EPA-600/2-76-276
December 1976
Environmental Protection Technology Series
SELECTED APPLICATIONS OF
INSTRUMENTATION AND AUTOMATION IN
WASTEWATER-TREATMENT FACILITIES
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-276
December 1976
SELECTED APPLICATIONS
OF
INSTRUMENTATION AND AUTOMATION
IN
WASTEWATER-TREATMENT FACILITIES
by
Allen E. Molvar
Raytheon Company
Portsmouth, Rhode Island 02871
Contract No. 68-03-0144
Project Officer: Joseph F. Roesler
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, 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.
11
-------
FOREWORD
The Environmental Protection Agency was created because of
increasing public and government concern about the dangers of pollu-
tion 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 Environmental 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.
Control strategies potentially applicable to wet and dry-weather
wastewater-treatment facilities were evaluated during the course of
this study. The evaluation included various levels of instrumentation
and automation which could be utilized in the implementation of these
control strategies. Cost/benefit analysis indicates that many untried
control schemes are economically attractive because of the low payback
periods. Furthermore, this study concludes that despite current
concepts, smaller (1 to 5 mgd) plants can afford and need significantly
greater amounts of automatic control. Finally, direct digital control
and computerized control can be economically justified in large dry-
weather treatment plants and storm-water control networks.
Francis T. Mayo, Director
Municipal Environmental
Research Laboratory
ill
-------
ABSTRACT
The application of modern control systems to the operation of wastewater-treatment
plants is discussed in this report. Control strategies for the commonly used wet- and
dry-weather treatment processes and their collection systems are described. Wher-
ever possible, the benefits derived from, as well as the operating problems associ-
ated with, the actual or proposed control strategies are documented. Cost/benefit
analysis indicates that many untried feedforward mass proportional control schemes
are economically attractive because of the low payback periods. Furthermore, this
study concludes that despite current concepts, the smaller (1 to 5 mgd) plants can
afford and need significantly greater amounts of automatic control. However, a lack
of reliable field-proven analytical sensors for all of the important parameters appears
to be the principal obstacle impeding the implementation of more sophisticated control
strategies.
Centralized control with semigraphic display should be used in virtually every treat-
ment plant since it saves on operating labor, improves operation, and increases the
safety of wastewater treatment. Automatic data acquisition systems are also cost
effective and should be used in all medium and large sized plants. Direct digital con-
trol and computerized control can only be economically justified in large dry-weather
treatment plants and large storm-water control networks.
This report was submitted in fulfillment of contract 68-03-0144 by Raytheon Company
under the sponsorship of the U. S. Environmental Protection Agency. It covers a
period from 30 January 1972 to 31 May 1975, and work was complete as of 31 May
1975.
IV
-------
CONTENTS
Page
Foreword iii
Abstract iv
List of Figures vii
List of Tables xi
Acknowledgements xiv
I. Introduction 1
n. Conclusions 4
HI. Recommendations 7
IV. Automatic Measurements 10
V. Automatic Control 17
VI. Control of Dry-Weather Treatment Processes . . 21
Influent Pumping and Pretreatment Control ... 21
Primary Sludge Pumping 24
Control of Dissolved Oxygen in Aeration Basins . . 32
Food-to-Microorganism Control 61
Trickling Filters 66
Secondary Sludge Pumping 70
Chlorine Disinfection 74
Control of Anaerobic Digesters 82
Sludge Conditioning for Vacuum Filtration ... 90
Sludge Dewatering 94
Incineration 98
Neutralization of Acids and Bases 101
Phosphorous Removal by Lime Treatment ... 128
VII. Control of Wet-Weather Treatment Processes and
Collection Systems 146
Combined Sewer Overflow Treatment Plant Small
(Hypothetical 150 Acre) Catchment Area . . . 148
Offline Storage and Chlorination Treatment Plant
(3000 Acre Drainage Area) 163
Combined Sewer Overflow Systems (Hypothetical
Large Drainage Area of 100,000 Acres) .... 173
v
-------
CONTENTS (Cont)
Vin. Cost/Benefit Analysis
IX. Modern Control Systems
X. Instrumentation Layouts
Appendices
A. Food to Microorganism Control Detail 287
B. Notation for Instrument Control Loops 291
References 294
vi
-------
FIGURES
1. Distribution of measuring instruments observed during
user survey 14
2. Performance summary for measuring devices in wastewater
treatment 15
3. Basic types of process control loops 19
4. Raw sewage pumping control 22
5. Screen control 23
6. Sludge density control 29
7. Sludge interface control. 31
8. Centrifugal blower performance curves showing effects of
inlet guide vanes 39
9. Flow-proportional DO control for completely mixed
activated sludge 41
10. PI feedback DO control for completely mixed activated
sludge 43
11. Flow proportional feedforward DO control with feedback
trimming 45
12. TOC feedforward DO control with feedback trimming ..." 47
13. Respiration rate DO control with feedback trimming ... 48
14. Variation in oxygen demand with basin length 50
15. Activated sludge system 62
16. Variation in BOD load in Baltimore 63
17. Flow proportional recirculation control 68
18. Flow proportional sludge pumping 73
19. Simple flow-pacing chlorination control loop 77
20. Typical compound chlorination control loop 78
21. Double compound loop C\2 control 79
22. Digester temperature control 84
23. Digester pH control 85
24. Combined pH, temperature, and digester gas control for
single-stage unit 88
25. Combined pH, temperature, and digester gas control for
two-stage unit 89
VII
-------
FIGURES (Cont)
26. Feedback pH control based on filtrate pH 92
27. pH control of conditioning chemical solutions 93
28. Effect of increasing sludge solids concentration (dry basis)
on requisite conditioner dose 95
29. Flow proportional chemical feed system 95
30. Mass loading chemical feed system 96
31. Control system for a sludge incinerator 100
32. Carbonates provide valuable buffering in range of pH 6 to 8 . 102
33. This configuration combines a thorough blending with fast
response of pH to reagent flow 104
34. Slurry control valve should be mounted so that solids will
settle away from it 107
35. Three-way valve is actuated by an alarm on effluent pH,
preventing discharge of off-limit materials 109
36. Construction details of a measuring electrode 110
37. Construction details of a flowing reference electrode ... Ill
38. An air purge of submersible electrodes can eliminate leakage
problems altogether 113
39. A properly scaled divider can effectively linearize an equal-
percentage valve characteristic 118
40. Two equal-percentage valves may be sequenced to act as a
single wide-range valve 119
41. This nonlinear function greatly improves pH control . . . 122
42. Flow-feedforward systems for fixed stroke (A) and variable
stroke (B) metering pumps 125
43. pH feedforward system is capable of a wide range through
sequenced equal-percentage valves 126
44. Small tank is a hydraulic analog of larger stream .... 127
45. Theoretical solubility of phosphates in lime solutions as a
function of pH 130
46. Actual measurement of effluent phosphorus from a single-
stage lime treatment process as a function of pH, compared
to theoretical relationship 131
47. Control of lime slaking 134
48. Control of dry lime feed. . 135
49. Upper record shows uncontrollable pH caused by water
flowing at a constant rate into slurry tank 136
50. Linearization of head signal from a flume or rectangular
weir may be performed with a square-root extractor and
multiplier 139
Vlll
-------
FIGURES (Cont)
51. Instability in this record was caused by an accumulation
of a coating on electrode surfaces, which was removed by
an ultrasonic cleaner 140
52. pH control can be combined with dosage control for effective
backup 141
53. Titration curve of 10-3 MH3PO4 with Ca(OH)2 142
54. pH control applied to recarbonation 145
55. Satellite treatment plant 150
56. Satellite plant performance 155
57. Satellite treatment plant instrumentation 161
58. Offline storage and disinfection plant 163
59. Offline storage and disinfection plant instrumentation . . . 170
60. Combined sewer overflow control 178
61. Total annualized wastewater-plant costs 182
62. Wastewater-plant labor costs 183
63. Automatic control payback periods vs plant size for typical
municipal wastewater treatment 184
64. Various methods of chemical additive changing 191
65. Phosphorus precipitation by lime addition in hydraulic load
following control mode 199
66. Total load following 201
67. Phosphorus precipitation by alum addition through hydraulic
load following control mode 204
68. Phosphorus precipitation by alum addition through total load
following control mode 206
69. Hydraulic load following controls for prechlorination . . . 210
70. Total load following controls for prechlorination .... 212
71. Hydraulic load following controls for disinfection and
dechlorination 216
72. Total load following control for disinfection and dechlorina-
tion 219
73. Hydraulic load following control for aeration 224
74. Total load following based on feedback control 226
75. Total load following based on feedforward control .... 227
76. Cascade temperature control of single-stage digester . . . 233
77. Combined pH and temperature control of single-stage
digester 237
78. Underground pipe excavation, backfill, anchors 246
79. Polymer addition system for flow augmentation 248
80. Stroke of metering pump sets dosage of reagent A .... 253
IX
-------
FIGURES (Cont)
81. Dosage calculated by divider may be recorded and is easily
set from a remote source 253
82. Dynamic compensation is required if actual flow at injection
point differs from measured flow in time 254
83. Water composition signal needs dynamic compensation and
protection against analyzer failure 256
84. Feedback trim may be accomplished in two ways .... 257
85. Typical semigraphic display panel in a central control room . 261
86. Control room layout 262
87. This computer control system fails to provide dosage control. 268
88. One mgd activated sludge wastewater-treatment plant,
instrumentation diagram 277
89. Ten mgd activated sludge wastewater-treatment plant,
instrumentation diagram 280
90, One mgd physical/chemical wastewater-treatment plant,
instrumentation diagram 283
91. Ten mgd physical/chemical wastewater-treatment plant,
instrumentation diagram 286
-------
TABLES
1. Instrument Performance 13
2. Summary Of Primary Sludge Pumping Control Strategies . 33
3. Possible Manipulation And Final Control Elements ... 40
4. Summary Of DO Control Strategies for Plug Flow Aeration
Basins With Uniform Oxygen Transfer 52
5. Summary Of DO Control Strategies for Plug Flow Basins
With Separately Adjustable Oxygen Transfer Devices . . 53
6. Summary Of Automatic DO Control for Completely Mixed
Aeration Basins 60
7. Results Of Computations Of Simulated Control Procedure
Under Specified Plant Conditions 65
8. Trickling Filters 69
9. Disinfection Via Chlorine Addition, 81
10. Anaerobic Digester Control Strategies 91
11. Sludge Dewatering—Vacuum Filter 98
12. Wet-Weather Detention And Chlorination 172
13. Approximate Costs Of Wastewater-Treatment Plants . . 186
14. Cost Of Chemical Additives 188
15. Savings In Chemical Additive Use By Practice Of Load
Following 189
16. Chemical Additive Doses 192
17. Number Of Parallel Units 193
18. Expected Life And Yearly Maintenance Requirements By
Type Of Instrument 197
19. Benefits Of Hydraulic Load Following In Phosphorus
Removal Via Lime Addition 198
20. Capital And Operating Costs Of Hydraulic Load Following
Controls 199
21. Cost/Benefit Analysis Results For Hydraulic Load Following
In Phosphorus Removal Via Lime Addition 200
22. Capital And Operating Costs Of Total Load Following Controls
For Phosphorus Precipitation By Lime Addition ... 202
XI
-------
TABLES (Cont)
23. Cost/Benefit Analysis Results For Total Load Following
In Phosphorus Removal Via Lime Addition 203
24. Benefits Of Automation 204
25. Capital And Operating Costs Of Hydraulic Load Following
Controls 205
26. Cost/Benefit Analysis Results For Hydraulic Load Following
In Phosphorus Removal Via Alum Addition 206
27. Capital And Operating Costs Of Total Load Following Controls
For Phosphorus Precipitation With Alum Addition . . 207
28. Cost/Benefit Analysis Results For Total Load Following In
Phosphorus Removal Via Alum Addition 208
29. Potential Benefits Of Automation 209
30. Capital And Operating Costs of Hydraulic Load Following
Controls 211
31. Cost/Benefit Analysis Results From Hydraulic Load
Following Control Of Prechlorination 212
32. Capital And Operating Costs Of Total Load Following
Controls 214
33. Cost/Benefit Analysis Results For Total Load Following
Control Of Prechlorination 215
34. Potential Economic Benefits Of Automation Of
Chlorination 215
35. Capital And Operating Costs Of Hydraulic Load Following
Controls 217
36. Cost/Benefit Analysis Results For Hydraulic Load Following
Control Of Disinfection And Dechlorination .... 218
37. Capital And Operating Costs Of Total Load Following
Controls 220
38. Cost/Benefit Analysis Results For Total Load Following . 221
39. Annualized Costs Associated With Fixed-Rate Aeration . 222
40. Potential Benefits Of Automation ($/Year) 223
41. Capital And Operating Costs Of Hydraulic Load Following
Controls 225
42. Cost/Benefit Analysis Results For Hydraulic Load Following
DO Control 226
43. Capital And Operating Costs Of Two Total Load Following
Controls 228
44. Cost/Benefit Analysis For Total Load Following Via
Feedback Control 229
xii
-------
TABLES (Cont)
45. Digester System Costs 231
46. Capital Cost Savings From Digester Automation .... 232
47. Capital Cost Savings Through Automation (Percent Of Total
Digester Investment) 233
48. Capital And Operating Costs Of Temperature Control . . . 234
49. Cost/Benefit Summary Of Temperature Control of
Digestion 236
50. Capital And Operating Costs Of pH And Temperature
Control 238
51. Cost/Benefit Summary Of pH And Temperature Control Of
Digestion 239
52. Capital And Operating Costs Of pH, Temperature, And
Methane Control 241
53. Cost/Benefit Summary Of pH, Temperature, And Methane
Control Of Digestion 242
54. Capital And Operating Costs Of Polymer Addition System. . 245
55. Potential Yearly Savings From Centralized Computer
Control 263
56. Control Building Cost 265
57. Signal And Control Wire Transmission Costs 265
58. Semigraphic Panel And Display Costs 266
59. Cost/Benefit Summary Of Centralized Control 266
60. Cost/Benefit Summary Of Centralized Computer Control . . 262
Kill
-------
ACKNOWLEDGEMENTS
The author is greatly indebted to the assistance of the subcontractors, Teetor-Dobbins,
Crawford & Russell, C. E. Maguire, Inc., Victory Tool & Die Works and the Foxboro
Company. In particular, the author is indebted to Dr. William E. Dobbins, Mr. Bela
Liptak, Mr. Russell H. Babcock, and Mr. F. G. Shinskey. This report would not
have been possible without the assistance of the Project Officer, Mr. Joseph F.
Roesler.
xiv
-------
SECTION I
INTRODUCTION
During the past three decades, automatic process control has had a dramatic impact
on most industrial processes. In particular, many chemical processes must be auto-
matically controlled in order to turn out a satisfactory product; consider, for example,
polyethylene production, where unsatisfactory control results in a reactor full of soot
rather than high-grade plastic. Today's managers look toward instrumentation and
automatic control as a means of increasing profits by:
• Maximizing productivity and minimizing off-spec products
• Reducing energy consumption and operating labor
• Automatically regulating the principal variables
• Computerizing supervisory control.
Despite the large number of successful applications of instrumentation and automatic
control in the chemical, food, petroleum, and other related industries, the wastewater-
treatment profession has been very slow to instrument and automate various waste-
water-treatment processes. In fact, most treatment processes are designed as high-
capacitance self-regulating processes that can be operated over a wide range of
conditions without any loss in efficiency. This design philosophy evolved from the
general lack of quick-time analytical instruments and automatic corrective control.
For example, the activated sludge process in a municipal plant must accept wide influ-
ent flow rate and organic strength variations; accordingly, oversized aeration basins
and secondary clarifiers that can absorb these variable loadings must be used.
Although some wastewater-treatment organizations have attempted to catch up with the
process industry in the use of instrumentation and control, much of the practical
-------
information (particularly as it relates to economic and technical feasibility) is not
widely known.
One primary obstacle impeding a logical approach to instrumentation and control of
wastewater-treatment plants stems from the lack of a comprehensive national state-
of-the-art survey as a reliable source of baseline information. In one effort to plug
this information gap, a team of experts in the areas of treatment process requirements,
plant design, costs, analytical instruments, control devices, and computerized automa-
tic techniques has performed a comprehensive study of the current state of the art.
During the first phase of this project, the team conducted a survey of instrumentation
practices in some 50 wet- and dry-weather wastewater-treatment facilities (1). In addi-
tion to this, all pertinent technical literature covering the base period of 1968 through
1973 was researched for instrumentation and automation information, and abstracts
were provided for all of the pertinent literature (2).
The results of the user survey and literature review have been published in separate
interim reports, thus, only the conclusions and recommendations will be repeated in
this document. This report draws upon the survey's findings, published literature,
case studies, and other pertinent information sources to develop the rationale for cur-
rent and contemplated monitoring and control strategies for the commonly used wet-
and dry-weather treatment processes. Instrumentation and automatic control devices
contemplated for use in wastewater-treatment projects should, except for safety rea-
sons, be justifiable on an economic basis. Accordingly, the use of a cost/benefit
analysis as a decision-making aid is demonstrated. Much cost data on the application
of instrumentation and automation for wastewater treatment and control are presented
in this report for use by the reader in such cost/benefit analyses. The suitability of
available instrument and automatic control devices is also appraised, and future
research needs are reported.
The state of the art begins with a review of automatic measuring devices and analytical
sensors that are needed to assess the pollution load and removal efficiencies. Then a
-------
basic philosophy of automatic open-loop, feedforward, and closed-loop control is pre-
sented. In the process control section, the principles of automatic control are brought
to bear on the commonly used wastewater-treatment processes for both wet- and dry-
weather facilities and their collection systems. For the reader's convenience, the
process control section is organized on a unit processes and operation basis, with each
subsection being virtually self-sufficient. Since the rapid advances of the control
industry make centralized and computerized control very efficient, a separate chapter
is devoted to this important topic.
After developing the process control strategies, the economic impact, payback periods,
and annualized cost savings are presented in such a manner as to develop the criteria
and demonstrate the most feasible procedures for finding cost-effective automatic con-
trol systems. In the final chapter, the principles developed are coupled "with operating
experiences to produce pragmatic instrumentation designs for 1 and 10 mgd biological
and physical/chemical wastewater-treatment plants.
-------
SECTION H
CONCLUSIONS
Based on this survey of 50 treatment facilities, nearly half the instrumentation
employed in treatment facilities measures flowrates and liquid levels, whereas ana-
lytical instruments represent about only one-quarter of the current instrumentation
used. Mechanical measuring devices dealing with position, speed, and weight account
for the remainder of the instruments observed in treatment plants.
The following measuring instruments for on-line continuous service in wastewater-
treatment activities are commercially available with sufficient reliability:
• Level • Conductivity
• Flowrate • Rainfall
• Temperature • Turbidity
• Pressure • pH
• Speed • Residual chlorine
• Weight • Free chlorine gas
• Position • Hazardous (flammable) gas.
Sludge density meters, sludge blanket level detectors, on-line respirometers, dis-
solved oxygen (DO) probes, and many automatic sampling systems use well-estab-
lished principles that are suitable for wastewater monitoring and control activities,
but require significant maintenance.
The investigations revealed that most treatment facilities use considerably fewer
instruments and automatic control devices than the closely related water supply and
chemical processing plants. The cost and performance data collected during the
survey of treatment facilities show that the average secondary plant allocates about
4
-------
3% of its construction cost for installed instruments, whereas water supply and chem-
ical treatment plants usually budget about 6% and 8%, respectively, for installed
instruments. Remote satellite wet-weather treatment facilities, which in theory
should operate unattended or at least with a minimal amount of operating manpower,
budgeted only about 2% for automatic control, and they usually required manual assist-
ance. Central computerized storm-water routing and in-line storage systems, how-
ever, employed an adequate number of instruments and automatic control devices,
and performed satisfactorily.
The primary reason for the paucity of instrumentation use in existing wastewater-
treatment plants is the unsatisfactory performance of many of the primary measuring
elements and analytical sensors. Since most of the measuring elements interface
directly with raw sewage, mixed liquor, or thickened sludge, these devices are sub-
ject to rapid fouling; accordingly, they need more frequent cleaning and calibration.
Several on-line measxiring devices for assessing the organic concentration (TOC, TOD,
COD, and respirometry) of wastewater are commercially available at this time.
These on-line analyzers require copious amounts of skilled maintenance, which is
usually unavailable in most wastewater-treatment plants, and finally, many instru-
ments have not been sufficiently evaluated in the field to form any conclusions.
Automatic control loops should be economically justifiable and cost effective. In this
report the payback period presented as a function of plant size was used for the evalu-
ation of automatic control loops. It is assumed that a payback period of 2. 5 years or
less is adequate to justify the purchase and installation of an automatic control loop
although longer payback periods may also be cost effective. The cost/benefit analy-
ses, which used actual field data wherever possible, show that most flowrate propor-
tional, feedback, and combination feedback feedforward control strategies are econom-
ically attractive for plants in the 1 to 5 mgd range. In spite of the favorable economics
of automatic process control, most wastewater-treatment plants use very little auto-
mation. The survey of 50 treatment facilities found that automatic chemical addition
control, residual chlorine control, and digester temperature control were used by
5
-------
about only one-third of the visited plants. At these plants, most managers considered
the automatic control systems cost effective since they do save energy and chemicals
and, at the same time, improve the plant operation.
Central control of a plant organizes its operation so that all important events, alarms,
and treatment information are displayed, and recorded in a centralized location.
Virtually all of the large facilities surveyed successfully utilized central control.
Moreover, a cost/benefit analysis shows that centralized control is economically
justifiable in plants as small as 1 mgd. Most new plants use automatic data acquisi-
tion systems; approximately 20% of the visited facilities used data-logging computers.
Compare this to the figure that only 10% of the plants used dissolved oxygen control.
However, direct digital computer control is not well established in dry-weather treat-
ment plants. Real-time computerized supervisory control of large storm and combined
sewer networks is cost effective because the vast number of variables and control points
exceeds human computational and decision-making capabilities.
Although the instruments and control systems in wastewater-treatment plants were
adequately maintained, satellite storm-water treatment facilities were supplied with
less than adequate maintenance. On the other hand, storm-water control centers that
typically receive storm-water and combined sewer network information were well
maintained and operated satisfactorily.
-------
SECTION III
RE COMME NDATIONS
Cost/benefit analyses clearly show that most automatic control loops can be cost
effective in smaller plants (i.e., 1 to 5 mgd). Therefore, greater amounts of auto-
matic control should be specified for the following processes, even in small plants:
• Prechlorination
• Aeration
• Digestion
• Disinfection
• Phosphorous removal
• pH adjustment.
Intensive applications of elaborate and novel logic schemes, computers, displays, and
recorders will not improve the effectiveness of wastewater treatment. Instead, well-
documented field evaluation programs are needed to help appraise the potentially useful
but untried control strategies developed in this work and elsewhere.
The need for more reliable operation to meet the quality standards and effluent permit
requirements will encourage greater use of instrument and automatic control.
Nearly 70% of the plants included in a recent survey of 50 plants had neither operation
nor cost data on their instrumentation. Wastewater-treatment plant management must
make every effort to improve recordkeeping practices. Development of a uniform,
easily practiced instrument recordkeeping system would facilitate improvements in
this area..
-------
Many misunderstandings and much confusion can be avoided in the future by having
design engineers use standard symbols on instrument drawings.
Since instrument purchasing and installation are steadily becoming more complex,
serious consideration should be given to new contractual procedures to ensure that
specified instruments and control systems are operating and effective when installed.
Since instrumentation and automatic control devices require both maintenance and
calibration, any plans for increased instrumentation must include plans for upgrading
the qualifications of the maintenance staff.
Instrument technician training programs sponsored by the private sector and public
funds should be expanded to combat the present-day shortage of qualified instrument
technicians in the wastewater-treatment field.
Cost/benefit analyses that make use of total annualized costs should be used as
decision-making aids when considering alternative and optional instrumentation
systems.
Environmentally oriented professional societies and government research agencies,
working together, should sponsor the writing of a comprehensive manual of recommended
practices relating to the instrumentation and automation of wastewater treatment.
As a suggested guide for future research and development, the following list of sensors,
control loops, and computer hardware and software sums up the most critical needs of
wastewater-treatment instrumentation and automation:
• Sensors:
• Rapid and automatic on-line organic monitoring devices
• On-line wet chemical analyzers for ammonia and total phosphates
• Flow meters for on-line use in sanitary, storm and combined storm systems.
-------
Control loops:
• Organic load equalization
• Food-to-microorganism ratio
• Breakpoint chlorination
• Mass proportional phosphate removal
Computer hardware and software:
• User-oriented language
• Uniform data formatting and reporting
• Standardized input/output requirements
• Centralized software library containing program routines useful for waste-
water treatment plant operation control and management.
-------
SECTION IV
AUTOMATIC MEASUREMENTS
INTRODUCTION
Inasmuch as an automatic control system should respond to changes of the influent
loading and ambient conditions, the automatic measuring devices that detect these
changes are a significant part of the automatic control system. The analytical sen-
sors and other measuring devices act as the eyes and ears of the control system and
provide the input signal to the controller. The universal presence of uncertainty in
any physical or analytical measurement must be clearly understood at the onset of an
error analysis. These errors come from the device itself as well as from the stan-
dards used in calibrating the instruments. Discounting the significance of errors that
arise from the calibration, the important sources of instrument errors in wastewater-
treatment applications are:
• Noise that introduces a significant amount of spurious signals
• Response time or dynamic lag
• Interactions that cause a loss of measurement integrity
• Deterioration of the measuring device.
INSTRUMENT NOISE
A survey (1) of approximately 50 wastewater-treatment facilities has provided a
large amount of data on the in-service history and operating experiences of many
process-measuring instruments. Most of the output signals from dissolved oxygen
(DO) probes, flowmeters, pH probes, and total organic carbon analyzers can
contain an appreciable amount of noise in relationship to the true signal. Incomplete
10
-------
mixing, poor shielding, poor sampling, and poor instrument designs accounted for
most of the noise observed in the on-line measuring devices. Fortunately, good engi-
neering practices and simple filtering systems reject most of the unwanted noise
components and, as a result, noise is not a significant problem with the current
instrumentation systems used in wastewater-treatment projects.
RESPONSE TIME
Response time errors or dynamic lags may also contribute to the uncertainty of a
measurement. If the sample characteristics are changing with respect to time, a
significant lag or long response time results in a signal whose value depends on the
time history of the sample over the previous interval of time. As an example, imag-
ine that a cold thermometer, reading 20°C (68°F), is suddenly plunged into boiling
water. The thermometer will not immediately read 100°C (212°F). In fact, a certain
lag period is required to reach the 100°C mark. During the lag period the thermome-
ter is producing an inaccurate measurement due to its dynamic lag. The instantaneous
error depends on the actual temperature, how fast the temperature is increasing, and
the characteristics of the thermometer itself. The difference between the actual and
indicated temperatures becomes negligible only after steady-state conditions have been
reached and maintained for a reasonable period of time.
Measurement lag is dangerous in automatic control systems because it can cause
overshoots and process instability. Some analytical sensors that measure the com-
position of a flowing stream operate on a batch principle, which accentuates the
response time problems. During the user survey, none of the observed automatic on-
line analyzers—with the possible exception of those measuring respiration rates— had
excessively long response times, taking into account the process time characteristics.
Several of the analytical systems employed sampling networks which did introduce
considerable lag times into the overall measurement system.
11
-------
INTERACTIONS
Although interaction errors may arise from coupling and feedback to other measuring
systems and power sources, interference by other chemical constituents is the most
serious interaction problem in the use of analytical sensors in wastewater-treatment
service. The reader should remember that analytical sensors are required to deter-
mine the trace concentration of a single component from a complex multi-component
mixture. However, interaction problems should be largely diminished with the avail-
ability of better and more specific selective-ion probes.
MAINTENANCE
The physical or chemical deterioration of measuring devices, as well as other changes
in their characteristics, may cause a corresponding change in their response time and
indication; this also constitutes a significant error source. Since most measuring
devices in wastewater service interface directly with raw sewage, mixed liquors, or
thickened sludge, these devices are subject to rapid deterioration by fouling from
solids deposition, slime buildup, and precipitation. Accordingly, they need frequent
cleaning and calibration. Moreover, the operating experiences with measuring devices,
which were accumulated during the user survey and are presented in Table 1, indicate
that instruments used for wastewater applications require more maintenance than their
industrial counterparts. The survey team found that most treatment plants supplied
approximately 90% of the maintenance resources needed. Although most plants have
reasonably well-qualified maintenance staffs, any plans that call for adding sophisti-
cated instruments must also provide for upgrading the staffs qualifications.
COMMERCIALLY AVAILABLE, USEFUL INSTRUMENTS
Figure 1 indicates that flow and level devices account for nearly half the instrumenta-
tion employed in existing treatment facilities (1). Analytical instruments represent
approximately a quarter of the instruments observed, while position, speed, weight,
and other mechanical-type measurements total about 15%.
12
-------
Table 1. INSTRUMENT PERFORMANCE
VARIABLE
LEVEL
FLOW
DENSITY
ANALYSIS
MISC.
CONTROL
INSTRUMENT
Bubbler
d/p Trans.
Float 6 Cable
Optical
Flume & Weir
Venturi, etc.
Propellers
Pos. Displace.
Magnetic
Nuclear
Mechanical
pH and ORP
Dissolved O2
Res. Chlor.
Turbidity
Conduct.
Chlorine Gas
Explosive Gas
BOD, TOC, etc.
Temp.
Press.
Speed
Weight
Position
Sampling
Rainfall
Level
Flow
Sludge
Air Flow
Dosage
Res. Chlorine
DO
APPLICATION
Tanks & Wet Wells
Digesters & Sludge
Tanks fc Wet Wells
Sludge Blanket
Major Flows
Air and liquids
Clean liquids
Gases
Liq. and Sludge
Med. & Thick Sludge
Med. & Thick Sludge
Aqueous Liquids
Aqueous Liquids
Aqueous Liquids
Fairly Clean Liquid
Aqueous Liquids
Airspace
Airspace
Wastewater
All
All
Engines, etc.
Sludge or C12
Sloice Gates
Liquid Streams
Storm Waters
Wells 6 Basins
All Fluids
Sludge Separation
Aeration
Chlorination
Aeration
TYPICAL
COST
$200
700
400
IK
2K+
800+
1K+
500+
2K+
5K
—
2K
2K
5K
3K
IK
3K
3K
—
300
200
—
2K
IK
4K
500
TYPICAL MAINTENANCE
FRQ/YR.
12
0.6
24
—
1.4
4
7
2*
12
48
MH/YR.
STP IND
8 4
5 5
60 5
— —
2 -
20 6**
10 10
80* 10
12 8
51 40
Excessive
300
100
365
—
200
24
12
50 29
60 —
140 —
— —
60 -
50* —
12+ 50
Excessive
1*
5
—
24*
18*
0.5
24*
8* 1
4 4
— —
60* —
30* —
20 —
50* -
HELIAB.
(MTBF)
1-2 yrs.
1-5 yrs.
. 2-2 yrs.
. 1-5 yrs.
.5-5 yrs.
2 mo. -5 yrs.
1 mo.-l yr.
1 mo.-l yr.
. 5-10 yrs.
1-3 yrs.
1-6 mos.
1-4 mos.
1-9 mos.
.2-lyr.
1-6 mos.
1-4 mos.
.5-lyr.
.2-lyr.
. 1-1 mo.
.5-2 yrs.
.1-5 yrs.
.6-5 yrs.
.6-2 yrs.
.1-lyr.
.1-lyr.
1-5 yrs.
TYPICAL
USE
5-15 yrs.
5-15 yrs.
2-20 yrs.
2-8 yrs.
5-30 yrs.
5-30 yrs.
1-8 yrs.
1-5 yrs.
5-20 yrs.
8 yrs.
2 yrs.
.5-5 yrs.
.1-5 yrs.
4 yrs.
4 yrs.
4 yrs.
8 yrs.
8 yrs.
.3-1 yr.
5 yrs.
5 yrs.
5 yrs.
10 yrs.
lyr.
4 yrs.
12 yrs.
NOTE:
STP = TREATMENT PLANT
IND = INDUSTRIAL,
SEE TEXT
* Estimated
** d/p Converter only
Based on the field survey results (which are summarized in Figure 2), instrument
users indicated that the following commercially available measuring instruments pos-
sess sufficient reliability for on-line use in wastewater-treatment facilities: level,
flow rate, temperature, pressure, speed, weight, position, conductivity, rainfall,
turbidity, pH, residual chlorine, free chlorine gas, and fire-hazardous (flammable)
gas. Suspended solids analyzers may also be added to the list, based on the experi-
ence at Palo Alto, California (3).
13
-------
PROCESS INSTRUMENTS ONLY(NONLABORATORY)
Figure 1. Distribution of measuring instruments observed during user survey.
Sludge density meters, sludge blanket level detectors, on-line respirometers, DO
probes, and many automatic sampling systems use well-established principles that
are suitable for wastewater monitoring and control activities but, based on the survey,
require significant maintenance. These instruments need improved maintenance
characteristics before they will become more widely used.
INSTRUMENT NEEDS
In spite of the many successful flow-measuring devices used in treatment plants, the
accurate and reliable monitoring of storm-water flow poses special problems. High
transient flows, large operating ranges, high suspended solids, and frequent collis-
ions with large debris are merely some of the obstacles that an acceptable in-sewer
storm-water flowmeter must overcome. Consequently a suitable storm-water flow-
meter needs to be developed that will accurately produce the flow rate data required
for sewer regulation.
14
-------
10
15
20
25
NO. OF CASES
30 35
40
BUBBLER TYPE LEVEL DETECTORS
DIFFERENTIAL PRESSURE, LEVEL DETECTOR
FLOATS
ALL OTHER LEVEL DETECTORS
WEIRS AND FLUMES
VENTURIS, ORIFICES, NOZZLES, ETC.
MAGNETIC FLOW RATE
| OTHER FLOW RATE METERS
NUCLEAR RADIATION DENSITY METERS
TRANSMITTING RAIN GAUGES
TEMPERATURE
PRESSURE
ROTATIONAL SPEED
WEIGHT
POSITION
TURBIDITY
CONDUCTIVITY
pH ANDORP
THALLIUM DO PROBE
MEMBRANE DO PROBE
~~l RESIDUAL CHLORINE
OTHER ANALYTICAL SENSORS
GAS MONITORS
SAMPLING SYSTEMS
UNSATISFACTORY
FAIR
SATISFACTORY
Figure 2. Performance summary for measuring devices in wastewater treatment.
15
-------
Some of the most important process parameters (such as total organic carbon and
suspended solids) have never been successfully monitored on an automatic basis in
wastewater-treatment plants. Current studies at the EPA-DC Pilot Plant at Blue
Plains have indicated that Dohrmann-Envirotech TOC analyzer and the Biospherics
suspended solids analyzer are satisfactory. Work at Palo Alto (3) showed that the
Biospherics and the Keene suspended solids analyzers were comparable in perform-
ance. If treatment process efficiency and reliability are to improve, then the follow-
ing analytical sensors—as yet unavailable—are needed to provide real-time control:
• Rapid and automatic on-line total organic analyzers or substrate monitors
• On-line wet chemical analyzers for ammonia and all phosphorus forms.
16
-------
SECTION V
AUTOMATIC CONTROL
Since the flow rate and strength of raw sewage fluctuate diurnally and seasonally,
wastewater-treatment plants operate under variable loading. This may be the prime
cause of effluent quality variation. Storm events and oil, grease, or industrial chem-
ical dumps also upset wastewater-treatment plants. To minimize the impact of these
deleterious disturbances, most treatment plants utilize high-capacity self-regulating
processes that may unnecessarily increase capital and operating costs. However,
major failures such as biomass washout and digester souring still occur all too
frequently. Most investigators believe that the thoughtful application of automatic
process control will improve the operation, maintenance, efficiency, productivity,
and reliability of most wastewater-treatment facilities.
Ideally a control system should accomplish its desired objectives without incurring
any errors; in fact, this is impossible because our process knowledge and control
actions are imperfect. All that one can expect to accomplish is the desired task,
with a minimum amount of disturbances. The goals of wastewater-treatment plant
automation are as follows:
• Improve treatment reliability
• Reduce operating and maintenance costs
• Minimize effluent variability
• Detect problems and institute corrective measures
• Reduce equipment and structure sizes through increased productivity
17
-------
• Find control actions that are readily adaptable by new and existing treatment
plants
• Devise stable control systems.
Undoubtedly very few existing control schemes could simultaneously satisfy all of the
above goals but, with good engineering judgement, the best mix of controller goals
should guide the design engineer.
Inasmuch as most personnel associated with wastewater treatment do not have back-
ground in automatic control theory, a brief explanation of automatic control, its
various modes, and important terminology follows. An automatic control system
consists of the following:
• The process
• Instruments, sensors, flow-measuring devices, and an automatic
analytical device
• Transmission devices to carry the instrument signals
• Controller logic to implement the strategy
• Final control elements to execute the control strategy.
Most investigators view process control as an automatic adjusting of the manipulation
variable so as to maintain the desired balance in the process output stream. The
controller modes can be classified (in the order of their complexity) as:
• Open-Loop Control—This simply involves estimating the form or number
of actions necessary to accomplish a desired objective. It is a predictive
controller, since no check is made to determine whether or not the corrective
action taken has accomplished the desired objective. Figure 3 shows a typical
open-loop control system. Open-loop control is capable of perfect control;
however, if any one of the variables affecting the desired outcome deviates
from its predicted values in either quality or quantity, open-loop control will
not give perfect control. Accordingly, open-loop control is suitable only for
well-known highly repeatable processes such as mechanical equipment or time-
phase events.
18
-------
INPUT
PROCESS
OUTPUT
A. OPEN-LOOP CONTROL
FIXED
••••••••
PROGRAM!
CONTROLLER
MATERIALS NECESSARY
TO GIVE THE
DESIRED OUTPUT
INPUT
B. CLOSED-LOOP
FEEDFORWARD CONTROL
PROCESS
OUTPUT
MEASUREMENT
SENSOR
REFERENCE
POINT
. ADDITIVE
REGULATOR
C. CLOSED-LOOP
FEEDBACK CONTROL
INPUT
PROCESS
OUTPUT
MEASUREMENT MEASUREMENT
REGULATOR SENSOR
COMPUTATION
REFERENCE
POINT
Figure 3. Basic types of process control loops.
19
-------
• Closed-Loop Feedforward Control— One of the newest and most potentially
useful control modes, this has often been referred to incorrectly as open-loop
control. As shown in Figure 3, feedforward control is based on input measure-
ments and computations to predict the amount of additive necessary to produce
the desired product, whereas open-loop control relies only on predetermined
assumptions. Since the process variables are monitored and corrective
actions are taken continually, feedforward control is really a form of closed-
loop control in which a complete loop is formed by the measurement device,
the controller, the additive regulator, and the process.
• Closed-Loop Feedback Control— In feedback control, the controlled variable is
measured and then compared to the reference or desired value. If a difference
or error exists between the actual and desired levels, the automatic controller
will make the necessary corrective adjustments in the final control element.
Figure 3, which contains a typical feedback control loop, illustrates the differ-
ences between open-loop, closed-loop feedforward, and closed-lopped feedback
modes. Feedback control depends only on the repeatability of the controlled
variable measurements and the suitability of the controller's design and adjust-
ments. No prior or predictive knowledge or assumptions are needed for feed
control systems. Because of the technical and economic impracticability of
precisely predicting the number of corrective actions needed to achieve satis-
factory results with open-loop or feedforward control, feedback control is used
most frequently.
• Multi-Loop Control—When several open-loop, closed-loop feedforward, and
closed-loop feedback control loops are united in a single control strategy
appropriate for the pertinent process requirements, the result is known as a
multi-loop control system.
Sections VI and VH make use of all four types of automatic control systems in
developing process control strategy; and, in Section VIII, the merits of the alternate
control strategies are reviewed in the context of achieving the most cost-effective
control scheme.
20
-------
SECTION VI
CONTROL OF DRY-WEATHER TREATMENT PROCESSES
INFLUENT PUMPING AND PRETREATMENT CONTROL
Raw Sewage Pumping
The raw sewage system consists of a pumping station, with either wet and dry wells
or submergible pumps in a wet well. As the sewage flows into the pumping station
and the level rises, the pumps begin in sequence so as to pump the station down to a
nearly empty condition. The control system described herein is for the start/stop
control of two pumps with automatic alternation.
Although the measurement of flow into a pumping station is usually quite difficult due
to the piping configuration, the measurement of flow out of a pumping station is
relatively simply (by using any of the common flow elements). In this case the instru-
mentation consists of a magnetic flow element, a transmitter, a telemetry system, a
flow recorder, and a flow totalizer.
Since pump control normally is based on the level, a level transmitter utilizing a
bubble tube or diaphragm transmits its signal to a level controller containing a con-
tacting mechanism that adjusts the start/stop point on both pumps. It is common to
have both pumps start at different levels on a rising level in the wet well and then have
them continue to operate until they reach a common shutoff point.
Normally the pumps are sized so that a single pump can accommodate all but the most
extreme flows anticipated; even large flows may be handled but, in consequence, the
pump downtime is lengthened. As a result, the second pump is essentially a backup
pump in case the first pump fails or it becomes necessary to remove one of the pumps
21
-------
for service and/or repair. If these pump control systems are wired permanently into
specific contact points, the first pump will be used much more than the second one and
usually will wear out within a few years. This problem is simply resolved by using
an automatic alternator, with each sequence starting on a rising level and terminating
with a common shutdown on a falling level. A simple ratchet relay can be used to
alternate the lead pump's position at the end of each cycle. As shown in Figure 4,
hands-off automatic switches are located on each pump so that either one can be
removed from service or operated manually if required.
There is an almost infinite number of variations possible in this type of station, the
most common being to use a proportional-type level controller in conjunction with a
variable-speed motor-driven pump. Essentially this permits the flow going out of the
pumping station to match the flow coming in. The practical application of these level
control techniques, however, depends on the hydraulics of each station.
MULTI-POINT
Figure 4. Raw sewage pumping control.
22
-------
Screening
The purpose of screening is to remove large floating objects that might otherwise
damage the plant's machinery. Inevitably a great many organic solids are picked up
in the screening process.
As the screen becomes loaded with accumulated materials, a differential liquid level
develops between the upstream and downstream faces of the screen. This differential
level can be used to actuate control of the screen's cleaning mechanism. Using the
differential level makes the system unresponsive to the absolute level within the
channel, which can vary widely and independently from the differential level. It is
important to note that the cleaning cycle should be so interlocked with associated
devices that, once the cycle has been initiated, it must be completed in one operation.
The level measurement is accomplished by means of bubble tubes located on opposite
sides of the screen. Figure 5 illustrates the associated instruments and control
devices.
UPSTREAM
DOWNSTREAM
Figure 5. Screen control.
23
-------
Variations of this arrangement have been in use for years. The same fundamental
principle was followed in rather crude mechanical devices, consisting of heavy cast-
iron floats and chain arrangements in float wells located on either side of the screen.
PRIMARY SLUDGE PUMPING
Introduction
The treatment of wastewaters by gravity separation is among the oldest methods used
for solids/liquids separations. Decades of primary treatment plant operation have
produced much practical knowledge and many "cookbook" operating techniques that
lead to effective removal of the settleable solids in sewage. Approaches to success-
ful treatment plant operation must consider the following:
• The specific gravity of raw sludge is only slightly greater than 1. 0.
• Efficient sludge collection for subsequent treatment and disposal requires
that the minimum possible quantity of water be removed with the sludge.
• Sludge must be removed continuously or frequently in order to minimize
septicity and excessive loading on removal mechanisms.
The process dynamics of primary sedimentation are relatively simple as far as
control is concerned. The basic representation is as follows:
cj i^" //""\ o Of \ t~\ \
SL-K(QI,Ss>%R) (1)
where:
Q = influent rate
S = settleable solids concentration
s
= % removed
S = sludge settled formation rate
K = units constant
24
-------
Equation (1), which represents the sludge formation rate, shows that over long periods
the sludge formation rate will equal the removal rate. However, these rates are not
balanced over short intervals, since sludge is usually settled and allowed to compact
in the clarifier before being removed by pumping. For a reasonable period of time
(e.g., 12 to 24 hours), the following applies:
S = S_ (2)
P L
where:
S = sludge pumped
P
The principal operating objective of primary clarifiers involves keeping the removed
sludge above some minimum density level, which is usually expressed as percent
solids. This ensures the maximum practical capacity for subsequent treatment, since
excessive quantities of water overload sludge treatment facilities and seriously impair
their efficiency. This relationship is represented by the following:
D > minimum value ' (3)
s
where:
D = sludge density (% solids)
S
The relationships and practices discussed in this section have been either a common
practice or a common operating objective for so long that current technical literature,
with the exception of teaching texts, no longer deals with this process phase. However,
the implementation of these common operating practices by means of automatic instru-
mentation and control equipment is recent. The key to successful control is the avail-
ability of reliable measuring and control instrumentation.
Since the only final control elements available are the raw sludge pumps and diversion
valves, the control modes for raw sludge pump are limited to on-off and variable-
speed sludge removal. Most of the small-and medium-size installations that use
pumps rather than diversion valves operate these pumps on an on-off basis, since
25
-------
variable-speed pumping can result in low line velocity which favors clogging. Conse-
quently, the presently available control techniques are not especially enhanced by the
availability of variable-speed pumps as opposed to the on-off operation of constant-
speed pumps.
Percent of Time Control
The simplest and most commonly used method of sludge withdrawal control is the
operation of the sludge pumps, based on a percent of time. This can be expressed
as follows:
Sp = f (t) (4)
where:
t = time
In practice a percentage timer or interrupter is normally used. Such a device usually
has a fixed cycle length (e.g., 60 minutes), and the percent of on-time is adjustable.
On the basis of operational tests and practical experience, the plant operator sets the
percent of on-time (e.g., 20%). The 60-minute cycle would result in the raw sludge
pump's operating 12 minutes out of each 60 minutes. The pump would run for 12
minutes, shut down for 48 minutes, and then repeat the cycle.
Since the percent-of-time control scheme is an open-loop control strategy, variable
influents, settling conditions, and the behavior of the other sludge removal equipment
(as well as other factors that impact sludge density) will change the desired duty cycle.
The percent of on-time would then be manually adjusted up or down to produce the
desired concentration of solids in the raw sludge. The simple percent-of-time control
strategy suffers from being insensitive to plant equipment and process conditions,
and requires frequent operator attention in order to produce a high-quality sludge.
26
-------
Percent of Time Control With Flow Proportion
If the settleable solids content of the influent sewage were constant, then withdrawing
settled sludge at a rate proportional to the influent flow would be adequate, since this
method overcomes the influence of the largest variable flow. However, it does not
consider variations in settleable solids concentrations of the plant influent, changes in
the other sludge-processing equipment, and process upsets in the primary clarifier.
Because the implementation of flow-proportioned sludge withdrawal normally involves
operating the sludge pumps at variable speeds, incipient line settling and plugging can
be overcome by using the pulse duration pump control. With this technique the flow-
meter is equipped with a cam switch that has a fixed cycle (e. g., 30 seconds). The
portion of the 30-second cycle when a circuit is closed is proportional to the flow
measurement; for example, at 50% of flow span, the circuit will be closed for 15
seconds and open for 15 seconds. Since these time periods are too short for pump
operation, the amount of on-time must be accumulated on a timer which usually is set
for a time representing a fixed volume (e.g., 100, 000 gallons). When sufficient time
(wastewater volume) has accumulated, a second timer is actuated to operate the sludge
pumps for a fixed period.
If the variation in waste flow is wide, it may be desirable to combine time and flow
proportional control. This would be accomplished by having the sludge pumps operate
on flow proportional control during the day and early evening. During the extremely
low flow period at night, the sludge pumps would operate on straight timer control,
since a straight flow-proportioned control would allow the clarifier sludge to become
septic.
27
-------
To our knowledge, the flow proportional method of sludge pumping is not in service
in any existing facility. Nevertheless, the method has merit for small installations
where more elaborate devices requiring continuing maintenance and skilled operations
would probably fall into disuse.
Feedback Sludge Density Control
The commercial availability of radioactive sludge density meters opened new horizons
for sludge pumping control. Although the application of radioactive source density
measuring devices has become quite common, service and maintenance problems can
be difficult. With proper installation and a reasonable preventive maintenance program,
however, radioactive sludge density measuring devices provide useful information and
may be included in an automatic control loop.
Feedback sludge withdrawal control uses the actual sludge density to regulate pumping
so as to ensure that the density is above some minimum value. Sludge pumping control
can be accomplished by using a combination of timers in conjunction with the density
measuring devices, as shown in Figure 6. The sequence of operation is as follows:
• The raw sludge pump is started at a preset interval determined by a cycle
timer.
• A second timer holds in the pump control circuit until a fresh sample of
sludge is in the density element pipe section.
• The raw sludge pump will continue to operate until it measures a density
that is below a preset minimum value. At this point, the sludge pump is
stopped.
• The cycle is repeated at regular intervals, as determined by the cycle
timer.
28
-------
TO SLUDGE
TREATMENT
Figure 6. Sludge density control.
Feedback sludge density control offers the following advantages:
• Withdrawn sludge does not fall below the predetermined minimum density.
• Some sludge is pumped at regular intervals, which helps to minimize the
development of septic sludge in the clarifier.
• Excessive sludge loading cannot build up to the point where the sludge
removal mechanism would be damaged.
• Sludge pumping control automatically reacts to changes in the influent,
the clarifiers process conditions, and the sludge removal equipment.
The importance of minimizing the waste sludge cannot be overemphasized. Since the
volume of sludge varies as the percent solids and not percent liquid fraction, a 1%
change in percent solids will have a dramatic impact on sludge volume. For example,
a change in solids at raw sludge from 3% to 4% (97% to 96% water) will result in a
29
-------
volume reduction of approximately 33%. Regardless of the final method of sludge
treatment, the size of equipment and the cost of operation are adversely influenced by
a sludge that has a low percentage of solids. This is especially true in the case of
digesters.
Interface Control
To prevent sludge accumulation and subsequent solids carryover, the sludge waste-
water interface can be used to regulate sludge withdrawal. In the past, operators
did this manually by noting the staining of a piece of Turkish towel which they dipped
in the clarifier with a pole.
The recent development of an optical technique for measuring the position of the sludge
waste interface has made this a viable automatic control technique. A sensing device
consisting of a light source and a sensor that are a few inches apart is mounted on a
carriage, which is motor driven vertically so that it passes through the possible area
of sludge wastewater interface. Passing through the interface breaks the light path,
and the position of the sensor at that moment corresponds to the top of the interface or
the position of the settled sludge.
The availability of periodic measurements of the settled sludge interface permits
automatic feedback control of sludge pumping, as illustrated in Figure 7. The sludge
pumps start on a fixed cycle and shut down on the basis of the sludge water interface
falling below a preset limit. This technique is essentially the same as that used with
density measuring devices, but it ensures that the sludge blanket stays within a safe
region of the primary clarifier.
To effectively regulate withdrawn sludge density as well as the sludge blanket depth,
the sludge density and sludge interface control schemes can be combined. If a conflict
exists (such as a rising blanket and too low a sludge density), then an interlock control
system can alert the operator so that he can decide the course of action to take.
30
-------
TO MOTOR CONTROLS
OPTICAL SLUDGE
LEVEL DETECTOR
^TO SLUDGE
TREATMENT
Figure 7. Sludge interface control.
Application Notes
These control systems are all highly stable because they all operate as successive
batching techniques. Inasmuch as any system or components thereof can fail, the
impact of such failures must be examined. Automatic sludge pumping failures fall
into the following three general groups:
• Failures resulting in the complete suspension of sludge pump operations.
• Failures resulting in the partial suspension of sludge pump operations;
such failures occur intermittently and are due principally to primary
element failures.
• Failures resulting in continuously operating sludge pumpsj these failures,
too, are due principally to primary element failures.
31
-------
A complete failure of the first type is usually detected by the plant operator before
any major problem develops. If problems do occur, however, they consist usually of
overloading the scraper mechanisms, which can be detected by the motor load
detectors or torque switches.
Partial failures are much more difficult for the operator to detect because they are
intermittent. However, they ultimately produce the same results as complete failures
and are detected in the same manner.
Shutoff failures are the most important, since they can completely upset the sludge
treatment and disposal system. In the case of sludge digesters, the digester becomes
overloaded and digester failure can soon follow. The only practical method of detect-
ing a failure wherein the sludge pumps run continuously is to arrange a timer so that
an alarm sounds whenever a normal period of pump operation is exceeded. Fortunately,
shutoff failures are rare.
Automatic control of raw sludge pumping permits the process regulation of primary
clarifiers, as summarized in Table 2. These methods provide an economical solution
to the critical problem of properly controlling sludge pumping and represent a marked
improvement over manual operation, since they minimize the costs of pumping,
dewatering, and digesting sludge.
CONTROL OF DISSOLVED OXYGEN IN AERATION BASINS
Today most secondary treatment facilities use aerobic biological processes, such as
conventional activated sludge, step aeration, completely mixed, and contact stabiliza-
tion, to remove soluble and colloidal organic pollutants. One of the most important
requirements for achieving high BOD removals is maintenance of proper dissolved
oxygen (DO) levels in the aeration vessels. If the DO drops below a critical level
the effluent quality deteriorates. Excessive DO concentrations can hinder secondary
32
-------
Table 2. SUMMARY OF PRIMARY SLUDGE PUMPING CONTROL STRATEGIES
Control Method
Benefits and
Potential Savings
Advantages
Disadvantages
Time cycle
Time cycle with flow
proportion
Sludge density with
auxiliary timers
Sludge level with
auxiliary timers
Better separation and 10
to 30% labor savings
over manual control
Same as above
20 to 50% less labor
than manual opera-
tion; reduction of
water pumped with
sludge can extend
sludge handling
capacity
Same as above
Simple, inexpensive,
reliable
Reliable and responsive
to flow rate changes
Good, reliable separa-
tions; densities mea-
sured and consistently
controlled; most free
water kept out of
sludge
Good, reliable separa-
tions; sludge level
maintainable at best
level; low cost
Depends on experience
and judgement for
adjustments (can pass
sludge in overflow and
water in underflow)
Requires flow signal and
more instrumentation
than time cycle alone
Unrealistic repair
delays and downtimes,
although basic instru-
ment is good; expen-
sive; some water
drawn off with/sludge
Procedure not fully
developed; probe fail-
ure rate high and
should be better;
influence partly by
color and turbidity
solids settling and consume large amounts of energy (4). Experience shows that a DO
level of about 1. 0 mg/1 promotes process stability, high BOD removals, and minimum
aeration energy consumption.
Since the microorganisms consume oxygen according to their metabolic requirements,
the mixed liquor must be aerated to maintain a satisfactory DO level. Diffused aerator
devices and mechanical aerators are frequently used to transfer oxygen to the mixed
liquor, and power consumption is directly proportional to oxygen demand.
To maintain a constant DO level the oxygen supply rates should match the oxygen up-
take rates. Because the uptake rates change continually due to variations in the
organic loads, microorganism concentration, and degree of nitrification, it is difficult
33
-------
and time consuming to manually adjust the aeration equipment. Instead, most plants
transfer more oxygen than necessary to save manpower and to ensure an adequate DO
level.
Automatic DO control systems can modulate the rate of oxygen supply to the mixed
liquor by adjusting the aeration equipment's power consumption. A good DO controller
should hold the DO at the prescribed level in spite of changing loads and ambient
conditions. The control system, also needs to be stable and easy to maintain. A
nationwide survey of instrumentation experiences (1) found that 20% of the secondary
treatment plants used automatic DO control. Of the users about 70% were satisfied
with automatic DO controller performance, and they reported power savings from 10
to 40% and BOD removal increases of about 10%.
The subsequent text describes automatic DO control systems that have been success-
fully practiced in wastewater-treatment plants and some promising—but untried-
control strategies. Application notes and instrument diagrams are also included.
DO Control Systems
Aeration basins are classified as plug flow or completely mixed, based on flow
patterns, geometry and mixing characteristics. Plug flow vessels (large length-to-
width ratios) are used in the conventional activated sludge, step aeration, and contact
stabilization processes, whereas the completely mixed activated sludge processes
modification is practiced in rectangular or circular vessels. Aerated lagoons, which
are popular for industrial wastewater treatment, exhibit mixing characteristics that
fall in between plug flow and complete mix. Because of the difference in contacting
patterns and oxygen consumption rates, DO control strategies depend on the type of
aeration basin and the activated sludge process.
Inasmuch as most new treatment plants use the completely mixed activated sludge
process and plug flow basins behave like several completely mixed vessels in series,
34
-------
the following text explains the rationale behind DO control schemes for a completely
mixed aeration basin. Moreover, a diffused air system is used in the strategy
development. Extensions to plug flow aeration basins and aerated lagoons, as well as
to other types of aeration equipment, follow directly.
A material balance around a completely mixed aeration basin gives a relationship
between dissolved oxygen concentration, influent flow rates, respiration rate, and
airflow rates:
/dC\
v I —°|= FC - FC + kl (C - C ) U V - R V (5)
\ dt / i o a s o
where:
C = aeration basin DO concentration (mg/1)
o
C. = influent DO concentration (mg/1)
F = flow rate (liter s/min)
kl = transfer coefficient (liters)
a
U = airflow rate (liters/min) at standard conditions
C = saturation value of DO at system temperature and pressure (mg/1)
s
R = respiration rate or oxygen utilization rate (mg/1 min)~^
V = volume of aeration basin (liters)
Rearranging terms, the oxygen balance becomes:
dC C. - C
—- = kl (C -C)U-R9+ — (6)
dt a. s o' 6
where:
-5
35
-------
The respiration rate, R, is related to the substrate utilization rate and the endogenous
respiration rate. Eckenfelder (5) expressed this relationship by the following equation:
R = Q' S + (3 X (7)
r vss
where:
a - coefficient relating substrate removal to oxygen consumption
S = substrate removal rate
r
(3 = endogenous constant rate
X = mixed liquor volatile suspended solids
VSS
A substrate balance around the aeration basin yields a relationship between loadings
and removal rates:
VdS
, °= FS - FS - S V (8)
dt i o r
Several kinetic models such as two phase, second order and Monad (Michaelis-
Menton) have been proposed to describe substrate removal rates. In general, the
reaction rate increases with substrate concentration. This control system analysis,
however, will not depend on a particular kinetic model.
Equations (6), (7) and (8) were consolidated into the following overall relationship in
terms of the fundamental parameters:
dC ["(S. -S)dS~| C-C
—5 = ki
-------
Rearranging Equations (6) and (9) in such a manner as to give the airflow require-
ments as a function of system variables yields Equation (10):
U =
„
R e +
c - c.
O 1
e
- s -
o
dC
0
|
dt
1
kl (C - C )
J a s o
ds e \ c
0 i e- n i c
dt / vss
- C.
' ]
e
(lOa)
dc
dt
kl (C - C )
a s o
a Ob)
The airflow rate requirements, according to Equation (lOa), increase with respiration
rate, dissolved oxygen concentration, and the rate of DO change in the basin. Since
ambient and wastewater characteristics change C and kl , air requirements also
s a
vary according to these factors. Equation (lOb) shows that increased substrate load-
ing and higher mixed liquor volatile suspended solids also raise the airflow require-
ments. Similar equations and conclusions are obtained for all types of aeration
devices and process configurations but, instead of air requirements, the energy
consumption changes as a function of the previously mentioned parameters. These
findings agree with operational experiences observed in activated sludge plants and
aeration lagoons.
Many manufacturers supply oxygen transfer equipment; this equipment is broadly
classified as diffused air devices, mechanical aerators, or combinations of both.
Diffused aeration systems are further categorized, based on the type of dispersion
media as porous diffusers (fused aluminum oxide) or non-porous diffusers (spargers,
ridge and furrow). Updraft, downdraft, plate and brush are the major classes of
mechanical aerators. For all types of aeration equipment, the power consumption
is directly proportional to oxygenation output.
Selecting the manipulation variable involves examining the input/output relationships
of the aeration equipment for the purpose of attaining the control objectives. Available
final control elements which execute the control strategy also influence the choice of
37
-------
the manipulated variable. Specifically, the manipulation variable must not only
regulate the DO level, but it must reduce energy consumption as well. For example,
consider a centrifugal blower with operating characteristics, as shown in Figure 8.
Further assume this blower is being driven by a single-speed induction motor. Clearly,
airflow rate under standard conditions (equivalent to mass rate of air, kg/min) is a
prime manipulation variable; however, the point of application and the implementation
method are crucial. Throttling the discharge line will reduce the DO in the aeration
basin, but little or no power saving will be obtained since the throttled valve dissi-
pates the excess energy. Consequently, throttling the discharge of a centrifugal
blower is an unacceptable point of application. To achieve good DO regulation and to
effect power saving, the inlet (low pressure) line should be throttled. If the blower
was equipped with inlet guide vanes, adjustment of these vanes would result in even
greater power savings. The final control element would be a valve with an automatic
operator for the throttling case and an automatic positioner for the inlet guide vanes
case. Both methods would regulate airflow to the inlet side of the blower.
As a second example, consider lobe-lype blowers. Since these devices are designed
to work over a relatively narrow pressure range, throttling the inlet or discharge will
not effectively reduce the airflow rates, and the power consumption will remain about
the same. Lobe blowers, consequently, should be equipped with multiple-speed
driving motors for good DO control and power savings. However, in large plants that
use four or more blowers, simply turning the blowers on or off may be a suitable
method for regulating airflow rates and single-speed motors would suffice. A certain
minimum power should be applied to each aeration basin to ensure good mixing and to
keep the biomass in suspension. The aeration equipment must also operate within a
prescribed range for safety reasons (surging).
In order to select the manipulation variable and points of application, the type of
driving motor, plant size, system configuration, power consumption relationships for
keeping the biomass in suspension, final control elements and control objectives must
38
-------
120
I I
VANES WIDE OPEN
0 10 20 30 40 50 60 70 80 90 100 110 120
RATED VOLUME AT CONSTANT DENSITY (PERCENT)
u
Cf
3
c
a.
LU
CA
rc
O
0 10 20 30 40 50 60 70 80 90 100 110 120
RATED VOLUME AT CONSTANT DENSITY (PERCENT)
Figure 8. Centrifugal blower performance curves showing effects of inlet
guide vanes. (Numerals 1-5 indicate five different settings of
vane s.)
39
-------
be considered. Table 3 contains a list of possible manipulation variables and asso-
ciated final control elements for mechanical turbine and diffused aeration equipment.
Although the DO control loops can interface directly with final control elements,
cascade control loops (both outer and inner loops) are highly recommended because
they increase the system's responsiveness and stability. The outer loop adjusts the
air flow rate to the aerator. The inner loop compensates for secondary variations
and lessens the workload of the DO controller. For example, consider a diffused
aeration system where airflow rate at standard conditions is the manipulation variable.
Without an inner flow loop, changes in air flow demand could cause excessive demand
on its centrifugal blower and surging would result. If its inner loop controlled
pressure and sensed temperature, surging could be prevented. For some aeration
equipment, such as multi-speed blowers or multi-speed mechanical aerators, the
inner cascade loop is nothing more than a simple stepping relay. Other equipment,
such as centrifugal blowers, requires rather elaborate inner control loops to throttle
airflow rates and prevent surging.
Table 3. POSSIBLE MANIPULATION AND FINAL CONTROL ELEMENTS
Aeration Equipment
Manipulated Variable
Final Control Element
Air diffusers
Airflow rate
Submerged aerator
(turbine /or ifi ce)
Surface aerators
Airflow rate
Turbine speed
Immersion depth
Speed
Automatic inlet valve
Automatic inlet valve positioner
Variable-speed motor
Multiple-speed motor
On-off control for multiple
blower installations
Same as air diffusers
Multiple-speed motor
Adjustable weir
Adjustable blade depth
Multiple-speed motor
40
-------
Proposed Control Strategies
Measuring flow rates, MLVSS, respiration rates, and dissolved oxygen demands
would permit virtually perfect control. However, the wastewater characteristics
(a, |3, and kl ) continually change in most treatment systems, respiration rates are
a
very difficult to measure accurately, arid DO demand measurements are hindered by
noisy DO signals.
Since oxygen demand depends on flow rate to some extent, raw sewage flow rate was
the earliest proposed control parameter for regulating oxygen transfer rates. The
flow-proportioned control system (Figure 9) is an open-loop controller which auto-
matically maintains a constant ratio between the raw wasteflow and airflow at standard
conditions. If the strength of the sewage remains constant, a flow-proportional DO
SETTLEDf
SEWAGE
TO
SECONDARY
CLARIFIER
VENT
Figure 9. Flow-proportional DO control for completely mixed activated sludge.
41
-------
control system will track oxygen demand quite well. However, since most raw
sewage strengths change dramatically during a 24-hour period, the ratio must be
frequently readjusted, requiring considerable manpower. Instead, most facilities
2
simply choose a high ratio (typically 1. 0 to 1.5 feet air/gallon of wastewater) to
prevent septic conditions and reduce the adjustment manpower requirements. Aeration
power savings, accordingly, are quite small, usually about 10%. Inasmuch as raw
sewage flow provides only an indirect estimate of oxygen demand, which cannot dis-
tinguish between sanitary wastes, storm water, or infiltrated ground water, flow-
proportional DO control is not suitable for most municipal and industrial treatment
plants. It is useful only for small industrial plants that have constant strength waste-
water.
With the availability of in situ DO probes, feedback control of an aeration basin's
dissolved oxygen became practical. Inspection of Equations lOa and lOb shows that
airflow rates or power requirements vary directly with the DO operating level.
Additionally, Pontryagrin's Minimum Principle says the manipulation variable should
be either at the maximum or minimum value when an error exists; otherwise, the
steady-state control action is best. Accordingly, optimal control theory requires
error data also.
Good DO feedback controllers use high-gain proportional action to give fast responses
plus an integral mode to eventually generate suitable steady-state control in the absence
of changing oxygen demands, as illustrated in Figure 10. This control strategy, which
has many of the desired features indicated by optimal control theory, minimizes
secondary disturbances and increases responsiveness by means of a cascaded local
loop around the final control element. For diffused air this might be an airflow con-
trol loop, or a submergence depth control loop for surface aerators. When a positive
error exists (low DO), the feedback controller responds by increasing the airflow rate
42
-------
SETTLED,
SEWAGE r
r
( AIT
DO
AERATION BASIN
* ivAEyj * * i 4 t * *
TO
-^SECONDARY
CLARtFIER
AIR MANIFOLD*
Figure 10. PI feedback DO control for completely mixed activated sludge.
setpoint (power consumption) and vice versa regardless of the cause of the DO disturb-
ance. Mathematically, the manipulation variable follows Equation 11:
U = k (e) +
/ edt
•'o
(11)
where e equals the desired DO minus the actual DO and k , k represent the adjust-
1 2
able controller gains.
If process conditions give rise to sporadic-type DO readings, it may be necessary to
filter or smooth the probe reading prior to taking control actions, since feedback DO
43
-------
controllers respond to disturbances, regardless of causes. Other types of feedback
controllers, such as two position or differential gap, may also be useful in some
cases.
PI feedback DO controllers are the most popular form of DO regulation practiced
today in wastewater-treatment plants. During the nationwide plant survey (1), most
users of this control scheme reported 25% reductions in aeration energy consumption.
Of the six installations observed, four worked satisfactorily and two were considered
unacceptable. The principal problem with feedback DO control is the maintenance
requirements associated with the DO probe. About 60 manhours are needed annually
to inspect, clean, and calibrate each DO probe.
Feedforward control strategies can generate suitable control actions without incurring
sizable errors characteristic of sluggish feedback control systems. All feedforward
strategies, by definition, are responsive to incoming load conditions rather than
waiting for the resultant error before taking corrective action. Based on Equation
10, the forthcoming control strategies neglect the derivative terms and respond to
changing influent demands. The feedback control trims the feedforward strategy to
compensate for model inaccuracies, measurement errors, and changing ambient
conditions.
If a PI controller replaces the respiration rate and derivative terms in Equation lOa,
the airflow rate (power consumption) is given by:
U = k F (S. P.) + k e + k I edt (12)
-L £ o J
where k , k , k are adjustable controller gains and S. P. = the DO setpoint.
J. & i3
As illustrated in Figure 11, airflow rate regulation (power consumption) is propor-
tional to the influent flow rate and the error in the aeration basin's dissolved oxygen.
If the influent flow rate increases, the feedforward loop increases the airflow. When
the basin's DO level is low, the feedback loop reinforces this action; otherwise, the
44
-------
SETTLED
SEWAGE
1 AIT
DO
AERATION BASIN
,00
t »V J\ * » » » I I
TO
-^SECONDARY
CLARIFIER
~©
AIR MANIFOLD h
Figure 11. Flow proportional feedforward DO control with feedback trimming.
feedback control trims the airflow. Similar comments apply to falling influent flow
rates. This type of control tends to eliminate diurnal DO oscillations associated
with varying flows, although storm-water surges and groundwater infiltration may
cause false increases in airflow rates (power consumption).
Because most treatment facilities already have influent flow rate-measuring devices,
this control strategy requires the same equipment and maintenance as the feedback
DO control loop. In some plants, flow distribution networks may necessitate additional
flowmeters, but this appears unlikely. Under EPA sponsorship, evaluations of this
control strategy at Palo Alto, California, indicated that power consumption savings
were 11%.
45
-------
Flow and Mass Loading
Recent improvements in rapid on-line organic measuring systems make automatic
substrate determinations a near-term reality. On-line total organic carbon (TOC),
total oxygen demand (TOD), and respirometer (organic demand, OD) instruments are
available but require further demonstration. Total organic carbon data in conjunction
with influent flow rates can provide estimates of the organic loading (F * S.) on the
aeration basin. Since the oxygen demand increases with the organic loading and
influent flow rate, aeration power requirements can be determined by rearranging
Equation lOb, where PI control replaces the derivative terms:
U = k F (S.P.) + ct F S. + k e + k / edt (13)
-L 1 £t o J
O
where k » k , k are adjustable controller gains, and S.P. = the DO setpoint.
1 A o
As illustrated in Figure 12, the feedforward loop measures the total organic load in
the influent and regulates the airflow rate accordingly. When the load increases, the
feedforward loop calls for more air, but this action is trimmed on the basis of current
and past DO levels. If the actual DO is lower than desired, the feedback loop augments
the feedforward action or, if the DO level is greater than the setpoint, the feedback
loop opposes the forward loop. Although operating experience is necessary to select
3 ^
the appropriate a ratio (ft air: Ib of organic), 700 to 1000 ft air/lb of BOD is a
typical value for diffused aeration systems. The selected value, however, depends on
the operating DO level, mass transfer coefficients, type of treatment, and degree of
BOD removal.
Inasmuch as on-line organic monitors are new and still undergoing evaluation in
wastewater-treatment plants, this control strategy is untried. Engineering estimates
show, however, that 35% power savings are attainable for mass loading feedforward
feedback dissolved oxygen control systems.
46
-------
PRIMARY
EFFLUENT
INLET
I LOWER AIR LIMIT
1 FOR AERATION
BASIN
FYLM'NIMUMAIRFLOWRELAY
I
I INNER CASCADE LOOP
CENTRIFUGAL .
BLOWER
AIRFLOW RATE = »(Flow rate) (Infl. TOO
TO
SECONDARY
CLARIFIER
e dt
INTERLOCK
where e = error in 00
o,kj.l<2 = constants
inicnLutix I
j SURGE CONTROL
Figure 12. TOG feedforward DO control with feedback trimming.
Respirometers measure the oxygen utilization rate of the aeration basin. If the
aeration equipment dissolves oxygen at a rate equal to the utilization rate, the
aeration basin's DO level would remain essentially constant. Adding enough additional
oxygen to raise the incoming wastewater DO to the operating level ensures an exact
balance between oxygenation capacity and demand. Then the DO remains at the desired
level. From this concept, on-line respiration rates and flow data can regulate aeration
energy consumption according to the following equation, which is derived from Equation
lOa:
V + k3 / 6dt
(14)
47
-------
Including PI feedback control makes this strategy less dependent on modeling and
measurement errors; moreover, the integral mode eliminates permanent offsets.
Because respiration rate data are approximately equivalent to derivative action, this
strategy anticipates aeration basin DO changes and adjusts the airflow rate (energy
consumption) accordingly.
As shown in Figure 13, the airflow rate increases with respiration rates and may
increase or decrease with flow rates in the forward loop; the feedback loop either
reinforces or trims the forward loop control action in such a manner as to reflect the
current state of the aeration basin.
AIT V
SETTLED
SEWAGE
I AERATION BASIN
DO
it I * » »
TO
-i SECONDARY
CLARIFIER
AIR MANIFOLD t
Figure 13. Respiration rate DO control with feedback trimming.
48
-------
The oxygen demand equation (Equation 10) can be readily solved for the respiration
rate, R, if the airflow rate, residual DO, and transfer coefficient are known; more-
over, derivatives of the oxygen concentration can be numerically estimated. Practical
considerations, however, dictate using a small on-line computer or time-sharing with
a larger facility for estimating the respiration rate. Implementation of calculated
respiration rate control must follow the difference equation form of Equation lOa:
Co. + 1 - Co.
where
SP - Co. + e.
3 + 1 J +
edt
j + 1
N = 1
(15)
e (0) = 0
The PI control mode was added to ensure convergence to the setpoint. Although
influent flow rate, airflow rates, and transfer coefficients are needed to calculate the
respiration rate, when the expression for this is substituted into Equation lOa, these
terms cancel. Since Equation 15 is the digital approximation for three-mode propor-
tional integral derivative (PID) control, calculating the respiration rate is equivalent
to approximating the aeration basin's first derivative, and the resulting control
strategy becomes the classical feedback PID controller. Unlike the direct respiro-
meter measure technique, this method adds very little information about system
behavior and is not a true feedforward control strategy. The benefits and maintenance
requirements are virtually the same as PI feedback DO control.
49
-------
Conventional Activated Sludge—Plug Flow Configuration
Many secondary wastewater-treatment plants are built using a long, narrow, multi-
pass aeration basin which has become known as the plug flow configuration. In
actuality, the backmixing caused by the rolling action of aeration units makes this
type vessel behave like several completely mixed reactors in series. Consequently,
completely mixed analyses extend to the plug flow rather directly. The key question
becomes: how many tanks in series are needed to represent the plug flow basin ?
Previous investigators (6) have used from two to ten vessels to represent the conven-
tional activated sludge process. Since the oxygen demand changes along the length of
the operation basin, as shown in Figure 14, finding the optimum number and location
of DO probes is a difficult task. At the head or inlet of the aeration basin where the
return biomass and primary effluent are mixed, the organic concentration (and hence
the oxygen demand) is highest. It is important to supply enough DO so as not to in-
hibit stabilization; a DO of approximately 0.5 mg/1 should be sufficient. As biological
stabilization proceeds along the length of the aeration basin, the oxygen demand
2.0
<
a •-
>• S
5 *=
o <
1.5
1.0
0.5
0 50
LENGTH ALONG AERATION BASIN (PERCENT)
Figure 14. Variation in oxygen demand with basin length.
50
100
-------
diminishes. If properly designed, the oxygen demand of the outlet end is small and
a DO concentration between 1. 5 and 2. 0 mg/1 is desired.
The selection of a control strategy and DO probe placement depends heavily upon the
type and flexibility of the aeration system. When a uniform aeration intensity is used,
the oxygen profile is virtually fixed, and only the level (rather than the shape) can be
altered via automatic control. In the influent zone, oxygen demand is high and the
resultant DO level is low; farther down the aeration tank, the oxygen demand is lower
and the resultant DO is high. If a separate aeration source which can be regulated
exists for each pass, then more satisfactory control is possible because it can
regulate power consumption and the DO level for each pass. In the following para-
graphs some notes on the application of automated DO control to plug flow aeration
are provided.
Plug flow aeration basins are designed as single- or multiple-pass units, and the
oxygen transfer equipment is either uniformly applied or divided into separately
adjustable sections. Single fixed output systems have little flexibility and do not lend
themselves to automatic regulation. Although aeration theory maintains that the head
section has the highest oxygen demand and consequently should be the best location for
DO probes, alternate probe placement should be planned so that the operator by trial
and error techniques can find the optimum probe location. Recommended control
strategies for this type (listed in Table 4) are simple to tune and keep operational.
Separately adjustable aeration systems can be considered similar to several completely
mixed systems in series. Clearly the probes for the respective control systems should
be placed in their relevant zones of influence. Here, correct probe placement is
important, but not as crucial as with uniform aeration. Control strategies (listed in
Table 5) range from simple techniques requiring no special equipment to sophisticated
control systems needing sophisticated equipment and maintenance. Potential power
savings usually increase with the complexity of the control system. The best strategy
depends on many factors, such as plant size, power costs, and maintenance labor.
51
-------
Table 4. SUMMARY OF DO CONTROL STRATEGIES FOR PLUG FLOW AERATION
BASINS WITH UNIFORM OXYGEN TRANSFER
Description
Flow Proportional
PI feedback
DO control
Aeration
Power
Savings
5%
10%
Advantages
Simple, straight-
forward, easy to
maintain
Aeration power
responds to aera-
tion basin DO level
Disadvantages
Does not respond to
change in organic
Controls DO at a single
point only along the
aeration basin
Re co mmendation s
Suitable for small plants
with uniform oxygen
Good controller
for medium-size and
large plants
Economic analyses are applied as decision making aids in a subsequent section,
where specific control strategies are recommended.
One method of automatic DO control is flow ratio control, but adjusting aeration
equipment capacity in proportion to influent flow rate provides only marginal savings
since the wastewater composition undergoes considerable fluctuation in a 24-hour
period. Moreover, such a control system is prone to enormous errors during storm
events and when infiltration occurs. As in completely mixed systems, only about 10%
power savings can be expected and poor DO regulation is customary with this type of
control system.
In the case of feedback control, in situ DO probes, located in each zone of influence
coupled with PI control, pace aeration capacity in accordance with local DO levels.
For example, a four-pass diffused aeration system with each pass capable of separate
regulation can keep the dissolved oxygen level of each pass at the prescribed level.
In concept, each pass is treated as a separate completely mixed vessel. Sufficient
flexibility must be provided for placing the probe anywhere in the aeration pass. Plant
operators, by virtue of trial and error probe locations, will find the optimum probe
placement for each pass. When surface aerators with speed or individual submergence
controls are installed, the DO probe should be located in each aerator's zone of influence.
52
-------
Table 5. SUMMARY OF DO CONTROL STRATEGIES FOR PLUG FLOW BASINS
WITH SEPARATELY ADJUSTABLE OXYGEN TRANSFER DEVICES
Description
Aeration
Power
Savings
Advantages
Disadvantages
Recommendations
Flow proportional
control
5%
Simple, straightforward,
easy to maintain
Does not respond to
changes in organic
strength
Suitable for small uniform
aeration intensity plants
PI feedback
DO control
15%
Aeration power responds
directly to actual basin
DO level
May tend to oscillate due
to flow variations
Good controller for mod-
erate sized, uniform,
and separately adjust-
able aeration systems
01
CO
Flow proportional
with feedback
trimming
20%
Reacts to flow rate varia-
tions and adjusts the
aeration intensity pro-
file accordingly
May be difficult to tune
this controller. High
infiltration rates or
storm events may lead
to poor control
Applicable only for large,
separately adjustable
aeration systems
Organic loading
with feedback
trimming
25%
Responds to flow and
strength variations in
the influent stream
This untried strategy
requires an on-line
organic monitoring
system which may be
difficult to maintain
Useful in large, highly
automated plants, espe-
cially with F/M control,
that have separately
adjustable aeration sys-
tems
Power forward
with feedback
control
22%
Responds to flow and
variations in the first
aeration section; con-
sequently, DO demand
changes are anticipated
for the later sections
May not be any more
beneficial than PI
controller
Suitable for moderate
sized, separately
adjustable aeration
systems
-------
PI feedback DO control systems of separately adjustable aeration systems can reduce
aeration cost by 25%; each control loop requires 60 manhours of maintenance per year
(the same as completely mixed feedback loops).
A uniform aeration system has less flexibility since only the total power can be
regulated rather than a series of individual aerators. Automatic regulation of this
type of system ctmsequently can only raise or lower the entire DO profile. Inasmuch
as the plug flow is equivalent to the transportation lag, the DO probe should be located
in the leading section of the aeration basin; the changes in the oxygen demand are also
most severe in this section. However, the plant operators must still be able to move
the DO probe so as to find the best location. Some design engineers recommend
monitoring the middle and end sections DO, but rarely is this information used for
control purposes. The PI feedback control of uniform aeration systems reduces
power costs by approximately 20%, and 60 manhours of maintenance per system are
required.
Unlike completely mixed aeration basins which equalize the influent load, plug flow
basins keep the inflow segregated. Accordingly, the head section reacts rapidly to
changing influent oxygen demands and can be used to forewarn the later sections. In
other words, a feedforward control strategy in which the control information flow
follows the hydraulic flow pattern is well suited. Since plug flow reactors have spatial
as well as temporal variations in concentration and reaction rates, their behavior must
be described by partial differential equations. A material balance around the dif-
ferential section of a plug flow basin yields Equation 16, whose solution describes
the time and distance relationships to dissolved oxygen:
3C 9C
0 (16)
64
-------
where:
Q = space velocity (flow rate/normal area)
L = length or distance from inlet
t = time
r = respiration rate
At steady state:
8C
_±=0
3t
3C kl
= -iTif U
If U (L) = constant, then:
c ,TX kl
— - -LJ^+_£/c -CM!
L Q + Q (Co V U
It can be shown that 9C /9L > 0 under these circumstances and a DO gradient exists
o
in the aeration basin.
Under constant loading conditions and using second-order kinetics (e.g. , r = kSX),
then:
U(L) + akS. e~k/QLX-pkdX (17)
where:
X = average MLVSS
kd = endogenous respiration constant
k = second order rate constant
55
-------
Tapered aeration attempts to match the oxygen transfer to the oxygen demand accord-
ing to a relationship similar to Equation 17. It is informative to notice that increases
in influent concentration (S ) raise the entire aeration requirement, but the distribu-
in
tion profile remains virtually the same; on the other hand, flow rate changes alter the
shape of the air distribution network.
In summary, uniform aeration systems are difficult to control and the DO concentra-
tion must vary throughout the reactor. The head section always has a lower DO value
than the last sections. Lack of flexibility precludes the use of any elaborate feed-
forward control scheme. Flow ratio, feedback DO, or combinations of these two
should provide meaningful control of uniform aeration systems. The saving attain-
able with these control systems is somewhat less than in the completely mixed cases
since only single point control is possible.
For separately controllable aeration systems the following advanced but not yet demon-
strated control strategies will improve DO profile regulations. Flow and feedback DO
control will be discussed first. Equation 17 shows that the airflow (power consumption)
requirements at any given location change exponentially with respect to hydraulic flow
rate perturbations. Adding PI feedback control ensures elimination of measurement
and modeling errors:
U = kx exp (-C1/F) + k2e + kg f edt (18)
where:
C = kAXL , with A = cross-sectional area normal to flow direction.
If the aeration vessel has four separate control loops, tuning involves selecting 12
gains and four constants. Obviously, this controller is too complicated for other than
large plants.
56
-------
Engineering estimates indicate 20% power savings with this control strategy. Approxi-
mately 60 manhours per year of maintenance is anticipated for each control loop.
Since most plants that practice plug flow aeration use either two or four passes, the
total maintenance demand is the product of the number of loops (passes) multiplied by
the maintenance requirements per loop.
Another approach is to use feedforward substrate and flow with feedback DO trimming
for control. Increased influent substrate concentrations (S ) result in a proportional
in
increase in aeration intensity along the entire length of the aeration basin, as seen
from Equation 18. Flowrate increases, as discussed in the above control scheme,
necessitate an alteration of the aeration profile. Substrate measurements via TOC,
COD, TOD or respirometry and flowrate information are combined according to
Equation 17 into the following control, where PI feedback control is added to offset
measurement and model errors:
U = le S. exp (-C /F) + k e + k / edt
(19)
This control strategy adjusts the airflow (power consumption) for each pass or loop
in proportion to the influent substrate and exponentially with respect to the flowrate.
Rather than utilizing a substrate monitoring instrument, it may be possible to monitor
the power consumption or airflow rate for the first pass while under the control of PI
DO feedback regulation and feed it forward to the remaining aeration loops. If U.
denotes airflow rate or power consumption for the jth section, this control strategy
functions according to the following control law:
U] = kl] Uj - 1 ** (-C/F) + Yj
TT = 1r r> + lr
Ul Y C K2. /edt
57
-------
Engineering estimates show that this untried control strategy should reduce aeration
power by 25% for the substrate monitoring version, and by 20% for the feedforward
power consumption technique.
Even for completely mixed aeration systems proper DO probe location is essential
for all but the flow ratio control strategy since the control system acts on local DO
concentration sensed by the in situ probe. Accordingly the probe should be located
in a representative region of the aeration vessel. (The stagnant areas with low local
DO levels are unrepresentative regions and should be avoided.) Inasmuch as completely
mixed systems have large, well-mixed areas with uniform DO concentrations, finding
a representative region is usually easy.
Two methods are currently used for DO sensing: in situ and remote monitoring. In
situ methods are preferred because the time lags are essentially zero. All in situ
arrangements should have enough flexibility to allow operators to easily move the DO
probes to several locations in the aeration basin. If the DO probe is remotely located
.from the aeration basin, the elapsed time and subsequent oxygen depletion in trans-
porting the mixed liquor sample from the basin to the probe must be included in the
setpoint selection and control system tuning.
Although the DO control loops can interface directly with final control elements,
cascade control loops are highly recommended because they increase the system's
responsiveness and yet are stable. For some aeration equipment, such as multi-
speed blowers or multi-speed mechanical aerators, the inner cascade loop is nothing
more than a simple stepping relay. Other equipment, such as centrifugal blowers,
requires rather elaborate inner control loops to throttle airflow rates and prevent
surging.
All of the proposed control systems require varying amounts of maintenance. The
primary sensors, DO probes, organic analyzers, etc., must be kept calibrated to
58
-------
provide accurate data; moreover, controllers must be kept properly tuned. Table 6
summarizes the operational and maintenance characteristics of the proposed DO
control strategies for completely mixed aeration basins.
Since aeration basins exhibit imperfect mixing at the microscale level and available
DO probes exhibit some noise, instantaneous DO probe readings may not accurately
represent the true or average DO concentration in that region. Field experiences
support those comments. For example, in completely mixed aeration systems,
turbulent mixing conditions give rise to macro-fluid elements moving about the
vessel; accordingly, the DO fluctuates as these macro-elements pass across the DO
probe surface. The same phenomenon also occurs at a given cross-section of a plug
flow aeration basin (7). Since macro-mixing under turbulent conditions is a random-
type process and the noise associated with DO probes is also random, reasonably
the sporadic DO readings may be viewed as random events and the DO signal as the
sum of the true value and some random variable. Moreover, the available data indi-
cate that the randomness can be considered as white (random) noise with a zero mean.
The anticipated difficulty emerges when estimating the true DO value from the probe
signal. Estimators range from simple averaging delay and hold circuits to Kalman
filters and to nonlinear regression analyses. Choosing the best filter (estimation
procedure) is difficult. However, in line with the preferred philosophy of keeping a
control system as simple and easy to maintain as possible, the straightforward
estimators such as multipoint or time averaging, short-time rejection, delay and hold,
and low-pass filters are adequate for aeration basin DO control.
The quality of automatic regulation obtained from the proposed control systems de-
pends heavily on the adjustments made to their control modes. Without proper tuning,
all of the feedback controllers and some of the feedforward strategies will perform
poorly. It should be noted that the degree of difficulty increases with the number of
control modes. All the feedback controllers should be adjusted via the Ziegler-Nichols
method of ultimate sensitivity, with subsequent field experience providing added refine-
ments; the feedforward controllers should be adjusted by trial and error.
59
-------
Table 6. SUMMARY OF AUTOMATIC DO CONTROL FOR COMPLETELY MIXED AERATION BASINS
en
o
Description
Flow proportional
PI feedback
Flow proportional
with DO feedback
trimming
Organic loading via
TOC with feedback
trimming
Calculated respiration
rates with feedback
trimming
Aeration basin
respiration rates
with feedback
trimming
Aeration
Power
Savings
10%
25%
30%
35%
30%
35%
Confidence
Level
High
High
Medium
Low
Medium
Low
Advantages
Simple, straightforward
control strategy.
Control based on actual DO
values. Most commonly
practiced automatic DO
control. Easy-to-tune
controller.
Fast response to flow rate
changes. Readily adapt-
able to most plants.
Responds to influent load
changes (flow and organic)
before they upset the aera-
tion basin.
Uses aeration power levels
at constant DO values to
calculate respiration
rates. These measuring
devices are readily avail-
able and well behaved.
Since this control reacts
to respiration rate
change, it anticipates
DO variations and adjusts
aeration power accord-
ingly.
Disadvantages
Frequently yields poor DO
regulation, especially
with storm water or
ground water infiltration.
Slow response to upsets;
also may tend to oscil-
late. Many times, a
filter is required to
reject sporadic DO sig-
nals.
Moderately difficult to tune.
High infiltration or storm
events may lead to poor
control. Not any more
sensitive to oxygen
demand changes than PI
feedback controller.
This untried control strat-
egy requires an on-line
organic monitoring
instrument which may
prove difficult to main-
tain.
Requires a process com-
puter. Control may be
no better than simple
PID control.
High maintenance demands
and poor reliability of on-
line respirometers are
a serious drawback to
this strategy.
He co mm eudations
Useful only in small plants
(less than 1 mgd), or for
constant organic strength
wastewaters .
Suitable for most waste-
water-treatment facilities.
Very useful in moderately
large plants with dry-
weather flow and relatively
constant organic strength
influent.
Suitable only for large,
highly automated plant.
Should also be cascaded
with F/M control.
Useful in large plants with
real-time computer time
available.
-------
FOOD-TO-MICROORGANISM CONTROL
Introduction
With proper operation the activated sludge process can achieve 90% to 95% BOD reduc-
tions. For most municipal wastewater facilities, this means production of an effluent
BOD of about 20 mg/1 or less. To operate an activated sludge process efficiently, it
is necessary to control physical and biological parameters such as temperature, dis-
solved oxygen, hydraulic residence time, food-to-microorganism ratio (F/M), and
mixed liquor suspended solids. This section addresses control of one of the most
important biological parameters—the F/M.
In the activated sludge process, wastewater is contacted with microorganisms under
aerobic conditions whereby the microorganisms use the biodegradable pollutants as a
food source and purify the wastewater. It is essential for the microorganisms to form
a satisfactory floe so that they can be effectively separated; a fraction of the settled
biological solids is then recycled and the remaining mass is wasted.
It is generally accepted that BOD removal efficiency is dependent on the F/M. If this
ratio is too high, the bacteria undergo exponential growth and form dispersed floe that
does not settle and, as a consequence, the BOD removal is unsatisfactory. On the
other hand, for low F/M, unoxidized fragments of the floe remain in suspension and
cause a turbid effluent.
By means of the material balances on the substrate (BOD) and the biomass around an
activated sludge system (illustrated in Figure 15), it can be shown that the mean cell
residence time is related directly to the F/M:
61
-------
Q.F
AERATION TANK
* LA
COMPLETELY MIXED
VOLUME, V
Q + QD 1 Q + QQ - QUI ciu AI ** " **W
n | n ~ n ^^ n WHL •• ^^
CA -" SETTLING "
TANK
ns.cs
! !
SLUDGE
STORAGE
•CR ^ TANK ^nuu
(GRAVITY) ^
COMPLETELY MIXED
CA = MLVSS IN AERATION TANK F = BOD OF SEWAGE
CR = BIOMASS CONCENTRATION OF J} = FLOW RATE, SEWAGE
RETURN SOLIDS WITH SLUDGE QR = FLOW RATE, RETURN SLUDGE
STORAGE Qs = FLOW RATE. SLUDGE PUMPDOWN
Cs ' 3SSIS£S5SSSfm "• = ""^.WASTESLUOGE
Figure 15. Activated sludge system.
where:
Y = growth-yield coefficient mass of microorganisms/mass of substrate utilized
kd = microorganism decay coefficient, time"
U = F/M
6p, = mean cell residence time
Organic loading in most plants changes considerably during the day. For example,
the BOD loading for Baltimore (as reported by Keefer) (8) varies from 37 percent to
166 percent of the daily average in about an 8-hour period. (This can be seen in Fig-
ure 16.) Under such varying loads, it is not possible to maintain a constant F/M ratio
by adjusting the sludge wasting rate. This is because the growth rate of biomass is
much lower than the rate of increase in BOD loading. Under the loadings shown on
Figure 16, if the F/M ratio were at its desired value at 0800 it would be much greater
than the desired value at 1600, even with no sludge wasting during that period.
62
-------
180
ISO -
140 -
120
100
M 0200 04OO 0600 0800 1000 N 1400 I6OO 1800 2000 2200 M
TIME OF DAY
Figure 16. Variation in BOD load in Baltimore.
The diurnal spread in the F/M ratio can be greatly reduced by the provision of sludge
storage in the return sludge line, as shown by the flow diagram, Figure 15.
F/M Control Algorithm
A control procedure would involve continuous measurement of the flows and concentra-
tions. F (food) would be measured as TOC, COD or some other parameter which
could be correlated with BOD. The manipulated variables would be Q and Q . Q
R S W
could be automatically provided for as an overflow from the sludge storage tank. The
control algorithm, and its derivation, are given in the appendix.
63
-------
Example
Table 7 gives the results of computations made to simulate the control procedure
under the following plant conditions:
Average Q = 70 mgd
Average F = 144 mgA = 1200 Ib/MG
V = 30 million gallons
G = 3.5 days
Y=0.5
KD= 0.05
CS= CR assumed to be constant at 10,000 mg/L = 83,400 Ib/MG. (This would
correspond to an average sludge volume Index of 100.)
The diurnal variations in Q and F are in proportion to the values in Figure 16.
The results indicate that it would be necessary to stop the sludge return during the
period from midnight to noon. This is because the reduction in C^ brought about by
the natural washout of solids would be less than the reduction in CA necessary to main-
tain a constant F/M ratio. Thus it would be impossible to maintain a constant ratio
during this period. The sludge in the underflow from the final tank would be stored
up during this period and would be available to supply the needs for sludge return
during the period from noon to midnight.
It is noted that the quantity of sludge storage varied from zero to 4.52 million gallons,
assuming no sludge wasting. Thus, a storage tank capacity of about 5 million gallons
(one-sixth of the aeration tank volume) would have sufficed. If the settling quality of
the sludge were poor (sludge volume index greater than 100) Qg, QR and maximum
sludge storage volumes would all be greater than shown.
Analysis
The control strategy requires real-time BOD and mixed liquor volatile suspended
solids data, which are difficult parameters to measure accurately. Moreover, a
64
-------
Table 7. RESULTS OF COMPUTATIONS OF SIMULATED CONTROL PROCEDURE
UNDER SPECIFIED PLANT CONDITIONS
T
24
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Ave
Q
70
68
65
61
58
54
50
56
62
65
67
76
85
88
90
90
90
89
88
87
86
86
85
88
75
F
1626
1543
1460
1251
1001
1043
1084
792
500
417
334
584
792
1126
1460
1626
1751
1585
1418
1418
1418
1485
1547
1585
1200
CAO CR CAD
16,273 83,400 13,279
14, 548
13,279
12, 170
11,208
10,356
9621
8999
8337
7654
6997
6377
5755
7855
11,560
15,330
17,073
18,390
16,453
14,611
14, 393
14,227
14,895
15,341
12, 400
11,072
8803
6773
6759
6323
5174
3617
3162
2611
5178
7855
11,560
15,330
17,073
18,390
16,453
14,556
14, 393
14, 227
14, 895
15, 341
16,273
QR
0
0
0
0
0
0
0
0
0
0
0
0
27.0
47.4
55.6
39.7-
38.0
1.7
0.0
15.2
15.5
24.5
22.9
30.0
P
2.3833
2.31666
2.21666
2.08333
1.98333
1.8500
1.71666
1.91666
2.11666
2.21666
2.28333
2.58333
3.7833
4.5667
4.9033
4.3733
4.3166
3.0733
2.9833
3.4566
3.4333
3.7333
3. 6466
3.9833
e-P*
0.88865
0.90799
0.91178
0.91685
0.92069
0.92543
0.93098
0.92325
0.91558
0.91177
0.90924
0.89796
0.85416
0.82673
0.81533
0.83342
0.83539
0.87981
0.88311
0.86586
0.86671
0.85594
0.85904
0.84707
e
795.94
754.85
713.53
610.49
487.88
507.41
526.21
385. 66
244. 09
203.80
163.83
288. 21
20, 154
29,238
31,975
25,793
25,074
2274
697
12,987
13, 148
18,864
18,059
21,435
CAt
14, 548
13,279
12, 170
11,208
10,356
9621
8999
8337
7654
6997
6377
5755
7855
11,560
15, 330
17,073
18,390
16,453
14,611
14, 393
14, 227
14, 895
15,341
16,273
CA
15,410
13,914
12,725
11,689
10,782
9988
9310
8668
7996
7326
6687
6066
6805
9708
13,445
16,201
17,732
17,422
15,332
14,502
14,310
14,561
15,118
15,807
QS
12.9
11.3
9.9
8.6
7.5
6.5
5.6
5.8
5.9
5.7
5.4
5.5
9.1
15.8
23.5
25.2
27.2
18.9
16.2
12.5
17.4
19.3
19.6
11.0
vss
1.32
1.79
2.20
2.56
2.88
3.14
3.37
3.62
3.86
4.10
4.32
4.56
3.81
2.39
1.05
0.45
0.00
0.72
1.38
1.48
1.56
1.54
1.50
0.71
05
Ol
-------
mini- or micro-computer is necessary to implement on-line F/M control. Although
all of the above-mentioned equipment and analyzers are commercially available, no
wastewater plant currently uses on-line automatic F/M control, but several municipal
plants are planning to control the F/M ratio soon. At the present time, however, the
benefits of automatic F/M control remain unclear and controversial. Instantaneous
F/M control is not needed except where shock-unabsorbable loads occur. Instead,
most secondary treatment plants should maintain control at an average F/M with a
variation of ±25% allowed.
TRICKLING FILTERS
The trickling filter is the earliest and currently the most widely employed method of
secondary treatment. The wastewater stream is distributed over the top of the filter
packing by a spray arm that is either hydraulically or electrically driven. The dis-
tributed wastewater then flows down by gravity over the media, which has a fixed
biological film attached to its surface. Aerobic, facultative, and anaerobic bacteria
present in the biological film purify the wastewater by metabolizing the organic con-
taminants. As in the activated sludge process, removal efficiency depends upon the
DO content, nutrient level, pH, temperature, and organic loading/microorganism
ratio. BOD removal efficiencies range from 80% to 90% for well-designed and well-
operated systems. After passing through the filter, the effluent is collected and sent
to a secondary darifier. Trickling filters are usually classified as low rate, high
rate, or super rate, according to their loading rates.
Most of the design principles and operating practices of trickling filters were empir-
ically derived by a group of engineers and scientists acting under National Research
Council (NRC) sponsorship. The NRC found that trickling filter efficiency varied
inversely with load, and directly with filter surface area and the number of passes of
waste through the filter. Equation 20 summarizes the factors that affect operating
efficiency:
66
-------
100
E = - — HT (2°)
i + o.oo85/w7vF
where :
E = efficiency of removal of 5 -day BOD
W= weight, 5 -day BOD applied per 24 hours
V = filter volume acre per foot
F = number of effective passages through filter
For a single-stage filter with no circulation, F is equal to unity. However, when
recir culation is used, F changes as follows:
(21,
(1 + 0. 1R)
where :
R = recirculation ratio
Substituting this relationship into Equation 20 shows that BOD removal efficiency
increases with higher recirculation ratios.
As previously mentioned, trickling filter performance is a function of the DO, nutrient
levels, and recirculation rates. Recirculation keeps the media moist when flow is low,
improves distribution, retards the entry and egress of filter flies, and maintains a
sufficient hydraulic load to prevent clogging.
It is desirable to operate trickling filters at maximum BOD removal efficiency and
with a minimum amount of power consumption. Equation 20 shows that higher recir-
culation rates (up to flooding) will increase the BOD removal efficiency. Note, how-
ever, that the effectiveness decreases at high recirculation rates. Also, high recir-
culation rates consume large amounts of electrical energy in the recycle pumps.
67
-------
Recirculation rates may be held constant, kept above a certain minimum value, or
adjusted proportionately to influent flow rate. If constant recirculation is to be used,
then process control is unnecessary because process requirements will be on an
average basis. However, this mode of operation is unresponsive to changing flow rate.
When the total flow to the trickling filter is to be kept above some minimum value, the
total flow must be monitored so that, if the flow falls below a preset minimum, the
recirculation pump will automatically turn on; otherwise, the recirculation pump will
turn off. Minimum flow control systems prevent the trickling filter from starving
during low-flow periods such as nights, weekends, and holidays.
Flow proportional recirculation control, as illustrated in Figure 17, responds to the
flow rate variations, which are usually the most significant disturbances encountered
by trickling filters. A flow meter monitors the raw sewage flow and forwards this
SECONDARY
SETTLING
VARIABLE
SPEED DRIVE
Figure 17. Flow proportional recirculation control.
68
-------
information to a ratio controller that sets the desired flow rate for the inner recir-
culation flow control loop. A butterfly valve or variable-speed pump may be used to
vary the recirculated flow in proportion to the primary effluent. Flow-proportioned
control is relatively simple to implement and provides the following advantages:
• Maximum efficiency of maximum flows without hydraulic overloading
• Minimum necessary recirculated rates to prevent "starvation" at low flows.
The large capacity that is inherent in trickling filters with recirculation makes the
systems virtually immune to shock loads. In fact, trickling filters are frequently
used as "roughing" filters to protect downstream activated sludge processes from
shock loads.
The hardware necessary to accomplish the desired control strategies is readily avail-
able from commercial sources. All the proposed strategies are being practiced with
a high degree of success at many trickling filter plants. Table 8 highlights the bene-
fits of automating a trickling filter's recirculation rate, as well as some of its limita-
tions and recommended uses.
Table 8. TRICKLING FILTERS
Control Method
Benefits and
Potential Savings
Advantages
Disadvantages
Once through
Constant recirculation
Tliroughput kept above
minimum by con-
trolled recirculation
Flow proportional
recirculation above
minimum flow
Protects filter
Protects filter; saves
power
Provides maximum
efficiency
Simple
Simple; provides sub-
strate during low flow
Same as above
Permits high rate of
operation with pre-
dictable results
Limited to low-rate
filters
Higher equipment and
energy costs
Higher equipment costs
Requires variable-flow
recirculation system,
which in turn requires
somewhat more
maintenance
69
-------
SECONDARY SLUDGE PUMPING
Introduction
Solids separation by gravity is the usual technique for removing secondary sludge in
the effluent from the biological filters and the activated sludge aeration tanks. Gen-
erally, the comments made about primary clarifiers apply also to secondary clarifiers,
but, in addition, the successful operation of secondary clarifiers must take the follow-
ing into consideration:
• Sludge removal following the activated sludge process is limited by the return
sludge requirements.
• It is especially important that secondary sludge be removed promptly and either
returned to the process or processed for disposal. Secondary sludges are highly
unstable and will become septic quickly, which has a deleterious effect on the
activated sludge process.
• While sludges from biological filtration are small in volume (except during
periodic filter unloading), they must be removed regularly in order to minimize
potential septicity.
Control objectives, while similar to those for raw sludge, are dictated by the require-
ments of highly sensitive secondary treatment processes. Maintaining maximum
working sludge density must be tempered by the ever-present specter of septicity.
Secondary sludges associated with the activated sludge process are subject to a con-
dition known as "bulking," which will completely upset the entire secondary settling
process. Since the sludge will not settle, none of the usual subsequent processes will
function. This condition can be detected by a sludge level indicator, and appropriate
measures should be taken to correct it. It is evident that control is essential:
• Because of the close relationship between the secondary sludge removal process
itself and the activated sludge treatment processes
• In order to minimize the pumping of low solids secondary sludges
• Because the clarifier effluent must be void of settleable matter; otherwise, a
high chlorine demand and poor effluent quality will result.
70
-------
The final control element is a function of the secondary process. In the case of a
biological filter, the usual element is a sludge pump. However, when the secondary
process is activated sludge, the element can be a sludge pump, diverting valves, or a
combination of both.
Control for secondary sludge removal in the case of biological filters is essentially
the same as for raw sludge. The pump is started and stopped in response to the
selected control parameter, and is usually powered by a fixed-speed motor.
On the other hand, the activated sludge process requires a more sophisticated final
control element because the settled biomass must be returned in some regulated pro-
portion. If the return sludge is not sufficiently concentrated, the MLVSS and the
aeration tank residence time will decrease. Further attempts to build sludge age by
increasing the recycle ratio will overload the secondary clarifiers, which will lead to
additional process deterioration. Since the secondary process produces excess solids
above return requirements, an appropriate amount of sludge also must be wasted to
keep the process in equilibrium. The operating criteria for secondary clarifiers are
as follows:
• Secondary sludge must be returned to the aeration process as required.
• Excess secondary sludge must be removed for treatment and disposal.
• Because secondary sludges from the activated sludge process deteriorate
rapidly, secondary sludge must remain in the final clarifier as briefly as
possible.
Control Strategies
Control of secondary sludge withdrawal from a biological filter system is relatively
simple, since the process itself generates only small quantities of settleable solids
in the normal day-to-day operation. Low-rate filters typically slough their solids
seasonally, whereas high-rate filters continuously discharge small quantities of solids.
Inasmuch as the biological sludge produced is not returned to the process but, instead,
71
-------
is directed to treatment and disposal, a slight anaerobiosis of the sludge is no problem.
The usual control scheme consists of a timer to operate the sludge pump (similar to
the time cycle control strategy used for primary sludge pumping).
While there are many variations of the control scheme cited above, no installations
use the more sophisticated techniques, such as sludge waste interface detection or
control based on sludge density. Apparently, both experience and engineering judg-
ment indicate that such methods are unnecessary.
Successful operation of secondary clarifiers for the activated sludge process requires
producing an adequate supply of return sludge and a high quality effluent that is essen-
tially free from suspended solids. At times the dense sludge and high quality effluent
goals may conflict; when this happens, the operator must decide which goal should be
given the higher priority. Usually, producing an effluent free from settleable matter
is considered more important than withdrawing a dense sludge. The F/M control
section (described previously) details the specific designs and instrument loops
associated with returning the proper amount of sludge. Because of the similar physics,
equipment, and geometry used for primary and secondary clarifiers, the same control
strategies and instrument diagrams apply to both types. Different processing objec-
tives, however, change the desirability of specific control strategies.
Unlike primary clarifiers, most sludge withdrawal pumps for activated sludge pro-
cesses are variable-speed devices so that they can return sludge to the aeration basin
on a continuous basis. The most common technique consists of regulating sludge with-
drawal, wasting, and recycling on the basis of the influent flow (as illustrated in Fig-
ure 18). The influent flow is measured and transmitted to a ratio controller that sets
the index of the return sludge loop, consisting of a flow-measuring element, trans-
mitter, controller, and final control element. Waste sludge is also regulated by a
separate control loop that has the same equipment as the return sludge loop. Operating
experience, laboratory data, or an F/M controller will determine the fraction of
settled sludge returned.
72
-------
FROM INFLUENT
FLOW TRANSMITTER
TO MOTOR
CONTROL
TO SLUDGE
TREATMENT
Figure 18. Flow proportional sludge pumping.
In a well-operated activated sludge system, the settling properties and volume of
settled sludge change very slowly because the overflow rate of the clarifier is high
(that is, the capacitance of the clarifier is large). Accordingly, the flow proportional
system should provide satisfactory regulation. Flow proportion sludge recycle and
sludge wasting control saves 30% of the labor required by manual operation, and
effectively copes with flowrate perturbations.
One untried but potentially useful control strategy is to regulate the mass rate of
returned sludge. If a sludge density meter monitors the recycle stream, then the
product of sludge density and flow rate yields the mass rate at which microorganisms
are being returned to the aeration basin. This sludge return loop can be cascaded
with influent flow rates or incoming loads so as to maintain a proper F/M ratio.
73
-------
Another possibility involves using sludge level control to regulate the wasting loop and
either flow proportion or mass rate to modulate the return sludge. Although sludge
interface devices are unproven, once this sensor demonstrates enough reliability, the
strategy should be tested under field conditions.
Application Notes
Control system failures for secondary clarifiers result in the same types of problems
as experienced in primary clarifier control failures. In the case of activated sludge,
sludge pumping control failures strongly impact the treatment processes until ulti-
mately the process itself could also fail. Continuous sludge withdrawal and return
pump operation, which rarely occurs, will pull out the sludge blanket and eliminate
the benefits of sludge recycle. Downstream sludge thickeners and stabilizing devices
will soon become overloaded and fail.
Hardware availability and reliability are important factors, together with proper
installation and preventive maintenance, in selecting potential control strategies.
Suitable sludge flowmeters, sludge density monitors, and variable-speed sludge pumps
are commercially available. Accordingly, flow ratio and mass rate control strategies
can be readily implemented. Sludge interface detectors, although unproven in this
type of service, should be shortly available and sufficiently reliable to use in an auto-
matic control loop. Many municipal plants are successfully using sludge ratio, mass
rate, and time cycle control strategies for automating their sludge return and wasting
operations.
CHLORINE DISINFECTION
Introduction
Chlorine gas is widely used to disinfect industrial and municipal water and wastewater
because it is effective, easy to apply and inexpensive. It has a high toxicity for micro-
organisms responsible for waterborne diseases. In certain applications (such as
74
-------
storm-water treatment and remote package treatment systems), hypochlorite solutions
or bleaching power provides a safer, more flexible means of disinfection than gaseous
chlorine. Hypochlorite and bleaching powders, however, are more expensive and
deteriorate with time. Other disinfection agents such as ozone and ultraviolet irradia-
tion are used only occasionally because they are in general more expensive.
Chlorine gas, which is very soluble in water (0. 7% at 20° C and 1 atm), hydrolizes
rapidly to form hypochlorous acid when dissolved in water:
C12 + H2O ^: HOC1 + H+ + Cl"
At chlorine concentrations of less than 0.1% and pH values greater than three, the
hydrolysis goes virtually to completion. The rate of kill has been empirically corre-
lated to the 1.3 power of the residual chlorine concentration.
Although a rigorous accounting of all the forms and reactions of chlorine and compounds
present in wastewaters is beyond the intended scope of this report, it is instructive to
examine some of the principal reactions among chlorine, reducing agents, and ammonia.
When chlorine is added to water-containing reducing agents such as hydrogen sulfide
and nitrites, the hypochlorous acid reacts with the reducing agents to form chlorides,
and no useful disinfection results. After satisfying this demand, further chlorine
additions will result in the formation of chloramines. Chlorine stored in the form of
chloramines is available for disinfecting purposes (this is usually referred to as com-
bined available chlorine). Continued chlorination initiates the complete oxidation of
the chloramines to nitrogen, nitrate and nitrogen trichloride (breakpoint chlorination).
As a practical guide, because of the high ammonia concentrations present in waste-
water, only combined residual chlorine is used as a disinfectant.
Chlorine residual of 0.5 mgA after a 15-minute contact period is generally necessary
to ensure the adequate disinfection of sewage treatment effluent. The wastewater must
be thoroughly mixed with the chlorine and then processed in a properly designed
75
-------
contact chamber. Because of the variable sewage flowrates and reducing agent con-
centration, the chlorine feed rate must be adjusted continually in order to maintain
a constant residual chlorine level.
A good automatic chlorine control system should supply sufficient chlorine to ensure
adequate disinfection and should also minimize chlorine consumption. Furthermore,
the control system should be safe, easy to maintain, and compatible with the environ-
ment of a wastewater-treatment plant. Fortunately, these broad control objectives
can be accomplished with several widely used and well-proven chlorination control
strategies, which range from simple flow proportional to the complex compound plus
postcontact residual chlorine control. Chlorine feed rate is always the manipulated
variable, and the final control element may be a valve, a metering pump, or a chlorine
feed loop. The selected control strategy then sends a chlorine feed rate signal to the
final control element, which implements the final action.
Flow Proportional Control
The most widely used control strategy regulates chlorine feed in proportion to plant
flow, as depicted in Figure 19. To make this control strategy effective, the flowrate
should be measured directly at the head of the mixing chamber or as closely as
practical to the point of chlorine application. A common error has been to use the
flow measured at the plant's headworks rather than the flow at the chlorination site.
Although the long-term average flow will be the same, the immediate difference in
flowrate can be large. For example, returning back flush water or digester super-
natant will momentarily increase the down-stream flowrates. When properly designed,
the flow-proportioned control system will automatically adjust the chlorine feed rate
in accordance with the flow variation so as to maintain a constant ratio between the
chlorine added and the wastewater flow. The plant operator should take samples
periodically of the chlorinated effluent for bacterial and residual chlorine analysis,
and, on the basis of this information, he should manually adjust the ratio of chlorine
added to wastewater flow.
76
-------
H
SOLUTION-TYPE
CHLORINATOR
I
0
CONTACT CHAMBER
Figure 19. Simple flow-pacing chlorination control loop.
Flow-proportioned chlorination control is reliable and simple to implement and main-
tain; it also responds rapidly to flowrate perturbations. Since this strategy neglects
changing chlorine demands on an automatic basis, flow-proportioned chlorination is
not well suited to applications where the chlorine demand changes frequently or where
subsequent dechlorination processes are employed. Nevertheless, flow proportion
control is adequate for the many small plants that use small-to-moderate amounts
of chlorine.
Compound Chlorine Control
With the availability of reliable residual chlorine analyzers, the ratio between the
chlorine feed rate and wastewater flowrate can be adjusted automatically on the basis
of on-line residual chlorine measurements (Figure 20). Since the chlorine dose
depends upon the quantity of reducing agents present, the ammonia concentration, the
organic load, and the amount of suspended solids, the chlorine demand changes sig-
nificantly in the course of a day in many wastewater-treatment plants. Accordingly,
the amount of chlorine per volume of wastewater necessary to achieve a given residual
77
-------
RESCI2
SOLUTION-TYPE
CHLORINATOR
COMPOUND LOOP CONTROL
Figure 20. Typical compound chlorination control loop.
chlorine level also changes significantly in the course of a day. With the onset of a
high chlorine demand period, the residual chlorine drops below the desired level.
The residual chlorine analyzer then sends the concentration data to the controller,
where an increased chlorine flowrate command is generated and acted on by the final
control element. This feedback system automatically increases or decreases the flow
ratio until the desired residual chlorine level is obtained.
Compound chlorination control accurately modulates the chlorine application and so
maintains the desired residual chlorine level. Many wasterwater-treatment facilities
are successfully using this control strategy today. Although the automatic control
equipment is more expensive than simple flow-proportioned control equipment and
requires more maintenance, reliable equipment is commercially available, with its
higher costs offset by the chlorine gas savings obtainable by tighter control. Con-
sequently, compound chlorination control is considered suitable for medium-to-large
plants that use significant amounts of chlorine and/or experience time-varying chlorine
demands, and for plants that use dechlorination processes.
78
-------
Double Compound Control
Compound control, however, does pose some difficult problems because most stan-
dards or codes require that a prescribed residual chlorine be maintained after at
least a 15-minute contact time. Accordingly, feedback residual chlorine control
systems that have potential 15-minute lags are prone to instabilities. If a feedback
control system is to perform adequately, the loop time must be within a 3-to-5 minute
range; this means that the residual chlorine must be determined shortly after mixing.
The difficulty of relating the control residual to the residual at the end of the proper
contact period is best handled by a second residual chlorine analyzer that records the
residual after contact. To ensure an adequate residual, it may be necessary to use a
postcontact residual chlorine analyzer that readjusts the chlorine application rate at
the head of the contact chamber as shown in Figure 21.
Both the compound and double compound control strategies use pacing or elementary
feedforward as the primary action, and residual chlorine as a secondary trimmer.
Since pacing is a highly stable control mode and the chlorine demand normally changes
slowly, compound control loops are very stable.
SOLUTION-TYPE
CHLORINATOR
RES CI2
s~>. RES CI2
ARC
RES CI2
RES CI2
FE
F
rnniTACT CHAMBER
Figure 21. Double compound loop C12 control.
79
-------
Application Notes
The control dynanics of the chlorination process are dominated by both reaction time
and large delays. Control will be optimum with proper design to ensure that: 1) the
chlorine water stream is well mixed into the main flow, 2) the flow signal is properly
represented at the point of application, and 3) when an analyzer is used, the measure-
ment lag does not vary appreciably with the plant's throughput rate.
The success of the compound controllers depends on accurate residual chlorine mea-
surements; these analyzers require periodic maintenance and calibration. Accumulated
user experiences indicate that about 140 manhours a year must be allocated for servic-
ing each residual chlorine analyzer.
Chlorination control system failures can have a nearly catastrophic effect on the opera-
tion and safety of a wastewater plant. A failure that results in the cessation of chlorine
feed permits unchlorinated effluent to be discharged into the receiving water. Other
failures can result in overchlorination (residual chlorine analyzer failures are princi-
pally responsible for this type of failure). Chlorine will be fed into the wastewater up
to the capacity of the chlorination equipment, regardless of demand or residual. The
impact of gross overchlorination depends upon the nature and size of the receiving
water.
There are hundreds of flow proportional chlorination systems in service today. A
recent survey found that 94% of the users were satisfied with the system's performance.
Flow-proportioned control is particularly well suited to small plants, where it is
usually difficult to obtain skilled maintenance service.
About one-third of the existing wastewater-treatment plants use compound control to
regulate the gaseous chlorine feed rate, whereas only a handful of plants use double
compound control. Typically, the newer and larger plants are more apt to use com-
pound control because of their need for tighter residual chlorine control. A survey of
some 13 plants that practice compound control disclosed that 77% of them were sat-
isfied with the compound control loop's performance.
80
-------
The automatic residual chlorine control devices that are presently available are field
proven and ensure the proper chlorination of wastewaters, especially after secondary
treatment. Occasionally, chlorination control of raw sewage, storm water, or com-
bined sewage may fail because the residual analyzer becomes plugged with debris.
Compound control systems can pay for themselves in chlorine saving, while simulta-
neously ensuring a facility's compliance with discharge standards.
The data presented in Table 9 was obtained from the survey as reported by Molvar et
al.(l) and highlights the benefits and limitations of the chlorination control strategies.
Table 9. DISINFECTION VIA CHLORINE ADDITION
Control Method
Benefits and
Potential Savings (1 )
Advantages
Disadvantages
Fixed rate
Flow proportion
Residual chlorine feed-
back
Compound control
Double compound con-
trol
15% labor saving and
25% chlorine saving
over basic method
15% labor saving and
50% chlorine saving
over basic method
15% labor saving and
50% chlorine saving
over basic method
15% labor saving and
55% chlorine saving
over basic method
Simple and reliable
Reliable and well-
established
Well-established and
produces good
effluent residual
control
Well-established and
produces excellent
residual control
Best available control
of final residual;
especially useful
whenever nature and
strength of influent
vary widely
Can overdose or under-
dose as flow changes
Requires some mainte-
nance
Requires some mainte-
nance but more instru-
mentation; poor per-
formance when flow
changes rapidly
Requires some mainte-
nance and instrumenta-
tion
Requires considerable
maintenance and
instrumentation
81
-------
CONTROL OF ANAEROBIC DIGESTERS
Introduction
Anaerobic digestion is a complex biological process which converts organic matter to
methane and carbon dioxide in the absence of molecular oxygen. This process is
widely used in the stabilization of domestic and industrial wastewater sludges. The
goal of anaerobic digestion, as well as any other sludge stabilization process, is to
produce an easily dewaterable sludge which can be safely disposed of without environ-
mental nuisances or hazards. Anaerobic digesters, however, have a reputation of
being unstable, unreliable, and troublesome. This belief is primarily due to improper
operation and control, rather than any inherent instabilities of the anaerobic digestion
process.
Although relatively little is known about the metabolic processes that occur during
anaerobic digestion, certain operational concepts and practices have been established
and accepted. For example, such factors as volatile acids concentration, alkalinity,
pH, retention time, biomass concentration, loading rates, and temperature strongly
influence the stability and operational efficiency of anaerobic digesters. Process
control technology offers a method of improving the reliability of anaerobic digesters
by automatically regulating some of the above-mentioned factors. The automatic
monitoring associated with process control also would measure process conditions
and tend to eliminate human errors. The following sections develop control strategies
for the three most important process control parameters: temperature, pH, and
methane gas production.
Temperature
Since the reactions taking place in anaerobic digesters result from the metabolic
activity of heterogenous bacterial populations, the temperature effect on process
efficiency is determined by the response of the bacterial species present. Tempera-
ture ranges for the optimal growth of microorganisms can be divided into three regions:
82
-------
psychrophilic (<20°C), mesophilic (20 to45°C), and thermophilic (>45°C) (9).
Although microorganisms are active from -5°C to about 80 °C, specific microorgan-
isms often have narrow temperature ranges in which they will reproduce. In fact,
bacteria are classified as psychrophilic, mesophilic, or thermophilic bacterium
according to the optimal growth temperature. A common taxonomic criteria also
uses growth temperature ranges for a specific bacterium. Thus, temperature strongly
influences the ultimate population that will actively grow in the digesters.
Most authorities agree that economical operation of anaerobic digesters occurs in the
mesophilic and thermophilic zones, although much controversy exists on the pros and
cons of mesophilic vs thermophilic digestion (10,11).
The rate of a chemical reaction doubles for every 10 °C increase in temperature.
Over a narrow temperature range this is approximately true for biological reactions.
Also, it has been demonstrated that sudden temperature changes are detrimental to
anaerobic digesters (12). Since it is important to maintain the digester at a tempera-
ture which allows stabilization to proceed at the highest possible rate, good tempera-
ture control is essential.
Older digesters have heating coils embedded in the digestion tank through which hot
water is circulated to transfer energy to the digester liquor. Present designs pump
the sludge to an external heat exchanger where the digester liquor is recirculated,
resulting in higher thermal transfer efficiencies and in better mixing.
Since digesters are designed and operated as large-capacity processes with residence
times from 10 to 30 days, simple two-mode feedback temperature control, as illus-
trated in Figure 22, keeps the digester's temperature within acceptable limits. The
resistance or gas-filled temperature element, which must be located in a well-mixed
representative region, measures the digester's temperature. When the temperature
decreases below the lower setpoint (differential gap), the controller turns on the sludge
recirculating pump.' Accordingly, the digester's contents are heated and the tempera-
ture rises. When the temperature increases to the upper limit, the controller shuts
83
-------
SP
SLUDGE
FEED
GAS
RECIRCULATION
PUMP
SLUDGE
RECIRCULATION
PUMP
HEAT
EXCHANGER
Figure 22. Digester temperature control.
off the recirculation pump. A second on/off thermostat controls the boiler or heater
so as to maintain the proper hot-water temperature. Most commercially available
temperature control systems can satisfy the anaerobic digester's requirement of main-
taining a uniform temperature.
PH
pH is an important indicator of the condition of anaerobic digesters. Methane produc-
tion results from two major groups of microorganisms: the acid-forming group, which
is responsible for hydrolyzing the complex organics to simpler compounds (typically
volatile acids), and a second group, the methane formers, which are sensitive to pH.
Most reports indicate that the optimum pH is in the range of 6. 8 to 7.2. If the pH is
out of the 6. 8 to 7.6 range, the methane bacteria become inhibited and sludge stabil-
ization ceases.
84
-------
The three common causes of digester failures (hydraulic, organic, and toxic over-
loadings) result in pH changes. The magnitude of the pH change depends upon the size
of the overload and the alkalinity of the digester liquor. Hydraulic and organic over-
loadings produce a rapid decrease in pH due to the rapid buildup of volatile acids.
Toxic materials preferentially kill the sensitive methane-forming bacteria which
results in a gradual reduction in pH.
Since the methane bacteria become inhibited outside the 6. 8 to 7.2 range, automatic
pH control can eliminate the deleterious effects of pH upsets caused by hydraulic and
organic overbadings. In a healthy digester the formation of bicarbonate alkalinity
counterbalances the formation of volatile acids, and the pH remains constant. Because
it is difficult to assess the cause of pH disturbances, a feedback pH control system
(as shown in Figure 23) should hold the pH in an acceptable range. Usually lime is
used for pH control of digesters. However, unless high intensity mixers are used
proper dissolution and distribution of lime in digesters does not take place. It is not
ALKALINE
MATERIAL
GAS
SLUDGE
FEED
DIGESTER
GAS
RECIRCULATION
PUMP
DIGESTED -
SLUDGE
Figure 23. Digester pH control.
85
-------
good practice to use NaOH because it rapidly reacts with CO and could produce a
Lt
partial vacuum. Sodium bicarbonate or sodium carbonate are the chemicals of choice.
A recent stimulation study (13) shows that simple proportional control can keep the pH
within acceptable limits.
The control strategy depicted in Figure 23 shows a pH feedback control loop for a
well-mixed anaerobic digester. Since low pH's are responsible for most digester
problems, provisions have been made only for the addition of alkaline reagents. When
the pH drops below the set-point, the alkaline material is added at a rate proportional
to the pH error. The addition of sodium hydroxide raises the pH to the acceptable
range and thus provides a more favorable environment for stable digester operation.
The fouling nature of sludge liquor means the pH probes will become coated in a short
time. Consequently, it will be difficult to keep an automatic pH control system oper-
ating correctly. The addition of in situ ultrasonic probe cleaning and installing the
probe in a well-mixed region of the digester may eliminate or at least minimize the
pH probe fouling problem. Each proportional pH control system should require about
60 man hours of maintenance per year for cleaning and calibrating the pH probes and
tuning the proportional controller. Based on engineering estimates, automatic pH
control of anaerobic digesters may allow a 10% increase in sludge processing.
Methane Gas Production
The goal of stabilizing sludge by anaerobic digestion can be achieved in only one way—
the production of methane gas. The degree of organic removal is in direct proportion
to the amount of methane produced. When the gas production trend is downward, the
digestion process is failing. However, several events which are not related to pro-
cess failures may cause variations in gas production. For example, decreasing the
amount of sludge fed to the digester will clearly lower gas production. Also, tempera-
ture variations of only 2°F or 3°F will decrease gas production. In the forthcoming
86
-------
control development, it will be assumed that adequate temperature control exists so
that decreasing methane gas production indicates instability and impending digester
failure.
Since anaerobic digestion is a two-step process, a decrease in methane gas production
implies that the number or reaction rate of the methane-forming bacteria is inadequate
to convert the organic intermediates to methane and carbon dioxide. More methane
formers must be added to the anaerobic digesters or the reaction rate must increase.
In single-stage digesters, the methane population will increase by auto-catalytic
growth as a result of excess available food. Because methane formers are slow
growing, the digester cannot be fed; otherwise the process becomes inhibited by pH
depressions. Increasing the digester's temperature also increases the growth rate
and the rate of volatile acid destruction.
For single-stage digesters, accordingly, the rate of methane gas production can be
used to automatically modulate feeding rates, temperature, and pH, as shown in
Figure 24. If methane gas production falls and the rate of loading is unchanged and no
toxic materials are present, then it would be desirable to increase the temperature.
When the methane gas production drops, the measuring elements transmit this infor-
mation to the control system, increasing the digester's temperature inversely pro-
portional to the gas production rate. If this strategy does not correct the situation,
digester feeding is terminated when the gas rate decreases below a minimum value.
If the temperature control works, the methane gas production rates rise to their usual
values, the temperature is reduced gradually to its normal levels and the digesters
are once again fed at a uniform rate. Unfortunately, the methane formers grow slowly
and several days are necessary to restore a single-stage digester.
With two-stage digesters, the most effective control strategy involves recycling
settled sludge (methane bacteria) from the second to the first digester. This strategy
leads to a buildup of methane-forming bacteria when gas production drops. The
methane production data by means of a feedback controller regulates the sludge
87
-------
DIGESTER
GAS
(SEE FIGURE 25 FOR
TWO-STAGE UNIT CONTROLS)
FEED}
HOT WATER
SUPPLY
HOT WATER
RETURN
Q*--
Figure 24. Combined pH, temperature, and digester gas control for
single-stage unit.
88
-------
recycling, as illustrated in Figure 25. Since the methane production is a linear func-
tion of the methane bacteria concentration and the control objective is to keep the
methane production rate above some minimum value, a simple on/off controller should
provide adequate corrective control action. When the methane gas production drops,
the sludge recycle pump is turned on until the methane rate rises above the prescribed
value. The recycle sludge pump is then turned off. To be successful, the second-
stage digester must contain a sufficient supply of methane bacteria; otherwise, increas-
ing the sludge recycle rate will simply decrease the hydraulic retention time and per-
formance will deteriorate. In most cases it is desirable to cascade this control
scheme with temperature and pH controllers, as was done in the single-stage digester.
f (SEE FIGURE 24 FOR ORIGIN OF THIS SIGNAL AND FOR
REST OF CONTROLS THAT ARE SAME FOR BOTH FIGURES)
TO SOLIDS
DEWATERING
AND
DISPOSAL
ORGANISM RECYCLE
FEED-
DIGESTER
INTERFACE
SEPARATOR
Figure 25. Combined pH, temperature, and digester gas control for
two-stage unit.
89
-------
Both these methane gas production control strategies have never been tried in a waste-
water-treatment plant. Reliable methane or carbon dioxide analyzers and gas flow-
meters are readily available and have successfully demonstrated their suitability for
wastewater-treatment projects. Undoubtedly many plants are using methane gas pro-
duction rates in formulating their operating practices, but these control strategies are
being practiced on a manual rather than automatic basis. Simulation studies have
shown that methane bacteria recycling based on methane gas production rates will
prevent most digester failures automatically. In the absence of specific performance
data, the proposed control strategy for single-stage digesters should permit process-
ing about 15% more waste sludge than a manually controlled digester. For two-stage
digesters, the proposed control action should increase sludge processing by about 25%.
Since sludge stabilization accounts for approximately 40% of a treatment plant's capital
cost, anaerobic digesters, the most commonly used sludge stabilization process,
should be carefully considered during plant automation and process control. The pro-
posed control strategies summarized in Table 10 have the potential for saving capital
and operating expenses, and for increasing the digester's reliability.
SLUDGE CONDITIONING FOR VACUUM FILTRATION
Introduction
Chemical conditioners are required to effectively dewater activated sludge and pri-
mary and digested sludges by vacuum filtration. Polyvalent metal ions such as Al
(HI) and Fe (HI) or synthetic organic poly electrolytes (cationic, nonionic, or anionic)
are added to the sludges in order to structure them properly for dewatering. These
chemicals attach themselves to the discrete sludge particles and form a bridge to
other individual particles. The conditioned sludge now has sufficient structural
strength and porosity to allow the rapid escape of water under a vacuum-driving force.
A recent laboratory study (14) showed that the supernatant pH and the sludge solids
concentration significantly affect the conditioner dosage required to enhance sludge
90
-------
Table 10. ANAEROBIC DIGESTER CONTROL STRATEGIES
Description
Temperature
pH
Methane gas
production
(single stage)
Methane gas
production
(two stage)
Potential
Savings
10% more
throughput
15% more
throughput
15% more
throughput
25% more
throughput
Confidence
High
Low
Moderate
Moderate
Advantages
Maintains temperature
within 1°C of desired
value. Presently avail-
able equipment is satis-
factory.
Keeps pH within the opti-
mum range of 6. 8 to
7.6. Compensates for
organic overloading.
Acts as an early warning
system with automatic
corrective action for
toxic overloading.
Automatically compen-
sates for toxic over-
loading by introducing
new methane -forming
cells .
Disadvantages
pH probe fouls easily due
to nature of digester
sludge. In fact, this
strategy has never been
tried in a municipal
plant.
Slow recovery time
(digester is out of ser-
vice for this time).
Severe corrosion prob-
lems with gas metering.
Must ensure an adequate
supply of concentrated
sludge for return to first
digester. Severe corro-
sion problems with gas
metering.
Remarks
Temperature control should
be used on all digesters.
pH probes must be located
in a well-mixed region.
Ultrasonic cleaning may
be helpful. Untried.
This control strategy is
useful when toxic over-
loading occurs frequently.
Should be cascaded with
temperature control.
This control strategy is
useful when toxic over-
loading occurs frequently.
CO:
-------
dewatering by vacuum filtration. Since process disturbances can change the pH and
the mass loading of settled sludge, automatic control systems that adjust chemical
feed rates so as to keep the sludge cake production at a maximum level and to produce
a cake with a minimum moisture content (60% to 70%) are examined.
pH Control Systems
The hydrogen ion concentration will affect the surface charge of the sludge particle as
well as the properties of the conditioners. The optimum pH for effective conditioning
of sludge is 6.0 to 7.0 for Fe (III) and 4. 5 to 5. 5 for Al (III). Although a feedback con-
trol system that acts on the conditioned sludge's pH would provide ideal control (refer
to previous discussion), pH is very difficult to measure continuously in thickened waste
sludge. Accordingly several alternative pH control systems have been proposed.
Rather than measure the pH of the conditioning sludge, the automatic pH control sys-
tem responds to changes in the filtrate's pH, as depicted in Figure 26. Since the
solids do not affect the pH, measurements of the filtrate's pH are indeed a measure
CONDITIONER
SOLUTION
pH ADJUSTING REAGENTS (ACID)
SP
WASTE
SLUDGE
Figure 26. Feedback pH control based on filtrate pH.
92
FILTRATE
-------
of the slurry's pH. However, the system has a deadtime of several minutes, which
may cause some control difficulties. Fortunately, the alkalinity of the sludge slurry
and chemical demand changes so slowly that most of these systems work reasonably
well.
A second method of pH control regulates the pH of the conditioning chemical solutions,
as shown in Figure 27. The success of this method pivots on the constant alkalinity of
the sludge slurry and the ability of the flow proportional dosing system to perform
accurately.
Although none of the plants visited in a recent nationwide survey practiced any form
of automatic sludge-conditioning pH control, it is estimated that automatic pH control
would increase a vacuum filter's productivity by about 20% over a well-run manual
operation.
Dosage Control
The data reported by Tenney (14) clearly shows that a stoichiometric relationship
exists between the sludge mass and the dosage of chemical conditioners. Moreover,
DRY
CHEMICAL
FEEDER
»•!•<
-3 pH ADJUSTING
1 REAGENTS
pH
TO SLUDGE CONDITIONING VESSEL
Figure 27. pH control of conditioning chemical solutions.
93
-------
this relationship appears to be linear, as shown in Figure 28. If the sludge concen-
tration remains essentially constant, then the conditioning chemicals can be added on
a flow proportional basis, as illustrated in Figure 29. However, where the solids
concentration also changes, the conditioners should be added on a mass proportional
basis. This strategy would require a density-measuring device such as a nuclear or
ultrasonic sludge meter, as shown in Figure 30, which also depicts the other elements
of this control system. Both the flow proportional and mass proportional control sys-
tems require accurate flow rate-monitoring devices. Unfortunately it is difficult to
measure flow rates reliably at the low velocities usually used in transporting thickened
sludge. Specially designed flowmeters, which became available only recently (such as
BIF Solid Bearing Fluids Meters), should be used.
Automatic dosage control is rarely used in wastewater-treatment plants today; instead,
constant rate chemical feeding is used most frequently. Constant dosage control will
probably reduce chemical consumption by about 15% and increase throughout by 5 to
10%. Mass proportional dosage control should reduce chemical consumption by 20 to
25% and increase the filtering capacity by about 10%.
Before extensive efforts are undertaken to demonstrate these potentially useful but
untried control strategies (which also use new and relatively untried flow-measuring
elements), the extent of need must be carefully appraised, since today vacuum filtra-
tion is only one of many acceptable techniques for disposing of waste sludge.
SLUDGE DEWATERING
Because vacuum filtration reduces the water content of sludge from 95% to 20%, it is
widely used to reduce the water content of sludge prior to its final disposal by means
of landfilling or incineration. In today's wastewater-treatment plants, vacuum filters
are operated intermittently for several days/per cycle to dewater a backlog of waste
sludge. The fundamental filtration equation is:
94
-------
2000
Figure 28. Effect of increasing sludge solids concentration (dry basis) on requisite
conditioner dose. Increasing solids concentration (or surface area)
increases requisite conditioner dose proportionately. A definite stoich-
iometry exists between sludge solids concentration and requisite con-
ditioner dose, as shown by insert.
CONDITIONER
CHEMICAL
SOLUTION
SP
FE
\
\
1
r
a
=>
MAG
0
OR
PUMP SPEED
Figure 29. Flow proportional chemical feed system,
95
-------
CONDITIONER
CHEMICAL
SOLUTION
\
MAG
D
1
0
3
^/
H
Figure 30. Mass loading chemical feed system.
dV
dt
PA
|j(rcV + RmA)
(22)
where:
V = volume of filtrate
t = time
P = pressure
A = area
H = viscosity of filtrate
r = specific resistance
c = weight of solids/unit volume of filtrate
Rm = resistance of filter media
This equation shows that the performance of the vacuum filter as measured by the
filtration flow rate \—rr} depends on the stream variables of specific resistance (r)
96
-------
and solids concentration. Proper chemical conditions by means of several automatic
control systems should keep the resistance at a stable low value. Filter performance
also depends on the machine variables of pressure drop, renewed surface area rate
(rotational speed), and sludge feed rate.
From a control standpoint, the objectives of sludge dewatering are to maximize the
production of dewatered sludge and to minimize the blinding of filter media in spite of
any process disturbances. The latter objective is largely a function of the filter media
and is not a major problem in view of contemporary designs and materials. Process
control is necessary as a means of establishing sludge flow and the filter's rotational
speed. Usually the supply trough is maintained at a constant level by pumping the
sludge into the supply trough and subsequently returning the excess through a gravity
overflow.
In small and medium-size plants, the vacuum filter may operate only 2 or 3 days/week.
Sludge feed rate and rotational speed control eliminates the need for constant operator
attendance, and also results in a more uniform sludge cake. Unfortunately the limita-
tions involved in measuring low-velocity sludge flows entail a serious impediment to
success. Because the present control technique (of maintaining the sludge trough level
by pumping an excess and then returning it through an overflow) is both simple and
effective, possible improvements of this technique are not likely in the near future.
There is an alternate technique: starting and stopping the sludge feed pump in order
to maintain an acceptable range of levels in the supply trough, but this method triggers
a potential problem, the possibility that the feed line will become plugged during the
feed pump's off-time.
Hardware for the reliable level measurement, flow measurement, and final control
elements is only partially available. Unfortunately any limitation of the sludge flow
by either variable-speed pumps or valves promotes plugging of the feed line. A com-
puter offers no advantages to this process, and neither are complex formulations or
elaborate equipment sequencing required.
97
-------
The failure of vacuum filter control systems in present-day service results in flooding
the supply trough and/or starving the filter, thus taking the vacuum filter out of service
until the situation can be corrected. Normally no permanent damage results, however.
The current technique of pumping excess sludge is essentially self-regulating, and the
advisability of future improvements remains questionable because the added equipment
would require extensive maintenance and introduces operating problems of its own.
Table 11 highlights the advantages, savings, and limitations of vacuum filtration con-
trol schemes.
Table 11. SLUDGE DEWATERING-VACUUM FILTER
Control Method
Supply trough level con-
trol (overflow)
On-of f operation of
feed pump
Benefits and
Potential Savings
Uniform sludge cake
production
Uniform sludge cake
production
Advantages
Simple and effective
Simple and effective
Disadvantages
Excessive pumping;
complicates any chem-
ical addition
Increases probability
of plugged feed lines
Other control strategies that may be suitable and are currently being studied include
control of the rotation speed and the vacuum of the filter, controlling the feed rate and
addition of the coagulant based on the moisture content of the filtered sludge, and con-
trol of the feed of the iron coagulant based on the pH of the sludge conditioning tank.
INCINERATION
Sludge incineration is a combustion process that reduces the sludge volume and pro-
duces a sterile and easily disposed-of ash. Numerous proprietary incineration systems
and furnaces, each with its own control scheme, are commercially available today.
98
-------
This section touches on the important unit operations and processes that are automat-
ically controlled in most incineration systems. For.a more detailed examination, the
reader should consult the literature recommended by relevant manufacturers.
Control of an incinerator and associated air pollution control devices can be classified
by the following broad areas:
• Fuel feed
• Auxiliary fuel feed
• Combustion air
• Air pollution control.
In this instance the fuel would be raw and secondary waste sludge but, depending on its
treatment, drying (or other processing) may be necessary prior to incineration. Con-
trol of the incinerator proper, however, entails control of the following parameters:
• Fuel feed rate
• Auxiliary fuel feed rate
• Grate airflow rate
• Top airflow rate
• Ash chamber.
The instrumentation associated with air pollution control devices is unique to each
proprietary design. In general, though, it monitors and controls the following ele-
ments:
• Smoke density
• Water flow to scrubbers, etc.
• Stack gas composition (e.g., O^).
Starting and shutdown sequences are complex and usually involve an analog programmer.
Because of the potential hazards, elaborate safety shutdown systems are also included;
these relate particularly to the auxiliary fuel burner equipment. Similar systems relate
to specific types of air pollution control equipment in order to prevent equipment damage.
99
-------
FiTni'e 31 shows a typical set of control loops associated with commercially available
incineration systems. The only control involving the operation of the incinerator is
temperature. All other control loops relate to safety shutoffs, alarm systems and
air pollution control. These control systems are supplied as part of the incinerator
par-kage.
AIR
NOTES:
1) Not shown: alarm system.
2) Starting circuit also requires full purge.
3} Safety circuit shuts down incinerator on:
• Loss of main air
• Loss of power drive
• Loss of cooling water
• Loss of fuel pressure
• Loss of flame
• Overtemperature
• Overpressure.
Figure 31. Control system for a sludge incinerator.
100
-------
NEUTRALIZATION OF ACIDS AND BASES
Control of the neutralization of acids and bases is a very comprehensive subject.
Satisfactory control requires not only attention to the instruments and valves, but also
to process chemistry, piping, mixing, and vessel design. As a result, this treatise
can summarize only the recommended practices in all these areas.
General Chemistry and Reagents
The relationship between the amount of reagent added to a sample (or the flow of
reagent added to a stream) and the resulting pH is known as a "titration curve"
(Figure 32). The shape of this curve determines how well the effluent pH can be con-
trolled, and the effort necessary to control it within specified limits. The shape of
each curve is related to the strength of the acids and bases in the waste and the
reagent, as well as their relative concentrations. References which detail acid-base
chemistry as related to pH control are listed in the bibliography, particularly
reference 15.
Weak acids and bases are especially helpful in buffering the pH of wastewater. The
most common and useful natural buffers are the carbonates that occur in most water
supplies. Carbon dioxide dissociates water weakly to form hydrogen and bicarbonate
ions:
[H+] [HOD
J 3 J =10""*"" (23)
lc°2]
The bicarbonate ion further dissociates into hydrogen and carbonate ions:
[HCOC
3 1 = 10~ •""•"" (24)
101
-------
I
:
-
0.01N SODIUM CARBONATE
0.0
0.005 0.01
ACID ADDED ((-ion/1)
0.015
Figure 32. Carbonates provide valuable buffering in range of pH 6 to 8.
This double dissociation gives a titration curve with two intermediate plateaus in the
vicinity of pH 6.35 and 10.25. Figure 32 illustrates this for a 0.01N carbonate solu-
tion titrated with a strong acid.
Most water supplies contain a measurable amount of alkalinity, which is expressed as
ppm of CaCO . For waters of pH below 8, all of the reported alkalinity exists as the
O
bicarbonate ion. Alkalinity values are, therefore, directly convertible to bicarbonate
normality. The buffering effect of alkalinity in the vicinity of pH 7 can then be
estimated as a function of alkalinity.
The choice of a reagent for neutralizing a waste should be based on ease of handling
and effluent quality considerations, as well as cost and availability. Among the bases,
lime is by far the easiest to handle, being available dry in bags for small plants, and
in trucks or railroad cars for large consumers. It is also available in a 35% slurry
in tank trucks as a byproduct of acetylene manufacture.
102
-------
Due to the limited solubility of Ca(OH) (1.16 g/1 at 25°C), the maximum pH of a
&
lime solution is only 12.5. Consequently, lime does not represent the hazard to
workers or equipment that caustic does. However, it must be used as a slurry, which
complicates mixing and metering.
Lime is available in two chemical forms: 1) high-calcium lime (93 to 98% CaO), and
2) dolomitic lime (55 to 58% CaO; 37 to 41% MgO). The latter is not generally recom-
mended for effluent neutralization due to the very low solubility of the MgO and its
sluggish reaction rate (16). Even high-calcium lime presents a residence-time
problem due to its limited rate of reaction. Other alkaline reagents are limestone
(CaCO ), caustic soda (NaOH) and soda ash (Na?CO ).
3 " o
The most common acidic reagent is sulfuric acid (H SO ). Its principal advantages
£ TC
are its moderate cost and low corrosiveness in concentrated solution. Mild-steel
vessels and pipe may be used to carry the acid. Other acidic reagents are hydro-
chloric acid (HC1) and carbon dioxide (CO )
^
Dynamic Response
Effective control of effluent pH depends on the dynamic response of the system and
measuring device to a change in reagent flow. The best control will be achieved when
the deadtime or delay in that response is minimized. All other things being equal,
control effectiveness varies inversely with the square of the deadtime. Effective
mixing is one method of reducing deadtime.
Figure 33 depicts a properly designed neutralization vessel. Blending is achieved by
dropping the reagent directly into the influent as it falls into the vessel. The turbu-
lence imparted by the fall provides the motive force. Concentrated reagents—
especially sulfuric acid—evolve heat that is localized at the point of blending. Drop-
ping the reagent into the open stream of influent (as in Figure 33) avoids the severe
corrosion and boiling that would otherwise occur if the two streams met in a pipe.
103
-------
REAGENT
EFFLUENT
Figure 33. This configuration combines a thorough blending with fast
response of pH to reagent flow.
The residence time in a vessel is its active volume, divided by the effluent flow rate.
To make use of the entire volume, entry and exit should be diametrically opposite
each other. The measuring electrodes should be located at the exit so that their
potential truly represents the effluent pH. The agitator provides backmixing by
recirculating the neutral effluent back to the feed point. At the same time, it pumps
the thus-diluted feed to the point of discharge (and measurement) in a fraction of the
actual residence time. Deadtime, as established by tank geometry between the time
reagent is added and the measurement system senses the change, has been determined
theoretically (15). It is half the tank volume, divided by the agitator pumping rate for
cubic vessels laid out as in Figure 33. Verification is provided by Hoyle (17).
Hoyle (17) also demonstrates how feed entering one side of the vessel at the bottom
requires twice as much time to reach the surface at the other side than if surface
feeding is used. The reason for the difference is the downward circulation of the
agitator, which forces the flow upward along the walls. Feed introduced near the
104
-------
bottom must flow upward along the wall, then downward along the shaft, and imally
upward along the opposite wall before it can leave.
Propellers or axial turbines are recommended for mixing because they impart high
turbulence. They are also less expensive than the slower radial turbines since tiiuv
require no speed reduction. Shinskey (15) recommends about 2. 5 HP/1000 gallons o
vessel volume below 1000 gallons, 1.8 between 1000 and 10,000 gallons. Hoyie ana
others (17 and 18) support these recommendations.
Vortex formation must be avoided to properly utilize agitator horsepower. In a smaii
vessel (100 gallons or less), off-center or off-vertical mounting of the agitator may
suffice. In larger cylindrical vessels, vertical baffles are necessary, hi cubic tauns,
the corners tend to break up vortices although, in very large vessels, baffles may
also be required. Influent should not be introduced into corners, nor effluent removed
from them (especially if the vessel is baffled), as this adds deadtime to the response.
Deadtime is least when the depth, length, and width or diameter are equal.
Acid-base reactions are generally instantaneous, but occasionally a lag is encountered
in ionizing the reagent. The limiting factor is the rate of solution of relatively
insoluble reagents such as lime, limestone, and carbon dioxide. Particle size, pur.uy,
velocity, and pH all determine the rate of solution.
If insufficient residence time is allowed for the reaction to go to completion, the pH w_
the effluent at the control point may not be its final pH. If, for example, lime were
added to a pH 2 influent to control the effluent at pH 7, its value could eventually reaeu
9 at some point downstream. A readjustment of the control point to pH 5 could yiekt ^
final pH of 7, but only at a constant load. If the influent pH were to increase to 5 at
some later time, no lime would be added, and the final effluent pH would also be 5.
To minimize the difference between the controlled and final pH, residence time in a
vessel using lime should be 15 minutes or more. The shorter the residence time, the
greater these two values will differ. Shinskey (15) concludes that the slowest reacting
105
-------
component in lime reagent is the 1 to 2% CaCO that is always present. Increasing the
3
pH control point inhibits CaCO solution and thereby yields an ostensibly faster
O
reaction.
Although it may be possible to readjust the control point based on reagent flow in order
to yield a constant final pH, this has not yet been done in practice. However, such a
technique would apply to those processes with insufficient residence time to yield a
stable effluent pH. Another possible solution to this problem is given in a later section.
Concentrated soluble reagents are used at very low flowrates requiring small meter-
ing pumps or control valves with small orifices. Dynamic response of the reagent
flow to changes in the valve or pump control signal can be delayed if the piping down-
stream is allowed to drain freely. The loop seal in Figure 33 is used to prevent this
from happening. The only section of line that drains is downstream of the loop seal;
the seal, therefore, should be located at the point of entry into the vessel, whereas
the valve or pump may be located at any convenient place.
Suspended reagents should be continuously recirculated back to the slurry tank by a
centrifugal pump. The circulation loop should pass as closely as practicable to the
neutralization vessel, so that the slurry may drop directly into it. Figure 34 shows
that the spur from the circulating loop should proceed uphill into the horizontally
mounted valve, and then downhill into the neutralization tank. In this configuration,
the valve will be least likely to plug since the solids will drain away in both directions
when the valve is closed. Flushing the lime valve or downstream piping continuously
with water is not recommended, since its carbonate content (alkalinity) can form a
scale of CaCO where it meets the reagent.
O
The dry reagent may be metered directly onto the surface of the neutralization vessel,
although some delay associated with dissolution will be encountered. Often the slurry
is mixed in a separate tank that overflows into the neutralization vessel. Both the dry
feed rate and the water flow into this slurry tank must be manipulated by the effluent
106
-------
CO
LIME SLURRY
NEUTRALIZATION
VESSEL }
Figure 34. Slurry control valve should be mounted so that solids will
settle away from it.
pH controller, or the slurry concentration will vary. If the water flows at a constant
rate, changing the dry feed rate will affect the effluent pH only gradually because the
concentration of the slurry must change first. With this added delay, the pH con-
troller cannot keep up with even moderate variations in influent conditions.
More than one vessel may be desirable if effluent pH is to be controlled within the
limits of 6 to 9 or better at all times. There are several reasons for this:
• Without heavy buffering, pH in the neutral range is so sensitive to reagent
addition that rapid fluctuations and even cycling are often unavoidable.
• A sizable deviation from setpoint is usually needed before the controller can
adjust reagent delivery to balance a large change in load.
• Temporary overloads are not uncommon, because process vessels are peri-
odically emptied or cleaned, industrial chemicals are dumped into a municipal
sewer system, etc.
107
-------
An additional downstream vessel can attenuate the cycling and brief excursions in pH
resulting from a rapid change in load or momentary overload (19). However, another
pH measurement must be made at its discharge, with recording or alarming, although
control is applied upstream.
An additional upstream vessel (i.e., one without pH control equipment) cannot
attenuate cycling in controlled pH, but it can absorb load changes and temporary over-
loads as readily as a downstream tank. The upstream tank has the further advantage
of providing an opportunity for acid and basic wastes to neutralize one another, thereby
saving reagent. Plants treating both types of waste should, therefore, have an addi-
tional capacity upstream of the neutralization vessel.
Protection against an overload and component failure can also be provided by either
an upstream or downstream vessel. Protection requires the coordination of three
functions:
• Effluent discharge must be stopped when pH deviates beyond limits.
• Off-spec effluent must be recycled for additional treatment.
• Sufficient capacity for accumulating influent must be available to allow sufficient
time for normal operation to be restored.
Figure 35 illustrates a two-vessel system that incorporates the protective features
specified above. The two may be physically separate, or the neutralization vessel
may simply be a partitioned corner of the larger tank. The submersible pump dis-
charges or recycles effluent, depending on the position of the three-way valve. The
valve is actuated by an alarm on effluent pH, allowing discharge only when applicable
specifications are satisfied. Returning the recycle stream back to the point of influent
entry provides some degree of mixing in the surge tank.
To avoid completely emptying the neutralization tank, a switch should deenergize the
pump on low level. With the pump off, reagent flow may be shut off by suitable logic
to avoid any waste. At the same time, the pH controller should be transferred to
108
-------
RECYCLE
INFLUENT
T T-**
2
X1
X1
2
X1
X1
X
i V
\
PARTITION
X1
^
x-
o
O .
1 '
x
•^
X
X
^
^
^>^^>^^^^<^^
DISCHARGE
SURGE TANK
NEUTRALIZATION
TANK
Figure 35. Three-way valve is actuated by an alarm on effluent pH,
preventing discharge of off-limit materials.
manual; otherwise, a large transient will be sustained when the flow begins again.
The agitator may be deenergized with the pump to conserve power. A system like this
is described in greater detail in Shinskey (20, p 57).
Measurement
The measurement of pH has had a reputation of low reliability and high maintenance.
The glass measuring electrode has an extremely high impedance (>100 megohms),
and the circuit is therefore sensitive to electrical leakage. The reference electrode
must maintain a liquid junction with the process fluid and, consequently, is subject to
contamination. Being immersed in the process fluid also presents the possibility of
fouling. However, the major factor in many failures is undoubtedly a lack of under-
standing of the principles of pH measurement.
The hydrogen-ion-sensitive portion of a pH electrode is a thin glass membrane
(Figure 36). Behind this membrane is a solution buffered at pH 7. A new glass
electrode requires an hour or more to hydrate, then the membrane develops a potential
that is proportional to the difference between solution pH and the buffer atpH 7. To
complete the circuit a reference electrode is necessary. The most common reference
electrode used in the U. S. is a silver wire coated with silver chloride and immersed
109
-------
CENTER CONDUCTOR
(WHITE)
SILVER
CHLORIDE
BEAD
SHIELD
(BLACK INSULATION
OVER BRAIDED SHIELD)
SHIELDED
.ELECTRODE CABLE
GLASS-TO-METAL SEAL
-(TOTAL GLASS CONSTRUCTION)
INTERNAL SOLUTION
pH-SENSITIVE MEMBRANE
Figure 36. Construction details of a measuring electrode.
in a KC1 solution saturated with AgCl. It is intended to develop a potential identical
to the internal cell of the glass electrode.
Connection with the process occurs through a liquid junction at the tip of the electrode.
A "flowing" electrode contains a reservoir of KC1 electrolyte, which flows through a
small orifice or porous plug at typically 1 ml/day. A flowing electrode is shown in
Figure 37. Care must be taken to maintain this now at all times, or the junction may
become contaminated and cause an error in voltage. Temperatures below 19° C will
cause both KC1 and AgCl to crystallize from a 4. OM solution, tending to plug the
liquid junction.
Nonflowing electrodes are completely filled with a saturated solution or gel of KC1 and
AgCl. They require neither reservoirs nor pressurization, thereby eliminating many
maintenance and potential problems. Their accuracy is quite adequate for most waste-
neutralization systems. Their life is limited, however, in that some seepage of
electrolyte does occur, with a gradual contamination developing. Process solutions
110
-------
INTEGRAL SEALED LEAD
RESERVOIR _
FILLING SCREW
GLASS
TUBE
ELECTODE
LEAD
RETAINER
COTTON
RETAINER
DIALLYLPHTHALATE
CAP
SEAL
ELECTROLYTE
SOLUTION
KCI SATURATED
WITH AgCI
PACKED COLUMN OF
SILVER CHLORIDE
CERAMIC JUNCTION
Figure 37. Construction details of a flowing reference electrode (21).
containing sulfides or other ions with a great affinity for silver usually bring about
more rapid deterioration.
Because the glass and reference electrodes have similar internal cells, no temperature
error exists at pH 7 (0 mV), since waste neutralization is concerned primarily with
accuracy in the pH 6 to 8 range, where the error is minimal. If temperature com-
pensation is required it is readily available by inserting a resistance temperature
detector into the electrode chamber. This resistance varies the gain of the pH
amplifier.
In industrial installations the solution being treated is grounded through vessels,
piping, etc. Because its ground potential may differ from that at the pH meter,
ground currents could flow through the electrodes and cable. These currents are
most likely to seek the path of lowest resistance; in the absence of a better conductor,
111
-------
this would be through the reference electrode. Although the ground currents tend to
be ac, they can be rectified by the reference cell, producing a dc millivolt error. A
solution ground wire is recommended to carry these ground currents from the solution
to the instrument ground through a capacitor.
Possible Causes for Failure of Electrodes and Assemblies
In the event of failure of the electrodes or assemblies the operator will observe one of
these possible symptoms: insensitive electrode, calibration drift, or slow electrode
response.
When a pH measurement seems unusually constant compared to its past behavior, a
malfunction is to be suspected. The usual diagnosis for this failure is electrical leak-
age which is caused by liquid leakage into the assembly. Field experience with sub-
mersible electrode assemblies has demonstrated how difficult it is to keep moisture
out. Process solution could enter the lower chamber through an ineffective seal, or
moist air could be drawn into the terminal box. Extreme variations in ambient con-
ditions are probably responsible for most of the trouble. A reduction in ambient
temperature will cause a partial vacuum within the assembly, along with condensation
of moisture. If a leak does exist, more moist air will be ingested, with further con-
densation. Since moisture can enter but not escape, it accumulates. The moisture
changes the resistance of the cables and terminal strips significantly, so as to cause
a zero cell potential (pH = 7.0).
Packets of silica gel have been used to dry the air inside the assembly, but these are
effective only in the absence of leaks. If a leak exists, the packets accumulate the
moisture and short-circuit any terminals they touch.
Submersible assemblies (like the one shown, in Figure 38) are most susceptible to
leakage. Their superstructure is often at a very low or widely varying ambient tem-
perture, while the submerged portion is at a solution temperature. The leakage of
either solution or atmospheric moisture into the assembly can be eliminated with the
112
-------
REFERENCE
SOLUTION
TERMINAL
BOX
TO OTHER
ASSEMBLIES
•TEE
INSTRUMENT
— AIR
PURGE
FLOWMETER
1/4" PLASTIC
TUBING
COUPLING
Figure 38. An air purge of submersible electrodes can eliminate leakage
problems altogether.
air purge shown in Figure 38. The bubbler hanging below the electrodes will build a
back pressure of air equivalent to the liquid depth. Then any penetration will allow
only air to leak outward, rather than moisture to leak inward.
Instrument air is typically dried to, at most, a dewpoint of 4. 5°C at 690 kPa (40°F at
100 psig), which corresponds to -20. 5°C (-5°F) at an atmospheric pressure. Thus,
condensation resulting from this source is not to be expected.
When the pH measurement is responsive, yet its calibration seems to drift from day
to day, the reference electrode is suspect. The most probable cause is a plugged
liquid junction, probably caused by the backflow of process fluid into the electrode, or
by the precipitation of silver chloride from the reference solution. This contamination
with the process solution will develop a variable electrode potential.
If backflow of process fluid appears suspect, then the solution is to relocate the follow-
ing reference electrode within the reaction vessel so that proper flow is assured. A
flowing reference electrode requires about 3 feet of head above the vessel liquid level
113
-------
to maintain a proper flow. If the vessel is pressurized, the electrode must be pres-
surized that much higher. (Vendors have kits available for this purpose.) On the
other hand, if a flowing electrode is plugged with silver chloride, the electrode, its
tubing, and reservoir should all be flushed free of crystals and refilled with 1.0 Molar
solution. Restandardization of the pH meter may be required for a few days following
the change.
A contaminated nonflowing reference electrode cannot ordinarily be repaired, but
must be replaced.
Calibration drift can also be caused by an open solution ground or defective ground
capacitor. It may also appear in the first hour after a dry glass electrode is installed,
as the membrane hydrates.
Slow response of the electrode is another problem and the distinction between insensi-
tivity and slow response is important. When the electrode assembly is moved from
one solution to another, a failure to indicate the proper change in pH is considered
insensitivity. If the change is indicated but only after several minutes have elapsed,
sensitivity is there but response is slow.
In a control loop, insensitivity appears as a constant pH, while slow response causes
oscillation of an increasingly or abnormally long period. Its primary cause is a film
covering the glass membrane, which does not restrict the electron flow by electrical
impedance but, rather, insulates the electrode from the varying composition of the
solution. The ion flow from the solution to the electrode is impeded by the coating.
Coatings can develop from a variety of sources. Precipitation of such solids as
CaSO., CaCO , and the hydroxides of heavy metals causes fouling. Greases, fats,
4 o
and polymers already in the waste can also cause problems. Even a thin film of these
substances may delay response significantly. Usually flowing reference electrodes
are not susceptible to fouling, but glass and nonflowing electrodes are.
114
-------
Installation and Maintenance
The pH meter or pH-to-current converter should first be calibrated with a portable
potentiometer or electrode simulator connected in place of the electrodes. Only after
the instrument is calibrated to the millivolt range corresponding to the desired pH
scale should the electrodes be connected. A new glass electrode should be given an
hour or more to hydrate and after that the electrode should not be allowed to dry out
while not in use.
Standardization should always be made with the electrodes in a buffer solution of known
composition. Distilled water should not be used for standardization because it is too
easily contaminated. The pH meter allows a one-point calibration only, but always
use two different buffers to confirm electrode sensitivity.
Whenever moisture is found in an electrode assembly or junction box, it should be
removed with acetone, methanol, or a similar water-miscible, volatile solvent,
followed by air-drying.
Coatings on the electrodes may be removed by periodic immersion of the entire assem-
bly into a washing solution. Water-miscible solvents are suggested for greases and
fats. A 10% (or thereabouts) solution of H SO or HC1 is extremely effective in
2 4
removing salts and hydroxides.
When fouling is rapid, an ultrasonic cleaner (available from most instrument vendors)
is recommended. This has been proven effective against clay, hydroxides, and
calcium sulfate and phosphate precipitates. They do not seem to be effective on
elastic polymers such as latex.
Mechanical cleaning is not recommended in that it is not as effective as the above
methods in removing thin or hard films, and the probability of breakage is increased.
The presence of carbonates in alkaline waters reduces the sensitivity of the pH of
water to the addition of strong acids and bases. An alkaline water containing 100 ppm
115
-------
CaCO equivalent at pH 8 may contain only 2 ppm when the pH is reduced to 5 or less.
O '
Below this level, virtually all bicarbonate ions are converted to carbon dioxide, which
is then lost to the atmosphere. Neutralization with a hydroxide cannot restore this
lost buffering capacity.
In the absence of buffers, the sensitivity of pH to reagent addition may be quite high.
This means that a very small error in the ratio of reagent to influent flow can change
the effluent pH significantly. For example, to provide effective control within a pH
range of 6 to 8, the system would tolerate a 10% mismatch in terms of chemical
equivalents between a basic reagent and an acid waste at pH 4, a 1% mismatch for a
pH 3 waste, a 0.1% mismatch for a pH 2 waste, and so on. This example should
indicate not only the sensitivity of the lightly buffered pH curve, but the effect of the
initial concentration of the waste on control.
This extreme sensitivity makes control both difficult and necessary. The high gain
of the pH curve requires a low controller gain if continuous cycling is to be avoided,
but a low controller gain leaves the process susceptible to upsets in influent condi-
tions. For these reasons, the process should be designed to be as controllable as
practicable, following the guidelines given previously. Of equal importance is the
application of control techniques and devices especially designed or selected for pH
neutralization.
Final Control Elements
Final control elements such as metering pumps and control valves are discussed
below.
Metering pumps with variable-speed drives are quite satisfactory for manipulating
the flow of strong reagents. They are both linear and responsive, and are available
in a sufficient variety of materials to deliver most commercial solutions used in
neutralization. They can be sensitive to vapor lock or plugging with suspended solids.
116
-------
Perhaps their most stringent limitation in waste neutralization is a rangeability of
only about 20:1. Below about 5% of full speed, the motor tends to stall. When the
reagent demand falls below this, limit cycling of the effluent pH will result. If an
adjustable stroke is combined with adjustable speed, the rangeability is extended to a
multiple of their individual rangeabilities. However, the gain of the pump then varies
with stroke, affecting the performance of the control loop. Attempting to extend the
rangeability by sequencing pumps of different size is also encumbered by the gain
change with size. As a result, wide-range delivery is probably best accomplished
with control valves.
Reagents are most commonly delivered through one or more control valves supplied
from a head tank (in the case of solutions) or a recirculating loop (in the case of
slurries). Valves are available with rangeabilities from 35:1 up to 100:1 or more,
and in linear and equal-percentage (logarithmic) characteristics. For throttling
slurries, ball valves are recommended—they have inherent equal-percentage char-
acteristics.
A valve is said to be linear if a linear relationship exists between the stem position
and the flow of liquid through the valve at a constant pressure. For equal-percentage
valves, equal changes in valve stem position result in the same percentage change in
liquid flow. In theory an equal-percentage valve will never shut off, but manufacturers
provide modifications which allow the valve to shut off.
A linear characteristic is desirable for all applications except neutralization of a
single, dominant weak acid or base. In that case, buffering varies with reagent
demand. Fortunately, the variation in the gain of an equal-percentage valve in propor-
tion to reagent flow will compensate this effect. In all other cases, the gain of the
titration curve is either constant or not singularly related to reagent demand.
When the valve being used has the wrong characteristic for the process, characteriza-
tion should be added. It may be implemented with a nonlinear positioner, a diode func-
tion generator, or a specially configured analog divider.
117
-------
A valve positioner compares the control signal to the valve stem position, and acts
on the motor to make them agree. Its most important function is to eliminate the
hysteresis that is common to all actuators. Positioners should be used on all control
valves used in waste-neutralization processes. Additionally, some positioners are
available with contoured cams capable of positioning the stem as a nonlinear function
of the control signal.
Figure 39 illustrates the nonlinear characteristic obtained by connecting the controller
output (x) for an equal-percentage valve into both the numerator and denominator of a
specially scaled divider. The divider will develop a biased signal which can be used
to linearize an equal-percentage valve. When the divider output (y) is sent to the
(linear) valve positioner, the overall relationship between the flow and controller
output is nearly linear as shown. The divider is capable of higher gains than a con-
toured positioner and can, therefore, provide more exact linearization.
100
OUTPUT, % 50
50:1 EQUAL PERCENTAGE
CHARACTERISTIC
100
Figure 39. A properly scaled divider can effectively linearize an equal-
percentage valve characteristic.
118
-------
When both acid and basic reagents are used for neutralization, the valve positioners
require different calibration. Both valves should be closed at 50% controller output,
with the base valve fully open at 0% and the acid valve fully open at 100%. During an
air-supply failure, however, both valves should close. Additionally, a solenoid valve
on the air supply to the positioners can then be used to close both valves in case of a
power failure or when the operator wishes to disable the system for maintenance, etc.
If both valves are equal percentage, the required function for linearization takes the
form of an "S, " with the highest gain at midscale. This curve can best be obtained
with a diode function generator.
Rangeability may be extended orders of magnitude by sequencing equal-percentage
valves. Consider the need to manipulate flow over a 1000:1 range from 0.01 to 10
gpm. A 50:1 valve can throttle only to 0. 2 gpm, but another 50:1 valve could cover
the range from 0. 01 to 0.5 gpm. Figure 40 shows the combined flow range on semi-
logarithmic coordinates.
FLOW GPM
0.5
0.2
O.I
0.01
SMALL
"VALVE
t
LARGE
'VALVE
50
CONTROL SIGNAL, %
100
Figure 40. Two equal-percentage valves may be sequenced to act as a
single wide-range valve.
119
-------
The respective valve positioners must be calibrated to give full stem travel over the
ranges of the control signal indicated on Figure 40. In addition, the valves must not
be open at the same time. A pressure switch set to close when the control signal
exceeds the full output of the smaller valve and to open below the minimum flow of the
larger valve can actuate three-way solenoid valves to select which control valve will
be open.
Feedback Control
The accuracy with which reagent flow must match influent demand places the burden
of meeting effluent pH limits squarely on the feedback controller. Although feed-
forward control may be used in some situations and may be helpful, it cannot approach
the accuracy of reagent delivery needed. All of the foregoing recommendations
regarding vessel layout, mixing, etc., are intended to make feedback control more
effective.
An ideal feedback controller can be described by the following equation:
(*• \
if de \
e + - / edt + D ^ (25)
R J dt / ^ '
Output m is proportional to the deviation e between measurement and setpoint through
the proportional band P, expressed in percent. The output is also related to the time
integral of the deviation through the integral or reset time constant R. Finally, the
output is also affected by the rate of change or time derivative of deviation through the
derivative time constant D.
The primary controlling modes are proportional and integral. Derivative is helpful in
accelerating response to upsets, but its effectiveness is limited by sensitivity to noise
(uncontrollable rapid fluctuations) in the measurement signal.
Integral is necessary to reduce the deviation to zero and, the smaller the integral
time, the faster the deviation will be reduced. R cannot be reduced below the deadtime
120
-------
in the loop, however, or an undamped oscillation will result. Theoretically, R and D
can both be set equal to the deadtime times 2/V, as developed by Shinskey (22). How-
ever, variations in the slope of the titration curve, reaction rate, and electrode
response can lead to instability with such close settings. In practice, the two settings
are best spaced 4:1 apart, such that R is 4/rr times deadtime and D is deadtime
divided by TT. In situations where derivative is not or can not be used due to a high
noise level, R ought to be increased to two deadtimes for stable response.
The proportional mode combines stability and responsiveness, but P must be set high
enough to avoid continuous cycling. Most pH curves are sufficiently sensitive in the
neutral range that stability requires a very wide proportional band, often exceeding
200 or 300%. As a result, the controller is not especially responsive to upsets, and
large deviations tend to persist. This problem is due not only to the sensitivity of the
pH curve, but also to its extreme nonlinearity. Small deviations from setpoint
require the slightest corrective action, whereas even moderate deviations may require
drastic changes in reagent delivery.
In cases where the proportional band cannot be set wide enough for stability due to its
limited range, a limit cycle will appear. This cycle tends to be of a regular period
and nearly uniform amplitude, usually—although not always—symmetrical about the
setpoint. Any variation in its amplitude would indicate a change in the titration curve.
Adding a compensating nonlinear function to the controller tends to satisfy both the
extreme proportional-band requirement and the nonlinearity of titration curves. Con-
trollers that have simple nonlinear functions imposed on the deviation signal prior to
action by the three conventional control modes are commercially available. The func-
tion consists essentially of three straight line segments that are symmetrical about
zero deviation, as shown in Figure 41. The width of the gap is adjustable, as well
as the gain within it, to permit matching to the titration curve; gain outside the gap is
unity.
121
-------
ERROR SIGNAL, e
Figure 41. This nonlinear function greatly improves pH control.
In practice, the gap width is set to quench the limit cycle that would otherwise develop
with a nominal proportional band of 100%. The proportional band is then set to
recover rapidly from deviations exceeding the gap, but also to avoid limiting cycling
outside the gap. The gain within the gap is set to avoid cycling within it. Too low a
gain setting, however, may promote a slow cycle that would be equal in amplitude to
the gap width, as explained by Shinskey (15).
A companion to the nonlinear controller is available, which automatically adjusts the
gap width for applications where the titration curve varies extensively. This instru-
ment is described by Shinskey (23). Called an adaptive controller, it senses an
oscillatory or drifting condition of the pH loop and restores stable control.
The same pH deviation on which the nonlinear controller operates is sent to the adap-
tive controller. Its output is, in turn, connected to a remote gap-width adjustment in
the nonlinear controller. Oscillation at or near the natural period of the pH loop will
cause the adaptive controller to gradually expand the gap until stability is restored;
122
-------
the gap will then remain at the value until another undesirable condition appears.
When a steady or slowly drifting deviation develops, the adaptive controller will
gradually narrow the gap until zero deviation is achieved.
The rate at which the gap is changed is based on the magnitude of the deviation and the
integral time of the adaptive controller. A low-frequency gain adjustment is available
to reduce the rate of gap closure compared to the rate of expansion. This additional
parameter was found necessary for stability due to the extreme difficulty in controlling
pH precisely at the setpoint. Considering that it is required only to control within a
band (e. g., 6 to 8.5), small amounts of offset are tolerable, whereas small-amplitude
cycling may not be.
A third adjustment tunes the adaptive controller to the natural period of the pH loop.
It is important that the period be nearly constant when an adaptive controller of this
type is used. Consequently, in cases where electrodes foul rapidly, continuous
cleaning (as with an ultrasonic cleaner) should be provided.
The adaptive controller is especially useful when the influent is comprised of a
multiplicity of wastes, including both strong and weak acids and bases. Batch opera-
tions, cleaning and rinsing, and periodic shutdown of parts of the production may cause
these agents to come and go in varying proportions. As a result, the nonlinear con-
troller may rarely be properly adjusted without automatic adaptation. Persistent
limit cycling, during which acid and basic reagents are alternately added, can be
avoided with adaptation. A properly adjusted nonlinear-adaptive control system
requires only a few cycles to expand the gap to the point extinction.
Feedforward Control
Feedforward control is defined in Reference 24 as: "Control in which information
concerning one or more conditions that can disturb the controlled variable is converted
into corrective action to minimize deviations of the controlled variable. "
123
-------
With regard to waste neutralization, the "conditions" mentioned above are the demands
for reagent needed to neutralize the influent. To achieve neutrality:
VA=FBNB (26)
where F and N represent the flow rate and ihe chemical equivalents of acid (A) and
base (B). Either may be the waste or the reagent. To apply feedforward control,
Equation 26 is solved for the reagent flow in terms of the variable flow and equivalent
concentration of the waste, as well as the fixed equivalent concentration of the reagent.
If the waste flow can vary rapidly but its composition cannot, then the feedforward
system reduces to a simple flow ratio of reagent to influent, with the ratio adjusted
by the effluent pH controller. This system requires a linear influent-flow measure-
ment and a linear reagent-flow control loop or final element. The principal drawback
to these linear systems is that their rangeability is limited to that of the final element
and influent flowmeter.
When a metering pump is the final element, either of the two arrangements shown in
Figure 42 may be used. If the pump has only variable speed or stroke (but not both),
a multiplier is required to combine the influent-flow signal with feedback trim from
effluent pH. If both stroke and speed are remotely manipulable, the feedforward
signal may manipulate one while feedback adjusts the other. Their multiplication is
inherent in that reagent flow is the product of stroke and speed.
The interlock between the reagent flow and the influent pump described previously is a
flow-feedforward system in which only two values of flow are used. Regardless of the
reagent flow at the time the pump is deenergized, it will drop to zero. When the pump
is restarted, the reagent flow will return to its last value, which is the most probable
new demand of the influent.
Plants wherein wastes are directed to the neutralization facilities under gravity head
do not impose sudden changes in flow on those facilities. Influent flow cannot change
124
-------
INFLUENT
FLOW
FIXED
STROKE
EFFLUENT
pH
SPEED
METERING PUMP
INFLUENT
FLOW
VARIABLE
STROKE
EFFLUENT
"pH
(SPEED
METERING PUMP
Figure 42. Flow-feedforward systems for fixed stroke (A) and
variable stroke (B) metering pumps.
without first raising or lowering the level of liquid in the channels feeding the facility.
The capacity of these channels, therefore, absorbs the flow variations effectively.
However, spills, rinses, and sudden discharges tend to be carried through the chan-
nels almost as if they rode a conveyor. Arriving at the neutralization facility only
moderately diluted or blended with the remainder of the influent, they impose sudden
reagent demands on the control system. An influent surge tank, as shown in Figure
35, is recommended to distribute these upsets. But where such a tank does not exist
or can not be installed, feedforward control is the only choice remaining to diminish
their effect on effluent pH.
Equation 26 could be solved directly by the control system in terms of reagent flow,
if influent concentration were known directly. Analyzers that can make the determina-
tion of total acidity or alkalinity are available, but their dynamic response tends to be
slower than the process itself. As a result, they are useful only in the isolated case
of an especially slow feedback-loop response.
125
-------
Influent pH is the measurement normally used for feedforward control (19). Convert-
ing Equation 26 to logarithms will indicate the relationship between reagent flow and
influent pH, if ionization is complete.
log F + log N = log F + log
log FA - PHA = log FB - PHB
(27)
Equation 27 is implemented by using an equal-percentage (logarithmic) valve to relate
reagent flow to influent pH. The schematic for such a system is shown in Figure 43.
The actual equation solved by the system is:
s = m ± K (pH - r)
(28)
where s is the signal to the valve(s), m is the feedback controller output, and K is a
gain adjustment. The term "r" is a feedforward reference value that represents the
influent pH at which there is no feedforward contribution to the valve position. The
sign preceding K is positive for a basic influent and negative for an acidic influent.
For completely ionized influents, K is adjusted to produce a tenfold change in reagent
flow per unit change in pH. [Its numerical value depends on the rangeability of me
valve(s) and the span of the pH measurement.] Single weak acids and bases that
change their pH only 0.5 unit per decade concentration would require K to be doubled.
ADAPTIVE
SIGNAL
GAP
INFLUENT
pH
EFFLUENT
pH
NONLINEAR
CONTROLLER
EQUAL-PERCENTAGE
VALVE(s)
Figure 43. pH feedforward system is capable of a wide range through
sequenced equal-percentage valves.
126
-------
Mixtures of strong and weak components, however, vary the relationship between pH
and concentration so widely (15) that a field adjustment to the best average value is
the final recourse. This factor severely limits the improvement that feedforward
control brings to the neutralization process (19).
When manipulating equal-percentage valves for pH control, the gain change caused by
the valve characteristic adversely affects the stability of the feedback loop. To over-
come this problem, the feedforward component of valve position [i. e., K(pH - r] is
used to set the gap width of the nonlinear controller. Thus, when the influent pH is
moderate (calling for little reagent), the gap will be narrow (compensating for a
diminished valve gain). Figure 43 indicates this as an adaptive loop in that the con-
troller is automatically adjusted as process conditions change.
At present, the hydraulic analog control has not been proven in practice. Its success
is contingent on the satisfactory performance of certain hydraulic operations that still
require some development. Nonetheless, it is included here for possible further
examination by ffiose who are confronted with a particularly unyielding neutralization
problem.
The problem is the lack of sufficient residence time for neutralization with lime in a
conduit having no backmixing. Figure 44 illustrates the addition of a small stirred
LIME
SLURRY
Figure 44. Small tank is a hydraulic analog of larger stream.
127
-------
tank with sufficient residence time, to which a sample of the treated stream is sent.
The pH in the tank is controlled by adding a reagent to the main stream.
Whether such a configuration is workable depends on uniformly distributing the
reagent across the stream and obtaining a representative sample of the treated
stream. Even a 1% variability in either of these functions could cause a 1 or 2 pH
deviation in the final condition of the effluent. Distribution may be more difficult than
sampling in that it must be effective over the entire range of reagent flow. The weir
will help distribution, and downstream baffling should improve the blend. Locating
the sample point too far downstream, however, will add to the deadtime of the control
loop.
PHOSPHOROUS REMOVAL BY LIME TREATMENT
Owing to the increasing concern over the rate of eutrophication of our surface waters,
many states have instituted standards for phosphorus removal. The areas most
affected are wherever both industry and agriculture are concentrated along lakeshore
and riverfront, particularly in areas bordering on the Great Lakes. Typical among
standards instituted by these states is Indiana's Water Quality Standards SPC1R-3,
requiring 80% removal or 1. 0 mg/1 phosphorus, whichever is more stringent.
In order to meet these or any standards, adequate measuring and controlling systems
must be applied to the treatment process. One of the available phosphorus removal
processes is the precipitation of calcium phosphate salts through addition of lime.
The design aspects of this process are taken primarily from the "Process Design
Manual for Phosphorus Removal" (25).
General Theory and Process Dynamics
The process design manual cites the three principal forms in which phosphorus may
enter wastewaters: orthophosphate, polyphosphate, and organic phosphorus com-
pounds. The organic phosphorus compounds largely break down during biological
128
-------
treatment into the orthophosphate. Polyphosphates come primarily from detergents
and corrosion inhibitors, but they, too, ultimately hydrolyze into orthophosphates.
Consequently, phosphorus control in wastewaters is centered around the removal of
the orthophosphate ions by precipitation.
The solubility of salts is governed by an equilibrium between their ionic components.
Consider the precipitation of calcium orthophosphate:
3Ca2+ + 2P03~ - Ca(P0) J (29)
When the product of the activities of the individual ions raised to the powers of their
proportion in the solid reaches the solubility-product constant, the solution is satur-
ated with respect to that solid:
[Ca2+]3 [ PC3" ) 2 = 10"26 (30)
Bracketed terms refer to ionic activity in units of mols/liter. Under the dilute condi-
tions usually found in wastewaters, activity and concentration are nearly identical.
Therefore, the bracketed terms may be considered as concentration in mols/liter with
little error in most cases.
Stoichiometric equations such as Equation 29 can be used to estimate the dose of
calcium required to react with the orthophosphate. Solubility product equations such
as 30 indicate the excess quantity of calcium needed above the Stoichiometric demand
to drive the precipitation reaction toward completion.
The multi-basic nature of the phosphate species complicates the picture, however.
As a function of pH (to be described below), a hydrogen-phosphate ion forms, which
may also be precipitated by calcium ions:
Ca2+ + HPO2~ ^ CaHOP, J (31)
4 4
129
-------
The solubility of this salt is given as:
2
-7
(32)
Figure 45 illustrates the theoretical solubility of phosphates in lime solutions, plotted
as total dissolved P on a logarithmic scale against pH. Assuming an initial concentra
tion of 30 mg/1 P, no precipitation will result from lime addition until pH 6. 5. If
more phosphorus is present, precipitation will begin at a lower pH or, if less, at a
higher pH. The curve representing mg/1 P vs pH for CaHPC>4 precipitation is fixed;
only the point of intersection with the initial concentration is dependent on that initial
concentration.
As pH is raised further by lime addition, the rate of CaHPO precipitation (the slope
2_
of the curve) diminishes. Above pH 8, enough PO ions are formed to begin
100
50
:
p
mg/l
0.5
INITIAL
CONCENTRATION
CaHP04
PRECIPITATES :
„_ Ca3 (P04)2
— PRECIPITA
CIPITATES
-
pH
Figure 45. Theoretical solubility of phosphates in lime solutions as a
function of pH.
130
-------
precipitation of Ca (PO ) . Butler (26) calculates the pH of a solution saturated with
o 4 &
both CaHPO and Ca (PO ) as 8.06.
~t O ^E ^
As the pH increases above this point, [ PO I concentration increases, which reduces
r 2+1
[ Ca j concentration. Consequently, CaHPO no longer precipitates above pH 8.06.
r 2+1
Lime addition above pH 10 begins increasing Ca again, causing total P to drop
sharply.
However, Figure 45 describes a theoretical relationship for the solubility of only the
orthophosphates. Other forms of phosphorus besides orthophosphate may or may not
precipitate, resulting in a different level of total P. Albertson and Sherwood (27)
have found that 9 mgA PO (which would be 3 mg/1 P) at pH 9 and 4 mg/1 PO (1.3
mg/1 P) at pH 10, are soluble after lime treatment of domestic sewage.
Figure 46 indicates lower solubilities of total P at pH 9 to 11 than theory indicates.
All the points represent data taken from the EPA pilot plant at Lebanon, Ohio. The
reduction in the phosphorus level below the solubility of calcium phosphates must be
due to other reactions. The wide variation in the data points indicates the variability
2.0
THEORETICAL
SOLUBILITY
OF Ca3 (P04)2
FROM FIGURE45
9.0
9.5
10.0 10.5
CLARIFIER pH
11.0
11.5
Figure 46. Actual measurement of effluent phosphorus from a single-stage
lime treatment process as a function of pH, compared to
theoretical relationship.
131
-------
of these undefined reactions or the components in the water that determine their
course. The disagreement between this data and that cited above by Albertson and
Sherwood (27) indicates that other characteristics of the water or the treatment plant
have some influence over the phosphorus level in the effluent.
The conclusions that can be drawn from the above information are these:
• pH control is necessary for effective phosphorus removal.
• Phosphorus concentration in the effluent cannot be correlated exactly with pH
from one plant to another, nor even from time to time within the same plant.
It has been demonstrated that lime addition to a pH of 9.5 or 10 is necessary to reduce
effluent phosphorus levels below 1.0 mg/1. If the effluent were to be discharged to a
receiving body of water at these high pH levels, damage to aquatic life could result.
Hence, the pH must be reduced below 8.5 (or whatever applicable standards require)
by adding an acid of some type. Addition of CO to lower the pH (recarbonation) is
Zt
widely practiced.
If the wastewater alkalinity is already high (>150 mg/1 CaCO ), lime addition to
O
pH 9.5 to 10 will precipitate enough CaCO along with the phosphates to form a settle-
O
able floe. After clarification, the pH is adjusted to 8 or thereabouts by recarbonation.
The data in Figure 46 came from a single-stage plant such as this.
With wastewaters of low alkalinity, lime addition to pH 9.5 will not precipitate enough
CaCO for good floe formation. In these situations, the pH must be raised by lime to
3
11 in order to precipitate Mg(OH) and as much CaCO as possible, along with the
£t O
phosphates. Then a first-stage recarbonation to pH 10 precipitates more CaCO to
3
promote floe formation. After clarification, a second stage of recarbonation adjusts
the pH to its final value of 8 or thereabouts. No precipitation takes place during the
second stage, since the minimum solubility of calcium in a system saturated with
CO occurs about pH 10.
£t
Lime treatment for phosphorus removal may be applied to the raw wastewater prior
to biological treatment. CO generated during biological treatment is absorbed, and
£t
132
-------
may lower the effluent pH to acceptable levels, so that no neutralization is required.
A pH of less than 9. 5 is generally accepteable prior to biological treatment (25).
The success or failure of chemical waste-treatment processes hinges to a great
extent on the accuracy and responsiveness of the chemical feeding system. This is
particularly true when a slurry is being fed. If failures are frequent due to plugging,
etc., controls are usually bypassed and feed rates left at some constant maintainable
value. This practice results in the overfeeding and waste of chemicals, and usually
creates other problems. But a well-designed feeding system can be both reliable and
responsive, and will operate indefinitely under automatic control.
In tonnage quantities, pebble lime (CaO) is slaked with water to form hydrated lime
[Ca(OH) ] in an apparatus similar to that shown in Figure 47. Because the reaction
£1
is exothermic, heat is released as steam. Consequently, enough water must be
supplied to make up the loss as well as provide the desired concentration of slurry.
In order to avoid having to adjust feeding the slaker over the entire range of lime
rates demanded by the waste-treatment process, the slaker should be operated on-off.
As the level in the slurry tank reaches a low point, the lime feeder and water flow are
started at preset rates. When the level reaches the high limit, both are stopped.
Because the lime-slurry control valve may be only 1/2 inch in size, a vibrating
screen should be used to take out the gravel always present in pebble lime. The lime
slurry should be continuously circulated to avoid plugging. The control valve should
be mounted at the highest point in its branch, so that solids will not settle there when
the valve is closed. In addition, there should be no horizontal sections in the branch
containing the valve. The valve should be mounted as close to the point of discharge
into the reaction vessel as possible, to minimize any delay in response to the control
signal. Avoid flushing continuously with water though, as CaCO tends to form a
O
scale wherever the water meets the lime.
The control valve should be a ball valve for rangeability and resistance to plugging.
To minimize wear from sand in the slurry, the piping should be selected for a 5 psi
133
-------
LIME
VIBRATING
SCREEN
LIME SLURRY CONTROL
BALL VALVE
CIRCULATING
LOOP
Figure 47. Control of lime slaking.
or less drop across the valve. Since control ball valves are not available in sizes less
than 1/2 inch, the slurry concentration should be selected with this in mind.
A magnetic flowmeter is recommended since it offers an unobstructed path and pro-
vides a linear measurement of flow. It must be mounted in a vertical or sloping upflow
section, as shown In Figure 47. A flow controller closes the loop to the valve; how-
ever, it may in turn be set by a pH controller or other signal, as described later in
this section.
Smaller plants will find hydrated lime in 50-lb bags easier to use. Typically, the dry
lime is metered into a slurry tank, as shown in Figure 48. The slurry then overflows
into the reaction vessel at a rate proportional to the flow of water into the slurry tank.
134
-------
H
SPEED
CONTROL
LIME
FEEDER
CONTROL
"SIGNAL
Hi
i
i y
V
1
>S
•WATER
SLURRY TO
REACTOR
Figure 48. Control of dry lime feed.
To simplify the controls, some contractors fail to put a control valve on the water,
leaving it to enter at a constant rate. This practice unfortunately creates a serious
time lag; when the lime feeder is stopped, for example, the lime in the slurry con-
tinues to overflow until the tank is depleted. This renders control over the reaction
vessel virtually impossible.
But if a linear control valve on the water is made to follow the same signal as the
lime feeder, the concentration of the slurry will stay nearly constant. When the con-
trol signal changes, the rate of overflow will then change, and the reaction will be
responsive and controllable.
Figure 49 compares the records of pH control of a neutralization vessel. In the top
record, water into the slurry tank is flowing at a constant rate. In the lower record,
water flow as well as lime feed rate is manipulated by the pH controller. The second
pH measurement shown in both records represents the effluent from the clarifier
downstream of the neutralization vessel. The superior degree of control achieved
135
-------
C I
-
n
fa*
\a*LCA fJt&Tt
^i/j-.He - A' 'i
t •"
i
• ;
•
Figure 49. Upper record shows uncontrollable pH caused by water flowing at a constant rate into slurry
tank; when water flow was manipulated with lime flow, lower record was produced.
-------
with the latter system is obvious. (Note that time proceeds from right to left on the
charts.)
Lime calcining or exhausts from combustion engines or submersible burners are
economical sources of CO for large users. If the source is dedicated CO genera-
£t £l
tion for recarbonation, then the control signal simply manipulates the fuel flow or
burner mechanism. Since any combustion system has a low fuel limit to avoid flame-
out, the rangeability of these sources is not wide.
If the exhaust from a calciner or internal combustion engine is used as a CO source,
its availability is fixed by other considerations. Whatever CO is not used for
Lt
recarbonation must be vented to the atmosphere. The control signal would then
manipulate dampers in the feed line and stack, opening one while closing the other.
Since the CO from this source may vary between 6 and 18%, automatic control from
a continuous signal such as pH must be used, rather than simple pacing or flow control.
Carbon dioxide may be purchased as a liquid under pressure in cylinders. Feeding
then requires vaporization and metering of the dry gas. For rates less than 1000 Ib
per day or so, enough heat is available from the surroundings for evaporation. Higher
rates require a source of heat and temperature or pressure control.
Commercial feeding equipment is available from several vendors, who also furnish
chlorinators. The feeding equipment usually has built-in metering components, so
that flow is proportional to a standard control signal. The response of the process to
control is usually satisfactory if the feeder is located not more than 100 feet from the
recarbonation vessel.
Process Control
Since wastewater flow rates tend to fluctuate diurnally (if not more often), feeding
chemicals at a constant rate cannot be satisfactory in the long term. If the composition
137
-------
of the waste and the chemicals are both reasonably constant, however, results will
be favorable provided dosage is maintained uniformly.
Dosage control requires first an accurate measurement of wastewater flow. This
signal must then be linearized (if it is nonlinear) and applied to the chemical feeder
or flow controller through a ratio or dosage adjustment. Lime feeders (such as that
shown in Figure 48) usually have a numerical but uncalibrated proportioning adjust-
ment. Carbon-dioxide feeders typically have an uncalibrated proportioning adjust-
ment, but a gas flow indicator is available to verify the dosage.
The most positive—but also most expensive—method of chemical feeding uses a flow-
meter and controller, as shown in Figure 47. This controller may be set to deliver
a fixed flow rate, or may be set in proportion to wastewater flow. The proportioning
or ratio setting may even be calibrated in terms of mg/1 lime dosage if the lime
slurry is of uniform composition. A positive calibration like this is very helpful to
operators in making adjustments based on laboratory analyses of either raw waste-
water or treated effluent.
Figure 50 shows how to convert an open-channel flow measurement into a linear
signal. Rectangular weirs and Parshall flumes produce a head that is related approxi-
mately to the 2/3 power of the flow. The differential pressure (dp) transmitter sends
the head signal to a square-root extractor, which produces the 1/2 power of head;
this signal, when multiplied by head, yields the 3/2 power, which is linear with flow.
The exact relationship between head and flow varies from 1.52 to 1.58 power,
depending on the width of the throat. Using the 3/2 power results in an error of less
than ±1%.
At this writing, pH control represents the most effective control method over
phosphorus concentration in a lime-treatment system. Although the relationship
between the phosphorus content and pH is not exact, as the scatter in Figure 46
indicates, the pH nonetheless governs its removal. Secondly, it is the only available
measurement which indicates that a sufficient quantity of lime has in fact been added
138
-------
>
H
n n
AIR tX— G
/-.,-, ,-, * ^.^.
i
h /• to N3/z= F, TO CHEMICAL
" * 7/ CONTROLS
; \ hV2
// J r-
tr ^i /
s
BUBBLER
HWEIR
"fr"""^
0 X
Figure 50. Linearization of head signal from a flume or rectangular weir
may be performed with a square-root extractor and multiplier.
to affect precipitation of Ca (PO ) . Thirdly, pH control can adjust the dosage to
O TT Lt
compensate for variations in wastewater alkalinity and phosphorus content.
Although pH measurement and control have not found favor in treating water, there
were very good reasons for past failures. Failure modes for pH-measuring systems
are cited and recommended practices to improve reliability are listed in References
15 and 28. If these recommendations are followed, the pH measurement can be as
reliable as any other in the plant.
A particular problem in phosphate removal and recarbonation is the fouling of
electrodes due to precipitation. A coating of sludge can build up on the electrode
surfaces, which can destabilize control in a few hours. Fortunately, ultrasonic clean-
ing is quite effective on these deposits.
Figure 51 is a record of the pH and lime-valve position in a laboratory-scale phosphate
removal process. Approximately 12 hours after startup, an oscillation began to
develop, and expanded until the valve was closing fully each cycle. The reason for
the increasing cycle was that the coating insulated the electrodes from the process,
creating an increasing delay in response. When an ultrasonic cleaner mounted on the
139
-------
:
LIME-PHOSPHATE NEUTRALtZATION
\
\\
V
*
11
fr I
ULTRASONIC
CLEANER ON
.
VALVE
POSITION
TIME
Figure 51. Instability in this record was caused by an accumulation of a coating on electrode surfaces,
which was removed by an ultrasonic cleaner.
-------
electrode assembly was energized, the electrode response improved almost imme-
diately and stable control was restored. (Note that time proceeds from right to left
in the record.)
In actual practice, the cleaner is continuously energized, so that deposits cannot
accumulate. The laboratory test simply demonstrated that the cleaner could also
remove accumulated deposits, thereby indicating its effectiveness.
Figure 52 illustrates a pH control system wherein the pH controller sets the lime
dosage rather than setting the lime flow directly. This system can be operated in the
constant-dosage mode by placing the pH controller in manual. In fact, an indicator
connected to the output of the pH controller can be calibrated in any convenient dosage
units such as mg/1 or Ib/million gallons. Then an electrode failure will not result
in complete loss of control but simply fall back to the next best system; i. e., con-
stant dosage. (Pneumatic controls are shown, although electronic may be used as
well. Conversion from electronic to pneumatic is usually required somewhere in
the loop, in that the pH and magnetic flowmeters are electronic while pneumatic
valves are customarily used.)
LIME
FLOW
WASTEWATER
FLOW
WASTEWATER ^
\
***~l**~<^^+**l+mr
•w
0
f p
o
HJ
TO
FLOCCULATOR
RAPID MIX
Figure 52. pH control can be combined with dosage control for effective
backup.
141
-------
The titration curve of lime against pure phosphoric acid in Figure 53 shows the range
between pH 9 and 10 to be virtually linear. Therefore, a conventional proportional-
plus-reset controller may be used in contrast to the nonlinear controller recommended
by the manufacturer (28) for neutralization of acids and bases. In two-stage plants
(first stage, pH elevated to 11 with lime; second stage, pH reduced with CO ) where
£
the pH must be elevated to 11, the slope of the curve is lower, facilitating control.
However, precise control is more important at pH 11, or considerable overdosage
could result.
The titration curve for wastewater would differ markedly from Figure 53 below pH 8,
owing to alkalinity absent in the phosphoric acid sample. The same alkalinity would
also affect the Ca/P mol ratio in that bicarbonate ions will consume lime as the pH
is raised. Considering that wastewaters typically have a higher alkalinity than
phosphorus content, this curve is given only as a guide. The alkalinity moderates
the curve, tending to make it become more linear.
pH 8
C«3 (PO,I2
PRECIPITATES
CaHPO,
PRECIPITATES
0.5 I 1.5
Ca/P MOLE RATIO
_Q
Figure 53. Titration curve of 10 MH PO with Ca(OH)
o 4
142
-------
The manufacturer (28) indicates that the residence time between the point of lime addi-
tion and the pH measurement should be 15 minutes or more for complete reaction.
This limitation applies to neutralization, but for adjustment to pH levels above 9 with
lime only two minutes is required (15). Where less than two minutes is available in a
rapid mix, pH electrodes should be moved part way into the flocculator to provide the
necessary reaction time.
The pH required to produce a given phosphate in the effluent is a function of the
calcium-ion level. In systems treating wastewaters of variable hardness and alkalinity,
monitoring the effluent calcium-ion level may allow occasional adjustment to the pH
o O_i_
setpoint. Since log f PO increases with pH and log I Ca 1 increases with pCa, an
increase of 0.1 pCa should allow a reduction of 0.1 pH while yielding the same total
phosphorus. Frequent or automatic readjustment of the pH in this manner could
result in lime savings proportional to the variations in calcium content of the waste-
water.
Other methods of phosphorus control have also been considered, but, at this writing,
an ion-selective electrode for orthophosphate has not been commercially marketed,
although there is a demand for such an electrode.
Automatic spectrophotometric chemical analyzers are available for measuring ortho-
phosphate. They may be operated batchwise or continuously, yielding results with a
delay of only 5 minutes or so following introduction of a sample. Control could be
applied based on this analysis, with the PO controller adjusting the setpoint of the
pH controller. However, phosphorus compounds other than PO will escape analysis.
Although PO control would be more precise than pH control or pH + pCa control, it
still will not provide total P control. The last requires a more difficult and time-
consuming analysis, which does not lend itself as readily to control. Experience in
treating a given waste may indicate that total phosphorus and orthophosphate are
identical or at least correlatable, in which case the simpler orthophosphate analysis
is adequate.
143
-------
An orthophosphate analysis of the wastewater prior to lime addition could be used to
set the lime dosage in a feedforward manner. But the lime dosage is primarily
determined by wastewater alkalinity and hardness, not orthophosphate concentration.
Therefore, unless the prior history of the alkalinity and hardness of the water is taken
into consideration, dosage control based on orthophosphate concentration times waste-
water flow would be only a slight improvement over dosage control based on flow
alone. Both are inferior to pH control.
Recarbonation of treated waters with a pH greater than 10 can reduce alkalinity by
precipitation of CaCO . However, recarbonation must be controlled precisely to
O
achieve minimum alkalinity. The pH corresponding to minimum calcium solubility is
approximately 10. It has been reported (25) that a pH of 10.3 gave the most CaCO
3
precipitation at the Lake Tahoe tertiary treatment plant. Any individual plant may
have its own optimum pH, based on the wastewater characteristics.
Insufficient recarbonation will not remove all possible CaCO , while excess recarbona-
o
tion will redissolve some calcium by forming more HCO ions. Consequently, what-
o
ever the optimum pH, it must be controlled precisely because it will yield the fastest
settling floe and the least alkaline effluent. Calculations of calcium solubility give
-4 -4 -4
concentrations of 2 x 10 M at pH 10.5, 1.5 x 10 M at pH 10, and 1.8 x 10 M at
pH 9.5, indicating the significance of precise control.
A conventional proportional-plus-reset controller may be used to control the recarbona
tion pH by manipulating the CO source, as shown in Figure 54. Both ultrasonic
£
cleaning and automatic temperature compensation of the pH electrodes should be pro-
vided for the precipitation at pH 10. However, neither of these features is necessary
for recarbonation to a final effluent pH below 8.5.
144
-------
PHOSPHATE
CLARIFIER
I
pH
ROM
-— «-"-
dl
1
LI— I
,/s
TOCLARIFIEF
OR FILTER
CO,
CARBONATOR
Figure 54. pH control applied to recarbonation.
145
-------
SECTION VII
CONTROL OF WET-WEATHER TREATMENT PROCESSES
AND COLLECTION SYSTEMS
Combined sewers make up approximately half of the sewer systems in the United
States. Some 1300 U. S. towns and cities are serviced by sewer systems that overflow
during rainstorms, when the combined volume of sanitary waste, infiltration, and
storm water exceeds the capacity of permanent facilities.
To accommodate these excess flows and to prevent the flooding and polluting of sewered
areas, most combined sewers direct the excess flows through overflow structures into
neighboring watercourses. The overflow structures route the small dry-weather
flows through intercepting sewers to treatment facilities. (Automatic mechanical
regulating gates have been used widely to make improved separations between dry-
weather and storm flows, with varying degrees of success (29).)
Although they may only make up a small fraction of the total annual volume of sewage
handled, overflows are a major source of water pollution. Whether from sewers or
treatment plants, overflows (bypasses) usually account for a disproportionately greater
quantity of the biologically active and toxic materials being delivered to nearby streams
(30). Since the often-thought-of direct solution to combined sewer overflow pollution,
that is, separation, would be extremely costly and require several decades to achieve,
sanitary engineers have proposed alternative measures that would be effective sooner
at only 40% to 70% of the separation cost (31). These methods (based on various com-
binations of storage and treatment processes, followed by disinfection) are discussed
at length in this section. Another advantage of the alternative measures over sewer
separation is that all flows into the receiving body (or bodies) of water will then be
treated, whereas the effluent from a separate storm sewer usually is not treated (32).
146
-------
Tests have shown that some form of primary treatment is usually sufficient to reduce
organic and solids pollution levels to allowable limits. Since it rains only 4 to 15% of
the time, storm-water facilities are designed with low capital and high operating
costs (just the reverse of the criteria used for most treatment plants).
The topography and physical layout and size of the collecting watershed, the sewer
system, and the availability of land at the proposed site all enter into the selection of
storage and treatment methods; e.g., off-line storage, routing and storage in the
sewer, handling in the satellite treatment plants, or centralized treatment. Because
all treatment plants provide some degree of storage, the relative storage size and
treatment plant rate desired depend largely on the nature of the expected overflow.
The influent to an overflow treatment plant results from a transient storm-water flood
that is far in excess of normal variations in dry-weather sewage flow. Overflows in
both combined and separated sewers are characterized by a large and rapid rise in
sewage flow and strength (33). It is this sudden mass of highly contaminated water
that must be prevented from polluting local watercourses.
As a specific treatment approach, microstraining has been shown to be effective in
reducing solids and BOD from overflows (34, 35). Other methods for handling over-
flows include in-line storage, retention basins, and centralized overflow treatment
and are discussed in References 29, 34, and 36 through 39.
The concept of automating storm and combined sewer systems will be explained below
by considering three hypothetical cases. Each case will consider a different size
catchment area. The treatment systems discussed in each case are hypothetical
solutions to the problem and other treatment systems may also be applicable to handle
the storm waters within a catchment area. The three catchment areas selected for
this study are 1) small (150 acres), 2) medium (3000 acres) and large (100,000 acres).
147
-------
COMBINED SEWER OVERFLOW TREATMENT PLANT SMALL (HYPOTHETICAL 150
ACRE) CATCHMENT AREA
In the small catchment area micro-straining offers a practical method to achieve the
physical treatment of combined sewer overflows. The hypothetical treatment facility
that makes use of such a method is generally relatively small and self-contained, and
usually is located at an overflow point that is relatively remote from a central treat-
ment facility.
In this section the combination of storage and microstraining as a suitable alternate
method for treating combined sewer overflows will be discussed. The basic instru-
mentation for such plants is reasonable and well established, and will operate effec-
tively if properly installed and maintained.
The process objectives are as follows:
• To eliminate pollution (caused by combined sewer overflows) by collection and
sufficient treatment to allow a discharge into the receiving bodies without
causing their degradation
• To handle the specified design overflow with minimum capital and operating
costs
• To provide a safe, practical, and efficient facility
• To consider a treatment plant that utilizes microstrainers and is suitable for
a relatively small catchment area.
Background
The overflow event is marked by a relatively sudden variation in flow rate and in the
concentration of pollutants. From a hydraulic viewpoint alone, overflows can be
several hundred times the magnitude of the normal dry-weather flowrate (40).
Pollutants classifiable as suspended solids and BOD or COD are usually strongest at
the start of the overflows and may contain difficult contaminants such as grease or
salinity, but this depends entirely on local conditions. The overflow itself is initiated
by either rain or thaw; consequently, typical rain and runoff patterns must be
examined as part of the overall problem.
148
-------
The quantity of precipitation and subsequent events in the sewer system must be
measured to a finer time scale than is usually customary in computing hydrographs
for hydraulic engineering studies (41). These studies have to be reasonably accurate
for an entire seasonal cycle as a minimum. In any case, a statistically valid picture
of the expected overflow must be produced if the facility is to be designed to cope
properly with the event.
Because the overflow reaching the treatment facility is always a highly dynamic
occurrence an initial storage volume is desirable. Initial storage also provides the
means for grit and other solids sedimentation that can reduce the burden on subse-
quent treatment facilities. The solids accumulation can be removed by the most con-
venient and economic method after the overflow has subsided. If treatment by a
downstream continuously operating plant is preferred to treatment by the overflow
facility, the entire content in storage can be returned to the downstream plant as soon
as the overflow has ceased.
The usefulness of a microstrainer in overflow treatment for small catchment areas
has been described in U. S. EPA publications (34, 35) and is summarized below. The
microstrainer operated at flow rates of 35 to 45 gallons/minute/square foot with
differentials of 24 inches of water. At these rates, the suspended solids in the com-
bined sewer overflow were reduced from an influent range of 50 to 700 mg/1 to an
effluent range of 50 to 40 mg/1 and below. At the higher influent levels of suspended
solids, the removal performance was enhanced, yielding effluent concentrations of
approximately 10 mg/1. The conventionally used percentage removal performance
criteria are not valid for microstraining the combined sewer overflow. The volatile
suspended solids reduction paralleled the reduction of total suspended solids.
The highest concentration of suspended solids frequently occurs when the overflow
rate is highest. The concurrence of a high suspended solids concentration and a high
overflow rate results in a very high potential of contaminant loading per unit time into
the reciving stream. The microstrainer was unusual in that it removed a much
149
-------
greater percentage of the suspended solids when their concentration was higher. A
reasonably sized microstrainer, then, can limit the pounds of suspended solids per
unit time entering the stream from a combined sewer outfall.
The organic matter in the combined sewer overflow was highly variable—ranging
from 10 to 2000 mg/1 as BOD,., and 20 to 400
5
reduced organic matter by some 25% to 40%.
from 10 to 2000 mg/1 as BODr, and 20 to 4000 mg/1 as COD. The microstrainer
5
Satellite Overflow Treatment Plant
This hypothetical proposed facility is considered to be semimanned in that it is suffi-
ciently automatic to detect and handle any overflow but requires routine and conscien-
tious maintenance, as well as manual cleanup and replenishment after an overflow.
As outlined in Figure 55, the facility consists of the following:
• An inlet structure
• The storage and surge reduction volume (if found to be necessary)
• Treatment facilities (microscreens)
• A disinfection system with contact provisions
• A disposal system and miscellaneous pumps, equipment, and controls
• The station itself (consisting of buildings, etc.).
CONTACT BASIN
BAR SCREENS
STORAGE VOLUME
FLOAT
Figure 55. Satellite treatment plant.
150
-------
The inlet structure, which accepts the overflow, contains gates for station isolation
and will also include pumps if necessary. Coarse racks and other devices to handle
large objects and protect the downstream machinery will be located either here or in
the intermediate structure. A sensing device to detect the overflow start may also be
located in the inlet structure.
The storage and surge reduction volume accepts the peak flow rate of the overflow as
effluent is withdrawn through the microstrainers at a maximum rate. A storage
volume may not be necessary in all cases, but if storage is provided some sedimenta-
tion will occur. The storage volume is designed to be kept nearly empty between over-
flows and has a residual volume that will retain sludge and isolate it from the micro-
strainers.
The residual volume, tank washings, etc., are usually pumped back to the interceptor
or main treatment plant as sludge. (Sediment might be removed by earth-moving
machinery at less cost, in some cases.) The storage tank also contains an emergency
overflow to prevent flooding the station. Scum, grease, and other clogging materials
require special handling to prevent fouling the microscreens, but the cure would have
to vary from situation to situation.
The intermediate structure separates the storage volume from the microscreens.
This structure contains gates to moderate the flow from the storage volume to the
microstrainers and also allows isolation of one or more microstrainers. It may con-
tain bar screens, pumps, or chemical dosage systems, including flowmetering.
The disinfection system is required to treat the sanitary component of the overflow for
health considerations. The disinfectant addition rate must be maintained proportional
to the effluent rate, and the treatment ratio is readjusted by a suitable residual
analyzer system. The contact tank is sized to ensure sufficient residence time for
disinfection under maximum flow conditions, and also provides a source for reasonably
clean water for screen washing and station cleanup. Standby water is required if the
151
-------
contact basin runs dry or becomes contaminated by solids, due to an overflow or
screen failure. If the standby water source consists of city water, a backflow pre-
ventive device is also required.
Sludge disposal pumps and auxiliaries are required to wash down a facility during and
after an overflow and to assist in manual operation. Controls will be required, too,
for the throughput rate {via level control), microstrainer, operation, disinfection,
and alarms. Recording and communications systems are also necessary.
The microscreen section, containing two or more machines, has its own control system,
which varies the drum speed and backwash flow in response to the head loss. A
source of spray water and a drain for drum washings are both required. A high differ-
ential alarm detects a hydraulic overload condition.
Microstrainers have a limited maximum flow capacity (typically between 10 and 40
gallons/minute/square foot of the submerged strainer area) when operated at high
pressure differentials (42). So long as that differential is not exceeded, they are
self-regulating. The water level on the outer or downstream side of a microstrainer
is maintained at about one-fourth the diameter of the drum in order to maximize the
differential and to ensure that the material on the screen does not dry out. The up-
stream level is kept nearly up to the backwash overflow level to maximize the effec-
tive filter area and head across the screens (42).
When the input level to the microstrainers is kept at a fixed value, some means must
be provided to control the flow from the storage volume as the storage level rises to
provide more capacity for the overflow. Where head is available, automatically posi-
tioned gates are suitable; otherwise, pumps could be used in the usual fashion for
level control. The microstrainer controls that are usually furnished (to control the
drum speed and backwash rate in response to the differential) are less important when
the input level is controlled, but they do provide the means for setting lower drum
speeds manually if desired for improved solids retention or better backwashing.
152
-------
It may be desirable to improve the solids removal capability through pretreatment by
adding a coagulant but such a procedure, though feasible, would depend on field trials.
In any case, a suitable and continuing sampling and testing program is required to
control performance, both in the design phases and after the plant is built and running.
Disinfection systems based on applying disinfectant in ratio with effluent flow, with a
residual analyzing trim, are well established in practice and need no discussion here.
The rangeability of such a system is limited, however, and the effluent flow rate
variation should not be greater than about 10:1, using a single disinfectant feed system.
(This is based on a chlorinator rangeability of 20:1 and a dosage rate variation of 2:1.)
From an operational standpoint, a station should be sized so that the treatment portion
runs continuously and all overflows are contained in an intermediate storage. This
would use the treatment system most economically. Sizing the station by using
extensive storage to handle all storms is usually impractical; the station should be
designed to treat overflows as efficiently as possible within a maximum practical
storage capacity. The treatment efficiency will be maximum if, as soon as the stor-
age volume has accumulated to a certain minimum, the treatment system is made to
run at its maximum rate and continues to run at that rate until the overflow ceases or
the volume in storage is at or near its minimum.
Facility Sizing
The 150-acre drainage area is an indefinite measure for overflow treatment plant
sizing. The overflow facility cannot be sized for fixed flow rates but, in any area, will
depend upon the:
• Typical local rainfall—its duration, intensity, and frequency data (more
detailed in space and time (43)
• Dynamic analysis of the catchment area
• Dynamic analysis of the sewer system between the catchment area and the pro-
posed plant
• Dynamic interaction between the proposed facility and the existing system.
153
-------
A computer-type program is the logical method for performing sizing calculations,
but only after data has been collected in the area.
A design runoff rate of 1 cubic foot/sec-acre (44), produced by a 50% collection (runoff
coefficient =0.5) from a 2 inch/hour continuous rainfall, is equivalent to a continuous
runoff rate of approximately 100 mgd from a 150-acre catchment area. The following
discussion is patterned on an estimate of a suitably sized plant for a noncontinuous
situation.
A 2 inchAour storm of 1-hour duration produces 54,300 gallons per acre. If the
runoff coefficient is 0.5 from the 150-acre watershed, then the total runoff is about
4 mg. For the hypothetical example, assume an overflow where, due to the rainfall
distribution (in space and time) and the sewer characteristics, about 2.8 mg reach
the satellite treatment plant, with a peak influent rate of approximately 46 mgd
(Figure 56).
Further assume a 2 mg storage volume, followed by a 25 mgd capacity treatment sys-
tem. As noted in Reference (36), there is a tradeoff between storage volume and
treatment rate or in other words the plant size and the storage volume are interrelated.
In the reference cited, the ratio of volume to rate ranges from 0 to 0.286 day; that is,
the plant will be operating on an average basis that amount of time. In this example,
the time is chosen as 0.08 day.
Treatment of the accumulated overflow is arbitrarily assumed to start 30 minutes
after the overflow begins filling the storage volume, when the stored level has reached
10% of capacity. Without level control (which allows the water to flow to the micro-
strainers through a fixed restriction at a rate proportional to the level), 65% to 70%
of the storage capacity is used and treatment lasts about 6 hours. With level control
(which allows the water to flow to treatment at the maximum rate of 25 mgd), less
than 40% of storage is used and treatment lasts only 2.6 hours. For these reasons,
the volume and treatment rate selected are a reasonable choice for an overflow of
such dimensions.
154
-------
OVERFLOW RATE
• (PLANT INFLUENT)
100
50
OVERFLOW
STORAGE ONLY
FULLI2MG)
STORAGE FED
BACK TO SEWER
HOURS
6 HOURS
Figure 56. Satellite plant performance.
The use of level control in this hypothetical case greatly increases the potential
capacity of the station. Precise capacity is hard to estimate without more factual
information about the overflow since an actual situation is almost certain to be non-
linear in several ways, but a suitable combination of representative field data and a
dynamic analysis should be sufficient. (An analog computer or its digital equivalent
should be considered for this type of calculation.)
Control Objective
Overflows of excessively high rate and duration can exceed the storage volume and the
throughput rate of the treatment facility. In these cases the excess flows should be
allowed to overflow in the least objectionable manner before the facility itself is
flooded.
155
-------
Disinfection is easily controlled through standard methods. The feasibility of closed-
looped control of chlorine residual has previously been demonstrated for continuous
plants and, with some modification, is applicable to intermittent automatic service.
Because of the remote nature and multiplicity of installations implicit in the use of
satellite treatment plants, it is most desirable from an economic viewpoint to auto-
mate these plants so that they can operate largely unattended, with no loss in either
reliability or performance. The principal objectives, then, are that:
• Instruments and instrument systems shall be chosen as components of the
instrument arrangements proven to be workable in wastewater systems.
• Control shall be either pneumatic or electronic, as dictated by system capa-
bility. The high reliability and low cost of pneumatic systems must not be
ignored.
• Instrument systems selected must be capable of a full and accurate perform-
ance upon energization after long shutdown periods.
• Equipment shall be selected for reliability, minimum maintenance, and per-
formance—in that order. Accuracy is considered less important than repeat-
ability and reliability and need be no more than nece^ary for system function-
ing as a whole.
From the viewpoint of facility functioning, it is desirable to monitor the throughput
rate and the concentrations of the constituents of interest as a function of time. For
the purpose of operating the station (either automatically or manually), the following
measurements are desirable:
• Influent channel level to initiate the station operation
• Storage level to warn of a possible overflow and to pace the throughput rate
(if desired); also an alternate for station startup
• Individual strainer differential to control the drum speed and backwash rate
and to warn of a possible overflow
• Influent or effluent flow rate to compute the strainer performance, to control
the disinfection rate, and to maintain a historical record
• Disinfectant process concentration measurements in the effluent to ensure com-
pliance with standards
156
-------
• Organic strength to ensure compliance with standards
• Solids concentration to ensure compliance with standards and to indicate the
strainer performance
• Salinity concentration or a conductivity measurement
• Facility internal status checks, illegal entry by unauthorized persons, fire and
flood alarms.
For control, the mechanical and hydraulic measurements, chlorine residual, and
salinity are the only ones practical at this time (45). Automatic turbidimeters or
nephelometers for solids determination are in a marginal performance status,
although they have had some success in estimating solids content. Automatic sampling
is available for accumulating samples for manual laboratory analyses (especially when
such automatic sampling is based on sound principles), but automatic instruments for
suspended solids or organic content (especially those suited for unattended locations)
are not yet commercially available.
Automatic controls are necessary for the rotary strainers, the disinfection system,
the level control to the strainers, and station management if the satellite station
(whether manned or not) is to be able to handle overflows as they occur.
Microscreens are not installed singly, since doing this might put the facility out of
service when a screen was down for maintenance. It may prove desirable to limit by
automation the machines in service so that only a minimum number will operate under
minimum flow.
The disinfection system may not be suited for operation at very low flowrates. Con-
sequently, some means for maintaining the facility effluent rate above a certain
minimum is desirable; it is for this reason that level control between the initial stor-
age and the microscreens is similarly desirable. For the same overflow event, level
control (or pumps) between the storage and screens could allow some 20% reduction
in the storage volume. The exact benefit would depend on the process parameters,
especially the nature of the overflow incident.
157
-------
A signal for treatment initiation is advisable to start facility operations. Normally
the facility operates either in a standby or an active mode. Manual maintenance and
post-event cleanup are assumed. The signal for treatment initiation could be sent to
a remote site for automatic notification.
Plant Operation
In a normal operation between overflows, the facility remains in a standby condition.
The inlet gate is open, the storage tank is empty, all motors are shut off, and the
screening and disinfection systems are in condition to start. Wherever necessary (as
in an instrument that depends on electrochemical probes), a small flow of clean water
may be maintained through the instrument to keep it in readiness.
The facility itself is secured against intruders, fire, freezing, etc., with all neces-
sary alarms and detectors and all instruments in a standby condition. Routine main-
tenance is performed as necessary, and possibly various systems are exercised after
extended dry periods.
At the start of an overflow, the level in the storage tanks starts to rise, the recording
instruments start, and an alarm is Initiated at a remote control center. After the
level has risen sufficiently to start flowing to the screens, a screen or screens are
started at minimum speed, and backwashing at the top of the screen begins at a
minimum flow. As the flow continues through the screens, the effluent flow is sensed
and the disinfectant sampling systems start. As the overflow continues, the level at
the input to the screens rises, the flow to the screens increases, the head loss across
the screens increases, the drums turn faster, and the backwash flows increase. The
increased effluent flow increases the disinfection rate, although it is modified by the
chlorine residual control system. The sampling system collects effluent, either at a
fixed rate or in ratio to the main effluent rate. As the level in the storage volume
continues to rise, the level at the input to the screens reaches its operating maximum
158
-------
and is prevented from rising further by the level control system and the automatic
control gates. The screens then operate at constant influx until the overflow begins
to end. In this way the storage volume accumulates the overflow up to its maximum
capacity, without causing an overflow at the input to the screens.
At the end of the overflow as the level in the storage begins to drop again, the screens
and disinfectant systems slow down after the level has dropped below the input level to
the microscreens. Eventually the screens stop and the station returns to a standby
position until it can be cleaned up, replenished, and made ready for the next overflow
event.
The facility will operate reliably in this way only if maintenance and servicing are
correctly carried out to the degree determined by the original design, as well as by
experience.
Instrumentation
Instrumentation for a satellite facility must be simple to operate and maintain, highly
reliable (under existing conditions of service and maintenance available), failsafe,
quickly repairable or replaceable, and inexpensive. All criteria can be met by using
the conventional electronic or pneumatic instrumentation common to the process
indus tries.
Instrument interconnections must be of high quality and installed in accordance with
good practice. Wherever explosive hazards are absent, NEMA 4 housings, sealed
conduits, and noncorrodible materials should be used for electrical systems.
Wherever conditions indicate that electric power is potentially hazardous, pneumatic
or intrinsically safe systems are strongly recommended. Pneumatic systems must
be designed to resist both corrosion and leakage.
The instrument panel must be properly designed for use and maintenance, with all
connections well marked and secure but easily accessible. Power supplies for
electronic systems must have a conservative safety factor and be self-protective.
159
-------
Instrument air systems must produce clean and dry air. In any case, the system
should have a guarantee of at least 18 months between failures, assuming that a
specified and controlled maintenance program is set up and followed.
Due to long periods of inaction, the recorder will be shut off when the station is idle.
The chart could be automatically marked with a date/time stamp when the overflow
begins, or an intermittent strip chart could be used. The recorder has a uniform
percentage marking for all three inputs, and the time scale is measured from the
starting time for that occurrence.
The level measuring systems for the storage volume and micros trainer input (LT-2
and LIC-3) are analog devices and should be of the bubbler type, with flow control
relays, manual purge valves, and some means for automatically shutting them off when
the facility is idle. The inlet level detector (LS-1) is properly an inverted bell or
diaphragm box, capable of operation without an air supply. Conductive or capacitive
probes must be carefully applied since intermittent usage, dirt, and moisture have
caused many such installations to fail in the past. Redun" Jicy is advisable here.
The level control system (loop 3 of Figure 57) senses the level at the common inlet
basin for the microstrainers and raises or lowers the motorized sluice gate as
necessary to maintain that level. A position indicator (Zl-3) shows the gate position
on the instrument panel from which an operator can control it. Also shown in Figure
57 is an interlock (LS-2B) that would keep the sluice gate closed whenever the level in
the storage volume was too low, and could also start the microstrainers. LIC-3 is a
narrowband proportional controller. The proportional band, sluice gate size and
speed, and controlled basin volume should all be sized by a control engineer.
The controlled gate driven by LIC-3 must have a feedback slidewire suitable for wet
conditions; in addition, it must be designed for frequent actuation and for long periods
of idleness. It is assumed that the design control engineer and the commissioning
instrument engineer have ensured that the loop controlling the gate (or whatever
device is used) is not too sensitive and will not immediately wear itself out.
160
-------
SAMPLE
PUMP
Figure 57. Satellite treatment plant instrumentation.
The microstrainer controls (dPIC-3) are sized ana lurnished by the machine supplier.
The controller is either proportional or proportional-reset with anti-windup. The
backwash valve must be capable of handling dirty water and producing a backwash flow
that varies linearly with the controller output.
Residual chlorine analyzers are reliable when correctly installed and maintained, but
they must be handled properly. The sample stream must be representative, fresh,
and reliable. The relatively large sample pump (typically 1 to 2 hp) is a good source,
but the sample must still be reliable after passing through an adequate strainer. (The
strainers customarily supplied with these analyzers are designed for clean water
service and consequently are apt to prove inadequate for this service.) The sampling
point must be properly chosen (46) and some provision must be made to prevent the
sensing cell or probes from degrading while the facility is idle.
161
-------
The disinfectant feed system is a conventional chlorinator system or its equivalent.
The disinfectant rate is kept in ratio to the flow (FT-5), but that ratio is adjusted by
the residual analysis (AIC-4). The feed device has a minimum rate at which it adds
disinfectant to the contact basin, so that the feed device should be set to run only when
a certain minimum effluent flow is present.
Disinfectant addition control is available as a package that includes a scale (WI-6) and
flowmeter (FI-6) in the case of chlorine systems and needs no further discussion here.
The typical chlorinator can be modified either at the factory or in the field to accept
an analyzer trim signal, as well as the flow signal.
For flow measurement, a Parshall flume with a stilling well is the normal choice,
since it has a rangeability in excess of 20:1 and is simple, rugged, and inexpensive.
Instrument Costs
This hypothetical facility provides a 2 mg storage and 25 mgd treatment rate. Note
that this size facility, starting empty and beginning the treatment when 0.2 mg has
been accumulated, accommodates a constant flow of approximately 50 mgd before
overflowing at the end of 1 hour and is therefore comparable to a 50 mgd plant (where
such level control is not used).
The associated capital cost for installed instrumentation is approximately $40,000
(excluding the connections to remote facilities, microscreens, gates, and chlorinators)
and will vary only slightly with the facility's size until growth is accomplished by
installing duplicated treatment units. (The instrumentation costs related to micro-
strainers are included in the cost of the microstralners.) The annual maintenance
cost for satellite systems is not linearly related to size, and typical maintenance
costs are about $30,000 per year for a district comprising up to six satellite facilities.
Wherever more than four satellite facilities exist in one locality, communication and
computerization should enter the picture, and highly specialized technicians (whether
on the payroll or contracted) may be required.
162
-------
OFFLINE STORAGE AND CHLORINATION TREATMENT PLANT (3000 ACRE
DRAINAGE AREA)
This hypothetical facility (outlined in Figure 58) consists of the following elements:
• Inlet structure
• Pumping station
• Storage and short-time sedimentation units
• Disinfection system with contact provisions
• Pumpback and sludge cleaning facilities
• Miscellaneous pumps, controls, and equipment
• Buildings necessary to house the facility.
In this concept excess stormflow is captured in a holding basin where gross solids
settle out. The liquid is chlorinated before discharge to effect microorganism control.
AUXILIARY
WATER SUPPLY
TO TRAY
CHLORINE GAS
MIXES WITH WATER
VACUUM RELIEF
VENT TO
ATMOSPHERE
CHLORINE
CONTAINER
r**pH^^»^BM"™'
INFLUENT t II /
-G-I / -k
FROM
REGULATOR
MAIN WATER
SUPPLY TO INJECTOR
DETENTION/SEDIMENTATION CHAMBER
fa-
EFFLUENT TO RECEIVING BODY
TANK DRAIN TO INTERCEPTOR
Figure 58. Offline storage and disinfection plant.
163
-------
This station is considered to be semimanned; during nonoperating periods the day
shift is covered by a custodian/maintenance man. The initiation of the station's
operation is automatic, with necessary personnel be ing summoned in order to be present
throughout any storm event. The actual control of the station, too, is automatic, with
personnel watching for equipment failures and performing station cleanup upon cessa-
tion of the storm event.
This type of treatment facility is structured for a medium-sized drainage area that
has inherent regulation due to in-system storage. The satellite facility must be able
to respond to rapid changes in flowrate due to the small area served and a lack of
inherent system storage. Depending on the nature and size of the area served, it is
possible for a very short storm to be over before the installation is activated. It is
also quite possible for a small storm not to overflow the plant if the storm's volume
is less than the plant's storage capacity.
The inlet structure accepts the overflow from the combined system, probably through
a regulator station. The structure contains gates, scr 3ns, and a pumping station.
The pumping station delivers the storm-water overflow into a series of short-time
detention/sedimentation tanks. As the combined sewage is pumped from me inlet
structure to the treatment tank chambers, chlorine or other disinfecting agents are
added. After short-term sedimentation and disinfection, the plant effluent is dis-
charged into the adjacent watercourse.
The prevailing process philosophy is that storm-water overflows often carry a greater
pollutional load than the community's raw sewage. Potential pollutions! loading is of
such magnitude that short-time detention for the removal of gross solids, followed by
disinfection, provides a dramatic reduction in pollutional loading on adjacent water-
courses (37, 47).
Upon cessation of the storm event, any overflow remaining in the sedimentation cham-
bers, along with sludge, screenings, etc., is pumped back into the interceptor for
ultimate delivery into the central treatment facility with the exception of the few
164
-------
facilities that have on-site sludge handling capability. The sedimentation chambers
are not normally equipped with sludge scrapers. The station cleanup and pumpback
at the end of a storm event are usually performed by the station personnel who were
present during the storm. Some plants such as Spring Creek and Jamaica Bay, N. Y.,
have automatic cleanup capability.
Following a storm event, the facility is pumped dry, cleaned up, and generally
restored to standby status. Because the facility's operation is essentially batch in
nature, continuous scrapers are not necessary in the sedimentation basins. These
basins could be cleaned in any number of ways, including hand labor and hydraulic
flushing.
The disinfection system is designed to maintain a desired residual at the end of the
contact period by applying the disinfecting agent in proportion to the flow and by
adjusting the dosage through use of an automatic residual chlorine analyzer control
system. Chlorine dosages range from over 4 mg/1 at low flows to 2 mg/1 at high
flows; the objective here is to provide a 1 mg/1 residual as the treated effluent leaves
the facility.
Facility Sizing
Sizing the facility usually is based on the premise that the worst conditions will be
encountered whenever the duration of the storm equals the time of the concentration.
Whenever rain falls in a given area, it takes time for the water to flow from the
periphery of the catchment area to the control point. (The control point is defined as
that point which all water from the catchment area must pass.) This total time is
known as the "time of concentration. " If a storm ends before the flow from the
farthest point has reached the control point, the area provides temporary storage and
the maximum runoff rate at the control point is less than if the storm had persisted.
It is generally recognized that, the shorter the duration of a storm, the greater will be
the expected average intensity. Consequently, the most critical storm is one whose
165
-------
duration equals the time of concentration. A storm of longer duration has a lower
intensity, while one of shorter duration will never reach its peak (as far as control
works are concerned) before the storm is over. As a result the plant design is a
function of the topography and a coefficient of runoff as well as of rainfall. A small
hilly area that is heavily built up may have a runoff equal to a much larger area of
moderate topography and building density. The result of such unpredictable variables
is that facility size depends only partially on the drainage area. The design of such
facilities must be structured around the rate of flow rather than the total flow; in
brief, each facility presents a unique design problem. The design of the storage
capacity necessary for a facility of this type depends on the minimum settling time
desired, as well as the maximum design flow.
Control Objectives
The functions of instrumentation contained in this class of facility are as follows:
• To provide control of treatment facilities, chemical feed pumps, and proper
sequencing of the sedimentation chambers
• To monitor the station and equipment status
• To indicate and record important parameters for facility and process evalua-
tions.
In a broad sense, the functions denoted above are true of all combined sewer overflow
facilities. However, in view of the specific instance at hand and the present state of
the art, station records and monitoring functions do provide important aids to evaluat-
ing such facilities.
The problems involved in achieving these objectives are many, principally because
this type of installation operates discontinuously. This in turn causes a number of
difficulties not typical of conventional treatment facilities, such as:
• The equipment stands idle and inoperative between storm overflow events.
• The sampling systems, chemical feed systems, etc., require careful and
continuous maintenance if they are to operate when needed.
166
-------
Since the facility is capable of only partial removal of floating and settleable solids,
over-design events deteriorate the effluent inversely as the peak rate during the peak
time. Because the process is batch in nature and requires fairly complex sequencing
controls as well as chlorination control, process control is necessary.
The manipulation variables are station flow and chlorine flow. Station flow is a
manipulation variable in that the flow can be diverted around the station, or diverted
to parallel tanks, or both.
Plant Operation
Normal operation between storm events or overflows requires that the facility be
maintained in a standby condition. This means that the inlet gates are open, the fine
and coarse screens are not running, all motors are off, and the disinfection system
is shut down. The residual chlorine analyzer is flushed with potable water to keep the
electrode system in an operable condition, and the normal lighting and ventilation
circuits are kept energized.
The facility is secured against intruders, vandals, and fire by means of the necessary
security and alarms. The facility size and complexity necessitates routine daily
custodial service and maintenance procedures. Routine facility exercises are gen-
erally not essential because of the frequency of operation and the presence of operators.
However, routine maintenance is essential for all mechanical equipment, and most
especially for the instrumentation.
At the start of any overflow event, the system overflows through the inlet gates into
the screen chamber, where the rising level initiates the screen drives. The overflow
continues into the wet well of the pumping station, where the rising level starts the
first pump at a low speed. As the level continues to rise, the speed increases over
the control band. A continuing rise in the level starts the second pump at a low speed,
etc. Upon decreasing level, the pumps decrease in speed and finally drop out in
reverse order.
167
-------
Each pump is equipped with a flowmeter that in turn operates a chlorine feeder in
proportion to the flow. In this way, the potential operating difficulties associated with
summing systems and sequential feeder operation are eliminated. A flow summing
system requires a flowmeter and transmitter for each measured flow. The signals
from each flow are fed into a battery of summing relays or summing amplifiers. The
signal representing the total is sent to the readout and feed equipment. However, the
signal representing each flow has an elevated zero (e.g., 3 Ibs, 4 ma dc, and 10 ma
dc), which requires that an appropriate zero signal be supplied to the summing relays
or amplifiers for each measured flow that is zero. Because the most probable point
of error for any instrument is zero, this signal is usually provided by an external
generator; a programmer determines to which summing relay or amplifier the signal
is provided.
In addition to all of the complexities inherent in accomplishing the above, the addition
of summing relays or amplifiers provides one additional point where an error can
develop. For these reasons, the most accurate and reliable system in which multiple
pumps are involved uses a separate flowmeter and feeder for each measurement.
This type of measurement and control system also results in each flow-measuring
device operating over the flow range (usually only 3:1). If pumped flow signals are
added or if the entire flow is passed through a single flowmeter, the required range-
ability is extreme and accuracy is very poor. (For example, if three pumps of equal
capacity are used, the required rangeability of the flow-measuring systems is approxi-
mately 9:1).
Controlling the chlorine feeders can be accomplished by continuous modulation or
impulse duration. Neither system has any particular advantage over the other except
where specific types of feeders are included in the evaluation.
168
-------
Typical Control Instrumentation
Instrumentation for an offline storage and disinfection facility is selected on the
premise that such a facility is often unmanned. Properly selected instruments will
yield records that permit an operator to evaluate whether or not the equipment is per-
forming correctly. Such a concept helps to eliminate the redundancy required by a
completely unmanned facility. The occasional presence of personnel is required
because of the size and complexity of the mechanical equipment, not because of the
instrumentation. Facilities of the size normally required do not lend themselves to
a completely unattended operation. The choice of operating media for instrumentation
is influenced by the following conditions:
• The presence of high humidity and possibly corrosive vapors
• Discontinuous operation
• A lack of adequate maintenance.
Certain instrumentation, notably analytical measurements and particularly residual
chlorine, requires special care to ensure that the equipment is protected from
moisture and corrosion.
As the facility goes on line, the following sequences (as shown in Figure 59) occur:
• The increasing level is detected by LS-1; this in turn activates LA-1 and
initiates the treatment function.
• The level in Hie wet well is measured and transmitted by LT-2 to level
recorder controller LRC-2. The signal from LRC-2 controls the pump speed
so as to hold the wet well levels within acceptable limits.
• The pumped flow from the wet well to the detention/sedimentation chamber is
measured and transmitted by FT-3, recorded, and totalized by FR-3 and
FQ-3, where a contact paces the chemical feeders.
• The chlorine residual is measured and transmitted by A-4 and AT-4 to
recorder controller ARC-4. The controlling signal is fed back to the chemical
feeder in order to trim its feed rate so as to achieve the desired chlorine
residual.
• Additional alarm detectors can be added as required to any of the measuring
or control signals.
169
-------
INFLUENT
FROM
REGULATOR
EFFLUENT TO RECEIVING BODY
TANK DRAIN TO INTERCEPTOR
Figure 59. Offline storage and disinfection plant instrumentation.
Since the pumping station incorporates multiple pumps, a sequencing and proportional
level control system is required. As the level rises, successive pumps start and
gradually increase in speed with the rising level. As a storm event comes to an end,
the pumps decrease in speed and then drop out as the level decreases. The details
of these control systems are a function of the type of pump drive and type of instru-
mentation.
The chlorination system is the other principal variable that can be incorporated into
a control strategy. A number of problems, which are unique to this type of facility,
interfere with complete design flexibility. Some of these are:
• Flow-measuring devices are usually quite difficult to adapt to these facilities.
• Chlorine feed equipment is expected to operate intermittently.
• Because these facilities are usually partially underground, the use of cylinder-
stored liquid chlorine imposes a safety hazard to operating personnel.
170
-------
• Chlorine (in the form of hypochlorite) may be in storage for extended periods
of time with a resultant loss of strength.
• The monitoring and sampling equipment constitutes a serious maintenance
problem because of intermittent operations.
Flow measurement is quite troublesome because design constraints usually result in
long control weirs on either the inlet or outlet. During low flows the weirs are not
properly ventilated, and the level change during high flows is small due to their
length. Pipelines entering the treatment facility do not always flow fully, making
application of either open-conduit or closed-conduit primary devices almost
impossible. Wherever the plant inflow is delivered to the detention tanks by pump,
conventional venturi-type devices or magnetic flowmeters can be used. Wherever
large open channels are encountered, combining Pitot magnetic flowmeters and level
devices provides a means of approximating the flow.
The essence of these problems is that storm-water chlorination on a flow-proportional
basis is not usual or routine in comparison to experiences in sewage treatment plants.
Although flow-proportional control is used commonly in sewage treatment plants, it
may not be a viable choice for storm-water installations.
Residual control of chlorination is now desirable, since it directly controls the final
variable and compensates for changes in chlorine demand. However, in highly
transient flows serious time delays can occur between a residual chlorine measure-
ment and an automatic correction. Another problem is maintaining the sampling sys-
tem and residual measuring equipment in a ready state during the long periods between
storm events, but this can be done by continuously circulating potable water through
these units during shutdown.
The application of compound chlorine control assumes a viable means of measuring
the station influent rate and residual. Unfortunately, the generally unsatisfactory flow
arrangements for station influent seriously limit this technique. Although individual
evaluation is essential in determining the practicability of compound loop control, it
is used successfully in many chlorination and detention storm-water treatment facilities.
171
-------
Table 12, which lists the advantages, benefits, confidence levels, and limitations of
the chlorination control strategies, should guide the reader in selecting the most
feasible chlorination control strategy.
Table 12. WET-WEATHER DETENTION AND CHLORINATION
Control Method
Benefits and
Potential Savings
Advantages
Dis advantages
Chlorination in propor-
tion to flow
Residual chlorine feed-
back control
Compound loop control
(Residual chlorine
feedback with fine
chlorine trim)
Effective chlorination
with minimum Clg
consumption
Effective chlorination
with minimum C12
consumption
Effective chlorination
with minimum C12
consumption
Simple and effective
Effective
Very responsive to
changes in both
flow and demand
Difficult to measure
flow; does not compen-
sate for any changes
in chlorine demand
Questionable value
because of highly
transient flows and
irregular time lags
in chlorine analyzer
Requires a properly
operating analyzer
and chlorinator; very
subject to malfunc-
tioning
Instrumentation Maintenance
The philosophy governing operation and maintenance must consider the fact that this
facility consists of batching or discontinuous processing operations. Most waste
treatment facilities never completely shut down, once they start. But in this facility,
the most critical time for the instrumentation is during startup—the instrumentation
must work.
Accordingly the Instrumentation must receive regular routine maintenance, even
though the facility has not been activated for some time. Such routine maintenance
should adhere to the following (approximate) schedule:
• Weekly—Manually synthesize a storm event and activate the analytical instru-
mentation (however, this is not necessary if the installation is actually operat-
ing on a once-a-week frequency)
172
-------
• Monthly—Inspect and service each device; check the operation of the sampling
systems
• Semiannually—Inspect and service the compressed air systems.
Normally a facility of this size is located in a metropolitan area of such magnitude
that a central municipal wastewater-treatment facility already exists. In this case,
arrangements should be made with the pertinent city department to have a competent
instrument serviceman available, and he should be assigned to service the facility
instruments on a regular basis.
COMBINED SEWER OVERFLOW SYSTEMS (HYPOTHETICAL LARGE DRAINAGE
AREA OF 100, 000 ACRES)
The combined sewer system servicing a large area usually accommodates storm
flows by diverting the excess flow to adjacent watercourses. In the traditional system,
the overflow occurs at regulator stations that are so arranged that, during dry
weather and for flows up to two or three times dry-weather flows, they are diverted
to an interceptor for subsequent delivery to the treatment plant. This arrangement
requires a regulating device that operates so as to divert to adjacent watercourses
all flows in excess of the interceptor capacity.
Regulators are generally designed as self-actuated devices. Historically they have
not functioned very efficiently because of: 1) a lack of operating power inherent in
self-actuated regulators, and 2) inadequate maintenance.
Even if these devices did operate with complete satisfaction, they still would not
dramatically reduce the stream pollution caused by combined sewer overflows. It has
been suggested that use of the storage capacity of the combined sewer system be
made, thereby reducing the frequency and duration of overflows. Recent efforts in
this area have examined and demonstrated the practicability of doing this (38, 39).
However, such a proposal does involve a number of conditions that must be satisfied.
These are as follows:
173
-------
• The regulator devices must operate reliably.
• Some means for completely stopping the overflow must be provided.
• Monitoring a number of variables is required.
• A supervisory program for the operation of regulator stations must be avail-
able or developed from experience, math models, etc.
Satisfying the need for reliable regulator devices, as well as for some means of com-
pletely stopping the overflow, requires the use of externally powered gates, dams,
etc. This requirement in turn can be met by using electric, hydraulic, or pneumatic
actuators. A number of conventional sluice gates can also be adapted to this service.
Variables to Be Monitored
The use of the system storage capacity implies knowledge of hydrologic events
occurring in the total drainage basin. Storms occurring in such remote areas may
impose a heavy hydraulic load but be displaced in time. Knowledge of these events
does permit anticipatory actions such as the dewatering of interceptors and the raising
of overflow barriers. Local measurements, while necessary, do not permit such
system preparation. These conditions require the monitoring of rainfall and sewer
levels at areas remote from the point of regulation.
The regulation facility requires monitoring the following variables:
• The interceptor level
• The combined sewer level
• The equipment status (e. g., opened, closed, etc.).
It is commonly recognized that the flow should be monitored, but the obstacles to in-
line storm water and combined flow measurement are almost insurmountable on both
a physical and economic basis for the following reasons:
• Combined sewage contains large quantities of suspended and floating solids that
make the maintenance of conventional flow-measuring devices a nearly con-
tinuous process.
174
-------
• The application of standard designs of flow-measuring devices and incidental
hydraulic requirements make each installation customized and seriously
infringe on the adaptability of standard devices to existing installations.
• Devices such as magnetic flowmeters and sonic meters are costly.
As a result, the storage capacity and flows must be inferentially determined from
level, and this requires an evaluation of applicable methods. The methods of level
measurement that must be considered include the following:
• Float and cable (and variations)
• Capacitance
• Bubble tube.
The classic float and cable instrument is a poor choice for this service since it accu-
mulates solids on the float, requires a direct mechanical connection to the readout or
transmission device, and also requires extra maintenance when used for this type of
service. The same is generally true of related devices that use a ball and arm.
A type of capacitance device that may be applicable for flow measurements in sewers
has one plate located above the flow and parallel to the surface of the flow. The flow
itself is the second plate through a ground connection since the plate areas are con-
stant and capacitance is a measure of the distance between the plates. There are just
two practical problems: the relationship between the level and the capacitance is
nonlinear, and the span is small.
The classic bubble tube appears to be the most practical level detection device for
this application. It is completely flexible, it is a static-measuring system, and it is
readily adaptable to all transducer types. With a little imagination, this tube can be
adapted to practically any installation (either proposed or existing). It is quite
possible to operate this device from bottled compressed air, nitrogen, CO , or from
£
a small compressor.
175
-------
Transducers and Transmission
The transducing of measurements and equipment status may be accomplished in
several ways. In the case of equipment status (e.g., a valve opening or closing, a
motor running, etc.), the presence or absence of a signal or two different states of a
signal suffice. Wherever a continuously monitored and modulated signal is required,
there are two common methods of signal conditioning:
• Modulating the signal to the impulse duration
• Modulating the signal to the variable-rate frequency shifts.
These signals in turn are usually converted to audio tone signals for transmission by
leased telephone circuit or microwave. The choice, which is a function of a number
of considerations, is largely related to communication link requirements.
Regulator Station Automation
The regulator station requires the following measurements for monitoring and con-
trolling combined sewage overflows:
• The level in the combined sewer upstream of the control device (e.g., the
sluice gate, etc.)
• The level in the interceptor
• The level in the combined sewer downstream of the control device (e. g., the
sluice gate, etc.)
• The status of control elements [e.g., the sluice gate (whether opened or
closed), etc.]
• The station power supply and security status.
In turn, these measurements require the following equipment items:
• Three bubble tubes, with associated level transmitters and a compressed air
supply
• One combined sewer control element (e.g., the sluice gate, fabridam, etc.)
176
-------
• One interceptor control element (e.g., the sluice gate, valve, etc.)
• Signal conditioning devices (as required)
• Communication equipment
• Local control stations.
While many combinations of arrangement and equipment choice are possible, the
actual arrangement and equipment choice are governed largely by existing conditions
and so can be depicted only in a general manner. A schematic diagram representing
a typical arrangement is shown in Figure 60.
177
-------
CATCHMENT AREA
COMBINED SEWERS
CONTROL POINTS
INTERCEPTOR
RECEIVING BODY
LEGEND
LEVEL TRANSMITTER
OPTIONAL LEVEL TRANSMITTER
RAINFALL GAUGE TRANSMITTER
Q CONTROL GATE
Figure 60. Combined sewer overflow control.
178
-------
SECTION VIII
COST/BENEFIT ANALYSIS
INTRODUCTION
Two objectives of this report are to show the economic feasibility of automation as
compared to manual operation and to compare the economics of competing control
strategies where appropiate.- Cost/benefit analysis such as payback period appears
suited to accomplish these objectives. This technique is described by Peters and
Timmerhaus (48) for industrial applications. Payback period has the advantage that
it allows a business to determine the length of time required to recoup its investment
based on the savings and profits from the installation and thereby, with one simple
figure, management can make a reasonable decision as to the effectiveness of the
investment. Since these cost estimates are being used for comparative purposes, it
was not necessary to obtain the best accuracy and therefore the data is limited to
preliminary design estimates.
Since municipal wastewater treatment facilities do not, in general, operate to make a
profit, the payback period as applied to municipalities has been modified to reflect the
time necessary to pay for the installation of all equipment for automation based on the
savings from that automation. To properly explain this concept, the following example
is used. When a cost/benefit analysis comparing an automated versus a manual method
was performed on a municipal wastewater treatment plant, a potential gross savings of
$50,000 resulted because of automation. The automation requires the use of an analyzer,
controller and final control element. The cost and life expectancy of these instruments,
for our purposes, will be assumed as a:
179
-------
• $10, 000 instrument value with 5 years life expectancy
• $50, 000 instrument value with 10 years life expectancy
• $10, 000 instrument value with 15 years life expectancy.
In addition to the capital investment needed, there would be yearly operating and
maintenance expenses. This yearly cost in our example is estimated to be 800 man-
hours or, at $10 a manhour, $8000/year. If one assumes an interest rate of 6% for
municipal loans with 10 years or longer duration and 8% for 5 years or shorter dura-
tion, the list of assumptions is now sufficient to illustrate the required steps in a cost/
benefit analysis via the total annual cost technique.
First, one should convert all the dollar figures into annualized values. The benefits
are already in that form ($50, 000/year) so that only the capital cost figures need to
be converted. This is done as follows (49):
total annual cost = annual capital recovery cost + annual labor cost + annual
replacement part cost*
where:
P = amount of money needed at present time
i = prevailing interest rate, fraction
N = useful life of equipment, years
LC = annual labor cost, $/year
RC = annual cost of replacement parts, $/year
Substituting values in the above equation results in an annual cost of $18, 340.
*The equipment is assumed to be incapable of being salvaged after its useful life.
180
-------
The net yearly savings is the gross yearly savings less the annual cost or $50, 000 -
18,340 = $31,660. This represents the yearly savings of 45% in terms of added
instrumentation cost ($31, 660/70, 000 = 0.45). Another way of expressing this rela-
tionship is by saying that the payback period for this amount of instrumentation is
2.2 years. This data may be summarized as: total capital cost/yearly net savings =
$70, 000/(50, 000 - 18, 340) = 2. 2 years.
Since large amounts of capital are needed for the construction and upgrading of
municipal and industrial wastewater plants, accurate and realistic economic analyses
are essential for an intelligent allocation of resources. Specifically, a return on an
investment and a payback period are the traditional financial criteria used as decision-
making aids in selecting the most desirable investments. Although the particular
criteria depend upon the economic climate (for instance, the interest rate or avail-
ability of funds), it shall be assumed in this report that current payback periods of
approximately two and one-half years or less are needed to justify the installation of
optional instruments and automatic control equipment. In some cases where the
equipment be.com.es integral to operation of the plant (e.g., computers) or where the
equipment is expected to have a long life, management may then find payback periods
in excess of two and one-half years justifiable. When some form of instrumentation
or automatic control is essential and several competitive approaches are available,
payback periods can be used to select the most desirable alternative.
This section details the general methodology of the cost/benefit analysis and shows the
application to the alternate control strategy selection for some of the more important
wastewater-treatment processes. These techniques can be readily adapted to other
treatment processes and instrumentation schemes by the reader. Also, the cost/
benefit analyses may be easily updated to reflect new costs and interest rates. (See
Figures 61, 62 and 63.)
181
-------
/
TOTAL
CAPITAL COST
($106)
100
YEARLY
OPERATING COST
($106)
0
TOTAL ANNUALIZED COSTS
$/MG TREATED $106/YEAR
I >
CAPITAL ASSUMED TO BE OBTAINED AT 6% FOR 20 YEARS
10
WASTEWATER-TREATMENT PLANT CAPACITY (MOD)
IOO
\—I Mill 'O.OI
NOTE: The type of plant described is a multistage biological process providing the following reductions:
OD-95%, SS-100%, P-100%, and N-85%.
Figure 61. Total annualized wastewater-plant costs.
-------
Go
CO
ANNUAL
LABOR COST
(SI 000)
10001 1
TOTAL NUMBER
OF
OPERATORS
I I I I I 11 1000
100
I—I—I I I I 11
I—I—I I I III
AVERAGE TOTAL COST PER EMPLOYEE YEAR: $8000 TO 815,000.
FOR INSTRUMENT MAINTENANCE TECHNICIANS, THIS COST IS HIGHER (-$20,000)
IOO
I 10
PLANT CAPACITY (MGD)
IOO
Figure 62. Waste water-plant labor costs.
-------
PAYBACK PERIOD
(YEARS)
/
0.01
HYDRAULIC LOAD FOLLOWING -.
TOTAL LOAD FOLLOWING
BY MEANS OF ON LINE
ANALYTICAL SENSORS
10 '100
PLANT SIZE (MGD)
(T) PHOSPHOROUS REMOVAL BY LIME
(?) PHOSPHOROUS REMOVAL BY ALUM
? PRECHLORINATION
DISINFECTION
AERATION
IT) DIGESTION (FIGURES 24 AND 25)
CENTRALIZED CONTROL
COMPUTER CONTROL
I I I INI
i
Figure 63. Automatic control payback periods vs plant size for typical
municipal waste water treatment.
-------
Basis for Economic Comparisons
Several alternate control systems are evaluated for each of the main unit operations
in industrial or municipal waste water-treatment plants. Some of these control schemes
are superior to the manual (uncontrolled) mode of operation in that they can guarantee
consistent and acceptable effluent quality, whereas others can demonstrate an economic
advantage because of the resulting increase in equipment life and the reduction in:
• Equipment size • Consumption of chemicals
• Operating labor • Consumption of electric power
• Maintenance requirements • Quantity of sludge produced.
The balance between costs and benefits is also affected by the size of the plant, with
the larger operations being in a better position to justify the initial cost of a higher
level of automation. There is no dollar value attached to the consistent meeting of
effluent quality regulations by automation devices, nor is there a cost penalty assigned
to the inconsistent performance of manually operated plants. These are treated as
extra bonuses or penalties that are difficult to be quantitatively evaluated. It is
difficult to quantify any economic advantages gained because of increases in effluent
quality.
Throughout this section, the dollar values used are 1974 first-quarter dollars, which
correspond to 362 on the Marshall & Stevens scale and to 151 on the Chemical Engineer-
ing cost index. Unless otherwise noted, the plant lives are assumed to be 25 years.
The interest rates are assumed to be 6% for instruments with a life expectancy above
10 years, and 8% for instruments with a life expectancy below 10 years. Chemical
coats used in the cost benefit analyses are representative averages. No provisions
have been made with regard to volume, method of shipment, storage, or plant location
cost differentials. As a result, the methodology in this section should be applied as a
guideline.
185
-------
Evaluation of Cost Benefits
Table 13 provides the approximate operating and capital costs of wastewater-treat-
ment plants (50) together with their total numbers in the U. S. Usually less than 4% of
the listed capital cost is spent on instrumentation, and more than half of the operating
cost is spent on labor. A recent nationwide survey of wastewater-treatment facilities
found that most secondary plants spend about 3.3% of their capital cost on instrumenta-
tion (1). Finally, treatment plant capital and operating cost data, expressed as a
function of plant size (as in Figure 61), clearly shows the economy of scale. This leads
to a popular misconception which is that only large plants (i. e., 50 mgd) can afford the
luxury of instrumentation and automatic control budgets. However, the higher unit
treatment costs of smaller sized plants may make it feasible for medium-sized plants
(i. e., 5 mgd) to reap even greater benefits from instrumentation.
Table 13. APPROXIMATE COSTS OF WASTEWATER-TREATMENT PLANTS
Costs
Yearly operating costs ($10°/year)*
6
Capital cost ($10 )
Number of domestic municipal and
industrial plants in this size range
Plant Capacity (mgd)
10
0.3
8.0
500.0
50
1.0
25.0
150.0
100
1.5
45.0
100.0
*Accordtng to Reference 51, the components of total operating and maintenance
costs in the 1965 to 1968 period in municipal plants were: 1) labor—60%, 2)
electricity—14%, 3) chemicals—4%, and 4) others—22%.
The installed cost of wastewater-treatment equipment is substantial and well documented
(52). The use of state-of-the-art instrumentation may reduce the quantity and/or size
of these devices by guaranteeing their optimized utilization and by eliminating their
intermittent use. One example of quantity reduction would be the continuous—instead
of intermittent—use of vacuum filters, while one example of size reduction would be
186
-------
the feedforward control of pH, which allows for a substantial reduction in the size of
the neutralization tank. Similarly, the size of pumps can be reduced and their life
increased if they are operated continuously instead of on and off.
Well-designed control systems make the various pieces of equipment work only as
hard as necessary. This (in the case of aerators, for example) can result in a 10%
increase in equipment life.
Automation can also eliminate the misoperation-type of failures due to human error.
Burning-out motors, flooding, or running the equipment dry are all prevented by well-
designed instrumentation, which pays off in lower maintenance expenses, less equip-
ment downtime, and higher overall efficiency.
Compared to all other operating expenses, labor costs are the highest in wastewater-
treatment plants. If the decision-making process (leading up to a manual adjustment
made by an operator) can be reproduced by instrumentation, the automated system
will perform more consistently, accurately, and reliably, in addition to relieving the
operator of that task. While increased levels of instrumentation will reduce the
operating labor costs of the plant, more sophisticated instruments and analyzers
would require a higher level of maintenance attention. The probable overall saving is
treated as an additional benefit from automation and is not included in the quantitative
cost/benefit analysis for the various unit operations. However, the potential reduc-
tions in operating labor costs via the labor-saving role of centralized and computerized
control are significant and are discussed subsequently in this section in detail.
Table 14 lists some of the unit costs and dosages of chemical additives (53). On an
average, chemicals account for less than 10% of the total operating costs. Automation
can substantially reduce the total use of chemicals by continuously monitoring and
matching only the actual demand. However, only a small number of chemical addition
loops are actually controlled on a demand basis other than flow proportional. In fact,
the field survey of user experiences (1) observed no cases of significant chemical
187
-------
Table 14. COST OF CHEMICAL ADDITIVES*
Cost/Ton ($)
Typical Dose (mg/liter)
Corresponding
Yearly
Cost ($)
1 MOD Plant
10 MOD Plant
100 MGD Plant
Hydrated
Lime
20
250
8,000
80,000
800,000
Ferric Chloride
100
100
16,000
160,000
1,600,000
Chlorine
100**
15
2,500
25,000
250,000
*Other chemical costs /ton ($): alum $80 (costs are based on a liquid alum
addition at $80/ton of 22% dry AlgOg), ammonia $70, caustic (100%) $140,
sulfuric acid (100%) $40, HC1 (36%) $40, SO2 liquid $75, and polymers $1000
to $2000.
**Cost does not include shipping or cylinder rental charges.
savings by means of demand control with the exception of chlorine. The increased
popularity of physical/chemical treatment and chemical advanced wastewater treat-
ment will undoubtedly increase the significance of chemical saving control systems.
In plants where chlorination, flocculation, coagulation, phosphorous removal, and/
or neutralization are also practiced, the resulting savings can be very substantial.
BASIC ASSUMPTIONS COMMON TO ALL UNIT OPERATIONS
Load Following
All wastewater-treatment facilities experience cyclic (diurnal) variations in their
hydraulic and organic loads. In order to lessen the use of chemical additives, the
best control method is to ratio the additive flow to the pollutant load. A halfway
measure ratios the rate at which the additive is charged to the volumetric flow rate
of the incoming wastewater, while the least desirable technique charges the additive at
a constant rate.
188
-------
In order to evaluate the benefits of hydraulic load following, it is necessary to approxi-
mate the amount of chemical additives that will be conserved through use of this method.
The percentage of these chemicals is estimated by a comparison with the constant feed
method, where the addition rate is set for the maximum load and is left unaltered dur-
ing the day. The relationship between the peak and average flows is as follows (54):
F = 1. 84 (F )
p a'
0.92
where:
F = peak flow
P
F = average flow
a
The peak-to-average ratio decreases with increasing flows. These are listed in
Table 15 for 1, 5, 10, 50, and 100 mgd average flows.
Table 15. SAVINGS IN CHEMICAL ADDITIVE USE BY PRACTICE OF
LOAD FOLLOWING
Average
Plant
Flow (Fa)
(mgd)
1
5
10
50
100
500
Hydraulic Load Following
Peak-to-Average
Ratio of Flow
(K - F /F )
p a'
1.84
1.62
1.53
1.34
1.27
1.18
Percent of
Savings Over
Constant Rate
Charging
,«, K-l
(/u" K )
46
38
35
24
21
15
Total Load Following
Peak-to-Average
Ratio of Chemical
Additive Use
(A = 0.75K2 + 0.25K
= FpCp/FaCa)
3.00
2.38
2.13
1.68
1.53
1.34
Percent of
Savings Over
Constant Rate
Charging
f*"1)
( A }
67
58
53
40
35
26
189
-------
If the daily hydraulic load variation is assumed to be sinusoidal (6) around the average
flows, then setting the chemical addition rate continuously at a level corresponding to
the plant's peak flow will result in unnecessary overcharging of the chemical additive
by the ratio of F /F , as shown in Figure 64 and Table 15.
pa
The concentration of a wastewater stream also varies according to a diurnal schedule
(6). This variation can also be considered sinusoidal in character and is in phase
with the hydraulic load variation. In other words, at low hydraulic loads the concen-
tration is also likely to be low, and at peak volumetric flows the pollutant concentration
is also likely to be high.
The savings in the consumption of chemical additives, therefore, will be even greater
when the feature of pollutant concentration load following is added on top of the hydraulic
flow ratioing. This amplification effect can be made quantitative by assuming that the
pollutant concentration variation equals 75% of the hydraulic load variation (55). This
can be expressed as:
C = (0.75(K-1) + 1) Ca
P
where:
C = peak concentration
P
Ca = average concentration
K= F /F
p a
If the chemical additive was charged at a constant rate, its flow would correspond to
the peak values of both the pollutant concentration and hydraulic load:
constant rate setting = (F ) (C ) = (KF ) (0. 75(K-1) + 1) Ca
P P a
If it is assumed that, with the total load following, the rate of chemical additive usage
corresponds to F Ca, then the peak-to-average chemical use ratio is:
a
2
F Cp/F Ca = 0.75K + 0.25K
P a
190
-------
DOSE
DOSE
DOSE
REQUIRED AMOUNT
CONSTANT RATE ADDITION
HYDRAULIC LOAD FOLLOWING
TOTAL LOAD FOLLOWING
TIME
TIME
Figure 64. Various methods of chemical additive changing.
191
-------
The percentage savings accomplished by the total load following is provided in Table
15, while the three methods of chemical additive charging all illustrated in Figure 64.
Chemical Additive Doses
After approximating the percentage savings potential of the various control strategies,
the next logical step is to determine the quantities of chemicals to which these percent-
ages apply. Table 16 lists some of the more important unit operations, together with
typical dosages of chemical additives (assuming a constant rate of charging) and their
daily cost for various plant sizes.
The daily cost of makeup is based on the cost data in Table 14, and the daily cost of
chemical sludge disposal is estimated on the basis of $5/wet ton, with alum treatment
Table 16. CHEMICAL ADDITIVE DOSES
Unit Operations
Odor control (prechlorination) (57)
C12 dose (ppm)
Daily C12 usage (Ib/day)
Daily C12 cost ($/day)
Phosphorus precipitation by lime (56)
Lime dose (ppm)
Daily lime consumption (tons /day)
Daily makeup cost ($/day)
Daily sludge disposal cost ($/day)
Daily total cost ($/day)
(58,59)
Phosphorus precipitation by alum
Alum dose (ppm)
Daily alum consumption (tons /day)
Daily makeup cost ($/day)
Daily sludge disposal cost ($/day)
Daily total cost ($/day)
Disinfection via chlorination (57)
Cl£ dose (ppm)
Daily Cl£ usage (Ib/day)
Daily C12 cost ($/day)
Plant Size (mgd)
1
5
42
5
500
2.1
42
15
57
200
0.83
66
1.25
67.25
15
126
15
5
5
210
25
500
10.5
210
75
285
200
4.15
330
6.25
336.25
15
630
75
10
5
420
50
500
21
420
150
570
200
8.3
660
12.50
672. 50
15
1260
150
50
5
2100
250
500
105
2100
750
2850
200
41.5
3300
62
3362
15
6300
750
100
5
4200
500
500
210
4200
1500
5700
200
83
6600
125
6725
15
12,600
1500
192
-------
resulting in 0.25 ton of wet chemical sludge after each mg treated, and lime producing
3 tons of wet chemical sludge per each mg treated (56).
One can use the data in Table 16 to calculate quantitative costs/benefit. For example,
to calculate the potential daily savings in operating a 50 mgd phosphate precipitation
unit that uses alum by converting its control strategy from hydraulic load following to
total load following, Tables 15 and 16 contain the data required. In Table 15, the
percentage of alum saved is found to be 40 - 24 = 16%. In Table 16, the daily cost
savings is found to be (0.16) (3,362) = $540. In figuring a cost/benefit analysis, it is
necessary to balance this savings against the cost of added instrumentation.
Parallel Trains of Equipment
Because the instrumentation cost is a function of the number of parallel trains of
equipment rather than the wastewater-treatment plant capacity, some assumptions
are desirable in this respect. Table 17 lists some of the major pieces of equipment
and the probable number of parallel units that are likely to exist in plants of various
sizes. These quantities form the basis of the cost/benefit analysis of various control
strategies (60).
In Table 17 it can be noted that the quantity of instrumentation (not the size of pipe-
line items) is about the same for 1, 5, and 10 mgd plants. Therefore, wherever the
benefits of automation are proportional to plant capacity, it will be much easier to
justify a higher level of automation for a 10 mgd plant than for a 1 mgd one.
Table 17. NUMBER OF PARALLEL UNITS
Unit Operations
Primary clarifier
Aeration tank
Secondary clarifier
Sludge dewatering (vacuum filter)
Sludge digester
Plant Capacity (mgd)
I
2
2
2
2
2
5
2
2
2
2
2
10
2
2
2
2
2
50
4
4
4
4
4
100
8
8
8
8
8
500
20
20
20
20
20
193
-------
Assumptions on Sensors and Final Control Elements
In order to establish a common basis for comparing the costs of more sophisticated
control systems to manual (or operator-controlled) installations, it is essential to
define the types of instruments that must be added for automatic operation. It is
also necessary to make some assumptions about the types of instruments that already
exist in a manually operated plant, since the cost of these devices does not enter into
a relative economic evaluation of the various control strategies. Such instruments
will be distinguished on the control system drawings by two concentric circles.
A transmitter signal will be assumed to exist and be freely available for all control
strategies. This assumption applies to all flow-ratioed additive control strategies,
including deodorizing, neutralizing, precipitating, coagulating, flocculating, or
disinfecting control systems.
Variable speed metering pumps and adjustable chlorinators will be assumed to be
common to all control strategies. When it is sufficient to use such indirect flow data
as chlorinator or metering pump setpoints, this information will be assumed to be
freely available to all control strategies.
When direct and independent flow measurements are required, this will be assumed to
be an extra-cost item. For example, if it is desirable to install a magnetic flow-
meter to detect the flow rate of lime slurry, this will be treated as an additional-
cost item for the particular control strategy involved.
A transmitter signal that represents the individual airflows to each aeration tank will
be assumed to be freely available.
When sludge flow information is obtained indirectly from pump speeds, this data will
be assumed to be freely available. When direct flow data is required for the control
strategy, the cost of the sensor (such as a magnetic flowmeter) will be assumed to be
extra. All magnetic flowmeters in sludge service will be assumed to be 6 inches in
size and mounted in 8-inch diameter pipes (unless otherwise noted). For sludge
service the magnetic flowmeters will be estimated to be uncalibrated weather-proof
194
-------
units with fiberglass lining and ultrasonic cleaners. The 1974 uninstalled costs that
follow include the cable and the general-purpose converter required to generate a 3
to 15 psig output signal: 1) four inches $3r30, 2) six inches $3380, and 3) eight
inches $3900. Unless otherwise noted, the 6-inch size will be assumed to be applicable
at a rounded-off unit price of $3500. Finally, the digester gas flow rate will be con-
sidered an extra-cost item for the control strategies needing this information.
When the final control element is a pump or a feeder, it is assumed to exist and be
freely available; when it is a valve, it is assumed to be an extra-cost item.
For sludge service the control valves will be estimated as ductile iron Veebal valves.
Their 1974 uninstalled costs are as follows: 1) four inches: $865, 2) six inches:
$1005, and 3) eight inches: $1400. Unless otherwise noted in this text, the 6-inch
size will be assumed to be applicable at a rounded-off unit price of $1000. This is
the right size for an 8-inch sludge pipe installation (61).
If a control strategy requires measuring the interface between the sludge and effluent
or detecting the percentage of solids in a flowing stream, the sensors required will
be treated as extra-cost items.
All chemical analyzers required by a particular control strategy will be considered
extra-cost items.
Cost of Instrument Maintenance
The following assumptions were used to arrive at an hourly rate for instrument
repair, tuning, and maintenance:
• Maintenance is performed by in-house technicians.
• The base salary (and corresponding skill) of these technicians does not
exceed the $7/hour rate.
195
-------
• The level of benefits and the ratio of nonproductive overhead are those of a
typical municipal operation.
• There is a scheduled preventive-maintenance program in effect to guarantee
that the idle time of maintenance personnel does not exceed 10%.
Based on these assumptions, a $10/hour rate has been used throughout this section.
Table 18 lists the estimated yearly maintenance requirements of the various types of
instruments (1, 62) based on actual service in municipal wastewater-treatment plants
and industrial facilities.
In small plants, it may be more economical to rely on contract instrument maintenance
at approximately $25/hour rather than to provide a full-time in-house maintenance
capability.
For the purposes of this document, the instruments are grouped into three classes
(refer to Table 18 for specific values) (1, 62):
• Class I—Over 15 years life expectancy
• Class II—Ten years (or a 5 to 15 range) life expectancy
• Class III—Under 5 years life expectancy.
REVIEW OF UNIT OPERATIONS
For phosphate precipitation through lime addition the cost of additional instruments
needed for hydraulic and total load following will be compared to the economic benefits
of reduced lime usage and the reduced chemical sludge disposal cost.
When lime is charged at a fixed rate corresponding to the maximum flow rate and the
maximum pollutant concentration during the daily cycle, it results in the maximum
expense in lime makeup, CO use and sludge disposal. The yearly total cost for
z
various plant sizes (Table 16) is as follows:
196
-------
Table 18. EXPECTED LIFE AND YEARLY MAINTENANCE REQUIREMENTS
BY TYPE OF INSTRUMENT
Instrument Type
Panel-mounted devices
Annunciators
Controllers
Converters
Indicators
Programmers
Recorders
Switches
Final control elements
Position (control valves)
On-off valves, pumps
Variable -speed feeders, pumps
Transmitters
Flow
Level
Others
Pressure
Temperature
Analyzers
CH4
Chromatograph
COD, TOC, TC
Combustibles
Conductivity
C02
DO
NDIR
Nuclear sludge density
02 in gas
pH and ORP
Phosphate
Refractive index
Maintenance
(hours /year)
2
8
10
4
20
8
2
12
4
16
8
6
10
4
6
50
150
150
25
60
25
GO
75
50
40
50
150
40
Lifespan*
(years)
15
10
10
15
5
10
15
10
15
10
15
15
10
15
15
5
5
5
5
10
10
5
5
10
5
5
5
5
Instrument Type
Residual chlorine
Respirometer, BOD
Turbidity
UV
Sensors
Magnetic
Orifice
Position (displaced gas)
Position (displaced liquid)
Propellors
Venturi
Weir, flume
Level
Bubbler
Capacitance
d/p
Float and cable
Nuclear
Pressure
Bourdon
d/p
Temperature
Bimetallic
Filled
TC, RTD
Weighing systems
Position (sluice gates)
Speed
Maintenance
(hours /year)
140
150
60
60
12
5
50
12
10
20
2
8
6
5
GO
10
2
4
4
6
8
GO
30
10
Lifespan*
(years)
5
5
10
5
15
15
5
10
10
15
15
15
10
15
15
10
15
15
15
10
15
10
5
5
*5 means 0 to 5, 10 means 5 to 15, and 15 implies 15 or more.
-------
1 mgd $ 20, 500
5 mgd $ 102, 500
10 mgd $ 205, 000
50 mgd $1,025,000
100 mgd $2,050,000
The potential total savings resulting from the application of the hydraulic load follow-
ing and total load following control strategies is given in Table 19 (multipliers from
Table 15).
Table 19. BENEFITS OF HYDRAULIC LOAD FOLLOWING IN PHOSPHORUS
REMOVAL VIA LIME ADDITION
PlantSize
(mgd)
Savings Through
Hydraulic Load Following ($/year)
Savings Through Total Load
Following ($/year)
1
5
10
50
100
(0.46) (20,500)= 9,400
(0.38) (102,500)= 39,000
(0. 35) (205, 000) = 72, 000
(0.24) (1,025,000) =245,000
(0.21) (2, 050, 000) = 430, 000
(0.67) (20,500) - 13,700
(0.58) (102,500) = 59,400
(0. 53) (205, 000) = 108, 500
(0. 40) (1, 025, 000) = 410, 000
(0. 35) (2, 050, 000) = 718,000
Figure 65 describes the instruments required for hydraulic load following. The
corresponding capital and operating costs (63) are contained in Table 20. The annual-
ized cost of one unit of this control system (64) based on a 6% interest rate, is $1695
for a useful life of 10 and 15 years as shown in Table 20.
It is assumed that one of these systems will be required for each primary clarifier
in the plant. Therefore, based on Table 17, the annualized cost as a function of
plant capacity is as follows:
198
-------
WASTEWATER FLOW (( FT
TO RAPID MIX TANK
(ONE PER PRIMARY CLARI FIERI
(SEE TABLE 17)
LIME SLURRY PUMP
Figure 65. Phosphorus precipitation by lime addition in hydraulic load
following control mode.
Table 20. CAPITAL AND OPERATING COSTS OF HYDRAULIC
LOAD FOLLOWING CONTROLS
Cost Components
10 years depreciation period
FT-2
FY-3
FRC-4
CV-5
15 years depreciation period
FE-1
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
Capital Cost
($)
1200
380
760
720
3060
1400
300
350
1400
1700
5150
8210
Req'd Maintenance
(hours/year)
10
10
16
12
12
15
75 hours/year
$750
199
-------
1 mgd $ 3,390
10 mgd $ 3,390
50 mgd $ 6,780
100 mgd $13,560
Considering the yearly cost of instrumentation listed above and the resulting yearly
savings (given in Table 19), this control system can be justified for all plant sizes.
The annualized cost of the required instrumentation, expressed as a percentage of the
resulting annual savings, is shown in Table 21.
Table 21. COST/BENEFIT ANALYSIS RESULTS FOR HYDRAULIC LOAD
FOLLOWING IN PHOSPHORUS REMOVAL VIA LIME ADDITION
Plant
Size
(mgd)
1
5
10
50
100
Annualized
Instrument
Cost ($)
3,390
3,390
3,390
6,780
13,560
Chemical
Savings
Resulting
From
Improved
Control
($/year)
9,400
39, 000
72, 000
245, 000
430, 000
Percent of
Total Savings
Spent on
Instrumentation
36
8.7
4.7
2.8
3.1
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
(2 x 8210)7(9400 - 3390) =2.8 years
(2 x 8210)7(39,000 - 3390) = 0.46 year
(2 x 8210)7(72,000 - 3390) = 0.24 year
(4 x 8210)7(245, 000 - 6780) = 0. 14 year
(8 x 8210)7(430, 000 - 13, 560) = 0. 16 year
Figure 66 describes the instruments required for a total load following, based on the
pH trimback on the hydraulic flow ratio. The corresponding capital and operating
costs (63) are shown in Table 22.
The annualized cost of one unit of this control system (64), based on an 8% interest
rate for components with 5 years lifespan and on 6% for all others as shown in Table
22 is $3734.
200
-------
FYYMSAGESETPQjNT
LIME SLURRY
Figure 66. Total load following.
One of these systems will be used before each primary clarifier and, therefore, the
annualized total cost as a function of plant capacity is:
1 mgd $ 7,468
lOmgd $ 7,468
50 mgd $14, 936
100 mgd $29, 872
Considering these yearly costs and the resulting yearly savings (given in Table 19),
this control system can also be justified for all plant sizes except the 1 mgd plant.
The annualized cost of the required instrumentation, expressed as a percentage of the
resulting annual savings, is shown in Table 23.
201
-------
Table 22. CAPITAL AND OPERATING COSTS OF TOTAL LOAD
FOLLOWING CONTROLS FOR PHOSPHORUS
PRECIPITATION BY LIME ADDITION
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
5 years depreciation period
pHE-6
pHT-7 (with ultrasonic cleaning)
10 years depreciation period
FT-2
FY-3
FRC-4
CV-5
pHRC-8
15 years depreciation period
FE-1
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
510
1,000
1,510
1,200
500
760
720
820
4,000
1,400
500
600
2,400
3,000
7,900
13,410
50
10
10
10
16
12
16
12
30
166 hours/year
Yearly cost of
replacement parts
is estimated as
$340
$2,000
202
-------
Table 23. COST/BENEFIT ANALYSIS RESULTS FOR TOTAL LOAD FOLLOWING
IN PHOSPHORUS REMOVAL VIA LIME ADDITION
Plant
Size
(mgd)
1
5
10
50
100
Annualized
Instrument
Cost ($)
7,468
7,468
7,468
14, 936
29, 872
Chemical
Savings
Resulting
From
Improved
Control
($/year)
13,700
59, 400
108,500
410, 000
718,000
Percent of
Total Savings
Spent on
Instrumentation
55
12.5
6.9
3.7
4.2
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
(2 x 13, 410)/(13, 700 - 7468) = 4. 3 years
(2 x 13,410)7(59,400 - 7468) = 0.58 year
(2x 13,410)7(108,500 - 7468) =0.27 year
(4 x 13,410)7(410, 000 - 14, 936) = 0. 14 year
(8 x 13, 410)7(718, 000 - 29, 872) = 0. 16 year
Phosphate Precipitation Through Alum Addition
The cost of additional instruments needed for hydraulic and total load following is
compared here to the economic benefits of reduced alum usage and to the cost reduc-
tion of chemical sludge disposal.
When alum is charged at a fixed rate corresponding to the maximum flow rate and to
the maximum pollutant concentration expected in a single day, this mode of operation
results in the maximum expense for alum makeup. Sludge disposal may generate
additional cost benefits but no economic analysis was performed. The yearly total cost
(Table 16) is $24,500 for a million gallons treated per day.
The potential savings, resulting from the application of the hydraulic load following
and total load following control strategies, is listed in Table 24. The multipliers used
are from Table 15.
Figure 67 describes the instruments required for hydraulic load following. The
corresponding capital and operating costs (63) are tabulated in Table 25. The annual-
ized cost of one unit, using 6% interest, is $1695 for the useful lines shown in Table
25.
203
-------
Table 24. BENEFITS OF AUTOMATION
Plant Size
(mgd)
Savings Through Hydraulic
Load Following ($/year)
Savings Through
Total Load Following ($/year)
1
5
10
50
100
(0.46) (24,500) = 11,300
(0.38) (122,500) =46,600
(0.35) (245,000) =85,500
(0.24) (1,225,000) =294,000
(0.21) (2,450,000) =514,000
(0.67) (24,500) = 16,400
(0.58) (122,500) = 71,000
(0.53) (245,000) =130,000
(0.40) (1,225,000) =490,000
(0.35) (2,450,000) =858,000
WASTEWATER
FLOW
ALUM
SOLUTION
E/P
RATIO
MAG
TO
RAPID MIX
TANK
Figure 67. Phosphorus precipitation by alum addition through hydraulic load
following control mode.
If one of these loops for each clarifier (Table 17) is used, the annualized cost as a
function of plant capacity varies from $3390 for a 1 mgd plant to $13,560 for a 100
mgd plant. The potential yearly savings resulting from this control strategy is shown
in Table 24, and the payback periods and other relevant data are furnished in Table 26.
204
-------
Table 25. CAPITAL AND OPERATING COSTS OF HYDRAULIC
LOAD FOLLOWING CONTROLS
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
10 years depreciation period
FT-2
FY-3
FRC-4
CV-5
15 years depreciation period
FE-1
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
1200
380
760
720
3060
1400
300
350
1400
1700
5150
8210
10
10
16
12
12
i5.
75 hours/year
$750
Figure 68 describes the instruments required for total load following, based on detect-
ing both the flow rate and phosphate concentration in the raw sewage. This feedforward
control strategy is limited by the unavailability of a reliable phosphorus analyzer (1).
If it is assumed that the phosphorus analyzer costs $7500 and is out of service 20% of
the time and that, during this period, only hydraulic load following is practiced, then
the capital and operating costs (63) for this system are as shown in Table 27.
The annualized cost of one unit of this control system, based on an 8% interest for
components with 5 years lifespan and on 6% for all others, as shown in Table 27, is
205
-------
Table 26. COST/BENEFIT ANALYSIS RESULTS FOR HYDRAULIC LOAD
FOLLOWING IN PHOSPHORUS REMOVAL VIA ALUM ADDITION
Plant
Size
(mgd)
1
5
10
50
100
Annualized
Instrument
Cost ($)
3,390
3,390
3,390
6,780
13,560
Chemical
Savings
Resulting
From
Improved
Control
($/year)
11,300
46,600
85,500
294,000
514, 000
Percent of
Total Savings
Spent on
Instrumentation
30
7.3
4.0
2.3
2.6
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
(2 x 8210)/(11, 300 - 3390) = 2. 1 years
(2 x 8210)7(46,600 - 3390) = 0.38 year
(2 x 8210)7(85, 500 - 3390) = 0.20 year
(4 x 8210)7(294, 000 - 6780) = 0. 11 year
(8 x 8210)7(514, 000 - 13,560) = 0. 13 year
ALUM
SOLUTION
E/P
RAW e-
SEWAGE^
MAC
'SETPOINT
'MULTIPLIER
PHOSPHORUS
TO
RAPID MIX
TANK
Figure 68. Phosphorus precipitation by alum addition through total load
following control mode.
206
-------
Table 27. CAPITAL AND OPERATING COSTS OF TOTAL
LOAD FOLLOWING CONTROLS FOR PHOSPHORUS
PRECIPITATION WITH ALUM ADDITION
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
5 years depreciation period
AT-6 (colorimetric orthophosphate
analyzer)!^
10 years depreciation period
AT-6 (sampling system, analyzer
house, etc.)
FT-2
FY-3
FRC-4
CV-5
15 years depreciation period
FE-1
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
7,500
2,000
1,200
500
760
720
5,180
1,400
500
600
2,600
3,200
8,300
20,980
150
50
10
10
16
12
12
30
290 hours/year
Yearly cost of
replacement parts
is estimated as
$350
$3,250
207
-------
$6685 (64). If one of these systems is used in front of each primary clarifier (Table
17), the annualized total cost as a function of plant size is $13,370 for a Imgd plant
and $53,480 for a 100 mgd plant.
The yearly savings can be obtained from Table 24, based on the assumption that the
total load following control strategy will be in operation for 80% of the time, and only
hydraulic load following will be practiced during the remaining 20% of the time. There-
fore, the projected savings are given in Table 28, and one can conclude that this
control strategy can be justified only for plants that are larger than 1 mgd.
Table 28. COST/BENEFIT ANALYSIS RESULTS FOR TOTAL LOAD FOLLOWING
IN PHOSPHORUS REMOVAL VIA ALUM ADDITION
Plant
Size
(mgd)
1
5
10
50
100
Annualized
Instrument
Cost ($)
13,370
13,370
13,370
26, 740
53,480
Savings
Resulting
From
Improved
Control
($/year)
15,360
66,300
121,100
451,700
790,000
Percent of
Total Savings
Spent on
Instrumentation
87.0
20.0
11.0
6.0
6.8
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
(2 x 20, 980)/(15, 360 - 13, 370) = 21 years
(2 x 20, 980)/(66, 300 - 13, 370) = 0. 79 year
(2 x 20, 980)/(121, 100 - 13, 370) = 0. 39 year
(4 x 20, 980)/(451, 700 - 26, 740) = 0. 20 year
(8 x 20, 980)/(790, 000 - 53, 480) - 0. 23 year
Prechlorination for Odor Control
In this cost/benefit analysis, the potential savings in chlorine consumption resulting
from automation is compared to the cost of the required instrumentation.
The main purpose of prechlorination is the destruction of such odor-causing com-
pounds as hydrogen sulfide. This can be accomplished by adding oxidizing compounds
that will selectively remove sulfur (ferrous chloride or chlorine). For the purposes
of this cost/benefit analysis, it will be assumed that prechlorination will be practiced
at a dose of 5 ppm Cl (Table 16). In installations where the total chlorine demand is
208
-------
high or is unpredictably variable, the use of an Fed additive may be a better choice,
^
because less chemical is required for the selective removal of sulfur and because
overchlorination—which can slow down subsequent biological treatment—is less likely
to occur.
When the chlorinator is set manually for charging chlorine at a fixed rate that
corresponds to the product of maximum wastewater flow rate and maximum chlorine
demand in the daily cycle, it results in the highest consumption of Cl . The yearly
£
total cost for various plant sizes (Table 17) is as follows:
1 mgd $ 1, 825
5 mgd $ 9,125
10 mgd $ 18,250
50 mgd $ 91, 250
100 mgd $182, 500
The potential total savings, resulting from the application of hydraulic load following
and total load following control strategies, is given in Table 29 (the multipliers used
are from Table 15.)
Table 29. POTENTIAL BENEFITS OF AUTOMATION
Plant
Size
(mgd)
Savings Through Hydraulic
Load Following
($/year)
Savings Through Total
Load Following
($/year)
1
5
10
50
100
(0.46) (1,825) = 840
(0.38) (9,125) = 3,460
(0.35) (18,250) = 6,380
(0.24) (91,250) = 21,900
(0.21) (182,500) = 38,400
(0. 67) (1, 825) = 1, 220
(0.58) (9,125) = 5,300
(0.53) (18,250)= 9,650
(0.40) (91,250)= 36,500
(0.35) (182,500) = 64,000
209
-------
Figure 69 describes the added instrumentation required for hydraulic load following.
As explained previously, it is assumed that the chlorinator and the wastewater flow
signal exist and are freely available to all control strategies. The capital and main-
tenance costs (63) for each of the prechlorination control systems are shown in Table
30. The annualized cost of a unit, using a 6% interest rate, is $226 (64) for a useful
life of 10 and 15 years, as shown in Table 30. On the basis of using one of these loops
for each clarifier (Table 17), the annualized cost as a function of plant capacity is as
follows:
1 mgd $ 452
5 mgd $ 452
10 mgd $ 452
50 mgd $ 904
100 mgd $1,808
F£h
3/2
MODULATED
CONTROL
SETPOINT
ASSUMES THAT THIS IS AN EXISTING
D/P CELL THAT DETECTS A WEIR LEVEL
THROUGH AN AIR BUBBLER SYSTEM
RAW SEWAGE f-
CI2 SOLUTION
TO
PRIMARY
CLARIFIER
CI2 SOLUTION DIFFUSER
Figure 69. Hydraulic load following controls for prechlorination.
210
-------
Table 30. CAPITAL AND OPERATING COSTS OF HYDRAULIC
LOAD FOLLOWING CONTROLS
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
10 years depreciation period
FY-1
FY-2
15 years depreciation period
Installation materials
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
320
200
520
100
180
250
530
1050
10
$100
These costs are lower than the projected yearly savings (Table 29) for all plant sizes.
The annualized cost of the required instrumentation, expressed as a percentage of the
resulting annual savings, is shown in Table 31, together with the projected payback
period.
Total load following control is accomplished by modulating the Cl feed rate in
^
accordance with both waste water flow and concentration of residual chlorine. Figure
70 describes a control system in which the chlorinator is paced by the raw sewage
flow signal to give an approximate Cl feed rate, and then is feedback trimmed by
&t
oxidation reduction potential (ORP). The ORP is an acceptable measurement for
feedback trimming because it measures the relative state of oxidation or reduction of
the sewage. The odor-causing anaerobic bacteria function best under negative ORP
211
-------
Table 31. COST/BENEFIT ANALYSIS RESULTS FRO HYDRAULIC LOAD
FOLLOWING CONTROL OF PRECHLORINATION
Plant
Size
(mgd)
1
5
10
50
100
Annual! zed
Instrument
Cost ($)
452
452
452
904
1,808
Savings
Resulting
From
Improved
Control
($/year)
840
3,460
6,380
21,900
38,400
Percent of
Total Savings
Spent on
Instrumentation
54.0
13.0
7.1
4.1
4.8
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
(2 x 1050)/(840 - 452) - 5.4 years
(2 x 1050)7(3460 - 452) = 0. 7 year
(2 x 1050)/(6380 - 452) = 0. 36 year
(4 x 1050)/(21, 900 - 904) = 0.20 year
(8 x 1050)7(38,400 - 1808) = 0.23 year
RAW SEWAGE
TO
PRIMARY
CLARIFIER
Figure 70. Total load following controls for prechlorination.
212
-------
conditions. Another reason for using ORP control is that residual chlorine analyzers
cannot be used, because a Cl residual would be wasteful of chlorine and would also
£u
retard the biological activity in the downstream aeration tank.
The capital and operating costs for the system (Figure 70), as estimated by Liptak
(63) are shown in Table 32. The annualized cost of one unit of this control system,
based on an 8% interest rate for components with 5 years lifespan and on 6% for all
others, as shown in Table 32, is $2529 (64). If one of these systems is used in front
of each primary clarifier (Table 17), the annualized total cost as a function of plant
size is $5058 for a 10 mgd or less plant, and $20,232 for a 100 mgd plant.
Table 33 contains the data regarding the overall cost/benefit analysis (the projected
yearly savings is based on Table 29). From the cost/benefit analysis, it can be
concluded that this control strategy can be economically justified only for plants that
are larger than 5 mgd.
Postchlorination for Disinfection
In this cost/benefit analysis, the potential savings in chlorine consumption resulting
from automation are compared to the cost of the required instrumentation.
The purpose of disinfection is to prevent the spread of waterborne diseases by eliminat-
ing pathogenic organisms. This is guaranteed by adding chlorine in sufficient quantities
to provide a residual chlorine concentration of under 1 ppm. Because the chlorine
itself is toxic, many areas of the nation require neutralization with SO before the
£
effluent leaves the wastewater-treatment plant in order to protect the receiving surface
waters.
When the chlorinator is set manually for charging chlorine at a fixed rate that
corresponds to the product of maximum wastewater flow rate and maximum chlorine
demand (assumed to be a dose of 15 ppm) in the daily cycle, it results in the highest
consumption of Cl . The total yearly costs for various plant sizes were presented in
£t
Table 16..
213
-------
Table 32. CAPITAL AND OPERATING COSTS OF TOTAL LOAD
FOLLOWING CONTROLS
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
5 years depreciation period
ORPE-5
ORPT-4
10 years depreciation period
ORPE-5 (sampling system)
ORPRC-3
FY-1
FY-2
15 years depreciation period
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
510
1000
1510
1000
820
320
200
2340
300
400
1600
1900
4200
8050
50
10
15
16
10
10
111 hours/year
Yearly cost of
replacement parts
is estimated as
$290
$1400
214
-------
Table 33. COST/BENEFIT ANALYSIS RESULTS FOR TOTAL LOAD
FOLLOWING CONTROL OF PRECHLORINATION
Plant
Size
(mgd)
1
5
10
50
100
Annualized
Instrument
Cost ($)
5,058
5,058
5,058
10,110
20,232
Savings
Resulting
From
Improved
Control
($/year)
1,220
5,300
9,650
36,500
64,000
Percent of
Total Savings
Spent on
Instrumentation
100
96
52
28
32
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
Not applicable
Excessive
(2 x 8050)/(9650 - 5058) = 3.5 years
(4 x 8050)/(36, 500 - 10, 116) =1.2 years
(8 x 8050)/(64, 000 - 20, 232) =1.6 years
The potential total savings resulting from the application of the hydraulic load follow-
ing and total load following control strategies to chlorination (and dechlorination) are
given in Table 34 (the multipliers used are from Table 15).
It is assumed that plants practicing fixed-rate chlorination also practice fixed-rate
dechlorination, using liquid SO at $75/ton. The combined economic benefits of load
Zi
following controls for both a Cl addition and an SO addition are assumed to be 10%
2 2i
higher than for Cl alone.
£t
Table 34. POTENTIAL ECONOMIC BENEFITS OF
AUTOMATION OF CHLORINATION
Plant Size
(mgd)
1
5
10
50
100
Savings Through
Hydraulic Load Following ($/year)
C12 Only
(0.46) (5,480)= 2,520
(0.38) (27,400)= 10,400
(0.35) (54,800)= 19,100
(0.24) (274,000) = 66,000
(0. 21) (548, 000) = 115, 000
Cl and SO
£* £1
2,770
11,400
21,100
72,700
126,000
Savings Through
Total Load Following ($/year)
C12 Only
(0.67) (5,480) = 3,660
(0.58) (27,400)= 15,900
(0.53) (54,800)= 29,000
(0.40) (274,000) = 137,000
(0.35) (548,000) = 191,000
Cl and SO
2 2
4,010
17,400
31, 800
150, 000
209,000
215
-------
It is assumed that the chlorinator, the wastewater flow signal, and the sulfonator
used in the dechlorination step all exist and are freely available to all control strate-
gies. Figure 71 describes the added instrumentation required for the hydraulic load
following control strategy. The capital and maintenance costs for the new instruments
are given in Table 35.
The annualized cost of one unit, using a 6% interest rate, is $664 (64) for the useful
lives indicated in Table 35. Based on using one of these systems downstream of each
secondary clarifier (Table 17), the annualized cost as a function of plant capacity is
$1328 for plants of 10 mgd and less and $5312 for 100 mgd plants. These costs are
lower than the projected yearly savings (Table 34) for all plant sizes. The annualized
cost of the required instrumentation, expressed as a percentage of the resulting annual
savings, is shown in Table 36, together with the projected payback periods.
ci,
H20
SO, SOLUTION
CONTACT-CHAMBER
PLUG FLOW REACTOR
DISCHARGE
TO
RECEIVING
WATERS
Figure 71. Hydraulic load following controls for disinfection and dechlorination.
216
-------
Table 35. CAPITAL AND OPERATING COSTS OF HYDRAULIC LOAD
FOLLOWING CONTROLS
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
10 years depreciation period
FY-1
FY-2
FY-3
FY-4
15 years depreciation period
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
320
200
320
320
1160
200
300
600
900
2000
3160
10
10
10
$300
Total load following is accomplished by modulating both the Cl and SO feed rates in
£ 2i
accordance with both wastewater flow and concentration. Figure 72 describes a con-
trol system in which both the sulfonator and the chlorinator are paced by the effluent
flow rate signal to give an approximate Cl feed rate, and then are trimmed in
£t
accordance with residual chlorine concentration. The precontact residual chlorine
analyzer (ARC-101) adjusts the chlorinator dosage, and the postcontact chlorine
analyzer (ARC-102) trims the setpoint of ARC-101 in a cascade manner. When, due to
a sudden upset or misoperation, ARC-102 cannot maintain its setpoint below some
preset limit, its output signal is sent through to the sulfonator, activating the SO
£i
charging loop (which is normally inactive).
217
-------
Table 36. COST/BENEFIT ANALYSIS RESULTS FOR HYDRAULIC LOAD FOLLOWING
CONTROL OF DISINFECTION AND DECHLORINATION
Plant
Size
(mgd)
1
5
10
50
100
Annualized
Instrument
Cost ($)
1,328
1,328
1,328
2,656
5,312
Savings
Resulting
From
Improved
Control
($/year)
2,770
11,400
21, 100
72, 700
126, 000
Percent of
Total Savings
Spent on
Instrumentation
48.0
11.6
6.3
3.7
4.2
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
(2x3160)7(2770 - 1328) = 4. 4 years
(2x3160)7(11,400 - 1328)= 0.59 year
(2x3160)7(21,100 - 1328) = 0.32 year
(4 x 3160)7(72, 700 - 2656) = 0. 18 year
(8 x 3160)7(126, 000 - 5312) = 0. 21 year
The capital and operating costs for the system (shown in Figure 72) were obtained
from Liptak (63) and listed in Table 37. The annualized cost of one unit of this control
system, based on an 8% interest rate for components with 5 years life span and on 6%
for all others, as shown in Table 37, is $7466 (64). If one of these systems is used
after each secondary clarifier (Table 17), the annualized total cost as a function of
plant size is $14, 932 for plants less than 10 mgd, $29,864 for 50 mgd plants and
$59,728 for 100 mgd plants. Table 38 contains the data regarding the overall cost/
benefit analysis. The projected yearly savings are based on Table 34. From the cost/
benefit analysis, it can be concluded that, unless the installation of this type of control
system is required by law, it can be economically justified only for plants that are
larger than 5 mgd.
Aeration
The cost of additional instruments needed for control of DO in activated sludge for
hydraulic and total load following control strategies will be compared to the economic
benefits of reduced power consumption and increased equipment life.
218
-------
CI, SOLUTION
CASCADE
MIXING DEVICE
CONTACT-CHAMBER
PLUG-FLOW REACTOR
SO, SOLUTION
DECHLORINATOR
Figure 72. Total load following control for disinfection and dechlorination.
219
-------
Table 37. CAPITAL AND OPERATING COSTS OF TOTAL LOAD
FOLLOWING CONTROLS
Cost Components
5 years depreciation period
AT-6 (amperometric Cl2 analyzer
with pneumatic output signal and
accessories)
AT-6 (same as above)
10 years depreciation period
AT-6 and AT-7 (sampling system
and housing)
ARC-8
ARC-9
FY-1
FY-2
FY-3
FY-4
AY-5
15 years depreciation period
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
Capital Cost
($)
3,500
3,500
7,000
3,000
820
820
320
200
320
320
320
6,120
750
1,000
2,000
2,400
6,150
19,270
Req'd Maintenance
(hours/year)
140
140
30
16
16
10
10
10
10
382 hours/year
Yearly cost of
replacement parts
is estimated as
$430
$4,250
220
-------
Table 38. COST/BENEFIT ANALYSIS RESULTS FOR
TOTAL LOAD FOLLOWING
Plant
Size
(mgd)
1
5
10
50
100
Annual! zed
Instrument
Cost ($)
14,932
14,932
14,932
29,864
59,728
Savings
Resulting
From
Improved
Control
($/year)
4,010
17,400
31,800
150,000
209, 000
Percent of
Total Savings
Spent on
Instrumentation
100
86
47
20
28
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
Not applicable
Excessive
(2 x 19, 270)/(31, 800 - 14, 932) = 2. 3 years
(4 x 19, 270)/(150, 000 - 29, 864) = 0. 64 year
(8 x 19,270)7(209,000 - 59,728) = 1.03 years
When the aeration is set at a fixed rate that corresponds to the maximum flow rate and
maximum BOD concentration during the daily cycle, it results in a minimum equipment
lifespan and in a maximum consumption of power. The total cost of this operation
mode equals the sum of the associated operating and capital expenses.
The yearly operating cost can be calculated on the basis that the continuous power
consumption (64) for each mgd capacity unit is 20 kW, and that the cost of electricity
ranges from 3£ to 5£/kWh, as a function of the quantity used. This is summarized in
Table 39.
The annualized capital cost of the aeration equipment can be calculated by using the
following equation (65):
annual capital cost =
(1150) (P)
0.81 / ;(1 + I)
= 140 (P)
0.81
where:
I = present-day Marshall & Stevens cost index of 362
I = 1968 Marshall & Stevens cost index of 258
68
221
-------
Table 39. ANNUALIZED COSTS ASSOCIATED WITH FIXED-RATE AERATION
Plant
Size
(MGD)
1
5
10
50
100
500
Yearly Operating Costs
P
(Power
Used in
kW)
20
100
200
1,000
2,000
10, 000
Yearly
Consumption
of Electric
Power
(10G kWh)
0.175
0.875
1.75
8.75
17.5
87.5
Unit Cost of
Power
Considering
Quantity
Discount
(£/kWh)
5.0
4.5
4.0
3.5
3.0
2.5
Total
Power
Cost
($/year)
8,750
39,600
70,000
306,000
525, 000
2,190,000
Annualized
Capital
Costs
140 (P)0.81
($/year)
1,650
5,740
10,250
37,800
66,300
244,000
Annualized
Total Costs
($/year)
(Columns
5 & 6)
10,400
45,340
80,250
343,800
591,300
2,434,000
I = assumed interest rate of 6%
^ = assumed life expectancy of 20 years
P = fixed maximum electric power requirement in kW (as shown in second
column in Table 39)
The annualized operating, capital, and total costs of fixed-rate aeration are shown in
Table 39 for the various plant sizes. The potential total savings that result from the
application of hydraulic load following, total load following through DO feedback con-
trol, and total load following through DO trimmed feedforward TOG control strategies
are shown in Table 40. The multipliers are derived from Table 15 and the annualized
costs from Table 39.
Figure 73 describes the instruments required for hydraulic load following. The
corresponding capital and operating costs taken from Liptak (63) and Table 18 are
listed in Table 41. The annualized cost of a single unit of this control system, based
•222
-------
Table 40. POTENTIAL BENEFITS OF AUTOMATION ($/YEAR)
Yearly Savings
Savings in operating
costs
Savings in annualized
capital costs
Total annualized savings
Plant
Size
(mgd)
1
5
10
50
100
500
1
5
10
50
100
500
1
5
10
50
100
500
Benefits of Improved Control
Hydraulic Load Following
Increased Lifespan
(5% - 21 Years)
Reduction in Electric
Power Consumption (10%)
$ 872
$ 3,255
$ 5,316
$15,312
$23,870
—
$ 9
$ 56
$ 96
$ 240
$ 364
—
$ 881
$ 3,311
$ 5,412
$15,552
$24,234
—
Total Load Following
Feedback (DO)
Increased Lifespan
(109? = 22 Years)
Reduction in Electric
Power Consumption (25%)
$ 2,110
$ 8,566
$13,780
$45, 600
$68,400
—
$ 60
$ 185
$ 293
$ 855
$ 1,290
—
$ 2,170
$ 8,745
$14,075
$46,435
$69,690
—
Feedforward (TOC
Trimmed by DO)
Increased Lifespan
(15% = 23 Years)
Reduction in Electric
Power Consumption (35%)
$ 2,900
$ 12,000
$ 19,400
$ 63,700
$ 95,000
$298,000
$ 113
$ 253
$ 400
$ 1,300
$ 1,740
$ 5,500
$ 3,013
$ 12,253
$ 19,800
$ 65,000
$ 96,740
$303,500
to
to
CO
-------
SETTLED
SEWAGE
AERATION TANK
No. 1
i I I I I I I 1
AIR -
BLOWER
J
(m
t
IER
i i
AEF
L t
1ATI
t i
)RS
1 c
c )V
i
VENT
Figure 73. Hydraulic load following control for aeration.
on a 6% interest rate and the data given in Table 41, is, according to Molvar (64),
$1026. It is assumed that one of these systems will be required for each aerator in
the plant. Therefore, based on Table 17, the annual!zed cost as a function of plant
capacity is $2052 for plant sizes of 1 to 10 mgd, increasing to $8208 for 100 mgd.
Considering the yearly cost of instrumentation listed above and the resulting yearly
savings shown in Table 40, this control system can be justified only for plants that
are larger than 5 mgd. The annualized cost of the required instrumentation, expressed
as a percentage of the resulting annual savings, is shown in Table 42.
Figures 74 and 75 describe two total load following control strategies. The first is
based on feedback DO control in addition to the hydraulic load following system
described in Figure 73, while the second also includes feedforward TOC. The
224
-------
Table 41. CAPITAL AND OPERATING COSTS OF HYDRAULIC LOAD
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/years)
10 years depreciation period
FY-1
FY-2
FRC-3
CV-4
15 years depreciation period
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
380
200
760
600
1940
200
300
750
1000
2250
4190
10
5
16
12
10
53 hours/year
$530
corresponding capital and operating costs obtained from Liptak (63) and Table 18 are
given in Table 43. The annualized costs for a single unit of each control system are
calculated on the basis of an 8% interest rate for components with 5 years lifespan
and 6% for all others, with the data given in Table 43, according to Molvar (64):
annual cost of feedback control = $3110
annual cost of feedforward control = $8290
If one of these systems is installed for each aeration tank (Table 17), the annualized
total cost as a function of plant size for both the feedback and feedforward control
strategies is as follows:
225
-------
Table 42. COST/BENEFIT ANALYSIS RESULTS FOR HYDRAULIC LOAD
FOLLOWING DO CONTROL
Plant
Size
(mgd)
1
5
10
50
100
Annual! zed
Instrument
Cost ($)
2,052
2,052
2,052
4,104
8,208
Savings
Resulting
From
Improved
Control
($/year)
886
3,300
5,400
15,600
24,000
Percent of
Total Savings
Spent on
Instrumentation
>100
61
37
26
34
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
Not applicable
6.5 years
(2 x 4190)/(2494 - 2052) = 2. 5 years
(4 x 4190)7(7461 - 4104) = 1.5 years
(8 x 4190)7(11, 168 - 8208) = 2.1 years
SETTLED
SEWAGE
AIR
Figure 74. Total load following based on feedback control.
226
-------
SAMPLE LOOP
SETTLED t
SEWAGE
L I I I I I I I I
Figure 75. Total load following based on feedforward control.
Size
1 mgd
5 mgd
10 mgd
50 mgd
100 mgd
500 mgd
Feedback
$ 6,220
$ 6,220
$ 6,220
$12,440
$24,880
Feedforward
$ 16,580
$ 16,580
$ 16,580
$ 33,160
$ 66,320
$165,800
Based on the assumptions that: 1) the cost of electric power will remain discounted
to large users (Table 39), and 2) a complete TOC analyzer is to be dedicated to each
227
-------
Table 43. CAPITAL AND OPERATING COSTS OF TWO TOTAL LOAD
FOLLOWING CONTROLS
Cost Components
5 years depreciation period
AE-5 (galvanic DO with cleaner)
AT- 6 (amplifier-converter)
AT-8 (TOC analyzer)
10 years depreciation period
AT-8 (sampling system, housing,
etc.)
FY-1
FY-2
FRC-3
CV-4
ARC-7
AY-9
15 years depreciation period
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
Feedback Scheme
(Figure 74)
Capital
Cost ($)
800
1,000
1,800
380
200
760
600
820
2,760
400
600
1,600
2,100
4,700
9,260
Req'd Maintenance
(hours/year)
60
10
10
5
16
12
16
15
144 hours/year
Yearly cost of
replacement parts
is estimated as
$360
1,800
Feedforward Scheme
(Figure 75)
Capital
Cost ($)
800
1,000
10,000
11, 800
2,000
380
200
760
600
820
300
5,060
600
750
2,200
2,800
6,350
23,210
Req'd Maintenance
(hours/year)
60
10
150
25
10
5
16
12
16
10
20
334 hours/year
Yearly cost of
replacement parts
is estimated as
$660
4,000
228
-------
aeration tank, the potential benefits of feedforward control (Table 40) are less than
its annualized cost in all cases except for 50 mgd and 500 mgd plants. If either one
or both of these assumptions are changed, the economic justification of feedforward
TOG control can become feasible for larger than 5 mgd plants. For these reasons,
the feedforward control strategy will no longer be considered in this analysis.
In evaluating the benefits of the feedback control strategy, it is assumed that it will
be in service 90% of the time while, for the remaining 10% of the time, only hydraulic
load following will be practiced. Based on Table 40, the projected savings are:
1 mgd - (0.1) (691) + (0.9) (2,170) =$ 2,022
5 mgd - (0.1) (3,311) + (0.9) (8,743) =$ 8,200
10 mgd- (0.1) (5,412) + (0.9) (14, 075) = $13, 500
50 mgd - (0. 1) (15,681) + (0.9) (46,455) = $43,460
100 mgd - (0.1) (24, 218) + (0. 9) (69,690) = $65,140
Considering the data in Table 44, this control strategy can be justified only for plants
that are larger than 5 mgd.
Table 44. COST/BENEFIT ANALYSIS FOR TOTAL LOAD
FOLLOWING VIA FEEDBACK CONTROL
Plant
Size
(mgd)
1
5
10
50
100
Annualized
Instrument
Cost ($)
6,220
6,220
6,220
12,440
24, 880
Savings
Resulting
From
Improved
Control
($/year)
2,022
8,200
13,500
43,365
65, 140
Percent of
Total Savings
Spent on
Instrumentation
>100
75
46
28
38
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
Not applicable
9.35 years
(2 x 9260)/(8, 680 - 6, 220) = 2.5 years
(4 x 9260)7(28, 600 - 12,440) = 1. 21 years
(8 x 9260)7(43, 000 - 24, 880) = 1. 84 years
229
-------
Anaerobic Digestion
In this unit operation, the cost of additional instruments is compared to the economic
benefits of increased sludge stabilization capacity. Digester capacity is increased by
extending its availability and reducing the frequency and duration of digester failures.
The digestion process biologically converts organic materials into methane and carbon
dioxide (13). The steady-state material balance between the amount of organic mate-
rials fed and the quantity of methane produced by the methane-forming bacteria can be
upset by three types of digester failure causes:
• Hydraulic Overloading—Will wash out the methane-forming microbial
population at a faster rate than they are produced.
• Organic Overloading—Will result in the accumulation of volatile acids,
which tend to lower the pH and inhibit the methane-forming organisms.
• Toxic Overloading—Causes the death of me thane-forming organisms.
Several process variables can signal the approach of a digester upset or failure (for
instance, a sudden increase in hydraulic or organic load, low pH, or reduced CH
production), and several manipulated variables are available to the instrument
engineer as "handles on the process," which can be used to restabilize the operation.
These include:
• Adjusting the operating temperature
• Modifying the pH by adding a basic reagent
• Changing the sludge feed rate
• Recycling some of the microorganisms that get separated in two-stage
digesters.
The costs and benefits of the control strategies based on these considerations will be
compared here to the performance of uncontrolled digesters.
230
-------
The benefits of automated operation will be determined first by establishing the total
volume of the digesters and then by estimating the percentage reduction in digestion
capacity, which can be obtained by better use through automation. Table 45 provides
the basis on which various digester system costs can be estimated (52). As a first
approximation, it can be said that single-stage digester systems represent 30%, and
two-stage digester systems 40%, of the total capital investment in a wastewater-treat-
ment plant. Table 46 shows the savings in capital costs for various plant sizes,
assuming a 10%, 15%, 20%, 25%, and 30% increase in their throughput due to auto-
mation.
Approximately 8 Btu/hours are required for each cubic foot of digester volume to
maintain its operating temperature at the desired 35° C. This heat requirement is
about 25% of the total energy content of the generated digester gas (66).
Table 45. DIGESTER SYSTEM COSTS
Design Adjustment, FQ
Type
One Stage
Two Stage
FD
1.000
2.358
Regional Adjustment, FR
Region
1
2
3
4
5
6
7
8
9
10
FR
1.000
1.061
0.984
1.032
1.073
1.024
1.062
1.024
1.065
1.031
Notes: Installed reactor cost includes purchased cost of reactor, auxiliaries,
handling and setting, piping, concrete, steel, instrumentation, electrical, insula-
tion, paint, and indirect costs (prime contractor engineering and construction
overhead):
Installed reactor cost ($)—(Installed base cost) (F ) (F )
D ti
Annual maintenance—Will be approximately 2.47% of installed reactor cost.
Operation cost—Reactors require approximately 0.25 operator per shift.
231
-------
Table 46. CAPITAL COST SAVINGS FROM DIGESTER AUTOMATION
Plant
Size
(mgd)
1
5
10
50
100
1
5
10
50
100
Single-Stage Digesters
10%*
$0. 047
0.153
0.246
0.780
1,320
$ 4,850
15,700
25,300
80,500
136, 000
15%*
$0. 071
0.230
0.370
1.170
1.980
$ 7,300
23,700
38,100
120,000
204,000
20%*
$0. 094
0.305
0.490
1.560
2.640
±
$ 9,650
31,400
50,600
161,000
272,000
25%*
Camtal C
$0.117
0.382
0.615
1,950
3.300
^.nnualizec
$ 12,500
39,300
63,200
200,000
340,000
30%*
ost Saving
$0. 141
0.460
0.738
2.340
3.950
Savings*'
$ 14,550
47,000
75,700
241,000
406,000
Two-Stage Digesters
10%*
s mo6)
$0. 064
2.204
0.328
1.040
1.760
" ($/year)
$ 6,600
21,000
33,700
107,000
181,000
15%*
$0. 096
0.306
0.492
1.560
2.640
$ 9,900
31,500
50,600
161,000
272,000
20%*
$0.128
0.408
0.656
2.080
3.520
$ 13,200
42,000
67,400
214,000
362,000
25%*
$0.160
0.510
0.820
2.600
4.400
$ 16,500
52,500
84, 200
268,000
453, 000
30%*
$0.192
0.610
0.980
3.110
5.260
$ 19,800
62,700
101,000
321,000
540,000
to
CO
to
*Percent increase in throughput (Table 47).
**Based on a 6% interest rate and 15 years depreciation period.
-------
Figure 76 illustrates the temperature control system, which consists of a continuously
operating external sludge circulation system, controlled by a temperature cascade
loop. The continuous operation is advantageous because it keeps the digester tempera-
ture at a fixed value (instead of cycling between limits) and because it contributes to
good mixing and agitation.
Table 47 lists the estimated savings for the various types of control strategies.
Table 47. CAPITAL COST SAVINGS THROUGH AUTOMATION
(PERCENT OF TOTAL DIGESTER INVESTMENT)
Control Type
Temperature control only (Figure 76)
pH control only
Methane control only
Temperature and pH control (Figure 77)
Temperature, pH, and methane control
(Figures 24 and 25)
Percent Savings
Single Stage
10
15
15
20
25
Two Stage
10
15
25
20
30
HOT WATER
FEED
SLUDGE
RECIRCULATION
PUMP
HWR
Figure 76. Cascade temperature control of single-stage digester.
233
-------
The cost of an on-off cycling temperature control system is not substantially lower
than the cascade loop described in Figure 76. Capital and maintenance costs are
shown in Table 48 (these are the same for both single- and two-stage digestion).
The annualized cost of a single unit for temperature control, using a 6% interest rate,
is $1263 based on the information given in Table 48. Using one of these systems for
each digester (Table 17), the annualized cost as a function of plant capacity varies
from $2526 for a 1 mgd plant to $10,104 for a 100 mgd plant.
Table 48. CAPITAL AND OPERATING COSTS OF TEMPERATURE CONTROL
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
10 years depreciation period
TRC-2
TRC-3
CV-4
15 years depreciation period
TT-1
TT-5
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
820
760
500
2080
400
350
275
500
900
1250
3675
5755
16
16
10
6
6
60 hours/year
$600
234
-------
These costs are lower than the yearly savings projected for all plant sizes (see the
two 10% columns in Table 46). The annualized cost of the required instrumentation,
expressed as a percentage of the resulting annual savings, is shown in Table 49,
together with the projected payback periods.
In addition to good temperature control, it is possible to further improve the auto-
matic control of digesters by continuously monitoring and controlling the pH of the
circulated sludge. This will prevent the pH from dropping to the point where the
growth of methane-forming microorganisms is inhibited by excessive acidity. The
control system necessary to accomplish this is shown in Figure 77. pHE 6 is mounted
in an easily isolated bypass line of a well-mixed and continuously flowing sludge
sample. Ultrasonic cleaning should increase the mean time between the required
preventive-maintenance checks. The nonlinear controller (described previously)
should also contribute to a stable control performance.
The capital and maintenance costs shown in Table 50 are the same for both single-
and two-stage digestion systems. The annualized cost of a single unit, using an 8%
interest rate for devices with 5 years life and 6% for all others, is $3790 using the
data given in Table 50. Using one of these systems for each of the digesters (Table
17), the annualized cost as a function of plant capacity varies from $7580 for a 1 mgd
plant to $30,320 for 100 mgd plants.
These figures are lower than the yearly savings projected for all plant sizes if the
increase in digester capacity realizable is 20% or more (see Table 46). The annual-
ized cost of the required instrumentation, expressed as a percentage of the resulting
annual savings, is shown in Table 51, together with the projected payback periods.
In addition to the pH and temperature, the quantity of methane gas generated is an
important indicator of digester performance. If, at a constant feed rate (fixed
hydraulic and organic loading), the methane production drops off, it can be taken as
an early signal of a toxic overloading episode. Automatic controls can respond to this
235
-------
Table 49. COST/BENEFIT SUMMARY OF TEMPERATURE CONTROL OF DIGESTION
Digester
Type
Single
stage
Two stage
Plant
Size
(mgd)
1
5
10
50
100
1
5
10
50
100
Annualized
Instrument
Cost
($/year)
2,526
2,526
2,526
5,052
10, 104
2,526
2,526
2,526
5,052
10, 104
Savings
Re suiting From
Improved
Control ($/year)
4,850
15,700
25,300
30,500
136,000
6,600
21,000
33,700
107,000
181,000
Percent of
Total Savings
Spent on
Instrumentation
52.0
16.0
10.0
6.3
7.4
38.0
12.0
7.5
4.7
5.6
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
(2 x 5755)7(4850 - 2526) = 5 years
(2 x 5755)/(15, 700 - 2526) = 0. 87 year
(2 x 5755)7(25,300 - 2526) = 0.5 year
(4 x 5755)7(80,500 - 5052) = 0. 31 year
(8 x 5755)7(136, 000 - 10, 104) = 0. 36 year
(2 x 5755)7(6600 - 2526) = 2.9 years
(2 x 5755)7(21, 000 - 2526) = 0. 62 year
(2 x 5755)7(33,700 - 2526) = 0. 37 year
(4 x 5755)7(107,000 - 5052) = 0.22 year
(8 x 5755)7(181, 000 - 10, 104) = 0. 27 year
to
CO
Oi
-------
FEED
BASIC REAGENT
METERING PUMP
I SLUDGE
'pHRCW 1 CIRCULATION PUMP
HWS
HWR
NONLINEAR
Figure 77. Combined pH and temperature control of single-stage digester.
by increasing the operating temperature in order to increase the growth rate of
methane formers or by recycling more methane bacteria from the second-stage
separator.
If the dropoff in methane production occurred as a result of increased hydraulic feed
rate (resulting in bacteria washout), the automatic instrumentation can reduce or
stop the feed, in addition to the corrective actions specified above for toxic overloading.
The instantaneous response to organic overloading is an increase in methane production,
which lasts until the resulting low pH starts to inhibit the growth of methane formers.
Consequently, organic overloading is best controlled in a feedforward manner by
measuring the incoming flow and concentration and keeping its product nearly constant.
In the absence of such information, pH can be used as a fairly sensitive feedback
control to slow or stop the digester feed.
237
-------
Table 50. CAPITAL AND OPERATING COSTS OF pH AND
TEMPERATURE CONTROL
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
5 years depreciation period
pHE-6 (with ultrasonic cleaning)
pHT-7
10 years depreciation period
TRC-2
TRC-3
CV-4
pHRC-8
YP-9
15 years depreciation period
TT-1
TT-5
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
510
1.000
1,510
820
760
500
900
1.000
3,980
400
350
600
700
2,400
3,000
7,450
12,940
60
10
16
16
10
16
20
6
6
20
180 hours/year
Yearly cost of
replacement parts
is estimated as
$300
$2,100
238
-------
Table 51. COST/BENEFIT SUMMARY OF pH AND TEMPERATURE CONTROL OF DIGESTION
Digester
Type
Single
stage
Two stage
Plant
Size
(mgd)
1
5
10
50
100
1
5
10
50
100
Annualized
Instrument
Cost
($/year)
7,580
7,580
7,580
15,160
30, 320
7,580
7,580
7,580
15,160
30,320
Savings
Resulting From
Improved
Control ($/year)
9,650
31,400
50,600
161,000
272,000
13,200
42,000
67,400
214,000
362,000
Percent of
Total Savings
Spent on
Instrumentation
79.0
24.0
15.0
9.4
11.0
57.0
18.0
11.0
7.0
8.4
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
(2 x 12, 940)/(9650 - 7580) = 12. 0 years
(2 x 12, 940)/(31, 400 - 7580) = 1. 1 years
(2 x 12, 940)/(50, 600 - 7580) = 0.6 year
(4 x 12, 940)/(161, 000 - 15, 160) = 0. 35 year
(8 x 12, 940)7(272, 000 - 30, 320) = 0. 43 year
(2 x 12, 940)7(13, 200 - 7580) = 4.6 years
(2 x 12, 940)7(42, 000 - 7580) = 0. 75 year
(2 x 12, 940)7(67, 400 - 7580) = 0. 43 year
(4 x 12, 940)7(214, 000 - 15, 160) = 0. 26 year
(8 x 12, 940)7(362, 000 - 30, 320) = 0. 31 year
CO
-------
The control system that takes all three measurements into consideration is shown in
Figure 24 for a single-stage digester, and in Figure 25 for a two-stage digester. In
addition to the previously discussed components, this system contains a digester gas
composition analyzer (FID or NDIR) and a flow sensor that will yield information on
total methane production.
In case methane production drops, the operating temperature can be increased (a
single-stage unit), and/or additional methane-forming microorganisms can be recycled
(a two-stage unit).
Switches are provided on both pH and methane production to signal low values. These
switch actuations can be used automatically to slow or terminate digester feeding or to
alarm the operator.
The capital and maintenance costs are listed in Table 52. Two-stage digestion systems
require one additional device, the CV-20 valve (shown on Figure 25). The annualized
cost for a single unit, using an 8% interest rate for devices with 5 years lifespan and
6% for all others, as shown in Table 52, is $7052. Similarly, the two-stage annualized
cost is $7290. Using one of these systems for each digester (Table 17), the annualized
cost as a function of plant capacity is given in Table 53.
These figures are lower than the yearly savings projected for all plant sizes except
the single-stage digesters in 1 mgd plants (see the 25% and 30% columns in Table 46).
The annualized cost of the required instrumentation, expressed as a percentage of the
resulting annual savings, is shown in Table 53, together with the projected payback
periods.
240
-------
Table 52. CAPITAL AND OPERATING COSTS OF pH, TEMPERATURE,
AND METHANE CONTROL
Cost Components
5 years depreciation period
pHE-6 (with ultrasonic cleaning)
pHT-7
AE-12 (CH analyzer)
AT-13
10 years depreciation period
AE-12 (sampling system)
THC-2
TRC-3
CV-4
pHRC-8
YP-9
FY-11
FY-14
FRC-15
FSL-16
FAL-17
pHSL-18
pHAL-19
CV-20
15 years depreciation period
TT-1
TT-5
FT-10 (with orifice included)
Installation materials
Control panel section
Engineering and design
Installation and startup labor
Total installed cost per loop
Total annual operating cost per loop
Single Stage
Capital
Cost ($)
510
1,000
4,000
1,000
6,510
1,000
820
760
500
900
1,000
200
380
760
75
50
75
50
6,570
400
350
500
840
900
3,000
4.000
9,990
23,070
Req'd Maintenance
(hours/year)
60
10
50
10
16
16
10
16
20
5
10
16
2
2
2
2
6
6
10
30
299 hours/year
Yearly cost of
replacement
parts is esti-
mated as $510
3,500
Two Stage
Capital
Cost ($)
510
1,000
4,000
1,000
6,510
1,000
820
760
500
900
1,000
200
380
760
75
50
75
50
1.000
7,570
400
350
500
840
900
3,000
4.000
9,990
24, 070
Req'd Maintenance
(hours/year)
60
10
50
10
16
16
10
16
20
5
10
16
2
2
2
2
10
6
6
10
30
309 hours/year
Yearly cost of
replacement
parts is esti-
mated as $510
3,600
241
-------
Table 53. COST/BENEFIT SUMMARY OF pH, TEMPERATURE, AND
METHANE CONTROL OF DIGESTION
Digester
Type
Single
stage
Two stage
Plant
Size
(mgd)
1
5
10
50
100
1
5
10
50
100
Annual! zed
Instrument
Cost
($/year)
14, 104
14, 104
14, 104
28,208
56,416
14,580
14,580
14,580
29,160
58,320
Savings
Resulting From
Improved
Control ($/year)
12,500
39, 300
63,200
200,000
340,000
19,800
62,700
101,000
321,000
540,000
Percent of
Total Savings
Spent on
Instrumentation
100
36
22
14
17
74
23
14
9
11
Payback Period
(No. of Loops) (Installed Loop Cost)
(Yearly Savings— Yearly Costs)
Not applicable
(2 x 23, 070)/(39, 300 - 14, 104) =1.8 years
(2 x 23, 070)/(63, 200 - 14, 104) = 0. 94 year
(4 x 23,070)7(200,000 - 28,208) = 0.53 year
(8 x 23, 070)7(340, 000 - 56, 416) = 0. 65 year
(2 x 24, 070)7(19, 800 - 14, 580) - 9. 2 years
(2 x 24, 070)7(62, 700 - 14, 580) = 1.0 year
(2 x 24,070)7(101,000 - 14,580) = 0.56 year
(4 x 24, 070)7(321, 000 - 29, 160) = 0. 33 year
(8 x 24, 070)7(540, 000 - 58, 320) = 0. 40 year
IS
to
-------
FLOW-AUGMENTING EFFECT OF POLYMERS IN WASTEWATER SYSTEMS
Introduction
The purpose of this analysis is to quantitatively evaluate the potential benefits of flow
augmentation through the use of polymers in wastewater transportation systems and
then to compare these benefits to the costs of the required polymer addition system.
It has been reported (67) that the addition of polymers to the water flow in open
channels will reduce the friction, thereby increasing the velocity of flow. For the
purposes of this analysis, the polymer and the concentration to be used are assumed
to be Separan AP-30 and 25 ppm, respectively. It is also reported (66) that the flow
augmentation effect of this polymer at that concentration is a 25% increase in flow
velocity.
Basis of Evaluation
The effect of flow augmentation by adding polymers will be evaluated by analyzing its
influence on a typical sewer interceptor. The interceptor can be viewed as a "bottle-
neck" in a gravity flow collection system. It will be assumed that the distance
between the interceptor and the wastewater-treatment plant is 0.75 mile or 4000 feet.
Other parameters were assumed or based on information given in "Wastewater Engi-
neering" (67):
• Pipe diameter 72 inches
• Manning roughness (N) 0. 013
-4
• Slope 1.8x 10
• Design flow velocity 2 fps
• Design flow 58 cfs
The flood height of the .relief well is assumed to be 4 feet. Under storm conditions
that result in flooding the well, this extra 4 foot head adds to the energy gradient of
the system, resulting in a total maximum flow of 150 cfs.
243
-------
If the addition of 25 wppm of Separan AP-30 results in a 25% increase in flow, then
the total augmented maximum flow is 188 cfs. The gain of this 38 cfs equals the
capacity of 42" diameter pipe under flooded conditions. Therefore this cost/benefit
analysis will compare the costs of the:
i Construction of a new parallel main, 4000 feet long and 42" in diameter
• Total cost of the polymer addition system.
Assumptions
• The useful life of the 42" parallel main is assumed to be 30 years.
• The interest rate for municipal loans is assumed to be 6%.
• The 42" pipe is assumed to be maintenance free.
• The useful life of the polymer addition system components is in the 10 to
15 years range, as listed specifically in Table 54.
• It is assumed that the polymer addition system will require maintenance as
indicated in Table 54.
• It is assumed that flooding conditions exist 1% of the time, or 80 hours a year.
• It is assumed that a suitable building already exists near the interceptor to
house the equipment required for the polymer injection.
Cost of 42" Parallel Main
It is assumed that the 42" parallel main can be installed in a relatively unpopulated
area with no loss of efficiency due to the breaking of hardtop road surfaces or to the
interferences caused by traffic, etc. If these assumptions are not correct, the cost
of this system could easily be doubled.
• The material cost at $15/foot of 42" diameter concrete pipe is
4000 x 15 $ 60, 000
• The concrete pipe labor time required at three manhours (MH)/
foot of pipe is 12,000 MH. At $18/MH, this gives a total cost
of $216,000
244
-------
Table 54. CAPITAL AND OPERATING COSTS OF POLYMER
ADDITION SYSTEM
Cost Components
Capital Cost
($)
Req'd Maintenance
(hours/year)
10 years depreciating period
Feeder (F-l) for 500 Ibs/hour
capacity
Pump (P-l) for 200 gpm at 10 to 20
feet of head
On-off valve XCV-1
Five level switches with interlock
and time delay
15 years depreciation period
Bin (T-l), 1000 ft3 capacity
Mixing tank (T-2) provided with a
5 HP agitator (5000 gallon capacity)
Total installed capital cost
Total annual operating cost
8,000
5,000
720
1,500
15,220
20,000
40,000
60,000
75,220
20
10
12
30
50
122 hours/year
$1,220
The excavation and backfill labor cost is estimated in accordance
with Figure 78. The required 7200 MH at $18/MH will give a
total cost of $130,000
The cost of sand for backfill is estimated at $3/cubic yard for
a total quantity of 10,000 cubic yards
Therefore the total installed cost is
$ 30.000
$436,000
The annualized cost of the 42" parallel main is calculated on the basis of a 30 years
life span, 6% interest, and zero maintenance cost as:
annual cost = (436,000)
(0.06) (1.06)
30
(1.06) -1
30\
(436,000) (0.0728) = $31,700
245
-------
For self-sustaining soil-SK-A:
4w + (hw) + br
CY/LF =
27
For loose soil-SK-B:
(hw) + h2
CY/LF =
27
where:
CY = cubic yards of earth
LF = linear feet
• For single lines in a trench, W = outside dimension of line + 1' - 0".
• For multiple runs in a trench, W = sum of outside dimensions of lines + No. of runs x 0' • 6".
• Minimum W = 7-0".
• Allow % CY concrete for each anchor.
• Show calculation for W and CY/LF below or on a supplementary sheet and identify the
cross section calculated by number.
For purposes of this example, let:
IOxB+102
27
= 6CY/FT
Then 4000 LF requires 24,000 CY of excavation and backfill.
Figure 78. Underground pipe excavation, backfill, anchors.
246
-------
Cost of Polymer Addition System
It is assumed that polymer is added only when flooding conditions occur and that it is
added at the fixed rate required to produce a 25 mg/1 concentration in a 188 cfs flow
f*
(188 x 62.4 x 25/10 -0.3 Ib/second = 1050 Ib/hour.
It is also assumed that a suitable building already exists to house the equipment
required for the polymer injection.
Assuming that polymer will be used for 80 hours each year and that its cost is $0.25/
Ib, the total yearly polymer cost is: Cp = 1050 x 80 x 0.25 = $21,000/year.
The metering pump (P-l) is sized to deliver 1050 pounds of polymer per hour. This
corresponds to about 210 gpm when handling 1% concentration liquid. The polymer
storage bin (t-1) is sized to hold a 3 month supply of additive or 20 x 1050 = 21,000
3 3
Ibs. Assuming a 30 Ib/ft polymer density, a 1000 ft bin will be more than adequate.
(see Figure 79.) The mix and feed tank (t-2) is sized to hold a 15 minute supply of
1% polymer solution. A 5000 gallon tank is sufficient for this purpose [(100) (1050)/
(4) (8.3) = 3150J .
The tapwater charge valve (XCV-1) and the polymer feeder (F-l)are both opened by
the low-level switch (LSL-1) and are closed by the high-level switch (LSH-2). XCV-1
is sized to pass 100 gpm (this is equivalent to 50,000 Ibs/hour). In order to arrive at
a 1% solution concentration, 500 Ibs of polymer needs to be added every hour. Conse-
3
quently, the feeder would be set to deliver approximately 10 Ibs or 0. 3 ft of polymer/
minute. The mixing tank agitator is also started by LSL-1 and is stopped by LSH-2,
but only after a 10 minute time delay. LSH-3 starts and stops P-l while LSHH-4 and
LSHH-5 actuate alarm points in the treatment plant's control room.
Table 54 listed the various applicable cost elements (52, 57). These are added to the
yearly polymer cost given previously to arrive at the total annualized cost:
247
-------
LOCATED ON J
TREATMENT PLANT _f
PANEL ~\
T-1
POLYMER
STORAGE
BIN
(1000 FEET3)
SIZED TO PASS
100GPM
T-2
MIX AND FEED
TANK
(5000 GALLONS)
CHARGING
PUMP OF 200 GPM
CAPACITY
P-1
GRAVITY FLOW
TAP
WATER
Figure 79. Polymer addition system for flow augmentation.
total annual cost = 15,220 f (°' °6) ^° \+ 60,000 /<°-06> (W*
(1.06) -1
= $30,490
Conclusions
The annualized cost of installing a 42" parallel main is $31,700 and the annualized
cost of a polymer addition in order to achieve the same result is $30,490. On this
basis it can be concluded that, if the various assumptions made in this analysis are
applicable, the flow-augmenting effect of a polymer addition can be cost competitive
with the installation of a new main.
248
-------
Favorable to the flow augmentation technique is the consideration that the required
system can be quickly installed and that its installation would not interfere with
traffic or with other aspects of the daily life of a community. The main disadvantage
of this approach is the substantial use of polymer, which in itself is a limited resource.
It does appear that the polymer addition strategy, at best, should be considered only
as a temporary solution for application wherever the construction of parallel mains
would take too long. The polymer charging equipment can be built into a skid so that
it will be portable and reusable elsewhere, once it has fulfilled its original function
in a particular location.
249
-------
SECTION IX
MODERN CONTROL SYSTEMS
ADVANCED CONTROL TECHNOLOGY APPLIED TO WATER AND WASTEWATER
TREATMENT
Introduction
The two forms of control used almost exclusively in water- and waste-treatment plants
are dosage control and simple feedback regulation. Dosage control (also known as
pacing, flow proportioning, or ratio control)-is used when on-line measurements of the
controlled variables are unavailable. An example of dosage control would be the feed-
ing of flocculant in proportion to the flow of the waste stream in order to regulate efflu-
ent turbidity. If the quality of the wastewater and the efficiency of the clarifier are
reasonably uniform, the effluent turbidity should not change appreciably.
When the controlled variable can be automatically measured, simple feedback control
is usually employed. This would be the case wherever the controlled variable is liquid
level, pH, residual chlorine, or the like. Although feedback control can provide the
precision lacking in dosage control, it also introduces some problems. If the gain of
the feedback control loop becomes too high, oscillations will develop; sustained cycling
will result in off-specification effluent and can consume chemicals out of proportion to
the plant requirements. If the control-loop gain is too low, variations in wastewater
flow or composition can cause the controlled variable to deviate substantially from
acceptable limits.
The simple feedback controller can be adjusted for only one set of plant conditions so,
if the plant or waste characteristics change appreciably, too high or too low a loop gain
250
-------
could result because of nonlinear input/output relationships. Occasionally, the pro-
cess is particularly difficult to control because the controlled variable is very sensi-
tive to corrective action (as in pH neutralization), or very slow to respond (due to long
sampling delays, poor mixing, etc.), or both. In these cases, even if the plant char-
acteristics are constant, moderate variations in wastewater flow and composition can
produce an off-specification effluent.
Some industrial treatment plants have an influent equalization basin of several hours
capacity which can minimize input variations and thereby facilitate control. But in
places where such a capacity is unavailable or the cost is prohibitively high, feedfor-
ward control may be the only practical alternative.
This section will discuss the applications of feedforward control to minimize the effects
of these influent disturbances, and of adaptive control to compensate for parametric
variations within the plant. For a detailed discussion of the theory behind these
methods, the reader is referred to work of Shinskey (15, 22). The use of digital com-
puters to implement these and other control methods will then be examined.
Feedforward Control
Feedforward control can be defined as control in which information concerning one or
more conditions that can disturb the controlled variable is converted into corrective
action to minimize deviations of the controlled variable. Inasmuch as water flow rate
is one of the above conditions, dosage control systems that use water flow to set chem-
ical flow are, in fact, feedforward systems. Whether subsequent deviations in the con-
trolled variable (i. e., composition) are continuously recorded or only observed from
occasional laboratory analyses is beside the point.
Dosage control systems may take two forms. In the first, the dosage is set but not
measured or recorded. In the second, the dosage is calculated from the measure-
ments of two flow rates and, as such, may be controlled and recorded. The calculation
251
-------
of dosage requires a measurement of the manipulated flow, which is not required by
the first system.
Figure 80 shows two dosage control systems of the first type. Since the flow of
reagent A is the product of stroke and speed, the stroke setting determines the dosage.
Reagent B is controlled by a ratio controller—this device multiplies the water flow
signal by the setting on the ratio dial to generate the flow setpoint. Since the flow
measurements are in the squared form, the dial must have a square-root scale. Fig-
ure 81 shows a calculation of the ratio of bandwidth being recorded and controlled;
here again, square-root scales are required. Any familiar units may be used, such
as mg/1 or Ib/million gallons.
Occasionally the measurement of the water flow will be separated from the point of the
reagent addition by an overflowing vessel. Figure 82 illustrates two such possibilities.
If the inflow to the first vessel is measured with chlorine being added to its overflow,
the rate of overflow will lag behind the measurement. In this case, a compensating
lag is required in the feedforward path. Its time constant should be set to equal the
volume change from zero to maximum flow, divided by maximum flow.
If, however, the only measurement available is at the vessel exit, the flow at the point
of injection will lead the measurement. Lead-lag compensation is then required.
Although a pure lead would be ideal, such a device cannot be made since the ideal
function requires infinite gain. The maximum gain available in a real compensator is
the lead-to-lag ratio. The lead time constant should be set to match the change in
vessel volume from zero to maximum flow divided by maximum flow. The lag time
should then be reduced to the lowest value attainable which does not amplify the flow
noise excessively.
In actual practice, the compensating time constant varies with the overflowing volume
and overflowing rate rather than the maximum values that were selected for setting the
compensators. If the overflowing volume varied linearly with the flow, the maximum
values would be valid for all flow rates, but such is not the case. It is possible to use
252
-------
SPEED
STROKE
REAGENT B
METERING PUMP
WATER
SETPOINT
Figure 80. Stroke of metering pump sets dosage of reagent A, while
ratio controller for reagent B has a calibrated ratio or
dosage dial.
(B/W)2 MEAS/j^\ REMOTE SET
.- dp
REAGENT B
WATER
Figure 81. Dosage calculated by divider may be recorded and is easily
set from a remote source such as a digital computer.
253
-------
LEAD
LAG
CI2 GAS »•
^ 1
CHLORINATOR
J
CUSOLN rt
(d
(DIFFUSER
\
p)
WEIR
"X^-v^v.^.^^^
Figure 82. Dynamic compensation is required if actual flow at injection
point differs from measured flow in time.
actual measurements of volume and flow to manipulate the settings adaptively but, in
most cases, the improvement realizeable is not worth the cost or effort. A number
of techniques are available for linearizing the head measurement across the weir for
conversion to a usable flow signal (49).
If the particular component requiring treatment in the influent water is measurable,
the composition may be used to adjust the dosage. As an example, the alum dosage
for phosphorous removal can be based on the amount of orthophosphate in the influent.
In most cases, however, the composition measurement is unavailable, nonspecific,
insufficient, or too unresponsive to be meaningful. An example of nonspecific mea-
surement is pH. A neutralization process must have base-adjusted to total acid in the
influent, but pH indicates only the ionized acid. Consequently, feedforward systems
using influent pH measurements have not been as successful as one would like. How-
ever, Shinskey (19) reports on the performance of a neutralization facility where efflu-
ent specifications could not have been met without feedforward control from the influent
pH.
Referring back to the phosphorus removal process, phosphorus loading alone does not
determine alum dosage requirements. Alkalinity, other forms of phosphorus, and
254
-------
solids content also have their effect so, even if adequate orthophosphate measurements
were available, dosage still could not be adjusted with absolute certainty.
If a sample of the influent must be removed for treatment prior to analysis, the result-
ing signal will lag behind true process conditions. Thus, a total acidity measurement
from a titrated sample would not be useful in most cases for feedforward pH control,
because the effluent pH electrodes would be able to respond before a change in influent
acidity would be indicated by the titrator.
Figure 83 illustrates a feedforward system using feed composition information. Two
precautions are worth noting:
• Dynamic compensation may be required for composition even if not for flow.
» There must be a provision for operating the system without the analyzer in
service.
The latter is particularly important since analyzer reliability is not high, and the sys-
tem cannot function with a zero input.
Both of these requirements can be provided by the controller shown in Figure 83. By
connecting its output back to its measurement input, the output will follow the setpoint
in the steady state. Furthermore, the dynamic response can be varied by selecting
appropriate settings for the controller modes. Finally, upon analyzer failure the con-
troller can be placed in manual, holding its output at the last valid measurement.
Dynamic compensation will differ from that used in the flow signal path because the
composition of the entire vessel is affected—not only the overflowing portion. Con-
sequently, the lag will usually be much longer than that applied to the flow signal.
If an on-line analysis of the water quality is available, it will be more usefully applied
to the effluent than the influent. For example, it would be impossible to meet effluent
pH specifications when using only influent flow and pH for control. A pure feedforward
system such as this is not capable of the extreme accuracy required.
255
-------
MEAS
SPEED
STROKE
REAGENT
WATER
Figure 83. Water composition signal needs dynamic compensation and
protection against analyzer failure.
When drawing water from a large source, the flow rate may change rapidly when valves
are moved, but influent composition is usually slow to change. Consequently, feed-
forward control from flow may be necessary while a composition input is not, if an
effluent analysis is available for feedback. This logic breaks down to some extent
when considering the specialized case of neutralization, however, since influent pH
can change rapidly from spills and rinsings. In any case, an effluent quality measure-
ment does reveal the performance of the facility, while the influent composition does
not.
Feedback trim from effluent quality is typically applied as shown in Figure 84. If the
reagent flow is not transmitted, as with a metering pump or chlorinator, the upper
arrangement must be used. If the measurement is available, either configuration is
acceptable.
Adaptive Control
Adaptive control can be defined as a control action whereby automatic means are used
to change the type or Influence (or both) of control parameters in such a way as to improve
the performance of the control system. This definition includes a self-adjusting control
256
-------
AC W EFFLUENT
QUALITY
j DOSAGE
REAGENT
WATER
HOW "^ SETPO.NT
A) NONCALCULATED DOSAGE CONTROL
WATER ^
FLOW
T
SET /ACV, EFFLUENT
DOSAGEV y QUALITY
I
! t
REAGENT REAGENT
FLOW VALVE
OR PUMP
B) CALCULATED DOSAGE CONTROL
Figure 84. Feedback trim may be accomplished in two ways, depending
on whether dosage is calculated.
system, as well as any form of automatic adjustment. The term "adaptive" connotes
the control system conforming to the changing characteristics of the process being
controlled. By "parameters" is meant the adjustments normally made by hand (e. g.,
proportional, integral, derivative, lag, gain, etc.).
Adaptive control has not been applied extensively because of the difficulty in adjusting
the controller settings by means of a signal. Pneumatic control modes are developed
mechanically, with a lever for the proportional band and re stricter s for reset and
derivative. Automating these adjustments is too clumsy and difficult to be worthwhile,
although pneumatic multipliers have been used to adjust loop gain in certain applica-
tions.
The mode settings of a typical electronic controller are also introduced mechanically,
as the shaft rotation of a logarithmic potentiometer. Shinskey (19) cites an adaptation
performed with a nonlinear electronic controller, which is achieved by varying the
257
-------
width of its gap from a remote signal. This has an effect similar to a proportional
band adjustment and happens to be ideally suited to pH control situations. It was used
in a pH control loop to cancel the effect of the variable gain of equal-percentage valves
that were needed to manipulate the reagent flow over a very wide range.
Electronic controllers are now available with modes that can all be adjusted remotely.
Although intended primarily to allow a front-of-panel adjustment of a rack-mounted
controller, the modes can be set from any 0 to 10 volt signals. Proportional band,
reset time, and derivative time are inversely related to the applied voltage. Remote
mode adjustment has always been possible with digital computer control. Consequently,
most of the research and the few plant installations of adaptive systems have been con-
fined to digital computers.
For many processes, enough information is already available to program the adaption.
If the period of a loop varies inversely with flow, for example, proportional, reset,
and derivative should all vary inversely with flow. This can be called feedforward
adaptation, since the settings are calculated directly from a measurement of the vari-
able that alters the process characteristics. Varying the bandwidth of the nonlinear
controller as a function of valve position is another example of feedforward adaptation,
since the gain of the loop is known to be affected by the valve position in a predictable
manner.
A feedforward control system tries to hold a controlled variable at a desired value by
balancing the manipulated variable against a measurement of the load on the process.
How successfully it achieves this objective depends on how accurately the required
value of the manipulated variable can be calculated, based on the available information
on the load and its effect on the process.
Any errors in this calculation result in an imbalance, and hence a deviation, of the
controlled variable from its setpoint. This deviation is removed by a feedback con-
troller, which adjusts the feedforward calculation as necessary to compensate for
258
-------
whatever condition may have caused the error. In essence, the feedback controller
has adapted the gain of the feedforward loop to match the new process gain.
The top of Figure 84 shows how feedforward and feedback are combined to their mutual
benefit in controlling a flowing process. The output of the composition controller
adjusts the reagent-to-water ratio to compensate for variations in inlet composition,
side reactions, losses, etc. The multiplier also allows the gain of the feedback con-
trol loop to be changed directly proportional to water flow, thus achieving a feedfor-
ward adaptation of that feedback mode.
In some installations, there is no way of knowing the process gains and measuring
those factors that affect it. The most prominent example is the control of pH of plant
effluents comprised of a multiplicity of wastes with differing buffer characteristics.
The shape of the effluent titration curve may change continuously and randomly. The
pH controller can be adjusted only for the conditions that prevail at the time of adjust-
ment; a subsequent reduction in buffering will cause oscillation.
Since oscillations are undesirable, particularly if the controller is alternately adding
acidic and basic reagents, the controller must be detuned. The price paid for the
absence of oscillations is unresponsive control with all but the least buffered solutions.
In addition to being able to adjust the control modes from a signal, feedback adaptation
requires some means of discerning whether an adjustment is needed. If a loop is in the
steady state with no deviation, there is no way of knowing its gain. It has to be in a
transient or oscillatory condition before its gain and period are revealed.
Most feedback adaptation schemes mimic what an engineer would do in tuning a con-
troller. They introduce step changes in setpoint or controller output, examine the
resulting closed-loop response curve, and introduce appropriate changes in settings.
But where an engineer may adjust a given controller only once, an adaptive system
must function continuously (or at least periodically) if the process characteristics are
changing continuously. Needless to say, periodic disturbances intentionally introduced
to test the loop are generally undesirable.
259
-------
An effective adaptive scheme should maintain the loop in an acceptable adjustment by
using the normal upsets existing in the process. Even then, two or three oscillation
cycles are required before the adjustment is complete, just as when an engineer tunes
a controller. In addition, the adaptive device itself needs to be adjusted to avoid insta-
bility within its own feedback loop. As desirable as automatic tuning seems to be, it is
by no means a panacea. It ought to be reserved for those control loops that can't be
stabilized in any other way.
Shinskey (20) gives an example of a case history of a self-adjusting pH control system.
By eliminating extended intervals of cycling without sacrificing responsive effluent
control, the usage of lime and acid reagents was cut in half. This system is recom-
mended whenever variable buffering is encountered, but especially when both acid and
basic reagents are used.
On further work with the system (23), the gap width of a nonlinear controller was
adapted to compensate for variations in the titration curve. A discriminator circuit
compared the frequency of the pH deviation from the setpoint to an adjustable "cross-
over" frequency. If the frequency of the deviation was higher, the discriminator devel-
oped a positive error signal to expand the gap; if lower, a negative error was gener-
ated to close the gap. The crossover frequency was set so that the natural frequency
of oscillation would cause the gap to expand. The rate of expansion or contraction was
proportional to the pH deviation.
CENTRALIZED CONTROL
Introduction
For the purposes of this discussion, centralized control will be defined as a control
system that allows the plant operator to observe and control the plant operation from
a remotely located control room. In this control room the main operating parameters
are displayed, together with information on emergency or alarm conditions. Provided
260
-------
for the operator's use are: 1) regular analog controllers, which allow him to vary
such operating parameters as flows, temperatures, levels, and concentrations, and
2) on-off pushbutton-type remote controls, which allow him to start or stop pumps,
open or close valves, etc. The plant's overall status can be quickly observed from
the semigraphic panel sections, which show (by steady lights) the valves that are open
and the motors that are running. They also indicate (by flashing lights) any operating
parameters (e.g., levels, pH, etc.) that are abnormal (see Figures 8 5 and 86). Visual
observation of some unit operations can also be achieved remotely by switching a closed-
circuit television display to the corresponding channel. It is also assumed that the
analytical laboratory used for grab-sample evaluation (wherever automatic detectors
are not yet available) is located in the same building as the control room.
The benefits of central control are many but, in terms of a cost/benefit analysis,
labor savings are the most important. In order to provide some quantitative data, the
Figure 85. Typical semigraphic display panel in a central control room.
261
-------
SEMIGRAPHIC
SECTION
AS REQUIRED
MULTIPLE PUSHBUTTONS
IN ONE ACTUATOR
LEGENDS ON EMERGENCY
PUSHBUTTONS EXIT ONLY
SEMIGRAPHIC
OFFICES, WASHROOMS, LOCKERS,
LABORATORY.LAVATORY.ETC.
3'0"
5'6"
MN
. * L.
<
<\
«2M
X
>AV PANEL FRAMEW
^ COMPUTER ACCE
OPERATOR'S DE
OR TEMPERATURE (
I
FRONT
r»P
PANEL £7
AUXILIARY
COMPONENTS
CIRCUIT
DISTRIBUTION
PANEL
2'6"
3 TO"
\
AIR SUPPLY
TIE-IN
WATER COOLER
Figure 86. Control room layout.
262
-------
following operating labor estimates illustrate the potential savings of centralized
control:
Type of Control System
Local control
Central control
Centralized computer control
Operating Costs Spent on Labor
70%
50%
40%
Table 55 shows the potential total savings as a function of plant size.
The economic justification of centralized control will be developed on the basis of the
following assumptions:
• A central building is not available to house the control panel, and its cost at
$35/foot^ should be included as part of the total cost of centralized control.
• The instrumentation is electronic, and both the signal and 110 volt on-off con-
trol leads will be carried in armored conduits.
• The local control system already includes the various transmitters and con-
trollers and, therefore, the cost differential for centralized control should
consider only the extra cost of increased transmission distance, not the cost
of the instruments.
Table 55. POTENTIAL YEARLY SAVINGS FROM CENTRALIZED
COMPUTER CONTROL
Plant Size
(mgd)
1
5
10
50
100
Centralized Control
($/yr)
12,000
40,000
67,000
220, 000
400, 000
Centralized Computer
Control
($/yr)
18, 000
60,000
100, 000
330,000
600, 000
263
-------
• The cost of the nongraphic portion of the central panel is offset by the local
panels and support frames needed for the local control strategy.
On the basis of these assumptions, the cost of centralized control will include three
main contributing elements: 1) the building cost (given in Table 56), 2) the installed
cost of added transmission leads (listed in Table 57), and 3) the cost of a semigraphic
section and some miscellaneous alarm, pushbutton, and display devices (outlined in
Table 58).
In Table 59, both the savings and costs are summarized and are expressed in both
percentage and payback-period terms. Based on this evaluation, centralized control
can be justified for all plant sizes.
COMPUTER CONTROL
Introduction
The multipliers and dividers, leads and lags, etc. (shown in Figures 80 to 84), are
all computing functions. These functions may be performed (along with feedback con-
trol) either with dedicated analog instruments or by a digital computer. This section
will discuss the implementation of advanced control strategies by a digital computer.
Since a digital computer represents a substantial investment, it cannot be easily justi-
fied if called on to perform only a few tasks. This makes justification difficult for a
waste-treatment facility requiring control over only a few variables.
Furthermore, there is little incentive to produce a better quality effluent since it
ordinarily has no inherent value. However, if the treated water is to be reused, then
the added precision and flexibility of the digital computer may be warranted. This
would certainly be the case for a large-scale water-treatment plant, spread over many
acres.
264
-------
Table 56. CONTROL BUILDING COST*
Plant Size
(mgd)
1
5
10
50
100
Control
Panel
Length
(ft)
10
25
30
50
60
Control
Room Size
(ft x ft)
15 x 25
20 x40
30 x40
40 x 50
40 x 60
Control
Bldg Area
(ft2)
375
800
1200
2000
2400
Total
Cost
($)
13,100
28,000
42,000
70,000
84, 000
Annual
Cost
($)
1,020
2,200
3,300
5,500
6,600
*A depreciation period of 25 years and an interest rate of 6% are assumed. The
annual cost is calculated as:
total cost
(0.06) (1.06)
(1.06)25 - 1
25
= 0.078 (total cost)
Table 57. SIGNAL AND CONTROL WIRE TRANSMISSION COSTS*!
Plant
Size
(mgd)
1
5
10
50
100
Average
Assumed
Transmission
Distance
(ft)
150
250
300
500
600
No. of Wire Pairs
DC Signal
(Transmitters,
Controllers)
15
25
40
80
160
110 Volt
On-Off
(Pushbuttons,
Alarms, etc.)
40
60
100
200
400
Total
Transmission
Wiring
Requirement
(No. xft)
55 x 150
85 x 250
140 x 300
280 x 500
560 x 600
Total
Installed
Transmission
Cost
($)
1,500
3,850
7,600
25,000
60, 000
Annual
Cost
($)
118
300
600
1,950
4, 700
*A depreciation period of 25 years and an interest rate of 6% are assumed. The annual cost = 0.078
(total cost).
'The installed cost of a 3/4-inch conduit with eight pairs of No. 14 wire is estimated as $1.50/foot.
Each foot of required wire pair, therefore, is estimated as
265
-------
Table 58. SEMIGRAPHIC PANEL AND DISPLAY COSTS* t
Plant
Size
(mgd)
1
5
10
50
100
Length of
Semigraphic
Section
Required
(ft)
5
12
15
25
30
No. of Lights,
Pushbuttons,
Alarms, etc.
30
45
75
150
300
Total Cost of
Telephone and
Closed-Circuit
TV Network
/d>\
(•P)
2,000
3,000
5,000
7,000
10,000
Total
Cost
($)
7,900
16, 350
22,250
36,500
49,000
Annual
Cost
($)
1,070
2,220
3,030
4,960
6,650
The life of this material is assumed to be 10 years, and the corresponding
interest rate is assumed to be 6%. The annual cost = 0.136 (total cost).
'The semigraphic cost is assumed to be $1000/foot, and the installed unit cost
of the pushbutton-type devices is assumed to be $30 each.
Table 59. COST/BENEFIT SUMMARY OF CENTRALIZED CONTROL
Plant
Size
(mgd)
1
5
10
50
100
Total
Projected
Savings
($/yr)
12, 000
40, 000
67, 000
220, 000
400, 000
Total
Projected
Costs
($/yr)
2,208
4,720
6,930
12,410
17,950
Percent of
Savings
Used to
Cover
Expenses
19.0
12.0
10.0
6.0
4.5
Payback Period [Based on Totals From
Tables 56, 57 and 58
(Savings and Costs)]
22, 500/(12, 000 - 2, 208) = 2.2 years
48,200/(40,000 - 4,720) = 1.4 years
71, 850/(67, 000 - 6, 930) = 1.2 years
131,5007(220,000 - 12,410) = 0.6 year
193, 0007(400, 000 - 17, 950) = 0.5 year
266
-------
Direct Digital Control
Direct digital control (DDC) may be defined'as control action in which control is per-
formed by a digital device which establishes the signal to the final controlling element.
Historically, DDC was thought to be the most justifiable function of a process com-
puter, and payout was expected simply by replacing analog controllers. In actual
practice, DDC was found to exceed the cost of analog control, due principally to the
cost of input/output devices and the control stations necessary for backup.
Whether the computer is capable of better control than analog instruments depends on
the requirements of the process being controlled. Variables that respond rapidly to
manipulation (such as flow, pressure, and liquid level) can usually be controlled better
with analog instruments. The reason for this is that the computer does not control
continuously but intermittently—at intervals of 1 second or more. The actual scan
interval depends not only on access time, but also on the number of tasks assigned to
the computer. Analog controllers typically can respond in 0.1 second.
Backup is also a problem with DDC. A computer failure can mean the loss of control
throughout the plant. But if the faster loops are controlled by analog instruments, the
operators can control the slower ones manually until service is restored. To avoid a
complete loss of control over critical variables when the computer is out of service,
backup systems like the one shown in Figure 87 can be used. Residual chlorine con-
trol is performed in the computer, but chlorine flow is manipulated to meet the desired
dosage through analog instruments. Computer outage will cause the computer/manual
(C/M) station to transfer to manual, fixing the dosage at its last set value. At any
time the operator may adjust dosage from this station. He may also adjust it from the
computer console by placing the primary control function analyzer/computer (AC) to
manual. Since the computer console may be located some distance from the analog
controls, this provision is important.
When fall adaptation, or control over processes with relatively long or variable dead-
time is required, the computer is capable of superior performance. It should, therefore,
267
-------
COMPUTER
CONTROL
Figure 87. This computer control system fails to provide dosage control.
be saved for these more difficult and higher level operations. In addition to the con-
ventional control modes of proportional, integral, and derivative, and the adaptation
thereof, the computer is capable of an endless variety of algorithms. With processes
dominated by deadtime, keying the scan interval to the deadtime can result in a sub-
stantial improvement over faster scan. Various nonlinear modes (such as error-
squared) are also available. In many systems, the engineer is free to compose his
own control programs instead of being forced to draw from the conventional ones.
Digital Supervisory Control
Digital supervisory control can be defined as control action in which the control loops
operate independently subject to intermittent corrective action; e. g., setpoint changes
from an external source. Under the present context, the "external source" is a digital
computer, although it also could be simply a remote console supervised by the operator.
Supervisory control systems differ operationally from DDC systems in that the com-
puter manipulates the setpoint of an analog controller. Figure 84 will serve to compare
268
-------
the two. The DDC system with fallback to dosage control applies to the upper sche-
matic in Figure 84—only the C/M station is missing. The lower schematic could be a
digital supervisory system if the quality control function (AC) were performed in the
computer. This system also has the capability of fallback to dosage control because
of the dosage calculation made by the analog divider. A local setting of the dosage is
available at the dosage controller or at the computer, without the additional C/M sta-
tion required in the DDC system. However, the dosage controller is not a conventional
analog controller, with the facility for a computer-driven setpoint. It also contains
logic to transfer to local set upon computer failure and to signal the computer when-
ever the controller is placed in local set or manual operation.
Computer Justification
Obviously any in-depth discussions of the capabilities and potentials of computer con-
trol would not be feasible within this presentation. However, some general comments
on the right reasons and the wrong reasons for installing a computer can be included.
One of the right reasons is the need for solving complex mathematical equations that
a human operator could not handle (for example, the maintenance of the optimum food-
to-microorganism ratio in activated sludge or digester systems). Another good reason
to have a computer is the need for memory and for performing a large number of simple
tasks within a short period of time (for example, the regulation of storm-water treat-
ment as a function of several alternate strategies that are stored in the computer
memory and automatically selected for use by the computer as a function of the rate,
distribution, and duration of the storm). Other tasks that can be given to the computer
on a low-priority basis include effluent-monitor ing data storage, recordkeeping of all
types, inventory control, preventive-maintenance scheduling, and any other adminis-
trative and bookkeeping-type function.
One of the wrong reasons for the installation of a computer is to use it as an expensive
status symbol. Another is to use it as a replacement for conventional instrumentation
without improving the operation.
269
-------
Wherever a computer control system seems to provide the appropriate solution to a
process problem, a choice must then be made between packaged systems for standard
applications and custom systems for special requirements.
Packaged systems consist of pre-engineered assemblies of computer hardware and
software with suitable sensing actuating, as well as display devices that perform some
well-defined and commonly practiced functions. Packaged systems are available for
data logging and flow control tasks. Although developed and produced for a fairly wide
market, packaged systems do have some flexibility for adapting to process variations
and the like. . To protect their proprietary interests, many suppliers furnish a "User's
Manual" without any program documentation. Packaged systems sell for a clearly
stated price and are ideal for users who want the functions that this type of system
performs. They are not for users with serious and specialized process control prob-
lems, or for those who need a large general-purpose computer system.
For the user who has unique problems that are not addressed by packaged systems or
who needs a general-purpose computer system, the solution is to install a custom sys-
tem. Although a custom system is a unique assembly of equipment and computer pro-
grams, its component pieces (especially routine for data acquisition, limit checking,
and alarming) are standard. Because of its novel design, a custom system costs more
than a packaged system of comparable scope.
The potential yearly savings that can result from computer control has been listed in
Table 55. However, the reduction in labor requirements relative to centralized com-
puter control is not very substantial because some of the operating labor requirements
that are eliminated by the computer are replaced by an increased need for scheduled
preventive maintenance.
Present-day minicomputers are defined more by their price (i. e., less than $20,000)
than by their performance. However, prospective users should not be misled by the
low cost of the central processor.
270
-------
Because of the need for special wiring practices and for software development in addi-
tion to the hardware costs, even a small digital computer installation is likely to cost
$200,000. On the other hand, because many of the algorithms are repetitious and
because the size of the computer memory is not the main cost item, even a large com-
puter installation with several CRT displays will not cost more than $800,000. There-
fore, the installed cost of central computer control can be taken as the following for
various-sized plants:
1 mgd $200,000
5 mgd $300,000
10 mgd $400,000
50 mgd $700,000
100 mgd $800,000
Based on 10 years depreciation and 6% interest, the total yearly costs are:
1 mgd $ 27,200
5 mgd $ 40,900
10 mgd $ 54,500
50 mgd $ 95,000
100 mgd $109,000
Adding these costs to the annual cost of centralized control (Table 59) gives the total
cost of centralized computer control. These costs are compared to the projected
savings in Table 60 to show the economic justification and payback periods. It is
apparent that computer control cannot be justified unless a plant is larger than 10 mgd.
Although computer control of wastewater-treatment plants is still in its infancy, com-
puter control has become well established in industrial processing. But computer con-
trol costs vary widely in both industrial and wastewater processing. The benefits are
derived from well-recognized sources, but are not easily predicted. However, as
computers become used more frequently in wastewater-treatment projects, an histor-
ical background will be built up, and the cost benefits of computer control can then be
more quantitatively investigated.
271
-------
Table 60. COST/BENEFIT SUMMARY OF CENTRALIZED
COMPUTER CONTROL
Plant
Size
(mgd)
1
5
10
50
100
Total
Projected
Savings
(Table 55)
($/yr)
18,000
60, 000
100,000
330,000
600, 000
Total
Projected
Costs
<$/yr)
29,408
45, 620
61,430
107,410
126,950
Percent of
Savings
Used to
Cover
Expenses
>100
76
62
33
21
Payback Period
(yrs)
Not applicable
Excessive
471, 850/(100, 000-61, 430) = 12 years
831, 500/(330, 000-107, 410) =3. 7 years
993, 000/(600, 000-126, 950) =2.1 years
272
-------
SECTION X
INSTRUMENTATION LAYOUTS
INTRODUCTION
The treatment processes discussed in this section and illustrated in Figures 88 to 91
are typical of currently used instrument designs that may not be entirely suitable for
any one particular application. These instrument drawings illustrate current practices
in the selection and application of instrumentation to wastewater-treatment facilities.
The engineering firm of C.E. Maguire, Inc. (located in Waltham, Mass.) was com-
missioned to produce these instrument designs and the following narrative descriptions.
All instrumentation shown was selected deliberately to perform usefully and reliably
with minimum attention. The 1 mgd plants were sparingly instrumented, since it was
assumed that the usual minimum team of operators would have little difficulty in keep-
ing the various processes within allowable limits. The 10 mgd plants are much more
extensively instrumented, and with the intent of producing useful, reliable information
as one of the primary requirements. In all cases, the analyst's role in making peri-
odic and continuing determinations of substrate and solids concentrations and the like
is assumed. Completing a typical daily report form is also assumed to be one of the
operator's basic duties; consequently, all practical instrumentation to give him the
necessary data has been provided.
The existence of a central control room was also assumed, together with the implica-
tion that all other field instrumentation would have to perform satisfactorily under
corrosive, wet, and often explosive conditions. Recording was separated from the
control function, and all signals were assumed to be standard and analog (usually 4 to
20 ma dc or 3 to 15 psi). Separate single-channel recorders are satisfactory for
273
-------
small installations but, wherever there are more than a few recorded or controlled
functions, multi-channel recorders are preferable since they show reliably the relative
change of variables with time, and with each other. (This is not usually achieved with
the customary single-channel recorders.)
RECOMMENDED INSTRUMENTATION
Wherever alarms are shown, highly reliable and industrially proven annunciators are
implied. Moreover, their proper utilization is assumed so that the operator is warned
when—and only when—a potentially dangerous or costly situation develops.
Level measurement, particularly of wastewater, is most reliably determined by
bubble tubes, fitted with constant-flow air supply controllers. Level difference is
usually obtained from two bubblers, using a force-balance, non-over-rangeable device
of the differential pressure type. The level in digesters and other closed vessels can
be reliably measured by force-balance diaphragm-type instruments, but the pressure
within the vessel cover must also be considered. Proven optical detectors are avail-
able for detecting sludge blanket levels, and sonic devices may also be practical.
Large flows of wastewater are measurable by Parshall flumes, Venturis, or magnetic
flowmeters. The flumes are simplest and most trouble-free, the Venturis must use
water-purged connections, and so the magnetic meter is somewhat more practical,
although provision must be made for electrode maintenance. Orifices and flow tubes
are suitable for air and gas flows, while direct and bypass rotameters should be con-
sidered for a flow indication of clean materials.
The radiation density meter remains the best automatic instrument for primary and
thickened sludges, but installations should be designed carefully for instrument cali-
bration, servicing,. and removal. Density can always be determined rapidly by an
operator, using a hydrometer or beam balance and a grab sample; in some cases,
this may be the best answer.
274
-------
The temperature of wastewater is best measured by a platinum resistance bulb or a
well-constructed linearized thermistor assembly. The installation of a quality system
must be well designed for ruggedness, corrosion-resistance, and serviceability, and
can read out to a tenth of a degree with accuracy to ± two-tenths. Filled systems are
satisfactory for digesters, while incinerators will have thermocouples supplied with
the system.
The weight of sludge on conveyor belts is measurable by many well-proven designs
(usually available with the conveyor). Weight may also be of interest in inventory con-
trol of chlorine and other materials, but hydraulic systems and more expensive elec-
tronic systems are replacing mechanical scales.
pH, dissolved oxygen, and other electrochemical probes are often a problem, espe-
cially with raw sewage, but mechanical (and perhaps sonic) cleaning systems are
promising. The probes should always be located where they are exposed to a repre-
sentative section of the process and are easily accessible for inspection and servicing
without the need for tools. The probe, probe holder, cable, and junction box or trans-
mitter must be designed to withstand the environment.
Turbidity analyzers, based on the nephelometry of a free surface, are suitable wher-
ever a constant sample stream can be maintained, but they are impractical for raw
sewage or sludges and, consequently, are used off-line (i. e., a small representative
side stream must be fed to them). Residual chlorine analyzers for both free and total
residual are well established and these, too, are used off-line.
Automatic samplers of the type that scoops a small sample at intervals out of a good-
sized side stream (e. g., 30 to 50 gpm in a 2" pipe) have been shown to be effective
and reliable. However, sample systems employing tubing, peristaltic pumps, an
intermittent vacuum, etc., usually do not work on wastewater at all. The "copious
side stream" type also adapts itself well, in many cases, to delivering a representa-
tive stream through the laboratory for use by technicians, as well as laboratory-type
analyzers.
275
-------
Note also that no sluice gates or large control valves are shown as final control ele-
ments. This is because large valves and gates waste energy and because sluice gates
in particular must not be used in conventional control loops without adjusting the
limited duty factor of the gates.
All of the instruments discussed in this section are readily available and can be used
with reasonable assurance of a good performance. Their selection and application,
however, should be performed, or at least reviewed, by persons who are familar with
process rather than laboratory instrumentation, and who are aware of the difference
between the two types.
Instrumentation recommendations contained herein are based on the assumption that
qualified personnel are available at the applicable site to service and maintain simple
analog process-type instrumentation.
ONE MGD ACTIVATED SLUDGE PLANT (Figure 88)
Since the influent to this kind of plant consists of raw sewage, a level alarm (LDA-1)
is recommended wherever plugging of the initial screen is possible and of consequence.
An airflow meter (FI-1) is recommended wherever air is used to help conserve the
air and to make consistent flow adjustments possible. A manual sample point after the
comminutor is noted, and total flow is measured with a Parshall flume. The flow is
recorded and totalized in a conventional manner, [if the influent flow experiences
radical, abrupt changes (as from on-off or high-speedAow-speed pumps), the flume
should be moved to wherever the flow changes less radically. ] There is no instru-
mentation associated with the primary clarifier.
The settled influent, mixed with return sludge, is aerated with good mixing and aera-
tion before the dissolved oxygen (DO) probe is reached at AT-3. The DO value is fed
back to the DO controller (AIC-3), where a signal is generated to vary the airflow by
means of proportional-plus integral controller FRC-3, which maintains the desired
value (the setpoint of the DO controller). A minimum value device (FY-3) ensures
276
-------
to
NOTE: FOR SYMBOLS AMD ABBREVIATIONS,
SEE IS* STD. 55.1
CONTROL SYSTEM IUN BE PNEUMATIC
OR ELECTRONIC.CONVERTERS ARE NOT
SHOWN.
Figure 88. One mgd activated sludge wastewater-treatment plant,
instrumentation diagram
-------
that the airflow will provide good mixing and prevent flooding of the air diffusers in
spite of any process upset or manual change of AIC-3 or FRC-3. The required air-
flow (shown on YRC-3) varies -with the flow rate and aeration needs of the influent
and, by putting the influent flow and DO on the same chart, DO control failures become
immediately apparent. The airflow could well be recorded on the same chart, if
desired.
Return sludge is pumped at a rate proportional to influent to the plant. FFIC-3 and
FIC-4 form a cascaded loop for flow stability, similar to the airflow arrangement with
AIC-3 and FRC-3. The flow of return sludge is not recorded (although it can be and it
is indicated on the control panel), since functioning of the return sludge system is con-
sidered routine.
Good practice recommends a few spare recording points on a control panel, and one of
these might be used occasionally for return sludge flow. Any failure of the sludge
return system is detected and brought to the operator's attention by the sludge blanket
alarm (LA-4). The rest of the system is completely conventional, with the chlorinator
paced by the influent flow signal from FT-2 in an open-loop feedforward arrangement.
This is usually satisfactory, although the ratio of chlorine flow rate to influent flow
rate will not sense a large flow to the holding tank, which could cause a slight but
temporary increase in the chlorination ratio (pounds of chlorine: gallons of effluent).
TEN MGD ACTIVATED SLUDGE PLANT (Figure 89)
Influent to this kind of plant is also handled conventionally. Two instruments that can
be especially useful are a temperature recorder and a plugged screen alarm. The
temperature measurement is intended to show thermal upsets, but the need for TR-1
and LDA-1 depends on conditions peculiar to each site. Flow indicators FI-1A and
FI-1B are recommended, since they allow the operator to know how much chlorine or
air he is using and thus conserve material and maintain constant flows. The measure-
ment flume and the transmitter (FT-2) provide a signal for recording and totalizing
with the addition of a high-flow alarm (FA-2).
278
-------
Flow equalization is provided by the storage wet well and the variable-speed pumping
system. Expected variations in influent rate and the desired degree of smoothing
dictate the necessary storage well size and pump flexibility, using conventional control
engineering methods. The instrumentation is designed to provide variable-speed float-
ing control with nonlinear forcing at the extremes; in other words, the pumps will
change speed only gradually except when the wet well approaches either a full or empty
condition, in which case the pumps are driven to a high or low speed in order to keep
the wet well level within bounds. The size and shape of the wet well, pumping flexi-
bility, influent variability, and instrumentation all must be considered as a single
design.
The operation of the flow equalization system is evident when changes occur. When
the equilibrium is stable, the pump output equals the influent rate, and the level in the
wet well nears midpoint. An increase in the influent rate is essentially integrated by
the wet well, and the pump rate rises in proportion to the level increase in the wet
well (caused by an increase in the wet well input over the output). The rate at which
the pump output changes in response to the level change is determined by the gain
setting of the derivative relay (FY-2B) and that of the controller (LIC-2); this is indi-
cated by the middle slope of the curve on the plot of depth vs outflow in Figure 89.
The scaling of the level and flow transmitters, the cross section of the wet well, the
pump rangeability, the gain of FY-2B and LIC-2, the integration rate of LIC-2, the
characteristics of FY-2A, and the breakpoints in the response curve of LIC-2 all form
a complicated relationship that is best designed and tested by means of simulation.
The reset and derivative value are particularly sensitive, especially since the wet
well and LIC-2 form two integrations in one loop (a potentially unstable situation
unless the time constants are quite different).
The low-pass filter (FY-2A) is necessary to exclude noise and other higher frequency
disturbances that cannot indicate a true change in the wet well content. The derivative
relay (FY-2B) anticipates influent changes and may or may not be necessary, although
it can reduce the wet well size requirement; this relay should not be replaced by a
279
-------
rwrnotrr ircstouAt. CHLOJHN*
1-0
00
o
NOTE: FOR SYMBOLS AND ABBREVIATIONS.
SEE ISA STD. SS.I
CONTROL SYSTEM MAYBE PNEUMATIC
OR ELECTRONIC.CONVERTERS ARE NOT
SHOWN.
•AERATION SYSTEM MAY BE MECHANICAL
OR DIFFUSED AIR.
Figure 89. Ten mgd activated sludge wastewater-treatment plant,
instrumentation diagram
-------
conventional derivative action in a controller (LIC-2) that would be influenced by any
setpoint change as well as any input signal change. The performance of the flow equal-
ization process is continuously displayed and then stored in a form usable for future
analysis by the three-channel recorder (FR-2/LR-2/SR-2). The flow signal from
FT-3 could be substituted for the speed signal, as is shown subsequently in the dis-
cussion of the 1 mgd physical/chemical plant.
The level switch arrangement (LS-2B) controls the operation of the aerators/agitators
so that they slow down and stop as the wet well level drops, thus preventing excessive
bottom scouring and/or the aerators running with too little coverage.
Primary sludge is withdrawn in the conventional manner by using a radiation-type
density meter (DE-7) and timers. A sludge blanket level probe could provide equiva-
lent control, but the density meter, when used with FE-7, provides a semiquantitative
record of the weight of withdrawn solids. The aeration system is intended to operate
similarly tothe 1 mgd activated sludge plant discussed previously, with the suggested
addition of a feedforward loop (via FFRC-3B) to make the aeration proportional to the
flow from the pumps.
Secondary sludge recycle is flow proportioned, as in the previous example, while a
sludge blanket probe ensures wasting in the case of a secondary sludge buildup. More-
over, in practice the control valve is usually prevented from closing entirely to ensure
continuous wasting. The operation of the final clarifier is monitored by the turbidity
recorder (AR-4A), which is useful for indicating trends and upsets but cannot be
depended on for a reliable reading of the solids content.
Chlorination is conventional, with the open-loop flow pacing from FT-3 readjusted by
closed-loop control from ARC-4B. As usual, the final residual chlorine in the effluent
must be determined from samples taken after the chlorine contact tank, since the values
from AT-3 are required for control and will be higher than the final values.
The operation of the thickener to concentrate the sludge and feed the digester is not
shown; neither are systems to warn of escaping gas and/or control the withdrawal of
281
-------
waste sludge or gas. The single exception to this is a back-pressure valve to ensure
that digester pressure is always positive and sufficient to support the digester cover.
A conventional method to circulate and heat the digester contents is shown, using a
thermostatic device (TIC-8) which controls the hot-water pump.
The digester gas analyzer(s) indicated by AT-8 are useful for sensing changes in diges-
tion, rather than merely the average composition. Suitable analyzers might monitor
the methane content (via flame ionization detectors) or simply density, but they work
if—and only if—the sample transport and conditioning system is well designed.
ONE MGD PHYSICAL/CHEMICAL PLANT (Figure 90)
For a discussion of influent monitoring and flow equalization, refer to the previous
sections. The physical/chemical plant begins to differ from the activated sludge plant
with the feed system for the solids contact unit.
Lime dosage can be feedforward and yet remain accurate and reliable provided delays
and nonlinearities are kept small. If the pacing signal for the lime feeder is taken
from FIC-2 rather than FE-2, the advantage in time response largely compensates for
the unavoidable delay between a signal to the variable-speed lime feeder motor and the
moment of lime-slurry delivery into the rapid-mix tank. The lines among the dry lime
feeder, the small mixer, and the rapid-mix tank must be short, and the intensive-mix
tank must be small. (The time of hydration as well as the time of lime reaction must
also be verified as part of the system's design,) The addition of polymers or other
coagulants is also controlled by means of a feedforward loop; this has become well-
established practice.
Sludge removal is controlled by timers and an in-line density meter (as in the previous
example), with the addition of a multiplying relay (FY-7) to indicate the mass flow of
solids removal continuously.
A cascaded feedback system for recarbonation is illustrated in Figure 90, where the
setpoint of the flow controller (FIC-3) is manipulated by the pH controller (ARC-3).
282
-------
KJ
00
NOTE: SOURCt OF MCKW45H
WATER FOB ANY GIVEN
CARSON FILTER IS
EFFLUENT FROM OTHER
CARBON FILTERS IN
OPERATION
CONTROL SYSTEM MAY BE PNEUMATIC OR
ELECTRONIC. CONVERTERS *»E NOT SHOWN.
Figure 90. One mgd physical/chemical wastewater-treatment plant,
instrumentation diagram
-------
A submerged combustion unit could be used instead but, while this might eliminate the
need for a mechanical mixer, it could also cause a possible decrease in response and
flexibility.
The filtration system is conventional. Although only one is shown, several filters are
employed, each with its own control valve, and all feed a common clear well. Simple
feedback controls pace the filtration rate and pump speed to accommodate the input,
while the filter condition is indicated for each filter by LDI-3. The backwash is initi-
ated either manually or by a timer (in a system not shown in detail in Figure 90.
Overall filtration efficiency is shown by a turbidity analyzer (AI-6), with alarm AA-6
to warn of extreme values. The AT-6 output could also be recorded to indicate trends
continuously. (The need for AT-6 results from the requirements of the carbon beds
for a feed free of solids.)
Only one of several carbon beds is shown. These beds are connected either in series
or parallel. Gross plugging and plugging trends are recorded and alarmed by PDR-5
and PDA-5. If a suitable automatic water quality analyzer were commercially avail-
able, this would be the place to use it but, with the present state of the art, carbon
bed control and backwashing can be controlled manually, or automatically by a pro-
grammed timer, based on the laboratory tests of samples collected manually in whole
or in part.
The effluent chlorination and control of the thickener, centrifuge, etc. are not shown
in Figure 90.
TEN MGD PHYSICAL/CHEMICAL PLANT (Figure 91)
The instrumentation for this complicated and extensive facility is designed to provide
the most effective, useful, and reliable performance, as well as the production of
meaningful data. Centralized control (in one or several centers) of some 10 to 12
unit operations is assumed. It is further assumed that: 1) no process intelligence at
frequencies greater than 0.1 Hz or overall accuracy greater than 2% need be collected,
284
-------
and 2) in many cases, the interaction of measurable values, as well as the same vari-
ables measured singly, represents desirable data. Consequently, considering the cost
per point of obtaining useful information, multi-channel (sampling type) self-balancing
potentiometric records are least costly and thus are recommended for recording up to
six points per recorder. One alternative is to use a computerized data collection sys-
tem, but this is slightly more expensive (in both capital and operating cost) and is also
somewhat less reliable overall, so that its selection must hinge on other factors such
as increased flexibility and automatic data processing.
The operation of the water purification processes is straightforward and has been
described above. Two residual chlorine analyzers are shown, arranged in cascaded
feedback loops to ensure an adequate final residual at all times. The savings in chlo-
rine provided by the second analyzer (AT-6B) plus the assurance of meeting effluent
standards justifies the cost. These analyzers have been well proven in wastewater
service. Although backwashings, thickener overflows, and vacuum filtrates are
returned to the head of the plant, there is no sludge recirculation. Control of the
thickener and vacuum filter is assumed to be conventional as supplied by the equipment
manufacturer, and it is described elsewhere.
The output of the vacuum filter battery is indicated, recorded, and totalized conveni-
ently and reliably from the conveyor belt scale (WT-11). This information provides a
good performance index for management, as well as a convenient and responsive indi-
cation of filter performance for the operator (he is no longer required to watch every
filter so often, once he has put it into service).
The operation of the incinerator and scrubber is merely indicated since these systems
are usually provided as complete packaged systems. However, reliable flue gas oxy-
gen analyzers with sampling systems are now commercially available. These operate
under plant conditions with little or no maintenance and, in addition, provide an output
so dependable that an operator can rely on the information to set the fuel/air ratio to
his burners so as to ensure their utmost performance and efficiency.
285
-------
str POINT
-®
00
cr>
NOTE: SOURCE Of BACKWASH
WATER FOR ANY OWEN
CARBON FILTER IS
EFFLUENT FROM OTHER
CARBON FILTERS IN
OPERATION.
NOTE: FOR SYMBOLS AND ABBREVIATIONS,
SEE ISA STD. SS.I
(fi)— SYMBOL DENOTES COMMON RECORDER
CONTROL SYSTEM MAY BE PNEUMATIC OR
ELECTRONIC CONVERTERS ARE NOT SHOWN.
Figure 91. Ten mgd
physical/chemical wastewater-treatment plant,
instrumentation diagram
-------
APPENDIX A
FOOD-TO-MICROORGANISM CONTROL DETAIL
SOLIDS BALANCE AROUND AERATION TANK
dC
A
dC
where
VR+ YQF
a = -
v (1)
Solution is
- 6 - (6 - CAO)e"Pt (2)
where
WYQF
C . n = initial value
AO
287
-------
To get aeration tank into the desired F/M ratio at time t, desired
CONTROL PROGRAM FOR Q
ti
Say t = 1 hour = 1/24 day
1. On the hour read
MGD Ib/MG
2. Assume Q = 0.0 and compute
XV
C equation (4)
p equation <1)
6 equation (3)
C equation (2)
A.
If computed C . > C A _, make Q^ = 0
A AD R
If C < C compute required Q as follows:
A ALJ R
1. Assume trial value of Q and compute B from equation (I)
R
2. Compute required e from
- CAO e'Pt
-
288
-------
3. Compute required Q from
R
6 (Q + K V) - YQF
Q = - - - (6)
4. With new value of Q_, repeat the computation until Q converges.
R K
5. Change Q to computed value.
R
PROGRAM FOR Q0
o
Rate of flow of solids from aeration tank
= (Q + QR)
C = average value for interval
CAO+CA
To remove solids from final tank at same rate, make
(Q + QR) C
Q = —- (7)
S CS
The value of C will not be independent of Q_ but will depend upon Q and the settling
S DO
qualities of the sludge. The problem is to maximize C and thereby minimize Q
o &
without making Q C < (Q + Q ) C .
D D XX A
IfQ C <(Q + Q )C , sludge will be accumulating in the final tank and Q should be
S S R A o
increased.
If Q - C >(Q + Q)C, excess water is being pumped from final tank into sludge
S S R A
storage, and Q should be decreased.
D
289
-------
VALUE OF SLUDGE STORAGE
Increase in sludge storage volume during interval = AV = (Q - Q )
OO O IX
290
-------
APPENDIX B
NOTATION FOR INSTRUMENT CONTROL LOOPS
The meaning of the instrument symbols can be determined using the information con-
tained in both Figure B-l and Table B-l. The abbreviations used in the symbols con-
sist of a combination of two or more letters. The first letter indicates the measured
variable. The second and third letters are modifiers and instrument function indi-
cators. The numbers that may be found below the identifying letters are the loop
numbers as assigned by the engineer. Detailed information concerning this coding
system is contained in Figure B-l and Table B-l.
291
-------
INSTRUMENT SYMBOLS
[Ml
-or
tx
9
T
WATER SURFACE
ANY PUMP
BLOWER
SCREEN
SLUICE GATE
CHECK (FLAP) VALVE
BUTTERFLY VALVE
ANY VALVE
DIAPH. ACTUATOR
SOLENOID ACT.
HANDWHEEL ACT.
ADJUST. OR VARIABLE
-* Xr-
ELECTRIC MOTOR
HYDRAULIC DRIVE
AIR PURGE UNIT
WEIR
FLUME
VENTURI OR FLOW TUBE
ORIFICE PLATE
POS. DISPL. METER
PITOT TUBE
CAPILLARY
ELECTRIC LINE
PNEU. LINE (3-15 PSD
INSTRUMENT IDENTIFICATION
/
LOCATK
(NONE) -
— V 1 r\-i
VI <-O7
}N
FIELD
' \
LOOP
NO.
ASSIGNED BY
\
FUNCTION
SEE TABLE
BEHIND PANEL
-ON PANEL
= "ON AUX. PANEL
ENGINEER
OO- TWO INSTR'S IN ONE
Figure B-l. ISA symbols
292
-------
Table B-l. INSTRUMENT ABBREVIATIONS
IDENTIFICATION LETTERS
Typical tag for Indicating Flow Controller FIC-3A;
A
B
C
D
E
F
H
I
J
K
L
M
N
P
Q
R
S
T
U
V
w
X
Y
Z
F I
first second
letter letter
Measured or
Initiated Variable
ysis (1)
icr (flame)
C
last
letter
3
lOOD
Mr
number
Modifier
Alarm
Special (3)
A
suffix
Instrument
Function
Alarm
Special (3)
Conductivity
Density or Sp. Gr.
Electric (General)
Flow
Hand (Manually Initiated)
Power
Time or Time Schedule
Level
Moisture or Humidity
Special (3)
Pressure (Vac.)
Running (Status)
Speed
Temperature
Multi-use
Weight, Force or Torque
Special (3)
Position
— Controller
Differential —
Element (measuring) Element
Ratio (2) —
- High
Indicator Indicator
Light (Pilot)
Special (3)
Totalizer
Recorder (4)
Switch
Well
Special (3)
Special (3)
Control Station
Low or Light
Special (3)
Totalizer
Recorder (4)
Switch
Transmitter
Valve
Well
Special (3)
Relay or Computer
(1) Type of analysis to be defined outside balloon as: pH, ORP, D. O. (dissolved
oxygen), R. C. (residual chlorine), TURB (turbidity), etc.
(2) As a modifying letter to designate (fraction) ratio, i. e., FFIC-Flow Ratio
Indicating Controller.
(3) As defined in Instrument List of each job.
(4) Or printer.
293
-------
REFERENCES
1. Molvar, A.E., J. F. Roesler, R.H. Wise and R. Babcock, "Instrumentation
Automation Experiences in Wastewater-Treatment Facilities," EPA report
to be published.
2. Molvar, A. E. and J. F. Roesler, "Selected Abstracts for Instrumentation
and Automation of Wastewater Facilities," National Technical Information
Service, No. PB-225 520/6 (1973).
3. Petersack, J. F. andR.G. Smith, "Advanced Automatic Control Strategies
for the Activated Sludge Treatment Process," Environmental Protection
Technology Series, EPA 670/2-75-039, May 1975.
4. Ryder, Robert A., "Dissolved Oxygen Control in Activated Sludge," 24th
Industrial Waste Conference 1969, Purdue University (1969), p 238.
5. Eckenfelder, W. W., "Water Qualify Engineering for Practicing Engineers,"
Barnes & Nobles (1970).
6. Smith, R. andR.G. Eilers, "Simulation of the Time Dependent Performance
of the Activated Sludge Process Using the Digital Computer," FWQA, Advanced
Waste Treatment Branch, Cincinnati (October 1970).
7. Bush, A.W., "Aerobic Biological Treatment of Wastewater," Oligodynamics
Press, Houston, Texas, (1971).
8. Keefer, C. E., "Sewage-Treatment Works," McGraw-Hill Book Co.,
New York (1940).
9. Stanier, R.Y., M. Doudoroff, andE.A. Adelberg, "General Microbiology,"
2nded., MacMillan, London, (1964).
10. Kotze, J. P. , P.G. Thiel, and W.H. Hottingh, "Anaerobic Digestion H,
The Characterization and Control of Anaerobic Digestion," Water Research,
Vol. 3 (1969), pp 459-494.
11. Lohmeyer, G. T., "A Review of Sludge Digestion," Sewage Ind. Wastes,
31 (1959), pp 221-235.
12. Brown, J. M. and B. Kinchusky, "Digester Indigestion From High Temper-
ature," Water Pollution Control Federation, J 37 (1965), pp 416-417.
294
-------
13. Graef, S. P. and J. F. Andrews, "Automatic Control of the Anaerobic
Digester," presented at the 1972 national meeting of the AIChE in St. Louis.
14. Tenney, M. et al, "Chemical Conditioning of Biological Sludges for Vacuum
Filtration," Journal of WPCF, Volume 42, No. 2 (February 1970), ppRl-R20.
15. Shinskey, F.G., "pH and plon Control in Process and Waste Streams,"
Wiley, New York (1973).
16. Parsons, W. A. "Chemical Treatment of Sewage and Industrial Wastes,"
National Lime Association, Washington (1965).
17. Hoyle, D. L., "The Effect of Process Design on pH and plon Control,"
presented at 18th ISA-AID Symposium, San Francisco (May 3, 1972).
18. Bethlehem Regulator Selection Handbook, The Bethlehem Corporation,
Bethlehem, Pa.
19. Shinskey, F.G. andT. J. Myron, "Adaptive Feedback Applied to Feed-
forward pH Control," ISA Paper No. 565-70.
20. Shinskey, F.G., "Adaptive pH Controller Monitors Nonlinear Process,"
Control Eng. 57 (February 1974).
21. "pH Electrodes and Holders," Technical Information Sheet 43-1 la,
The Foxboro Company, Foxboro, Mass. (December 1972).
22. Shinskey, F.G. "Process-ControlSystems," McGraw-Hill Book Company,
New York (1967), p 102.
23. Shinskey, F.G. "AdaptiveNonlinear Control System," U.S. Patent 3,794,817
(February 26, 1974).
24. SAMA Standard PMC 20.1-1973, "Process Measurement and Control
Terminology," Process Measurement and Control Section, Inc., SAMA,
New York.
25. "Process Design Manual for Phosphorous Removal," prepared by Black &
Veatch, Consulting Engineers, for U.S. EPA (October 1971).
26. Butler, J.N. "Ionic Equilibrium: A Mathematical Approach," Addison-
Wesley, Reading, Mass. (1964), p 507.
27. Albertson, O.E. andR.J. Sherwood, "Phosphate Extraction Process,"
Dorr-Oliver, Inc., Stamford, Conn. (1968), p 9.
28. "Control Systems Design Manual for the Neutralization of Acids and Bases,"
The Foxboro Company, for U.S. EPA (1974).
29. EPA publication 11024 DMS, 05/70, "Engineering Investigation of Sewer
Overflow Problem" (Roanoke), p 1.
295
-------
30. Ibid, pp 95 and 96.
31. PHS publication 126, 11/64, "Pollutional Effects of Stormwater and Over-
flows From Combined Sewer Systems."
32. Reference 29, item 15, p 3.
33. EPA publication 11024, 01/71, "Storm and Combined Sewer Pollution
Sources and Abatement" (Atlanta), pp 59 and 60.
34. EPA publication R2-73-124, 01/73, "Microstraining and Disinfection of
Combined Sewer Overflows—Phase 2" (Philadelphia), p 44.
35. EPA publication 670/2-049, "Microstraining and Disinfection of Combined
Sewer Overflows—Phase HI (Philadelphia).
36. EPA publication 11023 EYI, 04/72, "High Rate Filtration of Combined
Sewer Overflows" (Cleveland).
37. EPA publication 11023, 08/70, "Retention Basin Control of Combined
Sewer Overflows" (Springfield, Illinois).
38. EPA publication 11020 FAQ, 03/71, "Dispatching System for Control of
Combined Sewer Losses" (Minneapolis).
39. EPA publication 11022 ELK, 12/71, "Maximizing Storage in Combined
Sewer Overflows" (Seattle).
40. Reference 29, p 167.
41. Reference 33, pp 8 and 41.
42. Reference 34, p 1.
43. Reference 33, p 4.
44. Reference 34, p 3.
45. EPA survey, "State of the Art of Instrumentation in Municipal Waste Water
Plants" (presently in process).
46. Fischer & Porter Company bulletin R90-49-31, "Closed Loop Chlorination
Control."
47. EPA publication 11022 DMU,07/70, "Combined Sewer Regulator Overflow
Facilities."
48. Peters, M.S. andK.D. Timmerhaus, Plant Design and Economics for
Chemical Engineers, McGraw-Hill Book Co., New York (1968) p 243.
49. Liptak, E.G. (editor), Environmental Engineers1 Handbook, Volume n,
Chilton, Pa. (1974), p 500.
296
-------
50. Ibid, Volume I, p 1486.
51. Michel, R. L., "Cost and Manpower for Municipal Wastewater Treatment
Plant Operation and Maintenance, 1965-1968," Journal WPCF (November
1970).
52. U.S. EPA, "Capital and Operating Costs of Pollution Control Equipment
Modules," EPA-R5-73-023 (July 1973).
53. Reference 50, Section 5.11.
54. Boyle Engineering and Lowry and Associates, "Master Plan Trunk Sewer
Facilities," for County Sanitation District No. 3 of Orange County, Calif.
(June 1968).
55. Reference 50, Section 2.12.
56. Reference 50, Section 5.11.
57. Reference 50, Section 5.23.
58. Reference 50, Section 6.22.
59. Roesler, J. F., "Factors in the Selection of a Control Strategy," Office of
Research and Monitoring, U. S. EPA, Cincinnati.
60. Smith, R., "Wastewater Treatment Plant Control," Joint Automatic Control
Conference, St. Louis (August 1971).
61. Reference 49, Section 1.2.
62. Upfold, A. T., "Manhour Ratings Standardized for Instrument Maintenance,"
Instrumentation Technology (February 1971).
63. Liptak, B.C., "Process Instrument Costs—How to Estimate," Chemical
Engineering, Reprint No. 101.
64. Molvar, A.E., "Instrumentation and Control of Biological Wastewater
Treatment Plants," prepared by Raytheon Company (February 1972).
65. Barnard, J. L., "Treatment Cost Relationships for Industrial Waste
Treatment."
66. Derick, C. and K. Logic, "Flow Augmenting Effects of Additives on Open
Channel Flows," EPA-R2-73-238 (June 1973).
67. "Wastewater Engineering," Metcalf and Eddy, Inc., McGraw-Hill Book Co.
(1972), p 63.
297
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
REPORT NO.
EPA-600/2-76-276
2.
I. RECIPIENT'S ACCESSIOI»NO.
4. TITLE AND SUBTITLE
SELECTED APPLICATIONS OF INSTRUMENTATION AND
AUTOMATION EN WASTEWATER-TREATMENT FACILITIES
5. REPORT DATE
December 1976 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Allen E. Molvar
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Raytheon Company
Submarine Signal Division
W. Main Road, Portsmouth, RI 02871
10. PROGRAM ELEMENT NO.
1BB043
11. CONTRACT/GRANT NO.
68-03-0144
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory
Office of Research and Development
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
See also EPA-600/2-76-198, "Instrumentation and Automation Experiences
in Wastewater-Treatment Facilities"
lf Implication of modern control systems to the operation of wastewater-treatment plants is
discussed. Control strategies for the commonly used wet- and dry-weather treatment processe
and their collection systems are described. Wherever possible, the benefits derived from, as
well as the operating problems associated with, the actual or proposed control strategies are
documented. Cost/benefit analysis indicates that many untried feed forward mass proportional
control schemes are economically attractive because of the low payback periods. This study
concludes that despite current concepts, the smaller (1 to 5 mgd) plants can afford and need sig-
nificantly greater amounts of automatic control. A lack of reliable field-proven analytical
sensors for important parameters appears to be the principal obstacle impeding the implemen-
tation of more sophisticated control strategies. Centralized control with semigraphic display
should be used in treatment plants since it saves on operating labor, improves operation, and
increases the safety of wastewater treatment. Automatic data acquisition systems are cost
effective and should be used in medium and large sized plants. Direct digital control and com-
puterized control can only be economically justified in large dry-weather treatment plants and
large storm-water control networks.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTJFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
*Automation, Automatic Control,
Automatic Control Equipment, Data
Processing, Digital Computers,
Instruments, *Waste Treatment,
Wastewater, Process Control,
Centralized Control
Activated Sludge,
Process Control Theory
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
312
20. SECURITY CLASS (This page I
UNCLASSIFIED
22. PRICE
EPA Form 222O-1 (9-73)
298
•&U.S. GOVERHHEin HINTING OffICl: 1977-757-065/5450 Region No. 5-M
------- |