are
typically operated at 40 kV (negative) and about 250 uA. The product of
negative ion density and gas treatment time in the corona section is
Not * 0.6 x 1013 sec/m3,
at an electric field of approximately 4 x 10s volt/m. The electric field for
particle charging could be increased with the use of improved high voltage
insulation.
The collector section is a set of concentric cylinders (overlapping
length 28 cm), alternately grounded and biased negative. The total collecting
area is 1.18 m2 (12.7 ft2). For a volume flow of 6.2 m3/min (220 ACFM), the
specific collection area is about 11.3 m2/(m3/sec) (58 ft2/1000 ACFM). The
collector cylinders have a nominal spacing of 0.6 cm. After wet-flushing
loose particulate, the cylinders will hold about 6 kV without arcing. In
order to increase the collecting electric field, the voltage is increased
typically to 15 kV, with steady arcing carrying about 1.5 mA. Then the col-
lecting electric field is about 24 x 10s volt/m.
Early tests of the ESP produced little permanent collection of diesel
particulate. Preliminary mass train sampling (non-isokinetic) indicated an
overall mass removal efficiency in the neighborhood of 30 to 40%. Later
measurements (with cascade impactors and a backup filter) of the mass re-
moval efficiency of the ESP gave values of about 26%. While the ESP achieved
only low collection of diesel soot, there was on the other hand a substantial
agglomeration of the particles. Electrical aerosol size analyzer data showed
roughly 50 to 70% removal of particles of aerodynamic diameter on the order
of 0.1 ym. This phenomenon was easily confirmed by visual observation of
some macroscopic agglomerates. Microscopic examination of particles at the
ESP outlet showed a preponderance of very loose and fluffy agglomerates on
the order of 25 ym in actual size. Settling velocity experiments with large
agglomerates indicated that the aerodynamic diameter was roughly 1/10 the
physical particle diameter. This was consistent with later cascade impactor
measurements of the mass median diameter (mmd) at the ESP outlet. Subsequent
testing of the ESP was designed to emphasize its role as an agglomerator,
with the ESP followed by some other device for trapping the agglomerated
diesel soot.
The electrical resistivity of the diesel soot is of concern in regard
to surface conduction through sooty deposits on insulating standoffs in the
ESP. The resistivity of a b'ulk sample of diesel soot collected at the ESP
outlet was measured in moist air using an ASME PTC-28 test cell. The soot
was compacted as little as possible. With temperature and moisture content
comparable to those of the diesel exhaust gas, the resistivity of the fluffy
agglomerated soot measured about 10 ohm-cm.
Several common cleaning fluids were bench-tested for effectiveness in
cleaning the ESP. The fluids were tested both on bulk samples of diesel soot
and on soot layers impacted on metal parts. Perchloroethylene, a common dry
385
-------
cleaning fluid, was chosen for cleaning the ESP because it is effective, non-
flammable, and of relatively high boiling point and low toxicity.
The ESP was occasionally flushed after the device had cooled down over-
night. About 1 gallon (4 liters) of perchloroethylene was pumped directly
from a supply container through four spray nozzles in the top of the ESP.
The dirty fluid flowed directly from an outlet in the bottom of the ESP into
a take-up container. Teflon insulators were thoroughly cleaned, but there
always remained a thin film of soot on metal surfaces which could not be re-
moved by any cleaning fluid (without rubbing).
One problem in the aftertreatment of diesel exhaust is the collection
and handling of the large volume of low bulk density soot (roughly 120 kg/m3).
An advantage of wet flushing with perchloroethylene is that the collected soot
is left in a highly compacted state when the dirty cleaning fluid is evapo-
rated (to be followed by condensation in a closed system). The measured bulk
density of dry residue from the wet-flushing procedure was 1300 kg/m3.
ESP/CYCLONE
Three series of tests were conducted on three different ESP/cyclone
combinations, with the diesel truck running at 88 km/hr (55 mph). First, a
Fisher-Klosterman XQ-4 industrial cyclone was connected in the exhaust line
after the ESP. The cyclone was calibrated by the manufacturer under ambient
conditions to have a DSQ cutpoint of 2.0 ym for a volume gas flow of 5 m3/min
(180 ACFM). By Lapple's law, the DSQ cutpoint extrapolates to 1.6 ym under
actual operating conditions (175°C, 2.5 x 10~5 Pa*sec, 6.2 m3/min, 36.5
m/sec). Performance of the ESP/cyclone combination was tested by extracting
samples of the exhaust gas through cascade impactors. Three test stations
were used to accumulate data simultaneously at the outlets of the diesel
truck, the ESP, and the cyclone. Average results of several tests are shown
in Figure 2.
The agglomeration of carbon soot by the ESP is demonstrated in Figure 2
in that the aerodynamic mmd at the ESP outlet is increased by roughly one
order of magnitude to about 2.5 ym. (On the basis of electrical aerosol size
analyzer data combined with cascade impactor data, the mmd of the primary
diesel particulate is known to be approximately 0.3 ym.(l)) The total mass
loading of the impactor stages plus backup filter was 37-0 mg/scm at the
truck outlet and 27.2 mg/scm at the ESP outlet. The ESP alone achieved about
26% mass removal efficiency.
Figure 2 shows that about 56% of the diesel particulate at the ESP out-
let had an aerodynamic diameter greater than the extrapolated DSQ cutpoint
of the cyclone. However, the total mass loading at the cyclone outlet was
34.7 mg/scm, 28% greater than at the ESP outlet. There was evidently sub-
stantial hydrocarbon condensation in the gas stream passing through the
cyclone. Figure 2 shows also that the mmd of the diesel particulate was
greatly reduced (to about 0.9 ym) in passing through the cyclone. This was
not due to collection of large particles in the cyclone because only a very
small amount of diesel soot was retained in the cyclone catchpot. The de-
crease in mmd could have resulted both from hydrocarbon condensation and from
386
-------
100
s
o
V
w
Ui
HI
u
cc
111
a.
Ill
<
3
5
U
60
40
20
Tl
X?'"
?-'• .4
/
, • , I
J I
I I I I I
0.2
1.0
AERODYNAMIC DgQ,
10
20
4181-303
Figure 2. Impactor testing of ESP/cyclone combination cumulative percent mass of
aerodynamic diameter less than the DSQ cutpoint
387
-------
breakup of the fluffy, agglomerated soot particles in the vortex of the
cyclone, in reentrainment from the cyclone catchpot, or in the impactor jets.
The concept of an ESP/cyclone combination was tested further with a
small five-stage cyclone system designed and calibrated for in situ stack
sampling. In this case, a gas sample at the ESP outlet was extracted through
the five-stage cyclone at 170°C and 0.03 m3/min (1 ACFM). The corresponding
calibrated D50 cutpoints are 8.3, 3.9, and 2.5 ym for the first three
cyclones, and an estimated 1.1 and 0.5 ym for the last two cyclones. A
total of 136 mg of diesel particulate was collected in the cyclones (57%)
and on the backup filter (43%). The measured mass distribution is displayed
in Table 1. The five-stage cyclone data are compared with the cascade
TABLE 1. CUMULATIVE PERCENT MASS > D5o OF
DIESEL SOOT AGGLOMERATED BY THE ESP
8.3 urn 3.9 urn 2.5 urn 1.1 ym 0.5 ym
Cascade Impactors
(From Figure 2) 17% 35% 49% 61% 67%
Five-Stage Cyclone 24% 35% 44% 50% 57%
impactor data, where the cumulative percent mass of aerodynamic diameter
greater than the D50 cutpoint is extracted from the ESP OUTLET curve of
Figure 2. The five-stage sampling cyclone collects roughly the expected
amount of agglomerated diesel particulate.
Finally, tests were conducted on various combinations of sampling
cyclones, with diesel exhaust gas samples (at the ESP outlet) pulled through
the cyclones at much higher flow rates than that for which they were de-
signed. The concept under investigation was that if such a cyclone could
collect agglomerated diesel particulate efficiently at a flow rate of .55
to .85 m3/min (20 to 30 ACFM), then the total exhaust gas flow out of an
ESP agglomerator could be passed through a parallel array of perhaps 10 such
cyclones.
The tests were conducted with the diesel truck running at 88 km/hr (55
mph) and cyclone sampling gas temperatures approximately 170°C (340°F). The
cyclone combinations were tested at flow rates of .55 and .80 m3/min (20 and
28 ACFM). Particulate masses collected in the cyclone catchpot and in a
backup filter were measured. The three cyclones used were designated 9, 3,
and 4, with specifications as follows:
Cyclone 9 - Large IP (inhalable particulate) Cyclone
D50 = 15 ym at .015 m3/min, 22°C
388
-------
Cyclone 3 - Small IP Cyclone (mounted inside Cyclone 9)
D50 = 2.5 ym at .015 m3/min, 22°C
Cyclone 4 - Cyclone 4 of the Five-Stage Sampling Cyclone
D50 = 0.6 ym at .03 m3/min, 22°C
D50 - 1.1 ym at .03 m3/min, 170°C
Results of testing the cyclones at high flow rates are given in Table 2,
These tests indicate that an ESP agglomerator followed by cyclonic trapping
TABLE 2. SUMMARY OF TESTING SMALL CYCLONES
AT HIGH GAS FLOW RATES
Cyclone
9
9 and 3
9 and 3
4
Flow Rate
m3 /min
0.55
0.55
0.80
0.80
Pressure Drop
kPa (in. H20)
1 (4)
56 (14)
96 (24)
96 (24)
Mass in Catchpot
%
23
52
52
37
of agglomerated particulate is a viable concept for aftertreatment of diesel
exhaust. The mass removal efficiency of such a device is at least 50%. The
mass removal efficiency can be improved by extensive engineering and testing
to match the cyclone performance to the characteristics of the agglomerated
diesel particulate. Methods for disposal of the trapped diesel particulate
require further investigation. An ESP will require cleaning sooner or later,
by wet flushing or air jet scouring. Soot collected in the cyclone catchpot
will have to be removed and compacted. During the tests of the ESP in com-
bination with other devices, it was necessary to clean the collector cylinders
by wet flushing with perchloroethylene after 16 to 20 hours of operation with
the diesel truck running at 88 km/hr.
ESP/GRAVEL BED FILTER
The gravel bed filter, previously tested on primary diesel exhaust(2),
was connected in the exhaust line after the ESP to test the concept of using
a granular filter to trap diesel particulate that had been agglomerated in
the ESP. Three series of tests were conducted with different configurations
of the gravel bed and at different diesel truck speeds.
The ESP/gravel bed filter was tested first at a diesel truck speed of
88 km/hr (55 mph). The gravel bed consisted of a 10 cm depth of 2 mm steel
389
-------
shot supported on a screen in a drum of diameter 45 cm (18 in.)- Cascade
impactor data for the ESP/gravel bed combination are shown in Figure 3. The
combination was tested with and without stirring of the surface of the gravel
with a rake to break up the cake of carbon soot that forms on the surface.
Stirring resulted in a large puff of collected carbon soot passing through
the exhaust line. The curve labeled WITH STIRRING represents average data
for runs which included one stir in the middle of each 30 minute data col-
lection run. The curve labeled WITHOUT STIRRING represents average data for
runs taken over 3 days, with the gravel bed stirred once each day before the
beginning of testing. This daily stirring typically reduced the engine back-
pressure from a maximum value of 17 kPa (68 in. HzO) to about 12 kPa (48 in.
Had). The mass removal efficiency did not vary measurably with the change
in engine backpressure.
TABLE 3. TOTAL MASS LOADINGS MEASURED FOR
THE ESP/GRAVEL BED COMBINATION
Total Mass Loading
Sampling Point mg/scm
Truck Outlet 37.7
ESP Outlet 27.7
Gravel Bed Outlet
with stirring 20.7
without stirring 5.4
The total mass loading of the impactor stages plus backup filter is
summarized in Table 3. The ESP alone achieved about 26% mass removal effi-
ciency, as in previous cascade impactor testing of the ESP/cyclone combi-
nation. The overall mass removal efficiency, without stirring the gravel
bed, was 86%. The data in Figure 3 show that the mind of the agglomerated
diesel soot (about 3.5 ym) greatly decreases when the gas stream passes
through the gravel bed. This is the result of larger particulate being
trapped in the gravel bed.
Figure 4 shows electrical aerosol size analyzer data for the ESP/gravel
bed combination, with the diesel truck running at 88 km/hr. These data show
the effect of the treatment devices on the very fine particulate matter that
is predominant in the primary diesel exhaust. In the aerodynamic size range
around 0.1 ym, the ESP alone achieved 60 to 70% removal of particulate mass.
The addition of the gravel bed resulted in only a small improvement in mass
removal in this small size range. The main function of the gravel bed is to
trap the particulate mass which has been shifted to a larger size range by
the ESP.
390
-------
100
g
Q
v>
v>
u
cc
LU
D
80
60
40
20
~\ 1 1—I—r—r
0.2
-~1P
I I I
I t I
1.0
AERODYNAMIC DSQ, jum
10 20
4181-304
Figure 3, Impactor testing of ESP/gravel bed combination cumulative percent mass of
aerodynamic diameter less than the DSO outpoint.
391
-------
100
80
>
o
u
u.
LL.
60
DC
V)
1
40
20
0.04
<
GRAVEL BED OUTLET
ESP OUTLET
i i i I
I
0.10 0.20
GEOMETRIC MEAN DIAMETER, jum
0.40
4181-298
Figure 4. Mass removal efficiency in small size ranges of diesel paniculate.
392
-------
_ Secondly, the ESP/gravel bed combination was tested (without stirring)
with the diesel truck running at 24 km/hr (15 mph). These tests were in-
tended to simulate lower speeds and smaller displacement engines. The engine
backpressure remained steady at about 2.5 kPa (10 in. H20). With the decrease
in truck speed, the exhaust gas volume flow decreased from approximately
6.2 m /min (220 ACFM) to 2.2 m3/min (80 ACFM) at 95°C (205°F). The partic-
ulate loading at the diesel truck outlet did not remain constant, but in-
creased by roughly 75% to about 65 mg/scm. At the same time, the ESP/gravel
bed combination collected more particulate mass per unit volume of gas flow
at the lower gas flow. The overall mass removal efficiency was 83%, about
the same as at 88 km/hr.
Finally, because the ESP/gravel bed combination had demonstrated high-
efficiency cleaning of diesel exhaust gas for a practical operating time,
the gravel bed was reduced to a more practical size. A gravel bed diameter
of 20 cm was chosen, to correspond to the ID of the ESP and to achieve a
reduction in filter area by a factor of 5. The diesel truck was operated at
88 km/hr. The mass removal efficiency was tested by mass train sampling at
the diesel truck outlet and at the gravel bed outlet. Test results for
gravel beds of depth 5 cm and 10 cm are summarized in Table 4.
TABLE 4. MASS REMOVAL EFFICIENCY OF THE ESP
IN COMBINATION WITH AN 8-INCH DIAMETER
GRAVEL BED FILTER
Gravel Bed
Depth
cm
Engine
Backpressure
kPa (in. H20)
Outlet Mass Loading
Truck Gravel Bed
mg/scm mg/scm
Efficiency
5 10 (40) rising to 34.9 15.3 56
15 (60) in 3 hours
10 12 (48) rising to 37.9 11.6 69
16 (64) in 3 hours
These tests indicate that a compact gravel bed filter can be contained
within the housing of the ESP, directly below the collecting cylinders, and
can achieve mass removal efficiency of 70%, or better, at acceptable levels
of engine backpressure. The gravel bed filter can be cleaned simultaneously
with the ESP by wet flushing with perchloroethylene. The mass removal
efficiency of the ESP/gravel bed combination is largely independent of the
running speed of the diesel truck. The higher mass loadings at lower exhaust
gas flow rates are compensated by collection of greater particulate mass per
unit volume gas flow by both the ESP and the granular filter. However,
lower exhaust gas flow rates reduce the scouring of the insulating standoffs
in the ESP and lead to more frequent cleaning periods.
393
-------
ESP/FIBER FILTER
Two types of ESP/fiber filter combinations were tested. First, the
deep bed filter, previously tested on primary diesel exhaust(2), was loaded
with fiberglass batts and connected in the exhaust line after the ESP.
Then, the concept of agglomerating and trapping diesel particulate in a
single compact device, with a removable cartridge filter, was tested with
an electrified filter(1). This two-stage device was designed to transform
high axial gas flow past a disk electrode in the corona charging section
into a much lower radial gas flow in the collector section. The inner and
outer sections of the collector assembly were loaded with various types of
fiberglass filter media. The device had both electrical and mechanical
forces acting to collect diesel soot.
The performance of these devices exhibited the behavior typical of all
fiber filters tested: high mass removal efficiency (60 to 70%) but a life-
time of at most a few hours due to rising engine backpressure. The elec-
trified filter experienced also deterioration of electrification because
of diesel soot buildup in the electrified fiberglass filter. Two general
conclusions derive from a variety of laboratory tests involving the after-
treatment of diesel exhaust with fiber filters: (1) high collection effi-
ciencies and short operating periods will result from a dense cake of carbon
soot forming on the surface of the filter material and causing a rapidly
increasing pressure drop; and (2) exhaust gas leakage around the edges of
the filter material, resulting from high pressure drop, will be a problem
in the utilization of any replaceable cartridge filter.
CONCLUSION
Conventional electrostatic precipitation is effective in the agglomer-
ation of diesel particulate matter. An ESP can be used with either cyclones
or a granular filter to obtain mass removal efficiency in the range of 50 to
85% at moderate engine backpressure and for a practical operating time
between maintenance periods. Such a combination device can be made suf-
ficiently compact to be used on stationary diesel engines or on large, heavy-
duty diesel vehicles engaged in fleet operation where periodic maintenance
can be performed at a central garage.
ENDNOTES
1. M. G. Faulkner, E. B. Dismukes, J. R. McDonald, D. H. Pontius, and A. H.
Dean, "Assessment of Diesel Particulate Control: Filters, Scrubbers,
and Precipitators," EPA-600/7-79-232a (NTIS No. PB80-128655), U.S. EPA,
IERL, Research Triangle Park, North Carolina, October 1979.
2. M. G. Faulkner, J. L. DuBard, J. R. McDonald, D. C. Drehmel, and J. H.
Abbott, "Studies of Particulate Removal from Diesel Exhausts with
Mechanical Techniques," at Third Symposium on the Transfer and Utili-
zation of Particulate Control Technology, Orlando, Florida, 1981.
394
-------
STUDIES OF PARTICULATE REMOVAL FROM DIESEL
EXHAUSTS WITH MECHANICAL TECHNIQUES
By: M. Greg Faulkner, James L. DuBard, Jack R. McDonald
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35255
Dennis C. Drehmel, James H. Abbott
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
ABSTRACT
A series of tests are discussed which were designed to characterize the
collection of particulate emissions from diesel exhaust by several different
mechanical methods. The source of particulate emissions is a 5.7 liter GM
diesel truck mounted on a chassis dynamometer. The control devices which
are discussed include fiber filters, gravel bed filters, and trap/cyclones.
The overall mass collection efficiencies, fractional mass collection effi-
ciencies, and operating characteristics of these devices were determined by
measurements of inlet and outlet total mass loadings and particle size
distributions.
A device containing a fiber filter was investigated with three different
filter materials: long fiber glass wool, batts derived from fiberglass
insulation, and stainless steel fiber mats. Collection efficiencies as
high as 90% were achieved, coupled with a quick pressure rise culminating
in gas sneakage. A device containing a gravel bed of 45.7 cm diameter and
10.2 cm depth was investigated with 2 mm steel shot. Efficiencies ranged
from 45 to 70% and increased with increasing system backpressure.
INTRODUCTION
The superiority in fuel economy of the diesel engine compared to a
gasoline engine of the same displacement has created a widespread market
lor diesel powered light duty vehicles due to the increasing cost of fuel
In addition, automobile manufacturers view the diesel engine as a possible
olufion to'the Federal Government's regulation requiring that the average
fuel mileage for each auto manufacturer be at least 11.6 km/1. (27.5 ml/gal.)
by the yeaf 1985. Unfortunately, one characteristic^ the Diesel engine is
a high level of combustion by-products, the most obvious of which is
particulate matter*.
The particulate effluent of the
median diameter (mmd) of the particles
*EPA has defined particulate matter as .everything collected on a filter at
52°C a25°F) after dilution with ambient air in a dilution tunnel.
395
-------
ranges from 0.3 to 0.5 ym, with about 70% occurring at diameters less than
1.0 ym(l). The small size of the particulate places it in a range where the
collection efficiency of both filtration and electrostatic devices is re-
duced. The problem of particle collection is magnified by the large
quantities of particulate produced, amounts ranging from 0.14 g/km for the
Volkswagen Rabbit to 0.53 g/km for the General Motors 5.7 liter engine(2).
The ultimate goal for particulate emissions set by EPA is 0.125 g/km(2),
which would require that 76% of the particulate matter from the GM 5.7 liter
engine be removed from the exhaust stream. Over a 8000 km (5000 mi.)
service interval, this would amount to about 3.2 kg of particles. Using
120 kg/m3 for the value of the bulk density of this collected material(1).,
this corresponds to about 27 liters (7 gal.).
Southern Research Institute (SoRI) has been studying non-regenerative
aftertreatment devices. These devices include both devices which have been
specifically developed for light duty vehicles and stationary-source devices
which have been adapted to vehicular sources. The devices which were tested
as part of this study can be classified as either electrostatic or purely
mechanical in nature. The mechanical devices include a deep bed filter(3)
which can accommodate up to a 15 cm depth of fibrous filter material, a
granular bed filter, and a barrier filter. Also in this catagory are con-
densation traps developed by Eikosha Co. (Japanese Aut-Ainers)(3) which
depend on gas stream cooling to condense particles on a relatively coarse
wire mesh. Data on the mechanical devices are presented following a brief
description of the test layout and particle sampling techniques used. Data
on the electrostatic devices are presented elsewhere(4).
TEST CONDITIONS
The diesel exhaust used in the control device tests was furnished by
GM 5-7 liter diesel engines. At the SoRI facility in Birmingham, Alabama,
a 1979 Chevrolet diesel pickup truck mounted on a chassis dynamometer pro-
vided the exhaust. For tests performed at the Department of Transportation
(DOT), Transportation Systems Center, Cambridge, Massachusetts, an engine
mounted in one of the engine dynamometer test cells was used.
The majority of the tests were run with the truck operating at a steady
road speed of 88 km/hr (55 mph) in order to provide a high load for the
control device. At DOT the engine was operated at 1700 rpm to achieve this
same condition. This provided a gas flow rate of about 7.6 actual m3/min
(270 acfm). For lower speed runs, the truck was operated at 24 (15) or 32
km/hr (20 mph). These speeds corresponded to flow rates of 2.4 and 3.2
m3/min, respectively. The other engine speed used at DOT was 900 rpm
corresponding to 48 km/hr (30 mph). Total mass and particle size distri-
bution were measured before and after each control device to characterize
the particle collection of the control device. The total mass loading in-
formation was obtained by using Gelman 47 mm stainless steel filter holders
with glass fiber filters.
The particle size distribution data were obtained using modified
University of Washington Mark V cascade impactors having seven stages with
stainless steel inserts coated with Apiezon H grease to prevent particle
396
-------
bounce. These instruments were used to obtain particle size distributions
on a mass basis over the size range from about 0.2 to 4 ym. The impactors
were either mounted in ovens close-coupled to the exhaust pipe or were
wrapped in heating jackets which allowed them to be operated at the same
temperature as the exhaust gas.
In addition to the cascade impactors, a Thermo Systems Model 3030 Elec-
trical Aerosol Size Analyzer (EASA) was used to determine concentrations and
electrical mobility derived size distributions of particles in the 0.01 to
0.3 ym size range. A Climet Model 208 optical particle counter was used to
monitor concentrations in the fine particle size range. The SoRI SEDS III
sample extraction, conditioning, and dilution system was used as an inter-
face between these instruments and the exhaust stream. The system removes
condensable vapors from the sample gas at elevated temperatures followed by
controlled dilution to particle concentrations within the operating ranges
of the measurement instruments. The optical data were not used for detailed
sizing information but provided quick indications of efficiency in several
size bands.
DEEP BED FILTER
The first device to be fabricated and tested was the deep bed filter
shown schematically in Figure 1. The dimensions of this device are based on
a theoretical study which predicts 82% collection of diesel particulate
when used with a 10 cm thick fiber mat of 10 ym fiber diameter and a porosity
of 0.99(3). This device has a very large face area due to the downward gas
flow. This geometry results in a greatly reduced flow velocity through the
filter mat which enhances particle capture by the filter fibers and reduces
reentrainment. In addition, should the filter be subject to blinding, the
large face area should increase the lifetime of the device. The geometry
would also be convenient for use with interchangeable filter elements.
The filter box was fabricated at Southern Research Institute for use in
a joint testing program with the Automotive Research Laboratory of the DOT
Transportation Systems Center in Cambridge, Massachusetts. A special stain-
less steel fiber mat was ordered from Bekaert Steel Wire Corp. in Belgium.
However, when it became apparent that this material would not arrive in time
for the first DOT test, screening tests were conducted at SoRI to determine
substitute materials. Fiberglass batts derived from roll insulation and
long fiber spun glass (angel hair) both demonstrated efficiencies in the
80-90% range and were used in the DOT test. The fiberglass batts had a
fiber diameter of about 15 ym and, for the sample tested, a porosity on the
order of 0.99. However, since the material was very compressible, the
porosity was highly variable. The angel hair had a 20 ym fiber diameter and
an undefinable porosity since it came compressed in 0.45 kg (1 Ib) bags and
could easily be pulled apart with the fingers.
The tests were run at engine speeds equivalent to 48 and 88 km/hr (30
and 55 mph). The pressure drop across the control device was monitored and
the testing was discontinued if the pressure became too high or rose quickly
and then leveled out or dropped, an indication that the gas was finding an
alternate path through or around the filter material.
397
-------
Initially, the deep bed filter demonstrated about 90% collection effi-
ciency at 88 km/hr and 95% efficiency at 48 km/hr using 15 cm of fiberglass
batts for the filter material. These values were degraded, however, as the
pressure drop across the filter rose and gas sneakage was induced. On-site
equipment modifications reduced the amount of sneakage present but were in-
capable of completely eliminating it. Figure 2 shows the pressure drop across
the filter for one of the tests. In this example the pressure curve starts
low, begins to rise quickly as the cake builds up, and then starts to level
off. This pressure behavior proved to be typical for every run in the test
series. Examination of the filter material showed some discoloration in the
layers of fiberglass which implies that the particles were sifting through
the mat. The front surface was very black and showed signs of building a
cake of particulate. The edges, however, were also very black, which implies
the presence of gas sneakage, an effect that would explain the leveling out
of the pressure curve of Figure 2. Collection efficiencies measured after
the pressure drop curve had leveled off were quite low. Figure 3 shows the
collection efficiencies calculated from data taken early and late in the
same test shown in Figure 2. The overall efficiency for the early data is
89%. The data taken late in the test, after the backpressure had leveled
off, showed a 6% efficiency.
The behavior of the spun glass in the deep bed filter was very similar
to that of the fiberglass batts. Initial efficiencies using about 15 cm of
the material were in the 80-90% range for the 88 km/hr equivalent engine
speed. This material appeared to be less susceptible to sneakage than the
fiberglass batts. This is probably due to the less rigid form of the spun
glass which should allow the material to flow to fit the container. Two
runs were made with this material. In the first test, the pressure curve
started to break over, an indication of the onset of sneakage. In the
second run with spun glass, no break in pressure rise was evident. This
run was terminated due to high backpressure (23 kPa or 92 in. HaO). The
efficiencies were essentially the same for both runs as the measurements
were made before the onset of sneakage in the first run.
The second test with DOT was held in May 1980. During this test, the
deep bed filter was run using the stainless steel fiber mat which had been
ordered from Belgium for the November 1979 test. This material was composed
of 12 urn stainless steel fibers which were spun into a 0.64 cm (0.25 in.)
thick mat which was then sintered on both sides to give it some rigidity.
The calculation called for a 10 cm thick bed but only 2.5 cm (four layers)
were used in these tests. The first test was run at 88 km/hr equivalent
speed with an upward gas flow to determine if the soot could be dislodged
from the face of the filter mat by rapping the filter box. However, the
test was cut short when the backpressure rose from 0.9 to 1.4 kPa (3.6 to
5.6 in. HaO) in 25 minutes and then leveled out for the duration of the 1
hour test. After the test, the filter box was opened and the filter
material was examined. There was obvious soot trails on the sides of the
fiber bed indicating severe sneakage. There was no evidence of soot between
the layers. The soot cake on the first surface of the fiber mat was very
dense and tenacious. It could not be shaken off of the surface. Attempts
to remove the cake by using a backflow of compressed air were only slightly
successful but the compressed air permanently compressed the filter material.
398
-------
Successful cake removal could only be achieved by scraping the face of the
filter. The cake appeared denser here than it did with the fiberglass batts
used in the November test but this is probably due to the smooth surface of
the stainless steel. A retest of this device using a downward gas flow
resulted in mechanical failure: pressure on the filter collapsed the
support screen in the filter box.
GRAVEL BED FILTER
The next device examined was a gravel bed filter. It was anticipated
that the gravel bed would agglomerate the particulate matter so that it
could be collected in a large particle separator such as a cyclone. A
gravel bed offers the advantages of limited dust capacity (which should
promote fast filling and subsequent reentrainment) and rugged filter
material (which will allow stirring or other agitation of the filter bed
to break up clogging and induce loss of collected particles). The device
was constructed from a removable lid type 114 liter (30 gal.) drum having
a 45 cm (18 in.) inside diameter. A platform covered with a steel screen
supported the bed, which consisted of industrial grade steel shot. An
exhaust pipe was welded to the side of the barrel near the bottom and
another was welded to the lid. A stirring device consisting of a rake
whose teeth extended 3 to 4 cm into the shot bed was added later in the
test.
Initial tests were conducted with an upward gas flow to allow obser-
vation of the bed while the truck was running. The initial collection
efficiency using 5 cm of 2 mm shot measured 22%. After 6 hours of operation,
the backpressure dropped and localized boiling of the shot was observed. The
efficiency was then 5%.
At this point the system was replumbed to a down draft. The bed was
changed to 10 cm of 2 mm shot and was covered by a screen. The results of
this test are shown in Figure 4. After 6 hours of operation, the unit was
opened and inspected. Some rearrangement of the shot had occurred due to
the direct blast from the inlet air. Consequently the screen was removed,
the shot bed was smoothed out, and the screen was recut and replaced. It
was observed at this point that, although the particulate had dispersed
itself through the shot bed fairly evenly, it was also piling up on the
top surface. This surface deposit was easily broken up when the bed was
smoothed out and releveled. This mild stirring of the shot caused the
break in the AP curve shown in Figure 4. However, the pressure quickly
resumed its upward trend and reached 27 kPa (108 in. H20) after 10 hours
of operation. Inspection again showed a thick surface layer of particulate.
The collection efficiency measured at the end of the test was 72%.
At this point the system was modified to remove the screen on top of
the bed, install a baffle so the gas stream would not dig holes, and install
a rake so the top 3 cm of the shot could be stirred to break up the top
layer. By stirring at 1 hour intervals, the pressure drop could be con-
strained to the 10-12 kPa (40-50 in. H20) range with efficiencies of about
40%. The emissions increased briefly each time the bed was stirred. Al-
though the ratio of large to small particles increased considerably during
399
-------
the stirring puff, no appreciable agglomeration effects were observed
during the non-stirring portion of the operation cycle. Therefore the
gravel bed filter must be considered a particle collector rather than an
agglomerator.
AUT-AINER
The Aut-Ainer, made by Eikosha Co. in Japan, is being developed
specifically as a diesel exhaust control device. Figure 5 is a conceptual
sketch of an Aut-Ainer with three bands of stainless steel mesh for soot
collection. This mesh is a coarsely woven belt of 0.2 x 0.4 mm flat wire
which has been rolled around the central tube and sandwiched between per-
forated plates. The device depends on cooling condensation for collection
of particulate. The cooling is supplied by the central tube which acts as
a ram air tube. Heat transfer is augmented by several perforated plates
located across the gas flow and connected to the cooling tube. The exhaust
gas flow is expected to tear agglomerates off of the wire mesh and swirl
them through the cyclone to a catch bag on the side of the device, re-
sulting in an effectively self-cleaning device. The only maintenance
required would be periodic changing of the soot collection bag.
Three models of Aut-Ainer were tested. The first was the one shown in
Figure 5. The first two bands of steel mesh are 5 cm thick and the third
is 2.5 cm thick. The mixing of the cooling air with the exhaust at the end
of the Aut-Ainer initially caused some concern as the effects on the con-
densable hydrocarbons in the exhaust stream were unknown. This difficulty
was overcome by extending the cooling air tube down the center of the
sampling port tube so that undiluted exhaust could be sampled. After this
section, a baffle was placed in the exhaust pipe to promote mixing and
another sampling tube was placed downstream to allow sampling of the diluted
exhaust. The amount of dilution present was determined by examining the con-
centrations of CO, C02, and NOX present before and after dilution and
deriving a correction factor based on the assumption that the quantities
of these gasses were invariant.
The first Aut-Ainer was examined during the May, 1980, DOT test.
According to the manufacturer, who was at DOT for the first week of testing,
the Aut-Ainer requires about 12 hours of running-in time to obtain maximum
efficiency. After 20 hours of running at 1700 rpm (88 km/hr equivalent),
an efficiency of 12% was obtained for this speed. The manufacturer also
pointed out that the device was designed for an engine in the 2-3 liter
displacement range rather than the 5.7 liter GM engine being used. By
scaling speeds and displacements, 48 km/hr (900 rpm) on the 5.7 liter
engine can be considered the equivalent of the desired test speed of 88
km/hr for a 3 liter engine. Therefore a test was run at 900 rpm which
yielded an efficiency of 32%. The efficiencies quoted are overall figures
derived from filter and impactor data. Figure 6 shows size dependent effi-
ciencies for both engine speeds. Ultrafine data indicate that the device
was condensing but not collecting large quantities of small particles.
The second Aut-Ainer tested was designed after the DOT test by K. Aoi
of Eikosha Co., who believed that a larger design would be more suitable
400
-------
for an engine of 5.7 liter displacement. This device is essentially the
same as the one shown in Figure 5 except that the three sections of mesh
are 18 cm long for a total of 54 cm and that there is a bleed-off pipe
connected to the back of the cyclone. The outlet used a double pipe which
carried the cooling air past the sampling ports before allowing mixing. At
the suggestion of Aoi the catch bags were enclosed so that the exhaust
through the bags could be examined. The flow rates at the inlet and regular
outlet were close to 7.6 actual m3/min (270 acfm). The flow through the
bags barely registered on the water manometer used with the pitot tube.
Based on an estimated 0.13 cm (0.05 in.) deflection, a flow rate of 0.9
actual m /min (31 acfm) was obtained. This is only 14% of the total flow.
The large capacity Aut-Ainer was tested at SoRI using the conditions
for which it was designed, i.e., a 5.7 liter vehicle operating at 88 km/hr.
Filter data taken at the regular gas outlet indicated that the device was
collecting 35% of the particulate after 40 hours of operation. EASA data
taken at this point showed a negative efficiency for particles of less than
0.7 ym diameter, which suggests that this device, like its smaller counter-
part, was condensing vapors to form small particles.
The filter samples taken from the exhaust through the catch bags were
pale yellow. This implies that only negligible quantities of soot passed
these bags which would indicate a high efficiency if opacity measurements
were made. However, these filters showed a larger weight gain than the
very black filters taken from the regular exhaust. The collection efficiency
through the bags measured 22% with negative efficiency again occurring for
the small particles. The temperature of the gas at this point was 84°C
(184°F) compared to 132°C (271°F) at the regular outlet and 174°C (345°F)
at the inlet. Since the temperature was less than 100°C, the filters were
desiccated for several days and reweighed. The changes in weight were
within the error of the weighing procedure which suggested that the weight
gain was due to condensed hydrocarbons. Gas chromatography data taken from
the residue collected on the filters support this by showing a distinct
shift toward lighter, lower condensation temperature organic compounds in
the sample taken from the catch bag exhaust filter compared to a sample
taken from the regular outlet filter.
A different Aut-Ainer designed by S. Masuda was also tested. This
device has no air pipe but is designed to maximize its surface area for
cooling purposes. The device is 5 cm thick by 40 cm wide and contains
three 27 cm long sections of the same steel mesh. There is no cyclone in
this device and therefore any agglomerates which are reentrained should
be discharged through the exhaust outlet. According to Masuda, the device
was built for a 1 liter engine having a maximum gas flow rate of 3.5 m /min.
Consequently, the truck was run at 32 km/hr which gave a gas flow of about
3.2 mVmin.
After 20 hours of operation the efficiency was 36% with a backpressure
of less than 3 kPa. EASA data indicate that the device is not agglomerating
at all but is producing a condensation fog of very small particles.
401
-------
BARRIER FILTER
The barrier filter tested was a ceramic monolith made by Corning. The
device was run at an engine speed equivalent to 48 km/hr (30 mph) and showed
efficiencies of 92% using impactors and 97% using the electrical aerosol
analyzer. However, its lifetime is very limited due to the rapidly rising
pressure drop which reached 35 kPa (140 in. H20) in 7 hours.
CONCLUSION
The use of a fine fiber filter to capture diesel soot does not appear to
be practical due to the tendency toward formation of a cake on the face of
the filter. This cake of particles reduces gas flow through the filter which
results in an increase in the pressure drop across the device. The pressure
continues to increase until either it impedes the operation of the engine or
a lower pressure route around the filter mat is established. Once a route
for gas sneakage is established, the collection efficiency of the device
drops sharply.
The three Aut-Ainers tested were unable to achieve acceptable collection
efficiency levels. In addition, there was no evidence that these devices
were agglomerating particulate. Instead, they produced a fog of small par-
ticles as hydrocarbon vapors present in the diesel exhaust were condensed in
the control device.
The granular bed filter is a concept which may prove applicable to the
diesel exhaust problem. Although the collection efficiency was not as high
as that of a fiber filter, the gravel bed has the important advantage of
being cleanable, whether by mechanical or chemical means. The ceramic
filter demonstrated a high collection efficiency but has a limited lifetime.
This device is undergoing development by EPA at Ann Arbor, Michigan, and by
several automobile manufacturers to extend its usable lifetime by re-
generation.
ENDNOTES
1. McCain, J. D., and M. G. Faulkner. Assessment of Diesel Particulate
Control: Particle Size Measurements. EPA-600/7-79-232c (NTIS PB80-
224256), U.S. Environmental Protection Agency, IERL, Research Triangle
Park, North Carolina, December 1979.
2. Federal Register, 44 (33) February 1, 1979.
3. Faulkner, M. G., E. B. Dismukes, J. R. McDonald, D. H. Pontius, and
A. H. Dean. Assessment of Diesel Particulate Control: Filters,
Scrubbers, and Precipitators. EPA-600/7-79-232a (NTIS PB80-128655),
U.S. Environmental Protection Agency, IERL, Research Triangle Park,
North Carolina, October 1979.
4. DuBard, J. L., M. G. Faulkner, J. R. McDonald, D. C. Drehmel, and J. H.
Abbott. Studies of Particulate Removal from Diesel Exhausts with
Electrostatic and Electrostatically-Augmented Techniques, at the Third
Symposium on the Transfer and Utilization of Particulate Control
Technology, Orlando, Florida, 1981.
402
-------
0.10 m
EXHAUST GAS
OUTLET BELOW
FILTER
RECTANGULAR
FIBROUS FILTER
ELEMENT
EXHAUST GAS
INLET ABOVE
FILTER
4181-317
Figure 1, Deep bed filter.
20 40 60 80 100 120 140 160 180 200 220 240 260 280
TIME, minutes ini-284 A
Figure 2. Pressure drop across fiberglass batt
403
-------
1 WW
o
rff
OC
i 1°
a.
t—
z
111
o
OC
Ul
a.
1
£ '"• i r- 1 i w i| • i — i~n~nw^i i > i i i 1 1 1-
: * ••••, :
00 •«
^o o •••
o
0 0
°°00OOC>
o -
0
o
- °o -
O IMPACTOR BEFORE BREAKDOWN °OQ0
A EASA
• IMPACTORS AFTER BREAKDOWN
• EASA AFTER BREAKDOWN
I I 1 1 1 1 1 ll 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
u.u
(J
Ul
O
tL
90.0 £
l-
ui
o
OC
Ul
a.
99.0
10-2
10-1 100
PARTICLE DIAMETER, jum
101
4181-315
Figure 3. Efficiencies before and after onset of sneakage for deep bed filter
using fiberglass batts.
30
25
ui
cc
=) « 20
^ S
oc
B.
2
15
10
5
0
120
100
! 80
60
40
20
0
STIRRING OF GRAVEL
I
I
I
I
4567
TIME, hours
8 9
1.0
0.75
>-
o
10
UJ
O
0.50 E
0.25 •"
4181-316
Figure 4. Gravel bed fitter efficiency and backpressure.
404
-------
AIR COOLING —
EXHAUST IN
STAINLESS STEEL MESH'
TT
I F
:^-j l i
EXHAUST OUT
Figure 5. Aut-Ainer.
SOOT CONTAINER
4181 -2 97A
100
o
10
z
ui
u
oc
UJ
a.
.
-
™
«•
-
m
m
P m ,
r l i II
0 88 kph
A 88 kph
• 48 kph
• 48 kph
i i l l 1
T - a°
T.i|* iAo9 . Md^Q bo^ t
9 *• A OJRXrV!)
••0 0^0 o^iir-
—
IMPACTORS :
EASA -
IMPACTORS
EASA
ill) i i l l l l 1 1 1 l 1 1 1 1 111
0.0
>
O
UJ
O
LL>
90.0 uj
g
UJ
u
K
UJ
a.
99.0
10-2
10-1
ioo
101
PARTICLE DIAMETER,
4181-314
Figure 6. Aut-Ainer efficiency for 88 and 48 km/hr speeds.
405
-------
UPDATE ON STATUS OF CONNECTICUT'S CONTROL PROGRAM
FOR TRANSPORTATION-RELATED PARTICULATE EMISSIONS
By: H. Ledger Chamberlain
Environmental Programs Department
Northeast Utilities
Berlin, CT
John H. Gastier
Bureau of Planning and Research
Connecticut Department of Transportation
Wethersfield, CT
ABSTRACT
Connecticut's State Implementation Plan (SIP) indicated transportation
related sources contribute more than half of total suspended particulate
(TSP) emissions.
Our earlier paper, presented at the 1979 Symposium, discussed implica-
tions of this finding and identified additional efforts needed in two areas:
measures to reduce and control emissions; and programs to better define and
document the transportation sector's contribution to TSP emissions.
Control programs include both exhaust controls for exhaust emissions and
fugitive controls for reentrained particulate material.
The success and cost of these controls are of interest both as means of
achieving air quality standards and as a potential source of reductions for
"offset" and "banking" programs. Potential interaction of mobile and
stationary source controls exists.
Connecticut's unique program is presented. The results of extensive
testing to further define transportation-related TSP emissions in Waterbury,
Connecticut are summarized and preliminary conclusions regarding control
strategies are discussed.
TEXT
At the last symposium we presented a paper C-' which briefly described
the TSP situation in Connecticut and then concentrated on the role which
transportation played in bringing about that situation. In it we detailed
the NAAQS compliance status of the state, the mode and extent of the trans-
portation contribution to ambient TSP levels, and measures being taken to
ameliorate the situation. Also presented were reservations about the
technical analyses upon which the official description of the situation was
based. This present paper updates progress which has been made and indicates
areas where additional effort is required.
On the basis of readings from the statewide HiVOL monitoring network
operated by the Connecticut Department of Environmental Protection, the city
of Waterbury has been designated as noncompliance for the primary TSP
406
-------
standard. A secondary standard noncompliance designation has been applied
to the entire state. <••'' Despite this, tremendous progress has been made in
the past decade. The number of violations of the primary twenty-four hour
standard has decreased from twenty-one in 1971 to one in 1979. During the
same time period the number of secondary violations has decreased from
thirty-four to seven. The severity of the violations has decreased as well.
This improvement has been due in large part to a vigorous program of
stationary source control.
The goals of Connecticut's particulate control program may be defined
as: 1) achieving the primary standard in Waterbury; 2) achieving the
secondary standard statewide; and 3) development of an accommodative SIP
strategy. An accommodative SIP is one which will decrease pollutant levels
so that they are demonstrably below the standards, thus leaving a margin to
accommodate new growth.
The entire Connecticut portion of the New Haven-Hartford-Springfield
interstate Air Quality Control Region was originally given a primary non-
compliance designation, based on monitoring at the Waterbury permanent
monitoring site. Subsequent examination of TSP levels at other sites in the
region indicated that the primary nonattainment designation should be limited
to the city of Waterbury. (2) The Waterbury monitoring site is strongly
oriented to transportation sources, being located within an interchange of
Interstate Route 84, immediately south of the highway mainline. The
Connecticut Route 8 expressway is a short distance to the west of the site.
The Waterbury Central Business District (CBD) lies to the north of the site
and 1-84.
The results of a number of analyses indicate that the high TSP levels
encountered at the Waterbury site are the result of a one time situation.
The apparent cause was a large construction project, consisting of widening
and realigning 1-84 immediately east of the monitoring site. This project
involved a massive rock "cut and borrow" operation which began in November,
1976. Total completion of the project is scheduled for 1981. Although all
legal controls and numerous special measures were employed there is
tremendous potential for fugitive emissions from such operations as blasting,
loading, on-site rock crushing, storing, unloading, and placement of material
In many instances the material must be handled many times. The following
table shows quantities of material removed (in thousand cubic yards) from
the 1.2 mile project for each year of its duration. For reference, the
TABLE 1. MATERIAL QUANTITIES REMOVED FROM 1-84 REALIGNMENT PROJECT
Year
1975
1976
1977
1978
1979
1980
Earth
(1000CY)
0
111
312
167
186
4
Rock
(1000CY)
0
56
406
261
31
6
Total
(1000CY)
0
167
718
428
217
10
Total
Peak Mo. Interchange Site
(1000CY) [TSP]24 [TSP]A
0
83(Dec)
80 (May)
68 (July)
62(0ct)
6 (Jan)
180
256
255
249
174
84.7
86.5
81.3
80.0
69.6
Typical Site
[TSP]24 [TSP]A
173
135
147
237
147
65.5
60.1
70.0
62.3
49.8
407
-------
annual geometric mean and second high twenty-four hour TSP concentrations
(in micrograms per cubic meter) from the Waterbury interchange site and a
more typical Waterbury site are included.
A separate, but complementary, study was conducted by CONNDOT in an
attempt to quantify particulate emissions from various construction activities.
A HiVOL sampler was placed approximately twenty-five feet from the operations
being monitored. Typical TSP concentrations are included in the following
tabled3)
TABLE 2. TSP CONCENTRATIONS IN VICINITY OF SELECTED CONSTRUCTION OPERATIONS
Operation [TSP] (ug/m3)
Rock Crushing/Trucking 1000
Paving 150 - 200
Hauling (Dry) 500
Hauling (Damp) 70
While the precise contribution of the 1-84 project to ambient TSP levels
cannot be determined, the information presented above shows that a very large
source of particulates existed in Waterbury for more than four years. Traffic
was maintained on 1-84 and most local streets throughout the construction
period. Thus a mechanism was provided for reentrainment and transport of
particulate matter throughout the area, as was reflected in TSP concentrations
at the interchange monitoring site.
For a short time in 1979 the DEP operated a network of TSP monitoring
sites encompassing the entire Waterbury area. Results of the monitoring
indicate that the extremely high TSP levels measured at the interchange site
are not representative of the city as a whole. We believe that the site's
proximity to 1-84 (it is a Zone A site), to the realignment project, and to
the Waterbury CBD, resulted in the monitoring of a non-typical TSP episode
With termination of construction activity the cause of the episode will have
been eliminated. Reclassification of the city as primary attainment should
then be possible, based on three consecutive years of monitored data.
Demonstration of secondary standard compliance and, thereafter, an
accommodative growth margin will require reduction of monitored TSP concentra-
tions at other sites more typical of the character of Connecticut than
Waterbury. However, there are two possible sources of error to be encountered
in determining the state's actual and equitable control requirements. They
are passive sampling error and macroscale atmospheric transport.
Passive sampling error is the result of sampling methodology. As is
standard practice, DEP operates its HiVOL samplers every sixth day, with the
monitors being tended only once during each sampling cycle. In some critical
areas samplers are operated every third day. During the inactive portion of
the sampling cycle particles may collect on the filters, introducing a
positive bias to the readings. This is known as passive sampling and the
error introduced can be significant. Studies by DEP have shown that it can
exceed twenty percent and is largely dependent on TSP concentration. Recent
work by Sweitzer indicated average passive loading of 102 tests at two rural
408
-------
Illinois sites as 16.3% and 17.2%. (4)
The DEP has begun using HiVOL samplers equipped with devices which cover
the filter during passive periods. Data from these samplers show a decrease
in ISP levels. The exact magnitude cannot be determined because of normal
year-to-year variation and the lack of colocated monitors using a normal
operating schedule. However, DEP has estimated a correction factor of 13%(2)
The implications of correcting for or eliminating passive sampling error
are obvious. Application of DEP's correction relationship to Connecticut's
monitor readings eliminates several of the secondary standard violations and
reduces the primary violation to a secondary violation.
Efforts to account for macroscale atmospheric transport have thus far
been minimal, having been almost exclusively devoted to oxidants and acidic
precipitation precursors. These efforts must be expanded to include partic-
ulate transport. This is an especially critical issue for Connecticut
because of its proximity to the New York City metropolitan area. Computer
modeling and monitored data both provide evidence that a significant portion
of the particulates in Connecticut were transported from the New York area.
The magnitude of the transport phenomenon is surely debatable, but its
implications are obvious. Connecticut must implement controls which must be
more severe than otherwise and must compensate for emissions over which it
has no control. This imposes a competitive economic disadvantage on
Connecticut's industries and businesses. These disadvantages would be
compounded by a cut-off of federal funds should the NAAQS not be attained.
Consideration of passive sampling and transport will reduce the magnitude
of our task. However, considerable disagreement remains concerning the
portion of Connecticut's particulate emissions attributable to transportation
and what measures should be undertaken to control them.
According to the emissions inventory prepared by DEP for the 1979 SIP
revisions transportation sources were responsible for fifty-three percent of
Connecticut's particulate emissions in 1976.<-2) Assumed emission factors were
used to calculate the emissions from transportation sources, i.e., motor
vehicles. These emission factors consist of three components:
- exhaust or tailpipe emissions;
- tire wear; and
- reentrained road dust.
Tailpipe emissions consist of combustion products, primarily carbon
particles and large aggregates of heavy hydrocarbon molecules. Lead-containing
particles are emitted by vehicles burning leaded gasoline. Tire wear emissions
are produced by friction between tires and the roadway surface. They consist
of particles of rubber and other tire constituents. Reentrained road dust
consists of particulate matter which is forced into suspension by vehicle-
induced turbulence and natural air currents. The particles may be of any
origin.
409
-------
The emission factors used in inventory calculations are contained in
the following table.
TABLE 3. VEHICULAR TSP EMISSION FACTORS
Tailpipe Tire Wear Reentrainment Total
Year (gm/mile) (gm/mile) (gin/mile) (gm/mile)
1976 0.37 0.20 0.68 1.25
1982 0.25 0.20 0.68 1.13
1987 0.21 0.20 0.68 1.09
Vehicular emission factors are most commonly expressed as mass emitted per
unit of distance traveled, usually grams per mile. The decrease in the
tailpipe emission factor over time is due primarily to the phase-out of
leaded gasoline.
We continue to have serious reservations concerning both the form and
magnitude of these emission factors. All of the emission components are
constant, rather than varying with the major factors which influence them.
Our previous paper discussed these influencing variables and reviewed
programs needed to better define vehicular emissions and measures to control
them.
Much is already known about the age and type mix and the distribution
of significant operating parameters of the vehicle population in Connecticut.
This information is used to calculate emissions of other pollutants. What is
not known is the effect that all of these variables have on the three TSP
emission components. In most cases direct or inverse proportionality may be
deduced, but the magnitude of the effects is unknown. Controlled testing
and monitoring of all three emission components is necessary in order to
establish emission factors upon which to base control strategies.
However, this is not enough. A significant portion of the emitted
particulate mass consists of particles large enough to settle out of
suspension rapidly, and thus cannot contribute to ambient TSP concentrations.
Typical particle size/density distributions must therefore be developed.in
order to account for particle settling.
A reentrainment emission factor of 5.15 gm/mile has been recommended
by EPA. To its credit, the DEP correctly assumed that it is too large if
particle settling were not to be taken into account. It was assumed that a
portion of the emissions was suspendable and the remainder was not. The
suspendable emission factor of 0.61 gm/mile was developed using a simplified
half-life analysis. The error inherent in this approach is that particle
size is a continuum, not divisible into two portions at a discrete separa-
tion point.
Much study is required to determine typical particle size/density
distributions which will complement the emission factors for each component,
particularly reentrainment. From this information a model can be developed
which will account for dispersion as well as particle settling.
410
-------
Shortly after^ur previous paper was presented, both the Federal
Highway Administration (FHWA) and the EPA indicated that they were planning
to conduct programs aimed at better defining particulate emissions from
transportation sources. Connecticut opted to await the results of these
efforts rather than undertake the considerable expense of conducting dupli-
cate studies. Unfortunately, both federal agencies cut back or eliminated
these programs due to budget restrictions, leaving much to be accomplished.
In our previous paper we outlined several projects being conducted or
proposed in the Waterbury area in an attempt to gain a better understanding
of the area's TSP situation. Two of these are worth noting.
ConnDOT is participating in an interagency demonstration program
examining the relationships between air quality and economic development.
Agencies on the state, regional, and local level are taking part. An
interesting concept being investigated is the use of transportation-related
pollutants, particularly TSP, as offsets and/or banking credits for use by
stationary sources. An example is a business gaining a TSP offset by imple-
menting TSP-reducing transportation, such as vanpools, for its work force.
A further example is implementation of less polluting transportation measures
by a political body, thereby establishing "banked" emission rights which
could be used to entice industry to locate or expand within the area. Such
concepts may well be of interest in shaping the efforts of both stationary
and transportation sources to achieve a common goal.
The other project is a monitoring study conducted by DEP at the inter-
change monitoring site in Waterbury. Daily TSP concentrations were measured
for a period of several months, together with corresponding meteorological
parameters and traffic counts on three streets in the vicinity of the
monitor. Some data has been obtained from this study, but little analysis
has been conducted as yet. However, preliminary investigation indicates
that there is a positive proportionality between traffic volume and TSP
concentration. The correlation however appears to be extremely weak. We
hope that more detailed analyses of these data will provide some insight
into the influence of vehicular traffic on TSP concentrations.
Connecticut's future programs for control of transportation-related
particulate emissions will be influenced by several perspectives which have
changed since our last paper was presented. We believe that monitoring
results, when completely reviewed, will show that Connecticut is in compli-
ance with the primary TSP standard statewide. However, attainment of the
secondary standard must be achieved as expeditiously as possible. A com-
prehensive program for achieving the secondary standard will be formulated
for the 1982 SIP revisions. Control efforts required of Connecticut will
be less severe, as accounting for passive sampling error and interstate
transport becomes more widespread.
As discussed earlier, the results of several studies indicate that
Connecticut's primary nonattainment area (Waterbury) was the result of close
proximity of a large one-time construction project to a sensitive monitoring
site. Areas where monitored data indicate secondary standard violations will
be carefully analysed to determine sources of ambient particulates and
411
-------
appropriate control measures. Again, an accommodative SIP is a state goal.
The efforts needed to define vehicular emission factors for particulates
have been discussed. Our decision-makers can only interpret the decision
that previously scheduled federal programs were expendable as an indication
that our earlier concerns were an over-reaction. We have already urged that
one of the two federal agencies involved reconsider their cancelled programs
because the scope of the problem is not limited to Connecticut. The task is
also too complex and expensive for a single state to address adequately. We
reiterate our plea that programs to quantify vehicular emissions be reinstat-
ed with the j3£me priority as originally intended.
Efforts will be devoted to developing methodologies for analysing TSP
in EIS preparation and in the state's Indirect Source Permit program. Here
again, results from federal programs will be sought because more compelling
problems will preempt available resources.
Attention will be given to various scenarios which are at present
unpredictable. Many may force us to completely re-evaluate our present
course. One such scenario is the adoption of new TSP standards for respir-
able particulates. Another could be an extensive amendment of the Clean
Air Act by this session of Congress.
Finally, several trends have recently been taking shape which cause us
concern. Significant increases in the use of coal and wood for home heating
may increase TSP levels throughout the state. This could require compensating
reductions in emissions from the transportation sector. The increased
ambient TSP levels would also result in additional particulates settling
onto roadways and being reentrained, thus compounding the problem.
Concern for fuel conservation has prompted a significant increase in
the use of Diesel passenger vehicles. General Motors is projecting that
twenty percent of its 1985 automobile sales will have Diesel engines. The
particulate emissions from these vehicles are many times those of conventional
vehicles. This trend is likely to have an adverse effect on progress toward
attainment of Connecticut's particulate air quality goals.
References
1. Gastler, J.E., and Chamberlain, H.L., Status of Connecticut's Control
Program for Transportation-Related Particulates, Second Symposium on the
Transfer and Utilization of Particulate Control Technology,
EPA-600/9-80-039D, September, 1980.
2. Connecticut Department of Environmental Protection, State Implementation
Plan for Air Quality, June 22, 1979.
3. Connecticut Department of Transportation, A Study to Determine
Particulate Concentrations From a Portland Cement Concrete Pavement
Recycling Program, August, 1980.
4. Sweitzer, T.A., Characterization of Passively Loaded Particles on HiVOL
Samples, JAPCA, 30: 1324-1325, December, 1980.
412
-------
AUTHOR INDEX
AUTHOR NAME
Albrecht, P.R.
Anderson, M.H.
Arce-Medina, E.
Ariman, T.
Armstrong, J.A.
Bakke, E.
Balfour, W.D.
Bamberger, J.A.
Bergmann, L.
Berlant, MJ,
Bernstein, S.
Beutner, H.P.
Bickelhavspt, R.E.
Boericke, R.R.
Bohn, R.
Borenstein, M.
Brookman, E.T.
Bump, R.L.
Bush, P.V.
Calvert, S.
Carries, D.
Carr, R.C.
Chamberlain, H.L.
PAGE
IV-116
11-334
11-76
III-290
IV-188, IV-252
1-236
III-119
III-398
1-323
11-218
11-405
111-71, III-228
1-165
III-353
IV-344
111-90
IV-125
11-425
1-157
III--1, 111-10, IV-156
IV-13 5
1-118
IV-406
413
-------
AUTHOR INDEX (cont.)
AUTHOR NAME . PAGE
Chambers, R. 1-45
Chiang, T. III-250, III-261
Chou, K.H. IV-73
Cowen, SJ. IV-264
Crippen, L.K. 1-148
Crowson, F. III-438
Crynack, R.R. II-242
Czuchra, P.A. IV-55
Dalmon, J. H-390
Demski, R.J. 1-341
Dennis, R. 1-1, IH-140
Dietz, P.W. III-449, III-459
Donovan, R.P. 1-11
Drehmel, D.C. III-341, IV-210
DuBard, J.L. IV-383
Durham, M. H-54, IV-285
Ensor, D.S. 1-176, IV-242
Eskinazi, D. , III-238
Faulkner, M.G. 11-199, IV-395
Feldman, P.L. IV-3
Ferrigan III, JJ. 1-197
Finney, W.C. 11-358
Fjeld, R.A. 11-179
Fortune, O.F. 414 1-82
-------
AUTHOR INDEX (cont.)
AUTHOR NAME PAGE
Frazier, W.F. III-171
Gardner, R.P. III-128
Gaunt, R.H. 1-216
Gehri, D.C. 1-333
Gentry, J.W. III-406
Giles, W.B. III-468
Hardison, L.C. 111-33
Harmon, D.L. IV-317
Hawks, R.L. III-221
Helfritch, D. 1-75
Henry, F. III-301
Henry, R.F. IV-63
Hesketh, H.E. IV-222
Hoenig, S.A. III-382
Hovis, L.S. 1-23
Hyde, R.C. 1-129
lionya, K. HI-181, IH-321
Jaworowski, R.J. 1-185
Jensen, R.M. 1-138
Joergensen, H.J. 11-370
Johnson, C.A. !-352
Kalinowski, T.W. III-311
Kanaoka, C. HI-280
Kirstein, B.E. 415 III-373
-------
AUTHOR INDEX (cont.)
AUTHOR NAME PAGE
Kolnsberg, HJ. IV-179
Krishnamurthy, N. IV-232
Ladd, K. 1-55, 1-65
Lagarias, J.S. 1-272
Landham, Jr., E.G. 1-237
Langan, W.T. III-211
Lawless, P.A. 11-25, 11-35, 11-44
Leith, D. III-270
Leonard, G.L. 11-120
Maartmann, S. 11-130
Mahoney, D.F. 1-206
Mappes, T.E. Ill-ISO
Martin, D. IV-145
Masuda, S. 11-189, 11-380
Mathai, C.V. IV-200
Mazumder, M.K. 11-160, 11-169
McCrillis, R.C. IV-306
McElroy, M.W. 1-94
McLean, K.J. 1-265, 11-304
Menegozzi, L. 11-404
Menoher, C. Ill-Ill
Mitchner, M. 11-97
Moore, W.E. IV-105
Mormile, D. , IV-363
416
-------
AUTHOR INDEX (cont.)
AUTHOR NAME PAGE
Moslehi, G.B. 11-109
Mosley, R.B. n-1, 11-13
Musgrove, J.G. III-193, III-201
Noonan, P.M. IV-326
Oglesby, H.S. JII-80
Ostop, R.L. 1-107
Parker, R. 111-51, IV-2
Parquet, D. III-363
Parsons, Jr., E.L. 1-303
Pattern, J.D. HI-160
Pearson, G.L. 1-120
Pedersen, G.C. 111-60
Petersen, H.H. 1-291
Piulle, W. 1-253
Potokar, R.W. III-417
Prem, A. HI-21
Presser, A.M. IV-26
Pyle, B.E. II-66
Raemhild, G.A. 11-349
Reardon, F.X. III-102
Rimberg, D.B. 11-262
Rinaldi, G.M. IV-95
Rinard, G. H-283, 11-295
Rubow, L.N. IV-83
-------
AUTHOR INDEX (cont.)
AUTHOR NAME PAGE
Rugg, D. ll~213
Samuel, E.A. Il~U9
Schliesser, S.P. 11-252
Semrau, K.T. I11'43
Shilling, N.Z. n-230
Smith, W.B. x-96
Snaddon, R.W.L. IV~74
Sparks, L.E. 11-314, 11-326
Spawn, P.D. IV-335
Starke, J. HI-428
Stevens, N.J. I"313
Sullivan, K.M. H-141
Tatsch, C.E. IV-353
Teller, A.J. III-393
Thompson, C.R. H-415
Urone, P. IV-275
VanOsdell, D.W. 1-35
Viner, A.S. IV-168
Wakabayashi, A. III-332
Wang, H.H. IV'36
Wang, J.C.F. IV-373
Wegrzyn, J. IV-46
Weyers, L.L. 1-226
Wilks, W.H. 418 IV-15
-------
AUTHOR INDEX (cont.)
AUTHOR NAME PAGE
Williamson, A.D. IV-297
Yamamoto, T. H-87
Yung, S. IV-1, IV-155
Zarfoss, J.R. "-208
*USGPO: 1982—559-092/0432
419
------- |