&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

-------