EPA-600/2-76-235
November 1976
Environmental Protection Technology Series
PHYSICAL-CHEMICAL TREATMENT OF A
MUNICIPAL WASTEWATER USING POWDERED CARBON
No. I
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-600/2-76-235
November 1976
PHYSICAL-CHEMICAL TREATMENT OF
A MUNICIPAL WASTEWATER USING
POWDERED CARBON
NO. II
by
Don E. Burns
Richard N. Wallace
Darrell J. Cook
Eimco BSP Division of Envirotech Corporation
Salt Lake City, Utah 84110
Contract No. 68-01-0183
Project Officer
James J. Westrick
Wastewater Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal
Environmental Research Laboratory—Cincinnati, Ohio,
U. S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the
contents necessarily reflect the views and policies
of the U. S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute
endorsement or recommendation for use.
-------
FOREWORD
The Environmental Protection Agency was created because of
increasing public and government concern about the dangers of
pollution to the health and welfare of the American people.
Noxious air, foul water, and spoiled land are tragic testimony
to the deterioration of our natural environment. The complexity
of that environment and the interplay between its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in
problem solution and it involves defining the problem, measuring
its impact, and searching for solutions. The Municipal Environ-
mental Research Laboratory develops new and improved technology
and systems for the prevention, treatment, and management of
wastewater and solid and hazardous waste pollutant discharges
from municipal and community sources, for the preservation and
treatment of public drinking water supplies, and to minimize
the adverse economic, social, health, and aesthetic effects of
pollution. This publication is one of the products of that
research; a most vital communications link between the researcher
and the user community.
This report summarizes the results of a pilot plant program
which developed the use of chemical clarification and powdered
activated carbon adsorption for the treatment of municipal waste-
water. The development of such alternative processes provides
additional tools for use by water pollution control agencies in
their efforts to maintain and improve the quality of the environ-
ment.
Francis T. Mayo
Director
Municipal Environmental
Research Laboratory
Cincinnati
111
-------
ABSTRACT
Salt Lake City municipal wastewater was treated in a nominal
100 gpm pilot plant by chemical coagulation-precipitation,
powdered activated carbon adsorption and granular media
filtration. Chemical-primary sludge was gravity thickened
and vacuum filter dewatered. Spent carbon was gravity thick-
ened, vacuum filter dewatered and thermally regenerated in
a fluidized bed furnace. Solids-contact units were used for
chemical treatment and carbon contacting.
The pilot plant was operated over a 16 month period to demon-
strate treatment effectiveness under diurnal flow conditions.
Reuse of thermally regenerated carbon was practiced for 3
months.
A high quality effluent was consistently produced, similar
to that expected for conventional biological treatment
followed by tertiary treatment for phosphorus and suspended
solids removal.
Spent powdered carbon was effectively regenerated with fixed
carbon recovery of 90 percent on the average.
Soluble organic materials were found to be removed by a com-
bination of chemical coagulation, anaerobic biological
activity and adsorption on powdered carbon.
Lime and alum-primary sludges were effectively dewatered by
vacuum filtration.
This report was submitted in fulfillment of Project No. 17020
HNH, Contract No. 68-01-0183 by the Eimco BSP Division of
Envirotech Corporation under the sponsorship of the Environ-
mental Protection Agency. Pilot plant work was completed
December, 1973.
IV
-------
CONTENTS
Page
Disclaimer ii
Foreword iii
Abstract iv
Figures vi
Tables xiv
Acknowledgements xviii
Sections
I Conclusions 1
II Recommendations 3
III Introduction 4
IV Pilot Plant Facilities 8
V Operating Procedures 27
VI Pilot Plant Performance 36
VII Sludge Treatment 100
VIII Powdered Carbon Regeneration igg
IX Carbon System Response 240
X PAC-PCT Process Design Recommendations 258
XI Economic Analysis 279
XII References 291
XIII Glossary 293
XIV Appendices 297
v
-------
FIGURES
No. Page
4-1 Overall Photo of PAC-PCT Pilot Plant 9
4-2 Process Flow Diagram 10
4-3 Eimco Reactor-Clarifier, Solids- 12
Contact Chemical Treatment Unit
4-4 Sludge Treatment System 15
4-5 Chemical Sludge Surge Tank and 16
Thickener
4-6 Carbon Make-up System 19
4-7 Granular Media Filter 21
4-8 Powdered Carbon Regeneration System 23
4-9 Composite Sampling and Process 26
Turbidity Monitoring Device
6-1 Chronology of Plant Operation 38
6-2 Effluent History Curve for Operating 40
Period I (July 13-September 30, 1972)
6-3 Effluent History Curve for Operating 44
Period II (January 6-February 20, 1973)
6-4 Effluent History Curve for Operating 47
Period III (March 28-June 7, 1973)
6-5 Effluent History Curve for Operating 50
Period IV (June 10-August 13, 1973)
6-6 Soluble Chemical Oxygen Demand Removal ^
Pattern During Initial Start-up of
Chemical Contactor with FeClo
6-7 Effect of Aeration on Iron 54
Precipitation
vi
-------
Figures (continued)
No. Page
6-8 Effluent History Curve for Operating 58
Period V (August 14-September 15, 1973)
6-9 Effluent History Curve for Operating 61
Period VI (September 28-November 8,
1973)
6-10 Daily Variation in Lime Primary 65
Sludge Level and Effluent Suspended
Solids at Constant Flow
6-11 Daily Variation in Lime Primary Sludge 66
Level and Effluent Suspended Solids at
2:1 Diurnal Flow
6-12 Typical Hourly Variations in Lime 68
Primary Sludge Concentration and
Effluent Turbidity
6-13 Daily Variation in Alum Primary Sludge 71
Level and Effluent Suspended Solids
For Chemical Contactor
6-14 Typical Diurnal Flow Patterns 72
6-15 Typical Hourly Variations in Alum 73
Primary Sludge Concentration and
Effluent Turbidity
6-16 Daily Variation in Alum Primary Sludge 75
Concentration at Bottom Sample Tap
6-17 Daily Variation in Ferric Primary 76
Sludge Concentration and Effluent
Suspended Solids
6-18 Daily Variation in Carbon Sludge Level 78
and Effluent Suspended Solids for
Carbon Contactors
6-19 Typical Hourly Variation in Carbon 80
Sludge Concentration
6-20 Daily Variation in Carbon Sludge 82
Concentration at the Bottom Tap
6-21 Effect of Filtration Rate on 84
Suspended Solids Removal Efficiency
VII
-------
Figures (continued)
No. Page
6-22 Terminal Head Loss Pattern for 86
Coarse Bed Filter
6-23 Effect of Polyelectrolyte on Terminal 87
Head Loss Pattern
6-24 Tri-Media Filter Bed Design 89
6-25 Typical Normalized Head Loss 91
Build-up Pattern
6-26 Terminal Head Loss versus Bed 92
Depth Pattern
6-27 Typical Head Loss versus Bed 96
Depth Patterns
7-1 Lime Primary Sludge Thickening - 105
Thickener Solids Profile
7-2 Lime Primary Sludge Thickening - 107
Laboratory Predicted Maximum Under-
flow Relationships
7-3 Lime Primary Sludge Thickening - no
Effect of Polyelectrolyte on
Predicted TSL
7-4 Lime Primary Sludge Thickening - 111
Effect of Polyelectrolyte on Settling
Rate
7-5 Lime Primary Sludge Thickening - 112
Comparison of Laboratory and Pilot
Plant Results
7-6 Alum Primary Sludge Thickening - 117
Thickener Solids Profile
7-7 Alum Primary Sludge Thickening - 120
Maximum Laboratory Underflow
Concentration
7-8 Alum Primary Sludge Thickening - 122
Predicted TSL versus Feed Suspended
Solids Concentration at various Under-
flow Concentrations
vin
-------
Figures (continued)
No.
7-9 Alum Primary Sludge Thickening -
Predicted TSL versus Feed Suspended
Solids Concentration at Various
Underflow Concentrations
7-10 Ferric Primary Sludge Thickening - 127
Thickener Solids Profile
7-11 Carbon Sludge Thickening - 134
Thickener Solids Profile
7-12 Carbon Sludge Thickening - 135
Thickener Solids Profile
7-13 Carbon Sludge Thickening - 135
Thickener Solids Profile
7-14 Carbon Sludge Thickening - Maximum 139
Underflow Concentration
7-15 Carbon Sludge Thickening - Labora- 149
tory Predicted TSL versus Feed
Suspended Solids Concentration @
120 g/£ Underflow Concentration
7-16 Carbon Sludge Thickening - Labora- 141
tory Predicted TSL versus feed
Suspended Solids Concentration @
140 g/£ Underflow Concentration
7-17 Carbon Sludge Thickening - Labora- 142
tory Predicted TSL versus Feed
Suspended Solids Concentration @
Maximum Underflow Concentration
7-18 Lime Primary Sludge Filtration - 150
Pilot Plant Results
7-19 Lime Primary Sludge Filtration, 152
Pilot Plant Results - Effect of
Feed Suspended Solids Concentration
on Cake Percent Moisture
7-20 Lime Primary Sludge Filtration, Leaf 156
Test Results - Effect of High Feed
Suspended Solids Concentration
IX
-------
Figures (continued)
No. Page
7-21 Lime Primary Sludge Filtration, Leaf 158
Test Results - Effect of Feed Suspended
Solids Concentration
7-22 Lime Primary Sludge Filtration, Leaf 159
Test Results - Cake Moisture Content
7-23 Alum Primary Sludge Filtration, Leaf 160
Test Results - Effect of Media and
Chemicals
7-24 Alum Primary Sludge Filtration - 167
Pilot Plant Results
7-25 Alum Primary Sludge Filtration, Leaf 169
Test Results - Effect of Elapsed Time
on Cake Weight
7-26 Alum Primary Sludge Filtration - Effect 17.2
of Feed Suspended Solids and Lime on
Cake Moisture Content
7-27 Alum Primary Sludge Filtration - 173
Cake Moisture Content
7-28 Alum Primary Sludge Filtration, Leaf 174
Test Results - Effect of Lime Dosage
7-29 Alum Primary Sludge, Laboratory 176
Results - Effect of Sludge pH on
Aluminum Solubility
7-30 Ferric Primary Sludge Filtration, Leaf 178
Test Results - Effect of Lime Dosage
7-31 Ferric Primary Sludge - Effect of 179
Lime Dosage
7-32 Carbon Sludge Filtration, Pilot Plant 188
Results - Effect of Form Time on
Solids Capture
7-33 Carbon Sludge Filtration, Pilot Plant 189
Results - Cake Moisture Content
7-34 Carbon Sludge Filtration - Virgin 19°
Carbon Results
x
-------
Figures (continued)
No. Page
7-35 Carbon Sludge Filtration, Leaf Test
Results Effect of Form Vacuum
7-36 Carbon Sludge Filtration, Pilot Plant 193
Results - Effect of Feed Suspended
Solids Concentration
8-1 Fluidized Bed Furnace - Campaign
Number 1
8-2 Fluidized Bed Furnace - Campaign
Number 2
210
8-3 Comparative Equilibrium Adsorption
Isotherm Tests: Regeneration Run
Number 5
91 7
8-4 Fluidized Bed Furnace - Campaign
Number 3
O ~\ C.
8-5 Comparative Equilibrium Adsorption
Isotherm Tests: Regeneration Run
Number 6
8-6 Comparative Equilibrium Adsorption
Isotherm Tests: Regeneration
Run Number 7
8-7 Comparative Equilibrium Adsorption 220
Isotherm Tests: Regeneration Run
Number 8
8-8 Comparative Equilibrium Adsorption 221
Isotherm Tests: Regeneration Run
Number 9
8-9 Comparative Equilibrium Adsorption 222
Isotherm Tests: Regeneration Run
Number 11
8-10 Comparative Equilibrium Adsorption 226
Isotherm Tests: Regeneration Run
Number 10
8-11 Relationship Between Iodine Number 227
and Volatile Solids
XI
-------
Figures (continued)
No. Page
8-12 Relationship Between Regenerated and 229
Spent Carbon Percent Volatile Solids
8-13 Relationship Between Regenerated and 230
Spent Carbon Iodine Numbers
8-14 Furnace Feed Rate versus FBF 233
Operating Factor
9-1 SCOD Removal-Adsorption Response: 245
Pilot Plant Results Using Virgin
Carbon
9-2 SCOD Removal-Adsorption Response; 248
Combined Results of Current and
Previous Studies
9-3 SCOD Removal-Adsorption Response: 249
Pilot Plant Effluent SCOD Response
to Carbon Dosage
9-4 SCOD Removal-Adsorption Response: 251
Laboratory Equilibrium Adsorption
Isotherms
9-5 SCOD Removal-Adsorption Response: 254
Comparison of Regenerated and
Virgin Carbon
9-6 Ultraviolet Light Adsorbance versus 257
SCOD Concentration
11-1 Estimated Treatment Costs versus 290
Plant Size - With Sludge Incineration
and Carbon Regeneration
A-l Sludge Thickening Test Apparatus 304
A-2 Sludge Dewatering Test Apparatus 306
B-l Graphical Technique for Laboratory 310
Gravity Thickening Results
B-2 Reduced Lab Thickening Test Data 312
C-l Laboratory Study 316
XII
-------
Figures (continued)
No. Page
C-2 UV versus COD Process Curves 317
C-3 UV versus BOD Process Curves 313
C-4 UV versus TOC Process Curves 319
C-5 Selectivity Due to Treatment Effects 321
C-6 Specific Treatment Point Correlation 323
C-7 Diurnal Changes in the Nature of 324
Organic Concentrations
C-8 ABS and Urine UV Spectra 325
C-9 UV Spectra Shift Due to Change in 326
Relative Organic Concentration
xzn
-------
TABLES
No. Page
5-1 Scheduled Analytical Tests of Process 34
Stream Composite Samples
6-1 Operating Period I: Average Operating 39
Conditions and Treatment Results July 13
to September 30, 1972
6-2 Operating Period II: Average Operating 43
Conditions and Treatment Results
January 6 to February 20, 1973
6-3 Operating Period III: Average Opera- 46
ting Conditions and Treatment Results
March 28 to June 7, 1973
6-4 Operating Period IV: Average Opera- 49
ting Conditions and Treatment Results
June 10 to August 13, 1973
6-5 Plant Grab-Sample Soluble Iron 51
Profiles
6-6 Operating Period V: Average Operating 57
Conditions and Treatment Results
August 14 to September 15, 1973
6-7 Operating Period VI: Average Opera- 60
ting Conditions and Treatment Results
September 28 to November 8, 1973
6-8 Pertinent Operating Conditions for 77
Periods Shown in Figure 6-18
7-1 Lime-Primary Sludge Thickening 104
Average Pilot Plant Results
7-2 Lime-Primary Sludge: Laboratory
Thickening Test Results
7-3 Lime-Primary Sludge Thickening:
Effect of Polyelectrolyte
xiv
-------
Tables (continued)
No. Page
7-4 Alum-Primary Sludge Thickening: 115
Average Pilot Plant Results
7-5 Alum-Primary Sludge: Laboratory 118
Thickening Test Results
7-6 Alum-Primary Sludge: Laboratory 119
Thickening Test Results
7-7 Ferric-Primary Sludge Thickening; 126
Average Pilot Plant Results
7-8 Ferric Chloride-Primary Sludge: 128
Laboratory Gravity Thickening Test
Results
7-9 Ferric Chloride-Primary Sludge: 129
Laboratory Gravity Thickening Test
Results
7-10 Spent Carbon Sludge Thickening: 132
Average Pilot Plant Results
7-11 Carbon Laboratory Gravity 137
Thickening Test Results
7-12 Lime-Primary Sludge: Vacuum 145
Filtration Pilot Plant Results
7-13 Arbitrary Filter Cake Discharge 149
Code
7-14 Lime-Primary Sludge Filtration: 153
Ratio of Pilot Plant to Leaf Test
Results
7-15 Lime-Primary Sludge: Leaf Test 155
Result Using Polyelectrolyte
7-16 Alum-Primary Sludge: Pilot Plant 162
Vacuum Filtration Results
7-17 Alum-Primary Sludge Filtration: Ratio 168
of Pilot Plant to Leaf Test Results
7-18 Ferric-Primary Sludge: Vacuum ]_81
Filtration Pilot Plant Results
xv
-------
9-3 Regression Analysis Log X/M on Log Ce
Tables (continued)
No. Page
7-19 Carbon Sludge: Vacuum Filtration 184
Pilot Plant Results
7-20 Spent Carbon Sludge Regeneration:
Ratio of Pilot Plant to Leaf Test
Results
7-21 Suggested Vacuum Filter Design and 197
Performance for PAC
8-1 FBF Operating Conditions: 201
Campaign #1
8-2 Carbon Characteristics: 203
Campaign #1
83- FBF Operating Conditions: 208
Campaign #2
8-4 Carbon Characteristics: 209
Campaign #2
8-5 FBF Operating Conditions: 213
Campaign #3
8-6 Carbon Characteristics: 214
Campaign #3
8-7 FBF Operating Conditions: 224
Virgin Carbon
8-8 Carbon Characteristics: 225
Run #10 and Virgin Carbon
9-1 Carbon Treatment System: Summary 241
of Single-stage Data
9-2 Carbon Treatment System: Summary 242
of Two-Stage Counter-Current Data
246
10-1 Chemical Treatment: Recommended 265
Design Parameters
10-2 Alum-Primary Sludge: Design 267
Recommendation for Sludge Disposal
xvi
-------
Tables (continued)
No. Page
10-3 Carbon Treatment: Process Design 273
Recommendation
278
10-4 Carbon Regeneration System:
Recommended Process Design
11-1 Assumed Average Raw Wastewater and 279
Effluent Water Qualities
11-2 Assumed Unit Cost for Economic 280
Analysis
11-3 Design Parameters and Major Equip- 281
ment Sizing for Chemical Treatment
11-4 Estimated Alum Treatment Costs 283
11-5 Design Parameters and Major Equip- 285
ment Sizing for Powdered Carbon
Treatment
11-6 Estimated Carbon Treatment Costs 287
11-7 Summary of Physical-Chemical Treatment 289
Costs - With Alum Treatment, Sludge
Incineration and Carbon Regeneration
B-l Laboratory Gravity Thickening Data 313
Reduction: Graphical Construction
Relationships
314
B-2 Error in Predicted Thickener Solids
Loading at Various Desired Underflow
Concentrations
xvn
-------
ACKNOWLEDGEMENTS
The support and indulgence of the Salt Lake City Commissioner
of Water and Sewers, city engineers and municipal sewage
pumping plant personnel is acknowledged.
The contract manager for Eimco, Mr. G. L. Shell, who
administered this and a previous research contract on the
same subject, is deserving of a special acknowledgement.
Acknowledgement is due Messrs. Keith. Mounteer, Steve Nelson,
Ron Sorensen, and Doug Hendry who nursed the pilot plant
through start-up and kept it operating under adverse con-
ditions.
A special acknowledgement is also due Mrs. Glenna Bitton who
typed the first draft of this report and prepared many other
reports associated with this contract, and Mrs. Anne Medlock
who retyped the entire report into its final form.
The support, counsel and patience of the project officer,
Mr. James Westrick and Mr. Jesse M. Cohen of the Wastewater
Research Division, Municipal Environmental Research Laboratory,
Cincinnati, Ohio, Environmental Protection Agency, is acknowl-
edged with sincere thanks.
xvi 11
-------
SECTION I
CONCLUSIONS
With the completion of this study essentially all of the
originally defined project objectives were achieved. Specific
pertinent conclusions deduced from the results of this study
are as follows:
1. When employing alum or FeCl3 pretreatment, two stage
counter-current powdered carbon contacting (at 100 mg/£
carbon) and typical diurnal flow conditions, the
powdered activated carbon - physical chemical treatment
(PAC-PCT) process evaluated consistently produced a high
quality effluent of less than 5 mg/5, chemical oxygen
demand, 5 mg/£ suspended solids and 0.3 mg/£ phosphorus.
2. When properly operated, the solids-contact treatment
units employed in the chemical and powdered carbon con-
tacting-clarification steps performed very well during
diurnal flow conditions.
3. Suitable design and operating conditions were identified
for the critical final liquid-solids separation step of
granular media filtration as evidenced by long term pro-
duction of highly clarified pilot plant effluent.
4. Powdered carbon regeneration using a fluidized bed
furnace (FBF), with off-gas recycle, resulted in fixed
carbon recoveries, on the average, in excess of 90
percent.
5. Thermally regenerated powdered carbon did not exhibit
full recovery of adsorptive capacity but when reused in
the PAC-PCT process no significant loss of treatment
effectiveness was found.
6. Acceptable recoveries of fixed carbon was not achieved
with the FBF unit used when employing the originally
designed bed injection gas principal of operation.
-------
7. Soluble organics (COD) were found to be removed by three
mechanisms in the PAC-PCT process; namely, chemical
coagulation, anaerobic biological activity and adsorption
on activated carbon.
8. For the system studied, use of alum or ferric chloride
in the chemical treatment step resulted in significant
removal of soluble COD, whereas use of lime did not.
9. No significant treatment or cost benefit of counter-current
staging of powdered carbon contacting was precisely de-
fined.
10. Lime and alum-primary sludge mixtures, produced in the
chemical treatment step, were found to be effectively con-
centrated by gravity thickening and dewatered by vacuum
filtration.
11. Ferric-primary sludge thickened well but was found very
difficult to dewater by vacuum filtration.
12. Laboratory and pilot plant gravity thickening test re-
sults were not consistently correlated. Short comings in
the laboratory procedures and/or inconsistent operation
of the pilot plant units were considered primary causes.
13. Laboratory leaf tests and pilot plant vacuum filtration
test results were found to be precisely correlated.
14. The use of ultraviolet light absorption monitoring of
carbon system feed or effluent was not found to provide
useful real-time operational intelligence.
15. Alum-powdered carbon treatment of 10 mgd of Salt Lake
City municipal wastewater, with thermal regeneration
and reuse of carbon, to produce an effluent containing
less than 5 mg/£ COD, 5 mg/£ SS and 0.5 mg/£ phosphorus
was estimated to cost 33.4 £/1000 gal. If powdered
carbon was not regenerated but used only once, the total
treatment cost was estimated to be 36.2 C/1000 gal.
-------
SECTION II
RECOMMENDATIONS
1. It is strongly recommended that the Environmental Pro-
tection Agency (EPA) promote, and if necessary, fund
construction and preliminary operation of a 3 to 10 mgd
PAC-PCT plant to treat a municipal wastewater similar to
Salt Lake City, Utah. Such a project would commercialize
PAC-PCT technology developed by the EPA and others by
providing full scale demonstration of treatment effective-
ness and also allow determination of any undefined opera-
tional or performance problems.
In addition, if a flexible thermal regeneration system
were constructed, full scale optimization of powdered
carbon regeneration could be accomplished.
2. It is recommended that a cost sensitivity study of the
PAC-PCT process be conducted to identify cost sensitive
areas which might be justifiable 'future development
efforts.
3. It is recommended that chemical-primary sludge dewatering
alternatives be evaluated on a real time basis. Such
alternatives should include, at least, vacuum filtration,
centrifugation and pressure filtration.
4. It is recommended that a pilot plant study be conducted
comparing parallel granular and powdered carbon systems
following chemical treatment-clarification of a raw
municipal wastewater.
-------
SECTION III
INTRODUCTION
Physical-Chemical Treatment (PCT) of wastewater, as dis-
cussed herein, consists of chemical coagulation-precipita-
tion followed by powdered activated carbon (PAC) contact-
ing. PCT was originally developed as a tertiary method
for reclamation of secondary biological effluents. In
recent years, PCT has been evaluated as a method for treat-
ment of raw wastewaters. Because of well developed re-
generation technology, major emphasis has been placed on
studying granular activated carbon (GAG) contacting. The
use of PAC has several real and potential advantages when
compared to GAG.1 They are as follows:
1) The cost of powdered carbon is substantially
less.
2) Powdered carbon adsorption will equilibrate with
soluble wastewater organics in a small fraction
of the time.
3) Powdered carbon might be supplied on demand to
meet varying feed organic strength.
4) Powdered carbon systems require a fraction of
the carbon inventory.
5) Powdered carbon contacting systems are more
amenable to control of undesirable biological
activity.
Environmental Protection Agency (EPA) funded studies have
indicated that thermal regeneration of PAC is technically
feasible and may be economically justified.2, 3 In 1969,
Eimco Division of Envirotech Corporation and EPA negoti-
ated a contract to conduct an extensive pilot plant study
for the purpose of demonstrating the operability and treat-
ment effectiveness of the PAC-PCT process applied to a
municipal wastewater. A nominal 100 gpm constant flow
pilot plant was constructed in Salt Lake City, Utah, in
-------
1969. This pilot plant was operated during 1970 and the
first half of 1971. The broad objectives of this study were
as follows:
1) To determine the relative treatment and cost effec-
tiveness of lime, alum and ferric chloride chemicals
for precipitating phosphorus and removal of suspend-
ed solids in solids-contact treatment units.
2) To determine the effect of carbon dosages and num-
ber of counter-current stages of contacting in
solids contact treatment units on the removal of
soluble chemical oxygen demand (SCOD).
3) To demonstrate the effectiveness of carbon regen-
eration by a fluidized bed furnace system and the
reuse effectiveness of regenerated carbon.
4) To compile process operating and efficiency data
suitable for design and economic evaluation of a
full scale PAC-PCT plant.
Most of the above objectives were achieved during this study.
Lime, alum and ferric chloride were all found to be effective
in the chemical pretreatment step. Alum treatment of 10 mgd
of Salt Lake City wastewater for phosphorus removal and clar-
ification was estimated to be less expensive than ferric
chloride or lime treatment. The PAC system effluent SCOD
was found to be relatively insensitive to the carbon dosage
employed. Use of two-stage counter-current contacting was
found to require less carbon than single-stage contacting.
However, precision of pilot plant test results was such that
a high degree of statistical significance was not determined
for this difference. Regeneration of spent carbon in the
fluidized bed furnace system resulted in about 40 to 60 per-
cent loss of carbon.
The results of this study demonstrated the operability and
treatment effectiveness of the PAC-PCT process. Estimated
economics indicated that the process may be an economic al-
ternative to presently available treatment approaches for
certain wastewaters. However, several areas needed further
study to develop new and/or more reliable design data. A
follow-on contract was negotiated for an additional year of
studies based on those areas needing further work. Due to
the unsuccessful performance of the fluidized bed carbon re-
generation system, as originally designed and installed, a
3-month contract extension was negotiated. The fluidized
bed furnace was slightly modified, allowing achievement of
efficient regeneration and the opportunity to recycle and
re-use regenerated carbon.
-------
OBJECTIVES
Objectives of the current study were as follows:
1) Demonstration of treatment effectiveness by long
term pilot plant operation under diurnal flow con-
ditions. Specifically, the effect of diurnal flow
on the performance of the solids-contact treatment
units was to be determined. It was also desired to
demonstrate the long term consistency of pilot plant
effluent quality.
2) Continued development of the fluidized bed furnace
system to obtain fixed carbon losses of less than
25 percent. The quality of regenerated carbon, as
determined by re-use in the pilot plant, was to be
determined.
3) Obtain pilot plant performance results for gravity
thickening and vacuum filtration dewatering of chem-
ical-primary sludges (lime, alum and ferric chlor-
ide) . Laboratory thickening and vacuum filter leaf
test results would also be obtained. Correlations
between laboratory and pilot plant results would be
defined.
4) Evaluate the utility of certain continuously moni-
toring water quality evaluation instruments for
assisting operation of the PAC-PCT process. Speci-
fic instruments to be used would be ultra violet
light absorption, orthophosphate, turbidity and pH
measuring devices.
APPROACH
The contract work was divided into two phases of operation.
The first phase was to last approximately 3 months. During
this time, the pilot plant, which had been dormant for app-
roximately 9 months, was renovated and modified. Modifica-
tion of existing facilities and some new construction were
completed. In addition, process evaluation and control in-
struments were purchased and installed.
After the pilot plant was made operable, several weeks were
devoted to training plant operating personnel. During the
remainder of Phase I, the pilot plant was operated under
constant flow conditions with lime pretreatment and single-
stage carbon contacting. Major emphasis was placed on pro-
ducing spent carbon for regeneration operations. The purpose
was to expedite development work on the fluidized bed furnace
system.
-------
Phase II of the contract work comprised a 12-month period
which included a 3-month extension. The major purpose of
Phase II plant operation was to obtain long term operational
and performance data. The pilot plant was operated under
diurnal flow conditions with various peak to minimum flow
ratios. During three major pilot plant operating campaigns,
which were approximately 3 months each, either lime, alum or
ferric chloride pretreatment of the raw municipal wastewater
was used. This schedule was to allow near equal compilation
of continuous gravity thickening and intermittent vacuum fil-
tration data on the chemical-primary sludges.
Two-stage counter-current contacting of the carbon was used
during most of Phase II, since it had previously been identi-
fied as the most economical mode of operation.1 A carbon
dosage of about 100 mg/Jl was used. If pilot plant effluent
SCOD concentrations of less than 15 to 20 mg/£ had not been
obtained, then a higher carbon dosage of approximately 300
mg/£ would have been used.
Granular media filter bed design would be evaluated in an
attempt to maximize efficiency of filter operation.
A major activity during Phase II would be the compilation of
carbon regeneration and re-use data. Of course, this activ-
ity had to be preceeded by successful development of the
fluidized bed furnace system.
-------
SECTION IV
PILOT PLANT FACILITIES
A photo of the PAC-PCT pilot plant facility is shown on
Figure 4-1. The liquid treatment process flow diagram for
this facility is shown on Figure 4-2.
Raw Salt Lake City municipal wastewater was chemically treat-
ed to precipitate phosphorus and coagulate suspended solids.
The chemically treated wastewater was solids-contacted and
gravity settled. The clarified wastewater constituted the
PAC treatment system feed.
Chemical contactor underflow sludge was sent to a chemical-
primary sludge treatment system. This system consisted of
continuous gravity thickening, thickened sludge storage and
intermittent vacuum filtration.
Carbon was used for the sole purpose of removing soluble or-
ganic material. The carbon contacting and clarification
system consisted of single or two-stage counter-current con-
tacting in solids-contact treatment units followed by granu-
lar media filtration. Spent carbon was removed as carbon
contactor underflow and discharged to a gravity thickener.
Thickened spent carbon was stored prior to being thermally
regenerated in a fluidized bed furnace system. All carbon
containing streams were recycled to the first-stage carbon
contactor (e.g., thickener overflow, vacuum filter filtrate
and granular media filter backwash water).
The following subsections are a detailed description of the
major components of the pilot plant.
RAW WASTEWATER FLOW
The pilot plant influent was coarse screened and comminuted
Salt Lake City raw wastewater. About 350 gpm was diverted
from the Salt Lake City pump station discharge header to the
pilot plant raw wastewater sump. Flow not pumped to the
pilot plant was returned to the pump station wet well through
a gravity overflow. During the first 7 months of pilot plant
-------
FIGURE 4-1: OVERALL PHOTO OF PAC-PCT PILOT PLANT
-------
Alum
or
Fed,
FIGURE 4-2:
PROCESS FLOW DIAGRAM
Polymer Lime
CHEMICAL
CONTACTOR
RAW SEWAGE
Coarse Screened
and Comminuted
Overflow
Plant
Effluent
Note:
® Through (gs
Stream Sampl-
ing Locations
CLEARWELL
T 1 T T T
Vacuum Filter
Filtrate
Carbon
Thickener Carbon Regen-
Overflow eration
System
Spillage
f
l^-l
Backwash
Pump
GRANULAR
MEDIA FILTER
FILTER
BACKWASH
COLLECTION
TANK
-------
operation, all rags and most of the grit in the raw waste-
water were allowed to go to the chemical treatment unit.
Because of rag clogging problems in the thickened sludge
handling system pumps, a "de-ragging" screen was placed in
the raw wastewater sump for the last 8 months of operation.
This screen had 1/2-inch square openings.
The raw wastewater pump was sized for a flow in excess of
200 gpm. It was driven by a 7-1/2 Hp, 180-volt, DC shunt
wound motor. The motor received a 0-180 volt signal from
an electronic control device, which inturn received a signal
from a cam-operated pneumatic-electrical controller. The
cam was on a 24-hour clock. The cam shape was a variable
allowing any pattern of diurnal flow variation desired. The
actual pattern of flow used was identical to that for Salt
Lake City municipal wastewater.
Flow from the raw wastewater pump to the chemical treatment
unit passed through a magnetic flow meter. A signal from
this flow sensor was transmitted to a flow indicating, re-
cording, totalizing and control unit.
CHEMICAL TREATMENT SYSTEM
Chemical Treatment Unit
Chemical treatment and clarification of raw wastewater was
accomplished in a 12-ft diameter by 10.5-ft side water depth
Reactor-Clarifier treatment unit.*
Alum or FeCl3 was fed to the wastewater at B and flash mixed
at U (capital letters refer to locations on Figure 4-3). If
polyelectrolyte was used it was added at D. When lime treat-
ment was used, the raw wastewater flow bypassed the flash-
mix tank and went directly to the draft tube E. Lime was
added at C. This resulted in the lime contacting copious
amounts of recycled solids simultaneously with the raw waste-
water.
A slow speed turbine F pumped raw wastewater, chemicals and
previously formed chemical precipitates up the draft tube
and out into the reaction zone. The pumping turbine capacity
was adjustable in the range of about 5 to 12 times the aver-
age hydraulic throughput of the unit. Reaction zone slurry
*Manufactured by Eimco BSP, Division of Envirotech Corporation,
Salt Lake City, Utah
11
-------
FIGURE 4-3:
EIMCO REACTOR-CLARIFIER, SOLIDS-
CONTACT CHEMICAL TREATMENT UNIT
SYMBOLS
A. Influent Pipe
B. Inorganic Coagulant Feed
C. Lime Feed
D. Polymer Feed
E. Draft Tube
F. Pumping Turbine
G. Sample Point, Reaction Zone
H. Sludge Rake
I. Sludge Thickening Cone
J. Expanding Area Clarification Zone
K. Liquid Surface Skimmer
L. Float Baffle
M. Peripherial V-Notched Launder
N. Acid Feed Point
0. Effluent to First Carbon Contactor
P. Effluent to Second Carbon Contactor or Drain
Q. Reaction Zone Cone
R. Sludge Slowdown Line
S. Timer Controlled Sludge Valve
T. Sample Taps
U. Chemical Addition Flash Mixer
W. Flow Splitter Box, Effluent Sample Point
X. Acid Neutralization Flash Mixer
Y. Baffle Plate
Z. Scum Box
12
-------
flowed under the reaction zone cone Q into an increasing
area upflow clarification zone. Settled solids were moved
to the center of the unit by sludge rake H and either drawn
up the draft tube or removed through the thickening cone I.
Partially thickened sludge was automatically removed from
the treatment unit by gravity flow through line R, and an
air operated, timer controlled sludge valve S. Chemical
sludge blowdown was routed to the chemical sludge handling
system.
Since the chemical treatment unit received raw wastewater,
provision was made for clarification zone surface skimming
by mechanism K. Floatable materials were kept from passing
over the peripherial launder M (a V-notched weir) by baffle
L. Effluent neutralization was accomplished by addition of
concentrated sulfuric acid to the effluent in a flash mix
tank at location X. Clarified effluent exited the unit at
splitter box W, which divided flow to one or two carbon con-
tactors or to drain. A sample of clarifier effluent was
continuously withdrawn at point X and conveyed by gravity
flow to a composite sampler and process turbidimeter.
Except for the following modifications, the solids-contact
unit was unchanged from that used in previous contract work.
The original square turbine blades were replaced with long
rectangular blades which extended from the periphery of the
draft tube to next to the rake drive shaft. This modifica-
tion eliminated ragging problems previously experienced.
During alum and ferric chloride treatment campaigns, a
baffle plate Y was installed in the solids-contact units.
It was anticipated that this baffle would hydraulically iso-
late the turbulent reaction zone and sluge outlet region.
Higher underflow sludge concentration than was experienced
during previous contract work was expected.
Flash mixing units were installed to improve the reaction
of inorganic coagulants and neutralizing acid with the waste-
water. During previous contract work coagulant was injected
into the influent pipe just before the draft tube and acid
was added to the launder 180 degrees from the effluent
splitter box.
The sludge blowdown line R was changed from a 2 to a 4-inch
diameter pipe. This modification was for the purpose of min-
imizing clogging of the smaller line with rags, grit and for-
eign objects.
13
-------
Chemical Sludge Handling System
The chemical sludge handling system is shown on Figure 4-4.
Primary-chemical sludge was automatically blown down at
various time frequencies from the solids-contact unit to
a surge tank at a rate of about 200 gpm for about 8 seconds
duration. A liquid level control operated Moyno Pump fed
the gravity thickener at a rate of 3 to 5 gpm until the surge
tank was emptied. In May 1973, a 3.5-ft diameter tank was
placed inside the original 8-ft diameter thickener. Figure
4-5 is a picture of the surge tank-thickener arrangement.
Sample taps were located at 6-inch intervals for slurry pool
level monitoring and sampling. The overflow from the thick-
ener was passed through a coarse screen to a surge tank.
From here the overflow was automatically pumped through a
1-inch totalizing water meter to the chemical treatment unit
reaction zone liquid surface.
The thickener drive rotated the rakes and pickets at 16 rph.
This resulted in a 6.7 and 2.9 fpm rake tip speed for the
8 and 3.5-ft diameter thickeners, respectively.
Originally, thickener blowdown was automated to coincide
with blowdown from the chemical treatment unit. The duration
of blowdown was adjusted by the operators to result in a de-
sired volume of sludge. Because of the inability to main-
tain a desired near constant sludge pool depth in the 3.5-
diameter thickener using the automated approach, manual blow-
down was used.
Thickener underflow was routed, by gravity, to a 560-gal
sludge storage and inventory tank. This vessel contained a
radial flow turbine mixer to insure uniform suspension of
the sludge prior to obtaining samples for laboratory analy-
sis. The volume of sludge was determined by depth measure-
ments. All chemical-primary sludge produced was collected,
volume inventoried and samples obtained for solids analysis.
When vacuum filtration studies were programmed, the inven-
toried sludge provided the feed. When vacuum filtration was
not programmed, the sludge was wasted to drain after volume
inventorying and sampling. The vacuum filter used was the
same unit used for the carbon regeneration system and will
be described in a later section.
Chemical Storage and Handling
High purity alum and ferric chloride was purchased in dry
form. Alum was obtained in 100-lb bags and the ferric chlor-
ide in 350-lb drums. These chemicals were intensely mixed
14
-------
Level Control
Switch for
Moyno Pump
\
FIGURE 4-4:
SLUDGE TREATMENT SYSTEM
Flow Meter
Recycle to
Chemical
Contactor
55 Gal. Drum
Surge Tank
Submersible
^ : __ Pump with
Pressure
Control Switch
dla x6'
SWD Surge Tank
Moyno Pump
With Variable
Drive
Sludge from
Chemical
Contactor
dia x 6'
SWD Chemical
Sludge
Thickener
3' dia x 3'
Face Vacuum
Filter
Moyno Pump
with Variable
Drive
4'-3" dia x6
SWD Sludge
Storage Tank
Recycle
Belt Wash
-------
FIGURE 4-5: CHEMICAL SLUDGE SURGE TANK AND THICKENER
-------
with tap water in a 55-gal drum to allow complete solution.
The concentrated solution was then gravity fed to a 560-gal
rubber lined tank where it was mixed with dilution water to
achieve the desired feed concentration. The stock feed
solution was fed to the flash mix tank by positive displace-
ment metering pump through PVC piping.
Concentrated sulfuric acid was purchased in bulk quantities
and stored in a 1000-gal covered mild steel tank. The acid
was fed to the flash mix neutralization tank by a positive
displacement metering pump through black iron piping. An
automatic acid feed system consisted of a process pH meter,
recorder and controller. The pH probe was located in a con-
tinuously flowing sample of neutralization tank overflow.
This placement resulted in a control response lag time of
less than 3 minutes. The controller had an adjustable low pH
set point, below which the acid feed pump would turn off.
As the pH increased above the set point, the acid feed pump
would turn on. Manual trim of the acid pump rate allowed
minimizing frequent OFF-ON pump action.
High purity hydrated lime (Ca[OH]2) was obtained in 50-lb
bags. Dry lime was continuously slurried with water and
pumped through plastic tubing by a positive displacement
pump to the chemical treatment unit draft tube. The lime
slurry pump was operated at a constant rate sufficient to
prevent deposition in the feed lines. Lime dosage was ad-
justed manually by a stroke adjustment on the dry lime feed-
er drive. During diurnal flow operation the dry lime feeder
drive motor speed was automatically paced to plant flow.
A 560-gal (4.24-ft diameter) polymer make-up tank was pro-
vided for dissolving dry polymer into pilot plant effluent
or tap water and storage of polymer solution during use.
Polymers were dissolved by aspiration into the make-up water.
High speed mixers (1750 rpm) were used to aid the solution
of polymers and provide intermittent mixing of the tank con-
tents. Polymer solution was fed by positive displacement
metering pumps to the chemical treatment units.
All chemical feed pumps, carbon feed pumps and process stream
samplers were driven by DC motors. The same voltage was
applied to all motors and came from a single electronic con-
troller. This controller received a signal via a pneumatic-
electrical converter from the raw wastewater flow recorder-
controller. Thus all reagent pump speeds were proportioned
to raw wastewater flow.
17
-------
CARBON TREATMENT SYSTEM
Carbon Contactors
Carbon treatment of chemically clarified raw wastewater was
accomplished in one or two 10-ft diameter by 10«5-ft SWD
solids-contact treatment units. These units were identical
to the chemical treatment unit (Figure 4-3) except that they
had no surface skimmers, flash mixers, or deflection plates,
the turbines were not modified, and the effluent launders
were of the submerged orifice type. The units were plumbed
to be operated either in parallel, as single-stage carbon
contactors, or in series, as a two-stage counter-current
contacting system.
Carbon feed slurry (virgin or regenerated) was fed to the in-
fluent line of the carbon contactors by a positive displace-
ment metering pump. When the two carbon contactors were
operated in series, the virgin carbon slurry was fed only to
the influent line of the second-stage unit. Underflow from
the second-stage unit was intermittently pumped by a Moyno
Pump to the influent line of the first-stage unit. Spent
carbon was automatically blown down from the first-stage
unit by gravity flow to the spent carbon slurry thickener
(to be described later).
Carbon Make-Up and Feeding
Carbon was purchased in 35 or 50-lb bags. The type of virgin
carbon used was Aqua-Nuchar.*
The virgin carbon make-up and feed tank is depicted on Figure
4-6. The 840-gal rubber lined and covered tank was outfitted
with a two prop axial flow mixing impeller. During and just
after addition of virgin carbon the air above the liquid was
sucked through a water scrubber to remove carbon dust parti-
cles.
Routine carbon make-up procedure entailed first filling the
tank about one-fourth full of tap water. The water door
spray flow and vacuum pump were then started. Carbon was
manually poured, very slowly, into the tank. After the de-
sired weight of carbon had been added and wetted, tap water
was added to a predetermined depth to provide a 5 percent
by weight carbon feed slurry.
*A product of Westvaco, Covington, Virginia
18
-------
FIGURE 4-6:
CARBON MAKE-UP SYSTEM
Vacuum Pump
Virgin
Carbon
Water Door
Air
19
-------
Regenerated carbon slurry of about 15 to 20 percent total
solids was stored in either a 560 or 840-gal tank containing
axial flow mixing turbines. When required, a measured
volume of regenerated carbon slurry of known concentration
was transferred to the carbon make-up and feed tank and di-
luted with tap water to about 5 percent by weight slurry
concentration. From the known total and fixed carbon solids
values, the carbon slurry feed rate was established to pro-
vide a desired fixed carbon dosage to the carbon contactors.
Granular Media Filtration (GMF) Station
Effluent from one of the carbon contactors flowed by gravity
to the granular media filter. The filter was 3.5-ft diameter
by 6-ft high and of conventional downflow gravity design with
a coal-sand filter bed. The specific details of the depth
and size of media used will be presented in Section VI of
this report.
Figure 4-7 is a schematic of the granular media filtration
station. In the filtering mode, influent flows through a
U-tube to the filter inlet. The dual-media filter bed rested
directly on a false bottom of Pora-Septa* media. Flow was
collected in the filter outlet plenum and routed to an eff-
luent weir box located at an elevation which insured that
liquid in the filter bed was never at less than atmospheric
pressure. Filter flow rate was indicated by an orifice plate
by-pass flow meter and an orifice plate and differential
pressure cell arrangement for remote read-out. A sample of
filter effluent was continuously pumped from the effluent weir
box to a composite sampler and process turbidimeter.
Filtered water flowed from the effluent weir box to a 5000-
gal (12-ft diameter) clearwell which provided treated water
storage for plant use, filter backwash, and heat exchanger
water for the carbon regeneration system.
A column of pressure taps spaced at 6-inch intervals was lo-
cated on the filter housing. Pressure differentials between
the outlet plenum and each pressure tap were automatically
measured and recorded once each hour for a duration of 8 to
10 minutes. The filter was manually backwashed when the total
filter headloss (including underdrain) reached a value of
about 7 feet of water. The operational sequence during back-
washing was presented in a previous report.1 Backwash water
Manufactured by Multi-Metal Wire Cloth Incorporated,
Tappan, New York
20
-------
Influent
To Backwash
Equalization Tank
FIGURE 4-7:
GRANULAR MEDIA FILTER
1)
4J
P
0
_£
01
d IT)
r
1 i
i
i
:c
1 G UlUa 1 1 L- Hi-i-cL. LX. XL-aJ.
inverter
___.
X '
A i
i i
i i
1 r--
1 I i
i i_
i
1 *
1
I
1
[1
LI
Multi
Recor
Orifice
Plate
Indicators
|
3.5 ft dia
Filter
To Process
Sampler
Pneumatic
Electrical
I
i L°
|
C
±ru
E
V
IV
/
ert
er
ffluent
alve
1
1
u
n
1 1
1 — n
r
Weir Box
Drain Down
Valve
-------
was collected in a 3600-gal (10-ft diameter) equalization
tank which was equipped with a low speed, radial flow, mixing
turbine. Equalization tank contents were pumped to the re-
action zone of the first-stage carbon-contactor at a rate
of about 5 gpm.
Samples from the filter outlet plenum and just above the
coal-sand interface were continuously passed through two
Hach Low Range CR turbidimeters.* The electrical output sig-
nal from these turbidimeters, as well as flow orifice plate
and pressure tap pneumatic-electrical converters were con-
tinuously fed to a multiple-point recorder for display and
recording.
Carbon Regeneration System
The regeneration system consisted of four major unit opera-
tions: sludge concentration by gravity thickening, vacuum
filter dewatering, regeneration in a fluidized bed furnace
and off-gas carbon capture by wet venturi scrubbing. This
system is shown on Figure 4-8.
The spent carbon gravity thickener was a 5-ft diameter by
5-ft SWD unit. Thickener underflow was transferred to a
carbon inventory and sampling tank (about 500 gal). A small
Moyno Pump was used for this transfer operation. The volume
of thickened carbon slurry was measured. Then, after
thorough mixing, a representative sample was obtained for
laboratory tests. The inventoried, thickened carbon was
then transferred to a 1700-gal conical storage tank. To
prevent bridging in the conical storage tank, carbon was re-
cycled externally at a rate of about 5 gpm from the bottom
outlet to the top of the tank. The storage tank also con-
tained four air nozzles to allow for air mixing.
Vacuum filter dewatering was accomplished with a 3-ft dia-
meter by 3-ft face continuous belt filter."1" The vacuum
filter station consisted of the filter, flocculation tank,
vacuum pump and a single filtrate pump and receiver, all
mounted on a platform. All components of the vacuum filter
which contacted carbon slurry were made of corrosion resis-
tant thermoplastic. A rubber lined chemical flocculation
tank had two mixing compartments to achieve chemical disper-
sion and floe growth. A 55-gal drum, high speed mixer and
variable speed positive displacement diaphragm chemical feed
pump comprised the conditioning chemical make-up and feeding
facility.
* Manufactured by Eack Chemical Company, Ames, Iowa
+ Manufactured by Eimca ESP Division of Envirotech
Corporation, Salt Lake City, Utah
22
-------
1SJ
oo
FIGURE 4-8:
POWDERED CARBON REGENERATION
SYSTEM
Slowdown From
Carbon
Contactor
Recycle
Belt Wash
t ,_ To Inventory
Tank
Offgas
Blower
-------
The vacuum filter cake discharged into a hopper which had a
capacity of about 300 Ib of wet carbon filter cake. A ribbon
screw advanced the cake along the bottom of the hopper to the
throat of the furnace feed pump. The ribbon screw reconsti-
tuted the filter cake to a form of paste which the Moyno Pump
readily transported to the fluidized bed furnace.
All carbon bearing streams (e.g., vacuum filter filtrate,
belt wash, overflow, spillage and cleanup) were collected via
a floor drain sump and pumped to a 3600-gal (10-ft diameter)
inventory tank.
The outer dimensions of the carbon regeneration furnace were
5.25-ft diameter by 14-ft high. * The outer steel shell was
lined with suitable fire and insulating brick. A major modi-
fication was made to furnace internal geometry mid-way in
the contract period. An additional modification was made in
the method used for controlling fire box temperature and
fluidized bed 02 levels. These changes will be discussed in
Section VIII of this report. Thermocouples and pressure taps
were used to monitor temperature and pressure profiles. Stain-
less steel tuyeres (nozzles) were provided to retain the sand
and provide distribution of fire box gases into the sand bed.
The fuel used was natural gas. Combustion air was provided
by a 200-SCFM capacity blower. Hot gas flow from the fire
box at a superficial velocity of just over 1-ft per second
was required to maintain the 15 by 30 U.S. Mesh sand bed in
a fluidized state at operating temperatures.
The hot furnace off-gas, containing the regenerated carbon,
was quenched to about 200°F by the addition of about 15 gpm
of recycled water to the furnace outlet duct. The cooled
gases were then passed through a 6 and then a 4-inch diameter
wet venturi scrubber. These scrubbers, operated in series,
captured essentially all of the regenerated carbon. Scrubber
water flow was about 30 and 15 gpm to the 6 and 4-inch Ven-
turis, respectively. Scrubber water underflow was passed
through a cyclone separator (a 55-gal drum) for sand and
scale removal and then to a 1700-gal conical decant tank which
provided some separation of carbon. Decant tank overflow
was pumped through a heat exchanger prior to reuse as quench-
ing and scrubber water. Pilot plant effluent was used for
heat exchanger water. The fluidized bed furnace and venturi
scrubbing system were highly instrumented for safety and ease
of operation.
* Manufactured by BSP Division of Envirotech,
Brisbane, California
24
-------
COMPOSITE SAMPLER AND PROCESS TURBIDIMETER
An automatic sample compositor, and a turbidity measuring
and recording system was developed for this project. (See
Figure 4-9.) All process stream samples (see Figure 4-2 for
location) were piped to a central location. Each sample
stream passed upward into a hopper shaped container, sized
to maintain solids in suspension. Sample continuously over-
flowed the hopper via a 1-inch diameter horizontal pipe to
a common waste drain. Normally closed solenoid valves were
located in a vertical line just below the horizontal pipes.
Outlets from all five solenoids emptied into a common trough
which drained into a flow-through process turbidimeter. A
multiple cam timer programmed each solenoid to be opened for
12 to 15 minute intervals once each hour. Twenty-four hour,
flow proportioned composite samples were obtained from the
hopper container mentioned previously- Approximately 50 m£
of each sample was collected with a dip type sampler each
7 to 20 minutes. The actual frequence of sample pickup was
directly proportional to raw wastewater flow. Samples were
transferred to refrigerated storage vessels by gravity flow
through plastic hoses. Composited samples were refrigerated
at about 34°F until removed for analysis. The raw wastewater
was sampled separately with a dip type sampler located at
the raw wastewater sump.
LABORATORY FACILITIES
Analytical laboratory facilities were located at the pilot
plant site. The laboratory was equipped to conduct all
analysis reported except sulfides and dissolved metal analy-
sis. These analyses were made at the Envirotech Research
Center, Salt Lake City, Utah.
25
-------
FIGURE 4-9:
COMPOSITE SAMPLING AND PROCESS
TURBIDITY MONITORING DEVICE
RPM Proportional
To Plant Flow
Drain
Sample to
Refrigerator
(by gravity)
Process
Turbidimeter
Turbidity
Indicator
I
1
|
x<
\
o
Multi-Cam
Timer
Turbidity
Recorder
Sample
Inlet
26
-------
SECTION V
OPERATING PROCEDURES
The basic approach to pilot plant operation consisted of the
following steps: establish desired operating conditions,
operate the plant, collect operation data, collect samples,
analyze samples, and reduce data. No step was initiated un-
til the preceeding steps had been performed satisfactorily.
The desired operating conditions were established from the
contract statement of work, feedback from reduced data,
and/or operational planning. Some of these conditions were
pretreatment chemical type and dosage, diurnal flow pattern,
carbon dosage and number of carbon contacting stages, etc.
Plant operators would routinely monitor effluent clarity,
chemical dosages, sludge profiles, underflow sludge blowdown
rates, etc., to verify that specified operating conditions
were being met. After these conditions were met, operating,
inventory and mechanical data were observed and recorded.
When these data substantiated that specific operating con-
ditions were attained, sample collection was started.
The project chemist and plant operating engineer would
determine daily if samples were representative, prior to
laboratory analysis.
The operation data, inventory numbers, and laboratory results
were then reduced. Actual reagent dosages, sludge production,
pollutant removals, etc., were computed. Occasionally the
need for additional data, special studies, and operational
changes were identified by this routine data reduction.
CHEMICAL FEEDING
Reagent chemical stock flow to each unit was routinely measur-
ed (£/min) anc^ recorded about 6 times per day. In addition
all reagent stock solutions or slurries were volume inven-
toried once each day.
27
-------
Samples of alum and ferric chloride feed solutions were
occasionally analyzed for aluminum or ferric iron to deter-
mine actual concentrations. These concentrations were com-
pared with chemical makeup data to verify the reliability
of the latter.
Available lime determinations were occasionally made when
using lime. The total number of bags used and their con-
tents weight was recorded when added to the dry lime feeder.
Routine grab samples of the dry lime feed were weighed and
the feed rate in g/min recorded.
The number of bags used for carbon makeup were recorded along
with the manufacturer's listed bag weight (35 or 50 Ib). On
one occasion, a weight check indicated that the actual car-
bon weight was about 15 percent less than the manufacturer's
listed weight for a given bag. Samples of the carbon feed
slurry were occasionally grab sampled and suspended solids
determined to verify the computed feed slurry concentration
based on makeup data. When using regenerated carbon the
feed slurry was made up from concentrated slurries of stored
regenerated carbon by dilution to the desired feed strength.
SOLIDS-CONTACT UNITS
Instantaneous flow to the chemical treatment unit was auto-
matically measured and displayed on a recorder. Flow to
subsequent units was determined by subtracting routinely
measured process sample flows from the displayed plant
influent flow.
The approach to operation of the solids-contact treatment
units was presented in a previous report.1 A brief reit-
eration follows. It is important to keep in mind that the
type of unit used could be operated either as a solids-
contact or a solids-contact and sludge blanket clarifier.
Sludge blanket clarification is defined as maintenance of
at least two or more feet of slurry in the clarification
zone. With reference to Figure 4-3, this means a slurry
height at least two feet above the lower extremity of the
reaction zone skirt. During this current study no sludge-
blanket operation was specified, only solids-contact treat-
ment - with gravity sedimentation.
Operation of the solids-contact units basically involved
routine monitoring of sludge concentration profiles within
the three units. Two operational specifications were given
to the operators. The first was to maintain an approximate
specified solids concentration within the reaction zone.
28
-------
The second was to keep the top of the sludge slurry at or
slightly below the lower extremity of the reaction zone skirt.
An indirect method for determining sludge concentration was
used. This method consisted of collecting two liters of
sludge in a 2-liter graduated cylinder. The sludge was allow-
ed to subside quiescently for either 5 or 10 minutes. The
time was chosen such that the slurry interface was subsiding
at a rate of less than about 1-inch per minute. For the car-
bon and lime sludges, 5 minutes were adequate. For the alum
and ferric sludges, 10 minutes were required. At the desired
time interval, the volume of settled sludge was observed
and the percentage of the original sample volume recorded.
This measurement was called the 5 or 10 minute settled sludge
volume.
Approximately six times per day (twice per shift), the oper-
ators would determine settled sludge volume for samples from
all three solids-contact units. These samples were taken
at tap locations G and T on Figure 4-3. The values obtained
were compared with the previously mentioned two operational
specifications. If the slurry was at an undesirable high
level, the rate of sludge withdrawal (blowdown) was increased.
If the slurry level was too low, the rate of sludge with-
drawal was reduced or temporarily shut off.
Infrequently, the pumping turbine and sludge rake rotational
speeds were adjusted by the operators, if too high or too
low a sludge concentration was observed for the reaction
zone samples. Adjustments in turbine and rake speeds were
very infrequent, sometimes not being required for 30 to 60
days. Sludge blowdown rate adjustments were more frequently
required, sometimes being daily and other times being only
weekly or monthly.
GRANULAR MEDIA FILTER
As indicated in Section IV, the granular media filter (GMF)
was highly automated for backwashing and monitoring of per-
formance (e.g., flow rate, headless and turbidity). The
approach to operation of the filter station was mainly one
of visual observations for operational problems, calibration
of sensors and determining the need for backwashing.
About six times per day operators observed filter flow rate
and headloss and analyzed effluent grab samples for turbidity
and light transmittance. The operator determined the need
for backwashing by observation of either a total bed headloss
greater than about 85 inches of water or break-through of
carbon fines. Penetration of carbon into the sand layer was
visually observed through a plexiglass window. Carbon fine
29
-------
penetration into the sand coupled with an effluent quality
deterioration trend, observed by continuous turbidity moni-
toring, indicated break-through. After manual initiation of
the automatic backwash controller, the operator usually
observed the filter backwash sequence. Rates and durations
of initial water level drain down, air and water flows were
frequently noted. Filter bed expansion during water back-
wash was routinely noted. Abnormalities in the "quality of
media fluidization" or the presence of any "mud balls" were
also noted.
Just prior to and shortly after backwashing the incremental
headless profile through the filter bed was frequently ob-
served by the operator. A water manometer was used. Head-
loss profile at the end of a filter cycle indicated the
solids removal patterns. Headless profile at the start of
the filter run was used to verify that effective backwashing
was routinely being obtained.
GRAVITY SLUDGE THICKENERS
Operation of the chemical-primary and spent carbon sludge
thickeners consisted of routine monitoring of sludge pool
depths. Depths were denoted as being the highest sample tap
at which copious amounts of sludge was observed. The
operators were instructed to maintain the sludge level be-
tween two specified sample taps. This was accomplished by
manual operation of thickener blowdown.
At about 4-hr intervals, 200-mA grab samples of clarifier
underflow sludge and thickener overflow were obtained and
composited over a 24-hr period. Suspended solids analysis
of these samples indicated thickener solids captures.
When specified, 5 gal of 4-hr grab composite samples of
clarifier underflow sludge were obtained by the operators.
These samples were used to perform laboratory thickening
tests. Usually, on the same day, a grab sample from each
pilot plant thickener sample tap was obtained and total sus-
pended solids determinations conducted. This data constitu-
ted the thickener solids profile (concentration versus depth).
VACUUM FILTER
Prior to the start of pilot plant vacuum filter operation for
each type of sludge studied, filter leaf tests were run on
composite sludge samples of gravity thickener underflow. A
0.1 sq ft filter leaf was used to determine the necessary
type and dosage of chemical conditioners and an appropriate
filter media type.
30
-------
Just before the start of a vacuum filter run, suspended
solids of stored and mixed gravity thickener underflow was
determined. Chemicals were made up, and dosage and flow re-
quirements determined. The sludge was then pumped approxi-
mately 40 ft to the flocculator where chemicals were added
to and mixed into the sludge.
In order to maintain a constant drum submergence, and thus
constant form and dry times, sludge was added to the vacuum
filter faster than it could be dewatered. The excess sludge
overflowed to recycle or waste. Drum rotation speed was
started at the slowest possible level to achieve the thickest
possible cake. The speed was then increased incrementally,
taking data at each speed, until the cake was to thin to dis-
charge. At this point, the filter run was concluded. This
approach resulted in identification of the maximum possible
yield for a dischargeable filter cake.
At each vacuum filter drum speed, the following samples were
collected: composited samples of sludge feed before and
after chemical addition, cake samples using a "cookie cutter"
having an area of 0.085 sq ft, and a composited sample of
filtrate. The difference between the suspended solids in the
feed before and after chemical conditioning indicated solids
increase due to the addition of chemicals. The difference
between the suspended solids in the feed sludge before chem-
ical conditioning and the filtrate indicated vacuum filter
solids capture. Solids analysis of the filter cake samples,
taken by the "cookie cutter", were used to compute vacuum
filter yield and cake moisture content.
FLUIDIZED BED FURNACE
The operation of the carbon regeneration, system was on a
batch basis. Carbon losses were determined by conducting a
solids balance across the system. The quantity and charac-
teristics of spent carbon were determined prior to regenera-
tion run. All carbon not reaching the furnace was collected
and quantified. The difference was the carbon fed to the
furnace. Recovered carbon was collected, and the quantity
and characteristics determined. Carbon losses were based on
fixed carbon analysis. The characteristics of regenerated
carbon were measured by iodine number, molasses number, and
equilibrium adsorption isotherm tests and by reuse in the
pilot plant.
Safety procedures dictated the following start-up sequence
be followed for the fluidized bed furnace: scrubber water
on, burner air flow on, pilot gas flow on, pilot ignition
established, and finally burner gas flow on. The furnace
31
-------
was then brought up to operating temperature very slowly.
This normally took 3 to 4 days when starting at ambient
temperature.
To avoid carbon combustion in a fluidized bed furnace, there
should be less than about 1 percent oxygen by volume in the
fluidizing gas in the bed. This can be accomplished by
burning air and gas in the firebox stoichiometrically. How-
ever, construction materials are not available to handle the
resulting 3200°F temperature. Temperature in the firebox
can be controlled by using excess combustion air. Excess
air for cooling results in excess oxygen in the fluidized
sand bed where spent carbon is added. The original designers
of the fluidized bed unit felt that gas could be injected
into the bottom of the sand bed to scavenge the excess oxygen
so that carbon would not be burned. This principal of opera-
tion will be referred to as bed injection gas (BIG). Tempera-
ture control can also be obtained by burning gas and air
stoichiometrically and using an additional gas which contains
little or no oxygen for cooling. These were the two methods
of firebox temperature control used in this study.
When bed injection gas was used the following procedure was
followed to ready the furnace for operation. The air and gas
flow rates were set to provide the desired fluidization velocity.
Proportions of these flows were determined from previous
experience. Burner gas was then adjusted to provide a hot
gas flow to the bed with an oxygen content of less than 1
percent by volume. The off-gas recycle flow was trimmed
to maintain a firebox temperature under 1090°C, yet maintain
bed fluidization.
Water was injected into the bed prior to feeding carbon in
order to hold the bed temperature to within about 100°C of
the desired operating temperature. This water injection was
used during final trimming of the oxygen which was measured
in the stack gas.
The vacuum filter capacity was about four times greater than
the furnace capacity and, therefore, was run first to fill
the feed hopper and then as needed. When the oxygen had
been trimmed and the feed hopper was filling, the feed to
the furnace was started.
The feed pump to the furnace was automatically controlled
to maintain a set bed temperature. Minimal operator atten-
tion was needed during the run because of the automatic
features of the furnace control. These features would shut
the furnace down and/or sound an alarm to alert the operator
if any abnormal condition occurred. Some of these conditions
32
-------
were: low scrubber water flow, low burner air flow, flame
out or high bed temperature.
During the course of a regeneration run, pressures and temp-
eratures at various points through the furnace were recorded
(i.e., firebox, bottom of bed, upper bed, free-board). The
stack gas was routinely analyzed for oxygen content. The
flow through the heat exchanger was regulated to maintain a
scrubber water temperature of 40 to 60°C.
At the conclusion of a regeneration run, the carbon cake feed
was allowed to run out and the feed hopper, pump and piping
was flushed with water. The furnace was shut down according
to the manufacturer's instructions. About 3 to 4 days were
required to cool the furnace and sand bed to a temperature
approaching ambient.
Collection of regenerated carbon was accomplished by the
following technique. Carbon in the scrubber water decant
tank was allowed to settle and then clean supernatant de-
canted. Water was run through the scrubber system flushing
out all deposited carbon to the decant tank. All deposition
points (i.e., sand separator, launders, etc.) were flushed
with water and all carbon returned to the decant tank. The
carbon in the decant tank was again allowed to settle and
clear supernatant again decanted. The remaining carbon was
transferred to an inventory tank where it was mixed, volume
inventoried and sampled. The regenerated carbon was then
pumped to the carbon makeup and feeding tank as needed.
LABORATORY OPERATIONS
The spectrum of analytical tests routinely run on auto-
matically composited process stream samples is shown in
Table 5-1. Sampling and laboratory test procedures are
presented in Appendix A. These routine analyses were run 4
days per week. In addition to process stream samples, daily
grab composites of all solids-contact reaction zone and blow-
down sludge samples, and thickener overflows required sus-
pended solids analysis. All chemical stock feeds, and
chemical-primary and. spent carbon thickener blowdowns were
inventoried, sampled and appropriate mineral or suspended
solids determinations made when scheduled.
Operation of the vacuum filter and the carbon regeneration
system were on an intermittent basis. Each of these opera-
tions generated about 20 to 40 samples for suspended solids
and, for a few samples, mineral analysis. Spent and regen-
erated carbon samples were analyzed for volatiles, ash and
fixed carbon solids and other adsorptive characteristics.
33
-------
TABLE 5-1: SCHEDULED ANALYTICAL TESTS OF PROCESS STREAM COMPOSITE SAMPLES
u>
Raw
Water Quality Parameter* Wastewater
Turbidity a
Suspended Solids a
Light Transmittance
Total & Soluble Phosphorus a
Total & Soluble Ortho- a
Phosphate
Chemical Oxygen Demand a
Soluble Chemical Oxygen a
Demand
Biochemical Oxygen Demand b
Soluble Biochemical Oxygen
Demand
pH a
Alkalinity d
Total Hardness d
Calcium Hardness d
Sulfides b
Total & Soluble Iron or c
Aluminum
a - Routine analysis on 24 or 48 hour
Chemical
Treatment
a
a
a
a
a
a
a
b
-
a
d
d
d
b
c
composites
Carbon
First
Stage
a
a
a
-
-
—
a
-
b
a
-
-
-
b
c
Contactor GM
Second
Stage
a
a
a
-
-
—
a
-
b
a
—
-
-
b
c
Filter
a
a
a
a
a
a
a
b
—
a
d
d
d
b
c
b - No more frequently than once per week
c - Intermittant during FeCl3 or alum
campaign
d - Twice per week during lime campaign
* - Soluble pollutants were defined as those passing
a 0.45
micron membrane
filter
-------
The laboratory personnel had to handle these and other
special laboratory tests as time was available.
Laboratory gravity thickening and vacuum filtration leaf tests
were conducted by plant operators and/or project engineers.
35
-------
SECTION VI
PILOT PLANT PERFORMANCE
INTRODUCTION
The purpose of this section is to present and discuss the
liquid treatment performance of the PAC-PCT pilot plant.
The liquid phase of the plant includes the unit operations
of chemical treatment, carbon contacting and granular media
filtration (see Figure 4-2). Pilot plant effluent quality
and consistency will be emphasized. Detailed discussions of
sludge treatment, carbon regeneration and carbon system re-
sponse are presented in subsequent Sections.
Pilot plant performance for various operating conditions
will be presented and discussed. The data presented repre-
sents all available performance data except for some data
collected in November and December 1972. During November
and December 1972, a special study to identify the effect
of anaerobic biological activity on soluble organic removal
in the carbon contactors was conducted. The two carbon con-
tactors were run in parallel, one with and the other without
carbon feed. Results are presented in Section IX.
The effect of diurnal flow variations and operation on the
performance of the solids-contact units will be discussed
relative to effluent quality, sludge depth and solids con-
centration profiles, and down time due to treatment upsets.
The performance of granular media filtration will be discuss-
ed relative to the effects of bed design and operating con-
ditions on suspended solids removal efficiency and patterns.
It should be recalled that an objective of this study was to
evaluate the effectiveness of various filter bed designs.
During the first half of the study, a very coarse coal media
was used. This filter bed produced less than satisfactory
suspended solids removal which affected pilot plant effluent
quality.
36
-------
PILOT PLANT PERFORMANCE
Figure 6-1 is a chronology of plant operation indicating
major treatment variables and unit operations. Six oper-
ating periods were identified on the basis of uniform
chemical treatment, flow variation (i.e., diurnal pat-
tern) and/or carbon contactor operation (i.e., number
of stages and carbon dosage). The operating periods
will be presented chronologically. Average operating
conditions, and influent and effluent quality, were
determined for each period. Pilot plant effluent
history curves were plotted to show the consistency of
treatment achieved.
Operating Period I
Operating Period I was from July 13 to September 30,
1972, during Phase I of the contract work. The plant
was operated at constant flow with sufficient hydrated
lime added to achieve an operating treatment pH of 10.8
to 11.2. Constant flow was used because diurnal flow
equipment was not yet installed. High carbon dosages and
parallel single-stage carbon contacting were used in order
to produce the maximum possible amount of spent carbon for
regeneration studies. Development of the carbon regenera-
tion system was a major objective of Phase I studies. The
effluent from only one of the carbon contactors was filter-
ed through a "course bed" granular media filter.
Table 6-1 presents average pertinent plant operating con-
ditions and raw wastewater, chemical contactor, and plant
effluent water quality. Figure 6-2 is a plant effluent
history curve showing suspended solids, total and soluble
COD and total phosphorus.
The granular media filter was operated intermittently from
August 8 to September 17 because of new instrumentation
installation, and electrical and mechanical problems.
Because of these problems, filtered plant effluent data
was not available for the entire operating period. Hence,
plant effluent total COD is not plotted for the entire
period because total COD analysis was not performed on
carbon contactor effluents due to the presence of carbon
suspended solids which contribute to COD.
Upon start-up of a chemical treatment solids-contactor,
it is sometimes difficult to keep the dilute suspensions
of fresh floe out of the clarification zone. After ex-
periencing this problem, the flow was reduced by half to
allow the chemical treatment unit to build a sludge
37
-------
FIGURE 6-1:
CHRONOLOGY OF PLANT OPERATION
CO
CO
Flow - Peak s Minimum
Gravity Thickening
Vacuum Filtration Runs
Carbon System - Staging
Virgin or Regenerated Carbon
Carbon Dosage - mg/1
CC-1
CC-2
Granular Media Filtration
Carbon Regeneration Runs
Contract Phase
-------
TABLE 6-1: OPERATING PERIOD I: AVERAGE OPERATING CONDITIONS AND
TREATMENT RESULTS JULY 13 TO SEPTEMBER 30, 1972
to
Chemicalss Hydrated Lime:
H2S04:
Average Hydraulic Loading;
Carbon Contacting Mode;
Operating Conditions
@ 460 mg/£ (10.8 to 11.2 treatment pH)
@ 2030 Ib/MG =244 mg/£
Chemical Contactor 0.94 (0.49^) gpm/sq ft
Carbon Contactor 0.66 (0.32a) gpm/sq ft
GM Filter 2.1 gpm/sq ft
Single stage
Carbon Dosage:
Average
Process Treatment Results
Turbidity, JTU
Suspended Solids, mg/A
% Transmittance
Total P, mg/S, P
Soluble Total P, mg/Jl P
Total aP04, mg/Jl P
Soluble, aP04, mg/i P
COD, mg/&
SCOD, mg/£
pH, units
Methyl Orange Alkalinity53
Phenolphtalein Alkalinity^
Total Hardness^
Calcium Hardness*3
First Contactor
Second Contactor
Raw
Wastewater
33
87
-
4.50
3.16
2.82
2.25
136
30
7.5
293
417
255
Chemical
Effluent
4
14
-
0.29
0.14
0.10
0.04
-
-
10.2
116
66
239
197
280 mg/£
300 mg/£
Neutralized
Chemical
Effluent
4
10
-
-
0.15
0.06
0.05
55
35
7.1
65
317
284
Plant
Effluent
3
5
97
0.20
0.11
0.11
0.08
14
11
7.2
58
328
250
a - 7/29 to 8/9 only
b - mg/£ as CaCOs
-------
FIGURE 6-2: EFFLUENT HISTORY CURVE FOR OPERATING PERIOD I
(J1UV 13 - SltmiBSR 30, 1972)
10 15 20 25 30 1
22 25 30 1 5
July Aug
40
-------
inventory. This took several days and is indicated as the
lower hydraulic loadings presented in Table 6-1.
The neutralized chemical treatment contactor effluent was not
analyzed for total phosphorus because it was equal to that
of the chemical treatment contactor effluent. The pH values
shown in Table 6-1 are for stored composite samples and not
instantaneous measurements. The 10.2 pH value reported for
the chemical treatment contactor effluent is 0.6 to 1.0 unit
below the actual operating pH.
A slight solubilization of COD was experienced across the
chemical treatment step (i.e., increase in SCOD). The car-
bon contactors removed about 72 percent of the SCOD fed to
them. Carbon contactor effluent suspended solids averaged
32 to 36 mg/£, most of which was carbon.
During this operating period, the pilot plant removed 94
percent suspended solids, 96 percent total phosphorus, 90
percent total COD, and 21 percent hardness. Water stability
analysis were made in an attempt to quantify the very stable
character of several grab samples of chemical contactor eff-
luent. The "Marble Test" was used. This test determines
the difference, in pH units, between the observed pH and the
pH of stability (pHs). If the difference is positive there
is a depositing tendency and if the difference is negative
(observed pH less than pHs) there is a tendency of dissolving
calcium carbonate depositions.
Analysis of five grab samples collected on three different
days indicated a range of Marble Test stability of from
minus 0.08 to plus 0.02 pH units with an overall average of
0.00 pH units.
After 3 months of continuous operation using lime, no signifi-
cant mineral deposition had occurred on the chemical contac-
tor tank walls or launders. Some deposition of calcium car-
bonate was found on the inside of the reaction zone skirt
near the issue of the pumping turbine. This deposition
caused no operational or mechanical problems.
Generally, the average effluent qualities were considered
good. However, the poor granular media filter performance
caused considerable variability in the effluent quality.
Operating Period II
The second pilot plant operating period was from January 6
to February 20, 1973. The pilot plant was operated at con-
stant flow due to instrumentation failure. New equipment
41
-------
was installed and debugged in October, 1972. As mentioned
before, special studies were conducted in November and Decem-
ber, 1972. This accounts for the time lapse between operat-
ing Period I and II.
During Period II, sufficient hydrated lime was added
to produce an operating treatment pH of 10.8 'to 11.2.
Parallel single-stage carbon contacting was used. Effluent
from one of the carbon contactors was filtered through a
"coarse bed" granular media filter.
Table 6-2 presents pertinent operating conditions and a tab-
ulation of average plant influent and effluent and chemical
contactor effluent water quality. As indicated in Table 6-2,
different carbon dosages were used in the two carbon contac-
tors. Figure 6-3 is the plant effluent history curve of this
operating period. Due to operational problems and filter
failure, the filter was only operated about 1/3 of this oper-
ating period. However, even when the filter was on stream
it did a poor job of filtration. This can be seen in
Figure 6-3 by considering the high filter effluent suspended
solids and the large difference between total and soluble
COD. Soluble COD removal was generally very good and con-
sistent, whereas particulate removal was poor and variable.
During this period neither solubilization nor net removal
of SCOD was experienced across the chemical treatment step.
The carbon contactors removed about 72 percent SCOD and over-
flowed 20 to 30 mg/£ suspended solids.
The granular media filter removed only 33 percent of the sus-
pended solids fed to it. Because of this poor suspended
solids removal, the total COD of the plant effluent averaged
41 mg/£. From Table 6-2 it is seen that 31 ing/5, of the total
COD in the plant effluent was caused by suspended solids
passing through the filter.
Because of poor granular media filter performance, the plant
was only able to remove 81 percent suspended solids and 68
percent total COD. Had 95 percent suspended solids removal
been achieved, 88 percent removal of COD would have been
experienced. However, 96 percent removal of total phosphorus
was achieved even with poor filter performance.
As in the previous operating period the wastewater was again
slightly softened by solids-contact lime treatment as indi-
cated by a 26 percent reduction in total hardness. Marble
stability test results again showed that a very stable eff-
luent was produced by the solids-contact chemical treatment
unit. The average difference between the observed pH and
42
-------
TABLE 6-2:
OPERATING PERIOD II: AVERAGE OPERATING CONDITIONS AND
TREATMENT RESULTS JANUARY 6 TO FEBRUARY 20, 1973
OJ
Operating Conditions
Chemicals: Hydra ted Lime @ 490 mg/£ (10.8 to 11.2
Operating pH)
H2S04 @ 2080 Ib/MG = 250 mg/£
Average Hydraulic Loading: Chemical Contactor 0.84 gpm/sq ft
(constant flow) Carbon Contactor 0.56 gpm/sq ft
GM Filter 3.0 gpm/sq ft
Carbon Contacting Mode: Single Stage
Carbon Dosage: First Contactor 183 mg/£
Second Contactor 134 mg/£
Average Process
Treatment Results
Turbidity, JTU
Suspended Solids, mg/£
% Transmittance
Total P, mg/£ P
Soluble Total P, mg/£ P
Total a P04 , mg/£ P
Soluble a P0,,j, mg/£ P
COD, mg/£
SCOD, mg/£
BOD, mg/£
pH, units
Methyl Orange Alkalinity,
Phenolphthalein Alkalinity, a
Total Hardness, a
Calcium Hardness, a
Raw
Wastewater
47
99
4
3
2
2
127
41
84
7
342
—
484
272
.88
.62
.54
.07
.4
Chemical
Effluent
7
27
0.
0.
0.
0.
—
—
—
10.
142
86
251
211
26
13
09
05
3
Neutralized Plant
Chemical Effluent Effluent
3
9
-
0.21
0.13
0.08
46
41
25
8.0
64
—
387
299
4
19
90
0
0
0
0
41
10
5
8
49
—
360
298
.21
.13
.09
.06
.0
mg/£ as
-------
FIGURE 6-3: EFFLUENT HISTORY CURVE FOR OPERATING PERIOD
(JANUARY 6 - FEBRUARY 20. 1973)
44
-------
pH at stability was 0.01 pH units, with a range of only 0.01
pH units. These results are for eight grab samples collected
of six different days.
Except for the presence of excessive carbon suspended solids,
the quality of the plant effluent was considered very good.
As seen in Figure 6-3, about 1 week after plant start, a
very consistent plant effluent SCOD was produced.
Operating Period III
Operating Period III was from March 28 to June 7, 1973. The
plant was operated with alum pretreatment and diurnal flow.
Alum dosage and peak flow rates were set based on previous
plant operating experiences. The carbon contactors were run
in a two-stage counter-current mode. The carbon contactor
effluent was filtered through a tri-media filter.
Table 6-3 presents pertinent operating conditions and a tab-
ulation of average plant influent, chemical treatment eff-
luent and plant effluent quality for this operating period.
Figure 6-4 shows the plant effluent history curves. Since
the filter was not brought on stream until April 13, data
for the final carbon contactor was plotted prior to this
date. The substantial drop in suspended solids and total
phosphorus after the filter was brought on stream is obvious.
Two carbon dosages were used during this operating period.
From April 15 to May 5, the average carbon dosage was 294
mg/£. Before and after this period the carbon dosage averaged
104 mg/£. At the higher dosage, lower COD and SCOD values
were achieved. COD and BOD5 data for the two different dos-
age periods are broken out in Table 6-3 to show the effect
of the higher carbon dosage.
During this operating period two diurnal flow variations were
used, 2:1 and 3:1 peak to minimum flow. As seen in Figure
6-4, the 3:1 variation did significantly effect the plant
effluent quality and consistency.
During the 2:1 peak:minimum flow period, the average plant
flow was 35 gpm. The peak flow ranged from 40 to 44 gpm and
the minimum flow was 20 to 24 gpm. During this period, the
plant effluent averaged 4 mg/£ SS, 6.3 mg/& COD and 2.7 mg/£
SCOD. During the 3:1 peak:minimum flow period, the average
plant flow was 41 gpm. The peak flow ranged from 55 to 58
gpm and the minimum flow was about 20 gpm. During this
period the plant effluent averaged 13 mg/Jl SS, 11 mg/& COD
and 5 mg/£ SCOD. The slight difference in effluent SCOD
45
-------
TABLE 6-3: OPERATING PERIOD III: AVERAGE OPERATING CONDITIONS
TREATMENT RESULTS MARCH 28 TO JUNE 7, 1973
Chemicals:
Average Hydraulic Loading:
Carbon Contacting Mode:
Carbon Dosage:
Operating Conditions
Alum @ 14.0 mg/£ - Al
Chemical Contactor
First Stage Carbon
Second Stage Carbon
GM Filter
Peak to minimum
Two-stage counter-current
0.38 gpm/sq ft
0.51 gpm/sq ft
0.44 gpm/sq ft
2.0 gpm/sq ft
2 :1 and 3:1
294 mg/£ and 104 mg/£
Average
Process Treatment Results
Raw
Wastewater
Chemical
Effluent
Plant
Effluent
Turbidity, JTU
Suspended Solids, mg/£
% Transmittance
Total P, mg/£ P
Soluble Total P, mg/£ P
Total aP04, mg/£ P
Soluble aP04, mg/£ P
COD', mg/£
SCOD, mg/£
BOD, mg/£
pH, units
53
174
5.03
3. 87
2.91
2.52
143
38
112
11
29
0.67
0.14
0.28
0.08
41
23
26
2
9
97
0.2
0.1
0.1
0.07
6/lla
2.0/5.03
7.1
7.3
7.5
a - 294 mg/£ of PAC/104 mg/£ of PAC
-------
FIGURE 6-4: EFFLUENT HISTORY CURVE FOR OPERATING PERIOD IE
(MARCH 28 - JUNE 7, 1973)
47
-------
was due to a substantially higher carbon dosage during the
first part of this period. (See Figure 6-4.)
The substantial effect of flow variation noted above is felt
to have been caused by one or two reasons. The first relates
to the significantly higher peak hydraulic loading on the
chemical treatment unit clarification zone. During the 3:1
flow variation period, the peak hydraulic loading was about
0.57 gpm/sq ft, whereas during the 2:1 flow variation period,
the peak hydraulic loading was about 0.42 gpm/sq ft. The
40 percent higher peak hydraulic loading resulted in more
chemical floe escaping the clarification zone. The second
possible reason for deteriorated granular media filter per-
formance may have been due to the actual diurnal flow varia-
tion experienced. Because from 4 to 5 gpm sample streams
were taken from the overflows of the three preceding solids-
contact units, the actual diurnal flow variation to the filter
was well in excess of the 2:1 and 3:1 peak:minimum plant feed
flow. The filter actually experienced 4:1 and 6:1 diurnal
flow variations. The high variation may have contributed to
the deteriorated performance.
From the plant effluent phosphorus history curve shown in
Figure 6-4, it is seen that during the period, May 2 to 8,
abnormally high values of effluent phosphorus were experienc-
ed. The reason for this abnormality in phosphorus removal
was determined to be due to low alum dosages. Average daily
alum dosages were found to be about 40 percent of the average
dosage for Period III of 14 mg/£ aluminum. The average plant
effluent phosphorus value shown in Table 6-3 does not in-
clude values for the period May 2 to May 8.
During Period III, overall plant SCOD removal was 90 percent.
The alum treatment step removed 44 percent and the carbon
step removed 56 percent of the total SCOD removed. Suspend-
ed solids were reduced by 95 percent, COD by 94 percent, and
total phosphorus by 96 percent. In general, the plant eff-
luent was of very good quality.
Operating Period IV
Operating Period IV was from June 10 to August 13, 1973.
During this period FeCl3 pretreatraent and 2:1 peak:minimum
diurnal flow pattern were used. The carbon contactors were
operated in two-stage counter-current mode and the effluent
was filtered through a tri-media filter bed.
Table 6-4 presents pertinent operating conditions and a tab-
ulation of average influent, chemical treatment effluent,
and plant effluent quality. Figure 6-5 is a plant effluent
history curve for this period.
48
-------
TABLE 6-4: OPERATING PERIOD IV: AVERAGE OPERATING CONDITIONS AND
TREATMENT RESULTS JUNE 10 TO AUGUST 13, 1973
Chemical:
Average Hydraulic Loading:
Carbon Contacting Mode
Carbon Dosage
Operating Conditions
FeCls @ 96 mg/£
Chemical Contactor
First Stage Carbon
Second Stage Carbon
GM Filter
Peak to minimum
Two-stage counter-current
0.38 gpm/sq ft
0.52 gpm/sq ft
0.46 gpm/sq ft
2.1 gpm/sq ft
2:1
95 mg/£
Average
Process Treatment Results
Raw
Wastewater
VD
Chemical
Effluent
Plant
Effluent
Turbidity
Suspended Solids, mg/£
% Transmittance
Total P, mg/£
Soluble Total P, mg/£ P
Total aP04, mg/£ P
Soluble aPO4, mg/£ P
COD, mg/£
SCOD, mg/£
BOD, mg/£
49
118
4. 93
3.80
3.76
3.35
128
35
97
42
26
0.52
0.11
0.38
0.05
28
19
7
3
3
99
0.29
0.20
0.19
0.13
3
2
3
pH, units
7.2
7.3
7.6
-------
FIGURB 6-5: EFFLUENT HISTORY CURVE FOR OPERATING PERIOD
(JUNE 10 - AUGUST 13, 1973)
10
June
30 1 10
August
50
-------
The two periods of abnormally high plant effluent phosphorus
were determined to be due to low FeCl3 dosages.
As seen in Figure 6-5, there was about a 10 day period be-
fore plant effluent SCOD reached the low value of about
2 mg/£. Obviously the chemical treatment unit was not re-
moving substantial SCOD until after about 10 days, or by
June 19. This fact is shown in Figure 6-6, which indicates
that no removal of SCOD was experienced for the first 5 days
and then increasing SCOD removal, up to 50 percent, by the
12th day. As will be discussed shortly, anaerobic biological
activity existed in the chemical treatment unit. It is poss-
ible that anaerobic biological activity also removed some
SCOD in the chemical treatment unit. This would explain why
it took almost 2 weeks before substantial SCOD removals were
observed after start-up of the unit for the FeCl3 campaign.
During previous contract work, a "post precipitation" problem
was experienced when FeCl3 was used as the coagulant for raw
wastewater in a solids-contact treatment unit. The problem
was conjectured to have been due to reduction of ferric to
ferrous iron in the chemical treatment unit and subsequent
oxidation back to ferric iron in composite samples and plant
effluent clearwell. The cause of the reduction of iron was
suggested to be due to anaerobic conditions existing in the
chemical treatment unit. This same problem reappeared dur-
ing the current study, but with an added negative effect.
Chemical treatment unit average composite sample effluent
turbidity shown in Table 6-4 for Period IV was 42 JTU's.
During this same period, the average of at least six grab
sample turbidities each day was about 9 JTU's. This differ-
ence is similar to that reported for previous studies.1
TABLE 6-5: PLANT GRAB-SAMPLE SOLUBLE IRON PROFILES
Date August 8 August 28
Process Stream;
Raw Wastewater 0 0
Chemical Treatment Effluent 11 7.0
First Stage Carbon Effluent — 5.5
Second Stage Carbon Effluent — 2.5
Plant (GMF) Effluent 0 0
51
-------
FIGURE 6-6: SOLUBLE CHEMICAL OXYGEN DEMAND REMOVAL PATTERN DURING
INITIAL START-UP OF CHEMICAL CONTACTOR WITH FeCl.,
Elapsed Time, days
-------
Table 6-5 presents two plant soluble iron profiles. These
show high concentrations of soluble iron in the chemical
treatment unit effluent. It was subsequently removed in the
carbon and granular media filter steps. The only possible
external source of soluble iron would have been in the FeCl3
feed stock solution. This was considered to be unlikely
since high purity FeCl3 powder was used (95 percent purity).
To verify this belief, soluble iron determinations were made
of the flash-mix effluent which contained the FeCl3 feed.
No soluble iron was found. Thus it was concluded that all
iron reporting to the chemical treatment unit was insoluble
ferric precipitate (floe). Since the effluent pH was about
7.0 to 7.5, it was assumed that any soluble iron present
was in the ferrous form. This assumption was substantiated
by the following bench scale test. A grab sample of chemical
treatment unit effluent was split into two aliquots. One
of the aliquots was filtered through a 0.45 micron membrane
filter in the absence of air. The filtered and unfiltered
samples were split into two aliquots and one of each aerated
vigorously and one of each left sitting quiescently. Turbidi-
ties of the four samples were determined versus elapsed time
and were plotted in Figure 6-7. It is readily apparent that
gross turbidity increases were observed for all samples.
Vigorous aeration resulted in near instantaneous precipita-
tion of a turbidity exerting material. Soluble iron deter-
minations after turbidity equalization showed none of the
initially present 11 mg/£ of soluble iron. It was therefore
assumed that soluble iron was originally present in the
ferrous form.
The pertinent implication of the above assumption and pre-
viously mentioned observations is that iron was reduced
from the ferric to ferrous form within the•chemical treat-
ment unit.
An additional problem encountered during this period was the
appearance of a black, finely divided precipitate in the
chemical clarifier underflow sludge. This precipitate was
not detected in the clarifier effluent. As will be dis-
cussed in Section VII, the presence of this black precipi-
tate gave the sludge a septic appearance. However, no
signs of gross septicity, such as "rotten egg" odors or gas-
eousness, were found. Gravity thickener performance was not
adversely affected, but the finely divided precipitate
created a difficult problem with the vacuum filtration of
ferric-primary sludge. Upon acidification of samples of
filter cake gross sulfide odors (rotten egg) were observed.
Several grab sample profiles of all process streams, thick-
ener overflow and vacuum filter feed and filtrate all
53
-------
FIGURE 6-7: EFFECT OF AERATION ON IRON PRECIPITATION
ui
140
11 mg/1 initial
p mg/1 Final "
Unfiltered Aerated
Filtered Aerated
Unfiltered Standing
Filtered Standing
10
20
30
40 50 60
Elapsed Time, min
80
90
100
-------
indicated nominal to non-existent amounts of soluble sul-
fides. The only plausible explanation of where the copious
amount of sulfide could have been produced was that it
occurred in the chemical treatment unit and thickener. This
implies the presence of anaerobic biological activity in
general and of anaerobic sulfur reducing bacterial activity
specifically.
In summary, it is concluded that anaerobic biological activ-
ity in the chemical treatment unit resulted in the reduction
of iron and sulfur with the resultant precipitation of FeS
and leakage of soluble iron in the effluent. As previously
noted some biological removal of SCOD was also probably
experienced.
With FeCl3 treatment and two-stage counter-current carbon
contacting, the pilot plant removed 94 percent SCOD. The
chemical treatment step removed 48 percent and the carbon
step 52 percent of the total SCOD removed. A 97 percent
reduction in total COD was achieved. A 94 percent reduction
of total phosphorus was achieved.
Granular media filtration was very effective. Note that
the total COD values shown in Figure 6-5 are almost always
equal to SCOD values.
During operating Period IV, the pilot plant consistently
produced excellent quality effluent.
Operating Period V
Operating Period V was from August 14 to September 15, 1973.
The major treatment variation during this period was the
use of a reduced FeCl3 coagulant dosage and a polyelectro-
lyte floe strengthener.
During Period IV, the ratio of FeCl3 fed to phosphorus was
3.7 moles per mole. Though as much as 1/3 of the ferric iron
may have been solubilized, reducing this ratio to 2.5, the
0.29 mg/£ total phosphorus in the effluent was considered
quite low. Thus during Period V it was decided to signifi-
cantly reduce the FeCl3 dosage.
Previous experience had indicated that 0.25 mg/& of Altasep
2A2* was an optimum dosage for 50 to 75 m.g/£ FeCl3 dosages.1
*A product of Atlas Chemical Industries Incorporated,
Wilmington, Delaware.
55
-------
Using this polyelectrolyte at about $1.50/lb, a reduction of
7.5 mg/£ FeCl3 at SC/lb would be required for a cost trade
off. For Period V a target of about 70 mg/£ FeCl3 was set
and as seen in Table 6-6 the average dosage was 62 mg/£.
The carbon contactors were operated in the two-stage counter-
current mode. From August 24 on, regenerated carbon was
used.
Table 6-6 presents pertinent operating conditions and average
influent, chemical treatment effluent and plant effluent
quality. Figure 6-8 is a plant effluent history curve. As
was experienced during the previous operating Period (IV)
very consistent treatment was achieved. Comparison of aver-
age plant effluent qualities during Periods IV and V (Tables
6-4 and 6-6) shows that the only significant change was an
increase of phosphorus from 0,29 to 0.67 mg/&. Comparison
of chemical treatment unit and plant effluent phosphorus data
shown in Table 6-6 indicates a slight reduction in total
phosphorus across the two-stages of carbon contactors and
the granular media filter. However, a substantial increase
in soluble phosphorus occurred. It can only be presumed
that the anaerobic conditions in the carbon contactors re-
sulted in additional reduction of Fe+3 resulting in release
of phosphorus. The same phenomenon was experienced during
Period IV (Table 6-4) .
The iron post-precipitation problem identified during Period
IV persisted during Period V. As noted previously (Table
6-5) soluble iron was removed across the carbon contacting
stages and the filter. An equilibrium adsorption isotherm
test indicated that substantial soluble iron was adsorbed
on virgin carbon.
On August 24, reuse of thermally regenerated carbon was
started. As seen in Figure 6-8, a slight increase in plant
effluent COD was experienced when compared to the virgin
carbon period.
During Period V the pilot plant removed 90 percent of the
SCOD. The chemical treatment step removed 52 percent and
the carbon step removed 48 percent of the total SCOD re-
moved. Suspended solids removal was 90 percent and 98 per-
cent removal of total COD was experienced. Only 86 percent
removal of total phosphorus was achieved because of a re-
duced FeCl3 dosage. Again the granular media filter was
very effective in producing low levels of suspended solids
in the effluent. The final effluent was of excellent qual-
ity and the operation very consistent for this period.
56
-------
TABLE 6-6: OPERATING PERIOD V: AVERAGE OPERATING CONDITIONS AND
TREATMENT RESULTS AUGUST 14 TO SEPTEMBER 15, 1973
Operating Conditions
Chemicals: FeCl3;
Polymer:
Average Hydraulic Loading:
Carbon Contacting Mode
Carbon Dosage
@ 62 mg/&
@ 0.25 mg/£
Chemical Contactor
First Stage Carbon
Second Stage Carbon
GM Filter
Peak to minimum
Two-stage counter-current
0.39 gpm/sq ft
0.52 gpm/sq ft
0.45 gpm/sq ft
2.7 gpm/sq ft
2:1
98 mg/£
Ln
Average
Process Treatment Results
Raw
Wastewater
Chemical
Effluent
Plant
Effluent
Turbidity, JTU
Suspended Solids, mg/£
% Transmittance
Total P, mg/A P
Soluble Total P, rag/a P
Total aP04, mg/£ P
Soluble aPO4, mg/£ P
COD, mg/&
SCOD, mg/£
BOD, mg/£
43
166
4.94
3.67
4.31
3.76
133
30
58
35
35
0.80
0.24
0.51
0.10
28
16
16
2
4
99
0.67
0.61
0.63
0.59
3
3
3
pH, units
7.2
7.2
7.3
-------
FIGURE 6-8: EFFLUENT HISTORY CURVE FOR OPERATING PERIOD V
(AUGUST 14 - SEPTEMBER 15, 1973)
15
August
25
30 1 5*
September
J-r-
10
15
58
-------
Operating Period VI
On September 15, the pilot plant was taken off stream to ter-
minate the FeCl3 chemical treatment campaign. Plant modifi-
cations were implemented to allow for returning to lime chem-
ical treatment. As will be developed in later Sections,
additional lime sludge thickening experience was considered
desirable. In addition, plant SCOD removal data for lime
chemical treatment ahead of two-stage counter-current carbon
treatment was desired. On September 18, the chemical treat-
ment unit was started up with lime and on September 20, acid
(H2S04) neutralization was started. Instrumentation, equip-
ment, and operational problems were numerous. The seven
months idle time of some instruments, controls and equipment
resulted in extensive maintenance, calibration and repair
work. Coupled with this was the fact that the plant opera-
ting staff at the time had no experience with lime treatment.
This was due to reassignment of two key senior technicians
versed in plant operations and normal turn over of part time
and full time operation personnel. Consequently, retraining
of the new operating personnel was required. After about
one week (on September 27), the decision was made to abort
the lime chemical treatment study. Time was of the essence
and.the major priority at that time was to regenerate and
evaluate the reuse of carbon. Erratic operation of the
chemical treatment unit using lime could not be tolerated,
even for a few weeks.
The last operation, Period VI, was from September 28 to
November 8, 1973. Alum pretreatment with polymer addition
was used with a 2:1 diurnal flow pattern. The carbon con-
tactors were again operated in the two-stage counter-current
mode.
Table 6-7 presents pertinent operating conditions and aver-
age plant influent, chemical treatment effluent and plant
effluent quality- Figure 6-9 is the plant effluent history
curve for this operating period. Since emphasis was placed
on carbon regeneration which required considerable labora-
tory time, no phosphorus analyses were done. Operational
and sampling problems were experienced from October 9 to
October 15, and thus, no laboratory analyses were made.
Since low carbon dosages were used, residual COD was slight-
ly higher than for previous periods. However, 82 percent
removal of SCOD was achieved. The chemical contactor re-
moved 56 percent of the total SCOD removed while 44 percent
was removed in the carbon system. Suspended solids and
total COD removals were 97 and 94 percent, respectively.
59
-------
TABLE 6-7: OPERATING PERIOD VI: AVERAGE OPERATING CONDITIONS AND
TREATMENT RESULTS SEPTEMBER 28 TO NOVEMBER 8, 1973
o
Chemicals: Alum:
Polymer:
Average Hydraulic Loading:
Carbon Contacting Mode :
Carbon Dosage:
Average
Process Treatment Results
Turbidity, JTU
Suspended Solids, mg/£
% Transmittance
COD, mg/£
SCOD, mg/Ji
BOD, mg/£
Operating Conditions
@ 9. 6 mg/£ - Al
@ 0.25 mg/£
Chemical Contactor
First Stage Carbon
Second Stage Carbon
GM Filter
Peak to minimum
0.49
0.63
0.55
3.5
2:1
Two-stage counter-current
62
Raw
Wastewater
59
263
179
44
187
gpm/sq ft
gpm/sq ft
gpm/sq ft
gpm/sq ft
and 3 : 1
mg/£ Fixed
Chemical
Effluent
8
26
49
24
37
Carbon
Plant
Effluent
8
8
97
11
8
6
pH, units
7.1
7.3
-------
FIGURE 6-9: EFFLUENT HISTORY CURVE FOR OPERATING PERIOD
(SEPTEMBER 28 - NOVEMBER 8, 19735
30
Sept
5 10
October
15
20
25
30 1
November
61
-------
SOLIDS-CONTACT TREATMENT UNITS PERFORMANCE
Solids-contact treatment units were employed for two of the
three liquid wastewater treatment unit operations used in
the PAC-PCT system studied. Unit processes of chemical pre-
cipitation, coagulation, flocculation, solids-contacting, ad-
sorption, biological organic removal, and gravity sedimenta-
tion were accomplished in the two unit operations broadly
identified as chemical treatment-clarification and carbon
contacting-clarification. To say that the successful, or
unsuccessful, performance of the PAC-PCT pilot plant de-
pended to a large degree on the performance of the solids-
contact units would be an understatement.
The purpose of this sub-section is to evaluate two aspects
of the performance of the solids-contact units used. Opera-
tional stability and clarification effectiveness for the
three general types of applications evaluated will be dis-
cussed. These three applications were lime treatment, alum
or FeCl3 chemical treatment of a raw municipal wastewater
and carbon treatment of the chemically treated effluent.
The need to emphasize this topic is based on a few negative
pilot plant experiences and a few published negative judge-
ment comments generally based on very indirect experi-
ence. ^ , 5 , 6 , 7 On two occasions known to the authors, nega-
tive pilot plant experiences were the result, in part, of
using inappropriate pilot plant units offered by an equip-
ment manufacturer. In another instance with a prototype
unit, an inadequately designed sludge underflow pump caused
a major problem.
On the positive side, it should be noted that hundreds of
solids-contact treatment units, of various manufacturers,
have been successfully applied to municipal and industrial
wastewater applications. Typical industrial wastewater
applications are for treatment of metalurgical wastes (e.g.,
metal finishing, fluoride removal, rinse water treatment),
steel mill wastes and pulp and paper wastes (e.g., white
water, color removal).
Most of these applications are not at constant flow but
rather for stepped-variable flow.
The major concern with solids-contact treatment in munici-
pal wastewater applications is the presence of a large in-
ventory of sludge in the unit. Fear of excessively high
shock hydraulic loadings or gross thermal upsets are common
apprehensions. A real problem of iron and sulphur reduction
62
-------
was experienced in this study. This problem resulted from
the detention and mixing of a large inventory of primary
sludge at near neutral pH when ferric chloride was used. No
operational or performance problems were experienced when
using alum treatment at near neutral pH, or lime treatment
at a pH of 10.0 to 11.2.
To enable evaluation of the effect of flow, process and oper-
ational variations on the clarification performance, and
operational stability of the specific solids-contact units
used in this study (see Figure 4-3), approximate sludge pool
depths and concentrations were monitored six times each day-
Sludge pool depth was approximated by observing and recording
the highest sample tap at which sludge appeared. Sample taps
were located at about 13, 31, and 58 inches up the tank wall
from the tank bottom. These taps were designated as the
bottom, middle, and top.
Sludge concentration was routinely determined by an in-
direct 5 or 10 minute settling test. The procedure used was
as follows: A 2 liter graduated cylinder was filled to 2.0
liters of sludge. The sludge was allowed to quiescently
settle for 5 to 10 minutes. The volume of settled sludge
was observed and the percent volume it represented noted.
These values were reported as the 5 (or 10) minute percent
settled sludge volume. The time was chosen such that
the subsidence rate of the sludge-liquid interface at that
time was very small (e.g. less than about 0.2 5,/min) . For
alum or ferric-primary sludges, 10 minutes was determined
to be required, whereas for lime-primary and carbon sludges
5 minutes was sufficient.
Suspended solids concentrations were found to be directly
related to percent settled sludge volumes. For alum or
ferric-primary sludges the relationship was 0.05-0.15 g/&
per percent settled sludge volume, with an average of 0.10.
An average of 0.15 g/£ per percent settled sludge volume
was found for carbon sludges with a range of from 0.09 to
0.22. For lime-primary sludge the average was 0.19 g/£ per
percent settled sludge volume with a range of from 0.06 to
0.47.
The 5 or 10 minute settling test also provided an approximate
and normally conservative estimate of the sludge interface
subsidence rate. The percent settled sludge volumes re-
ported were directly related to a height the sludge inter-
face fell in the given time period. The 2.0 liter mark on
graduate cylinders used was about 1.3 ft or 16 inches.
Thus a 50 percent settled sludge volume meant that the
sludge interface had settled 16-(16 x 50/100) or 8 inches
during the test period. A 25 percent settled sludge volume
63
-------
meant the sludge interface had settled 16-(16 x 25/100) or
12 inches during the test period. The general expressions
for sludge interface subsidence rates, in inches/minutes,
for the two time periods used was:
5-minute test — 3.2 (1 - %Vol/100)
10-minute test — 1.6 (1 - %Vol/100)
Recall that 1.6 inches/minutes equals 1.0 gpm/sq ft. The
operational utility of this approximation of sludge settling
rate can be appreciated by the following example: For a
known clarification zone geometry, the average rise rate at
any sample depth is known for a given flow to the clarifier.
For illustrative purposes, assume that in the carbon contac-
tor the peak daily rise rate at the middle (31 inches deep)
sample tap was 1 gpm/sq ft (1.6 inches/minute) . The maximum
desired percent settled sludge volume for a 5 minute test
at this sample tap would have been
1.6 = 3
.2/1 - % Vol\
\ 10°/
or % Vol = 50
If the sludge pool concentration was in excess of 50 per-
cent volume, the slurry pool would have been expected to
expand (increase its depth) to a point where about 50 per-
cent volume existed.
Lime-Primary Sludge Application
The chemical contactor-clarifier was operated under two flow
patterns during lime treatment studies. First was constant
flow at about 1.0 gpm/sq ft and the second at 2:1 peak:
minimum diurnal flow with a peak of about 1.3 gpm/sq ft and
an average of about 1.0 gpm/sq ft. Figures 6-10 and 6-11
show the daily high and low sludge levels (sample taps) and
the effluent suspended solids values for the two flow
patterns. Comparison of effluent suspended solids concen-
tration ranges and averages indicated little difference in
the range of suspended solids, but do indicate slightly
lower suspended solids values during the constant flow
period. The daily variation in sludge pool depth was con-
sistently greater for the 2:1 diurnal flow period (Figure
6-11). The sludge pool interface rose and fell at least
2-1/2 ft and possibly as much as 3 to 4 ft. This resulted
in a 3 to 4 ft sludge blanket existing at the peak flow of
about 1.3 gpm/sq ft. At low flow, about 0.60 gpm/sq ft, no
sludge blanket existed. Recall that a sludge level at the
64
-------
FIGURE 6-10:
DAILY VARIATION IN LIME
PRIMARY SLUDGE LEVEL AND
EFFLUENT SUSPENDED SOLIDS
AT CONSTANT FLOW
g
•t
to
T3
0)
T>
C
(1)
a
CO
0
a
E,
rO
a;
I M
0)
B • •
25
Maximum
Mimimum
Aua
30 1 5 10 15
September 1972
65
-------
FIGURE 6-11:
DAILY VARIATION IN LIME PRIMARY SLUDGE LEVEL AND EFFLUENT
SUSPENDED SOLIDS AT 2:1 DIURNAL FLOW
axmum
0 Minimum
it
November 1972
December 1972
-------
middle tap was required for accomplishing solids-contacting
without sludge-blanket clarification. Sludge-blanket opera-
tion did not exist until the sludge pool interface rose
about to the top tap.
During the constant flow period, shown in Figure 6-10, a
sludge-blanket existed for very short periods on just a few
days. The daily variations in sludge pool levels during
this constant flow period, with an average hydraulic loading
of about 1.0 gpm/sq ft, were obviously not due to flow varia-
tions. Rather, they represent operational variations, mainly
in sludge blowdown frequencies (rates). At constant flow
(1.0 gpm/sq ft), there were several week durations when daily
sludge pool depth variations were as great as those shown in
Figure 6-11 for a 2:1 diurnal flow pattern. Such a period
occurred during most of January 1973, when failure of diurnal
flow control instrumentation necessitated constant flow oper-
ation. During this period poor operation of the solids-con-
tact unit resulted in average daily effluent suspended solids
ranging from 5-80 mg/&. The January 1973, monthly average
was a high 31 mg/£. One probable cause of operational diffi-
culties during January 1973, was the colder than normal am-
bient (-5 to -10°F) and wastewater (11 to 13°C) temperatures.
On any given day, if the hydraulic loading, lime dosage,
pumping turbine and sludge rake speeds were all constant
and the sludge blowdown rate was at a constant value such
that sludge withdrawal equalled solids production, then the
concentration of sludge at any sample point in the solids-
contact unit should have been constant. During this study,
there were few days when all the above operational variables
were constant. When they all were constant, very stable
sludge level and constant suspended solids concentrations at
the various sample taps were observed.
Figure 6-12 shows sludge concentration variations during a
given day when lime dosage, sludge blowdown frequency and
turbine and rake speeds were constant. The only imposed
variable was diurnal flow. It is readily apparent that a
consistent and somewhat predictable pattern of sludge level
and concentration variations resulted from the diurnal flow
effect. As flow was reduced to the minimum the sludge level
dropped below the middle sample tap (31 inches) and approach-
ed the bottom tap (13 inches), and reaction zone taps. As
the flow approached the daily peak the sludge pool expanded
and the interface gradually rose over a 6 hour period. At
about 6:00 P.M., the sludge pool interface was over 5-ft
from the tank bottom and a fairly uniform concentration
existed. The sludge-blanket was stable as indicated by the
very low effluent turbidity.
67
-------
FIGURE 6-12: TYPICAL HOURLY VARIATIONS IN LIME PRIMARY SLUDGE
CONCENTRATION AND EFFLUENT TURBIDITY
oo
ATOP Tap
iddle--
8AM 12
Clock Time
-------
The 6:00 P.M., test data for the top tap sample indicated
about 38 percent settled sludge volume. For the 5 minute
test used, this indicated the sludge pool interface settling
rate would be approximately:
3.2 (1 - 38/100) = 1.98 inch/min or 1.24 gpm/sq ft
The actual rise rate at the top tap location was 1.25
gpm/sq ft at 6:00 P.M. From the known geometry of the
solids-contact unit and the 5 minute percent settled sludge
volume data in Figure 6-12 at 6:00 P.M., it could be pre-
dicted that had the flow rate been increased to about 150 gpm
and sustained there, the sludge-blanket would have expanded
to the overflow launder and thus into the effluent. Had
the operator expected such a high flow, he would have immedia-
tely reduced the turbine speed to minimize hydraulic tur-
bulence and started manual blowdown of sludge to reduce the
sludge inventory in the unit.
Obviously, the critical hydraulic design condition for this
application relates to the sludge settling characteristics
at the sustained high flow level expected. Previous contract
studies indicated that at a constant loading of 1.3 gpm/sq ft
stable operation and good suspended solids removal was ach-
ieved.1 The maximum possible hydraulic loading for this
application was not determined. It is the authors' judge-
ment that a peak 2 hour hydraulic loading of about 1.5 gpm/
sq ft would be possible.
Evaluation of daily results such as shown in Figure 6-12
indicated that during constant plant flow period (@ about
1.0 gpm/sq ft) the sludge pool concentration was very uni-
form. The difference between maximum and minimum 5 minute
percent settled sludge volumes varied no more than 2 to 10
percent daily. During November and December 1972, when
the plant was subjected to a 2:1 diurnal flow pattern, as
shown in Figure 6-12, the difference between maximum and
minimum 5 minute percent settled sludge volumes varied from
20 to 50 percent and averaged about 30 percent. On those
days when lime dosage, turbine or rake speeds or blowdown
frequencies were changed substantially, larger variations
in the difference between maximum and minimum values were
experienced.
Alum or FeCl3~Primary Application
The treatment of Saxt Lake City raw municipal wastewater
with alum or FeCl3 resulted in different operating character-
istics than lime treatment. Based on previous contract work
(reference 1) and other relevant field experience by one
69
-------
of the authors, the solids-contact unit was modified, as
indicated in Figure 4-3, to include a sludge baffle plate.
This baffle plate was for the purpose of separating the re-
action zone and sludge concentrating zone of the solids con-
tact unit. The objective of the baffle plate was to in-
crease underflow solids concentration by reducing hydraulic
turbulence in the sludge withdrawal zone. Another variation
from previous contract work was to externally rapid-mix the
alum or FeCl3 just prior to the solids-contact unit. Thus,
the main functions of the solids-contact unit in the current
study was to flocculate via solids-contacting and clarify
via gravity sedimentation.
Figure 6-13 shows the daily variation in sludge pool depth
(sample tap location) and effluent composite suspended solids
concentration for pilot plant operating Period III. The
average alum dosage during this period was 14 mg/£ as aluminum.
The 2:1 and 3:1 peak:minimum diurnal flow patterns used are
shown in Figure 6-14. The effective clarification area of
the solids-contact unit used was about 100 sq ft.
The data in Figure 6-13 shows that fairly consistent perfor-
mance of the solids-contact unit was experienced. It is,
however, obvious that clarification performance deteriorated
substantially during the 3:1 diurnal flow period. Both the
range and overall average effluent suspended solids concen-
tration about doubled when comparing the 3:1 and 2:1 diurnal
flow operation. As opposed to the other two periods shown
in Figure 6-13, during the 3:1 diurnal flow period (without
polyelectrolyte), there was almost always a sludge-blanket
of at least 1-2 ft, even at minimum daily flows (about 0.22
gpm/sq ft) .
Operational stability was considered poorer for the 3:1
diurnal flow period. During this 28 day period, the chemical
treatment unit effluent was diverted from the carbon con-
tactor due to unacceptably high suspended solids on five
different occasions for from 2 to 7.5 hour duration. The
average down time was just under four hours. No down time
was required during the other two periods.
A clearer picture of the effect of diurnal flow variations
can be seen by considering the fairly typical daily varia-
tions of sludge pool concentration and effluent turbidity
shown in Figure 6-15. The sludge pool solids concentration
is seen to be fairly uniform at all clarification zone
sample locations for all three conditions shown.
On May 17, the sludge inventory was at such a high level
that a few hours after peak daily flow (about 0.57 gpm/sq ft)
70
-------
FIGURE 6-13: DAILY VARIATION IN ALUM PRIMARY SLUDGE LEVEL AND EFFLUENT
SUSPENDED SOLIDS FOR CHEMICAL CONTACTOR
L>iurnal Flowi
H Maximjam
O Minimiim
28 31 5 10 15
Karch 1973 April 1973
20
25
30 5 10
May 1973
-------
FIGURE 6-14: TYPICAL DIURNAL FLOW
PATTERNS
3.0
to
O
-P
to
3
O
0)
-M ^3
a e
03 -H
-P C
03 -H
.
4-1 r-l
O -H
tO
O Q
•H
-P
1.0
12
4AM
SAM 12
Clock Time
4PM
8PM
-------
FIGURE 6-15: TYPICAL HOURLY VARIATIONS IN ALUM
PRIMARY SLUDGE CONCENTRATION AND
EFFLUENT TURBIDITY
100-
;A April
£D May
0 June 3
27, 1973 i
, 1973
1973
0)
0)
0)
•p
C
0)
u
M
-------
the sludge-blanket rose and began to overflow the effluent
launders. The basic cause of the problem was not a high flow,
but rather poor operation. The sludge inventory (depth and
concentration) had been allowed to become too great. On this
specific day (May 17), the sludge pool interface rose in the
clarification zone at a rate of about one ft/hr. Had opera-
ting personnel been aware of the high sludge inventory situa-
tion and had checked the units performance more than once
during the immediately preceding 8 hour shift, the clarifica-
tion failure might have been averted.
Inspection of Figure 6-15 shows that use of 0.25 mg/£ of
anionic polyelectrolyte at the 3:1 diurnal flow resulted in
an extremely stable sludge pool and effluent clarity. The
variation in maximum and minimum daily sludge concentration
(i.e., 10 minute percent settled sludge volume) is readily
apparent in Figure 6-15. Figure 6-16 shows these variations
over the entire operating period. The general effect of di-
urnal flow pattern and use of polyelectrolyte is apparent.
From the results presented in Figure 6-13 through 6-16 it can
be concluded that neither effective or reliable clarification
was achieved at a sustained peak daily hydraulic loading of
0.57 gpm/sq ft, without the use of polyelectrolyte. Effective
and reliable clarification was achieved at a sustained peak
hydraulic loading of 0.42 gpm/sq ft. These conditions are
consistent with perviously reported results.1
On June 8, 1973, use of FeCls in the chemical treatment solids-
contact unit was commenced. A 3:1 diurnal flow with a peak
daily hydraulic loading of 0.57 gpm/sq ft was attempted.
After about a week of futile effort it was concluded that
this was an excessively high peak hydraulic loading. The
plant flow was changed to a 2:1 diurnal flow pattern with
a peak hydraulic loading of about 0.45 gpm/sq ft and accept-
able performance resulted.
The daily sludge pool interface depth variations were essen-
tially the same as for alum treatment shown in Figure 6-13
(2:1 flow). Daily variations in sludge pool concentration
(10 minute percent sludge volume) and effluent suspended
solids values are shown in Figure 6-17. The effectiveness
of polyelectrolyte in densifying and increasing the settling
rate of the sludge and thus reducing concentration variations
is obvious. As noted on Figure 6-17 and discussed for opera-
ting Period IV, the effluent suspended solids concentration
shown included post-precipitation iron compounds. Thus it
was not possible to determine the exact cause of the variable
effluent suspended solids concentrations.
74
-------
FIGURE 6-16: DAILY VARIATION IN ALUM PRIMARY SLUDGE CONCENTRATION
AT BOTTOM SAMPLE TAP
01
80
o>
E
^60
o
3
rH
0}
TD
0)
-P
-P
0)
40 --
20 --
0
3:1 Diurnal Flow
2:1 Diurnal Flow
28 1 10
April 1973
20
30 5 10
Mav 1973
20
30
June
-------
FIGURE 6-17: DAILY VARIATION IN FERRIC PRIMARY SLUDGE CONCENTRATION
AND EFFLUENT SUSPENDED SOLIDS
15 20 25 30 1 5 10 15 20 25 30 1 5 10 15
July 1973 August 1973 September 1973
-------
Carbon Application
Solids-contact units were the heart of the carbon treatment
system for soluble organic removal. The solids-contact units
provided mixing, contact time, sludge inventory and rough
liquid-solids separation. Unit processes performed were ad-
sorption, anaerobic biological activity and gravity sedimen-
tation. The clarification effectiveness of the solids-con-
tact units under constant flow conditions had been extensively
evaluated previously.1
To further evaluate the operational stability and clarifica-
tion effectiveness of the carbon contactors, results for
four various sets of operating conditions will be presented
and discussed. Figure 6-18 shows daily sludge pool interface
depth variations and effluent composite sample suspended
solids concentrations for these four sets of conditions.
Table 6-8 presents pertinent operating conditions. For peak:
minimum diurnal flow variations up to 2.3:1, very stable
sludge pools existed. Maximum to minimum sludge pool depth
variations were 1 to 3 ft. At the higher diurnal flow varia-
tion of 2.9:1, the variation may have been slightly larger.
TABLE 6-8:
PERTINENT OPERATING CONDITIONS FOR
PERIODS SHOWN IN FIGURE 6-18
Period
Dates
February
1973
November
December
1972
July
August
1973
May
1973
Pretreatment
Chemical
Contacting
Mode
a)
Lime
SS
Lime
SS
FeCl3
2SCC
Alum
2SCC
Stage
Flows: gpm/sq ft
1st
1st
Peak
Average
Minimum
Peak: Minimum
Flow
--
0.56
--
1:1
0.90
0.72
0.45
2:1
0.62
0.52
0.27
2.3:1
0.79
0.55
0.27
2.9:1
a) SS-single stage; 2SCC-two stage counter-current
77
-------
FIGURE 6-18:
DAILY VARIATION IN CARBON SLUDGE LEVEL AND EFFLUENT
SUSPENDED SOLIDS FOR CARBON CONTACTORS
oo
Pretreatment
Chemical
Ferric Chloride
Diurnal Flow
# of Stages
2(lst stage)
2 1st stage)
0 Maximum
<:) Minimum
15 20 10 15 20 30 5 25 30 1 5 8
Feb 1973 Nov 197? nec 1972 July Aug 1973
9 15 20
May 1973
-------
The lowest variability and overall average effluent suspended
solids concentration were experienced during the constant
flow period. These data were not, however, directly compar-
able to the other periods shown since the granular media
filter performance would contain carbon fines, and chemical
floe, within the carbon-contacting clarification system, due
to recycle of backwash to the carbon contactor. If these
fines were to build up, or not report to the spent carbon
sludge treatment system, clarifier effectiveness would be
expected to diminish.
Except for the first few days during mid-November 1972, shown
in Figure 6-18, when carbon inventory was being gradually
reduced, reasonably good clarification was experienced.
The variability of effluent suspended solids concentration
during the higher two diurnal flow periods was not sub-
stantially greater than for the lower (2:1) diurnal flow
period. Overall average effluent suspended solids con-
centrations were, however, significantly higher. Data in
Table 6-8 indicate that peak hydraulic loadings were not
excessive. During the July to August 1973, period when
FeCl3 pretreatment chemical was employed, some post-precipi-
tation of iron may have occurred. At the peak hydraulic
loading of 0.62 gpm/sq ft, efficient removal of iron floe
would not have been likely. The high effluent suspended
solids concentrations shown for the May 1973, period in
Figure 6-18 may have been partly due to a pretreatment
effect; namely, 30 to 50 mg/£ of suspended solids in the
carbon-contactor feed. This chemical floe did not settle
in the chemical clarifier at the peak hydraulic loading of
0.57 gpm/sq ft. Because of this, it is unlikely that the
chemical floe would have settled in the carbon clarifier
which was operating at a higher peak hydraulic loading of
0.79 gpm/sq ft.
In general, it would have to be concluded that the clarifi-
cation efficiency of the carbon treatment solids-contact
units was not impressive. It will be shown that a properly
designed and operated filter station effectively removed
relatively high carbon-contactor effluent suspended solids
concentrations.
The sludge pool concentration uniformity and stability, for
typical days during the four previously discussed operating
periods, is shown in Figure 6-19.
The effect of increasing diurnal flow ranges on sludge pool
concentration, which is inversely related to sludge volume,
is apparent. It is also obvious that in general the re-
action zone and sludge pool concentrations were approximate-
ly equal.
79
-------
FIGURE 6-19:
TYPICAL HOURLY VARIATION IN CARBOH
SLUDGE CONCENTRATION
70
60-
50-
40
1:1 Diurnal Flow
February 14, 1973
70"
0)
CP
'D
40-'
O Bottom Tap
Reaction Zone
2:1 Diurnal Flow
November 19, 1972
c
(i>
u
2:1 Diurnal Flow
July 25, 1973
30
12
12
Clock Time
80
-------
That the single day data present in Figure 6-19 was indeed
typical is verified by the data presented in Figure 6-20.
Figure 6-20 shows daily differences in maximum and minimum
5 minutes percent settled sludge volumes (<* suspended solids
concentration) for the four operating periods described in
Table 6-8 and Figure 6-18. It is interesting to note that
in December 1972, at a peak daily hydraulic loading of 0.90
gpm/sq ft and very low sludge inventory conditions, a very
stable sludge pool existed.
Summary Discussion
The. results of this study .show that properly designed, sized
and operated solids-contact treatment units are applicable
to physical-chemical treatment of municipal wastewater.
Potential problems due to the presence of biodegradable
material and diurnal flow patterns can be averted. Prudent
operation is, however, just as important as prudent design
of the units.
Of the applications evaluated in this study, carbon treatment
was the most stable operation. It should be appreciated,
however, that a high degree of clarification effectiveness
by the carbon contactor-clarifiers was not a process re-
quirement.
The results indicated that when polyelectrolyte was used
with alum or FeCl3 treatment the sludge depth and concen-
tration variations were less than for lime treatment without
polyelectrolyte. FeCl3 treatment sludge depth and concen-
tration variations were apparently less than for alum treat-
ment sludge.
As with any gravity sedimentation unit the peak hydraulic
loading is a critical design parameter. It should be app-
reciated that the critical rise rate in a solids-contact
unit is as dependent upon the settling properties of the
sludge pool as it is on the free settling properties of the
feed suspended solids and precipitated material. Maximiza-
tion of the chemical treatment clarifier underflow is very
desirable. Maximization of hydraulic loading, at the ex-
pense of underflow sludge concentration may not minimize
total treatment costs.
During this study a uniform volumetric rate of sludge blow-
down was utilized. The results presented indicate that a
non-uniform rate would have more likely produced a higher
underflow solids concentration. Had sludge been blowndown
only during the low flow period of diurnal flow, higher
underflow suspended solids concentrations should have resulted.
81
-------
FIGURE 6-20: DAILY VARIATION IN CARBON SLUDGE CONCENTRATION AT
THE BOTTOM TAP
oo
NJ
Pretreatment
Chemical
0
15 20
Feb 1973
15 20
Nov 1972
Dec July 1973
May 1973
-------
Solids-contact units do not provide automatic treatment.
They must be operated. An understanding of the chemistry,
physics and hydraulics of the process, by the operator, is
just as important as an understanding of the mechanical
aspects, if not more so.
GRANULAR MEDIA FILTER PERFORMANCE
Previous contract experiences had indicated two problem
areas concerning the filter performance and operation.1
First, use of too fine a coal media resulted in only nominal
penetration of suspended solids into the filter bed. The
vast majority of suspended solids were removed in the upper
few inches of coal resulting in excellent effluent quality
but very short cycle times when high feed suspended solids
were encountered. The second problem area previously ex-
perienced concerned backwashing. Because of the filter
housing design (bed depth and free board) excessive backwash
water was required (30 to 50 percent in excess of estimates),
In addition, on one occasion a "mud ball" problem was ex-
perienced.
Based on the results of this previous study, it was re-
commended that evaluation of coarser coal media and the use
of polyelectrolyte be made. It was also recommended that
backwashing techniques be critically evaluated to define
procedures which would prevent operational problems.
During the current study, two filter bed designs were evalua-
ted. During Phase I a very coarse coal media was used and
during Phase II a tri-media bed was used. Backwashing tech-
niques were also characterized and the long term effective-
ness evaluated.
Coarse Media Filter Bed
During the first 6 months of plant operation with lime pre-
treatment, a coarse bed filter was used. This filter bed
consisted of 14 inches of 2.4 to 4.8-mm anthracite coal over
10 inches of 0.60 to 0.84-mm sand. It was anticipated that
the sand layer would maintain a high quality effluent.
In general the suspended solids removal performance of the
coarse bed filter was very poor. Figure 6-21 shows four
monthly average percent suspended solids removals at four
filtration rates. Monthly average effluent suspended solids
concentrations ranged from 6 to 21 mg/£. It is apparent
from inspection of Figure 6-21 that as filtration rates
increased performance deteriorated considerably. As will
-------
Suspended Solids Removal, percent
it^ LTl
O O
00
-------
be discussed later, use of anionic polyelectrolyte was found
to improve performance substantially.
The suspended solids removal pattern in the filter bed was
variable. Figure 6-22 shows typical terminal head loss
patterns. At low filtration rates, some penetration of sus-
pended solids into the coal-sand interface occurred. For
the runs shown, about 50 percent suspended solids removal
was achieved. Apparently most of the suspended solids that
penetrated through the coal layer were not removed by the
smaller sized sand media.
A series of filter runs were conducted, at constant filtra-
tion rates, to demonstrate the effect of using polyelectro-
lyte on the suspended solids removal pattern and performance.
Results shown on Figure 6-23 indicate that at the higher
polyelectrolyte dosage increased amounts of suspended solids
were removed in the coarse coal layer. Without polyelectro-
lyte (Run #130), only about 20 percent removal of 24 mg/£
feed suspended solids was experienced. With polyelectrolyte,
50 to 70 percent removal of 17 and 20 mg/£ feed suspended
solids was achieved. The effluents with 9 and 6 mg/£ of
suspended solids were considered marginally acceptable.
The coarse bed filter was routinely backwashed using the
following procedures:
1) Lower the water level by draining for 4 to 8
minutes to the surface of the coal.
2) Air-scour at about 3.5 SCFM/sq ft for 2 to 3
minutes.
3) Water wash at about 20 gpm/sq ft for 5.5 to 7.5
minutes.
A higher water backwash rate would have been desirable, but
limited pumping capacity made it impossible. Bed expansion
during water backwash was only about 10 percent (@ 13 to
16°C). Actually the coal layer was not observed to expand
significantly, whereas the sand layer expanded by about 20
percent. The coal media was not fluidized, but relative
rolling of filter media grains was observed. Apparently the
coal media was coarse enough to allow flushing of suspended
solids removed in the sand layer.
Use of the above backwash procedure for over six months did
not result in any "mud ball" formations or long-term buildup
in clean bed head loss (i.e., immediately after backwashing).
It was concluded that this backwash procedure was effective
-------
FIGURE 6-22: TERMINAL HEAD LOSS PATTERN FOR COARSE
BED FILTER
70 - -
60 —
:; & , 3. .:Q": gpm/sq f t _L. i
J^S-'gpm/sq ftlu
0
5 10 15
Bed Depth, inches from floor
86
-------
FIGURE 6-23;
EFFECT OF POLYELECTROLYTE ON TERMINAL
HEAD LOSS PATTERN
80
2.4 - 4.8 mm Coal
3j gpm/sq ft
IPoljy.
3:30 42
3.9 gpm/sq ft)
No Poly.
# 2(10 @ ;29 hijs
3 gpm/sq ft
.|0_mg/l Po|y
Clean Bee
3 gpm/sq ft
Typical
Heiadlods
5 10 15 20
Bed Depth, inches from floor
87
-------
for the filter bed used. For a 2 gpm/sq ft filtration rate
the above water backwash condition resulted in recycling less
than 5 percent of the filtered water for cycle times greater
than or equal to 21 hours.
Tri-Media Filter Bed
Having evaluated two extremes in the size of coal for the
dual-media filter beds used during previous and current con-
tract studies, it was decided to try a medium sized coal.
Results for the coarse media filter bed design indicated
that the 0.60-mm top size sand had not consistently pro-
duced the desired effluent suspended solids concentration of
5 Tag/I.
To prevent gross mixing of coal and sand layers there was a
limit "to the minimum sand top size which could be used. Pre-
vention of gross intermixing at normal water backwash rates,
requires that the largest coal size be no more than about
2.5 times the smallest sand size. As a result of these
limitations use of a third layer of garnet sand (density
4.2 g/cc) was considered necessary to have a layer of media
in the filter bed with a particle size considerably less
than 0.6-mm. A tri-media (or multi-media) filter bed was
designed and installed resulting in the approximate media
depths and sizes shown in Figure 6-24. Each layer of media
was made of commercially available cuts which were back-
washed at about 30 percent bed expansion and then scalped.
About 11 inches of standard 60 to 80 U.S. Mesh Idaho Garnet*
was installed, backwashed three times and 3 inches (about
27 percent) scalped off. One thousand Ibs of 20 by 30 U.S.
Mesh filter sand was then installed. After backwashing at
about 30 percent bed expansion considerable intermixing of
the garnet and sand was observed. There was, however, a
distinguishable layer of garnet of about 6 inches. About
4.5 inches of sand was skimmed off. About 25 inches of
anthracite coal+ (1.2 to 1.5-mm effective size, <1.7 uni-
formity coefficient) was then installed, backwashed three
times to a total bed expansion of about 30 percent and 5 to
6 inches scalped off. Slight intermixing at the coal-sand
interface was observed. The resulting filter bed, as shown
in Figure 6-24, was used for the last 7 months of this study
except that in late April about 5 to 6 inches of coal was
lost during an unattended backwashing.
* A product of Idaho Garnet Abrasive Company, Kellogg,Idaho
+ A product of Reading Anthracite Coal Company, Pottsville,
Pennsylvania
88
-------
FIGURE 6-24: TRI-MEDIA FILTER BED DESIGN
40
30 --
in
Q)
,C
u
C
o
0
20 -
O
u
c
a
-P
w
•H
p
10 -
—
1
-
1
-
-
[-
-j-
C
-
-
o
-
~
a
h
i_
1
5 to
-
7
-
-
-
7
-
-
-
-_
-
-
t
-
~
tim
\
t
Sand
0
.75
- o.
to
0
.84 mm
Garnet
25 to 0.5
nim
\
-
_l_
L
-
-
-t-
I
-
-
-
-
.. - ,. i _: -'.--
L
1.
L..
-
-(-
-
-
-|-
-
1 1
ill
I
Tl i :
.
1 i ' I
L
_
/
^ — coal —
intermixed
zone
f
-
\
V
/
- L ... L
4
L
47
r
-6
-5
in
o.
rt
EH
CO
0}
: ^
fi
— 3
-4-2
89
-------
It should also be noted that on February 13, 1973, the orig-
inal all-plexiglass filter housing was fractured beyond re-
pair. A new steel filter with a 2 ft wide plexiglass window
over the full bed depth was fabricated, installed and started-
up on April 13, 1973.
The following sub-sections discuss the performance of the
tri-media filter during the last four operating periods
(April to November 1973).
April 1973: During April 13 to 28, 1973, the pilot plant
feed flow was as depicted in Figure 6-14, at a 2:1 peak to
minimum diurnal flow. Due to substantial effluent sample
flows from the three preceding solids-contact-clarifiers
(about 13 gpm constant flow) the actual flow variation to
the filter was much higher. The approximate peak, average
and minimum daily filtration rates were 3.0, 2.4, and 0.6
gpm/sq ft, respectively.
At these filtration rates and for the other plant operating
and performance conditions (see Table 6-3), the filter sus-
pended solids removal performance was good. An average sus-
pended solids removal of 90 percent was achieved resulting
in an average effluent suspended solids of 2.8 mg/£ (1 to
4 mg/£ range) . At a terminal headloss of 6 to 7 ft of water,
filter cycle times of from 4 to 20 hours with an 8 hour
average was experienced.
Figure 6-25 shows the typical, normalized total filter head-
loss development pattern experienced for this period of
operation. The very rapid increase in headloss buildup near
the end of the cycle is indicative of a suspended solids re-
moval pattern where a majority, or all, or the filter media
pores at some given depth are completely clogged.
Figure 6-26 shows three typical terminal headloss versus
bed depth patterns observed. Two features of the patterns
shown were judged desirable. First, few suspended solids
were removed in the bottom 6 to 12 inches of the filter bed,
thus implying a low suspended solids concentration in the
filter effluent. Secondly, substantial removal of suspended
solids occurred in both the top half of the coal layer and
the top few inches of the sand layer as indicated by the
substantial headloss at these locations. Since the void
sizes in the coal layer are substantially larger than in
the sand layer, the near equivalent headloss observed implies
that substantially more suspended solids were removed in
the coal layer than in the sand layer.
90
-------
FIGURE 6-25:
TYPICAL NORMALIZED HEAD LOSS BUILD-UP
PATTERN
w
M
O
ffi en
TI e
0) Dj
pq tn
\
•H O
(0 (N
-P S
O
(U
N
•H
e
>-i
o
O
C
Filtrate Volume, gal/sq ft
91
-------
FIGURE 6-26: TERMINAL HEAD LOSS VERSUS BED DEPTH PATTERN
Garnet - Sand
Coal
80 -
April 1973
O# 425
Q# 426A
A# 426B
Reference Operating Period III
see Tafcle 6-3
Coal - Sand
Inteirf ace
fctjal Clean
.§: gpm/sq ft ;
10 20
Bed Depth, inches from floor
92
-------
The terminal headless patterns shown in Figure 6-26 supports
the inferred pore clogging suspended solids removal pattern.
As noted above, the substantial suspended solids removal at
the top few inches of sand suggests that all voids in the top
coal layers were not clogged (i.e., substantial suspended
solids penetrated into the sand layer). During this opera-
ting period a few filter runs were terminated, at less than
6 to 7 ft headloss, due to breakthrough of suspended solids
into the effluent. These breakthroughs typically occurred
during later morning when the filtration rate would increase
by a factor of 5, from a minimum of about 0.6 to a maximum of
3 gpm/sq ft, over a period of 4 to 5 hours. These break-
throughs were graphically indicated by the continuous re-
cording of effluent turbidity as measured by a Hach CR-low
range nephlometer unit.*
The tri-media filter bed was routinely backwashed using the
following procedure:
1) Lower the water level by draining for 8 to 15
minutes to the surface of the coal.
2) Air-scour at about 3.5 SCFM/sq ft for 1.5 to 2.5
minutes.
3) Water backwash at about 30 gpm/sq ft for 2.8 to
1 minute.
The average backwash used about 110 gal/sq ft. At an average
filtration rate of 2.4 gpm/sq ft and a cycle time of 8 hours,
this resulted in a somewhat high recycle of 6.8 percent of
the filtrate. At 30 gpm/sq ft and about 17°C water tempera-
ture, the total filter bed expansion was about 30 percent.
On April 24, 1973, about 5 inches of coal media was flushed
out of the filter with the backwash water. The cause appar-
ently was due to a substantial depth of the coal layer rising
as a plug to the backwash outlet. In the senior author's
experience this was an anomalous observation for a 3.5 ft
diameter pilot plant filter.
The recycle of lost coal media to the carbon-contacting
system and subsequently to the spent carbon thickeners re-
sulted in the nuisance of frequent underflow pump clogging
for several days.
*Manufactured by Hach Chemical Company, Ames, Iowa
93
-------
May 1973: During the last half of plant operating Period III,
the peak and average plant flows were increased to increase
the diurnal variation to about 3:1 peak .-minimum flow. This
resulted in the filter experiencing about a 6:1 peak:minimum
diurnal flow variation. Of significance also was the doubling
of filter feed suspended solids due to the higher overflow
rate in the carbon-contactor-clarifiers.
During the period April 28 to May 6, 1973, the filter was
subjected to peak, average and minimum daily filtration
rates of about 4.6, 2.9 and 0.7 gpm/sq ft. Under these
operating conditions, only 70 to 80 percent removal of an
average feed suspended solids of about 60 mg/£ was achieved.
Filter cycle times were erratic, varying from 1.5 to 15 hours.
Breakthrough of suspended solids into the effluent was the
rule rather than the exception. Terminal filter headloss
patterns indicated little removal of suspended solids in the
coal layer. The vast majority of terminal headloss existed
at the coal-sand interface.
On May 6, 1973, the filter flow was cut in half resulting in
peak, average and minimum daily filtration rates of 2.3, 1.5
and 0.4 gpm/sq ft. Feed suspended solids remained about 60
mg/£ (40 to 100 range). Suspended solids removal performance
remained somewhat erratic in the 75 to 85 percent removal
range. The majority of headloss buildup continued to occur
at the coal-sand interface.
During the period May 10 to 15, 1973, several filter runs
were conducted adding about 1/4 mg/£ anionic polyelectrolyte
just ahead of the filter. Effluent suspended solids concen-
trations were reduced by about half. Terminal filter bed
headloss patterns were somewhat erratic but generally indi-
cated substantial suspended solids removal in the coal layer.
No real breakthrough of suspended solids was experienced.
Very short filter cycles of just a few hours (2 to 5) were
typical.
On May 21, 1973, polyelectrolyte feed was started to the
chemical treatment unit in an effort to obtain lower eff-
luent suspended solids and more stable operation. A slight
improvement in filter suspended solids removal performance
was experienced. Filter cycle times increased to a range
of 17 to 25 hours. Terminal filter bed headloss data in-
dicated that most suspended solids were still being removed
at the coal-sand interface.
In general it was concluded that operation of the pilot plant
under the stress of relatively high flows used during May
caused relatively poor filter performance. Use of polyelec-
trolyte just ahead of the filter did result in a more uniform
94
-------
effluent quality (i.e., no breakthrough of suspended solids),
but prohibitively short cycle times resulted.
June-September 1973: During operating Period IV and V
(6/10/73 to 9/14/73) the performance and operation of the
filter was exceptionally good. Part of the reason was the
nominal peak, average, and minimum daily filtration rates
of about 3.3, 2.5, and 1.0 gpm/sq ft.
During Period IV (see Table 6-4) no polyelectrolyte was
added. Chemical treatment was achieved using FeCl3. Feed
suspended solids to the filter averaged 65 mg/£ (20-120
range) and the effluent suspended solids averaged 3 mg/£
(1-11 mg/£). This amounted to an average of 95 percent
removal.
Filter cycle times ranged from 9 to 38 hour and averaged 23
hours. Terminal total filter bed headloss was in the 5 to
6 ft of water range.
The majority of suspended solids removal occurred at the coal
sand interface as indicated by the major headloss buildup
at this location. Figure 6-27 shows the typical terminal
headloss pattern experienced during Period IV. Twenty-four
out of the 30 terminal headloss patterns observed exhibited
this major headloss buildup at the coal-sand interface.
Though penetration of suspended solids to the interface was
the rule, few filter runs were terminated due to solids
breakthrough into the effluent.
During Period V (see Table 6-5) polyelectrolyte was used
with FeCl3 chemical treatment. Filter feed suspended solids
averaged 76 mg/£ (50 to 100 range) and effluent suspended
solids averaged 4 mg/£ (1 to 8 mg/£). This resulted in an
average of 95 percent removal.
Filter cycle times ranged from 10-38 hours and averaged about
23 hours. Terminal total filter bed headloss was in the
range of 5 to 6 ft of water.
During Period V the majority of suspended solids were re-
moved in the coal layer as indicated by the major headloss
buildup at this location. Figure 6-27 shows the typical
terminal headloss pattern indicative of this situation.
Of some 20 terminal headloss patterns observed, 17 showed
this pattern. It can be concluded that the use of polyelec-
trolyte in the chemical treatment unit resulted in a more
readily "filterable" floe appearing in the filter feed.
This observation emphasizes the fact that consideration of
95
-------
FIGURE 6-27: TYPICAL HEAD LOSS VERSUS BED DEPTH PATTERN
Major S_S_ Removal
Coal-Sand :
Interface
Period V
Majjor SS
Retrieval in
coal layer
Typical Clean Bed
Head Loss @> 2.5 gpm/sq ft
10 15 20
Bed Depth, inches from floor
96
-------
the "total" treatment system must be taken into account
when evaluating the effectiveness of any given unit operation,
Backwashing procedures used during Periods IV and V were
identical to those noted previously for Period III. The
use of about 100 gal of backwash water per sq ft when ex-
periencing average filtration rates and cycle times of 2.5
gpm/sq ft and 23 hours, resulted in recycle of about 3.2 of
the filtrate.
During one backwashing, when the total bed expansion was
observed to be 22 percent, the coal layer was observed to
have expanded about 17 percent and the garnet-sand layer
25 percent. The water fluidization properties of the two
layers of media were obviously similar.
The backwash procedure used was judged to be effective in as
much as no "mud ball" problems or increase in clean bed head-
loss pattern were experienced during this 3 month period of
operation.
During previous contract work (Reference 1) using FeCl3
treatment, a substantial coating of the coal media with a
reddish-brown (iron compound) precipitate was observed. No
such coating was observed during the current study.
September to November 1973: This last plant operating Per-
iod, IV, was from September 28 to November 8 during which
time alum and polyelectrolyte were used in the chemical treat-
ment step. Regenerated and reused carbon was used in the
carbon contacting step. The peak, average and minimum daily
filtration rates were approximately 4.7, 3.5, and 1.1 gpm/sq
ft.
At the higher filtration rates, suspended solids removal
performance deteriorated somewhat. Filter feed suspended
solids averaged 38 mg/£ (16 to 109 mg/£) and the effluent
suspended solids averaged just under 8 mg/£ (3 to 19 mg/A).
The average percentage removal was thus only 79 percent.
Performance was judged to be fair to good at the higher fil-
tration rates.
Filter cycle times ranged from 8 to 31 hours and averaged
about 18 hours. Terminal total bed headless was typically
in the range of 5 to 6 ft of water. Under these operating
conditions approximately 2.9 percent of filtered water was
used for backwashing and then recycled back to the first
stage carbon contactor.
97
-------
Typically most suspended solids were removed in the coal
layer. Of some 19 terminal headless patterns observed, 13
showed patterns similar to that indicated for Period V in
Figure 6-27. Five out of the 19 observations indicated
more uniform distribution of terminal headloss as shown
previously in Figure 6-26. Comparison of suspended solids
removals patterns for Periods III and VI, and considering
differences in filtration rates, indicates that use of poly-
electrolyte in the chemical treatment unit resulted in a
more readily "filterable" floe appearing in the filter feed.
Backwashing of the filter media continued to be most effec-
tive. After 7 months of operation, no "mud balls", media
fouling or change in clean bed headloss were observed. At
the end of this study, the media was removed from the filter.
It appeared to be quite clean and free from foreign material.
No stagnant zones or "dead spots" were found in the filter
bed.
Summary Discussion
The coarse coal filter bed evaluated during Phase I did not
produce acceptable performance. The generally poor suspended
solids removals and tendencies for solids breakthrough were
unacceptable. The use of polyelectrolyte just ahead of the
filter improved performance to a marginally acceptable level.
The amount of polyelectrolyte needed, however, would result
in a prohibitive cost.
In general the performance of the tri-media filter bed was
rated good to excellent. At. relatively high filtration
rates and under certain pretreatment conditions performance
deteriorated. Use of polyelectrolyte in conjunction with
inorganic coagulants (alum or FeCls) in the chemical treat-
ment step resulted in a more readily "filterable" suspension.
Slightly higher filtration rates should be possible when
using polyelectrolyte in the chemical treatment step.
Performance and operational results indicated that the "floe
strength" of filtered suspended solids was fairly weak.
Breakthrough of suspended solids at less than terminal head-
loss and at increasing filtration rates were partial in-
dications of a "weak floe". Because of these observations
it would be strongly recommended that any filter station,
for this application, be designed to minimize sudden or
drastic changes in filtration rate.
Based on the results of this study, the authors recommend
that a tri-media filter bed, or a dual media filter bed with
slightly smaller sand media be used. Use of polyelectrolyte
in the chemical treatment step is also recommended, especially
98
-------
if inorganic coagulants are employed. Peak filtration rates
should be restricted to less than 4 gpm/sq ft, and, if
possible, an average filtration rate of 2.5 gpm/sq ft is
suggested. Facilities for feeding about 1/2 mg/£ anionic
polyelectrolyte in-line just ahead of the filter should be
provided for but only used if needed to protect plant eff-
luent quality. Auxiliary air-scour prior to water back-
washing is also strongly recommended.
The granular media filter is the last liquid-solids separa-
tion step in the PAC-PCT flowsheet. It must continuously
and reliably produce a highly clarified effluent. Conser-
vative sizing is fully justified in the opinion of the
authors.
99
-------
SECTION VII
SLUDGE TREATMENT
This section presents results and discussions of pilot plant
and laboratory sludge dew.atering studies. Gravity thickening
and vacuum filtration of the three chemical-primary sludges
and spent carbon sludge are evaluated.
GRAVITY THICKENING
Routine daily data collection around the gravity thickening
units entailed determination of feed and overflow suspended
solids concentrations, overflow and underflow volumes and
average thickener sludge depth. Underflow suspended solids
concentrations were determined on an intermittent basis,
whenever the thickened sludge inventory and storage tanks
were full. The total pounds of thickened sludge removed over
a period of time (days) was determined. Thickener solids
capture was consistently in the high 90 percent range, typi-
cally 99 percent. Thickener solids loading (TSL) in Ib of
dry solids per day per sq ft of thickener area was, therefore,
determined on the basis of underflow solids only- Thus TSL
values reported are an average for a fraction of a day to
several days. The observed TSL, and thickened feed and under-
flow suspended solids varied significantly from average
values during periods of inventory adjustments to the chemical
treatment unit and/or thickener. Typical sludge inventory
adjustment periods lasted no more than one to two days.
The approach used to identify pilot plant thickener perfor-
mance was to plot feed and underflow suspended solids con-
centration, sludge depth and average TSL for each day of
operation for all sludges. Periods of reasonably stable
thickener operation (TSL and sludge depth) were identified.
For these stable operation periods, average pertinent chemi-
cal treatment unit and thickener operating condition and
thickener performance data are reported.
The method of conducting and the approach to analyzing the
results of laboratory thickening tests are contained in
Appendixes A and B, respectively. As indicated in Appendix B,
100
-------
the graphical construction method used to reduce laboratory
data resulted in a variance of predicted TSL. The precise
extent of these variances was not determined. The labora-
tory results presented for each test made are predicted TSL
values at various underflow suspended solids concentrations
(g/Jl) and maximum obtainable underflow suspended solids con-
centration.
Lime-Primary Sludge Thickening
Lime-primary sludge was produced in the chemical treatment
unit under a relatively narrow range of operating conditions.
Hydraulic loading (0.5 to 1.0 gpm/sq ft), pH (10.8 to 11.2)
and pumping turbine speed (24 to 34 rpm) were all closely
controlled. The only major operational variable change was
constant versus diurnal flow.
Monthly average lime-primary sludge production results ranged
from 6100 to 7000 Ib of dry solids per MG of plant flow. The
weighted average sludge production was 6600 Ib/MG. The
monthly average lime dosages ranged from 410 to 510 mg/£ of
Ca(OH)2, with a weighted average of 453 mg/£. This resulted
in a weighted average sludge production of 1.7 Ib of chemical
and primary sludge per Ib of lime used (1.66 to 1.86 range).
Sludge production values are consistent with those previously
reported.1
The following assumptions were made in an effort to identify
constituents of the lime-primary sludge. These assumptions
are based on changes in pertinent water quality parameters
caused by lime treatment.
1) Reduction in bicarbonate alkalinity resulted from
precipitation of CaCC>3,
2) Reduction in non-calcium hardness resulted from
precipitation of Mg(OH)2 and
3) Reduction in phoshporus resulted from precipitation
of Ca5OH(P04)3.
Based on the above assumptions computed average lime-primary
sludge constituents were:
Sewage Solids 13.3
Mg(OH)2 8.3
CasOH(P04)3 9.4
CaC03 69.0
TOTAL 100.0%
101
-------
The average computed sludge production value was found to be
95 percent of the observed value of 6600 Ib/MG. A calcium
balance across the chemical treatment unit indicated some
unaccounted calcium precipitate. It is possible that some
unreacted lime, Ca(OH)2, and/or calcium-organic compounds
were present in the sludge accounting for the slight diff-
erence between computed and observed sludge production
values. Another probable explanation was that CO2 was ab-
sorbed in stored samples of chemical treatment unit effluent
resulting in erroneously high bicarbonate alkalinity deter-
minations on the chemical treatment unit effluent sample.
That this may have occurred was indicated by stored composite
sample pH ' s being several tenths of a pH unit lower than
daily averaged grab sample pH values. It is, therefore,
suggested that the unaccounted for calcium compound in the
lime-primary sludge was
Based on the above analysis the lime-primary sludge was about
87 percent chemical precipitates. A very large portion was
CaCC>3. If the CaC03 value could be recovered by a classifi-
cation and recalcination procedure achieving 100 percent
recovery of lime, approximately 80 percent of the plant's
required lime dosage could be satisfied.
Laboratory test data indicated that the overall lime-primary
sludge solids density was 2.0 to 2.6 g/cc. Sludge slurry
density was found to be 1.00 plus 0.000553 times the slurry
concentration in g/A.
No attempt was made to determine the particle size and dis-
tribution of sludge solids . The CaC03 precipitate formed in
municipal wastewater treatment is much more finely divided
than that formed in potable water treatment by lime soften-
ing. Presumably, the presence of condensed phosphates and
soluble organic compounds inhibit particle growth.
Pilot Plant Results
During most of the approximately six months of lime treat-
ment, feed and underflow to and from the thickener was auto-
mated. The thickener was fed at 20 to 60 minute intervals
at about 200 gpm for 6 to 10 second durations. Thickener
blowdown was at the same time interval and at about 150 gpm
for 1 to 3 second durations.
During the last 3 weeks of lime treatment, the thickener
feed was pumped at a rate of about 5 gpm or less. Also
during this period, pickets were attached to the sludge rakes.
10;
-------
Table 7-1 presents average operating and performance data for
four periods of reasonably stable thickener operation and
performance. The effect of chemical contactor flow variation
on underflow solids concentration was quite dramatic. The
underflow suspended solids concentration of the chemical con-
tactor was reduced to less than 1/2 at 2:1 peak to minimum,
diurnal flow compared to constant flow operation. Constant
pumping of the chemical contactor underflow resulted in a
slight additional reduction in underflow suspended solids
concentrations.
In general, thickener performance at a TSL of 15 to 20 lb/day/
sq ft was considered poor. Thickener hydraulic loadings were
well below 50 gpd/sq ft resulting in excellent suspended
solids capture of 99+ percent. Typical thickener solids pro-
files shown in Figure 7-1 illustrate the poor thickening per-
formance .
Laboratory Test Results
Twenty-seven laboratory gravity thickening tests were run
from November, 1972, to February, 1973.
Table 7-2 presents laboratory test results. The test number
denotes the date the sludge sample was grab-composited. For
example, #0116 was January 16, 1973, and #1129 was November
29, 1972. Test #'s 0220A, B and C and 0129 were for single
grab samples. Sample #0220 was partially thickened and then
two portions diluted with supernatant to result in lower
suspended solids concentrations.
Numerous attempts were made to define reasonably precise
relationships between TSL and feed suspended solids for
various desired underflow suspended solids. None were found!
Figure 7-2 shows maximum underflow suspended solids concen-
trations obtained and the predicted TSL to achieve same for
various feed suspended solids concentrations. The linear
regression line shown, which has a correlation coefficient
of 0.96, does not include data points (g) , which were obtained
for single grab samples of sludge.
The maximum underflow suspended solids concentration is seen
to increase directly proportional to the feed suspended
solids concentration. This type of response has been report-
ed by others for similar and different types of sludges.
It was thought that for a given particle size, distribution
and density and chemical character of sludge, that the maxi-
mum obtainable underflow suspended solids concentration
103
-------
TABLE 7-1:
LIME-PRIMARY SLUDGE THICKENING
AVERAGE PILOT PLANT RESULTS
Calendar
Chemical
Dates
Contactor:
8/27/72
to
9/27/72
11/8/72
to
12/1/72
1/4/73
to
1/31/73
2/1/73
to
2/22/73
Flow:
Average gpm/sq ft
Peak gpm/sq ft
Peak : Minimum
Lime Dosage, mg/£
Sludge Production,
Ib/MG
Turbine Speed, rpm
Reaction Zone-SS, g/£
Bowdown-SS, g/£
1.0
1.0
1:1
425
6600
34
6.8
177
1.0
1.2
2:1
410
6100
34
3.1
73
0.85
0.85
1:1
510
7000
26
2.2
88
0.93/0
0.93/0
1:1
470
6700
26
2.8
67
.75
.75
Gravity Thickener:
Solids Loading,
Ib/day/sq ft 20
Solids Capture, % 99 +
Blowdown-SS, g/£ 207
Sludge Depth, ft ~5
18
99 +
109
~5
17
99 +
108
5
-15
99 +
84
4-5
104
-------
FIGURE 7-1: LIME PRIMARY SLUDGE THICKENING
THICKENER SOLIDS PROFILE
T 6 8
Total Solids, percent by weight
105
-------
TABLE 7-2: LIME-PRIMARY SLUDGE: LABORATORY THICKENING TEST RESULTS
Test
#
0220Ca
0116
0220Ba
0218
0206
1129
0204
0215
0114
0107
0213
0109
0125
1117
0220
0123
0129
0121
1127
0130
0111
0220Aa
0118
Feed
SS,
g/£
39.5
40.4
46.9
56.3
57.5
61.0
61.8
62.9
64.4
65.5
66.5
70.4
75.3
78.0
78.4
84.1
84.3
92.3
94.5
97. 8
99.2
114
123
Thickener Solids Loading
Ib/day/sq ft @ Underflow SS (g/H)
80
30
7.5
41
19
71
35
48
27
86
37
46
95
--
—
--
__
—
—
—
—
—
—
—
100
17.1
5.6
24
10
24
15
21
15
37
13
16
30
30
76
24
106
62
—
—
—
—
—
—
120
11.6
4.8
15 = 6
7.2
14
9.6
12
9.4
26
83
11
16-
12+
34
18
42
23
29
73
88
44
—
—
140
9.4
—
11.8
6.1
11
--
9.4
7.3
21
6.6
8.4
12-
6.4
22
12
26
12
17
34
38
22
72
72
160
— —
9.9
—
__
--
—
6.3
19
5.8
—
9.8
5.6
14.5
9.5
18
8.1
13
22
23
14
36
29
170 180
« _ w M
__
__
__
18
__
— —
8.0
—
8.8
15
6.7
11
19
9.9
12
21
20
of:
200 Max.
8.2
— —
8.2
—
8.9
__
8.1
—
__
__
7.7
— —
4.8
__
—
12.8
5.5
9.7
— —
3.5
9.9
17 14.6
15 11.5
Max.
Under-
flow
SS, g/£
159
—
180
—
151
—
158
—
--
—
155
—
175
--
202
196
—
—
188
—
223
223
Critical
Point in
inch/7
inch
1.5
2.1
1.7
1.1
1.5
3.6
2.1
3.4
2.5
2.4
2.3
1.7
1.4
2.5
2.9
2.0
1.1
2.0
2.8
1.7
2.0
2.5
1.8
Single grab samples
-------
FIGURE 7-2: LIMB PRIMARY SLUDGE THICKKRIMG - LABORATORY
PREDICTED MAXIMUM UNDERFLOW RELATIONSHIPS
22ft
30 50 70 90 110 130
Feed Suspended Solids Concentration, g/£
107
-------
would be constant. Apparently the thickening properties are,
in part, effected by the feed suspended solids concentration.
It is probable that the laboratory test procedures used had
an effect. As feed suspended solids concentrations increased,
the final volume of thickened sludge in the 2 liter graduated
cylinder increased slightly. 'For example, as feed suspended
solids increased from 60 to 120 g/S, the regression curve on
Figure 7-2 indicates that the final volume of thickened
sludge increased 45 percent (from 0.38 to 0.55 liters).
This deeper sludge layer may have provided increased com-
pressive forces resulting in a higher final suspended solids
concentration.
Also shown on Figure 7-2 is the predicted TSL required to
achieve the maximum underflow suspended solids concentrations.
The linear regression line shown does not have a statisti-
cally significant positive slope (at a 95 percent confidence
level) . This fact indicates that as feed suspended solids
concentration increases, higher TSL's are not possible if
maximum underflow suspended solids concentration is desired.
A single series of laboratory tests were made to determine
the effect of a polyelectrolyte on thickening response. The
results of this are shown in Table 7-3.
TABLE 7-3:
LIME-PRIMARY SLUDE THICKENING:
EFFECT OF POLYELECTROLYTE
Test
No.
222-1
222-2
222-3
222-4
Feed
SS
31
31
31
31
Max.
Under-
Thickener Solids Load- Flow
ing (Ib/day/sq ft) @ SS,
Underflow SS (g/SL) of: g/SL
50 60 80
22 13 8.3
-- 42 19
-- 65 29
— — 61
100
6.8
13
18
48
120
-- »89
11 >116
14 123
42 >116
Criti- Polymer
cal Dosage
Point Ib/Ton
inches/ Dry
7 inch Solids
1.2
1.1
1.4
2.6
0.6
1.0
1.9
3.9
The polyelectrolyte used was anionic AltaSep 2A2. No screen-
ing of other potentially effective polyelectrolytes was
made. Previous studies had indicated that the Altasep 2A2
108
-------
was very effective in increasing the settling rate of rela-
tive dilute slurries (about 800 mg/&) resulting from lime
treatment.l The dramatic effect of polyelectrolyte is shown
on Figure 7-3. It should be noted, however, that the feed
suspended solids concentration was quite low (31 g/£). The
effect of polyelectrolyte on predicted TSL for higher feed
suspended solids would be expected to be less dramatic. As
seen in Table 7-3, the maximum underflow suspended solids
concentration reached did not seem to be greatly effected by
use of polyelectrolyte. The major effect is seen to be a
dramatic increase in initial sludge subsidence rate as shown
on Figure 7-4.
Comparison of Pilot Plant and Laboratory Results
It was originally intended to collect laboratory and pilot
plant data so that a direct comparison could be made. The
intended approach was to, before the fact, identify days of
-reasonably stable pilot plant thickener operation. Perfor-
mance data would be collected and a grab-composite sample
obtained for laboratory tests. Because of an inability to
predict operational problems before the fact, judgement was
not always sound. Of the 19 laboratory tests conducted for
the above purpose, only five comparisons indicated similar
laboratory and pilot plant results. The lack of precise
correlation between pilot plant and laboratory results pre-
sumably was due, in part, to variability of plant operating
conditions, sludge sampling, laboratory test procedures and/
or laboratory data reduction procedures.
In general, the laboratory predicted TSL and underflow sus-
pended solids concentrations were considerably in excess of
pilot plant results. For example, the ratio of underflow
suspended solids concentrations achieved by the pilot plant
to that predicted by the laboratory results, at the actual
pilot plant TSL, ranged from 0.50 to 1.09, with an average
of 0.75 for the 19 comparisons made. The ratio of pilot
plant TSL to laboratory predicted TSL, at the pilot plant
underflow suspended solids experienced, was much more vari-
able.
The lack of a definable precise correlation based on one on
one comparisons of pilot plant and laboratory results promp-
ted a different approach. Inspection of pilot plant per-
formance data in Table 7-1 shows that the average TSL was
just under 18 Ib/day/sq ft. The laboratory predicted under-
flow suspended solids concentrations at a TSL of 18 lb/day/
sq ft were determined, from plots of data in Table 7-2, and
then plotted vs feed suspended solids as per the circles on
Figure 7-5. The linear regression curve shown for these
109
-------
FIGURE 7-3:
LIME PRIMARY SLUDGE THICKENING -
EFFECT OF POLYELECTROLYTE ON PREDICTED TSL
1 234
Polymer Dosage, lb/ton/dry solids
110
-------
FIGURE 7-4: LIU! PRIMARY SLUDGE THICKENING - EFFECT
1,2-, I, , .Q^fVLY^LECTflffTJ^^^
Polymer Dosage, Ib/ton dry solids
111
-------
FIGURE 7-5: LIME PRIMARY SLUDGE THICKENING - COMPARISON
OF LABORATORY AND PILOT PLANT RESULTS
210
- A-
170
to -P
CO <4-l
O to
SO)
I"
•O
5
O
1-1
TJ
90
50
T
Prom Reference #1
Figure 35
A
T
A Pilot Plant
(T) Laboratory
120
Feed Suspended Solids Concentration,
-------
data has a correlation coefficient of 0.91. The triangles
on Figure 7-5 are pilot plant results from Table 7-1. It is
readily apparent that laboratory predicted values are gener-
ally in excess of pilot plant performance. The limited num-
ber of pilot plant data points on Figure 7-5 does not allow
statistical definition of a scale-up factor.
For comparative purposes the O data point on Figure 7-5 is
shown. This data point was obtained from Figure 35 of the
previous contract report for laboratory test results.1 The
closeness of fit with current contract results implies a
similarity in the two sets of data.
Summary Discussion
As previously noted pilot plant performance was less than de-
sired and expected. Initially it was felt that the current
results could not be explained, based on previous contract
experience. After evaluation of the data it becomes apparent
that the reason for less than desired thickener underflow
suspended solids concentration was the reduced chemical
clarifier underflow suspended solids concentration experienc-
ed due to diurnal flow and poor plant operation. This ob-
servation indicates the need to precisely determine the
clarifier underflow to be experienced prior to designing pro-
totype thickeners. It may be more rational to use a signifi-
cant safety factor in predicting clarifier underflow suspend-
ed solids concentration (i.e., thickener feed), than to use
a safety factor in predicting thickener performance.
A precise correlation between pilot plant and laboratory
test thickening results was not defined. The ratios of
pilot plant to laboratory test predicted underflow suspended
solids concentrations shown on Figure 7-5 were 0.75, 0.91
and 0.76 for the three data points shown. These data were
for similar pilot plant and laboratory test TSL values.
This observation suggests an approach to using the labora-
tory test data for predicting pilot plant results. The
following is an example of this suggested approach.
Assume a thickener feed suspended solids concentration of
80 g/£. If maximum thickener underflow is desired, Figure
7-2 indicates that a TSL of 8 Ib/day/sq ft should be used.
The laboratory data predicted the underflow would be 178
If the pilot plant to laboratory predicted ratio existing
for data on Figure 7-5 were applicable, a scale-up factor
of about 0.75 would apply- Thus the expected underflow
would be 0.75x178 or 134
113
-------
A thickener operating with a 3 ft deep sludge layer, a 134
g/H underflow suspended solids and a TSL of 8 Ib/day/sq ft
would have a sludge SRT of less than 3 days,
Alum-Primary Sludge Thickening
Alum-primary sludge was produced and thickened for just over
a 3 month period during April, May and October of 1973.
Analysis of thickened sludge inventory results and chemical
clarifier feed and effluent suspended solids indicated that
the sludge contained approximately 15 to 20 percent, by
weight, chemical solids. Slurry densities ranged from 1.05
g/cc for about 20 g/£ slurries to 1.10 g/cc for 60 g/H
slurries.
Pilot Plant Results
The pilot plant was operated with diurnal flow variation
during the entire alum treatment period. The chemical clari-
fier blowdown was at a rate of about 200 gpm for 6 to 10
seconds at 60 + 20 minute intervals. The blowdown flowed by
gravity to a 150 gal surge tank from which the sludge was
pumped to the thickener at less than 5 gpm. During April
and May, 1973, thickener blowdown was automated. The
weighted average duration of blowdown was 4.5 seconds
(range 3 to 10) and the weighted average frequency of blow-
down was each 75 minutes. During October sludge was manually
blowndown from one to five times each day by plant operators
in order to maintain the sludge level between two specified
sample taps. Thickener blowdown rate was about 150 gpm for
both automatic and manual operation.
Table 7-4 presents average results for three periods of
reasonably stable thickener operation and performance. The
effect of going from a 2:1 to 3:1 peak:minimum diurnal flow
variation on chemical clarifier underflow suspended solids
concentration was substantial. The slightly higher average
hydraulic loading was probably partly responsible for the
reduction in underflow suspended solids from 27 to 16 g/£.
Comparison of April and October, 1973, results indicate that
the use of 0.25 mg/£ of anionic polyelectrolyte in the
chemical unit did not result in expected higher underflow
suspended solids concentration.
It is interesting to note that the underflow suspended
solids concentrations of 15 to 27 g/£ achieved were from 4
to 15 times greater than experienced during previous con-
tract work at constant flow conditions. Obviously the
baffle plate shown on Figure 4-3 was very effective in iso-
lating the sludge withdrawal region from the mixing re-
action region.
114
-------
TABLE 7-4
ALUM-PRIMARY SLUDGE THICKENING
AVERAGE PILOT PLANT RESULTS
Calendar Date
4/10/73
to
4/26/73
5/13/73
to
5/21/73
10/17/73
to
11/9/73
Chemical Contactor:
Flow:
Average gpm/sq ft 0.35
Peak gpm/sq ft 0.45
Peak:Minimum 2:1
Alum Dosage, mg/£ 140
Polymer Dosage, mg/£ 0
Sludge Production, Ib/MG 1700
Turbine Speed, rpm 21
Reaction Zone-SS, g/£ 1.5
Slowdown-SS, g/£ 27
0.43
0.60
3:1
120
0
1100
21
1.7
15.6
0.43+0.10
0.55+10
2:1
120
0.25
1200
21-KL8
0.93
15.3
Gravity Thickener:
Solids Loading,
Ib/day/sq ft
Sludge Level, ft
Solids Capture, %
Blowdown-S S, g/ £
1.7
99
75
7-3
7.1
1-2
99
28
1-2
7.8
2-4
99
28
2-4
115
-------
As the thickener data in Table 7-4 show, during April the
TSL was a very low 1.7 Ib/day/sq ft. Thickener underflow
suspended solids were an impressive 75 g/£. Of importance,
however, is the relative long sludge SRT of 3-7 days.
During start-up of the thickener, about 2 weeks was required
to buildup a sludge depth of 4 ft. At that time, thickener
blowdown was started and the sludge level gradually reduced
to a 1 ft depth over a 2-1/2 week period. After 1-1/2 weeks,
the sludge turned septic and by the end of the 2-1/2 week
period was black, very odorous and releasing minute gas
bubbles.
On April 26, the 8 ft diameter thickener was taken off stream.
A 3.5 ft diameter by 6 ft deep unit was fabricated, installed
inside the 8 ft unit and brought on stream in early May. No
septicity problems were experienced during subsequent opera-
tion at approximate sludge SRT's of 1 to 4 days during May
and October, 1973.
Data in Table 7-4 indicates that the use of polyelectrolyte
in the chemical unit did not improve pilot plant thickener
performance. This observation was contrary to what was ex-
pected.
Figure 7-6 presents several thickener solids vs depth pro-
files determined during the period polyelectrolyte was used.
It is seen that about 2 to 3 ft of sludge was required to
reach a maximum thickened solids concentration.
Laboratory Test Results
Laboratory results are presented in Tables 7-5 and 7-6. All
tests shown were run on 24 hour grab-composite samples. In-
spection of the "critical point" data indicates an average
of 1.2 for Table 7-5 and 2.2 for Table 7-6. Consequently,
predicted TSL in Table 7-5 are relatively more conservative
than in Table 7-6 (see Appendix B for discussion of critical
point) .
As indicated for the lime-primary sludge results, there is
a general increase in maximum obtainable underflow suspended
solids concentrations as feed suspended solids increases.
This data, plotted on Figure 7-7, shows that no precise re-
lationship exists for all combined data. There is a defin-
ite trend in the April and May results (series 400 and 500
test #'s). Correlation coefficients for the two sets of
data shown was found to be 0.89 and 0.95, respectively.
The data obtained when polyelectrolyte were used in the
chemical unit (Table 7-6 and Q's on Figure 7-7) are not
well correlated. For this data, the average feed suspended
116
-------
FIGURE 7-6: ALUM RIIABT SLUDGE THICKENING -
SOLI
1.0 2.0
Total Solids, percent by weight
3.0
117
-------
TABLE 7-5: ALUM-PRIMARY SLUDGE: LABORATORY THICKENING TEST RESULTS
oo
Test
#
512
506
410
412
510
429
426
405
501
422
424
520
417
419
325
Feed
SS,
g/i
11.5
13.9
15.2
16.1
17.0
17.7
19.4
20.5
22.1
24.3
25.2
29.5
32.7
41.7
50.2
Thickener Solids Loading,
Ib/day/sq ft @ Underflow SS
20
27
42
—
—
67
107
—
—
—
—
—
—
—
__
30
6.9
24.0
13.0
21.0
24.0
31.0
38.0
—
58.0
52.0
109.0
—
—
—
40
5.3
15.0
8.1
9.9
11.0
16.4
23.7
23.0
35.0
33.0
47.0
78.0
74.0
—
50
__
7.1
4.5
5.13
5.4
6.7
12.7
13.0
20.0
19.0
24.0
52.0
37.0
56.0
60
__
—
__
—
—
__
—
9.0
9.8
11.8
14.0
43.0
19.0
28.0
35.0
(g/D of:
70
__
—
—
—
—
—
—
—
--
--
10.9
32.0
14.0
16.0
18.0
Max .
5.3
5.0
4.0
5.3
4.0
5.4
10.3
7.4
7.4
10.8
11.5
-12
13.8
9.2
9.5
Max .
Under-
flow
SS,g/l
43
60
55
49
57
55
54
70
68
64
67
98
71
82
81
Criti-
cal
Point in
Inch/7 Inch
0
0
1
0
1
1
1
1
0
1
1
1
1
1
2
.9
.9
.0
.9
.2
.0
.0
.8
.9
.8
.5
.0
.6
.7
.1
without polyelectrolyte to chemical contactor
-------
TABLE 7-6: ALUM-PRIMARY SLUDGE: LABORATORY THICKENING TEST RESULTS
Test
#
1101
1030
1023
1106
1028
1104
1009
Feed
SS,
g/i
10.1
12.4
12.6
14.9
15.8
22.3
24.9
Thickener Solids Loading,
Ib/day/sq ft @ Underflow SS (g/1) of:
20 25 30 35
22 15 10.3
36 — 12.0 9.8
73 — 23.0
75 35 20.0 13.8
64 28 17.0
60.0
67.0
40 50 Max.
Q O
8.4 — 8.2
12.6 9.9 9.9
13.3
10.1 — 10.7
24.0 15.0 15.0
21.0 12.6 13.5
Max.
Under-
flow
SS,g/l
33
42
51
36 '
39
50
48
Criti-
cal
Point
Inch/7 Inch
2.1
2.3
1.8
2.2
2.3
2.3
2.5
with polyelectrolyte to the chemical contactor
-------
FIGURE 7-7: ALUM PRIMARY SLUDGE THICKENING - MAXIMUM
LABORATORY UNDERFLOW CONCENTRATION
100
CO
w
I
g
i
0
20 30
Feed Suspended Solids Concentration, g/£
40
-------
solids was 15.3 g/£ and the average maximum underflow sus-
pended solids was 43
Interrelationships between predicted TSL, feed suspended
solids and underflow suspended solids are shown on Figures
7-8 and 7-9. It is apparent that relatively precise re-
lationships were found. Comparison of predicted TSL curves
for underflow suspended solids of 40 g/£ and the maximum g/£
on Figure 7-8 and 7-9, for the feed suspended solids in the
15 to 25 g/£ range, indicates the use of polyelectrolyte
in the chemical clarifiers resulted in substantially higher
TSL's. The difference diminished to a very small value at
25 g/&. It should be noted, however, that in the 15 to 25
g/£ feed suspended solids range, maximum underflow suspended
solids during April, without polyelectrolyte, were consider-
ably higher than during October with polyelectrolyte (-60 vs
43 g/£) .
Comparison of Pilot Plant and Laboratory Results
Eighteen one on one comparisons between pilot plant and lab-
oratory test results were made. Of these, fourteen labora-
tory test results predicted that the pilot plant should have
achieved maximum obtainable underflow suspended solids.
This observation is realistic for the April data when the
pilot plant thickener was grossly underloaded. That this
condition existed is indicated by the VV data point on
Figure 7-8, which represents average pilot plant data from
Table 7-4. During the operating period indicated, the ratio
of pilot plant to laboratory underflow suspended solids con-
centrations was 75/65, or 1.2. From Figure 7-8 it is seen
that the laboratory predicted TSL to reach maximum under-
flow suspended solids was just over 10 Ib/day/sq ft, which
was about 6 times the average pilot plant TSL. It was this
observation that prompted installation of a new pilot plant
thickener of about one-fifth the area of the original 8 ft
diameter unit.
Only one laboratory test was run during the May operating
period indicated in Table 7-4. Thus a direct comparison is
impossible. The April laboratory results on Figure 7-8 in-
dicates that a TSL of 7.1 (the May data average) an under-
flow suspended solids of just over 40 g/& should have been
achieved. The 28 g/H actually achieved is 70 percent of
the 40 g/£.
Seven laboratory tests during October, 1973, indicated that,
at the average TSL of 7.8 (see Table 7-4) , an average max-
imum underflow of 43 g/£ should have been achieved. The
28 g/£ actually achieved is 65 percent of the predicted
121
-------
FIGURE 7-8; ALUM PRIMARY SLUDGE THICKENING - PREDICTED TSL VERSUS FEED
SUSPENDED SOLIDS CONCENTRATION AT VARIOUS UNDERFLOW CONCENTRATIONS
CO
April 5 -_26,_. 19t3
Feed Suspended Solids Concentration, g/£
-------
FIGURE 7-9 : ALUM PRIMARY SLUDGE THICKENING - PREDICTED
TSL VERSUS PEED SUSPENDED SOLIDS CONCENTRATION AT VARIOUS
UNDERFLOW CONCENTRATIONS
5 10 18 20
Feed Suspended Solids Concentration,
25
123
-------
43 g/£. The average pilot plant data for the October period
is shown on Figure 7-9 as a M. - It can be seen that to pro-
duce an underflow of 28 g/^/tihe laboratory data predicted
a TSL of just over 23 Ib/day/sq ft. The actual TSL of 7.8
Ib/day/sq ft is 33 percent of the predicted 23 Ib/day/sq ft.
Summary Discussion
Pilot plant results indicated that sludge SRT's in excess of
3 to 4 days should be avoided to prevent septic sludge con-
ditions. Pilot plant results also indicated that use of
polyelectrolyte in the chemical unit did not increase pilot
plant chemical unit or thickener underflow suspended solids
concentration. This result was contrary to expectations
and to laboratory test results.
Comparison of pilot plant and laboratory test results were
similar to that for the lime-primary sludge data. That is
to say, for an underloaded thickener, at a given TSL the
pilot plant underflow suspended solids concentration was 0.65
to 0-70 of that predicted by laboratory test results. This
observation implies that had the pilot plant thickener been
operated at a TSL of 10 to 12 rather than the 7.8 Ib/day/sq
ft shown in Table 7-4, that the underflow suspended solids
concentration would have been the same. The validity of this
conjecture is doubtful.
Ferric-Primary Sludge Thickening
The pilot plant was operated for just over three months using
ferric chloride in the chemical treatment unit. Analysis of
thickened sludge inventory data and chemical treatment unit
feed and effluent suspended solids concentrations indicated
that the ferric-primary sludge was about 15 to 30 percent
chemical solids. It was conjectured in Section VI that
considerable ferric iron was reduced to soluble ferrous iron
within the chemical treatment unit. After a few weeks of
operation, the chemical clarifier underflow sludge turned
black due to the presence of a finely divided black precipi-
tate. Qualitative analysis of the sludge solids indicated
the presence of considerable sulfides which were released
as H2S upon acidification. Quantitative analysis for sol-
uble sulfides indicated only trace amounts. It was con-
cluded that considerable FeS was present in the ferric-
primary sludge. The exact amount was not determined.
Though the sludge was black, it had no other characteristics
of septicity (e.g., odors or gas bubbles).
124
-------
Pilot Plant Results
Average pilot plant operating and performance data are pre-
sented in Table 7-7. It is seen that the only operational
variation was the use of a reduced ferric chloride dosage
and 0.25 mg/£ of anionic polyelectrolyte during the last
month of operation. Contrary to results with alum treatment,
the use of polyelectrolyte produced a substantial improvement
in the chemical unit underflow suspended solids concentration
(from 11 to 23 g/£) .
The only thickener operational difference between the first
and second time periods shown in Table 7-7 was the method of
thickener blowdown. During the first period blowdown was
automated. At about 20 to 40 minutes intervals approximately
150 gpm was blowndown for 1 to 5 second durations. Because
of this automated feature, minimal operator attention was
given to the thickener. Consequently, wide variations in
sludge depth were experienced. In addition, frequent clogg-
ing of the thickener feed pump occurred. This caused chemi-
cal clarifier blowdown to overflow the surge tank directly
into the 430 gal thickener at about 200 gpm for 6 to 10
seconds each 20 to 40 minutes. The result was a high level
of turbulence in the thickener. The two very steep total
solids vs depth profiles shown on Figure 7-10 were typical
during this operating period. Thickener underflow suspended
solids ranged from 8 to 20 g/H and averaged only 15
During the second operating period in Table 7-7, thickener
blowdown was manually accomplished. Sludge depth was main-
tained at a specified level, plus or minus 1 ft. In addi-
tion, the thickener feed pump operation was frequently moni-
tored, minimizing hydraulic surges to the thickener. Thick-
ener blowdown suspended solids data in Table 7-7 show that
more careful operation of the thickener resulted in a doubl-
ing of concentration from 15 to 31 g/&. A slight improve-
ment in solids capture was also experienced.
During the last operating period, when polyelectrolyte was
added to the chemical unit, thickener performance was dramat
ically improved. This improvement is shown by results in
Table 7-7 and sludge solids profiles on Figures 7-10.
Laboratory Test Results
The results of twenty laboratory tests are presented in
Tables 7-8 and 7-9. These tests were conducted on samples
with a very limited range of feed suspended solids concen-
trations. Consequently, any interrelationship between TSL
and underflow suspended solids and feed suspended solids
125
-------
TABLE 7-7
FERRIC-PRIMARY SLUDGE THICKENING
AVERAGE PILOT PLANT RESULTS
Calendar Dates
6/10/73
to
7/20/73
7/21/73
to
8/8/73
8/27/73
to
9/17/73
Chemical Contactor:
Flow:
Average gpm/sq ft
Peak gpm/sq ft
Peak : Minimum
Fed 3 Dosage, mg/£
Polymer Dosage, mg/&
Sludge Production, Ib/MG
Turbine Speed, rpm
Reaction Zone-SS, g/£
Blowdown-SS, g/£
0.38
0.50
2:1
98
0
1100
12-16
1.0
11
0.38
0.50
2:1
98
0
1200
12-16
1.0
11.2
0.39
0.50
2:1
62
0.25
1000
23
1.3
24
Gravity Thickener:
Solids Loading,
Ib/day/sq ft
Sludge Depth, ft
Solids Capture, %
Blowdown-SS, g/Jl
6.3
erratic
95 +
15
6.7
4
97+
31
5.8
2-4
99
56
126
-------
FIGURE 7-10;
FERRIC PRIMARY SLUDGE THICKENING
THICKENER SOLIDS PROFILE
August 27, 1973
to
September 17, 1973
June 10, 1973
to
«$Uly 20, 1973
234
Total Solids, percent by weight
127
-------
TABLE 7-8: FERRIC CHLORIDE-PRIMARY SLUDGE:
LABORATORY GRAVITY THICKENING TEST RESULTS
NJ
00
Test
#
621
705
619
805
612
626
617
701
624
610
726
708
703
Averages
Feed
SS,
g/i
2.2
5.7
6.1
7.0
7.3
7.3
8.0
8.0
8.2
9.0
9.4
10.9
11.0
: 7.7
Thickener Solids Loading,
10
9.2
28
12
36
21
23
16
—
65
—
—
__
—
Ib/day/sq
20
4.5
14
3.4
18
7.7
6.4
3.4
9.0
24
8.8
—
49
39
ft
30
3.9
7.2
1.7
11
3.0
2.4
1.7
5.2
10
3.8
16
29
23
@ Underflow
40
3.7
5.7
—
8.1
2.1
1.8
--
4.3
4.8
2.5
10
17
14
SS
50
_ —
5.1
__
6.8
--
1.6
--
3.9
3.4
2.1
8.1
13
8.0
(g/D of:
Max.
3.6
4.6
1.5
6.3
1.8
1.6
1.6
3.7
2.6
2.0
6.5
11.0
4.6
4.0
Max.
Under-
flow
SS,g/l
43
57
36
61
48
52
33
57
68
57
75
58
71
55
Criti-
cal
Point in
Inch/7 Inch
1.3
1.6
1.3
1.4
1.2
1.3
1.9
2.3
1.1
1.2
1.9
1.3
0.9
1.4 +
-------
TABLE 7-9:
FERRIC CHLORIDE-PRIMARY SLUDGE:
LABORATORY GRAVITY THICKENING TEST RESULTS
Thickener Solids Loading
Test
No.
907
911
830
913
910
Feed
SS
g/£
20.8
21.3
21.7
25.5
26.1
(Ib/day/sq ft) @ Under-
flow
30
89
95
94
123
202
SS (g/£)
40
45
40
38
31
53
of:
50
28
22
20
--
22
Max
23
18
18
19
20
Max.
Under-
flow
SS,
g/£
60
57
52
47
53
Critical
Point
inch/
7 inches
2.7
1.8
2.2
2.1
2.1
Aver- 23.1
ages
20
54
2.2
concentrations cannot be defined. For this reason, average
results are shown in Tables 7-8 and 7-9.
Comparison of average maximum obtainable suspended solids
concentrations from Tables 7-8 and 7-9 indicate no differ-
ence. The higher feed suspended solids concentrations and
polymer dosage for Table 7-9 results apparently had no eff-
ect on maximum suspended solids concentration. Recall that
for both lime and alum-primary sludges, increased feed sus-
pended solids results in increased underflow solids concen-
trations. The average predicted TSL's to reach maximum
underflow solids when using polyelectrolyte in the chemical
unit is seen to be five times higher than when polyelectro-
lyte was not used. This direction of change was expected.
Evaluation of data in Table 7-8 indicates a general positive
trend of increasing maximum underflow solids concentrations
and predicted TSL at those concentrations as feed suspended
solids increases. The data ranged from about 1.5 lb/day/
sq ft at 35 g/H to 5 Ib/day/sq ft at 70 g/fc. The high TSL
value for test #708 is obviously out of line.
129
-------
Comparison of Pilot Plant and Laboratory Results
For the first two plant operating periods in Table 7-7, an
average TSL of 6.5 Ib/day/sq ft was experienced. An average
predicted underflow solids concentration of 33 g/£ was found
for all laboratory tests shown in Table 7-8. This value com-
pares well with that shown for the second pilot plant time
period in Table 7-7. It should be recognized, however, that
the average feed solids for the laboratory test data was
7.7 g/A compared to over 11 g/fc for the pilot plant.
When polyelectrolyte was being used, laboratory test results
predicted that maximum underflow would be achieved at the
pilot plant TSL. Pilot plant results in Table 7-7 indicate
maximum predicted underflow solids concentration was achieved.
For a one on one comparison of 18 days of results, the range
of pilot plant to laboratory test underflow solids concentra-
tion was from 0 . 2 to 1. 0 . In general, because of limited
range of data compiled, no useful correlation was found be-
tween pilot plant and laboratory test results.
Summary Discussion
Pilot plant operating experience indicated that, at TSL's of
4 to 9 Ib/day/sq ft and sludge SRT's of 1 to 4 days, no
septicity problems were encountered.
Both pilot plant and laboratory test results indicate that
the addition of polyelectrolyte to the chemical unit had a
dramatic effect on underflow solids concentration. Since
no meaningful correlation was found between pilot plant and
laboratory results, it is impossible to suggest a definitive
scale-up factor for the laboratory test results. It would
seem reasonable to expect a prototype thickener to produce
near maximum underflow solids concentration of 50+ g/£
from a feed solids concentration of about 24 g/H at a TSL
of 10 to 15 Ib/day/sq ft,
Spent Carbon Sludge Thickening
All spent carbon produced during this study was thickened in
a 5-ft diameter by 5-ft deep pilot plant gravity thickener.
The sludge rakes were rotated at 16 rph resulting in a
rake tip speed of about 4 fpm.
There were three general types of spent carbon sludges pro-
duced. The first was once used virgin carbon with lime pre-
treatment. The second was once used virgin carbon with alum
or ferric chloride pretreatment. The third type was spent
regenerated carbon with alum or ferric chloride pretreatment.
130
-------
The once used virgin carbon had. 7 to 15 percent volatile
solids, 7 to 17 percent ash and 69 to 86 percent fixed carbon
The spent regenerated carbon had about 13 percent volatile
solids, 21 percent ash, and 66 percent fixed carbon. The
volatile solids were presumably composed of adsorbed organics
and biologically generated solids.
The ash material was a combination of original ash (about
3 percent), chemical solids from pretreatment and for re-
generated carbon some fine sand (less than 60 U.S. Mesh)
from the fluidized bed furnace.
Data from some 32 laboratory gravity settling tests indicated
that carbon sludge (slurry) density was 1.000 plus 0.00017
times g/£ of suspended solids. The term 0.00017 is about
one-half of what would be expected for pure activated carbon
solids with a particle density of 1.5 g/cc. This lower value
was obviously due to the presence of considerable alum or
ferric floe and biological solids.
Pilot Plant Results
As will be shown later in this Section, the pilot plant
thickener was grossly underloaded. Consequently the unit re-
quired little operator attention. Due to often times spora-
dic operation of the carbon regeneration system the thick-
ener was frequently used to store thickened carbon.
Five reasonably stable thickener operation and performance
periods were identified. The results for these periods are
shown in Table 7-10. Scrutiny of the carbon contactor oper-
ating data indicated only one identifiable effect on blow-
down solids concentration. Going from period 2 to 3 a
drastic reduction in blowdown solids concentration of from
44 to 11 g/£ was experienced. The cause had to be the sub-
stantial decrease in carbon dosage or the increase in dirunal
flow variation. Previous operating experience at moderate
to low carbon dosages (150 to 75 mg/£) and constant plant
flow did not result in low blowdown solids concentrations.1
It is concluded that the low blowdown carbon solids exper-
ienced during period 3 in Table 7-10 was due to the relativ-
ely high diurnal flow variation. There is no obvious reason
for the lower than expected blowdown carbon solids shown for
period 4.
Gravity thickener performance results shown in Table 7-10
indicate that consistently high blowdown carbon solids con-
centration was achieved for various operating conditions.
Variations shown for feed carbon solids, TSL or sludge age
(SRT) did not affect blowdown carbon solids concentration.
131
-------
TABLE 7-10:
SPENT CARBON SLUDGE THICKENING:
AVERAGE PILOT PLANT RESULTS
OJ
to
PERIOD #
Calendar Dates
Carbon Contactor:
Flow:
Average gpm/sq ft
Peak gpm/sq ft
Peak: Minimum
Pretreatment Chem.
Carbon Dosage, mg/£
Turbine Speed, rpma
Reaction Zone-SS, g/£
Blowdown-SS, g/£
Gravity Thickener:
Solids Loading,
Ib/day/sq ft
Sludge, Depth, ft
Solids 'capture, %
Blowdown-SS, g/£
Approximate SRT, Days
11/01/72
to
11/30/72
0.76
0.90
2:1
Lime
120
25
7.8
47
3.8
2-3
99+
155
5-8
4/14/73
to
4/30/73
0.65
0.75
2:1
Alum
294
33
8.6
44
7.8
2-3
99+
151
2-4
5/5/73
to
6/7/73
0.55
1.04
3:1
Alum
86
38
6.6
11
1.9
2-3
98+
154
10-15
6/10/73
to
8/03/73
0.52
0.82
2:1
FeCl3
110
30
8.4
18
2.4
2-3
97+
100-154
5-10
10/22/73
to
11/08/73
0.60
-0.90
2:1
Alum
70
25
11
58
1.7
1-2
99+
142
5-10
27 inch Diameter Turbine
-------
The very good solids capture achieved was due to the very low
hydraulic loadings of less than 20 gpd/sq ft. Thickener
overflow suspended solids ranged from 50 to about 300 mg/£.
The high carbon SRT's shown resulted in some anaerobic acti-
vity within the thickener. No significant odor problems
were experienced relative to thickener operation. When
thickened stored spent carbon was mixed in the 500 gal
inventory tank prior to sampling, gross "rotten egg" odors
were experienced. Full scale plant operation would not re-
quire such a mixing operation.
Thickener solids versus depth profiles are shown in Figures
7-11, 7-12, and 7-13. It is seen that from 1 to 2 ft of
depth was required to reach near maximum carbon solids concen-
trations. The profiles in Figure 7-13, for period 5 in
Table 7-10, indicate that maximum possible concentrations
may not have been achieved.
Laboratory Test Results
Table 7-11 presents results from 31 laboratory thickening
tests. These tests were all conducted on grab-composite
samples obtained when alum or FeCl3 coagulants were used in
pretreatment. Three tests were made during the period of the
study when regenerated carbon was reused.
Figure 7-14 shows a general trend of higher maximum solids
concentrations for increasing feed solids concentrations.
The data are, however, quite erratic at low carbon feed
solids values.
Figures 7-15, 7-16, and 7-17 show the interrelationship be-
tween predicted TSL and feed carbon solids for various under-
flow carbon solids values. It is obvious that the responses
shown are different than for the alum-primary or lime-prim-
ary sludges previously discussed. The data show that as
feed carbon solids concentrations decrease below 30 to 40
g/£, that higher predicted TSL are indicated.
Evaluation of the data shown on Figures 7-15, 7-16, and 7-17
indicated no significant effect due to alum versus ferric
pretreatment. However, it is obvious that regenerated car-
bon can be thickened at significantly higher TSL's to pro-
duce a given underflow carbon concentration.
Comparison of Pilot Plant and Laboratory Results
Because the pilot plant thickener was loaded at a low level
during nearly all operating periods, a direct comparison of
pilot plant performance with laboratory results is impossible,
133
-------
FIGURE 7-11: CARBON SLUDGE THICKENING
THICKENER SOLIDS PROFILE
5 -
!
~ i
j^"^ '
; i , : ;
: :
[
5
1 ! !
1 1
i : '
' 1
I
1
i
: ; ' 'j "
• J
• ' j
: 1
r
I ; '
.
ii
': '• I • ' ' ' Hi
' ' ;
i
1 .'"'•!,
, ' i !
! ; :
; ; : i i : •
) ' '
:\ :
V
\
1 1
-*~1 —
1 ;
1 '
far
! '
1^ ^^
Total Solids, percent by weight
134
-------
FIGURE 7-12:
CARBON SLUDGE THICKENING
PROFILE
- THICKENER SOLIDS
'Period #4
Table 7-10
0
5 10 15
Total Solids, percent by weight
135
-------
FIGURE 7-13:
CARBON SLUDGE THICKENING
THICKENER SOLIDS PROFILE
40
30 4
Period # 5
Table 7- 10
c r:5 4
10 15'''20
Total Solids, percent by weight
136
-------
TABLE 7-11: CARBON LABORATORY GRAVITY THICKENING TEST RESULTS
00
-J
Test
#
621
803
512
610
529
619
510
520
701
731
708
603
1025
703
522
531
628
1106
426
424
1010
422
410
429
415
Feed
SS,
g/£
6
7
8
8
10
13
14
14
18
19
20
20
21
23
23
23
29
28
30
35
36
39
43
44
48
.0
.3
.2
.5
.6
.2
.2
.3
.2
.5
.2
.2
.8
.1
.3
.9
.2
.8
.2
.3
.8
.4
.1
.4
Thickener Solids Loading
Ib/day/sq ft @ Underflow SS (g/&) of:
40
64
82
46
20
94
60
__
—
107
—
66
96
53
115
86
—
—
--
—
—
—
—
__
—
—
60
51
52
34
15
54
30
67
__
51
86
27
44
21
48
37
53
64
63
35
40
82
54
60
52
—
80
47
44
30
—
50
24
55
54
32+
50
18
30
15 +
33
26
29
31
39
17
19
40
24
25
19
38
100
44
—
—
. —
—
—
48
50
27
39
16
26
—
28
—
22
23
31
12
13
26
15
--
10.4
20
120
— .
--
—
—
—
—
45
—
25
—
—
—
—
__
—
19
—
27
—
—
21
—
9.1
8.1
15
140 Max.
43
40
30
13
41
20
44
50
24
35
16
24
14
25
22
17
19
25
11
10
20
12
8.0
7.5
13
Max. Critical
Underflow Point in
SS, g/£ inch/7 inch
120
104
82
76
96
106
135
95
130
122
101
119
88
122
103
149
127
135
113
123
133
124
133
126
139
2
1
2
1
1
1
2
2
1
1
1
2
2
2
3
1
1
2
1
2
2
2
1
1
2
.1
.4
.0
.5
.7
.5
.0
.1
.7
.4
.5
.1
.0
.2
.0
.5
.7
.9
.8
.4
.3
.2
.7
.4
.2
-------
TABLE 7-11 (CONT.); CARBON LABORATORY GRAVITY THICKENING TEST RESULTS
OJ
00
Test
#
419
1030
626
624
417
401
Feed
SS
g/i
51.8
57.6
64.6
66.9
• 76.9
94.5
Thickener Solids Loading,
Ib/day/sq ft @ Underflow SS (a/1) of:
80
40
94
29
94
--
__
100
20
45
14
24
70
—
120 140
15
26 20
10.4
13
26 16
38 19
Max.
13
17
9.3
11.3
11
12
Max .
Under-
flow
SS,g/l
130
161
143
126
179
182
Criti-
cal
Point
Inch/7 Inch
2.4
2.3
1.5
2.1
1.9
2.0
-------
U>
s
OS
o
Maximum Undertow Concentration,
CO O tO
O O O
rfk.
o
M
CJ
o
00
o
*J
(D
CD
a
c
m
•a
CD
P.
(D
a
oo ^-4-tT
a
CO
o
o
o«.
(DO
0
rt-
4
P
rt-
H-
O
-B8-
U3
f=
O-
o
-i-L-l 1(1'
-r-O-1
*^I
M
O
M
-d
I
M
*.
-4-
W }W
-^-;r-
-M I ! l-fh
-j-r ^4-i3
! ^" i" "t" •" , i 'rf-n
^Ff~
H-BH-
3-t4^-l-
.^LL.l.T±i±I±1Lr
4-l-:--r -^---[Ff
; j, , ,
^-Hg o
-^i
-wo
__^d ^
^nB CQ
-^S
M8S
l_T_-jg W
JMi-3
-ffl
i
O
teg
I
a
-------
FIGURE 7-15: CARBON SLUDGE THICKENING - LABORATORY PREDICTED TSL VERSUS FEED
50 SUSPENDED SOLIDS CONCENTRATION @ 120 g/i UNDERFLOW CONCENTRATION
10
20 30 40 50 60
Feed Suspended Solids Concentration, g/£
70
80
-------
FIGURE 7-16: CARBON SLUDGE THICKENING - LABORATORY PREDICTED TSL VERSUS FEED
SUSPENDED SOLIDS CONCENTRATION @ 140 g/J?. UNDERFLOW CONCENTRATION
0
20 30 40 50 60
Feed Suspended Solids Concentration, g/£
-------
£>•
to
50
FIGURE 7-17: CARBON SLUDGE THICKENING - LABORATORY PREDICTED TSL VERSUS FEED
JUSFENDED SOLIDS CONCENTRATION @ MAXIMUM UNDERFLOW CONCENTRATION
Feed Suspended Solids Concentration, g/H
-------
The highest pilot plant loading indicated in Table 7-10
was 7.8 Ib/day/sq ft. At this TSL, the laboratory results
in Figures 7-14 and 7-17 indicate that a maximum underflow
solids concentration of about 130 g/£ should have resulted
for the 44 g/£ feed carbon solids. The predicted 130 g/£
is significantly less than the 151 g/H experienced.
In general, pilot plant underflow carbon solids concen-
trations experienced were well in excess of laboratory pre-
dicted values.
Summary Discussion
Laboratory test results indicate that at a TSL of 10 lb/day/
sq ft, a maximum underflow carbon solids concentration
should be produced regardless of feed carbon concentration.
Pilot plant results indicate that this underflow carbon con-
centration could be as high as 150 g/£. A prototype gravity
thickener designed at a TSL of 10 Ib/day/sq ft with a 3-ft
deep thickened sludge layer would result in a carbon SRT
of less than 3 days. This relatively low SRT should not re-
sult in significant odor problems.
143
-------
VACUUM FILTRATION
Gravity thickened sludges were intermittently dewatered on
the 3-ft diameter by 3-ft face belt filter described in
Section IV. The general approach to operation of the filter
station, including routine sampling, is presented in Section
V. The primary objectives of the vacuum filtration dewater-
ing studies were to determine required chemical conditions,
the maximum filter yield and minimum cake moisture obtainable
for the sludges dewatered. A secondary objective was to
determine the correlation between pilot plant and laboratory
leaf test results.
Lime-Primary Sludge Filtration
Laboratory leaf test experience during the previous contract
eliminated the need for determining the type of media to be
used for this sludge. Polypropolene - 852F (POPR-852F)*
was used for all tests. A 0.10 sq ft laboratory circular
leaf was outfitted with new media at the same time a new
media was installed on the pilot plant. This leaf was used
for all pilot plant simulation studies.
During the first three vacuum filter runs erratic vacuum was
experienced. The vacuum filter bridge block was removed and
inspected. It was noted that the bridge block openings were
improperly located resulting in vacuum being applied before
a sector was completely submerged. A properly machined
bridge block was installed; this eliminated the problem of
fluctuating vacuum.
It should be noted that the drum submergence was 33 percent,
when the sludge level was at 1/4 inch above the lower ex-
tremity of overflow pipe. The cake form time (FT) was less
than 33 percent of the cycle time and was determined by the
location of bridge block openings and drum submergence.
Six pilot plant vacuum filter runs were conducted. Twelve
series of leaf tests were conducted. Six were for simula-
tion of pilot plant operating conditions and six were to
obtain background data on high feed solids concentrations
and the use of polyelectrolyte conditioner.
Pilot Plant Results
Table 7-12 presents pertinent operating and performance data
for lime-primary sludge vacuum filtration. The cake form
Media designations are those of Eimco BSP Division of
Envirotech Corporation, Salt Lake City, Utah.
144
-------
TABLE 7-12: LIME-PRIMARY SLUDGE: VACUUM FILTRATION PILOT PLANT RESULTS
M
Run #
Operating Conditions
Form Time (FT), mina
Dry Time (GD) , ™in
Vacuum, in. of Hg
Solids :
Feed, g/£
Filtrate, g/£
Filter Cake:
Thickness, in.
Discharge Code*-1
Dry Weight (W) ,
Ib/sq ft
6D/W, min sq ft/lb
Performance :
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
1110-1
.
1.46
2.13
17
111
0.38
3/32
F-P
0.216
9.9
8. 88
64
99.6
1110-2
2.01
2.97
16
111
0.34
3/32
F-P
0.232
12
6.93
67
99.6
1110-3
2.67
3.98
15
111
0.40
1/8
G
0.295
14
6.63
65
99.6
112-1
2.86
4.25
7-15
113
0.58
3/16
E
0.348
12
7.30
69
99.5
112-2
2.86
1.03
12-20
113
0.61
3/16
G
0.322
3.2
6.76
72
99.5
112-3
2.16
3.19
7-15
113
0.48
7/32
E
0.332
9.6
9.22
67
99.5
112-4
1.46
2.17
7-15
113
0.48
1/8
G
0.234
9.3
9.62
66
99.5
112-5
1.15
1.71
9-17
113
0.41
3/32
F-G
0.201
8.5
10.5
71
99.5
116-1
3.60
1.3
10-18
91
1.3
1/16
E
0.260
5.0
4.33
74
98.6
aForm Time/Cycle Time = 0.287 through Run #116, then =0.26
bSee Table 7-13 for Code
-------
TABLE 7-12 (CONT.): LIME-PRIMARY SLUDGE: VACUUM FILTRATION PILOT PLANT RESULTS
Run #
Operating Conditions:
Form Time (FT) , min
Dry Time(QD) , min
Vacuum, in. of Hg
Solids:
Feed, g/£
Filtrate, g/£
Filtrate Turbidity,
JTU
Filter Cake:
Thickness, in.
Discharge Code
Dry Weight (W) ,
Ib/sq ft
QD/W, rain Ib/sq ft
Performance :
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
116-2
2.08
3.1
7-13
91
0.82
1/16
G
0.196
16
5.65
74
99.1
118-1
2.91
4.40
14-
116
0.29
1/8
E
0.268
16
5.52
72
99 +
118-2
2.81
0.90
17+
116
0.23
1/8
G
0.281
3.2
6.00
74
99 +
118-3
1.73
2.63
15
116
0.32
7/64
F-
0.249
11
8.64
69
99 +
118-4
0.92
1.39
15-
116
0.30
3/32
G
0.204
6.8
13.30
71
99 +
222-1 222-2
2.72
4.08
11
62
120
1/8-
G
0.181
23
3.99
75
99 +
2.45
3.70
11+
71
42
3/16
G+
0.27
14
6.66
74
99 +
222-3 222-4
2.12
3.20
12-
71
42
3/16
G+
2 0.271
12
7.67
73
99 +
1.34
2.02
13+
72
21
1/8
G-
0.223
9.1
9.99
73
99 +
-------
TABLE 7-12: (CONT.): LIME-PRIMARY SLUDGE: VACUUM FILTRATION PILOT PLANT RESULTS
Run #
Operating Conditions :
Form Time (FT), min
Dry Time (60) / min
Vacuum, in. of Hg
Chem. Dose,a Ib/TDS
Solids:
Feed, g/S,
Filtrate Turbidity
JTU
Filter Cake:
Thickness, in.
Discharge Code
Dry Weight (W) ,
Ib/sq ft
GD/W, min. Ib/sq ft
Performance :
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
222-5
1.39
2.1
13
— _
72
15
5/32
--
0.290
7.2
12.5
73
99+
222-6
1.28
1.94
14-
_ »
72
15
1/8
—
0.246
7.9
11.5
73
99 +
222-7
.78
1.18
13
1.0
72
15
3/32
--
0.201
5.9
15.5
74
99+
222-8
.79
1.19
14+
1.8
60
23
1/8-
—
0.271
4.4
20.6
75
99 +
223-1
2.73
4.13
8+
0
44-
61
1/8
F
0.160
26
3.52
78
99+
223-2
1.39
2.10
11-
1.2
44
78
3/32+
F
0.159
13
6.86
78
99 +
223-3
1.27
1.93
14
2.8
44
47
1/8
G+
0.186
10
8.79
76
99+
223-4
1.08
1.63
16
4.9
44
77
1/8
G+
0.197
8.3
10.9
76
99+
aAnionic Polyelectrolyte
-------
times (FT's) shown were computed from the measured drum cycle
time (CT), minute/revolution, and drum submergences. FT, dry
time (eD), and CT relationships are shown as footnotes in
Table 7-12. Filter dry cake weight (W) and moisture content
(Me) were determined on cake samples taken with a 0.0845 sq
ft area circular "cookie cutter." Values reported are nor-
mally the average of two weight determinations with 2 to 4
"cookie cutter" samples per determination. Samples were
taken across the drum face at the mid-sector point and/or
along the drum circumference within a given sector.
The quality of filter cake discharge, shown in Table 7-12,
was noted by operating personnel according to the code pre-
sented in Table 7-13.
The vacuum filter solids handling capacity in this subsection
is reported as FFR and expressed as Ib of dry cake solids
(dried 105°C) per hour per sq ft. Full scale filter yield
(Y) is related to FFR by the following expression:
Y = (FFR) (FT/CT) (SUF)
where FT/CT represents effective drum submergences and SUF
is a suitable "scale-up factor" or possibly a "safety
factor." By reporting results as FFR the sometime variation
in effective submergence is factored out.
Inspection of Table 7-12 indicates that, in general, filter
cakes of 1/8 inch thickness or more exhibited "good" dis-
charge. There was a "doctor blade" mounted just below the
discharge roller, but its use was not judged to be a nec-
essity for effectively discharging cakes of 1/8 inch or more
thickness.
Figure 7-18 shows form filtration rate (FFR) as a function
of cake FT. Theoretically for a uniformed sized particle
suspension the slope of these lines would be -0.50. The
results for feed solids concentration of 111 to 116 g/£ were
found to have a -0.59 slope. The results for a feed solids
concentration of 72 g/£ and 1.0 to 1.4 Ib/ton of dry sludge
solids (TDS) of anionic polyelectrolyte are seen to have a
-0.85 slope.
The data in Figure 7-18 show a increase in FFR for increas-
ing feed solids concentration.
It is also apparent that a feed with 60-71 g/£ dewatered at
rates similar to those for a feed with 111-116 g/fc, if 1.0
to 1.4 Ib/TDS of anionic polyelectrolyte was used. The
polyelectrolyte used was AtlaSep 2A2.
143
-------
TABLE 7-13: ARBITRARY FILTER CAKE DISCHARGE CODE
Symbol
B+
B
B-
E+
E
E-
G+
G
G-
F+
F
F-
P+
P
P-
X
Designation
Beautiful
Beautiful
Beautiful
Excellent
Excellent
Excellent
Good
Good
Good
Fair
Fair
Fair
Poor
Poor
Poor
Wouldn't
Discharge
Number
Weight
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
149
-------
FIGURE 7-18:
30
20 -
+J
M-t
cr
IH
10
S 8
En
CL,
0)
•P
g 4
•H
JJ
fO
•H
Cn
e 2 ,
LIME PRIMARY SLUDGE FILTRATION -
PILOT PLANT RESULTS
-0.85 Slope
-0.59 Slope
-0.50 Slop
Symbol
Run #
Feed SS. g/£
O 111072 111
A 011273 113
O 011873 116
D 022273 72 with polyelectrolyte
0 011673 91
X 022273 62
0 3 0.4 0.6 0.8 l.'O 2TO
Form Time (FT), min
3.0 4 i0 6.0
150
-------
The filter cake Mc was also found to be affected by feed
solids concentration. Figure 7-19 shows MC as a function of
a "simplified correlating factor," 0D/W. At Q^/W greater
than 10 min.sq ft/lb, minimum Mc's were achieved. The Mc
data at this Qp/W value also indicates that a decrease of
about 2 percent Mc was experienced for an increase in feed
solids of about 40 g/£.
The 70-90 g/£ feed solids data shown in Figure 7-19 indicates
that the use of polyelectrolyte did not result in any sub-
stantial increase in filter cake Mc.
Leaf Test Results
Each time the pilot plant vacuum filter was operated, a leaf
test was conducted at near identical operating conditions of
FT, eD, and vacuum level of a grab-composite of pilot plant
feed. Table 7-14 presents the ratio of pilot plant to leaf
test results. It is readily apparent that a high degree of
correlation between the two existed. The average ratio of
pilot plant to leaf test FFR is seen to be 1.15. For those
runs without polyelectrolyte, the average and standard de-
viation was found to be 1.22 and 0.30, respectively. For
the tests with polyelectrolyte, the average and standard
deviation was 1.03 and 0.15. The more precisely correlated
results with polyelectrolyte may have occurred because these
runs were conducted on two consecutive days by the same
personnel„
The average and standard, deviations of the ratios of pilot
plant to leaf test cake Mc for all data and those runs with-
out polyelectrolyte were identical at 0.98 and 0,025. The
results with polyelectrolyte were more precise with an aver-
age of 0.99 and a standard deviation of 0.016. The correla-
tion of pilot plant and leaf test cake thickness was very
good at 0.99.
The high degree of correlation between pilot plant and leaf
test results indicated in Table 7-14 indicated that leaf
test results could be used confidently to predict pilot plant
performance.
As discussed in the lime-primary sludge thickening part of
this Section, a more concentrated pilot plant underflow SS
concentration was expected,. Evaluation of previous contract
work had indicated that attainment of thickened lime-primary
sludge of 180 g/£ should be obtainable.1 As seen in Table
7-12, the highest vacuum filter feed solids concentration
experienced was 116 g/&. To extend the range of results for
151
-------
FIGURE 7-19
01
80
tn
•rH
0)
76
XI
-P
CD
O
0) -
Cb '
U
•P
I 68
c
o
u
0)
a S4
"^
o
S
60
LIME PRIMARY SLUDGE FILTRATION, PILOT PLANT RESULTS - EFFECT OF FEED
SUSPENDED SOLIDS CONCENTRATION ON CAKE PERCENT MOISTURE
Feed SS = 44 g/£
with iPolyraer
Feed SS =1 70 - 90 g/£
(Awith Poajyraer. Awithout x
» —=s«=__ « - ..polymer)
A^T^Tts j —-&^-~l *-
Feed SS = 110 - 120 g/£ i
2 4 6 8 10 12
Simplified Correlating Factor (0D/W)f min sq ft/lb
14
16
-------
TABLE 7-14
LIME-PRIMARY SLUDGE FILTRATION:
RATIO OF PILOT PLANT TO LEAF TEST RESULTS
Form
Run # Filtration
Rate
1110-1
1110-2
1110-3
112-1
112-2
112-3
112-4
112-5
116-1
116-2
118-1
118-2
118-3
118-4
222-1
222-2P
222-3P
222-4P
222-5P
222-7P
222-8P
223-1
223-2P
223-3P
223-4P
Average
Std. Deviation
1.
1.
1.
1.
0.
1.
1.
1.
1.
1.
1.
0.
1.
0.
0.
0.
0.
1.
1.
0.
1.
1.
1.
0.
1.
1.
0.
27
40
62
11
87
19
05
13
41
77
18
98
68
89
91
88
87
00
18
94
31
01
14
90
08
15
26
Cake
Moisture
Content
0.
0.
0.
0.
1.
0.
0.
0.
0.
0.
1.
1.
0.
1.
1.
0.
0.
1.
0.
1.
1.
0.
0.
1.
0.
0.
0.
970
971
956
972
000
971
904
973
949
974
014
000
986
014
000
987
987
000
987
014
000
975
987
013
962
983
025
Cake
Thickness
1.
1.
1.
1.
1.
1.
1.
0.
0.
1.
1.
0.
1.
0.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
0.
0.
00
00
14
00
00
17
00
75
50
00
00
67
17
86
00
00
00
33
25
00
00
00
00
00
00
994
169
153
-------
the current study, leaf tests were conducted at higher feed
solids concentrations. In addition, the effect of polyelec-
trolyte on filter yields was evaluated by leaf tests.
Prior to vacuum filter run #222 (see Table 7-12) two series
of leaf tests were conducted using polyelectrolyte condi-
tioner. The result of these two tests are presented in
Table 7-15. From the results for test series #214 it is
seen that for a lime-primary sludge at about maximum concen-
tration, polyelectrolyte has a substantial affect on the FFR.
Extrapolation of data for test #214-3, indicates that, at a
FT of 0.4 min, a FFR of about 100 Ib/hr/sq ft could be ach-
ieved. Cake thickness would be in excess of 1/8 inch. For
a conventional belt filter this would mean a yield in excess
of 25 Ib/hr/sq ft.
The results for test series #221 presented in Table'7-15
show a less dramatic and limited effect of polyelectrolyte
on FFR. For test series #221, a feed solids concentration
of 118 g/£ was diluted to 88 g/£ to accommodate the addition
of various polyelectrolyte dosages. No consistent increase
in FFR was experienced for polyelectrolyte dosages above 1.1
Ib/TDS. In addition, the FFR at 1.1 Ib/TDS was less than
for the 118 g/£ feed solids concentration without polyelec-
trolyte addition. This result is similar to the results
for pilot plant runs #222 and #223 in Table 7-12. It
appears that the higher the feed solids concentration, the
larger the increase in filter yield for a given polyelectro-
lyte dosage.
It is germane to note that the leaf test results for test
#221, in Table 7-15, indicated no significant reduction in
filter cake MC with polyelectrolyte dosages up to 3.2 lb/
TDS. Test #214 shows an increase in Mc. The reduction in
feed solids concentration and OD/W values could be the reason
for this observation and not the use of polyelectrolyte.
Two series of leaf tests were conducted at feed solids con-
centrations of 232 and 189 g/£. The results of FFR versus
FT are presented in Figure 7-20. The ratio of 9D/FT during
these tests was 1.5. The average filter cake Mc for these
tests was 66 percent (62 to 70 percent range). Also shown
on Figure 7-20 is the curve for pilot plant results at feed
solids concentration of 111 to 116 g/&. The pronounced
effect of high feed solids concentration on FFR is obvious.
Summary Discussion
Lime-primary filter cakes of 1/8 inch thickness were
found to exhibit "good" discharge characteristics. Filter
154
-------
TABLE 7-15: LIME-PRIMARY SLUDGE: LEAF TEST RESULT USING POLYELECTROLYTE
Run #
214-1 214-2
214-3
221-1
221-2
221-3
221-4
221-5
221-6
221-7
Operating Conditions:
Form Time(FT),min
Dry Time(6D),min
Vacuum, in. of Hg
Chem. Dose,a
Ib/TDS
Solids:
Feed, g/Jl
Filtrate, g/£
i— •
un
Ul
Filter Cake:
Thickness, in.
Dry Weight (W) ,
Ib/sq ft
9D/W,min sq ft/lb
Performance :
1.
2.
20
0
230
~~ —
1/4
0.
3.
50 1.50
27 2.27
20
0.56
215
~~ ™~
5/16
708 0.829
2 2.7
1.50
2.27
20
1.24
194
— ~"
7/16
1.32
1.7
1.50
2.27
20
0
118
0.1
1/8+
0.278
8.2
1.50
2.27
20
0
88
0.1
3/32
0.181
12.5
1.50
2.27
20
1.1
88
0.1
1/8
0.239
9.5
1.50
2.27
20
2.3
88
0.1
1/8+
0.240
9.5
1.50
2.27
20
2.7
88
0.1
3/32
0.228
10.0
1.50
2.27
20
3.2
88
0.1
1/16
0.262
8.7
1.50
2.27
20
4.5
88
0.1
1/4
—
— _
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
28.
61
*~ *™
3 33.2
63
— _
52.8
66
__> —
11.1
71
99 +
7.24
73
99 +
9.56
72
99 +
9.60
73
99 +
9.12
73
99 +
10.5
74
99 +
--
--
99 +
1 Anionic Polyelectrolyte
-------
FIGURE 7-20: LIME PRIMARY SLUDGE FILTRATION, LEAF TEST RESULTS -
EFFECT OF HIGH FEED SUSPENDED SOLIDS CONCENTRATION
----1-t-f-
i'ii
(tails mean unacceptable discharge)
0.8 1.0
Form Time (FT), min
-------
cakes just under 1/8 inch were marginally discharged with
the use of a "doctor blade".
The average of all pilot plant and leaf test data for W and
thickness indicated a filter cake density of 23 Ib of dry
solids per cu ft of cake.
The effect of feed solids concentration on FFR over a broad
range had to be determined by leaf tests. Figure 7-21 pre-
sents all leaf test results. The solid curve shown indicates
the FFR expected for a filter cake thickness of about 1/8
inch and vacuums in the 11 to 17 inches of Hg range. This
curve represents the maximum yield obtainable. The dashed
curve is from the previous contract report.l Though cake
thickness and vacuum are slightly different, the previous
and current results are obviously similar. Of significance
is the dramatic effect of feed solids concentration above
120 g/£. The obvious implication of these results is that
maximization of thickening prior to the dewatering step is
very beneficial. The economic trade-offs between clarifier,
thickener, and vacuum filter operation will be left for
others to determine.
The significant effect of feed solids concentration on filter
cake Mc is seen in a composite of leaf test data shown on
Figure 7-22. At QD/W values above about 4 min.sq ft/lb,
about 10 percentage points less MC was experienced for feed
solids concentrations of 189 to 232 g/£ compared to 111 to
116 g/&. The economic impact of a 10 percentage point re-
duction of MC on a subsequent incineration operation would
be very substantial.
Alum-Primary Sludge Filtration
During the period of April to June, 1973, alum was used in
the chemical pretreatment step. During this period, nine
pilot plant vacuum filter runs were conducted on gravity
thickened sludge. Leaf tests were conducted to simulate
pilot plant performance and to determine the effects of
several operating variables„
Preliminary Leaf Tests
Preliminary leaf tests were conducted to determine a suit-
able filter media and chemical conditioning to achieve a
high degree of solids capture and good filter cake dis-
charge. Figure 7-23 presents results from some of these
tests. FFR's are reported on the basis of Ibs of dry alum-
primary solids per hour per sq ft. Scrutiny of these results
indicated that NY-317F provided the highest FFR's
157
.
-------
FIGURE 7-21; LIME PRIMARY SLUDGE FILTRATION, LEAF TEST RESULTS
- EFFECT OF FEED SUSPENDED SOLIDS CONCENTRATION
60
50
tr
w
40
4J
CO
Oi
C
O
•H
-P
2 20
-P
r-H
•H
fcj
E
M
O
10 -
T-UT M-fe
-L
I ' Reference #1
::a/16 in!Cake
: 26 in of Hg
•polymer j /
•1/8 in CakeY
11 - 17 in of Hg
„ Represents Maximum Yield for
^~| Dischargeable Cake
% ^without
\-polymer
50 100 150 200
Feed Suspended Solids Concentration, g/£
158
-------
FIGURE 7-22
LIME PRIMARY SLUDGE FILTRATION, LEAF TEST RESULTS - CAKE
MOISTURE CONTENT
78
M
*£>
tn
-H
Q)
X)
-p
c
0)
u
O
-P
C
0>
•p
C
o
u
Q)
M
3
-P
en
-H
o
a
-0
58
(Solid Symbols, with..Polymer_)L
A
Feed
189
200
232
113-121
72
88
200
O
A
D
0
468
Simplified Correlating Factor
10 12
/W) , rain sq ft/lb
14
16
-------
FIGURE 7-23: ALUM PRIMARY SLUDGE FILTRATION, LEAF TEST
RESULTS - EFFECT OF MEDIA AND CHEMICALS
Spleed^SS, g/£/Lime,J3ercent Joy \
0,8
0.3 0.4
0.6 008 1.0 2.0
Form Time (FT), min
4.0
0 S.'O 10
160
-------
with the exception of the results for 50 percent lime con-
ditioning. The NY-317F produced both excellent solids cap-
ture and very good discharge of thin (1/8 inch) filter cakes
as opposed to the other two media. Thus, a NY-317F belt
was obtained for the pilot plant vacuum filter.
The preliminary leaf tests indicated that at least 20 per-
cent lime conditioning chemical was required to produce a
dischargeable cake. Conditioning of a 44 g/& feed with 2 lb/
TDS anionic polyelectrolyte did not result in a dischargeable
cake. It was decided to use lime as the conditioning chem-
ical at dosages in excess of 20 percent by weight.
Pilot Plant Results
Table 7-16 presents results of alum-primary sludge vacuum
filtration tests. Runs #419 and #426 were conducted on
fairly septic alum-primary sludge. Sludges filtered during
Runs #531 and #604 were produced during a period when about
0.2 mg/£ of anionic polyelectrolyte was used in the chemical
contactor-clarifier unit.
Two problems were encountered during alum-primary sludge
filtration studies. The first resulted from the fairly fra-
gile nature of the vacuum filter nylon belt. The 3-ft dia-
meter by 3-ft face vacuum filter thermoplastic drum was sent
back to the supplier, just prior to this series of tests, to
be dismantled and rebonded. Elimination of vacuum leaks
which resulted in achieving less than full vacuum capacities
during the lime-primary sludge filtration tests was the pur-
pose. The vacuum filter drum was returned in a shabby con-
dition with numerous burrs and sharp edges. Operating
personnel removed most of the burrs and sharp edges, but a
few remaining ones caused tears in the media. Data in
Table 7-16 indicate that starting with Run #521, reduced
solids captures were experienced. By Run #607 most holes
had been patched and solids captures are seen to return to
about 99 percent.
The second operational problem resulted from inadequate vat
mixing conditions. Alum-primary sludges of 50 to 70 g/fi,
were much more viscous than were 200 g/& lime-primary
sludges. Lime conditioning of the alum-primary sludge pro-
duced large floe particles which rapidly settled in the
filter vat. On one occasion a 50 percent difference in
solids concentration was found for samples taken at a high
and low vat level. It was often felt that the flocculator
mixer had to be run at excessive speeds to maintain a uni-
form suspension in the flocculator overflow, where feed
samples were obtained. In an attempt to achieve a more
161
-------
TABLE 7-16: ALUM-PRIMARY SLUDGE: PILOT PLANT VACUUM FILTRATION RESULTS
CTl
to
Run # 419-1 419-2 419-3
Operating Conditions
Form Time (FT) a,min
Dry Time(eD) /min
Vacuum, in. of Hg
Lime Dose, % by wgt
Sludge, pH
Solids:
Feed, g/fc
Feed & Lime, g/£
Filtrate, g/5,
Filter Cake:
Thickness, in.
Discharge Code,*3
Dry Weight (W) ,
Ib/sq ft
«D/W,min sq ft/lb
Performance :
Form Filtration Rate
(FFR)c ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
.
2.7
4.09
18 +
40
10
73
84
0.32
3/16-
G
0.246
16.6
4.75
74
99 +
1.43
2.16
18 +
38
10
73
81
0.18
1/8 +
G
0.162
13.3
6.13
76
99 +
0.88
1.34
18 +
39
8
73
73
0.29
3/32
F
0.121
11.1
8.25
75
99+
419-4
2.70
4.09
18+
100
12
73
103
0.12
5/16
G
0.38
10.7
6.05
78
99 +
419-5
0.30
0.46
17 +
100
—
73
--
_ __
5/64
F
4 0.102
4.5
14
77
99 +
426-1 426-2 426-3 427-1
0.59
0.88
20
44
11
60
73
0.37
1/16
G
0.121
7.3
10.1
74
99 +
0.88
1.32
20
46
11
60
77
0.16
1/16-
F
0.0865
15
4.55
73
99 +
2.67
4.03
20
52
11
60
87
0.16
1/8
E
0.169
24
2.62
73
99 +
2.70
4.09
20
37
11
53
65
""
5/32-
G
0.196
21
3.55
76
—
427-2
1.38
2.08
20
37
11
53
65
""
5/32-
G
0.200
10
7.09
76
— —
aForm Time/Cycle Time = 0.26
k>See Table 7-13
GFFR's are reported on the basis of Lbs of Dry Alum-Primary solids and do not include
the lime conditioning solids.
-------
TABLE 7-16 (CONT.): ALUM-PRIMARY SLUDGE: PILOT PLANT VACUUM FILTRATION RESULTS
OJ
Run #
427-3
427-4
516-1
516-2
516-3
521
525-1
525-2
531-1
531-2
Operating Conditions:
Form Time (FT) ,amin
Dry Time(0D),min
Vacuum, in. of Hg
Lime Dose,% by wgt
Sludge, pH
Solids:
Feed, g/£
Feed & Lime, g/£
Filtrate, g/A
Filter Cake:
Thickness, in
Discharge Code
Dry Weight (W) ,
Ib/sq ft
eD/W,min sq ft/lb
Performance :
Form Filtration Ra
(FFR) Ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
0.83
1.26
20
37
11
53
65
—
1/8-
G
0.160
5,2
te
9.43
75
•= —
0.36
0.54
20
31
10.8
53
65
— -
1/16
G
0.107
5
14.5
76
— —
1.38
2.09
17
36
9.7
34
43
0.30
3/64
G
0.0817
26
2.81
76
99
2.70
4.09
17
36
10
34
43
0.49
3/64 +
G
0.118
35
2.07
75
99
1.11
1.67
17
10
9
34
38
0.27
1/32-
G
0.079J
21
3.86
75
99
1.4
2.1
19
47
—
21
28
6.1
1/32
?
2.76
4.17
18
55
11.8
44
62
0.17
1/4
E
3 0.0856 0.344
25
2.75
73
72
12
5.31
77
99+
0.96
1.45
20
38
11.7
44
62
0.09
5/32
E
0.186
7. 8
8.25
77
99 +
1.48
1.99
16
20
9.9
62
71
1.0
3/32 +
--
0.142
14
5.03
76 +
98 +
1.19
1.59
15
20
10.1
62
71
5.4
3/32
__
0.110
15
4. 84
76 +
92
Starting with Run #531, FT/CT =0.28
-------
TABLE 7-16 (CONT.): ALUM-PRIMARY SLUDGE: PILOT PLANT VACUUM FILTRATION RESULTS
Run #
531-3
531-4
531-5
604-1
604-2 604-3
607-1
607-2
607-3
607-4
Operating Conditions:
Form Time (FT) , mi n
Dry Time(9D),min
Vacuum, in. of Hg
Lime Dose,% by wgt
Sludge, pH
Solids:
Feed, g/£
Feed & Lime, g/H
Filtrate, g/£
Filter Cake:
Thickness, in.
Discharge Code
Dry Weight (W) ,
Ib/sq ft
QD/W,min sq ft/lb
Performance:
2.02
2.71
19
10-15
9.5
62
66
1.0
1/16
—
0.151
19
2.01
2.69
19
25-35
10.9
62
72
2.3
3/64
—
0.188
15
0.94
1.26
19
25-35
10.4
62
66
7.5
7/64
—
0.156
8.1
2.7
4.1
21
18
10.3
59 +
63+
12
1/32
—
0.0739
55
2.7
4.1
21
27
10.4
—
67
4.3
1/32-
--
0.0856
48
2.7
4.1
21
27
10.3
—
67
7.2
1/32-
—
0.106
39
3.02
4.03
21
35
10.8
49
58
0.25
1/8
G+
0.186
22
1.54
2.07
21
30
10.7
49
57
0.12
3/64
G
0.118
18
0.93
1.25
21
35
10.7
49
53
0.19
1/16
G-
0.0911
14
2.98
3.99
19
15
8.9
49
50
0.17
1/32
P-
0.0782
51
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
4.21
75
98+
4.83
76 +
97
9.35
76-
88
1.53
80
81
1.68
76
94
2.07
74
89
3.12
75
99 +
3.95
75
99 +
5.43
74
99 +
1.54
74
99 +
-------
TABLE 7-16 (CONT.): ALUM-PRIMARY SLUDGE: PILOT PLANT VACUUM FILTRATION RESULTS
CTl
Run #
Operating Conditions
Form Time (FT) , min
Dry Time(eD) , min
Vacuum , in. of Hg
Lime Dose, % by wgt
Sludge pH
Solids:
Feed, g/£
Feed & Limef g/£
Filtrate, g/£
Filter Cake:
Thickness, in.
Discharge Code
Dry Weight (W) ,
Ib/sq ft
@D/W, min sq ft/lb
Performance:
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
611-1
.
3.0
4.00
19
27
10.7
41
48
0.42
1/8-
G+
0.152
26
2,60
79
99
611-2
0.89
1,19
20
27
10.8
41
48
0,47
1/16
F-
0.086
14
4.95
79
99
611-3
1.52
2.03
18-
27
10.7
41
48
2.4
3/32
G+
0.119
17
4.01
78
99
-------
uniform filter vat suspension, the vat. overflow method was
changed after Run #525. Sludge was fed to the vat at a rate
in excess (about 25 percent) of what could be dewatered.
The excess sludge overflowed the front edge of the vat into
the belt wash trough and to drain. Thus sludge was fed at
the center back of the vat, flowed underneath the drum and
the excess flowed over the entire length of the front edge
of the vat. This method of operation improved vat mixing
somewhat.
Figure 7-24 shows FFR versus FT for several of the pilot
plant runs presented in Table 7-16. The tails on some data
points represent unacceptable cake discharge. The FFR's re-
ported in Table 7-16 and subsequent graphs are on the basis
of Ibs of dry alum-primary sludge solids and do not include
the lime solids.
Most of the data in Figure 7-24 are seen to fit the -0.5
slopes shown. The data is not extensive or consistent enough
to define precise trends in the effect of feed solids or per-
cent lime used. The data for the 60 g/£ feed solids of sep-
tic sludge is obviously quite low considering the approximate
50 percent lime was used. In general, higher feed solids
and/or higher lime dosages resulted in higher FFR's.
The Mc's in Table 7-16 range from 73 to 80 percent. These
values are reported as a percent of total filter cake solids,
including lime solids. There is no obvious trend in the data
of lower Mc for higher 9D/Ws. After presentation of leaf
test results, Mc will again be discussed.
Leaf Test Results
Each pilot plant test was simulated with a leaf test. The
ratio of pertinent pilot plant to leaf test performance par-
ameters are presented in Table 7-17. Scanning the results
and oberving the standard deviations indicates that the
ratios are quite variable. After the previously noted
change in filter vat overflow method, Runs #531 to #611, the
variability of the data was reduced slightly.
A probable cause of some of the variation was found to be
due to the time lag between sample procurement and conduct-
ing the leaf test. Simulation leaf tests for Run #419 were
conducted after 1 day and for Run #516 after 6 days time
lag. The erratic ratios of pilot plant to leaf test per-
formance for these and other runs prompted an evaluation of
of time lag on FFR. Figure 7-25 shows results for several
leaf tests. The effect of time between sample procurement
and conducting the leaf test is seen to be inconsistent,
166
-------
FIGURE 7-24: ALUM PRIMARY SLUDGE FILTRATION - PILOT PLANT RESULTS
20
at
-P
4-!
8
6 -_
Pi
fc,
o, 4 H
-p
o
•H
-P
fO
M
2 -
g
^t
O
Feed SS, g/£ /percent lime
0.1
0.2
JT\
A44/~47
^- -
(tails mean unacceptable discharge)
0.4 O.G 0.8 1.0
Form Time (FT), min
2.0
4.0 6.0
-------
TABLE 7-17:
Alum-Primary Sludge Filtration
Ratio of Pilot Plant to Leaf Test Results
Run
#
419-1
419-2
419-3
419-4
419-5
426-1
426-2
426-3
427-1
427-2
427-3
427-4
516-1
516-2
516-3
525-1
525-2
531-1
531-2
531-3
531-4
531-5
604-1
604-2
604-3
607-1
607-2
607-3
607-4
611-1
611-2
611-3
419-611
Average
Std. Dev
531-611
Average
Std. Dev
Form
Filtration
Rate
1.62
1.47
__
0.41
__
0,86
0.50
0.65
0,59
0.79
0-74
1.14
0.86
1.02
2.01
0.99
0.74
1.12
0.83
1.06
0.66
1.11
0.52
0.46
0.59
0.91
0.91
1.09
0.45
0,85
0.85
0.92
0.89
. 0.35
0.82
. 0.23
Filter Cake
Moisture
Content
1.04
1.04
__
0.99
__
0.99
0.99
1.01
1.00
1.01
1.01
—
1.00
1.01
0.99
1.01
0.99
1.00
1.00
0.96
1.01
1.00
1.10
1.06
_„
1.01
1.00
0.99
1.04
1.01
1.03
1.01
1.01
0.03
1.02
0.03
Filter Cake
Thickens s
2.0
2.0
1.5
0.36
1.5
0.67
0.50
1.0
0.62
0,83
1.0
1.0
l.'O
0-75
2.01
1.0
1.0
1.0
1.5
0.67
0.25
1.16
0.5
0.33
0.5
1.0
0.5
1.33
0.33
1.0
1.0
1.0
0.96
0.47
0.80
0.39
Filter Cakea
Discharge
1.0
1.5
6
0.7
~-
1.5
0.7
1.0
0.6
0.7
1.0
1.8
1.5
1.5
2.0
__
__
—
—
—
—
__
„_
__
— .
0.8
1.1
2.6
0.2
0.3
1.0
1.2
1.4
1.2
1.1
0.8
See Table 7-13 for discharge code
168
-------
FIGURE 7-25
0.5
0.4 -I
tr
0.3
+J
x:
Cn
•H
-------
but generally FFR decreases with time. Thus, many of the
high ratios of pilot plant FFR to leaf test FFR shown in Tab-
le 7-17 might be due to a time lag effect. After Run #516,
all simulation leaf tests were conducted on the same day as
the pilot plant run and generally within 1/3 to 2 hours after
sample procurement.
The only other reasonable explanation of the variable data
in Table 7-17 would be inconsistent mixing of the lime slurry
and sludge, and/or poor sampling. The latter is unlikely
and the former almost impossible to evaluate after the fact.
The questionable lime mixing effectiveness prompted a brief
study to compare pilot plant mixing with bench scale mixing
procedures. This comparative study was conducted during two
pilot plant vacuum filter runs as follows. During operation
of the pilot plant lime mixer-flocculator, five different
grab-composite sludge samples of flocculator feed (without
lime added) and discharge (with lime added and mixed) were
collected. Routine simulation leaf tests, at pilot plant
operating conditions, were conducted on the flocculator dis-
charge samples. About 1.5 liters of the flocculator feed
samples were subjected to laboratory scale lime conditioning
by adding the same dosage of lime as a 30 to 80 g per 100
to 200 m£ of tap water paste. Mixing was accomplished with
a large spatula. Simulation leaf tests were conducted on
these five laboratory lime conditioned samples.
Flocculator feed samples for this comparative study con-
tained from 41 to 49 g/£ of suspended solids. Lime dosages
were 27 to 30 percent by weight and form times ranged from
0.9 to 5.5 minutes. The ratio of dry time to form time used
was 1.5.
The average ratio for the five leaf test results on pilot
plant flocculator discharge samples to laboratory scale lime
conditioned flocculator feed samples were as follows:
Dry cake weight, W =0.94
Cake moisture content, Mc = 0.98
Cake thickness = 0.96
Cake discharge = 1.03
It is apparent that near identical vacuum dewatering results
were obtained when lime mixing-flocculation was accomplished
by the continuous flow pilot plant unit or by closely con-
trolled batch laboratory procedures. This observation would
tend to negate the previous suggestion that the variability
of results in Table 7-17 was due to inconsistent mixing of
the lime slurry and sludge in the pilot plant unit.
170
-------
The filter cake Mc comparison between pilot plant and leaf
test results was fairly precise and indicated near identical
results. Figure 7-26 shows Mc results for high and low feed
solids and lime dosage data as a function of the simplified
correlating factor. The data indicates that high feed solids
and lime dosage results in slightly lower MC compared to low
lime dosage. For low lime dosages, the effect of alum-primary
feed solids concentration of Mc is nominal. Figure 7-27
shows filter cake Mc versus eD/w for moderate lime dosages
of 30 to 44 percent. The data are seen to generally range
between 72 to 76 percent Mc and are not significantly affec-
ted by alum-primary feed solids concentrations.
Because pilot plant results did not show a definable trend in
the effect of feed solids concentration and percent lime con-
ditioning on FFR, several leaf tests were conducted in an
attempt to do so. Figure 7-28 shows the results of these
tests. Note that filter yield and not FFR is shown and that
the yield is for the operating conditions shown. The wavy
lines indicate where acceptable cake discharge was experienc-
ed. As previously noted, lime dosages greater than 20 per-
cent were required to produce dischargeable cakes.
The data in Figure 7-28 generally indicate that above some
lime dosage level, a very rapid increase in yield is experie-
nced for an incremental increase in lime dosage. For ex-
ample, the steep part of the curves show that for an incre-
mental increase of 10 percent lime (e.g., from 20 to 30 or
30 to 40 percent), a 2 to 3 fold increase in yield was ex-
perienced. This observation suggests that an economic op-
timum might exist. It would appear that facilities should
be provided for feeding 40 to 50 percent lime, but that
actual operation may require as little as 25 to 30 percent
lime on certain occasions.
Other Studies
Pilot plant and leaf test data were evaluated in an attempt
to identify a general relationship between the level of lime
addition and the increase in sludge solids. Pilot plant
data indicated an average of about 0.6 g increase in solids
for each 1.0 g of lime added (0.3 to 0.9 range). Data from
several laboratory scale tests indicated a 1.0 g increase in
solids for each g of lime added. Since the FFR's were re-
ported on the basis of dry alum-primary solids, the ratio of
feed solids before and after lime addition had to be used in
computing FFR's. The observed variability in suspended
solids increase with addition of lime may have imposed an
added variability in the FFR results, thus precluding de-
finition of the precise effects of feed solids and lime
171
-------
FIGURE 7-26:
ALUM PRIMARY SLUDGE FILTRATION - EFFECT OF FEED SUSPENDED SOLIDS AND
LIME ON CAKE MOISTURE CONTENT
88
-P
4^
tn
•H
0)
S
84
68
Pilot Plant ind ;Le^f :
10 15 20 25
Simplified Correlating Factor (
30 35
, min sq ft/lb
45
-------
FIGURE 7-27: ALUM PRIMARY SLUDGE FILTRATION - CAKE MOISTURE CONTENT
88 -,
i—
~j
u>
68
10 15 20 25 30
Simplified Correlating Factor (9 /W), min sq ft/lb
-------
FIGURE 7-28:
ALUM PRIMARY SLUDGE FILTRATION, LEAF TEST
RESULTS - EFFECT OF LIME DOSAGE
10
tr
w
\
M
5
13
rH
0)
•H
>H
Operating! Conditions
33% Subme^gance
i 20 in Hg
L EOETO- lime"—--l---5--»
1 1/8 inch Idischargealble cake
- AS-g/l
(with Pojlymer)"v
77 g/£nh41 g/X,
20 30
Percent Lime, by weight
174
-------
on performance parameters.
Note that the vacuum filter feed sludge pH's indicated in
Table 7-16 are in excess of 9. Alum sludge, resulting from
coagulation reactions, is an amphoteric material which is
dissolved at pH's in excess of 9. This condition is sub-
stantiated by data presented in Figure 7-29 which shows the
grams of soluble aluminum present in an alum-primary sludge
sample at various solids concentrations as a function of
slurry pH. Essentially all alum present was dissolved at a
pH of 11.7. At pH's of 10.7 to 11.0, the amount of aluminum
dissolved was inversely proportional to the solids concen-
tration. In practice, soluble aluminum would report to the
vacuum filter filtrate and recycled to the chemical treatment
unit. Presumably this recycled soluble aluminum could act as
an adjunct to the alum added. With lime conditioning, any
phosphorus released upon aluminum solubilization would be
precipitated as a calcium-phosphorus compound. Though a
phosphorus mass balance around the vacuum filter was not
made, it is presumed that most calcium-phosphorus precipitate
reported to the filter cake. It would be important to design
and .operate the vacuum filter to achieve high solids captures
to minimize recycling phosphorus material to the chemical
treatment unit.
Summary Discussion
Vaccum filter cakes of 1/8 .inch discharged easily from the
nylon media (NY-317F) with lime conditioning at dosages in
excess of about 20 percent by weight. Nylon media would be
recommended even though it is fragile and alignment diffi-
culties may occur on large machines.
All pilot plant and leaf test data were used to determine an
average filter cake dry solids density. A value of 21 Ib/cu
ft was determined.
A precise designation of lime dosage required to produce a
finite filter yield is impossible due to variations in feed
solids concentrations and characteristics (e.g., septicity,
percent alum sludge, etc.). For applications similar to this
study, lime, feeding facilities should be provided for 50 per-
cent by weight lime dosage.
In the authors' judgement, the data presented indicated that
for lime dosages of 30 to 40 percent by weight and feed solids
concentrations in excess of 40 g/&, that a full scale filter
175
-------
SS 6 001 / uiruiTumiY
176
-------
yield of 2 Ib of dry alum-primary solids/sq ft/hr could be
achieved. A filter cake Mc of about 75 percent, based on
total filter cake solids, would be expected.
Based on the filter vat mixing problems encountered in this
study, care should be exercised in providing good mixing
conditions. Lime conditioning of the alum-primary sludge
prior to gravity thickening might be worth considering. The
effect of lime addition prior to gravity thickening on TSL
and performance would have to be evaluated.
Ferric-Primary Sludge Filtration
In Section VI and earlier in this Section (Ferric-Primary
Sludge Thickening), it was suggested that the anaerobic con-
ditions existing in the chemical treatment unit resulted in
precipitation of FeS. The FeS precipitate did not adversely
effect the performance of the chemical treatment unit or
the gravity thickener, but its presence had a devastating
effect on vacuum filtration dewatering.
Of six pilot plant runs attempted, three were aborted due
to an inability to produce a dischargeable filter cake.
Several leaf test series were attempted in an effort to
determine suitable chemical conditioning to produce a dis-
chargeable cake.
Preliminary Leaf Tests
Preliminary leaf tests indicated a cake discharge problem.
Screening of several media, mostly nylon types, failed to
indicate any media with discharge properties superior to
those of NY-317F medium. NY-317F was used for alum-primary
sludge vacuum filtration and thus a new medium was not re-
quired.
Figure 7-30 shows a typical response of FFR versus percent
lime added. Lime in excess of 30 percent was required to
produce a marginally acceptable cake discharge. Because of
marginal leaf test performance results, problems with pilot
plant operation was expected.
A controlled bench scale test was conducted to show the pH
and solids increase responses due to lime addition to the
ferric-primary sludge. Figure 7-31 shows that the obvious
effect of lime addition is to increase pH arid uniform].y in-
crease feed solids concentration.
177
-------
FIGURE 7-30: FERRIC PRIMARY SLUDGE FILTRATION, LEAF
TEST RESULTS - EFFECT OF LIME DOSAGE
0.2
10 20 30 40
Percent Lime, by weight
50
Note; Tails mean poor Discharge,
178
-------
FIGURE 7-31: FERRIC PRIMARY SLUDGE - EFFECT OF
LIME DOSAGE
Lime Dosage, percent by weight
10 20 30
10
Ca(OH) Added,
179
-------
Pilot Plant Results
Table 7-18 presents the limited pilot plant data obtained.
It is readily apparent that low FFR's and thin cakes were
the rule. Though the data indicates that very thin cakes
were found to be dischargeable, it is not normal practice
to design full scale units to continuously discharge such
thin cakes.
The relatively low FFR's obtained were probably partly due
to the moderate to low levels of lime dosage used. A plot
of the data in Table 7-18 as FFR versus FT indicated very in-
consistent relationships.
About midway through these tests, the filter media belt was
cleaned with dilute acid. It. was hoped that previous use
with lime conditioning might have caused clogging or blind*-
ing of the media which could be eliminated by an acid wash.
No perceptable effect on performance was found.
Leaf Test Results
Because of the generally poor results and limited pilot plant
runs, limited leaf test simulations were conducted. These
tests were done only for pilot plant Runs #726 and #801.
For eight comparisons, the ratio of pilot plant to leaf test
FFR ranged from 0.97 to 1,70, averaging 1.40. The MC ratios
were much more consistent and averaged 0.97.
In addition to three aborted pilot plant runs, on four other
occasions leaf tests were made in an attempt to obtain a
dischargeable filter cake at lime dosages less than 50 per-
cent by weight. No success was experienced. It had become
apparent that the presence of finely divided precipitate
(FeS) precluded producing a dischargeable cake when using
only lime as the conditioning chemical.
Limited leaf tests were conducted in an attempt to identify
a type and dosage of polyelectrolyte which would flocculate
the troublesome precipitate. These results were negative.
The use of a primary coagulant, FeCl3, was finally tried as
an adjunct to the lime conditioning chemical. Acceptable
discharge and FFR's were obtained using the combination of
40 percent lime and 10 percent FeCl3.
Based on these results, the last pilot plant Run, #912, was
made using both chemicals. The results in Table 7-18 show
that acceptable performance was achieved.
180
-------
TABLE 7-18: FERRIC-PRIMARY SLUDGE: VACUUM FILTRATION PILOT PLANT RESULTS
Run # 726-1
Operating Conditions:
Form Time (FT), min
Dry Time (6D) , min
Vacuum, in. of Hg
Lime Dose, % by wgt
Sludge, pH
Solids:
Feed, g/£
Feed & Lime, g/£
Filtrate, g/£
Filter Cake:
Thickness, in.
Discharge Codeb
Dry Weight (W) ,
Ib/sq ft
SD/W, min sq ft/lb
Performance :
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture, %
Solids Capture, %
2.9
4.1
20
43
10.3
29.4
34.8
0.78
1/8
E
0.183
22
3.20
66
97
726-2
2
2
20
46
11
29
34
0
.1
.9
.2
.4
.8
.94
3/16
0
17
4
69
97
E
.175
.22
726-3
1.7
2.16
21
49
11.2
29.4
34.8
4.5
3/32
E
0.148
15
4.41
67
85
726-4
1
1
21
53
11
29
34
2
.27
.33
.3
.4
.8
.7
1/16
0
12
4
68
91
G
.115
.59
730-1
3.05
4.29
20
25
11.9
48.2
53.6
0.9
1/32 +
E
0.135
32
2.39
68+
98
801-1
3.00
4.2
20
--
11.9
49.3
49.4
2.2
3/32
E
0,165
25
2.74
69
95+
801-2
2
2
20
—
11
49
49
8
.10
.9
.9
.3
.4
.6
1/16 +
0
26
2
70
83
F
.112
.85
801-3
1.54
2.2
21
—
11.9
49.3
49.4
0.36
1/32
P
0.0809
27
2.61
70
99 +
FT/CT = 0.28
b
See Table 7-13
-------
TABLE 7-18 (CONT.): FERRIC-PRIMARY SLUDGE: VACUUM FILTRATION PILOT PLANT RESULTS
h-'
CO
Run # 801-4
Operating Conditions:
Form Time (FT), mina 1.19
Dry Time(9D) , min 1.7
Vacuum, in. of Hg 23
Lime Dose, % by wgt • — •
FeCls Dose, % by wgt --
Sludge, pH 11.9
Solids:
Feed, g/S, 49 = 3
Feed & Lime, g/£ 49.4
Filtrate, g/£ 3.0
Filter Cake:
Thickness, in.
Discharge Code^ X
Dry Weight (W) , 0.0557
Ib/sq ft
9D/W, min sq ft/lb 31
Performance :
Form Filtration Rate
(FFR) Ib/hr/sq ft 2.33
Cake Moisture, % 74
Solids Capture, % 94
808-1
3.01
4.23
21
28.1
33.3
2.3
1/32-
F
0.0425
0.72
70
820-1
2.72
3.81
20
30
64.3
64.8
0.62
nil
X
0.0302
12
0.5
73
99
912-1
3.30
4.63
22
28
12
8.4
57.3
68.8
0.17
1/8
E
0.193
24
2.92
72
99+
912-2
1.56
2.19
22
28
12
8.4
57.3
68.8
0.28
3/32
P
0.116
19
3.72
75
99+
912-3
0.9-7
1.36
22
28
12
8.4
57.3
68.8
3/32
E
0.166
8.2
8.55
73
aFT/CT =0.28
bSee Table 7-13
-------
Summary Discussion
The presence of a finely divided black precipitate in the
ferric-primary sludge precluded achieving effective filtra-
tion with lime conditioning. A combination of lime and FeCl3
conditioning was found to provide acceptable results, but
only at a predicted chemical cost of about $20/ton of dry
solids.
Spent Carbon Sludge Filtration
Vacuum filtration dewatering of thickened spent carbon was
done primarily to produce regeneration furnace feed. Pre-
vious experience had indicated solids capture was poor unless
cationic polyelectrolyte was used.1 It was also conjectured
that the filter media used was too open and that a tighter
media should be used. The media suggested was polypro-
polene-873 (POPR-873). This media was used during this
study.
Thirteen filter runs were conducted. Eleven of these runs
were in conjunction with carbon regeneration.
Five series of leaf tests were conducted simulating pilot
plant operating conditions. Other leaf tests were conducted
to obtain background information on certain operational var-
iables .
Pilot Plant Results
Pilot plant operating and performance data are presented in
Table 7-19. Data for most runs are averages for several
data collection series. For example, during Run #72872, the
vacuum filter was operated and data collected three times
with similar feed and operating conditions. The performance
results shown in Table 7-19 for Run #72872 are an average
of these three operating times,
As noted in Table 7-19, when the filter cake thickness was
less than 3/8. inch, less than "good" cake discharge was ex-
perienced. For filter cakes less than 1/2 inch thick, use
of the "doctor blade" was judged to be necessary.
Solids capture results indicate a gradual decrease from the
low nineties to about 80 percent over about a year.
Without exception, leaf test results always indicated solids
captures of greater than 99 percent. During pilot plant
operation, substantial deposits of carbon solids were
routinely observed at the drum headbolt-heads and at the
183
-------
TABLE 7-19: CARBON SLUDGE: VACUUM FILTRATION PILOT PLANT RESULTS
CO
Run #
72872
80172
81672
81672
91272
91272
92872
92872
122
872-1
122
872-2
Operating Conditions:
Form Time (FT), a min
Dry Time, (Go) , min
Vacuum, in. of Hg
# of Data Points
Solids :
Feed, g/£
Filtrate, g/£
Filter Cake:
Thickness, in
Discharge Code^
Dry Weight (W) ,
Ib/sq ft
eD/W,min sq ft/lb
Performance :
1.35
1,87
5
3
94
5.4
1/2-
G
—
34
0.64
0.89
6
2
173
8.4
1/2
G
--
65
0.67
0.90
6
6
104
10.5
5/16
G
__
47
0.67
0.90
6
1
147
~~ — '
3/8
G+
__
60
0,75
1.05
11
9
136
21
1/2
G
--
53
1.35
1.88
11
3
152
29
7/8
G
__
52
0.85
0.27
11
4
122
22
1/2 +
G
--
47
0.95
0.30
12
3
124
29
1/2 +
G
--
43
0.40
0.128
9-12
1
154
16
1/4
G
0.384
0.33
1.09
0.349
9-12
1
154
12
3/8
G
0.644
0.54
Form Filtration Rate
(FFR)lb/hr/sq ft
Cake Moisture, %
Solids Capture, %
28-
77
94
56
77-
95
40
77-
90
51
75
^
-------
TABLE 7-19 (CONT.): CARBON SLUDGE: VACUUM FILTRATION PILOT PLANT RESULTS
CO
Run #
Operating Conditions
Form Time (FT) ,amin
Dry Time(eD) ,min
Vacuum, in. of Hg
Chem. Doseb Ib/TDS
# of Data Points
Solids:
122
872-3
:
3.04
0.96
9-12
1
Feed, g/£. 154
Filtrate. g/H
Filter Cake:
Thickness, in.
Discharge Code
Dry Weight (W) ,
Ib/sq ft
QD/W, min sq ft/lb
Performance :
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture , %
Solids Capture, %
17
1
G
0.937
1.0
19-
—
89
010
373-1
0.50
0.16
12-15
__
1
160
31
1/4
F
0.347
0.46
41
78
81
010
373-2
0.86
0.275
11-16
--
1
161
23
3/8
G
0.459
0.60
32-
78
86
010
373-3
2.99
0.957
10-15
--
1
160
11
3/4
G+
0.840
1.1
17
78
93
40473
0.67
1.01
7+
--
4
160
30
1/4-
F
50573
1.4
2.1
10
--
3
134
41
7/32
F
61573
2.8
4.2
5
6.5
5
120
15
1/2
E
71
973-1
2.8
4.06
4
9.2
1
(120)
"™ "*"
3/4
__
71
973-2
2.8
4.06
11
8.4
1
(120)
*"" "*"
9/8
__
1.054 1.498
33
30
75
81
13
13
74+
70
19
15
76-
88
3.9
23
76 +
"°" ~™
2.7
32
77
•""""
b
Starting with Run #40473, FT/CT = 0.26
Cationic Polyelectrolyte
-------
TABLE 7-19 (CONT.): CARBON SLUDGE: VACUUM FILTRATION PILOT PLANT RESULTS
CO
Run #
Operating Conditions
Form Time (FT), min
Dry Time(eD) , min
Vacuum, in. of Hg
Chem. Dose, Ib/TDS
# pf Data Points
Solids :
Feed, g/£ (
Filtrate, g/£
Filter Cake:
Thickness, in.
Discharge Code
Dry Weight (W) ,
Ib/sq ft
9D/W, min sq ft/lb
Performance :
Form Filtration Rate
(FFR) Ib/hr/sq ft
Cake Moisture, %
71
973-3
.
1.4
2.03
15
6.4
1
120)
^ ™~
5/8
—
0.824
2.5
35
77
71
973-4
2.8
4.06
4
4.9
1
122
2.3
1/2
E
0.707
5.7
15
75
71
973-5
0.70
1.02
6
0
1
120
12
1/4
F
0.393
2.6
34
75
71
973-6
2.8
4.06
5
4.0
1
121
16
5/8+
G
0.475
8.5
10
75
91473
2.8
4.2
12
—
3
160
^ ^
1/2-3/8
—
—
11
12
--
100873
2.8
0.84
15
--
4
170
— —
3/4
—
1.3
12
14
73
Solids Capture, %
-------
belt-drum edge location. These depositions indicated sub-
stantial, leakage at these locations. Once a deposition was
built up, no leakage of solids would be expected. Thus, the
very poor captures experienced during some runs may have been
due to some other cause(s). The only consistent relationship
of solids capture and a machine operation variable found is
indicated in Figure 7-32. The data show that at very low FT's
(i.e., high drum rotational speed), very low solids captures
resulted. Had dry cake weight (W) been plotted instead of
FT, a more linear response would have been indicated. The
response shown in Figure 7-32 was not found to exist for all
results, thus it should not be interpreted as a general ob-
servation.
It is possible that the poor solids captures experienced was
due to the type of filter media used. Leakage of solids
through a monofilament can occur, particularly when no vacuum
is applied and the section of the filter drum covered with
clean media is entering the slurry. This type of phenomenon
is consistent with the results shown in Figure 7-32. At
lower form times or high drum rotational speeds, more oppor-
tunitites for "leakage" of solids would occur per unit time
or quantity of sludge processed.
Evaluation of all available Mc results indicated quite vari-
able data, but a general trend toward higher Mc of low 9D/W's
was evident. The data was "smoothed" by grouping QD/W values
and averaging MC values. The results are shown in Figure
7-33. The "smoothed" data shows a decreasing trend in MC
up to a GD/W value of about 5 minutes sq ft/lb. Also shown
in Figure 7-33 are results for a filter run using virgin
carbon. The spent (once used) carbon, with its associated
organics and chemical floe, dewatered to lower Mc's than
virgin carbon.
As indicated above, a pilot plant filter run (#102073) was
conducted using virgin carbon. The results of this run are
shown in Figure 7-34. During this run, a vacuum of 6 inches
of Hg was used and the ratio of Q^/FT was 1.5. The 1/4 inch
filter cakes produced at FT's less than 0.5 minutes were
judged to exhibit "good" discharge properties.
Scrutiny of all available pilot plant results (summarized
in Table 7-19) indicated that a wide variation in vacuum
levels were experienced. The main reason for the variation
in vacuum levels and the generally low levels experienced
was due to the nature of the dewatered carbon filter cake
and the type of machine used. The filter cake routinely
exhibited gross cracking due to shrinkage. Use of very
short 0D's, of 30 percent of FT or 8 percent of CT, reduced
187
-------
FIGURE 7-32: CARBON SLUDGE FILTRATION, PILOT PL/iNT RESULTS - EFFECT OF FORM
TIME ON SOLIDS CAPTURE
00
CO
100 1
-p
a
0)
u
M
0)
ft
-p
ft
Rj
u
to
T3
•H
rH
o
90 -
Run
Virgin Carb&n
1.0
Form Time (FT), rain
-------
FIGUFE 7^33:
OO
CARBON SLUDGE FILTRATION, PILOT PLANT RESULTS - CAKE MOISTURE
CONTENT
79
tn
•rH
0)
0)
u
M
OJ
77
-p
c
QJ
O
u
a)
•P
01
•-75
74
L i\ i i i[
ft i \' :
' ! :V
.-.;:::.: iiq-^
_,_^___ ._ _ _., ,_
i
____T |_p. ^ , ^
i i
' ! 1 ' ' ,
1 i
_. _J
: ' ' ; ; ! i _
----- -- -
A^
^^^~J
•HSHi
^
: --- -
—
--- --
-•-.
~r ""
/erage for
l^l»^ (T\i
."«q^-i
2 -
Simplified Correlat
Virgin Caj
— 1
i!*^(fc-fr\. ,.i
^^»^
4
ing Factor
A
0
rlpon
~^— ~_
5
(eD/w) , m
Lime ?ret]
Alum & F©3
i
i
_ 6
in SQ ft/1'
"eatment
•ric Pretr«
s^-——-------~
:atment
f . , ^
o
-------
FIGURE 7-34: CARBON SLUDGE FILTRATION - VIRGIN CARBON RESULTS
80
70
60 --
^50 4-
M
s:
40
30 4
En
trf'
Oi
m
o 20
-H
•P
•H
CM
S
M
O
Pu
10
3/16 in Cake Thickness
0
1/4
Pilot Plant
Feed SS = 107
78 % M
Simulation
Leaf Tests
Feed SS = 116 g/£
79% M
— 1 - — r- - -— — — r - ' - T— - T - — — "
0.3 0.4 0,6 0.8 1.0
Form Time (FT), rcin
2.0
3.0
4.0
6 0
-------
cake shrinkage and thus cracking. However, Mc was high (see
Table 7-19 and Figure 7-33). Due to the subsequent regener-
ation furnace operation, Mc should be minimized. On one
occasion an attempt to reduce cake cracking was made using a
"drag blanket." The "drag blanket" was a piece of gunny sack
(burlap) which was kept wet. No observable reduction in cake
cracking tendencies was observed.
The vacuum filter station used had only one vacuum receiver
for both the form and dry cycles. Consequently, when the
filter cake cracked during the dry cycle the vacuum level
was reduced, not only for drying but for cake forming. Ob-
viously separate form and dry vacuum receivers would have
been desirable.
Three series of leaf tests were conducted to determine the
effect of form vc.cuum level on FFR. The results of these
tests are shown in Figure 7-35. For a non-compressible
cake the slopes should have theoretically been 0.50. The
lower values indicated in Figure 7-35 indicate that the
spent carbon filter cakes were compressible. It would appear
that a conservative estimate of the slopes shown would be
0.3. Thus the relationship between FFR and vacuum level
(AP) would be:
FFR = K (AP)°-3 (D
For a given sludge, K remains constant and thus equation (1)
can be used to predict FFR at various levels of form vacuum,
if FFR is known at a given vacuum level.
No consistent relationship between feed solids and FFR is
apparent from data presented in Table 7-19. If however the
effect of varying form vacuum levels is factored out, a
relationship can be shown for results obtained when lime was
used in the chemical pretreatment step. Figure 7-36 shows
predicted FFR at several feed solids concentrations for a
form vacuum of 10 inches of Hg. The FFR values in Table
7-19 were adjusted using equation (1). The line shown in
Figure 7-36 indicates that, if 33 percent effective submer-
gence is assumed, an increase in filter yield of one lb/hr/
sq ft would be expected for each increase of approximately
one percent in feed solids concentration.
Also shown in Figure 7-36 is a. data point for virgin carbon
Run #102073 xyX . It is apparent that use of virgin carbon
in the pilot^plant resulted in no deterioration of its de-
watering rate, This observation is understandable when con-
sidering that the once spent carbon properties were very
similar to those of virgin carbon. For example, the virgin
191
-------
FIGURE 7-35:
CARBON SLUDGE FILTRATION, LEAF TEST RESULTS
EFFECT OF FORM VACUUM
5 7
Form Vacuum, inches of Hg
192
-------
U)
FIGURE 7-36: CARBON SLUDGE FILTRATION, PILOT PLANT RESULTS - EFFECT
OF FEED SUSPENDED SOLIDS CONCENTRATION
20
40 60 80 100 120 140
Feed Suspended Solids Concentration, g/£
160
-------
carbon was 3 percent volatile solids, 3 percent ash and 94
percent fixed carbon; whereas the once spent carbon was 7
percent volatile solids, 7 percent ash and 86 percent fixed
carbon.
Inspection of data in Table 7-19 indicates that after Run
#40473 a substantial reduction in filtration performance,
namely FFR, was experienced. All filter runs after this time
were conducted on carbon spent in the pilot plant when alum
or ferric chloride were used. During this period of time
the carbon was exhausted to higher levels than before. For
example, volatile solids in the spent carbon was more than
doubled. In addition, lower carbon dosages were being used
and higher levels of solids were entering the carbon con-
tactors. The combination of a higher volatile fraction and
higher ash (chemical floe) resulted in fixed carbon values
of less than 70 percent in the spent carbon.
Results for Runs #50573 and #91473 indicate that a FFR of
about 12 Ib/hr/sq ft was experienced for a vacuum level of
about 11 inches of Hg. Using Equation (1) to predict the
effect of vacuum level, a FFR of over 14 Ib/hr/sq ft could
be expected at a vacuum of 20 inches of Hg. Results of Run
#61573, when 6.5 Ib/TDS of a cationic polyelectrolyte was
used, indicate that a FFR of 15 Ib/hr/sq ft was experienced
at a vacuum level of 5 inches of Hg.
Extrapolation to 20 inches of Hg would result in predicted
FFR of about 23 Ib/hr/sq ft. These predictions were based
on limited data and thus a substantial factor of safety
should be used in any scale-up.
Run #71973 was conducted using cationic polyelectrolyte
(Dow C-31) conditioning chemical. The FFR versus FT response
for the data obtained during this run was not well corre-
lated. It is apparent that the use of polyelectrolyte at
dosages above 6 Ib/TDS was effective in increasing FFR.
Run #100873 shown in Table 7-19 was for filtration of regen-
erated and reused carbon. The FFR is seen to be 14 lb/hr/
sq ft for the 3/4 inch cake produced. A single leaf test
series on a similar spent (regenerated) carbon sludge indi-
cated a FFR of about 10 Ib/hr/sq ft for production of a 1/2
inch thick cake. Both the leaf test and Run #100873 indi-
cated Mc's Of 73 percent for 9D/W values in excess of 1 to
3 minutes=sq ft/lb
194
-------
Leaf Test Results
Some leaf test results have already been discussed. Table
7-20 presents results from 18 comparisons of pilot plant
and leaf test simulations. As found for other sludges, the
average pilot plant FFR was significantly higher than the
leaf test FFR and the Mc about one percentage point lower.
Summary Discussion
Due to the multiplicity and variance of treatment variables,
and in some cases limited data, it would be unwise to general
ize about the vacuum dewatering of spent carbon sludge. It
can be noted, however, that filter cakes of from 1/4 to 1/2
inch thick were found to be dischargeable. FFR was shown
to be linearly related to the feed solids concentration. In
addition, FFR was found to be a function of the vaccum level
used.
It was also found that FFR and/or Mc were dependent on the
type of pretreatment chemical and carbon used in the pilot
plant (e.g., virgin or regenerated).
Table 7-21 presents the authors" interpretation of the de-
watering results presented and discussed. Prototype opera-
tions would most likely be for the case where the most limit-
ed data was obtained (i.e., alum pretreatment and regenera-
ted carbon). Results for these conditions in Table 7-21
indicate that maximization of filter feed solids concentra-
tion should be strived for and that the use of a cationic
polyelectrolyte should be implemented.
195
-------
TABLE 7-20:
SPENT CARBON SLUDGE FILTRATION:
RATIO OF PILOT PLANT TO LEAF TEST
RESULTS
Run #
72872-1
80172-1
80172-2
81672-2
81672-3
81672-4
81672-5
10373-1
10373-2
10373-3
102073-la
102073-2
102073-3
102073-4
102073-5
102073-6
102073-7
102073-8
Form
Filtration
Rate
0.97
0.67
0.96
0.97
0. 97
1.02
1.21
1.17
1.10
1.13
0.94
1.30
1.38
1.27
1.22
1.62
1.82
1.69
Filter Cake
Moisture
Content
1.00
1.01
0.99
0.97
0.97
1.00
0.99
1.01
1.00
0.99
0.99
0.99
1.00
0.96
0.99
0.96
0.99
1.00
Cake
Thickness
1.00
0.80
1.00
1.00
1.20
1.00
1.25
0.50
1.50
4.00
0.82
1.00
1.40
1.25
2.00
1.33
1.00
1.33
Virgin Carbon Run
196
-------
TABLE 7-21: SUGGESTED VACUUM FILTER DESIGN
AND PERFORMANCE FOR PAC
Operating Conditions:
Feed SS
Vacuum
Effective Submergence
Cake Thickness
120 g/£
20 inches of Hg
33 percent
1/2 inch
Pretreatment
Chemical
None
Lime
Alum or FeCl-,
Alum or FeCl-,
Yield, Moisture ©D/W
Carbon Ib/hr Content, min sq ft/
Source sq ft Percent Ib
Virgina 17
Virgin 18
Virgin 7 . 6
Regenerated 5
78 >l-2
76- >5
76- >5
73 >3
Not Once Spent
197
-------
SECTION VIII
POWDERED CARBON REGENERATION
Demonstration of efficient and effective thermal regeneration
of carbon_using a fluidized bed furnace was a major purpose
of this study. There were two main objectives. The first
was to demonstrate efficient recovery of fixed carbon. The
second was to evaluate the effectiveness of the regeneration
process. Effectiveness was evaluated by comparison of carbon
characteristics before and after regeneration and by reuse
of regenerated carbon in the PAC-PCT pilot plant. It was
decided that regeneration effectiveness would not be evalua-
ted until an efficiency of at least 75 percent recovery of
fixed carbon was obtained. Because of this priority of ob-
jectives, before 75 percent recovery was achieved, the
carbon adsorptive characteristics were analyzed for by iodine
and molasses numbers only. After 75 percent recovery was
achieved, equilibrium adsorption isotherm tests and pilot
plant reuse studies were made.
The goal of 75 percent recovery was set because of two
factors. First, others had reported that 85 percent, or
higher, recovery was possible using a fluidized bed fur-
nace.2, 3 Secondly, it was shown in the previous contract
report that recoveries over 60 percent might make regenera-
tion economically feasible.1
During Phase I of this study the priority activity was evalu-
ation of fluidized bed furnace (FBF) operating variables to
minimize carbon loss. Because of the unsuccessful results
achieved, FBF development activity continued into Phase II.
Major changes to the fluidized bed furnace (FBF) established
the operating campaigns to be reported here. The first,, or
Campaign #1, included 3 runs using the furnace as it existed
at the end of the previous contract. The bed injection gas
(BIG) principal of operation and two different carbon feed
points were used. Campaign #2 consisted of 2 runs conducted
after the furnace was modified to include an expanding area
sand bed chamber, 6 additional gas distribution tuyeres, and
an additional higher carbon feed point. Campaign #3 consisted
198
-------
of 6 runs using the furnace as it existed during Campaign
#2, but with use of an off-gas recycle (OCR) principal of
operation.
Before the furnace could be started, major repair and re-
work was necessary. During the last run of the previous
contract, the burner chamber refractory was melted, thus a
new refractory lining was required. This same run resulted
in several melted or deformed tuyeres which had to be re-
placed. A larger burner was installed to allow increased
flow capacity- The long shut down between contracts, re-
sulted in considerable corrosion damage to most major
electrical instruments and considerable rework was required.
A crushed silica sand was obtained locally, and hand-
screened to 15 by 30 U.S. Mesh prior to installation. A
schematic of the FBF is shown on Figure 8-1.
Results of previous contract studies indicated an unaccept-
able recovery of carbon. A short period of regenerated
carbon reuse had indicated no substantial loss in effective-
ness for removing SCOD in the PAC-PCT pilot plant. Toward
the end of the previous contract work, the two feed points,
shown in Figure 8-1 had been installed, but time was not
available for determining their effectiveness. The reason
for considering the use of a higher feed point was to allow
more time for the bed injection gas to react with excess
oxygen entering the sand bed from the firebox.
Studies conducted by others had indicated that lower opera-
ting temperatures resulted in lower carbon losses.2 There-
fore, it was decided to operate at as low an operating temp-
erature as allowable with the BIG principal of operation.
Because of the need to auto-combust natural gas within the
fluidized sand bed, the minimum safe operating temperature
was about 1350°F.
CAMPAIGN #1
The first run was conducted using techniques and procedures
developed during previous contract work. For this run the
18-inch high feed point was used and the bed temperature
was set at 1470°F which was 60°F lower than previous work.
The bed fluidized at about 1.2 ft/sec, which was the design
capacity of the unit. During the run slight gas and air
flow adjustments were made to maintain the stack gas oxygen
at less than 0.5 percent by volume.
Table 8-1 presents the FBF operating conditions during the
three runs of Campaign #1. The pretreatment chemical refers
to the chemical treatment step ahead of the carbon contactors
199
-------
FIGURE 8-Is PLUIDIZED BED
FURNACE-CAMPAIGN NUMBER 1
Sight Glasa
Dewatered
Carbon Cake
Injection
T/C (Thermocouple)
One of Six
Natural Gas
Iniectlon
Nozzles
Natural Gas
Air
200
-------
TABLE 8-1: FBF OPERATING CONDITIONS: CAMPAIGN #1
Run Number
Run Date, (1972)
Pretreatment Chemical
Operation Principal
Feed Point (from Bed
Floor) , in.
Static Sand Bed Depth, in:
Before run
After run
Sand Size, US Mesh
Gas Flow, SCFM:
Burner Air
Burner Gas
Bed Injection Gas
Gas Velocity, ft/sec:
Bottom of Bed
Freeboard^
Run Length, hr
Feed Rate, dry Ib/hr
Moisture in Feed, % by wgt
Avg Temperature, °F:
Freeboard
Sand Bed
Firebox
Avg Pressure, in. of H20:
Freeboard
Bottom of Bed
Firebox
Avg 02 Content of Stack Gas
% by volume
1
August
16-17,
Lime
BIGa
18
31
22
15x30
140
3
10
1.2
1.2
22
65
77
1370
1470
1920
-4
28
42
0.2
2
September
12-13,
Lime
BIG
36
46
40
15x30
100
1.2
10
0.9
0.8
49
38
77
1380
1470
1980
-5
38
59
0.2
3
September
28-29,
Lime
BIG
36
40
36
•=
120
1.6
>10
1.0
0.9
36
35
78
1400
1490
2070
-3
29
64
-
a - bed injection gas
b - includes water vapor
201
-------
It is identified because 20 to 30 mg/£ of suspended solids
were present in the overflow from this unit and thus entered
the carbon contactors. This concentration of chemical floe
is significant when considering that 50 to 300 mg/£ carbon
was fed to the carbon contactors. Operation principal refers
to either bed injection gas, BIG, or off-gas recycle, OCR.
Feed point is the elevation above the bottom of the bed
that carbon entered the FBF. Sand bed depth was measured
directly before start-up and after shut down. During opera-
tion, it was measured indirectly by observation of reference
points of known height on the inside of the unit. These
were viewed through an overhead observation port. This
method was used to identify both the initial static and ex-
panded bed depth during operation. Because the freeboard
zone was filled with a red glowing carbon haze during
carbon feeding, this visual observation could only be made
prior to or after carbon feeding. Burner air flow, burner
gas flow, and bed injection gas flow were manually adjusted
and measured by orifice obstruction manometers. These
measurements were sometimes at or beyond the limits of the
existing devices. At these times, they were approximated.
The values reported in Table 8-1 are arithmetic average of
measurements taken about every 1 to 3 hours during a run.
Gas velocities were computed and adjusted to operating
temperature. The gas velocity in the freeboard includes
water vapor contributed by the feed. This was significant
at times. Run length is actual feed time or clock time,
minus any down time.
Percent moisture in the feed was obtained from solids analy-
sis of vacuum filter cake samples. Temperatures were
measured within the bed by thermocouples and continuously
printed on a multi-point recorder. Pressures were measured
by diaphragm type pressure meters.
Table 8-2 presents the carbon recovery and characteristics
results of the three runs during Campaign #1. Fixed carbon
recovery during Run #1 was only 49 percent. A large quan-
tity of ash appeared in the product, 35 percent by weight.
Almost half of this ash was identified as being sand. Sand
loss from the furnace during both prerun start-up and during
the run amounted to about 30 percent by volume. Some of
this sand loss could be accounted for by intermittent high
velocities attained prior to the run, the high attrition
rate of the locally obtained sand, and the large, almost
explosive, volume expansion caused by the water in the
carbon feed. The volume of flow attributed to the water
vapor was half of that of fluidizing gas streams in the
freeboard. Since the carbon was fed at a single point,
rapid vaporization of this large amount of water may have
202
-------
TABLE 8-2: CARBON CHARACTERISTICS : CAMPAIGN #1
tsJ
O
Run Number
Run Date
Fixed Carbon
Recovery %
Solids, % by wgt
Volatiles
Ash
Fixed Carbon
Iodine Number
Molasses Number
Acid Solubles,
% by wgt
1
August
16-17, 1972
Spent
5.1
9.9
85
776
16
—
49
Regenerated
1.4
34.6
64.1
832
41
20.1
2
September
12-13, 1972
62
Spent Regenerated
12.8 1.2
6.4 23.0
80.8 75.8
700 862
18 44
• — —
3
September
28-29, 1972
56
Spent
3.6
5.0
91.4
796
44
— ,
Regenerated
0.
15.
83.
1020
201
— —
7
8
5
See Table 8-8 for virgin carbon properties
-------
had a significant effect on the travel of the carbon; prob-
ably causing short circuiting of carbon to the surface of the
bed. If this phenomenon existed it could_have resulted in
sand being literally thrown out of the unit.
The computed feed rate of 65 Ib/hr during Run #1 was consis-
tent with previous experience.
To give even more time for the oxygen in the gas flow from
the firebox to be consumed by the bed injection gas, it was
decided to move the feed point up another 18 inches, to the
36-inch feed point during Run #2. This seemed to be the
only variable that could be changed to increase the recovery
from that of Run #1. Sand was added to the unit to a total
depth of 46 inches. This insured that the feed point would
be covered during operation, even if some sand were lost
during start-up. Because of the high sand losses during
Run #1, it was decided to drop the fluidization velocity
during Run #2 to 0.9 ft/sec.
The recovery of fixed carbon during Run #2 was better than
for Run #1, but still only 62 percent, as shown in Table
8-2. The sand losses were lower, 19 percent by volume, with
a corresponding reduction of ash in the product, 23 percent
by weight. Freeboard velocities of 1.0 ft/sec or less, if
reasonably uniform, should not have carried sand out of the
unit. The carbon feed rate dropped to 38 Ib/hr during Run
#2.
A substantial spent carbon inventory within the pilot plant
units necessitated that the third run be made. There was
no rationale for changing operational variables, therefore,
it was decided to conduct Run #3 using the same conditions
as for Run #2. A decision had been made just prior to Run
#3 to seek expert advice on the FBF design and operation in
hopes of establishing some meaningful direction to pursue
further obviously needed development work.
Recovery of fixed carbon during Run #3 was 55 percent. Sand
loss was low, 10 percent by volume, with a corresponding
smaller ash content in the product, 16 percent by weight.
This lower level of sand loss was possibly due to having
purged out fines during the previous runs.
It was decided at this point that because sand seemed to
be continuously carried over into the product carbon even
at low fluidization velocities, a simple hydrocyclone should
be installed on the scrubber underflow. This was done and
during subsequent runs it performed very well.
204
-------
During the three runs of Campaign #1, freeboard temperatures
were 90 to 110°F below those for the sand bed. Thus, it was
unlikely that burning of substantial amounts of either car-
bon or natural gas occurred above the bed. This would infer
that carbon was burning within the bed, and that an inade-
quate amount of oxygen was being consumed prior to the
carbon feed point.
Consultation with others indicated that inadequate mixing
was taking place in the FBF sand bed and that mixing could
be improved by increasing fluidi.zation velocity. Since
the burner equipment was sized for the existing furnace,
with an approximate maximum velocity of 1.2 ft/sec at the
bottom of the bed, and since high sand losses were being
experienced at these velocities it was decided that the
bed should be reshaped to that of an inverted truncated
cone. This shape would provide a small area at the bottom
of the bed with a resultant higher fluidization velocity
and increased mixing at a given volumetric flow rate. The
area at the freeboard would be the same as before and would
thus have the same tendency for sand loss. The decrease in
the bottom area allowed for an increase in velocity of
approximately 80 percent. The area change was from 7.08
sq ft to 3.95 sq ft. To get better ditribution of the
fluidizing gas it was recommended by others that six addi-
tional tuyeres be installed in the bed floor. This new
tuyere pattern is shown on the inset of Figure 8-2. The
outside 6 tuyeres contained 3 orifices each and the remain-
ing 7 tuyeres contained 6 orifices each. These orifices
were 0.272 inches in diameter. This allowed for good sand
retention with no gas flow and a small headloss at operating
gas flows. An operating headloss of .10 inches of water
was originally designed for.
Additional changes made to the furnace at this time were a
new feed point at the 4-ft bed depth level, 2 new thermo-
couples (one at the 4-ft level and one at the 5-ft level)
and a larger sight glass on the top of the furnace. The
final furnace configuration prior to Campaign #2 is shown
on Figure 8-2.
It was also recommended by others that a high silica sand
with a wide particle size distribution be used in place of
the uniform (15 by 30 U.S. Mesh) locally obtained sand.
Thus about 3-ft of recommended 16 by 80 U.S. Mesh high
silica sand was procured and installed prior to Run #4.
205
-------
FIGURE 8-2: FLUIDIZED BED
FURNACE-CAMPAIGN NUMBER 2
Sight Glass
Six New
\ Tuyeres
P/I (Pressure
Indicator)
One of Six
Natural Gas
Injection
Nozzles
Natural Gas
Air
206
-------
CAMPAIGN #2
After a delay of several months due to late deliveries of
materials and to other project priorities, two runs were
conducted using the BIG principal of operation and the modi-
fied FBF. The new higher carbon feed point (at 48 inches)
was used to allow maximum time for the excess oxygen to be
consumed. Combustion air and gas were adjusted to provide
a recommended 2.5 fps sand bed bottom velocity. The furnace
operating conditions and results for Runs #4 and #5 are pre-
sented in Tables 8-3 and 8-4, respectively. The fixed carbon
recovery results in Table 8-4 indicate the very poor recover-
ies experienced.
Prior to Run #4 about 34 inches of 16 by 80 U.S. Mesh high
silica sand was placed in the FBF. Visual observation of
the fluidized bed prior to and after carbon feeding during
this run indicated that the bed was fluidized to about the
4-ft level. After Run #4 only about 24 inches of sand was
left in the FBF. Most of the sand blown out of the FBF was
caught in the newly installed hydrocyclone separator. The
substantial increase in ash indicated in Table 8-4 suggests
that copious amounts of fine sand passed through the hydro-
cyclone separator.
Prior to Run #5 two additional feet of 16 by 80 U.S. Mesh
high silica sand was added to the FBF. At the lower fluid-
ization velocity indicated in Table 8-3, considerably less
sand was lost. Of the 19 percent ash found in the regenera-
tion product, some 8.4 percent was found to be acid soluble.
Thus, a nominal amount of sand found its way into the re-
generated carbon recovery system. This observation was
substantiated when upon passing the regenerated carbon
through a 60 U.S. Mesh screen prior to reuse in the pilot
plant, very little sand was collected.
Average freeboard and sand bed temperature in Table 8-3 in-
dicate that during Run #5 substantial heat was generated
in the freeboard region. Presumably either carbon or natural
gas and oxygen was present above the fluidized bed. This
observation was not found for Run #4. The higher fluidiza-
tion velocity and/or the shallower bed during Run #4 may
have been the reason.
Inspection of Table 8-4 shows that almost full recovery of
iodine and molasses numbers were achieved during Runs #4 and
#5. Figure 8-3 shows the regenerated carbon that was re-
covered had SCOD removal properties essentially equal to
virgin carbon.
207
-------
TABLE 8-3: FBF OPERATING CONDITIONS: CAMPAIGN #2
Run Number
Run Date, (1973)
Pretreatment Chemical
Operation Principal
Feed Point (from Bed Floor) , in.
Static Sand Bed Depth, in:
Before run
After run
Fluidized Bed Depth, in.
Sand size, US Mesh
Gas Flow, SCFM:
Burner Air
Burner Gas
Bed Injection Gas
Gas Velocity, ft/sec:
Bottom of Bed
Freeboard^
Run Length, hr
Feed Rate, dry Ib/hr
Moisture in Feed, % by wgt
Avg Temperature, °F:
Freeboard
Sand Bed
Firebox
Avg Pressure, in. of H~0:
Freeboard
Bottom of Bed
Firebox
4
April
4-5,
Lime
BIGa
48
34
24
48-60
16x80
140
2.5
>10
2.5
1.3
12.5
71
75
1190
1390
1960
-3
36
44
5
May
2-3,
Alum
BIG
48
48
42
-
16x80
120
-
8.7
1.9
•vl.O
19.2
58
74
1480 ± 100
1380 ± 100
1900
-4
_
53
Avg 02 Content of Stack Gas,
% by volume
0.2
0.1
a - bed injection gas
b - includes water vapor
208
-------
TABLE 8-4: CARBON CHARACTERISTICS : CAMPAIGN #2
Run Number
Run Date
Fixed Carbon
Recovery %
Solids, % by wgt
Volatiles
Ash
Fixed Carbon
Iodine Number
Molasses Number
Acid Solubles, % by wgt
Aluminum, % by wgt
4-
Spent
5,7
6.9
87.4
679
25
--
~" "*"
4
April
5, 1972
39
Regenerated
0.3
44.7
55.1
988
108
--
""
2-
Spent
6.6
7.3
86.1
588
5
--
« w
5
May
3, 1972
36
Regenerated
1.2
19
79.8
955
77
8.4
1.6
See Table 8-8 for virgin carbon properties
-------
FIGURE 8-3: COMPARATIVE EQUILIBRIUM ADSORPTION
ISOTHERM TESTS: REGENERATION RUN NUMBER 5
0.200
C4
o
PS
o
o
o
w
bO
U
0)
i-l
1
PS
o
«H
3
bO
O
__j ; ^ i _.„ ' i
•?ped=Pilo
Ch^mic 411Clarifpr
-JEffliWt
: August 28, 1973|
CO Virgirt
-------
MODIFICATION OF FBF OPERATING PRINCIPAL
After having exhausted all reasonable, alterations of physi-
cal and operation variables, including those recommended
by others, the contractor concluded that the BIG principal
of FBF operation would not work for the unit being used.
This conclusion was officially stated after Run #4. At
this time (late April, 1973) only about three months of
pilot plant operating funds were left in the contract.
Reuse of regenerated carbon had not yet been attempted!
Based on successful results achieved by others using a
similar FBF regeneration system (reference 2), but with a
fundamentally different method of oxygen control, the con-
tractor decided to modify the existing unit in an attempt
to, at least, demonstrate that high carbon recovery was
possible. It was recognized that little if any regenerated
carbon reuse could be achieved with the limited time avail-
able. Because of this fact negotiations were undertaken
with the EPA to explore the possibility of an extension
to the contract. A 3 month extension was obtained with the
major purpose being to demonstrate high carbon recovery
and to evaluate the effectiveness of regeneration by reuse
of carbon in the pilot plant.
The relatively minor modifications in the FBF unit required
to allow operation using the off-gas recycle, OGR, princi-
pal can be seen by comparing Figure 8-4 with 8-2. Sections
IV and V contain additional information about the modifi-
cation and the approach to operation with the OGR principal.
CAMPAIGN #3
This Campaign spanned a 4-1/2 month period from June to
November, 1973, which included the 3 month contract exten-
sion. During this period six regeneration runs were made,
one using virgin carbon feed. Tables 8-5 and 8-6 present
FBF operating conditions and carbon characteristics for the
five runs using spent carbon. The only major operational
variations were sand bed temperature and fluidization vel-
ocity. Operating conditions were set immediately prior to
any given run and were based on feedback from previous runs,
The following is a brief discussion of each run, indicating
the rationale used to set operating conditions, fixed car-
bon recovery, carbon characteristics and any anomalies or
problems encountered.
Run #6 was the first using the OGR operating principal.
Relatively low fluidization velocity and sand bed tempera-
ture were used in hopes of maximizing fixed carbon recovery
211
-------
FIGURE 8-4: FLUIDIZED BED
FURNACE-CAMPAIGN NUMBER 3
Sight Class
P/I (Pressure
Indicator)
Natural Gas
Air
212
-------
TABLE 8-5: FBF OPERATING CONDITIONS: CAMPAIGN #3
Run Number
Run Date, (1973)
Pretreatment Chemical
Operation Principal
Feed Point (from Bed
Floor) , in.
Sand Bed Depth, in:
Static
Fluidized
Sand Size, US Mesh
Gas Flow, SCFM:
Burner Air
Burner Gas
Off Gas Recycle
Gas Velocity, ft/sec:
Bottom of Bed
Freeboardb
Run Length, hr
Feed Rate, dry Ib/hr
Moisture in Feed, %
by wgt
Avg Temperature, °F:
Freeboard
Sand Bed
Firebox
Avg Pressure, in. of
H20:
Freeboard
Firebox
Avg 02 Content of Stack
Gas, % by volume
6
June
17-18,
Alum
OGRa
18
32
-
16x30
80
8
70
2.0
0.8
30.7
21
76
1130
1250
1900
-4
57
0.2
7
July
18-20,
FeCl3
OGR
18
32
44+
16x30
90
9
60
2.2
1.0
28.3
27
76
1330
1400
1970
-3
64
0.3
8
Sept
11-13,
FeCl3
OGR
18
36
48+
16x30
85
9.2
75
2.2
l.Oc
21.5
37
76a
1170
1250
1960
-3
72
0.3
9
Oct
8-10,
FeCl3
OGR
18
-
—
16x30
90
10
80
2.7
1.1
51.7
16
73
1470
1550
2000
-1
70
0.0
11
Nov
2-3,
Alum
OGR
18
-
-
16x30
93
10
90
2.9
1.1
46.0
13
72
1470
1550
2000
0
68
0.1
a - off gas recycle
b - includes water vapor
c - estimated
213
-------
TABLE 8-6: CARBON CHARACTERISTICS : CAMPAIGN #3
Run Number
Run Date
Fixed Carbon
Recovery %
Solids,
% by wgt
Vol atiles
Ash
Fixed
Carbon
Iodine
Number
Molasses
Number
Acid Solubles
% by wgt
Iron,% by wgt
Aluminum, %
by wgt
6
7 8
9
June July September
17-18, 1973 18-20, 1973 11-13, 1973
100
Spent
8.7
9.2
82.2
506
15
—
__
_~
Regen-
erated
1.5
16.5
82.0
788
59
9.7
__
__
78 9
Spent
12.7
15.5
71.8
311
10
__
--
__
Regen-
erated Spent
3,0 16.0
17.7 19.1
79.4 69.9
684 158
69 7
10.2
3.2
. — —
3
Regen-
erated
4.5
26.0
69.5
527
28
19.0
12.4
— —
October
8-10, 1973
100
11
November
2-3, 1973
76
Regen-
Spent
13.4
18.0
68.6
266
6
—
— —
—
erated
1
32
65
723
56
16
7
0
.8
.7
.5
.1
.0
.4
Spent
13.1
23.3
63.6
274
8
— ,
—
— —
Regen-
erated
2.8
26.0
71.2
662
58
15.8
5.7
18.8
See Table 8-8 for virgin carbon properties
-------
Results in Table 8-6 indicate that 100 percent recovery of
fixed carbon was achieved.
Operation of the FBF was very stable during Run #6. The
increase in ash which was not acid-soluble ash in the re-
generation product compared to the spent carbon (Table 8-6)
indicated that some sand was probably still being lost from
the FBF unit and not being captured by the hydrocyclone.
Comparison of regenerated carbon iodine and molasses numbers
with previous runs indicated less than full recovery of
these properties. A comparative equilibrium adsorption iso-
therm with virgin carbon (shown on Figure 8-5) also indica-
ted a significant loss of SCOD adsorptive capacity. It
was assumed that the low sand bed temperature of 1250°F was
the cause of not recovering full adsorptive capacity, Thus,
it was decided to increase the sand bed temperature during
the next Run, #7.
Table 8-6 indicates that only 78 percent recovery of fixed
carbon was achieved during Run #7. Comparison of Run #7
and #6 operating conditions indicate only two significant
differences. First, the sand bed temperature was increased
from 1250 to 1400°F. Secondly, the carbon contacting
system feed pretreatment chemical was different, being alum
for Run #6 and FeCI^ for Run #7. It was not considered
likely that a 150°F increase in sand bed temperature would
have caused such a drastic reduction in fixed carbon re-
covery. Thus, it was conjectured that the presence of
ferric flee and/or adsorbed iron (see Section VI, Operating
Period IV) might have catalyzed the burning of fixed car-
bon. To eliminate the possible temperature effect the
next Run, #8, would be run at the lower temperature of
1250°F.
Analysis of spent carbon characteristics in Table 8-6 and
comparison with previous regeneration runs, indicated that
carbon was finally being exhausted to a high degree. Vola-
tile solids, iodine and molasses numbers all indicated a
relatively high degree of carbon exhaustion. Regenerated
carbon characteristics showed a. substantial recovery of
adsorptive properties, but not full recovery. A comparative
equilibrium adsorption isotherm with virgin carbon (see
Figure 8-6) showed a significant loss of SCOD adsorptive
capacity at equilibrium SCOD values of 10 to 15 mg/£.
During Run #7 little sand was lost from the FBF and only a
nominal amount appeared in the recovered product.
215
-------
FIGURE 8-5; COMPARATIVE EQUILIBRIUM ADSORPTION
ISOTHERM TESTSj REGENERATION RUN NUMBER 6
U
Q
O
O
CO
O
e
bfl
h
O
i Cnemjic
Eff3|ueht| I i !
August 21, 1973
Regenerated Cprbon
4 6 8 10 20
Equilibrium SCOD (Ce), mg/1
216
-------
FIGURE 8-6: COMPARATIVE EQUILIBRIUM ADSORPTION
ISOTHERM TESTS: REGENERATION RUN NUMBER 7
a:
e
Q
O
O
W
bo
C
-------
Though the regenerated carbon adsorptive properties were
less than 100 percent recovered, it was still decided to
reuse the product in the PAC-PCT pilot plant. Thus, in mid-
August regenerated carbon from Runs #5, #6 and #7 were mixed
together and used for carbon treatment system feed.
The start of Run #8 was planned for August 16. However, on
August 15, with the FBF at operating temperature, an ab-
normally high pressure drop of 150 inches of water developed
across the tuyeres and sand bed. The combustion air blower
flow was reduced to only about 50 SCFM. It was feared that
either the tuyeres had sustained major damage or that the
sand bed had been "cemented". The latter suspicion stemmed
from the observation that after Run #7, which was the first
using a ferric chloride pretreatment spent carbon, the sand
bed grains were uniformly coated with a dark reddish-brown
precipitate. This coating could have been either iron oxide
or iron phosphate. It was felt that had momentary general
or localized sand temperature exceeded about 1600°F, that
iron phosphate fusion might have occurred and resulted in
cementation of sand grains around or in the orifices of
the tuyeres.
Because of the above mentioned high pressure drop problem,
the FBF had to be cooled down to allow removal of sand and
inspection of the tuyeres. No obvious signs of tuyere
damage, orifice clogging or sand grain fusion were found!
The sand bed was put back into the FBF, with about 4 inches
of new 16 by 30 U.S. Mesh high silica sand, and the unit
started-up again. Pressure drops returned to normal, how-
ever, full combustion air flow was still not obtained!
The air blower was dismantled and considerable packing
material between the multiple stages found missing. The
blower was repaired, the FBF brought back up to operating
temperature and Run #8 started - after about a 4 week loss
of time.
Recall that Run #8 was to be conducted at the reduced bed
temperature of 1250°F to observe if the poor fixed carbon
recovery during Run #7 might be due to the higher bed temp-
erature (1400°F) or possibly due to the presence of sub-
stantial amounts of iron in the spent carbon. As seen in
Table 8-6, a good recovery of 93 percent fixed carbon was
achieved. It was, therefore, concluded that the presence
of iron compounds in the spent carbon did not drastically
effect fixed carbon recovery, at least at a 1250°F opera-
ting temperature.
218
-------
Iodine and molasses numbers for Run #8 indicated that the
spent carbon was exhausted to a high degree. They also
indicated that, though a substantial recovery was experienc-
ed, the regenerated carbon values were very low. A com-
parative equilibrium adsorption isotherm test again indicated
substantial loss of SCOD adsorptive capacity (see Figure
8-7) .
Since regenerated carbon was being reused in the pilot plant,
and poor SCOD adsorptive properties could be detrimental
to performance, it was decided to shift priorities from
seeking to maximize fixed carbon recovery to maximizing re-
covery of adsorptive properties. Thus, a higher sand bed
temperature was scheduled for Run #9.
In addition to a higher operating temperature, the fluid-
ization velocity was increased. Table 8-6 shows that a
fixed carbon recovery of 100 percent was achieved. Re-
covery of adsorptive properties was judged to be very good,
considering that the spent carbon feed was once regenerated
and reused carbon. Figure 8-8 shows SCOD removal results.
The high recovery of fixed carbon for Run #9 seemed to
negate previously expressed concern about the presence of
iron compounds in the carbon and that higher operating
temperatures might be causing increased carbon losses.
Apparently other operational or process variations were, at
least in part, the cause of inconsistent recoveries of
fixed carbon and possibly adsorptive properties.
Run #11 was conducted during the final week of pilot plant
operation. The spent carbon feed was a mixture of once and
twice regenerated and reused carbon with alum pretreatment
chemical. Table 8-6 shows that only 76 percent recovery
of fixed carbon was achieved. Some carbon was burnt during
the run because of a malfunctioning main gas line control
valve. On at least two occasions, this valve closed and
the combustion air blower remained on. This resulted in
air, with high oxygen content, going directly to the sand
bed. Carbon feeding continued at a decreasing rate for
several minutes until the bed temperature reached a level
which shut the feed pump off. The amount of carbon burned
due to this problem was undeterminable, but probably less
than 2 to 4 percent of the total fixed carbon fed during
the run.
Observation of carbon characteristics in Table 8-6 and the
equilibrium adsorption isotherm test results on Figure 8-9
indicates that substantial recovery of adsorptive properties
was achieved. Especially when considering that a mixture
219
-------
0)
fa
0>
fti
Q
O
O
02
bJD
Carb
a)
x
•H
fa
tn
rH
cd
>
O
0)
FIGURE 8-7: COMPARATIVE EQUILIBRIUM ADSORPTION
ISOTHERM TESTS: REGENERATION RUN NUMBER 8
0.400
0.200-
Feed=*Pilo
Chertical! Clarifjier
Effluent
September 26, 1973
Regenerated
! Caibdn !
0.004
60 80
Equilibrium SCOD (C ), mg/1
220
-------
FIGURE 8-8: COMPARATIVE EQUILIBRIUM ADSORPTION
ISOTHERM TESTS: REGENERATION RUN NUMBER 9
0.400
H
Si
O
0>
&
Q
O
O
CO
bO
-------
I
0)
bfl
k
O
0.35
FIGURE 8»9l COMPARATIVE EQUILIBRIUM ADSORPTION
ISOTHERM TESTS; REGENERATION RUN NUMBER 11
0.20
November 14, 1973
i i i i ' i T I
T—-—"-1 "—1-
4 8 8 10
Bqullibrium SCOD (C
222
-------
of once and twice regenerated carbon had been spent in the
pilot plant carbon contacting system.
A special regeneration run was conducted October 20, 1973,
using vacuum filtered virgin carbon as the regeneration
system feed. The FBF operating conditions and results for
this Run #10, are presented in Tables 8-7 and 8-8. FBF opera-
ting conditions are seen to be similar to those for Runs #9
and #11 except that the sand bed temperature was 100°F lower,
1450°F. Table 8-8 shows excellent fixed carbon recovery of
98 percent. No significant change in adsorptive properties
were found for molasses number and SCOD adsorption tests.
Figure 8-10. It is interesting to note that an increase in
iodine number from 942 to 1040 was experienced. It is
normally expected that thermal regeneration of carbon re-
sults in a reduced iodine number due to collapsing or block-
age of micropores.
DISCUSSION OF RESULTS
Evaluation of the results of this study indicate wide varia-
tions in the effectiveness of the regeneration system.
Variability of the degree of carbon exhaustion, pretreatment
chemical used, FBF operating conditions (fluidization vel-
ocities, temperature, sand bed depth, etc.) affected per-
formance. There were, however, certain identifiable trends
in some data which will be discussed here.
Figure 8-11 shows a relatively precise inverse relationship
between volatile solids and iodine number for both spent and
regenerated carbon. Only one data point departed signifi-
cantly from this precise trend; for Run #2 an abnormally
high volatile solids was measured and this is not shown
on Figure 8-11.
Such a precise relationship between volatile solids and
iodine number for spent carbon was not expected. The vola-
tile solids determination used measured both adsorbed organ-
ics and any biomass present in the spent carbon. The vari-
able nature of carbon contacting system operation (e.g.,
number of stages, dosage and SRT) and the length of time of
storage of gravity thickened carbon (e.g., 1 to about 20
days) would have been expected to effect the amount of bio-
mass present, and thus prevented determination of such a
precise relationship as shown on Figure 8-11.
There is a second definable trend in the data shown on
Figure 8-11. It appears that the more exhausted the carbon
(i.e., high volatile solids and/or low iodine number), the
more difficult high recovery of adsorptive properties was
223
-------
TABLE 8-7: FBF OPERATING CONDITIONS: VIRGIN CARBON
Run Number
10
October
Run Date, (1973)
Operation Principal
Feed Point (from Bed Floor, in.
Sand Size, US Mesh
20-21
18
15x30
Gas Flow, SCFM:
Burner Air
Burner Gas
Off Gas Recycle
Gas Velocity, ft/sec;
Bottom of Bed
Freeboard^
80
11
90
2.6
1.0
Run Length, hr
Feed Rate, dry Ib/hr
Moisture in Feed, % by wgt
39.3
17.7
78
Avg Temperature,
Freeboard
Sand Bed
Firebox
F:
1350
1450
1970
Avg Pressure, in. of H20:
Freeboard
Firebox
-3
63
Avg 02 Content of Stack Gas,
% by volume
0.5
a - off gas recycle
b - includes water vapor
224
-------
TABLE 8-8: CARBON CHARACTERISTICS: RUN #10 AND VIRGIN CARBON
10
to
Run Number
Run Date
Fixed Carbon
Recovery %
Solids, % by wgt
Volatiles
Ash
Fixed Carbon
Iodine Number 9
Molasses Number
Acid Solubles,
% by wgt
Iron, % by wgt
Aluminum, % by wgt
10
October
20-21, 1973
Feed
4.1
3.3
92.5
42
85
1.9
0.5
0.1
98
Regenerated
1.7
9.6
88.7
1040
83
2.2
1.3
0.1
August
2, 1972
Virgin
Carbon
3.0
2.7
94.3
1025
--
September
27, 1972
Virgin
Carbon
1.9
3.9
94.2
980
100
--
May
30, 1973
Virgin
Carbon
3.8
2.8
93.4
977
--
Average
3.2
3.2
93.6
981
--
-------
FIGURE 8-10: COMPARATIVE EQUILIBRIUM ADSORPTION
ISOTHERM TESTS: REGENERATION RUN NUMBER 10
p-4
Removal
Q
0
O
CfJ
ho
u^
Eq
Carbon
0)
X
•H
fe
Cn
O
0)
ho
£H
O
0.400
Oo200
0.100
0.080
Feed = Pilot Plant
Chemical Clarifier
Effluent
November 14, 1973
O!Regefaer^t©d .vrgii
ICairbOn !
6 8 10
Equilibrium SCOD (C ), mg/1
226
-------
FIGURE 8-11: RELATIONSHIP BETWEEN
IODINE NUMBER AND VOLATILE SOLIDS
Indicates Regeneration
Run Number
0
200
400 GOO
Iodine Number
227
-------
to achieve. This observation is seen more clearly in
Figures 8-12 and 8-13. Two factors are evident. First,
the higher the degree of exhaustion, the higher the incre-
mental recovery of adsorptive property. Secondly, the
higher the degree of exhaustion, the further from full re-
covery of adsorptive properties. The second factor would
seem to indicate the existence of a rate limiting phenomenon
in the FBF regeneration process.
Any discussion of the significance of the effects of thermal
regeneration variables on recovery of adsorptive properties
has to be a relative one. There is no precise and accurate
universal test for adsorptive properties! In general, the
various indirect measurements used in this study indicated
that when using the OCR principal of operation, a significant
loss of molasses number, iodine number and the ability to
adsorb SCOD from chemically treated Salt Lake City waste-
water was experienced.
In the final analysis, the effectiveness of regeneration can
only be determined by reuse in the PAC-PCT process. As will
be shown in Section IX, little difference in SCOD removal
could be found when using regenerated carbon in the pilot
plant, compared to virgin carbon.
It should be noted that comparative equilibrium adsorption
isotherms for regeneration Runs #9 and #11 are not valid
comparisons. The comparison should have been made with
actual pilot plant feed carbon (i.e., once regenerated car-
bon) and not virgin carbon.
Using a uniquely different process for carbon regeneration,
others have reported that the regeneration product was re-
covered in a slurry which had several thousand mg/Ji of SCOD
in the bulk solution.9 During this present study, SCOD's
of from 35 to 172 mg/£ were found for regenerated carbon
slurries. Recycling this SCOD material, which is probably
nonadsorbable, would result in adding about one mg/£ SCOD
to the wastewater. Thus, a process SCOD removal problem
would not exist.
The source of SCOD material in the FBF product slurry was
undetermined. It could have resulted from poor oxygen and
fuel control in the FBF unit causing volatilized, and poss-
ibly pyrolized, organics to be scrubbed out of the off-gas.
A major area of concern about the regeneration system per-
formance is the build up of inert, or ash, material upon
regeneration and reuse. For the present study, there were
three basic sources of inerts:
228
-------
FIGURE 8-12: RELATIONSHIP BETWEEN
REGENERATED AND SPENT CARBON PERCENT VOLATILE SOLIDS
6.0
Indicates Regeneration
Run Number
5 6 78 9 3D 1214
Spent Percent Volatile Solids
229
-------
IV)
OJ
o
•§
O
M
•d
©
•5
bfi
0)
OS
FIGURE 8-13: RELATIONSHIP BETWEEN
REGENERATED AND SPENT CARBON IODINE NUMBERS
400
100
200 300 400
Spent Iodine Number
500
600
800 1,000
-------
1. Sand from attrition and/or carry over from the FBF,
2. Chemical precipitates entering the carbon contacting
system from the chemical clarifier overflow and
3. Combustion products from the burning of carbon and
volatile organics.
Though sand carryover was a problem with the FBF used, it
would be hoped that better design of a prototype, including
scrubber water screening and/or classification (e.g., hydro-
cyclone), would minimize this problem.
Essentially all of the chemical compounds, coming from the
chemical clarifier overflow floe, present in the FBF product
should be acid soluble. In prototype operation, facilities
for acid treatment of FBF product should be provided to allow
elimination of these chemical inerts.
The amount of inerts generated from combustion of carbon and
organics in the FBF should be nominal. For example, if 10
percent fixed carbon losses were experienced, the ash result-
ing would amount to only about 0.3 Ib ash/100 Ib of regenera-
tion product. The contribution from combusted organics
would be of a similar amount.
In regard to the above discussion which implies little con-
cern over build up of inerts in a carbon regeneration and
reuse system, a rather real problem does exist. In Section
VI it was indicated that efficient suspended solids removal
by the granular media filter resulted in containment and
build up of carbon fines in the carbon contacting system.
It is probable that a substantial build up of carbon fines
will result from attrition of carbon in the carbon contact-
ing and regeneration operations. In addition, normal attri-
tion of sand and refractory in the FBF will probably add to
the fines to be handled in the carbon contacting and clarifi-
cation system. The best approach to removal of fines from
the carbon system would be to, when necssary for control,
recycle granular media filter backwash to the chemical
clarifier. The exact amount of blowdown of fines required
is not precisely definable, but it should be no more than a
few percent of the carbon used.
Evaluation of FBF capacity, Ib dry feed solids/hr, results
for this study indicates widely varying data. For example,
when employing the BIG operating principal (Runs #1 thru
#5), average feed rates ranged from 35 to 71 Ib/hr. When
employing the OCR operating principal (Runs #6 thru #11),
average feed rates ranged from 13 to 37 Ib/hr. The FBF
231
-------
feed rate capacity is related to heat input (combustion gas),
heat losses (in off-gas and radiation), feed moisture con-
tent, thermal efficiencies and other factors. The near
random variation of these parameters during this study pre-
cluded definition of cause-effect relationships with FBF
feed rate.
An FBF operating factor was determined which was found to be
precisely related to FBF feed rate. The following rationale
was used to define this "FBF Operating Factor". Feed rate
should be directly proportional to heat input which can be
represented by combustion gas flow rate. Part of the ther-
mal efficiency of the unit is related to the amount of gas
which must be heated to operating temperature. Thus, feed
rate should be inversely proportional to the sum of com-
bustion air and off-gas recycle flow rates. The feed rate
should be directly related to heat consumed. It can be
shown that the heat consumed is directly related to the
difference in firebox and sand bed temperatures, and other
factors.
Based on the above rationale the following "FBF Operating
Factor" was computed for each regeneration run:
(T - T
FBF f ' FB
of =
Q + Q
a ogr
Where :
Q = combustion gas flow, SCFM
Q = combustion air flow, SCFM
d
Q r = off-gas recycle gas flow, SCFM
TFB = firebox temperature, °F
TB = sand bed temperature, °F
A plot of FBF feed rate versus this "Operating Factor" is
shown on Figure 8-14. A linear regression correlation co-
efficient of 0.88 was computed, indicating a fairly precise
fit of the data. A plot including the BIG results would
show approximately the same relationship and precision. It
is not the intent of the authors to imply a general linear
relationship, but rather only for the range of data shown.
It should also be noted that the parameters used to compute
the "Operating Factor" are not mutually exclusive variables.
The data in Figure 8-14 implies that to maximize FBF capacity,
232
-------
co
X!
CD
CD
0
£1
O
40
30
251
20
15
10-
5-
0-
20
FIGURE 8-14: FURNACE FEED RATE
VERSUS FBF OPERATING FACTOR
i r ' i
25 30 35 40
'TBF Operating Factor", °F
45
233
-------
maximizing heat input (Qg), minimizing off-gas recycle
(i.e., operate at highest practical firebox temperature) and
the lowest practical sand bed temperature is necessary- The
choice of sand bed temperature would have to take into
account the effectiveness of regeneration obtainable.
In summary, it is concluded that the FBF regeneration
system using the OCR principal of operation efficiently
regenerated carbon. Fixed carbon recoveries of 76 to 100
percent were experienced with an overall average in excess
of 90 percent. Some loss of carbon adsorptive properties
was experienced. The limited time during which efficient
recoveries were experienced precluded being able to identify
operating conditions which would maximize recovery of ad-
sorptive properties for the system studied.
OPERATING PROBLEMS
During the course of this study numerous operating problems
were experienced. Not all problems affected results pre-
sented and discussed, but they are considered worthy of dis-
cussion .
Perhaps the most frustrating problem, because of its un-
definable nature, was short-circuiting of firebox gases
around the tuyere arch (a castable refractory member).
Short-circuiting was indicated by the existence of hot spots
running from the firebox to the sand bed regions of the
steel shell. Several probable causes of short-circuiting
were considered. Improper installation of refractory mater-
ials was a prime suspect. Probably just as significant was
numerous heat-up and cool-down cycles and various modifica-
tions made to the FBF unit. Several new carbon feed points
were jack hammered through the sand bed zone refractory
wall. Six holes were also jack hammered through the tuyere
arch. The existence of firebox gas short-circuiting did
not seriously effect performance when using the OCR operation
principal. Under this condition, firebox gas presumably
contained only nominal amounts of oxygen (e.g., less than
0.5 percent by volume). However, when using the BIG opera-
ting principal, firebox gas contained as much as 12 to 17
percent oxygen by volume. Short-circuiting of gas could
have resulted in carbon being exposed to oxygen at ignition
temperatures at which it would have been burned. It is un-
likely that short-circuiting was so extensive as to have
caused the 40 to 60 percent losses of carbon experienced.
The second major operational problem related to the measure-
ment of oxygen. A hand Orsat unit, with a 0 to 7 percent
oxygen by volume scale, was routinely used. The device was
234
-------
considered reasonably accurate but somewhat imprecise in
the FBF operating range of less than one percent oxygen by
volume. Attempts were made to use a continuously sensing
and indicating device. The need for near continuous cali-
brations and erratic performance resulted in the device
being abandoned.
A more serious problem existed due to the location of sampl-
ing for oxygen measurements. The sample point used was at
the off-gas stack, after the venturi scrubbers. This loca-
tion was necessary when using the BIG operating principal.
Adjustment of air and gas flow rates prior to carbon feeding
allowed achievement of a desirable net exit oxygen level.
Whether or not this desired oxygen level existed in the FBF
freeboard or upper reaches of the sand bed as well as in the
off-gas stack was not practically determinable.
For OCR operation, the off-gas stack sample point was,
after the fact, determined to not be the most desirable.
Prior to feeding carbon, measurement of oxygen at this loca-
tion was appropriate. However, if adjustments in combustion
gas or air flow rates were required during carbon feeding,
the net effect of these adjustments on firebox oxygen con-
tent was not determined. It would have been more appropriate
to have monitored the oxygen present in the firebox gas.
Another annoying problem experienced was plugging of the FBF
carbon feed line. This line was originally installed as a
2-inch diameter pipe. At 60 Ib/hr dry carbon feed rate, the
velocity was about 2.9 fpm. During some runs, this line
plugged two or three times. Plugging necessitated reduction
of fluidizing gas flow to below minimum fluidization flow,
dismantling of the feed line plumbing and rodding of the
carbon into the hot sand bed.
After the seventh regeneration run, the two inch diameter
feed line was replaced by a 1-inch diameter line with long
radius elbows. The four fold increase in velocity reduced
feed line plugging problems substantially.
Another carbon feed related problem existed in the vacuum
filter cake hopper. A ribbon screw mixed the cake, forming
a paste, which was readily pumped by a Moyno pump. Filter
cake frequently bridged the six inch throat, just above the
ribbon screw, resulting in no carbon reaching the pump.
Interruption of FBF feed due to bridging resulted in rapid
increase in sand bed temperature. The bridging tendency in
the filter cake hopper necessitated frequent operator
attention.
235
-------
Another problem frequently encountered was back flow of sand
into the 6-1/4-inch diameter BIG nozzles. After Run #3
individual gas flow meters were installed on each nozzle.
This allowed routine monitoring of gas flow and thus mini-
mized poor distribution of BIG.
On occasion, small quantities of sand were found on the
firebox floor. This resulted from back flow of sand through
the tuyere orifices. The extent of back flow was considered
nominal and not a significant problem.
A previous study using a similar FBF unit with OCR operation
indicated extensive corrosion problems in the OGR system.
Based on their experience, OGR plumbing used in this study
was reinforced fiberglass piping. No corrosion problems
were experienced. The type of OGR blower used in the
current study had mild steel rotors. These rotors were
severely corroded after about two regeneration runs and
necessitated replacement of the blower.
COMMENTS ON "BIG" OPERATING PRINCIPAL
As previously noted, the authors concluded that the BIG
operating principal with the FBF unit used in this study did
not result in acceptable recovery of fixed carbon. The
fundamental problem was defined as being non-preferential
combustion of natural gas in the lower portion of the sand
bed, prior to the carbon feed point. There are at least two
factors which might explain why this problem was insurmoun-
table in this study.
The first factor relates to the comparative flame propa-
gation rates of carbon and the natural gas fuel used.
Flame propagation rate is the velocity at which a flame
will travel in a fuel air mixture. Data on the flame propa-
gation rate for the carbon used in this study could not be
found. It was determined that the flame propagation rate
for pulverized soft coal was considerably in excess of that
for natural gas.1 ° It is thus inferred that, were carbon
and natural gas (BIG) to have been simultaneously exposed
to localized excess oxygen, that the carbon would have been
preferentially burned to a higher degree than would the
natural gas. That localized pockets of gas with excess
oxygen might have existed in the sand bed is entirely con-
ceivable when considering the second factor to be discussed;
namely, the quality of sand bed fluidization. A large body
of published knowledge exists concerning fluid-bed tech-
nology. Several observations were made during operation of
the pilot plant FBF unit with the BIG operating principal
which might allow bringing to bear some of this technology
236
-------
in an attempt to understand why unsuccessful carbon recovery
results were experienced.
Almost without exception the pressure drop across the bottom
18-inches of the fluidized sand bed fluctuated widely during
operation. The absolute value of pressure drop varied with
fluidization velocity, bed temperature and size of sand, but
the fluctuations were consistently about +6 inches of water.
When visually observing the fluidized bed, through a sight
glass mounted on top of the unit, it was also observed that,
almost without exception, the top of the bed would vacilate
over about a ft depth. Typically the top of the bed would
stay at about the 4-ft marker for several seconds and then
suddenly, and somewhat violently, surge to the 5-ft marker
and stay there for about a second. The entire FBF unit
could be felt to pulse or shudder when this occurred. Some
sand grains were thrust upward, well into the freeboard
zone during these bed surges. These observations are in-
dicative of a "slugging" fluidized bed.11 Slugging occurs
due to the growth of rising gas bubbles when experiencing
a uniformly reduced pressure region and also partly due to
coalescence of bubbles. Slugging typically occurs when
fluidization velocity is considerably greater than the mini-
mum particle fluidization velocity.
The type of tuyere system used could also have had an effect
on slugging. Ideally, the greater the number of uniformly
spaced orifices and higher the pressure drop, the more uni-
form the distribution of gas flow and the smaller the gas
bubbles formed at the orifice outlet. Presumably it was
this factor that prompted the recommendation of installation
of additional tuyeres to improve the quality of bed fluid-
ization and mixing.
A second set of observations related to the presence of fuel
(natural gas) in the FBF freeboard zone. With the FBF at
operating temperatures and levels of fluidization,, bursting
of bubbles was observed at the surface of the sand bed.
Obviously, the bubbles were not seen, but rather the motion
and color of sand at the bed surface. When a bubble would
burst at the bed surface, the color of sand would glow more
intensely in that region. Bubbles would break the bed sur-
face forming 6 to 14-inch diameter circles of sand discolor-
ation. As a bubble would break at the bed surface, a flash
of light usually occurred at that spot. This obviously
indicated ignition of natural gas. Whether the bubbles or
the freeboard under certain operating conditions was docu-
mented by temperature profile print out.
The observation of "bursting bubbles" and burning of natural
gas in the freeboard zone substantiated that slugging was
237
-------
occurring and that the quality of fluidization was typical
of gas fluidized sand beds.
The reasons for slugging being generally observed are not
clearly understood by the authors. As previously noted,
normal fluidization velocities especially at the bottom
of the variable area bed container, were considerably in
excess of minimum fluidization velocities. This condition
typically results in slugging.11 In addition, the relatively
large bed depth to diameter ratio would have exaggerated
slugging. Another factor which might have promoted slugging
was the relatively low designed tuyere headless of ID-inches
of water. Others suggest that a minimum of 14-inches of
water be used for design.11
During carbon feeding, a red haze of carbon existed in the
freeboard zone making it impossible to view the surface of
the fluidized bed. Considering that carbon, with it's
associated water, was fed at a single point and that the
vaporized water volume was as high as 50 percent of firebox
gas flow, it is quite probable that gross slugging occurred
just above the carbon feed point. This factor plus the
typical non-random nature of gas fluidized sand beds might
have resulted in gross short-circuiting of carbon to the
surface of the sand bed. It is probable that carbon short-
circuiting resulted in residence times much shorter than
indicated by average gas velocities and depth of travel.
Another factor which might have caused short-circuiting
in the freeboard zone and thus promoted slugging was the
relatively shallow freeboard depth. As can be seen in
Figure 8-2 when the sand bed was fluidized to the 5-ft
level, there was only about a 4-ft freeboard.
Based on the authors' experience presented in this report
and their interpretation of a limited amount of published
literature on the subject of fluidized bed processes, they
seriously doubt that the BIG operating principal could be
made to work in a prototype carbon FBF unit. There are,
however, several suggestions to be made which substantially
increase the probability of success.
The bed injection fuel used should have a flame propagation
rate well in excess of carbon if such a fuel is economically
available. Perhaps the operating fluidization velocity
should be limited to minimize slugging. Much reduced bed
depth to diameter ratios for prototype units should also
reduce the tendency for slugging. Deeper freeboard would
also be suggested to reduce any tendency for slugging and
also to minimize sand carryover. Perhaps increased numbers
-------
of tuyeres with smaller orifices and higher pressure drops
would also reduce the slugging tendency. It is possible that
a special tuyere could be designed which could accomplish
more efficient mixing of firebox gas (with its excess oxygen)
and bed injection fuel.
Considering the substantially higher capacity per unit area
when using the BIG operating principal it would be hoped
that those truly skilled in the art of fluidized bed pro-
cesses will actively pursue further development in this
area.
239
-------
SECTION IX
CARBON SYSTEM RESPONSE
The singular purpose of carbon treatment was removal of
soluble organics from the wastewater. To describe the car-
bon system and to form a basis for comparing treatment
effects, a model of the carbon system response is desirable.
This response model should interrelate, at least, carbon
system feed and effluent soluble organics and carbon dosage.
The identification of a carbon system response model was not
an original objective of this study.
Periods of reasonably stable carbon system operation and
performance were identified before pertinent system variables
could be quantified. The first step was to plot effluent
quality, carbon dosage and carbon system solids retention
time (SRT) for each day of pilot plant operation. Routinely
collected solids-contact unit solids profile data indicated
uniform solids concentration in the carbon contactor units.
Therefore an approximate carbon SRT was calculated as
follows.1
SR _ (Reaction Zone Solids) (Volume of Carbon Slurry)
(Flow Rate) (Carbon Dosage)
The numerator is an approximation of the inventory of carbon
(pounds) within the solids-contact units and the denominator
is the carbon fed (Ib/day).
Various treatment campaigns were determined on the basis of
staging, chemical pretreatment, carbon dosage, effluent
quality, SRT and use of regenerated or virgin carbon. SCOD
was used as the measure of organic material.
Tables 9-1 and 9-2 present average SCOD, SRT and carbon dos-
age data. Reduced data shown indicate incremental organic
reductions and organic removals for the various campaigns.
During pilot plant operation, granular media filtration of
all carbon contactor effluent was impossible (e.g., when
carbon contactors were run in parallel). Thus, the effluent
240
-------
TABLE 9-1: CARBON TREATMENT SYSTEM: SUMMARY OF SINGLE-STAGE DATA
Campaign #
VLl-la
VL1-2
VL1-3
VL1-4
VL1-5
VLl-6
VL1-7
VL1-71
VL1-8
Observed
Fixed
Carbon
Dosage
(M) ,mg/a
319
298
0
98
0
31
198
141
141
Data
SCOD, mg/£
Raw
Waste-
water
(Cr->
r
28. 8
26.8
42.3
42.3
46.0
46.0
40.3
35.8
35.8
Carbon
System
Feed
-------
TABLE 9-2: CARBON TREATMENT SYSTEM: SUMMARY OF TWO-STAGE COUNTER-CURRENT DATA
Cam-
paian
#
VA2-9a
VA2-10
VA2-11
VF2-12
VF2-13
RF2-14
RA2-15
RA2-16
i
Observed Data
Fixed
Carbon
Dosage
(M) ,
106
294
103
108
75
101
72
28
SCOD, mg/£
Raw
Waste-
water
(Cr>
39.0
40.1
40.7
34. 3
30.3
32.5
42.8
47.6
Carbon
System
Feed
(Co>
23.8
23.0
22,7
19.6
17.3
17.9
24.0
24.3
Inter-
mediate
Feed
(C±)
10.0
7.0
13.1
8.7
6 .6
7.1
17.4
15.9
Eff-
luent
4. 2
2.0
5.0
2.7
3.0
3,8
7,0
10.6
Reduced Data
SCOD Removed,
mg/£
1st
Stage
(X")
13.8
16.0
9.6
10.9
10.7
10.8
6.6
8.4
2nd
Stage
(X')
5.8
5.0
8.1
6.0
3.6
3.3
10.4
5.3
All
Sta-
ges
(X)
19.6
21. C
17.7
16.9
14.3
14.1
17,0
13,7
Organic
Removal
g/g
All
Sta-
ges
(X/M)
0.185
0.071
0.172
0.156
0.191
0.140
0.236
0.489
2nd
Stage
(X'/M)
0.055
0.017
0.079
0.056
0.048
0.033
0.144
0.189
Carbon
SRT,
days
14
6,1
15
9.8
20
13
18
30
Run
Length
days
13
19
32
49
21
23
34
7
NJ
*VA2-9 = Virgin Carbon; Alum P re treatment; 2_ Two-Stage Counter-Current Operation;
9-Run #9~
-------
SCOD concentrations shown in Table 9-1 are from the final
carbon contactor and not the granular media filter. This
approach is consistent with previously reported results.1
With few exceptions, the effluent SCOD's of the granular
media filter and carbon contactor effluents were equal.
The campaign numbers in Table 9-1 and 9-2 were coded to
indicate certain operational variables. The first letter
indicates the use of virgin (V) or regenerated (R) carbon.
The second letter indicates chemical pretreatment, lime (L),
ferric chloride (F) or alum (A). The first number indicates
single-stage (1) or two-stage counter-current (2) operation.
The number after the hyphen indicates the chronological
campaign number. For example, Campaign VA2-10 is virgin
carbon, alum pretreatment, two-stage counter-current con-
tacting and campaign number 10.
CARBON SYSTEM RESPONSE MODEL
Several types of models have been used to describe real time
powdered carbon-organic removal system responses.1'12'13
In a previous report the authors used both the classical
Freundlich adsorption model and an empirical model as
follows:l
- = b —
M a X M
Where:
X = organics removed (C -C ), mg/£
M = fixed powdered carbon dosage, mg/£
C = system feed organics, mg/£
C = system effluent organics, mg/£
a,b = emperical constants
This emperical model, X/M vs C /M, was first used by
others. 2
Correlation of system feed arid effluent SCOD's and carbon
dosages was found to be very imprecise when using the
Freundlich model; whereas, use of the X/M vs CO/M model
resulted in very precise data correlations. Because of the
variability of the powdered carbon system performance re-
sults, the use of neither model allowed determining
statistically significant treatment effects, even though
the actual data suggested substantial effects due to, for
example, the number of counter-current contacting stages.1
243
-------
Both of the above models have unreal extrapolated boundry
conditions. The X/M vs CO/M model predicts the unreal con-
dition of being able to achieve zero effluent SCOD at a
large, yet finite, carbon dosage. Extrapolation of the
Freundlich model indicates an asymptotic approach to zero
effluent SCOD at infinitely large carbon dosages.
Extrapolation of the Freundlich model to very low carbon
dosage, approaching zero, predicts that the carbon system
effluent SCOD will"approach the influent SCOD. On the other
hand, extrapolation of the X/M vs CO/M model to zero carbon
dosage predicts that the system effluent SCOD will be sig-
nificantly less than the system feed SCOD. As will be
shown later in this Section this latter boundry condition
was observed for the powdered carbon system used.
Because of the factors discussed above, there was no clear
basis for choosing one or the other model to evaluate the
results obtained during this study. Due, in part, to the
following reasons, the decision was made to use the Freund-
lich model.
1. Most published articles dealing with the application
of powdered carbon treatment of municipal waste-
waters have used the Freundlich model.
2U Laboratory equilibrium adsorption responses are
usually precisely correlated by the Freundlich
model.
3. Use of the X/M vs CO/M model provides no new in-
sight into the evaluation of treatment effects on
the powdered carbon system response.
Figure 9-1 is a Freundlich plot of organic removal data for
single and two-stage counter-current operation using virgin
carbon. The lines shown are linear regressions of log X/M
on log Ce. Table 9-3 presents several such regression
equations for organic removal data in Table 9-1 and 9-2.
It is apparent from the poor correlation coefficient in
Table 9-3 and the relative slope in Figure 9-1 that the
single-stage curve does not realistically describe the
system response. The slope is much too flat relative to
the laboratory equilibrium adsorption isotherms and to the
two-stage results. In general, it appears that organic re-
movals, X/M, for two-stage counter-current treatment are
generally higher than for single-stage treatment.
244
-------
FIGUEE 9-1: SCOD REMOVAL-ADSORPTION RESPONSE:
PILOT PLANT RESULTS USING VIRGIN CARBON
J_l
eS
Remov
Q
0
O
03
tu
•n
Q)
fc
arbon
O
OJ
X
h
o
(D
K
IH
O
0.31
6 8 10 S3 40
Carbon Contactor Effluent SCOD (Ce), mg/£
245
-------
TABLE 9-3
REGRESSION ANALYSIS LOG X/M ON LOG C
Contact
Stage
Single Stage
(Virgin Carbon)
Two-Stage Counter-Current
(Virgin Carbon)
Two-Stage Counter-Current
(Virgin & Regenerated
Carbon)
First Stage of 2SCCa
(Virgin & Regenerated
Carbon)
Second Stage of 2SCC
(Virgin & Regenerated
Carbon)
First Stage & Second
Stage (Virgin &
Regenerated Carbon)
Regression Correlation
Equation Coefficient
1= 0.073 C °'37
M e
f=°-053Ce°-87
i=0.049Ce°-90
|= 0.018 C "•"
M e
| = 0.009 C 1'31
M e
| = o.oio c 1-23
M e
0.50
0.76
0.90
0.71
0..1
0.92
2SCC = Two-Stage Counter-Current
246
-------
The cross-hatched area in Figure 9-1 represents the range
of laboratory equilibrium adsorption isotherm test results
obtained during this study. Organic removal during these
tests was presumably by an adsorption mechanism only.
Organic removal in the pilot plant carbon system is seen
to be considerably higher than predicted for adsorption by
the laboratory adsorption test results. This fact rein-
forces the thesis that a removal mechanism (biological or-
ganic removal) in addition to adsorption existed in the
carbon contactor.
Figure 9-2 is a plot of all single-stage and two-stage
counter-current data from a previous study and the present
study with regression curves.1 The purpose for presenting
all available data was to increase the range and hopefully
the precision of the data correlation. In a general sense
the response shown is realistic because it does show the
expected difference between single and two-stage systems.
At low Ce's, the difference in X/M between single and two-
stage systems is large. As Ce becomes large, the differences
in X/M becomes increasingly smaller.
It is unrealistic to believe that only an adsorption mech-
anism is operative in the system studied. The differences
in pilot plant and laboratory organic removal can best be
explained as due to the presence of anaerobic biological
activity. The impact of any other treatment effects on the
carbon system response is not evident in Figures 9-1 and 9-2,
The following subsection is an attempt to elucidate various
treatment effects,
ANALYSIS OF TREATMENT EFFECTS
Number of Stages
Figure 9-3 is a plot of fixed carbon dosage (M) versus eff-
luent SCOD (Ce) for all pilot plant data. The regression
curves shown (see Table 9-3) were calculated using an aver-
age feed of 37.3 mg/£ SCOD for the single-stage results and
21.6 mg/£ SCOD for the two-stage counter-current contacting
results.
At face value the results shown in Figure 9-3 suggest that
the number of stages used (two versus one) resulted in a
different system response. At virgin carbon dosages of
75 mg/£ or more, the average system effluent contained
3.4 mg/£ SCOD (standard deviation of 1.2) for two-stage
counter-current contacting. For single-stage contacting,
at carbon dosages of 98 mg/£ or more, the system effluent
247
-------
FIGURE 9-2! SCOD REMOVAIr-ADSORPTION RESPONSE:
COMBINED RESULTS OF CURRENT AND PREVIOUS STUDIES
f£= 0.026 C°-7«1
o.oi-
6 8 10 20
Carbon Contactor Effluent SCOD (Ce),
40
mg/i
60 80
248
-------
Cn
g
CD
U
Q
O
U
CO
W
S-l
0
-p
U
0
U
c
O
£!
J-l
(C
U
FIGURE 9-3:
SCOD REMOVAL - ADSORPTION RESPONSE:
PILOT PLANT EFFLUENT SCOD RESPONSE TO CARBON DOSAGE
40
30 -
20
10
[7] Single-stage, Virgin PAC
X =
M
0.073
X = 0.047 C
0.37
e
(Average Co = 37.3)
Two Stage Counter-Current
Virgin PAC
Regenerated PAC
0.90
(Average C0 = 21.6)
O
100 200 300 400
Fixed Carbon Dosage, mg/£
500
249
-------
contained an average of 8.4 mg/£ SCOD (standard deviation of
1.1). These average effluent SCOD concentrations are statis-
tically significantly different at a probability level of
0.99.
The fact that average carbon system feed SCOD differed sub-
stantially for the two contacting modes (37.3 vs 21.6 mg/£)
would tend to negate the validity of any inference that the
system responses shown in Figure 9-3 were effected by the
number of contacting stages,
Evaluation of SCOD removal data in Tables 9-1 and 9-2 in-
dicates that for virgin carbon dosages greater than 75 mg/£
the single-stage contacting mode removed 77 percent of the
SCOD (standard deviation of 1.8). The two-stage counter-
current system removed from 78 to 91 percent SCOD with an
average of 84 percent removal (standard deviation of 5.0).
The use of two-stages of contacting resulted in only slightly
improved percent removal of SCOD.
It must be concluded that the results of this study do not
allow quantification of the effect of number of contacting
stages on the powdered carbon-organic removal system re-
sponse .
Chemical Pretreatment Effects
On Figure 9-3 there is an obvious and statistically signifi-
cant difference in system feed SCOD concentration between
single and two-stage results. This difference can be related
to chemical pretreatment effects. All single-stage operation
was done with hydrated lime pretreatment and all two-stage
operation was done with inorganic coagulants.
In a previous report (reference 1) it was concluded that there
was no significant effect on organic removal by different
chemical pretreatments. This conclusion was made on the
basis of laboratory work only. When actual pilot plant re-
sults from the current study are evaluated, the effect of
chemical pretreatment becomes apparent.
Figure 9-4 is a plot of laboratory equilibrium adsorption
isotherm tests run on the effluent from the chemical treat-
ment unit. The cross-hatched area represents the range of
laboratory adsorption equilibrium isotherm tests run on
untreated raw wastewater during previous studies.1 The
data show that an apparent unadsorbable fraction was re-
duced from 10 to 20 mg/£ SCOD for lime or no treatment to
3 to 7 mg/y, SCOD with alum or FeCl3 pretreatment. Figures
9-3 and 9-4 show that the feed SCOD concentrations produced
250
-------
FIGURE 9-4: SCOD REMOVAL —ADSORPTION RESPONSE:
LABORATORY EQUILIBRIUM ADSORPTION ISOTHERMS
0.400
0.200
73
0)
0
<0
p
o
o
ra
be
0.100
.080
.060
O
.040
en
rt
o
Q)
p;
o
•p o.oio
etf
0.007
0.005
0.003
iii ii;.i I ^Niilliii; I
4 6 8 10 20
Equilibrium SCOD (Ce), mg/Ji
40
251
-------
by alum or FeCl^ pretreatment are considerably lower than
SCOD feed concentrations produced by lime pretreatment. It
can be inferred from data on Figure 9-3 that, because of
the removal of apparently unadsorbable material, lower eff-
luent SCOD (Ce) was achieved with the inorganic coagulants.
Previously reported laboratory data indicated that both high
pH lime and inorganic coagulant pretreatment significantly
reduced SCOD; however, equilibrium adsorption isotherm re-
sponses were not significantly affected. Evaluation of
SCOD data in Table 9-1 and 9-2 indicates that lime pretreat-
ment (pH 10.8 to 11.2) results in from about 25 percent de-
crease to 25 percent increase in SCOD. Alum or FeCls pre-
treatment consistently resulted in 40 to 50 percent removal
of SCOD. It is reasonable to expect that such significantly
different pretreatment results would affect subsequent car-
bon treatment responses. It is the authors' conjecture that
alum or ferric chloride treatment removes unadsorbable and/
or very difficult to adsorb organic material, whereas, lime
treatment does not. Whether or not this conjecture is true,
the single-stage data in Figures 9-1 arid 9-3 should be
shifted to lower equilibrium SCOD values to ascertain any
effect of staging.
Biological Effects
A biological organic removal mechanism has been inferred
based on data presented in Figure 9-1, where organic removal
in the carbon system was seen to be considerably higher than
predicted by laboratory equilibrium adsorption isotherm tests,
The existence of a biological organic removal mechanism is
also verified by data from Campaigns VLl-3 through 6 (Table
9-1). During Campaigns 3 and 5, no carbon was fed to the
carbon contactor and yet about 30 percent removal of SCOD
was observed. During these campaigns about 150 to 200 Ib of
spent carbon slurry, of about 8 q/H concentration, was main-
tained in the carbon contactors. Campaigns 4 and 6, with
carbon dosages of 98 and 31 mg/£, were run parallel to
Campaigns 3 and 5, respectively. Using biological organic
removal data from Campaign 3 and 5, it is inferred that 66
and 82 percent of the total SCOD removal achieved in
Campaigns 4 and 6, respectively, was by a biological mechan-
ism.
Unfortunately, there was no parallel operation with signifi-
cant difference in SRT which would have allowed making an
absolute identification of SRT effects. However, by compar-
ing Campaigns VLl-1 and 2 with Campaigns VLl-7' and 8, an
252
-------
effect of SRT can be seen for single-stage operation.
Going from VL1-1 and 2, to VLl-7" and 8 it is evident that
equivalent effluent quality was achieved at a carbon dosage
of 141 mg/£ and an SRT of 7.1 days as was achieved at a
carbon dosage of 309 mg/£ and an SRT of 2.8 days.
Digressing for a moment, it is germane to note that when all
pertinent treatment variables are held nearly constant, the
carbon system response was precisely reproduced. Comparison
of SCOD removal for parallel Campaigns VLl-1 with VL1-2 and
VLl-7' with VL1-8 demonstrates this precision.
By comparing Campaign VF2-12 with Campaign VF2-13 an effect
of SRT can also be seen for two-stage operation. Going
from VF2-12 to VF2-13 it is evident that by reducing carbon
dosage from 108 mg/£ to 75 mg/£ and increasing SRT from 9,8
days to 20 days equivalent effluent quality of about 3 mg/&
SCOD was achieved.
Regeneration Effects
A reduction in SCOD adsorptive capacity of regenerated car-
bon was shown in Section VIII by iodine number data and by
comparative equilibrium adsorption isotherm tests.
Figure 9-5 is a Freundlich plot of all virgin and regenera-
ted carbon data (two-stage counter-current contacting). Re-
gression curves are shown for the virgin carbon data and
for virgin plus regenerated carbon data. In the same range
of equilibrium SCOD concentration, regenerated carbon or-
ganic removals were slightly below virgin carbon organic
removals. The limited number of data points and the rela-
tively high variance of the virgin carbon data precludes
identification of a statistically significant difference.
The net effect of a slight loss of adsorptive capacity is
minimized when reusing regenerated carbon in the plant due
to the presence of a substantial biological organic removal
mechanism.
Comparison with Previous Results
In a previous contract report, the first and second stage
organic removal responses were drastically different.1 In
the present study, organic removal models are approximately
the same for both stages. During this study high concen-
trations of carbon fines (50 to 70 mg/&) were allowed to
overflow the first-stage contactor-clarifier into the
second-stage unit. By so doing, a continual seeding of
the second-stage with active biomass was probably achieved.
Data for two-stage contacting in Table 9-2 indicate that
253
-------
d
0)
fa
o
K
Q
O
O
br
fa
I
0)
PS
o
bfi
h
O
FIGURE 9-5: SCOD REMOVAL-ADSORPTION RESPONSE:
COMPARISON OF REGENERATED AND VIRGIN CARBON
1.0-
0.8-
0,6-
0.2-
o.i-
0.03-
0.06-
0 v
0.02-
Virgin Carbon Only
CQ=21.6
Virgin and Regenerated
Carbon
Regenerated Carbon
Carbon Contactor Effluent SCOD (Ce), mg/n
254
-------
relatively high SRT's were used. On numerous occasions
excessive odor levels were found eminating from the second-
stage contactor as well as the first-stage unit. This was
not experienced during previous contract studies.1 In the
previous study adsorption and biological removal was in-
ferred to be operative in the first-stage and predominantly
adsorptive removal in the second-stage. In the present
study adsorption and biological removal was apparently occurr-
ing in both stages.
As rioted earlier in this Section, pilot plant results from
the present study definitely indicated the presence of a
chemical pretreatment effect when comparing lime and alum
or ferric chloride usage. Such an effect had been con-
jectured previously, but it was not identified as having
a significant effect on the effluent SCOD quality obtainable
with the PAC-PCT process,
Results from both the previous and present studies qualita-
tively indicated that more efficient use of carbon can be
made using a two-stage counter-current contacting system
compared to a single-stage system. However, the variance
of results in both studies precluded defining the difference
with reasonable statistical significance. This short-corning
is understandable when one considers that staging would effect
only the adsorptive removal response. As indicated previous-
ly in this Section, adsorption could account for as little
as 18 to 34 percent of the total SCOD removal by the PAC-PCT
system, To determine the effect of a treatment variable
which has an impact on such a small portion of total SCOD
removed from a relatively low strength, wastewater would re-
quire very precise data.
ULTRAVIOLET MONITORING STUDY
Most carbon system response models indicate that carbon dos-
age requirement to produce a constant effluent SCOD is
linearly related to the feed SCOD concentration. If one
could instantaneously measure SCOD, either directly or in-
directly, carbon dosage could be varied directly with feed
SCOD changes. In addition it would be desirable to have
real time monitoring of carbon system effluent to alert
operating personnel of undesirably high system effluent SCOD.
Such an operation should result in production of a very con-
sistent quality effluent.
Unfortunately, instrumentation for direct real-time deter-
mination of SCOD on a plant operation scale is not commer-
cially available. Evaluation of available indirect methods
indicated that ultraviolet light (UV) absorption might be
255
-------
of utility. A correlation of UV absorbance of 254 nm and
TOG had been determined for a variety of treated and un-
treated waters and wastewaters.14 Based on this correlation
and the desire for instantaneous organic monitoring, a
commercially available UV instrument was installed at the
pilot plant. The purpose of the UV instrument installation
was to demonstrate the usefulness of UV-absorbance as an
operational tool to continuously monitor organic concentra-
tion. The primary objective was to determine the correla-
tion between UV absorbance at 254 nm and dissolved organics,
and to use UV monitoring to operate the carbon treatment
system. After two months of data collection, it became
evident that a poor correlation was achieved (Figure 9-6).
A laboratory study was designed to identify the variables
responsible for the poor correlation observed. (Refer to
Appendix C) .
The basic problem of achieving a general UV absorbance-organ-
ic concentration correlation results from the changing
relative concentrations of the organic solutes present in
the wastewater. In the PAC-PCT process, relative concen-
tration changes result from at least three basic causes:
(1) chemical treatment selectivity, (2) activated carbon
adsorption selectivity and (3) natural diurnal organic con-
centration changes.
In a selective PAC-PCT treatment process, general UV absor-
bance-organic concentration correlation prospects are minimal,
However, a working correlation should be possible at specific
treatment points (e.g., after chemical or carbon treatment)
if very consistent treatment effects are achieved.
To identify a working correlation at a specific treatment
point, the PAC-PCT plant's final effluent was continuously
monitored. Analysis of data for three months of monitoring
showed a fairly constant effluent of about 3 mg/£ SCOD (1 to
5 range) and 0.4 absorbance (0.2 to 0.8 range). Since the
effluent SCOD was such a low and relatively constant value,
it was impossible to identify a correlation between UV and
SCOD.
It has been generally observed that, after backwash of a
granular media filter, poor quality effluent is usually
experienced for a short period of time. When this "bad
water" was experienced during this study the UV printout in-
creased drastically for a short period of time, indicating
UV is responsive to this type of treatment upset.
The use of UV absorbance to monitor the plant effluent for
the purpose of determining carbon dosage was judged to be an
unrealistic venture.
256
-------
K)
Ln
\ ft4 - --
20 30 40 50
Soluble Chemical Oxygen Demand, mg/£
60
70
-------
SECTION X
PAC-PCT PROCESS DESIGN RECOMMENDATIONS
The purpose of this section is to present the authors'
recommendations for design of the various unit operations
and processes included in the PAC-PCT municipal wastewater
treatment system studied. These recommendations were formu-
lated from the results presented in Sections VI through IX,
previous contract work (reference 1) and other relevant
experiences of the authors. It is pertinent to note that
the following subsections will contain some judgement and
opinions of the authors.
The design recommendations presented here will be used in
the next section as the basis for an economic analysis of
the PAC-PCT system studied.
The unit operations and processes to be discussed will
include:
1. Preliminary treatment
2. Chemical treatment
3. Chemical-primary sludge treatment
4. Powdered carbon treatment
5. Carbon regeneration
Also included will be comments concerning operation of the
PAC-PCT system.
PRELIMINARY TREATMENT
Preliminary treatment normally consists of screening, grit
removal and comminution. During the present study, there
was very ineffective preliminary treatment of the raw waste-
water being used. Therefore, it was demonstrated that the
type of solids-contacting chemical treatment unit used would
operate successfully without the removal of either grit or
fibrous material. It is, however, recommended that both
grit and fibrous material be removed prior to the chemical
treatment step to avoid possible clogging and wear of pipes,
pumps, and mixing turbines.
258
-------
Although preaeration of the raw wastewater was not evaluated
during this study, it is suggested. It might reduce the
strong anaerobic conditions that can exist in the chemical
treatment unit, especially when alum or FeCl3 is used.
The following preliminary treatment steps are recommended
for Salt Lake City or similar wastewaters.
1. Mechanically cleaned bar screens with 2-inch
openings
2. Aerated grit chambers with grit removal and washing
3. Comminution
Conventional design and sizing criteria are suggested. Raw
wastewater flow measurement, indication, recording and in-
tegration (totalizing) are also considered a must in order
to enable pacing subsequent chemical and reagent feed flow
rates, thus maintaining constant dosages.
CHEMICAL TREATMENT
Choice of Chemicals
The choice of treatment chemical is often based on engineer-
ing judgement and/or preliminary laboratory or pilot plant
testing results.
Lime would probably be the chemical of choice for a waste-
water low in alkalinity and high in phosphorus concentration.
Phosphorus removal by lime treatment is directly related to
pH and independent of influent phosphorus concentration. For
a low alkalinity wastewater raising the pH can be achieved
at reasonably low lime dosages.
If a wastewater is high in alkalinity and low in phosphorus
concentration, then either alum or ferric chloride would
probably be the chemical of choice. Neither alum nor ferric
chloride treatment are adversely affected by high alkalinity
except that slightly increased dosages would usually be re-
quired due to operation at a pH in excess of the optimum.
At a given effluent soluble phosphorus concentration, the
amount of either alum or ferric chloride required is directly
proportional to the feed phosphorus concentration.
Another alternative would be to use the combination of lime
and FeCl3 at a pH range of about 9 to 10. In this approach
the lime provides for pH adjustment (phosphorus insolubiliza-
tion) and the FeCl3 for coagulation of suspended solids (in
eluding chemical precipitates),
-------
A desire to remove soluble organics in the chemical treat-
ment unit would also effect the choice of chemicals. During
this study, it was observed that use of alum or ferric
chloride treatment resulted in 40 to 50 percent SCOD removals;
whereas lime treatment resulted in no consistent effect on
SCOD removal. The observed SCOD removal values for lime
treatment ranged from +25 to -25 percent, indicating net
removal at times and net solubilization of particulate matter
at other times. The results in Section IX strongly implied
that alum or ferric chloride treatment removed some appar-
ently unadsorbable SCOD and thus minimized PAC-PCT system
effluent SCOD.
Sludge dewatering and disposal alternatives are key considera-
tions in the choice of chemical. During this study, lime-
primary sludge was found to thicken and dewater more easily
than either alum or ferric-primary sludges. However, five
to six times as much sludge (by weight) was produced by lime
treatment than by alum of ferric chloride treatment. For
lime treatment of wastewaters containing a relatively high
concentration of magnesium, single-stage lime treatment is
appropriate. On the other hand, for wastewaters where the
magnesium concentration is low, two-stage lime treatment-
recarbonation is considered appropriate. Lime is added to
the first stage to produce a high pH level in order to pre-
cipitate significant magnesium hydroxide for coagulation.
The excess calcium added (in the lime) to achieve the high
pH is normally precipitated by recarbonation prior to a
second stage solids-contacting clarifier. The additional
CaC03 sludge produced in the second stage must be disposed
of. Recalcination of the CaC03 sludge has been found to be
economically justified in some applications,
The above comments provide some insight into relevant con-
siderations concerning the choice of treatment chemical.
The only rational approach to choosing a treatment chemical
is to assess the relative total capital and operational and
maintenance costs for both the wet end and the sludge treat-
ment and disposal alternatives available.
A previously reported economic analysis of lime, alum or
ferric chloride treatment of Salt Lake City raw municipal
wastewater indicated that alum was the least expensive.l
This analysis included chemical-primary sludge incineration
and land disposal of the ash produced. The results of the
current study substantiate that alum is the economic choice
for Salt Lake City.
It is the authors' judgement that designing for continuous
use of polyelectrolyte coagulation-flocculation aid cannot
be economically justified. It is, however, readily acknow-
260
-------
ledged that use of polyelectrolytes allows operation of
the chemical treatment clarifier at significantly increased
hydraulic loadings. Increased sludge thickening rates
would also be expected in both the clarifier and a gravity
sludge thickener. It is, therefore, recommended that the
chemical treatment clarifier and gravity sludge thickener
be designed to routinely operate without the use of poly-
electrolyte; but that a polyelectrolyte feeding system be
installed. Doing so will afford the plant operator some
flexibility and operational alternatives, especially during
periods of treatment upsets, unexpected high wastewater flows
and/or mechanical problems (e.g., a clarifier being off-
stream) .
Chemical Mixing
Achievement of effective and economical coagulation requires
rapid initial dispersion of coagulant throughout the entire
wastewater flow. For both alum and ferric chloride coagu-
lation the flash mixer should be designed to produce a
mean temporal velocity gradient, G, of 800 to 1000 sec"-'-,
with a turbine tip speed of about 5 ft/sec. A nominal
detention time of 30 seconds at peak flow should be adequate.
For lime addition, it is very desirable to react the slaked
lime slurry with recycled sludge prior to contacting the
raw wastewater. Adding lime to the recycle sludge prevents
gross scaling near the lime feed point. The recycle solids
offer a large surface area on which the chemical precipitates
can form thus promoting rapid, complete chemical reactions
and producing a more stable effluent.
Solids recycle can be achieved either by internal or ex-
ternal pumpage of sludge. The method used during this study
was internal recycle. This method is favored because of
successful experiences during this study and because it
avoids the potential problem of plugging in external sludge
pumping and plumbing systems.
When using alum or ferric chloride as primary coagulants,
if desired, polyelectrolyte should be dispersed with an in-
line mixer at a relatively low G value (100 to 200 sec"-*-).
In no case should the polyelectrolyte merely be added to
the surface of the flocculation (or reaction) zone since
very poor dispersion occurs, lowering effectiveness and in-
creasing the cost of using polyelectrolyte. Manufacturer's
recommendations for polyelecltrolyte make-up, storage and
handling should be followed.
261
-------
Type of Equipment
A solids-contact treatment unit, with internal sludge re-
circulation, is recommended by the authors for lime treat-
ment. This type of solids-contact unit can achieve a high
degree of sludge recycle and, therefore, produce a stable
water which minimizes scaling problems. For lime treatment,
high upflow velocity in the draft tube is required to pick
up and recirculate sludge. The sludge rakes can be a standard
design normally used in wastewater treatment; however, torque
capacity for lime treatment applications should be at least
twice that required for alum or ferric chloride treatment
applications. The chemical treatment unit must have a sur-
face skimming device to remove accumulated floatables. The
skimmed floatables may be combined with the blowdown solids
from the chemical treatment unit or concentrated separately
in a scum concentrator, depending on the amount of floatables.
For alum and ferric chloride treatment, no benefit was de-
fined, either in the laboratory or full-scale, for solids-
contact treatment. However, the solids-contact unit used in
this study provided a substantial sludge age which was par-
tially responsible for the presence of anaerobic biological
activity in the chemical treatment unit. This biological
activity along with chemical coagulation resulted in as much
as 40 to 50 percent removal of SCOD.
It is recommended that the classical three step (flash mix,
flocculation and sedimentation) chemical treatment approach
be used for alum and ferric chloride treatment. This
recommendation is based on economics. Since no advantage or
disadvantage was quantified for a solids-contact unit, the
less expensive three step approach should be used.
A major problem with ferric chloride treatment in the solids-
contact unit during this study was due to the existence of
anaerobic reducing conditions. Under these conditions, it
was found that ferric iron was reduced to soluble ferrous
sulfide. This ferrous sulfide reported to the ferric-
primary sludge making vacuum filter dewatering virtually
impossible. Some soluble ferrous iron overflowed the chemi-
cal treatment unit and was presumably subsequently oxidized
and post-precipitated as ferric hydroxide., Reduction of
ferric to ferrous iron results in inefficient use of ferric
chloride. Pre-aeration of the raw wastewater or oxygen
addition to the solids-contact unit might eliminate the
reduction of ferric iron to ferrous iron.
An important factor in the operation of the chemical treat-
ment unit relates to sludge discharge or blowdown capacity.
During the operation of any type of clarification device,
26:
-------
there will be times when the sludge level will rise. In
order to quickly reduce the sludge level, a sludge blowdown
capacity must exist which will allow rapid removal of excess
sludge to a surge tank. A surge tank is desirable, especially
to reduce hydraulic upset of a gravity sludge thickener. It
is recommended that the surge tank have a surge capacity equal
to a volume of a three foot depth in the clarifier clarifica-
tion zone. The ability to transfer this volume in no more
than one hour is recommended.
Process Monitoring
Effective, efficient and reliable chemical treatment of waste-
water requires conscientious and consistent operator atten-
tion. Suitable tools and aids are required to provide the
operator with knowledge of system performance and input con-
ditions at any given time. In addition, collection, record-
ing and filing of these conditions will provide a historical
data base which is necessary for new operator training and
engineering evaluation of the system.
Grab and flow-proportional composite sampling facilities
should be provided to enable sampling of chemical treatment
unit feed, effluent and underflow streams. Grab sample
facilities after flash mixing and flocculation are also de-
sirable. Composite samples can be subjected to analysis
for specific pollutants to assess average system performance
over an operating period (normally one or two days). Grab
samples are required to enable assessment of instantaneous
performance or system conditions.
As a minimum, continuous monitoring of chemical treatment
clarifier effluent for clarity is suggested. The best
practical approach is to use a continuous flow through tur-
bidimeter with indication and recording of effluent turbidity.
Recording of effluent turbidity allows assessment of trends
in addition to the absolute level.
The primary treatment variable in a lime treatment system
is pH. It is recommended that continuous monitoring and
recording of pH shortly after mixing and reaction of lime
(prior to clarification) be used. An electronic signal from
the pH indicator can be integrated with one from the raw
wastewater flow meter and the result used to automatically
pace lime feed to flow and a desired treatment pH. Diffi-
culties were encountered in this study using such an auto-
mated system. These difficulties were considered to be due
to the specific instrumentation used.
263
-------
Flow and slurry concentration monitoring, indication and
recording of chemical treatment clarifier underflow should
be used. Such information will greatly assist the operator
in programming and controlling sludge withdrawal to maximize
concentration and thus minimize volume. Magnetic flow meters
and sonic sludge density devices are suggested.
Suitable facilities for routinely monitoring chemical in-
ventories and instantaneous feed rates are considered very
desirable.
Recommended Design
Table 10-1 is a tabulation of the authors' recommended de-
sign for cost effective alum treatment for Salt Lake City
or similar wastewaters.
CHEMICAL-PRIMARY SLUDGE TREATMENT
Evaluation of chemical-primary sludge treatment and disposal
alternatives and strategies was not within the scope of
this study. Thus only those sludge treatment alternatives
studied will be discussed at any length.
Sludge disposal methods such as lagooning or landfill
should be considered for locations where land area is avail-
able, and climate and other environmental conditions are
favorable. If local conditions preclude landfilling of con-
centrated sludges, incineration is the obvious alternative.
Both landfill and incineration costs will be presented in
Section XI to show economic effects. Recalcination of lime
sludge is a possibility, but at present is not considered
economically feasible for Salt Lake City or similar waste-
waters .
With the possible exception of land spreading of sludges,
all other alternatives are very cost sensitive to sludge
concentration. In general, the more concentrated the sludge
the lower the total sludge treatment and disposal costs.
Gravity thickening and vacuum filtration of chemical-primary
sludges were the only two methods evaluated to determine, in
part, the efficacy of concentrating the sludges studied.
Gravity Thickening
Operation and design of the chemical treatment unit has a
pronounced effect on gravity thickening performance. It
was shown in Section VII that, in general, the greater the
chemical treatment unit blowdown sludge concentration, the
greater would be the thickener underflow sludge concentration.
264
-------
TABLE 10-1
CHEMICAL TREATMENT3:
RECOMMENDED DESIGN PARAMETERS
Chemical;
Type Alum
Alum Dosage, mg/H as Aluminum 13
Polyelectrolyte :
Type Slightly Anionic
Dosage, mg/£ 0.25
Flash Mix:
Type Vertical Turbine
Retention Time, min 0.5
(at peak flow)
Mean Temporal Velocity 800-1000
Gradient, G, sec~l
Clarifier:
Type Flocculating Clarifier
Hydraulic Loading, 0.4
gpm/sq ft (peak)
Weir Loading, gpd/ft 10,000
Flocculation Detention Time, 15
min. (at peak flow)
Flocculating Mean Temporal Velocity 20-60
Gradient, G, sec"-*-
aOf Salt Lake City or similar wastewater
Routine use not recommended
265
-------
The chemical treatment unit blowdown sludge concentration
was effected, at least, by diurnal flow variations and by
use of polyelectrolyte.
To minimize thickener feed flow variation, it is recommended
that intermittent sludge blowdown from the clarifier be to
a surge tank. The gravity thickener would then be fed from
the surge tank by a feed pump at a near constant rate.
It is recommended that the gravity thickener be designed to
achieve maximum possible underflow concentration. However,
excessively long SRT's should be avoided. SRT's in excess
of about 3 days may result in septic conditions for alum or
FeCl3~primary sludges and could adversely effect thickener
performance and subsequent dewatering operations.
It is recommended that polyelectrolyte feed to the gravity
thickener be designed for. Polymer would be used only
when needed to maintain a maximum feed concentration prior
to vacuum filtration (if used).
Pickets on the thickener rake arms are recommended. Pickets
provide channels for water released from the thickening
sludge.
Vacuum Filtration
In general, vacuum filtration yield is directly proportional
to the feed sludge concentration. Also, cake moisture con-
tent will decrease as feed sludge concentration increases.
Thus, it was recommended in the previous subsection to de-
sign the gravity thickener to produce the maximum possible
underflow concentration.
Vacuum filtration dewatering of lime-primary sludge was
found to be easy and effective without the use of chemical
conditioning. The use of polyelectrolyte for conditioning
dilute lime-primary sludges was observed to substantially
increase filtration yield. It is, therefore, recommended
that polyelectrolyte feeding facilities be provided for
lime-primary vacuum filtration systems to provide some
operational and performance flexibilities.
About 30 to 40 percent lime {Ca(OH)2} by weight was required
to effectively dewater alum-primary sludge. Adequate vat
mixing of lime conditioned alum-primary sludge is critical
to prevent solids settling out.
As indicated previously, the presence of FeS precipitate
in the ferric-primary sludge made vacuum filtration dewater-
ing of ferric-primary sludge nearly impossible. If FeS
266
-------
formation can be avoided then vacuum filtration dewatering,
with lime conditioner, should be possible.
Recommended Design
In the design of any sludge thickening and vacuum filtra-
tion dewatering system, operational strategies for handling
various types of system upsets must be considered. The
desirability of surge capacity ahead of the gravity thick-
ener has already been discussed.
The vacuum filtration station should be designed for excess
capacity either by planning for less than continuous opera-
tion or by provision of stand-by units. Ample spare parts,
including filter media, should be provided.
In general, the authors recommend designing for filtration
rates significantly lower than average achievable rates.
The filter station must be able to process all sludge pro-
duced, in a reasonable time frame and for conditions of
poorest predictable dewatering properties.
Table 10-2 is a tabulation of recommended design criteria for
gravity thickening and vacuum filtration of alum-primary
sludge for Salt Lake City or a similar wastewater.
TABLE 10-2
ALUM-PRIMARY SLUDGE:
DESIGN RECOMMENDATION FOR SLUDGE DISPOSAL
Gravity Thickener
Raking constant speed with
pickets
Solids Loadings, Ib/day/sq ft 10
Predicted Underflow Solids, g/l 50
Vacuum Filter
Type standard belt filter
with flocculator
Media Type Nylon (NY 317-F)C
Maximum Lime Feeding Capacity,
% by wgt 50
Yield, Ib/hr/sq ft b 2
Cake Moisture Content, % by wgt 75
Based on alum and primary sludge solids only
Based on total cake solids
°Eimco BSP Division of Envirotech Designation
267
-------
CARBON TREATMENT
Carbon Contactor
One of the most important design requirements for a carbon
treatment system is to establish the PAC dosage necessary
to achieve a given effluent quality. Since adsorption,
biological and coagulation removals all occurred in the
PAC-PCT process studied, there is no obvious way to estab-
lish a carbon dosage. The reponse of other wastewaters to
PAC-PCT treatment can only be conjectured. Therefore, all
that will be done in the next section is to show the relative
economics of using various carbon dosages.
Since pilot plant results indicated that as little as 26
percent of SCOD removed across the entire plant was removed
by adsorption, it would appear that a lower grade, and/or
lower cost carbon than used in this study could be used.
About 43 percent of the removal was by a biological mech-
anism in the carbon contactors.
All of the results reported thus far for these pilot plant
studies were for a relatively weak municipal wastewater.
The response that may be obtained from other stronger waste-
waters might be quite different than those observed during
this study- It should be emphasized that long term field
data (pilot plant and/or full scale) is necessary to identify
all of the soluble organic removal mechanisms operative in
a PAC-PCT system.
The effect of counter-current staging is still not well de-
fined. It was suggested in a previous report that two-
stage counter-current contacting enhanced stable plant
operation.1 The second stage adds liquid-solids separa-
tion reliability (at least redundancy) and longer SRT's.
Two stages are more costly and, of course, involve twice
the equipment to operate and maintain. When a low level of
adsorptive removal is sufficient, a single-stage contactor
would be the less expensive approach.
If the major SCOD removal mechanism were adsorption, it is
probable that two-stage counter-current contacting of the
adsorbent would be the most economical approach.
The solids-contact treatment units used in this study were
judged to be excellent powdered carbon contactors. High
concentrations and large volumes of carbon slurries were
kept in contact with the wastewater. The units acted, in
part, as suspended growth anaerobic bio-reactors.
263
-------
Because of the very small particle size (95 percent smaller
than 325 U.S. Mesh) and density (1.52 g/cc) of powdered
carbon, requirements for low hydraulic loadings would be
expected. Powdered carbon has been shown to be self floccu-
lating at slurry concentrations above 1000 mg/£. Peak over-
flow rates of 0.7 - 0.9 gpm/sq ft appear to be reasonable
based on the results of this study.
Once through flash mixing, flocculation (contact time) and
sedimentation systems have been used for laboratory and
pilot plant scale PAC studies. These systems did not in-
clude PAC recycle and therefore did not experience signifi-
cant biological activity and thus removal of some SCOD by
this mechanism. Usually considerable amounts of coagulating/
flocculating chemical additives are required for the once
through approach.
The authors consider the use of solids-contacting type of
units for PAC contacting/clarification as the method of
choice.
Biological removal of SCOD, in the solids-contact units, was
accomplished predicated by maintaining powdered carbon SRT's
in the order of several days. Maintaining SRT value in excess
of more than 3 days during warm summer months and about 7
days during cold winter months was undesirable due to the
production of offensive gaseous odors such as E^S. The need
for exercising control of SRT may require deviation from use
of standard solids-contactor unit sizes. Specifically the
volume of the reaction zone (i.e., size of mechanism) would
be based on the desirable carbon SRT for a given application.
It would, of course, not be considered desirable to reduce
the nominal carbon-influent wastewater flow contact time to
less than about 10 to 15 minutes. Otherwise adsorptive
equilibriums would not be closely approached.
There were numerous full scale engineering design considera-
tions not evaluated in the current study. For example, re-
quirements for virgin carbon make-up, handling and feeding
systems, materials of construction, mixers, pumps, etc.,
were not systematically evaluated. It is suggested that
carbon make-up, handling and feeding technology implemented
by large potable water treatment operations using PAC for
taste and odor control may be directly applicable to waste-
water applications. The same suggestion may be valid for
mixers, pumps, etc., also.
It can be noted that after some two years of use the mild
steel solids-contact treatment units showed no obvious signs
of erosion and/or corrosion. The surfaces of these units
were originally prepared by sand blasting and painting with
269
-------
one coat of a phenolic fortified alkyd primer and then two
coats of coal tar epoxy paint, eight mils dry thickness each.
Granular Media Filtration
Filtration is the final liquid-solids separation step in the
PAC-PCT system studied. Design for achievment of no more
than 5 mg/£ average suspended solids is recommended. This
will allow minimizing loss of expensive carbon and assure a
clean, esthetically pleasing system effluent.
To produce a low suspended solids effluent consistently and
reliably, conservative hydraulic loading is recommended.
Peak filtration rates should be no more than 3 to 4 gpm/sq ft.
The suspended solids in the filter feed were observed to pro-
duce a "weak floe" whe.n removed by filtration. Therefore
sudden and/or large changes in filtration rate should be
avoided to prevent breakthrough of suspended solids into the
filter effluent. In addition the limitation of terminal head-
loss to no more than 8 ft of H20 would be recommended.
Use of anionic polyelectrolyte, added by in-line mixing just
ahead of the filter inlet, was found to result in a slightly
"tougher11 floe which was less susceptible to premature break-
through. It is, therefore, recommended that provisions be
made for feeding polyelectrolyte when rapidly varying or
high flow rate conditions are anticipated.
Considerable experience in the area of designing and applying
granular media filters in wastewater applications have con-
vinced the authors of the efficacy of using auxiliary air-
scour, prior to water backwashing. Achievement of effective
backwashing is a prerequisite to any successful filter opera-
tion and performance especially in wastewater applications.
In a previous report it was recommended that filter cycle
times in excess of about 24 hours should be avoided.l The
purpose was to minimize potential "mud ball" formations which
might not be destroyed by normal backwashing procedures. This
recommendation is still valid.
Backwash water should be collected in a mixed surge tank and
then slowly recycled back to the first stage of the carbon
contacting system. Recycling to the carbon system is neces-
sary to prevent losses of expensive carbon, but can cause a
build-up of carbon and inert fines. Inert fines (inorganic)
would be produced by carbon regeneration operations. These
carbon and inert fines must eventually be bled out of the
270
-------
closed circuit carbon treatment system. A suggested approach
is to periodically recycle backwash water to the chemical treat-
ment clarifier feed stream.
Recommendations for filter bed media and backwash procedures
will be presented in a subsequent subsection.
Process Monitoring
The solids-contact units recommended for PAC contacting-
clarification require the same operational attention and
approach as previously suggested for the chemical treatment
unit.
The suggestion for continuous real time monitoring of carbon
contactor effluent clarity needs to be qualified. The use
of turbidimetric devices utilizing white light scattering
and/or transmittance principals of operation are not applic-
able to PAC slurry concentrations of more than about 20 to
30 mg/£ suspended solids. Of the two alternate principals
of operation the light transmission devices are preferred.
Even then some sort of dilution of the clarifier effluent
sample stream with suspended solids free water may be re-
quired in order to provide a reasonable degree of precision.
The efficacy of the UV monitoring of carbon system feed and
effluent for determination of soluble organics was discussed
in Section IX. As indicated there, the use of UV monitoring
was not considered as a useful operational tool for the PAC
system studied.
Continuous sensing and recording of carbon system effluent
UV absorption can provide a historical system performance
record. An indication of trends in increasing or decreasing
UV absorption could be useful operational intelligence in
assessing daily or weekly variations in carbon system per-
formance.
The most important operational strategy concerning the
carbon system relates to maintenance of some level of bio-
logical activity; but not a level which will result in pro-
duction of undesirable odors.
If lime is used in the chemical treatment step, close mon-
itoring and control of carbon system feed pH is desirable.
Wide variations in pH can significantly reduce biological
activity for several hours to several days, depending on the
severity and duration of a pH variation.
271
-------
Maintenance of a well mixed (i.e., uniform suspension) carbon
slurry in the carbon contactors is considered desirable, even
though it may result in relatively dilute underflow concentra-
tions. Maintaining a well mixed slurry is considered con-
ducive to enhancement of biological activity.
The basic approach recommended for controlling biological
activity in the carbon contactors is the same as used in the
rational operation and control of the activated sludge pro-
cess, i.e., control of SRT. The basic assumption required
is that the active bio-mass present is directly related to
the total suspended solids (the same as for activated sludge
systems).
Control of SRT necessitates routine knowledge of the inven-
tory of solids in the carbon system and the flux of carbon
through the system. The flux of carbon (Ib/day) can be
determined from the rate of removal of carbon solids in the
contactor-clarifier underflow or the feed rate of carbon to
the system. The inventory of carbon (Ibs) is determined by
suspended solids profiles and slurry volume measurements.
The carbon inventory value (Ibs) divided by the carbon flux
rate (Ib/day) indicates the carbon system SRT.
During this study operating charts were posted relating carbon
system inventory to reaction zone suspended solids and carbon
slurry depth (i.e., sample tap). Also posted was a chart
relating carbon flux rate to feed carbon dosage and system
flow rate. Use of these charts and routinely collected
operational data allowed essentially instantaneous determina-
tion of an approximate carbon system SRT.
It is anticipated that the granular media filter station
will be designed for automatic backwashing. Initiation of
backwashing will be based on one of the following criteria:
1. Effluent turbidity exceeds a predetermined level.
2. The filter pressure drop reaches the maximum
available.
3. The length of filter run reaches a predetermined
value.
4. Operating personnel desire to manually initiate a
backwash cycle.
The first three criteria can be automatically evaluated by
process evaluation instrumentation. Effluent clarity can
effectively be determined by continuous flow through process
turbidimeters. The measured turbidity should be indicated
and recorded. The pressure drop across the filter media can
also be automatically measured by differential pressure cells
and the resulting measurement should also be indicated and
recorded.
272
-------
It is also recommended that facilities be installed to, when
desired, determine pressure profiles (i.e., pressure versus
bed depth) of clogged filters. If excessive penetration of
solids into the filter bed is observed, the operator can be
alerted to the need to start feeding or increase the feed rate
of polyelectrolyte.
Frequent monitoring and assessing the effectiveness of back-
washing is recommended. Backwash rate may have to be adjusted
due to seasonal variations in water temperature to insure
achievement of a minimum desirable filter bed expansion which
is recommended to be 20 percent. The frequency of backwashing
(i.e., cycle time) can be very simply monitored by installa-
tion of a counting device which is actuated by a valve opera-
tor. Routine monitoring of clean bed headloss (immediately
after backwashing) is desirable to determine any long term
trends of increasing headloss. This observation would pos-
sibly indicate inadequate backwashing or accumulation of coal
fines at the surface of the coal layer.
Recommended Design
Table 10-3 presents key design recommendations for a carbon
contacting-clarification and granular media filtration
system for treating Salt Lake City or similar wastewater.
TABLE 10-3
CARBON TREATMENT:
PROCESS DESIGN RECOMMENDATION
Carbon Contactor-Clarifier
Type Solids contact with in-
ternal solids recycle
Peak Hydraulic Loading, gpm/sq ft 0.8
Slurry Concentration, g/£ 10
Predicted Underflow Concentration, g/& 30
Approximate Carbon SRT, days 3-7
Minimum Reaction Zone Hydraulic 15
Retention Time, Minutes
Granular Media Filtration
Peak Hydraulic Loading, gpm/sq ft 3-4
Average Backwash Recycle, % of 3
Filtrate Flow
Backwash Procedure:
Air-Scour Rate, SCFM/sq ft 3-5
Air-Scour Duration, minutes 2-10
273
-------
Table 10-3 (continued)
Backwash Procedures (continued):
Water Volume, gal/sq ft 70-90
Bed Expansion, % of Static Bed Depth >20
Filter Media, Size Range, mm
Coal 1.5-2.7
Sand 0.75-0.84
Garnet 0.25-0..50
The filter media sizes and types indicated in Table 10-3
should not be considered absolute. They simply indicate a
design found in this study to produce acceptable performance.
CARBON REGENERATION
Consideration of treatment and cost effectiveness of alterna-
tive methods of powdered carbon regeneration is not within
the scope of this study. Therefore, only recommendations
relating to the use of a fluid bed furnace approach will be
presented.
Due to limited knowledge of the authors in the area of
fluidized bed furnace technology, it would be presumptuous
to make any recommendations without qualification. Certain
relevant comments and observations are,, however, in order.
Thickening and Dewatering
It was shown in Section VII that the vacuum filtration yield
(Ib of dry solids/hr sq ft) for spent carbon slurries was
directly proportional to feed solids concentration. It is
therefore recommended that spent carbon slurries be gravity
thickened to the maximum possible concentration prior to
vacuum filtration dewatering. Though a rigorous cost analy-
sis was not made, the authors conjecture that one would show
the above recommendation to be the most cost effective.
During this study the spent carbon gravity thickeners were
located in an enclosed building. The units were routinely
underloaded and were often used to accumulate and store spent
carbon prior to a regeneration run. No significant odor
problems were experienced. When thickener underflow was
stored in an inventory tank for up to 2 to 3 weeks and then
vigorously mixed, considerable sulfide odors were observed.
It would be expected that a full scale operation would not
necessitate storage of spent carbon for periods of 2 to 3
weeks and thus odors would not be an operational problem.
274
-------
As noted in Section VII, carbon sludge vacuum filter yields
were lower when using alum or ferric chloride pretreatment
than when using lime pretreatment. During the current study
only limited closed circuit operation of the carbon system was
achieved. It is unlikely that a steady state solids and
carbon condition was reached. It is therefore suggested that
a conservative interpretation of carbon dewatering performance
results can be made.
Poor solids capture was often experienced during this study,
especially when polyelectrolyte was not used. Therefore,
routine use of a cationic polyelectrolyte is therefore
recommended in order to maximize solids capture.
It was found that the spent carbon filter cakes cracked
excessively during the dry cycle. The vacuum filter used
had a single form and dry vacuum receiver resulting in re-
duced operating vacuum levels. It is strongly recommended
that separate form and dry vacuum systems be used to allow
maximum vacuum levels for cake formation and maximum filter
yields.
During this study transport of filter cake from the vacuum
filter to the fluid bed furnace was accomplished by collec-
tion in a hopper, reconstitution as a paste with a ribbon
screw and transfer in a pipeline via a Moyno type pump. This
transport system worked well normally, but on occasion bridg-
ing of filter cake in the hopper throat (6 inches wide by 4
feet long) was experienced. Reconstitution of cake to a
paste and maintenance of at least 12 ft/min pipe velocity
was considered necessary to effectively transfer cake.
Fluid Bed Furnace
Results from this study indicate that high recovery of fixed
carbon is possible with the fluidized bed furnace used. An
average recovery of just over ninety percent was experienced
(range from 76 to 100 percent). Regenerated carbon had
slightly reduced equilibrium adsorption capacity for iodine,
molasses and waste soluble COD, when compared to virgin
carbon.
In general it was concluded that the fluid bed furnace system
evaluated was capable of effectively regenerating powdered
carbon spent in the PAC-PCT process studied. There is no
basis for concluding that regeneration effectiveness was
maximized or that regeneration costs were minimized in this
study. Additional study is required to accomplish these two
objectives.
275
-------
It was found in this study that positive control of oxygen
in the fluid bed was necessary to achieve successful perfor-
mance. The approach used during this study to monitor oxygen
was considered poor and it is recommended that a more highly
instrumented and automated approach previously used by others
be considered.2
Based on the results of this study use of the bed-injection-
gas principal of operation cannot be recommended. The poten-
tial cost benefit associated with use of this operating
principal suggests, however, that additional development
studies may be warranted.
In recommending the off-gas-recycle principal of operation
for a fluid bed furnace approach to powdered carbon regenera-
tion, it is recommended that the following design objectives
be strived for:
1. Maximize fire box operating temperature consistent
with available materials.
2. Minimize fluid bed operating temperature consistent
with achievement of an acceptable level of regnera-
tion effectiveness.
3. Maximize the quality of bed fluidization (e.g.,
maximum numer of small orifices in tuyeres; maxi-
mum economical bed depth to diameter, etc.).
During this study it was found desirable to remove furnace
sand and other foreign material from the recovered carbon by
cyclone 'and screening operations. These steps are also
recommended for a full-scale system. Doing so will help pre-
vent plugging of pipes and pumps and minimize the build up
of inerts within the closed-loop carbon regeneration and
reuse system.
Inorganic precipitates which enter the carbon contacting
system via the effluent from the chemical treatment clarifier
will report to the spent carbon slurry (i.e., regeneration
system feed). Most of these materials will be converted to
metal oxides in the furnace operation and thereafter remain
with the regenerated reused carbon in the same form. Build
up of these materials will ultimately necessitate blowdown
from the closed-loop system.
Blowdown of accumulated inorganic ash can be accomplished by
wasting along with associated carbon solids. An alternative
would be to acid treat recovered regenerated carbon and
recycle the leachate to the chemical treatment step.
276
-------
Design Recommendations
Basically the carbon regeneration system should be designed
for continuous operation. Recognizing the need for down
time for equipment repairs and maintenance it is recommended
that about 15 percent excess capacity be designed into the
furnace and about 25 percent excess capacity into the vacuum
filter.
Based on the results of this study Table 10-4 presents
recommended process design criteria for the powdered carbon
regeneration system major components.
277
-------
TABLE 10-4
CARBON REGENERATION SYSTEM:
RECOMMENDED PROCESS DESIGN
Gravity Thickener
Solids Loadings, Ib/day/sq ft
Predicted Underflow Concentration, g/£
Vacuum Filter
Type
Drum Submergence, %
Media Type
Vacuum Level*, in. of Hg
Airflow Rate, ACFM/sq ft
Polyelectrolyte:
Type
Dosage, Ib/TDS
Yield, Ib/hr/sq ft
Cake Solids, %
Filtrate Solids, g/£
Fluidized Bed Furnace (Off Gas Recycle)
Solids Loadings, Ib/hr/sq ft
Free Board Velocity, ft/sec
Firebox Temperature, °F
Operating Temperature, °F
20
120
Standard Belt Filter
33
Polypropylene
20
Cationic
5-10
5
27
3
1.2
2000
1250
*Use separate form and dry vacuum receivers
273
-------
SECTION XI
ECONOMIC ANALYSIS
The following economic analysis is presented for the purpose
of showing the estimated costs of a PAC-PCT system for treat-
ing a municipal wastewater similar to Salt Lake City, Utah.
To enable assessment of economy of scale, estimated costs
will be presented for average design flows of 5, 10 and 50 mgd.
Peak to average design flows of 1.5 to 1.0 were assumed.
The cost estimates are presented for the chemical (i.e., alum)
treatment step, the carbon treatment step and then chemical
and carbon treatment. Doing so enables identification of
the costs for removal of suspended solids and phosphorus by
chemical treatment separately from the cost for removal of
soluble organics (BOD5 and SCOD).
Table 11-1 presents the assumed pertinent raw wastewater and
system effluent water qualities. As footnoted in Table 11-1
the actual effluent organics to be expected from a PAC-PCT
system will depend on specific raw wastewater soluble organics
and their amenability to such a treatment process.
TABLE 11-1
ASSUMED AVERAGE RAW WASTEWATER AND
EFFLUENT WATER QUALITIES
Parameter, Raw
Units Wastewater
BOD 5, mg/£
Suspended Solids, mg/£
Turbidity, JTU
Total COD, mg/£
Soluble COD, mg/£
Phosphorus, mg/£ as P
125
150
50
175
45
6
Chemical
Treatment
Effluent
25
25
10
40
25
1
Plant
Effluent
5a
5
3
ioa
6a
0.5
aWill depend on specific raw wastewater soluble organics
279
-------
Major process design parameters and equipment type and sizing
were based on those presented and discussed in Section X.
Assumed unit costs for chemicals and operating costs are
presented in Table 11-2.
TABLE 11-2
ASSUMED UNIT COST FOR
ECONOMIC ANALYSIS
Chemicals - (as of January, 1974; FOB Salt Lake City)
Alum 3.5C/lb
Lime l.OC/lb
Polyelectrolyte $1.50/lb
Powdered Carbon lO.OC/lb
Capital Costs
All equipment costs were estimated as installed equipment
by the authors' firm and amortized at 6 percent for 25
years.
Operating Costs - (as of January, 1974)
Power $4.00/MG
Ash and Sludge Cake Haulage $5.00/ton
Maintenance 1% of capital costs/year
CHEMICAL TREATMENT COSTS
Equipment design parameters and assumed operating schedules
used to size equipment for the chemical treatment step are
shown in Table 11-3. It is apparent that at least two
(parallel) flocculating clarifiers are used for all plant
sizes. For the 5 and 10 mgd plant flow sizes the solids
handling systems have only a single train of equipment. It
should also be noted that different solids handling system
operating schedules were assumed for the three plant flows.
The approach invoked here was that larger plants with larger
investments in capital equipment and facilities should strive
for approaching near continuous solids handling operating to
minimize total treatment costs.
Table 11-4 presents estimated alum treatment costs. Pre-
treatment costs shown include bar screening, aerated grit
removal, comminution and flow measurement lumped together.
Raw wastewater pumping was not assumed to be required.
280
-------
CO
TABLE 11-3
DESIGN PARAMETERS AND MAJOR EQUIPMENT
SIZING FOR CHEMICAL TREATMENT
Item
Alum Dosage, mg/£
{A12 (S04)3-17H20>
Alum Flash Mixing:
Retention Time, minutes @ peak flow
Mixer :
Type
Horsepower
Tank
Number
Diameter x depth, ft
Flocculating Clarifiers:
Peak Hydraulic Loadings, gpm/sq ft
Number of Units
Diameter of Units, ft
Flocculator Horsepower
Gravity Sludge Thickener:
Solids Loading, Ib dry solids/day-sq ft
Number of Units
Diameter x depth, ft
P
5
125
h
- All
5
1
7x9
0.4
2
100
2
10
1
30 x 10
LANT FLOW,
10
125
%
Vertical Radial Flow
7.5
1
10 x 10
0.4
2
140
3
10
1
45 x 10
mgd
50
125
h
Turbine -
40
2
13 x 13
0.4
4
215
5
10
2
68 x 10
-------
Table 11-3 (continued)
PLANT
Item
FLOW,
10
mgd
50
NJ
CO
to
Vacuum Filtration:
(Belt Filter)
Lime Dosage, percent by weight
Yield (including chemical), Ib dry
solids/hr-sq ft
Operating Schedule:
Days/Week
Hours/Day
Number of Units
Diameter x face width,
ft
Operating Schedule:
Days/Week
Hours/Day
Number of Units
Diameter of Units, ft
40
2.8
5
8
1
12 x 18
Sludge Incineration:
Type of Unit
Solids Loading, Ib dry solids/hr-sq ft 8.1
Feed Moisture Content, percent by weight 75
BTU Values, BTU/lb dry solids 3300
5
8
1
16.75
40
2.8
5
16
1
12 x 18
40
2.8
7
20
3
12 x 16
Multiple Hearth Furniture -
8.1 13
75 75
3300 3300
5
16
1
16.75
7
20
1
22.25
-------
TABLE 11-4
ESTIMATED ALUM TREATMENT COSTS
Capital Costs
(1000 of $)
Flow, mgd
Pre treatment
Chemical Contact
Gravity Thickening
Vacuum Filters
Incineration
TOTAL
Contingency (50%)
Total Capital Cost
Amortized Cost,
C/1000 gal
(25 years @ 6%)
Operating Costs
(C/1000 gal)
Treatment Chemicals
Sludge Conditioning
Incineration (Fuel,
Power, Labor)
Power
Maintenance, @ 1%
Cap/Year
Supervision & Labor
Ash or Sludge Cake
Haulage
Total C/1000 9al
Operation
Total C/1000 gal
Alum Treat-
ment with
Incineration
5
86
748
67
527
825
2253
1127
3380
14.5
3.7
0.6
2.7
0.2
1.7
3.0
0.2
12.1
26.6
10
122
1065
78
549
825
2639
1320
3959
8.5
3.7
0.6
2.1
0.2
1.0
2.5
0.2
10.3
18.8
50
385
3220
184
1368
1250
6407
3204
9611
4.1
3.7
0.6
1.3
0.2
0.4
0.8
0.2
7.2
11.3
Alum Treat-
ment without
Incineration
5
86
748
67
527
-
1428
714
2142
9.2
3.7
0.6
-
0.2
1.1
3.0
2.2
10.8
20.0
10
122
1065
78
549
-
1814
907
2721
5.8
3.7
0.6
-
0.2
0.7
2.5
2.2
9.9
15.7
283
-------
The contigency cost shown in Table 11-4 under "capital costs"
is an estimate of costs for engineering, administration and
legal in addition to interconnecting piping, valves, instru-
ments and buildings.
The effect of economy of scale and increased operating
schedules on capital costs for larger plant sizes is apparent
when comparing the estimated costs for 5 and 50 mgd plant
sizes. For the 5 mgd plant amortized capital costs are 55
percent of the total costs. For the 50 mgd plant size amor-
tized capital costs are reduced to 37 percent of total costs.
Comparison of total costs for the 10 and 5 mgd plant sizes
indicate that, respectively, 20 and 33 percent increases in
costs are indicated for including sludge incineration.
Analysis of the cost data shown indicates that the vast
majority of increased cost for incineration is due to capital
costs requirements.
CARBON TREATMENT COSTS
Equipment design parameters and assumed operating schedules
used to size equipment for carbon treatment and regeneration
are shown in Table 11-5. It should be noted that the equip-
ment sizing (i.e., loading) parameters indicated are maximum
desired values. The actual operating schedules indicated
results in actual loadings less than the values shown, in
some cases.
Table 11-6 presents estimated carbon treatment costs, all
with and at 5 and 10 mgd without regeneration. The capital
costs shown for the "without regeneration" case could be
reduced slightly if spent carbon were wasted to the chemical-
primary solids handling system, rather than handled separately
as shown in Table 11-6.
Analysis of the estimated costs in Table 11-6 shows that for
a carbon dosage of 100 mg/£ it is cost effective to use
carbon once and throw it away for plant flows less than 5
mgd. For a carbon dosage of 300 mg/£ it is cost effective
to use carbon once and throw it away for plant sizes less
than about 3 mgd.
Because of the equipment intensive carbon regeneration system
there is a substantially greater economy of scale for the
case where regeneration and reuse is employed than for when
it is not.
284
-------
TABLE 11-5
DESIGN PARAMETERS AND MAJOR EQUIPMENT
SIZING FOR POWDERED CARBON TREATMENT
PLANT
Item
FLOW,
10
mgd
50
CO
Ul
Contacting Clarification
Solids Contact Units:
Number of Stages
Peak Hydraulic Loading, gpm/sq ft
Number of Units
Diameter of Units, ft
Granular Media Filters:
Peak Hydraulic Loading, gpm/sq ft
Number of Units
Diameter of Units, ft
Backwash Collection Tank:
Diameter x Depth, ft
Mixer Type
Mixer Horsepower
PAC Regeneration
PAC Dosage
Gravity Thickener;
Solids Loadings, Ib/dry solids/
day-sq ft
Number of Units
Diameter, ft
Side Water Depth, ft
2
0.8
2
96
3
2
34
2
0.8
2
135
3
3
38
2
0.8
4
210
3
14
40
20 x 12 20 x 14 20 x 16
- - - Vertical Radial Flow Turbine - - -
2.5 2.5 2.5
100
20
1
18
8
300
20
1
28
8
100 300
20 20
100 300
1
22
8
1
40
8
20
1
50
9
20
1
90
14
-------
Table 11-5 (continued)
Item
100
300
100
300
100
300
M
CO
Vacuum Filtration:
(Belt Filter)
Cationic Polyelectrolyte
Dosage, Ib/TDS
Yield, Ib dry solids
Operating Schedyle:
Days/Week
Hours/Day
Number of Units
Size of Units:
Diameter, ft
Face Width, ft
PAC Thermal Regeneration:
(Fluid Bed Furnace)
Solids Loading, Ib dry solids/
hr-sq ft
Feed Moisture, percent by wgt
Bed Temperature, °F
Fire Box Temperature, F
Spacial Velocity, ft/sec
Number of Units
Diameter of Units, ft
5
6
5
8
6
8
5
6
5
8
10
12
5
6
5
20
6
6
5
6
5
20
8
14
5
6
7
24
8
14
5
6
7
22
10
16
77
1500
2000
1.2
1
20
77
1500
2000
1.2
2
20
77
1500
2000
1.2
1
20
77
1500
2000
1.2
2
20
77
1500
2000
1.2
2
20
77
1500
2000
1.2
7
20
-------
TABLE 11-6
ESTIMATED CARBON TREATMENT COSTS
-J
Regeneration and Reuse
Plant Flow, MGD
Carbon Dosage, mg/£
Capital Costs (1000 of $)
100
Carbon Treatment
Granular Media Filtration
Gravity Thickening 55
Vacuum Filtration 195
Thermal Regeneration 800
Total
Contingency (50%)
Capital Costs
Total
Amortized Cost ,
(C/1000 gal)
2047
1023
3076
13.2
Operating Costs (C/1000 gal)
Carbon Makeup
Regeneration (Fuel,
Power, Labor)
Power
Supervision & Labor
Maintenance^ 1%/Cap/Yr
Total £/1000 gal
Operation
Total Treatment Costs
(C/1000 gal)
1.5
0.2
3. 0
1.8
7.3
20.5
5
300
668
329
74
330
1600
3001
1501
4502
19. 3
b 2.5b
3.4
0.2
3.0
2.6
11.7
31.0
100
60
170
800
2607
1304
3911
8.4
0.8
1.5
0.2
2.5
1.1
6.1
14.5
YES
10
300
997
580
97
270
1600
3544
1772
5316
11.4
b 2.5b
3.4
0.2
2.5
1.5
10.1
21.5
NO
100
115
270
1600
7886
3943
50
300
2985
2916
147
700
5600
12,384
6174
11,829 18,558
5.1 8.0
0.8
0.9
0.2
0.8
0.5
3.2
8.3
b 2.5b
2.4
0.2
0.8
0.9
6.8
14.8
100
55
195
1247
624
1881
8.0
5
300
668
329
74
330
1401
701
2102
9.0
8.0° 24.0°
0.2 0.2
3.0 3.0
1.1 1.2
12.3
20.3
28.4
37.4
100
60
170
1807
904
2711
5.8
8.0
0.2
2.5
0.8
11.5
17.3
10
300
997
580
97
270
1944
972
2916
6. 3
C 24. 0(
0.2
2.5
0.9
27.6
33.9
25 years @ 6%
10% Makeup
'100% Makeup
-------
TOTAL PAC-PCT TREATMENT COSTS
Table 11-7 and Figure 11-1 show the estimated costs for a
PAC-PCT system which includes sludge incineration and powdered
carbon thermal regeneration and reuse. As indicated in Table
11-1 such a system is suggested to be capable of producing an
effluent quality of comparable 6005 and phosphorus concentra-
tions and lower in COD concentration than a conventional
biological treatment system followed by chemical treatment
for phosphorus removal and granular media filtration for sus-
pended solids removal. As previously noted the efficacy of
the PAC-PCT system treatment effectiveness is dependent on
the specific raw wastewater soluble organic characteristics
(i.e., their amenability to removal by carbon treatment).
Evaluation of the estimated costs for increasing plant sizes
in Table 11-7 and Figure 11-1 indicates that similar economy
of scale exists for the chemical and carbon treatment systems.
This observation is presumably due to the similarity in wet
end and sludge and regeneration equipment schemes.
Figure 11-1 shows that above a plant size of about 20 mgd
only a nominal economy of scale is realized.
Based on the economic analysis presented it is the authors'
judgement that the PAC-PCT process evaluated represents a
potentially cost effective alternative for treatment of
certain municipal wastewaters.
288
-------
TABLE 11-7
SUMMARY OF PHYSICAL-CHEMICAL TREATMENT COSTS -
WITH ALUM TREATMENT, SLUDGE INCINERATION AND CARBON REGENERATION
CO
Plant Flow, MGD
Chemical Treatment Costs
Capital Costs - Amortized
25 years @ 6% (0/1000 gal)
Operating Cost (0/1000 gal)
Total 0/1000 gal
Carbon Dose, mg/£
Carbon Treatment Costs
Capital Costs - Amortized
25 years @ 6% (0/1000 gal)
Operating Cost (0/1000 gal)
Total 0/1000 gal
Chemical and Carbon Treatment
Costs
Capital Cost - Amortized
25 years @ 6% (0/1000 gal)
Operating Cost (0/1000 gal)
Total 0/1000 gal
5
14.5
12.1
26.6
100 300
13.2 19.3
7.3 11.7
20.5 31.0
27.7 33.8
19.4 23.8
47.1 57.6
10
8.5
10.3
18.3
100 300
8.4 11.4
6.1 10.1
14.5 21.5
16.9 19.9
16.4 20.4
33.3 40.3
50
4.1
7.2
11.3
100 300
5.1 8.0
3.2 6.8
8.3 14.8
9.2 12.1
10.4 14.0
19.6 26.1
-------
FIGURE 11-1:
ESTIMATED TREATMENT COSTS VERSUS PLANT
SIZE - WITH SLUDGE INCINERATION AND
CARBON REGENERATION
60 t-
50
H
rt
ho40
o
o
o
•O-
0
o
20
10
Chemical + Carbon Treatment (100 rng/£)
Chemical Treatment
— , 0-
I
10 20 30 40
Plant Design Flow, mgd
50
290
-------
SECTION XII
REFERENCES
1. Burns, D.E. and Shell, G.L., "Physical-Chemical Treat-
ment of A Municipal Wastewater Using Powdered Carbon"
Environmental Protection Technology Series, EPA-R2-73-
264 (August, 1973).
2. Shuckrow, A.J., Dawson, G.W. and Bonner, W.F., "Physical-
Chemical Treatment of Combined and Municipal Sewage"
Environmental Protection Technology Series, EPA-R2-73-
149 (February, 1973).
3. Berg, E.L., Villers, R.V., Masse, A.N. and Winslow, L.A.,
"Thermal Regeneration of Spent Powdered Carbon Using
Fluidized Bed and Transport Reactors" Chemical Engineer-
ing Progress, Vol. 67, pp. 154-164, (1970).
4. Shuckrow, A.J., Bonner, W.F., Presecan, W.L. and
Kazmierczak, E.J., "A Pilot Study of Physical-Chemical
Treatment of the Raw Wastewater at the Westerly Plant
in Cleveland, Ohio" presented at the International
Association on Water Pollution Research Workshop,
Vienna, Austria (September, 1971).
5. O'Brien and Gere Engineers, Inc., "Town of Clay Chemical-
Physical Investigation" private communication, Syracuse,
New York (January, 1971).
6. Gulp, R.L. and Gulp, G.L., Advanced Wastewater Treatment,
Van Nostrand Reinhold Company, New York, New York (1971).
7. Burns, D.E.., Inter-Company Communication, March 6, 1972.
8. Talmage, W.P. and Fitch, E.B., "Determining Thickener
Unit Area" Industrial and Engineering Chemistry, Vol.
47, No. 1, pp. 38-41 (January, 1955).
291
-------
9. Knopp, P.V., and Gitchel, W.B., "Wastewater Treatment with
Powdered Activated Carbon Regeneration by Wet Air Oxida-
tion" presented at the 25th Purdue Industrial Waste
Conference (May, 1970).
10. Perry, R.H., Chemical Engineer's Handbook, McGraw-Hill,
Inc., (1963).
11. Dayio, K., and Lenenspiel, 0., Fluidization Engineering,
John Wiley and Sons, Inc., New York (1969).
12. Beebe, R.L., "Activated Carbon Treatment of Raw Sewage
in Solids-Contact Clarifiers" Environmental Protection
Technology Series, EPA-R2-73-183 (March, 1973).
13. Davis, D.S., and Kaplan, R.A., "Removal of Refractory
Organics from Wastewater with Powdered Activated Carbon,"
Journal WPCF, Vol. 35, No. 3, p. 442 (1966).
14. Dobbs, R.A., Wise, R.H., and Dean, R.B., "The Use of
Ultraviolet Absorbance for Monitoring the Total Organic
Carbon Content of Water and Wastewater," U.S. Department
of the Interior, FWPCA Advanced Waste Treatment Research
Laboratory Report, Water Research, Vol. 6, p. 1173 (1972).
15. Standard Methods, 13th Edition, American Public Health
Association Inc., Washington, D.C. (1971).
16. ASTM Standards, Part 9, American Society for Testing
and Materials, Philadelphia, Pa. (1964).
17. Kynch, G.J., "A Theory of Sedimentation," Trans. Faraday
Soc. 48, 166, (1952).
292
-------
SECTION XIII
GLOSSARY
Abbreviations:
ABS alkyl benzene sulfonate
BIG bed injection gas
BOD5 five-day, 20°C-biochemical
oxygen demand
cc cubic centimeter
COD chemical oxygen demand
CT GMF cycle time
cu ft cubic feet
DC direct current
EPA Environmental Protection
Agency
FBF fluidized bed furnace
FFR form filtration rate
FT vacuum filter cake form time
ft feet
fpm feet per minute
fps feet per second
GAC granular activated carbon
gal gallon
G mean temporal velocity
gradient
GMF granular media filter
gpd gallons per day
gpm gallons per minute
Hp horsepower
hr hour
in. inch
JTU Jackson turbidity units
kg kilogram
£ liter
lb pound
log logarithm
M meter
Mc cake moisture content
mgd million gallons per day
mg milligram
MG million gallons
293
-------
multiple hearth furnace
min minute
m]_ milliliter
mm millimeter
N normality
nm nanometer
OGR °ff 9as recycle
AP incremental pressure
PAC powdered activated carbon
PCT Physical-Chemical Treatment
pH negative logarithm of the
hydrogen ion concentration
PM post meridian
PVC polyvinyl chloride
p/I pressure indicator
rpm revolutions per minute
RZ reaction zone
SCOD soluble chemical oxygen
demand
SCFM standard cubic feet per
minute (@ 20°C, sea level)
sec second
sq square
SRT solids retention time
SS suspended solids or
single-stage
STOC soluble total organic carbon
SWD side water depth
TOG total organic carbon
TSL thickener solids loading
UV ultraviolet light
W vacuum filter dry cake weight
wgt weight
Y vacuum filter yield
2SCC two-stage counter-current
K constant in Equation (1)
294
-------
Symbols:
@ at
°C degrees centigrade
Cr raw wastewater solbule
organic concentration
C0 feed soluble organic
concentration
Ci intermediate stage soluble
organic concentration
Ce effluent soluble organic
concentration
°F degrees fahrenheit
M powdered carbon concentration
# number
X organics removed (Co-Ce)
S-Q vacuum filter cake dry time
less than
approximately
% percent
295
-------
English Metric Conversion Table:
Form Filtration Rate (FFR), Ib/hr/sq ft
feet (ft)
gallon (gal)
gallons per minute (gal/min)
Horsepower (Hp)
hydraulic loadings (gpm/sq ft)
inch (in.)
inches per minute (in/min)
million gallons (MG)
million gallons per day (mgd)
pound (Ibs)
pound per hour (Ib/hr)
pounds per ton of dry solids (Ibs/tds)
simplified correlating factor
(min-sq ft/lb)
standard cubic feet per minute (SCFM)
thickener solids loading (Ibs/day/sq ft)
yield (Ibs/hr/sq ft)
4.89 kg/hr/sq M
0.3048 M
3.785 £
3.785 £/min
1.34 kilowatts
0.679 £/sec/sq. M
2.54 cm
2.54 cm/min
3785 metric tons
or M3
3785 metric tons/
day on M3/day
454 gins or 0.454 kg
0.454 kg/hr
0.05% or 500 ppm
0.205 min-sq M/kg
0.0283 SCM/min
4.89 kg/day/sq M
4.89 kg/hr/sq M
296
-------
SECTION XIV
APPENDICES
APPENDIX A
SAMPLING AND LABORATORY TEST PROCEDURES
SAMPLING
Collection of samples for analytical testing was achieved
using one of the following methods:
1. Automatic composite sampling,
2. Routine grab-composite sampling,
3. Non-routine grab sampling, and
4. Reagent feed grab sampling.
Flow proportioned composite samples of each process stream
were collected and transferred to a refrigerator using an
automatic sampler. These samples were composited over a
24 hour period. Two 24 hour composites were usually com-
bined in the laboratory for analysis as 48 hour composites.
Composite samples of the solids-contact units reaction zone
and underflow slurries and thickener overflows were collec-
ted by manually obtaining 200 m£ grab samples on a time
proportionate basis (normally each 4 hours). These samples
were composited for a 24 hour period.
Concentrations of feed chemicals, spent and regenerated car-
bon slurries, and sludges were monitored by taking grab
samples of well mixed inventoried volumes and analyzing for
suspended solids or specific chemicals. Grab samples were
obtained to determine instantaneous concentrations of total
or suspended solids at various depths in thickeners and
solids-contact units. Grab samples of process streams were
also used to determine sulfide, iron and aluminum concentra-
tions .
297
-------
ANALYTICAL PROCEDURES
In certain of the following subsections the term "soluble"
is used. By arbitrary definition this term refers to that
portion of a given pollutant or material contained in an
aqueous sample which will physically pass through the open-
ings in a 0.45 micron membrane filter.
Turbidity
Turbidity of grab and composite samples were determined with
a Hach Chemical Company, Model 2100 laboratory turbidi-
meter. This device is a nephelometer (90 degree light
scattering) which was standardized with a plastic rod.
Turbidity was reported as Jackson Turbidity Units (JTU).
Suspended Solids
Suspended solids were defined as material retained on a 0.45
micron membrane filter. Concentrated samples of chemical-
sewage sludge and carbon slurries were filtered through
glass mat filter pads (Type GF/B). A calibrated 30 m£
syringe was used to measure sludge sample volumes. The
Standard Methods procedure for Nonfilterable Residue was
used.15
Phosphorus
Total phosphorus was determined by the persulfate digestion
method.15 Orthophosphate was determined by the ascorbic
acid method.l 5
Chemical and Biochemical Oxygen Demand
Total and soluble chemical oxygen demand (COD, SCOD) and 5-
day, 20°C biochemical oxygen demand were determined accord-
ing to Standard Methods. Total COD was not determined
on carbon contactor effluent samples since one mg/£ of
carbon was found to exert about one mg/Jl of COD.
pH
The pH data reported were obtained with a laboratory pH
meter. Process pH meters and recorders were used only to
automatically monitor and control lime feed and acid neut-
ralization during lime treatment.
Hardness, Alkalinity, Sulfides
These analysis were performed according to Standard
Methods.l 5
298
-------
Iron, Aluminum, Calcium, Magnesium
These analysis were made by Atomic Adsorption.
Available Lime Index
"The available lime index" of hydrated lime was determined
using the procedure presented in "Standard Analysis of
Limestone, Quicklime, and Hydrated Lime".17
Gaseous Oxygen, Carbon Dioxide and Carbon Monoxide
Oxygen was determined with a Bacharach 0-7 percent by
volume oxygen Indicator, Model CPD, using Fyrite 02 indi-
cator solution.
Carbon dioxide was determined with a Bacharach 0-20 percent
by volume CO2 Indicator, Model CND, using Fyrite CO2 indica-
tor solution.
Carbon monoxide concentration as percent by volume was
determined with a Bacharach Universal Gas Sampler, Model
19-7016, using CO indicating tubes no. 19-0195.
Total Suspended Solids, Ash, and Volatile Solids
Total suspended solids of carbon slurry samples were deter-
mined according to Standard Methods. Ash content of carbon
solids was determined according to the procedure presented
in ASTM, Part 28 (1964), D1506-59, p. 750. Volatile solids
of carbon samples were determined according to the procedure
presented in ASTM, Part 28 (1964), D1620-60, p. 811, with
the exception that a temperature of 550°C was used, rather
than 950°C, to account for organic volatiles only.16
Volatile and ash content of chemical sludges were determined
according to Standard Methods.15
Iodine Number
The iodine number was defined as the milligrams of iodine
adsorbed by one gram of fixed carbon when the iodine concen-
tration of the residual filtrate was 0.02 normal.
Procedure:
1. Pulverize an air dried sample of carbon in a mortar
and pestle.
299
-------
2. Dry for a minimum of 3 hours at 105°C.
3. Weigh enough dried pulverized carbon to result in
the removal of iodine from 0.1 Normal to about
0.02 Normal.
4. Transfer the weighed sample into a dry, glass-
stoppered, 250 m£, Erlenmeyer flask.
5. To the flask add 10 m£ of five percent by weight
HC1 acid and swirl until carbon is wetted.
6. Place flask on hotplate, bring contents to boil
and allow to boil for only 30 seconds.
7. After allowing flask and contents to cool to room
temperature add 100 mi of standardized 0.1 Normal
iodine solution to the flask.
8. Immediately stopper flask and shake contents
vigorously for 30 seconds.
9. Immediately filter by gravity through a folded
filter paper.
10. Discard the first 20 or 30 m£ of filtrate and
collect the remainder in a clean beaker. Do not
wash the residue on the filter paper.
11. Pipette 50 m£ of the filtrate into a 250 nu
Erlenmeyer flask.
12. Titrate with standardized 0.1 Normal sodium thio-
sulfate solution until the yellow color has almost
disappeared.
13. Add about 2 m£ of starch solution and continue ti-
tration until the blue indicator color just dis-
appears .
14. Record the volume of sodium thiosulfate solution
used.
15. Calculate the iodine number as follows:
X _ A- (2.2 x B x m£ thiosulfate solution)
M fixed carbon (g)
C = N2 x m£ thiosulfate solution
50
300
-------
Iodine Number =
Where:
X/M
A
X
M D
mg iodine adsorbed per g of carbon
Nn x 12693.0
B = N2 x 126.93
C = residual filtrate normality
N, = normality of iodine solution
N2 = normality of sodium thiosulfate
solution
D = correction factor (obtained from a
residual filtrate Normalization graph
using term C)
The capacity of a carbon for adsorbing any solute is depen-
dent on the concentration of the solute in the liqid contact-
ing the carbon. Thus, the concentration of iodine in the
filtrate must be specified. The amount of carbon sample
to be used in the determination is governed by the activity
of the carbon. If the residual filtrate normality, C, is
not within the range of 0.008N to 0.035N, the above procedure
must be repeated using a different size sample. To maximize
the accuracy of the test, the potassium iodide to iodine
weight ratio should be 1.5 to 1.0 in the standard iodine
solution.
Molasses Number
Using 100 as a standard for virgin Aqua Nuchar A, a compara-
tive molasses number was determined according to the follow-
ing procedure:
1. Prepare a stock molasses solution as follows:
a. Dilute 20 m of molasses to 250 m£ with dis-
tilled water.
b. Filter the entire solution with Whatman GF/B
5.5 cm filter paper.
(This is good for 24 hours only and should be
refrigerated).
2. Weigh 0.50 g each of fixed carbon to be tested
and virgin carbon (molasses number 100-arbitrary)
previously dried at 105°C.
301
-------
3. Add the samples to separate 150 m£ beakers and
pipette 50.0 m£ of stock molasses solution to the
beaker.
4. Stir and heat the mixture to 90°C and then cool.
5. Filter the samples through 5 . 5 cm Whatman GF/B
filter paper.
6. Determine the absorbance of each filtrate with a
spectrophotometer at 425 nm, using distilled water
as zero absorbance.
7. Calculation:
Molasses Number = 100 x
A
Sample
Where: A = absorbance
LABORATORY PROCEDURES
Equilibrium Adsorption Isotherm Test
The adsorptive response of soluble wastewater organics on
virgin or regenerated carbon was determined using twenty-
four hour grab composite samples of chemical treatment unit
effluent. Contacting of the carbon was accomplished with
a six place Turbitrol Jar Test Apparatus. All adsorption
tests were conducted at a room temperature of 68 to 75 °F.
The step by step procedure followed was:
1. Acid rinse all glassware.
2. Prepare carbon samples as follows:
a. Tare 100 m£ beakers
b. Add a desired weight of oven dried (105°C)
fixed carbon to each of the beakers and then
add 20 to 30 m& of distilled water.
c. Cover beakers with watch glasses and place on
a hot plate. Bring contents to a boil, remove
from the hot plate and cool to near ambient
temperature.
3. Add 1.0 to 1.5 liter samples of wastewater to each
of six beakers and mix at 100 rpm.
302
-------
4. Measure the initial temperature and pH of each
sample, adjusting the pH to 7.0 to 7,5 if necessary
with concentrated H2S04 or NaOH.
5. To the mixing samples, affect a quantitative trans-
fer of the degassed carbon (from Step 2), recording
the amount of distilled water used for rinsing. A
blank wastewater sample is always run without any
carbon.
6. Mix samples for one hour at 100 rpm (preliminary
tests had indicated that at least 95 percent SCOD
removal was obtained at about 20 minute contact
time) .
7. At the end of one hour measure pH and temperature
of each sample.
8. Filter a sample from each beaker through a 0.45
micron membrane filter and determine the COD con-
centration .
Standard procedure involved running triplicate COD tests on
the blank and duplicate tests on each carbon contact sample.
The difference between the blank and carbon contact average
organic concentration was considered adsorbed organics. The
organic removal (mg/£ adsorbed organics per mg/£ carbon) was
plotted versus the average equilibrium organic concentration
(mg/£ on log-log graph paper).
Gravity Thickening Tests
Laboratory gravity thickening tests were conducted by the
following procedure. The laboratory test device used is
shown on Figure A-l.
1. Obtain a 24 hour grab composite sample of sludge
blowdown. If tests are to be run at different
initial solids concentrations, then a sample of
overflow was also obtained for dilution of the
sludge.
2. Pour 2 liters of representative sludge into a 2
liter graduated cylinder for which the volume
versus depth relationship is known.
3. Gently mix the contents to insure uniform suspension,
A rubber stopper on the end of a rod works well. If
a conditioning chemical is used, it is added during
this gentle mixing.
303
-------
FIGURE A-lj SLUDGE THICKENING TEST APPARATUS
Picket Thickener
Mechanism
Ring Stand-
Two-Liter
Graduated
Cylinder
304
-------
4. Immediately insert the picket thickener mechanism
into the sludge and start it rotating at 6 revolu-
tions/hr. Also, immediately start a clock timer.
5. Observe and record sludge interface height (m£) at
appropriate time intervals such that a smooth curve
is identified. The test is normally run until no
further thickening is observed for a several hour
period.
6. Remove the picket mechanism and decant as much clear
supernatant as possible, recording the remaining
volume of sludge.
7. Filter all of the thickened sludge through a Whatman
#1 filter paper, using a Buchner funnel. The fil-
trate is saved for a specific gravity determination.
8. Dry the filtered sludge at 105°C to constant weight
to determine the mass of sludge solids (supernatant
from Step 6 is assumed to contain no suspended
solids).
From the above data, slurry and solids densities can be
calculated.
Vacuum Filter Leaf Tests
Leaf tests were conducted on composite sludge samples. The
filter circular leaf used had an area of 0.10 sq ft. The
general arrangement of laboratory equipment for conducting
the leaf test is shown on Figure A-2. The procedure
followed was:
1. Place about 1 to 2 liters of a representative
sample of sludge in a 2 liter beaker.
2. If conditioning chemicals are used, they are gently
mixed into the sludge with a large spatula.
3. Apply a vacuum to the filter leaf.
4. Immerse the properly sealed and conditioned filter
leaf into the mixed sludge. The vacuum level is
immediately adjusted to the desired level and left
there for the desired cake form time (FT).
305
-------
FIGURE A-2: SLUDGE DEWATERING TEST APPARATUS
Flexible
Vacuum Hose
To Vacuum
(15-20" Hg)
CO
o
2-Liter
Sludge
Sample
Filter
Leaf
(0.10 ft2)
-------
5. Remove the leaf from the sludge and maintain the
desired vacuum for a desired cake dry time (9 ) .
a. If the cake cracks, note the time.
b. Note the time of any reduction in vacuum
level during dry time.
6. Turn off vacuum at end of dry time.
7. Note any difficulty in removing the cake from leaf
8. Place all cake on a tared evaporation dish and
weigh immediately; this will give wet filter cake
weight (W ).
w
9. Measure filter cake thickness (T ).
10. Dry the filter cake to constant weight at 105°C;
this will give the dry filter cake weight (W ).
11. Determine the filtrate volume (contents of vacuum
flask) and filtrate suspended solids and/or
turbidity.
12. Determine sludge suspended solids of samples
collected before and after addition of any con-
ditioning chemicals such as lime.
The following are typical vacuum filter performance para-
meters which can be computed from the above data:
1. % Solids = 100 x (feed SS - filtrate SS)
Capture feed SS
% Moisture M W - W^ . . , . , . , _,_
r , . = c = w D = wet weight-dry weight
content wet weight
w ^
(Note: dry weight includes any conditioning
chemicals)
Filter Cake = W = dry cake weight (W ) .
,.,•,, >-> I J-il
Wei^ht leaf area
Form = dry filter cake weight (W)
Filtration form time (FT)
Rate (FFR)
(report as Lb of dry solids/hr•sq-ft)
307
-------
Full-Scale (Y) = Yield
Filtration = FFT submergence
Rate
= FFR x FT
Cycle Time (CT)
= W_
CT
Note: If conditioning chemical adds suspended
solids (SS) to feed slurry, the FFR should
be multiplied by the ratio of original feed
SS to original feed SS plus added SS to ob-
tain FFR and Y in terms of original feed SS,
Dry Cake = dry filter cake weight
Density leaf area and cake thickness
= W
cake thickness (T )
c
Typical vacuum filter performance parameter relationships
for a given feed sludge are:
1. At constant vacuum:
FFR = K, (FT)~k
Where: K = a constant
k = a constant
2. At constant cake thickness:
FFR = K2 (FT~Z
Where: K2 = a constant
z = a constant (theoretically - 1.0)
3. At constant form time (FT):
FFR = K3 (AP)Y
Where: K3 = a constant
Y = a constant (theoretically for a
noncompressible cake = 1/2)
308
-------
APPENDIX B
REDUCTION OF LABORATORY GRAVITY
THICKENING TEST DATA
All laboratory batch settling test data were reduced using
a standard graphical technique employed by the Sanitary
Engineering Technology Department, Eimco BSP Division of
Envirotech Corporation, Salt Lake City, Utah. The basis
for this technique is found in Kynch's theory. 1 7 This
graphical technique is depicted in Figure B-l and is made
by the following procedure:
1. Draw a linear initial sludge interface subsidence
slope as per line AB.
2. Draw a final linear slope after the knee of the
curve as per line CD.
3. Construct a bisect (EF) of the obtuse angle at E,
the intersection of tangent lines AB and CD.
4. Where the bisect (EF) crosses the sludge subsidence
curve, at point G, construct line GH perpendicular
to the bisect line EF and tangent to the curve.
Line GH is called the operating line and presumably identi-
fies the interrelationship between predicted underflow con-
centration and thickener solids loading. Underflow sus-
pended solids (g/&) equals the grams of dry sludge solids
present in the sample divided by the sludge interface height
(liters). Thickener solids loading (TSL) at a desired under-
flow suspended solids concentration is computed as follows:
TSL = Cf H0
t
x
Where:
Cf - feed or initial suspended solids in Ib/cu ft
H = initial sample height in ft
o
309
-------
FIGURE B-l:
GRAPHICAL TECHNIQUE FOR
LABORATORY GRAVITY THICKENING RESULTS
tn
i-i
0)
-P
-H
Cn
•H
Q)
U
(0
in
J-l
0)
4->
C
H
Cn
Ti
CO
100 200 300 400
Elapsed Time (tv), min
500
310
-------
t = time in days related to the desired underflow
suspended solids (interface height) by the
operating line GH.
If, as on Figure B-l, there is a discontinuity in the sub-
sidence curve near the starting interface height, a correc-
tion must be made to the observed tx. Such a discontinuity
is sometimes due to nonquiescent slurry conditions at elap-
sed time zero. For the data on Figure B-l, the correction
is seen to be 23 minutes.
Laboratory thickening data reported in Section VII were ob-
tained by making "graphical" constructions and computing
the TSL for several underflow suspended solids concentrations
and tx's defined by the operating line GH. At tx values
less than point G, the actual subsidence curve was assumed
to be the operating line.
Judgement is normally applied in choosing the plotting scale
on Figure B-l. Normal procedure, when using a two liter
graduated eyeliner, is to plot the subsidence curve on 10
by 7 inch arithmetic graph paper. The ordinate, interface
height, is the 10 inch scale and the abscissa, time, is
the 7 inch scale. A critical judgement is the location of
the actual curve on the graph (i.e., the choice of a time
scale). The standard procedure used was to locate the curve
such that the point G, defined as critical point, is located
at about 33 percent of the 7 inch abscissa or at 2.3 inches.
The effect of the relative location of the "critical point"
on the abscissa on predicted TSL can be indicated by the
following example.
On Figure B-l the time scale is 100 min/in. resulting in the
"critical point" being located at 1.6 inches on the 7 inch
abscissa. The relationship between predicted TSL and de-
sired underflow suspended solids, as defined by the opera-
ting line GH, is shown in Figure B-2 as curve B. This data
on Figure B-l was replotted twice, once at one-half and once
at double the time scale shown. "Graphical" construction
analysis for these two replottings resulted in predicted
TSL versus underflow suspended solids relationship shown on
Figure B-2 as curves A and C. Table B-l shows pertinent
parameters of this analysis for the three different plots.
From Figure B-2, it is apparent that the choice of abscissa
plotting scale has a pronounced effect on the predicted TSL
at desired underflow suspended solids. The variation is
greater as desired underflow suspended solids approaches
the maximum obtainable concentration. Table B-2 indicates
the substantial percent change in predicted TSL at various
underflow concentrations.
311
-------
FIGURE B-2:
REDUCED LAB THICKENING TEST DATA
50 55 60 65 70
Underflow Suspended Solids Concentration, g/SL
312
-------
TABLE B-l
LABORATORY GRAVITY THICKENING DATA REDUCTION:
GRAPHICAL CONSTRUCTION RELATIONSHIPS
Curve Symbol (See Figure C-2)
Abscissa Scale, min/inch
Initial Slope:
Vmin
inch/inch
T Correction, min
X
Final Slope:
£/min
inch/inch
Critical Time:
min
inches
% of 7 inches
Operating Line Slope:
£/min
i n r.h /in ch
A
200
-0.012
11.4
-23
-8xlO~5
-0.080
200
1.0
14
-0.0039
-0.99
B
100
-0.012
-6.20
-23
-8xlO~5
-0.040
162
1.6
23
-0.0018
-0-89
C
50
-0.012
3.13
-23
-8xlO~5
-0.020
137
2.7
39
-0.00080
-0.80
313
-------
TABLE B-2
ERROR IN PREDICTED THICKENER SOLIDS
LOADING AT VARIOUS DESIRED UNDERFLOW CONCENTRATIONS
Underflow SS % Change in TSL per inch
Concentration Deviation of Critical Time
% of Maximum3 from 2.3 inches
80
85
90
95
100
6.1
13
17
19
17
Maximum = 72 g/£
An obvious conclusion based on the above example is that, if
a conservative estimate of predicted TSL at a desired under-
flow suspended solids concentration is desired, then the
abscissa plotting scale should be compressed (i.e., high
values of min/inch). For this report, however, it is indeed
unfortunate that a more rational and expedient method of
laboratory thickening test data reduction was not available.
It was essentially impossible to reduce over one hundred
sets of thickening test data by hand and have all the criti-
cal points fall at a given location on the abscissa. Thus
the laboratory results reported in Section VII have a data
analysis error included.
314
-------
APPENDIX C
UV LAB STUDY
A laboratory study was designed to identify some of the
variables responsible for the poor correlation observed be-
tween UV absorbance and organic concentration. The specific
variables considered were changes in the nature of the sol-
uble organic fraction due to treatment selectivity effects
and natural diurnal changes in relative organic concentra-
tions .
LABOPATORY STUDY
Figure C-l shows the approach used to simulate the PAC-PCT
treatment effects. Laboratory simulation was necessary
because this study was made during a period of pilot plant
shut down when the project chemist had time available.
Grab samples of raw wastewater were taken at high, average
and low flow periods of the day. Organic concentrations
of the wastewater were observed to be variable, low in the
morning and high in the afternoon and evening. Each sample
was treated with hydrated lime to a pH of 11.0 and settled.
Supernatant was neutralized with H2SC>4 to pH 7.0 and con-
tacted with Aqua Nuchar A carbon at dosages of 150, 300 and
10,000 mg/£. A sample from each treatment step (raw waste-
water, lime treated, 150, 300 and 10,000 mg/£ carbon) was
filtered through glass-fiber filter paper (GF/B) to remove
particulate matter. The UV absorbance was determined for
each sample at a wavelength of 254 nm. COD was considered
the primary organic measure used because of testing accur-
acy and precision.
RESULTS
A plot of UV absorbance versus COD, BOD and TOC for each
sample (high, medium and low organic concentration) is
shown on Figures C-2, C-3, and C-4. From these plots, it
is evident that UV absorbance versus organic concentration
is not directly proportional for the entire process. An
explanation of these nonlinear results may be possible by
considering the four general types of organic material in
315
-------
FIGURE C-l: LABORATORY STUDY
Raw Wastewater
Lime
Treatment
pH 11.0
Filtration
->- UV Absorbance, COD, BOD , TOC
Neutralization
pH 7.0
Filtration
UV Absorbance, COD, BOD, TOC
Powdered Activated
Carbon Treatment
150,300
& 10,000
mg/l
Filtration
UV Absorbance, COD, BOD, TOC
316
-------
5-
S S 4
o
IM
0) -P
O hfi
§ § 3-
0
in
1-
FIGURE C-2: UV VERSUS COD-PROCESS CURVES
7pm Raw
6am Raw
10 000
Lime
(Carbon Dosage, mg/i)
10
20
30
I
40
I
50
I
60
I
70
I
80
COD/mg/£
-------
FIGURE C-3: UV VERSUS BOD PROCESS CURVES
OJ
h-'
CO
5-
4-
0
O
C
m
0) H->
O bH
c c
rt
tn cJ
2-
5pm Raw
6am Raw
Lime
150 (Carbon Dosage, mg/O
300
10,000
10
20
BOD, mg/£
30
40
-------
FIGURE C-4: UV VERSUS TOG PROCESS CURVES
u>
5-
4-
0
s a
o
o he
a c
d o
•Q-H
f-i 0)
0 fc
cc as
3-
1-
5pm Raw
6am Raw
Lime
(Carbon Dosage, mg/l )
10,000
10,000
10
12
14
16
TOC, mg/l
-------
the fraction of the wastewater which will pass through a
GF/B glass-fiber filter paper:
1. Colloidal solids
2. Solutes easily adsorbed on carbon
3. Solutes difficult to adsorb on carbon
4. Solutes unadsorbable on carbon
The change in UV adsorbance with the change in COD from raw
wastewater to lime treated water is probably due to selec-
tive removal of organic colloids. The effect of particulate
organics has been shown by using various filter pore sizes
and comparing UV adsorbance, TOC and turbidity.1 In the
present study, coagulation-flocculation probably removed
some soluble organic material and thus the UV absorbance.
The knee of the curves, on Figures C-2, C-3 and C-4 is lo-
cated in a transitional region between colloid removal and
difficult to adsorb solute removal. The transitional region
was caused by the preferential removal of easily adsorbed
solutes. As the organic solutes become increasingly diffi-
cult to adsorb, the slope of the process curves continue to
decrease until the slope becomes fairly constant approaching
the unadsorbable solute region. The unadsorbable solutes,
which vary in concentration with time, are probably respon-
sible for a considerable variation in UV adsorbance with
respect to organic concentration. This variation is evidenc-
ed by the massive carbon dosage data shown on Figure C-2.
One sample containing 18 mg/£ COD showed a UV absorbance of
zero while another sample containing only 4 mg/& COD showed
a UV absorbance of 0.25
It would appear that the nature of the soluble organic frac-
tion changes significantly due to treatment effects, thus,
making it virtually impossible to define a precise general
correlation.
TREATMENT SELECTIVITY
Evidence of the changing nature of the soluble organic frac-
tion and of the selectivity of carbon and coagulation-flocc-
ulation was demonstrated with the use of a consecutive dilu-
tion technique. Consecutive dilutions were made of each
sample (raw wastewater, lime treated, 150, 300, 10,000 mg/&
carbon) in a process curve and a UV absorbance was deter-
mined for each dilution. A plot of UV absorbance versus COD
on Figure C-5 shows a linear response for each sample. By
320
-------
FIGURE C-5: SELECTIVITY DUE TO TREATMENT EFFECTS
NJ
H
5-
Q)
O
O
6 •*
« m
CM
fi
Q) -P
O bfl
2 s
cd oj
0 >I
CO 0)
3-
• Raw
O Lime Treatment
Q 150 mg/l PAC
A 300 mg/l PAC
X Massive mg/l PAC
Dilution Curve
Raw
150 mg/l
300 mg/l
Process Curve
10,000 mg/I
10
20
30
r
40
50
I
60
70
COD
-------
reducing the COD concentration and not varying the relative
concentration of the different organic solutes, the UV ab-
sorbance varied directly with respect to COD. By reducing
COD concentrations by coagulation-flocculation and carbon
contacting, a nonlinear curve (Figure C-2) resulted, in-
dicating the selective effect of coagulation-flocculation
and carbon treatment. This selectivity is responsible for
the inability to achieve a correlation between UV absorbance
and organic concentration for the entire process.
Figure C-6 indicates a UV-COD concentration correlation is
possible by considering the specific treatment points (i.e.,
lime treatment, 150 mg/£ or 300 mg/& carbon) of Figure C-2.
However, it must be kept in mind that this is only a working
correlation for this particular wastewater after specific
treatment effects rather than a true general UV-organic
concentration correlation.
NATURAL RELATIVE ORGANIC CONCENTRATION CHANGES
While UV scanning various raw wastewater samples, it was
noted that the knee of the curve and the plateau region
shifted (Figure C-7). The curve shift was probably due to
natural diurnal changes in relative concentration of organic
solutes. To confirm this hypothesis, a test was designed
to demonstrate curve shift. Various concentrations of
detergents (ABS type) and urine were scanned. ABS and urine
were used because they can both be present in municipal
wastewater. Typical results, as presented in Figure C-8,
indicate detergents show a knee in the curve from 235 nm to
250 nm. Urine shows a knee from 250 nm to 270 nm. Various
detergent-urine concentration ratios were scanned. They
showed that as the detergent concentration increased relative
to the urine concentration, the curve shifted to the left
putting the standard 254 nm wavelength well within the
plateau region. But as the urine concentration increased
relative to the detergent concentration, the curve shifted
to the right putting the standard 254 nm wavelength on the
knee of the curve (Figure C-9). By so doing, the probability
of obtaining a precise UV absorbance-organic concentration
correlation is greatly reduced. Although detergents and
urine are not the only species adding to the UV absorbance
character of the wastewater, the above tests do demonstrate
the effect of changing relative concentrations of typical
wastewater organics on UV absorption characteristics.
DISCUSSION
The basic problem of achieving a general UV adsorbance-
organic concentration correlation results from the changing
relative concentrations of the organic solutes present in
322
-------
FIGURE C-6: SPECIFIC TREATMENT POINT CORRELATION
5-1
00
M
U)
4-
f-l
0)
o
Q) 4J
O be
q e
rt o
X>iH
f-l (D
o >
Ul tti
3-
10
Lime Treatment to pH 11.0
^50 mg/I PAC
300 mg/ PAC
20
30 40 50
COD, mg/l
60
70
80
-------
U>
NJ
0>
O
s
o
IH
O
Cfl
0.1 -
0.2 -I
0.3-
0.7
230
FIGURE C-7: DIURNAL CHANGES IN THE
NATURE OF ORGANIC CONCENTRATIONS
Defined By Tangent Bisection
240
250
260 270 280
Wave Length, nm
310
-------
FIGURE C-8: ABS AND URINE UV SPECTRA
U)
NJ
Ul
0.3-
0.4-
0.5-
0.6-
r-l
H 0.7-
0
0.9
1.1-
1.2-
1.3-
1.4
230
ABS
240
250
260 270 280
Wave Length^, nm
300
310
-------
FIGURE C-9: UV-SPECTRA SHIFT DUE TO
CHANGE IN RELATIVE ORGANIC CONCENTRATION
N)
o
O.o n
0.1-
0.2-
0.3-
0.4-
-------
the wastewater. In the PAC-PCT process, relative concentra-
tion changes result from at least three basis causes:
1. Chemical treatment selectivity,
2. Carbon adsorption selectivity, and
3. Natural diurnal organic concentration changes.
Application of UV absorbance to monitor organic concentra-
tion is dependent on the treatment scheme used and the
nature of the wastewater evaluated. If the wastewater or
treatment process stream's relative organic concentrations
are fairly constant and a nonselective treatment process
(i.e., gravity clarification) is used, then a precise UV
absorbance-organic concentration correlation might be
possible for the process influent and effluent. In a
selective treatment process, such as PAC-PCT, general UV
absorbance-organic concentration correlation prospects are
minimal. However, as previously shown a working correlation
should be possible at specific treatment points (i.e., lime
treated, 150, or 300 mg/£ carbon) if very consistent treatment
effects are achieved.
As reported in Section IX an attempt to determine such a
working correlation between COD and UV resulted in a less
than satisfactory response.
327
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-235
2.
3. RECIPIENT'S ACCESSI Of* NO.
4. TITLE AND SUBTITLE
PHYSICAL-CHEMICAL TREATMENT OF A MUNICIPAL WASTEWATER
USING POWDERED CARBON - No. II
5. REPORT DATE
November 1976 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Don E. Burns, Richard N. Wallace, Darrell J. Cook
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Eimco BSP Division of Envirotech Corporation
Salt Lake City, Utah 84110
10. PROGRAM ELEMENT NO.
1BB043
11. CONTRACT/GRANT NO.
68-01-0183
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
See also EPA-R2-73-264, "Physical-Chemical Treatment of a Municipal Wastewater Using
Powdered Carbon" (PB-224 494)
16. ABSTRACT
Salt Lake City municipal wastewater was treated in a nominal 100 gpm pilot plant by
chemical coagulation-precipitation, powdered activated carbon adsorption and granular
media filtration. Chemical-primary sludge was gravity thickened and vacuum filter de-
watered. Spent carbon was gravity thickened, vacuum filter dewatered and thermally re-
generated in a fluidized bed furnace. Solids-contact units were used for chemical
treatment and carbon contacting. The pilot plant was operated over a 16-month period
to demonstrate treatment effectiveness under diurnal flow conditions. Reuse of ther-
mally regenerated carbon was practiced for 3 months.
Soluble organic materials were found to be removed by a combination of chemical coagu-
lation, anaerobic biological activity and adsorption on powdered carbon. Spent pow-
dered carbon was effectively regenerated with an average fixed carbon recovery of 90
percent. Lime and alum-primary sludges were effectively dewatered by vacuum filtra-
tion. A high quality effluent was consistently produced, similar to that expected for
conventional biological treatment followed by tertiary treatment for phosphorus and
suspended solids removal.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
*Chemical removal (sewage treatment),
*Sludge, Dewatering, *Activated carbon
treatment, Adsorption, ^Thermal recovery
methods, Filtration, Pilot plants
Physical-chemical
treatment, Powdered
carbon, Powdered acti-
vated carbon, Carbon
regeneration, Fluidized
bed furnace
13B
J. DISTRIBUTION STATEMEN1
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)'
UNCLASSIFIED
21. NO. OF PAGES
346
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
328
U. ,. GOYEPIirEIIT PBIUTIIIG OFFICE' 1977 — 757-056/5419
------- |