United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
EPA-600/7-79-244a
November 1979
EPA Alkali Scrubbing
Test Facility:
Advanced Program,
Fourth Progress Report;
Volume 1. Basic Report
Interagency
Energy/Environment
R&D Program Report
.
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
$WlW^
(v/^ Repository Material EPA-eoo/7-79-244a
^T Permanent Collection November 1979
^
^^
EPA Alkali Scrubbing Test Facility:
Advanced Program, Fourth Progress Report;
Volume 1. Basic Report
'Env'iTonmentel
\0
2
-JQ CO
o £
c o
CD
by
Harlan N. Head and Shih-Chung Wang
Bechtel National, Inc.
50 Beale Street
San Francisco, California 941 19
Contract No. 68-02-1814
Program Element No. EHE624
s 18/9
"^ -2 >
.
Q
to <
3 Q_
CTLU
3
CD
X
-
(§
CM
EPA Project Officer: John E. Williams
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, NC 2771 1
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
NOTICE
This report was prepared by Bechtel National, Inc. as an account of work spon-
sored by the Environmental Protection Agency (EPA). Neither the EPA nor Bechtel,
nor any person acting on behalf of either:
a. Makes any warranty or representation, expressed or implied, with respect
to the accuracy, completeness, or usefulness of the information contained
in this report, or that the use of any information, apparatus, method, or
process disclosed in this report may not infringe privately owned rights;
or
b. Assumes any liabilities with respect to the use of, or for damages re-
sulting from the use of, any information, apparatus, method or process
disclosed in this report.
ii
-------
ABSTRACT
This Fourth Progress Report presents the test results from late-November
1976 through June 1978 of the Advanced Test Program on a prototype lime/
limestone wet-scrubbing test facility for removing S02 and particulates from
coal-fired boiler flue gases. The test facility is located at TVA's Shawnee
Power Station, Paducah, Kentucky. Tests were conducted on two parallel scrub-
ber systems: a venturi/spray tower (35,000 acfm P 300°F), and a Turbulent Con-
tact Absorber or TCA (30,000 acfm @ 300°F).
Forced oxidation with two scrubber loops was developed for the venturi/spray
tower system. This system was successfully demonstrated with limestone, lime,
and 1imestone/MgO slurries. An open pipe air sparger discharging into an
agitated tank was successfully used to oxidize the slurry. Oxidized slurry
solids have improved dewatering and handling properties.
Attempts to oxidize the venturi/spray tower bleed stream were successful only
when magnesium ion was present to buffer the slurry pH and to increase the
liquor sulfite concentration for oxidation.
Forced oxidation with limestone slurry in a single scrubber loop was demonstra-
ted on the TCA system. In this system, more efficient oxidation was achieved
with an air sparger than with an eductor.
Additional testing on the TCA system included limestone testing with low fly
ash loadings, limestone type and grind testing, automatic limestone feed control
testing, limestone reliability testing with high fly ash loadings, limestone
testing with Ceilcote egg-crate type packing, lime and limestone testing with
added MgO, and flue gas characterization testing to determine inlet and outlet
participate mass loadings, size distribution, and SOg concentration.
Auxiliary studies during this test period included laboratory reactivity
studies of various limestone types and grinds, development of pressure drop
correlations for both scrubbers, trace element analyses in the slurry liquor,
and waste solids dewatering and characterization studies.
m
-------
ACKNOWLEDGMENTS
This report was prepared as a team effort by the following Bechtel personnel:
Dr. Harlan N. Head, Project Manager
Dr. Shin-Chung Wang, Technical Manager
A.M. Abdulsattar D.Y. Kawahara
D.A. Burbank, Jr. T.M. Martin
G.A. Dallabetta R.R. McKinsey
C.L. DaMassa D.T. Rabb
D.G. Derasary L.S. Reider
J. Hing C.H. Rowland
Dr. Gary Rochelle of the University of Texas has acted as technical consultant
on this project.
The "Fifth TVA Interim Report of Corrosion Studies: EPA Alkali Scrubbing Test
Facility", which is reproduced in Appendix K, was prepared by G.L. Crow and
H.R. Horsman of TVA's Division of Chemical Development at Muscle Shoals, Alabama.
Acknowledgment and appreciation are extended to the TVA staff at the Shawnee
Test Facility and to TVA's Emission Control Development Projects Group at Muscle
Shoals, Alabama, who are responsible for operation, maintenance, and engineering
modification of the facility.
Special appreciation is expressed to John E. Williams, EPA project officer, who
has guided this program from its onset.
-------
CONTENTS
Notice
Abstract
A cknowledgments
Illustrations
Tables
Section
1 SUMMARY
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
1.10
1.11
1.12
1.13
1.14
1.15
Two-Scrubber-Loop Forced Oxidation Test Results
on the Venturi/Spray Tower System
Bleed Stream Oxidation Test Results on the
Venturi/Spray Tower System
Limestone Test Results on the TCA with
Low Fly Ash Loading
Limestone Type and Grind Test Results on the
TCA System
Test Results of Automatic Limestone Feed Control
on the TCA System
One-Scrubber-Loop Forced-Oxidation Limestone Test
Results on the TCA with Air Eductor
Limestone Reliability Test on the TCA with High
Fly Ash Loading in the Flue Gas
Limestone Test Results with Ceil cote Egg-Crate
Type Plate Packing on the TCA
One-Scrubber-Loop Forced-Oxidation Limestone Test
Results on the TCA with Air Sparger
TCA Limestone/MgO and Lime/MgO Test Results
Flue Gas Characterization Test Results
Laboratory Limestone Reactivity Test Results
Pressure Drop Correlations for Spray Tower and TCA
Analysis of Trace Elements
Results of Waste Solids Dewatering and
Characterization Studies
11
iii
iv
xii
xv
1-1
1-3
1-6
1-8
1-9
1-10
1-11
1-13
1-15
1-16
1-17
1-19
1-22
1-23
1-24
1-25
-------
Section Page
1.16 Operating Experience 1-32
2 INTRODUCTION 2-1
3 TEST FACILITY 3-1
3.1 Scrubber Selection 3-1
3.2 System Description 3-6
3.3 EPA Pilot Plant 3-9
4 ADVANCED TEST PROGRAM 4-1
4.1 Advanced Test Program Objectives and Completed 4-1
Schedule
4.2 Analytical Program 4-2
4.3 Data Acquisition and Processing 4-5
5 VENTURI/SPRAY TOWER TWO-SCRUBBER-LOOP FORCED-OXIDATION 5-1
LIMESTONE TEST RESULTS
5.1 System Description 5-3
5.2 Discussions of Test Run Results 5-10
5.3 General Operating Characteristics of the Two- 5-22
Scrubber-Loop System with Forced Oxidation
5.4 Summary of Findings 5-24
6 VENTURI/SPRAY TOWER TWO-SCRUBBER-LOOP FORCED-OXIDATION 6-1
LIMESTONE/MgO TEST RESULTS
6.1 System Description 6-3
6.2 Discussions of Test Run Results 6-3
6.3 Summary of Findings 6-6
7 VENTURI SPRAY TOWER TWO-SCRUBBER-LOOP FORCED-OXIDATION 7-1
LIME TEST RESULTS
7.1 System Description 7-5
7.2 Discussions of Test Results 7-5
7.3 Summary of Findings 7-17
VI
-------
Section Page
8 VENTURI/SPRAY TOWER BLEED STREAM OXIDATION LIMESTONE/MgO 8-1
TEST RESULTS
8.1 System Description 8-3
8.2 Discussions of Test Run Results 8-3
8.3 Summary of Findings 8-7
9 TCA LIMESTONE TEST RESULTS WITH LOW FLY ASH LOADING 9-1
9.1 System Description 9-1
9.2 Discussions of Test Run Results 9-1
9.3 Summary of Findings 9-6
10 TCA LIMESTONE TYPE AND GRIND TEST RESULTS 10-1
10.1 Test Program 10-1
/
10.2 Results 10-4
10.3 Summary of Findings 10-7
11 TCA AUTOMATIC LIMESTONE FEED CONTROL TESTING 11-1
11.1 Test Program Objectives 11-1
11.2 Basis for the Limestone Feed Control Scheme 11-2
11.3 Experience 11-3
11.4 Alternatives 11-4
12 TCA ONE-SCRUBBER-LOOP FORCED-OXIDATION LIMESTONE 12-1
TESTS WITH AIR EDUCTOR
12.1 System Description 12-2
12.2 Discussion of Test Results 12-9
12.3 General Operating Characteristics of the One- 12-14
Scrubber-Loop System with Forced Oxidation
Using an Air Eductor
12.4 Summary of Findings 12-16
vii
-------
Section Page
13 TCA LIMESTONE TEST RESULTS WITH HIGH FLY ASH LOADING 13-1
13.1 Limestone/High Fly Ash Run 715-2A 13-1
13.2 Limestone/High Fly Ash Run 716-2A 13-3
13.3 Limestone/High Fly Ash Reliability Run 717-2A 13-3
14 CEILCOTE SUPPORT PLATE PACKING TESTS ON THE TCA SYSTEM 14-1
14.1 Introduction 14-1
14.2 Test Plan 14-2
14.3 Results of Runs 718-2A and 719-2A 14-2
15 TCA ONE-SCRUBBER-LOOP FORCED-OXIDATION LIMESTNE TESTS 15-1
WITH AIR SPARGER
15.1 System Description 15-3
15.2 Discussions of Test Results 15-7
15.3 Summary of Findings 15-11
16 TCA LIMESTONE/MgO TEST RESULTS 16-1
16.1 Discussions of Test Run Results 16-2
16.2 Summary of Findings 16-7
17 TCA LIME/MgO TEST RESULTS 17-1
17.1 Discussions of Test Run Results 17-1
17.2 Summary of Findings 17-7
18 FLUE GAS CHARACTERIZATION 18-1
18.1 Special Test Runs 18-1
18.2 Flue Gas Mass Loading Tests 18-16
18.3 Air/Slurry Test Results 18-17
19 LIMESTONE REACTIVITY STUDY 19-1
19.1 Laboratory Test Program 19-1
19.2 Limestone Reactivity by the HC1 Method 19-3
vm
-------
Section Page
19.3 Limestone Reactivity by the S02 Titration Method 19-8
19.4 Comparison of and Recommendations for the HC1 19-11
and S02 Titration Methods
19.5 Pilot Plant Test Program 19-13
20 FLUE GAS PRESSURE DROP CORRELATIONS 20-1
20.1 Spray Tower Pressure Drop Correlation 20-1
20.2 TCA Pressure Drop Correlation and Holdup Testing 20-4
21 ANALYSIS OF SLURRY LIQUOR FOR SELECTED TRACT METALS, 21-1
NITROGEN, AND ORGANIC CARBON
21.1 Analysis Program 21-1
21.2 Trace Metals 21-2
21.3 Nitrogen Oxides and Total Organic Carbon 21-5
21.4 Flue Gas Oxidant Monitoring Program 21-6
21.5 Summary 21-6
22 WASTE SOLIDS DEWATERING AND CHARACTERIZATION 22-1
22.1 Lamella Gravity Settler Thickener Test Program 22-2
and Results
22.2 Cylinder Settling and Funnel Filtration Tests 22-10
22.3 Filter Leaf Tests 22-13
22.4 Hydrometer and Pipette Particle Size Distribution 22-22
Tests
22.5 Hydroclone Tests 22-29
22.6 Scanning Electron Microscope (SEM) Analysis of 22-38
Oxidized Solids
22.7 Gypsum Crystallization 22-41
23 OPERATING EXPERIENCE DURING LIME/LIMESTONE TESTING 23-1
23.1 Scrubber Internals 23-1
23.2 Oxidizers 23-8
IX
-------
Section Page
23.3 Reheaters 23-11
23.4 Fans 23-12
23.5 Pumps 23-13
23.6 Alkali Addition Systems 23-15
23.7 Dewatering Systems 23-16
23.8 Instrument Operating Experience 23-23
23.9 Materials and Equipment Evaluation 23-26
24 REFERENCES 24-1
-------
Appendices Page5*
A Converting Units of Measure A-l
B Scrubber Operating Periods B-l
C Properties of Raw Material C-l
D Database Tables D-l
E Test Results Summary Tables for the Venturi/Spray E-l
Tower
F Graphical Operating Data from the Venturi/Spray F-l
Tower Tests
G Average Liquor Compositions for the Venturi/Spray G-l
Tower Tests
H Test Results Summary Tables for the TCA H-l
I Graphical Operating Data from the TCA Tests 1-1
J Average Liquor Compositions for the TCA Tests J-l
K Fifth TVA Interim Report of Corrosion Studies K-l
L Test Data for Waste Solids Dewatering and L-l
Characterization
M Particulate Mass Loading Test Results M-l
(*) Appendix pages are in Volume 2.
xi
-------
ILLUSTRATIONS
Figure Page
3-1 Schematic of Venturi Scrubber and Spray Tower 3-3
3-2 Schematic of Three-Bed TCA 3-4
3-3 Test Facility Mist Eliminator Configuration 3-5
3-4 Flow Diagram for Solids Dewatering System 3-8
4-1 Schedule for Advanced Test Program 4-3
5-1 Flow Diagram for Two-Scrubber-Loop Forced-Oxidation 5-8
Tests on the Venturi/Spray Tower System
5-2 Arrangement of the Venturi/Spray Tower Oxidation 5-9
Tank with Sparger
8-1 Flow Diagram for Bleed Stream Oxidation in the 8-4
Venturi/Spray Tower System
10-1 Test Program Arrangement 10-2
12-1 Flow Diagram for One-Scrubber-Loop Forced Oxidation 12-5
with Air Eductor in the TCA System Using One Tank
12-2 Flow Diagram for One-Scrubber-Loop Forced Oxidation 12-7
with Air Eductor in the TCA System Using Two Tanks
12-3 Arrangement of the TCA Oxidation Tank with Eductor 12-8
in Vertical Position
14-1 Pressure Drop Across 4 Bar Grids and 23 Layers of 14-5
Ceil cote Support Plate Packing
15-1 Flow Diagram for One-Scrubber-Loop Forced Oxidation 15-4
with Air Sparger in the TCA System Using One Tank
15-2 Flow Diagram for One-Scrubber-Loop Forced Oxidation 15-5
with Air Sparger in the TCA System Using Two Tanks
15-3 Arrangement of the TCA Oxidation Tank with Air Sparger 15-6
18-1 Mean Differential Mass Loading Versus Aerodynamic 18-8
Particle Size for all TCA System High Fly Ash Loading
Runs
xii
-------
Figure Page
18-2 Mean Differential Mass Loadings versus Aerodynamic 18-9
Particle Size for all TCA System Low Fly Ash Loading
Runs
18-3 Mass Percent Penetration Versus Aerodynamic Particle 18-10
Size for all TCA High Fly Ash Loading Runs
18-4 Mass Percent Penetration versus Aerodynamic Particle 18-11
Size for all TCA Low Fly Ash Loading Runs
20-1 The Effect of Gas Velocity and Slurry Flow Rate on the 20-2
Flue Gas Pressure Drop Across the Four Slurry Headers
in the Spray Tower
20-2 The Effect of Header Position and Number of Headers on 20-3
the Flue Gas Pressure Drop at 14 gpm/ft2 Total Slurry
Flow Rate
20-3 TCA Bed Pressure Drop (3 Beds, 4 Grids) - 0 in. Total 20-6
Bed Height of Nitrile Foam Spheres
20-4 TCA Bed Pressure Drop (3 Beds, 4 Grids) - 15 in. Total 20-7
Bed Height of Nitrile Foam Spheres
20-5 TCA Bed Pressure Drop (3 Beds, 4 Grids) - 22.5 in. 20-8
Total Bed Height of Nitrile Foam Spheres
20-6 TCA Bed Pressure Drop (3 Beds, 4 Grids) - 30 in. Total 20-9
Bed Height of Nitrile Foam Spheres
20-7 Difference Between TCA Total Pressure Drop and Bed 20-11
Pressure Drop
20-8 Comparison of Experimental Data and Predicted Values of 20-14
Bed Pressure Drop - Three-Stage TCA with 1 1/5-inch Dia.,
6.5 gram Solid Nitrile Foam Spheres
20-9 Comparison of Experimental Data and Predicted Values of 20-15
Bed Pressure Drop - Three-Stage TCA without Spheres
22-1 The Lamella Gravity Settler Thickener 22-4
22-2 Filter Leaf Filtrate Rate vs. Cloth Air Permeability, 22-18
Lime with Fly Ash Slurry (60 second cycle)
22-3 Particle Size Distributions of Oxidized Slurry with 22-24
Low Fly Ash Loading from the Venturi/Spray Tower System
22-4 Particle Size Distribution by the Pipette Method for 22-26
Oxidized Slurry with Low Fly Ash Loading from Run 857-1A
xiii
-------
Figure Page
22-5 Particle Size Distribution by the Pipette Method for 22-27
Oxidized Slurry with Low Fly Ash Loading from Run 858-1A
22-6 Results of Static Cylinder Settling Tests with Oxidized 22-28
Slurry with Low Fly Ash Loading for Runs 857-1A and 858-1A
22-7 Particle Size Distributions from Hydroclone Testing 22-35
(Dorrclone P50A at 18 gpm Feed Rate and 27 psi Pressure Drop)
22-8 Particle Size Distributions from Hydroclone Testing 22-36
(Dorrclone P50A at 15 gpm Feed Rate and 19 psi Pressure Drop)
22-9 Particle Size Distributions from Hydroclone Testing 22-37
(Dorrclone P50A at 12 gpm Feed Rate and 12 psi Pressure Drop)
22-10 Scanning Electron Microscope Photographs of Slurry from 22-39
Forced-Oxidation Run 854-1A (200x magnification)
22-11 Scanning Electron Microscope Photographs of Slurry from 22-40
Forced-Oxidation Run 854-1A (500x magnification)
xiv
-------
TABLES
Table Page
4-1 Field Methods for Batch Chemical Analysis of Slurry and 4-4
Alkali Samples
5-1 Results of Forced-Oxidation Tests with Two-Scrubber-Loops 5-4
on the Venturi/Spray Tower System Using Limestone Slurry
6-1 Results of Forced-Oxidation Tests with Two-Scrubber-Loops 6-2
on the Venturi/Spray Tower System Using Limestone Slurry
with Added Magnesium Oxide
7-1 Results of Two-Scrubber-Loop Forced-Oxidation Lime Tests 7-2
on the Venturi/Spray Tower System
8-1 Results of Forced-Oxidation Tests on the Venturi/Spray 8-5
Tower Bleed Stream Using Limestone Slurry with added
Magnesium Oxide
9-1 Summary of TCA Limestone Tests with Low Fly Ash Loading 9-3
10-1 Limestone Analysis for Runs 707-2A to 712-2A 10-3
10-2 Summary of TCA Limestone Type and Grind Tests (with fly 10-5
ash, no MgO addition)
12-1 Summary of One-Scrubber-Loop Forced-Oxidation Limestone 12-3
Tests with Air Eductor on the TCA System
13-1 Limestone Tests with High Fly Ash Loading Using Three 13-2
Tanks in Series on the TCA System
14-1 Limestone Testing on the TCA System with Ceilcote Plate 14-3
Packing
15-1 Summary of One-Scrubber-Loop Forced-Oxidation Tests with 15-2
Air Sparger on the TCA System
16-1 Summary of Limestone Tests with MgO Addition on the 16-3
TCA System
17-1 Summary of Lime Tests with MgO Addition on the TCA System 17-3
18-1 Flue Gas Characterization Program Results for the TCA 18-3
System
18-2 Flue Gas Characterization Program Results for the Venturi/ 18-4
Spray Tower System
xv
-------
Table
18-3 TCA System Air/Slurry Test Results 18-18
18-4 Venturi/Spray Tower Air/Slurry Test Results 18-19
18-5 Run Conditions for the Air/Slurry Tests on the TCA System 18-20
18-6 Run Conditions for the Air/Slurry Tests on the Venturi/ 18-20
Spray Tower System
19-1 Limestones Used in Reactivity Tests 19-2
19-2 Reactivity Comparison by Source (HC1 Method) 19-6
19-3 Reactivity Comparison by Grind (HC1 Method) 19-7
19-4 Utilization Comparison by Source (S02 Titration) 19-10
19-5 Utilization Comparison by Grind (S02 Titration) 19-12
20-1 Data from Liquid Holdup Tests on the TCA 20-16
21-1 Results of Slurry Liquor Analysis 21-3
22-1 Summary of Dewatering Characteristics of Shawnee 22-12
System Bleed
22-2 Cloth Properties 22-15
22-3 Filter Leaf Test Results on Lime Slurry with High Fly 22-16
Ash Loading
22-4 Filter Leaf Test Results on Limestone Slurry with 22-17
Low Fly Ash Loading
22-5 Filter Leaf Test Results with Blinded Cloths 22-20
22-6 Filter Leaf Test Results at Optimum Filtration Cycle 22-21
22-7 Results of One-Half Hour Tests with Dorrclone P50A 22-32
Hydroclone - Range of Operation -
23-1 Mist Eliminator Wash System 23-3
23-2 Frequency of Pump Maintenance 23-14
23-3 Filter Maintenance Requirement - December 1976 to May 1978 23-18
23-4 Filter Cloth Service 23-19
23-5 Centrifuge Vibration Measurements 23-21
xvi
-------
Section 1
SUMMARY
This is the fourth progress report on an Advanced Test Program under the
sponsorship of the Environmental Protection Agency (EPA) to test prototype
lime and limestone wet-scrubbing systems for removing sulfur dioxide (S02)
and particulate matter from coal-fired utility boiler flue gases. Test re-
sults during the period from late-November 1976 through June 1978 are pre-
sented in this report. Results of earlier testing have been published in EPA
report Nos. EPA-650/2-75-047 (June 1975), EPA-600/2-75-050 (September 1975),
EPA-600/7-76-008 (September 1976), and EPA-600/7-77-105 (September 1977).
The program is being conducted at a test facility operating on flue gas from
Boiler No. 10 at the Tennessee Valley Authority (TVA) Shawnee Power Station,
Paducah, Kentucky. Bechtel National, Inc. of San Francisco is the major
contractor and test director, and TVA is the constructor and facility
operator.
There are two parallel scrubbing systems being operated during the Advanced
Test Program:
t A venturi followed by a spray tower
• A Turbulent Contact Absorber (TCA)
Each system is capable of treating approximately 10 MW equivalent (maximum
1-1
-------
rates of 35,000 acfm* @ 300°F for the venturi/spray tower and 30,000 acfm
@ 300°F for the TCA) of flue gas normally containing 1500 to 4500 ppm of S02.
Flue gas can be withdrawn from either the upstream or downstream side of the
Boiler No. 10 particulate removal equipment, allowing testing v/ith either
high fly ash loadings (3 to 6 grains/dry scf) or low fly ash loadings (0.04
to 0.2 grain/dry scf).
Major areas of testing during this reporting period were:
• Two-scrubber-loop forced-oxidation testing on the venturi/
spray tower system with the following alkalis:
- Limestone slurry, high and low fly ash loadings in
the flue gas
- Limestone slurry with added magnesium oxide (MgO),
high fly ash loadings in the flue gas
- Lime slurry, high and low fly ash loadings in the
flue gas
• Bleed stream oxidation testing on the venturi/spray tower with
limestone slurry and added KgO and with high fly ash loadings
in the flue gas
• Limestone testing on the TCA with low fly ash loadings in the
flue gas
• Limestone type and grind testing on the TCA
• Automatic limestone feed control testing on the TCA
• One-scrubber-loop forced-oxidation testing using an air eductor
on the TCA with limestone slurry and with high fly ash loadings
in the flue gas
* Although it is the policy of the EPA to use the metric system for quantita-
tive descriptions, the British system is used in this report. Readers who
are more accustomed to metric units are referred to the conversion table in
Appendix A.
1-2
-------
• Limestone reliability testing on the TCA with high fly ash load-
ings in the flue gas
• Limestone testing with Ceilcote egg-crate type plate packing on
the TCA
• One-scrubber-loop forced-oxidation testing using an air sparger
on the TCA with limestone slurry and with high fly loadings in
the flue gas
• Limestone testing on the TCA with added MgO and with high fly
ash loadings in the flue gas
• Lime testing on the TCA with added MgO and with high fly ash
loadings in the flue gas
• Flue gas characterization testing on both scrubber systems to
determine inlet and outlet particulate mass loadings, size dis-
tribution, and $03 concentrations
In addition to the major tasks of testing listed above, auxiliary studies were
conducted in the following areas:
t Laboratory studies of the reactivity of various limestone types
and grinds
• Development of pressure drop correlations for the spray tower
and the TCA
• Analysis of trace elements in the slurry liquor which may cata-
lyze or inhibit the oxidation of sulfite
0 Waste solids dewatering and characterization studies
Major conclusions obtained from the work during the reporting period are sum-
marized below.
1.1 TWO-SCRUBBER-LOOP FORCED-OXIDATION TEST RESULTS ON THE VENTURI/SPRAY
TOWER SYSTEM
Following the development of forced-oxidation techniques at the EPA IERL-RTP
pilot plant on a 0.1-MW scale, forced oxidation had been successfully demon-
1-3
-------
strated at the Shawnee Test Facility on 10-MW prototype scrubbers. From
early January 1977 through mid-May 1978, the venturi/spray tower system was
operated in a two-scrubber-loop configuration with forced oxidation of cal-
cium sulfite to calcium sulfate in the low-pH, venturi slurry loop (the first
of two independent slurry recirculation loops). A total of 23 runs were made
with limestone slurry (see Section 5), including 15 runs with high fly ash
loadings and 8 runs with low fly ash loadings in the flue gas. Twenty runs
were made with lime slurry (see Section 7), including 10 runs each with high
and low fly ash loadings. And finally, 6 runs were made using limestone
slurry with added MgO (see Section 6), all with high fly ash loadings in the
flue gas.
A summary of results of these runs follows:
0 Forced oxidation of sulfite to sulfate in the first of two
independent scrubbing loops was successfully demonstrated in
the two-scrubber-loop venturi/spray tower system with both lime
and limestone slurries. Successful demonstration was culminated
with a lime and a limestone one-month reliability test, both
using flue gas with high fly ash loadings.
• The operating characteristics and test results for lime and lime-
stone are similar in many aspects. Under the one-month reliabil-
ity test conditions of 18,000 to 35,000 acfm gas rate, 600 gpm
venturi slurry rate, 1600 gpm spray tower slurry rate, 15 percent
venturi slurry solids concentration, 18 ft oxidation tank level,
and 5.5 oxidation tank pH, both the lime and limestone runs
achieved 97 percent or higher average sulfite oxidation at air
stoichiometric ratios ranging from 1.4 to 2.8 atoms oxygen/mole
S02 absorbed. The filter cake solids content averaged 85 per-
cent or higher.
• For both lime and limestone tests, ranges of conditions under
which near complete sulfite oxidation (greater than 95 percent)
was demonstrated were an oxidation tank pH.of 4.5 to 5.5, an oxi-
dation tank level of at least 14 ft, and an air stoichiometry of
at least 1.5 atoms oxygen/mole S02 absorbed. Filter cake solids
contents of 80 percent or greater i
-------
Sulfite oxidation of 90 percent or higher was required to
obtain maximum improvement in waste solids dewatering charac-
teristics.
In a lime test, the degree of sulfite oxidation was not affec-
ted when the venturi slurry solids concentration was decreased
from the normal 15 percent to 8 percent.
For lime slurry, oxidation efficiency appeared to drop off at
pH's greater than 6. Limestone slurry normally has pH values
below 6.
Forced oxidation was achieved by simple air/slurry contact in
the oxidation tank. Within the ranges of test conditions, air/
slurry contact efficiency did not appear to be affected by using
an air sparger with 130 1/8-inch holes, an air sparger with 40
1/4-inch holes, or a 3-inch diameter open pipe was used. Air/
slurry contact was primarily achieved by the two-turbine agitator
operated at 56 rpm and rated at 17 brake Hp.
Independent lime addition to both scrubber loops was necessary
to have smooth venturi and spray tower inlet pH control. With
limestone operation, addition of fresh limestone slurry to the
spray tower hold tank alone was sufficient to maintain the ven-
turi inlet pH at the desired level.
In both lime and limestone operations, the venturi inlet slurry
liquor constantly exhibited a gypsum saturation level of about
100 percent when oxidation was forced within the venturi loop.
This was probably the result of the abundance of gypsum crystal
seeds produced by forced oxidation. The 100 percent saturation
level is well below the incipient scaling level of 135 percent.
In a limestone run operating without a gypsum desupersaturation
tank after the oxidation tank, neither sulfite oxidation nor
gypsum saturation in the venturi loop was adversely affected.
Lime utilization in the spray tower averaged about 88 percent.
Overall lime utilization averaged 98 percent, demonstrating the
advantage of a two-scrubber-loop system.
Limestone utilization was improved with the two-scrubber-loop
operation to at least 80 percent and as high as 98 percent, de-
pending on the venturi inlet pH which averaged from 4.5 to 5.5.
In a limestone test, operation with the agitator turned off in
the oxidation tank was not successful. Air sparging alone in
the oxidation tank could not keep the solids from settling.
In a series of lime tests, oxidation efficiency did not appear
to be affected when the oxidation tank level v/as dropped from
18 ft to 14 ft at 1.8 air stoichiometry and 5.5 pH. At a 10-ft
tank level, an air stoichiometry approaching 4 atoms oxygen/mole
1-5
-------
SOg absorbed was required to achieve near complete oxidation.
t Spray tower scaling (mostly by calcium sulfite) may result
when the recirculated slurry solids concentration in the
spray tower drops below about 7 percent (low fly ash loading).
This is especially true for lime operation when the spray
tower outlet pH stays above about 5.5. A low slurry solids
concentration in the spray tower also penalizes S02 removal
efficiency in limestone scrubbing because of the reduced dis-
solved alkali concentration.
• The anticipated S02 removal enhancement was achieved in the
two-scrubber-loop venturi/spray tower limestone system by
adding MgO in the spray tower slurry loop. Under typical
operating conditions, average SO^ removal was 96 percent at
2250 ppm average inlet SOo concentration with 5150 ppm effec-
tive magnesium ion concentration* in the spray tower. Under
the same operating conditions but without MgO addition, S02
removal averaged 86 percent at 2550 ppm inlet S02 concentration.
• Magnesium ion does not enhance 3^2 removal in a scrubber loop
in which oxidation is forced, because sulfite ion is oxidized
into nonscrubbing sulfate ion. S02 removal in the venturi loop
(oxidation loop) was determined in a run with MgO addition to
be 29 percent, which is typical of removal efficiency with lime-
stone slurry in the absence of magnesium ion.
• The oxidation efficiency appeared to be unaffected, if not
improved, by the presence of magnesium ion.
• For runs with MgO addition (8000 ppm effective magnesium ion
in the bleed slurry liquor), filter cake solids content averaged
85 percent at 98 percent oxidation, essentially the same as that
without MgO addition.
1.2 BLEED STREAM OXIDATION TEST RESULTS ON THE VENTURI/SPRAY TOWER SYSTEM
From mid-May 1978 through mid-June 1978, four runs were made on the venturi/
spray tower system (see Section 8) in which one hold tank was used for both
the venturi and spray tower recirculation slurries. The system bleed slurry
* Effective magnesium ion concentration is defined as the total magnesium
ion minus that magnesium ion concentration equivalent to total chlorides.
Magnesium chloride has no effect on S02 removal.
1-6
-------
was taken from the hold tank and sent to a separate tank in which oxidation
was forced by simple air/slurry contact. Limestone slurry with high fly ash
loading was used in all four runs. Magnesium oxide was added to the effluent
hold tank to maintain an effective magnesium ion concentration of 5000 ppm.
Major conclusions from these runs were as follows:
• Forced oxidation on the bleed stream outside the scrubber
slurry loop is not expected to work with lime or limestone
slurries (without added MgO). Slurry pH rise in the oxida-
tion tank, caused by the dissolution of residual alkali,
prevents solid calcium sulfite from dissolving and oxidizing
in the liquid phase. The pH rise also reduces the rate of
oxidation reaction.
• Bleed stream oxidation of limestone slurry is feasible in the
presence of a sufficient quantity of magnesium ion. Magnesium
ion tends to buffer the pH rise and promotes the availability
of dissolved sulfite for oxidation. Higher liquor sulfite con-
centration also allows the oxidation to take place at a higher
pH.
• Average sulfite oxidations greater than 95 percent were achieved
with an air stoichiometry of 1.6 atoms oxygen/mole SO? absorbed,
an oxidation tank level of 18 ft, an oxidation tank pR of 5.4 to
6.3, and 5000 ppm effective magnesium ion concentration. Oxida-
tion was consistently high even though the pH rose as high as
6.7.
t Oxidized slurry from all runs had good dewatering properties.
Filter cake solids concentration averaged about 84 percent.
• Oxidation of the bleed stream outside of the scrubber slurry
loop did not interfere with the S0£ removal enhancement by the
magnesium ion, as was experienced when oxidation was accomplished
within the scrubber loop.
• In two runs, 30 gpm of oxidized slurry was recycled from the oxi-
dation tank back to the effluent hold tank in an attempt to
reduce the pH rise in the oxidation tank. This recycle stream
proved to be unnecessary as the oxidation tank pH was only 0.1
to 0.2 pH unit higher than the effluent hold tank pH without the
recycle.
Additional testing has been planned in the future to fully characterize the
limestone/MgO bleed stream oxidation system. The possibility of bleed stream
1-7
-------
oxidation on a lime/MgO system will also be investigated.
1.3 LIMESTONE TEST RESULTS ON THE TCA WITH LOW FLY ASH LOADINGS
IN FLUE GAS
Six limestone runs were conducted on the TCA system from late November 1976
through early February 1977 (see Section 9). All runs were made with flue
gas having low fly ash loadings. The main objective of this series of tests
was to compare the TCA performances for runs with high and low fly ash load-
ings. Another objective was to investigate the scrubber inlet slurry pH and
S02 removal, at low fly ash loading, as functions of the slurry solids con-
centration, hold tank residence time, number of hold tanks, and limestone
stoichiometry. A summary of results of these runs follows.
t At the same limestone utilization, the scrubber inlet pH for
runs with low fly ash loadings tended to be 0.3 to 0.4 pH unit
higher than runs with high fly ash loadings. This is consis-
tent with earlier observation during venturi/spray tower lime
tests. In those tests, at a controlled scrubber inlet pH of 8,
lime utilization averaged about 93 percent for runs with low
fly ash loadings, compared with about 88 percent for runs with
high fly ash loadings. This difference was believed to be
caused by the acidic fly ash which releases its acidity from
absorbed S03 under scrubber conditions, even though the fly
ash pond liquor at Shawnee is alkaline after a prolonged period
of leaching.
• Within the range of slurry solids concentrations tested (8 to
15 percent), higher solids concentration gave slightly higher
scrubber inlet pH and S02 removal. The effect is more pronounced
at higher residence time (12 minutes) with three tanks in series.
0 At the same limestone stoichiometry (1.1) and total residence
time (12 minutes), three hold tanks in series gave higher S02
removal and scrubber inlet pH than a single hold tank.
1-8
-------
1.4 LIMESTONE TYPE AND GRIND TEST RESULTS ON THE TCA SYSTEM
A series of six tests were conducted on the TCA system from March 4 through
April 2, 1977 (see Section 10), to compare the reactivity of limestones from
two different sources: Fredonia, Kentucky, and Long View, Alabama. The
effect of ground limestone size was also investigated. All tests were made
with flue gas having high fly ash loadings. Significant findings of these
tests follow.
At the same nominal grind of 70 percent less than 325 mesh,
Long View limestone was slightly more reactive than Fredonia
limestone. At a residence time of 4.1 minutes, stoichiometric
ratio of 1.2 moles Calcium (Ca)/mole S02 absorbed, and a liquid-
to-gas ratio of 50 gal/Mcf, S02 removal was 55 percent at 3350
ppm inlet S02 concentration for Long View, compared with 49
percent removal at 3100 ppm inlet S02 concentration for Fredonia.
However, a severe limestone feed line plugging problem was expe-
rienced with the Long View limestone.
• S02 removal was sensitive to the limestone grind, especially at
a Tow residence time. At 50 gal/Mcf liquid-to-gas ratio, 1.2
stoichiometric ratio, and 4.1 minutes residence time, SOo removal
increased from 49 to 72 percent, both at 3100 ppm inlet M)p, when
the grind size of the Fredonia limestone was reduced from 59 per-
cent less than 325 mesh to 96 percent less than 325 mesh. At a
higher residence time of 12 minutes, S02 removal increased from
54 percent at 3400 ppm inlet S02 to only 64 percent at 3250 ppm
with the corresponding limestone grinds.
Thus, co obtain the desired degree of S02 removal, consideration must be given
to the cost trade-off between the use of coarse and fine ground limestone. Use
of coarse limestone requires higher limestone stoichiometry, resulting in more
waste sludge to be disposed of. Use of fine ground limestone, on the other hand,
requires higher capital and operating costs for the ballmills, but lower stoich-
ometry with less waste sludge generated.
1-9
-------
1.5 TEST RESULTS OF AUTOMATIC LIMESTONE FEED CONTROL ON THE TCA
SYSTEM
The feasibility of automatic limestone feed control in a limestone wet-
scrubbing system was investigated on the TCA from August 30 to September 9,
1977 (see Section 11). The benefits of automatic limestone feed control
include:
• Minimizing the limestone usage to meet the S02 emission
standard
• Reducing manual operation and operator errors
a Reducing the scrubber operating costs
a Improving reliability by reducing the potential for plug-
ging with soft mud-type solids caused by excessive limestone
stoichiometry
The control scheme tested at Shawnee was based on the material balance concept.
In this scheme, a desired stoichiometric limestone feed is maintained in rela-
tion to the amount of S02 entering or absorbed in the scrubber according to the
following equation:
Limestone (60 percent slurry) addition rate, gpm
= G x (S02I - Kout) x KSR
where: G = flue gas flow rate, acfm @ 300°F
S02I = inlet S02 concentration measured by Du Pont analyzer, ppm
Kout = a manually adjustable constant related to desired outlet
S0£ concentration, ppm
KSR = a manually adjustable constant proportional to the stoichio-
metric ratio
= (unit conversion factor) x (stoichiometric ratio)
= 2.34 x 10'8 x (stoichiometric ratio)
1-10
-------
The gas rate, G, and inlet S02 concentration, S02I, may vary within a wide
range depending on the boiler load and the sulfur content in the coal. There-
fore, the control scheme included the following overrides:
t If the outlet S02 concentration as measured by the Du Pont
analyzer exceeds a set maximum, the limestone feed rate will
be stepped up to a preset maximum.
0 If the outlet S02 concentration drops below a set minimum,
the limestone feed rate will be stepped down to a preset
minimum.
The first provision minimizes the infringement of the S02 emission standard;
while the second allows limestone savings by avoiding unnecessary S02 removal.
The experience with this first-generation control design at Shawnee, although
favorable and limited in time, has revealed several areas for improvement.
One specific area that needs improvement relates to the methods of control
override mentioned above. These methods of override result in either the over-
feeding or underfeeding of limestone with a consequent loss in scrubber perfor-
mance and sometimes system scaling. The modification proposed would add a
proportioning device which would regulate the limestone feed rate, at the
overriding conditions, in proportion to the difference between the outlet S02
setpoint and the prevailing S02 outlet concentration. Additional tests, includ-
ing some made with the proposed modification, are planned in the future Shawnee
test program.
1.6 ONE-SCRUBBER-LOOP FORCED-OXIDATION LIMESTONE TEST RESULTS ON THE
TCA WITH AIR EDUCTOR
As an alternative to the types of air sparger used in the venturi/spray tower
1-11
-------
system, an air eductor was used to accomplish air/slurry contact in the TCA
system. Tests were conducted on the one-scrubber-loop TCA system with lime-
stone slurry having high fly ash loadings. No one-scrubber-loop forced-
oxidation tests were made with lime slurry because sulfite is the major SC^
scrubbing species in a lime scrubber. Forced oxidation within the scrubber
loop converts the sulfite to sulfate in the slurry liquor, resulting in
poor S02 removal.
A total of 15 runs were made. These included 11 runs made from June 24
through August 29, 1977, with the air eductor mounted in a vertical position
over the oxidation tank, and 4 runs from September 9 through October 4, 1977,
with the air eductor in a horizontal position beside the oxidation tank
(see Section 12). Major conclusions from these runs were as follows:
t Under the base case conditions of 30,000 acfm gas rate, 1200
gpm TCA slurry rate, 1600 gpm eductor slurry rate, 15 percent
recirculated slurry solids with fly ash loading, and 8- to 12-ft
slurry level in the effluent hold tank (oxidation tank), sulfite
oxidation better than 90 percent was achieved at an eductor inlet
pH range of 4.95 to 5.45 and air stoichiometric ratios as low as
2.0 atoms oxygen/mole S02 absorbed.
• The dewatering and handling characteristics of slurry solids oxi-
dized to 90 percent or higher in a one-scrubber-loop system were
as good as those in a two-scrubber-loop system.
• With a single-tank configuration and limestone addition to the
effluent hold tank, low hold tank pH hence, low SOo removal was
necessary to achieve good sulfite oxidation. The §02 removal
could be improved, however, by limestone addition to the TCA
inlet slurry stream instead of the hold tank.
• Operation in a two-tank configuration allowed for a higher pH
to the TCA for good SOp removal and a lower pH to the eductor
for good sulfite oxidation. Sulfite oxidations of up to 98
percent were achieved with satisfactory S02 removal (in the
range of 80 to 90 percent).
0 S02 removal in a one-scrubber-loop forced-oxidation system tended
to be a few percentage points higher than the predicted S02 removal
1-12
-------
under similar operating conditions without forced oxidation.
Air stripping of carbon dioxide (CO?) from the liquor and addi-
tional slurry agitation by air, both of which promote the lime-
tone dissolution rate, are believed to cause the SC^ removal
enhancement.
• No significant difference in sulfite oxidation efficiency was
observed between operations with the air eductor in the vertical,
top-entry position or operations with the air eductor in the
horizontal, bottom-entry position.
• Sulfite oxidation was unsatisfactory (about 60 percent) when
the eductor slurry flow rate was reduced from 1600 to 1200 gpm.
Decreased agitation in the oxidation tank at the lower flow
rate, rather than the reduced air stoichiometric ratio, is be-
lieved to be the main reason for poor oxidation.
• Air eduction by the eductor was as predicted by the manufac-
turer. Air flow rate was sensitive to the depth of the eductor's
submergence in the slurry, ranging from 600 scfm at zero submerg-
ence to 200 scfm at 12-ft submergence, at a 1600 gpm slurry rate.
A total of 200 scfm of air was sufficient to achieve good sulfite
oxidation at 1600 gpm slurry flow.
• Air/slurry contact was hampered by the configuration of the educ-
tor discharge hold tank (20-ft diameter tank with only 8- to 12-ft
normal liquid depth). A smaller diameter but deeper tank would
probably have given better oxidation efficiency.
t Erosion of the air eductor with limestone slurry having high
fly ash loading (in limestone/fly ash service) was a problem.
The eductor body was constructed of neoprene-lined carbon steel.
The rubber near the nozzle had eroded in a circular pattern after
1500 operating hours, and the carbon steel body was eroded through
after an additional 550 hours. The stellite nozzle and whirler
showed only minor erosion.
1.7 LIMESTONE RELIABILITY TEST ON THE TCA WITH HIGH FLY ASH LOADING
IN THE FLUE GAS
From October 20 through November 21, 1977, a month-long limestone reliability
test was conducted on the TCA system (see Section 13). Flue gas with high fly
ash loading (3 to 6 grains/dry scf) was used in the test. The objectives of the
test were to meet the current EPA New Source Performance Standard (NSPS) of
1-13
-------
1.2 Ib S02/MM Btu and 0.10 Ib particulate matter/MM Btu, while achieving
reliable system operation with respect to scaling and plugging.
After two exploratory runs (October 7 to 20, 1977), the following test condi-
tions were selected to meet the objectives of the reliability test:
Gas rate = 20,000 to 30,000 acfm @ 300°F
Slurry rate = 1000 gpm
Recirculated slurry solids concentration = 15%
Total residence time (three series tanks) = 14.4 minutes
Limestone stoichiometry (controlled) = 1.2 mole Ca/mole S02
absorbed
Total static bed height (three beds) = 22.5 inches
The gas flow rate was varied according to the No. 10 Boiler load which normally
fluctuated between 100 and 155 MW.
Reliable operation of the TCA system was demonstrated during the one-month test,
with a scrubber availability of 99.3 percent. Average S02 removal for the en-
tire run was 87 percent at 2800 ppm average inlet S02 concentration. This
corresponds to an average emission of 1.0 Ib S02/MM Btu, well within the EPA
standard of 1.2 Ib S02/MM Btu. However, due to wide fluctuations in inlet
S02 concentration and slow system or operating response time, the S02 emissions
at times exceeded the EPA standard for periods greater than the three hours
allowed by existing EPA regulations.
Average particulate loading was 0.043 grain/dry scf corresponding to an average
emission of 0.083 Ib particulate/MM Btu. Although the EPA standard of 0.10 Ib
particulate/MM Btu was exceeded in a few measurements; however, on the average,
this standard was not exceeded.
1-14
-------
1.8 LIMESTONE TEST RESULTS WITH CEILCOTE EGG-CRATE TYPE PLATE
PACKING ON THE TCA
An egg-crate type packing support plate manufactured by Ceilcote Company was
tested as TCA internals as an alternative to the 1-5/8 inch diameter, 6.5 gram
solid nitrile foam spheres. The plate dimensions were 2 ft x 2 ft x 2 inch,
with 1-3/16 inch square openings. They were made of fiberglass filled polypro-
pylene. Two one-week tests were conducted from November 24 through December 9,
1977 (see Section 14) to study the performance of the Ceilcote plates with re-
spect to S02 removal, particulate removal, and turn-down capability.
Twenty-three layers of Ceilcote plates (46 inches in height) were installed
between the second and third bar grids. The number of plate layers was selec-
ted to give an estimated pressure drop of about 8 inches H20 (including the
four grids) which is normally obtained with 15 inches total static height
of foam spheres (three beds) operating at 30,000 acfm gas rate and 1200 gpm
slurry rate.
Under the typical operating conditions at full gas load (30,000 acfm gas rate,
1200 gpm slurry rate, 15 percent slurry solids with high fly ash loading, 12
minutes residence time, and 1.2 moles Ca added/mole S02 absorbed), S02 removal
with 23 layers of Ceilcote plates averaged 76 percent at 3200 ppm average in-
let S02 concentration. This S02 removal was slightly lower than the 80 per-
cent predicted for 15 inches of foam spheres under the same operating conditions.
When the gas rate was turned down to 18,000 acfm (with other test conditions
unchanged), the S02 removal for the Ceilcote plates was 86 percent at 3050
ppm S02, considerably higher than the 78 percent predicted for the foam spheres.
Thus, the Ceilcote plates possess better turndown capability than a TCA unit
1-15
-------
operating with spheres under equivalent conditions.
Particulate removal efficiencies at full gas load for the Ceilcote plates
and for the foam spheres were similar (about 98.9 percent). At 18,000 acfm,
the removal efficiency for the Ceilcote plates decreased to 98.4 percent,
lower than the 99.1 percent experienced in a previous TCA test with 15 inches
of spheres and at 20,000 acfm. However, pressure drop across the Ceilcote
plates was only 2.9 inches HgO compared to about 4.5 inches ^0 for the
three sphere beds. ;
No plugging occurred during the two Ceilcote plate test runs. Minor scale
formation was observed along the edges of some plates. However, the overall
duration of both runs was not sufficient to determine reliability.
In summary, Ceilcote plates give much better S02 removal efficiency under
turndown conditions compared with sphere beds of similar operating conditions.
1.9 ONE-SCRUBBER-LOOP FORCED-OXIDATION LIMESTONE TEST RESULTS ON THE
TCA WITH AIR SPARGER
Because of the erosion problem with an air eductor (see Subsection 1.6), fur-
ther one-scrubber-loop forced-oxidation testing on the TCA was conducted
using an air sparger having 40 1/4-inch diameter holes, similar to the one
used in the venturi/spray tower system. As in the testing with the air
eductor, only limestone slurry with high fly ash loading was used. Seven
runs were made from December 9, 1977 through January 24, 1978, using either
one or two tanks. In addition, one run was made from May 31 to June 8, 1978,
with limestone slurry and added MgO (see Section 15). Important conclusions
from these tests are listed below:
l-ifi
-------
t In similar oxidation tank environments, higher sulfite oxi-
dation efficiency was achieved in the venturi oxidation tank
than in the TCA oxidation tank (95 to 100% versus 90 to 95%).
The poorer performance on the TCA system was attributed to the
weaker agitation in the TCA oxidation tank.
• With two tanks in series, 94 percent sulfite oxidation was
achieved at an air stoichiometry of 1.7, oxidation tank pH of
5.45, and an oxidation tank level of 18 ft.
• Higher air stoichiometry (about 1.9) was required to achieve
a similar degree of oxidation (94 percent) in a one-tank configu-
ration because of higher pH.
t Operation in the two-tank mode gave better limestone utilization
and SOp removal than the one-tank mode.
• As expected, MgO addition did not enhance the S02 removal when
oxidation was forced within the scrubber loop.
Further single-loop forced-oxidation tests on the TCA system will be conducted
when a higher speed agitator is installed in the oxidation tank.
1.10 TCA LIMESTONE/MgO AND LIME/MgO TEST RESULTS
From late January 1978 through the end of the report period, the TCA system
was operated with lime or limestone slurry with added MgO and without forced
oxidation (see Sections 16 and 17). Flue gas with high fly ash loadings
was used in all tests. These tests were conducted to try to resolve some
of the inconsistent results obtained in a similar test series made earlier
in April through November 1976. In those earlier tests, air leakage into
the scrubber downcorner was suspected during some of the runs, such air leak-
age was probably the cause of higher than expected sulfite oxidation and
gypsum saturation.
Unfortunately, coals from many different sources were burned in Boiler No. 10
1-17
-------
in the first quarter of 1978 during the coal miner's strike, resulting in
as much as tenfold variation in inlet SC^ concentrations (320 to 3400 ppm).
The wide fluctuations in the inlet S02 concentration not only created pro-
cess control problems, but also caused difficulties in the data interpreta-
tion and in the run-to-run comparison of test results. Therefore, the
inconsistent results obtained earlier were never resolved.
1.10.1 Limestone/MgO Tests
The general conclusions that can be drawn from the limestone/MgO tests are
listed below:
• Higher effective Mg"1"1" concentration is needed in a limestone
system than a lime system to achieve a similar degree of S02
removal enhancement. This is probably due to the lower pH
(5 to 6) inherent in the limestone slurry, where the sulfite-
bisulfite equilibrium favors a shift toward nonscrubbing
bisulfite species.
• Under typical TCA operating conditions, S02 removal improved
by about 15 percentage points with 9000 ppm effective Mg++
concentration.
• Addition of MgO at the levels tested did not always result
in a gypsum-subsaturated mode of operation. Lower sulfite
oxidation generally gave lower gypsum saturation.
• No correlation appeared to exist between gypsum saturation
and SOo make-per-pass. However, gypsum scaling potential tended
to be higher (incipient scaling gypsum saturation level tended
to be lower) at higher $©2 make-per-pass.
1.10.2 Lime/MgO Tests
The general conclusions that can be drawn from the lime/MgO tests are listed
be!ow:
1-18
-------
• At low inlet S02 concentrations encountered (1600 to 2000 ppm),
sulfite oxidation was 10 to 15 percentage points higher than
normal. Higher 02/SOp ratio in the inlet flue gas probably
caused the higher oxidation.
t Improvement in S02 removal with 2000 ppm effective Mg++ concen-
tration was not discernible when the inlet S02 concentration
was low (1600 to 2000 ppm). In this case, gas-phase mass trans-
fer may become rate-controlling. High sulfite oxidation at the
lower inlet S02 concentrations would also render Mg"1"*" ion ineffec-
tive (sulfite concentration in liquor reduced) in S02 removal
enhancement.
0 As in the limestone/MgO tests, lower sulfite oxidation tended
to give lower gypsum saturation. No correlation appeared to
exist between S02 make-per-pass and gypsum saturation. At simi-
lar gypsum saturation levels, gypsum scaling potential appeared
to be higher at higher S02 make-per-pass. (The incipient scal-
ing gypsum saturation level tended to be lower at higher S02
make-per-pass).
• Within the range of the hold tank residence times tested (3.0
to 4.1 minutes), lower residence time resulted in lower gypsum
saturation under a similar sulfite oxidation level.
1.11 FLUE GAS CHARACTERIZATION TEST RESULTS
The results of flue gas particulate measurements from special test programs
and routine sampling tests are presented in Section 18. Summary results
from these tests are presented below.
1.11.1 Special Test Program
A special flue gas characterization test program was set up for the TCA from
late January through early March 1977. In addition, a special run was con-
ducted in mid-June 1977, completing a similar test series on the venturi/
spray tower system which was reported previously (Reference 4).
1-19
-------
The effects of major operating variables on the particulate mass removal
and participate size distribution were studied in these special tests.
Variables investigated for both scrubber systems included gas rate, slurry
recirculation rate, inlet flue gas fly ash loading (high or low), and MgO
addition. In addition, venturi pressure drop, slurry solids concentration,
and mist eliminator configuration were studied for the venturi/spray tower,
and mist eliminator wash scheme for the TCA.
For the TCA system, the average total outlet mass loading ranged from 0.042
to 0.065 grain/dry scf with high fly ash loading in the inlet gas, and was
0.026 grain/dry scf with low fly ash loading in the inlet gas.
Mass penetration for the TCA with high fly ash loading in the inlet gas
ranged from 40 to 90 percent for 0.1 micron aerodynamic particle size, and
less than 2 percent for aerodynamic particle sizes greater than 5 microns.
With low fly ash loading in the inlet flue gas, the mass penetration was 25
to 70 percent for 0.1-micron particles, and 10 percent for particles greater
than 5 microns.
Mass emission from the TCA with high fly ash loading in the inlet flue gas
averaged 0.025 grain/dry scf for particles with actual diameters less than
2 microns (except at low gas or low slurry rate, which gave 0.037 grain/dry
scf), and averaged 0.028 grain/dry scf for particles with actual diameters
greater than 2 microns (except for the base case run with continuous mist
eliminator bottom wash at 0.055 grain/dry scf and the MgO addition run at
0.038 grain/dry scf). With low fly ash loading in the inlet gas, the emis-
sion averaged 0.003 and 0.025 grain/dry scf for particles with actual dia-
meters less than 2 microns and greater than 2 microns, respectively.
1-20
-------
The single special test, conducted on the venturi/spray tower system during
the report period, was made with spray tower-alone operation (venturi plug
wide open) with flue gas having low fly ash loading. Inlet particulate mass
loading averaged about 0.15 grain/dry scf with 85 percent mass removal. The
venturi pressure drop of 2.5 inches K^O (corresponding to the venturi plug
wide open) is believed to have contributed significantly to this high removal;
however, there is no way to isolate the spray tower from the venturi to
verify this assumption.
1.11.2 Routine Mass Loading Test Results
Particulate mass loadings measured routinely at Shawnee during the other
venturi/spray tower and TCA test blocks showed results similar to those
obtained during the special test program.
In general, outlet mass loading decreased for both scrubber systems with
increasing flue gas pressure drop. The venturi/spray tower system generally
yielded slightly lower outlet mass loading than the TCA, because a majority
of runs were operated with a 9 inches F^O venturi pressure drop. The TCA
outlet mass loading exceeded the EPA standard of 0.10 Ib/MM Btu (correspond-
ing to about 0.052 grain/dry scf assuming 30 percent excess air) when flue
gas pressure drops were below about 8 inches f^O.
1.11.3 Air/Slurry Test Results
Tests with air and slurry were conducted during the No. 10 Boiler outage in
April and May 1977 to determine the scrubber mass emissions contributed by
the slurry entrainment.
1-21
-------
However, the dust loadings in the inlet air were significant, and sometimes
exceeded the measured outlet mass loading. It could only be concluded that
the actual slurry emissions should be less than the measured values.
Within the range of operation, measured emissions ranged from 0.001 to 0.003
grain/dry scf for the venturi/spray tower (with 8 percent slurry solids con-
centration) and 0.001 to 0.005 grain/dry scf for the TCA (with 15 percent
slurry solids concentration). These ranges should be the upper limits of the
emissions contributed by the slurry.
1.12 LABORATORY LIMESTONE REACTIVITY TEST RESULTS
Laboratory tests of limestone reactivity were conducted at Shawnee to study
the effect of limestone source and grind on reactivity. The results of these
tests are presented in Section 19. Limestones from 11 different quarries were
tested; each stone was ground to two nominal sizes - 75 percent less than 200
mesh and 90 percent less than 325 mesh.
Two methods were used to test the reactivity:
• HC1 Method: 5 grams of dried limestone powder were added to an agi-
tated 300 ml solution of HC1 at pH 3 and 25°C and the rise of pH with
time was recorded.
• S02 Method: First 0.2 mole of limestone was dissolved in two liters
of distilled water at 50°C until a constant pH was reached. Then a
constant stream (0.273 g-mole/hr) of pure SOg gas was bubbled through
the slurry and the pH drop with time was monitored.
Some conflicting results were obtained when comparing the two methods. The
S02 method, although somewhat more involved than the HC1 method, is believed
to be more appropriate for evaluating limestone reactivity because it more
closely resembles actual S02 wet-scrubbing process.
1-22
-------
Due to lack of equipment, no attempt was made to correlate the limestone
reactivity values with physical properties (hardness, average pore size,
pore size distribution, porosity, etc.).
Tests have been planned for both the IERL-RTP Pilot Plant and the Shawnee
Test Facility to study how SOp removal, pH, and stoichiometry are related
for different limestone type and grind.
1.13 PRESSURE DROP CORRELATIONS FOR SPRAY TOWER AND TCA
Data obtained from the special pressure drop testing on the Shawnee spray
tower and TCA are presented in Section 20. Equations that were used to
fit the data are presented below.
1.13.1 Spray Tower Pressure Drop Correlation
For flue gas/slurry operation, the following equation was obtained:
AP = 0.0062V1'17 exp (0.084L)
where:
AP = flue gas pressure drop across four spray headers (4 ft vertical
distance between headers), in. 1^0
v = superficial gas velocity in the spray tower, ft/sec at 125°F
L = slurry flow rate to four headers, gpm/ft of spray tower cross-
sectional area (Shawnee spray tower cross-sectional area = 50 ft )
1-23
-------
1.13.2 TCA Pressure Drop Correlation
For air/water operation, the fitted equation was:
AP = (AP)NS -i- 0.0363 v Hs°'69 exp (0.014 L)
with:
(AP)NS = 0.095 v exp (0.021 L)
where:
AP = TCA pressure drop for three beds and four grids, in. F^O
(AP)NS = TCA pressure drop for four grids (no spheres), in. 1^0
v = superficial air velocity, ft/sec at 90°F
L = water flow rate per unit scrubber cross-sectioanl area, gpm/ft
(TCA cross-sectional area is 32 ft2)
H
-------
(see Section 21). It is hoped that, through this trace element monitoring
program, any unexpected excursions in sulfite oxidation level would be re-
flected in unusual variations of trace element concentrations from the normal
levels. However, no unusual excursion in sulfite oxidation has been encoun-
tered since the monitoring program was started.
The trace metals being monitored are Fe, Mn, Mo, Ni, Ti, and V. Nitrite/
nitrate and total organic carbon (TOC) are also being analyzed. Typical con-
centrations of these elements in ppm are:
£e Mn Mo M II 1 N02/N03 TOC
Venturi inlet 5 42 5 1 0.4 1 45 5
TCA outlet 7 7 5 0.4 0.3 1.5 15 2
The high value of 42 ppm Mn at the venturi inlet is typical during two-loop
forced oxidation and may be the result of more soluble oxidized manganese
compounds.
1.15 RESULTS OF WASTE SOLIDS DEWATERING AND CHARACTERIZATION STUDIES
Section 22 presents the results of special v/aste solids dewatering and charac-
terization studies, such as the evaluation of a Lamella gravity settler thick-
ener, filter leaf tests, hydrometer and pipette particle size distribution
tests, evaluation of hydroclones, scanning electron microscope (SEM) analysis,
and gypsum crystallization study. Results of routine cylinder settling tests
and Buchner funnel filtration tests are also presented.
A brief summary and conclusions of these studies are given below.
1-25
-------
1.15.1 Lamella Gravity Settler Thickener Test Results
Testing of a Lamella inclined plate settler thickener, Model LGST No. 3 manu-
factured by Parkson Corporation, was conducted from July through November 1977.
The tests were designed to determine the maximum clarification and thickening
capacity of the unit with different slurry types and feed rates. The unit con-
sisted of a rectangular clarification tank containing a series of inclined
parallel plates, which was mounted on a circular thickening tank containing a
picket fence-type rake. The inclined plate pack had a total settling area of
115 sq.ft. with 57.5 sq.ft. utilized as effective clarification area. The
thickener tank was 4 ft in diameter and 10 ft in height. The overall height
of the unit was 20 ft.
The following conclusions were drawn as a result of the tests:
• The Lamella settler requires about one-sixth the land area
and has a smaller slurry inventory, compared with a conven-
tional clarifier having equivalent settling area.
• Oxidized slurry at 15 weight percent solids has surface load-
ing rates (gallons of slurry per minute per ft ) 5 to 10 times
that of unoxidized slurry.
• Flocculant addition greatly increased the surface loading rate
of both oxidized and unoxidized slurry.
• After initial modifications, the Lamella settler operated free
from scaling, plugging, and mechanical problems.
• In comparison with conventional settlers, the Lamella settler
is superior for clarification but offers little advantage
for thickening.
1.15.2 Cylinder Settling and Funnel Filtration Test Results
Cylinder settling and Buchner funnel filtration tests were conducted routinely
1-26
-------
at Shawnee to monitor and detect any gross changes in the slurry dewatering
characteristics. The initial settling rate indicates the rate at which
solids fall during the unhindered portion of settling in a clarifier.
Ultimate settled solids concentration in the cylinder represents that achiev-
able in a thickener or settling pond. Funnel test cake solids concentration
represents that obtainable on a vacuum filter-type dewatering device. The
test results are summarized below:
• The average range of the initial settling rate was 0.7 to
1.2 cm/min for an oxidized slurry and 0.2 to 0.4 cm/min for
an unoxidized slurry, regardless of high or low fly ash
loading.
• The initial settling rate of unoxidized slurry is probably
limited by calcium sulfite particles, while fly ash is prob-
ably the limiting factor for oxidized slurry.
• The average range of ultimate settled solids and funnel fil-
tration solids was 70 to 76 weight percent for unoxidized
siurry.
• In a two-scrubber-loop oxidation with limestone slurry and
high fly ash loading, the addition of MgO (8000 ppm effective
Mg +) slightly reduced the initial settling rate, ultimate
settled solids, and funnel test cake solids.
• For limestone slurry with 5000 ppm effective Mg++ and high
fly ash loading, oxidation in the bleed stream reduced the
initial settling rate to half that of other oxidized slurries.
The ultimate settled solids were somewhat lower for the slur-
ries with magnesium, but the funnel test cake solids were not
significantly affected.
0 The reduced initial settling rate when magnesium is present
is probably the result of an increase in liquor viscosity and
density due to the increased total dissolved solids.
1.15.3 Filter Leaf Test Results
Tests were conducted with a standard filter leaf having a cloth area of 0.10
ft2. The objectives of the test were to identify promising filter cloths
1-27
-------
for future evaluation on the Shawnee rotary drum vacuum filter and to deter-
mine the filtration characteristics of sludges generated under different
operating conditions.
Twenty-three cloths were evaluated, ranging in air permeability from 0.8 to
100 cfm/ft2. These included Technical Fabricators TFI9162, Lamport 7512-SHS,
and Ametek STE-F9D8-HJO (air permeability range from 7 to 20 cfm/ft2), which
were normally used on the Shawnee vacuum drum filter. Dry cake production
rate for these three cloths ranged from 73 to 112 lbs/hr/ft2 for lime slurry
n
with high fly ash loading and 108 to 134 Ibs/hr/ft for limestone slurry
with low fly ash loading. Cake solids concentration for the three cloths
ranged from 61.8 to 66.9 weight percent and from 44.6 to 45.3 weight percent
for the respective slurry types. None of the other cloths tested showed
significantly improved dry cake production rate or cake solids concentration.
Weight percent solids in the filtrate increased with the air permeability
from 0.0 to 0.36 for lime and from 0.0 to 0.09 for limestone, all within
the acceptable limit.
Results of tests at the optimum filtration cycle showed that oxidized cake
solids were between 70 and 80 weight percent, regardless of alkali type or
fly ash content. Unoxidized lime slurry with low fly ash loading had much
lower cake solids of about 50 weight percent.
1.15.4 Results of Particle Size Distribution Tests with Hydrometer
and Pipette
Use of conventional settling rate data to assess the effects of process varia-
bles on the gypsum particle size distribution generated during forced-oxidation
1-28
-------
testing can result in misleading interpretation because of the presence of
other materials, particularly fly ash, in the slurry. Even with low fly ash
loading in the flue gas at Shawnee, the product solids contain about one
weight percent fly ash. The fly ash normally settles at a much slower rate
than the gypsum produced at Shawnee, masking the true settling rate of the
latter. At Shawnee, the limit of settling rate appears to be 1.5 to 2.5
cm/min for oxidized slurry. Therefore, once the true settling rate of gypsum
exceeds this limit, any further increase in the average size of gypsum parti-
cles due to changes in process variables will not be observed.
Two particle size analysis methods, potentially capable of analyzing gypsum
particle sizes in the presence of fly ash, have been briefly assessed. The
hydrometer method (ASTM D422-63, reapproved 1972, "Standard Method for
Particle-Size Analysis of Soils") involves placing an ASTM hydrometer (151H
or 152H in ASTM Specification E 100) in a 1 liter graduated cylinder and
observing the hydrometer reading with the time of slurry settling. In the
pipette method, a 10-ml pipette is used to periodically withdraw slurry
samples from the settling cylinder at a constant depth, and the of samples
are analyzed for solids composition and weight to calculate the size distri-
bution using Stokes1 Law.
On the basis of the test results, the hydrometer test gives a good indication
of the overall size distribution but no information on the fly ash/gypsum
distribution. The pipette method appears to be the most promising method
for segregating the gypsum size distribution from the fly ash.
1-29
-------
1.15.5 Hydroclone Test Results
A limited series of tests was conducted in September and November 1977 to
determine the suitability of using a hydroclone as a dewatering device for
the FGD slurry. Tests with a small Dorr-Oliver Doxie 5 hydroclone were
abandoned early in the program because of plugging problems. Tests with a
larger porcelain Dorr-Oliver Dorrclone P50A unit (nominal 15 gpm capacity)
were more successful.
For the Dorrclone P50A hydroclone, underflow solids concentration was excel-
lent, at about 55 weight percent, when oxidized limestone slurry with high
fly ash loading was used. However, the overflow contained unacceptably
high solids concentration at 2 to 3 weight percent. Chemical analysis showed
that the fines in the overflow were predominantly fly ash, in the range of
60 to 70 weight percent. The feed slurry concentration was 15 weight percent.
The pressure drop across the unit ranged from 10 to 29 psi at 11 to 19 gpm
feed rates. About 80 percent of the feed flow went to the overflow. Size
distribution analysis by hydrometer method showed that, within the range of
tests, the pressure drop and feed rate did not appreciably affect the particle
size distributions in the overflow and underflow. The underflow size distri-
bution was only slightly higher than the feed. The overflow solids were
about 90 percent less than 15 microns.
Based on the limited test results, the hydroclone is judged to be unsuitable
as a dewatering device for waste lime/limestone slurry because of high solids
content in the overflow. Relatively high energy requirements (high pressure
drop) and possible erosion problems would also make it unattractive for this
service.
1-30
-------
1.15.6 Scanning Electron Microscope (SEM) Analysis of Oxidized Solids
Solids samples from the venturi/spray tower two-scrubber-loop forced-oxidation
system were observed under the scanning electron microscope. These samples
were taken from a lime test with high fly ash loading. The spray tower solids
(10 percent oxidized) were primarily calcium sulfite hemihydrate crystals of
a few microns in size in the shape of plates and rosettes. Solids from the
venturi loop, system bleed, clarifier underflow, and filter cake (all 95 per-
cent oxidized) were primarily calcium sulfate dihydrate ranging in size
from 20 to 100 microns. These solids were chunky and exhibited no appreciable
change in size or morphology once the oxidation step was completed.
1.15.7 Gypsum Crystallization Study
Some theoretical studies were conducted on the application of gypsum crystal-
lization technology to the flue gas desulfurization systems. Because of
larger interstitial voids, large crystals (i.e., gypsum solids) will filter
more easily than small ones. Therefore, it is desirable to increase the
gypsum crystal size and reduce the number of crystal nuclei.
There are two ways nuclei can form:
• Primary Nucleation. A spontaneous clustering of solute mole-
cules into larger stable aggregates in a supersaturated solution.
• Secondary Nucleation. The formation of nuclei by crystal-crystal
and crystal-sol ids collisions. This phenomenon becomes increas-
ingly important as both turbulence and slurry solids content
increase.
The modeling assumed that the primary nucleation obeys a "power law" model and
that the secondary nucleation is a function of slurry solids content.
1-31
-------
On the basis of limited data from the venturi/spray tower two-scrubber-loop
forced-oxidation, the modeling results suggested the following: when the
oxidation tank solids residence time was halved by reducing the slurry
solids content from 15 percent to 8 percent, the primary nucleation is
predominant in the venturi oxidation tank. The data showed that both the
final settled solids content and the initial settling rate decreased when
the solids residence time and the slurry solids concentration were halved.
1.16 OPERATING EXPERIENCE
Section 23 presents the operating experience at Shawnee during lime and lime-
stone testing. A brief summary of this experience is given below.
1.16.1 Scrubber Internals
The strong correlation between high alkali utilization and clean mist elimi-
nator operation has been further confirmed during the report period. From
June 1977 to the end of the report period in June 1978, the venturi/spray
tower mist eliminator was cleaned only once (in December 1977) after 4183
operating hours. Since that time the mist eliminator has operated for 3308
hours without plugging till the end of the report period. The TCA mist
eliminator was cleaned in June 1978 at the end of the report period after
7671 operating hours.
The 316L stainless steel TCA support grids have been in slurry service for
about 5 years with no evidence of significant erosion.
1-32
-------
Six-gram, 1-5/8 inch diameter solid nitrile foam spheres continued to give
satisfactory service with no significant failures. Acceptable levels of
initial ball shrinkage and wear have been observed. The average rate of
decrease in diameter was 0.07 inch per 1000 hours in 7700 hours of service.
The continual decrease in diameter required periodical addition of new
spheres to maintain the desired static bed height.
The four slurry spray nozzles in the TCA (Spraco No. 1969F, 316L stainless
steel) continued to give excellent service at 5 psi pressure drop both in
terms of wear and nonclogging characteristics.
Bete No. ST48FCN, stellite-tipped spray nozzles have been used successfully
as the spray tower slurry nozzles. New nozzles were installed in December
1976. Visual inspection showed negligible pitting and relatively uniform
wear after 2649 hours of service.
Venturi internals continued to erode significantly in several locations,
especially the venturi plug guide vanes and the shroud. However, the
operability of the venturi has not been seriously affected. In June 1978,
a commitment was made to fabricate and install a new venturi plug drive
mechanism. The shroud splash guard was rebuilt. A new shaft was fabricated
for the plug limit switch mechanism and new guide vanes and blocks were
installed. The old shaft had been in service since March 1972.
A trap-out funnel was installed in the spray tower below the bottom spray
header for the two-scrubber-loop forced-oxidation testing. Accumulations
of solids on the underside (gas side) of the funnel have been a common
occurrence. Spray tower operation, however, has not bean ->;!vor-33ly affected.
1-33
-------
1.16.2 Oxidlzers
The air sparger in the venturi/spray tower oxidation tank consisted of an
octagonal ring about 4 ft in diameter and made of 3 inch diameter 316L
stainless steel pipe sections welded together. The first air sparger had
130 1/8-inch diameter holes on the underside. This was later replaced by
one with 40 1/4-inch diameter holes. Some erosion and plugging of the air
holes were experienced, particularly with the 1/8-inch holes. On October 4,
1977, the 40 1/4-inch hole air sparger ring was removed and replaced with
a single 3-inch diameter pipe extending to the center of the tank and dis-
charging air downward through an elbow. No plugging or erosion was observed
for the 3-inch air pipe. Huraidification of the air was not necessary when
the air pipe was used.
During December 1977, the 40 1/4-inch hole sparger previously used in the
venturi/spray tower system was installed in the TCA oxidation tank. Because
of insufficient air compressor capacity, only a limited number of oxidation
tests were conducted on the TCA with this air sparger.
The air eductor used on the TCA system for forced oxidtion was a Model ELL-10
special eductor manufactured by Penberthy-Houdaille Industries. The unit was
normally operated at 1600 gpm slurry flow. The materials of construction
were stellite for the nozzle and whirler and neoprene-lined carbon steel for
the eductor chamber and exit throat. The neoprene lining in the jet throat
was observed to be chipped off after only 620 hours of service in limestone/
fly ash slurry. After about 1800 hours of operation, bare carbon steel was
exposed. Tests were terminated after 2055 hours when an epoxy patch on the
bare carbon steel failed and the bare carbon steel body eroded through. The
1-34
-------
stellite nozzle and whirler showed only minor evidence of erosion.
1.16.3 Reheaters
The fuel-oil-fired reheaters with external combustion chamber were manufac-
tured by Bloom Engineering Co. These reheaters have been in service since
March 1974 in the venturi/spray tower and May 1975 in the TCA. Both units
have operated reliably with only occasional flameout and equipment problems,
During the test period, accumulation of dried slurry solids was common in
and around the old Hauck inline reheater shell.
1.16.4 Fans
The reliability of the 316L stainless steel induced-draft fans, manufactured
by Zurn Industries, has been fair to good during the current operating period.
Maintenance has included occasional steam cleaning and replacement of out-
board bearings and pillow blocks five times for the two fans during the
last 18 months.
1.16.5 Pumps
The Allen-Sherman-Hoff slurry pumps were neoprene-lined pumps with Nelson
liquid drive. The pump packing consisted of graphite impregnated asbestos,
the sleeves were chrome-plated steel, and the pump seals were of air-flush
type (Centriseal). The pumps provided satisfactory service with a moderate
maintenance requirement. The most frequent reason for maintenance was re-
packing. Generally, wide flow variations called for by the test conditions
1-35
-------
and operations at low flow rate accelerated the repacking frequency.
The positive-displacement variable-speed Moyno pumps were used for fresh
lime and limestone slurry feed. Periodic replacements of the worn-out stator
and rotor were required.
1.16.6 Alkali Addition Systems
The lime slaker system has given good reliability over the last 18 months of
intermittent operation. Maintenance consisted principally of periodic slaker
cleaning and the replacement of the slaker grit screen seven times. The
screen slaker motor, drive shaft, and bearings were replaced once in Febru-
ary 1978. The process control elements were cleaned and serviced in March
1978.
The limestone dry-grinding system, installed in 1970, has given acceptable
performance with moderate maintenance requirements, considering the age of
the system.
1.16.7 Solids Dewatering System
The Ametek 3 ft diameter x 6 ft rotary drum vacuum filter required moderate
maintenance, except for the filter cloth. During the 19 month report period,
the drum speed control was repaired 3 times, the snap-blowback cake discharge
mechanism was repaired twice, and the filter cloth was replaced 15 times.
Operating experience indicated a relationship between cloth life and the
technique by which new cloth was fitted to the drum. A controlled amount
of looseness in the fit between the dividers appeared to be desirable for
1-36
-------
cake discharge and non-blinding.
The Bird 18 inch x 28 inch solid bowl continuous centrifuge was disassembled
and inspected on March 22, 1977, after about 4000 hours of operation since
the last service by Bird Machine Co. Major wear sites were found at the
two center feed ports at the periphery of the screw conveyor, and at the
plows on the solids discharge end. Limited stellite hardface welding was
performed onsite at that time. The feed pipe was replaced in March 1978.
Another inspection in June 1978, 6460 hours after the previous servicing, re-
vealed serious wear in several places, requiring factory repair. The current
plans call for factory repair by Bird Machine Co. using tungsten carbide hard-
facing on the conveyor tips instead of the stellite used previously.
1.16.8 Instrument Operating Experience
Service requirements for the Uniloc Model 321 L submersible pH electrode
assemblies consisted of periodic cleaning and buffering of the electrodes,
generally every 2 or 3 days, to ensure accuracy. The electrodes were placed
in 50°C water bath to minimize the effect of thermocycling and thus service
requirements when they were not in use.
The Du Pont Model 400 UV split-beam photometers for flue gas S0£ concentra-
tion measurement have been accurate and reasonably trouble-free in the last
18 months of service. Cleaning of the sample cell and sample line was re-
quired once every 1 or 2 months, and the particulate filter once every 3 to 4
weeks. Ultraviolet lamp failure occurred infrequently and was caused by uncon-
trollable momentary power fluctuations due to the switching of station power.
1-37
-------
Performance of the Teledyne Model 9500 inlet flue gas oxygen analyzer was
unacceptable. Frequent failure of the micro fuel cell was the major problem.
Attempts by Du Pont technicians to convert a Du Pont Model 400 UV photometer
for S02 to analyze N0£ were unsuccessful.
Foxboro 2800 series and 1800 series magnetic flow meters operated reliably
with acceptable accuracy. Periodic scale cleaning was required to improve
accuracy and sensitivity.
Both Dynatrol Model CL-10HY U-tube density meters and Ohmart radioactive den-
sity meters provided acceptably accurate and dependable service.
1.16.9 Materials Evaluation
The TCA tower rubber lining was generally in good condition although the sur-
faces where scale was often removed exhibited scabs from chipping tools. There
was no evidence of rubber lining separation from the metal wall.
Schedule 80 PVC 1120 (ASTM D-1785) piping of 6-inch and 8-inch sizes has per-
formed satisfactorily in the TCA main slurry line service.
Blisters were observed in some sections of the rubber-lined pipe during a
March 1977 inspection. These blisters have not posed an operating or main-
tenance problem to date.
A test panel (20 x 15 x 3/8 inch) of carbon steel coated with Fluorelastomer
CXL-2000, made by Pullman-Kellogg, failed after about 2000 hours of service in
the TCA inlet gas duct where flue gas cooling slurry spray impinged across the
panel face.
1-38
-------
A section of U-shaped Bondstrand FRP pipe was in excellent condition after
about 8800 hours of service in the suction line of the TCA main slurry pump.
The Valtek 6-inch butterfly control valve was judged conditionally acceptable
for slurry service. The valve disk does wear and will need replacement.
The Fisher 2-inch butterfly control valve has been in intermittent service
since September 1976. The fishtail disk was judged to be in excellent condi-
tion. The valve body appeared to be wearing faster than the disk in the
throttling service.
The Durco 6-inch butterfly valve, installed in late 1975, was in excellent con-
dition after about 3 years of slurry service. The valve service was primarily
in the open position.
Visual inspection of the 8-inch Hayward strainers in the TCA slurry service
showed further erosion at the basket support ledge and at the lower half of
the strainer outlet during an inspection in November 1977. Similar erosion
occurred in the 6-inch size units in the venturi/spray tower.
The Elliott strainers in the venturi/spray tower slurry line did not show
further erosion of the strainer body.
1-39
-------
Section 2
INTRODUCTION
In June 1968, a program was initiated under the direction of the Environ-
mental Protection Agency (EPA)* to test prototype lime and limestone wet-
scrubbing systems for removing S02 and particulates from flue gases. A
test facility was built which operates with flue gas from coal-fired Boiler
No. 10 at the Tennesee Valley Authority (TVA) Shawnee Power Station, Paducah,
Kentucky. Eechtel Corporation of San Francisco was the major contractor
and test director, and TVA was the constructor and facility operator.
Initially, the test facility consisted of three parallel scrubber systems:
a venturi followed by a spray tower, a Turbulent Contact Absorber (TCA),
and a Marble-Bed Absorber. Testing of the Marble-Bed Absorber was dis-
continued in July 1973 because of operational problems. These systems were
chosen for their ability to remove both S02 and particulates from the burning
of medium- to high-sulfur coal. Each system has a capacity of approximately
10 MW equivalent of flue gas (35,000 acfm 0 300°F for the venturi/spray tower
and 30,000 acfm @ 300°F for the TCA). The systems operate with flue gas con-
taining 1500 to 4500 ppm of S0£ obtained either upstream (3 to 6 grains/dry
scf of particulates) or downstream (0.04 to 0.20 grain/dry scf of particulates)
from Boiler No. 10 particulate removal equipment.
* The National Air Pollution Control Administration prior to 1970
M
-------
The original testing program at the facility lasted from March 1972 to
October 1974. Results are presented in the Summary of Testing through
October 1974 (Reference 1). During the original program, emphasis was placed
on demonstrating reliable operation. The most significant reliability problem
was associated with scaling and/or plugging of mist elimination surfaces. The
TCA mist elimination system consisted of a wash tray in series with a chevron
mist eliminator, both with underside washing. Long-term operability of this
system was demonstrated in limestone service in an 1835-hour test at a scrubber
gas velocity of 8.6 ft/sec.* The venturi/spray tower mist elimination system
consisted of a chevron mist eliminator with underside washing. At the end of
the original testing program, long-term operability of the venturi/spray tower
had not been demonstrated.
In June 1974, the EPA, through its Office of Research and Development and
Control Systems Laboratory, initiated a 3-year Advanced Test Program at the
Shawnee Facility. Bechtel Corporation continued as the major contractor and
test director, and TVA as the constructor and facility operator.
The results of advanced testing from October 1974 through April 1975 at the
Shawnee Facility are presented in a First Progress Report (Reference 2).
During this period, successful operation of a chevron mist eliminator with
intermittent top and bottom wash was demonstrated in the venturi/spray tower
system in lime service at a superficial scrubber gas velocity of 8.0 ft/sec.
In the TCA in limestone service, plugging of the combined wash tray/mist elimi-
* In this report, all gas velocities and liquid-to-gas ratios are at scrubber
operating conditions, i.e., saturated gas at scrubber temperature. The
scrubber temperature is approximately 125°F under normal operating conditions.
The gas velocities are all superficial velocities.
2-2
-------
nator system was a continual problem at velocities greater than 8.6 ft/sec.
Tests were interrupted in May 1975 for a 6-week scheduled maintenance outage
on Boiler No. 10.
The results of advanced testing at the Shawnee Facility from June 1975
through mid-February 1976 are presented in a Second Progress Report (Refer-
ence 3). During this period, the wash tray/mist eliminator system on the
TCA was discontinued in favor of a chevron mist eliminator similar to the one
used in the venturi/spray tower. Roth systems usually operated at their maximum
fan capacity (9.6 ft/sec superficial gas velocity in the spray tower and 12.5
ft/sec in the TCA). Significant results were:
• The discovery that the mist eliminator is more easily kept clean at
high alkali utilization
• The demonstration that above 85 percent alkali utilization an inter-
mittent bottom wash will keep the mist eliminator free of soft solids
(raw water at 1.5 gpm/ft for 6 minutes every 4 hours)
• The demonstration that below 85 percent alkali utilization a continu-
ous bottom wash will limit soft solids accumulation to less than 10
percent mist eliminator restriction (combined raw water and clarifier
return liquor at 0.4 gpm/ft )
• The demonstration that intermittent top wash prevents scale forma-
tion on top mist eliminator blades (raw water for 4 minutes every 8
hours at 0.5 gpm/ft )
• The demonstration on the TCA that three hold tanks in series improves
limestone utilization. At 85 percent S02 removal, limestone utiliza-
tion was improved from about 60 percent with a single hold tank to
about 75 percent with three tanks in series
• The demonstration in an 1143-hour run that the venturi/spray tower in
lime service can operate smoothly with a gas rate following a daily
boiler cycle from 40 to 100 percent of full load
The results of advanced testing from mid-February 1976 through November 1976
are presented in a Third Progress Report (Reference 4). During this period,
short (6 to 8 hours) factorial tests were run on both scrubber systems using
2-3
-------
lime slurry, limestone slurry, and limestone slurry with added MgO to enhance
S0£ removal. Both these factorial tests and the longer tests discussed below
were fitted to mathematical models for predicting SC^ removal as a function
of major process variables, such as liquid-to-gas ratio, scrubber inlet liquor
pH, and effective magnesium ion concentration.
On both systems, longer runs of 4 to 8 days each were made to explore operating
variables under several modes of operation. These included runs with lime and
limestone alkali, with and without added MgO, using flue gas with and without
fly ash.
In addition, a series of venturi/spray tower lime runs of 2 to 6 days each
allowed collection of scrubber inlet and outlet data characterizing flue gas
mass loading, particle size distribution, and SOg concentration.
The Third Progress Report also presented an introductory discussion of lime/
limestone wet scrubbing chemistry.
This Report presents the results of testing at the Shawnee Facility from
November 1976 through June 1978. Test runs during this period were usually
of about 2 to 8 days' duration.
During this period, an emphasis was placed on forced-oxidation testing in which
calcium sulfite solids (CaSf^'l/Zl^O) are converted to gypsum (CaS04'2H20) by
contacting the slurry with air. In forced-oxidation testing, 95 mole percent
or more of the SC^ in the waste solids product is in the form of gypsum.
These high-gypsum solids, with or without fly ash, have better dewatering and
solids handling characteristics than solids containing large amounts of unox-
idized calcium sulfite. Forced-oxidation solids from the Shawnee rotary drum
2-4
-------
vacuum filter typically contain only about 15 percent free moisture, compared
with 40 to 50 percent for unoxidized solids. The forced-oxidation solids re-
semble moist soil having reasonable bearing strength, whereas unoxidized solids
are thixotropic. The initial settling rates of forced-oxidation solids are
about 10 times higher than those for calcium sulfite solids.
Forced oxidation v/as accomplished at Shawnee either by air sparging a slurry
tank or by pumping slurry through an air eductor. The oxidation configurations
tested included staged scrubbers (venturi and spray tower operating with separ-
ate effluent hold tanks with oxidation of venturi hold tank slurry), one stage
(TCA) oxidation, and bleed stream oxidation. All of these oxidation equipment
configurations are discussed in Sections 5, 8, 12, and 15.
The venturi/spray tower was used almost exclusively for forced-oxidation test-
ing during this reporting period. Most venturi/spray tower runs were limestone,
lime, or limestone/MgO tests with two-stage forced oxidation by air sparging
the venturi effluent hold tank. These runs were made with both low and high fly
ash loadings. During May and June 1978, a few additional limestone/MgO tests
were performed with bleed stream oxidation by air sparge. In March 1977, three
lime bleed stream oxidation tests were performed using an eductor; these latter
runs were unsuccessful.
About 25 percent of the TCA operation was devoted to forced-oxidation testing,
all with limestone or limestone/MgO. Because EPA pilot studies indicated that
calcium sulfite solids are needed for adequate S02 removal in a one-stage forced-
oxidation system, lime testing with forced oxidation was not performed on the
TCA.
2-5
-------
Prior to December 1977, oxidation on the TCA system was accomplished by an
eductor. Bleed stream eductor oxidation tests performed between March and
May 1977 proved unsuccessful. Subsequent eductor oxidation of scrubber hold
tank slurry was successful. However, air sparging of hold tank slurry was
found to be cheaper, more reliable, and less energy intensive than eductor
oxidation. Hence, TCA forced-oxidation tests made after December 1, 1977
were with hold tank sparging.
Testing of the TCA without forced oxidation included the following:
• Runs with limestone, limestone/MgO, lime, and lime/MgO
0 Limestone runs with low fly ash loading
t Limestone type and grind tests using coarse limestone from
the Longview, Alabama quarry, and coarse and fine limestones
from the Fredonia, Kentucky, quarry
• Automatic limestone feed control tests
• Tests with Ceilcote support plate packing instead of spheres
t Flue gas characterization tests with limestone to determine
the mass loading, particulate size distribution, and 503
concentration for both the scrubber inlet and outlet
During a 10-week Boiler No. 10 outage in April and May 1977, air/slurry tests
were performed on both systems to quantify the effect of slurry entrainment
on scrubber mass emissions. Three one-month long term tests were made. One
of these, Run 717-2A, was on the TCA system with limestone and without forced
oxidation. The other two were on the venturi/spray tower with two-stage
forced oxidation, Run 819-1A with limestone alkali and Run 863-1A with lime.
All three of these tests were at variable gas load and high fly ash loading to
simulate commercial operation.
2-6
-------
Other special topics covered in this report include:
• Laboratory limestone reactivity tests
• Spray tower flue gas/slurry and TCA air/water pressure drop
correlations
• Evaluation of trace elements
• Evaluation of waste solids dewatering characteristics for
slurries with and without forced oxidation. This includes
laboratory settling and filtration tests, special separation
equipment (Lamella inclined plate settler and hydroclone),
scanning electron microscope of solids, and a gypsum
crystallization study
2-7
-------
Section 3
TEST FACILITY
Two parallel scrubbing systems were being operated during the Advanced Test
Program. Scrubbers incorporated in these systems were:
9 A venturi followed by a spray tower
• A Turbulent Contact Absorber (TCA)
Each system operates independently and is capable of treating approximately
30,000 acfm of flue gas from the coal-fired Boiler No. 10 at TVA's Shawnee
Power Station. This gas rate is equivalent to approximately 10 MW of power
plant capacity.
Boiler No. 10 normally burns a medium- to high-sulfur bituminous coal which
produces S02 concentrations of 1500 to 4500 ppm in the flue gas. Ductwork
to th? scrubbers is tied in both upstream and downstream of the Boiler No. 10
particulate removal equipment allowing scrubber operation on flue gas with
high fly ash loadings (3 to 6 grains/dry scf) or low fly ash loadings (0.04
to 0.20 grain/dry scf).
3.1 SCRUBBER SELECTION
The major criterion for scrubber selection was the potential for removing both
S02 and particulates at high efficiencies (defined for the Shawnee Facility as
3-1
-------
S02 removal greater than 80 percent and particulate removal greater than 99
percent). Other criteria considered in the selection of the scrubbers were:
• The ability to handle slurries without plugging or excessive
seali ng
• Reasonable cost and maintenance
• Ease of control
• Reasonable pressure drop
The venturi/spray tower and the TCA were chosen to meet these criteria.
The adjustable-throat venturi scrubber in the venturi/spray tower system was
manufactured by Chemical Construction Company (Chemico). The venturi scrubber
removes the bulk of the particulates. But because the residence time in a
venturi scrubber is low, S02 removal with lime/limestone slurry is inadequate.
The spray tower that follows the venturi scrubber provides additional capacity
for the removal of S02-
The TCA was manufactured by Universal Oil Products (UOP). It operates with up
to three beds of low-density spheres, nominally 1-1/2 inches in diameter, that
are free to move between retaining grids. As the incoming flue gas contacts
the slurry in these beds, both S02 and particulates are removed.
Figures 3-1 and 3-2 (drawn with major dimensions to scale) show the two scrub-
ber systems and the mist eliminators. The chevron mist eliminator used on
the two scrubber systems during this testing period is depicted to scale in
Figure 3-3. In two-stage forced-oxidation testing on the venturi/spray tower
system, a catch funnel as shown in Figure 3-1 is used to separate the spray
tower effluent slurry from the venturi effluent slurry.
3-2
-------
GAS OUT
CHEVRON MIST
ELIMINATOR
SPRAY TOWER
INLETSLURRY
MIST ELIMINATOR
WASH WATER
MIST ELIMINATOR
WASH LIQUOR
ADJUSTABLE PLUG
VENTURISCRUBBER
x"
APPROX. SCALE
VENTURI EFFLUENT SLURRY
SPRAY TOWER
EFFLUENT SLURRY
Figure 3-1. Schematic of Venturi Scrubber and Spray Tower
-------
GAS OUT
MIST ELIMINATOR
WASH WATER
CHEVRON MIST
ELIMINATOR
RETAINING
BAR-GRIDS
GAS IN
A AA
O .
°° 0°0
o o o
3*9.2 Q.O
o
o o ° o
MIST ELIMINATOR
WASH LIQUOR
INLET SLURRY
MOBILE PACKING
SPHERES
5'
I, 1
APPROX. SCALE
EFFLUENT SLURRY
Figure 3-2. Schematic of Three-Bed TCA
3-4
-------
THREE-PASS, OPEN-VANE. 316L SS
CHEVRON MIST ELIMINATOR
(HORIZONTAL CONFIGURATION)
GAS FLOW
6 in.
Figure 3-3. Test Facility Mist Eliminator Configuration
3-5
-------
The cross-sectional area of the spray tower is 50 ft2 in both the scrubbing
section and the mist eliminator section. The cross-sectional area of the TCA
scrubber is 32 ft2 in the scrubbing section and 49 ft2 in the mist eliminator
section.
3.2 SYSTEM DESCRIPTION
The Shawnee Test Facility contains five major areas:
t The scrubber area (including tanks and pumps)
• The operations building area (including laboratory area,
electrical gear, centrifuge, and filter)
• The thickener area (including tanks and pumps)
• The utility area (including air compressors, air dryer,
limestone storage silos, mix tanks, gravimetric feeder,
and pumps)
• The pond area
The test facility has been designed so that a number of different scrubber
internals and piping configurations can be used with each scrubber system.
For example, the TCA scrubber can be operated with one, two, or three beds
of spheres or with only the support grids. Waste solids separation can be
achieved with a clarifier alone or with a clarifier in combination with a
filter or a centrifuge. Either system can be operated with a single hold
tank or with up to three tanks in series.
The Advanced Program tests have been conducted in closed-liquor-loop opera-
tion. A closed liquor loop is achieved when the amount of makeup water input
to the system is equal to the amount of water normally leaving the system via
3-6
-------
the discharge solids and the humidified flue gas. In this report, it is
assumed that a closed liquor loop is achieved when the discharge solids con-
centration is 38 weight percent or higher and no separate liquor is purged.
In tests with forced'oxidation, solids discharge concentrations of 80 percent
and higher were routinely achieved with no separate liquor purge, creating an
extremely tight water balance.
Flow diagrams depicting scrubber loop configurations for forced-oxidation
testing are presented in subsequent sections of this report. For the venturi/
spray tower system, Figure 5-1 shows the two-scrubber-loop configuration, and
Figure 8-1 shows the configuration for bleed stream oxidation. For the TCA
system, Figures 12-1 and 12-2 show one-tank and two-tank configurations for
oxidation by an eductor, and Figures 15-1 and 15-2 show the corresponding con-
figurations for oxidation by sparging.
Figure 3-4 is a flow diagram of the solids dewatering system for either the
venturi/spray tower or the TCA. Typically, the venturi/spray tower system
has a clarifier followed by a filter, and the TCA has a clarifier followed
by a centrifuge.
In May and June 1976, the ductwork to the scrubber systems was modified so
that gas can be withdrawn from the boiler either before or after the steam
plant particulate removal equipment. In the former configuration, all the
entrained particulate matter (fly ash) is introduced into the scrubber; in
the latter, the gas to the scrubber contains only a residual amount of fly
ash. The gas flow rate to each scrubber is measured by venturi flow meters
and controlled by dampers on the induced-draft fans. The concentration of
S02 in the inlet and outlet gas streams is monitored continuously by Du Pont
3-7
-------
BLEED FROM
SCRUBBER SYSTEM
CO
i
00
PROCESS
WATER
HOLD
TANK
TO
SCRUBBER
SYSTEM
FILTER
OR
I CENTRIFUGE
FILTRATE
OR
CENTRATE
}
WATER
RESLURRY
TANK
POND
Figure 3-4. Flow Diagram for SolIds Dewaterlng System
-------
photometric analyzers.
The scrubbing systems are controlled from a central graphic panelboard, where
all significant process variables are on digital display. Important process
control variables are continuously recorded. Trend recorders are provided for
periodic monitoring of selected data sources. Chemical composition of major
streams and scrubber inlet liquor pH are determined several times a day.
3.3 EPA PILOT PLANT
There are two smaller scrubbing systems (300 acfm each) at the EPA Industrial
Environmental Research Laboratory in Research Triangle Park, North Carolina.
These small pilot-scale scrubber systems are capable of simulating the Shawnee
scrubber systems with excellent agreement in the lime/limestone wet-scrubbing
chemistry. Preliminary data are generated on the pilot-scale system to deter-
mine the validity of new concepts and to guide the selection of those promising
concepts that should logically be investigated on the larger scale Shawnee
units. Examples of studies originating at the EPA pilot plant and then later
investigated at the Shawnee Test Facility include gypsum unsaturated operation,
increasing limestone utilization, forced oxidation, and magnesium-enhanced lime
and limestone scrubbing.
3-9
-------
Section 4
ADVANCED TEST PROGRAM
This section contains a description of the Shawnee Advanced Test Program,
which began in June 1974 and extended through June 1978.
4.1 ADVANCED TEST PROGRAM OBJECTIVES AND COMPLETED SCHEDULE
The objectives of the Advanced Test Program are:
To demonstrate process reliability, with an emphasis on mist
elimination systems
To investigate advanced process and equipment design variations
for improving system reliability, economics, and energy require-
ments
To evaluate process variations for a substantial increase in
alkali utilization for limestone systems
To evaluate the effect of increased magnesium ion concentration
on improving control, reducing gypsum saturation, and increasing
S02 removal
To evaluate the efficiency and reliability of lime and limestone
scrubbers under conditions of widely varying flue gas flow rate
and inlet S02 concentration
To evaluate system performance and reliability without fly ash in
the flue gas
To determine the effectiveness of forced oxidation in producing
an improved throwaway sludge product
To determine the practical upper limits of S02 removal efficiency
for both limestone and lime scrubbing systems
4-1
-------
• To perform reliability demonstration runs on advanced process
and equipment design variations
• To characterize stack gas emissions, including inlet and outlet
particulate mass loading and size distribution, slurry entrain-
ment, and total sulfate emissions
0 To evaluate methods of automatic control
• To provide material for the field evaluation of sludge disposal
processes (the field evaluation is being conducted under a
separate EPA program)
• To evaluate corrosion and wear of alternative plant equipment
components and materials
t To develop a computer program for the design and cost comparison
of full-scale lime and limestone systems
A bar chart of the Advanced Test Program, showing the time spent on each of
the defined objectives, is presented in Figure 4-1. The period covered by
the Fourth Progress Report was from November 1976 through June 1978.
4.2 ANALYTICAL PROGRAM
During the testing, samples of slurry, flue gas, limestone, lime, and coal were
taken periodically for chemical analyses, and samples of flue gas were taken for
particulate mass loading determinations. A summary of the analytical methods
for determining important species in the slurry solids and slurry liquor is pre-
sented in Table 4-1. A laboratory procedures manual (Reference 5) was issued in
March 1976. A listing of the compositions of the raw material used in the test-
ing program is presented in Appendix C.
Four Du Pont photometric analyzers were used for continuous S02 gas analyzing
one each at the inlet and outlet of each scrubber. Scrubber inlet liquor pH
4-2
-------
Figure 4-1
Schedule for Advanced Test Program
-------
Table 4-1
FIELD METHODS FOR BATCH CHEMICAL ANALYSIS OF
SLURRY AND ALKALI SAMPLES
SPECIES
Sodium
Potassium
Calcium
Magnesium
Sulflte
Total sulfur
Carbonate
Chloride
FIELD METHODS
SOLIDS
Primary Method Backup Method
X-ray fluorescence'1' Atomic absorption
..
Amperometrlc tltratlon
X-ray fluorescence^1' Tltratlon
C02 evolution
LIQUIDS
Primary Method Backup Method
Atomic absorption
Atomic absorption
Atomic absorption
Atomic absorption
Amperometrlc tltratlon
Tltratlon
Potent lometrlc tltratlon Mercuric nitrate tltratlon
(1) The X-ray fluorescence method was discontinued 1n August 1977. Since that time, the primary methods have been atomic absorption for solid
calcium, and tltratlon for total sulfur In solids.
-------
was continuously monitored with Universal Interloc pH analyzers. Both scrubber
inlet liquor pH and outlet liquor pH are monitored periodically (every 2 hours
for inlet, and every 8 hours for outlet) by the laboratory. A modified EPA
particulate train (manufactured by Aerotherm/Acurex Corporation) was used to
measure mass loading at scrubber inlets and outlets.
4.3 DATA ACQUISITION AND PROCESSING
Data recorded by onsite personnel were sent to the Bechtel home office in
San Francisco. Bechtel home office personnel analyzed, summarized and plotted
the data on a run-by-run basis.
An onsite analytical data acquisition system, designed and (in part) installed
by Radian Corporation, recorded the results of laboratory analyses on printed
summary sheets. A minicomputer was used to perform important calculations and
print the resultant data.
Onsite, on a daily basis, selected raw analytical and operating data for each
8-hour shift were entered into a National CSS, Inc. (NCSS) computer data file
and were verified. San Francisco personnel accessed this data file and loaded
it into a NOMAD database on the NCSS system. The database stored the data in
a hierarchial order according to scrubber, test run, time, and sample point
number. User-written computer programs associated with the database performed
further calculations on the raw data, e.g., calculation of percent sulfite
oxidation and stoichiometric ratio from solids analytical data. Other user-
written programs prepared tables that presented the data for specified times or
presented run-averaged data for a specified test sequence. The database run
4-5
-------
summary tables for November 1976 through June 1978 are presented in Appendix
D. Selected hourly test data and scrubber inspection reports were transmitted
daily by telecopier from the test facility to San Francisco. Daily data
packages* which provided detailed documentation of each day's operation, were
mailed to San Francisco.
4-6
-------
Section 5
VENTURI/SPRAY TOWER TWO-SCRUBBER-LOOP
FORCED-OXIDATION LIMESTONE TEST RESULTS
In lime and limestone wet-scrubbing systems for removing S0£ and participate
from coal-fired boiler flue gas, disposal of the waste solids product has
been a major problem both technically and economically.
The waste solids consist primarily of calcium sulfite hemihydrate, calcium
sulfate dihydrate (gypsum), and fly ash. The relative amounts of sulfite
and sulfate depend on the degree of oxidation in the scrubber system. In
most medium to high-sulfur coal applications, as in the Shawnee Test Facil-
ity, natural oxidation of sulfite to sulfate in the scrubber system amounts
to only 10 to 30 percent, and calcium sulfite is the predominant material
in the waste sludge.
Calcium sulfite wastes present a serious disposal problem because of the
difficulty of dewatering. The slurry can be dewatered to only about 50 to
60 percent solids, producing an unstable, thixotropic material unsuitable
for landfill. Where space is available, ponding of the untreated sulfite
sludge has been practiced. Rut the pond area may be impossible to reclaim,
and in many locations sufficient space is not available.
Beginning in April, 1975, studies conducted by the EPA at the 0.1-MW IERL-RTP
pilot plant (Reference 6) have shown that calcium sulfite can be readily
5-1
-------
oxidized to gypsum by simple air/slurry contact in the hold tank of the
scrubber slurry recirculation loop. The oxidized slurry settles by a
factor of 10 faster than the unoxidized, sulfHe-base slurry, and can be
easily dewatered to 80 to 90 percent solids, producing a more tractable
material resembling moist soil which may be and suitable for landfill.
Furthermore, because of the superior solids dewatering properties, forced
oxidation results in a smaller volume of waste solids (even though the molec-
ular weight of gypsum is higher than that of calcium sulfite hemihydrate),
and hence lower solids disposal cost.
Preliminary economic evaluations by TVA (Reference 7) have shown that forced
oxidation is a more economical method of producing landfill materials than
other methods, such as commercial fixation of sulfite sludge with additives
and blending sulfite sludge with fly ash.
In the United States, where natural gypsum is widely available, gypsum pro-
duced from scrubber systems may be unable to compete extensively with natural
gypsum as raw material for the cement and wall board industries. Thus, the
incentive in the United States has been to develop simplified forced-oxidation
procedures directed only toward improving waste solids handling and disposal
properties. As a disposal material, the gypsum sludge can have high fly ash
content; moreover, the oxidation reaction need be carried only to about 95
percent completion.
Based on the findings at the 0.1-MW IERL-RTP pilot plant, a program was ini-
tiated in January 1977 at the Shawnee Test Facility to develop procedures
for forced oxidation on the larger 10-MW prototype scrubbers. This program
has since continued as the major part of the Shawnee Advanced Test Program.
5-2
-------
The results of forced-oxidation testing on the venturi/spray tower system
are presented in this section and Sections 6 through 8. TCA forced-oxida-
tion test results are given in Sections 12 and 15.
In this section, the results of two-scrubber-loop forced-oxidation limestone
tests on the venturi/spray tower system are reported. A total of 15 runs
were made (January 4 through March 9 and October 18 through December 15,
1977) with high fly ash loading in the flue gas, including a one-month
reliability run (Run 819-1A). In addition, eight runs were conducted
(August 12 through October 4 1977), using flue gas with low fly ash loading.
Except for the one-month reliability run, each test normally lasted 5 to 6
days, which was judged to be sufficient time to reach kinetic equilibrium
and to allow collection of adequate run data.
Major test conditions and important test results for these runs are summarized
in Table 5-1. A log of the scrubber operating periods is given in Appendix
B. Properties of coal and limestone used during these tests can be found
in Appendix C. Appendix D presents computer data base run summaries. Detail-
ed operating conditions and results are summarized in Appendix E. Selected
operating data are graphically presented in Appendix F. Average scrubber
liquor compositions and the corresponding calculated percent gypsum satura-
tions are given in Appendix G.
5.1 SYSTEM DESCRIPTION
During the month of December 1976, the venturi/spray tower system was modified
to allow operation in a two-scrubber-loop configuration with forced oxidation
of calcium sulfite to calcium sulfate (gypsum) in the hold tank of the first
5-3
-------
Table 5-1
RESULTS OF FORCED-OXIDATION TESTS WITH TWO SCRUBBER LOOPS
ON THE VENTURI/SPRAY TOWER SYSTEM USING LIMESTONE SLURRY
Major Test Conditions
Fly ash loading
Gas rate, acfm 0 300°F
Venturl liquor rate, gpm
Spray tower liquor rate, gpm
Venturl percent solids redrculated (controlled)
Residence times, min: Oxidation tank^
Oesupersaturatlon tank
Spray tower EHT
Venturl Inlet (oxidation tank) pH (controlled)
Venturl pressure drop, 1n. H,0
A1r rate to oxidation tank, scfm
Clarified liquor returned to^
Selected Results
Percent S02 removal
Inlet SOg concentration, ppm
Spray tower percent solids redrculated
Spray tower Inlet pH
Spray tower limestone stolcMometrlc ratio
Spray tower Inlet liquor gypsum saturation, %
Spray tower sulflte oxidation, %
Overall sulflte oxidation, X
Overall limestone utilization, X
Venturl Inlet liquor gypsum saturation, X
Venturl Inlet liquor sulflte concentration, ppm
Air stolchlometry, atoms 0/mole SO, absorbed
/ 1 f\\ *•
Filter cake solids, wtXuu'
Onstream hours
801 -1A
High
25,000
400
1300
15
17
7
19.4
4.5
9
400<4>
S.T
62
3400
5.8
5.7
1.20
75
13
93
91
100
29
4.7
81
263
802-1 A
«H^MSw
High
25,000
400
1400
15
17
7
18
5.0
9
400<4>
S.T
60
3500
6.1
5.5
1.37
110
16
95
83
110
24
4.7
86
256
803-1A
High
25,000
600
1400
15
11.3
4.7
18
5.0<3>
9
400<4>
S.T
67
3150
5.4
5.65
1.47
120
16
96
87
105
21
4.7
81
138
804-1 A
High
25.000
600
1400
15
11.3
4.7
18
5.0
9
400<4>
V. & S.T.
71
3150
7.4^
5.8
1.45
105
15
97
88
105
22
4.4
85
151
805-1A
High
25,000
600
1400
15
11.3
4.7
18
5.0
9
400<4>
V.
78
3450
15.6
6.15
1.53
25
11
98
87
95
16
3.7
81
112
806-1 A
High
25,000
600
1400
15
11.3
4.7
18
4.5
9
250<4>
V.
73
3400
15.1
6.15
1.30
20
10
99
95
95
16
2.5
84
45
806-18
High
25.000
600
1400
15
11.3
4.7
18
4.5
9
150<4>
V.
74
2850
14.5
6.1
1.25
55
15
98
95
95
15
1.7
82
45
806-1C
High
25,000
600
1400
15
11.3
4.7
18
4.5
9
100<4>
V.
73
3250
14.2
6.0
1.25
50
17
97
96
100
15
1.0
79
54
806- ID
High
25,000
600
1400
15
11.3
4.7
18
4.5
9
50<4>
V.
80
3100
15.9
6.15
1.35
15
15
67
91
105
130
0.50
79{11)
41
Note: Footnotes for this table are listed at the end of Table 5-1 (continued).
-------
Table 5-1 (continued)
RESULTS OF FORCED-OXIDATION TESTS WITH TWO SCRUBBER LOOPS
ON THE VENTURI/SPRAY TOWER SYSTEM USING LIMESTONE SLURRY
Major Test Conditions
Fly ash loading
Gas rate, acfm @ 300°F
Venturi liquor rate, gpm
Spray tower liquor rate, gpm
Venturi percent solids recirculated (controlled)
(2)
Residence times, min: Oxidation tankv
Desupersaturation tank
Spray tower EHT
Venturi inlet (oxidation tank) pH (controlled)
Venturi pressure drop, in. HgO
Air rate to oxidation tank, scfm
Clarified liquor returned to^ '
Selected Results
Percent SOp removal
Inlet SOp concentration, ppm
Spray tower percent solids recirculated
Spray tower inlet pH
Spray tower limestone stoichiometric ratio
Spray tower inlet liquor gypsum saturation, %
Spray tower sulfite oxidation, %
Overall sulfite oxidation, %
Overall limestone utilization, %
Venturi inlet liquor gypsum saturation, %
Venturi inlet liquor sulfite concentration, ppm
Air stoichiometry, atoms 0/mole S02 absorbed
Filter cake solids, wt%(10)
Onstream hours
807- 1A
High
25,0-0
600
1400
15
11.3
4.7
18
4.5
9
0
V.
79
3350
16.6
6.25
1.34
8
7
18
91
135
555
0
66
808-1A
High
25,000
600
1400
15
11.3
4.7
18
4.5
9
/ A \
150<4>
V.
76
2850
15.5
6.25
1.20
23
16
97
98
100
19
1.7
82(1D
65
809-1A
Low
2,5,000
600
1400
15
11.3
4.7
13.4
4.5
9
150^5^
S.T
82
2450
7.9
5.7
1.30
105
21
98
98
95
35
1.85
137
810-1A
Low
25,000
600
1400
15
11.3
4.7
13.4
5.0
9
150^5^
S.T
83
2700
8.1
5.8
1.35
105
22
98
96
95
25
1.65
.(13)
162
811-1A
Low
25,000
600
1400
15
11.3
4.7
13.4
5.5
9
150^5^
S.T
85
2600
8.0
5.9
1.40
105
25
97
96
105
45
1.65
.(13)
184
812-1A 813 1A
Low Low
25,000 25,000
600 600
1400 1400
15 15
11.3 11.3
0 0
13.4 13.4
5.5 5.5
9 9
1 K\ 1 £,}
150(5) 150(5)
S.T S.T
93
2350
7.0
5.95
1.93
100
(14)
25 -u '
98
81
105
60
1.70
86
141 7
-------
Table 5-1 (continued)
RESULTS OF FORCED-OXIDATION TESTS WITH TOW SCRUBBER LOOPS
ON THE VENTURI/SPRAY TOWER SYSTEM USING LIMESTONE SLURRY
Major Test Conditions
Fly ash loading
Gas rate, acfm 9 300°F
Venturl liquor rate, gpm
Spray tower liquor rate, gpm
Venturl percent solids redrculated (controlled)
Residence times. m1n: Oxidation tank'2'
Desupersaturatlon tank
Spray tower EHT
Venturl Inlet (oxidation tank) pH (controlled)
Venturl pressure drop, In. H.O1
Air rate to oxidation tank, scfm
Clarified liquor returned to'7'
Selected Results
Percent S02 removal
Inlet SOj concentration, ppm
Spray tower percent solids redrculated
Spray tower Inlet pH
Spray tower limestone sto1ch1ometr1c ratio
Spray tower Inlet liquor gypsum saturation, %
Spray tower sulflte oxidation, X
Overall sulflte oxidation, 1
Overall limestone utilization, %
Venturl Inlet liquor gypsum saturation, %
Venturl Inlet liquor sulflte concentration, ppm
Air stolchlometry, atoms 0/mole SO, absorbed
Filter cake solids, wt*'10'
Onstream hours
814-1A
Low
25,000
600
1400
15
8.8
4.7
13.4
5.5
150<5'
S.T
91
2450
8.4
5.95
1.78
100
27
98
83
105
25
1.65
88
136
815-1A
Low
35,000
600
1400
15
8.8
4.7
13.4
5.5
S.T
86
2500
8.4
5.85
1.98
105
26
96
67
105
40
1.70
87
306
816-1A
Low
35,000
600
1400
15
11.3
4.7
13.4
5.5
2U>'5>
ST.
86
2350
7.7
5.75
1.68
105
27
98
83
100
25
1.80
86
142
817-1A
High
35,000
600
1400
15
11.3
4.7
16.8
5.5
210<6'
V.
83
2500
8.9
5.9
1.60
100
21
97
82
105
25
1.75
86
188
S16-1A
High
35.000
600
1600
15
11.3
4.7
14.7
5.5
7.5-9
210<6'
V.
86
2550
9.6
5.9
1.64
100
19
98
81
105
25
1.70
86
141
819-1A
"(?)
600
1600
15
11.3
4.7
14.7
5.5
S9
210'6'
V.
86
2950
10
5.85
1.65
100
21
98
81
100
25
1.45-2.80
87
840
819-18
3«
600
1600
15
11.3
4.7
14.7
_(9)
S9
210<6'
V.
85
3000
9.6
5.9
1.65
110
19
98
83
105
25
1.45-2.80
86
126
Notes:
Gas rate was varied from 18,000 to 35,000 acfm to follow the boiler load.
Oxidation tank level was 18 ft for all runs except Runs 814-1A and 815-1A which were at 14 ft.
Actual venturl Inlet pH averaged 4.8.
Used a sparger ring located 6 Inches from tank bottom. The sparger ring had 130 1/8-Inch holes on the bottom side.
Used a sparger ring located 6 Inches from tank bottom. The sparger ring had 40 1/4-Inch holes on the bottom side.
A1r discharged downward through a 3-Inch diameter pipe with an open elbow at center of oxidation tank about 3 Inches from tank bottom.
Spray tower loop (effluent hold tank) or venturl loop (oxidation tank).
Controlled by proper flow split of clarified liquor returned to venturl and spray tower loops.
Spray tower limestone stolchlometry controlled at 1.6.
Clarlfler and filter 1n series were used for solids dewaterlng 1n all runs except as noted.
Values may not be representative due to changes 1n oxidation In short runs.
Excludes last half of run when filter was out of service.
Clarlfler only was used.
Test with oxidation tank agitator turned off. Terminated after 7 hours because the air sparging alone did not keep the solids suspended.
-------
scrubber (venturi) slurry loop. The first scrubber (venturi) loop was
operated at a relatively low pH (4.5 optimum) to provide favorable condi-
tions for oxidation while the second scrubber (spray tower) loop was
operated at a higher pH (5.9) for good S02 removal.
A typical flow diagram of this set-up is shown in Figure 5-1. To separate
the venturi and spray tower scrubber loops, a catch funnel was installed
beneath the bottom spray header of the spray tower. To eliminate slurry
entrainment through the catch funnel, the bottom spray header had to be
turned upward.
The hold tank in the first scrubber stage (venturi) recirculation loop was
used as the oxidation tank. This tank was 8 ft in diameter and could be
operated at either a 10-, 14-, or 18-ft slurry level. It contained an air
sparger ring arranged as shown in Figure 5-2. The sparger ring was made
of straight 3-inch 316L SS pipe pieces welded into an octagon of approxi-
mately 4 ft in diameter, and located 6 inches from the bottom of the tank.
The initial sparger ring had 130 1/8-inch diameter holes pointed downward.
This sparger was later replaced with another one that contained 40 1/4-inch
diameter holes. The sparger ring was fed with compressed air to which
sufficient water was added to assure humidification. The air had to be
humidified to prevent local supersaturation and scaling at the sparger holes.
In more recent tests, the sparger ring was replaced by a 3-inch diameter pipe
with an open elbow discharging air downward at the center of the tank about
3 inches from the tank bottom (not shown in Figure 5-2). Humidification of
air was found to be unnecessary when the 3-inch diameter pipe was used.
The oxidation tank had an agitator with two axial flow turbines, both pumping
5-7
-------
FLUE GAS
FIDE GAS
ALKALI
VENTURI
t
£
SPRAY TOWER
SPRAY TOWER EFFLUENT
HOLD TANK
MAKEUP WATER
COMPRESSED
VENT AIR
OXIDATION
TANK
ILARtFtED LIQUOR FROM SOUPS DE WATER ING SYSTEM
WATER
OVERFLOW
DESU PER SATURATION
TANK
BLEED TO
SOLIDS
DEWATERING
SYSTEM
Figure 5-1. Flow Diagram for Two-Scrubber-Loop Forced-Oxidation
Tests on the Venturi/Spray Tower System
-------
OUTLET
SPARGER
BAFFLE
OXIDATION TANK
PLAN VIEW
COMPRESSED AIR
AGITATOR
COVER
OUTLET
' —
BAFFLE
SPARGER WITH
130 1/8-inch HOLES OR
40 1/4- inch HOLES
(DOWNWARD DISCHARGE)
Figure 5-2,
L
VENT
L_
INLET
OXIDATION TANK
AGITATOR
COMPRESSED AIR
ELEVATION VIEW
01 2345
i i I I I f
SCALE, FEET
Arrangement of the Venturi/Spray Tower
Oxidation Tank with Sparger
-------
downward. Each turbine was 52 inches in diameter and contained four blades.
The bottom turbine was 10 inches above the air sparger. The agitator ro-
tated at 56 rpm and was rated at 17 brake Hp.
A 10-ft diameter desupersaturation tank, operating with a 5-ft slurry level,
followed the oxidation tank to provide both time for gypsum precipitation
and air-free pump suction.
Provision was made to add alkali to both scrubber loops. In limestone test-
ing, however, fresh limestone slurry was added only to the spray tower loop.
Clarified liquor from the solids dewatering system could be returned to
either scrubber loop or to the mist eliminator bottom wash circuit.
The first stage (venturi) scrubber loop was fed with slurry bleed from the
second stage (spray tower) downcomer. Slurry was bled from the first stage
(venturi) scrubber loop to a clarifier followed by a rotary drum vacuum
filter for dewatering.
5.2 DISCUSSIONS OF TEST RUN RESULTS
In the limestone tests with high ash loadings (Runs 801-1A through 808-1A,
and 817-1A through 819-1B), the flue gas was withdrawn from the Boiler No. 10
duct upstream of the particulate removal equipment. Fly ash loadings in the
flue gas ranged from 3 to 6 grains/dry scf. The majority of the fly ash was
collected in the first scrubber loop (venturi), leaving the slurry in the
second loop (spray tower) relatively free of fly ash.
For the limestone testing with low fly ash loadings (Runs 809-1A through
816-1A), the flue gas was taken from the Boiler No. 10 duct downstream of
5-10
-------
the particulate removal equipment. This flue gas contained particulate
concentrations of 0.04 to 0.20 grain/dry scf.
Generally, the results with high and low fly ash loadings are similar, with
good sulfite oxidation and good filter cake being achieved in both cases.
Table 5-1 summarizes the results of two-scrubber-loop forced-oxidation lime-
stone tests with high and low fly ash loadings.
5.2.1 Startup Test Run 801-1A
In the initial test with high fly ash loading, Run 801-1A achieved 93 percent
average sulfite oxidation by simple air sparging in the venturi stage hold
tank. Test conditions for Run 801-1A were chosen to approximate those used
in the tests at the IERL-RTP pilot plant (Reference 6) and to provide the
best oxidation tank environment for oxidation to occur. These conditions
included 4.5 pH for optimum reaction rate in the oxidation tank, 18-foot
maximum slurry level in the oxidation tank, and an air stoichiometric ratio
of 4.7 atoms oxygen/mole S02 absorbed, which corresponded to the maximum air
compressor capacity of 400 scfm.
The filter cake solids concentration averaged 81 percent which was, the
highest ever recorded at Shawnee at the time. In the subsequent runs, when
sulfite oxidation was above about 90 percent, filter cake solids concentra-
tions higher than 80 percent were consistently achieved.
Although good oxidation was achieved for this run, the S02 removal was poor
and averaged only 62 percent at 3400 ppm average inlet S02 concentration.
5-11
-------
5.2.2 Modifications of Initial Test Conditions to Improve SC^ Removal
In Runs 802-1A through 805-1A, operating conditions were changed to improve
$02 removal.
Increasing spray tower liquor rate from 1300 to 1400 gpm and oxidation tank
pH from 4.5 to 5.0 (by higher limestone addition rate to the spray tower
hold tank) during Run 802-1A gave essentially no improvement in S02 removal.
Marginal increase in S02 removal from 60 to 67 percent was achieved when the
venturi liquor rate was increased from 400 to 600 gpm in Run 803-1A.
In Runs 804-1A and 805-1A, the solids concentration in the spray tower recir-
culation loop was increased by returning the clarified liquor from the dewater-
ing system to the venturi loop rather than the spray tower loop. This change
increased the solids in the spray tower loop from 5.4 percent to 15.6 percent,
with solids controlled at 15 percent in the venturi loop. The higher solids
concentration provided more limestone surface to dissolve in the spray tower
loop and the spray tower inlet liquor pH increased from 5.65 to 6.15 (pH con-
trolled at 5.0 in the venturi loop). Overall S02 removal increased from 67
percent to 78 percent. Subsequent runs were made with clarified liquor return
to the venturi loop.
5.2.3 Optimization of Air Stoichiometric Ratio
Runs 801-1A through 805-1A were made at an air flow rate to the sparger in the
oxidation tank of 400 scfm (maximum air compressor capacity) corresponding to
an air Stoichiometric ratio of about 4.5 atoms oxygen per mole of S02 absorbed.
Runs 806-1A and 807-1A were made to determine the minimum air rate required for
5-12
-------
nearly complete sulfite oxidation. The following tests were made at 4.5 oxida-
tion tank pH and 18-foot tank level, using the 130-hole air sparger:
Effect of Air Stoichiometry on Sulfite Oxidation for
Run
806-1A
806-1B
806-1C
806-1D
807-1A
Limestone Slurry
Air Rate,
scfm
250
150
100
50
0
(4.5 pH) with High Fly Ash Loading
Air Stoichiometry,
atoms oxygen/mole S02 abs.
2.5
1.7
1.0
0.5
0
Percent
Sulfite
Oxidation
99
98
97
67
18
The break in oxidation appeared to occur somewhere between 1.0 and 0.5 air
Stoichiometry. This compares with an air Stoichiometry of about 2.6 required
in tests at the IERL-RTP pilot plant run under similar conditions without fly
ash (Reference 6). It has been postulated that either an oxidation catalyst
such as manganese, possibly introduced with the fly ash, or better mass
transfer due to mechanical agitation, accounts for the excellent oxidation.
Although such a catalyst is suspected, one has not yet been identified in
the Shawnee system.
Based on these tests, an air rate of 150 scfm, corresponding to an air Stoi-
chiometry of 1.7 atoms oxygen per mole S0£ absorbed, was chosen as a standard
rate for subsequent runs made at 25,000 acfm gas rate. Run 808-1A was made
at this air rate to confirm the good oxidation (97 percent).
5-13
-------
5.2.4 Limestone Tests with Low Fly Ash Loading
Following a series of tests with lime slurry (see Section 7), testing with
limestone was resumed but with low fly ash loading in the flue gas (Runs
809-1A through 816-1A).
In general, the results with high and low fly ash loadings were similar.
High sulfite oxidation and filter cake solids content were achieved in both
cases. The majority of fly ash was removed by the venturi and ended up in the
first (oxidation) slurry loop. It is postulated that, even with low fly ash
loading in the flue gas, the oxidation catalysts introduced with fly ash may
still be present in sufficient quantity for good oxidation.
During the testing with low fly ash loading, the spray tower slurry solids
concentrations were sufficiently high (about 8 percent) for good S02 removal,
even though all the clarified liquor was returned to the spray tower loop.
(The venturi slurry solids concentration was controlled at 15 percent as in
the case of high fly ash loading.) This allowed the continuous washing of
the bottom side of the mist eliminator with a diluted clarified liquor, which
permitted operation of the spray tower at higher limestone stoichiometries
without mist eliminator fouling.
5.2.5 Effect of Oxidation Tank pH
Runs 809-1A through 811-1A were made to determine if the slurry liquor pH in
the oxidation tank affected the degree of sulfite oxidation. Over a pH range
of 4.5 to 5.5 no adverse effect was seen. Sulfite oxidation remained 97 per-
cent. Higher pH levels, where a drop in oxidation might be expected, were
5-14
-------
not investigated as they would be outside of the practical operating range for
a two-scrubber-loop limestone system. The increase in pH also resulted in
increased SC^ removal up to 85 percent in Run 811-1A.
5.2.6 Effect of Desupersaturation Tank
Based on the normal slurry recirculation rate in the venturi loop of 600 gpm,
the slurry residence time in the oxidation tank was 11.3 minutes. The oxida-
tion tank overflowed into a desupersaturation tank which provided an additional
4.7 minutes residence time, for a total of 16 minutes in the venturi recircula-
tion loop. In Run 812-1A, the desupersaturation tank was removed from the
recirculation system with no adverse effect on either sulfite oxidation or
gypsum saturation in the venturi loop. However, the desupersaturation tank
was returned to service for subsequent runs because it provided a convenient
surge for both the recirculation pump and the bleed pump.
5.2.7 Effect of Oxidation Tank Agitator
Run 813-1A was made with the oxidation tank agitator turned off to determine
the effect on oxidation of agitating the tank with the sparged air only. The
run was ended after 7 hours because air sparging alone could not keep the
solids in suspension.
5.2.8 Effect of Oxidation Tank Level
Most runs were made with an 18-foot (maximum) slurry level in the oxidation
tank. However, the tank level was dropped to 14 feet (8.8 minutes oxidation
5-15
-------
tank residence time at 600 gpm) in Run 814-1A and sulfite oxidation was still
maintained at 98 percent with no adverse effect on other operating conditions.
5.2.9 Effect of Flue Gas Flow Rate
Runs 801-1A through 814-1A were made at a reduced flue gas flow rate under
the assumption that high S0£ removal, in the range of 85 percent, could not
be achieved at full gas rate because of the reduced pH in the venturi loop
for forced oxidation.
Beginning with Run 815-1A, the flue gas flow rate was increased from the
reduced rate of 25,000 acfm (at 300°F) to the maximum rate of 35,000 acfm.
Air flow rate to the oxidation tank was also increased proportionally from
150 to 210 scfm to maintain the same air stoichiometry of about 1.7. S02
removal for Run 815-1A was 86 percent under the test conditions listed in
Table 5-1, which is about 5 percentage points below S02 removal achieved on
an identical run (Run 814-1A) at the lower flue gas flow rate. Sulfite oxi-
dation remained high at 96 percent for Run 815-1A despite the higher sulfite
throughput in the oxidation tank. All subsequent runs were made at the
higher flue gas flow rate.
5.2.10 System Control
In the limestone tests with two scrubber loops, the control philosophy was to
hold the venturi inlet pH (oxidation tank pH) at 5.5 by adjusting the limestone
slurry feed rate to the spray tower effluent hold tank. Control in this manner
proved to be difficult, and wide fluctuations were experienced in both pH and
5-16
-------
limestone stoichiometry. This was especially true when the venturi inlet pH
approached the spray tower pH. For example, in Run 815-1A, the venturi inlet
pH varied between 4.9 and 6.3, with corresponding fluctuations in the limestone
stoichiometric ratios of 1.1 to 1.9 in the venturi loop and 1.2 to 2.7 in the
spray tower loop.
In Run 815-1A, the oxidation tank level was 14 feet, which was satisfactory
for forced oxidation (96 percent oxidation at an air stoichiometric ratio of
1.7). In Run 816-1A, the fluctuation in venturi inlet pH was reduced to a
range of 5.2 to 5.8 with a corresponding reduction in fluctuation in lime-
stone stoichiometry by increasing the oxidation tank level to the maximum
of 18 feet. This increased hold tank residence time.
In Run 819-1B, the control philosophy was changed: the limestone stoichio-
metry in the spray tower was controlled at 1.6 moles calcium per mole SC^
absorbed, and the venturi inlet pH was allowed to vary. With direct control
on the spray tower stoichiometry, the fluctuation in venturi inlet pH was
5.2 to 5.8, no greater than in the previous runs with venturi inlet pH control.
On the basis of these runs, control of limestone stoichiometry in the primary
scrubbing loop (spray tower) is recommended over control of pH in the
oxidation loop (venturi).
5.2.11 Mist Eliminator Operation
In previous runs with limestone slurry and high fly ash loadings (Runs 805-1A
through 808-1A), problems with mist eliminator plugging occurred. In these
runs, the spray tower solids concentration was maintained at 15 percent. This
5-17
-------
required that the clarified liquor from the solids dewatering system be returned
to the venturi loop and which allowed only enough makeup water in the spray
tower system for an intermittent mist eliminator underside wash. Such a wash
was inadequate at the limestone utilizations experienced in the spray tower (60
to 80 percent) and the mist eliminator plugged within a matter of days.
Beginning with Run 817-1A (again with high fly ash loading), the mist elimi-
nator was washed continuously on the bottomside at 0.4 gpm/ft2 with clarified
liquor diluted with available makeup water. Excess clarified liquor was re-
turned to the venturi loop. This wash scheme, coupled with a sequential top
wash, proved adequate and the mist eliminator no longer plugged. The continu-
ous wash diluted the spray tower solids concentration to about 9 percent. At
9 percent solids concentration, S02 removal was 83 percent at 2500 ppm inlet
S0£ concentration.
In Run 818-1A, the slurry recirculation rate in the spray tower loop was in-
creased from 1400 gpm to the maximum controlled rate of 1600 gpm. With this
modification, S02 removal was increased to 86 percent at 2550 ppm inlet con-
centration.
5.2.12 Effect of Air Sparger Design
Runs 801-1A through 808-1A were made with the original air sparger which con-
tained 130 1/8-inch diameter holes (see Section 5.1). These small holes were
prone to plugging, even though the compressed air had been prehumidified to
minimize the scale and solids deposits on the wet-dry interfaces. Erosion
of these small holes also occurred.
In Runs 809-2A through 816-1A, a similar sparger having 40 1/4-inch diameter
5-18
-------
holes was used. This second sparger was inspected at the end of Run 813-1A
after about 1800 hours of service with both lime and limestone runs, and was
found to be almost completely plugged. A 3-inch diameter, 6-inch long 304
SS spool piece connecting the air line to the sparger was found to be
severely corroded with holes in several spots. Most of the sparging air
must have been emitted through the holes in this spool piece, probably
causing the plugging of the inactive 1/4-inch diameter holes on the sparger
ring. This spool piece was replaced by one with 316L SS material.
Because of the conditions mentioned above, air was emitted mainly on one side
of the oxidation tank for a good portion of the testing. Air dispersion
must have been accomplished primarily by the agitator. Despite this maldis-
tribution of air, better than 95 percent sulfite oxidation was achieved in
all runs affected.
Starting with Run 817-1A, the sparger ring was replaced with a 3-inch pipe
with an elbow on the end to direct the air downward at the center of the tank
about 3 inches above the tank bottom and one foot below the bottom agitator
blade. Sulfite oxidation remained 97 percent or higher at the same air stoi-
chiometry of about 1.7.
Based on the success with the open 3-inch pipe, it was concluded that the
agitator plays a primary role in dispersing the air. The agitator used in
these tests has been described in Subsection 5.1. Prehumidification of air
was found unnecessary with the open 3-inch pipe.
5-19
-------
5.2.13 Limestone Long-Term Reliability Run 819-1A
During November 1977 Run 819-1A, a one-month limestone slurry reliability run,
was made with a two-scrubber-loop configuration on the venturi/spray tower
system and with forced oxidation in the venturi loop. This run operated for a
total of 840 hours (35 days). The run was designed to demonstrate operating
reliability of the scrubber system with respect to scaling and plugging and to
determine if the EPA New Source Performance Standards for SC^ and particulate
emissions could be met.
To simulate variable boiler load, the flue gas flow rate was varied between
18,000 and 35,000 acfm (4.8 and 9.4 ft/sec spray tower superficial gas veloc-
ity) as the boiler load varied between 100 and 155 MW. Flue gas with high
fly ash loading was used. The venturi plug was fixed at a position to give
9 inches 1^0 pressure drop across the venturi at a full 35,000 acfm flue gas
flow rate. The actual venturi pressure drop ranged from 3 to 9 inches.
Slurry recirculation rates were held constant at 600 gpm and 1600 gpm in the
venturi and spray tower loops, respectively. The venturi inlet pH was con-
trolled at 5.5 by controlling the limestone feed rate to the spray tower hold
tank.* The oxidation tank was maintained at an 18-ft slurry level and an air
flow rate of 210 scfm discharged through a 3-inch pipe.
During the run, the scrubber was shut down for 18 hours due to a boiler out-
age, 17-1/2 hours total for three scheduled weekly inspections, plus 3-1/2
hours of unscheduled downtime for a total of 39 hours. Based on the 3-1/2
hours of unscheduled downtime, the scrubber system had an availability of
* As discussed in Subsection 5.2.10, this mode of control was later changed
to limestone stoichiometric ratio control in the spray tower (Run 819-1B),
5-20
-------
99.6 percent. The unscheduled downtime included 3 hours to clean a partially
plugged slurry return pipe from the venturi to the oxidation tank and 1/2
hour to clean a plugged mist eliminator bottom wash nozzle.
The plugged mist eliminator nozzle was discovered after 391 hours of opera-
tion. The mist eliminator in the vicinity of the plugged nozzle was severely
restricted by slurry solids (7 percent overall mist eliminator restriction.)
The nozzle was cleaned but the mist eliminator was not disturbed. By the
end of the run (840 hours), the mist eliminator restriction had dropped to
3 percent, demonstrating that a restricted area can be self-cleaning.
For the entire run, the SC^ removal averaged 86 percent at 2950 ppm average
inlet S02 concentration. This removal efficiency corresponds to an average
emission of 1.1 Ib SO£/MM Btu which meets the EPA New Source Performance
Standard of 1.2 Ib S02/MM Btu. However, fluctuations to unusually high
inlet SOp concentrations were experienced, and the standard was at times
exceeded for periods greater than the three hours allowed by EPA regulations.
The outlet particulate loading ranged from 0.021 to 0.063 grain/dry scf with
an average of 0.042 grain/dry scf. Assuming 30 percent excess air to the
boiler, the average outlet particulate loading corresponds to 0.08 lb/MM
Btu, which meets the EPA New Source Performance Standard of 0.1 Ib/MM Btu.
However, a few of the outlet particulate loading measurements exceeded the
standard.
Sulfite oxidation averaged 98 percent during the run with the air stoichio-
metric ratio varying between 1.4 and 2.8 atoms oxygen/mole S02 absorbed.
The filter cake solids concentration average 87 percent. Overall limestone
5-21
-------
utilization was 81 percent while the spray tower limestone utilization was
61 percent, demonstrating the advantage of a two-scrubber-loop system in
achieving high alkali utilization.
To summarize, the operating reliability of the venturi/spray tower system in
a two-scrubber-loop configuration with forced oxidation in limestone slurry
service has been demonstrated with a system availability of 99.6 percent.
However, under the conditions selected, the system was unable to continually
meet EPA New Source Performance Standards for SC^ and particulate emissions
even though the average emissions for the run met the standards.
5.3 GENERAL OPERATING CHARACTERISTICS OF THE TWO-SCRUBBER-LOOP
SYSTEM WITH FORCED OXIDATION
5.3.1 Water Balance
Because the dewatering properties of the oxidized sludge are better, less
liquor leaves the system with the waste sludge (better than 80 percent
solids compared with about 50 to 60 percent solids with unoxidized sludge)
and the overall water balance in the slurry system is tighter. The tighter
water balance results in higher dissolved solids in the slurry liquor and
less available water for mist eliminator wash.
5.3.2 Control of Slurry .Soljjjs. Concentration
The slurry solids concentrations in the two scrubber loops are interre-
lated by the system water balance. Major water loss from the scrubber
system occurs in the first (venturi) scrubber loop where the flue gas is
5-22
-------
humidified. The Shawnee tests have normally been run by holding the
venturi loop solids concentration at 15 percent and letting the spray
tower solids concentration vary. The spray tower slurry solids concentra-
tion has depended primarily on the fly ash loading in the flue gas and
on the loop to which the clarified liquor from the solids dewatering
system is returned. To a certain degree, the spray tower solids concen-
tration can be controlled by the proper distribution of the clarified
liquor (usually in excess of that required for the mist eliminator wash)
returned to the venturi and spray tower loops.
With limestone, low spray tower slurry solids concentration (about 5 to 6
percent) contributes to low S02 removal (60 to 70 percent), while the
effect is much less pronounced with lime. The only adverse effect observed
with lime (see Section 7) was a slight tendency to scale in the bottom part
of the spray tower when the solids concentration dropped below about 7
percent.
5.3.3 Chloride Concentrations
Chlorides enter the scrubber systems as hydrochloric acid (HC1) in the flue
gas. The chlorides are absorbed and concentrated in the scrubber liquor
as a major component of the total dissolved solids. Chloride ion in the
scrubber liquor tends to reduce the liquor pH (at constant limestone
stoichiometric ratio) and consequently reduce the SO^ removal efficiency.
Because of tighter water balance with forced oxidation, chloride ion concen-
tration has tended to be higher than in previous operation without forced
oxidation. Chloride ion concentration in the venturi loop during forced-
5-23
-------
oxidation testing ranged from 2,000 to 11,000 ppm, varying mainly with the
chloride concentration of the coal burned. Chloride ion concentrations in
the spray tower loop depended on where the clarified liquor from the dewater-
ing system was returned. With clarified liquor returned to the spray tower
loop, spray tower chloride ion concentration averaged about one-half the
concentration in the venturi loop. With clarified liquor returned to the
venturi loop, the spray tower chloride ion concentration was about one-fifth
the venturi concentration. With the low spray tower chloride ion concentra-
tions in the latter case, higher pH, and consequently higher S02 removal
efficiency, was achieved in the spray tower loop.
5.4 SUMMARY OF FINDINGS
Following the development of forced-oxidation techniques at the IERL-RTP
pilot plant on a 0.1-MW scale, forced oxidation was successfully demon-
strated at the Shawnee Test Facility on 10-MW prototype scrubbers. The
following summary is based on the Shawnee limestone test results from the
venturi/spray tower system operating on a two-scrubber-loop configuration:
• Forced oxidation in the first of two independent scrubbing
loops was successfully demonstrated in the two-scrubber-loop
venturi/spray tower system with limestone slurry. Successful
demonstration was culminated with a one-month reliability
limestone test using flue gas with high fly ash loading.
• Under the one-month reliability test conditions of 18,000 to
35,000 acfm gas rate, 600 gpm venturi slurry rate, 1600 gpm
spray tower slurry rate, 15 percent venturi slurry solids
concentration, 18-ft oxidation tank level, and 5.5 oxidation
tank pH, 98 percent average sulfite oxidation was achieved at
air stoichiometric ratios ranging from 1.4 to 2.8 atoms oxygen/
mole SOo absorbed. The filter cake solids concentration
averagea 87 percent.
5-24
-------
Ranges of conditions under which nearly complete sulfite oxida-
tion (greater than 95 percent) was demonstrated were an oxida-
tion tank pH of 4.5 to 5.5, oxidation tank level of at least
14 ft, and an air stoichiometric ratio of at least 1.5 atoms
oxygen/mole SO? absorbed. Filter cake solids contents greater
than 80 percent were consistently achieved under these conditions.
Forced oxidation of the slurry within the scrubber loop dramat-
ically improved the dewatering characteristics of the waste
solids. Sulfite oxidation of 90 percent or higher was required
for maximum improvement.
Forced oxidation was achieved by simple air/slurry contact in
the oxidation tank. Within the ranges of test conditions, an
air sp?rger with 130 1/8-inch diameter holes, an air sparger
with 40 1/4-inch diameter holes, and a 3-inch diameter pipe
all seemed to give the same air/slurry contact efficiency.
Air/slurry contact was primarily achieved by the two-turbine
agitator operated at 56 rpm and rated at 17 brake Hp.
Slurries with high or low fly ash loadings oxidized equally well.
With forced oxidation, the venturi inlet slurry constantly ex-
hibited a gypsum saturation of about 100 percent, which was be-
low the incipient scaling level of 135 percent. This was undoubt-
edly caused by the abundance of gypsum crystal seeds produced by
forced oxidation.
Limestone utilization was improved with the two-scrubber-loop
operation to at least 80 percent and as high as 98 percent, de-
pending on the venturi inlet pH which averaged from 4.5 to 5.5.
The venturi inlet pH was controlled by limestone addition to
the spray tower hold tank.
In limestone scrubbing, a low slurry solids concentration in
the spray tower reduced the percent S02 removal. Furthermore,
a solids concentration of 7 percent or higher was required to
prevent calcium sulfite and sulfate scaling.
Operation without the desupersaturation tank in the venturi
slurry loop did not have an adverse effect on either sulfite
oxidation or gypsum saturation in the venturi loop.
Operation with the agitator turned off in the oxidation tank
was not successful. Air sparging alone in the oxidation tank
could not keep the solids from settling.
5-25
-------
Section 6
VENTURI/SPRAY TOWER TWO-SCRUBBER-LOOP
FORCED-OXIDATION LIMESTONE/MgO TEST RESULTS
From March 1 through May 10, 1978, a series of six runs was made in which
MgO was added to the spray tower hold tank along with the limestone slurry.
The primary purpose of the MgO addition was to enhance S02 removal efficiency
in the spray tower loop by increasing the dissolved sulfite ion concentration
for S0£ scrubbing. In a two-scrubber-loop configuration, the magnesium ion
concentration in the venturi loop is higher than that in the spray tower loop
because of the water loss in humidifying the flue gas in the venturi loop.
But because the sulfite ion is converted into nonscrubbing sulfate ion by
forced oxidation, the higher magnesium ion concentration in the venturi loop
does not enhance S02 removal in the venturi loop. The secondary purpose of
the MgO addition was to determine whether the presence of magnesium ion had
an effect on oxidation efficiency.
Table 6-1 presents the major test conditions and important test results for
the six limestone runs with MgO addition. More detailed information for these
runs is given in Appendices B through G.
6-1
-------
Table 6-1
RESULTS OF FORCED-OXIDATION TESTS WITH TWO SCRUBBER LOOPS
ON THE VENTURI/SPRAY TOWER SYSTEM USING LIMESTONE SLURRY WITH ADDED MAGNESIUM OXIDE
Major Test Conditions
Fly ash loading
Flue gas rate, acfm & 300°F
Slurry rate to venturl, gpm
Slurry rate to spray tower, gpm
Venturl percent solids reclrculated (controlled)
Residence times, m1n: Oxidation tank
Desupersaturatlon tank
Spray tower EHT
Venturl Inlet (oxidation tank) pH (controlled)
Spray tower limestone sto1ch1ometr1c ratio (based on solids)
Effective Kg** concentration (S.T. loop), ppm
Venturl pressure drop, 1n. H20
Oxidation tank level, ft
Air rate to oxidation tank, scfjir1'
(2\
Clarified liquor returned tov '
Selected Results
Percent S02 removal
Inlet S02 concentration, ppm
Spray tower percent solids reclrculated
Spray tower Inlet pH
Spray tower limestone stolchlometrtc ratio (based on total slurry)
Spray tower Inlet liquor gypsum saturation, %
Spray tower sulflte oxidation, X
Effective Mgf+ concentration (S.T. loop), ppm
Overall sulflte oxidation, S
Overall limestone utilization, X
Venturl Inlet liquor gypsum saturation, t
Venturl Inlet liquor sulflte concentration, ppm
A1r stolcMometry. atoms 0/mole S02 absorbed
Filter cake solids, wtv '
Onstream hours
820-lft
High
35,000
600
1600
15
11.3
4.7
14.7
5.5
-
5000
9
18
210
V.
96
2250
6.0
6.05
1.16
100
30
5150
98
92
130
50
1.70
oc
85
462
820-18
High
35,000
600
1600
15
11.3
4.7
14.7
-
1.6
5000
9
18
150
V.
2500
8.3
5.9
1.28
105
17
4985
92
90
130
950
1.10
oo
Ot
137
820-1C
High
35,000
600
1600
15
11.3
4.7
14.7
• (5.
1.6
5000
9
18
0
V.
91
2750
10.5
5.9
1.52
90
20
4700
36
82
145
5585
0
134
821-1A
High
35,000
600
0
15
11.3
4.7
•
5.5
"
5000
9
18
210
V.
.
'
-
-
-
-
-
-
-
-
-
-
"
822-1A
High
35.000
600
1600
15
11.3
4.7
14.7
" (5)
1.6
5000
9
18
210
V.
91
2750
8.0
5«7(;
.75
1.55
100
21
4895
97
79 •
125
735
1.45
85
232
822-18
High
35,000
600
1600'4>
15
11.3
4.7
14.7
(5)
5000
9
1 D
IB
210
V.
90
2400
5f
.6
5CC
.39
1.21
110
23
4845
98
93
130
410
1.70
85
85
Notes:
Air discharged downward through 3-1nch diameter pipe with an open elbow at center of oxidation tank about 3 Inches from tank bottom.
Venturl loop (oxidation tank).
CsiarSyf^werndturneSeSffSeforT0 SlS.tSlTI-iVll SSin SO, removal with venturl alone. Venturl S02 removal averaged 29*.
In runs with control by spray tower sto1ch1ometr1c ratio, the venturl Inlet pH averaged 5.0.
Run failed due to low MgO dissolution rate 1n spray tower effluent hold tank.
-------
6.1 SYSTEM DESCRIPTION
The scrubber system configuration and the arrangement of the oxidation tank
for the two-scrubber-loop forced-oxidation limestone tests with MgO addition
were the same as those for the limestone tests without MgO addition. These
are shown in Figures 5-1 and 5-2, and described in Subsection 5.1.
Dry MgO powder was fed to the spray tower hold tank by a screw feeder. The
3-inch diameter pipe with an open elbow (not shown in Figure 5-2) was used
instead of the air sparger ring.
6.2 DISCUSSIONS OF TEST RUN RESULTS
6.2.1 Effect of Magnesium Ion on SO? Removal
Run 820-1A was made under identical conditions to Run 818-1A (see Table 5-1)
except for the addition of MgO. Effective magnesium ion concentration*
averaged 5150 ppm in the spray tower. The anticipated removal enhancement
was achieved; the average S02 removal was 96 percent at 2250 ppm average
inlet SOp concentration for the run with magnesium compared with 86 percent
removal at 2550 inlet ppm for Run 818-1A without magnesium.
The spray tower inlet slurry liquor was 100 percent saturated with gypsum and
no scale was observed. This condition was typical of all the tests in the
limestone/MgO, forced-oxidation test block.
* Effective magnesium ion concentration is defined as the total magnesium ion
minus that magnesium ion concentration equivalent to total chlorides. Mag-
nesium chloride has no effect on S0 removal.
6-3
-------
In Run 821-1A, an attempt was made to determine the SC^ removal in the venturi
by turning off the slurry recirculation to the spray tower. Unfortunately,
the MgO added to the spray tower hold tank would not dissolve without slurry
recirculation and S02 absorption in the spray tower. The run was therefore
ended.
A second attempt was more successful. Run 822-1B was an extension of Run
822-1A, in which the spray tower slurry recirculation was turned off once a
shift for only 30 minutes. This short time period did not significantly upset
the system balance. Average S02 removal in the venturi loop was found to be
29 percent, which is typical of removal efficiency with limestone slurry in
the absence of magnesium ion. Thus, it has been demonstrated that magnesium
ion does not enhance S02 removal in a scrubber loop with forced oxidation.
Run 822-1A was made in an effort to improve removal efficiency by minor changes
in piping configuration to locate makeup and bleed streams at their optimum
locations in the venturi slurry recirculation loop. The bleed from the spray
tower loop was sent to the desupersaturation tank instead of to the oxidation
tank as shown in Figure 5-1. Since the spray tower bleed stream contained
higher sulfite ion concentration than the slurry liquor in the oxidation tank,
introduction of this stream at a point in the venturi loop after the oxidation
tank should have promoted SOg removal in the venturi. Also, the bleed to the
solids dewatering system was taken from the oxidation tank instead of the
desupersaturation tank as is shown in Figure 5-1. Improvement in SO^ removal
efficiency, if any, was too small to observe.
6-4
-------
6.2.2 Effect of Magnesium Ion on Oxidation Efficiency
In this test block, the 3-inch pipe was used for discharging air into the
oxidation tank, and an oxidation tank level of 18 feet was maintained.
Runs 820-1A, B, and C were a series to explore the air stoichiometry required
to achieve nearly complete oxidation. The results were as follows:
Air Stoichiometric Ratio, Percent
Run atoms oxygen/mole S0? absorbed Sulfite Oxidation
820-1A 1.7 98
820-1B 1.1 92
820-1C 0 . 36
The oxidation efficiency for Run 820-1R was marginally acceptable at an air
Stoichiometric ratio of 1.1. Although sulfite oxidation averaged 92 percent,
it fluctuated widely, indicating that barely enough air was available. During
the last 40 hours of Run 820-1B, the air Stoichiometric ratio increased to 1.3
as the inlet 862 concentration decreased, and the oxidation was steady at 98
percent. Thus, sulfite oxidation efficiency appeared to be unaffected, if not
improved, by the addition of MgO.
Filter cake solids concentration at 98 percent oxidation averaged 85 percent,
demonstrating that MgO addition does not adversely affect filtration
characteristics of oxidized sludge. However, initial settling rate declined
somewhat (see Subsection 22.2). This series of runs also demonstrated the
effect of forced oxidation on solids filtration characteristics. Filter cake
solids concentration decreased from 85 percent to 63 percent as the oxidation
of sulfite decreased from 98 percent to 36 percent.
6-5
-------
An additional observation in this series of runs was that overall limestone
utilization increased from 82 to 92 percent as the air rate to the oxidation
tank was increased from 0 to 210 scfm. Based on the results of Runs 820-1A
and 1C, it appeared that forced oxidation promoted higher limestone utilization.
6.3 SUMMARY OF FINDINGS
The following is a summary of findings based on the test results:
The anticipated SOg removal enhancement was achieved. Under
typical operating conditions, average overall S02 removal was
96 percent at 2250 ppm average inlet SOo concentration with
5150 ppm effective magnesium ion concentration in the spray
tower in which oxidation was not forced (Run 820-1A). Under
the same operating conditions, but without MgO addition,
overall SOp removal averaged 86 percent at 2550 ppm inlet
S02 concentration (Run 818-1A, Table 5-1).
Magnesium ion does not enhance SC^ removal in a scrubber loop
in which oxidation is forced, because sulfite ion is oxidized
into non-scrubbing sulfate ion. SOo removal in the venturi
loop (oxidation loop) was 29 percent (Run 822-1B with inter-
mittent spray tower operation), which is typical of removal
efficiency with limestone slurry in the absence of magnesium
ion.
The oxidation efficiency appeared to be unaffected, if not
improved, by the presence of magnesium ion. Under typical
conditions, sulfite oxidation was 98, 92, and 36 percent at
1.7, 1.1, and 0 air stoichiometry, respectively (Runs 820-1A,
820-1B, and 820-1C, respectively). Nearly complete oxidation
appeared to be possible at an air stoichiometric ratio as low
as 1.3.
Under the conditions tested, filter cake solids concentration
at 98 percent oxidation averaged 85 percent, demonstrating
that MgO addition does not adversely affect the filtration
characteristics of oxidized sludge. The initial settling
rate, however, is reduced when magnesium is present (see
Subsection 22.2).
6-6
-------
Section 7
VENTURI/SPRAY TOWER TWO-SCRUBBER-LOOP
FORCED-OXIDATION LIME TEST RESULTS
In this section, the results of forced-oxidation lime tests with a two-
scrubber-loop configuration on the venturi/spray tower system are presented.
A total of 20 lime runs were made. These included 10 runs with high fly
ash loading in the flue gas (4 runs from March 10 through April 1, 1977,
1 run from October 11 through October 18, 1977, and 5 runs from December 16,
1977 through February 27, 1978) and 10 runs with low fly ash loading (9
runs from June 21 through August 11, 1977 and 1 run from October 6 to Octo-
ber 11, 1977). Tests with high fly ash loading included a one-month lime
reliability run (Run 863-1A). Except for the one-month reliability run,
each test normally lasted 5 to 6 days which was judged to be sufficient
time to reach kinetic equilibrium and to allow collection of adequate run
data.
In many aspects, the operating characteristics and results of the two-
scrubber-loop forced-oxidation lime tests were similar to those of lime-
stone tests presented in Section 5.
Table 7-1 summarizes the major test conditions and important test results
for the two-scrubber-loop forced-oxidation lime runs. Additional detailed
information for these runs can be found in Appendices B through G.
7-1
-------
Table 7-1
RESULTS OF TWO-SCRUBBER-LOOP FORCED-OXIDATION LIME TESTS
ON THE VENTURI/SPRAY TOWER SYSTEM
Major Test Conditions
Fly ash loading
Gas rate, acfra 9 300°F
VentuH liquor rate, gpm
Spray tower liquor rate, gpm
Venturl percent sol Ids reclrculated (controlled)
Residence times, m1n: Oxidation tank
Oesupersaturatlon tank
Spray tower EHT
Venturl Inlet (oxidation tank) pH (controlled)
Spray tower Inlet pH (controlled)
Venturl pressure drop, 1n. H,0.
b
Oxidation tank level, ft
A1r rate to oxidation tank, scfm
IO\
Clarified liquor returned tov°;
Selected Results
Percent S02 removal
Inlet S02 concentration, ppm
Spray tower percent solids reclrculated
Spray tower Hme stolchlometHc ratio
Spray tower inlet liquor gypsum saturation, %
Spray tower sulflte oxidation, *
Overall sulfite oxidation, %
Overall lime utilization, %
Venturl Inlet liquor gypsum saturation, %
VentuH Inlet liquor sulflte concentration, ppm
A1r stolchlometry, atoms 0/mole S02 absorbed
Filter cake solids, w«(10)
Onstream hours
851-1A
High
25,000
600
1400
15
11.3
4.7
18
4.5
8.0
9
18
150<5>
V. & S.T.
78
3300
6.0
1.13
80
15
97
98
95
39
1.45
81
110
852-1A
High
25,000
160
1400
15
42
17.7
18
8.0
(m1n)
18
150<5>
V. & S.T.
70
3400
8.3
1.13
50
13
83
96
105
22
1.55
73
74
853-1A
High
25,000
600
1400
15
11.3
4.7
18
>4.5<3>
8.0
(tnln)
18
ISO'5'
S.T.
77
3450
6.4
1.14
80
14
97
95
100
35
1.40
78
161
854- 1A
High
25,000
600
1400
15
11.3
4.7
18
5.2
8.0
9
18
150<5>
S.T.
82
3150
6.1
1.1S
55
11
96
97
100
35
1.40
79
166
855-1 A
Low
25,000
600
1400
15
11.3
4.7
18
4.5<4>
8.0
9
18
S.T.
83
2350
7.3
1.14
95
18
98
98
105
24
1.90
75^11)
158
856-1 A
Low
25,000
600
1400
15
11.3
4.7
18
5.0
8.0
9
18
150<6> .
S.T.
88
2500
7.9
1.16
90
12
97(9)
98
95
36
1.65
83C11'
209
857-1 A
Low
25.000
600
1400
15
11.3
4.7
18
5.5
8.0
9
18
150<6>
S.T.
84
2650
7.9
1.14
90
11
97
99
100
64
1.65
78
128
858-1 A
Low
25,000
600
1400
11.3
47
,/
1 O
18
5.5
Brt
.U
1 ft
150<6>
V.
83
2750
15.2
11 ^
. 1J
oc
85
1 A
14
97
98
110
20
1C A
.60
81
ICO
162
_ •
Note: Footnotes for this table are listed at the end of Table. 7-1 (continued).
-------
Table 7-1 (continued)
RESULTS OF TWO-SCRUBBER-LOOP FORCED-OXIDATION LIME TESTS
ON THE VENTURI/SPRAY TOWER SYSTEM
Major Test Condtions
Fly ash loading
Gas rate, acfm @ 300°F
Venturl liquor rate, gpm
Spray tower liquor rate, gpm
Venturi percent solids recirculated (controlled)
Residence times, min: Oxidation tank
Desupersaturation tank
Spray tower EHT
Venturi inlet (oxidation tank) pH (controlled)
Spray tower inlet pH (controlled)
Venturi pressure drop, in. HgO
Oxidation tank level, ft
Air rate to oxidation tank, scfm
/Q\
Clarified liquor returned tov '
Selected Results
Percent SOg removal
Inlet SO, concentration, ppm
Spray tower percent solids recirculated
Spray tower lime stoichiometric ratio
Spray tower inlet liquor gypsum saturation, %
Spray tower sulfite oxidation, %
Overall sulfite oxidation, %
Overall lime utilization, %
Venturi inlet liquor gypsum saturation, %
Venturi inlet liquor sulfite concentration, ppm
Air stoichiometry, atoms 0/mole SO- absorbed
Filter cake solids, wt%'10'
Onstream hours
859-1A
Low
25,000
600
1400
15
11.3
4.7
18
5.5
8.0
9
18
135<6'
S.T.
92
2950
7.2
1.16
60
10
76
97
95
99
1.20
74
115
859-1B
Low
25,000
600
1400
15
11.3
4.7
18
5.5
8.0
9
13
150<6>
S.T.
92
2700
7.6
1.16
90
12
99
99
105
33
1.50
78
70
859-1C
Low
25,000
600
1400
15
11.3
4.7
18
5.5
8.0
9
18
100<6>
S.T.
92
2700
7.3
1.15
90
20
69
96
105
82
1.0
71
120
859-1D
Low
25,000
600
1400
15
11.3
4.7
18
5.5
8.0
9
18
0
S.T.
93
2600
7.1
1.17
50
10
30
90
115
85
0
55
72
860-1A
Low
25,000
600
1400
15
11.3
4.7
12.6
5.5
8.0
9
18
150<6>
S.T.
95
2200
7.0
1.18
70
10
99
99
100
20
1.75
82
133
861-1A
Low
25,000
600
1400
15
11.3
4.7
18
5.5
8.0
9
18
150<7>
S.T.
92
2700
7.3
1.16
90
19
98
99
95
35
1.50
86
117
I
CO
-------
Table 7-1 (continued)
RESULTS OF TWO-SCRUBBER-LOOP FORCED-OXIDATION LIME TESTS
ON THE VENTURI/SPRAY TOWER SYSTEM
Hajor Test Conditions
Fly ash lading
Gas rate, acfm 9 300°F
Venturl liquor rate, gpm
Spray tower liquor rate, gpm
Venturl percent solids reclrculated (controlled)
Residence times, m1n: Oxidation tank
Desupersaturatlon tank
Spray tower EHT
Venturl Inlet (oxidation tank) pH (controlled)
Spray tower Inlet pH (controlled)
Venturl pressure drop, In. H^O
Oxidation tank level, ft
Air rate to oxidation tank, scfm
Clarified liquor returned to'8'
Selected Results
Percent S02 removal
Inlet SO- concentration, ppm
Spray tower percent solids redrculated
Spray tower lime stolchlometrlc ratio
Spray tower Inlet liquor gypsum saturation, %
Spray tower sulflte oxidation, %
Overall sulflte oxidation, %
Overall Hme utilization, %
Venturl Inlet liquor gypsum saturation, %
Venturl Inlet liquor sulflte concentration, ppm
A1r stolcMometry, atoms 0/mole S02 absorbed
Filter cake solids, wtS*10'
Ons t ream hours
862-1 A
High
35,000
600
1400
15
11.3
4.7
IS
5.5
8.0
9
18
210<7>
V.
85
2650
17.3
1.11
85
18
97
98
105
30
1.65
86
162
863-1 A
High
(1)
600
1600
15
11.3
4.7
14.7
5.5
7.8
S9
18
210<7>
V. & S.T.
88
2950
10.4
1.11
100
21
97
98
105
35
1.40-2.75
85
779
864-1A
High
35,000
600
1600
15
11.3
4.7
14.7
5.5
7.8
9
18
210<7>
V. « S.T.
94
2250
9.5
1.13
100
24
98
99
100
25
1.75
85
115
865-1A
High
35,000
600
1600
15
6.3
4.7
14.7
5.5
7.8
9
10
290/210* 7'
V. & S.T.
89
2300
10.1
1.10
95
26
89
98
100
65
2.10
80
254
866-1 A
High
35,000
600
1600
15
6.3
4.7
14.7
5.5
7.8
9
10
350<7>
V. & S.T.
94
1700/2400
9.9
1.15
90
26
98/81
98
95
40
3.85/2.70
81
159
86 7-1 A
High
35,000
600
1600
15
8,8
4.7
14.7
5.5
7.8
9
14
210«7>
V. & S.T.
89
2300
11.8
1.18
85
20
98
99
90
40
1.80
86
137
1
2
3
4
5
6
7
8
9
(10
(11)
Gas rate was varied from 18,000 to 35,000 acfm to follow the boiler load.
Actual pH range was 4.6-6.2 (avg. 5.6).
Actual pH range was 4.3-5.2 (avg. 4.8).
Actual average pH was 4.7.
Used a sparger ring located 6 Inches from tank bottom. The sparger ring had 130 1/8-Inch holes on the bottom side.
Used a sparger ring located 6 Inches from tank bottom. The sparger ring had 40 1/4-Inch holes on the bottom side.
Air discharged downward through a 3-Inch diameter pipe with an open elbow at center of oxidation tank about 3 Inches from tank bottom.
Spray tower loop (effluent hold tank) or venturl loop (oxidation tank).
Excluded a period 7/2/77-7/5/77 In which oxidation dropped to as low as 621 for unknown reason.
Clarlfler and filter In series used for solids dewatering In all runs.
Intermittent filter operation.
-------
7.1 SYSTEM DESCRIPTION
The scrubber system flow configuration and the arrangement of the oxidation
tank for the two-scrubber-loop forced-oxidation lime tests were essentially
the same as those for the limestone tests, which have been shown in Figures
5-1 and 5-2 and described in Subsection 5.1.
There were only two minor changes in flow configuration when lime, instead
of limestone, was used. In the lime tests, fresh lime slurry (alkali) was
added to both scrubber loops to maintain pH control. However, fresh lime-
stone slurry was added only to the spray tower loop in the limestone tests.
In these tests, sufficient residual undissolved limestone remained to
provide pH control in the venturi loop. In lime testing, the clarified
liquor from the solids dewatering system was never returned to the mist
eliminator wash circuit as was done in limestone testing. Because of
normally high alkali utilization (see Section 8, Reference 3), intermit-
tent top and bottom mist eliminator wash with makeup water only was suffi-
cient to keep the mist eliminator clean during lime tests.
7.2 DISCUSSIONS OF TEST RUN RESULTS
7.2.1 pH Control During Initial Runs
During Run 851-1A, the first lime run with high fly ash loading (see Table
7-1), it became apparent that pH control in the venturi loop was different
with lime than with limestone. In limestone testing, the venturi inlet pH
was maintained by adding excess limestone to the spray tower loop. Residual
limestone in the bleed from the spray tower loop to the venturi provided
sufficient buffer to keep the pH from dropping rapidly in the venturi loop.
7-5
-------
Under normal venturi operating conditions with lime (600 gpm slurry flow rate
to the venturi, 9 inches H20 pressure drop), pH could not be maintained in
the venturi by lime addition to the spray tower loop alone without raising
the pH excessively in the spray tower. After about 30 hours of operation of
Run 851-1A, the control method was changed to provide separate control of the
venturi and spray tower inlet pH at 4.5 and 8.0, respectively, by independent
lime feed to the venturi and spray tower loops.
In Runs 852-1A and 853-1A, unsuccessful attempts were made to control the ven-
turi inlet pH by changing the venturi slurry recirculation rate and venturi
pressure drop. The tests were made primarily to see whether independent lime
addition to the venturi loop could be avoided. Operation in this manner proved
to be unstable, and sulfite oxidation dropped drastically when the venturi inlet
pH increased to above 6.
Finally, in Run 854-1A, the venturi inlet pH control method was changed back
to that with an independent lime feed to the venturi loop. The venturi and
spray tower inlet pH's were controlled at 5.2 and 8.0, respectively, by separ-
ate lime feed pumps. This mode of pH control provided smooth operation.
Sulfite oxidation averaged 96 percent for Run 854-1A, and SOg removal was 82
percent at 3150 pom average inlet S02 concentration.
7.2.2 Effect of Oxidation Tank pH
After lime Run 854-1A, lime testing was interrupted by a scheduled Boiler No.
10 maintenance outage from April 2 through June 13, 1977. Flue gas with low
fly ash loadings (0.04 to 0.20 grain/dry scf) was used when lime testing
resumed.
7-6
-------
As with limestone, runs were made (Runs 855-1A through 857-1A) to determine
if oxidation tank pH affected sulfite oxidation. Over the pH range of 4.5
to 5.5, no adverse effect on oxidation was observed. During these tests,
percent S02 removal ranged from 83 to 88 percent at about 2500 ppm inlet S02
concentration. Subsequent runs were made at a venturi inlet pH of 5.5.
During Run 856-1A, sulfite oxidation dropped for unknown reasons to as low
as 62 percent and then gradually recovered to 98 percent in a three-day
period (July 2 to 5, 1977). This drop in oxidation was also observed in the
same time period on the TCA system (Run 802-2A, Section 12). Although the
reason for this drop in oxidation was never determined, it was postulated
that the presence of oxidation inhibitors (possibly introduced by makeup
water from the Ohio River) or a drop in concentrations of oxidation catalysts
(e.g., trace metals and organic matter) might have been the cause.
7.2.3 Effect of Venturi Slurry Solids Concentration
In Run 858-1A, the effect of dropping the venturi loop slurry solids concentra-
tion from 15 to 8 percent (reduced oxidation tank residence time based on
system bleed rate) was investigated. Sulfite oxidation was not affected and
remained high at 97 percent.
In this run, clarified liquor from the dewatering system and makeup water in
excess of the mist eliminator wash were added to the venturi loop rather than
to the spray tower. This resulted in a buildup of the spray tower slurry
solids concentration from 8 to 15 percent. With lime slurry, the higher spray
tower slurry solids concentration did not improve S02 removal, as had been
observed with limestone slurry (see Run 805-1A, Section 5). This difference
7-7
-------
between lime and limestone was expected, because lime is more soluble
than limestone.
Subsequent runs with low fly ash loading were made at 15 percent venturi
slurry solids and with clarified liquor returned to the spray tower loop.
7.2.4 Effect of Air Rate to the Oxidation Tank
As with limestone, tests were made to determine the minimum air required
for nearly complete sulfite oxidation. Results were as follows:
Effect of Air Stoichiometry on Sulfite Oxidation
for Lime Slurry (5.5 pH) with Low Fly Ash Loading
Air Rate, Air Stoichiometry, Percent Sulfite
Run scfm atoms oxygen/mole SOg abs. Oxidation
859-1B
859-1A
859-1C
859-1D
150
135
100
0
1.5
1.2
1.0
0
99
76
69
30
In these tests, the break from nearly complete oxidation occurred between 1.2
and 1.5 air Stoichiometry, somewhat higher than with limestone but still ex-
cellent. In the limestone tests, a sparger ring with 130 1/8-inch holes was
used. In the lime tests, a ring with 40 1/4-inch holes was used. Both of
these spargers plugged, as has been discussed in Section 5. Oxidation tank
pH was higher in the lime tests (5.5) than in the limestone tests (4.5).
7-8
-------
Any combination of these differences may have contributed to the increase in
air stoichiometry required for near complete oxidation.
7.2.5 Effect of Spray Tower Hold Tank Residence Time
In Run 860-1A, the spray tower hold tank residence time was dropped from 18
minutes to 12.6 minutes, from a tank level of 10 feet, 9 inches to one of 7 feet,
6 inches. The higher tank level had been used previously to keep the bleed pump
taking suction from the spray tower downcomer from cavitating. Before this run,
the suction point on the downcomer was lowered to allow lower tank level and more
reasonable residence time.
This change had no adverse effect on sulfite oxidation or SQ^ removal. Sul-
fite oxidation was 99 percent and S02 removal was 95 percent at 2200 ppm in-
let S0 concentration.
7.2.6 Effect of Air Sparger Design
Following lime Run 860-1A, a series of limestone tests with low fly ash loading
was conducted (see Runs 809-1A through 816-1A, Section 5). Lime testing re-
sumed on October 6, 1977. Starting with lime Run 861-1A, a 3-inch diameter pipe
with an open elbow was used to introduce air into the oxidation tank. Previous
lime runs had used either a sparger with 130 1/8-inch holes (Runs 851-1A
through 854-1A) or a sparger with 40-1/4 holes (Runs 855-1A through 860-1A).
Some plugging and erosion of the holes was experienced with these rings.
As with limestone tests (see Subsection 5.2.12), oxidation efficiency during
lime tests was as good with the 3- inch air pipe as with the spargers. For example,
7-9
-------
in Run 861-1A with the air pipe, 98 percent sulfite oxidation was achieved
at an air stoichiometry of 1.5 atoms oxygen/mole SO^ absorbed. Based on the
success with the open 3-inch pipe, it was concluded that, as with limestone
runs, the agitator plays a primary role in dispersing the air. The agitator
used in the oxidation tank has been described in Subsection 5.1.
7.2.7 Control of Spray Tower Slurry Solids Concentration
As has been mentioned in Subsection 5.3.2, the slurry solids concentrations
in the scrubber loops are interrelated by the system water balance. Virtually
all the fly ash is captured in the slurry in the venturi loop so that the spray
tower slurry is essentially free of fly ash. To compensate for the fly ash,
the venturi loop is normally controlled at a higher slurry solids concentra-
tion than the spray tower loop.
Slurry solids concentration in the venturi scrubber loop was controlled at
15 weight percent in a majority of the lime runs. Slurry solids concentra-
tion in the spray tower varied from 6 percent to almost 20 percent, depending
on whether the clarified liquor from the dewatering system was returned to the
venturi loop or the spray tower loop. Beginning with Run 863-1A, the return-
ing clarified liquor was split between the two scrubber loops to control the
spray tower slurry solids concentration at about 10 percent. This solids
level was a compromise between low solids (below about 7 percent), where scal-
ing and a drop in S02 removal are experienced, and high solids (above about 15
percent), where it becomes more difficult to keep the mist eliminator clean.
7-10
-------
7.2.8 Effect of Flue Gas Flow Rate
Runs 851-1A through 861-1A were made at a reduced flue gas flow rate of 25,000
acfm (at 300°F) because it was assumed that forced oxidation would reduce
S02 removal efficiency in the venturi loop at lower pH to the extent that an
overall S02 removal efficiency of at least 80 percent could not be achieved.
This proved not to be the case. Beginning with Run 862-1A, the flue gas flow
rate was increased to the maximum achievable of 35,000 acfm (at 300°F), which
corresponds to a spray tower superficial velocity of 9.4 ft/sec. In this run
S02 removal averaged 85 percent at 2650 ppm inlet S02 concentration. Air flow
rate to the oxidation tank was increased proportionally with flue gas from
150 scfm to 210 scfm to maintain the same air stoichiometry. Sulfite oxida-
tion for Run 862-1A remained high at 97 percent despite the higher sulfite
throughput in the oxidation tank.
7.2.9 Effect of Spray Tower Slurry pH on Sulfite Scaling
The pH drop across the spray tower depends on the S02 removal and the inlet
S02 concentration. In Runs 851-1A through 862-1A, the spray tower inlet slurry
liquor pH was controlled at 8.0. In some of these runs, when the inlet S02
concentration was low (corresponding to low S02 make-per-pass), this level of
inlet pH resulted in an outlet pH approaching 6.0 and caused sulfite scaling,
especially in those runs at low slurry solids concentration. It has been
generally observed that sulfite scaling does not occur if the spray tower
outlet pH stays below about 5.5. Therefore, during Run 863-1A, the inlet pH
was adjusted downward slightly at 7.8. In this run and in subsequent runs,
small patches of scale were observed to appear and disappear in a cyclic manner.
7-11
-------
This cyclic appearance of scale did not interfere with scrubber operation.
No effect on S02 removal was discernible as a result of the slight adjustment
in inlet pH.
This slight adjustment in pH is significant in that it demonstrates the need
for good pH control in commercial installations. This adjustment is one of
several operating adjustments that can be made to eliminate a scaling problem.
7.2.10 Effect of Oxidation Tank Level
In Runs 864-1A through 867-1A, the effect of oxidation tank slurry level (and
consequently air and slurry residence times) was explored. In these tests,
air was discharged into the bottom of the oxidation tank through an open 3-
inch pipe as previously described. Major test conditions are listed in Table
7-1. All four runs were made at an oxidation pH of 5.5. The effect of the
tank level is summarized below:
Oxidation Air Stoichiometry, Percent
Run No. Tank Level, ft atoms oxygen/mole S02 abs. Sulfite Oxidation
864-1A 18 1.75 98
867-1A 14 1.80 98
865-1A 10 2.10 89
866-1A 10 3.85/2.70 98/81
Oxidation efficiency was high at 18- and 14-foot tank levels but dropped off at
a 10-foot level. In Run 866-1A, high oxidation efficiency was achieved at a
10-foot tank level by increasing air stoichiometry. Part of Run 866-1A was
made at lower air stoichiometry, with subsequent loss in oxidation efficiency.
7-12
-------
These runs demonstrated that 98 percent sulfite oxidation can be achieved
at 14- to 18-foot tank levels at an air stoichiometry of 1.8 atoms oxygen/
mole SC>2 absorbed. At a 10-foot tank level, an air stoichiometry approach-
ing 4 is required.
It must be pointed out that the oxidation at a 10-foot tank level is not
directly comparable with those at 14 and 18 feet because the top turbine
of the agitator is located at the 11-foot level. In 10-foot slurry level
tests, the top turbine is not in contact with the slurry, and a different
agitation pattern results.
Filter cake solids concentration during these tests was about 85 percent
when sulfite oxidation was 98 percent. In test periods when oxidation
efficiency dropped below 90 percent, the filter cake solids concentration
tended toward a lower range of 80 percent.
7.2.11 Lime Long-Term Reliability Run 863-1A
From mid-December 1977 through mid-January 1978, Run 863-1A, a one-month
lime slurry reliability run, was made with the venturi/spray tower system
in a two-scrubber-loop configuration with forced oxidation in the venturi
scrubber loop. Onstream operation for this run totaled 779 hours (32 days).
As with the limestone reliability Run 819-1A (see Subsection 5.2.13), the
run was designed to demonstrate operating reliability of the scrubber
system with respect to scaling and plugging, and to determine if the EPA
New Source Performance Standards for $03 and particulate emissions could
be met.
7-13
-------
As in the limestone reliability test, the flue gas flow rate was varied be-
tween 18,000 and 35,000 acfm (4.8 and 9.4 feet/sec spray tower superficial
gas velocity) as the boiler load varied between 100 and 155 MW. Flue gas
with high fly ash loading was used. The venturi plug was fixed at a position
to give 9 inches h^O pressure drop across the venturi at a full 35,000 acfm
flue gas flow rate. The actual venturi pressure drop ranged from 2 to 9
inches ^0. The slurry recirculation rates to the venturi and spray tower
were held constant at 600 and 1600 gpm, respectively. The venturi inlet pH
was controlled at 5.5. The oxidation tank level was 18 feet, and the oxida-
tion air flow rate was 210 scfm discharged through a 3-inch pipe. As pre-
viously discussed, the spray tower inlet slurry pH was adjusted downward
from 8.0 to 7.8 to eliminate an observed sulfite scale buildup.
During the run, the scrubber was shut down for a total of 57 hours; 46 hours
were due to boiler outages, 7-1/2 hours were for scheduled scrubber inspec-
tions, and 3-1/2 hours were unscheduled downtime. This resulted in a scrub-
ber availability of 99.6 percent, excluding the interruptions due to boiler
outages and the scheduled inspections. The unscheduled downtime included
2 hours for mist eliminator cleaning and 1-1/2 hours for air compressor
repair.
Average S02 removal for the entire run was 88 percent at 2950 ppm average
inlet S02 concentration. This corresponds to an average emission of 0.9 Ib
S02/MM Btu, well within the EPA standard of 1.2 Ib S02/MM Btu. However,
because of unusually wide fluctuations in inlet SOg concentration and slow
system response time, the S02 emissions at times exceeded the EPA standard
for periods greater than the three hours allowed by EPA regulations.
7-14
-------
The fluctuations in inlet S02 concentration, ranging up to 4700 ppm at one
point, resulted from the wide variety of coals being burned during the 1977-
78 coal strike. Normally, inlet S02 concentration ranges between 2000 and
3000 ppm. These high S02 concentrations were beyond the removal capacity of
the venturi/spray tower system with its limited slurry recirculation rates
(liquid-to-gas ratios of 21 and 57 gal/Mcf at a 35,000 acfm full gas flow
rate in the venturi and spray tower, respectively).
Average particulate loading was 0.046 grain/dry scf (0.034 to 0.059 range),
corresponding to an average emission of 0.09 Ib particulate/MM Btu (assum-
ing 30 percent boiler excess air). This figure is below the EPA emission
standard of 0.10 particulate/MM Btu. However, in a few cases, the EPA
standard was exceeded.
Sulfite oxidation averaged 97 percent during the run, with the air stoichio-
metric ratio varying between 1.4 and 2.8 atoms oxygen/mole S02 absorbed.
The filter cake was excellent throughout the run, and solids concentration
averaged 85 percent. Lime utilization was 90 percent in the spray tower
and 98 percent overall, reflecting the high utilization to be expected in
a two-scrubber-loop system.
At the first scheduled inspection after 160 operating hours, the mist elimi-
nator was found to be 15 percent restricted by solids. After a review of
the history of the mist eliminator exposure, the restriction was attributed
to excess calcium carbonate from the previous limestone run (limestone
stoichiometric ratio of 1.65 in the spray tower) and a failure to activate
the intermittent underwash for the first eight hours of the reliability
run. At the beginning of the reliability run, the mist eliminator under-
7-15
-------
wash had been changed from continuous with diluted clarified liquor
(needed for the limestone run conditions) to intermittent with makeup
o
water (1.5 gpm/ft^ for 6 minutes every 4 hours -- satisfactory for lime
runs). The mist eliminator was cleaned and the run was continued. This
mishap broke a record of 4138 hours of operation under widely varying
conditions without cleaning the mist eliminator.
At subsequent inspections at 399 operating hours and at the end of the run,
the mist eliminator was entirely clean.
In summary, the operating reliability of the venturi/spray tower system
in a two-scrubber-loop configuration with forced oxidation in lime slurry
service has been demonstrated with a system availability of 99.6 percent.
However, under the conditions selected, the system was unable to continually
meet EPA New Source Performance Standards for S02 and particulate emissions
even though the average emissions for the run met the standards.
7.2.12 General Operating Characteristics of the Two-Scrubber-Loop
Forced-Oxidation System with Lime Slurry
The operating characteristics of the two-scrubber-loop forced-oxidation
system with lime slurry was generally the same as that with limestone slurry
The operating characteristics with limestone slurry have been discussed in
Subsection 5.3.
A major difference for the lime slurry was that separate control of the ven-
turi and spray tower inlet pH was required by independent lime feed to each
of the scrubber loops. This method of pH control for lime was necessary
7-16
-------
to avoid wide fluctuations in the venturi and spray tower pH (see Subsection
7.2.1) because lime slurry is less buffered than limestone slurry. With
limestone slurry, the venturi inlet pH could be controlled by adding excess
limestone to the spray tower hold tank only.
7.3 SUMMARY OF FINDINGS
The operating characteristics and test results of the two-scrubber-loop
forced-oxidation lime tests were similar to those of limestone tests in
many respects. The same conclusions could often be drawn for both
lime and limestone operations (see Subsection 5.4).
Findings from lime testing on the venturi/spray tower system included the
following:
• Forced oxidation of sulfite to sulfate in the first of two
independent scrubbing loops was successfully demonstrated in
the two-scrubber-loop venturi/spray tower system with lime
slurry. Successful demonstration of the forced oxidation was
culminated with a one-month reliability lime test using flue
gas with high fly ash loading.
• Under the one-month reliability test conditions of an 18,000
to 35,000 acfm gas rate, 600 gpm venturi slurry rate, 1600 gpm
spray tower slurry rate, 15 percent venturi slurry solids
concentration, 18-foot oxidation tank level, and 5.5 oxida-
tion tank pH, an average sulfite oxidation of 97 percent was
achieved with air stoichiometric ratios ranging from 1.40 to
2.75 atoms oxygen/mole S02 absorbed. The filter cake solids
concentration averaged 85 percent.
0 As with limestone tests, ranges of conditions under which near-
ly complete sulfite oxidation (greater than 95 percent) was
demonstrated were an oxidation tank pH of 4.5 to 5.5, oxidation
tank level of at least 14 feet, and an air stoichiometry of at
least 1.5 atoms oxygen/mole S02 absorbed. Filter cake solids
contents of about 80 percent or higher were achieved under
these conditions.
• The degree of sulfite oxidation was not affected when the
venturi slurry solids concentration was dropped from the normal
7-17
-------
15 percent to 8 percent by weight.
• Oxidation efficiency appeared to drop off at pH's greater
than about 6.
• Sulfite oxidation of 90 percent or higher was required to
obtain maximum improvement in waste solids dewatering
characteristics.
• As with limestone tests, forced oxidation was achieved by
simple air/ slurry contact in the oxidation tank. Within
the ranges of test conditions, an air sparger with 130 1/8-
inch diameter holes, an air sparger with 40 1/4-inch diameter
holes, and a 3-inch diameter pipe all seemed to give the same
air/slurry contact efficiency. Air/slurry contact was pri-
marily achieved by the two-turbine agitator operated at 56 rpm
and rated at 17 brake Kp.
• Slurries with high and low fly ash loadings oxidized equally
well.
• Independent lime addition to both scrubber loops was necessary
to have smooth venturi and spray tower inlet pH control.
• As in the limestone tests, the venturi inlet slurry liquor con-
stantly exhibited a gypsum saturation level of about 100 per-
cent when oxidation was forced within the venturi loop. This
was probably the result of the abundance of gypsum crystal
seeds produced by forced oxidation. The 100 percent saturation
level is well below the incipient scaling level of 135 percent.
t Lime utilization in the spray tower averaged about 88 percent.
Overall lime utilization averaged 98 percent, demonstrating
the advantage of a two-scrubber loop system.
t Spray tower scaling (mostly by calcium sulfite) may occur when
the recirculated slurry solids concentration in the spray tower
drops below about 7 percent (low fly ash loading). This is
especially true when the spray tower outlet pH stays above
about 5.5.
t Intermittent mist eliminator wash on both the top and bottom
sides with makeup water was enough to keep the mist eliminator
clean, with a lime stoichiometry of 1.10 to 1.18 in the spray
tower.
• Oxidation efficiency did not appear to be affected when the
oxidation tank level was dropped from 18 to 14 feet at 1.8 air
stoichiometry and 5.5 pH. At a 10-ft tank level, an air stoi-
chiometry approaching 4 was required to obtain nearly complete
oxidation.
7-18
-------
Section 8
VENTURI/SPRAY TOWER BLEED STREAM
OXIDATION LIMESTONE/MgO TEST RESULTS
Forced oxidation within the scrubber loop requires a compromise between
the conditions needed for good oxidation (low pH with an optimum of 4.5)
and those needed for good S0£ removal (high pH). Such would not be the
case if it were possible to oxidize the slurry bleed stream by simple
air/slurry contact. Bleed stream oxidation would be especially desirable
for retrofitting the forced-oxidation system to existing commercial lime/
limestone flue gas desulfurization facilities because it would have no
effect on the scrubber operation.
However, tests at the IERL-RTP pilot plant (Reference 6) showed that forced
oxidation of a scrubber bleed stream required higher air stoichiometry
compared with oxidation within the scrubbing loop (5.3 versus 3.5). Fur-
thermore, the properties of the oxidized solids were inferior to those
obtained by oxidation within the scrubbing loop.
Bleed stream oxidation tests conducted at the Shawnee Test Facility in
March 1977 confirmed these observations. Tests were conducted using a Pen-
berthy ELL-3 air eductor. A 120-gpm slurry stream was pumped from the oxi-
dation tank through the air eductor, which aspirated air at 30 to 57 scfm
and discharged the air/slurry mixture back to the oxidation tank. Bleed
rate to the oxidation tank was about 3 gpm. Three tests were conducted
8-1
-------
with limestone slurry with a high fly ash loading which was bled from
the TCA system and three tests were conducted with lime slurry with a low
fly ash loading which was bled from the venturi/spray tower system.
In all tests, the pH of the slurry recirculating through the eductor rose
to a level of 7 to 8 because of the dissolution of residual alkali and no
oxidation occurred. Evidently, the dissolution rate of solid calcium sul-
fite (03503) at such a pH level was too slow for significant sulfite oxida-
tion to occur.
In a series of batch oxidation tests conducted on the same equipment during
the boiler outage of April-May 1977, complete oxidation was achieved by con-
tinuously adding 93 percent sulfuric acid to maintain a slurry pH of 5 to 6
in the oxidation tank. Typically, the oxidation rate was about 2.5 x 10"4
g-mole sulfite oxidized per liter per minute. However, sulfuric acid addi-
tion would not seem to be a commercially desirable procedure.
Only marginal improvement was observed in slurry settling rate, settled den-
sity, and filter cake solids content of the oxidized slurry from the batch
tests. The marginal improvement is in agreement with the results observed
in the IERL-RTP pilot plant bleed stream tests.
Despite the generally unfavorable results, batch oxidation tests at the
Shawnee Laboratory indicated that nearly complete sulfite oxidation could be
achieved by simple air sparging of lime or limestone slurry when magnesium
ion was present in concentrations of 1000 ppm or higher. Magnesium ion
apparently has two effects: it tends to buffer the pH rise from dissolving
residual alkali in the waste slurry solids, and it tends to promote dissolved
sulfite availability, allowing oxidation to take place at a higher pH.
8-2
-------
Starting in mid-May 1978, bleed stream forced oxidation with limestone slurry
and added MgO was successfully demonstrated in a month-long series of tests
on the venturi/spray tower system. Oxidized slurry from these tests had
good dewatering properties, and filter cake solids concentration averaged
about 84 percent. Thus, it is commercially feasible to improve the quality
and reduce the volume of waste solids in installations that incorporate
magnesium ion in the slurry liquor by simple air/slurry contact of the bleed
stream.
8.1 SYSTEM DESCRIPTION
The venturi/spray tower system was arranged as shown in Figure 8-1 for the
bleed stream oxidation tests. Both the venturi and the spray tower slurries
discharged into a single hold tank to which limestone and MgO were fed. A
bleed stream was taken from the spray tower downcomer to take advantage
of the low pH at that point. The bleed stream was discharged to the oxida-
tion tank. All tests were conducted at an 18-ft oxidation tank level.
Bleed from the oxidation tank was dewatered by a clarifier and a filter
in series.
8.2 DISCUSSIONS OF TEST RUN RESULTS
Four bleed stream oxidation runs were made on the venturi/spray tower system
using limestone with added MgO. All tests were conducted with approximately
5000 ppm effective magnesium ion concentration in the slurry liquor. Major
test results are reported in Table 8-1. More detailed information for these
tests can be found in Appendices B through G. Percent S02 removal was high
8-3
-------
FLUE GAS
MAKEUP WATER
COMPRESSED
VENT AIR
I
EFFLUENT
HOLD TANK
CLARIFIED LIQUOR
FROM SOLIDS
DEWATERING SYSTEM
BLEED TO
SOLIDS
DEWATERING
SYSTEM
Figure 8-1. Flow Diagram for Bleed Stream Oxidation in the Venturi/Spray Tower System
-------
Table 8-1
RESULTS OF FORCED-OXIDATION TESTS ON THE VENTURI/SPRAY TOWER BLEED STREAM
USING LIMESTONE SLURRY WITH ADDED MAGNESIUM OXIDE
oo
i
tn
tejor Test Conditions
Fly ash loading
Flue gas rate, acfm (? 300°F
Slurry rate to venturi.gpm
Slurry rate to spray tower, gpra
Percent solids redrculated (controlled)
EHT residence time, m1n.
Spray tower Inlet pH (controlled)
Scrubber limestone stolchlometric ratio (control led) (based on solids)
Effective Mg++ concentration, ppm
Venturl pressure drop, 1n. H20
Oxidation tank level, ft
A1r rate to oxidation tank, scfir/ '
Recycle flow from oxidation tank to EHT, gprn
Selected Results
Percent SO, removal
L
Inlet SO, concentration, ppm
c.
Scrubber percent solids recirculated
Scrubber inlet liquor pH
Oxidation pH
Limestone utilization, 51 (based on total slurry)
Sulfite oxidation in oxidation tank, %
Sulfite oxidation in scrubber inlet slurry, J
Gypsum saturation in scrubber Inlet liquor, t
Gypsum saturation In oxidation tank, %
Effective Kg** concentration in scrubber inlet liquor, ppm
Oxidation tank liquor sulflte concentration, ppm
A1r stoichlometry, atoms 0/mole S02 absorbed
Filter cake solids, wtX^
Onstream hours
823-1 A
High
18,000
600
1600
15
11.2
5.3
-
5000
9
18
110
30
94
2600
13.3
5.25
6.30
36
98
86
120
115
4990
65
1.55
83
205
824-1A
High
35,000
600
1600
15
11.2
-
1.9
5000
9
18
210
30
88
2600
14.1
5.25
5.90
38
97
49
85
90
5215
105
1.60
85
159
825-1A
High
18,000
600
1600
15
11.2
-
1.4
5000
9
18
110
0
95
2500
14.7
5.45
5.65
64
97
39
105
115
5380
220
1.60
85
229
826- 1A
High
„ ^««(3)
26,500V '
600
1600
15
11.2
-
1.4
5000
9
18
210
0
89
2750
15.2
5.35
5.45
61
96
29
105
115
4970
230
2.0
84
246
Notes:
Air discharged downward through 3-inch diameter pipe with an open elbow at center of oxidation tank about 3 Inches from tank bottom.
tj Clar1f1er and filter in seYies used for solids dewaterlng in all runs. * * « cnn »rfm
3) Desired flow rate was 35,000 acfm but problems with the venturl plug lifting mechanism limited the rate to 26,500 acfm.
-------
as expected in runs with MgO. Oxidized slurry solids in all runs had good
dewatering properties, averaging about 84 percent solids in the filter
cake.
In Runs 823-1A and 824-1A, conducted at 18,000 acfm and 35,000 acfm, respec-
tively, 97 to 98 percent sulfite oxidation was achieved at an air stoichiom-
etry of about 1.6 atoms oxygen/mole S02 absorbed. Oxidation was consistently
high even though the oxidation tank pH averaged 6.3 in Run 823-1A and at times
*
rose as high as 6.7.
In these runs, 30 gpm of oxidized slurry was recycled from the oxidation tank
back to the scrubber hold tank through the oxidation tank overflow. The pur-
pose of this recycle was to suppress the slurry pH rise in the oxidation
tank (caused by the dissolution of residual alkali) to obtain good oxidation.
However, the opposite occurred. The hold tank pH was depressed, requiring
excess limestone feed to maintain a scrubber inlet pH of 5.3. The net result
was a limestone utilization of less than 40 percent for these runs. This
phenomenon was attributed to the limestone blinding by CaS03 precipitated in
the scrubber slurry loop. Sufficient CaS03 crystal seeds were not available
in the scrubber loop because the slurry had high gypsum content caused by the
recycle stream. The blinded limestone (in high excess quantity) eventually
dissolved in the oxidation tank resulting in the high pH observed.
Runs 825-1A and 826-1A (at 18,000 acfm and 26,500 acfm, respectively) were
conducted without this recycle. In both of these runs, the pH difference be-
tween the scrubber hold tank and the oxidation tank averaged only 0.1 and 0.2,
with the oxidation tank pH averaging 5.65 or less. Oxidation greater than 95
percent was easily achieved in both runs with air stoichiometric ratios of 1.6
8-6
-------
and 2.0, respectively. Time was not available in the test block to determine
minimum air stoichiometry. Control of limestone feed was poor in these runs,
resulting in relatively low limestone utilization (64 and 61 percent, respec-
tively).
This short series of runs has shown that in systems containing magnesium ion,
the slurry bleed stream can be readily oxidized. Furthermore, oxidation of
the bleed stream does not interfere with enhancement of 502 removal by the
magnesium ion as was experienced when oxidation was accomplished within the
scrubber loop (see Subsection 6.2.1).
8.3 SUMMARY OF FINDINGS
Findings from bleed stream oxidation tests included the following:
t Bleed stream oxidation is not expected to work with lime or limestone
slurries without magnesium oxide addition. Slurry pH rise, caused
by the dissolution of residual alkali, prevents solid calcium sulfite
from dissolving and oxidizing in the liquid phase. The pH rise also
reduces the oxidation rate.
• Bleed stream oxidation of limestone slurry is feasible in the presence
of magnesium ion. Magnesium ion tends to buffer the pH rise from
dissolving residual alkali in the waste slurry solids, and it tends to
promote dissolved sulfite availability, allowing oxidation to take
place at a higher pH.
• Average sulfite oxidation of 96 percent or higher was achieved with
1.6 air stoichiometry, 5.4 to 6.3 oxidation tank pH, 18-ft oxidation
tank level, and 5000 ppm effective magnesium ion concentration. Oxi-
dation was consistently high even though oxidation tank pH rose as
high as 6.7.
• Oxidized slurry in all runs had good dewatering properties, averaging
about 84 percent filter cake solids concentration.
• Oxidation of the bleed stream did not interfere with the enhancement
of S02 removal by the magnesium ion as was experienced when oxidation
was accomplished within the scrubber loop.
• In two runs (Run 823-1A and 824-1A), 30 gpm of oxidized slurry was
recycled from the oxidation tank back to the scrubber effluent hold
8-7
-------
tank in an attempt to reduce the pH difference between the two tanks.
However, the opposite occurred. The hold tank pH was depressed, re-
quiring excess limestone feed to maintain a pH of 5.3, resulting in
a very poor limestone utilization (less than 40 percent). This pheno-
menon was attributed to the limestone blinded by CaSOo precipitated
in the scrubber slurry loop. Sufficient CaSOj crystal seeds were not
available because the slurry had high gypsum content caused by the
recycle stream. The 30 gpm recycle flow proved to be unnecessary in
the succeeding two runs (Run 825-1A and 826-1A). Without the recycle
flow, the pH difference between the two tanks averaged only 0.1 and
0.2 pH units with the oxidation tank pH averaging 5.65 or less.
Additional testing has been planned in the future to fully characterize the
1imestone/MgO bleed stream oxidation system. Better operational control is
required to improve the limestone utlization. Higher limestone utilization,
however, is not expected to have an adverse effect on the oxidation efficiency
because of the reduced amount of residual alkali and the correspondingly less
possibility of pH rise in the oxidation tank. The possibility of bleed stream
oxidation on a lime/MgO system must also be investigated.
8-8
-------
Section 9
TCA LIMESTONE TEST RESULTS WITH LOW FLY ASH LOADING
A series of six limestone runs was made from late November 1976 through early
February 1977 using flue gas with low fly ash loading in the TCA system. Two
major objectives for this test block were: 1) to compare system performance
with high and low fly ash loadings in the flue gas and 2) to investigate scrub-
ber inlet slurry pH and SOp removal, with low fly ash loading, as functions
of percent solids recirculated, effluent hold tank residence time, the number
of hold tanks, and limestone stoichiometric ratio.
9.1 SYSTEM DESCRIPTION
The TCA operated during this test block in a single scrubber loop configuration
with either a single effluent hold tank or with three hold tanks in series.
Limestone slurry was fed to the first hold tank. Slurry bleed was taken either
from the single hold tank or the third of the three hold tanks in series. De-
watering was accomplished by a thickener followed by a centrifuge. All slurry
liquor from the dewatering system was returned to the scrubber loop.
9.2 DISCUSSIONS OF TEST RESULTS
Runs 701-2A through 706-2A constituted this test block. The major test
9-1
-------
conditions common to all were:
TCA gas velocity 12.5 ft/sec
Liqirid-to-gas ratio 50 gal/Mcf
Fly ash loading low (0.04 to 0.20 gram/scf dry)
The results of this test block, along with the major operating conditions,
are summarized in Table 9-1. Detailed test results and operating conditions
are reported i-n Appendices H, I, and J.
9.2.1 Effect of Fly Ash
Evaluation of the effect of fly ash on S02 removal performance was not possible
possible because earlier limestone runs with fly ash were made using a different
type of mobile spheres and different liquid and gas rates. However, a limited
comparison of the scrubber inlet slurry pH at the same limestone utilization
indicated that fly ash tends to depress the pH by 0.3 to 0.4 pH unit. This
behavior is consistent with the observation made earlier during lime runs and
utilization testing. In those tests, at a controlled scrubber inlet pH of 8,
lime utilization for the tests with low fly ash loading averaged about 93
percent, compared with about 88 percent for runs made with fly ash under similar
test conditions (Reference 4). This difference was believed to be caused by
the acidic fly ash which releases its acidity from absorbed $03 under scrubber
conditions, even through the fly ash pond liquor at Shawnee is alkaline after
a prolonged period of leaching.
9.2.2 Effect of Recirculated Solids
Run 701-2A was the first and the base case run. In addition to the test
9-2
-------
Table 9-1
SUMMARY OF TCA LIMESTONE TESTS WITH LOW FLY ASH LOADING
10
I
CO
Gas rate, acfm @ 300°F
Liquor rate, gpm
Percent solids recirculated
Residence time, min.
Number of hold tanks
Stoichiometric ratio (controlled)
Solids dewatering system
Scrubber inlet pH
Stoichiometric ratio
Scrubber inlet S02> ppm
Percent S02 removal
Inlet percent sulfate saturation 9 50°C
Percent sulfite oxidation
SCL make-per-pass, m-mol/1
Scaling
Centrifuge cake solids, %
701-2A
30,000
1200
8
4.1
1
1.2
Cl & Ce
5.80
1.21
2700
87
45
12
12.5
Yes (sulfite)
48
702-2A
30,000
1200
15
4.1
1
1.2
(1)
5.85
1.16
3000
81
20
7
13.0
No
58
703-2A
30,000
1200
8
4.1
1
1.1
Cl & Ce
5.65
1.12
2950
74
25
7
12.0
No
54
704-2A
30.000
1200
8
12
3
1.1
Cl & Ce
5.7
1.10
2700
77
65
13
11.5
No
48
705-2A
30,000
1200
15
12
3
1-1
ci
5.85
1.12
2850
84
60
11
12.5
No
~
706-2A
30,000
1200
15
12
1
1.1
Cl
5.55
1.13
2850
81
120
18
12.0
No
—
Notes:
(1) Centrifuge only for the first 90 hours and clarifier only thereafter,
(2) Clarifier (Cl) or clarifier followed by centrifuge (Cl & Ce).
-------
conditions previously mentioned, other pertinent test conditions for this
run were:
Percent recirculated solids = 8 percent
Stoichiometric ratio = 1.21 moles Ca/mole 502 absorbed
Residence time =4.1 minutes
No. of hold tanks = 1
During Run 701-2A, 862 removal averaged 87 percent and inlet slurry pH averaged
5.80 at an average inlet S02 concentration of 2700 ppm.
For Run 702-2A all conditions were the same as those in Run 701-2A except the
recirculated slurry solids concentration was increased from 8 to 15 percent.
S02 removal for this run averaged 81 percent and inlet slurry pH averaged
5.85 units at an average inlet S02 concentration of 3000 ppm. The improvement
expected in S02 removal performance was masked by the higher inlet S02 concen-
tration during Run 702-2A.
The same effect was investigated at a 12-minute residence time with 3 effluent
hold tanks. Run 704-2A was made with 8 percent solids whereas Run 705-2A
was made with 15 percent solids. All other conditions were the same. During
Run 705-2A, the scrubber inlet liquor pH averaged 5.85, up from an average 5.7
during Run 704-2A. The S02 removal averaged 84 percent at an average inlet
S02 concentration of 2850 ppm, compared with 77 percent S02 removal at an
average inlet S02 concentration of 2700 ppm for Run 704-2A.
Hence, the effect of increased recirculated solids is an improved S02 removal
performance. The effect is more pronounced at the larger (12 minutes) resi-
dence time with 3 tanks in series.
9-4
-------
9.2.3 Effect of Limestone Stoichiometric Ratio
Maintaining all other conditions as those in Run 701-2A, the limestone
Stoichiometric ratio was dropped from 1.2 to 1.1 for Run 703-2A.
The results were an average S02 removal of 74 percent and an average inlet
pH of 5.65 at an average inlet S02 concentration of 2950 ppm. As expected,
S02 removal performance deteriorated for this run, and the rate of scale
deposition on the mist eliminator vanes was significantly less than in Run
701-2A.
9.2.4 Effect of Hold Tank Configuration
Runs 705-2A and 706-2A differed only in the hold tank configuration. Run
7052A was made with three tanks in series; Run 706-2A was made with one tank.
During Run 706-2A, the scrubber inlet pH averaged 5.55, compared with 5.85
for Run 705-2A. S02 removal for Run 706-2A averaged 81 percent at an average
inlet S02 concentration of 2850 ppm, compared with 84 percent S02 removal at
the same average inlet S02 concentration for Run 705-2A.
Thus, the overall effect, as seen in earlier tests (see Section 8, Reference 3),
was a slight deterioration in performance in changing from three tanks to one
tank.
9.2.5 Gypsum Unsaturated Operation
Runs 701-2A through 705-2A were operated under conditions which resulted in
very low gypsum saturation. During these runs the sulfate saturation of the
inlet liquor ranged from 20 to 65 percent. The low sulfate saturation was
9-5
-------
most likely due to the extremely low sulfite oxidation (7 to 13 percent) which
may have been the result of operating at relatively high inlet pH (5.65 to 5.85),
In addition, Run 701-2A had indications of sulfite precipitation within the
scrubber, resulting in reduced liquor sulfite available for oxidation in the
effluent hold tank.
9.3 SUMMARY OF FINDINGS
The operating characteristics of limestone scrubbing with high and low fly
ash loadings in the flue gas are very similar. The following is a summary
of findings and conclusions for this test block:
• At a given limestone stoichiometric ratio, high fly ash loading
in the flue gas tends to depress the pH by about 0.3 to 0.4 pH
unit.
0 At a given limestone stoichiometric ratio, increasing recircu-
lated slurry solids concentration from 8 to 15 percent will
increase S02 removal efficiency.
• Increasing limestone stoichiometric ratio improves S02 removal
efficiency but makes it more difficult to keep the mist eliminator
clean.
• At a given limestone stoichiometric ratio and hold tank residence
time, three hold tanks in series results in higher S02 removal
efficiency than a single hold tank.
Except for the first conclusion, the other results were also observed for runs
with high fly ash loading.
9-6
-------
Section 10
TCA LIMESTONE TYPE AND GRIND TEST RESULTS
A limited set of six tests was conducted on the TCA system during the period
from March 4 through April 2, 1977 to investigate the effect of limestone
type and grind on system performance. System performance was evaluated in
terms of S0£ removal and scrubber inlet slurry pH at a controlled stoichio-
metric ratio of 1.2 moles Ca/mole S02 removed. Limestone type, limestone
grind, effluent hold tank (EHT) residence time, and effluent hold tank
configuration were varied.
10.1 TEST PROGRAM
The independent variables evaluated in this test series were:
• Limestone grind -- fine (96 percent less than 325 mesh)
versus coarse (65 to 69 percent less than 325 mesh)
• Limestone type — Fredonia White versus Longview, Alabama
• Effluent hold tank residence time -- 4.1 minutes versus
12 minutes
• Effluent hold tank configuration -- single tank versus
three tanks in series
To allow a comparison among the independent variables within the time available,
the test series was arranged according to Figure 10-1. Table 10-1 shows the
10-1
-------
Figure 10-1
Test Program Arrangement
Run 707-2A
FREDONIA FINE
96% <325 mesh
R.T. = 4.1 min
(1 tank)
Decrease
Grind Fineness
Increase
Residence
Time
i
Run 711-2A
FREDONIA FINE
96% < 325 mesh
R.T. = 12.0 min
(1 tank)
Increase
Grind Fineness
Run 708- 2 A
FREDONIA COARSE
69% < 325 mesh
R.T. = 4.1 min
(1 tank)
i
Run 709-2A
Change LONGVIEW COARSE
Limestone type 65% < 325 mesh
R.T. =4.1 min
(1 tank)
Increase
Residence
Time
f
Run 710-2A
FREDONIA COARSE
69% <325 mesh
R.T. = 12.0 min
(1 tank)
Change
EHT
Configuration
Run 712-2A
FREDONIA FINE
96% <325 mesh
R.T. = 12.0 min
(3 tanks)
-------
Table 10-1
LIMESTONE ANALYSIS FOR RUNS 707-2A TO 712-2A
Bahco Analysis
Micron Size
3
6
9
12
16.5
22
27
30
Screen Analysis
Mesh Size
400
325
200
100
50
30
Number of Sampl es
Fredonia Fine
wt% Less Than
39.5
57.9
67.8
73.8
80.0
84.7
87.6
88.9
wt% Less Than
93.1
95.9
99.3
99.4
100.0
100.0
2
Fredonia Coarse
wt% Less Than
21.1
31.4
38.4
43.6
49.4
54.8
58.6
60.4
wt% Less Than
64.6
69.1
80.8
81.0
99.9
100.0
6
Longview Coarse
wt% Less Than
16.1
26.8
34.4
40.0
47.0
48.0
57.4
59.2
wt% Less Than
61.7
64.8
74.6
74.8
99.3
100.0
2
Chemical Analysis
CaO 54.14 wt%
MgO 0.48
C02 39.56
Acid insolubles 2.18
Number of Samples 6
53.18 wt% 54.23 wt%
0.74 • 0.66
39.54 39.99
2.27 1.27
11 4
10-3
-------
results of size and chemical analyses of the limestone grinds. Both Bahco
and screening procedures were used for size analyses.
10.2 RESULTS
Pertinent results and the major test conditions are presented in Table 10-2.
Detailed test results and operating conditions are reported in Appendices
H, I, and J. Discussion of the effect of each variable tested is presented
bel ow.
10.2.1 Effect of Grind
The effect of grind with Fredonia White limestone was investigated at two
effluent hold tank residence times, 4.1 and 12 minutes. Runs 707-2A and
708-2A were conducted at 4.1 minutes' residence time. As indicated in Table
10-2, use of the more finely ground Fredonia limestone resulted in higher $6
removal by 23 percentage points and higher inlet slurry pH by 0.4 pH unit.
Similar tests, Runs 710-2A and 711-2A, at 12 minutes residence time resulted
in higher S02 removal by 10 percentage points and higher inlet slurry pH by
0.25 pH unit when using the finer ground limestone. Thus, it appears that
the benefits of fine grinding are greater at the lower residence time than
at the higher residence time. This behavior is consistent with the fact
that S02 removal in wet-limestone scrubbing systems is profoundly influenced
by the rate and the amount of time available for limestone dissolution.
10-4
-------
Table 10-2
SUMMARY OF TCA LIMESTONE TYPE AND GRIND TESTS
(with fly ash, no MgO addition)
I-J
en
Major Test Conditions*1^
Limestone type
Average grind, wt% <325 mesh
Gas rate, acfm @ 300°F
Liquor rate, gpm
Percent solids recirculated
Effluent residence time, min.
Limestone stoichiometric ratio (controlled)
Selected Results
Scrubber inlet pH
Stoichiometric ratio
Percent S02 removal
Inlet S02 concentration, ppm
S02 make-per-pass, m-mol /liter
Inlet liquor % gypsum saturation 1? 50°C
Percent sulfite oxidation
Onstream hours
Rotes:
(1) Clarifier only used for solids dewatering in all
(2) Run 709- 2A was terminated prematurely due to sev<
707-2A
Fredonia
96
30,000
1200
15
4.1
1.2
5.6
1.23
72
3100
11.5
100
17
113
runs.
3 re limes torn
708- 2A
Fredonia
69
30,000
1200
15
4.1
1.2
5.2
1.15
49
3100
8.0
115
14
128
! feed line olu
709-2A
Longview
65
30,000
1200
15
4.1
1.2
5.3
1.18
55
3350
9.5
110
20
56<2>
ooina nrohlpmt
710-2A
Fredonia
69
30,000
1200
15
12
1.2
5.4
1.21
54
3400
9.5
110
16
162
711-2A
Fredonia
96
30,000
1200
15
12
1.2
5.65
1.25
64
3250
11.0
80
15
93
712-PA
/It crt
Fredonia
96
30,000
1200
15
12 (3 tanks)
1.2
5.85
1.21
69
3050
11.5
60
18
117
-------
10.2.2 Effect of Limestone Type
To compare the effect of limestone type, limestone from Longview, Alabama, was
tested in Run 709-2A. This limestone was ground to a nominal 70 percent less
than 325 mesh.
Comparison of Runs 708-2A and 709-2A indicates a 6 percent increase in S02
removal and 0.1 unit improvement in inlet slurry pH in going from the Fredonia
coarse to the Longview coarse limestone. However, even though S02 removal im-
proved with the Longview limestone, severe limestone feed-line plugging problems
were experienced. This required premature termination of Run 709-2A. The plug-
ging was attributed to coarsely ground insolubles in the Longview limestone.
10.2.3 Effect of Effluent Hold Tank Residence Time
Results from Runs 707-2A, 708-2A, 710-2A and 711-2A were also evaluated in
terms of the effect of effluent hold tank residence time for a given limestone.
Runs 707-2A and 711-2A were compared to evaluate the effect of an increase
from 4.1 to 12.0 minutes in effluent hold tank residence time for the Fredonia
fine limestone. S02 removal dropped by 8 percentage points at the higher
residence time and the inlet slurry pH was higher by 0.05 pH unit. This
behavior, which was contrary to expectations, may have been the result of
inadequate control of variables other than residence time.
For the coarse Fredonia case, Runs 708-2A and 710-2A were compared. It was
found that increasing the residence time from 4.1 to 12.0 minutes improved
S02 removal by 5 percentage points and the inlet slurry pH by 0.2 pH unit.
This behavior was in line with expectations.
10-6
-------
The relative merits of an increased residence time for the two limestone
grind classes could not be quantified from these runs because of the anomalous
behavior during the tests of finely ground limestone.
10.2.4 Effect of Effluent Hold Tank Configuration
As indicated by comparison of Runs 711-2A and 712-2A, three tanks in series
improved the S02 removal by 5 percentage points and the inlet slurry pH by
0.2 pH unit.
10.3 SUMMARY OF FINDINGS
For the variables evaluated in this limited test series, the fineness of
grind had the greatest effect on improving S02 removal. Effluent hold tank
residence time, limestone type, and effluent hold tank configuration all had
lesser effects.
To achieve a desired degree of SOg removal, consideration must be given to
the cost trade-off between the use of coarse-ground limestone and fine-ground
limestone. Use of coarse limestone requires higher limestone stoichiometry,
resulting in more waste sludge to be disposed of. Use of fine limestone, on
the other hand, requires higher capital and operating costs for the ballmills,
but lower limestone stoichiometry with less waste sludge generated.
A more extensive limestone type and grind testing program is currently scheduled
in the Shawnee Advanced Program.
10-7
-------
Section 11
TCA AUTOMATIC LIMESTONE FEED CONTROL TESTING
A four-week period in the TCA test program, beginning after the April-June
1977 Boiler No. 10 outage, was reserved for the testing of an automatic lime-
stone feed control system. However, electronic stability problems with the
"track-and-hold" module developed during Run 713-2A, and the test block was
deferred to a later date until these problems could be solved. A later
attempt during Run 714-2A was somewhat more successful. Summarized results
are included in Appendices H and I.
11.1 TEST PROGRAM OBJECTIVES
The objectives of the study were to determine the feasibility of automatic
limestone feed control in a limestone wet-scrubbing system and to observe
other benefits that may be gained. Benefits include:
0 Minimizing the limestone usage to meet the S02 emission standard
t Reducing the number of manual operation and operator errors
• Reducing the scrubber operating costs
• Improving reliability by reducing the potential for plugging with
soft mud-type solids caused by excessive limestone stoichiometry
11-1
-------
11.2 BASIS FOR THE LIMESTONE FEED CONTROL SCHEME
In a lime wet-scrubbing system, the lime addition rate can be controlled by
pH variations in the scrubber feed slurry. This type of control is not
feasible for the limestone system because the slurry pH is well buffered
in the normal operating range of 5.6 to 6.0. In view of the above limitation,
the limestone feed control scheme proposed for testing at Shawnee was based
on the material balance concept, i.e., maintaining a desired stoichiometric
limestone feed in relation to the amount of S02 entering or absorbed in the
scrubber. The basic control scheme is represented by the following equation:
Limestone (60 percent slurry) addition rate, gpm = G x (S02I-Kout) x KSR
where G « flue gas flow rate, acfm @ 300°F
S02I = inlet S0 concentration measured by Du Pont analyzer (wet), ppm
Kout = a manually adjustable constant related to desired outlet S02
concentration, ppm
KSR = a manually adjustable constant proportional to the stoichiometric
ratio
= (unit conversion factor) x (stoichiometric ratio)
= 2.34 x 10~B x (stoichiometric ratio)
The factor G(S02I-Kout) represents the amount of S02 absorbed per unit time.
Thus, at a set value of KSR which is proportional to the desired stoichiometric
ratio, the limestone addition rate is automatically adjusted to maintain the
desired stoichiometry.
In practice, inlet S02 concentration, S02I, and flue gas flow rate, G, may vary
within a wide range depending on the sulfur content in the coal and on the boiler
11-2
-------
load. Therefore, the proposed control scheme included overrides which can be
activated when the following situations arise:
• If the outlet S02 concentration exceeds a set maximum, the lime-
stone addition rate will be stepped up to a preset maximum.
• If the outlet S02 concentration drops below a set minimum, the
limestone addition rate will be stepped down to a preset minimum.
The former provision minimizes the infringement of the S02 emission standard;
the latter allows limestone savings by avoiding unnecessary S02 removal.
11.3 EXPERIENCE
Experience with this first-generation design, though quite limited, has
indicated several areas for improvement. One such area involves setting the
limestone flow rate when override conditions (violations of EPA NSPS or excess
removal) are encountered. At present, the system merely reverts to either a
minimum or maximum limestone flow rate setting depending on the override condi-
tion encountered, regardless of the deviation between the S02 outlet set-point
and the actual S02 outlet concentration. This shortcoming of the present de-
sign generally results in underfeeding and overfeeding of limestone, with
resultant loss in performance and sometimes system scaling. The modification
proposed would add a proportioning device to regulate the limestone flow rate
(at the override conditions) in proportion to the difference between outlet
S02 set-point and prevailing S02 outlet concentration. Other modifications,
though somewhat complex, include stoichiometry override and/or pH override.
11-3
-------
11.4 ALTERNATIVES
The Shawnee control scheme presented above can be altered to test other
schemes. Tests of the following schemes are planned:
0 Limestone addition rate = G x (S02I-Kout)KSR, at constant G
and subject to overrides
• Limestone addition rate = 6 x (S02I-Kout)KSR, at constant G
and without overrides
• Limestone addition rate = G x (S02I-Kout)KSR, at variable G
(following boiler load) and subject to overrides
• Limestone addition rate = G x (S02I-Kout)KSR, at variable G
and without overrides
• Limestone addition rate - G x S02I x KSR, at variable G
and subject to overrides
• Limestone addition rate = G x S02I x KSR, at variable G
and without overrides (this scheme is used by Research-
Cottrell at Rickenbacker Air Force Base)
It may not be possible to make all of these tests in a specific time period
allowed. However, if the tests are successful, the testing may be conducted
concurrently with other TCA test blocks. The successful operation of the
automatic limestone feed control system will be gauged according to the
achievement of the following goals:
• To attain reliable (nonscaling) operation of the scrubber and
the mist eliminator
• To minimize (1) the number of times the control overrides occur
and (2) the number of instances of operator adjustments
t To minimize required limestone consumption while meeting the
emission standard
11-4
-------
Section 12
TCA ONE-SCRUBBER-LOOP FORCED-OXIDATION
LIMESTONE TESTS WITH AIR EDUCTOR
Beginning in late June 1977, the TCA system was operated as a single scrubber
stage with forced oxidation in the scrubber slurry loop. As an alternative
to the type of air sparger used in the venturi/spray tower system, an air
eductor was used to bring about air/slurry contact. Tests were conducted with
limestone slurry and with high fly ash loadings. Because sulfite is the major
S02 scrubbing species in a lime scrubber, no one-scrubber-loop forced-oxidation
tests were made with lime slurry. Forced oxidation within the scrubber loop
converts the sulfite to sulfate in the slurry liquor, resulting in poor S02
removal. With limestone slurry, one-scrubber-loop forced oxidation was
successfully accomplished with good SQ^ removal.
The one-scrubber-loop configuration is of prime interest commercially because
most commercial installations, both operating and planned, are of this type.
Modification of these installations for forced oxidation would require as a
minimum a compressor (or blower) plus an air sparger in the scrubber hold tank
(Section 15), or an eductor and a pump, as will be described later.
A total of 15 runs were made using limestone slurry with high fly ash loadings.
These included 11 runs made from June 24 through August 29, 1977, with the air
12-1
-------
eductor mounted in a vertical position over the oxidation tank, and 4 runs
from September 9 through October 4, 1977, with the eductor in a horizontal
position beside the oxidation tank. Each test run normally lasted 5 to 6 days.
Table 12-1 summarizes the major test conditions and important test results
for the 15 runs. Additional detailed information can be found in Appendices
B through D, and H through J.
12.1 SYSTEM DESCRIPTION
An air eductor, Penberthy-Houdaille Model ELL-10 Special, was used to provide
air/slurry contact. The high velocity of the slurry pumped through the educ-
tor nozzle aspirates air into the liquid stream. The high shear developed in
the throat of the eductor breaks air into minute bubbles which are ejected
into the oxidation tank, aerating the slurry in the tank.
The main feature of an eductor is its ability to create smaller air bubbles
than can be obtained with an air sparger, and hence a higher mass transfer
coefficient. However, an air eductor consumes more energy than an air sparger
and is more prone to severe erosion in high-velocity slurry service.
The eductor used was capable of educting 600 scfm of air from the atmosphere
at zero back pressure with a slurry flow rate of 1600 gpm. The eductor nozzle
and the whirler within the nozzle were made of stellite, and the body was made
of neoprene-lined carbon steel.
Two flow configurations were used in the one-scrubber-loop forced oxidation
tests. In the one-tank configuration (Figure 12-1) slurry was recirculated
from the TCA effluent hold tank (oxidation tank) through the eductor and back
12-2
-------
Table IZ-l
SUMMARY OF ONE-SCRUBBER-LOOP FORCED-OXIDATION LIMESTONE TESTS
WITH AIR EDUCTOR ON THE TCA SYSTEM
Major Test Conditions
Fly ash loading
Gas rate, acfm 9 300°F
Slurry flow rate to TCA, gpm
Slurry flow rate to eductor, gpm
Percent sol Ids redrculated
EHT (oxidation tank) residence time, min.' '
Downcomer (low pH) tank residence time, min.^1'
EHT level, ft
Eductor mounting position in EHT
Eductor discharge point, ft from EHT bottom
Limestone stoichiometric ratio controlled at
TCA inlet pH controlled at
EHT (oxidation tank) agitator speed, rpm
Limestone addition poinv '
Air flow rate to eductor, scfm
Selected Results
Percent S02 removal
Inlet S02 concentration, ppm
Percent sulfite oxidation
Air stoichiometry, atoms 0/mole S02 absorbed
TCA inlet pH
Eductor inlet pH
Percent limestone utilization
Gypsum saturation in TCA inlet liquor, %
Onstream hours
801 -2A
High
30,000
1200
1600
15
12
-
6.1
Ver.
6<4>
1.2
-
45
EHT
-600
70
2750
53
6.4
5.8
5.8
80
105
102
801-2B
High
30,000
1200
1600
15
15.7
.
8.0
Ver.
6<4>
1.2
-
45
EHT
530
70
2750
63
5.6
5.8
5.8
80
105
47
802-2A
High
30,000
1200
1600
15
23.5
_
12
Ver.
10
1.2
-
45
EHT
485
76
2900
56
4.5
5.75
5.75
82
100
142
803-2A
High
30,000
1200
1600
15
23.5
_
12
Ver.
10
_
5.0
45
EHT
465
62
2950
91
5.2
5.05
5.05
90
105
165
804-2A
High
30,000
1200
1600
15
23.5
12
Ver.
10
_
5.0
68
EHT
410
62
3000
93
4.5
5.0
5.0
93
105
166
805-2A
High
30,000
1200
1600
15
23.5
.
12
Ver.
10
' _
5.3
68
TCA
470
72
3050
93
4.4
5.35
5.25
84
105
140
806-2A
High
30,000
1200
1200
15
23.5
12
Ver.
10
.
5.3
68
TCA
290
70
3000
58
2.8
5.35
5.3
91
110
115
Note: Footnotes for this table are listed in Table 12-1 (continued).
-------
Table 12-1 (continued)
SUMMARY OF ONE-SCRUBBER-LOOP FORCED-OXIDATION LIMESTONE TESTS
WITH AIR EDUCTOR ON THE TCA SYSTEM
ro
Major Test Conditions
Fly ash loading
Gas rate, acfm U 300°F
Slurry flow rate to TCA, gpm
Slurry flow rate to eductor, gpm
Percent solids recirculated
EHT (oxidation tank) residence time, min.'1'
Downcomer (low pH) tank residence time, mln.^
EHT level, ft
Eductor mounting position in EHT
Eductor discharge point, ft from EHT bottom
Limestone sto1ch1ometric ratio controlled at
TCA inlet pH controlled at
EHT (oxidation tank) agitator speed, rpm
Limestone addition point' '
Air flow rate to eductor, scfm
Selected Results
Percent SOg removal
Inlet SOg concentration, ppm
Percent sulfite oxidation
Air stolen iometry, atoms 0/mole S02 absorbed
TCA inlet pH
Eductor Inlet pH
Percent limestone utilization
Gypsum saturation 1n TCA Inlet liquor, 5!
Onstream hours
807-2A
High
30 ,000
1200
1600
15
23.5
2.9
12
Ver.
10
-
5.4
68
EHT
475
83
2700
96
4.3
5.45
5.1
82
100
185
808- 2A
High
30,000
1200
1600
15
23.5
2.9
12
Ver.
10
1.3
.
68
EHT
200<5>
71
2900
93
2.0
5.1
4.95
63
125
162
809- 2A*3)
High
30,000
1200
1600
15
15.7
1.9
8
Ver.
6(4)
1.3
-
68
EHT
300<5>
77
3200
96
2.5
5.2
5.0
63
115
149
810 2A(3)
High
30,000
1200
1600
15
15.7
1.9
8
Ver.
6(4)
.
5.4
68
EHT
300<5>
84
2900
98
2.5
5.5
5.15
77
110
130
811-2A
High
30,000
1200
1600
15
23.5
2.9
12
Hor.
0.8
1.3
-
68
EHT
270
87
2800
96
2.3
5.85
5.45
60
100
164
812-2A
High
30,000
1200
1600
15
15.7
1.9
8
Hor.
0.8
1.3
-
68
EHT
350
81
3000
93
2.9
5.5
5.3
56
100
141
813-2A
High
30,000
1200
1600
15
23.5
-
12
Hor.
0.8
-
5.3
68
TCA
310
80
2850
95
2.8
5.35
5.15
90
90
130
814-2A
High
30,000
1200
1200
15
15.7
-
8
Hor.
0.8
-
5.3
68
TCA
145
74
2550
61
1.6
5.35
5.3
85
115
119
Notes:
1)
2)
3)
(4)
(5)
Residence times are based on slurry flow rate to TCA.
EHT (effluent hold tank) or TCA (pump suction on the TCA Inlet slurry line).
Filter in series with clarlfier was used during Runs 809-2A and 810-2A. Filter cake solids contents were 85 and 88%, respectively.
At discharge of 4 ft long by I0-1nch diameter pipe attached beneath the eductor.
Air flow controlled by a sliding plate in the air Intake line to eductor.
-------
MAKEUP WATER
FROM SOLIDS DEWATERING
SYSTEM
BLEED TO SOLIDS
DEWATERING SYSTEM
OXIDATION
TANK
Figure 12-1. Flow Diagram for One-Scrubber-Loop Forced Oxidation
with Air Eductor in the TCA System Using One Tank
12-5
-------
into the hold tank. The hold tank was 20 ft in diameter and operated at
slurry levels from 6 to 12 feet. Limestone makeup was added either to the
hold tank or to the pump suction of the scrubber recirculation line. In the
two-tank configuration (Figure 12-2) a small downcomer tank (7-ft diameter)
was used to receive the slurry from the scrubber. Slurry was then pumped
from the downcomer tank through the eductor to the large hold tank (oxidation
tank) where limestone was added. A slurry tie line connecting the two tanks
equalized the slurry level in both tanks. The slurry flow rate through the
eductor was maintained at a higher level than the slurry flow rate to the
scrubber. Thus, there was always a slurry backflow through the tie line
from the hold tank to the downcomer tank. The two-tank configuration had
the advantage that the pH of the slurry passing through the eductor was
lower and thus the chemical reaction rate was higher.
In both one-tank and two-tank configurations, tests were made with the eductor
mounted in both vertical and horizontal positions. In the vertical position,
as shown in Figures 12-1 and 12-2, the eductor was mounted above the effluent
hold tank (oxidation tank) in an off-centered position, and the air/slurry
mixture was discharged vertically downward. Figure 12-3 shows the detailed
arrangement of the oxidation tank with the eductor in a vertical position.
With the eductor in a horizontal position (not shown in figures), air/slurry
mixture was discharged into the oxidation tank through the tank wall in a
horizontal, radial direction (at about 4 o'clock position in the plan view
of Figure 12-3) and at an elevation 10 inches above the tank bottom.
Normally, slurry was bled from the oxidation tank to a clarifier for dewatering.
The clarifier operated with an underflow solids concentration of about 35 to
45 percent. In two runs (Runs 809-2A and 810-2A), the clarifier was followed
12-6
-------
MAKEUP WATER
FLUE GAS
LIMESTONE
CLARIFIED LIQUOR^
FROM SOLIDS DEWATERING
SYSTEM
BLEED TO SOLIDS
DEWATERING SYSTEM
FLUE GAS
\
LT
o o o
00000
oo o
oo o o o
ooo
o o o o c
TCA
AIR (ATM. PRESS.)
EDUCTOR
(PENBERTHY
ELL-10 SPECIAL)
TANK
OOWNCOMER
HOLD TANK
Figure 12-2. Flow Diagram for One-Scrubber-Loop Forced Oxidation
with Air Eductor in the TCA System Using Two Tanks
12-7
-------
AGITATOR
DOWNCOMER
BAFFLE
AIR (ATM. PRESS.)
-TO EDUCTOR
PLAN VIEW
INLET
AIR (ATM, PRESS.)
SLURRY
DOWNCOMER
BAFFLE
TO EDUCTOR
ELEVATION VIEW
01234 5
SCALE, FEET
Figure 12-3. Arrangment of the TCA Oxidation Tank with Air Eductor
in Vertical Position
12-8
-------
by a rotary drum vacuum filter.
12.2 DISCUSSION OF TEST RUN RESULTS
In a one-scrubber-loop forced-oxidation system using limestone slurry, a
compromise must be made in the scrubber liquor pH between a higher pH
desired for good S02 removal and a lower pH required for good sulfite
oxidation. Although the optimum oxidation reaction rate occurs at about
4.5 pH (Reference 3), it has been found that the oxidation rate is adequately
fast up to a pH of about 6. Thus, the oxidation pH range is compatible
with the limestone S02 scrubbing pH range of 5 to 6.
12.2.1 Effect of Tank Level in the Initial Runs
Run 801-2A was made with a one-tank configuration and a vertical eductor
position (Figure 12-1). With 12 minutes residence time in the effluent hold
tank normally used in the past operation without forced oxidation, the slurry
level in the 20-ft diameter tank was only 6.1 feet. This was an awkward
height-to-diameter ratio for forced oxidation. Good air/slurry contact was
difficult to maintain, and the air bubble residence time was short in the
shallow tank. Average sulfite oxidation was only 53 percent, with an air
stoichiometric ratio of 6.4 atoms oxygen/mole S02 absorbed. Air/slurry
mixture from the eductor discharged at the slurry surface in the tank
through a 4-ft long by 10-inch diameter pipe attached beneath the eductor.
In Run 801-2B, sulfite oxidation was increased to 63 percent by raising the
tank level to 8 feet and discharging the eductor 2 feet below the slurry sur-
face. The 2-ft back pressure on the eductor discharge reduced the educted
12-9
-------
air rate from 600 to 530 scfm, with a corresponding decrease in the air
stoichiometry from 6.4 to 5.6.
The oxidation tank level was further raised to 12 ft in Run 802-2A. The
4-ft long spool piece on the eductor discharge was removed, and the eductor
was still discharging at 2 feet below the slurry surface. Sulfite oxidation
was still poor at 56 percent.
These initial runs were all made at a controlled limestone stoichiometry of
1.2, which corresponded to an oxidation tank pH of 5.8.
12.2.2 Effect of Low pH
In the next two runs (Runs 803-2A and 804-2A), still with a one-tank configura-
tion and a vertical eductor position, better than 90 percent sulfite oxidation
was achieved by dropping the oxidation tank pH (TCA inlet pH) to 5.0. These
were the first successful one-scrubber-loop forced-oxidation runs in the TCA
system at the Shawnee Test Facility. As expected with the low pH, the S02
removal was only 62 percent for these runs at 3000 ppm inlet SOg concentrations.
Assuming that air/slurry contact time was limited by the poor TCA hold tank
configuration, the lower pH must have increased the reaction rate sufficiently
for near complete oxidation to take place. At the IERL/RTP pilot plant, where
the hold tank had a height-to-diameter ratio approaching one and the mixing was
probably more uniform, no limiting effect of pH on oxidation was observed up
to about pH 6.0 (Reference 6).
Run 804-2A differed from Run 803-2A by an increase in agitator speed from 45
rpm to the maximum obtainable of 68 rpm. The effect was to increase the average
12-10
-------
oxidation slightly (from 91 to 93 percent) and to reduce time-dependent
fluctuations. Subsequent runs were all made at 68 rpm agitator speed.
12.2.3 SOg Removal Improvement
To improve S02 removal and still maintain sufficiently low pH in the oxidation
tank for good oxidation, the limestone addition point was moved from the efflu-
ent hold tank to the pump suction on the scrubber inlet feed line, and the
scrubber inlet pH was increased to 5.3 (Run 805-2A). The net result was an
increase in S0£ removal to 72 percent (at 3050 ppm inlet concentration) with
a continuation of 93 percent sulfite oxidation. Addition of the limestone to
the scrubber feed line allowed the air/slurry contact to take place at the
lowest pH in the scrubber system.
This run was the most successful in the series of limestone runs with a single
hold tank and a vertical eductor position. The slurry circulation rate through
the eductor was 1600 gpm which gave an air rate of 470 scfm, corresponding
to an air stoichiometry of 4.4 atoms oxygen/mole S02 absorbed.
12.2.4 Effect of Decreased Slurry Rate to the Eductor
In Run 806-2A, the eductor feed rate was reduced to 1200 gpm, with a correspond-
ing reduction in educted air to 290 scfm. Sulfite oxidation dropped to only
58 percent, indicating poor mixing in the oversize effluent hold tank with the
reduced eductor discharge flow. The air rate was adequate, as indicated by
good oxidation in later runs at a 1600 gpm eductor feed rate and air rates as
low as 200 scfm.
12-11
-------
12.2.5 Effect of TCA Inlet pH in a Two-Tank Configuration
Runs 807-2A through 810-2A were made with a two-tank configuration and with
vertical eductor position (Figure 12-2). With the two-tank mode of operation,
fresh limestone was added to the effluent hold tank (oxidation tank), and the
eductor inlet pH was maintained at a lower level than the TCA inlet pH (oxi-
dation tank pH). The TCA inlet pH was varied from 5.1 to 5.5 while the educ-
tor inlet pH changed only from 4.95 to 5.15. No systematic effect of pH on
sulfite oxidation was observed, but SOg removal increased from 71 to 84 percent
as the pH was increased.
12.2.6 Effect of Air Rate
The air rate to the eductor was dropped from 475 scfm in Run 807-2A to 200
scfm in Run 808-2A with only a 3 percent reduction in sulfite oxidation (96
percent oxidation reduced to 93 percent). The air rate was controlled at a
constant eductor slurry rate of 1600 gpm by restricting the eductor air inlet.
The 200 scfm air rate, corresponding to an air stoichiometry of 2.0 atoms
oxygen/mole S02 absorbed, was the lowest air rate in the single-loop oxidation
series with eductor in which better than 90 percent oxidation was achieved.
12.2.7 Effect of Slurry Tank Level
Of the four runs with a two-tank configuration and a vertical eductor mounting
(Runs 807-2A through 810-2A), two were made at a 12-ft oxidation tank level
12-12
-------
(23.5 minutes' residence time) and two were at an 8-ft tank level (15.7 minutes'
residence time). Sulfite oxidation occurred equally as well at the low tank
level (96 and 98 percent) as at the high level (96 and 93 percent). These
runs were made over an eductor inlet pH range of 4.95 to 5.15.
12.2.8 Effect of Eductor Mounting Position
Runs 811-2A through 814-2A were made with the eductor mounted in a horizontal
position as previously described in Subsection 12.1. In these tests, the
eductor discharged against a back pressure close to the liquid head in the
oxidation tank.
Runs 811-2A, 812-2A, and 813-2A were conducted under similar conditions as
Runs 808-2A, 809-2A, and 805-2A, respectively, except for the difference in
the eductor position. No major trend in sulfite oxidation was observed in
these runs with a change in eductor position. Percent sulfite oxidations
in both cases were in the mid 90's.
In Run 814-2A, the slurry flow to the eductor was reduced to 1200 gpm to
.observe the effect on sulfite oxidation. The oxidation tank level was also
reduced from 12 to 8 ft to obtain a reasonable air flow rate. Actual educted
air flow rate was only 145 scfm with the eductor discharging against 7.2 ft
of slurry head in the oxidation tank. Sulfite oxidation dropped to only 61
percent, apparently because of both poor mixing (see Subsection 12.2.4) and
insufficient air.
12-13
-------
12.3 GENERAL OPERATING CHARACTERISTICS OF THE ONE-SCRUBBER-LOOP SYSTEM
WITH FORCED OXIDATION USING AN AIR EDUCTOR
12.3.1 Water Balance
In the TCA system, forced oxidation did not significantly change the water
balance. In a majority of runs only a clarifier was used for dewatering,
and the underflow waste solids concentration was run at 35 to 45 percent,
the same as without forced oxidation.
During Runs 809-2A and 810-2A, however, the rotary drum vacuum filter was
used in series with the clarifier, and the oxidized filter cake averaged
over 85 percent solids concentration. During these runs, the water balance
was tighter than normal when clarifier only was used.
12.3.2 Enhancement of S02 Removal
S02 removal during the one-scrubber-loop forced-oxidation tests has tended
to be a few percentage points higher than predicted by the S02 removal
model fitted to previous Shawnee TCA limestone runs without forced oxidation
(Reference 4). The reason for this removal enhancement has not yet been
determined. However, one speculation is the limestone dissolution rate has
been increased because of the reduction of carbonate in the liquor by air
stripping of C02 and the reduction of bisulfite by oxidation. Increased
agitation in the effluent hold tank by the eductor discharge plume may also
have promoted the limestone dissolution rate.
12-14
-------
12.3.3 Chloride Levels
Chloride-ion concentration during the TCA limestone forced oxidation tests
ranged from 1000 to 3000 ppm, which is typical of previous tests without
forced oxidation. As previously mentioned, the water balance and the tight-
ness of the liquor loop were not significantly changed.
12.3.4 Eductor Performance
Air eduction by the eductor was as predicted by the manufacturer. The air
eduction rate was sensitive to the back pressure (i.e., slurry submergence of
the eductor) and ranged from 600 scfm at zero submergence to 200 scfm at 12 feet
submergence at a 1600 gpm slurry rate. At 1600 gpm, the slurry pressure drop
across the eductor was 38 psi. As seen from the two-tank runs in the TCA
system, over 90 percent sulfite oxidation could be achieved at an air rate as
low as 200 scfm.
The major operational problem in using the eductor for air/slurry contact
resulted from the configuration of the eductor discharge hold tank. Because
the residence time in the eductor is extremely short, most of the oxidation
takes place in the hold tank. The 20-ft diameter hold tank was too large and
normally operated at only a 6.1 ft slurry level to achieve 12 minutes' resi-
dence time. This slurry level was too shallow to allow time for the oxidation
reaction to take place. At 8- and 12-ft slurry levels, over 90 percent oxi-
dation could be achieved, but only by reducing the pH to increase the reaction
rate.
If the eductor were discharged into a smaller but deeper tank, such as the
oxidation tank in the venturi/spray tower system, good oxidation could
-------
probably be achieved at a higher pH and a lower slurry rate.
The eductor body was constructed of neoprene-lined carbon steel. The nozzle
and the whirler were made of stellice. The rubber lined body was not satis-
factory in slurry service with high fly ash loadings. After 1500 hours, the
rubber had eroded in a circular pattern, presumably at the point of impact
of the slurry from the nozzle. In a few spots, the rubber had worn to bare
steel. Repair with Epoxylite 203 was unsuccessful. Without the rubber
lining, the bare steel had eroded through after an additional 550 hours.
The stellite nozzle and whirler showed only minor evidence of erosion.
12.4 SUMMARY OF FINDINGS
Nearly complete forced oxidation using an air eductor was achieved in the one-
scrubber-loop TCA system with limestone slurry having high fly ash loading.
The following is a summary of findings based on the results of these tests:
• Under the base case conditions of 30,000 acfm gas rate, 1200 gpm
TCA slurry rate, 1600 gpm eductor slurry rate, 15 percent recircu-
lated slurry solids (with fly ash), and an 8- to 12-ft slurry level
in the effluent hold tank (oxidation tank), sulfite oxidation of
better than 90 percent was achieved at eductor inlet pH range of
4.95 to 5.45 and air stoichiometric ratios as low as 2.0 atoms
oxygen/mole S02 absorbed.
• The dewatering and handling characteristics of slurry solids
oxidized to 90 percent or higher in a one-scrubber-loop system
were as good as those in a two-scrubber-loop system.
• Operation in a two-tank configuration allowed for a higher pH to
the TCA for good SO^ removal and a lower pH to the eductor for
good sulfite oxidation. Sulfite oxidations of up to 98 percent
were achieved with satisfactory S02 removal (80 to 90 percent).
• With a single-tank configuration and limestone addition to the
effluent hold tank, low hold tank pH (hence, low SO? removal) was
necessary to achieve good sulfite oxidation. The $62 removal
could be improved, however, by limestone addition to the TCA inlet
slurry stream instead of the hold tank.
12-16
-------
• S02 removal tended to be a few percentage points higher than the
predicted SOp removal under similar operating conditions without
forced oxidation. The reason for this removal enhancement is
believed to be air stripping of C02 from the liquor, which
increases the limestone dissolution rate.
t No significant difference in sulfite oxidation efficiency was
observed between operations with the air eductor in the vertical,
top-entry position or in the horizontal, bottom-entry position.
0 Sulfite oxidation was unsatisfactory (about 60 percent) when
the eductor slurry flow rate was reduced from 1600 to 1200 gpm.
Decreased agitation in the oxidation tank at the lower flow rate,
rather than the reduced air stoichiometric ratio, is believed to
be the main reason for poor oxidation.
• Air eduction by the eductor was as predicted by the manufacturer.
Air flow rate was sensitive to the depth of the eductor's submer-
gence in the slurry, ranging from 600 scfm at zero submergence to
200 scfm at 12 ft submergence, at a 1600 gpm slurry rate. A flow
of 200 scfm of air was sufficient to achieve good sulfite oxidation
at 1600 gpm slurry flow.
• Air/slurry contact was hampered by the configuration of the eductor
discharge hold tank (20-ft diameter tank with only 8 to 12 ft normal
liquid depth). A smaller diameter but deeper tank would probably
have given better oxidation efficiency.
• Erosion of the air eductor (in limestone/fly ash service) was a
problem. The eductor body was constructed of neoprene-lined carbon
steel. The rubber near the nozzle had eroded in a circular pattern
after 1500 operating hours, and the carbon steel body had eroded
through after an additional 550 hours. The stellite nozzle and
whirler showed only minor evidence of erosion.
12-17
-------
Section 13
TCA LIMESTONE TEST RESULTS WITH HIGH FLY ASH LOADING
The ultimate objective of this test block was to identify those test con-
ditions that would allow compliance with the EPA S02 and particulate emis-
sion standards and at the same time permit reliable system operation. The
efforts towards this objective culminated in Run 717-2A. Run 717-2A was
preceded by two exploratory runs, Runs 715-2A and 716-2A.
13.1 LIMESTONE/HIGH FLY ASH RUN 715-2A
The test conditions for Run 715-2A (see Table 13-1 and Appendices H and I)
were selected such that both the EPA S02 and particulate emission stan-
dards could be met. Three tanks in series were chosen to give good lime-
stone utilization. A total static sphere height of 22.5 inches (3 beds at
7.5 inches/bed) was selected in order to meet the particulate emission
standard. At this bed height and at a 30,000 acfm gas rate, the slurry flow
rate had to be reduced to 1000 gpm from the normal 1200 gpm to avoid
flooding.
S0£ removal for the run averaged 86 percent at an average inlet S02 concen-
tration of 3000 ppm and a limestone utilization of 81 percent. Average outlet
particulate mass loading was 0.039 grain/dry scf.
13-1
-------
Table 13-1
LIMESTONE TESTS WITH HIGH FLY ASH LOADING USING THREE TANKS IN SERIES
ON THE TCA SYSTEM
Major Test Conditions
Gas rate, acfm @ 300°F
Liquor rate, gpm
Percent solids recirculated
Total residence time (3 tanks), min.
TCA inlet pH controlled at
Stoichiometric ratio controlled at
Total static bed height (3 beds), inches
Selected Results
Percent S02 Removal
Inlet S02 concentration, ppm
Average outlet mass loading, gr/dry scf
Percent sulfite oxidation
TCA inlet pH
Percent limestone utilization
Percent gypsum saturation in TCA inlet. liquor
Mist eliminator restriction, percent*1'
On stream hours
715-2A
30,000
1000
15
14.4
5.75
—
22.5
86
3000
0.039
21
5.75
81
100
0
161
716-2A
27,000
1000
15
14.4
5.75
—
30
91
2700
0.025
25
5.8
86
105
0
135
717-2A
Variable^2*
1000
15
14.4
—
1.2
22.5
87
2800
0.043
22
5.9
81
90
1
747
Notes;
(1) Continuous mist eliminator bottom wash with diluted clarified liquor. Wash rate was
0.4 gpm/ft2 for Runs 715-2A and 716-2A and 0.3 gpm/ft2 for Run 717-2A.
(2) Gas rate varies between 20,000 and 30,000 acfm, depending on the No. 10 boiler load.
13-2
-------
Both the S02 and participate emission standards were met during Run 715-2A,
although not by a comfortable margin.
13.2 LIMESTONE/HIGH FLY ASH RUN 716-2A
Although the particulate emission during Run 715-2A met the EPA standard,
the particulate emission level would probably give very poor opacity if
extrapolated to a 500 MW size stack. It is estimated that to get a 20
percent opacity {ror a 500 MW stack), an outlet mass loading of about 0.02
grain/dry scf, or less, would be required.
Run 716-2A (see Table 13-1 and Appendices H and I) was made with 30 inches
total static bed height (3 beds at 10 inches/bed) to observe whether an out-
let mass loading of 0.02 grain/dry scf, or less, could be achieved. The flue
gas flow rate had to be reduced from 30,000 to 27,000 acfm to avoid flooding
of the scrubber at 1000 gpm slurry flow rate and 30 inches total bed height.
Average outlet particulate mass loading decreased to 0.025 grain/dry scf
for Run 716-2A, short of the 0.02 grain/dry scf target. S02 removal aver-
aged 91 percent at 2700 ppm inlet S02 and 86 percent limestone utilization.
The improvements in both S02 and particulate removal resulted, of course,
from the higher scrubber pressure drop (10 to 12.5 inches H20 for Run 715-2A
versus 12.8 to 16.1 inches H20 for Run 716-2A).
13.3 LIMESTONE/HIGH FLY ASH RELIABILITY RUN 717-2A
During October/November 1977, Run 717-2A, a one-month limestone slurry reli-
ability run, was made in the TCA system. This run was conducted over 747
13-3
-------
hours (31-1/8 days), and was designed to demonstrate operating reliability
of the scrubber system with respect to scaling and plugging and to determine
if the EPA New Source Performance Standards for $62 and particulate emissions
could be met. The fan damper, the outlet duct to the reheater, the bottom
bar grid, and the scrubber walls below the bottom grid were cleaned before
the run. Detailed operating conditions and test results for Run 717-2A are
presented in Appendices H and I.
The run was made without forced oxidation and with flue gas containing high
fly ash loading. Three, hold tanks in series with a total residence time of
14.4 minutes were used to improve the limestone utilization. The flue gas
flow rate was varied between 20,000 and 30,000 acfm (gas velocity between
8.4 and 12.5 ft/sec) according to the No. 10 Boiler load, which normally fluc-
tuated between 100 and 155 MW. The slurry flow rate was constant at 1000
gpm; the recirculated slurry solids concentration was 15 weight percent.
The limestone stoichiometric ratio was controlled at 1.2 moles Ca/mole S0£
absorbed by the automatic limestone feed system. Three beds were used, with
7.5 inches static height of nitrile foam spheres in each bed.
A total of three scheduled inspections were made during the run. These in-
spections normally lasted less than 5 hours. During the third inspection on
November 14, the scrubber was down for 5 extra hours so that a problem on the
limestone feeder weigh belt could be resolved. This was the only scrubber
unavailable time during Run 717-2A, giving a scrubber availability of 99.3
percent excluding the downtime for scheduled inspections.
Average SOg removal for the entire run was 87 percent at 2800 ppm average
inlet S02 concentration. This average removal efficiency was higher than
13-4
-------
the 84 percent required to meet the 1.2 Ib S02/MM Btu emission standard.
However, the standard was exceeded frequently for periods greater than the
three hours allowed by EPA regulations due to wide fluctuations in inlet S02
concentration and slow system or operating response time.
At the controlled limestone stoichiometric ratio of 1.2, the TCA inlet pH
averaged 5.9 and the limestone utilization averaged 81 percent. Sulfite
oxidation was 22 percent.
The outlet particulate loading during Run 717-2A ranged from 0.026 to 0.069
grain/dry scf, with a run average of 0.043 grain/dry scf. Assuming a 30
percent boiler excess air, the outlet particulate loading should be 0.052
grain/dry scf or less to meet the emission standard of 0.10 Ib particulate/
MM Btu.
The mist eliminator was entirely clean at the beginning of the run. Routine
inspection showed that it was 1 percent restricted by solids after 161 on-
stream hours, 2 percent after 350 hours, entirely clean after 587 hours, and
1 percent after 747 hours at the end of the run. There was no significant
increase in solids deposits on the scrubber internals during the run.
In summary, reliable operation of the TCA system was demonstrated during
Run 717-2A. However, as in the venturi/spray tower reliability limestone
Run 819-1A, the scrubber performance did not meet the EPA new source S0£
and particulate emission standards at all times, even though the run-averaged
S0£ and particulate emissions met these regulations.
13-5
-------
Section 14
CEILCOTE SUPPORT PLATE PACKING TESTS ON THE TCA SYSTEM
14.1 INTRODUCTION
After completion of the TCA long-term limestone reliability run, Run 717-2A,
two weeks were needed to complete the installation of an air-sparging system
for forced-oxidation testing. The Shawnee Project Steering Committee decided
that during these two weeks a series of short tests would be conducted using
scrubber internals other than the 1-5/8 inches, 6.5 gram solid nitrile foam
spheres.
An egg-crate type plate (packing support plate) manufactured by Ceilcote Com-
pany was selected. The plate dimensions were 2 feet x 2 feet x 2 inches, with
a 1-3/16 inch square opening. They were made of fiberglass filled polypropy-
lene. The nitrile foam spheres in the TCA were removed with the four bar
grids remaining in the scrubber. Twenty-three layers of Ceilcote plates (46
inches in height) were installed between the second and third bar grids.
The number of plate layers was selected to give an estimated pressure drop of
8 inches 1^0 (including the four grids) at a 30,000 acfm gas rate and a 1200
gpm slurry flow rate. This is a typical pressure drop obtained under these
flow rates for a three-bed, four-grid TCA with 5-inch static height of spheres
per bed. This arrangement allows a direct comparison of scrubber performances
14-1
-------
between the plate packing and the mobile spheres. The objective of these
tests was to obtain preliminary data on the Sf^ and participate removal effi-
ciency for the selected packing. Due to the short duration of the tests,
it was not expected that sufficient information for assessing the reliability
of this packing with respect to scaling and plugging would be obtained.
14.2 TEST PLAN
Two one-week test runs were made. The first test (Run 718-2A, see Subsection
2.4) was made under the following typical TCA operating conditions so that the
S0£ and particulate removal data could be directly compared with those of a
TCA scrubber:
• 30,000 acfm gas rate
• 1200 gpm slurry flow rate
0 High fly ash loading
• 15 percent solids in recirculated slurry
0 12 minutes slurry residence time
• 1.2 limestone stoichiometric ratio on automatic control
The second test (Run 719-2A) was made under the same conditions as the first,
except that the gas rate was reduced to 18,000 acfm to observe the turndown
capability of the plate packing.
14.3 RESULTS OF RUNS 718-2A AND 719-2A
S02 removal for Run 718-2A, as shown in Table 14-1, averaged 76 percent at 3200
ppm average inlet S02 concentration. The inlet pH averaged 5.75 and limestone
14-2
-------
I
oo
Table 14-1
LIMESTONE TESTING ON THE TCA SYSTEM
WITH CEILCOTE PLATE PACKING
Major Test Conditions
Gas rate, acfm @ 300°F
Liquor rate, gpm
Percent solids recirculated
Residence time, min.
Stoich. ratio controlled at
Height of plate packing, inches
Selected Results
Percent SO» removal
2
Inlet SO_ concentration, ppm
Average outlet mass loading, gr/dry scf
Percent sulfite oxidation
TCA inlet pH
Percent limestone utilization
Percent gypsum sat'n in TCA inlet liquor
Mist eliminator restriction, percent
718-2A
30, 000
1200
15
12
1.2
46
76
3200
0.042
21
5.75
75
110
0.05
719-2A
18,000
1200
15
12
1.2
46
86
3050
0.057
19
5. 75
86
105
0.5
-------
utilization averaged 75 percent. The flue gas pressure drop across the 23
layers (46 inches) of Ceilcote plates and four bar grids, measured in a series
of pressure drop tests conducted prior to the runs (see Figure 14-1), was 7.2
inches H20 or about 10 percent less than estimated before these tests.
Under similar operating conditions, a three-bed, four-grid TCA with 5 inches
static height of 1-5/8 inch diameter solid nitrile foam spheres per bed (about
9 inches H20 flue gas pressure drop) would give about 80 percent S02 removal,
(Reference 3) somewhat better than the plate packing.
Run 719-2A (see Table 14-1) was similar to Run 718-2A except that the gas rate
was reduced from 30,000 to 18,000 acfm to compare the turndown capacity of the
Ceilcote plate packing with that of the 1-5/8 inch diameter, 6.5-gram solid
nitrile foam spheres. Average S02 removal for this run was 86 percent at an
average inlet S02 concentration of 3050 ppm. The inlet pH averaged 5.75 and
limestone utilization averaged 86 percent. Flue gas pressure drop across the
plates and four bar grids as shown in Figure 14-1 averaged about 2.9 inches
H20.
Under the same operating conditions, a three-bed, four-grid TCA with 5 inches
of spheres per bed (about 4.5 inches H20 flue gas pressure drop) gives about
78 percent S02 removal. Thus, when compared with a TCA unit operating with
spheres, S02 removal is slightly worse at high gas rates, but the turndown
capacity is significantly better. The better turndown capacity can be attri-
buted to a reduction of the sphere activity, and therefore the gas-slurry
contact, as the gas flow rate is reduced.
14-4
-------
8 --
6 --
O
CM
5 - -
o.
O
cc
O
U)
IT
3
CO
w
111
EC
0.
4 - -
3 --
2 - -
1 --
} r
LIQUOR RATE
A
O
A
O
1,400 gpm
1,200 gpm
1,000 gpm
800 gpm
600 gpm
CEILCOTE PLATE
PACKING:
HEIGHT-46 inches
LAYERS-23
OPENING - 1 3/16 in. x 1 3/16 in.
SLURRY: SOLIDS CONC. - 15 wt. %
TEMPERATURE - 126 - 144° F
4-
18,000 20,000
22,000
24,000
26,000
28,000
30,000
GAS RATE, acfm @ 300 F
1 1 —
8
1 1
9
1 1
10
1 1
11
1 1 1
12
GAS VELOCITY, ft/sec @ 125 F
Figure 14-1.
Pressure Drop Across Four Bar Grids and Twenty-three
Layers of Ceil cote Support Plate Packing
14-5
-------
Outlet participate loading for Run 718-2A ranged from 0.038 to 0.046 and aver-
aged 0.042 grain/dry scf. Particulate removal averaged 98.9 percent, exactly
the same as it did under similar conditions for the TCA system with 5 inches
of spheres per bed (three beds) at an average particulate loading of 0.064
grain/dry scf.
The particulate loading was higher for Run 719-2A, averaging 0.057 grain/dry
scf (range: 0.050 to 0.065). Particulate removal averaged 98.4 percent,
lower than the 99.1 percent previously experienced by the TCA with 5 inches
of spheres per bed for three beds at nominally 20,000 acfm gas rate and con-
tinuous mist eliminator wash, all other conditions being the same (Reference 4),
No plugging occurred during these two runs. The mist eliminator was 0.5 per-
cent restricted at the end of the runs, somewhat less than the 1 percent re-
striction present before these runs. Minor scale formation was observed along
the edges of some of the plates. However, the overall duration of both runs
was not sufficient to determine reliability.
In summary, under similar operating conditions, Ceilcote plates gave better
turndown capacity for 502 removal than sphere beds.
14-6
-------
Section 15
TCA ONE-SCRUBBER-LOOP FORCED-OXIDATION
LIMESTONE TESTS WITH AIR SPARGER
The results of TCA one-scrubber-loop forced-oxidation limestone tests using
an air eductor for air/slurry contact have been reported in Section 12.
Erosion of the air eductor used in those tests was a major problem. Testing
was terminated after about 2050 hours when the neoprene-lined carbon steel
eductor body had eroded through.
In this section, the results of forced-oxidation tests using an air sparger
are reported. The TCA system was operated in a one-scrubber-loop configura-
tion employing either one or two tanks. The air sparger had 40 1/4-inch
diameter holes, similar to the one used in the venturi/spray tower system.
As in the testing with the air eductor, only limestone slurry with high fly
ash loading was used. Forced oxidation within the scrubber loop using lime
slurry would not be expected to yield good S02 removal, as has been mentioned
previously. Seven runs were made from December 9, 1977 through January 24,
1978, using limestone slurry with high fly ash loading. In addition, one
run was made with limestone slurry and added MgO from May 31 through June 8,
1978. Each test lasted about 6 days.
Table 15-1 summarizes the major test conditions and the important test results.
Detailed information concerning these runs can be found in Appendices B through
D, and H through J.
15-1
-------
Table 15-1
SUMMARY OF ONE-SCRUBBER-LOOP FORCED-OXIDATION LIMESTONE TESTS
WITH AIR SPARGER ON THE TCA SYSTEM
Major Test Conditions^2'
Fly ash loading
Flue gas rate, acfm @ 300°F
Slurry flow rate to TCA, gpm
Percent solids reclrculated
Residence times, m1n: Oxidation tank
Em-
Oxidation tank level, ft
A1r flow rate to oxidation tank
Limestone stolchlometrlc ratio (controlled)
TCA Inlet pH (controlled)
Effective Mg++ concentration, ppm
Limestone addition point
Total static height of spheres, Inches
Selected Results
Percent SO- removal
Inlet SO* concentration, ppm
Percent sulflte oxidation
Air stoichiometry, atoms 0/mole S02 absorbed
TCA Inlet pH
Oxidation tank pH
Limestone utilization, %
Gypsum saturation In TCA Inlet liquor, %
Effective Mg++ concentration, ppm
Onstream hours
815- 2A
High
30,000
1000
15
5.2
14.4
18
130
1.3
-
-
EHT
20
89
3000
40
1.0
6.25
-
80
110
-
75
816-2A
High
30,000
1000
15
5.2
14.4
18
180
1.3
-
-
EHT
22.5
91
2850
54
1.40
6.25
5.7
76
100
-
48
817-2A
High
20,000
1000
15
5.2
14.4
18
130
1.3
-
-
EHT
22.5
79
3000
94
1.70
5.8
5.45
81
100
-
161
818-2A
High
25,000
1000
15
5.2
14.4
18
130
1.3
-
-
EHT
22.5
85
3000
67
1.25
6.2
5.65
77
95
-
131
818-2B
High
25,000
1000
15
5.2
14.4
18
0
1.3
-
EHT
22.5 •
82
3300
24
0
6.2
-
81
110
-
140
819-2A
High
20,000
1000
15
4.9
-
17
130
1.3
-
-
Ox1d. Tk
22.5
75
2800
94
1.90
5.55
5.55
77
110
-
164
820- 2A
High
20,000
1000
"15
4.9
-
17
130
-
5.9
-
Oxid. Tk
22.5
79
2500
92
2.0
5.65
5.65
62
115
-
259
821 -2A
High
30,000
1200
15
4.1
-
17
170
1.2
-
5000
Oxid. Tk
15
84
2500
95
1.65
5.35
5.35
79
110
4960
182
ro
Notes:
(1
(2
Used a sparger ring located 8 inches from the bottom of oxidation tank. The sparger ring had 40 1/4-inch diameter holes on the bottom side.
ClaMfier only was used for solids dewaterlng except for Run 821-2A in which clarlfler and centrifuge were used. Centrifuge cake solids
concentration was 79 percent.
-------
15.1 SYSTEM DESCRIPTION
Two operating configurations were used in the one-scrubber-loop forced-oxidation
tests. With one tank, as shown in Figure 15-1, effluent slurry from the TCA
was discharged to the oxidation tank, where limestone (and MgO during Run
821-2A) was added, and the slurry was recycled back to the TCA. With two tanks
in series, as shown in Figure 15-2, TCA effluent slurry was discharged to the
oxidation tank and the slurry then overflowed to a second tank (effluent hold
tank), where limestone was added and slurry was recycled from the second tank
back to the TCA.
Although the one-tank configuration is simpler and modification of the exist-
ing full-scale installations to this forced oxidation scheme can be made
readily (with the addition of an air sparger and a compressor or blower), a
balance must be made in the single tank between the higher pH desired for
good SC«2 removal and the lower pH desired for good sulfite oxidation. The
two-tank configuration has the advantage of oxidizing the lower pH scrubber
effluent slurry before limestone is added. The two-tanks-in-series mode also
provides better limestone utilization.
The oxidation tank arrangement is shown in Figure 15-3. The tank was 7 feet
in diameter and was operated at a 17- to 18-feet level. All tests were con-
ducted with an air sparger ring made of straight 3-inch 316L SS pipe pieces
welded into an octagon of approximately 4-feet diameter. The ring was located
8 inches from the bottom of the tank and had 40 1/4-inch diameter holes pointed
downward. The sparger ring was'fed with compressed air to which sufficient
water was added to assure humidification.
15-3
-------
BLEED TO SOLIDS
. DEWATERING SYSTEM
FLUE GAS
MAKE UP WATER
1
FLUE GAS
LIMESTONE
CLARIFIED LIQUOR „
i
FROM SOLIDS DEWATERING
SYSTEM
+ * *(
k
KttUHUU
T V
•iv >N 1
^
OOO
££2,2.
ooo
00£O
0
o
ooo
oooo
v «•»•. ^
V
|
^ I \
\
g
-^1
CD
c
J
\
1
T
3
TCA
MjO_
WATER
COMPRESSED
AIB
OXIDATION TANK
Figure 15-1. Flow Diagram for One-Scrubber-Loop Forced Oxidation
with Air Sparger in the TCA System Using One Tank
15-4
-------
I I
FLUE GAS
MAKEUP WATER
FLUE GAS
LIMESTONE
CLARIFIED LIQUOR
FROM SOLIDS DEWATERING
SYSTEM
BLEED TO SOLIDS
DEWATERING SYSTEM
ssssssssww
T T
ooooo
oooooo
TCA
COMPRESSED
AIR
WATER^,
OVERFLOW
A A
EFFLUENT HOLD
TANK
OXIDATION
TANK
Figure 15-2. Flow Diagram for One-Scrubber-Loop Forced Oxidation
with Air Sparger in the TCA System Using Two Tanks
15-5
-------
INLET
AGITATOR
BAFFLE
SPARGER
COMPRESSED AIR
OXIDATION TANK
PLAN VIEW
J
OUTLET
BAFFLE ^
SPARGER WITH
40 1/4-inch HOLES
(DOWNWARD DISCHARGE) v
\
012345
»<
\
|
-
c
^=
I^H^V ^—
— 1
J
r=
—
\-
S
/
AGITATOR
/^ (37 rpm, 3 Hp)
-* OXIDATION TANK
* COMPRESSED AIR
'— ' _ INLET
r
SCALE. FEET
ELEVATION VIEW
Figure 15-3. Arrangement of the TCA Oxidation Tank with Air Sparger
15-6
-------
A major shortcoming of this oxidation system was that the agitator was rated
at only 3 Hp and rotated at 37 rpm (compared with 17 brake Hp and 56 rpm for
the venturi oxidation tank). This agitator was similar in configuration to
the agitator in the venturi oxidation tank with two axial flow turbines (49
inches in diameter) pumping downward. Because of the weaker agitation, runs
with similar oxidation tank environment (pH, air stoichiometry, tank level,
percent slurry solids, and limestone utilization) had lower oxidation effi-
ciency in the TCA oxidation tank than in the venturi oxidation tank.
A 20-Hp variable speed agitator is on order and will be used to determine the
relationship between oxidation tank agitation and air requirements.
A second shortcoming was that the existing Shawnee air compressor did not
have sufficient capacity to serve the venturi and the TCA oxidation tanks
simultaneously at full flue gas load. To circumvent this problem, several of
the TCA runs were made at reduced flue gas flow rates. An additional air com-
pressor has been ordered to correct this limitation.
A clarifier was used for solids dewatering in all runs except Run 821-2A.
In that run, a clarifier followed by a centrifuge was used.
15.2 DISCUSSIONS OF TEST RUN RESULTS
15.2.1 Air Stoichiometry
Runs 815-2A through 818-2B were made with two hold tanks in series as shown in
Figure 15-2. The primary effort during these runs was to identify the air
stoichiometric ratio required for nearly complete oxidation. In the first two
15-7
-------
test runs, operated at the maximum achievable flue gas flow rate of 30,000
acfm, it was found that the air compressor did not have a high enough capacity
to supply both the venturi/spray tower system and the TCA system. With an
air rate of 210 scfm to the venturi oxidation tank, only 180 scfm was avail-
able for the TCA system. Further tests were conducted at reduced flue gas
flow rates (20,000 to 25,000 acfm) to allow higher air stoichiometry at the
available air rate. The results of these tests conducted over an oxidation
tank pH range of 5.4 to 5.7 were as follows:
Air Stoichioraetric Ratio, Percent
Run atoms oxygen/mole S02 absorbed Sulfite Oxidation
817-2A
816-2A
818-2A
815-2A
818-2B
1.7
1.4
1.25
1.0
0
94
54
67
40
24
Thus, with two hold tanks in series, an air stoichiometric ratio of about 1.7
was required to achieve greater than 90 percent oxidation. Under similar con-
ditions in the venturi oxidation tank, higher oxidation efficiency (95 to 100)
percent) was achieved. This better performance in the venturi oxidation tank
was attributed to the superior agitation in the venturi oxidation tank.
Runs 819-2A and 820-2A were made with the oxidation tank as the only hold tank
as shown in Figure 15-1. In these runs, the pH in the oxidation tank was higher
as a result of limestone addition. Because of the higher pH, a higher air
stoichiometry was required. This effect can be seen by comparing Runs 817-2A
15-8
-------
(two hold tanks) and 819-2A (one hold tank) made at essentially the same
operating conditions. In both runs, 94 percent sulfite oxidation was achieved.
An air stoichiometric ratio of 1.7 atoms oxygen/mole S02 absorbed was used
in the run with two hold tanks (5.45 oxidation tank pH), and an air stoichio-
metry of 1.9 was required in the run with one hold tank (5.55 oxidation tank
pH). Run 820-2A, at still higher oxidation tank pH of 5.65, achieved only
92 percent oxidation at 2.0 air stoichiometry.
15.2.2 SQg Removal Efficiency
At given limestone stoichiometric ratios, S02 removal efficiency appeared to
be independent of oxidation efficiency, and was primarily a function of flue
gas flow rate, slurry flow rate, and inlet S02 concentration. This relation-
ship agreed closely with previously developed correlations for the TCA system
in limestone service without forced oxidation. At 3000 ppm inlet S02 concen-
tration and a limestone stoichiometric ratio controlled at 1.3 moles Ca/mole
of S02 absorbed, S02 removal efficiency ranged from about 90 percent at 30,000
acfm to about 80 percent at 20,000 acfm. In general, two-tank operation gave
better S02 removal than one tank.
15.2.3 Limestone Utilization
Limestone utilization was higher in the runs using two tanks in series than in
the single-tank runs. Again comparing Runs 817-2A and 819-2A, limestone utili-
zation was 77 percent with one tank and 61 percent with two tanks even though
the inlet pH was lower with one tank. S02 removal efficiency was also improved
from 75 percent with one tank to 79 percent with two tanks. The improvement can
15-9
-------
be attributed to higher residence time (19.6 minutes total with two tanks versus
4.9 minutes with one tank) and the approach to plug flow inherent with tanks in
series.
Because of the relatively poor $62 removal efficiency in Run 819-2A, the next
run (Run 820-2A) was made at a slightly higher pH. The oxidation tank pH was
increased from 5.55 to 5.65. $$2 removal increased only slightly from 75 per-
cent at 2800 ppm inlet S02 concentration to 79 percent at 2500 ppm. However,
the limestone utilization decreased from 77 percent to 62 percent.
15.2.4 MgO Addition
The addition of MgO should not enhance S02 removal in a scrubber loop with
forced oxidation (see Section 6). This was demonstrated in Run 821-2A.
Magnesium ion in the scrubber liquor improves S02 removal by increasing the
sulfite ion, an effective S02 scrubbing component, but forced oxidation
converts the sulfite to sulfate, which is nonreactive.
In Run 821-2A, with 4960 ppm effective magnesium ion concentration and with
forced oxidation, the S02 removal efficiency averaged 84 percent, which was
no higher than expected without MgO addition. In a previous run with MgO
addition the same as in Run 821-2A but without forced oxidation, S0>> removal
averaged 92 percent. Thus, the enhancement on S02 removal with MgO addition
is not achieved in a scrubber loop with forced oxidation.
15.2.5 General Operating Characteristics
The general operating characteristics of the one-scrubber-loop system with
15-10
-------
forced oxidation using an air sparger were the same as those using an air
eductor (see Subsection 12.3) with regard to the water balance and the chlo-
ride levels. In Run 821-2A, the liquor loop was tighter because of the use
of the centrifuge in series with the clarifier to minimize magnesium oxide
consumption.
An air sparger does not, of course, have the severe erosion problems associated
with an air eductor, and it also consumes less energy than an eductor due to
lower pressure drop. This latter advantage of air sparger on energy consump-
tion may also hold true when considering the sparger/agitator and eductor/
agitator combination, even though the eductor discharge plume provides signi-
ficant mixing in the oxidation tank.
15.3 SUMMARY OF FINDINGS
Despite oxidation tank agitator and air compressor limitations, forced oxida-
tion with an air sparger in a one-scrubber-loop TCA system was demonstrated.
The following is a summary of findings.
t With two tanks in series, 94 percent sulfite oxidation was achieved
at an air stoichiometry of 1.7, oxidation tank pH of 5.45, and an
oxidation tank level of 18 feet. Under a similar oxidation tank
environment, higher oxidation efficiency (95 percent or over) was
achieved in the venturi oxidation tank. The poorer performance on
the TCA system was attributed to the weaker agitation in the TCA
oxidation tank.
• Because of higher pH in a one-tank configuration, higher air stoichio-
metry (about 1.9) was required to achieve a similar degree of oxida-
tion (94 percent).
• Operation in the two-tank mode gave better limestone utilization and
S02 removal than operation in the one-tank mode.
• As expected, MgO addition did not enhance the SO? removal when oxida-
tion was forced within the scrubber loop (Run 82T-2A).
15-11
-------
Further tests will be conducted in the future when a higher speed agitator
for the oxidation tank and a new air compressor are installed on the TCA
system.
15-12
-------
Section 16
TCA LIMESTONE/MgO TEST RESULTS
Because of insufficient compressed air capacity for forced oxidation with an
air sparger in the TCA system (see Section 15), further testing with forced
oxidation of the TCA system was deferred until a new air compressor could be
installed.
Beginning in late January 1978 through the end of the report period, the TCA
system was operated without forced oxidation using lime or limestone slurry
with added MgO. The primary purpose of these tests was to resolve some of
the inconsistent results obtained in the earlier lime/MgO tests (Runs 601-2A
through 617-2A) and limestone/MgO tests (Runs 583-2A through 589-2A), in
which air leakage into the scrubber downcomer was suspected during some runs,
resulting in higher than expected sulfite oxidation and gypsum scaling (see
Sections 10 and 11, Reference 4).
In those earlier tests, made from April through November 1976, a 7-ft diameter
by 21-ft high effluent hold tank (D-204) was used in a majority of runs. The
scrubber downcomer inside the effluent hold tank (the internal downcomer) dis-
charged the scrubber effluent slurry at an elevation of 14 ft above the tank
bottom. Since the scrubber operated below atmospheric pressure, those runs
having a tank level less than 14 ft (to maintain low residence times) experi-
enced some air leakage via the downcomer. The air leakage was indicated by
16-1
-------
the fact that sulflte oxidation was 10 to 15 percentage points higher than
normal for those runs, resulting in high gypsum saturation and scaling prob-
lems.
To avoid the air leakage problem with the internal downcomer at tank levels
below 14 ft, an external downcomer was installed in August 1977. The external
downcomer branches off the existing internal downcomer and enters the efflu-
ent hold tank (D-204) radially near the tank bottom, thus preventing air
leakage. Blank plates are provided to allow either downcomer to be used.
16.1 DISCUSSIONS OF TEST RUN RESULTS
The results of limestone tests with MgO addition are presented in this section.
Lime tests with MgO addition are reported in Section 17.
A total of six limestone runs were made from late January through early May
1978, including two base case runs without MgO and four runs with MgO addition.
Flue gas with high fly ash loading was used in all runs. Table 16-1 presents
the major test conditions and important test results. Additional detailed
information can be found in Appendices B through D, and H through J.
Because of the shortage of coal caused by the coal miners' strike in the first
quarter of 1978, coals from different sources were burned in Boiler No. 10.
As a result, inlet S0£ concentration fluctuated as much as tenfold (320 to 3400
ppm), and the run-average concentrations were considerably lower than normal
in most of the runs. These wide variations in inlet S0£ concentration not only
created process control problems, but also caused some difficulties in the in-
terpretation of run data and in the run-to-run comparison of test results.
16-2
-------
Table 16-1
SUMMARY OF LIMESTONE TESTS WITH MgO ADDITION
ON THE TCA SYSTEM
CO
Major Test Conditions^
MgO addition
Fly. ash loading
Gas rate, acfm @ 300°F
t Slurry flow rate to TCA, gpm
Percent solids recirculated
EHT residence time, mln' '
EHT level, ft^
( 2}
Downcomer to EHTV '
Limestone stoichiometric ratio (controlled)
Limestone addition point* '
Effective Mg concentration, ppm
Selected Results
Percent SO- removal
Inlet S02 concentration, ppm
S02 make-per-pass, m-moles/ liter
TCA inlet liquor sulfite concentration, pprtr '
TCA inlet liquor gypsum saturation, %
Percent sulfite oxidation
Percent limestone utilization
TCA inlet pH
Effective Mg concentration, ppm
Onstream hours
590- 2A
No
High
30,000
1200
15
4.1
17
Ext.
1.2
pc
0
64
2100
7.3
770
110
45<5>
88
5.0
0
45
590-2B
No
High
30,000
1200
15
4.1
17
Int.
1.2
DC
0
77
2450
9.9
270
-120
31
85
5.3
0
92
591-2A
Yes
High
30,000
1200
15
4.1
17
Ext.
1.2
EHT
5000
92
2300
11.3
2000
105
43
80
5.5
4930
207
592-2A
Yes
High
30,000
1200
15
4.1
17
Ext.
1.2
EHT
9000
95
2850
13.8
4360
105
25
74
5.4
8870
293
593-2A
Yes
High
30 ,000
900
15
4.1
12.8
Ext.
1.2
EHT
9000
93
2850
18.8
4855
85
24
63
5.45
9000
260
594-2A
Yes
High
30,000
1200
15
4.1
17
Int.
1.2
EHT
9000
93
2850
13.8
5370
90
18
81
5.35
8950
99
Notes:
(1)
(2)
(3)
(4)
(5)
All runs were made with 3 beds (4 grids) and with nominally 5 inches/bed of 1 5/8 inch diameter nitrile foam spheres.
Clarifier and centrifuge were used for solids dewatering in all runs.
The effluent hold tank (D-204) was 7 ft in diameter and 21 ft high. Tests were made with either an internal or external
downcomer.
EHT or downcomer (DC).
Total sulfite-includes S03= and HSOs~.
Steady-state sulfite oxidation was not reached during Run 590-2A.
-------
16.1.1 Base Case Runs 590-2A and 590-2B
Runs 590-2A and 590-2B were base case limestone runs without MgO addition.
The operating conditions for these runs were the same except for the difference
in the downcomer used. Run 590-2A with the external downcomer was terminated
after only 45 hours of operation because of severe cold weather and freezing
problems. Because of the short duration of the run, steady-state natural
sulfite oxidation level was not reached from the nearly complete oxidation
level of a previous forced-oxidation test (Run 820-2A). Run 590-2B was a
continuation of Run 590-2A, except that the internal downcomer had to be used
because the external downcomer was frozen.
The results of Run 590-2B are compared below with those of a similar test,
Run 583-2A, which was a base case limestone run for the 1imestone/MgO tests
made in 1976 (Reference 4).
590-2B 583-2A
EHT residence time, min 4.1 3.0
EHT level, ft 17 12.5
Downcomer to EHT Int. Int.
Air leakage No Yes
Percent S02 removal 77 77
Inlet S02 concentration, ppm 2450 3050
S02 make-per-pass, m-moles/liter 9.9 12.5
Percent sulfite oxidation 31 34
TCA inlet liquor gypsum sat'n, % 120 145
Other operating conditions for the two runs were the same. Sulfite oxidation
in Run 583-2A was 34 percent, which was higher than normal because of the air
leakage. However, oxidation for Run 590-2B was also relatively high at 31
percent, probably due to the lower inlet S0£ concentration (higher 02/S02
ratio in the inlet flue gas). The higher TCA inlet liquor gypsum saturation
16-4
-------
of 145 percent for Run 583-2A was consistent with the higher sulfite oxidation.
16.1.2 Effect of MgO Addition
In Run 591-2A, magnesium oxide was added to maintain an effective Mg++ concen-
tration* of 5000 ppm. S02 removal increased from 77 percent at 2450 ppm without
MgO addition for Run 590-2B to 92 percent at 2300 ppm inlet S02 concentration.
At 9000 ppm effective Mg++ concentration (Runs 592-2A and 594-2A), average S02
removal was 93 to 95 percent at a higher inlet S02 concentration of 2850 ppm.
Run 591-2A is compared below with a similar test, Run 583-2B, made in 1976
(Reference 4).
591-2A 583-2B
EHT residence time, min 4.1 3.0
EHT level, ft 17 12.5
Downcomer to EHT Ext. Int.
Air leakage No Yes
Percent S02 removal 92 84
Inlet S02 concentration, ppm 2300 2900
S02 make-per-pass, m-moles/liter 11.3 13.5
Percent sulfite oxidation 43 30
TCA inlet liquor gypsum, sat'n, % 105 110
Other operating conditions for the two runs were the same. Sulfite oxidation
of 30 percent for Run 583-2B was relatively high because of air leakage. How-
ever, oxidation was even higher for Run 591-2A at 43 percent without air leak-
age into the downcomer. This may be partially due to the low inlet S02 concen-
tration of 2300 ppm (high 02/S02 ratio in the inlet flue gas). The scrubber
* Effective magnesium ion concentration is defined as the total magnesium ion
minus that magnesium ion concentration equivalent to total chlorides. Mag-
nesium chloride has no effect on S02 removal.
16-5
-------
inlet liquor gypsum saturation was about the same in both runs. However,
gypsum scaling was observed in Run 583-2B but not Run 591-2A.
Runs 592-2A and 594-2A were made under the same operating conditions, except
for the difference in the downcomer used. These runs are compared below with
Run 584-2A, made in 1976 (Reference 4).
592-2A 594-2A 584-2A
EHT residence time, min 4.1 4.1 4.1
EHT level, ft 17 17 17
Downcomer to EHT Ext. Int. Int.
Air leakage No No No
Percent SOo removal 95 93 94
Inlet S02 concentration, ppm 2850 2850 2950
S0£ make-per-pass, m-moles/liter 13.8 13.8 15.0
Percent sulfite oxidation 25 18 20
TCA inlet liquor gypsum sat'n, % 105 90 50
Operating conditions for Runs 594-2A and 584-2A were essentially the same.
Inlet S02 concentrations and S02 removals were about the same for all three
runs. Gypsum saturation during Run 584-2A was only 50 percent, compared
with about 100 percent for Runs 592-2A and 594-2A. All three runs were free
of gypsum scaling.
In general, MgO addition improved the S02 removal efficiency, but no corre-
lation appeared to exist between gypsum saturation and S02 make-per-pass.
Lower sulfite oxidation tended to give lower gypsum saturation.
16.1.3 Effect of Slurry Flow Rate
Run 593-2A was made at a reduced slurry flow rate of 900 gpm. The slurry
level in the hold tank was lowered from 17 to 12.8 ft to maintain the same
16-6
-------
residence time at 4.1 minutes. S02 removal was not significantly affected
by the reduced liquid-to-gas ratio and remained high at 93 percent with
2850 ppm inlet S02 concentration. As a result, S02 make-per-pass increased
to 18.8 m-moles/liter.
Run 593-2A is compared below with Run 585-2A, made in 1976.
593-2A 585-2A
EHT residence time, min 4.1 4.1
EHT level, ft 12.8 12.8
Downcomer to EHT Ext. Int.
Air leakage No Yes
Percent S02 removal 93 85
Inlet S02 concentration, ppm 2850 2900
S02 make-per-pass, m-moles/liter 18.8 17.5
Percent sulfite oxidation 24 28
TCA inlet liquor gypsum sat'n, % 85 105
TCA inlet liquor sulfite cone., ppm 4855 4190
The conditions for the two runs were essentially the same, except for the
difference in the downcomer used. Air leakage into the downcomer occurred in
Run 585-2A, giving higher sulfite oxidation and gypsum saturation. The higher
S02 removal for Run 593-2A may be attributed to the lower oxidation and higher
sulfite concentration. Slight gypsum scaling occurred during both runs. It
appeared that the gypsum scaling potential is higher at higher S02 make-per-
pass.
16.2 SUMMARY OF FINDINGS
Six test runs were made on the TCA system with limestone slurry and with flue
gas containing high fly ash loading. Of the six runs, two were made without
MgO addition, one with 5000 ppm effective Mg++ concentration, and the other
16-7
-------
three with 9000 ppm effective Mg++ concentration. These runs were made to
resolve some inconsistent results from earlier limestone/MgO tests (made in
1976) in which air leakage into the scrubber downcomer was suspected during
some runs, causing higher than normal sulfite oxidation and gypsum scaling.
Unfortunately, the inconsistent results obtained in 1976 were not resolved
because of the wide variations in the inlet S02 concentration mentioned
earlier. The following is a summary of findings.
0 Higher effective Mg++ concentration is required in a lime-
stone system than in a lime system to obtain a similar degree
of improvement in SO? removal efficiency. This is probably
due to the lower pH (5 to 6) inherent in the limestone system,
where the sulfite-bisulfite equilibrium favors a shift toward
bisulfite, which is not an S02 scrubbing species.
0 Under typical operating conditions, S02 removal improved by
about 15 percentage points with 9000 ppm effective Mg++ concen-
tration.
a Adding MgO at the levels tested did not always result in gypsum
subsaturated operation. Lower sulfite oxidation tended to give
lower gypsum saturation.
0 No correlation appeared to exist between gypsum saturation and
S02 make-per-pass. However, gypsum scaling potential tended to
be higher (incipient scaling gypsum saturation level tended to
be lower) at higher S02 make-per-pass.
16-8
-------
Section 17
TCA LIME/MgO TEST RESULTS
Lime testing with MgO addition was conducted from early February through
late June 1978. These tests were made in the same period as the limestone
testing with magnesium oxide addition reported in Section 16.
As with the 1imestone/MgO tests, the main purpose of the lime/MgO tests was
to resolve the inconsistent results among the earlier tests (Runs 601-2A
through 617-2A) made in 1976. In some of those earlier lime/MgO tests, air
leakage into the scrubber downcomer caused higher than normal sulfite oxida-
tion, and gypsum scaling was observed (see Sections 10 and 11, Reference 4).
The reason for air leakage in those earlier tests and the remedy to prevent
the air leakage for the runs reported in this section have been presented in
Section 16. The corrective action involved the installation of a scrubber
downcomer external to the effluent hold tank near the tank bottom to ensure
proper vacuum seal.
17.1 DISCUSSIONS OF TEST RUN RESULTS
Nine lime runs were made from early February through late June 1978. These
were two base case lime runs without MgO addition and seven lime runs with
17-1
-------
2000 ppm effective Mg++ concentration.* Flue gas having high fly ash load-
ings was used in each run. Major test conditions and important test results
are summarized in Table 17-1. Additional detailed information for these
runs are given in Appendices B through D, and H through J.
As has been mentioned in Section 16, coals from different sources were burned
in Boiler No. 10 during the coal miners' strike in the first quarter of 1978.
Inlet S02 concentration fluctuated as much as tenfold (320 to 3400 ppm), and
the run-average concentrations were significantly lower than normal during
most of the runs. The wide variations in inlet S02 concentration created pro-
cess control problems and caused difficulties in the data interpretation and
in the run-to-run comparison of test results.
17.1.1 Base Case Runs 618-2A and 618-2B
Runs 618-2A and 618-2B were base case runs without MgO addition. These runs
were compared with the subsequent lime/MgO runs to observe the effect of
magnesium ion. Run 618-2A was made with the internal downcomer. The scrubber
effluent was later routed to the external downcomer (Run 618-2B) when the
frozen slurry in the external downcomer was thawed (see Subsection 16.1.1).
Sulfite oxidations and percent S02 removals for these runs were relatively
high because of the low inlet S02 concentrations.
* Effective magnesium ion concentration is defined as the total magnesium ion
minus that magnesium ion concentration equivalent to total chlorides. Mag-
nesium chloride has no effect on S02 removal.
17-2
-------
Table 17-1
SUMMARY OF LIME TESTS WITH MgO ADDITION
ON THE TCA SYSTEM
Major Test Conditions' '
MgO addition
Fly ash loading
Gas rate, acfm 9 300°F
Slurry flow rate to TCA, gpm
Percent solids recirculated
(2)
EHT residence time, min '
EHT level, ft'2)
Downcomer to EHT' '
TCA inlet pH (controlled)
Lime addition polnt^
Effective Mg++ concentration, ppm
Selected Results
Percent SO, removal
L
Inlet SO, concentration, ppm
i
SO, itsake-per-pass, m-moles/Hter
(4)
TCA inlet liquor sulfite concentration, ppnr
TCA inlet liquor gypsum saturation, %
Percent sulfite oxidation
Percent lime utilization
Effective Mg concentration, ppm
Onstream hours
618-2A
No
High
30,000
1200
8
4.1
17
Int.
7.0
DC
0
90
2100
10.2
85
120
28
93
0
162
618-2B
No
High
30,000
1200
8
4.1
17
Ext.
7.0
DC
0
96
1900
8.9
80
120
30
94
0
88
619-2A
Yes
High
30,000
1200
8
4.1
17
Ext.
7.0
DC
2000
94
1850-
9.2
355
90
23
98
2360
306
620-2A
Yes
High
30,000
900
8
4.1
12.8
Ext.
7.0
DC
2000
91
1600
10.3
330
95
27
97
2090
106
621-2A
Yes
High
30,000
1200
8
3.0
12.5
Ext.
7.0
DC
2000
96
1950
10.3
310
105
29
98
2135
239
622- 2A
Yes
High
30,000
900
8
4.1
12.8
Ext.
7.0
EHT
2000
92
2600
16.4
/C\
940(5)
95
22
99
/ C \
4510(5)
91
622-2B
Yes
High
30,000
900
8
4.1
12.8
Ext.
7.0
DC
2000
85
2600
15.0
465
95
14
96
2150
125
623-2A
Yes
High
30,000
1200
8
3.0
12.5
Ext.
7.0
DC
2000
91
2550
12.3
410
70
14
94
2190
271
624- 2A
Yes
High
30,000
900
8
3.0
9.4
Ext.
7.0
DC
2000
82
2900
16.4
480
65
15
99
1950
269
I
CO
Notes:
(1)
(2)
(3)
(4)
(5)
All runs were made with 3 beds (4 grids) and with nominally 5 inches/bed of 1-5/8 inch diameter nitrile foam spheres.
were used for solids dewatering in all runs except Run 624-2A 1n which clarifier only was used.
The effluent hold tank (D-204) was 7 ft in diameter and 21 ft high. Tests were made with either an internal or external downcomer.
EHT or downcomer(DC).
Total sulfite includes SO-f and HSOs".
Steady-state concentrations were not reached for Run 622-2A from a previous high magnesium limestone run.
Clarifier and centrifuge
-------
17.1.2 Effect of MgO Addition
In Run 619-2A, MgO was added to maintain 2000 ppm effective Mg++ concentra-
tion in the slurry liquor. Other test conditions for the run were the same
as Run 618-2B. Due to the low inlet S02 concentrations of about 1900 ppm
for both Runs 618-2B and 619-2A, the expected improvement in S0£ removal
with 2000 ppm effective Mg++ concentration was not discernible. The S02
removal of 96 percent for Run 618-2B was already high at the low inlet S02
concentration without MgO addition. Run 619-2A showed only a slight increase
in S02 make-per-pass from 8.9 to-9.2 m-moles/liter. A possible explanation
might be that the gas phase mass transfer resistance became significant in
both runs at the low inlet S02 concentration, and that the increased mass
transfer and reaction rates in the liquid phase due to magnesium ion did not
significantly improve the overall rate.
Run 619-2A is compared below with Run 601-2A, which was made in 1976 (Refer-
ence 4) under identical operating conditions except that the latter had the
internal downcomer:
619-2A 601-2A
Downcomer to EHT Ext. Int.
Air leakage No No
Percent S02 removal 94 92
Inlet S02 concentration, ppm 1850 2900
S02 make-per-pass, m-moles/liter 9.2 13.5
Percent sulfite oxidation 23 14
TCA inlet liquor gypsum sat'n, % 90 50
TCA inlet liquor sulfite cone., ppm 355 525
Sulfite oxidation and gypsum saturation for Run 601-2A were much lower than for
Run 619-2A. The rate of S02 removal was much higher for Run 601-2A, as evidenced
by the higher S02 make-per-pass and liquor sulfite concentration.
17-4
-------
17.1.3 Effect of Downcomer Air Leakage
Runs 620-2A and 621-2A were conducted to compare with Runs 604-2A and 613-2A
respectively, which were made in 1976 (Reference 4). Air leakage into the
internal downcomer was suspected in Runs 604-2A and 613-2A of causing high
sulfite oxidation and gypsum scaling. The results of these four runs are
compared in pairs, as follows:
620-2A 604-2A 621-2A 613-2A
Slurry Flow Rate, gpm 900 900 1200 1200
EHT residence time, min 4.1 4.1 3.0 3.0
EHT level, ft 12.8 12.8 12.5 12.5
Downcomer to EHT Ext. Int. Ext. Int.
Air leakage No Yes No Yes
Percent S02 removal 91 73 96 85
Inlet S02 concentration, ppm 1600 3200 1950 3050
S02 make-per-pass, m-moles/liter 10.3 17.5 10.3 14.0
Percent sulfite oxidation 27 28 29 35
TCA inlet liquor gypsum sat'n, % 95 90 105 95
Because of air leakage into the internal downcomer, sulfite oxidation in Runs
604-2A and 613-2A was higher than normal (28 and 35 percent, respectively).
Unfortunately, oxidation for Runs 620-2A and 621-2A was also high (27 and 29
percent, respectively), because of the low inlet S02 concentrations (high
02/S02 ratio in inlet flue gas). Gypsum saturations in all four runs were
similar (about 95 percent). However, gypsum scaling occurred in Runs 604-2A
and 613-2A but not in Runs 620-2A a-nd 621-2A. Therefore, as with limestone/
MgO tests, higher S02 make-per-pass increases the gypsum scaling potential.
17.1.4 Effect of Inlet SO? Concentration
Runs 622-2A, 622-2B, and 623-2A were conducted to replicate Runs 620-2A and
17-5
-------
621-2A operating conditions, but at a normal inlet SC^ concentration of about
2600 ppm.
Run 622-2A was intended to replicate Run 620-2A, but the fresh lime slurry
was inadvertently added to the effluent hold tank rather than the external
downcomer. In addition, magnesium concentration did not reach a steady-state
level during Run 622-2A from a previous high-magnesium limestone run. There-
fore, a comparison between those two runs would be irrelevant.
As can be seen in Table 17-1, direct comparisons are possible between Runs
620-2A and 622-2B and between Runs 621-2A and 623-2A to study the effect of
inlet S02 concentration. Average sulfite oxidation was about 28 percent at
1600 to 1950 ppm average inlet $©2 concentrations; sulfite oxidation was only
one-half as great (14 percent) when the inlet S02 increased to about 2600 ppm.
Gypsum saturation was lower (70 to 95 percent) at the lower sulfite oxidation,
compared with about 100 percent saturation at the higher oxidation. S02 re-
moval was about 5 percentage points lower at the higher inlet S02 concentra-
tion, as expected. However, S02 make-per-pass was 20 to 50 percent higher
at the higher inlet S02 concentration.
17.1.5 Effects of Hold Tank Residence. Time and Slurry Flow Rate
Run 624-2A was originally designed to be compared with Run 620-2A for the
purpose of observing the effect of the hold tank residence time on sulfite
oxidation and gypsum saturation. However, because of the large difference
in the inlet S02 concentrations, direct comparison of these two runs would
not be meaningful.
17-6
-------
Referring to Table 17-1, Run 624-2A can be compared with Run 622-2B for the
effect of hold tank residence time, and with Run 623-2A for the effect of
slurry flow rate (liquid-to-gas ratio). These three runs had much closer
inlet S02 concentrations. Sulfite oxidations for all three runs were about
the same (14 to 15 percent).
Dropping the residence time from 4.1 minutes (Run 622-2B) to 3.0 minutes
(Run 624-2A) reduced the gypsum saturation from 95 to 65 percent. Reducing
the slurry flow rate from 1200 gpm (Run 623-2A) to 900 gpm (Run 624-2A) while
keeping the residence time unchanged did not significantly change the gypsum
saturation. Percent S02 removal decreased, as expected, from 91 percent at
an inlet S02 concentration of 2550 ppm to 82 percent at an inlet concentration
2900 ppm. S02 make-per-pass increased from 12.3 to 16.4 m-moles/liter because
of the reduced liquid-to-gas ratio from 50 to 37 gal/Mcf.
17.2 SUMMARY OF FINDINGS
Nine runs were made on the TCA system with lime slurry and with flue gas having
high fly ash loading. Of the nine runs, two were made without MgO addition
as base cases, and seven were made with 2000 ppm effective Mg++ concentration.
As with the 1imestone/MgO tests (Section 16), the lime/MgO tests were conducted
to resolve inconsistent results among earlier tests made in 1976. In those
earlier tests, air leakage into the scrubber downcomer during some runs caused
high sulfite oxidation and gypsum scaling. Unfortunately, in most of the
runs reported in this section, high sulfite oxidation was also encountered.
This high oxidation was due to lower than normal inlet S02 concentrations, and
the inconsistent results obtained in 1976 were not resolved. However, because
17-7
-------
of low S02 make-per-pass, gypsum scaling did not occur. The following
findings reflect results of the recent tests:
• At low inlet S02 concentrations (about 1600 to 2000 ppm), improve-
ment in S02 removal was not discernible with 2000 ppm effective
Mg concentration. Overall reaction rate may be controlled by
the gas phase mass transfer at the low inlet S02 concentrations.
• The lower inlet S02 concentrations (about 1600 to 2000 ppm) re-
sulted in sulfite oxidations 10 to 15 percentage points higher
than normal, probably because of the higher 02/S02 ratio in the
inlet flue gas. Higher oxidation would also render magnesium ion
ineffective in S02 removal enhancement.
• As in the 1imestone/MgO tests, lower sulfite oxidation tended to
yield lower gypsum saturation. No correlation appeared to exist
between gypsum saturation and S02 make-per-pass. At similar gyp-
sum saturation levels, gypsum scaling potential appeared to be
higher at higher S02 make-per-pass. (Incipient scaling percent
gypsum saturation tended to be lower at higher S02 make-per-pass.)
• Within the range of the hold tank residence time tested (3.0 to
4.1 minutes), lower residence time resulted in lower gypsum satura-
tion at similar sulfite oxidation level (Run 622-2B versus 624-2A).
17-8
-------
Section 18
FLUE GAS CHARACTERIZATION
18.1 SPECIAL TEST RUNS
An intensive testing period to explore the effect of major operating variables
on each system has been completed. Tests completed on the TCA and one test
on the venturi/spray tower are reported in this section. Previous venturi/
spray tower tests were reported in the Third Progress Report (Reference 4).
For a detailed discussion of the tests, the reader is referred to Reference 9.
18.1.1 Test Program
The test series on both the venturi/spray tower and TCA systems was designed
to determine the effect of major operating variables on particulate mass re-
moval and size distribution.
Variables investigated in the venturi/spray tower system were flue gas rate,
slurry rate, MgO addition, venturi pressure drop, mist eliminator configuration,
and percent solids recirculated. Variables investigated in the TCA system were
flue gas rate, slurry rate, MgO addition, and mist elimination wash scheme.
The mist eliminators on both systems were washed intermittently on top with
fresh water (8-hour sequential cycle with one of the six nozzles activated for
4 minutes every 80 minutes). The following three modes of bottom wash were
18-1
-------
evaluated:
• High-frequency intermittent fresh water (HI), 4 min/hr at
1.5 gpm/ft , shut off during testing
• Low-frequency intermittent fresh water (LI), 6 min/4 hr at
1.5 gpm/ft , shut off during testing
t Continuous diluted clarified liquor (C), at 0.4 gpm/ft2, left
on during testing
For one run on each system, flue gas to the scrubber was obtained downstream
of the electrostatic precipitator (ESP) to evaluate scrubber performance
using flue gas with low fly ash loading. Flue gas for all the other runs
came from a takeoff just downstream of the boiler air preheaters. Listed in
Table 18-1 are the run conditions for the TCA system tests, and in Table
18-2, the conditions for the venturi/spray tower system tests.
18.1.2 Sampling Locations
Samples were taken from both the inlet and outlet of each scrubber. The duct
at these locations was forty inches in diameter. Two eight-point traverses,
90° apart, were performed for each mass loading and size distribution test.
The inlet sample ports were located upstream of the scrubber inlet where the
gas temperature was approximately 300 to 330°F. The outlet sample port was
located after the reheater at the scrubber outlet where the gas temperature
was approximately 250°F. The sample ports were about two duct diameters up-
stream and six duct diameters downstream of the nearest obstruction.
18.1.3 Test Methods
The procedure used for measuring particulate mass loading was a modification
of EPA Method Five. A Hi-Volume sampling train manufactured by the Aerotherm
18-2
-------
Table 18-1
FLUE GAS CHARACTERIZATION PROGRAM RESULTS
FOR THE TCA SYSTEM
Controlled Variables
Gas Rate, acfm @ 300°F
Scrubber Gas Velocity, ft/sec
Slurry Rate, gpm
Percent Solids Recirculated
Fly Ash Loading
Alkali
Mist Eliminator Configuration
Mist Eliminator Wash Scheme (D
MgO Addition
Uncontrolled Variables
Gas Velocity at Mist. Elim.. ft/sec
TCA Beds AP Range, in. HO '
Mist Eliminator AP Range, in. H-O
Total Dissolved Solids, ppm
Flue Gas Measurements
Mass Loading, Inlet
Average, gr/dscf
Range, gr/dscf
Mass Loading, Outlet
Average, gr/dscf
Range, gr/dscf
Mass Loading Removal
Average, %
Range, %
Test Philosophy
Comments
TFG-2A
30,000
12. 5
1200
15
High
Limestone
chevron
HI
Yes
8.2
5. 6 - 7.4
0.47 - 0. 50
50,000 - 55,000
4.67
3. 55 - 5.61
0.055
0.034 - 0. 092
98.8
98.3 - 99.4
MgO addition.
Noticeable
scale particles
were found on
the outlet mass
loading filter.
TFG-2B
30,000
12.5
1200
15
High
Limestone
chevron
HI
No
8.2
7.0-9.0
0.45 - 0.60
7,300 - 9.500
6.07
5, 14 - 7.59
0.048
0. 037 - 0.058
99.2
98.9 - 99.4
Basic operat-
ing conditions.
HI wash
scheme.
TFG-2C
30,000
12.5
1200
15
High
Limestone
chevron
C
No
8.2
7.0 - 8. 5
0.48 - 0. 60
5,000 - 10,000
5.82
4.77 - 7. 51
0.064
0.058 - 0. 080
98.9
98.4 - 99.2
Basic operat-
ing conditions.
C wash scheme.
TFG-2D
20,000
8.4
1200
15
High
Limestone
chevron
C
No
5.5
4. 4 - 5.1
0. 18 - 0.20
5, 700 - 7, 400
6. 13
4.23 - 6.86
0.056
0.051 - 0.065
99. 1
98.7 - 99.2
Low gas rate.
TFG-2E
30.000
12.5
600
15
High
Limestone
chevron
C
No
8.2
4. 4 - 5.0
0.45 - 0. 55
4,600 - 6, 300
5.69
4. 39 - 6.61
0.065
0.054 - 0.088
98.8
98.4 - 99. 1
Low liquor
rate.
TFG-2F
30,000
12.5
1200
15
High
Limestone
chevron
no wash
No
8.2
5.6-7.2
0.44 - 0. 50
6,400 - 8,700
4. 53
3.67 - 5.43
0.042
0.035 - 0.046
99. 1
99.0 - 99.2
Basic operat-
ing conditions.
No wash.
706-2A
30, 000
12.5
1200
15
Low
Limestone
chevron
HI
No
8.2
8.0 - 10.0
0.40 - 0.60
7,000 - 9,000
0. 124
0. 065 - 0.225
0.026
0. 018 - 0.032
74.0
53.8 - 85.8
Low inlet fly
ash loading.
00
CO
(1) Wash Schemes: HI. high frequency fresh water bottom wash with the bottom wash off
during testing; C, continuous diluted clarified liquor bottom wash left on during testing.
"With both schemes a low frequency fresh water top wash was used.
-------
Table 18-2
FLUE GAS CHARACTERIZATION PROGRAM RESULTS
FOR THE VENTURI/SPRAY TOWER SYSTEM
Controlled Variables
Gaa Rate, acfm @ 330°F
ST Gas Velocity, ft/Bee
ST Slurry Rift, gpm
Venlurl 6 P, In. H.O
Percent Solids Recfrcutated
Fly Ash Loading
A Ik. II
Mist Ellm. Wash Scheme
MgO Addition
C.i Vel. it Mid Ellm. , ft/«ec
System aPRangt, In. H.O
Total Dissolved Solida, ppm
Flue Ge.8 Measurements
Man Loading Inlet
Maia Loading Outlet
RanRe, grains/dry sef
Masa Loading Removal
Average,*
Range. %
SO Inlet
Average, ppm
Range, ppm
SO. Outlet
Average, ppm
Range, ppm
SO- Removal
Average. %
Range, %
Ten Phltoaophy
Comments
•
VFO-1A
35.000
9.4
1.400
9
8
High
Ltm«
1.1
Yea
9.4
13.3 - 13.8
13,200 - 18,000
4. 41
1.66 - 5.50
0. 023 - 0. 046
99. i
99. 0 - 99. 4
10.9
3.4 - 14.8
2.9
74.9
54.9 - 99.2
MgO add'n.
VFC-1B
35, 000
9.4
1,400
600
9
8
Low
Lime
chevron
LI
No
9.4
13.1 - 11. 9
2,500 . 5.200
0.070
0.039 - 0.096
0.003 . 0.007
92.0
82. 1 - 96. 1
2.6
1.7 . 3.7
0.9
67.3
47.1 . 81.5
35,000
9.4
1,400
9
8
Low
Llm«
LI
No
9.4
13.6 - 14.0
7,000 • 9.500
0.367
0.118 - 0.609
0. 005
0. OOJ - 0. 009
98.3
99.5 - 99.3
5.4
3.6 - 7.2
2.7
1.6- 4. 1
50.9
39.1 - 64.3
Low Inlet l\ a.h oadln
TDS low due lo
non-equlllbra-
tlon of clarffler.
TDS raised by
CaCI, add'n.
Boiler upa«t
condltlona
caused higher
Inlet grain toad*
Ing.
VFO-1C
3?, 000
9.4
1,400
9
8
High
Lime
LI
No
9.4
13,7 - 14.3
7, 700 • 9, 600
6. J7
4.82 - 8.19
0. 019
0. 013 . 0.023
99.7
99. 5 - 99. 8
8.9
3.0 - 17. 1
3.6
0, 3 - 8. 3
62.2
27. 2 - 97. 1
condltlona.
VFO-1D
20, 000
5.4
1,400
9
8
High
Lime
LI
No
5,4
10.2 - 10.8
6, 000 . S, BOO
4.5)
3.68 . 5.78
0. 026
0.018 - 0.019
99.4
99. 1-99.6
9.3
4.5 . 11.5
4.1
0. 4 - 7, 6
51.9
37. 5 - 65. 5
8
VFG-1E
35.000
9.4
1,400
175
4.5-6.0
8
Hl,h
Lime
LI
No
9.4
9.5 - 10.5
5,400 • 8,000
5.11
3. 77 - 7. 90
0. 028
0.020 - 0.037
99.4
99. 1 - 99. 7
15.4
7.1 - 24.8
8.9
.2 - 13. 9
__
_.
turl.
Flo m«te
cauaed high flow
to vtjiturt. De-
aired 140 gpm.
SOj values not
almultaneous
due to equip
probs, no re-
movals calcu-
lated.
VFG-1F
35.000
9.4
0
9
8
Hl(h
Lime
LI
No
9.4
12.5 . 13.2
7.200 - 10,000
5.14
1.72 - 6.41
0. 027
0.022 - 0.037
99.5
99. 4 - 99. 6
2.1
0.4-4.5
0.3
. - .
11.0
25. 0 - 100.0
r
VFG-1C
15. 000
9.4
1.400
9
8
High
Lime
LI
No
9.4
14. 1 - 15.5
8,300 • 9,400
6.05
1.90 - 8.76
0. 021
0.018 - 0.024
99.6
99.4 - 99.8
8.9
2.5 - 15.8
4. 1
2 8 - 5. 1
66.9
66.7 - 67.1
condltlona. York
detnlater added.
Demlster
plugged after
55 hr». of oper.
atlon.
VFG-II
35. 000
9.4
1,400
9
IS
High
Lime
LI
No
9.4
13.7 . 14,3
6,600 . 9,000
5.18
1. 70 - 6. 06
0. 026
0.017 - 0.037
99.5
99.4 - 99.7
10.7
5.8 - 21.2
3.5
0. 5 - 10. 9
72.5
46.8 - 91.4
solids reclrcu-
lated.
VFC-1P
35, 000
9.4
1.400
2.8-1.2
8
High
Lime
LI
No
9.4
6. 1 - 9. 7
6,900 - 8,100
5.63
5.13-6.66
0.031 - 0.040
99.4
99.3 - 99.5
6. 0
4. 1 - 7. 8
1.7
. - 2. 3
64.9
43,9 - 85.9
VFG-1E at cor-
rcct condltlona.
VFC-1Q
35, 000
9.4
1,400
2.5
S
Lo«r
Lime
LI
No
9.4
7. ?. - 7. 9
1,000 . 1.660
0.148
0.102 - 0.195
0. 020
O.Olt - 0.02J
86.4
79.6 - 90. 7
--
..
--
with low lnl«t
Fly Ash Loading.
40 my/ate ru«ed
upstream of ven-
turi to protect
rubber lining.
00
(1) LI represents a low frequency, fresh water, top and bottom mist eliminator
wash scheme. The bottom wash was shut off durtng testing.
-------
Division of Acurex Corporation was used. The major differences in test
conditions between sampling at Shawnee and a typical Method Five application
were: (1) increased sampling flow rate (2 cfm versus 0.75 cfm, typical for
EPA Method Five) (2) filter temperatures in the 350°F to 400°F range to
prevent acid ($03) condensation, and (3) the first two impingers filled
with n&2^G^ solution to remove SC^ and to provide corrosion protection
for the pump and dry gas meter.
Particulate size distributions were measured with out-of-stack heated iner-
tia! impactors. A Brink Model BMS-11 impactor was used for inlet sampling
and a Meteorology Research Inc. (MRI) Model 1502 impactor was used for outlet
sampling.
To obtain representative size distribution measurements, sampling traverses
were made. The flow rate through the impactor was held constant to ensure
consistent impactor stage cut sizes. The sampling nozzle was chosen to pro-
vide average duct velocity at the nozzle inlet. Because of the uniformity of
the duct velocity profile, isokinetic sampling rate was maintained within
+_ 10 percent by sampling at the average duct velocity.
Because all size distribution measurements were taken outside of the stack to
facilitate temperature control, a fraction of the particulate was deposited in
the probe and nozzle. For the outlet size distributions, the estimated probe
fallout included a significant fraction of the particulate in the size range
of the first three stages. Thus, it was necessary to combine the probe wash
and the first three stages when computing grain loadings as a function of
particle size. For the inlet measurements, no correction was required.
18-5
-------
18.1.4 Results of Tests on the TCA System
Table 18-1 gives the results of the mass loading tests conducted on the TCA
system. The average outlet mass loading ranged from 0.042 to 0.065 grain/dry
scf for the high inlet fly ash loading tests and was 0.026 grain/dry scf for
the test with low fly ash loading.
The average scrubber outlet emission values with high inlet fly ash loading
were in the range of, and sometimes exceeded, the EPA New Source Performance
Standards for particulate emissions. Preliminary results of mass loading
versus opacity measurements indicated that removals in the 0.020 to 0.025
grain/dry scf range would be required to meet the EPA opacity standard for a
500-MW plant with a single stack (assuming a 70 ft/sec exit velocity). This
value is well below the mass emission requirements.
Nondispersive infrared and thermogravimetric analysis of the particulate on
a single outlet filter from the run with low fly ash loading indicated that
emissions from entrained scrubber reaction products were around 0.005 grain/dry
scf. On the basis of this single analysis, the outlet loadings appeared to
be predominately fly ash. This was confirmed by the preliminary results from
air/slurry entrainment tests on the TCA system which indicated a slurry (40
percent fly ash and 60 percent slurry reaction products) entrainment value
between 0.003 and 0.005 grain/dry scf. This value is in agreement with the
low reaction product emission value and supports the conclusion that over 90
percent of the TCA emission is fly ash.
The highest emission values experienced under base case conditions occurred
with a continuous diluted clarified liquor mist eliminator underwash (Run
18-6
-------
TFG-2C, 0.064 grain/dry scf outlet). Changing to an all fresh-water mist
eliminator wash scheme (Run TFG-2B, 0.048 grain/dry scf outlet) reduced
emissions.
Reducing the circulating slurry rate had no effect on the emissions (Run
TFG-2E, 0.065 grain/dry scf outlet), but reducing the gas rate (Run TFG-2D,
0.056 grain/dry scf outlet) tended to reduce the emissions. MgO addition
(Run TFG-2A, 0.055 grain/dry scf outlet) tended to increase emissions,
although this run did not have the highest emission rate since it used a
fresh-water mist eliminator wash scheme.
Inlet and outlet grain loadings as a function of aerodynamic particle size
were calculated using mean values of the measurements for each run. These
loadings are plotted in Figure 18-1 for runs with high fly ash loading and
in Figure 18-2 for runs with low fly ash loading. The inlet distributions
were similar to the venturi/spray tower measurements, confirming that both
scrubber inlet size distributions were the same even though they were taken
from different sides of the boiler outlet duct.
Mass penetration as a function of aerodynamic particle size is plotted in
Figure 18-3 for runs with high inlet fly ash loading and in Figure 18-4 for
runs with low inlet fly ash loading. The mass penetration for the high fly
ash loading runs varied from about 20 to 90 percent for 0.1-micron particles
and less than 4 percent for particles larger than 5 microns. For runs with
low fly ash loading, penetration was between 25 and 70 percent for 0.1-micron
particles and about 6 percent for particles larger than 5 microns.
For runs with high fly ash loading, emission of particles smaller than 2
microns actual diameter averaged 0.025 grain/dry scf, except for the run
18-7
-------
s
o
*
2
O
3
8
UJ
cc
til
10
i.o - -
.S 01 --
0 01 - -
0001 - -
NOTE: SOLID SYMBOLS ARE
OUTLETS. OTHERS ARE
INLETS
_l—I—,_4_
I ' '"I
I I '"I
0.01
LH •—
0.1 1.0 10
AERODYNAMIC PARTICLE DIAMETER, microns
100
Figure 18-1. Mean Differential Mass Loading Versus Aerodynamic Particle Size
for all TCA System High Fly Ash Loading Runs
18-8
-------
10
1.0 - -
Ol
8
Q
0.1 - -
CD
Q
O
\
2
£C
UJ
o.oi - -
0 001 - -
O.OO01
001
NOTE: SOLID SYMBOLS ARE
OUTLETS. OTHERS ARE
INLETS
£0
6
A
O
0
O
A_ •
•4-
+
+
01 1.0 10
AERODYNAMIC PARTICLE DIAMETER, microns
100
Figure 18-2.
Mean Differential Mass Loading Versus Aerodynamic Particle Size
for all TCA System Low Fly Ash Loading Runs
18-9
-------
00
•
. 10 •
Z !
g
cc
UJ
z
UJ
o.
(A
| 1.0 .
0.1 -
' Tf v ' 'B
" R & X a v "
o ° B s g o TI
o r A ° TF
_ p " A TK
8 * T'
6 V TF
t
Y
I
D CD
v v a
v r
V 17
0 o 5
9
* * * 1 * » 1 1 I I A t 1 t • 1 t I 1 1 1 l 1 I • t | 1 • 1 I I • I 1 I
1 | 1*1111
G - 2A ;
•G 2B
G -2C
•G - 20
G 2E
G -2F
•
•
•
•
-
-
'
0.01
0.10 1.0 10 100
AERODYNAMIC PARTICLE DIAMETER, microns
Figure 18-3. Mass Percent Penetration Versus Aerodynamic Particle Size
for all TCA System High Fly Ash Loading Runs
-------
100
03
I
O
DC
UJ
z
UJ
a.
\
I 1
0.1
'
d 6
0
0
0
6
S
0
o
i i -I
A
A
6 o
A703 2A
O704 2 A
0705 2A
D706 - 2A
10 •-
A
a
o
A
a
o
o
A
O
0
.. I
0.01
0.10 1.0 10 100
AERODYNAMIC PARTICLE DIAMETER, microns
Figure 18-4. Mass Percent Penetration Versus Aerodynamic Particle Size
for all TCA Low Fly Ash Loading Runs
-------
with low gas rate (TFG-2D) and the run with low slurry rate (TFG-2E), where
the average was higher at 0.037 grain/dry scf. In a TCA, this decreased
removal efficiency would be expected with slurry or gas turndown because of
decreased sphere activity in the beds. Emission of particulates with diameters
greater than 2 microns averaged 0.028 grain/dry scf, except for the base
case run with a continuous mist eliminator wash (TFG-2C) and the run with
MgO addition (TFG-2A), where the emissions averaged 0.055 and 0.038 grain/dry
scf, respectively. The low fly ash loading run averaged 0.003 grain/dry
scf less than 2 microns actual diameter and 0.025 grain/dry scf greater than
2 microns actual diameter.
From these data, it would appear that slurry entrainment after evaporation
in the reheater contributes mostly to the larger size fraction of the emissions.
For example, the continuous diluted clarified liquor underwash (TFG-2C versus
TFG-2B) consisting of water and dissolved solids significantly increased
emissions only in the greater than 2 micron fraction. Also, when the gas rate
(TFG-2C versus TFG-2D) or the liquor rate (TFG-2C versus TFG-2E) was reduced,
the concentration of fine particulate increased and the concentration of larger
particulate decreased. This operation probably resulted from a decrease in
slurry entrainment. MgO addition, which increased the total dissolved solids,
also resulted in an increase in emission for the larger sizes.
18.1.5 TCA Run Descriptions
All the tests on the TCA system run specifically for flue gas characterization
data are described below.
18-12
-------
Run TFG-2B began on February 4 and ended on February 10, 1977, after 128
operating hours. The run was the first in the series specifically designed
to collect flue gas characterization data at typical operating conditions.
The major test conditions and results are listed in Table 18-1. Flue gas
data had also been collected during Run 706-2A and are included in Table 18-1
for comparison.
For Run TFG-2B, the average inlet grain loading was 6.07 grains/dry scf and
the outlet was 0.048 grain/dry scf, a 99.2 percent removal.
Run TGF-2C began on February 10 and ended on February 14 after 93 hours of
operation. Test conditions were the same as for Run TFG-2B except that the
mist eliminator bottom wash scheme was changed from an intermittent wash with
makeup water to a continuous wash with diluted clarified liquor. The purpose
of the run was to observe the effect of this change on the outlet flue gas
measurements. The average inlet grain loading was 5.82 grain/dry scf and the
average outlet grain loading was 0.064 grain/dry scf, a 98.9 percent removal.
The average outlet grain loading is higher than the previous Run TFG-2B.
Run TFG-2D began on February 15 and ended on February 19 after 93 operating
hours. The purpose of the run was to observe the effect of reduced gas flow
on flue gas measurements. Test conditions were the same as for Run TFG-2C,
except that the gas rate was decreased from 30,000 acfm to 20,000 acfm. The
average inlet grain loading was 6.13 grains/dry scf and the outlet grain
loading was 0.056 grain/dry scf, a removal of 99.1 percent.
Run TFG-2E began on February 19 and ended on February 23 after 96 operating
hours. The test conditions differed from those of Run TFG-2C in that the
recirculating slurry rate was 600 gpm (cf. 1200 gpm for Run TFG-2C) and the
18-13
-------
effluent hold tank residence time was increased from 12 to 24 minutes. The
average inlet grain loading was 5.69 grains/dry scf and the outlet grain
loading was 0.065 grain/dry scf, a removal of 98.8 percent. This outlet
grain loading is higher than in the previous run.
Run TFG-2F began on February 23 and ended on February 28 after 112 operating
hours. The test conditions were identical to those for Run TFG-2C, except
that both top and bottom mist eliminator washes were turned off when sampling
the flue gas. The purpose of the test was to eliminate the effect of the
mist eliminator washes on the flue gas measurements. The inlet grain loading
averaged 4.53 grain/dry scf and outlet grain loading averaged 0.042 grain/dry
scf, a removal of 99.1 percent. The outlet grain loading was 0.02 grain/dry
scf less than Run TFG-2C which indicated, as expected, that continuous wash
did contribute to the value.
Run TFG-2A began on February 28 and ended on March 3 after 76 hours of oper-
ation. Test conditions were the same as for Run TFG-2B except that MgO was
added to maintain an effective liquor magnesium ion concentration of 9000 ppm.
The purpose of the run was to observe the effect on the flue gas measurements
of high concentration of dissolved solids in liquor. The inlet grain loading
averaged 4.67 grain/dry scf and the outlet grain loading averaged 0.055
grain/dry scf, a removal of 98.8 percent. The outlet grain loading was not
different from any of the previous runs, even with the large increase in
total dissolved solids in the liquor. Noticeable numbers of scale particles
were found on the outlet filter.
Run 706-2A began on January 14 and ended after 339 hours of operation on
February 4. The test conditions were the same as those during Run TFG-2B
18-14
-------
except that flue gas with low fly ash loading was used. The average inlet
grain loading was 0.124 grain/dry scf, and the average outlet grain loading
was 0.026 grain/dry scf, a removal of only 74 percent.
18.1.6 Flue Gas Characterization - Run VFG-1Q
Run VFG-1Q, a flue gas characterization run, was made on the venturi/spray
tower system. Previously unreported, this run was a continuation of the
testing program for the venturi/spray tower. The results are listed in
Table 18-2, which includes all the venturi/spray tower data for these tests.
Run VFG-1Q was conducted with the venturi plug wide open and without venturi
flow. This was done in order to determine the particulate removal capability
of the spray tower alone when operating on flue gas taken with low fly ash
loading. Lime was used as S02 absorbent for this run.
To cool the flue gas and protect the spray tower rubber lining, about 40 gpm
of water was used in the water sprays upstream of the venturi and in one of
the four venturi tangential nozzles. The flue gas flow rate was 35,000 acfm
(@ 200°F) and the slurry flow rate to the spray tower was 1400 gpm.
Inlet particulate mass loading averaged about 0.15 grain/dry scf. The parti-
culate removal was higher than expected at about 85 percent. The venturi
pressure drop averaged 2.5 inches H^O and (corresponding to the venturi plug
wide open) is believed to have significantly affected the particulate removal.
However, there is no way to isolate the spray tower from the venturi to verify
this assumption.
18-15
-------
18.2 FLUE GAS MASS LOADING TESTS
The results of mass loading tests routinely performed during the reporting
period on the venturi/spray tower and TCA systems are presented in Appendix M.
The procedure used for measuring particulate mass loading was a modification
of EPA Method Five described in Subsection 18.1.3.
In the venturi/spray tower, the outlet mass loading generally decreased as
the pressure drop increased. For example, during Run 852-1A, the average
outlet mass loading was 0.062 grain/dry scf with a system pressure drop of
3.3 inches H20. During Run 819-1A, the average outlet was 0.052 grain/dry
scf with a pressure drop of 6.0 inches H20. During Run 863-1A the average
outlet mass loading was 0.045 grain/dry scf with an average system pressure
drop of 13.7 inches H20, and for Run 866-1A the outlet grain loading was
0.023 grain/dry scf with 16.7 inches H20 pressure drop.
On several occasions, the venturi/spray tower slightly exceeded the EPA New
Source Performance Standard for mass emissions from coal fired power plants
(0.052 grain/dry scf assuming 30 percent excess air to the boiler).
The TCA showed the same relationship between the pressure drop and the outlet
mass loading. For example, during Run 719-2A, when the pressure drop of the
system averaged only 3.4 inches H20, the outlet mass loadings exceeded the
performance standard during each test and the average grain loading was
0.060 grain/dry scf. In Run 619-2A, the average outlet grain loading was
0.044 grain/dry scf with an average pressure drop of 9.4 inches H20. In
Run 618-2A, the outlet grain loading averaged 0.032 grain/dry scf with an
average pressure drop of 12.4 inches H20.
18-16
-------
The TCA exceeded the performance standard during several periods. In general,
the standard was exceeded when the pressure drop was below about 8 inches H20.
18.3 AIR/SLURRY TEST RESULTS
During the April to May 1977 boiler outage, air/slurry tests were performed
on both the venturi/spray tower and TCA systems in an attempt to better deter-
mine the effect of slurry entrainment on scrubber mass emissions. If air is
substituted for flue gas in the scrubber, particulate emissions from the
scrubber should result only from the entrained slurry. Unfortunately, the
inlet air dust concentration was high (0.0006 to 0.0098 grain/dry scf) and
oil from the oil-fired reheater was noted in the scrubber outlet probe wash
during the TCA tests. Even after correcting for the presence of the oil,
there was still no way to determine how much of the inlet dust had not been
removed.
Thus, the only conclusion that can be drawn is that slurry-related mass
emissions must be less than the measured outlet mass emissions. For the
range of conditions tested, mass emissions of 0.001 to 0.003 grain/dry scf
were measured for the venturi/spray tower system and 0.001 to 0.005 for the
TCA system. The results of the tests for the TCA system and for the venturi/
spray tower system are presented in Table 18-3 and 18-4, respectively.
18.3.1 Test Program
The test conditions are outlined in Table 18-5 for the TCA system and Table
18-6 for the venturi/spray tower system. The volumetric air flow rate was
18-17
-------
Table 18-3
TCA SYSTEM AIR/SLURRY TEST RESULTS
Date
4/7/77
4/7/77
4/7/77
4/8/77
4/8/77
4/8/77
4/11/77
4/1Z/77
4/12/77
4/12/77
4/13/77
4/14/77
4/14/77
4/15/77
4/18/77
4/19/77
4/20/77
6/1/77
5/2/77
6/3/77
Run
No.
TAS-2A*
H
TAS-2A.,.
TAS-2A'"
TAS-2B
TAS-2B
TAS-2B
TAS-2A
TAS-2D
TAS-2D
TAS-2D,,.
TAS-2A*
TAS-2A
TAS-2A
TAS-2F
TAS-2A
TAS-2A
TAS-2A
TAS-2F
TAS-2F,,,
TAS-2A
Inlet
Total
_ _
--
--
--
- .
0.0070
--
--
0.0014
0.0012
0.0011
--
--
--
0.0013
0. 0098
0.0015
0. 0006
Mass Loading, gr/dscf
Outlet
Total | Probe | Filter | Est. '
0.0056 0.0013 0.0043
0.0053 0.0007 0.0046
0.0023 0.0003 0.0020
0.0086 0.0074 0.0012 0.001
0.0045 0.0006 0.0039 0.005
0.0164 0.0147 0.0017 0.002
--
0.0259 0.0232 0.0027 0.003
0.0201 0.0164 0.0037 0.004
0.0044 0.0039 0.0005 0.001
0. 0016 0. 0004 0. 0012
- -
--
0.0044 0.0019 0.0025 0.003
0.0029 0.0017 0.0012 0.001
0.0022 0.0008 0.0014 0.002
--
0.0078 0.0002 0.0076
0.0045 0.0002 0.0043
0.0010 0.0000 0.0010
Reheater
Oil Valve
% Open
100
100
100
100
100
100
100
100
100
100
_ _
100
100
29
--
20
25
30
Comment
No known oil
No known oil
No known oil
Oil in probe wash
Slight oil in probe wash
Oil in probe wash
Boilerhouae air, no oil
Oil in probe wash
Oil in probe wash
Slight oil in probe wash
No oil
Ambient air, no oil
Ambient air, no oil
Trace of oil
Oil in probe wash
Oil in probe wash
No oil
Boilerhouse air, no oil
Ambient air, no oil
No oil
00
MU
00
Estimated oil free outlet loading. Calculated by determining the probe
wash/filter wt. ratio for oil free runs using slurry (runs used are marked
with an asterisk) and increasing the filter loading in the contaminated
samples by this fraction (0. 19). Due to the uncertainty in this value, the
loadings were rounded off.
* See note above.
-------
CO
I
Table 18-4
VENTURI/SPRAY TOWER AIR/SLURRY TEST RESULTS
Date
5/5/77
5/9/77
5/9/77
5/11/77
5/11/77
5/12/77
5/12/77
5/13/77
5/13/77
5/16/77
5/16/77
5/17/77
5/18/77
5/19/77
5/20/77
Run
No.
VAS-1A
VAS-1A
VAS-1A
VAS-1B
VAS-1B
VAS-1B
VAS-1C
VAS-1C
VAS-1C
VAS-1E
VAS-1E
VAS-1E
VAS-1F
VAS-1F
VAS-1G *
Mass Loading,
gr/dscf
Inlet
« «
--
__
0.0008
0.0007
--
--
--
--
--
--
0.0021
0.0014
0.0011
0.0026
| Outlet
0.0028
0.0017
0.0022
0.0023
0.0011
0.0015
0.0015
0.0021
0.0027
0.0027
0.0019
0.0029
0.0015
0.0009
0.0010
Reheater
Oil
Valve,
% Open
100%
100%
100%
35%
30%
33%
33%
35%
35%
30%
25%
40%
40%
35%
40%
* Maximum Venturi AP = 9" WQ
-------
Table 18-5
RUN CONDITIONS FOR THE AIR/SLURRY TESTS ON THE TCA SYSTEM
RecircuUted (lurry, gpm
Equivalent inlet flue gas rate, aefm
Superficial gas velocity, ft/aec
• TCA
• mist eliminator
Percent lolida reclrculated
Mist eliminator wash acheme
Bed Configuration
Run Philosophy
TAS-2A
1200
30.000
12.5
8. 2
15
None
J beds, S Inches spheres/bed.
Normal bed potltioni.
Measure (lurry entralnmenf at
high gas rate.
TAS-2B
1200
20,000
8.4
5.5
15
3-paaa, open-vane chevron
3 bedi, 5 Inches spheres/bed.
Normal bed positions.
Measure slurry entralnment at
low gas rate.
TAS-2D
1200
JO. 000
12.5
8.2
15
3-pass, open-vane chevron
2 beds, 5 inches spheres/bed.
Removed top bed.
Measure effect of no- ">f beds
on slurry entrainment.
TAS-2F
0
30,000
12.5
8.2
None
Normal bed positions.
to the mass loading.
00
t\3
O
Table 18-6
RUN CONDITIONS FOR THE AIR/SLURRY TESTS ON THE
VENTURI/SPRAY TOWER SYSTEM
Rectrculated slurry, gpm (V/ST)
Equivalent Inlet flue gas rate, acfm
Superficial gas velocity, ft/sec
» Spray tower
• Mist eliminator
Percent solids reclrculated
Venturi AP, in. H,O
Mist eliminator
Mist eliminator wash scheme
Run Philosophy
VAS-1A
600/1400
35,000
9.4
9.4
8.0
9
chevron
None
Measure slurry en-
tratnment at high gas
rate.
VAS-1B
600/1400
20,000
5.4.
5.4
8. 0
9
chevron
None
Measure slurry en-
tralnment at low gas
rate.
VAS-1C
0/1400
35,000
9.4
9.4
8.0
9
chevron
None
Measure effect of
venturi on slurry
entrainment.
VAS-1E
0/0
35,000
9.4
9.4
..
--
chevron
None
Measure reheater
contribution to the
mass loading.
VAS-1F
600/0
35,000
9.4
9.4
8.0
9
chevron
None
Measure effect of
spray tower on slurry
entrainment.
VAS-1G
600/0
35. 000
9.4
9.4
8. 0
14
chevron
None
Measure effect of
venturi pressure
drop on slurry en-
trainment.
-------
adjusted to provide a scrubber superficial air velocity equal to the scrubber
superficial flue gas velocity at the specified inlet flue gas rate if flue
gas had been used. After leaving the scrubber, the air was reheated and the
resulting particulate mass loading was measured.
18.3.2 Discussion of Test Results
As can be seen in Tables 18-3 and 18-4, the inlet air mass loadings were of
the same order of magnitude as the outlet mass loadings. Therefore, the
determination of how much of the outlet mass loading measurement represented
unremoved dust from the inlet air and how much was slurry entrainment was
not possible. However, the amount of slurry emissions was less than or equal
to the amount of outlet emissions since tjie latter consists of slurry entrain-
ment, plus any dust not removed from the inlet air.
For the TCA scrubber, there was an additional complication. During testing,
what appeared to be unburned fuel oil from the oil-fired reheater was found
in the probe washes of the outlet mass loading tests. No oil accumulations
had been experienced in previous testing on either system, and the source of
oil could not be determined before testing was terminated due to routine
maintenance scheduled during the boiler outage.
In early June 1977, three tests were run to investigate if the TCA oil problem
could be resolved, but the problem apparently cleared up by itself since no
oil was found in the outlet mass loading probe washes.
An oil-free slurry entrainment mass loading upper limit was calculated for
the TCA by noting that during tests, when no oil was found, the outlet mass
18-21
-------
loading probe wash averaged 19 percent of the filter mass loading (see note,
Table 18-3). Adding this percentage to the filter loading resulted in an
upper limit for slurry entrainment of 0.001 to 0.005 grain/dry scf. The
corresponding venturi/spray tower upper limit for slurry entrainment was
0.001 to 0.003 grain/dry scf.
An upper limit on the outlet moisture loading was estimated by assuming that
the mass emissions during these tests represented solids from a slurry (TCA
slurry was 16 weight percent total solids, venturi/spray tower slurry was 9
weight percent total solids) containing both suspended and dissolved solids.
An estimated upper limit on the outlet moisture loading of 0.03 grain/dry
scf (0.003 grain/dry scf upper limit of slurry solids entrainment for the
venturi/spray tower, divided by 9 percent slurry concentration) was obtained
for both systems. This appears to be a rather low moisture emission value
for a three-pass, open-vane, chevron type mist eliminator unless the moisture
loading to the mist eliminator is also very low.
18-22
-------
Section 19
LIMESTONE REACTIVITY STUDY
Tests of limestone reactivity were conducted in the Shawnee laboratory to
study the effect of limestone source and grind on reactivity and utilization.
Limestones from eleven different quarries were studied; each stone was ground"
to fine and coarse sizes. Two laboratory methods were used to test reactivity:
the HCL method and the S02 method. Results from the HC1 method are presented
in Subsection 19.2 and results from the S^ method are presented in Subsection
19.3. Recommendations for improvements in the laboratory methods are dis-
cussed in Subsection 19.4.
19.1 LABORATORY TEST PROGRAM
Limestones from different quarries vary in chemical composition and physical
properties such as hardness, average pore size and pore size distribution,
and porosity (void fraction). Both chemical and physical factors can affect
the reactivity. Due to lack of equipment needed to determine these physical
characteristics, it was not the attempt of this program to correlate the
limestone reactivity values with physical properties.
The eleven limestones studied and their chemical compositions (as determined
by TVA, Muscle Shoals) are listed in Table 19-1.
19-1
-------
Table 19-1
LIMESTONES USED IN REACTIVITY TESTS
.
10
1
t\>
Limestone Source
Citadel Stone Company
Commonwealth Edison
Cowan Stone Company
Fredonla Quarries
Georgia Marble
Hoover Incorporated
Long view Lime Company
Luttrell Mining Company
Rlgsby Sc Barnard Quarry
The Stone Man
Vulcan Material Company
Location
Selma, Alabama
Jollet, Illinois
Cowan, Tennessee
Fredonla, Kentucky
Sylacauga, Alabama
Nashville, Tennessee
Longvlew, Alabama
Knoxvllle, Tennessee
Cave In Rock, Illinois
Chattanooga, Tennessee
Knoxville, Tennessee
Chemical Composition, wt percent
Calcium 1 Magnesium
(Ca) | (Mg)
32.5 0.2
.37.9 1.4
36.0 l.S
35.8 2.5
38.8 0.5
36.6 1.2
38.2 0.7
38.8 0.2
38.4 0.2
36.4 1.0
38.1 1.1
Carbonate I Acid
(CO".,) 1 Insolubles
47.4 14.8
56.6 3.5
55.9 4.0
60,5 2.2
59.2 2.0
54.1 3.8
59.9 1.0
58.4 1.4
57.8 0.4
57.2 4.1
60.2 0.5
Grind, percent less than
-200
Mesh
94
65
80
75
71
92
72
77
69
72
97
-325
Mesh
93
78
91
90
78
84
77
82
90
78
95
-------
Since the reaction of limestone is a heterogeneous reaction, one would expect
the reactivity to vary with surface area, which can be changed by grinding.
Therefore, each limestone was tested at two different grinds. The limestones
as received (generally 1/4 to 1 inch in size) were ground and screened at the
International Fertilizer Development Corporation (IFDC) in Muscle Shoals,
Alabama. Two samples of each limestone were prepared: a coarse grind of
about 75 percent less than 200 mesh and a fine grind of about 90 percent less
than 325 mesh. The actual measured grind for each sample as determined by
wet screen analysis is given in Table 19-1.
19.2 LIMESTONE REACTIVITY BY THE HC1 METHOD
The first method used to test limestone reactivity involved the addition of
limestone to a solution of HC1. For each test, 5 grams of dried limestone
were added to a 300-ml solution of HC1 at pH 3 and 25°C, and the rise of pH
with time was recorded.
19.2.1 Data Analysis
The original HC1 method used the slope of the pH versus time curve at pH 4
for comparison of limestones. Due to the steepness of the curve, the slope
could not be determined accurately. In addition, a reactivity value at pH
4 is not representative of the reactivity at normal scrubber operating pH's
of 5 to 6. Inaccuracies also result if the initial pH's vary.
Due to these problems, an integral method of analysis was developed. When
limestone reacts with a strong acid below pH 7, the following reactions
19-3
-------
predomi nate:
CaC03(s) + 2H+ -+• Ca++ + H2C03
H2C03 -»- H+ + HC03~
The sum of these reactions is:
CaC03(s) + H+ —* Ca++ + HC03"
Thus, the rate of limestone dissolution equals the rate of hydrogen ion deple-
tion. If reactivity is defined as the change in CaC03 concentration with time,
then integration and rearrangement give [remembering that pH = -1
Reactivity ^ -d(CaC03) = -d(H+) = -d(10"PH) = ip(-PHl) -ip(-pH2)
dt dt dii t2 - tx
For an initial pH of about 3 at time zero and a final pH of 7 at time t2,
Reactivity ^ l(){"PHl)
or
Reactivity = Initial Hydrogen Ion Concentration of the
HCL Aliquot, mmole/1
Time to reach pH 7, min(19-1)
The strong effect of initial pH indicated by Equation 19-1 was experimentally
proven during initial testing. As a result, only those tests where the initial
pH was between 2.95 and 3.10 were used for the final anaylsis. Initial testing
also showed that better mixing gave higher reactivity values.
19-4
-------
Fifty-four tests were run on the twenty-two limestone samples and analyzed
with Equation 19-1. Replication was done on a random basis. Of the fifty-
four tests, eight were rejected because they were run during the off-shift
without supervision and values were consistently high.
19.2.2 Reactivity Comparison by Source
Since all the limestones were not of the same grind, relative comparisons
could not be made. However, for each grind specification, two to five
different limestones could be compared. These comparisons are given in
Table 19-2.
From the limited comparison possible, the Vulcan Materials Company limestone
was the most reactive in the fine grind class (nominal 90 percent less 325
mesh), whereas the Stone Man limestone was the most reactive in the coarse
grind class (nominal 70 percent less 200 mesh). The finely ground Vulcan
limestone was about twice as reactive as the Fredonia fine limestone routinely
used at Shawnee.
The least reactive limestones in the fine and coarse classes were the Fredonia
Valley and the Georgia Marble varieties, respectively.
19.2.3 Reactivity Comparison by Grind
As expected, the reactivity increased with the fineness of the grind. Results
from fine and coarse grind are rearranged by source and presented in Table
19-3. Once again, since all the limestones were not of the same grind, rela-
tive comparisons could not be made to evaluate the relative benefits of grind-
19-5
-------
Table 19-2
REACTIVITY COMPARISON BY SOURCE
(HC1 Method)
Limestone
The Stone Man
Commonwealth Edison
Georgia Marble
Longview Lime Co.
Hoover Incorporated
Luttrell Mining Co.
Vulcan Material Co.
Rigsby & Barnard
Cowan Stone Co.
Citadel Stone Co.
Fredonia Valley Quarries
The Stone Man
Rigsby & Barnard
Commonwealth Edison
Longview Lime Co.
Georgia Marble
Fredonia Valley Quarries
Cowan Stone Co.
Luttrell Mining Co.
Vulcan Material Co.
Citadel Stone Co.
Hoover Incorporated
Grind
Mesh
-200 -325
—
--
-_
—
__
—
—
—
—
72
69
65
72
71
75
80
77
97
94
92
78
78
78
77
84
82
95
90
91
93
90
__
—
—
—
—
._
—
--
__
__
—
Test
2,
6,
14,
52
7,
10,
16,
22,
4,
54
50
13,
11.
1,
51
5,
49
3,
9,
15,
53
8,
26
30
38
21,
34
24
17
19,
23
35
18,
27,
27
22,
39
32
No1
31.
, 40
, 36
28,
, 37
25,
29,
33,
s.
45
, 48
, 41
43
, 47
42
44
46
Average
Reactivity,
mmole/1-min
3
2
2
1
3
2
5
3
3
3
2
3
2
1
0
0
2
2
1
3
2
2
.16
.76
.10
.23
.27
.88
.44
.56
.17
.06
.66
.00
.12
.60
.94
.90
.24
.06
.68
.81
.85
.40
Rank
1
2
3
4
1
2
1
2
3
4
5
1
2
3
4
5
1
2
3
1
2
3
19-6
-------
Table 19-3
REACTIVITY COMPARISON BY GRIND
(HC1 Method)
Limestone
Source
Citadel Stone Co.
Commonwealth Edison
Cowan Stone Co.
Fredonia Valley Quarries
Georgia Marble
Hoover Incorporated
Longview Lime Co.
Luttrell Mining Co.
Rigsby & Barnard
The Stone Man
Vulcan Materials Co.
Gri nd
Mesh
-200 -325
94
65
80
75
71
92
72
77
69
72
97
93
78
91
90
78
84
77
82
90
78
95
Average
Reactivity
Test No. 's mmoles/1-min
53
54
1, 18, 25, 42
2, 26 •
3, 27
4/19, 28, 43
49
50
5, 20, 29, 44
6, 30
8, 32
7, 21, 31, 45
51
52
9, 22, 33, 46
10, 34
11, 35
12, 17, 36, 41
13, 23, 37, 47
14, 38
15, 39
16, 24, 40, 48
2.85
3.06
1.60
2.76
2.06
3.17
2.24
2.66
0.90
2.10
2.40
3.27
0.94
1.23
1.68
2.88
2.12
3.56
3.00
3.16
3.81
5.44
Percent
Increase
in Reactivity
7
42
35
16
57
27
24
42
40
5
30
19-7
-------
Ing. Instead, the percentage increase in reactivity with grinding for each
type of limestone was calculated. These percentage increases are shown
in Table 19-3.
Grinding was the most effective in increasing reactivity for the Georgia
Marble limestone and the least effective for the Stone Man variety. To ex-
plain this behavior, it is hypothesized that the low reactivity of the
Georgia Marble limestone resulted from a combination of low porosity and
small pore size. When this limestone was crushed to make smaller particles,
new surface was exposed at the points of fracture, and thus a strong depend-
ence on reactivity with particle size was observed. Unlike the Georgia
Marble variety, it is believed that the Stone Man limestone had relatively
large pores and therefore, exhibited a weak dependence on grinding. Because
of lack of equipment, measurements of mean pore diameter and pore volume,
usually determined using mercury intrusion tests, were not measured.
19.3 LIMESTONE REACTIVITY BY THE S02 TITRATION METHOD
The second method used to test limestone reactivity was an SOg titration of
the limestone slurries. First, 0.2 mole of limestone was dissolved in two
liters of distilled water (0.1 M CaCQ^) at 50°C until a constant pH was
reached. Then, a constant stream (0.273 g-mole/hr) of pure S0£ was added,
and the pH drop with time was monitored.
This method, though somewhat more involved than the HC1 method, was believed
to be more representative in evaluating limestones to be used in S02 scrubbing
systems because:
19-8
-------
• The reaction products (CaSC^, CaSO^ etc.) are representative
of the products formed during wet scrubbing of flue gas.
t Due to their insoluble nature, these reaction products may plug
the limestone pores and thus reduce the reactivity of the lime-
stone. Such effects would not be simulated in the HC1 method
where soluble reaction products were formed.
• The buffering effect, around a pH of 5.8, characteristic of
limestone scrubbing systems, were more nearly approximated by
this method.
• The acid produced by this method, sulfurous acid, was weak com-
pared to HC1, a strong acid. For limestone scrubbing systems,
sulfurous acid chemistry was dominant.
• The effects of subsaturation, supersaturation, and blinding
were more nearly approximated.
19.3.1 Data Analysis
A total of thirty-three SC^ titrations were run on the twenty-two limestone
samples. Replication was done on a random basis to insure validity.
Data were collected as a pH versus time curve. Results were compared in terms
of limestone utilization, as defined below:
Percent Utilization = Time to Reach pH End-Point x S02 Feed Rate x 100
at a Given pH Amount of Limestone
_ Time to Reach pH x 0.273 g-moles S02/hr x 100
0.2 g-moles of limestone
19.3.2 Utilization (Reactivity) Comparison by Source
Since all the limestones were not of the same grind, relative comparison
could not be made. However, for each grind class, two to five limestones
could be compared. These comparisons are given in Table 19-4.
19-9
-------
Table 19-4
UTILIZATION COMPARISON BY SOURCE
*(S02 Titration)
Grind
Mesh
Limestone Source -200 -325
The Stone Man
Commonwealth Edison
Longview Lime Co.
Georgia Marble
Luttrell Mining Co.
Hoover Incorporated
Rigsby & Barnard
Fredonia Valley Quarries
Vulcan Material Co.
Citadel Stone Co.
Cowan Stone Co.
Rigsby & Barnard
The Stone Man
Commonwealth Edison
Georgia Marble
Longview Lime Co.
Fredonia Valley Quarries
Luttrell Mining Co.
Cowan Stone Co.
Citadel Stone Co.
Vulcan Materials Co.
Hoover Incorporated
•• M
--
--
__
--
--
—
—
--
--
69
72
65
71
72
75
77
80
94
97
92
78
78
77
78
82
84
90
90
95
93
91
__
.-
--
--
--
__
--
--
__
--
--
Test No. 's
14, 14R
2,2R
20, 20R
6
10, 10R
7
12, 12R
22, 22R
16,16R
18
4
11,11R
13
1
5, 5R
19
21,21R
9
3,3R
17
15
8
Average
Utilization
at pH 5. 4, %
61.9
61.4
59.2
40.9
63.2
38.7
100.0
95.5
89.9
64.8
47.8
72.8
40.9
39.8
30.7
23.9
50.5
40.9
37.1
69.4
60.3
38.7
Rank
1
2
3
4
1
2
1
2
3
4
5
1
2
3
4
5
1
2
3
1
2
3
19-10
-------
From the limited comparison possible, the Rigsby & Barnard limestone was found
to be the most reactive in both the fine grind (nominal 90 percent less than
325 mesh) and the coarse grind (nominal 70 percent less than 200 mesh) classes.
Fredonia fine limestone used at Shawnee ranked second best in the fine grind
class.
19.3.3 Utilization Comparison by Grind
As in the HC1 method results, finer limestones had higher reactivities than
their coarser counterparts, as shown in Table 19-5.
As before, the benefits of grinding could not be evaluated for all the sam-
ples. If the variability of the size cuts is ignored, one finds that finer
grinding is most beneficial for the Longview and the Fredonia limestones
and least beneficial for the Hoover Incorporated and Citadel Stone varieties.
The percentage increases in utilization with finer grinding are given in
Table 19-5.
19.4 COMPARISON OF AND RECOMMENDATIONS FOR THE HC1 AND SO?
TITRATION METHODS
In general, the S02 method was judged to be superior to the HC1 method because
it more closely approximated S02 scrubbing with limestone (see Subsection 19.3),
Due to the disparity of the methods, the data were not comparable. Shortcom-
ings of the S02 method include the complexity of the equipment and the time
per test (one hour for the S02 method and five minutes for the HC1 method).
19-11
-------
Table 19-5
UTILIZATION COMPARISON BY GRIND
(S02 Titration)
Limestone Source
Citadel Stone Co.
Commonwealth Edison
Cowan Stone Co.
Fredonia Valley Quarries
Georgia Marble
Hoover Incorporated
Long view Lime Co.
Luttrell Mining Co.
Rigsby & Barnard
The Stone Man
Vulcan Materials Co.
Grind
Mesh
-200 -325
94
65
80
75
71
92
72
77
69
72
97
93
78
91
90
78
84
77
82
90
78
95
Test No. 's
17
18
1
2, 2R
3,3R
4
21.21R
22, 22R
5, 5R
6
8
7
19
20, 20R
9
10, 10R
11.11R
12, 12R
13
14, 14R
15
16, 16R
Average
Utilization
at pH 5. 4, %
69.4
64.8
39.8
61.4
37.1
47.8
50.5
95.5
30.7
40.9
38.7
38.7
23.9
59.1
40.9
63.2
72.8
100.0
40.9
61.9
60.3
89.9
Increase in
Utilization
at pH 5. 4
-7
35
22
47
25
0
60
35
27
34
33
79-12
-------
The HC1 method could be improved by increasing the amount of HC1 solution
at pH 3 from 300 ml to one liter and decreasing the limestone sample from
5 grams to 0.1 gram, thus diluting the slurry and making it less subject
to pH fluctuations. In addition, the slope method should be discarded in
favor of the integral method described in Subsection 19.2.
The S02 titration method assumes that there is a complete reaction between
the S02 and the limestone. Use of an S02 analyzer on the offgas would
improve this method. Utilization could then be calculated from the amount
of S02 absorbed rather than from the amount of S02 entering the system.
Both methods would be more reproducible if the limestone samples were more
uniform. During S02 titration tests, the initial pH of the limestone slurry
in water varied as much as 1.3 pH units for limestone samples taken from the
same container. This could be caused by nonuniform chemical or physical
properties, nonuniform grinding, and/or settling during transportation.
Better correlations with limestone type could be made if values of physical
properties such as hardness, density, porosity, and average pore size were
available.
19.5 PILOT PLANT TEST PROGRAM
To verify the ranking by reactivity or utilization of a few of the limestones
and to gain a better insight into the role of physical properties of the lime-
stone, a test program to conduct limestone type and grind tests on a pilot
scrubber has been developed. Salient features of this program are described
below.
19-13
-------
19.5.1 Limestone Preparation
The limestone samples will be procured by TVA and ground at the Allis-Chalmers
grinding test facility in Milwaukee, Wisconsin. Allis-Chalmers will also pre-
pare the samples for the Martin Marietta test laboratory in Baltimore, Maryland.
At this laboratory, chemical composition, petrographic features, specific sur-
face area, and reactivity will be determined.
19.5.2 Limestone Type
The four limestones to be tested include Fredonia, Stone Man (Tiftonia),
Georgia Marble, and Rigsby and Barnard. Two more limestones may be added later.
19.5.3 Test Program
Testing will be done on the pilot scrubber operated by the EPA at Research
Triangle Park (RTP), North Carolina. The RTP pilot plant will be run under
the following conditions:
t TCA scrubber with internals to provide 8 inches 1^0 of
pressure drop
• 8 percent solids in recirculating slurry
• No fly ash
• 12.5 ft/sec (3.8 m/sec) superficial gas velocity
0 50 gal/Mcf liquid-to-gas ratio
• 3000 ppm S02 inlet
• 5000 ppm Cl in liquor
• 9 minutes' residence time (single hold tank)
19-14
-------
Stoichiometry will vary with limestone type and grind. Selection of stoichio-
metry is discussed below.
The pilot plant test runs are expected to last ten days each. At a feed
Stoichiometry of 1.2 and an SC^ removal of 80 percent, the limestone feed
rate is 11.8 Ib/hr or 1400 lb/5-day period.
The following tests have been planned for the RTF scrubber*:
Weeks 1 and 2 Coarse Fredonia (high Stoichiometry)
3 and 4 Coarse Fredonia (medium Stoichiometry)
5 and 6 Coarse Fredonia (low Stoichiometry)
7 and 8 Very Coarse Fredonia
9 and 10 Fine Fredonia
11 and 12 Coarse Stone Man
13 and 14 Fine Stone Man
15 and 16 Coarse Georgia Marble
17 and 18 Fine Georgia Marble
19 and 20 Coarse Rigsby and Barnard
21 and 22 Fine Rigsby and Barnard
where nominally fine = 90% less than 325 mesh
coarse = 75% less than 200 mesh
very coarse = 60% less than 200 mesh
19.5.4 Limestone Type and Grind Control Method
A different control method will be utilized for the pilot plant tests. Ideally
the limestones should be compared at the same S02 removal. However, the pilot
plant is designed for stoichiometric control. An estimate of S02 removal
versus Stoichiometry will be made for each limestone by the following method.
Coarse (75 percent less than 200 mesh) Fredonia limestone will be tested at
*Two other limestones will be tested later.
19-15
-------
three different stoleniometries to give a curve of stoichiometry versus S02
removal. On the basis of this curve, an S02 removal control point will be
chosen. It will be assumed that S02 removal depends primarily on total
limestone surface area. The total surface area is the product of the speci-
fic surface area (nr/g, measured by Martin Marietta) and the limestone feed
rate (g/time). The limestone feed rate for each type and grind test will
be set so that total surface area is constant and is equal to the total sur-
face area of Fredonia coarse at the S02 removal control point. This should
give a constant S02 removal at varying stoichiometries. In practice, the
S02 removal from each test will differ from the selected S02 removal control
point. It will then be assumed that the S02 removal versus stoichiometry
curve for each type and grind will be parallel to the curve for Fredonia
coarse. Comparisons will be made on the basis of varying stoichiometry at
the S02 removal control point.
After testing at the RTP pilot plant, the results for selected limestone
types and grinds will be confirmed by tests at the Shawnee Test Facility.
19-16
-------
Section 20
FLUE GAS PRESSURE DROP CORRELATIONS
20.1 SPRAY TOWER PRESSURE DROP CORRELATION
The flue gas pressure drop across the four slurry headers in the spray
tower was measured as a function of superficial gas velocity, total slurry
flow rate, and number and position of the headers in use. These measure-
ments were taken during venturi/spray tower Run 820-1A (see Section 6)
before steady state was reached for the run.
Figure 20-1 shows the flue gas pressure drop as a function of superficial
gas velocity and slurry flow rate with all four spray headers in use.
Figure 20-2 presents the effect of the number and position of spray headers
p
on the flue gas pressure drop at 14 gpm/ft total slurry flow rate.
The top three headers sprayed slurry downward and the bottom header sprayed
slurry upward (see Figure 3-1). Each header had seven Bete No. ST48FCN noz-
zles. The vertical distance between the spray headers was approximately 4
feet. The pressure drops across the four headers were measured by pressure
differential indicator PDI-1012. The lower pressure tap was located 20 inches
below the centerline of the bottom header and the upper tap 7 inches above
the centerline of the top header.
Equation 20-1 gives flue gas pressure drop as a function of gas and slurry
20-1
-------
1.4 ••
1.2 -•
SLURRY FLOW RATE,
gpm/ft2
• 28
D 20
14
SPRAY HEADERS
USED
ALL 4
ALL 4
ALL 4
CO
oc
ui
a
<
UJ
1.0 --
OC
8,
0.8 -•
0.6 --
O
OC
u
<
Q.
O
£
oc
i
UJ
OC
O.
CO
< 0.4
O w^»
UJ
D
-I
U.
0.2
0.0
4-
•t
-¥•
4-
4.0
5.0 6.0 7.0 8.0 9.0
SPRAY TOWER SUPERFICIAL GAS VELOCITY, ft/sec @ 125° F
10.0
Figure 20-1
The Effect of Gas Velocity and Slurry Flow Rate on the
Flue Gas Pressure Drop Across the Four Slurry Headers
(7 Nozzles/Header) in the Spray Tower
20-2
-------
T
1.4 --
1.2 -•
SLURRY FLOW RATE,
gpm/ft^
14
14
14
SPRAY HEADERS
USED
TOP 2
ALL 4
BOTTOM 2
V)
oc
UJ
9 1.0
x
>
oc
a.
CO
O
OC
O
<
a.
O
CC
Q
UJ
OC
1
LU
OC
a.
O
ui
0.8 - -
0.6 --
0.4 --
0.2 --
0.0
4.0
5.0 6.0 7.0 8.0 9.0
SPRAY TOWER SUPERFICIAL GAS VELOCITY, ft/sec @ 125° F
Figure 20-2
The Effect of Header Position and Number of Headers
(7 Nozzles/Header) on the Flue Gas Pressure Drop
at 14 gpm/ft^ Total Slurry Flow Rate
20-3
-------
flow rates for the case when all four slurry headers were used (see Figure
20-1):
AP = 0.0063V1'17 exp(0.084L) (20-1)
where:
AP = flue gas pressure drop across four spray headers, inches ^0
v = superficial gas velocity in the spray tower, ft/sec at 125°F
•)
L = slurry flow rate to four headers, gpm/ft of spray tower
cross-seel iopal area (Shawnee spray tower cross-sectional
area = 50 ft2)
Equation 20-1 accounts for 92 percent variation in the data, with a standard
error estimate of 0.06 inch H20. The equation should allow for reasonable
extrapolation beyond the range of operation.
The effects of both header position and the number of headers on the pressure
drop were studied at 14 gpm/ft2 total slurry flow rate. Figure 20-2 indicates
that the pressure drop was the highest when only the top two headers were used,
next highest with all four headers, and lowest with only the bottom two headers.
This could be explained by the fact that the slurry holdup in the spray tower
between the pressure taps was highest when the top two headers were used, and
lowest when the bottom two headers were used.
20.2 TCA PRESSURE DROP CORRELATION AND HOLDUP TESTING
i
A series of air/water pressure drop and liquid holdup tests were made on the
TCA during the April-June 1977 boiler outage. Pressure drop data were used
to determine flooding conditions. Flooding is characterized by a large
20-4
-------
increase in pressure drop for a small increase in gas velocity, and by large
fluctuations in pressure drop at constant gas velocity. The data indicated
that flooding usually occurred at 8 to 10 inches H20 for the three-bed TCA
independent of ball depth. Liquid holdup in the scrubber ranged from 40 to
250 gallons.
The pressure drop and liquid holdup tests were made with 0, 5, 7.5, and 10
inches per bed (3 beds, 4 grids) of 6.5-gram, 1-5/8 inch nitrile foam spheres.
Water flow rates ranged from 19 to 44 gal/min-ft2. Superficial air velocities
varied from 7.5 to 13 ft/sec in the pressure drop tests, and from 1 to 11.4
ft/sec, in the liquid holdup tests. The cross-sectional area of the Shawnee TCA
is 32 ft2.
20.2.1 TCA Pressure Drop Test Results
The results of the pressure drop tests are presented in Figures 20-3 through
20-6. The figures show that pressure drop across the TCA beds increased with
increasing liquor flow rate and increasing air velocity.
Measured pressure drops for runs made at the same conditions (replicate
runs) were in good agreement. Measured pressure drops were the same regard-
less of whether the test conditions were reached by increasing or decreasing
the air velocities.
A comparison of Figures 20-3 through 20-5 with earlier TCA pressure drop tests
(Reference 2) using high-density polyethylene (HOPE) and thermoplastic rubber
(TPR) spheres shows good agreement (within 1/2 inch H20) at pressure drops
below 9 inches H20. However, the 5-gram HOPE spheres and the 5-gram TPR
20-5
-------
E
a
£
a
IU
2
• 1
Q 43.8gBl/min-ft2
V CT.Sgri/inin-ft2
O SUgri/mtn-ft2
D 2S.Ogri/min-ft2
A mSpl/min-ft2
0
0 v
W o
5 °
1 1 • —
i
•
T
O
0
I
D
A
1 —
• i
REPLICATES
REPLICATES
HYSTERESIS
0
1 • 1 1 1
TEST AT 37.5 gri/imn-ft2
5 1 ** i* '
o a ^ " "
0 a . n
D A ^
A
•
i i 1 1 1 1
7.0
8.0
9.0 10.0 11.0
SUPERFICIAL AIR VELOCITY. fp» * 87° F
12.0
13.0
14.000 16.000 18.000 20.000 22.000
AIR FLOW RATE, acfme 87° F
1
24.000
26.000
Figure 20-3.
TCA Bed Pressure Drop (3 Beds, 4 Grids) -
0 inches Total Bed Height of Nitrile Foam Spheres
20-6
-------
14
13 -
12x
11 •
10 •
9,
X
.s
a."
O
cc 9 .
0
1U
cc
i
£
o 8 '
u
m
7 •
6 •
5 •
4 •
7
0 43.8 gal/min-ft2 ^ REPLICATES A
V 37.5 gal/min-ft2 ^ REPLICATES
O 31.3 gal/min-ft2 O HYSTERESIS TEST AT 37.5 gal/min-ft2
O 25.0 gal/min-ft2
A 18.8 gal/min-ft2 f
x ;
0
•
* T
m
0 *
> ^^v -
TT
0 0
^
• T
0 ¥
0 T
^ or
0 v o
& o
05 n -
0 0 Q
0 n
? o n A
A
V o D A
* 0 A
0 ° D
g s
X
0 8.0 9.0 10.0 11.0 12.0 13.0
SUPERFICIAL AIR VELOCITY, fps @ 86° F
14,000 16,000 18,000 20,000 22,000 24,000 26,000
AIR FLOW RATE^cfm @ 86° F
Figure 20-4. TCA Bed Pressure Drop (3 Beds, 4 Grids) -
15 inches Total Bed Height of Nitrile Foam Spheres
20-7
-------
15
14 .
13
12 •
11 •
S
IU
9 •
7 •
5 •
7.0
0 43.9 gal/min-ft2
V 37.5 gal/min-ft2
O 31.3 gal/min-ft2
Q 25.0 gal/min-ft2
A 18.8 gal/min-ft2
D
A
f
O
a
A
8.0
14.000
1
16.000
0
¥
D
A
REPLICATES
REPLICATES
*
f
O
a
A
D
A
0
0
-*-
-4-
-I 1-
9.0 10.0 11.0 12.0
SUPERFICIAL AIR VELOCITY, fpi £ 87° F
1 I 1
13.0
18.000 20,000 22.000
AIR FLOW RATE, acfm @ 87° F
24,000
26,000
Figure 20-5. TCA Bed Pressure Drop (3 Beds, 4 Grids) -
22.5 inches Total Bed Height of Nitrile Foam Spheres
20-8
-------
IB
14
i • i i | i 1 1 ,—
0 43.8gal/min-ft2 4 REPLICATES _
V 37.5gal/min-ft2 V REPLICATES 0
O 31.3gal/min-ft2
Q 2S.O gal/min ft2
A 18.8 gal/min-ft2
13
V
0 T
I w
O
0 v
11
o: 10
CC
Q
UJ
cc
DC
CL
Q
CD
5
o n
0
V
$ f O
0
0
v a
A
0
o * °
w a A
' o ° B A
o B
H 8
H 1 1 1 1 1 1 1 h
7.0 8.0 9.0 10.0 11.0 12.0 13.0
SUPERFICIAL AIR VELOCITY, fps © 99° F
I 1 1 1 1 1 1
14,000 16,000 18,000 20,000 22,000 24,000 26,000
AIR FLOW RATE, acfm @ 99° F
Figure 20-6. TCA Bed Pressure Drop (3 Beds, 4 Grids) -
30 inches Total Bed Height of Nitrile Foam Spheres
20-9
-------
spheres experienced flooding at lower air velocities than the 6-gram nitrile
foam spheres.
Total pressure drop across the TCA was also measured. The difference between
the total pressure drop and the pressure drop across the three beds gives the
pressure drop across the scrubber entrance, scrubber exit, and mist eliminator.
This difference, as expected, was independent of liquor flow rate and increased
with increasing gas velocity; it did increase with decreasing total sphere
height. The data are presented in Figure 20-7. The lines shown in the figure
are sight averages and are accurate to +_ 0.1 inch ^0 at pressure drops below
flooding. Above flooding, the data scatter is much larger at about +. 0.5 inches
^0. This entrance/exit mist eliminator pressure drop should theoretically be
independent of sphere height, but the presence of spheres might affect the gas
distribution in the scrubber and thereby increase the pressure at the lower
bed pressure tap.
The pressure drop across the beds was measured between a point 28 inches below
the bottom bed, or 35 inches directly above the inlet duct, and a point 40 inches
above the top bed (see Figure 3-2 for a TCA schematic). The total pressure drop
was measured between a point in the inlet duct (22 feet from the scrubber inlet,
before a 90 degree turn) and a point at the scrubber outlet.
20.2.2 Correlation of TCA Bed Pressure Drop
The following equation for pressure drop across the three TCA beds and four
grids* has been fitted to 164 air/water data points:
* The grids consist of 3/8-inch diameter bars with a center-to-center
distance of 1-1/4 inches.
20-10
-------
7.0
8.0
9.0 10.0 11.0
SUPERFICIAL AIR VELOCITY, fps e 87° F
12.0
13.0
14,000
16.000 18.000 20.000 22,000
AIR FLOW RATE, acfm 0 87° F
24,000
26,000
Figure 20-7.
Difference Between TCA Total Pressure Drop
and Bed Pressure Drop
20-11
-------
AP = ( AP)NS + 0.0363 v Hs°*69 exp (0.014 L) (20-2a)
with
( AP)NS = 0.095 v exp (0.021 L) (20-2b)
where:
AP = TCA pressure drop for three beds and four grids, inches H20**
(Af>)NS = TCA pressure drop for four grids (no spheres), inches H20**
v = superficial air velocity, ft/sec at 90°F
L = liquor flow rate per unit scrubber cross-sectional area,
gpm/fr
.2;
HS = total static bed height of 1-5/8 inch diameter 6.5-gram nitrile
foam spheres, inch
The fitted ranges of the variables in Equations (20-2a) and (20-2b) are
given below:
AP =0.9 to 8.0 inches H20
v = 7.6 to 13.1 ft/sec
L = 19 to 44 gpm/ft2
H$ = 0 to 30 inches
All measured pressure drops in excess of 8.0 inches H20 were excluded from the
correlation because of incipient flooding. Any case for which Equation (20-2a)
* TCA cross-sectional area in 32 ft2.
** Measured from below the bottom grid to above the top grid.
20-12
-------
predicts a pressure drop greater than 8.0 inches H20 is likely to be in the
flooding region.
Measured TCA bed pressure drops with spheres (107 data points) and the corre-
ponding predictions from Equation (20-2a) are shown in Figure 20-8.
Equation (20-2a) accounts for 96 percent of the variation in the data with
a standard error of estimate of 0.24 inch H20.
Measured TCA bed pressure drops without spheres (57 data points) and the
corresponding predictions from Equation (20-2b) are shown in Figure 20-9.
Equation (20-2b) accounts for 90 percent of the variation with a standard
error of 0.15 inch H20.
For the overall (spheres and no spheres) data, Equation (20-2a) explains
99 percent of the variation, with a standard error of 0.21 inch H20.
The form of Equation (20-2a) was chosen to meet boundary constraints on
pressure drop for limiting values v, L, and H$. For example, the equation
predicts zero pressure drop at zero gas velocity.
20.2.3 TCA Liquid Holdup Test Results
Liquid holdup was measured as the difference in effluent hold tank level
between down conditions (slurry pump off) and test conditions. The measured
test data are given in Table 20-1. As measured, the liquid holdup included
piping and downcomer holdup as well as holdup in the scrubber. To estimate
piping and downcomer holdup, tests were made with zero air velocity and
no spheres. Holdup was approximately 264 gallons at these minimum conditions,
However, even with zero velocity and no spheres, there is some holdup in
20-13
-------
8
O, 7-.
O
cc
Q
UJ
DC
D
6 -•
cc
Q.
O
Ul
OQ
O
UJ
5 -•
O
UJ
£ 4
MEASURED BED PRESSURE DROP, in.
Figure 20-8.
Comparison of Experimental Data and Predicted
Values of Bed Pressure Drop - Three-Stage TCA
with 1-5/8 inch Dia., 6.5 gram Solid Nitrile
Foam Spheres
20-14
-------
1.0 1.5 2.0 2.5
MEASURED BED PRESSURE DROP, in, H2O
3.0
' Figure 20-9. Comparison of Experimental Data and Predicted
Values of Bed Pressure Drop - Three-Stage TCA
without Spheres
20-15
-------
Table 20-1
DATA FROM LIQUID HOLDUP TESTS ON THE TCA
Total
Bed Height,
in
0
0
0 '
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
15
IS
15
15
15
15
15
15
15
15
15
15
22.5
22.5
22.5
22.5
22.5
22.5
22.5
22.5
22.5
22.5
22.5
22.5
22.5
22.5
2Z.5
22.5
30
30
JO
30
30
30
30
30
30
30
30
30
30
30
30
30
Air
Flow Rate,
acfm
0
0
0
0
0
14,400
17,900
21,600
14. 400
17, 900
21,600
14,400
17.900
21,600
14,400
21,600
14,400
17,900
21,600
14,300
17,850
21,500
14,300
17,850
21,500
14, 300
17,900
21,500
14, 300
17.850
21.500
14, 300
17,900
21.500
14,300
17.900
21,500
14.300
17,900
21.500
14.300
21.500
14,300
17,900
21,500
14, 300
21,500
14, 700
18,000
22.000
14, 700
18. 000
22, 000
14, 700
18. 000
22,000
14, 700
22, 000
14, 700
18. 000
22, 000
18, 000
14, 700
Velocity,
fps
0
0
0
0 '
0
7.5
9.3
11.2
7.5
9.3
11.2
7.5
9.3
11.2
7.5
;i.2
7.5
9.3
11.2
7.5
9.3
11.2
7.5
9.3
11.2
7.5
9.3
11.2
7.5
9.4
11.2
7.5
9.3
11.2
7.5
9.3
11.2
7.5
9.3
11.2
7.5
11.2
7.5
9.3
11.2
7.5
11.2
7.6
9.5
11.4
7.6
9.5
11.4
7.6
9.5
11.4
7.6
11.4
7.6
9.5
11.4
9.5
7.6
Liquor Flow Rate,
gal/min-ft2
18.8
28.1
37.5
43.8
43.8
18.8
18.8
18.8
28.)
28.1
28.1
37.5
37.5
37.5
37.5
37.5
43.8
43.8
43.8
18.8
18.8
18.8
28.1
28.1
28.1
37.5
37.5
37.5
43.8
44.4
43.8
18.8
18.8
18.8
28. 1
28.1
28.1
37.5
37.5
37.5
37.5
37.5
43.8
43.8
43.8
43.8
43.8
18.8
18.8
18.8
28.1
28.1
28.1
37.5
37.5
37.5
37.5
37.5
43.8
43.8
43.8
43.8
43.8
Liquid Holdup,
gal
264
264
264
312
312
240
264
282
276
264
294
288
312
288
288
252
288
300
288
282
300
324
288
336
336
363
420
408
390
384
408
288
264
324
324
312
372
324
348
408
348
384
288
384
408
360
408
312
312
336
348
360
396
360
420
408
396
468
384
420
492
432
408
* Difference In tank level between having the pump on
and the pump off. Includes holdup in the piping and downcomer.
20-16
-------
the scrubber as the liquid travels the length of the scrubber. Assuming the
liquid falls directly from the inlet spray header to the outlet, the minimum
scrubber holdup was estimated as 64 gallons by taking the cross-sectional area
of the feed pipe (7.9-inch ID) multiplied by the distance between the inlet
header and the outlet (24.7 feet). Therefore, the piping and downcomer holdup
was estimated as 200 gallons (264 gallons holdup with zero air rate and no
spheres, minus 64 gallons estimated scrubber holdup at this minimum condition).
This estimated piping and downcomer holdup was subtracted from the measured
liquid holdup in all tests to estimate the net scrubber holdup.
Scrubber holdup ranged from 40 to 250 gallons over the range of conditions
studied. Volumes were accurate to +_ 25 gallons. Scrubber holdup under typical
operating conditions and the effect on holdup of changing major conditions
are summarized below:
Typical operating conditions of 11.2 ft/sec,
37.5 gal/min-ft2, and 15 inches total sphere
height
Effect of air velocity increase from 7.5 to
11.2 ft/sec
Fffect of liquor rate increase from 18.8 to
37.5 gal/min-ft2
Effect of sphere height increase from 0 to
30 inches
210-gallon scrubber
holdup
Holdup increase by
20 to 60 gallons
Holdup increase by
40 to 120 gallons
Holdup increase by
70 to 110 gallons
20-17
-------
Section 21
ANALYSIS OF SLURRY LIQUOR FOR SELECTED TRACE METALS,
NITROGEN, AND ORGANIC CARBON
21.1 ANALYSIS PROGRAM
On July 2, 1977, during forced-oxidation Run 856-1A on the venturi/spray
tower system, the sulfite oxidation decreased from 95 percent to a low of
62 percent for no apparent reason. The oxidation then gradually increased
to 95 percent by July 5. During the same period, a similar decrease occurred
in the TCA system, Run 802-2A, from 45 to 14 percent oxidation. No similar
unexplained decrease in oxidation efficiency was observed throughout the
forced-oxidation test program.
Since both systems reacted nearly simultaneously, it is postulated that a
contaminant was introduced with the flue gas or makeup water. This contami-
nant either inhibited oxidation directly, perhaps by acting as a radical
scavenger, or inhibited oxidation indirectly by reducing the catalytic acti-
vity of certain trace metals.
As a result of this phenomenon, a program of routinely analyzing trace metals,
nitrogen, and organic carbon was initiated. It is hoped that if the condi-
tions that inhibited oxidation recur, these tests will help identify the
source.
21-1
-------
Not only can the analysis of trace metals help identify an oxidation anomaly,
but it can also document the effects that the scrubber environment has on
the materials from which the scrubber system is constructed.
TVA's Power Service Center Laboratory in Chattanooga, Tennessee, performed
the nitrogen and carbon analyses. TVA's General Analytical Laboratory in
Muscle Shoals, Alabama, performed the trace metal analyses.
21.2 TRACE METALS
Table 21-1 presents the results of the analysis of the venturi inlet and the
TCA outlet (see Figure 2-1) slurry liquor samples for selected trace metals
as reported by TVA. These sample locations were picked so the effects of
natural oxidation could most easily be discerned.
Metals with the highest catalytic characteristics for sulfite oxidation were
selected. Because the venturi/spray tower operated with a tighter water
balance than did the TCA system, concentrations were generally higher in the
venturi/spray tower system.
21.2.1 Cobalt (Co), Chromium (Cr), Copper (Cu)
The concentrations of cobalt, chromium, and copper were relatively low and
invariant. In January 1978, these components were dropped and replaced
with titanium and vanadium.
21-2
-------
Table 21-1
RESULTS OF SLURRY LIQUOR ANALYSIS
SYSTEM
V/ST
V/ST
V/ST
V/ST
V/ST
V/ST
All V/ST
TCA
TCA
TCA
TCA
TCA
All TCA
TYPE OF RUN
Alkali
L
LS
LS
L
LS
LS
-
LS
LS
LS
LS
LS
-
f. , 1 Fly Ash
Ox1dat1onvi;| Loading
Yes Low
Yes Low
Yes High
Yes High
Yes High
Yes'2' High
Yes
Yes'3' High
No High
Yes'5' High
No High
No High
-
TRACE MFTAI <;
Co Cr
0.2 0.06
0.2 0.02
0.7 0.02
0.4 0.01
-
-
0.4 0.02
0.2 0.03
0.1 ' Tr'4'
0.1 0.01
-
-
0.1 0.02
Cu
0.03
0.02
0.08
0.08
-
-
0.05
0.04
Tr
0.02
-
-
0.02
Fe 1 Mn
0.02 10
0.2 28
0.2 62
0.5 36
23 55
2.1 53
4.7 42
5 16
0.3 3
1.6 7
6 3
26 10
7 7
Mo N
1 0.
1 0.
8 1.
5 1.
6 1.
9 1.
5 1.
4 0.
6 0.
4 0.
6 0.
7 0.
5 0.
1 T1 V
5
8
4
2 0.5 0.5
2 0.4 1.3
2 0.2 0.9
1 0.4 0.9
5
3
3 0.1 0.4
3 0.4 1.5
8 0,3 2
4 0.3 1.5
Nox
TOC p
H ADDITIVE
8.0
22
50
59
42
27
45
12
2
14
21
26
15
9 6.
D
3 5.9
5 7.8
3 6.0 MgO
6 5.3 MgO
5 6.6
1.5 4.6
2 5.3
2 5.3
3 5.5 MgO
4 5.2 MgO
2 5.2
ro
i
CO
Notes:
(1) Two-stage forced oxidation
(2) Bleed stream oxidation
(3) Air eductor used for oxidation
(4) Tr = Trace
(5) Air sparger used for oxidation
-------
21.2.2 Iron (Fe)
The iron concentrations averaged 7 ppm in the TCA liquor and 5 ppm in the
venturi liquor. Iron solubility increases with decreasing pH. This explains
why iron concentrations were higher in the TCA, where the pH averaged 5.2,
than in the venturi, where the pH averaged 5.5.
Iron concentrations in both systems were higher in runs made using MgO addi-
tive. Two effects influenced this. MgO runs were made at lower pH than
most other runs during the period covered in this study (see Table 21-1);
the presence of Mg++ enhances iron solubility via the common-ion effect.
21.2.3 Manganese (Mn)
The concentration of manganese in the venturi averaged 42 ppm, more than five
times the 7 ppm averaged by the TCA. Manganese concentrations were signifi-
cantly higher for runs with oxidation than for runs without oxidation,
probably because MnSOg is insoluble.
21.2.4 Molybdenum (Mo)
The average concentration of molybdenum was exactly the same, 5 ppm for both
systems. Molybdenum concentrations averaged 1 ppm for low fly ash runs and
5 ppm for high fly ash runs.
21.2.5 Nickel (Ni)
The average concentrations were higher in the venturi than in the TCA and
seemed to follow trends similar to manganese, where concentrations were higher
21-4
-------
for runs with oxidation. The most probable explanation is that nickel sulfite
is insoluble and the sulfate is soluble.
21.2.6 Titanium (Ti)
Measured concentrations of titanium were relatively low (less than 0.5 ppm)
and invariant.
21.2.7 Vanadium (V)
Higher concentrations of vanadium were recorded in the TCA (1.5 to 2 ppm),
with lower pH and no oxidation, than in the venturi (about 1 ppm).
21.3 NITROGEN OXIDES AND TOTAL ORGANIC CARBON
When the trace metals analysis program was initiated, it was decided that
the study should include the analysis of other chemical species that could
inhibit sulfite oxidation. Two classes of compounds present in the scrubber
environment that could contribute to such an effect are nitrogen oxides (NOX)
and organic carbon compounds.
The concentration of nitrogen oxides in the venturi averaged 45 ppm, three
times the 15 ppm recorded for the TCA. The tighter material balance of
the venturi is the most likely explanation for this.
Total organic carbon levels measured for both scrubbers and the raw water
were relatively low and invariant, usually less than 6 ppm.
21-5
-------
21.4 FLUE GAS OXIDANT MONITORING PROGRAM
Nitrogen oxide (NO) and oxygen (02) in the flue gas were analyzed for two
weeks during October 1976. Data collection was limited due to difficulties
with equipment and calibration problems. NO concentrations were measured
at about 400 ppm (+ 50 ppm) for both systems. No other data were obtained
from this program.
21.5 SUMMARY
The primary purpose for initiating the analysis program for trace metals,
nitrogen, and organic carbon was to attempt to identify those constituents
of the scrubber environment which might have an effect on sulfite oxidation.
No oxidation excursions occurred during this period, and thus no results
toward that end were obtained. However, the data obtained have significantly
contributed to an understanding of scrubber chemistry, particularly with re-
gard to the effects of pH. Continued routine analysis of the slurry liquor
will be performed for both systems, and the results will be reported in
future publications.
21-6
-------
Section 22
WASTE SOLIDS DEWATERING AND CHARACTERIZATION
The results of slurry solids dewatering tests at the Shawnee Test Facility
are presented in this section. These tests were done to monitor operation
of the dewatering equipment and to evaluate the settling and dewatering
characteristics of the discharge slurries.
The testing of two types of dewatering equipment is discussed. First, the
results of field testing of an inclined plate settler, the Lamella Gravity
Settler by Parkson Corporation, are presented; then, the results of tests
using a hydroclone as a means of separating gypsum from unreacted alkali and
sulfite are given. The hydroclone used was the Dorrclone P50A by Dorr-Oliver.
Results are given for routine cylinder settling and funnel filtration tests
performed during this reporting period on both the venturi/spray tower and TCA
systems. Also presented are filter leaf test results on slurries obtained
from tests with and without forced oxidation.
Two methods for characterizing the particle size distribution of slurry solid
were tested: the hydrometer method and pipette method. These tests were
applied to oxidized slurries and also to slurry streams separated in the
hydroclone testing.
22-1
-------
Scanning electron micrographs (SEM) were taken of various slurries to deter-
mine whether any significant particle size changes had occurred during various
testing operations.
Finally, a study of gypsum crystallization was done to determine the extent
to which discharged solids could be improved by increasing the crystal size
and reducing the fines content.
22.1 THE LAMELLA GRAVITY SETTLER THICKENER TEST PROGRAM AND RESULTS
Testing of a Lamella inclined plate settler, a compact sludge dewatering device
manufactured by the Parkson Corporation, was conducted at the Shawnee Test
Facility from July through November 1977. The tests were designed to determine
the maximum clarification and thickening capacity of the unit under different
test run conditions, such as the type of slurry feed, the feed rate, and the
addition of flocculant. Potential fouling problems were also identified.
22.1.1 Equipment Description
The Lamella Gravity Settler Thickener, Parkson Unit LGST No. 3, was 20 feet
tall and consisted of two tanks constructed of 316 stainless steel. The
upper rectangular tank was a high-rate gravity settler and contained a series
of parallel FRP Lamella plates inclined at 55 degrees (plate pack). The
lower circular tank was 4 feet in diameter and 10 feet in height and served
as the thickener; it contained a picket fence type sludge rake and an underflow
outlet. The total liquid capacity was approximately 1,000 gallons.
Each of the inclined plates in the plate pack had an equivalent settling area
22-2
-------
equal to its projection onto a horizontal plane. Each square foot of this
projected area was equal in settling capacity to a square foot of area in a
conventional settler. The unit being tested had a total settling area of
115 ft2, 57.5 ft2 of which was utilized as clarification area.
The settler thickener is shown in Figure 22-1. The slurry enters from a bottom-
less feedbox into the plate pack from the side, about halfway from the bottom
of the plate. Liquor flows upward, exiting at the top of the rectangular tank
through flow distribution orifices. This forces the slurry to be evenly dis-
tributed over all of the plates. The solids settle out onto the plate surfaces
and slide downward into the circular thickening tank. The thickened sludge
is then pumped out of the unit via a Moyno underflow pump.
22.1.2 Test Plan
A series of tests involving the Lamella Settler was completed using the
following: lime or limestone slurry, flue gas with low or high fly ash load-
ing, and oxidized or unoxidized slurry. The aim of the tests was to determine
for each type of slurry the maximum possible feed rate that would maintain
the underflow solids concentration at greater than 40 weight percent and the
overflow at less than 0.5 weight percent. Additionally, the solids concentra-
tion profile in the Lamella circular hold tank as a function of feed rate was
measured. Frequency of plugging of the plates and lines and the amount of
maintenance needed were also noted.
22-3
-------
*-OVER FLOW
FEED
LOW DISTRIBUTION
ORIFICES
PICKET FENCE
TYPE SLUDGE
SCRAPER
MOTOR
LAMELLA
PLATE PACK
Figure 22-1. The Lamella Gravity Settler Thickener
22-4
-------
22.1.3 Size Comparison of a Lamella Settler and a Conventional Clarifier
A comparison of the Lamella Settler with a conventional clarifier of equivalent
settling area, based on land requirements, shows that a Lamella Settler requires
about one-sixth the land area needed for an equivalent conventional clarifier.
The capacity test data from the Lamella Settler with oxidized slurry having
high fly ash loading was used for scaleup from the 10-MW Shawnee Test Facility.
Sizing estimates showed that 5,200 ft2 of settling area would be required for
a 500-MW plant. This is equivalent to an 80-ft diameter conventional clarifier
which requires 5,200 ft2 of land area. Based on size estimates obtained
from the Parkson Corporation, the manufacturers of the Lamella Settler, a
Lamella Settler Model 10400/55 with a 5,200 ft2 settling area should require
only 800 ft2 of land area, a savings of 4,400 ft2 over a conventional clarifier.
22.1.4 Summary of Tests
Below is a summary of the testing of the Lamella Settler, indicating the type
of slurry used, the run number, and the dates of the testing. In subsequent
subsections, major conclusions of the testing are discussed. Data from the
tests are presented in Appendix L.
Slurry Type Subsection Runs Dates of Test
Oxidized limestone 22.1.4.1 803-2A & 804-2A July 12 - July 20
high fly ash loading 818-1A Oct. 26 - Oct. 31
15 weight percent 819-1A Oct. 31 - Nov. 29
solids
Oxidized lime 22.1.4.2 859-1A July 26
low fly ash loading
15 weight percent
solids
22-5
-------
Oxidized limestone 22.1.4.3 809-1A Aug. 15 - Aug. 17
low fly ash loading
15 weight percent
solids
Unoxidized limestone 22.1.4.4 714-2A Aug. 31 - Sept. 9
high fly ash loading 714-2B
8 & 15 weight percent
solids
flocculant testing
Oxidized lime 22.1.4.5 862-1A Oct. 13 - Oct. 18
high fly ash loading
8 & 15 weight percent
solids
flocculant testing
Unoxidized limestone 22.1.4.6 717-2A Oct. 21 - Oct. 24
high fly ash loading
15 weight percent
solids
22.1.4.1 Oxidized limestone slurry with high fly ash loading
Several tests were done with this type of slurry during Runs 803-2A, 804-2A,
818-1A, and 819-1A.
• During Runs 803-2A and 804-2A, a feed rate of 17 gpm gave an over-
flow solids concentration of 0.32 weight percent (compared to less
than 0.1 weight percent normally obtained for the TCA clarifier)
and an underflow of 38 weight percent. However, at feed rates of
20 to 22 gpm, the overflow solids concentrations were 0.82 and 1.9
weight percent, respectively, exceeding the desired value of 0.5
weight percent or less. At the end of the testing, the entire
unit was found to be clean except for minor soft solids in the
weir box.
• A cylinder settling test on slurry from Run 818-1A indicated that
80 gpm would be the maximum feed rate to the Lamella Settler over-
flow solids. However, 20 gpm was the highest feed rate that yielded
satisfactory overflow solids of 0.3 to 0.6 weight percent. Feed
rates of 21.5 and 24 gpm had overflow solids of 0.8 and 1.2 weight
percent, respectively. The underflow solids concentration at 20 gpm
22-6
-------
was 34 to 57 weight percent.
• Slurry from Run 819-1A at a feed rate of approximately 9 gpm
yielded satisfactory results in this one-month reliability
test. The overflow solids concentration varied from 0.05 to
0.18 weight percent. The underflow solids varied from 35 to
54 weight percent.
• Throughout the test, the Lamella Settler operated smoothly;
there were no plugging problems or erratic behavior. Visual
inspection revealed that there was no scale buildup.
22.1.4.2 Oxidized lime slurry with low fly ash loading
• With slurry from Run 859-1A at a feed rate of 10 gpm, the under-
flow line plugged. This plugging was due to the high settling
rate of the essentially pure gypsum particles. To remedy this
problem, the diameter of the underflow line was increased and a
freshwater purge line was added.
22.1.4.3 Oxidized limestone slurry with low fly ash loading
t With slurry from Run 809-1A at a feed rate varying from 6 to 12
gpm, the overflow solids remained below 0.5 weight percent, rang-
ing from 0.07 to 0.45 weight percent. The underflow solids varied
from 28 to 41 weight percent.
• Throughout the testing, the settler performed well, with only minor
plugging of the underflow line. This plugging was corrected by
water purging the line.
22.1.4.4 Unoxidized limestone slurry with low fly ash loading
Both laboratory cylinder settling tests and Lamella Settler tests were done on
slurry from Run 714-2A and Run 714-2B. Laboratory cylinder settling tests on
both 15 and 8 weight percent slurries were done with the addition of 0 to 25
22-7
-------
ppm flocculant, American Cynamid's Superfloc 1204, and Nalco's nontoxic floc-
culant, Nalco 8861. Subsequently, the Lamella Settler was run with slurries
containing 8 weight percent solids without flocculant. The following results
were obtained.
• For 15 weight percent unflocculated slurry, the maximum feed
rate to the Lamella Settler for satisfactory overflow was 2 gpm.
• For 8 weight percent slurry flocculated with Nalco 8861, the max-
imum feed rate to the Lamella Settler for satisfactory overflow
was 44 gpm, though this may have been influenced by the limiting
capacity of the underflow pump (about 7 gpm).
• With no flocculant addition, slurries with 8 weight percent solids
had a higher mass loading than 15 weight percent slurries (2.7 to
5.7 times as great). This was a result of less hindered settling
in the more dilute slurry.
• The increase in mass loading rates of 15 weight percent slurries
by flocculant addition was small (1.5 to 2.5 times as great).
t The increase in mass loading rates of 8 weight percent slurries
by flocculant addition was large (28 times as great).
0 Nalco 8861 had a greater effect on mass loading rate of 8 weight
percent slurry than the Superfloc 1204 (3.3 to 4.1 times more
effective).
• Visual inspection of the Lamella Settler showed that the system
was free of solids buildup.
22.1.4.5 Oxidized lime slurry with high fly ash loading
0 A laboratory cylinder settling test for this slurry with no floc-
culant addition at 15 weight percent solids, from Run 862-1A, indi-
cated that the maximum feed rate to the Lamella Settler would be
10.2 gpm.
22-8
-------
• Slurry of 15 weight percent solids with no flocculant addition
at a feed rate of 10 gpm gave a satisfactory overflow of 0.07
to 0.51 weight percent solids. The underflow was 34 to 49 weight
percent solids.
• A laboratory cylinder settling test of this slurry with 9 ppm
Nalco 8861 flocculant and diluted to 8 weight percent indicated
the maximum feed rate to be 40 gpm.
• Slurry of 8 weight percent solids with Nalco 8861 addition at a
feed rate of 40 gpm gave overflow solids of 0.03 to 0.07 weight
percent and underflow solids of 41 to 52 weight percent.
22.1.4.6 Unoxidized limestone slurry with high fly ash loading
t With slurry from Run 717-2A at 15 weight percent solids, a
cylinder settling test indicated a maximum feed rate of 2.3 gpm.
In a test at 2.3 gpm feed rate, the solids concentration of the
overflow increased steadily from 0.6 to 6.8 weight percent during
the run. This run was originally intended for a long-term Lamella
reliability run. However, the run was discontinued because of the
high solids in the overflow and the inability of the pump to con-
trol flow below 2 gpm.
22.1.5 Solids Concentration Profiles
Solids concentration profiles in the Lamella Settler circular hold tank were
determined throughout the testing by taking side drawoff samples at nine levels
on the tank. The profiles for all the tests were fairly uniform. The data are
summarized in Appendix L.
Generally, the concentration profile was uniform and similar to the inlet solids
concentration for the oxidized slurry, and was uniformly near the underflow
concentration for the unoxidized slurry. It is postulated that the oxidized
slurry solids settle quickly and thus do not have sufficient time to diffuse
22-9
-------
across the thickener tank to the sampling valves, whereas the unoxidized
slurry solids settle slowly and consequently and have a better chance of
diffusing across the thickener tank to the sampling valves.
22.1.6 Conclusions
t The Lamella Settler compared with a conventional clarifier
of equivalent settling area requires about one-sixth the
land area, and has correspondingly smaller slurry inventory.
t Oxidized slurry at 15 weight percent solids has mass loading
rates anywhere from 5 to 10 times that of unoxidized slurry.
• Flocculant addition greatly increased the mass loading rate
of both oxidized and unoxidized slurry.
• After initial modifications the Lamella Settler operated free
from plugging, scaling, and mechanical problems.
• In comparison to conventional clarifiers, the Lamella Settler
is superior for clarification but does not appear to offer an
advantage for thickening.
22.2 CYLINDER SETTLING AND FUNNEL FILTRATION TESTS
Cylinder settling tests and funnel filtration tests were conducted on a routine
basis to monitor the dewatering characteristics of the slurries and to detect
any gross changes. The initial settling rate indicates the rate at which the
solids fall during the unhindered portion of settling. Ultimate settled
solids represent the highest achievable solids concentration in a thickener
or settling pond. Funnel test cake solids represent those achievable on a
filter type dewatering device.
22-10
-------
Data for the period from January 1977 through May 1978 are presented in
Appendix L, including alkali type, fly ash loading, pH, and MgO addition.
Results from earlier test blocks are presented in the Third Progress Report
(Reference 4).
Table 22-1 summarizes the settling data obtained, including the effect of
oxidation, fly ash loading, alkali, and additive concentration on initial
settling rate, ultimate settled solids, and funnel test cake solids.
The results during the testing period are similar to previous results. The
benefits of forced oxidation on the settling and dewatering characteristics
of the sludge are clearly evident. For runs without MgO addition, the initial
settling rate for an oxidized slurry (0.7 to 1.2 cm/min average range) was
much higher than for unoxidized slurry (0.2 to 0.4 cm/min average range),
s
regardless of whether the slurry had high or low fly ash loading. Again ex-
cluding magnesium runs, the average ultimate settled solids and funnel test
solids ranged from 70 to 76 weight percent for the oxidized slurry, whereas
for unoxidized slurry they ranged from 40 to 57 weight percent.
The initial settling rate of unoxidized slurry is probably limited by the
sulfite particles. In the case of oxidized slurry, the rate is probably
limited by the fly ash. Even in the case of slurry with low fly ash loading,
fly ash constitutes about one weight percent of the solids.
The addition of magnesium primarily affected the initial settling rate. For
oxidized slurry with the two-loop mode of oxidation and with 8000 ppm effective
magnesium ion concentration, the average initial settling rate, the ultimate
settled solids concentrations and the funnel test cake solids concentrations
were all slightly lower than those for oxidized slurry without magnesium.
22-11
-------
Table 22-1
SUMMARY OF DEWATERING CHARACTERISTICS
OF SHAWNEE SYSTEM BLEED
Fly Ash
Oxidation Loading AUal
Yes High LS
Yes High LS
Yes H1qh LS
Yes High LS
Yes High L
Yes Low LS
"es LOW L
- No High LS
No High LS
No High LS
No Nigh L
No High L
No Low LS
No Low L
Initial Settling Rate, cm/rain
1 Oxidation Mode Avg. 1 Rihge"
1-loop 1.1 0.6-1.3
2- loop 1.2 1.0-1.4
Bleed Stream 0.4 0.3-0.6
2- loop 0.8 0.2-1.2
2-loop 1.0 0.8-1.2
2- loop 0.9 0.6-1.2
2-loop 1.2 0.4-2.4
0.2 0.1-0.5
0.2 0.1-0.4
0.1 0.0-0.1
0.2 0.2-0.5
0.8 0.2-1.2
0.2 0.1-0.5
0.4 0.1-0.9
Ultimate Sett
Avg.
74
72
71
66
73
74
70
54
45
41
50
42
43
40
Ing Solids, wW
Range
67-84
62-86
61-84
46-73
61-85
61-87
60-81
41-67
30-60
32-46
48-66
31-52
33-54
30-55
Funnel Test Cake Solids. wtX Sli
Avg. 1 Range So
76 73-80 15
72 65-88 15
73 71-76 15
70 46-76 15
71 64-78 15
73 64-82 15
76 64-83 15
57 48-66 15
57 45-64 15
55 47-69 15
53 51-55 15
52 43-63 8
50 41-59 15
45 40-50 8
irry Effective Mg^
Ids Concentration, ppm
0
0
5000
8000
0
0
0
0
5000
9000
0
>000
0
0
PO
no
i
i—»
ro
Note: Values for forced oxidation runs are only from data where solids oxidation
1s greater than or equal to 90 percent.
-------
For bleed stream oxidized slurry, the initial settling rate was about half
that of other oxidized slurries. The ultimate amount of settled solids and
funnel test cake solids were in the same range as with other oxidized slurries.
For unoxidized slurry, the effect of magnesium was more pronounced. For
a limestone slurry with high fly ash loading, the average initial settling
rate was 0.20 cm/min with no magnesium and 0.05 cm/nrin with 9000 ppm magnesium,
a decrease of a factor of four. The amount of ultimate settled solids was
somewhat lower for the slurries with magnesium; the amount of funnel test
cake solids was about the same.
One factor affecting the initial settling rate but not the ultimate amount of
settled solids when magnesium is present is the increase in liquor viscosity
and density due to the increased total dissolved solids.
22.3 FILTER LEAF TESTS
Bench-scale filter leaf tests were conducted to identify promising filter
cloths for future evaluation on the rotary drum filter and to determine the
filtration characteristics of sludges generated under different operating
conditions.
All tests were conducted with a standard filter leaf having a cloth area of
0.10 ft . Rotary drum filter operation was simulated by immersing the filter
leaf, under vacuum, into gently agitated slurry for a time equivalent to the
filter cake form time and then removing the leaf from the slurry and draining
it under vacuum for an equivalent filter cake drying time.
22-13
-------
22.3.1 Cloth Evaluation Tests
Samples of 23 cloths and three previously blinded ones (saved after removal
from the drum filter) were evaluated by a filter leaf procedure designed to
simulate the Shawnee drum filter operating at about one rpm with a submergence
(area basis) of 33-percent. The 60-second complete filtration cycle was divided
into 20 seconds for cake form time, 34 seconds of drying time, and 6 seconds of
cake discharge time. No attempts were made to adjust cycle time to optimize
individual cloth performance.
A summary of the cloth properties is presented in Table 22-2. The cloths had
an air permeability range of 0.8 to 100 cfm/ft2 (air permeability is defined
as the air flow through the cloth, in cfm/ft2, under a pressure differential
of 1/2 inch 1^0 across the cloth). Evaluations were carried out on two types
of unoxidized clarifier underflow slurry with an initial solids concentration
of 20 weight percent: lime slurry with high fly ash loading from Run VFG-1P
(see Table 22-3) and limestone slurry with low fly ash loading from Run 706-2A
(see Table 22-4). Cloths numbered 8, 11, and 14 were of the types normally
used on the Shawnee rotary drum filter. None of the other cloths tested showed
significantly improved dry cake production rate or cake solids concentration.
For the lime slurry with fly ash, the filtrate rate increased with air permea-
bility (see Figure 22-2) up to a maximum filtrate rate of 53 gal/hr-ft2 for
cloths with permeability of 40 cfm/ft2 or higher. Cake production ranged from
n
73 to 116 Ib/hr-ft , generally increasing with cloth permeability. For the
limestone slurry without fly ash, filtrate rate (35 to 45 gal/hr-ft2) and cake
production (92 to 145 gal/hr-ft2) were both independent of cloth type, indica-
ting that cloth resistance was overshadowed by cake resistance.
22-14
-------
Table 22-2
CLOTH PROPERTIES
ro
ro
' t
01
Cloth
Numbe r
1
2
3
4
5
o
' 12 t
81
9
10(2I
11 '
12
13(2l
14( '
15
lit
17
18
19
20
21
22
23
Vend,,'"
E
E
A
L
E
A
E
TFI
A
E
L,
A
E
A
A
A
E
E
E
L
A
E
E
ID Number
NY-317F
POPR-925F
STE-GODO-AJO
0-2107
NY-330
XB-HG6D4A-J8
NY-384F
9162
WNH-G2E7-AG8
POPR-851
7512-SHS
ST-EF8D8A-G1
NY-304
STE-F9D8-HJO
WNG-NOG6-PH3
XBH-H4G8-IFO
POPR-873
POPR-852F
NY-529F
4048 -SHS
ST-EF8E6H-G2
PO-801HF
POPR-859
Air Permeability.
cim/tt1
0.8
1.0
2.9
3. 0
3.2
5.0
6.0
7.0
7,5
8. 3
10. 0
10. 0
14.0
20.0
24.0
24.0
30. 0
32.5
39.65
40. 0
48.0
50.0
100. 0
Material
Nylon
Polypropylene
Olefin
Polypropylene
Nylon
Polyester
Nylon
Polypropylene
Nylon
Polypropylene
Polypropylene
Olefin
Nylon
Olefin
Nylon
Polyester
Polypropylene
Polypropylene
Nylon
Polyester
Olefin
Polyethylene
Polypropylene
Weave
2/2 Twill
2/1 Basket
1/1 Plain
Plain
--
1/1 Plain
1/1 Plain
Plain
1/1 Plain
1/1 Plain
Plain
1/1 Plain
1/1 Plain
2/2 Twill
1/4 Satin
3/1 Twill
1/1 Plain
2/2 Twill
2/2 Twill
Twill
2/2 Twill
1/1 Plain
2/2 Twill
Yarn
multi- filament
irmlti -filament
multi- filament
--
multi-filament
multi-filament
multi- filament
mutti- filament
multl- filament
..
multi-filament
spun staple
multi-filament
multi- filament
multi- filament
mono- filament
multi-filament
multi-filament
.-
multi- filament
mono-filament
mono -filament
Thread
Count
220x 128
104 x 34
60 x 30
48 x 26
--
66 x 34
79 x 40
59 x 32
62 x 47
58 x 32
58 x 28
58 x 38
54 x 28
58 x 38
1 30 x 66
74 x 68
116 x 34
64 x 34
144 x 54
156 x 60
58 x 46
11Z x 48
68 x 30
Cloth
Finish
Heat Set
Heat Set
Scoured/Heat Set
--
__
Heat Set
-_
_-
Greige
Scoured/Heat Set
--
Grey
--
_-
--
Calende red
Heat Set
Heat Set
Scoured /Heat Set
Heat Set
Greige
Cloth Weight
ounces/yd
3. 8
8. 0
9. 0
14. 0
-•
9. 8
9. 0
S. 0
6. 8
8. 0
8." 5
6. 1
11. 6
°. 0
8. 3
5.0
6.2
9. 3
8. 0
9.2
6.2
11.0
8.5
(1) E = Evmco, A = Ametek, L - Lamport, TFI = Technical Fabricators.
(2) Sample of cloth normally used on the rotary drum filter.
-------
Table 22-3
FILTER LEAF TEST RESULTS ON LIME SLURRY
WITH HIGH FLY ASH LOADING151
Cloth
Number
1
2
3
4
5
6
7...
c(3)
9
10,..
n(3)
12
13...
14<3>
15
16
17
18
19
20
21
22
23
Air Permeability,
cfm/ft2
0.8
1. 0
2.9
3.0
3.2
5.0
6.0
7. 0
7,5
8.3
10. 0
10.0
14. 0
20. 0
24.0
24.0
30.0
32.5
39.65
40.0
48.0
50. 0
100. 0
Dry Cake Production,
lb/ft2-hrW
74
79
78
75
73
75
74
73
111
116
98
110...
_.' '
112
101
113
92
95
94
110
112
111
111
Filtrate Rate,
gal/ftz-hf *
35
33
32
35
37
43
32
46
46
46
42
48...
__^ '
51
56
44
51
49
53
52
52
56
52
Cake
Thickness, In.
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
3/8
3/8
5/16
3/8_.
__
3/8
11/32
3/8
5/16
5/16
5/16
3/8
3/8
3/8
3/8
Cake Solids,
65.6
57.7
60.1
64.8
66.6
64.0
65.6
66.9
65.1
59.1
61.8
^4' 8(2)
-- *
64.6
66.0
63.7
66.5
62.3
63.4
66.4
64.6
66.0
65.9
Filtrate Solids,
wt%
0. 00
0. 01
0. 01
0. 01
!o. 01
0. 05
0. 02
0.01
0. 02
0. 10
0. 02
0. 03 ..,.
--
0. 03
0. 13
0. 01
0.27
0. 07
0.11
0.14
0.23
0.25
0.36
ro
ro
i
en
1. Refer to Table 22-2.
2. Ran cloth on limestone slurry only.
3. Sample of cloth normally used on the rotary drum filter.
4. Based on a sixty-second cycle.
5. From Run VFG-1P.
-------
Table 22-4
FILTER LEAF TEST RESULTS ON LIMESTONE
SLURRY WITH LOW FLY ASH LOADING fe)
Cloth (
Number
1
2
3
4
5
6
7
11 (4)
12
13 (4)
14 l '
15
16
17
18
19
20
21
22
23
Air Permeability,
cfm/ft2
0. 8
1.0
2.9
3. 0
3. 2
5.0
6. 0
7. 0
7.5
8. 3
10.0
10.0
14.0
20. 0
24. 0
24.0
30.0
32.5
39.65
40.0
48.0
50. 0
100. 0
Dry Cake Production, Filtrat<
Ib/ft2-hr<5) gal/ft'
; Rate, Cake
'-hi«" Thickness, in.
145 36 5/8
131 39 5/8
92 36 7/16
105 37 1/2
104 41 15/32
108 38 1/2
92 35 7/16
106 44 1/2
108 41 1/2
114 45 . 9/16
102 45 1/2
134 41 5/8
109 39 " 1/2
99 44 1/2
135 42 5/8
133 39 5/8
92 39 7/16
107 40 1/2
116 38 17/32
121 38 9/16
119 43 9/16
Cake Solids,
wt%
48. 0
43.2
43.4
44:>>
43. 7
45.8
44.6
44. 5
43. 9
44. 7
41.6
42. 1
45.3
44. 9
41.0
44. 6
43.9
43.1
44.2
45.2
45.4
43.9
Filtrate Solids,
wt%
0. 00
0. 02
0. 07
°-°2(2,
i.'zo (3)
0.05
0.25 ( '
0.01
0. 77 * '
0. 10
0. 03
0. 00
0. 03
0. 01
0. 03
0. 06
0. 07
0. 05
0. 02
0. 09
0. 07
0. 84 ( '
r\>
ro
c
1. Refer to Table 22-2.
2. Ran cloth on lime slurry only.
3. Leak suspected during leaf test.
4. Sample of cloth normally used on the rotary drum filter.
5. Based on a sixty-second cycle.
6. From Run 706-2A.
-------
ro
ro
09
25
-*-
SO
CLOTH AIR PERMEABILITY, cfm/ft2
75
100
Figure'22-2. Filter Leaf Filtrate Rate versus Cloth Air Permeability,
Lime with Fly Ash Slurry (60 second cycle)
-------
Weight percent solids in the filtrate gradually increased with increasing
cloth permeability from 0.00 to 0.36 for lime and from 0.00 to 0.09 for lime-
stone. However, all of these values are within the acceptable operating range.
High filtrate solids values with limestone for cloths 6, 8, 10, and 23 were
attributed to leakage during these particular tests.
Tests with previously blinded cloths of types 8, 11, and 14 showed a reduction
in cake production rate to between 25 and 75 percent of the clean rate (see
Table 22-5).
22.3.2 Filtration Tests
Additional filter leaf tests were performed at Shawnee to evaluate the filtra-
tion characteristics of sludge generated during the following types of operation:
Alkali
Lime
Lime
Lime
Limestone
Limestone
Fly Ash Load ins
High
Low
Low
High
High
Forced
Oxidation
Yes
Yes
No
Yes
Yes
Independent
Scrubber Loops
2
2
—
1
2
Results of tests at the optimum filtration cycle for each type of operation are
presented in Table 22-6. Note that results for slurries from forced-oxidation
tests were similar, regardless of alkali, fly ash content, or number of scrubber
loops. Oxidized cake solids were between 70 and 80 weight percent and dry solids
22-19
-------
Table 22-5
FILTER LEAF TEST RESULTS WITH BLINDED CLOTHS
LIME SLURRY WITH HIGH FLY ASH LOADING
Cloth
Number
8
11
14
Total
Operating
Hours
866
349
530
Dry Cake Production,
Ib/ft2-hr<2)
New 1 Blinded
73 21
98 40
112
Filtration Rate,
eal/ft2-hr^2)
New
46
42
51
LIMESTONE SLURRY WITH LOW FLY ASH
Cloth
Number
8
11
14
Total
Operating
Hours
866
349
530
Dry Cake Production,
New | Blinded
108 56
108 88
134 43
Blinded
8
27
LOADING
Filtrate Rate,
Ral/ft2-hr<2)
New I
38
41
41
Blinded
21
29
29
Cake
w
New
66.9
61.8
64.6
Solids,
Blinded
57.6
67. 1
Cake Moisture,
New 1
44.6
44. 7
45.3
Blinded
42.5
44.6
43. 0
ro
ro
i
ro
o
(1) Refer to Table 22-2.
(2) Based on a sixty-second cycle.
-------
Table 22-6
FILTER LEAF TEST RESULTS AT OPTIMUM FILTRATION CYCLE
Alkali
Type
Lime
Lime'1'
Lime'1'
Limestone
Limestone
Fly Ash
Loading
High
Low
Low
High
High
Forced
Oxidation
Yes
Yes
No
Yes
Yes
Scrubber
Loops
2
2
--
1
2
Form
Time,
sec
5
5
5
5
5
Dry
Time,
sec
25
20
20
25
30
Cycle
Time.
sec
40
33
33
40
47
Initial
Solids,
wt.X
58.4
47.9
39.3
58.9
52.7
Cake
Solids,
wt.%
82.3
70.1
49.9
79.8
78.8
Cake
Thickness,
in.
3/4
3/8
3/8
7/16
5/16 - 7/16
Dry Solids
Loading,
Ib/ft2-hr
428
295
281
258
212
Filtrate
Loading,
gal/ftZ-hr
27.3
23.3
18.1
14.7
16.0
Submergence,
% of cycle
12.5
30.0
30.0
12.5
21.0
no
ro
(1) Worst results of several runs at optimum filtration cycle for this slurry.
-------
p
loadings were between 200 and 300 Ib/hr-ft , except for the high value of 428
p
Ib/hr-ft*- for two-loop forced oxidation with lime and high fly ash loading.
The unoxidized lime solids with low fly ash loading had much lower cake solids
content (about 50 weight percent). Appendix L presents detailed filter leaf
test results.
22.4 HYDROMETER AND PIPETTE PARTICLE SIZE DISTRIBUTION TESTS
One of the methods used to assess the ease of solids dewatering is to test the
initial settling rate of solid particles within the slurry. However, the
product solids at Shawnee are a mixture of CaS04'2H20, CaS03*l/2H20, fly ash,
unreacted lime or limestone, and other inert materials. All have independent
particle size distributions which superimpose to form the particle size
distribution of the slurry mixture. The use of settling rate data to assess
the effects of process variables on the CaS04*2H20 particle size distribution
generated during forced-oxidation testing can be masked by other materials
in the slurry, particularly fly ash. Even in operations with flue gas having
passed through the electrostatic precipitators at Shawnee, enough fly ash
gets past the precipitator to form about one weight percent of the product
solids.
This fly ash is extremely fine when compared with CaS04*2H20 particles, and
settles at a slower rate. The resulting limited settling rate limit appears
to be 1.5 to 2.5 cm/min at the Shawnee Test Facilty. Once the gypsum settling
rate reaches this limit, any increase in the average size of gypsum particles
due to changes in process variables will not be observed. Therefore, a test
to assess process variable effects on gypsum particle size in the presence
22-22
-------
of fly ash Is needed.
Two particle size analysis tests potentially capable of assessing gypsum par-
ticle sizes in the presence of fly ash were briefly investigated: the hydro-
meter test and the pipette test. The hydrometer test was ASTM designation
D422-63 (Reapproved 1972), "Standard Method for Particle - Size Analysis of
Soils". In this test, a liter of solution containing about 50 grams of
gypsum solids is placed in a 1-liter graduated cylinder in which an ASTM
hydrometer (151H or 152H in ASTM Specification E100) is placed. The hydro-
meter reading is observed with time of settling to give the particle size
distribution of the solids, which range in size from 5 to 75 microns. This
technique is used as a regular laboratory analysis by the Bechtel Pipeline
Division Slurry Laboratory.
Two samples from forced-oxidation testing were analyzed by the ASTM hydro-
meter method: one from Run 857-1A and one from Run 858-1A (both runs with
low fly ash loading). The hydrometer test size distributions obtained from
the two samples are plotted in Figure 22-3 (as circles). Over the size range
measured, no significant difference between the two size distributions is
noticeable. Even though the flue gas in these runs had passed through an
electrostatic precipitator, fly ash was visibly present in the slurry.
To identify which particle species was falling most rapidly (i.e., fly ash or
CaS04*2H20), a pipette sampling method was used (Reference 10). The pipette
method does not involve the use of a hydrometer. Instead a 10-ml pipetted sam-
ple is periodically removed from the settling slurry at some constant depth.
The sample is analyzed for weight of solids and weight percent CaS04*2H20.
From this information and Stokes1 Law, the corrected CaSO^h^O particle size
22-23
-------
100
90
80
70
60
50
40
30
20-
10
9
8
7
I
5
4
*
2
1.0
.9
.8
.7
.6
.5
.4
.3 •
O
f A
0°
A
O
O
O A
RUN 857-1A
O HYDROMETER
A PIPET
RUN 858 - 1A
• HYDROMETER
A PIPET
10
-f-
—I
20 30 40
PARTICLE STOKES DIAMETER,
50
60
Figure 22-3.
Particle Size Distributions of Oxidized Slurry with
Low Fly Ash Loading from the Venturi/Spray Tower System
22-24
-------
distribution can be calculated.
The overall size distributions from the pipette size distribution test are
also plotted in Figure 22-3 (as triangles). Agreement between methods was
good for Run 858-1A but not for Run 857-1A. This may have been caused by
the poor sampling technique associated with the pipette method.
The data from the pipette test are much more revealing when weight percent
CaSO^^h^O from each sample is plotted along with the overall size distribu-
tion. This has been done for a sample from Run 857-1A in Figure 22-4 and a
sample from Run 858-1A in Figure 22-5. These plots illustrate that the fines
are mostly ash (e.g., only 6 percent of the solids smaller than 17 microns are
CaS04*2H20 in Figure 22-4, and if the two plots were combined a CaS04*2H20
size distribution could be generated).
Static cylinder settling tests of the two slurries were run after they were
both adjusted to the same initial percent solids with clarified liquor. Both
the upper and lower interfaces observed in these settling tests are plotted
in Figure 22-6. Chemical analysis showed that the solids built up from the
bottom of the cylinder (the lower interface) were gypsum and that the solids
settling from the top (the upper interface) were essentially fly ash. This
is further proof that the initial settling rates presented for slurry from
the forced-oxidation tests normally represent the size distribution of the
fly ash and not the size distribution of the CaS04*2H20 solids.
The data from these comparisons prove that the initial settling rate test is
not a valid indication of the relative size distribution of the gypsum crystals
generated during forced-oxidation te'sting. The hydrometer test gives a good
indication of the overall size distribution but no information on the fly
22-25
-------
8 "
til
_j
*
£ «
g
I .
o
E
-00-
20 30 « 50
STOKES fARTICLE MAMETEH.
O
I »
t «
20 30 40 SO
STOKES f ARTICLE DIAMETER. "
Figure 22-4.
Particle Size Distribution by the Pipette Method for
Oxidized Slurry with Low Fly Ash Loading from Run 857-1A
22-26
-------
Q
Ul 40
20 30 40 50
STOKES PARTICLE DIAMETER.
S *
t 40
u
£
z
o w
o*
Z
20 30 40 50
STOKES PARTICLE DIAMETER, microns
Figure 22-5.
Particle Size Distribution by the Pipette Method for
Oxidized Slurry with Low Fly Ash Loading from Run 858-1A
22-27
-------
1000
900
RUN 857 - 1A
RUN858-1A
LOWER INTERFACE
5 10 15 202530354045505560
TIME, minutes
Figure 22-6. Results cf Static Cylinder Settling Tests on
Oxidized Slurry with Low Fly Ash Loading for
Runs 857-1A and 858-1A
22-28
-------
ash/gypsum distribution. The pipette method appears to be the most promising
method for separating the gypsum size distribution from the fly ash distribu-
tion; however, the disagreement between the overall size distribution generated
by this method and by the hydrometer method must be resolved before any real
confidence can be placed in the results of the pipette test. Further study
will be made in the future as time permits.
22.5 HYDROCLONE TESTS
Hydroclones are frequently used for solid-liquid separations because of their
small size, low capital cost, and simple design. To determine their applica-
bility as a dewatering device for FGD slurry, a limited test program was
conducted at the Shawnee Test Facility in September and November 1977.
The objective of the hydroclone testing was to achieve less than 0.5 weight
percent solids in the overflow and greater than 40 weight percent solids in
the underflow. Tests were conducted on oxidized limestone slurry from the
venturi/spray tower system. The slurry solids consisted primarily of calcium
sulfate (30 to 50 microns), fly ash (about 10 microns), and calcium sulfite
(about 5 microns). Hydroclone feed was taken from the desupersaturation tank
via the discharge of the venturi slurry recirculation pump outlet (see Figure
5-1). Both hydroclone overflow and underflow streams were returned by gravity
to the desupersaturation tank.
Tests with a small Doxie 5 hydroclone were discontinued because of plugging;
tests with a larger Dorrclone P50A were successful. Underflow solids concen-
tration was excellent (about 55 percent), but overflow solids were high (2 to
3 percent). Chemical analysis showed that 60 to 70 percent of the overflow
22-29
-------
solids was fly ash.
Based on the limited test results, the hydrometer Is judged to be unsuitable
as a dewatering device for waste lime/limestone slurry because of high solids
content in the overflow. A relatively high energy requirement (high pressure
drop) and possible erosion problems would also make it unattractive for this
service.
Results of the hydroclone test program are summarized in this section. De-
tailed test data are presented in Appendix L.
22.5.1 Test Results with the Doxle 5 Hydroclone
The first tests were performed with a Doxie 5 hydroclone manufactured by
Dorr-Oliver. The unit consisted of six 10-mm diameter cyclones, and was de-
signed to separate particles as small as 3 to 8 microns in diameter at a
pressure drop of 40 psi and a feed flow rate of 5 gpm.
Tests were made with oxidized limestone slurry with low fly ash loading from
Runs 815-1A and 816-1A in September 1977. During the first tests, the over-
flow solids concentration was 0.22 to 0.27 weight percent and the underflow
was about 30 weight percent. Pressure drop ranged from 25 to 43 psi, and the
feed flow rate ranged from 4.5 to 5.7 gpm. Although the separation was good,
the strainer on the Doxie inlet plugged every five to ten minutes. Difficul-
ties with the pump also complicated testing.
A second set of tests was conducted without the strainer. The overflow solids
concentration was about 5 weight percent, and the underflow was 25 to 30
weight percent.
22-30
-------
These tests indicated that the Doxie 5 hydroclone could separate the small
particles but could not handle the large particles also present in the Shawnee
slurry. Therefore a larger unit, Dorrclone P50A, with a higher separation
size was purchased.
22.5.2 Test Results with the Dorrclone P50A Hydroclone
More extensive testing was done on a Dorrclone P50A porcelain hydroclone
manufactured by Dorr-Oliver. The unit had two-inch I.D. feed and overflow
connections and a 12-mm (0.47 inch) underflow orifice (1.5 inch connection).
It was designed to separate 95 percent of particles greater than 30 microns
at a 22 psi pressure drop and a 15 gpm feed rate. A new Moyno pump was used
for these tests.
These tests were performed from November 8 to 11, 1977, with oxidized lime-
stone slurry with high fly ash loading from Run 819-1A at a nominal concentra-
tion of 15 weight percent solids. Tests were performed to determine the range
of pressures and feed flow rates, the flow separation, the solids separation,
and the differences in solids composition and size between the overflow and
the underflow.
The hydroclone operated a total of 17 hours with no sign of plugging, scaling,
or solids buildup. The interior remained smooth and free of erosion.
Several one-half hour tests were conducted to characterize the hydroclone
separation. The results of these tests are shown in Table 22-7. Pressure
drops ranged from 10 to 29 psi at feed rates of 11 to 19 gpm (maximum for the
pump). The outlets were at atmospheric pressure. About 80 percent of the
22-31
-------
Table 22-7
RESULTS OF ONE-HALF HOUR TESTS WITH DORRCLONE P50A HYDROCLONE
- RANGE OF OPERATION -
ro
ro
i
U)
ro
Inlet
Preaiure, pitg
(» Pressure Drop)
10
1Z
14
17
19
Zl
24
27
Z9
Feed Flow j 2
Rate, gpm '
11
12
13
14
15
16
17
IB
19
Overflow
gpm
8.8
9.4
10.4
11.0
11.8
1Z. 8
13.8
15.6
16.4
Flow Rate
wt% lolldi
3.35
3.14
2.41
Z.75
2.64
2.36
2.24
1.91
2.08
Underflow Flow Rate
gpm 1 wt% tolldi
2.2 51.27
2.6 S5.74
2.6 49.43
3. 0 S6.-66
3.2 56.87
3.2 54.93
3.2 54.55
2.4 46.6
2.6 53.75
Percent Flow
Overhead
80
78
80
79
79
80
81
87
86
Values for the venturi Inlet line
1. Measured with a magnetic flow meter.
2. Wt. percent solids was not measured.
were 15.0 and 16.3 wt. percent.
3. By difference (feed minus underflow). Overflow rate was also measured
by bucket but, due to the small size of the bucket, the overflow rate
was not felt to be very accurate.
4. Measured by bucket and stopwatch.
-------
total flow went to the overflow. The overflow had 1.9 to 3.4 weight percent
solids concentration with higher values at lower pressure drops, as expected.
Chemical analyses showed that fly ash was 57 to 62 percent of the overflow
solids. The underflow had 47 to 57 weight percent solids at flow rates of
2.2 to 3.2 gpm. Although the underflow solids concentrations were excellent,
the overflow solids concentrations greatly exceeded the maximum of 0.5 weight
percent desired.
The hydroclone was equipped with an underflow valve that could be throttled
to increase underflow percent solids. This valve was not used because the
underflow solids concentration was acceptable with the valve fully open.
The overflow was throttled to see whether the overflow solids concentration
could be reduced. This increased the pressure drop from 21 to 27 psi at 16.3
gpm feed, but the overflow solids remained at 2.5 weight percent. Fifty-six
percent of the total overflow solids was fly ash. The overflow rate was 7.9
gpm. The throttling of overflow diluted the underflow to 26 weight percent
solids at a flow rate of 8.4 gpm.
Three two-hour tests conducted at 12, 15, and 18 gpm showed little differences
in composition and particle size distributions in the overflow and underflow
solids. Operating variables, solids concentrations, and chemical analyses are
shown in Appendix L.
The chemical analyses show that the overflow solids were predominantly fly ash
(59 to 71 percent) with some gypsum (11 to 13 percent), calcium carbonate (4
to 7 percent), and calcium sulfite (2 to 6 percent).
22-33
-------
The underflow solids contained 56 to 58 percent gypsum, 18 to 32 percent fly
ash, 5 to 6 percent calcium carbonate, and 1 to 2 percent calcium sulfite.
The feed solids were 46 to 48 percent gypsum, 28 to 35 percent fly ash, 5 to
6 percent calcium carbonate, and 1 to 2 percent calcium sulfite. Material
balances around the hydroclone show that 80 percent of the flow went into
the overflow, but only 12 percent of the solids did. Of the solids, about
30 percent of the fly ash, 23 percent of calcium sulfite, 12 percent of the
calcium carbonate* and 3 percent of the gypsum remained in the overflow.
The high gypsum separation was a result of its larger particle size.
i
22.5.3 Particle Size Distribution
Particle size distribution analyses were made by the Bechtel Slurry Labora-
tory on slurry samples taken from the two-hour Dorrclone P50A hydroclone
tests. Overflow and underflow samples were taken at feed rates of 12, 15,
and 18 gpm. In addition, the feed was sampled at a flow rate of 15 gpm.
These samples were analyzed by the hydrometer method as described in Subsec-
tion 22.4. The Calgon dispersant (sodium hexametaphosphate) normally used
in this method was deleted because it reacted with the slurry.
Size distribution analyses for the three feed rates are plotted in Figures
22-7 through 22-9. The plots show that, in the range tested, particle size
distributions in the overflow and underflow did not vary with pressure drop
and feed rate. The solids size distribution in the underflow was slightly
higher than that in the feed. The difference was small because the overflow
solids were a minor fraction of feed solids. About 90 percent of the overflow
solids were smaller than 15 microns, and about 95 percent were smaller than
22-34
-------
100
90 ••
80 •-
70 •
GO ••
40 • •
30 • •
20 •
10 ••
18 6PM FEED RATE
O OVERFLOW
• UNDERFLOW
A FEED AT 15 GPM
O
O
O
O
O
O
h
•4 1 1 I I I
456 78910 15 20
PARTICLE STOKES DIAMETER, .
30 40 5060708090100
Figure 22-7. Particle Size Distributions from Hydroclone Testing
(Dorrclone P50A at 18 gpm Feed Rate and 27 psi Pressure Drop)
22-35
-------
too
90
80
70 •
40
30
20
10
15 GPM FEED RATE
O OVERFLOW
• UNDERFLOW
A FEED
-* 1 1 1—I I JO)
H > I I I I
4 56789 10 15 20
PARTICLE STOKES DIAMETER.
3040 5060708090100
Figure 22-8. Particle Size Distributions from Hydroclone Testing
{Dorrclone P50A at 15 gpm Feed Rate and 19 psi Pressure Drop)
22-36
-------
Ill
_l
*
100
90
80
70 ••
60 •
50 •
40
30
20
10
12 GPM FEED RATE
O OVERFLOW
• UNDERFLOW
A FEED AT 15 GPM
•4 1-
I I Q I
H 1 1 1 I I
4 56789 10 15 20
PARTICLE STOKES DIAMETER.
3040 5060708090100
Figure 22-9. Particle Size Distributions from Hydroclone Testing
(Dorrclone P50A at 12 gpm Feed Rate and 12 psi Pressure Drop)
22-37
-------
25 microns.
The Dorrclone was designed to separate 95 percent of the particles greater
than 30 microns in diameter. The particle size analysis shows that only 50
percent of the feed solids were greater than 30 microns. Thus, the Dorrclone
was designed to separate only half the solids from the liquor. As the actual
separation was 88 percent of the feed solids in the overflow, the Dorrclone
performed better than design. However, additional treatment of the overflow
from this device would be required to reduce the overflow solids concentra-
tion from 2 to 3 weight percent down to less than 0.5 percent required to
keep fines concentration from building up in the system.
22.6 SCANNING ELECTRON MICROSCOPE (SEM) ANALYSIS OF OXIDIZED SOLIDS
During forced-oxidation testing, the oxidized solids experience at least
three different environments: the oxidation tank, the clarifier, and then
the filter. In order to determine if any significant particle size changes
occur during the various operations, slurry samples from selected points in
the scrubber were taken during venturi/spray tower Run 854-1A (made in March
1977 with fly ash, two-stage forced oxidation). These samples were sent to
Dr. S.K. Seale at the TVA Laboratory in Muscle Shoals, Alabama, for scanning
electron microscope (SEM) analysis. Samples were taken from the spray tower
scrubbing loop, the venturi loop, the system bleed, the clarifier underflow,
and the filter cake.
Figures 22-10 and 22-11 are SEM photographs of the slurry solids at different
magnifications. The solids are 10 percent oxidized in the spray tower loop
and 95 percent oxidized at all other points. The crystals from the spray
22-38
-------
77-711 280X 1816 3/38/77 0888
Spray Tower Loop
50/i
Venturi Loop
Bleed
Clarifier Underflow
Filter Cake
Figure 22-10. Scanning Electron Microscope Photographs of Slurry
from Forced-Oxidation Run 854-1A (200x magnification)
22-39
-------
20M
Spray Tower Loop
Venturi Loop
Bleed
Clarifier Underflow
Filter Cake
Figure 22-11. Scanning Electron Microscope Photographs of Slurry
from Forced-Oxidation Run 854-1A (500x magnification)
22-40
-------
tower loop are primarily calcium sulfite hemihydrate (CaS03*l/2H20), while
the larger crystals in the other photographs are primarily calcium sulfate
dihydrate (CaSO^ZF^O), and ranged in size from 20 to 100 microns. A compar-
ison of subsequent samples (the bleed, the clarifier underflow, and the filter
cake) with samples from the venturi loop indicates that there is no change
in the particle size or morphology once the oxidation step has been completed.
22.7 GYPSUM CRYSTALLIZATION
22.7.1 Introduction
Some of the major considerations for any throwaway lime and limestone S02
scrubbing process are the ease of waste sludge dewatering and the safe and
economical disposal of discharged solids. During the summer of 1977, Dr. J.R.
Beckman of California State University, Northridge, investigated gypsum crys-
tallization technology as it applied to the Shawnee flue gas scrubbing systems.
This subsection summarizes the results of Dr. Beckman's studies.
Clarification followed by filtration has proved to be a viable method of
sludge disposal for the purpose of land disposal. Because of larger inter-
stitial voids, large particles will filter more easily than small particles.
Methods that can increase the gypsum crystal size and reduce the number of
fines (i.e., improve the solids quality) are therefore desirable.
22.7.2 Theory and Results
The smallest stable crystals capable of growth are called nuclei. There are
two ways by which nuclei can form:
22-41
-------
• Primary nucleation: the spontaneous clustering of solute molecules
into larger stable aggregates in a supersaturated solution.
• Secondary nucleation: the formation of nuclei by crystal-crystal
and crystal-solid collisions. This becomes of increasing impor-
tance as both turbulence and slurry solids content increase.
To increase the quality of solids, therefore* it is desirable to minimize the
number of crystal nuclei in the solution. Both primary and secondary nuclea-
tion sources contribute to the total number of nuclei. However, one will
usually be dominant for a given system.
According to Dr. Beckman, primary nucleation obeys a "power law" model of the
form:
n = k^1
where:
n = number of nuclei
kn = constant
G = diameter growth rate of crystals
i = kinetic parameter; constant
Dr. Beckman further states that secondary nucleation is a function of slurry
density:
n = kro
where:
kra = constant
MT = slurry solids content
j = kinetic parameter; constant
The results of modeling show that the quality of solids should improve with
22-42
-------
increased residence time and/or slurry density if primary nucleation predom-
inates. The opposite should occur if secondary nucleation predominates.
When the solids residence time was halved in the venturi loop, from 15 weight
percent in Run 857-1A to 8 percent in Run 858-1A, the final settled solids
dropped from 77 to 66 weight percent, and the settling rate dropped from 2.0
to 1.1 cm/min. It should be noted that fly ash interference affects the
observation of gypsum settling rate. On the basis of the finding that dewater-
ing becomes more difficult with lowered solids residence time and reduced
slurry density, Dr. Beckman concluded that primary nucleation is predominant
in the Shawnee system.
22-43
-------
Section 23
OPERATING EXPERIENCE DURING LIME/LIMESTONE TESTING
This section summarizes the operating experience during lime/limestone
testing, with both high and low fly ash loading in the flue gas and with
and without MgO addition at the Shawnee Test Facility from December 1976
through June 1978.
23.1 SCRUBBER INTERNALS
23.1.1 Mist Eliminator Operation
The three-pass, open-vane, chevron mist eliminator, made of type 316L stain-
less steel, continued in service from December 1976 to June of 1978 on both
the venturi/spray tower and the TCA.
From December 1976 to March 1977, moderate plugging was experienced in both
systems. Mist eliminator restrictions of 10 to 40 percent* typically occurred
after only a few hundred hours of operation. During that time, the wash
sequence and spray pattern were continually refined. In February 1977, a
continuous underwash system was made available for both the venturi/spray
tower and the TCA. Automatic control of the underwash flow was accomplished
* As estimated visually by TVA inspectors.
23-1
-------
by an instrument loop, consisting of a magnetic flow meter, controller,
recorder, I/P converter (ampere to pressure signal converter), and control
valve. By June 1977, the wash sequences and patterns as outlined in Table
23-1 were being used. Since that time plugging has been nearly eliminated.
It was found during the tests that operation at low-alkali stoichiometry
made it much easier to keep the mist eliminator clean. Below approximately
1.25 moles Ca/mole S02 removed, a periodic flush with fresh water was all
that was required. Above a 1.25 stoichiometry, the mist eliminator became
progressively more difficult to keep clean, and a continuous wash with diluted
clarified liquor was needed.
From June 1977 to June 1978 the venturi/spray tower mist eliminator was
cleaned once, in December, after 4183 hours of operation. At that time, the
mist eliminator was 15 percent plugged due to a short-term malfunction of the
wash system. The venturi/spray tower mist eliminator has operated 3308 hours
since that time without plugging; the June 1978 restriction was less than 1
percent.
The TCA mist eliminator was cleaned on June 21, 1978. After 7671 hours of
operation, it was only 12 percent restricted.
23.1.2 TCA Grid Supports
The 3/8-inch diameter, 316L stainless steel bar-grids, installed on 1-1/4 inch
centers in October 1973, have been in slurry service for approximately 5
years with no evidence of significant erosion. In 1977, two bars in the bottom
grid exhibited signs of deterioration. These bars were analyzed and determined
23-2
-------
ro
CO
oo
Table 23-1
MIST ELIMINATOR WASH SYSTEM
SCRUBBER SYSTEM
Gas Rowrate, acfm
Alkali
Bottomside M.E. Wash
Wash scheme
Number of nozzles
Nozzle location (inches below M.E.)
Size of nozzlesO)
Nozzle "On" time, min.
Nozzle "Off" time
On/Off sequence/nozzle, hrs.
Total makeup water, gpm
Makeup water/nozzle, gpm
Specific washrate (gpm/sq.ft.)
Ap/nozzle (across each nozzle), psi
Wash water (continuous basis), gpm
Topside M.E. Wash
Wash scheme
Number of nozzles(2)
Size of nozzlesO )
Nozzle "On" time, min,
Nozzle "Off" time, min.
On/off sequence/nozzle, min.
Flowrate (gpm/nozzle)
Ap/nozzle, psi
Covering area (ft2)
Specific wash rate (gpm/ft2)
Wash water (continuous basis) (gpm)
Spray
Tower
35,000
Lime
LI
10
10
1/2G35W
6
3 hr-54
4
75
7.5
1.5
50
1.9
LI
6
3/4H6W
4
76
80
8
13
15
0.53
0.4
Spray
Tower
35,000
Limestone
C
4
20
1/2G35W
Cont.
min. 0
_
20(3)
5.0
0.4
21
HI
6
3/4H6W
3
7
10
8
13
15
0.53
2.4
TCA
30,000
Lime
LI
2
31
1-1/4H190WSQ
6
3 hr-54 min.
4
75
37.5
1.5
41
1.9
LI
6
3/4HH71WSQ
4
76
80
8
13
14.5
0.55
0.4
TCA
30,000
Limestone
c
2
31
3/4HH71WSQ
Cont.
0
20(3)
10
0.4
20
HI
1 1 A
6
3/4HH71WSQ
3
7
10
8
13
14.5
0.55
2.4
LI = low intermittent
HI = high intermittent
C = continuous
(1) All nozzles manufactured by Spraying Systems Co.
(2) Sequential wash with one nozzle activated at a time.
(3) Clarified liquor diluted with makeup water.
-------
to be Type 409 stainless steel that were mistakenly installed during a sphere
change.
23.1.3 Foam Spheres
During this test period, the three beds of the TCA contained 6-gram nitrile
foam spheres obtained from Universal Oil Products. No serious sphere fail-
ures occurred during operation. Acceptable levels of initial sphere shrinkage
and wear were observed.
Some initial problems did occur because of poor quality control during the
manufacturing process. Approximately 10 percent of the nitrile spheres in
lot number 6689 contained cuts and splits; these spheres were replaced. Also,
in lot number 6874, undersized spheres were present which were small enough
to leave the beds by passing through the grids; these spheres were replaced.
From December 1976 to December 1977, the nitrile foam spheres were carefully
monitored as to their wear rate. This was done by sampling each bed and meas-
uring the minimum diameter of 25 spheres. Average weight measurements were
also made of these samples but were much more variable.
Foam spheres underwent an initial shrinkage*; in the first 1000 hours of
^service, the minimum diameters decreased by 6 to 12 percent. The subsequent
shrinkage rate for the next 6700 hours of service was measured as 0.05 inch
per 1000 hours. The average shrinkage rate for the total 7700 hours was 0.07
inch per 1000 hours.
Includes decrease in size due to abrasion.
23-4
-------
The entire evaluation of sphere wear took place while operating with flue
gas with high fly ash loading, which represents a more severe environment
than when operating with low fly ash loading. Mixing of the various beds to
achieve desired static heights precluded further evaluation of these spheres.
At the completion of Run 820-2A, on January 24, 1978, inspection of the sphere
beds revealed that each bed had a quantity of immobilized spheres that were
stuck together by scale deposits. The location and number of spheres involved
were different for each bed, and was an estimated 5 percent of the bottom
bed. This condition had seldom been observed in the TCA and has not occurred
since. Gas flow for Run 820-2A was low, 20,000 acfm, equivalent to 8.4 ft/sec,
and was the same as the previous run. The spheres in this run had decreased
in size, and some of them were forced through the top grid.
23.1.4 TCA Slurry Nozzles
The four main slurry spray nozzles in the TCA were Model 1969F, Type 316 stain-
less steel nozzles manufactured by Spraco. These continued to give excellent
service both in terms of wear and nonclogging characteristics. These nozzles
are normally operated at 5 psi pressure drop. Some erosion occurred at the
discharge orifice.
23.1.5 Venturi Internals
In January 1977, minor repairs were made to the venturi plug guide vanes and
the shroud. Repairs were also made to the rubber lining in the venturi
flooded elbow. An inspection of the venturi in March 1977 showed serious
23-5
-------
abrasion to a section of the guide vane structure that had previously been
repaired. The protective rubber covering the four guide vanes had failed to
varying degrees and was replaced with new covering. Other maintenance included
the replacement of the venturi bull nozzle in December 1977 and the repair
.of the venturi inlet gas duct in February 1978. Solids or scale deposits in
the venturi or the flooded elbow were not a significant problem.
In May 1978, due to a failure of the venturi plug guide vanes, the drive motor
was disconnected and a jack was temporarily installed to manually operate the
plug. In June, a new venturi plug drive mechanism was fabricated and installed.
The shroud splash guard in the plug shaft was rebuilt, a new shaft was fabri-
cated for the plug limit switch mechanism, and new guide vanes and blocks were
installed. The old shaft had been in service since March 1972.
23.1.6 Spray Tower Slurry Nozzles
Bete No. ST48FCN nozzles were used successfully as the main slurry spray
nozzles in the spray tower. The materials of construction for these nozzles
were 316 stainless steel for the body and stellite VI for the spiral diffuser.
New stellite diffusers were installed at the start of Run 801-1A. In March
1977, after a total of 2649 hours, 973 hours of which were in limestone service
and 1676 hours in lime service, the diffusers were judged to have negligible
pitting and relatively uniform wear, as shown by measured thickness of the ref-
erence area of four randomly selected diffusers:
4th level (upper) 0.254 inch
3rd level 0.223 inch
2nd level 0.220 inch
23-6
-------
1st level (lower) 0.217 inch
These figures illustrate that flow rates through the four headers were approx-
imately equal, except for a possibly reduced flow rate at the 4th level. The
metal used for these diffusers, stellite VI, consisted of 25 percent chromium,
5 percent tungsten, 1 percent carbon, and 69 percent cobalt.
23.1.7 Spray Tower Trapout Funnel
Separation of liquor flow leaving the venturi scrubber from that leaving the
spray tower was necessary to achieve independent scrubber loops for the two-
loop forced-oxidation testing. This flow division was accomplished by install-
ing a large trapout funnel in the spray tower below the bottom level of the
spray nozzles. The funnel and the associated slurry tanks required for the
oxidation system are shown in Figure 5-1. Flue gas moves upward around the
funnel periphery through the one-foot annular space at the wall of the spray
tower and then is deflected over the funnel by an annular ring that is attached
to the tower wall. The annular ring also serves to direct falling liquor spray
from above into the trapout funnel. There has been no evidence of corrosion
on the funnel which is constructed of Type 316L stainless steel.
During normal operation, accumulations of solids on the underside (gas side)
of the funnel were a common occurrence. On one occasion, at the end of Run
814-1A, nearly complete plugging occurred on one portion of the funnel. Growth
of solids was progressive and built up on the outer edge of the funnel to a
point only a few inches from the bottom of the deflector ring. There were no
problems with solids accumulation on the inside of the funnel, and no problem
was encountered with slurry flow through the downcomer.
23-7
-------
As a result of the changes in configuration, there exists an annul us space
between the trapout funnel downcomer and the original spray tower slurry
outlet, from which a lateral crossover pipe about 50 feet long takes the
slurry to the oxidation tank. The annul us and the gravity flow line have
been a source of periodic solids restriction. Cleaning of the area normally
requires a 2 or 3 man maintenance crew and 3 to 4 hours. The problem has been
tolerated because a permanent solution would require unacceptably expensive
configurational changes and future testing will allow the elimination of the
annul us and the crossover line.
23.2 OXIDIZERS
23.2.1 Air Sparger
The air sparging system used in slurry oxidation involved bubbling air into
the bottom of a slurry tank in conjunction with simultaneous slurry agitation.
At Shawnee the oxidation tanks were purposely tall and thin (7 to 8 feet in
diameter and 20 feet high) to provide a long air/slurry contact time.
The first air sparger consisted of an octagonal ring with 130 1/8-inch holes
pointed downward (see Figure 5-2). It was installed near the bottom of the
oxidation tank of the venturi/spray tower system in December 1976. During
\
the May 1977 outage, the sparger was inspected after 1745 hours of service.
Erosion of the air holes in the ring was noted as follows: 54 holes badly
eroded, 33 holes slightly eroded, and 65 holes plugged with hard scale. This
sparger was replaced by an identical ring having 40 1/4-inch holes. In Octo-
ber 1977, the new sparge ring was inspected after 2400 hours of operation.
23-8
-------
A spool piece connecting the sparge ring to the compressed air source had
seriously eroded. It was apparent from the solids in the sparge ring that
most of the air had been escaping from the holes in the spool piece for some
time. The spool piece material was determined to be Type 304 stainless
steel rather than 316L as intended. Despite the nonuniform distribution of
air in the oxidation tank, good oxidation was achieved during this test
period.
On October 4, 1977, the air sparger ring was removed and replaced with a
single 3-inch diameter pipe extending to the center of the tank. Oxidation
continued to be good with this arrangement, which indicates that the agitator
plays a key role in achieving good air/slurry contact. The air and slurry
were mixed with a 20 Hp axial flow agitator rotating at 56 rpm.
During December 1977 the sparger ring with 40 holes of 1/4-inch diameter was
installed in the TCA effluent hold tank. The ring, constructed of 316L stain-
less steel, was materially in good condition, except for minor erosion at
the 1/4-inch air holes. A conventional tank agitator (37 rpm, 3 Hp) was used
for mixing. Near complete oxidation was achieved with this configuration,
but it was suspected that the agitator did not have enough power to achieve
good air/slurry contact. Future tests will use a variable speed agitator.
Both venturi/spray tower and TCA sparge systems operated from the same Worth-
ington oil-free air compressor (100 Hp motor) under the following conditions:
50 psig, 270°F, and flow rates normally at 210 scfm per scrubber. The compres-
sor loading/unloading cycle was normally 4 seconds/10 seconds when operating
with 210 scfm air flow to the oxidizer. The oil-free compressor was chosen
to minimize the possibility of slurry contamination by oil containing oxida-
tion inhibitors.
23-9
-------
Maintenance on the sparge systems has been mainly confined to the compressor.
During the test period, the following maintenance was performed:
• The compressor would not unload in December 1977. Two of the
four loading valves were disconnected, allowing the compressor
to operate. The valves were repaired on January 1978 and
reconnected.
• In January 1978 the four unloading valves were rebuilt. The
Viton 0-rings and teflon outer rings were replaced. A compressor
discharge air hose ruptured and was replaced.
• The compressor cooling water jacket was back flushed with water
in June 1978 to remove scale and silt buildup.
The suction valve was serviced and the compressor cooling water jacket was
flushed periodically.
23.2.2 Penberthy Eductor
The eductor is a device that passes a high-velocity slurry through a con-
stricted nozzle, into an eductor chamber, and then through a moderately re-
stricted jet throat. In the chamber, the slurry induces a vacuum that draws
ambient air into the chamber via an entry pipe located at right angles to
the slurry flow path. The air is then entrapped and mixed in the slurry as
the fluid leaves tha eductor chamber and passes through the jet throat.
At Shawnee, a Penberthy Model ELL-10 Special eductor was tested. The materials
of construction were stellite for the nozzle and neoprene-lined carbon steel
for the eductor chamber and exit jet throat. The system was operated at 1600
gpm.
Test results indicated that near complete oxidation could be achieved, but at
the same time, serious erosion problems developed. The neoprene lining in
23-10
-------
the jet throat was observed to be chipped off after only 620 hours of
operation. After approximately 1800 hours of operation, bare carbon steel
was exposed, and when an epoxy patch failed after a total of 2505 hours of
operation, the tests were terminated.
Currently, no plans exist to resume testing of eductors. When compared with
sparge air systems, eductors offer no advantages in oxidation capability.
Secondary reasons for resuming testing include the materials erosion problem
and unfavorable operating and capital costs in comparison to sparge air sys-
tems (see Section 12).
23.3 REHEATERS
Flue gas from the scrubber is reheated to prevent condensation and corrosion
in the exhaust system, to facilitate analytical sampling, to protect the
induced-draft (ID) fans from solid deposits and droplet erosion, and to in-
crease plume buoyancy.
The original in-line, fuel-oil-fired units supplied by Hauck Mfg. Co. were
modi-FIed in March 1974 (venturi/spray tower) and May 1975 (TCA) to incorporate
a fuel-oil-fired external combustion chamber manufactured by Bloom Engineering
Co. Both units have operated reliably since the modifications, with only
occasional flameout and equipment problems. During the test period, mainte-
nance consisted of the following:
• The venturi/spray tower Honeywell oil controller failed in February
1977, and was replaced.
• The motor on the TCA reheater combustion air blower was replaced
in June 1977.
23-11
-------
t The venturi/spray tower reheater control failed In September
1977, due to a faulty UV fire eye, and was replaced. During the
same month, the TCA reheater air control failed due to a faulty
interlock relay and was replaced.
• The TCA Bloom reheater UV fire eye malfunctioned and was replaced
in February 1978. The venturi system Bloom reheater ignitor was
completely rebuilt.
• The TCA reheater oil flow transmitter, detector, and eccentric
beam failed in April 1978. The transmitter circuit was modified
to utilize a 10- to 50-milliamp, 75-volt transmitter, and the
system was returned to service.
t The TCA oil flow control valve failed in May 1978 and was removed,
completely cleaned, reset to factory specifications, and reinstalled-
The valve was worn and pitted.
• The TCA oil flow transmitter, which failed in April, was replaced.
• The ignitor for the venturi/spray tower reheater failed and was
replaced in May 1978.
During the test period, accumulation of dried slurry solids in and around
the reheater was common. Typically, gas ducts to the Hauck reheater shell
gained in buildup of scale close to the reheater entrance and developed a
buildup within the shell that extended upwards to the discharge of the Bloom
reheater. Other deposits occurred above the Bloom discharge and in the
discharge duct from the reheater. The solids were allowed to accumulate
to a thickness that would cause a significant pressure drop increase across
the length of the duct. At that point the solids were manually removed.
Cleaning was required about once every two months.
23.4 FANS
The 316L stainless steel fans at the Shawnee Test Facility are ID, centrifuge
fans manufactured by Zurn Industries. Reliability has been fair to good
during the current operating period, with the only system downtime due to
23-12
-------
the following fan problems:
• Occasional cleaning of the fan and fan dampers
• Replacement of the ID fan outboard bearing and pillow block,
five times for the two fans during the last 18 months
0 Replacement of the motor on the Beck damper position control
23.5 PUMPS
23.5.1 Allen-Sherman-Hoff Pumps
The Allen-Sherman-Hoff centrifugal pumps are neoprene lined, slurry service
pumps equipped with Nelson liquid drive units. The pump packing that is used
consists of conventional graphite-impregnated asbestos, the sleeves are 316
stainless steel, and the pump seals are of the air-flush type (centriseal)
used to prevent excess water input to the circulating slurry.
Considering the wide range of operating conditions, the pumps provided satis-
factory service with a moderate maintenance requirement. Table 23-2 shows
a tabulation of pump maintenance during the test period. The most frequent
reason for maintenance was repacking. Although there was no apparent pattern
associated with repacking frequency, the following general trends were observed:
• Wide flow variation as dictated by changing test conditions increased
repacking frequency.
• Operation of a pump at the low-flow end of the pump curve accelerated
the frequency of packing problems.
23.5.2 Moyno Pumps
Moyno pumps are positive displacement variable-speed pumps used principally
23-13
-------
Table 23-2
FREQUENCY OF PUMP MAINTENANCE
ham (Cipiclty. gp»)
Other
G-lpl
Main Slurry (650)
C-102
Thickener Underflow (50)
rrocen Miter (650)
C-104
Hi In Slurry (650)
C-IOS
Bltet (1001
C-107
HHt Ellmlnitor Kith (40)
Renleced I-R with Durco Kirk 11
Replaced V«H-dr1vc
G-109
Bleed (100)
54
ro
t
•£»
C-201
Mitn Slurrv (1200)
Bad motor ground repaired
G-?QZ
Thfcltiner Underflow (50)
G-203
ProcMl Water (650)
C-204
Main Slurry (1200)
57
C-205
Bleed (100)
G-2.Q7.
Hist Eliminator Uiiti (40)
Alkali Addition (Bl
1*
Procitl Miter (100)
Replaced coupling
G-202A
Thickener Umttrflotr (50)
Sludoe-to-H«5te (850)
jells.
Centr1fuo« Cake Reslurrt (351
0-409
Centrate Return (501
(HIP
Filtrate Return (50)
Replaced «1tn Durco Series E
Filler Cake Res lurry (35)
Service hater (400)
C-502B
Serylce Hater (4001
G-501A
Fuel Oil
Fuel 011
Replaced mechanical seal
-------
for supplying alkali makeup slurry for the scrubbers. Five such pumps located
near the alkali storage tanks served this purpose. A larger sized Moyno is
used for the recirculation of underflow from the TCA clarifier. This pump has
seen only limited service.
Moyno pump rotors are helical shaped and made of chrome plate steel for abra-
sion resistance. Stators are rubber-lined with a double helical groove. The
most frequent maintenance problem associated with these pumps is the replace-
ment of eroded pump stators. Occasionally, there was a loss of pump capacity
resulting from erosion of the rotor surface after what appeared to be a failure
of the chrome plating.
23.6 ALKALI ADDITION SYSTEMS
23.6.1 Lime
The lime addition system consisted of a storage silo, a screw feeder, a lime
slaker manufactured by Portec-Cahaba, a slaked-lime holding tank, and asso-
ciated feed pumps. Fresh water slaked the lime to approximately 20 weight
percent solids. The system gave good reliability over the last 18 months
of intermittent operation.
Maintenance consisted principally of periodic slaker cleaning and the replace-
ment of the slaker grit screen on seven occasions. In February 1978, in addi-
tion to a screen replacement, the screen shaker motor, drive shaft and bearings
were replaced. The slaker process control elements were cleaned and serviced
in March 1978.
23-15
-------
23.6.2 Limestone
The limestone addition system consists of a drying-grinding system, a dry
storage tank, a weigh belt feeder, a slurry tank, and an associated feed pump.
The dry-grinding system was installed in 1970 for a dry limestone injection
project which preceded the present program. Considering the age of the system,
it has given acceptable performance with moderate maintenance requirements
during this test period. The following significant maintenance was performed:
• The expanded metal and refactory was replaced in two-thirds of
the barrel in the burner section of the limestone dryer (April
1977).
• The limestone ball feeder circuit was repaired, and the magnetic
starter for the limestone feed belt motor was replaced (July 1977).
a The limestone grinding system dryer was restored to its original
condition after a section of refractory had fallen loose and a
portion of the inner liner had melted (October 1977).
• The refractory in the burner section of the limestone grinding
system dryer was patched several times. The initial patching with
rock wool insulating cement proved unsuccessful. "Krocrete B"
refractory was used for subsequent repairs (November 1977).
• The fuel oil controller on the limestone grinding system dryer was
replaced because equipment vibration was causing breaks in the cir-
cuit boards. A new controller with a more vibration resistant
mounting was installed (February 197.8).
• New Honeywell operating controls on the limestone grinding system
dryer and new gauges in the fuel oil and atomizing air lines were
installed (March 1978).
23.7 DEWATERING SYSTEMS
The primary dewatering of the purged slurry in both scrubbers was achieved
by clarifiers. Further dewatering of the clarifier sludges was accomplished
in the venturi/spray tower system by a filter and in the TCA system by a
centrifuge.
23-16
-------
23.7.1 Filter
An Ametek 3 ft x 6 ft rotary drum vacuum filter without cake wash was operated
at the facility for waste sludge dewatering and dissolved scrubbing additive
recovery. Cake discharge was effectively achieved by a snap-blowback cycle.
In December 1976, a variable speed drive was installed on the drum.
The feed to the filter was typically 15 gpm with 30 to 40 weight percent
solids content. The filtrate generally contained less than 0.02 weight
percent solids. The filter cake varied from 55 to 85 weight percent solids,
depending mainly on whether the sludge was unoxidized or oxidized.
With the exception of the filter cloth, the filter was a moderate maintenance
item. Table 23-3 gives a breakdown of maintenance categories, frequency,
approximate total manhours required, and approximate material costs.
Filter cloth replacement, as noted in Table 23-4, was a serious problem at
Shawnee. The causes of cloth blinding and fraying were not completely
understood. However operating experience indicated a relationship between
cloth life and the technique by which the cloth was fitted to the drum. A
certain amount of looseness in the fit between the dividers appeared to be
desirable for cake discharge and nonblinding. For this filter the looseness
was approximately 1-1/2 inches of extra cloth per filtration section (10
inches x 6 feet section). After being in service for a few hundred hours
the cloth shrunk to approximately 3/4-inch of looseness per section. A
three-man maintenance crew was used to change the cloth; one hammered in the
rope caulking while the other two workers folded over and held the desired
looseness of the cloth. The looseness evidently allowed the cloth to snap
the cake off when the air puff of the cake discharge cycle was applied to
23-17
-------
Table 23-3
FILTER MAINTENANCE REQUIREMENT
December 1976 to May 1978
ro
CO
00
Event
Drum Speed Control Repair
Snap Blowback Repair
Cloth Replacement
Frequency of
Occurrence
3
2
15
Estimated
Total Ons 1te
Labor (man-hours)
24
16
90
Estimated
Total Material
Cost ($)
1000
0
1500
-------
Table 23-4
FILTER CLOTH SERVICE
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
CLOTH TYPE*
Ametek
TFI
Lamports
Lamports
Lamports
Lamports
Lamports
TFI
TFI
Lamports
Ametek
Ametek
Lamports
Lamports
Ametek
Ametek
DATE INSTALLED
11/8/76
6/6/77
7/7/77
7/13/77
7/19/77
8/3/77
8/15/77
9/19/77
9/23/77
10/18/77
10/28/77
2/8/78
2/14/78
2/22/78
3/21/78
4/19/78
HOURS IN SERVICE
1721
292
127
142
302
249
540
187
501
187
2096
188
190
290
535
1197
COMMENT
Blinded
Blinded
Hole in Cloth
Blinded
Blinded
Blinded
Blinded
Blinded
Blinded
Blinded
Hole in Cloth
Bl i nded
Hole in Cloth
Blinded
Blinded
Hole in Cloth
* Ametek - Ametek olefin (STE-F9D8-HJO)
Lamports - Lamports Polypropylene (7512-SHS)
TFI - Technical Fabricators Incorporate, Polypropylene (9162)
23-19
-------
the given filter cloth section. The oxidized sludge exhibited less tendency
towards cloth blinding, and the Ametek olefin appeared to provide the most
satisfactory service of these cloths tested. The better service life of
Ametek was tentatively attributed to the looseness of the Ametek weave in
comparison with the Lamports and TFI weaves.
23.7.2 Centrifuge
A continuous centrifuge is used to dewater scrubber waste sludge from the
TCA system and to recover the dissolved scrubbing additives. Typical
operating conditions consisted of a feed stream flow of 15 gpm at 30 to 40
weight percent solids, a centrate of 0.1 to 3.0 weight percent solids, and
a cake of 55 to 65 weight percent solids for unoxidized slurry and 70 to 80
weight percent for oxidized slurry. Approximately 40 percent of the total
solids was fly ash; the remaining solids were predominantly calcium sulfate
and sulfite.
The machine is a Bird 18 inch x 28 inch solid bowl continuous centrifuge
which operated at 2050 rpm. The material of construction is 316L stainless
steel with stellite hardfacing on the feed ports, conveyor tips, and solids
discharge ports. The bowl head plows and case plows are replaceable. The
pool depth was set at 1-1/2 inches. No cake washing was performed in this
machine.
In January 1977, vibration readings were taken on the Bird centrifuge with
the lapse timer reading 5711 hours. The previous set of readings was taken
on June 23, 1976, when the timer reading was 2126 hours. The results, shown
in Table 23-5, indicate that the vibration had increased slightly.
23-20
-------
Table 23-5
CENTRIFUGE VIBRATION MEASUREMENTS
Maximum Vibration, mils
Centri f uge :
Motor:
Inboard bearing, horizontal
Inboard bearing, vertical
Outboard bearing, horizontal
Outboard bearing, vertical
Drive end, horizontal
Drive end, vertical
Inboard end, horizontal
Inboard end, vertical
2126 Hours
5.0
3.0
8.2
5.5
2.5
6.0
3.2
1.8
5711 Hours
3.8
4.2
9.5
8.5
4.6
10.0
5.0
3.5
23-21
-------
The centrifuge was disassembled on March 22, 1977, for inspection. Major
wear sites were found at the two center feed ports at the periphery of the
screw conveyor, and at the plows on the solids discharge end. Limited
stellite hardface welding was performed onsite. Approximately 4000 hours of
operation had been logged on the centrifuge since the machine was last ser-
viced by Bird.
The Bird centrifuge was dismantled and inspected, and the feed pipe was
replaced in March 1978. In June, the centrifuge was inspected after 6460
hours of operation since the previous servicing. The inspection was prompted
by the gradual and continued increase of centrate suspended solids to a level
of approximately 3 weight percent. The machine was judged to be in generally
fair condition, but certain components were in need of factory repair.
Serious wear was observed at the conveyor tips on the discharge end and at
the junction of the cylinder and the 10 degree section of the conveyor. Wear
was also present at the casing head plows and solids discharge head near the
discharge ports. The bowl and effluent head were in good condition.
The centrifuge will be shipped to the factory in July 1978 for an overhaul
that will include servicing the gear and bearing unit, rebuilding all worn
conveyor surfaces and discharge parts, adding hardfacing to the conveyor tips
and discharge ports, and replacing all seals, bushings, case plows, and dis-
charge plows as necessary.
In an attempt to improve performance and machine life, tungsten carbide hard-
facing will be applied to the conveyor tips, instead of the previously used
stellite.
23-22
-------
23.8 INSTRUMENT OPERATING EXPERIENCE
23.8.1 pH Meters
The process pH is monitored using Uniloc Model 321 submersible electrode
assemblies. Originally, Uniloc Model 320 flow-through meters were installed.
But because of line plugging problems and frequent sensing electrode breakage,
this type of sensor was abandoned. Service requirements for the submersible
assemblies consisted of periodic cleaning and buffering of the electrodes,
generally every 2 or 3 days to ensure accuracy. Also to minimize the service
requirement, the instrument electrodes were placed in water when the scrubbers
were not operating. On infrequent occasions, the electrode assembly failed due
to wet electrical contacts, bad reference, or sensing electrodes. To pre-
vent disruption of online pH monitoring, spare electrode assemblies were
maintained in standby at all times.
To ensure the accuracy of the process pH meters, a procedure was established
in which a laboratory-measured pH was taken once every two hours for comparison
purposes. This procedure permits a normal operation to within +^0.2 pH unit
of a desired set point, a basic requirement for obtaining accurate test data
at the Shawnee Test Facility.
23.8.2 S02 Analyzers
A Du Pont Model 400 UV split-beam photometer is used to measure S02 concentra-
tions. In the last 18 months, the instrument was accurate and reasonably
trouble-free. Maintenance requirements were limited to cleaning the sample
cell and sample lines approximately once every 1 or 2 months and cleaning the
23-23
-------
participate filter about once every 3 to 4 weeks. Ultraviolet lamp failure
was the only component problem ami was caused by uncontrollable and momentary
power fluctuations due to the switching of station power. The effective par-
ticulate filter for the instrument at Shawnee was a cylindrical chamber con-
structed of a fine mesh screen. The screen cylinder was surrounded by a solid
open-ended protective cylinder. The gas sample lines operated leak free; the
lines were Teflon tubing with heat tracing.
23.8.3 02 Analyzers
Oxygen in the inlet flue gas was measured with a Teledyne Model 9500 which used
a microfuel cell. Operating performance was not acceptable because of the fol-
lowing reasons:
t The Teledyne oxygen analyzer system was put into service on June 15,
1977, and worked properly until June 21 when the sequence mechanism
failed.
t The sequence mechanism was serviced and reinstalled on July 11, but
on July 13 the fuel cell failed.
• The analyzer with a new cell was put into service on August 22, and
on August 23 the sequence mechanism failed for the second time.
• On October 11 through 13, several resistors were replaced, and the
system was calibrated against a portable Teledyne analyzer. On
October 17, the fuel cell failed and was replaced.
• The system was disassembled and cleaned on November 18. The solenoid
valves were scaled and inoperative. On November 27, the fuel cell
failed and was replaced.
0 In December, the solenoid valves failed on the inlet sample system.
The fuel cell failed on December 21 and was replaced.
• The analyzer was inoperative most of January because of freezing
problems with the instrument scrubbing water supply.
• Two fuel cells failed in March. The water scrubbing system plugged
with rust particles and was cleaned on March 17 and again on March
20. A filter was installed in the water supply line on March 29.
23-24
-------
• The fuel cell failed on June 13 after 30 days of operation, and a
new cell was installed.
23.8.4 N02 Analyzer
In December 1976, a Du Pont Model 400 UV photometer was modified by Du Pont
technicians to analyze for N02. The system was initially installed at the
TCA inlet. For the duration of the tests, which lasted until February 1978,
no useful data were obtained due to the erratic output signal. Numerous
modifications and checks were performed by Du Pont representatives, but the
source of the problem was not located. The test was judged to be a failure.
Further testing was discontinued in consideration of cost and interest.
23.8.5 Magnetic Flowmeters
During the past test period, the Foxboro 2800 series and 1800 series magnetic
flowmeters were used exclusively and showed no serious problems. Scale
cleaning was required to improve accuracy and sensitivity, and on infrequent
occasions the magnetic converter required maintenance. Nearly all meters are
equipped with electric purge systems for scale removal. The system worked
satisfactorily except for a few meters that required periodic manual/chemical
scale cleaning. The most frequent cleaning was required on the bleed meters
and alkali addition meters. The frequency was approximately once every six
weeks and was normally performed at a convenient opportunity, e.g., the boiler
was off-line. Cleaning was accomplished by a HC1 acid wash and manual scrap-
ing of the electrode button with a soft wood stick. On large meters where a
man's hand can be inserted, fine sandpaper was used to remove tenacious scale.
23-25
-------
The electronic components of the meter have worked well. Of all the meters
at the test facility perhaps three per year need to have relay contacts
cleaned or repaired. However, the meters were considered reliable, acceptably
accurate, and relatively easy to service.
23.8.6 Density Meters
Both Dynatrol Model CL-10HY U-tube density meters and Ohmart radiation density
meters are used at Shawnee. Both types of meters provided acceptably accu-
rate and dependable service. The instruments were most useful in monitoring
trends in density. From an operations point of view, the U-tube meters had
some initial problems with line plugging. The cause was attributed to operator
error in setting too low a flow through the instrument. The appropriate cor-
rective measures consisted of manually setting the flow through the instrument
at 8 gpm. Care was taken to avoid high flow setting because the U-tube would
erode and fail.
23.9 MATERIALS AND EQUIPMENT EVALUATION
During the test period, the following equipment/materials were monitored:
• Lining and piping
• Control valves
• Strainers
23-26
-------
23.9.1 Lining and Piping
23.9.1.1 TCA tower lining
Hardness measurements were made of the tower lining in March 1977. Measure-
ments were made in areas in contact with spheres or areas receiving direct
spray impingement. The range of readings was 44 to 57 Durometer which corres-
ponds closely to the readings published in the Third Interim Report on Corro-
sion Studies (Reference 3, Appendix L). Generally, the lining was in good
condition although the surface where scale was often removed exhibited scabs
from chipping tools. There was no evidence of rubber lining separation from
the metal shell.
23.9.1.2 PVC pipe
A number of system modifications required the installation of plastic piping.
Schedule 80 PVC 1120 (ASTM D-1785) piping of 6- and 8-inch sizes was installed
for a large section of the main slurry line for the TCA. Fittings were all
of standard radius design. PVC piping also served the venturi/spray tower
oxidation tank system. The operation of the PVC piping was satisfactory from
a mechanical standpoint.
23.9.1.3 Rubber-Lined pipe
During the March 1977 boiler outage, blisters were observed in some sections of
the rubber-lined pipe. To date, these blisters have not precipitated pipe fail-
ures or blockage problems and therefore do not pose an operating or maintenance
problem at this time.
23-27
-------
23.9.1.4 Pullman Kellogg Test Panel
A test panel (20 inches x 15 inches x 3/8 inch) supplied by Pullman Kellogg
and made of carbon steel coated with fluoroelastomer CXL-2000 was installed
in the inlet gas duct of the TCA scrubber before the start of Run 705-2A in
January 1977. It was arranged to have cooling slurry spray impinge across
the panel face. Inspection of the panel on May 2 showed that the coating had
failed along the exposed flanged edge allowing erosion of the carbon steel
substrate. The panel had been exposed for approximately 2000 hours of service.
The panel was removed; the material was judged to be unsatisfactory in this
particular service and no further testing was planned.
23.9.1.5 Test Pipe - Bondstrand
A section of U-shaped Bondstrand fiberglass-reinforced plastic pipe was re-
moved from service in March 1977. It had been installed in the suction line
of the TCA main slurry pump. The estimated total service period for the
pipe was approximately 8800 hours. There did not appear to be any signifi-
cant wear of the pipe surface, and the entire section was considered to be
in excellent condition.
23.9.2 Control Valves
23.9.2.1 Valtek 6-inch butterfly control valve
Since the inspection of September 1976, further erosion had taken place in the
body of this valve. Previously observed grooves had deepened, and the worst
single hole located in the disk plane was 0.28 inch deep. Measurements at
23-28
-------
the outer edge of the stallite disk ranged from 0.22 to 0.24 inch. The valve
was judged conditionally acceptable for slurry service; the valve disc does
wear and will need replacement eventually.
23.9.2.2 Fisher 2-inch butterfly control valve
This valve had a fishtail disk which was judged to be in excellent condition.
Erosion had caused only slight rounding of sharp edges of the disk and pitting.
Minor erosion had also taken place in the body of the valve in the plane of the
closed disk position. The estimated maximum depth of wear was 0.5 mm. The
valve body appeared to be wearing faster than the disk in the throttling
service. During March 1977, the valve was moved from the venturi/spray tower
to the TCA.
23.9.2.3 Durco 6-inch butterfly valve
Located in the venturi/spray tower, this valve was considered to be in excel-
lent condition, except for four small irregular wear patterns at the disk edge
possibly caused by debris'trapped at the sealing edge during closure. The
valve body liner of abrasive-resistant polyethylene was in excellent condition.
Measurements of the disk thickness showed no change or loss of the coating
material. The valve service was primarily in the open position.
23.9.3 Strainers
Hayward strainers in both systems were inspected in November 1977. The 8-inch
size strainers on the TCA showed erosion at the basket support ledge and at
23-29
-------
the lower half of the strainer outlet. Similar erosion occurred in the 6-inch
size strainer used on the venturi/spray tower.
Elliott strainers located in the main slurry line to the venturi/spray tower
were one of the two original sets. There did not appear to be further erosion
of the strainer body at'the flange locations where this type strainer had
previously leaked.
All the strainers required infrequent maintenance (once or twice per year)
that consisted of basket replacement (most frequent) and bead-welding of the
eroded basket support ledges. The strainers were judged to be very useful and
well worth the cost from the standpoint of preventing scale/solids accumula-
tions in the spray nozzles.
23.9.4 Corrosion Studies
The Fifth Interim Report on Corrosion Studies at the test facility, written
by TVA, is presented in Appendix K. The report covers the operating period
from October 1976 through February 1978 and discusses corrosion and erosion
*
of the test facility equipment and test specimens.
23-30
-------
Section 24
REFERENCES
1. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Summary of
Testing through October 1974. EPA Report 650/2-75-047, June 1975.
2. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program, First Progress Report, EPA Report 600/2-75-050, September 1975.
3. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program, Second Progress Report, EPA Report 600/7-76-008, September 1976.
4. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program, Third Progress Report, EPA Report 600/7-77-105, September 1977.
5. Bechtel Corporation, Shawnee Chemical Procedures Laboratory Manual,
March 1976.
6. Borgwardt, R.H.. Sludge Oxidation in Limestone FGD Scrubbers, EPA
Report 600/7-77-061, June 1977.
7. Barrier, J.W., et al., Comparative Economics of FGD Sludge Disposal,
presented at the 71st Annual Meeting of the Air Pollution Control
Association, Houston, Texas, June 25-30, 1978.
8. Gladkii, A.V., et al., "State Scientific Research Institute of Industrial
Gas Cleaning (Moscow)" report for Protocol Point A-l, Development of
Lime/Limestone Scrubbing for Stack Gas Desulfurization, US/USSR Sulfur
Oxides Technologies Sub Group, 1974.
24-1
-------
9. Rhudy, R.G., and Head, H.N., Results of Hue Gas Characterization Testing
at the EPA Alkali Wet-Scrubbing Test Facility, presented at the Second
Fine Particle Scrubber Symposium, New Orleans, Louisiana, May 2-3, 1977.
10. Kilmer, V.J., and Alexander, L.T., "Methods of Mechanical Analysis of
Soils". Soil Science, 68:,15-24 (1949).
24-2
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-244a
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
EPA Alkali Scrubbing Test Facility: Advanced
Program, Fourth Progress Report; Volume 1.
Basic Report
5. REPORT DATE
November 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Harlan N. Head and Shih-Chung Wang
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Bechtel National, Inc.
50 Beale Street
San Francisco, California 94119
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-1814
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Periodic; 11/76 - 6/78
14. SPONSORING AGENCY CODE
EPA/600/13
is.SUPPLEMENTARY NOTES IERL-RTP project officer is John E. Williams, Mail Drop 61,
919/541-2483. Earlier progress reports are EPA-600/2-75-050, -600/7-76-008, and
-600/7-77-105.
is. ABSTRACT The report gives results of advanced testing (late-November 1976 - June
1978) of 30,000-35,000 acfm (10 MW equivalent) lime/limestone wet scrubbers for
SO2 and particulate removal at TVA's Shawnee power station. Forced oxidation with
two scrubber loops was developed on the venturi/spray tower system with limestone,
lime, and limestone/MgO slurries. Bleed stream oxidation was successful only with
limestone/MgO slurry. Forced oxidation with a single scrubber loop was developed
on the TCA system with limestone slurry. Other test blocks on the TCA were lime-
stone with low fly ash loadings, limestone type and grind, automatic limestone feed
control, limestone reliability, limestone with Ceilcote egg-crate type packing, lime/
MgO, and flue characterization.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Pollution
Flue Gases
Scrubbers
Tests
Calcium Oxides
Calcium Carbonates
Sulfur Oxides
Dust
Aerosols
Magnesium Oxides
Pollution Control
Stationary Sources
Particulate
13B
21B
07A,13I
14B
07B
11G
07D
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
344
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 O-73)
24-3
------- |