RESEARCH NEEDS
                      FOR
           AUTOMATION
                       OF
WASTEWATER TREATMENT
                 SYSTEMS
             PROCEEDINGS of a WORKSHOP
sponsored by the U.S. ENVIRONMENTAL PROTECTION AGENCY
         in cooperation with CLEMSON UNIVERSITY


-------
                      14S54
            RESEARCH NEEDS
                           FOR
              AUTOMATION
                            OF
 WASTEWATER TREATMENT
                     SYSTEMS
                PROCEEDINGS of a WORKSHOP
           held at Clemson, SC, September 23-25, 1974
Sponsored by the U. S. ENVIRONMENTAL PROTECTION AGENCY
          in cooperation with CLEMSON UNIVERSITY
        H. O. Buhr, J. F. Andrews and T. M. Keinath, editors
                       Clemson University
                    Clemson, South Carolina
                             1975

-------
            ORGANIZING COMMITTEE

Cochairmen:

   A. W. Breidenbach, Environmental Protection Agency
   J. F. Andrews, Clemson University

Arrangements:

   G. D. Barnes, City of Atlanta
   T. M. Keinath, Clemson University


Members:

   W. W. Eckenfelder, Jr., Vanderbilt University
   C. F. Guarino, City of Philadelphia
   J. F. Roesler, Environmental Protection Agency
   W. A. Rosenkranz, Environmental Protection Agency
   J. R. Trax, Environmental Protection Agency
   D. R. Wright, Environmental Protection Agency

-------
                                                  CONTENTS
FOREWORD
Page
.   5
WORKSHOP OBJECTIVES	   7
     Andrew W.  Breidenbach,  Environmental Protec-
     tion Agency

WORKSHOP SUMMARY 	   9
     William A. Rosenkranz, Environmental Protection
     Agency
WASTEWATER COLLECTION SYSTEMS
  Automated Collection Systems
  Instrumentation Needs 	  12
     Curtis  P.  Leiser,  Municipality  of Metropolitan
     Seattle
  Present Practice  and Research Needs in Wastewater
  Collection System Design and Operation	  18
     James J. Anderson, Watermation, Inc.
  Discussion	  19

  Report of Working on Wastewater Collection Systems .  21
     John A. Lager, Metcalf & Eddy, Inc.
     Harold Torno, Environmental Protection Agency

BIOLOGICAL TREATMENT PROCESSES
  Dynamics and Control of Biological
  Treatment Processes	  26
     John F. Andrews, Clemson University

  Automation of the Activated Sludge Process	  38
     Michael J.  Flanagan, Brown & Caldwell Consulting
     Engineers

  Discussion	  46
  Report of Working Party on Biological
  Treatment Processes	  52
     Paul H. Woodruff, Roy F. Weston, Inc.
     A. W. West, Environmental Protection Agency

PHYSICOCHEMICAL PROCESSES

   Field Experiences with a Pilot-Physical-Chemical
   Treatment Plant	  56
     Walter  W.   Schuk,  Environmental  Protection
      Agency

   Automation of Physical and Chemical Processes	  58
     Thomas M. Keinath, Clemson University
    Discussion	  64
   Report of Working Party on
   Physicochemical Processes	
     Wesley W. Eckenfelder, Jr., Vanderbilt University
     John Stamberg, Environmental Protection Agency
                                                 Page

  Report of Working Party on Physicochemical Processes  66
     Wesley W. Eckenfelder, Jr., Vanderbilt University
     John Stamberg, Environmental Protection Agency

SLUDGE PROCESSING, TRANSPORT AND DISPOSAL

  Automation of Sludge Processing,
  Transport and Disposal 	  70
     Bart T. Lynam, Raymond R. Rimkus and Stephen
     P.  Graef, The Metropolitan  Sanitary District of
     Greater Chicago
  Control of Sludge Handling:
  Some Successes and Problems	  72
     Joseph B.  Farrell.   Environmental  Protection
     Agency

  Discussion	  76
  Report of Working Party on Sludge Processing,
  Transport and Disposal	  79
     Richard I. Dick, University of Delaware
     John R. Trax, Environmental Protection Agency

COMPUTER APPLICATIONS IN AUTOMATION

  Current Practice in Instrumentation and Computer
  Application at the County Sanitation Districts of
  Los Angeles County 	  84
     Walter E. Garrison,  Kip Payne and Tim Haug,
     County Sanitation Districts of Los Angeles County
           Computer Applications in Automation ....
             William E. Dobbins. Teetor-Dobbins, P.C.
                                                   99
  66
   Discussion	106

   Report of Working Party on Computer Applications  . . 109
     Carmen  F. Guarino, Philadelphia Water  Depart-
     ment
     John M. Smith, Environmental Protection Agency

EVALUATION OF THE EFFECTIVENESS OF
AUTOMATION
   Evaluation of the Effectiveness of Automation  	114
     Joseph F. Roesler, Environmental Protection Agency
   Field Evaluation of the Effectiveness of Automation  ..118
     Allen E. Molvar, Raytheon Company
   Discussion	128
   Report of Working Party
   on Evaluation of Effectiveness  	130
     Walter  G.  Gilbert,  Environmental  Protection
     Agency
     James A. Mueller, Manhattan College

LIST OF PARTICIPANTS	133

-------
                                             FOREWORD
   Improvement of  the  operation  of  municipal wastewater
treatment systems is one  of the nation's major environmental
problems. Possible solutions to this problem  are increases in
the quantity and quality of personnel  involved in operations
and/or increasing use of automated systems. The need for an
increase in the quantity and quality of personnel is well known
and  has been the  subject of  numerous conferences;  the
prospect of improving operations through the use of automa-
tion,  however, is  not as  well recognized. There  are  many
reasons  for this, but one of the most important is that research
efforts expended on the automation of wastewater treatment
systems are  relatively minor when compared with research on
other aspects of water pollution control.
   There is  currently a great  interest  in  the automation of
wastewater treatment systems as evidenced by the attendance
of more than 200 persons at  an International Workshop  on
Instrumentation,  Control  and  Automation  of Wastewater
Treatment Systems which was  held in London during the fall
of  1973.  The  automation  systems  reported on  at  that
workshop for cities  such as  Atlanta, Chicago, London,  Los
Angeles, Paris,  Philadelphia,  etc.,  will  greatly  affect  the
performance of wastewater treatment systems valued at many
millions of dollars.  However, few of the papers presented at
the workshop were oriented toward research, which again is an
indication that relatively few researchers are currently engaged
in this  area. Most automatic control installations for waste-
water treatment systems are, of necessity, designed on  an
empirical  basis  because   of  a  lack of more fundamental
knowledge   concerning such  factors as dynamic  behavior,
control  strategies,  component  reliability and cost/benefit
analysis.
   The above statements illustrate the great need for research
on the automation of wastewater treatment systems. However,
in order to accomplish such research  in the  most effective
manner, it is necessary to first clearly define and establish
priorities for the research needed. Recognizing this,  the U. S.
Environmental Protection Agency requested Clemson Univer-
sity to organize  and conduct a Workshop on Research Needs
for Automation of Wastewater Treatment Systems.
   In  order to  insure the incorporation  of all viewpoints,
participants were invited from  government  regulatory and
research  agencies, universities, operating engineers and man-
agers  of large treatment systems, consulting engineering firms
and equipment manufacturers. Extensive and lively discussion,
which forms the heart of a workshop, was  encouraged by
dividing the workshop into three portions. The  first portion
consisted  of formal  presentations  of present  practice and
current research  on  each of six topics, by authorities on these
specific topics. These formal presentations, and the discussions
associated  with  them, served to set the  stage for individual
meetings of working parties on each of the six topics, where
special attention was  devoted  to  stating the problems and
specifying  research  needed  to  solve  these  problems. The
cochairmen of  the  working  parties  then  prepared brief
documents summarizing  the  discussion in their session and
orally presented these for discussion at a reassembly of all of
the Workshop participants. Finalization of these documents,
representing the  deliberations of each working party and the
viewpoints expressed at the reassembly, was accomplished at a
meeting of the cochairmen of the individual working parties in
Washington on November 19, 1974. Special thanks are due to
the cochairmen of  the  six  working parties for  their per-
formance of a difficult and time-consuming task.
   These proceedings represent the integrated best judgement
of the experts gathered at the workshop as to research needs
for the automation of wastewater treatment systems. It should
provide a firm foundation for the  development of a national
research program in  this important area.

                         John F. Andrews, Cochairman
                         Workshop on Research Needs
                         for Automation of Wastewater
                         Treatment Systems
April  1975

-------
                             WORKSHOP  OBJECTIVES
                                             Andrew W. Breidenbach
                 Director, National Environmental Research Center, Environmental Protection
                                         Agency, Cincinnati, OH 45268
   This workshop was developed to provide an opportunity to
discuss problem areas and research needs for the automation
of wastewater  treatment  plants.  The  discussion  generated
should have significant impact on  future research and should
ultimately  affect the  design and operation of wastewater
treatment plants.
   Clemson was selected as the site for this Workshop because
of its leadership in the area of control engineering as applied to
environmental systems. Recognizing automation as  an area of
important research, the United States Environmental Protec-
tion Agency at the National Environmental Research Center-
Cincinnati (NERC) initiated  a research program in automatic
control of wastewater treatment plants a little over two years
ago. Because of mutual interests, a jointly sponsored workshop
to define research needs seemed essential if we were to achieve
our goal of  having  fully automated  wastewater  treatment
plants on stream in the 1980's.
   Looking back only ten years, it  is apparent that automation
of wastewater  treatment systems  is a  recently  developed
technology that has yet to be fully exploited.
   Ten  years  ago,  the  dissolved oxygen (DO) probe  was
emerging from the  laboratory  for  application  to relatively
clean  water. The thought of placing a DO probe in the hostile
environment of a wastewater treatment plant appeared to be
an unworkable concept. Today,  it is a reality. A recent survey
sponsored  by  the  U. S. Environmental  Protection  Agency
found that  12%  of  the  50  plants surveyed have automatic
on-line DO control.
   Total organic carbon (TOC)  was  virtually an academic
curiosity ten years ago. At that time, the Robert  A. Taft
Laboratory had just received the  prototype of a commercial
TOC  analyzer.  However, less than ten years later, continuous
on-line  TOC analyzers have become  available on the  open
market.  The same can be said about  process control  com-
puters; they were  available ten  years  ago, but  they  were
expensive and difficult to program. Yet, some process  com-
puters can be programmed easily  in languages as simple as or
simpler than FORTRAN. The development of the 1C chip in
the sixties gave birth to micro-computers.  Properly applied,
these computers are  very effective for data acquisition and
process control in situations where  previously  they were not
cost-effective. Today, more than two  dozen  plants are on
stream or  will  be on stream with process computer installa-
tions. Some cities such as Seattle and Minneapolis have been
using a computer for several years to control the stormwater
flow  in their combined sewer systems. They are now expand-
ing their computer  installations to control other aspects of
wastewater processing as well.
   Although  progress is being made,  it is  being made at a
relatively slow pace. One of the questions we must face at this
workshop is: What is delaying the implementation of automa-
tion in wastewater treatment  plants? Is it the lack of suitable
sensors for automatic on-line control? For example, consider
the recently developed ammonia probe: It does not appear to
be  applicable  for automatic on-line  control  for  nitrogen
removal because the probe requires considerable maintenance
by  a  skilled  technician.  The  same  can  be  said for  the
continuous on-line  TOC analyzer. It  is known that many
plants in the country simply cannot afford to hire such skilled
technicians. The lack of sensors and trained technicians may
be the major deterrent to the automation of treatment plants.
If so, is there a solution  to these problems?
   To control a property of a process stream, an engineer must
be knowledgeable in several areas.  He  must be familiar with
the characteristics and the limitations of measuring  devices,
with the treatment plant's chemical  and biological reaction
kinetics and with  the process equipment in which these occur.
A knowledge of the plant's design limitations and its opera-
tional  stresses  in terms of the loadings and  environmental
changes to which  this equipment will be subjected is essential.
Control theory, computer technology, and systems analysis are
important tools in  this field. Thus, special training  is often
necessary for an effective  environmental systems and control
engineer.
   Until recently,  there  were very  few   doctoral  degrees
awarded in the area of systems analysis. Now, a whole new
field  has opened. Clemson University was one of the first
universities in  the  United States  to  apply systems  analysis
techniques to the automatic control of wastewater treatment
plants. The first doctoral degree in the area of environmental
systems control engineering was awarded here in the middle
sixties. This is a  new field that has grown very rapidly with
continuing growth potential for further development.
   If automated  treatment  is  to be applied,  the  existing
technology must  first be examined. An open attitude towards
automation  is  a  requirement.  Evaluation of this technology
and adaptation of it to our needs will then follow. Finally, we
must  improve the technology where necessary.  Thus,  the
development of automation will require resources, hard work,
and research.
   The first objective of this workshop is to instill enthusiasm
in each of us. We must convince  ourselves (and then others)

-------
that resources, hard work, and research will be well-spent on
automation of wastewater treatment plants. Today, I  am
confident that we have accomplished the first objective merely
by having carried out this workshop. Next, we must exchange
information and  experiences,  communicating  our successes
and failures.
   The second objective of this workshop, and one that vitally
concerns EPA's National Environmental  Research Center in
Cincinnati, is to assess the research necessary to bring about
successful and cost-effective automation of wastewater treat-
ment plants throughout the United States. We must establish a
channel of communication between the users of our research
(especially the designers, operators, engineers, and managers of
our  wastewater  treatment  plants)  and  the   persons  who
influence  the  drafting and enforcement of state and  federal
regulations. One of the ultimate research goals of the Center is
to develop fully  automated  wastewater  treatment systems.
This  includes  flow  routing  and  storage in the  collection
systems as well as processing technology such as biological
treatment, physical-chemical  treatment and  sludge handling
and disposal. We are interested in defining the most rewarding
areas for further research to extend the present state-of-the-art
and develop practical control  technology.  We need to evaluate
control techniques such as F/M control,  with on-line respir-
ometers  or feedforward  TOC  measurements.  We  wish to
discuss  treatment  plant  reliability  ideas such  as  control of
toxic wastes  entering the  plant. We wish  to explore the
possibility  of how best to  utilize  a process  computer for
automating a  wastewater treatment plant. We also wish to
discuss the effectiveness of various automatic control strategies
and  how  these control  strategies  would  improve  the  per-
formance  or  lower  the  cost  of  operating  a  wastewater
treatment plant. And finally, we would like to discuss methods
of improving  the plant operators' attitudes towards automa-
tion. Ultimately, it is the  operator who must live  with the
automated  plant.  His  attitude  and his approach towards
operating an automatic plant will certainly help determine the
success or failure of automation. What additional research do
you want to see done in  this area? What are the problems that
your particular group are experiencing on a day-to-day basis?
And  finally,  do  you  have  any  ideas that  would  prove
potentially profitable for automation of wastewater treatment
plants in  this country—ideas that  you would  like to see
developed, even though you have neither the time or the funds
to do so?
   If we have  identified and prioritized, through some logical
process, a  set  of research needs, we will have  taken a stride
toward more cost-effective wastewater treatment.

-------
                               WORKSHOP  SUMMARY
                                             William A. Rosenkranz
                                 Director, Municipal Pollution Control Division,
                           Environmental Protection Agency, Washington, DC 20460
   The automation of wastewater treatment systems offers a
number of potential benefits including improved performance,
reduction  in size  and  construction cost of new systems,
improved reliability, more efficient use of operating personnel,
and  minimized operating  costs. These  benefits  are  clearly
"potential" since application of instrumentation and automa-
tion  in the wastewater field is still minimal. Compared to most
industrial  processing, automation  of  wastewater treatment
systems is in its infancy. The purpose of this Workshop was to
define how to move this specialized technology progressively
through adolescence and into adulthood.
   The  philosophy   for addressing  wastewater management
systems requires change. A treatment system should no longer
be considered as a marginal water pollution  control facility,
but  rather as a production facility for wastewater refining or
renovation. The medical and chemical industries developed the
philosophy  of pushing  the  newest  technology  into their
respective fields many years ago. It appears  now  that basic
water pollution control technology is sophisticated  enough to
adopt such a philosophy and thus attain this new goal.
   In  order  to  accomplish the automation of wastewater
control and  treatment systems in the most effective manner, it
is necessary to first  clearly define the research and develop-
ment needed and then establish priorities for implementation.
More than one hundred participants attended the "Workshop
on Research Needs for Automation of Wastewater Treatment
Systems". In  order  to  insure  a broad coverage  of all view-
points, the participants represented government regulatory and
research agencies,  universities,  operating engineers and man-
agers of large treatment systems, consulting engineering firms,
and  equipment manufacturers. The first day of the Workshop
was  devoted to the presentation and  discussion of  current
practice and research activities directed towards automation of
wastewater treatment systems. Experts from both government
and  private industry gave these presentations in order to
identify needed research and development. From these presen-
tations, it can be concluded that most  instrumentation for
control of wastewater systems is, of necessity, designed on an
empirical basis  because  of  a  lack  of more  fundamental
knowledge  concerning  such  factors  as dynamic behavior,
control  strategies,   component  reliability  and  cost-benefit
analysis. The effect of automation on the design and operation
of wastewater recycle systems and manpower requirements for
treatment plant operation was also considered.
   Working  parties  were organized to address the following
 subject areas: (1) automation  of wastewater collection sys-
tems, (2)  automation of biological  treatment processes, (3)
automation of physical-chemical processes, (4) automation of
sludge processing, transport and disposal, (5) computer appli-
cations,  and (6) evaluation of the effectiveness of automation.
   A number  of research needs were identified which were
common to most or all of the six subject areas. These common
research needs were:
   (l)The  Workshop recommended that an information clear-
     ing-house dealing with instrumentation and automation
     be established within EPA. For example, a great deal of
     instrument testing on a specific case basis is being done,
     yet results are not available to  the technical community
     on a broad basis.  A central location for  gathering and
     dispensing such information would be of considerable
     assistance to  those planning and using instruments in
     municipal wastewater systems.
   '(2)Development  of  efficient and dependable  sensors  is
      needed. This is a prerequisite to the implementation and
     verification  of virtually all control  strategies. Some of
      the  sensors  needing  improvement or  development in-
      clude:  sludge blanket level indicator,  settling velocity
      indicator, respiration rate sensor, suspended solids sen-
      sor, on-line  replacement for the BOD test and on-line
      analysers for ammonia, nitrate  and phosphorous.
   (3)Performance  specifications should  be developed  for
      sensors  and   instrumentation  as  a  guide to the  user
      community.  This could  be in the form of a  testing
      protocol and procedures  for  evaluation  of  sensor and
      instrumentation packages.
   (4)With  respect  to  treatment  processes  and  treatment
      systems, it was determined that a logical progression of
      research and  development should be  as follows:
      a. Development of dynamic  mathematical  models for
        individual  processes. In several  cases, this has been
        partially accomplished although many of  the models
        should  be  further  refined  using  the   latest  data
        available.  An important long-term goal is incorpora-
        tion  of the individual process models into an overall
        mathematical model for treatment plants.
      b. Tentative control strategies based on computer simula-
        tion using mathematical models should be developed.
        A small amount of work is now on-going  in this area.
      c. Control strategies should be evaluated at pilot scale to
        select the  most promising  for future demonstration.
        Such evaluation  has frequently  been limited in the
        past by lack of adequate sensor capability.

-------
      d.  The  most  promising  control  strategies  should be
         demonstrated at  full scale, including cost/benefit
         analysis.
   (5)A protocol should be established to evaluate instrumen-
      tation and automation system design and selection. This
      protocol should take into account all relevant aspects of
      the problem  such as direct cost savings, system per-
      formance and reliability, minimization of the consump-
      tion of energy and other resources, and the man-machine
      interface.
   (6)Instrumentation  and  automation  control  strategies
      should be  expanded to  include  interrelationships  be-
      tween liquid and  solids  processing,  stormwater and
      dry-weather flow control and treatment, and eventually
      area-wide wastewater management.
   Each  individual working party  identified  research and
development  needs specific to  the assigned subject area.
Highlights from the six groups are as follows:
   The importance of addressing the area of instrumentation
and automation in terms of a total  system  was identified by
the working  party  dealing with automation of  wastewater
collection systems. It is very important that the demonstration
of total control systems with the necessary operating strategies
incorporating  the newest  technologies be accomplished. This
could  even include  sophisticated  weather  forecasting and
tracking. The  role and benefits of flow equalization need to be
better documented and more widely applied. Coupled with
automation of treatment processes, this technique may have
significant plant performance advantages.
   The  automation of biological treatment processes group
identified  a  need for great improvement  in information
exchange, particularly from the operator to the manufacturer
and from the manufacturer back to the operator. Development
of performance specifications and acceptance  standards  for
instruments was  also emphasized. In addition, the  develop-
ment of adequate  mathematical process models,  practical
operating control strategies, real-time  monitoring,  and im-
provements in the man-instrumentation interface were areas of
research that were identified.
   The  matter of dynamic models  for process control and
monitoring  functions  was also  highlighted  in the  area of
physical-chemical systems. Control strategies and instrumenta-
tion  development  to  facilitate  implementation  of control
strategies were  two key  needs  identified.  Specific needed
research was identified for several treatment  processes includ-
ing chemical clarification, the deep-bed  filtration process and
granular carbon adsorption.
   Working  party deliberations indicated that sludge process-
ing, transport  and  disposal should be placed at  the highest
priority level. The need for development of a variety of sludge
quality sensors was  identified, particularly sensors to  measure
 settleability, dewaterability and other similar parameters.
    The  computer applications working party indicated  that
 improved capability to measure flow is needed throughout the
 wastewater control and treatment  system. This is  a  difficult
 technical problem and,  although advances such as  the use of
 sonic devices have been made in recent years,  no  major
 technological breakthrough appears to be on the  horizon.
 Improved flow measuring capability is a key to total system
 management. Development of control  strategy models, includ-
 ing reliability testing,  was  highlighted as a research need.
 Determination of specifications for computer  selection  and
 utilization of a centralized computer for data acquisition and
 report preparation were areas deemed worthy  of research. The
 need for increased educational activities to produce personnel
 who  understand  both  the computer and  treatment  plant
 operations was also stressed.
    The group dealing  with evaluation  of the  effectiveness of
 automation,  indicated  that  a cost-effectiveness  evaluation
 protocol is needed so that meaningful  comparisons of instru-
 ment applications  and  control systems can  be  made. Such
 comparisons  are vital  to selection of control  systems for
 specific  cases and for evaluating cost-effectiveness of alternate
 systems of control  and treatment. The concept of using a
 man-in-the-loop to control several satellite plants through the
 use of a  centralized computer and terminals was  proposed as a
 potentially cost-effective technique.
    The Workshop also discussed the matter  of eligibility of
 instrumentation and automation within EPA's Construction
 Grants program.  Recognizing that many  treatment plants
 involving large sums of money  are funded under the EPA
 grants program, a  need  for  detailed guidance concerning
 eligibility conditions for instrumentation and automation was
 identified. Computer applications are of specific concern.
    Although not identified as a research need, the Workshop
 noted that the  United Kingdom has recently established an
 instrumentation and automation Working Party for the pur-
 pose  of establishing  and advancing the state-of-the-art. Based
 upon this information, the Workshop recommended that EPA
 explore the possibility of setting up  a similar  panel in  the
 United States in order that we can have a direct interchange of
 information, ideas and technology  with the United Kingdom
 working group and similar groups throughout the world.
   In  conclusion, the  Workshop  developed  many  specific
 research needs related to the instrumentation and automation
 of wastewater treatment systems. It indicated a need for an
 information clearinghouse, international exchange of data, and
 projected a  new  philosophy  of wastewater renovation as
 opposed  to processing wastewater  to  the  minimum  quality
 requirements.  The cost-effective application of instrumenta-
 tion and automation to wastewater management systems will
be a key to implementing this philosophy.
                                                          10

-------
                         AUTOMATION
                                     OF
WASTEWATER COLLECTION SYSTEMS
                          Workshop on Research Needs
                  Automation of Wastewater Treatment Systems
             11

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                            AUTOMATED COLLECTION  SYSTEM
                                  INSTRUMENTATION NEEDS
                                                 Curtis P. Leiser
                      Manager of Computer Services, Municipality of Metropolitan Seattle,
                                  410 West Harrison Street, Seattle, WA 98119
 INTRODUCTION
   One need only review recent publications and conference
 agenda of technical organizations such as ASCE, WPCF and
 APWA to appreciate the growing interest nationwide and, to
 some degree, internationally in wastewater collection system
 control and treatment research. Centralized collection system
 automation is in  its  infancy. The number  of operating
 computerized systems can be counted on one hand. However,
 many  more are on the drawing boards for consideration or
 construction in the near future. The computerized systems
 have generally struggled through technical and funding prob-
 lems typical of research and development work. To this date,
 the systems which have been developed are not only atypical
 in design but also contain  significantly different  degrees of
 monitoring, control and the  ultimate form of automation
 which involves optimized simulation or modeling.
   This paper will concentrate on  the hardware instrumenta-
 tion considerations of these systems. Primary reference will be
 to  the results and  recommendations developed during a
 six-year  demonstration grant  study  between  the Environ-
 mental Protection Agency and Seattle Metro which culminated
 in a report titled  "Computer Management of a Combined
 Sewer System" (1). It was found during this study  that there
 was a considerable amount of research that could be accom-
 plished  to  overcome many of the  instrumentation problems
 and shortcomings which were revealed during the development
 of Seattle's computer-controlled system. It is appropriate, at
 this time, to present a short summary of the Seattle system, its
objectives and its equipment. More detailed information can
be obtained from the list of references at the end of this paper
(2-6).

SEATTIi METRO SYSTEM
   In  January of 1971, the Municipality of  Metropolitan
Seattle (Metro) first placed into operation a monitoring and
control  facility  termed CATAD or "Computer Augmented
 Treatment and  Disposal System." This centralized control
 system encompasses approximately a  ten-square-mile com-
 bined  sewer  system  in  and around the City of Seattle. The
 main objectives of the CATAD system are:
   1. To utilize the maximum storage capability of trunk and
     interceptor lines within a combined sewer system built
     to ultimate capacity so that overflows caused  by storm
     inflow are reduced or eliminated.
   2. To regulate daily  flows  to treatment plants, thereby
     aiding in the stabilization of the treatment processes and
     effectively increasing the  dry-weather capacity of exist-
     ing plants.
   3. To select the overflow points which will cause the least
     harm to receiving waters,  beaches and marine life during
     intense storms when overflows cannot be avoided.
   4. To eliminate the need  for or  reduce the cost of total
     separation of combined sewers which  would be espe-
     cially costly and disruptive to commercial and industrial
     areas of the city.
   5. To monitor and control mechanical equipment within
     remote stations while accomplishing the above objectives
     and,
   6. To retain dry- and  wet-weather flow data of component
     collection systems for subsequent identification of infil-
     tration problems and for potential charges in accordance
     with those flows.
   At the time the construction of the computerized monitor-
ing and control system began, a consideration portion of the
remote  equipment  was  already installed and  operating.  It
seemed  logical  to  incorporate  as much of that system  as
possible and interface the computer-control equipment to the
existing regulator controls.  Thus, two objectives would be
achieved, (a)  a minimum  disruption  of  existing control
equipment which was proven to be reliable and (b) the local
control  system within each station  could serve as a backup
                                                       12

-------
                                                                                   WASTEWATER COLLECTION SYSTEMS
control capability in the event  of computer or communica-
tions equipment failure.
   The existing  control equipment was primarily pneumatic,
while newer station designs included  electronic control equip-
ment to facilitate computer interface installation. Modifica-
tions to older stations included new precision  sensors,  com-
munication and  control equipment, to adapt existing regulator
and pumping stations for remote  control capability.
   Key elements of the system are pictured in Figure 1. At the
heart of the  CATAD system is  the  real-time process  control
computer  and  background  devices which  allow  for  the
programming of that computer. Specially designed interfacing
or connection equipment was installed to allow the computer
to communicate through cables  to many mechanical devices.
At the central console, information  is collected from  remote
stations  and is  converted to  visual displays. Control push-
buttons, alarm  and status indication lights,  radio and  tele-
phone communication equipment  and logging teletypes  com-
plete the components of the  central console. In addition, a
wall  map showing the geographic  locations of the collection
system together with general status-indication lights for each
station  combine to offer  the human  console  operator  all
information and control features which could be expected of a
computerized wastewater collection system.
   A second interface "black box" connects a series of water
quality monitors to  the computer. These monitoring stations
are located on the Green-Duwamish River south of Seattle and
monitor  such conditions as temperature,  dissolved oxygen,
conductivity, pH and solar radiation.  These data are fed to the
computer and are logged on a permanent paper form as well as
stored  on magnetic  tape for later statistical  processing. The
information thus becomes available to enable either the human
operator or,  through programming,  the computer, to select
overflow points  which would provide the least impact upon
receiving waters.
   A third interface connects the computer  to two  satellite
consoles  which  are compact versions of the  central console.
These satellite consoles are able to display the same data and
to command remote station  actions  independently  of the
central console.  The  Renton satellite  console is the  newest
system  addition and  has  been  given  more flexibility by
utilizing a mini-computer rather than a fixed logic  terminal, to
perform many functions. The mini-computer was  expected to
be adapted for  use  in process  control  loops  during  the
expansion of Metro's Renton secondary treatment plant, but
the $100,000 estimated cost for  the  control interface aborted
that plan until interfacing prices become more realistic.
   Thirty-seven remote stations communicate to the computer
through the last  interface. A telemetry control unit (TCU) at
each remote station acts as  an interpreter to collect, convert
and assemble station information into a form which  can be
transmitted over telephone  lines to  the computer.  Figure  1
shows the types  of information collected from three different
representative  remote  stations passing  through  a series  of
electronic units for transmission to the computer. Commands
return  in  the  opposite  direction to  control regulation  or
pumping  stations  in the  combined  sewer  portion of  the
collection system.

INSTRUMENTATION AND SENSORS IN
CONTROL SYSTEMS
   The objective of the data collection system is (1) to make a
continuous, accurate  measurement of all information neces-
sary to monitor the  safety and  operation of a station and
calculate required  flow and storage data to facilitate logical
control decisions and (2) to  transmit  that information with
some degree of security to  the central computer  system.
Sensors, or measuring devices,  must be  present to continuously
monitor hydraulic, meteorologic, atmospheric and water qual-
ity  parameters  for  optimum  control  of a  combined sewer
system. A minimum system must be capable of monitoring the
hydraulic parameters described below.

Flow Calculations
   Obviously the most important parameter to measure in a
sewage collection system is flow. Generally the more flow data
accumulated from the various elements of a collection system,
the  better  will  be  the control and predictive modeling
capabilities of  that system. A  measurement  of  water depth
together with known sewer  configuration  data  can  provide
open-channel conduit flow values by incorporating one of the
variations of Manning's equation.  More accurate system  flows
are calculated by measuring depth and using metering sections
such as the Palmer-Bowles  flume where  space is available or
Parshall flumes  where head losses are not critical.
   A measured  depth may  also allow the calculation of flow
through sluice gates  which  may be manually or mechanically
driven.  Added  to  the measurement  of  depth must be  the
position of the  control gate in the stream  flow. By incorporat-
ing these values with standard orifice formulae, the flow may
be calculated with some degree of reliability. Flow calculations
however,  can  be complicated  by upstream  or  downstream
backwater and submergence effects.
   Different measurements are necessary to calculate pumped
flow in a collection  system. Four measuring techniques  are
currently  being utilized in Metro's CATAD system. Wetwell
water  level and pump rotational speed allows calculation of
flow through standard pump curves provided by the equip-
ment manufacturer. Where  the pumping station force main is
long enough to cause sufficient head loss, force main pressure
may be measured and flow calculated  with calibration curves.
More expensive but  direct-reading meters such as a magnetic
flow meter or a calibrated velocity meter  are additional means
of  obtaining flow information  from pumping  stations  or
submerged conduits.
   A continuous water depth measurement  at one or more
locations  will  also  provide system storage data for use in
controlling and minimizing overflows. A single depth measure-
                                                         13

-------
          JACKGROl
          PERIPHERAL
         .DEVICES
INTERFACE
M
                                                                                        LEGEND

                                                                              O COMPUTER SUB  SYSTEM

                                                                              CD CENTRAL CONSOLE SUB SYSTEM

                                                                              C3 RIVER  MONITOR SUB SYSTEM
                                                                              £\ SATELLITE CONSOLES

                                                                              [""I REMOTE TERMINALS

                                                                              —- TELEPHONE LINES
                                                                              	INTERFACE "BLACK BOX"
                                                                              	 CABLING	
                                                                                                        I
                                                                                                        >
                                                                                                        5
                                                                                                        o
V
                                        INTERFACE
                                                                                                                 m
                                                                                                                 33
                                                                                                                 3D
                                                                                                                 m
TELEPHONI
LINES
1
WATER QUALITY
MONITORS

• jn fcOMPUTEFf F TELEPHONE
1 V
INTERFACED


CWALL MAP^
— ALARMS a STATUS
— LIMITED COMMANDS
— PLANT DATA ENTRY
1 	
r
7 X
/WEST POINT >
/ SATELLITE
/ CONSOLE

i
-"
* mmmmmm

/^CENTi
"V CONSC
. . f 	
WES
REG
STA
}
i r ~
i
i
i
t i
.T POINT
ULATOR
TIONS
rcu
| A T ^\ (REMOTE- LOCAL)
IAL \ v
>LE } \
L
—CRT DISPLAYS \ /
— LOGGERS
	 Al ARMS
— STATUS

1 — CONTROLS
r~
1
1
L-


< /RENTON
\ / SATELLI
\ / CONSOL




GATE
POSITIONS
' 	 — ' WATER-TDE
LEVELS
-L 	 ' EXPLOSION
| HAZARD
\ viisc.
A ALARMS
L\

ACU
1
J
__•.
»






DETACHABLE
CABLES —

" ~l~
1
1
1
1
r

WEST POINT
PUMPING
STATIONS
TCU
(REMOTE-LOCAL)
\A— DETACH ABLE
/ CABLES —
ecu
KATAD- LOCAL)

PUMP
SPEEDS
1
WATER 1 	 	
LEVELS | 	 "
FORCE MAM
PRESSURES
EXPLOSION
HAZARD

MISC.
ALARMS





1
1
1
1
1
f
MENT SYSTEMS
r ~ - -
i
RENTON
PUMPING
STATIONS
TCU
(REMOTE -LOCAL)
\/
ecu
(CATAD-LOCAL)

PUMP
SPEEDS
WATER
LEVELS
FORCE MAM
PRESSURES
EXPLOSION
HAZARD

MISC.
ALARMS







-------
                                                                                  WASTEWATER COLLECTION SYSTEMS
ment within a storage tank is a simple method of calculating
storage. In circular conduits the calculation is more complex
but  can  be simplified  by  having  more than  one  depth
measurement available.
   Rainfall is the minimum meteorologic data to be collected
and  supplied to an automated control system. Rainfall data
from each sub-basin will allow the analysis of infiltration and
will  provide some predictive information for control functions
in a  combined sewer system. The rain data also provides basic
design information for future construction.

Non-Essential Measurement Devices
   A sewage  collection  control system  must be primarily
concerned with hydraulic measurements described previously,
but depending upon the level of sophistication of that system
the measurement equipment described below may or may not
be present.
   Continuous monitoring of water quality within the collec-
tion  system and in its receiving waters would be of value, in a
control  decision to select a point of overflow, for example.
Quality  parameters such as dissolved oxygen level, suspended
solids,  temperature and  biochemical oxygen  demand  are  a
sampling of the parameters that  would be especially valuable
in determining  control  and design  decisions.  Automatic
sampling equipment may be part of this system and might be
triggered  either  by  local  conditions or  remotely  by the
computer control center.  Sampling devices may be an  alter-
native to continuous monitoring  and if so, their operation and
performance should be monitored as if the sampling equip-
ment were an integral  part of the collection system.
   The collection system may contain some type of treatment
process  entirely  separate from  the  major treatment   plant
facilities at the  end  of the collection  system. Small  local
treatment plants or  similar  facilities  to treat overflows  or
bypasses from the  system should be  monitored for chlorine
residual, for  example,  or  other  appropriate  performance
criteria.
   Two  or more gaseous parameters  may be measured in the
collection system. Explosion hazard monitoring is a feature of
the  Metro Seattle  System.  Its purpose is two-fold:  (1)  to
provide  a safety warning to protect personnel and facilities and
(2) to supply information for tracing sources of illegal dumps
of hazardous volatile  material into the collection system. No
method of treating  or disposing of hazardous materials within
the  system  is  currently being  applied.  A second valuable
measurement  which  currently is  very  difficult  to  obtain
without the use of an elaborate series of chemical tests similar
to an auto-analyzer is hydrogen  sulfide.  The purpose of this
measurement is to monitor both odor production and corro-
sion  within the sewage collection system.

CATAD INSTRUMENTATION PROBLEMS
   As one  might expect, there are a great many difficulties
encountered during a  research and development effort on the
scale of the Seattle Metro CATAD control system. The many
instrumentation and control problems have been condensed to
three general areas:
   1. Looking back  on  the  development  effort,  a  timing
     problem  was  evident.  The  majority  of  design and
     construction work  was  done between  1968 and 1970
     while some major changes  were taking place  in  the
     electronics and computer industry. The CATAD  system
     was too far along to reap many of the benefits of these
     advancements.
   2. The tendency  of  engineers  or  designers  to rely on
     existing technology, in Metro's system, meant specifying
     equipment which was generally  being  successfully ap-
     plied in industrial situations at the time. This resulted in
     a  larger  burden  being placed on  the interfacing and
     conversion tasks. The goal of having a local independent
     control system  at  each  remote  station  was excellent.
     However, the dissimilarity between  equipment created a
     great number of difficulties which will be described later
     in this paper.
   3. A  final  source  of problems  was  the  selection of  a
     different  set  of designers  and  suppliers  for  remote
     station  industrial equipment and centralized computer
     control equipment. An aero-space firm, ready with all
    ' the latest electronics technology, was on one side of the
      interface while an industrial, wastewater-oriented design
      firm was ready with older, but safer and more  reliable
      equipment, at the other side of the interface.
   At this point, the  types of equipment problems  will be
discussed in  more  specific  terms, leading  into a  list of
comments on  the type of  research which should be contem-
plated to alleviate these problems.
   One  broad category  of problems  developed  due  to  the
application of sensitive instruments and sensors in a  hostile
environment for which they were  not designed. Since most of
the industrial controls would be adversely affected by  humid-
ity and  explosive or corrosive  atmospheres, a controlled
environment  contained  within an expensive structure was
generally provided. These remote stations, in the 1960's, cost
an average of $200,000 each.  The new station price has now
been inflated to an average of $500,000. These high costs limit
the number  of-sites  where monitoring and control can take
place. In  Seattle's system, there are between  two and three
hundred  overflow points (some admittedly very small); how-
ever,  only approximately  ten percent  of those  sites  are
monitored  and  controlled  by the  CATAD  system.  Some
sensors, such  as  gate-position indicating devices, would be an
excellent selection for a normal environment but in the  humid,
occasionally  explosive  atmosphere  of  the sewage collection
system, tests indicated that the  equipment could not cope with
these conditions without a  prohibitive amount of money and
space being allocated  to  provide a special enclosure. Normal
utilization voltages, radio-frequency and other electromagnetic
disturbances that the  heavy  industrial-type  equipment was
                                                         15

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
immune  to,  played havoc  with  sensitive  computer com-
ponents. Costly damage done to five-volt diode-transistor logic
(DTL)  by  misapplication  of  line  voltage  during  routine
servicing pointed out a need for additional technician training,
warning messages and protective devices wherever possible.
   When applying sophisticated  computer control to a sewage
collection system, one naturally thinks of improved precision
capability. To obtain five-percent precision, some sensors and
control  equipment,  relatively  new  on the market,  were
incorporated  into  the  system. These devices tended to be
candidates  for  extensive and  repetitive calibration  and de-
manded a great deal more maintenance than was routinely
scheduled for station equipment.
   Interfacing, or  tying together the  computer and  remote
industrial-grade  devices, was  another major  source of diffi-
culties. The fact that an interface existed was  immediate cause
for  problems  because  the  separation  resulted in frequent
disputes as to whose responsibility a particular problem was.
Since  the computer equipment generally requires a  specific
range  for analog measurements,  it  turned out  that  a great
number of transducers was  required because of a  distinct
absence of commonality  between  instruments  offered by
different manufacturers and sometimes even within the sepa-
rate lines of  instruments available from year to year from a
single manufacturer. In an early attempt to prevent excessive
voltage from harming  the solid state electronic equipment at
each remote  station, a general-purpose industrial  interposing
relay  was  installed at every  contact-sensing  point.  Unfor-
 tunately, the relay coil itself (approximately 50  per  station)
generated an extremely fast induced electrical surge during
 contact activity. This electrical "noise"  caused a disturbance
 within the solid-state logic, which occasionally was of a serious
nature and was both difficult to pinpoint and  to correct. After
 a series of studies of the electrical noise problems, remedial
 shielding measures were successfully implemented (7, 8).

 RESEARCH NEEDS
    A battery of well designed and  tested instrumentation and
 sensors must be developed. These devices should be modular,
 heavy-duty, exhibit low long-term  drift, and should be easy to
 calibrate, maintain and replace. They must be designed with
 enclosures  and other protective features suitable for a hostile
 sewer environment; meaning they should be  explosion-proof,
 resistant  to  electrical  noise,  power system  voltage, large
 magnetic  fields  and  power   failures.  The  instrumentation
 should be digitally oriented and standardized, with no special
 one-of-a-kind components, so that interfacing and replacement
 problems are minimized. Ideally, this instrumentation should
 have  small space  and power requirements so that $500,000
 structures are not necessary to contain it. The new equipment
 should also be thoroughly tested in non-critical field locations.
    A  parallel effort in  this area would be to establish uniform
 standards or guidelines for instrumentation and sensors to  be
 used in collection system monitoring and control applications.
The  Instrument Society  of America  and American  Public
Works Association have  made some  fine  efforts towards
surveying the instrument field recently (9) but there has been
little or no attempt to establish standards in computer control
applications in wastewater systems. The Environmental Protec-
tion Agency sponsored an  excellent summary of automatic
sampling  equipment (10). However, the report would have
made an even greater  contribution if it had set some standards
and enforcement guidelines. Sampler manufacturers, as well as
instrument suppliers, will  continue to manufacture an infinite
variety  of equipment until uniform industry-wide standards
and guidelines are set by the Environmental Protection Agency
or a similar funding agency on a national scale.
   Another  research  effort  tied  closely  to  the  previous
suggestion is an  analysis  of  the potential  for applying
pre-programmed mini-computers in monitoring and  control
applications  in the  collection system. The  mini-computers
would replace complex arrays of relays, which are expensive to
install and maintain,  and  prone to failure. Industrial control
manufacturers  have  been   making  greater  use  of  mini-
computers in various applications. However, this equipment
has not  commonly been applied to the wastewater collection
and treatment field.
   Flow  measurement in  the collection system needs a great
deal more research. Metro  has relied  heavily on pneumatic
bubblers  for water elevation measurement but these devices
cost an  average of $3-5  per month to operate. Sonar water
level sensors which operate for about ten cents per month have
been  employed by Metro with a mixed degree of success and
failure.  A thorough investigation of the proper application of
sonar and other innovative  methods of non-contact measure-
ment would provide  a beneficial service to  the wastewater
collection automation goal.  Depth measurement, however, has
a drawback in  that flow  calculation is then required through
the  use  of  largely  empirical Manning's,  orifice  or weir
formulae.  A  "characterization"  technique  should   be  re-
searched which would allow the direct output of flow from a
sensor.  By incorporating  LSI (Large Scale Integrated) elec-
tronic modules into the primary sensor or tapered capacitance
probe techniques with direct digital output,  a heavy calcula-
tion burden on  the  control  computer is  thus avoided. A
researching  of velocity  measuring sensors,  which might  be
applied  to open-channel flow within sewer systems, would be
beneficial in achieving more reliable flow calculation. Magnetic
and  ultrasonic flow meters have  been found very useful in
generating reliable flow data where a full pipe  is available, such
as  in  a  pump force  main.  But precalibrated  and tested
ultrasonic flow meters are  not currently available and some
additional research in the application  of these devices would
prove valuable.
   Machine vibration and heat measurement  in pumping unit
control systems has been  applied successfully at Metro as well
as in other agencies. However, a limited amount of application
information  is available on  vibration monitoring of low-speed
                                                           16

-------
                                                                                      WASTEWATER COLLECTION SYSTEMS
 equipment (less than 1,000 RPM).
   Considerable  research has been completed and workable
 equipment  is  now available to  handle petroleum  products
 dumped into open waters. But oil spills are a more frequent
 problem  in combined  sewage  collection systems. Reliable
 explosion-hazard  monitors can now  detect  the  presence  of
 volatile products. Research is necessary to detect other illegal
 petroleum fractions, and, even more important, to recommend
 methods  of handling  illegal discharges before they can enter
 and foul treatment processes.
   Precise gate position  measurement is another  area which
 requires some additional  research. Mechanically or  hydrauli-
 cally driven sluice gates are easier to monitor than inflatable
 gates. However,  some problems  still exist  which could  be
 improved through additional research. Metro has employed
 multi-turn  potentiometers connected  to gate  drive gears  to
 produce  an analog  signal proportional to  gate  travel. The
 aerospace industry has used slide  wire technology with a great
 deal  of success.  However, collection  systems would benefit
 from additional study and the incorporation of digital encod-
 ing to minimize calibration and drift problems.
   A great deal of research is required to determine what type
 of monitoring and control equipment is  necessary for treat-
 ment  facilities within the collection  system. Such facilities
 would, for example, provide sedimentation, chlorination or
 other form of treatment to combined sewage overflows or
 pump  station  by-passes.  Water  quality   monitoring and
sampling within the collection system to determine treatment
levels and influence storage and control decisions is currently
not being done in a continuous on-line mode. These research
efforts  are  quite  similar  to  treatment   plant  control  and
monitoring problems and will be left for later speakers in this
workshop to discuss.
CONCLUSIONS
   There have been a significant number of promising develop-
ments in  the computer and  electronics  fields which should
make wastewater collection system automation much easier in
the future. Direct digital output sensors,  noise-resistant solid-
state  electronics, photo-optic isolators and other devices now
need  to be assembled into a  completed  system to serve as a
guide to other agencies seriously considering automation. The
EPA's Taft Center should  serve as a storehouse for the latest
automation   technology and  be  the  headquarters  for   the
issuance  of  new  standards and  guidelines  for control  and
monitoring instruments and sensors. Seattle Metro's pioneering
efforts in wastewater collection  system  control have shown
that once the research and development hurdles are overcome,
the automation concept holds  great promise  toward solving
the  problem of handling  urban  drainage  within combined
sewers. Metro  will continue to cooperate with the  Environ-
mental Protection Agency and other  municipal agencies  in
furthering wastewater collection system automation efforts.
References
 1.  Leiser,  C. P.,  "Computer Management  of a Combined  Sewer    6.
    System",  USEPA Report No.  EPA-670/2-74-022, National  En-
    vironmental Research Center, Cincinnati, OH (1974).               7.
 2.  Gibbs, C. V. and Alexander, S. M., "CATAD System Controls for
    Regulation of Combined Sewage Flows", Water &  Wastes Eng., 6,
    No. 8,46-69(1969).                                         8.
 3.  Leiser, C. P., "Maximizing Storage in Combined Sewer Systems",
    Water Poll. Control Res.  Ser., Project No. 11022-ELK, Washing-
    ton, DC (1971).                                             9.
 4.  Gibbs, C. V.,  Alexander,  S. M. and Leiser,  C. P., "System for
    Regulation of Combined Sewage Flows", Jour. San. Eng. Div.,
    Proc. Amer. Soc. Civil Engr., 98, SA6, 951-972 (1972).             10.
 5.  Mallory, T. W.  and Leiser, C. P., "Control of Combined  Sewer
    Overflow  Events", Paper prepared  for the 1973  APWA  Public
    Works Congress and Equipment Show, Denver, CO, September,
    1973.
   Mallory, T. W., "Seattle's CATAD", APWA Reporter, 41, No. 18,
   26-27 (1974).
   Metropolitan  Engineers, "Preliminary Final  Report-CATAD
   Noise  Analysis", Report to the Municipality of  Metropolitan
   Seattle, June, 1973.
   Boeing  Company, "Electromagnetic Compatibility Study of the
   CATAD System", Report to the Municipality of  Metropolitan
   Seattle, June, 1973.
   American  Public Works Association,  "Feasibility of Computer
   Control of Wastewater  Treatment", EPA  Contract  14-12-580,
   December, 1970.
   Sheiley, P. E. and Kirkpatrick, G. A.,  "An  Assessment of Auto-
   matic Sewer Overflow Samplers", USEPA  Report No. EPA-R2-
   73-261, Washington, DC (1973).
                                                           17

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS


                  PRESENT PRACTICE  AND  RESEARCH  NEEDS IN

    WASTEWATER  COLLECTION SYSTEM  DESIGN  AND  OPERATION

                                               James J. Anderson
                                President, Watermation, Inc.,  1404 E. 9th Street,
                                             Cleveland, OH 44114
INTRODUCTION
   The task of defining research needs concerning collection
systems seems somewhat  difficult, especially in the light of
current  practice.  Most current  sewer design  is  based on
steady-state,  mostly linear relationships, and sewer systems
operate with only gravity in control.
   Because of my experience, I  will confine my remarks to
existing sewer systems.
   The major problems in existing sewers today, in order of
importance to the homeowner-taxpayer, are 1) stoppage, 2)
inadequate capacity,  and  3) discharge of untreated polluted
water.  In  most cases, the untreated discharge  is because of
inadequate  capacity,  resulting in  the placement of relief
outlets.
   Malfunctioning of sewer systems has primarily been a result
of design  methods,  the long life of underground structures,
and  unpredictable urban  growth and development. Current
design  methods have already  been  mentioned. Many  of the
sewers in use  today were sized with now antiquated methods,
fifty to one hundred  years ago. It is rather surprising how well
such  systems are  performing  today.  In  many cities, the
combined  sewers have  served effectively  because they were
sized on the basis of estimated runoff from the tributary area,
and not on the basis of estimated wastewater  flow.
   With the advent of the first major interceptor sewer and
treatment plant construction early in this century, and even in
recent large  systems  constructed in  the 1950's, engineering
judgement, the public, and  economics allowed the use of
interceptor sewers which captured only 90 to 95 per cent of
the dry-weather flow. Treatment plants that  were constructed
had very limited capacity to treat excess flows during runoff.
   Our  work  has been to provide controlled dispatching
techniques  to  better use existing sewer  systems and to
determine what minimum improvements can  be made to
nearly eliminate untreated overflow. In order to accomplish
this, we have developed dynamic analysis methods for  design
and  operation.  We have  also reviewed new and old  sewer
system components  and selected those which  can be effec-
tively used in  dynamic operation of sewers. As one client put
it, we are assisting Newton in his previously unassisted control
of sewers!

COLLECTION SYSTEM COMPONENTS
   Components which can be used in a controlled combined
sewer system  include  automated regulators, relief sewers, and
off-line storage. Wet-weather treatment can be accomplished
by utilizing excess capacity in the dry-weather treatment plant
as well as additional wet-weather treatment facilities.

CONTROL SYSTEM COMPONENTS
   The control system components include  a rain gauging
network, sewer level measurement, a system simulation model,
a system optimization model, data acquisition and control
equipment, and a process control computer.

DESIGN METHODS
   Because  we  propose to  take advantage of the dynamic
characteristics of rainfall, runoff, and of the sewer system, we
must use dynamic design methods. Very briefly, we use design
dynamic  rain events in the simulated model to generate the
hydrographs and materials balance in an existing sewer system.
The  output from  the simulation is used  as input  to  the
optimization model.  The  results of  the  optimization  are
combined  with  engineering  judgement, which cannot be
computerized,  to select  a final design.  The  expected per-
formance of the final design is then analyzed by simulating a
number  of design rain events,  ranging from  light to  heavy,
which represent the spectrum of local rain events over a  period
of years.

RESEARCH NEEDS
  Assuming that research will precede practice by a number
of years, and that dynamic design and control  methodology  is
in  its infancy, a number of research opportunities  can be
foreseen in the area of our experience.
  Among the most significant are:
1. Determine where dynamic design methods apply and where
  steady-state or handbook methods will suffice.
2. Develop  dynamic design methods not requiring the designer
  of a small system to use a computer.
3. Assess actual  hydraulic conditions found  in sewers and
  generate  new equations describing these.
4. Define mixing, dilution and  transport  of dissolved and
  suspended  materials  in  sewers.  Perhaps  develop new
  methods for obtaining dynamic materials balances and for
  tracing pollution loads through the sewer system.
5. Develop  techniques to  predict likely amount and distribu-
  tion for a short future period (hours) during rainfall, based
  on actual measurements, for real-time control.
                                                       18

-------
                                                                                  WASTEWATER COLLECTION SYSTEMS
6. Determine  the  tolerance of treatment systems to shock
   hydraulic and  pollution loading, to  assist in determining
   economic trade-off between additional sewer control and
   degraded plant performance.
CONCLUSION
   Of  course, there are many  other  general subject areas
needing  research.   In  addition  there  is need  for  various
hardware and devices. For example, on-line on-stream sensors
have been a matter of interest for some time.
   However, it seems to me that until there is more widespread
application of dynamic  design and control of sewer systems,
the hardware, devices, and sensors will remain on the shelf or
inoperative in  the field. The latter will not generate the need
for their existence.
   The conversion of knowledge-in-being to knowledge-in-use
seems to be a stumbling block in the application of dynamic
methods  to sewers. Perhaps  this matter might itself benefit
from research.
                                                DISCUSSION
L. A. Schafer:
  The  CATAD  system is an  outstanding  example  of a
monitoring and control system. In the area of future needs,
however, why is digital compatability stressed? Although this
may be desirable, development of digital-type instrumentation
has  proven to be expensive and generally unsuccessful in  the
instrument industry and can be split off as a separate problem.

C. P. Blakeley:
  There is a distinct need for an  early-warning system that
advises the waste treatment plant  operator of unusual pollu-
tants in the collection system. This may not be essential for
the  very  large municipalities  where dilution  of a chemical
dump  or  spill in the collection system negates detrimental
effects  on biological  treatment,  but  in  smaller systems-
probably up to 50 MGD and even larger—it may be of value.
  No  really good sensors are available. Conductivity measure-
ment  will  provide a  warning  if the pollutant  is ionized. A
sampler, activated by the  alarm could provide a means to
identify  the  nature  of the pollutant while  the  record  of
conductivity  provides a reasonable handle on  the  magnitude
and volume of the spill.
   H2S analysis  of the atmosphere  above the  lift station
wetwell could indicate unusual contaminants if, under normal
operation, aerobic conditions maintain  at the station.  Again, a
sampler can be activated by the analytical device.
   These devices are available and can operate  unattended for
extended  periods of time. Their  use now may  help plant
operators and may also influence  design in the future.

Stephen P. Graef:
   One 01 the problems  in a collection system  is  that of
obtaining flow-proportional samples on each of several inter-
ceptors leading  into  a wetwell.  Needed  is  a method for
estimating flow in each interceptor so that a flow-proportional
composite sample  can be obtained.  The device must  be
upstream from points of recycle.

Kelly M. PeU:
   An Army  report, due for  publication  in  October  1974,
concerns side-by-side evaluation of various types of samplers,
tested  on  six  different  classes of  water. Objectives  of  the
report  were  to provide  methodology to compare  samplers
under standard conditions, and provide a basis for a customer
to select a sampler best suited for his requirements. The report
may also provide  information for  manufacturers to build a
better sampler.
 Russell H. Babcock:
   The motivation for major  manufacturers to design  and
 produce instruments for the special conditions encountered in
 waste treatment  facilities is not present. The market is very
 small. According  to a  report  by  A. D. Little in November
 1966, treatment plant and equipment represents about 20% of
 the  total expenditure  for collection and treatment facilities.
 An  article  entitled "The Worst Public Works Problem" by
 Edward T. Thompson appearing  in Fortune for December
 1958 used  figures from  Black and Veatch  to estimate  that
 treatment plant and equipment represents  13.5% of the total
 expenditure. The state  of the art survey concluded that 3.3%
 of plant cost was  for instruments. This included  installation
 which  is not normally a profit  center for  a manufacturer.
 Product shipment records from the Department of Commerce
 show  about  2%  of  plant  cost  is for  instruments.  Most
 manufacturers will agree  that approximately   50%  of an
 instrument  system sale is spent on panels and their mounting,
 piping and wiring. This  is a very low profit item normally
 furnished only because it is essential to a sale.  These figures
 multiply out as follows:
    20% X 2% X  50% = 0.2% of total expenditures using A. D.
 Little's data.
    13.5% X 2%X  50% = 0.135%  of total  expenditures using
 Fortune's data.-
 These  figures are for  process instruments and do not include
 laboratory instruments or supervisory telemetry which are not
 part of the market that  can support development of special
 instruments. Conclusion:  the market is not sufficiently large to
 support an independent instrument program.
                                                          19

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
Harry Torno:
   Would Mr.  Leiser please describe the algorithm for fully
automated  control  under CATAD—from rainfall and flow
input through control actions taken?

Joseph F. Roesler:
   Can you describe the storm water treatment centers? What
is needed to eliminate  the storm water  discharge  into Puget
Sound? How can automation play a  role here?


Mohamed Elsahragty:
   In order to achieve the highest benefits from the automa-
tion of wastewater collection systems,  it  is very helpful to use
optimization techniques, like dynamic  and linear program-
ming,  to screen the alternative system  operating strategies.
These techniques can be used  to select the real-time operating
rules from the  infinite  number  of alternatives that  are
technically  feasible.  Existing  optimization models are either
very expensive to use  (require  several hours on  the  largest
computers commercially  available) or too simple  to capture
the main features of the system. It is my opinion that the lack
of adequate optimization models  is a serious limitation  to
achieving the maximum benefits from  the existing automated
wastewater collection systems at minimum cost.

John F. Andrews:
   Would you  comment on the question of improving plant
performance by regulating  plant inflow as against accepting
variable inputs and improving performance by better control
of the plant?
   Both  authors have  raised  the  importance of adequately
considering the interface between the collection system and
the treatment plant. In this connection, there are  those who
might leap to  the conclusion that it would be highly desirable,
through  a combination  of sewer control  and balancing  tanks,
to obtain a constant wastewater flow rate  and composition
into the treatment plant. On the other hand, there are those
who would  concentrate  entirely upon control systems for the
treatment plant for  handling  of the "wild" or uncontrolled
inputs from the collection system.
   The discusser would  like to point out that there may be a
"best"  waveform  and  input  frequency for  the  inputs to
wastewater  treatment plants  and this may be neither the
"wild" nor constant  input. For example, there are instances in
other fields of science and engineering where cyclic inputs at a
particular  frequency have  been  found to give  improved
performance over constant  inputs.  Examples are the use of
alternating  instead  of direct  current  in electrical systems,
certain  chemical reactors which perform best with  cyclic
inputs and the surface renewal  theory for oxygen transfer. The
possibility  therefore exists  that  there  may  be  a  "best"
waveform and  input frequency for  the inputs to wastewater
treatment plants.
 Joseph F. Roesler:
   What difficulties do you face in explaining  to  the "City
 Fathers"  the  type  of computer  that  must  be  used  to
 accomplish  the job? Would  it  be  feasible  to  describe  the
 planned system in terms of  functional use, but  yet in  the
 layman's language so that the City Fathers can understand?

 Jack Matson:
   Has the  question of cost effectiveness been studied? What
 are  the quantifiable benefits  of an automated wastewater
 collection system, and do they exceed the costs?

 CLOSURE
 Curtis P. Leiser:
   Reply to   L. A.  Schafer:  Data  transmission  is almost
 universally done in digital form. By providing all original data
 signals also in digital form, we bypass the  need to provide,
 install and maintain expensive  interfacing and conversion units.
   Reply  to R. H. Babcock: I cannot argue with the analysis.
 Perhaps EPA  involvement in research for standardized instru-
 mentation would offset development costs.
   Mr.  Torno   is  referred   to  EPA  publication  EPA-
 670/2-74-022, "Computer management of a combined sewer
 system," pp. 140-152.
   Reply  to J. F. Roesler:  There currently are  no overflow
 treatment centers in Seattle. A rotary screening facility at  the
 King Street outfall was once  proposed but rejected as too
 costly to operate. The EPA is currently sponsoring research on
 numerous overflow treatment techniques. Overflows could be
 eliminated by providing additional offline storage, construct-
 ing  parallel interceptors at  critical locations and by  limiting
 storm inflow  through surface stormwater retention practices.
 The  computer  would  be  very  useful  in  measuring  flow,
 overflow and  storage rates,  and controlling or recommending
 control actions in offline storage facilities.
   Reply to J. Matson:  The reference given  above  devotes an
 entire  chapter to costs of  automated  control.  Quantifiable
 benefits of improved water  quality are obviously difficult to
 measure. It is  easier to base  such  decisions on a comparison of
 alternatives  for achieving a  certain degree of overflow reduc-
 tion. Comparisons from the referenced report show that the
 first 60-80% reduction is  far less expensive  using  automated
 control.
James J. Anderson:
   Reply to S. P. Graef: We have used stage-discharge relation-
ships, taking advantage  of the computer's capability to solve
complex   equations,  to estimate  flow.  The  equations  of
hydraulic  flow have been well defined in the literature and can
be used with good engineering judgment for practical applica-
tions.
   Reply to M.  Elsahragty: We have optimized a number of
combined and separate  sewered areas to obtain cost-effective
combinations of relief sewers, storage, control, and treatment
                                                          20

-------
for overflow pollution reduction. Our early modelling was very
complex until we learned which parameters and considerations
were significant.
  Reply to J. F. Andrews: Perhaps a new "figure-of-merit" or
"annual utilization  rate"  might be  used to answer  this
question.  The  plant removal in  pounds per hour could be
accumulated annually  and the  total divided  by  the  design
removal capacity in pounds per hour times the  number of
hours of operation. High removals as a per cent  of influent
may be achieved at  light loads, but the excess treatment
capacity  has  been  lost  forever.  It  is  my  belief  that  a
combination of load control and  process control are required.

  Reply to J. F. Roesler:  At one time the word "computer"
meant  "automation = loss of jobs." It  still has  a negative
meaning to a  few people. By  and large, however, "City
Fathers" are  progressive people who see the computer as
                                                                               WASTEWATER COLLECTION SYSTEMS
another  of  the complex  tools used in municipal administra-
tion. In my opinion, computers are more readily accepted by
elected officials than by  the engineering profession which is
only slowly learning how to use them. There are a number of
primers on computer science which could be modified for use
in the water pollution control field.
  Reply  to J.  Matson:  The  capital cost of an optimized
automated  collection  system in the  Cleveland Southerly
District was approximately $50  million. These construction
costs  are the  least possible  to eliminate  90 per  cent  of
combined sewer overflow over a ten-year period. The value of
benefits  for this reduction  have not, and, in my opinion,
cannot be quantified. For example, how can a value be placed
on  the aesthetics of a  creek without raw  sewage flowing
through the city? The selection of the automated system was
based on a  tenfold reduction of overflow at a cost of about
one-tenth of other standard engineering approaches.
                                     Report  of Working Party
                                                      on
                                         RESEARCH  NEEDS
                                                     FOR
          AUTOMATION  OF  WASTEWATER  COLLECTION  SYSTEMS
                                                 John A. Lager
               Vice President, Metcalf & Eddy, Inc., 1029 Corporation Way, Palo Alto, CA 94303
                                                 Harold Tonio
                     Office of Research  and Monitoring, Environmental Protection Agency
                                             Washington, DC 20460
   The following outline presents the problems and associated
 research  needs  on automation  of  wastewater collection sys-
 tems as  determined by  the working  party. A section on
 objectives,  applicable to both  combined and separate sewer
 systems,  has been included to place the problems in perspec-
 tive.

 OBJECTIVES
 1. Reduce or eliminate untreated overflows or bypasses.
 2. Provide  measurements  and  data  logging  of  flow and
   pollution characteristics for:
   a. NPDES compliance
   b. Legal protection
   c. Normal state and federal reporting purposes
 3. Minimize flooding.
 4. Equalize  flow  and  pollutant  loadings  arriving  at  the
   treatment facilities.
 5. Improve system operation—that is,  provide  for the opti-
   mum  utilization of the  existing  sewer system and its
   inherent  hydraulic  capacity in  such  a  way  that  the
   treatment  facilities are most  effectively used and  that
   overflows or bypasses are minimized.
6. Detect and monitor infiltration and/or inflow.
7. Detect system abnormalities, including stoppages and  mal-
   functions in regulating devices or other system components,
   and hazardous conditions in the sewer system,  such as the
   presence of H2S or other gases.
8. Provide  a  .data  base for  the  cost-effective planning of
   improvements to, or extensions of, existing systems.
9. Save on costs and improve system utilization by  pumping at
   highest efficiency,  early detection of faults or malfunctions
   in  the  system,  and  source detection of contaminants
   entering  the system at random times or places.

PROBLEMS
   The  following  is  a list of problems  considered  to be
important by the working party if the stated objectives are to
be met. These problems are not necessarily in priority order.
1. Equipment limitations, particularly the sensors  required to
   meet specific needs, such as flow or  load equalization: Of
                                                       21

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
   primary importance are flow measurement capabilities and
   the necessary  sensors to operate regulating devices within
   the sewer system. Considerable attention was directed at
   the conference  toward  the  limitations of sensors.  The
   working party believes, however, that the sensors available
   on the market are adequate, except for  flow measurement
   devices, and that the primary problem is an engineering one
   of devising  the  proper equipment  to convert the sensed
   signal at a data point to data in a central computer system.
 2. Lack of understanding the impact of automation on system
   design  and operation: Most sewer systems are designed on
   the basis of peak flow assumptions when, in fact, the urban
   runoff phenomena are highly time-variant.  The  uneven
   nature of the rainfall  runoff  process makes it extremely
   attractive  to use automated systems in order to maximize
   flow through that  system and/or to minimize the impact of
   pollutant discharges from overflows.
 3. Inadequate rainfall forecasting: The primary problem is one
   of predicting on a very short-term (1- or 2-hour) basis what
   the impact of rainfall will be on a catchment: How intense
   will  the  rain  be?  From which direction will it come? In
   which  direction will it travel  as it  crosses the catchment?
   How quickly will  it  abate? And what is the impact of this
   rainfall travel  and  intensity  on  the  runoff  from  that
   catchment?
 4. Lack of real-time data on the cost-effectiveness  of  auto-
   mated control of collection systems, and hence inadequate
   justification  of the  use of an automated system:  The
   working  party believes, however, that there are opportun-
   ities for the joint automation of treatment facilities and the
   collection system  so that these systems  are jointly cost-
   effective.  These  opportunities  should be  given  careful
   consideration.
 5. The need to improve  existing regulators and their design
   and, in some  cases,  to provide better regulators, such as a
   replacement for fabridams: While some  work is under way
   in this direction (the  swirl concentrator is an example),
   much still remains  to be done on improvement of regulators
   and  the control and sensing devices associated with them.
   Moreover, the effect of the imposition of controls, such as
   rapidly closing gates, on the physical integrity of the system
   has been inadequately  defined. For instance, such gates can
   introduce hydraulic  transients and water hammer condi-
   tions,  or pressure  conditions within the piping that could
   cause serious damage to the system. Better definition is also
   needed of the limitations on control devices that must be
   applied when they are installed in collection systems.
 6. The general reliability  of automation  systems and the cost
   of  systems related  to their  overall  reliability has  been
   inadequately defined.  It is  possible,  of course, to design
   control and telemetry equipment that will  provide per-
   formance similar  to that used in submarines or missiles.
   However, the cost  may be prohibitive and unjustified.
 7. There  is a  general lack of standardization  of regulators,
  hardware, computer hardware, and the programs or even
  the programming languages used when developing models,
  for operating in an  automatic model wastewater collection
  system:  The performance requirements  for system com-
  ponents  are  not  adequately  defined,  nor  are  system
  specifications  available  so  that purchasers of computing
  equipment  are  adequately protected in case  of either
  software or hardware failure. In this regard, it  is important
  that there be an interface with other engineering disciplines
  which have already solved many of the problems in process
  control of systems. For example, the petroleum and textile
  industries run highly automated process control operations.
  Their  engineers  have largely solved  many of the same
  problems that  we will face, such as telemetry and analog-
  to-digital conversion devices (and their interface with small
  computing systems).
8. Maintenance and personnel requirements that are unique to
  automated wastewater collection systems are largely unde-
  fined, not only in  numbers but also  in the skills required
  when  such  automated systems are installed. Finally, while
  not a research  problem, the  working party believes that
  state and  federal organizations should better define  the
  standards on  combined sewer overflows and  stormwater
  discharges,  so  that  those who   must design collection
  systems  to deal with   pollution  abatement  from these
  sources know  what their objectives are and have something
  on which to base valid cost comparisons.

PRIORITY NEEDS
  The following needs have been ranked in priority sequence
as determined within the working party  through  open  discus-
sion and critique.
  1.  A determination should be made of the relative balance
    between the benefits to be accrued from the automation
    of the collection  system and from treatment  facilities
    improvements (automation of treatment facility or some
    other upgrading).  To do this, some definition of overflow
    standards  will have to  be considered in terms of their
    ultimate impact on receiving water quality.
2. A  set of uniform standards and guidelines for instrumen-
   tation and sensors  must be established. In this regard, the
   standards  developed, particularly  for sensors, should  be
   directed  toward  operation and  performance,  and not
   toward technical details on how a piece of equipment is to
   be manufactured. An excellent example is the  publication
   "Specifications for  an  Integrated Water Quality Data
   Acquisition System,"  (Eighth Edition,  January  1968)
   developed by  the Instrumentation Development Branch,
   Methods Development and Quality  Assurance Research
   Laboratory, USEPA, National Environmental Research
   Center,  Cincinnati,  Ohio. Computer standards  should
   include a language specification.
 3.  The demonstration, on some existing wastewater systems,
    of the integration between collection system  automation
                                                          22

-------
                                                                                  WASTEWATER COLLECTION SYSTEMS
   and plant performance in a plant that is at least partially
   controlled by the same computer system as the collection
   system,  is needed.  Present opportunities  for  doing this
   exist in  Seattle, and future opportunities will be available
   in  Cleveland, San Francisco, Detroit, and perhaps Phila-
   delphia.
4. Operating strategies must  be  defined for  a  variety of
   automated  systems, considering  both the  system as a
   whole  and  the  responses of  the  existing  treatment
   facilities to  the varying loads that could be expected in
   the treatment of combined sewer  overflows. These strate-
   gies should reflect the difficulty of operating a large-scale
   model on a  small process  computer.  One such strategy is
   now being developed by Colorado State University for the
   City of San Francisco and should be considered.
5. The  range  of system sizes or characteristics that lend
   themselves to  automation of the  collection system needs
   to be identified. Obviously, some  systems are too small to
   warrant automation and others have a physical makeup
   that is unsuitable for automation. Also, the software, in
   the  form  of  simplified  models to drive  the  control
   systems, needs to be developed. Such models as the EPA
   stormwater  management model are much too complicated
   and  require too much computing power to be useful as
   control  models.
6. There should be a determination whether the equalization
   of flow or pollutant loading is a valid objective in separate
   sanitary  sewer  systems that may become overloaded by
   virtue of infiltration or inflow,  or  by inadequate capa-
   cities.
7. The effects on various wastewaters of in-system storage
   for a  range  of time  periods must be  determined. For
   instance, hydrogen sulfide production,  solids deposition
   and resuspension,  odors,  and  the possibility of  slug
   loadings and their  influence on the treatment facilities,
   should  be  studied for  both  combined  and  separate
   sanitary wastewaters.
8. Existing combined sewer  regulators  should  be  further
   evaluated with a view toward automation.  Research in
   improving  these  existing  regulators  and  also  in  the
   development of new regulators, such as a replacement for
   the fabridam, needs to be accomplished. The fabridam, in
   spite  of its simplicity, places in the system some unde-
   sirable  hydraulic characteristics  and is not totally amen-
   able to automatic control (it is very difficult to determine
   the actual level of the fabridam crest).
 9. Improved  rainfall  prediction or  projection techniques
   (anticipatory  modeling) must be  developed.  The use of
   radar  as a predictive  device for short-term pattern  and
   direction-of-rainfall relationships should be  investigated.
10. A vigorous technology transfer program in the combined-
   sewer  area and opportunities for  programmed collection
   system control should be stressed. As a result of the past
   seven   years of  the  combined-sewer  research  program,
    much   technology  has  been developed which needs dis-
    semination  in  the same  way that  treatment  process
    technology is  broadcast  to the profession  by the Tech-
    nology Transfer group.
                                                          23

-------
                       AUTOMATION
                                   OF
BIOLOGICAL TREATMENT PROCESSES
                        Workshop on Research Needs
                Automation of Wastewater Treatment Systems
           25

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                   DYNAMICS AND  CONTROL
                       OF  BIOLOGICAL  TREATMENT  PROCESSES
                                                 John F. Andrews
                  Environmental Systems Engineering, Clemson University, Clemson, SC 29631
 INTRODUCTION
   The need for consideration of dynamic behavior in both the
 design and operation of biological processes used for waste-
 water treatment  is frequently greater than that for industrial
 processes because of the large temporal variations which occur
 in wastewater composition,  concentration, and flow  rate.
 However, our understanding of this dynamic behavior and how
 it may be modified through the application of modern control
 systems is in  its infancy. Gross process failures are all too
 frequent and even when these are avoided, it is not unusual to
 find significant variations in process efficiency, not only from
 one  plant to another but also from day-to-day and hour-to-
 hour in the same plant.
    Dynamic mathematical models are usually necessary for the
 description of time-variant phenomena, as commonly encoun-
 tered in biological processes, and increasing efforts are being
 devoted to their development. The models usually consist of
 sets of non-linear differential equations for which analytical
 solutions are not available and this is one of the major reasons
 why the dynamic behavior of biological processes has not been
 adequately considered  in past years. Being practical people,
 most environmental engineers  have  said  "Why  develop  a
 dynamic mathematical  model when it is not possible to obtain
 a  solution?" However, computer  simulation has  largely
 eliminated this bottleneck and the current problem is not so
 much one of being able to  obtain a solution as it is to insure
 that the model adequately describes the dynamic behavior of
 the process being simulated.
    When the dynamic behavior of a plant has been defined, the
  environmental engineer  then should become interested in
  modifying this  behavior so that it  will  conform to  some
  desired  behavior. As  illustrated in  Figure 1, this can  be
  accomplished through  both process design and the incorpora-
  tion of control systems. When the characteristics of the plant
  to  be  controlled are  fixed,  as for  existing plants, the
   improvements which can be obtained may be limited, fre-
   quently  because of a  lack of flexibility in the plant design.
However, in the design of a new plant it is possible to strike a
better balance between the effort and expense devoted to the
plant and that devoted to the control system. The major point
of significance  is  that  both  the  control  system  and  the
controlled plant affect the plant outputs and the two should
therefore be designed as an integrated system.
   A literature review of present practice and current research
on the dynamics and control of all biological processes is not
possible  within  the  space  constraints of  this paper  and
attention will therefore  be  devoted to only the  two more
common processes used in the U. S., these being the activated
sludge process and anaerobic digestion.
ACTIVATED SLUDGE PROCESS
   The activated sludge process is the  most commonly used
process for the treatment of wastewaters and consists of two
units, an aeration basin and a sedimentation basin. The three
major  inputs  to  the  aeration basin  are  the  wastewater,
concentrated activated sludge from the sedimentation basin,
and air. The microorganisms in the activated sludge react with
the organic pollutants in the wastewater and oxygen in the air
to produce more activated sludge, carbon dioxide, and water.
The  effluent from the aeration basin flows to the sedimenta-
tion basin  where the activated sludge is separated from the
liquid  phase.  The process effluent consists of  the clarified
overflow from the sedimentation basin. This basin also serves
to concentrate the solids  which settle to the bottom of the
tank for recycle to the aeration tank.
   The recycle of concentrated sludge from the sedimentation
basin to  the  aeration basin is an  essential  feature of the
process. Recycle serves the purpose  of both increasing the
concentration of microorganisms in the aerator and maintain-
ing  the organisms in a physiological condition such that they
will readily flocculate. However, recycle has also resulted in
 difficulties in understanding and modeling the process since it
                                                          26

-------
         TIME VARYING
         INPUTS
    (a)
                                                                  TIME VARYING
                                                                  OUTPUTS
                        INPUTS
UNCONTROLLED
PROCESS
                                                    OUTPUTS
                                      (b)
10


n 3

Hz
 '


-------
AUTOMATION OF WASTE WATER TREATMENT SYSTEMS
                         o
                         UJ  
-------
                                                                               BIOLOGICAL TREATMENT PROCESSES
           300
        BATCH  REACTOR  SIMULATIONS
        DEMONSTRATING  THE  EFFECT OF
         INITIAL STORED MASS  CONC.
                                                        initial  concentration
                                                        750 mg/1
                                                            no rapid  transfer of
                                                            substrate  to  stored mass
                                              8
     12
TIME,  HRS
16
20
24
Figure 3. Batch Reactor Simulations Demonstating the Effect of the
Initial Stored Mass Concentration

change in the settling characteristics of the sludge since these
are dependent upon, among other factors, the specific growth
rate  of the   sludge  which is  a  function  of the  sludge
concentration  in the aeration basin. The relative proportions
of stored,  active, and inert mass may also change if there are
substantial changes in the specific growth rate.
   Bryant (2) was the first to develop a dynamic model of the
secondary  settler to  express the  above interactions on a
quantitative  basis. He also incorporated  in  his  model an
expression  presented  by Pflanz  (3)  for prediction of  the
suspended solids in  the settler effluent as a function of the
solids flux for the settler. A modificatioof Bryant's model has
been incorporated in the process model developed by Busby
and Andrews (1).

Control Strategies
   In the conventional activated sludge process, the operator
has a relatively limited choice of control actions, these being:
(1) air flow rate, (2) sludge recycle flow rate, and (3) sludge
wasting rate. Of these, air flow rate is primarily of importance
in controlling process economics and does not appear to  have
     much  effect on process efficiency as long as the  dissolved
     oxygen in the reactor remains above some  minimum level.
     Exceptions  to this may be  encountered  in  the control  of
     filamentous  organism  growth and  the  use  of high purity
     oxygen in lieu of air. Common strategies for the control of air
     flow rate  include  the regulation of the air flow rate  in
     proportion to the wastewater flow rate or the regulation of air
     flow to maintain a constant dissolved oxygen concentration in
     the aeration basin.
        Variation of  the sludge recycle flow rate in proportion  to
     the wastewater  flow rate is  also a common control strategy.
     However, this is not always successful, especially when poorly
     settling sludges  are encountered, since it does not take into
     account possible changes in the concentration of sludge in the
     recycle flow. An increase in  the return  sludge flow rate can,
     within limits, increase the mass of sludge in the aeration basin;
     however,  it also increases the solids loading to the secondary
     settler as well as creating additional turbulence in the settler
     and can sometimes result in an increased carryover of solids in
     the process effluent.
        There  is  a net growth of sludge in  the  activated sludge
                                                        29

-------
                INFLUENT
              WASTE WATER
                                                            H
                                                            5

                                                            o
                                                            •n
                                                            S
                                                            m
                                                            3)
                                                            S
                                                            m
                                             CLARIFIED
                                              EFFLUENT '
                                             SEPARATOR
GATE VALVE

-------
                                                                                    BIO LOGICAL TREATMENT PROCESSES
process  and  sludge  must  therefore   be  intermittently  or
continuously wasted from the system.  One control strategy is
to waste that amount of sludge each day which will maintain a
constant mass of sludge in  the  aeration basin; another  is to
waste sludge whenever  the  sludge blanket level in the settler
exceeds a certain depth. Sludge wasting may also be used when
poorly  settling or bulking sludge  is encountered in order to
prevent the sludge  blanket  level from rising in the separator
until  sludge is  discharged  in the effluent.  However,  on a
long-term basis this may be the  wrong control action, since, if
the bulking is due  to an overload of  organic materials, as is
frequently the case, sludge wasting will decrease the amount of
sludge in the reactor  thus resulting in still  further overloading
with possible process  failure.
   Two or more of the above mentioned control actions may
be combined and an example of  this has been given by Brouzes
(4)  in  which  he controlled sludge  wasting to maintain a
constant specific growth rate of  the sludge.  Using a special-
purpose analog  computer, he regulated and measured the air
flow  rate  required to  maintain a constant dissolved oxygen
concentration  in  the  aeration  basin.  Assuming a constant
oxygen transfer efficiency, the  computer  then calculated the
oxygen uptake  rate  and related  this,  using empirical coeffi-
cients,  to  the  sludge  production rate. The computer then
calculated  and  controlled  the  rate  of  sludge  wasting to
maintain  a constant specific growth  rate of the sludge. A
safety override control was provided to actuate sludge wasting
whenever the sludge  blanket level exceeded a certain depth in
the settler.
   The  step feed activated  sludge process (Fig.  4) permits a
fourth  control action  to be  taken, this  being the ability to
regulate the points at which wastewater  is added along the
length of the aeration  basin.  The basin is  usually constructed
in the "folded" fashion shown  for economy of construction.
This control action is especially effective for poorly settling or
bulking sludge which can lead to  process failure by loss of the
activated sludge in the overflow from the sedimentation  basin.
Andrews and Lee (5) have  illustrated the value of step feed as
a control  action by computer simulation using a simplified
 dynamic   model.  Figure  5  shows the   transient effect  of
suddenly shifting from an  operational mode where all of the
 wastewater  is  admitted to  stage one  (see Fig. 4)   to an
 operational mode where  all of  the  wastewater  is  equally
 divided between stages two and three. The sludge in the  settler
 is rapidly transferred to the aerator and  there is also a rapid
 decrease  in  the  solids flux to  the  settler. Both  of these
 responses  would have the short-term effect (hours) of decreas-
 ing  the mass of solids carried  over in the effluent from the
 settler. These predictions are qualitatively verified through the
 field studies reported  by Torpey (6) in his work on the step
 feed process at the Bowery Bay plant  in New York City.
 Torpey also demonstrated that there was a long-term  (days)
 improvement in the settling characteristics by application of
 this control action.
   The increasing use of high-purity oxygen  processes in the
U. S., coupled with the increased installation  of computers  in
wastewater treatment plants, offers the possibility of control
from a variable more fundamental than any now in common
use.  This is the  specific  oxygen  utilization rate which  is
expressed as mass of oxygen utilized per unit  of time per mass
of sludge in the aeration basin. This is a direct indicator of the
"activity"  of the  microorganisms  in the aeration basin.  It
could be calculated on a continuous basis from  oxygen and
solids balances on  the aerator. The oxygen balance would  be
relatively  easy to  perform for the closed type of high purity
oxygen processes  since  these are,  in effect, on-line respir-
ometers.

ANAEROBIC DIGESTION
   The  anaerobic   digestion  process  is  widely used  for the
treatment  of  organic  sludges  from municipal  wastewater
treatment plants. The process has several significant advantages
over other methods of waste treatment and among these are
the formation of useful by-products such as methane gas and a
humus-like slurry  well suited  for land reclamation. Unfortun-
ately, even with these advantages the process has in general not
enjoyed  a good  reputation because  of  its poor  record with
respect to process stability as indicated  through  the years by
the  many reports  of "sour" or failing digesters. The  major
problems with the process appear to  lie  in the area of process
operation as evidenced by its more successful performance in
large cities  where there  is less variation in the influent sludge
and skilled operation is more prevalent. At the present  time,
operating practice  consists only of sets of empirical rules and
there is a significant need for a rational control strategy to put
process operation  on a quantitative  basis.

Dynamic Model
   The present dynamic model of the process has evolved over
the  past ten years  and is  discussed in more  detail in  the
publications of Andrews (7,  8), Andrews and Graef (9), and
Graef and Andrews (10, 11). The model, summarized in Figure
 6, was developed  from material balances on the biological,
 liquid, and gas phases of a continuous-flow, complete-mixing
 reactor. The components on which material balances are made
 are given below:
   Biological Phase
    1. Organisms
   Liquid Phase
    1. Volatile acids.
   2. Conservative toxic agent.
   3. Cations.
   4. Bicarbonate.
    5. Dissolved carbon dioxide.
    6. Methane.
   Gas Phase
    1. Carbon dioxide.
    2. Methane.
                                                            31

-------

Is  Ha
,.  1300

00
a:
o
   1200
> 200
       £
  150
  to  1100
   1000
    900
       to
           100
50
              0
                       t « 0; ALL FEED TO REACTOR I   '
                       t • 0 ;  FEED DIVIDED EQUALLY
                               JO REACTORS 2 AND 3
                                ORGANISM MASS
                                  SEPARATOR,
                                   "Sec
                                                                -i
    200 HG/L
    200 MG/L
KD - 0.005 HRS'
fl - 0.2 HRS"1
    0.5 GH X/GM S
    150 CU M/HR
    131* SQ M
V,,V2,V3,Vi, - 150 CU M
KE • 10
R - 0.25
                                                     S0
                                                     KS
                                                                 Y  -
                                                                 F  -
                                                                 A  -
        SOLIDS  FLUX TO
         SEPARATOR, SF
                                          ORGANISM MASS IN
                                          REACTORS,
                                                                                       2.0
                                                                                           o
                                                                                           o
                                                                                           O  m
                                                                                           CO  »
           0.5       1.0       1.5      2.0      2.5       3.0       3.5
              ELAPSED TIME AFJIR CHANGE  IN FEED  PATTERN,  HRS'
                                                                                       1.5
                                                                                           C?
                                                                                           f»
                                                                                           m
                                                                         1.0
                                                                            0.5
                                                                            CO
                                                                            o
                                                                            I
                                                                            TO

-------
                                                                     BIOLOGICAL TREATMENT PROCESSES
V, V6
PT.D


(C02)DO
Txo *
F-v .
Ka, K, K| , KJ_Q
x0 r
So „
Y
>A.KS,K,
F.V.D


GAS PHASE
dPC02 y PC02
dt T VG o v8 °
Q = OCOt+ OCH4+ QH20, QC02 « - DVT0
»
M T
o  s ^™" IX ~ X ) T MX ~ K 1
dt v Yx/s
^ ~ Ka HS
R»=YCOt/X>* X. ' OCH4= D V YCH4/X P X,

Q
PC02
z, „

(CO^,
H*
x, r
S' t

Figure 6. Summary of Mathematical Model and Information Flow
                                              33

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
   There are strong interactions between the phases as well as
internal to each phase. These interactions must be considered
if the model is to predict the dynamic response of the five
variables most commonly used for predicting process  condi-
tion which are: (1) Volatile acids concentration; (2) alkalinity;
(3)  pH; (4)  gas  flow  rate;  and (5) gas  composition. The
following relationships were  used to express these interactions
on a quantitative basis:
   1. Yield coefficients.
      a.  Moles organisms produced/mole volatile  acid utilized.
      b. Moles carbon dioxide produced/mole organisms pro-
         duced.
      c.  Moles methane produced/mole organisms produced.
   2. Kinetics of organism death due to a conservative toxic
      agent.
   3. Inhibition  function for  relationship  between organism
      growth rate and un-ionized acid concentration.
   4. Equilibrium  relationship  between  ionized  acid,  un-
      ionized acid and pH.
   5. Equilibrium relationship between dissolved carbon diox-
      ide, bicarbonate and pH.
   6. Charge balance on ionic species in solution.
   7. Henry's law.
   8. Mass  transfer equation for transfer  of carbon dioxide
      across the gas-liquid interface.
   The  model  has  been   kept as  simple  as possible  by
considering  the  conversion  of volatile acids  to  methane  and
carbon dioxide as the rate limiting step. It is also assumed that
there is no lag phase, endogenous respiration, or  inhibition by-
products. The model is also  restricted to a pH range of 6 to 8
and does not consider the precipitation or dissolution of solid
chemical phases such as calcium carbonate.
   Two key features of the model are the use of an inhibition
function in lieu of the  Monod function to relate volatile acids
concentration and specific  growth  rate  for  the  methane
bacteria and consideration of the un-ionized  fraction  of the
volatile  acids as both the  growth  limiting substrate and
inhibiting  agent.  The  use  of an  inhibition   function  is  an
important modification  since it  enables the model to predict
process  failure by  high concentrations  of volatile acids at
residence times  exceeding the washout residence time. Con-
sideration of the un-ionized fraction of the volatile acids as the
inhibiting agent resolves the conflict which has existed in  the
literature as to whether inhibition is caused  by high volatile
acids concentration or  low  pH. Since the concentration  of
un-ionized  acids  is a  function  of both  total volatile acids
concentration and pH, both are therefore of importance.
   Digital  computer simulation  studies  provide  qualitative
evidence for the validity of the model by predicting  results
which have been commonly  observed in the field. Among the
results  predicted  by the model are: (1) At  steady state,  an
increase in  the alkalinity concentration in the digester  results
in an increase in the operational levels of pH and volatile acids;
(2) failure of the process can occur through hydraulic, organic,
and  toxic material overloading; (3) the course of failure, as
evidenced by the behavior of the operational variables, pH,
alkalinity (HCOs), volatile acids  concentration  (S), and gas
composition  is qualitatively the same as that observed in the
field; (4) stopping or reducing the flow to the reactor, the
addition of base, or recycle  of sludge  from  a  second-stage
reactor are effective techniques for curing failing digesters.

Process Stability
   Hybrid computer simulations were used to analyze process
stability by simulating digester overloading and observing what
changes in design and operational  characteristics  provided the
best buffer  against process  failure. The analysis procedure
involved making a change in a  digester parameter, such as
increasing the residence time (0), followed by simulating larger
and larger  step increases in digester loading  until  failure
occurred. By plotting the  locus of points of critical substrate
loading rates vs.  reactor residence time or other  parameter, it
was possible to obtain a semiquantitative measure of digester
stability. An example of a stability analysis for the effects of
residence time and alkalinity is given in Figure 7.
   In addition to the increases in stability which are  obtained
by increases  in  residence time or alkalinity, stability also
increases sharply with  an increase in the concentration of
methane bacteria in the  digester. This increased concentration
can  be attained by increasing the influent substrate concentra-
tion (sludge thickening)  or by recycle of concentrated digested
sludge from  a second stage. It is significant that three of the
measures for improving  stability,  increased residence time,
alkalinity,  and  influent substrate  concentration, can  be at-
tained by sludge thickening.  Other simulations indicated that
process stability could be enhanced by the incorporation of
suitable control systems.

Control Strategies
   Operating  practice   for  the  anaerobic  digester  consists
primarily of manual procedures, with notable exceptions being
closed-loop temperature  control and the use of density control
on the feed  sludge.  However, an outstanding  step toward
automation  of these  manual  procedures is  the near-real-time
computer control system developed by Philadelphia (12). This
system uses  a remote terminal as an interface  between the
plant operator and the computer. The operator inputs his data
into the terminal and the computer then reacts to the input
data using the internal  program of empirical  and theoretical
algorithms and the previous data inputs to determine the new
operating conditions. Sludge charging rates, mixing require-
ments, etc., are printed out at the terminal and the  operator
then implements these instructions. This near-real-time control
represents a valuable first step toward automation.
   Using the dynamic model previously described with simula-
tion by the hybrid  computer, Graef and  Andrews  (10, 11)
have explored a  variety of control signals, controller modes,
and control actions for the anaerobic digester. The simulations
                                                           34

-------
    s.?
    (P C

    O o

    &§•
     i
     i.
     /«

     8
        3-
   <= >

   £E
   -n 2
   >• O


   C >
   xD -H
             60.0
OJ
en
g     ~O
o-  co m
  CD  "^3
  CN3
        3 <
        3 O
        o r"
           m
           m
         50.0
               fe
SO.PA
                                              9 * 15 Days
                                                   01
                                                     0 mM

                                                    10 ^m/L as Acetic

                                                   597 gm/L as C

                                                                   1
                  0        1000       2000      3000      4000       5000     6000

                  ITIAL STEADY STATE BICARBONATE ALKALINITY,HC03~, mg/L as CaC03
                                                                                    n
                                                                                    6
                                                                                    r


                                                                                    33
                                                                               m

                                                                               -I
                                                                               •D
                                                                               3)
                                                                               O
                                                                               O
                                                                               m
                                                                                    m
                                                                                    en

-------
,,2500-


faw-
8

tisoo-
 5100C-
   500-
j 15
s

I 10


?Ts
NO CONTROL
 S02-40gm/L
                                       10 gm/ L as ACETIC
                                 ON-OFF  GAS SCRUBBING
                                 STAGED  RECYCLE
                                   FIRST STAGE  QR • 50 L/DAY
                                   SECOND STAGE QR . 175 L/DAY
                                                                        60  -
                                                                      30 iS
                                                                             S  50
                                                                             UJ
                                                                             u
                                                                             «
                                                                             ^ 30
                                                                LU     CVJ
                                                                s    °
                                                             IOC;
                                                                en
                                                                =>
                                                                                20
           b.
                                            S02 * tOgrn/Las ACETIC
                                 JL
                                       J_
                                                                                 10
                                                                                7.0
                                                                                6.8
                                                                                6.6
                                                                                6.4
-400?
                                                                                                                                                       e
                                                                                                                                                       E
  3»oP
                                                                                                                                                   200|

                                                                                                                                                      Q.

                                                                                                                                                      <


                                                                                                                                                   100 I
                                                                                        S02 • 40  gm/Las ACETIC

                                                                                       —I	I	1	L
                                                                                                           J	L
                                                                                                                                        J	L
                                                                                                                 S02 * 40 gin/Las ACETIC   _
                                                                                                                                  S02 "27gm/L
                                                                                                 UIO CONTROL
                                                                                                            FEEDBACK CONTROL SIGNAL IS pH
                                                                                                            FIRST STAGE ON-OFF BAND
                                                                                                              OFF  pH - 7.00
                                                                                                              ON   pH « 6.75
                                                                                                            SECOND STAGE ON-OFF BAND
                                                                                                              OFF  pH - 7.00
                                                                                                              ON   pH « 6.65
                                                                                                             1      I       I      I	I
                               TIME, Days
                                                                                                          -TIIVIE, Days

-------
                                                                                    BIOLOGICAL TREATMENT PROCESSES
indicated that the most effective control strategy was directly
dependent  on the type of overloading. The recycle of gas from
which carbon dioxide has been scrubbed, a new control action
proposed as  a result of this research, and base addition, both
using pH  as  the feedback signal, were  best  suited for the
correction  of organic overloading. The simulated response of a
digester to this control action is presented  in Figure 8.  A
simple on-off controller mode was used and the dashed lines
indicate the  controlled response to a step forcing in influent
substrate concentration  insufficient to cause process  failure,
while the  solid lines indicate the  response to a step  forcing
which would cause failure.
   Failure  by an  overload  of toxic materials was  best  pre-
vented by  the recycle of concentrated sludge from a second
stage using  the  rate of  methane  production  as  a  feedback
signal. Although the  control  action proposed  is  not new,
having first been proposed by Buswell (13) in  the 1930's, the
proposed control signal, rate  of methaiie  production,  is new
and  should be one of the best indicators of digester  condition
with respect to overloading with toxic materials.  The rate  of
methane production can be easily calculated  from  the com-
mon measurements  of flow rate and composition of the gas
phase and  would be an excellent  indicator of the activity  of
the methane bacteria which are the most sensitive and critical
organisms in the digester. An  analogue can be  drawn between
the  use  of  the  rate of methane production as an activity
indicator in  the anaerobic digestion  process and the  use  of
oxygen utilization rate as a measure  of microbial activity in
the activated sludge process.

RESEARCH NEEDS
   The  dynamic models, process  stability characteristics, and
control strategies summarized herein are  by no  means com-
plete and  still require further development. Some additional
factors  which should be incorporated into the models are as
follows:
Activated Sludge Process
   1. An improved procedure for predicting the concentration
      of suspended  solids in  the  effluent  from the  secondary
      settler. An important fraction of the effluent BOD is due
      to these suspended solids.
   2. Establishment of a quantitative relationship between the
      settling  characteristics   of  activated  sludge  and  the
      process parameters. At present only an empirical relation-
      ship  between interface settling velocity and sludge age is
      available and  there is some  doubt as to the validity of
      this relationship.

Anaerobic Digestion
   1. Incorporation of the effects of changes in  temperature
      into  the dynamic model. The process has a reputation of
      being  unstable  with respect  to  sudden  changes  in
      temperatures; however,  this has not been expressed on a
      quantitative  basis.  Process  stability  with respect  to
     temperature changes would be of special importance for
     the thermophilic version of anaerobic digestion.
   2. Establishment of a quantitative relationship between the
     solids-liquid separation characteristics of the digested
     sludge and  the process parameters. This is of importance
     in connection with vacuum filtration or centrifugation.
     Also  of  importance  in  this  respect  would  be  the
     establishment of a quantitative relationship between the
     quality of  the supernatant, filtrate, or centrate and the
     process parameters.
   In addition  to the above specific research needs, it should
also be mentioned that the dynamic models, process stability
characteristics,  and control strategies summarized herein have
only been  validated by simple  laboratory experiments, litera-
ture  searches, and discussion with  knowledgeable  operations
engineers.  This is, at  best, only semi-quantitative validation
(responses  are  in the  right  direction  and  right order  of
magnitude) and both pilot and  full-scale field experimentation
will be needed for quantitative validation. In this respect, it
should be  emphasized that models  are evolutionary in nature
and  will  change as more  knowledge is gained  about the
process. A model which is quite adequate as a first approxi-
mation may be replaced at a later date by a more exact model
with  better  estimates  of the coefficients, fewer empirical
relationships, and inclusion of more variables. This evolutionary
nature of  models is not always recognized and can lead to
reluctance  on the part of an investigator to either modify or
discard a model  in the  same  fashion that investigators in past
years have sometimes  been  reluctant to  modify  or  discard
verbal hypotheses.
   Still another very significant research need is the combina-
tion of dynamic models for the individual processes into an
overall dynamic  model for a  wastewater treatment plant with
subsequent use of the model to explore  computer-compatible
control strategies for the plant.  A preliminary effort in this
direction  is the  work  of Bryant (2) who coupled dynamic
models  of the  primary settler, aeration basin,  secondary
settler,  and  chlorine  contact  basin. Such  a  model,  after
appropriate validation,  could then be used to  explore  inter-
actions  between  the individual  processes with the ultimate
objective  of establishing an optimal control strategy  for the
plant. The author has been working toward this objective  for
the past six years.

ABBREVIATED GLOSSARY
A   concentration of anions other than HCO5, CO 3, S" and OH
C   concentration of cations other than the hydrogen ion
F   liquid flow rate to reactor
HS  un-ionized substrate concentration
Q   total dry gas flow rate from reactor
S   total substrate concentration
S"   ionized substrate concentration
Ty  concentration of conservative toxic chemical agent
 V   reactor liquid  volume
VQ  volume of gas space in reactor
X   organism concentration
Z   net cation concentration
9   hydraulic residence time
                                                           37

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
REFERENCES
 1. Busby, J. B. and Andrews, J. F., "Dynamic Modeling and Control
    Strategies for the  Activated Sludge Process," Jour. Water Poll.
    Control Fed. (In Press).
 2. Bryant, J. O., Jr., "Continuous Time Simulation of the Conven-
    tional  Activated Sludge  Wastewater Renovation System," Ph.D.
    Dissertation, Clemson University, Clemson, SC (1972).
 3. Pflanz, P., "Performance of (Activated Sludge) Secondary Sedi-
    mentation Basins," Advances in  Water Pollution Research (Jen-
    kins, S. H., ed.), Pergamon Press, New York (1969).
 4. Brouzes, P., "Automated Activated Sludge Plants with Respiratory
    Metabolism Control," Advances  in Water  Pollution  Research
    (Jenkins, S. H., ed.), Pergamon Press, New York (1969).
 5. Andrews, J. F. and  Lee, C. R., "Dynamics and Control of a
    Multi-Stage Biological Process," Proc. Ivth Intern. Fermentation
    Symp.  (Terui, G., ed.), Society  of Fermentation Technology,
    Japan,  Yamada-Kami, Suita-shi, Osaka, Japan  (1972).
 6. Torpey, W. N-, "Practical Results of Step Aeration," Sew. Works
    Jour., 20, 781 (1948).
 7. Andrews,  J. F., "A  Mathematical  Model  for  the Continuous
    Culture of Microorganisms Utilizing Inhibitory Substrates," Bio-
    technol. Bioeng., 10, 707 (1968).
                    8.  Andrews,  J. F., "Dynamic Model of the Anaerobic  Digestion
                       Process," Jour. San. Eng. Div., Proc. Amer. Soc. Civil Engr., 95,
                       SA1, 95 (1969).
                    9.  Andrews, J. F. and Graef, S. P., "Dynamic Modeling and Simula-
                       tion  of the Anaerobic Digestion Process," Anaerobic Biological
                       Treatment  Processes, Advan. Chem. Ser., No.  105,  American
                       Chemical Society, Washington, DC (1971).
                   10.  Graef, S. P. and Andrews, J. F.,  "Mathematical Modeling  and
                       Control of Anaerobic Digestion,"  Water-1973 (Bennett,  G. F.,
                       ed.),AIChESymp. Ser. 70, No. 136, 101 (1974).
                   11.  Graef,  S. P.  and  Andrews,  J. G., "Stability  and Control of
                       Anaerobic  Digestion," Jour. Water Poll. Control Fed.,   46,  666
                       (1974).
                   12.  Guarino, C. F., Oilman, H. D., Nelson, M. D. and Koch,  C. M.,
                       "Computer Control of Wastewater Treatment," Jour. Water Poll.
                       Control Fed., 44, 1718(1972).
                   13.  Buswell, A. M., "Septic tank  to controlled digestion," Biological
                       Treatment  of Sewage and Industrial Wastes, Vol. II. (McCabe, J.
                       and  Eckenfelder,  W. W., eds.), Reinhold Publ. Co., New York
                       (1958).
             AUTOMATION OF  THE ACTIVATED  SLUDGE  PROCESS

                                                  Michael J. Flanagan
               Brown and Caldwell, Consulting Engineers, 66 Mint Street, San Francisco, CA 94103
INTRODUCTION
   The  conventional  activated sludge process consists of two
main units,  an aerobic  biological reactor (oxidation tank)
followed by a solids-liquid separator (solids concentrator/sedi-
mentation tank). As shown in Figure 1, process inputs include
(1) wastewater, usually effluent from a primary sedimentation
tank; (2) return activated sludge (RAS), which is a  concen-
trated suspension of organisms and substrate recycled from the
                   sedimentation tank; and (3) an air or pure oxygen supply to
                   provide  oxygen and mixing  in  the  oxidation  tank.  Process
                   outputs include (1) process effluent; (2) waste activated sludge
                   (WAS); and (3) carbon dioxide.
                        The optimum performance  goal of the activated sludge
                   process can  be stated as  the production  of a specific and
                   consistent  quality  of process  effluent  with  a  minimum
                            Air or
                              02
                              1	1
CO2
Primary
Effluent
i
OXIDATION
TANK
t
	 I


RAS
STORAGE


SEDIMENTATION
TANK
^
i
* — Activat
Under ft
(RAS +
*

Process

Effluent
ed Sludge
ow
WAS)
IS 	 ^ To
Disposal
                         I	I
Figure 1. Activated Sludge Process
                                                           38

-------
                                                                                   BIOLOGICAL TREATMENT PROCESSES
 consumption of manpower, materials and energy. In practice,
 this goal has not yet been attained.  However, a number of
 activated sludge treatment plants are currently being designed
 or  constructed which  incorporate modern automatic control
 systems  for  monitoring and controlling the activated sludge
 process.   Simplified  versions  of  these  control  systems are
 described in  this paper, together with suggested improvements
 for attaining  increased  process performance.
    The most critical step in the activated  sludge process is
 usually  the  removal  of  biological  solids in  the  secondary
 sedimentation tank.  Unfortunately, the importance of gravity
 sedimentation  in  the  activated sludge process is often  over-
 looked.  Provided that the oxidation  tanks have sufficient
 capacity, virtually complete utilization and conversion of the
 degradable organic matter to biological solids will take place.
 The process  effluent contains both dissolved and  suspended
 oxidizable organic matter. In many  plants, the major portion
 of the oxidizable organic  matter in the process effluent is due
 to  suspended  solids which have  not been  removed by the
 sedimentation process.
   Strong  interactions  exist  between  the  oxidation  and
 sedimentation  tanks.  For example,  the  flocculation  and
 settling  characteristics  of the oxidation  tank mixed liquor
 influence the quality  of the process  effluent and the solids
 concentration of the RAS. The solids concentration, in  turn,
 influences the  solids concentration that can  be maintained in
 the oxidation tank,  the sludge age and the solids flux to the
 sedimentation  tank. An  automatic control  system  for the
 activated sludge process should therefore monitor and control
 both process units and their mutual interactions.
PROCESS CONTROL
   The following basic control actions can be exerted in the
activated sludge process:

   1. Control of the point(s) of addition of primary effluent
     to the oxidation tank (i.e. step feed process).
   2. Control of an inventory of activated sludge in a storage
     tank  to  offset  the  effects  of  diurnal variations  in
     wastewater flow and strength.
   3. Control of the air or oxygen flow rate to satisfy the
     mixed-liquor oxygen demand.
   4. Control of the RAS flow rate.
   5. Control of the sludge wasting rate.
   The  extent  to which control actions  1  and  2  can  be
implemented is  determined by  the  configuration  of process
units and equipment in a given plant. However, control actions
3, 4 and 5  can be  readily incorporated  in  existing plants
without  making  any  major  modifications  to the  process
equipment.  Several  different control  systems that  include
some or all  of  control actions 2, 3, 4 and 5 are described in
this paper. All  of the  process sensors are presently  available
and either digital or analog logic can be employed to solve the
control equations, although, in most instances, digital logic is
the preferred technique.
   The  control  systems shown in Figures 4 through  9 all
employ a system of RAS pump control which is designed to
maintain the sludge blanket  in the associated sedimentation
tank at a preset level or within preset limits. The blanket levels
are set low to minimize the occurrence and duration of anoxic
(zero DO) conditions.
   Under conditions of limiting solids flux, the operator must
override the  automatic controls and reduce.the solids flux to
the sedimentation tanks by  taking the appropriate remedial
action.
   A description of each of the control systems is given below.
Instrumentation symbols and identification are in accordance
with Standard  S5.1  (1973)  of the  Instrument Society of
America.
Control  of Oxygen  and Air Dissolution  in Mixed  Liquor
(Figures 2 and 3)
   Control of air and oxygen dissolution in the mixed liquor is
an important parameter in the activated sludge process. The
normal strategy is  to add sufficient air or oxygen to meet the
time-varying  oxygen demand of the mixed liquor. Because
electrical  energy is  one of the major operating costs of the
activated  sludge process, there is an economic incentive to
minimize unnecessary aeration.
   Most activated  sludge plants use  air as an oxygen source.
The  preferred method for air flow control is to maintain the
desired dissolved oxygen (DO) level in each oxidation  tank.
The  design of a DO  control system depends on the oxidation
tank  configuration  and the  aeration  method.  Numerous
examples are provided in the literature. A DO control system
that  has been successfully applied by Brown and Caldwell to a
number of diffused-aeration type activated  sludge plants  is
given in Figure 2.
   In the high-purity oxygen process, the oxidation tanks are
covered and thus function as on-line respirometers. Four stages
are normally provided within each oxidation tank. A cryogenic
oxygen  plant  supplies  oxygen  to  the oxidation tanks in
proportion to the oxygen uptake rate as reflected by changes
in oxidation tank  gas  pressure. The oxygen is introduced to
each of the oxidation  tanks in the first stage only,  via the
first-stage  recycle  compressors, and  is exhausted as vent gas
from the  fourth stage. The system is designed  to utilize 90
percent of the oxygen supplied. A submerged turbine  aerator
is  installed in each oxidation tank stage to maintain DO and
keep solids in suspension in the mixed liquor.
   The  oxygen supplied to the  intake of the  first-stage
compressors is  controlled to maintain a constant gas pressure
in the oxidation tanks by modulation of a control valve on the
oxygen supply  header. The  oxygen  purity control system  is
designed to maintain constant oxygen purity in the gas venting
from each oxidation tank by modulation of the vent gas valve.
                                                          39

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                                                    V
                                         TM	@	&-
                  Centrifugal
                  Blower (typ)
                                          From
                                          Other
                                         Blowers
                                                                                                        Oxidation Tank
                                                                                                           Pass (typ)
                                                                                    To OfAer
                                                                                  >• Oxidation Tank
                                                                                    Air Headers
                                                                         -Header
 Figure 2. Diffused Aeration DO Control System
                       Effluenl
                                           Stage I
 Stage 2       Stay* 3
OXIDATION TAHK (Tfp)
 Figure 3. Oxygen Dissolution Control System
                                                               40

-------
Constant Mixed  Liquor Solids Concentration Control (Figure
4)
   A common method  of solids control in the activated sludge
process  is  to maintain a constant  mixed-liquor  suspended
solids (MLSS) level and thus a constant mass of sludge in the
oxidation  tank.  This  method  provides good results if the
strength  of  the incoming  wastewater remains  reasonably
constant. Under  these conditions, the amount of sludge that is
wasted  equals the  net sludge  growth  in  the  system.  This
method is not recommended for activated sludge systems that
treat  wastewaters with widely  fluctuating  characteristics be-
cause the  food-to-microorganism (F/M)  ratio  is not  held
constant.
   Two  infrared sludge blanket sensors are provided in  each
sedimentation tank  to provide on-off  control of the  WAS
pump and thus  maintain the  sludge blanket between preset
high-low levels.

Constant Food-to-Microorganism Ratio Control (Figures 5 and
6)
   Two constant F/M  ratio control systems are described. In
the  control  system shown  in Figure  5,  the  F/M  ratio is
calculated  by measuring the respiration rates of  the return
activated  sludge and the  mixed  liquor  (Genthe, Arthur and
Srinivasaraghavan (1).  In the control system shown in Figure
6, the F/M ratio is calculated by measuring the total TOC in
the primary effluent, the soluble  TOC in the process effluent
and the average  volatile suspended solids in the  mixed liquor.
                                                                                  BIOLOGICAL TREATMENT PROCESSES
   Both  F/M ratio  control  systems incorporate a  sludge
blanket level control system whereby the RAS  pumps  are
controlled  to  maintain  a  constant blanket  level  in  the
associated sedimentation tank.  The  sludge blanket measure-
ment system  comprises a hoist-driven infrared blanket sensor
which  is  track-mounted  on the sedimentation  tank  fixed
bridge.  Under  program  control, the  sensor  measures  the
blanket level  at 6 (say) locations during a radial traverse of the
tank, thus enabling the  sludge blanket profile and volume to
be automatically  computed. The sludge  blanket level/volume
signal so derived serves as the process variable  for the RAS
pump control system.
   A biochemical oxygen demand meter, such as the Badger
model  OD-2000, can be employed to measure respiration rates
of the  RAS  and the  mixed liquor.  These  measurements,
together with measurements of plant  flow and RAS flow, can
be employed  to calculate the F/M ratio.
   The  microorganism concentration (M)  can be obtained
from the oxygen demand  rate  of  the RAS. The  oxygen
demand  rate is  a  measure  of the viable  microorganism
concentration because respiration is  essentially endogenous.
When  the  RAS  is  returned and  mixed with the primary
effluent, the  RAS  biomass  is  provided with  the  organic
substrate present  in the  primary effluent. Measurement of the
oxygen demand of the mixed liquor downstream of the point
of  RAS-primary effluent  mixing  reflects  the additional
oxygen required by the biomass to utilize the new substrate in
support of new cell synthesis. Thus the difference between the
                                      Plant Flaw )	
ritnary _
ffluent

— u-"-J—
'
1
1
1
J^ao
©
OXIDATION TANK (Typ)
                                                                                                          Process
                                                                                                         Effluent
                                                                                      —  7°
                                                                                       Disposal
 Figure 4. Constant MLSS Control System
                                                         41

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
         Plant Flo*(Flnf)	»


               I	
       Primary

       Effluent
                                                                      To
                                                                     Disposal
 Figure 5. F/M Ratio Control System Using Respirometry
       Primary
       Efflutut
                                                                   Disposal
 Figure 6. F/M Ratio Control System Using TOC and MLVSS Measurements


                                                             42

-------
oxygen demand  of the mixed liquor and the RAS provides a
measure  of F, the metabolizable substrate in the primary
effluent. As  shown  in  Equation 1,  the  F/M ratio  can be
calculated from two  flow measurements and two respiration
rate or oxygen demand measurements.
                                            (1)
   F/M
   ODMLFlNF
                     OD
where:
      F/M
     K
                        RAS    RAS
the food to microorganism ratio (preset
by operator)

constant determined by operating experi-
ence for a given plant

oxygen demand of mixed liquor
                       BIOLOGICAL TREATMENT PROCESSES


 microorganisms, and  average MLVSS as a  measure of the
 microorganism  population. According  to Weddle and Jenkins
 (2), volatile suspended solids  is an excellent index of the viable
 microorganism  content  of activated sludge for the practical
 operating range of activated sludge plants treating domestic
 sewage. Because the hydraulic regime of most oxidation tanks
 is between complete mixing and plug  flow, several measure-
 ments of MLVSS may be required to obtain an average value
 of the MLVSS concentration  in the oxidation tank.
   Because  the total  TOC  measurement includes  oxygen
 demand  of substrate which cannot  be assimilated by  the
 microorganisms, a measurement is made of the soluble TOC in
 the process effluent, and  this value is subtracted from the
 primary effluent TOC value to give a measure of the substrate
 available to the microorganisms. The  F/M ratio  is given by
Equation 2 as follows:
                                                                F/M  =
     ^I)RAS =     oxygen demand of RAS
                                                        •E
                     Total TOC[NF  - Soluble TOC£FF

                              MLVSS
                                                                                                                ,(2)
                   influent flow
RAS flow which is indirectly controlled
by varying the controlled variable
      "INF
       RAS
   A second  method of  F/M ratio control is based  on the
measurement  of primary effluent TOC less the process effluent
soluble  TOC  as  a  measure  of substrate  available  to  the
                                                           where:
                                                            Total
                =  constant determined by operating experi-
                   ence in a given plant

                =  total TOC in the primary effluent
                                                            Solu- TOCEFF  =  soluble TOC in the process effluent
                                            MLVSS
                   mixed liquor volatile suspended solids
                                                                           To
                                                                         Disposal
 Figure 7. Hydraulic SRT Control System
                                                        43

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS


 Solids Retention Time Control (Figures 7 and 8)
    Two solids  retention  time (SRT) control  systems  are
 described. Solids  retention  time is  defined as  the average
 retention time of biological solids in  the system and includes
 solids in  the  secondary sedimentation tank and, if necessary,
 solids in the mixed liquor and RAS transportation facilities. It
 is calculated as the sum of the volatile suspended solids (VSS)
 in the oxidation and secondary sedimentation tanks divided by
 the sum  of pounds of VSS  intentionally wasted and  those
 unintentionally  lost over  the effluent weirs. A distinction is
 drawn between SRT and sludge age; calculation of sludge age
 is based only on the solids in the oxidation tank.
   The simplest SRT control method is known as the hydraulic
 control method (Figure 7) and was first proposed by Garrett
 (3)  in  1958.  More recent applications of this method have
 been described by Walker (4) and Burchett (5). If the solids in
 the  sedimentation tank overflow  are ignored and mixed-liquor
 wasting is employed, then the quantity of mixed liquor that
 must be continuously wasted is given by Equation 3 as follows:
                                                'ST
                                               SRT
                                                             sedimentation tank volume in cubic feet
                          solids retention time in days.
      rWML
   7.48    OT
                                       ST
(3)
 where:
       WML
       'OT
                                  SRT
                    waste mixed liquor flow in gpd
oxidation tank volume in cubic feet
   Despite the  fact that this process suffers from inflexibility
and  requires the wasting of a greater volume (approximately 4
times) of waste compared to RAS wasting, the control system
is very simple and requires only measurement and throttling of
the waste mixed liquor stream.
   A more exact and flexible method of SRT control is shown
in Figure 8 and is based on computing the total solids in the
activated sludge system and the total solids wasted from the
system, both intentionally and unintentionally. For example,
if the SRT is selected as 5 days, then 20 percent of the  total
solids in the system are wasted continuously. As  the influent
BOD varies, the WAS wasting rate varies but  the percent
removal  rate from the system remains constant. Furthermore,
because a fixed percentage of the total solids are wasted daily,
the  percentage of  viable organisms in  the  system is  not
important because if 20 percent  of the total solids are being
wasted, then 20 percent of the viable microorganisms are also
being wasted. The relationship between SRT and  F/M ratio is
shown in Equation 4:
      SRT4  =    y(F/M)-b                         (4)
where:
                                                                       y   =
                                                            cell growth constant
             and b   =    endogenous respiration rate
                  Process Effluent
                      Flow (Feff)	
                  Plan, Flo* (Finf)	-
                                                                 To
                                                                Disposal
Figure 8. SRT Control System
                                                         44

-------
    Thus it can be seen that if the SRT is held constant, the
 F/M ratio  remains  constant. Using  this  method of  SRT
 control, sludge is wasted from the RAS channel. The wasting
 rate, F^^g, is controlled in accordance with Equation 5:
                                                      (5).
hWAS
where:
      M

      SRT

      F
                    8.33  . F
                             EFF
•  ssEFF
       EFF
      SS
        EFF
      SS
        RAS
                   controlled WAS flow, mgd


                   total solids in system, Ibs

                   5-20 days (system set point)

                   effluent flow, mgd


                   process effluent suspended solids, ppm


                   RAS suspended solids, ppm
Combined SRT and F/M Ratio Control System
   The combined SRT and F/M ratio control system shown in
Figure 9 incorporates features provided  in Figures 5 and 8,
together  with RAS storage facilities. Although  this process
configuration suffers from the disadvantage  that the RAS is
pumped twice, it should be capable of responding to diurnal
                       BIOLOGICAL TREATMENT PROCESSES


 variations  in  BOD  loading in  addition to  maintaining the
 desired SRT. Basically, the SRT control  is used to control
 solids wasting and the F/M ratio control is used to control the
 rate  of return of RAS to the  oxidation tank to match the
 time-variant BOD load  in the primary effluent. If the level in
 the RAS storage tank reaches a preset high level, an override
 level  control system increases the sludge  wasting  rate to
 prevent the sludge level  from rising.

 CONCLUSIONS AND RECOMMENDATIONS
   Several  control systems for  the  activated sludge  process
 involving  conventional  measurement and  control techniques
 have  been described. The  performance and  productivity of
 most existing  activated  sludge plants can be increased by the
 installation of one or more of the control systems that have
 been  presented in this  paper. The degree to which  control
 systems can be retrofitted to existing plants is usually limited
 by a  lack of flexibility  in the plant design. For new plants,
 however, an opportunity exists  to incorporate useful  control
 systems.  It is recommended that process and  instrumentation
 diagrams (P and ID's) be prepared during the functional design
 stage  to  help ensure that a  proper balance is  established
 between the process and the control system.
 ,  The principal justification for automation is to ensure that
 treated effluents comply with discharge regulations for various
 pollutants.  Secondary advantages of automation  include in-
 creased performance  and  productivity,  minimized costs and
the like.  Presently,  the principal obstacle associated with
treatment plant automation is the people problem. It is natural
for people to think and act in terms of their own expertise and
to resist acceptance of the unfamiliar.  Because the need for
                           	Plant FH,.(F:nl)	-•
                                                          WAS Pump lT,f)
Figure 9. Combined SRT and F/M Ratio Control System
                                                         45

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
 treatment plant automation has only occurred in recent years,
 it is not surprising that most people associated with the design
 and operation of wastewater treatment facilities are skeptical
 of automation.  Poorly designed and maintained automatic
 control systems will tend to reinforce this skepticism.
   It appears reasonable to assume that, as effluent standards
 become  more stringent and wastewater reuse for non-potable
 purposes increases, automation will play an increasing  role in
 the operation of the nation's municipal wastewater treatment
 plants. Furthermore,  as  the market for wastewater  instru-
 mentation expands, we can expect new and improved on-line
 sensors,  designed specifically for the measurement of waste-
 water parameters, to become  increasingly available  in  the
 future. The measurement  problem is  compounded  by  the
 heterogeneous nature of  wastewater  and the tendency  for
 probes and sample lines to  foul up.  Measurements  such as
 pressure, temperature, flow and level that are common in  the
 process  industries  are  relatively  easy  since an extensive
 collection of sensors  has been developed by the  process
 industries and may be used  in wastewater treatment  plants.
 The central problem is in the area of analytical measurements.
   The use of digital data  transmission and control systems is
 becoming increasingly cost-effective for wastewater treatment
 plant control and data management. Currently, it is estimated
 that  digital systems are more cost-effective than hardwired
 systems  for new  plants or plant expansions larger than 10 mgd
 (6).  Another advantage  of  digital  control systems  is that
 advanced control strategies can be readily implemented. Most
 unit  processes   encountered in  wastewater  treatment  are
 characterized by  long dead-times and thus feedforward control
 can  be used to great advantage in minimizing process upsets
 and reducing process excursions.
   Dynamic mathematical  models can be  employed in digital
 computer control systems for  real-time optimizing  control.
 Brown and Caldwell (7) is currently designing a  distributed
 digital control system for  a 125 mgd pure oxygen plant;  the
 design calls for running process models in a supervisory level
 computer that in turn adjusts the control strategies in the unit
 process level computers for performance optimization.  As  has
 been pointed out by  Andrews (8), dynamic mathematical
 models are usually necessary to describe time-variant  waste-
 water treatment  processes  and  usually  consist  of  sets   of
 non-linear differential equations. However, construction of a
 dynamic mathematical model requires a number of approxima-
 tions about process variable  interrelationships. If the approxi-
 mations are incorrect or if unanticipated changes  occur in the
 process, the model is no longer useful.
   One method which can overcome this problem is to employ
 a self-learning adaptive control system. With such a system, it
 is necessary to specify the important process variables but it is
 not  necessary  to specify  the  interrelationships between the
 variables. Providing a good process data base is maintained and
 updated, the model develops and continuously updates process
 equations  which  produce  the  least  error  in  predictions.
 Examples  of  such predictions would  be process effluent
 chlorine demand and suspended solids concentration.
   The firm  of Adaptronics  Inc. of Virginia, has developed an
 adaptive nonlinear modeling system  for the predictive control
 of complex  multivariable  processes such as the basic oxygen
 furnace and hot strip steel  mill runout table cooling sprays.
 Background  information on nonlinear and adaptive control
 techniques is given in references 9  and 10.
REFERENCES
 1. Genthe, W. K., Arthur, R. M., and Srinivasaraghavan, R., "On-line
    Oxygen Demand Rates of Primary Influent and Effluent and Their
    Relation to 5-Day BOD", Presented at 45th Annual Conf., Water
    Poll. Control Fed. (1972).
 2. Weddte, C. L., and Jenkins, D., "The Viability and Activity of
    Activated Sludge", Water Res., 5, 621 (1971).
 3. Garrett,  M. T.,  Jr.,  "Hydraulic Control of Activated  Sludge
    Growth Rate". Sew. & Ind. Wastes, 30, 253 (1958).
 4. Walker, L. F., "Hydraulically Controlling Solids Retention Time in
    the Activated Sludge Process", your. Water Poll. Control Fed., 43,
    30 (1971).
 5. Burchett, M. E., "Physical Facilities for Controlling the Activated
    Sludge  Process by  Mean Cell  Residence Time",  Presented at
    Northern  Regional Conf., California Water  PolL Control Assn.
    (1973).
 6. Systems Control Inc., "Phase IV Report, Palo Alto Wastewater
    Treatment Plant Automation Project" (1974).
 7. Brown and Caldwell, "Process Control Report; Sacramento Re-
    gional Treatment Plant" (1974).
 8. Andrews, J. F.,  "Review Paper; Dynamic Models and Control
    Strategies for Wastewater Treatment Processes", Water Res., 5,
    261 (1974).
 9. Control Engineering, "Nonlinear and Adaptive Control Techniques",
    Proceedings  of the First Annual Advanced  Control Conference
    Sponsored by Control Engineering and the Purdue Laboratory for
    Applied Industrial Control (1974).
10. Anon, "Controls that Learn to Make their Own Decisions", Bus.
    Week, April 6, 1974.
                                                 DISCUSSION
J. O. Bryant:
   The need for good data has been stressed. Would the authors
comment on  the  status of instrumentation to obtain these
data?
Michael du Cros:
   What specifically  is  required  in  the  way of new and
improved on-line sensors? Is it new parameters, new analytic
techniques, better reliability, or a combination of these?
L. A. Schafer:
   Can you tell me of any wastewater processes now function-
ing that successfully utilize composition analysis of the liquid
stream (other than residual chlorine)?

Stephen P. Graef:
   Would Mr. Flanagan  please indicate the application status
of each of the control schemes illustrated in his paper? Where
have they been tried, or what must be done  to implement
                                                           46

-------
                                                                                    BIOLOGICAL TREATMENT PROCESSES
them on a full scale field unit?

John K. Nelson:
   I have three questions for Mr. Flanagan—
1. In measuring oxygen uptake rates for RAS, is consideration
   given to  solids concentration, so that variations  in sludge
   stability can be taken into account?
2. When measuring clarifier sludge blankets, how is allowance
   made for hydraulic expansion of the blanket, so that a false
   clarifier solids inventory is not obtained?
3. An  objective of process control  is to monitor process
   variables, and to use these variables to control the process.
   However, how does one "monitor" a factor K (Equations 1
   & 2) which is based on operating experience?

Richard I. Dick:
   The  critical nature of solids separation in control of the
activated sludge process is noted in the paper. However, the
control systems presented for maintaining MLSS, F/M,  and
SRT would not seem to  consider this importance of solids
separation. Rather, an unlimited capacity for solids separation
was tacitly  assumed. With real solids separators of limited
capacity, this  control strategy might result in overload of the
final tank. The control systems do  have high level blanket
regulations,  but these controls might  contradict the require-
ments for maintaining the required MLSS, F/M, or SRT.

Poul E. Sorensen:
   Six  years  ago  Westberg  from  Sweden  introduced  the
concept of  the  totally  controlled process, giving a constant
effluent BOD. This process required storage of sludge. One of
the possibilities for sludge storage is the step-feed process, but
has  this  process  in practice  been  operated  as  a totally
controlled process?

N. J. Biscan:
   I agree that sludge storage for additional recycle in times of
high loading is an important concept. It is one that needs more
field application and testing.
   In a recent study at Dow, "Optimizing  a Petrochemical
Waste   Bio-oxidation System  Through Automation"  (EPA
Grant  No.  S800 766),  we developed  an on-line analog F/M
control system based on repetitive total carbon measurements
of the feed  and of a diluted and homogenized aeration basin
mixed liquor sample. The  automated  control system gave an
updated value  of F/M every  12 minutes.  The F/M  signal
proportionally controlled the fraction of the recycle sludge to
be wasted.
   The system was tested on a  glycol feed in a 250 gallon pilot
plant.  The response  time for F/M to return to within  20
percent of  its  steady-state value  after  a  50  percent step
increase in feed concentration  was 17 hours for the controlled
system (with no sludge storage capabilities) versus 46 hours in
the  uncontrolled  system.  Although  there  would  be  faster
 response times in municipal systems due to the higher yield of
 bacteria, response times for  F/M to return to its steady-state
 optimum value  would still be on the order of hours if sludge
 storage facilities are not provided.
   I think that  F/M is one of the most important parameters
 to be  controlled in the activated sludge process. Whether the
 control action is taken on the basis of total carbon measure-
 ments, total oxygen  demand measurements, or respirometry
 (i.e., oxygen uptake)  measurements, our research needs should
 include field testing of incorporating sludge storage capabilities
 in the  F/M control system.

 M. Dolan McKnight:
   Separating sedimentation from the reactor by interposing a
 sludge storage tank seems worthwhile. How would the design
 engineer determine the  capacity of such a  tank, could the
 aeration tank size be minimized, and how would  this affect
 clarifier size? Should  sludge wasting be  constant or variable in
 such a system?

 C. S. Zickefoose:
   With respect  to the storage of sludge in the activated sludge
 process, the  plant under construction for the  City of Portland,
 Oregon will have the  capability to store up to 800,000 gallons
 in four  separate tanks  of return sludge and/or  high-strength
 liquors generated in-plant.
   The contents  of  these  tanks  are  aerated  and  can  be
 introduced to the aerators by pumps; by monitoring TOC and
 solids at various points in the process, in-plant loads and return
 flows can be introduced at optimum times during the day.
   The plant has a design capacity  of 100 MGD average  flow
 and is  due to start up  in the fall of 1974.

 Heinrich O. Buhr:
   The flow  diagrams  presented by Mr. Flanagan illustrate
 some  of the strategies currently  proposed for  control  of
 activated sludge plants.  It seems likely, however, that research
 on  detailed  process models  which fully  recognize the time-
 varying nature of treatment  plant operations might lead to a
 re-examination and reformulation of many of these concepts.
   As  an example,  consider SRT and F/M ratio control:  It is
 generally  agreed that it  would  be desirable to  maintain  a
 constant biological growth rate, or  F/M ratio, in the aeration
 system. Further, it is  often assumed that this may be achieved
 through  maintaining a constant SRT, based  on  relationships
 such as equation 4 in Mr. Flanagan's paper:
              SRT-1   =  y(F/M)-b
(4)
However, this equation is based on steady-state considerations,
and in a plant which is subject to diurnal flow variations the
relationship is true only on a "daily average" basis. In practice
F/M will vary throughout the course of the day, typically from
50% to 170%  of the average value, even when the sludge
wastage rate is held constant.
                                                          47

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
   If closer control of F/M is required, a more direct approach
than SRT control must be adopted. For this purpose, "food",
F,  may be  measured  by the techniques suggested  by Mr.
Flanagan, or estimated from  the air  supply  rate as done by
Brouzes (Andrews, Ref. 4); then mass of micro-organisms, M,
may be adjusted in  accordance  with the variation  in  F, to
maintain F/M constant. In the absence of sludge storage, the
most effective control of M may be achieved by manipulating
the return sludge (RAS) rate. Some schemes, however, propose
to  keep total RAS flow rate from the  sedimentation  tanks
constant, and to attempt F/M control by varying the  waste
(WAS) flow rate instead (cp.  Brouzes, and Figures 5 and  6,
Flanagan). A typical system  material balance will show that
the   expected  WAS flow  rate  for  long-term  steady-state
operation is  only about 3% of the RAS rate, so that a control
strategy which utilizes WAS as the manipulated variable, can at
best  exert  a marginal influence on the mass of solids in the
aerator. This is not to say that variation of sludge wasting rate
will not cause changes in biological growth conditions, but the
effect is exerted through slow  changes of sludge inventory on a
time  scale  measured  in days.  For F/M  control  on an
hour-by-hour time  scale, a more  forceful control action  is
required, and this usually means direct manipulation  of RAS
flow rate.
   The main purpose of sludge wasting  control should  be to
maintain sludge inventory at  a desired value on a day-to-day
basis  and  this may  be achieved by  simple  level control  as
illustrated  in Figure 4 (Flanagan). Even in this  case, however,
dynamic studies will show  that  the time when the sludge
blanket is at its highest (and the waste sludge pump will tend
to  switch  on)  may  not necessarily be the best  time to
withdraw sludge. For example, with diurnal inflow  fluctua-
tions and,  say, ratio RAS control,  high blanket levels will
occur when inflow is highest; however, since  the controDed
RAS  flow rate  will  also be relatively high at this time, the
recycle sludge concentration will be at its lowest for the day.
When using straightforward level control, waste pumps will
thus switch on during periods of low sludge  concentration,
which will  tend to  maximize, rather   than  minimize the
volumetric quantity of waste sludge to be withdrawn daily. A
preferable  strategy  would  be  one  which programs sludge
withdrawal  to  take advantage  of periods  of  high sludge
concentration.
   These points illustrate some of the information which can
be gained from  a study of dynamic process models. Research
in  this  direction is clearly a prerequisite for  the design of
adequate control strategies.

GustafObson:
   In  order  to  control a process  a  reasonable quantitative
performance index is needed. Of course we know that the goal
is the cleanest possible water at the smallest possible cost. The
problem is what parameters are really affecting the  effluent
quality. This also  relates directly to the choice of proper
models. More research is needed to establish what setpoints on
DO, MLSS,  F/M, settling characteristics, etc., should be used.
Thus a good model is needed for better understanding of the
quantitative  performance index, and both such an index and a
model is needed to obtain good control.

J. O. Bryant, Jr.:
   Few  treatment plants operate as they were designed—SRT,
MLSS, etc.,  usually cannot be  adequately,  and accurately,
predicted  during the  design phase. Hence, operation of most
facilities becomes a function of the operator's ability to "learn"
what the  best parameter values for his plant are. Adaptive,
non-linear control has in  reality been applied routinely by the
good operator.

Wayne C. Smith:
   As more and more batch-type discharges from industries are
going to an activated sludge system, how do we handle these in
the  models and  how do  we  predict the  effects of these
discharges on the treatment system, both as regards effluent
quality  and upsetting the system? Also, how do we include
non-domestic materials that degrade very slowly in this model?

James A. Mueller:
   With respect  to the pure oxygen system, three problems
exist with use of data to verify our  mathematical models,
namely:
   1) no gas flow measurements between stages
   2) no data on  the amount  of gas  lost from the various
      stages due to leakage, and
   3) when  dealing with  the pure oxygen system, our models
      must  not only include the biological growth functions
     but  also the system chemistry due to CO2  dilution of
     the gas phase.
This makes  our  models more complex and more difficult to
verify. However, system understanding is increased by making
the attempt.
   In general if mathematical models are to be effectively used
in design  and control of  biological treatment  systems, the
model  should include measureable parameters and  have  re-
ceived  experimental  or  field verification. In most cases a
simplified model which  approximately describes the system
will  be more effectively used than a complex model that more
closely  describes the  actual system. The next viable phase of
research in  the modelling area  appears to be verification and
practical utilization instead of complex model development.

P. M. Berthouex:
   The problem of pilot plant and field scale testing has been
mentioned.  I would  like to propose that we need more
partnerships between  model-builders and others. Dr. Andrews
has been in such partnerships and I'm sure would agree that
more of this involving others will be profitable. It seems that
                                                          48

-------
                                                                                  BIOLOGICAL TREATMENT PROCESSES
more people  using the iterative cycle theory—data—analysis-
theory are needed.
   Some data we at Wisconsin  have collected  shows  that
certain  mechanistic  models predict too  much variation  in
effluent quality. We have seen systems that show no significant
dynamics in the output in spite of fairly strong input dynamics
(10 to 1 diurnal variation in organic load). Of course, not all
plants show  such  stability. Can  we predict  which  plants
(processes) will exhibit strong dynamics?
   I  have heard you say before that a good model will predict
system  failure.  Do you think we need models to predict
system  recovery after failure?  My thought is that a control
strategy  designed to prevent failure may not be sufficient  to
bring a "sour" system back into service. Do you think we can
predict, or need to predict, the dynamics of recovery?
   Second, we need to ask, "How much control is needed?"
and  "How will load balancing help us?" The answer,  I expect,
will  be different for small plants  than for large plants. For
small plants it might be nice to know how to design so that no
elaborate control strategies are needed.
   Another area where attention should profit us is stochastic
analysis of actual operating data and to work toward stochas-
tic dynamic models. To some this approach seems to conflict
with mechanistic  modeling. Really they should complement
each other; each  has  strengths and  disadvantages.  We  need
more experience with  both, but of the  two  the stochastic
methodology is less well understood, probably because it is
less  intuitively  satisfying.  In  particular,   these models are
developed from "normal" operating dynamics.  I doubt they
will predict  failure and I know  they are not useful once a
system has gone far out of control.

Gustaf Olsson:
   There is  a clear need for a  whole  spectrum of models
describing, for  example,  an  activated sludge  plant. Which
description to use would depend on the purpose for which the
model is required.
   System identification  is a  very powerful  technique  to
establish what is the proper  degree of complexity  required.
Using this technique, theoretical models can be compared and
adjusted to  real plant performance. Not only input-output
relations (ie. the black box approach) but also parameters for
physical a priori  models  can  be calculated. Moreover,  more
reliable  models of  the inherent system  noise  can  be  estab-
lished.

 Lars Pallesen:
    The  topic of modeling has been  brought up several times
 this morning. In this  connectfon I would like to  make  the
 following point:  It  is   possible  to  drive  a  car without
 understanding why it  works,  that is, without knowing how
 each segment of an automobile functions in detail!  It may be
 interesting to understand how, say, the engine is working, but
 if the objective is to drive the car, such knowledge is  of limited
usefulness.  What you  do need to know is some  "macro"
aspects  of  its  total behavior—don't bother about  how  the
steering gear is  put together, but concentrate on learning how
the car responds to turning the steering wheel. In short: worry
selectively.
   Similarly, the kind of mechanistic modeling well exempli-
fied in the  paper  presented  by Dr.  Andrews,  is  clearly
scientifically very interesting. At the present stage of develop-
ment, however, these "micro" models appear to be unable to
completely explain the  behavior of sewage treatment plants,
and, in the mechanistic way of thinking, further refinements
are therefore needed-making the  models even more compli-
cated.
   Fortunately mechanistic models are not needed, if our goal
is  to  control. Much  less   complicated  empirical  models,
describing  macro  features will be adequate, particularly  if
models reflecting  the  stochastic  nature of wastewater treat-
ment systems are employed.
   This approach to the control problem requires, of course,
that real plant  data be available to do the modeling. The very
fact that  the  empirical modeling procedure forces the re-
searcher to  compare  his models with  real  data, must be
considered an inherent strength of the method.

Russell H.  Babcock:
    With respect to automation of biological wastewater treat-
ment processes, it should be noted  that present practice  in
other fermentation industries such as brewing and the produc-
tion of antibiotics is still very crude.  According to one major
instrument manufacturer, the manufacturing of antibiotics is a
simple  batch  process. Temperature  is controlled  and pH is
measured.  The principal control is by laboratory  analysis  of
successive  samples. Laboratory animals are still used as means
of testing the final product.
    Pilot plant  work is  underway which  includes DO control
and pH control. This work, however, is in its  infancy and has
not yet been  applied to routine production. The process will
remain batch with the instrumentation being used to minimize
 laboratory control.  Conclusion: most fermentation processes
are still controlled by the basic techniques of laboratory
 analysis which have been in use for many years.

 Carmen F. Guarino:
    The automation of anaerobic digestion should not be made
 complicated. The C02  content of the gas, plus the quantity of
 CO2 generated, goes  a  long  way  towards  automating the
 digestion process.

 K. J. Jacobson:
    In  response to Mr.  Babcock's comment, the unwillingness
 of the fermentation  industry  to implement warranted and
 available sophistication in control strategies should not be used
 to  excuse the  same  deficiency for wastewater  treatment.
 Sophisticated  control instrumentation is in fact being applied
                                                           49

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
to new ventures in the fermentation industry (protein manu-
facture),  and gradually  to  the classic ventures,  drug and
alcohol production.  At  Penn  we are  developing a  mini-
computer interface for a  pilot-size fermentor. At present data
acquisition and data handling capabilities are being established
for model development purposes. The eventual goal is auto-
matic control of the important (and previously unaccessible)
fermentation  parameters.  Continuous  measurement  of cell
mass using fluorometry and of glucose substrate  by enzymic
oxidation are being developed. These quantities, along with
the  more standard  measurements,  will  enable  automatic
calculation of growth rate and yield  coefficients for use  in
process control.
   The  dynamic  character of  biological processes has been
treated this morning, but one important addition has not been
discussed.  The Monod growth kinetics, or some  derivative
which  accounts  for inhibition, is empirically applied  and in
general is  a  steady-state model since  it ignores time lags. It
does not fit dynamic data  nearly as well as a  first-order
dynamic  kinetic  model,  as  has  been  shown  by  Young,
Bruleynd Bungay  (1). As has been amply pointed out today,
this model refinement is hardly necessary  for implementing
practical control strategies. Indeed steady-state  models can do
a remarkable job  of predicting actual plant performance  for
activated sludge  wastewater treatment  (2). On the other hand,
accurate predictive models will provide important insights  to
aid understanding and designing systems, and hence assist in
designing the most  useful control strategies.  Therefore, al-
though sophisticated dynamic  models may  never be used for
process  control,  their development  is an immediate and
essential research need in order to establish optimal designs for
processes and control strategies.
REFERENCES:
1. Young, T.  B.,  Bruley, D. F.  and  Bungay, H. R. Ill, "Dynamic
   Mathematical Model  of the Chemostat," Biotechnol. Bioeng., 12,
   747-769 (1970).
2. Landis,  J. G., Casey, J. P., Hartzog, D. G-, "Oxygen Activated
   Sludge:  Design by Waste Analysis and Modeling," presented at 47th
   Annual Conference, WPCF, Denver, CO (1974).

CLOSURE
John F. Andrews:
   Reply  to  J. O. Bryant, Jr.:  The author is not engaged in
research on the development of sensors and therefore cannot
comment in detail on the status of instrumentation for making
measurements in  wastewater  treatment plants. Throughout
this workshop there have been  many comments regarding the
nonavailability of  sensors  for  making these measurements.
However, there  are some  basic questions  which  should  be
asked before we rush into an accelerated program  on  sensor
development. Among these are:
    1. What measurements are really needed in order to exert
     effective control over the plant?
   2. How frequently should these measurements be made in
     order to exert effective control?
   3. How accurate do the measurements need to be in order
     to exert effective control?
   4. How much time delay, between the time the sample is
     taken and the time when the results are known, can we
     tolerate and still exert effective control?
   The answers to these questions can best be obtained by a
dynamic analysis  of  the  plant  and its associated control
system. As  an example,  there is a  tendency to  search for
continuous,  on-line  sensors whereas in many  instances the
dynamic response of a process  is sufficiently  slow so that
discrete measurements at  relatively long time intervals would
be adequate. The allowable period between samples could be
defined using sampled data  control theory  if the dynamic
behavior of the process could be quantitatively defined.
   Reply  to P. E.  Sorensen: The author  is  familiar  with
Westberg's work  which was  a pioneering effort  in dynamic
modeling and control of the activated sludge process. As given
in the paper, the author believes the step feed process offers
great potential for process control. However, the author knows
of no  plant currently  using the  potential of step  feed in an
automatic control loop. This is one of those areas where pilot
or field studies are needed.
   Reply to Wayne C. Smith: The specific oxygen utilization
rate (mass of oxygen utilized per unit mass of sludge per unit
of time)  should be a valuable means  of detecting batch
discharges of concentrated organic wastes or toxic wastes. The
dissertation  by Busby  (1) should be referred to for detail on
the use of this parameter as a signal for control purposes.
   Calculation of the specific oxygen utilization rate would be
relatively easy for the high purity oxygen  process where the
reactors are covered and the flow of oxygen into  the reactors
are metered. Coupling of the oxygen balance with a solids
balance would then permit calculation  of the specific oxygen
utilization  rate which should be a direct measure of sludge
activity.
   Reply to James Mueller: The  author is in strong agreement
with Dr. Mueller concerning the need  for model verification.
The  ease  and speed with which computer  simulations can
frequently  be  made  may  lead  to a neglect of this very
important aspect of model development and, in the extreme,
can result in one becoming so enamoured with the techniques
that the purpose for using them  is almost forgotten. This can
lead to the generation of large quantities of worthless results if
the model  is not  a  reasonable representation  of the real
system.
   Mathematical modeling, computer simulation and physical
experimentation are not exclusive but rather complement one
another and should therefore be used in an iterative manner. It
is obvious that results of physical experimentation can provide
better  numerical values for  parameters  in computer simula-
tions using mathematical  models; however, knowledge gained
in simulation  is also useful for modifying the mathematical
model, guiding physical experimentation, and establishing the
type and frequency of field observations needed. Expressing
                                                          50

-------
                                                                                    BIOLOGICAL TREATMENT PROCESSES
relationships in mathematical terms, computer simulation and
physical  experimentation are  all part of the same problem,
model development, and to  a large extent  can  proceed
simultaneously.
   The author is also in agreement  with  the need for model
simplification  for field use. A model which is too complex is
subject to either "misuse or disuse." Sensitivity analysis can be
used for  this purpose by indicating those variables which have
little effect on the outputs and can therefore be considered as
constants.
   A more detailed discussion of some of the above mentioned
points is given in (2).
   Reply to P. M. Berthouex: The author is in agreement with
the points raised by Dr. Berthouex. For example, the inputs to
wastewater treatment plants are comprised  of both determin-
istic and stochastic components and our most recent work is
considering both.
   The  development  of  models,   to  a  certain  extent,  is
dependent upon the tools with which the investigator is most
familiar.  Just  as  Dr. Berthouex  has suggested that we need
more partnerships between model builders and others, I would
also suggest that we need more partnerships between those
concerned with  deterministic models and those working with
stochastic models.
REFERENCES:
1.   Busby, J. B., "Dynamic Modeling and Control Strategies for the
    Activated Sludge Process," Ph.D. Dissertation, Clemson University,
    Clemson, SC (1973).
2.   Andrews, J. F., "Application of Some Systems Engineering Con-
    cepts and Tools to Water Pollution Control Systems," Proceedings
    of the Symposium on the Use of Mathematical Models in Water
    Pollution Control  (A. James, ed.), University of Newcastle upon
    Tyne, England (1974).

Michael J. Flanagan:
   Reply to Michael du Cros: A combination of  these, for
example, sample  conditioning  equipment for automatic wet
chemistry analyzers requires improvement. Real-time or near-
real-time  measurements  that  are  required include  viruses,
enzyme activity, nitrogen and bioassays.
   Reply  to  L. A.  Schafer:  Yes—a  number of  Brown and
Caldwell-designed plants  have successfully  employed on-line
sensors for the measurement of composition variables such as
pH, ORP, conductivity, DO, etc., for  up to 10 years.
   Reply to S. P. Graef:  All of the control systems shown,
except those  that  include F/M  ratio  control,  have  been
successfully  applied. We  are currently designing an activated
sludge plant that will incorporate the control concepts shown
in Figure  9.  Sludge  storage will  be accomplished  in the
oxidation tanks by  using the step-feed  mode  of operation.
Power-operated slide gates will be employed to modulate the
quantity and point  of addition of primary  effluent  to the
oxidation tanks.
   Reply to J. K. Nelson:
1. As we have not tried F/M ratio control using respirometry,
   I cannot give a definite answer to this question.
2. In a properly designed  clarifier, hydraulic expansion of the
   blanket  should not occur within normal operating  limits.
   Temporary expansion of  the blanket caused by rotation of
   the sludge  collector does not  affect  the  blanket level
   measurements  because  the level programmer only operates
   when the sludge collector is at right angles to the horizontal
   travel axis of the level sensor.
3. In Equation  1, for example, K is an empirically determined
   constant which relates  F/M, the computed variable, to the
   measured  variables  on  the right-hand side  of the equation.
   The computed  value of F/M serves as the process variable
   for controller  UIC  (Figure 5), the output from which is
   proportional to the difference between  the  desired and
   computed F/M ratios. I  would also point out that other
   measured  variables  may  be  required  in  actual practice;
   however,  the  basic measurements required  are  oxygen
   demand and flow.
   Reply to  R. I.  Dick: I agree-the activated sludge process
control systems given in the paper are not valid under limiting
solids  flux conditions.  The  way  we  would  approach the
problem is  to  generate  a  velocity-solids level  relationship
through a series of mixed liquor settling tests conducted daily
at different solids concentrations (v = a x c'n).  A computer
program is used  to produce  the limiting solids  flux and its
associated  limiting underflow and mixed-liquor concentrations
as a  function of sedimentation tank overflow rate and return
sludge  ratio.   In  actual  operation,  if  limiting  solids  flux
conditions  are approached, the operator will exercise remedial
action such  as reducing  the total  sludge inventory  in the
system, placing idle tanks in service, changing operating mode
to reduce solids flux,  etc. When normal  operating conditions
are restored, the  automatic  control system is placed  back in
operation.
   Reply to  P. E. Sorensen: Yes-at  the Renton Treatment
Plant in Seattle, Washington.
   Reply to M. D. McKnight: For a given plant and associated
F/M  ratio,  storage for M  can  be estimated from the  average
time  that F exceeds F average.
                                                          51

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                   Report  of Working  Party
                                                      on
                         RESEARCH  NEEDS  FOR  AUTOMATION
                       OF  BIOLOGICAL  TREATMENT PROCESSES
                                                Paul H. Woodruff
                      President, Roy F. Weston Inc., Weston Way, West Chester, PA 19380
                                                   A. W. West
                    Office of Enforcement and General Counsel, National Field Investigations
                        Center, Environmental Protection Agency, Cincinnati, OH 45268
INTRODUCTION
   Biological processes are expected to continue to be the
principal means of removing biodegradable dissolved organic
matter and  reducing the volatile solids content of putrescible
solids produced by wastewater treatment facilities. Biological
processes are  capable  of  high dissolved organic  removal
efficiencies  and generally have a substantial cost-effectiveness
advantage over alternative methods. Engineered systems utiliz-
ing biological processes for wastewater treatment have been
under development for the past  100 years. Much progress has
been made  during this period of time in adapting the process
to overcome the various treatment problems which have been
encountered. A considerable amount of research has  gone on,
the preponderance of it in the past thirty years. Yet, the full
potential of biological processes to produce the high-quality
effluent demanded  by  today's  standards  has not  been
achieved. One of the major problems in achieving reliable peak
performance is insufficient understanding of process dynamics
and a lack of adequate instrumentation to improve treatment
plant performance through adoption of automation  for  pro-
cess controls.
PROBLEMS
   The following are major problems relative to automation of
biological treatment processes, as brought out at the work-
shop.
  1.  Lack of dynamic models of biological processes (activated
     sludge,  aerated lagoon, trickling filter, stabilization ponds,
     rotating filter media, anaerobic sludge process, aerobic
     sludge  digestion,  and anaerobic sludge digestion) which
     will adequately  identify the  essential process  variables
     and, therefore, the needed process measurements.  One
     area which typifies this problem is the poor understanding
     of the  impact of variables in the biological  reaction
     process on solid-liquid separation.
  2.  Lack of accepted and field-tested process control strate-
     gies.
  3.  Poor understanding of biological treatment processes.
  4.  Lack of knowledge regarding available instrumentation
     and its application, adaptation and performance in this
     area.
 5.  Lack of incentive  for  manufacturers to develop new
    instrumentation.
 6.  Insufficient  appreciation of  the  benefits  of instrumen-
    tation and automation and the consequent inability to
    provide a strong incentive based on a strict cost/benefit
    analysis.
 7.  High-quality treatment results are obtained at the expense
    of relatively high consumption of energy and  natural
    resources.
 8.  Unavailability of necessary sensors in sufficiently  reliable
    and appropriate form to meet process monitoring  and
    control needs.
 9.  Lack of industry-wide standard instrumentation specifica-
    tions.
10.  Generally low operating and maintenance skills, particu-
    larly with respect to understanding and proper utilization
    of additional instrumentation and control equipment.
11.  Inadequate  communication among operators,  designers,
    equipment suppliers, researchers, and regulatory agency
    personnel.
12.  Lack of definition of necessary and discretionary monitor-
    ing instrumentation and automation  which can  be sup-
    ported  by  the  available  human  resources at a small
    treatment plant.

RESEARCH NEEDS
    The  following research needs  for  the  automation of
biological treatment processes have been listed in order of
priority:
 1.  Develop improved means of information exchange:
    There is a dire need for making better use of past and
    present experience. This can be done only by setting up
    substantially improved means of sharing ideas and experi-
    ence between those groups (operators, designers, equip-
    ment suppliers, researchers  and government  regulatory
    personnel) who are involved in this area of technology.
    There is a particular need to develop new mechanisms to
    encourage the rapid exchange of experience with existing
    instrumentation and automation.
    There is a need for development of a standardized plant
                                                       52

-------
                                                                                   BIOLOGICAL TREATMENT PROCESSES
   performance data-logging report. Standard formats should
   be developed for each type of biological process. In order
   to have such standardized reports gain uniform acceptance
   and use,  they would have to be adopted by EPA and their
   use required as part of the  NPDES permit and monitoring
   process.
   Feasibility of a private "Underwriter Laboratories" type
   of evaluating and  testing  organization should be con-
   sidered for independent  evaluation. Such an organization
   could be supported by users of the instrumentation that is
   tested.
2. Develop  a thorough assessment  of available instrumenta-
   tion:
   A  thorough assessment should be  made  of presently
   available  instrumentation with respect to its applicability
   for the purpose  intended, reliability, serviceability, and
   consistency.
   EPA has  recently sponsored a study that was intended to
   provide such an assessment. Many believe, however, that
   the scope of the study was  far too limited to provide
   conclusive information.
3. Develop adequate mathematical process models:
   a.   Functioning  of biological processes must be under-
       stood sufficiently well to relate process variables in a
       mathematical model which will accurately simulate
       the real-world  process. Such a model would contain
       many  sub-models and would  have  to consider such
       wastewater input characteristics as flow, temperature,
       dissolved organic strength, nature of dissolved organ-
       ics,  suspended solids,  dissolved solids, pH, nutrients,
       toxicants,  etc.  Other parameters, such as  the amount
       of active biomass, detention times, degree of mixing,
       gas  diffusion rates, organic removal kinetics, solid-
       liquid separation rates,  suspended solids  carryover
       from final clarifiers, clarifier underflow, solids con-
       centration, sludge recycle rate, etc., would also have
       to be considered. The preceding parameters apply to
       the  activated  sludge  process; additional parameters
       would  have to  be considered  for  other types of
       biological  processes.
   b.  The advantages  of  flow control and the  means of
       accomplishing it also need to be researched.
   c.   For the present state-of-the-art  a particularly vexing
        research need  is to improve the ability to predict and,
        therefore,  hopefully control  the secondary clarifica-
        tion processes. Basic  research  is required to better
        understand  the solid-liquid separation phenomenon.
        Sludge dewatering operations must be considered an
        integral part of the biological  process when part of
        the   liquid  portion  (e.g., filtrate  or  centrate)   is
        returned to the head of the treatment system; thus
        sub-models satisfactorily describing the several sludge
        dewatering methods must be developed.
    d.  Mathematical  models for nitrification-denitrification
       biological processes should also be developed.
   e.   Research is  obviously required to better understand
       and  manage  the  interactions of all elements of the
       total  treatment  system,  i.e., sources of wastewater,
       collection system, pretreatment,  primary treatment,
       secondary treatment, tertiary  treatment, reintroduc-
       tion of treated water to the environment and operat-
       ing, maintenance and management manpower.
4. Develop practical operating control strategies:
   A very important research need is to devise and test, by
   mathematical models and plant control tests,  the strat-
   egies for biological processes. A full-scale demonstration
   facility is requked.
5. Develop   means  to   provide  economic  motivation  for
   process  equipment  and instrumentation improvements
   and use (invention, development, application):
   a.   The lack of fundamental knowledge concerning bene-
       fits  vs.  costs  of  automated  biological  treatment
       processes is a major obstacle  to  the general applica-
       tion  of these automated  systems.  An  important
       research  need is to define the potential improvements
       in biological process performance in a quantitative
       manner and the potential costs or savings in biological
       treatment process construction  and operation. This
       cost-benefit  analysis is  of  utmost importance in
       making the decision to automate.
   b.   Automation  has considerable  potential for minimiz-
       ing  consumption of energy and  chemicals while, at
       the same time,  optimizing treatment  performance
       when  compared  to  manual methods. This potential
       should be addressed as part of the preceding research
       areas; however,  due to the current  national  and
       international concern for minimizing energy and raw
       material  consumption (even when used for environ-
       mental betterment), this area deserves special atten-
       tion.  Although not directly related to automation per
       se, basic research is needed on methods to increase
       the performance yield as a function of energy  and
       materials consumed. This research needs to consider
       the  total system,   le.,  human  labor,  energy,  raw
       materials, and pollutants created in the manufacture
       of energy and materials required for the treatment
       facility, as well as use of  same for operating purposes.
   c.   A thorough market research effort is needed to define
       the size and time  element  for  various instrumen-
       tation/automation components to speed up develop-
       ment where market forces seem to justify it.
6. Develop sensors for  real-time monitoring and control of
   biological processes:
   A detailed instrumentation-needs survey should be made,
   with  input  and  suggestions  solicited from  operating
   personnel on a national basis. At least the following types
   of sensors are needed:
   a.   Sludge  blanket indicator that   would give in-situ
                                                         53

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
         readings and would be readable over at least a six-foot
         span.
    b.   Automatic settleability indicator for process control
         use.
    c.   A respiration-rate sensor integrated with a suspended
         solids sensor to give unit oxygen uptake rates.
    d.   Reliable suspended solids sensors for a range of 0 to
         5,000 and 4,000 to 20,000 milligrams per liter.
    e.   On-line TOC analyzer.
    f.   An on-line replacement  for the BOD test.
    g.   On-line  analyzers for  ammonia,  nitrate and  phos-
         phorus.
    h.   Sensors  for achieving methanol dosage control on the
         denitrification  stage of nitrification-denitrification
         treatment.
 7. Develop standard specifications, acceptance standards and
    application techniques:
    Research  is  needed  to  develop  standards that include
    specifications for  application,  acceptance, and  mainte-
    nance that are related to users' needs.
 8. Develop  improvements in the man-instrumentation inter-
    face:
    Research  is  needed to clarify  the  operator's needs for
    instrument support  and  the best  means of extending his
    abilities through instrumentation. There is a strong need
    for  improved educational  materials for  operations and
    maintenance personnel relative to the application and care
    of instrumentation for process control.
 9. Develop  improved means of effecting solid-liquid separa-
    tion:
    This item is actually  not one that involves instrumentation
    and  automation  so  much as  that  it  deals  with an
    important component of biological processes. Significant
    improvement is needed for effecting improved solid-liquid
    separation for both the. forward-flow process  and for
    solids-processing flow streams. Return streams from solids
    processing can  drastically affect  biological  systems. (It
    should  also  be noted  that  present  solids  dewatering
    systems  are  inordinately costly.) Many  biological pro-
    cesses depend on  the  effective separation of the  biomass
    from the  wastewater. Insufficient  attention has been paid
    to  this portion  of biological processes. More research is
    needed  to better understand  the  relationship between
    waste characteristics, biological reactor performance, and
    sludge settling characteristics,  which in turn would im-
    prove prediction of  effluent suspended solids from final
    clarifiers.
10. Develop  instrumentation  and  automation  particularly
    suited to small treatment plants:
    Small plants present a particularly difficult problem with
    respect to instrumentation. This is due to several things.
    For one, instrumentation and automation equipment in a
    small plant usually accounts for a larger percentage of the
    total  construction costs,  and  secondly, operating  and
    maintenance personnel are  usually substantially less quali-
    fied  than  personnel available at larger plants. However,
    this may also suggest that automation  under the proper
    circumstances could provide a  means of improving treat-
    ment plant performance and help to compensate for the
    shortage of trained operators.
    Research is needed  to better adapt present instrumenta-
    tion and automation technology to small plants. The role
    of the operator and degree of  manual control  may  be
    more significant in small plants than large. Thus, adaption
    of present technology that still requires significant opera-
    tor involvement may represent a long-term optimum.
SUMMARY
    Priorities have been listed based on the following overall
philosophy. Improved communication  is fundamental to op-
timizing  present knowledge and intelligently  directing all
future activity. Likewise, better knowledge of present automa-
tion  capabilities/equipment  would  speed the  return  from
existing  investment.  Future development  of technological
improvements would seem to be best directed if basic system
models were available. They should  tell  us  what we need to
measure  and suggest  operating control  strategies. Once  we
know what we  need and  how it is to be used, we need to
determine the economic incentives to insure that all are aware
of economic feasibility and opportunities. Development of
sensors and needed man-instrumentation research would seem
the next logical priority, followed by specific consideration for
the many small treatment plants.
                                                           54

-------
                  AUTOMATION
                              OF
PHYSICOCHEMICAL PROCESSES
                  Workshop on Research Needs
          Automation of Wastewater Treatment Systems
     55

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                         FIELD EXPERIENCES
           WITH  A  PILOT PHYSICAL-CHEMICAL  TREATMENT  PLANT
                                               Walter W. Schuk
            Environmental Protection Agency, 5000 Overlook Avenue S.W., Washington, DC 20032
INTRODUCTION
   The  automation of process  monitoring and  control has
played a major part in research performed at the EPA-DC Pilot
Plant in Washington, DC. Pilot plant instrumentation includes
conventional analog controls and sensors, on-line  wet chemis-
try analyzers, and a sensor-based digital computer (IBM S/7).
   A recent pilot plant study compared the effects of manual,
analog,  and digital process control  strategies on the  operating
cost and product quality  of a sewage treatment process. The
pilot plant physical-chemical system, with programmed diurnal
flow variation, was selected for this  study.

PROCESS DESCRIPTION
   The  pilot plant  physical-chemical  process consisted  of
screening, two-stage (pH 11.5) lime precipitation with inter-
mediate  recarbonation,  dual-media  filtration,  breakpoint
chlorination, and downflow granular carbon adsorption. The
process was designed for  a  nominal capacity of  50,000 gpd
with an  optional diurnal  variation  of  3:1  maximum  to
minimum flow.

Test Procedure
   The  process  was operated for  three sequential two-week
periods. The process was controlled manually by pilot plant
operators for the first operating period,automatically controlled
by analog controllers for the second operating period, and
controlled by a sensor-based digital computer for  the third
operating period. Grab samples of the influent and effluent of
each stage of the process were collected and composited by
pilot plant  operators. Daily laboratory analysis of organic
carbon, phosphorus, and nitrogen concentrations produced
process  loading and removal efficiency data for the  three
operating periods. Chemical usage was calculated by  manually
measuring the volumes of solutions removed from the chemi-
cal feed tanks each day.
Control Description
   The control strategies applied can be divided into two basic
groups: those that are flow-proportional with manual updating
such as flocculant aid addition or sludge  wasting, and those
that are flow-proportional with on-line sensor-based  feedback
updating  such as pH control or free chlorine control. In the
presence  of diurnal flow variation, manual or  sensor-based
feedback control  of the process could not be accomplished
without automatic flow proportioning.
   With the exception of breakpoint  chlorination, the feed-
back control  signals were developed by analog proportional-
integral controllers or  the digital equivalent of proportional-
integral control produced by the  sensor-based  digital com-
puter.  For control  of breakpoint  chlorination,  feedforward
NH3 proportioning was added to the analog control algorithm,
and  a  steady-state  control  equation was used for digital
operation.

RESULTS
   The  product  quality  produced  by  the  three  control
methods  indicated only a  marginal difference  in  favor of
automatic controls for the operating periods studied.
   Due to an  extreme change in the influent water character-
istics caused by extended heavy rainfall during one portion of
the study, direct comparison of chemical  usage for the three
control modes was impossible.  As an alternate to measuring
the actual volume  of chemicals dispensed, the deviation of the
controlled variable from the set point was analyzed to estimate
the probable  cost effect. With the exception of breakpoint
chlorination, both automatic control strategies used approxi-
mately  10%  less  chemicals than  the manual  strategy to
maintain the system.
   At the time this test was performed, the  digital control
equation for breakpoint chlorination was designed for steady-
state process flow. Linear flow-proportioning was added to the
                                                       56

-------
                                                                                         PHYSICOCHEMICAL PROCESSES
equation to compensate for diurnal flow during the test. Later
testing of the breakpoint process showed a non-linear response
to  flow changes  (probably  caused  by  changes  in mixing
energy). As  a  result, the digital control  mode provided the
poorest operation of the three control modes tested. While
flow  changes also caused upsets during manual  and  analog
control, recovery was much faster than with digital control.
   The  manpower  required  to  operate  and  maintain  the
physical-chemical  process, excluding solids  handling,  was
significantly  greater  for manual control.  For either analog or
digital control, 24-hour operation of the pilot plant process
required approximately  12 hours per day of operator time,
and  8  hours  per  day  of maintenance  time. For  manual
operation,  operator time increased to 36 hours per  day, and
maintenance time decreased to  4 hours  per  day.  The net
difference  of  20 hours  per  day  was a  100%  increase in
manpower to produce an equal product quality.

CONCLUSIONS
   The development  of automation in physical-chemical pro-
cess  control  has both  improved  process operation and in-
creased  process  maintenance  problems.  The  best control
algorithm   cannot  improve  a  process when  supplied with
inaccurate  or unreliable data,  nor can optimum  results  be
expected when final control elements malfunction or change
characteristics.  Present knowledge of  breakpoint chlorination
is inadequate to develop an optimum control equation for this
process; however, for other processes, proven,  viable control
strategies   were  not  implemented because  of inconsistent
on-line  sensor  data.  For  example,  the  lime  feed control
strategy could not be used until  the  original pH  monitoring
system  and the lime  feed system were replaced with  equip-
ment  of different design. Prior to these changes, the excessive
maintenance  time required,  and  process upsets  caused by
equipment  failures, precluded^ automatic  pH control of the
lime clarification process.
   Equipment failure not only impacted process operation but
also  complicated  process  research efforts.  To  date, most
process  evaluation has depended on manual data analyses and
long-term  operation to overcome the  lack of reliable on-line
information.  It seems futile to  expend research effort on the
cost-effectiveness of automation, or the impact of automation
on product quality and reliability, when the data produced by
on-line sensors cannot be accepted as reliable.
   It  is also  frustrating to put  forth  the maintenance effort
necessary  to produce reliable automatic  process control and
then evaluate control effectiveness by manual analysis of grab
or  composite samples,  made  hours or even days after  the
samples were collected.
   To simply ask manufacturers for reliable instrumentation is
equally futile. At present, consistent specifications for sewage
treatment instrumentation are non-existent. Typically,  they
include some blend of the electrical code, the plumbing code,
and the whim of the originating engineer.


RECOMMENDATIONS
   The  process  just described   required accurate,  reliable
measurement of four parameters for optimum process control:
flow, pH,  free  chlorine and  ammonia,  and the  ability  to
accurately  dispense four chemicals: lime,  carbon dioxide,
chlorine and caustic soda. To expand this process to full-scale
operation,  improvements should be  made in  most of  the
equipment  just   named,  and  the   breakpoint chlorination
control strategy must be improved.  •
   Although recent improvements in flow metering may have
improved  the reliability of flow  measurement, as yet there is
no  documented  method of checking large flow meters other
than  by  shipping  them  to  a test  facility at considerable
expense and lost time. A self-cleaning pH assembly is needed
to reduce  both maintenance time and process upsets caused by
fouled  electrodes.  Both the  free  chlorine and  ammonia
analyzers  are time-consuming, high-maintenance items. The
throttling  of chemical feeds by control valves  and metering
pumps  must be  linear  and repeatable over a broad  range
(approximately 10:1) if flow-proportioning control is  to  be
used. Few valves or metering pumps  meet these  standards.
With the increasing incidence of power shortages, reduced line
voltages are occurring more often. This results in equipment
being operated under borderline conditions more  often and for
longer  periods of time. To maintain process control accuracy
and reliability, instrument power supplies must  be improved.
   As some areas  must  meet  effluent quality standards for
carbonaceous, phosphorus, and nitrogen contaminants, reliable
on-line  sensors must be developed  that can measure  these
parameters in raw wastewaters and plant effluents.
   One major problem exists that has not yet been mentioned.
In  order  for  automation to  become  an effective tool  for
wastewater treatment process control,  it is mandatory that a
program be implemented to  recruit and  train  personnel  to
operate and maintain  automated waste  treatment systems.
Without well-trained personnel, reliability in data evaluation
and achievement of rigid process control will be impossible.
                                                         57

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
         AUTOMATION  OF  PHYSICAL  AND  CHEMICAL  PROCESSES
                                               Thomas M. Keinath
                 Environmental Systems Engineering, Ctemson University, Clemson, SC 29631
 INTRODUCTION
   Physical and chemical wastewater treatment processes have
 found application both  in  biological and physicochemical
 wastewater treatment systems. Although several  of the P/C
 processes are  primarily employed only in physicochemical
 process trains  (chemical clarification and adsorption on acti-
 vated carbon), most are  commonly found in both types of
 systems; e.g.,  gravitational sedimentation  and chlorination.
 Accordingly, this treatise has not been limited solely to those
 processes commonly found  in  physicochemical treatment
 trains. All major physical and chemical wastewater treatment
 processes will  be considered, with the exception  of chlorina-
 tion, which has been considered in detail in a companion paper
 by W. W. Schuk.
   Minor physical treatment operations including flow routing
 and equalization, comminution, screening and degritting have
 also been omitted from this discussion. Automation or control
 of several purely chemical operations such as pH adjustment
 and recarbonation  have  purposefully  not been considered
 herein because they  have  been routinely automated through-
 out the water and wastewater treatment industry.
   It is  to be noted,  furthermore, that this treatise is  not
 meant to be an exhaustive review of the pertinent literature.
 Rather, literature citations have  been made only  to illustrate
 the  general scope of  research  that  is currently being con-
 ducted.

 CLARIFICATION AND  THICKENING OF BIOLOGICAL
 SLURRIES

 Present Practice and Current Research
   Feed-forward control of the activated sludge process can be
 achieved only when a descriptive dynamic mathematical model
 for the process is available. For  this  system  it is mandatory
 that the dynamic model for the aeration basin  be coupled with
 a  complementary dynamic model  for the  secondary settler.
 Development  of time-dependent models  for  the  aeration
 chamber has  occurred primarily during the  past six years.
 During this period these models  have evolved to  a relatively
 structured status. Nonetheless, they lack verification at both
 the pilot- and prototype-scale levels.
   Dynamic models for the secondary clarifier/thickener have
 not yet attained this level of development. Further, none of
 the existing models  have  been used as  part of  the control
 strategies that  currently are employed in  full-scale systems.
 Control  algorithms that recently have been used in practice
 generally focus on one of four control  objectives: (1) ratio
proportioning of the sludge recycle flow to the influent flow
rate; (2) MLSS  control; (3) F/M or PLI control; or (4) SRT
control. The control action normally taken is to adjust the rate
of recycle flow in response to some measured variable such as
MLSS, TOC, or sludge blanket level in the secondary clarifier.
Without  reference to a  dynamic model  of the secondary
clarifier, one cannot ensure  that a specific control action will
result in the desired response. This can be accomplished only
when a dynamic model of the secondary clarifier/thickener is
used to:  (1) predict concentration of the biological solids in
the  underflow;  (2) predict solids blanket  height; and  (3)
predict the solids  concentration profile  in the  settler in
response  to dynamic changes in the influent and underflow
rates of flow and the influent solids concentration.
   Any dynamic model of the secondary clarifier also must be
able to predict the concentration of suspended solids in the
overflow in addition to the three factors enumerated above.
The first attempt to develop a dynamic model for continuous
thickening was that of Bryant (1). Through application of the
Kynch analysis of zone settling and considering that solids are
transported to  the bottom of a settler by bulk flow and
gravitational  sedimentation, the  following  expression  was
derived:
             at
                        ac
where,
   z  =
   C  =
   u =
   Gs =
   t =
        vertical distance in settler
        concentration of solids
        bulk (downward) flow velocity
        gravitational solids flux
        time
Solution  of the  partial  differential equation was effected
through spatial lumping into ten ordinary differential equa-
tions. Bryant also incorporated an empirical relationship (2) to
simulate  the response  of the clarification  function of the
secondary settler.
   Alkema  (3) provided  a  slightly different  approach to
dynamic  modeling of the  continuous thickening process. The
model  which he  developed  provided for the movement of
concentration discontinuities between the predominant layers
in the sludge blanket.
   Tracy  (4) and Tracy and Keinath (5) were  the first to
implement the lumped-parameter model for the entire clarifier
both above and below the feed point. Furthermore, they were
                                                        58

-------
                                                                                           PHYSICOCHEMICAL PROCESSES
the first to provide for laboratory verification of the dynamic
model. They found that the model accurately  predicted the
response of the thickener  to  transient inputs for the under-
loaded  case.  Moreover, their model  appeared  to  give  an
adequate  representation of the  dynamic response  of the
clarifier when it was overloaded, or when it was forced from
an overloaded to an underloaded condition or vice versa.

Research Needs
    (1) The existing dynamic mathematical model for the
        secondary settler must be modified to account for the
        movement  of displaced water  upward through the
        settler.
    (2) A structured clarification  model must  be  coupled
        with  the  continuous  thickener model  to  enable
        prediction of effluent suspended solids  levels. This
        must, of course, account for scour of biological solids
        from the sludge blanket into the effluent.
    (3) The existing models have been developed in terms of
        one spatial variable (depth). These must be modified
        for geometrical effects; e.g. radial profiles in circular
        settlers and longitudinal profiles in rectangular set-
        tlers.
    (4) Laboratory, pilot- and full-scale verification must be
        provided for the models.

SETTLING   CHARACTERISTICS   OF   BIOLOGICAL
SLUDGES

Background and Current Research
   Solution of any dynamic model of  the secondary  clari-
fier/thickener can be  achieved  only  when an  appropriate
expression for the relation dGg/dC is available. Bryant (1),
Alkema (3), and Tracy (4) obtained expressions for this term
by experimentally determining a relation for the  initial inter-
facial settling velocity as a function of the concentration of
suspended solids. From this relation an expression for Gs vs. C
was derived as detailed by Dick (6).
   One recognizes, however, that in an activated sludge system
the settling characteristics and,  therefore, the term 3Gs/9C
change continually.  Consequently, if a dynamic model of the
activated  sludge process is to be  fully useful from a control
viewpoint, it is  imperative that the  expression for 9GS/9C be
updated  continuously either by  (1) off-line measurement of
the settling properties of the biological slurry or (2) prediction
of the settling properties  by reference  to various biological
processes  parameters. The  latter approach is of course prefer-
able as one can  then formulate  overall  feed-forward control
strategies that include the  dynamics of sludge settling charac-
teristics. Conversely, the former approach provides a means for
feedback control.
    A digital solids-liquid interface  settling monitor has been
 developed by George (7) for the off-line measurement of the
 settling properties of biological slurries.  The device, however,
has not  been subjected to field trials. Other approaches to
on-line  measurement of  settling characteristics have  yet to
proceed beyond the conceptual stage.
   Considerable research has heretofore been directed  toward
determining  the  biological factors  that affect sludge settling
relationships, both  clarification and thicking.  These  have
focused  on  delineating the effects  of  organic  loading  and
oxygen tension on sludge settleability and on effluent clarity.
   Lesperance (8), Logan and Budd (9), Stewart (10), Bisogni
and  Lawrence  (11) and Chao (12) showed  that for conven-
tional air activated sludge systems  several optimum ranges of
organic loading intensities (F/M or PLI) existed with respect to
sludge settleability. At very high and intermediate PLI's sludge
settleability  was observed to deteriorate due to filamentous
and  zoogleal bulking,  respectively. Similar  studies on high-
purity oxygen systems conducted by Jewel et al.  (13), Okum
(14), Chao (12), and Albertsson et al. (15)  showed that the
sludge settling characteristics were materially better than for
conventional air systems. Chao's studies showed, furthermore,
that the sludge settleability for the high-purity oxygen system
was  relatively constant over the entire PLI spectrum and was
not subject to either zoolgleal or filamentous bulking.
   The  effects of biological factors on clarification efficiency
were  also investigated by Chao (12). His studies  showed  that
the  suspended  solids concentration in  the  overflow  of the
secondary clarifier increased with  increasing organic  loading
intensities.  In  contrast,  his  studies on  high-purity  oxygen
activated sludge systems showed  that the clarification  effi-
ciencies  were  materially  poorer   than for  conventional air
systems.
   It must  clearly  be  recognized   that  virtually all  studies
conducted regarding the effects of  biological factors on sludge
settleability  and  effluent  clarity  were conducted  under
pseudo-steady  state conditions.  Consequently, these  studies
can only be employed to indicate trends. Essentially no studies
have been conducted relative to the dynamics of settleability
and the clarification of biological suspensions vis-a-vis various
process  parameters.  One  notable  exception  is the work of
Chudoba, et al (16-18) who studied the effect  of microbial
population  dynamics on sludge settleability in systems  that
were operated at steady state with respect to input conditions,
but  were operated  under different mixing conditions. While
their results were inconclusive, they were able to demonstrate
certain interesting trends in sludge settleability as a function of
population  dynamics. Moreover, they qualitatively determined
that  the propagation  of  filaments proceeded  much  more
rapidly than their suppression.
Research Needs
     (1)  Various off-line and on-line approaches for determin-
         ing by experimental measurement the expression for
         9GS/9C must be developed and evaluated.
     (2)  A  quantitative dynamic relationship must  be estab-
         lished between the clarification and settling character-
         istics of activated sludge and the process parameters.
                                                           59

-------
AUTOMATION OF WASTE WATER TREATMENT SYSTEMS
PRIMARY CLARIFICATION

Current Research on Clarifier Dynamics
   Bryant  et al.  (19),  in  their  transient simulation  of a
wastewater  treatment plant,  employed  a  "hybrid" clarifier
model. The model employed a correlation of the fraction of
suspended  solids removed  in  a  full-scale clarifier versus
overflow  rate  to yield  the  solids separation  (essentially a
steady-state model) while assuming a number of completely-
mixed tanks in series to provide the time response. Provision
was made for  sludge hold-up in  the clarifier, although the
solids  concentration in the sludge and  the withdrawal  rate
were not taken as functions of the hold-up. More recent
transient simulations of wastewater treatment plants by Naito
(20)  and Johnson  and Yang  (21)  employed steady-state
models for  clarifiers, neglecting, without justification,  their
contribution to the transient behavior. One might  conclude,
therefore, that primary clarifier models that are employed as
part of overall dynamic treatment plant models are primitive
at best. The reason that physical  phenomena associated  with
clarifier dynamics are not considered is probably because  they
are not understood.
   Steady-state operation of clarifiers has been studied exten-
sively. For the sizing of clarifiers a plug-flow model proposed
by Camp (22) is often employed to calculate the vessel surface
area. Because  the plug-flow model is not  realistic, investiga-
tions  in  recent years  have  ranged  from measurements of
mixing [e.g. Wills and Davis (23), Murphy (24)], to research
into the causes of mixing [e.g. Fitch and Lutz (25), Takamatsu
and Naito (26)], to the development of performance models
allowing for mixing.
   One of the earliest models proposed for primary clarifiers is
that of Hazen (27). Much of the recent research that has been
directed toward dynamic model development  has been  con-
ducted by  Silveston and co-workers (28-32). Their approach
has  been to apply stochastic methods and power spectrum
analysis  to  primary clarifier dynamics.  Smith (33) also has
provided a  stochastic model which is based on flow rate and
the clarifier surface area.

Current Research on Chemical Clarification
   Convery et al. (34) have described extensive  studies on
automation of  the lime clarification process at EPA's  Blue
Plains Pilot Plant.  The studies focused on  four potential
control schemes: (1) conductivity-ratio, (2) flow-proportional,
(3) pH feedback trim plus flow-proportional, and (4) alkalinity
feedback trim  plus  flow-proportional.  Although   the latter
control  scheme provided  the  best  results,  it was judged
unsuitable  for  general  application  because   of  difficulties
encountered in maintaining the alkalinity sensing system. The
pH feedback trim plus flow-proportional system was recom-
mended for general  use. In  similar  pilot-plant studies  con-
ducted at the Cleveland-Westerly Wastewater Treatment Facil-
ity (35), simple pH feedback control was employed.
   Each  of the  control  schemes detailed above  provide  for
control only of the chemical dosing function.  No control is
provided specifically for an  effluent quality parameter, e.g.,
suspended  solids or soluble orthophosphate concentration. To
accomplish this, one would first need  to mathematically
characterize the chemical precipitation and chemical coagula-
tion/flocculation  processes and  then  would  need to couple
these with  a suitable dynamic model for the primary clarifier.
   The  same  approach  would  be  required  in the case  of
chemical clarification using either aluminum sulfate or ferric
chloride. For  these  cases,  however, it  would be  much more
difficult  to properly describe the chemistry  and chemical
interactions of the  system. Perhaps the most feasible control
algorithm would be one of the following simplified strategies:
     (1)  flow-proportional with  pH feedback trimming—where
         the chemical feed and pH set points would be routinely
         updated using off-line jar coagulation studies.
    (2)  mass  proportional  to   suspended  solids  with   pH
         feedback trimming.
    (3)  mass  proportional to soluble orthophosphate with  pH
         feedback trimming.

Research Needs
    (1)  Development of a  comprehensive  dynamic mathe-
         matical model of primary clarification.
    (2)  Development of a simplified  but realistic dynamic
         model of  the  chemical precipitation  and chemical
         coagulation/flocculation processes  (pH-dosage   do-
         main).


FILTRATION

Present Practice
   Automation of the  deep-bed filtration  process  has  not
advanced materially  beyond  what has been common practice
in the water treatment  industry for several  decades. Control
schemes  that  are  currently  employed in practice  generally
relate to two functions: (1) backwash initiation;  and   (2)
influent flow splitting.
   Four  different  backwash  initiation control modes were
described by Convery, et al., (34). These modes,  which were
evaluated at the Blue Plains Pilot Facility, included headless,
high-level (influent), programmed time interval, and manual.
The  headless  control mode  consisted of a  headless sensor
which initiated the backwash cycle when  the headless across
the filter bed  exceeded a preset value (e.g., 9 feet of water).
Control using the high-level scheme was accomplished through
the use of a high-level controller which served to  open the
effluent control valve so as  to  maintain a constant level of
water above the  filter.  When the effluent control valve had
been opened entirely, the backwash cycle was initiated. For
these  two modes, time-delay  circuits prevented the premature
triggering of the backwash cycle  due to accidental or momen-
tary events.
                                                          60

-------
                                                                                        PHYSICOCHEMICAL PROCESSES
   The programmed time-interval  controller  simply  initiated
the backwash sequence after the expiration of a preset number
of  operating  hours.   An  operator manually  changed the
operating  time set-point depending either on the flow rate
through the plant or the rate of headloss build-up. Convery, et
al. (34) concluded that the programmed time-interval control
mode with  a  back-up headloss  indicator alarm  (which pre-
vented flooding when system upsets caused increased solids
loading) and a high-level alarm (which indicated equipment
failure)  provided  for  peak  operating efficiency at the lowest
possible operating cost.
   At the  Cleveland Westerly Advanced Wastewater Treatment
Facility (35)  individual filter  backwash  sequencing will  be
accomplished  by  continuously sensing and scanning  filter
headloss and filter effluent turbidity for each of the  filters in
operation. When either of these values exceeds a set-point, the
backwash sequence for the critical filters) will be sequentially
initiated according to the incoming error signals. Convery, et
al. (34)  emphasized,  however,  that  turbidity could  not
properly  be used as a control  criterion  because changes in
clarifier efficiency resulted in marked changes in filter effluent
turbidity.
   Apportionment  of influent  flow between  the  operating
filters was accomplished at the Blue Plains facility (34) by a
mechanical  splitter box which equally distributed the flow to
each of  the  filters.  At the Cleveland Westerly facility the
influent will be apportioned among the operating filters so as
to minimize the overall headloss.
 Current Research
    Much of the current research relative to the automation of
 deep-bed filtration has been directed toward the development
 of descriptive dynamic models for  the process. The majority of
 the developments in this research area have been contributed
 by  Ives and  co-workers  (36^4) although Deb  (45) has also
 been active  in the  development of a dynamic  model of  the
 process. Qualitatively, these models provide for  description of
 headloss  and  breakthrough turbidity as a function of  time.
 based  on measurements of a  variety  of  parameters  that
 characterize the filter media and influent turbidity.
    None of the dynamic models  have yet been fully verified
 for suspensions of particulates found in wastewaters, although
 Payatakes, et al (48) and Mehter, etal. (49) have conducted
 laboratory experiments toward  this end. When  this is accom-
 plished, however, the dynamic models could be employed in a
 feedforward manner  to  determine optimal control strategies
 which  will  maximize filter run  times,  thereby  minimizing
 backwash water requirements and recycle flows. The filtration
 model,  of  course,  should be  coupled  with the primary
 (chemical)  clarification  model because of the strong  inter-
 actions between them. A preliminary effort in this regard  was
 conducted  by Kriegsman (46) who  coupled the models of
 Bryant, etal. (19) for the primary clarifier and Diaper and Ives
(44) for  the  deep-bed filter to simulate the performance of,
and the interactions between, these processes.

Research Needs
    (1) The  existing dynamic mathematical model for the
        deep-bed  filtration  process must be adapted for the
        filtration  of suspensions of wastewater particulates.
    (2) The  deep-bed filtration model must be coupled with
         the primary clarification model.
    (3)  Laboratory-, pilot-, and full-scale verification must be
         provided for the coupled dynamic models.
    (4) Control  strategies  using  the coupled  clarification-
         filtration models must be developed.
    (5)  Backwash supervisory  control  strategies  must  be
         developed to prevent  cascading and minimize effluent
         deterioration effects during periods of high flow and
         to minimize backwash water storage requirements for
         periods of low flow.

 ADSORPTION

 Present Practice and Current Research
    In practice  (34, 35)  the adsorption process has  been
 controlled in  essentially the  same  fashion as has filtration.
 That is, control  has  been  exerted  only with respect  to
 distribution  of flow to the  adsorbers and initiation  of the
 backwash sequence. At the  Cleveland  Westerly plant  (35),
 furthermore, an activated  carbon  column that is exhausted, as
 determined  by TOC measurements on the influent  to and
 effluent from the column, will be automatically removed from
 operation for spent carbon discharge and refill with regener-
 ated and virgin make-up activated carbon. Removal and refill
 operations will be automatically sequenced,  monitored, and
 controlled.
    Research  conducted  on  automation  of  the adsorption
 process has  been  directed  primarily toward dynamic mathema-
 tical modeling. Keinath  and Weber (47) were the  first to
 provide a simplified dynamic model of the process when used
 for  the removal  of organic  contaminants  from wastewaters.
 Under contract to the Environmental Protection Agency, Tien
 and associates (48-51) developed a structured dynamic model
 which  attempted to account for  the adsorption of soluble
 organics, the filtration of particulates, and the biodegradation
 of  both  soluble  and particulate  organics in a packed-bed
 activated carbon  adsorber. Due  to the complex interactions
 between the  adsorption,  filtration, and biodegradation func-
  tions, this  dynamic model only  very qualitatively represents
  columnar adsorption dynamics. For municipal wastewaters, it
  is questionable whether  a fully quantitative dynamic model
  can realistically  be developed or  whether such a model, if
  developed,   would  be  particularly  useful  from  a  control
  viewpoint.  Because of the large damping capacity in columnar
  adsorbers,  simple  feedback  control strategies  may  well be
  entirely satisfactory.
                                                           61

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS


   In  the case of industrial wastewaters where chromato-           municipal wastewater treatment systems should be
 graphic  or  reversible  displacement  effects  are  observed,           formulated. For  example, the treatment  objective
 however, a  dynamic model  is mandatory such that  feed-           might  be to  minimize  carbon loading rates while
 forward control can be effected  to prevent solute  displace-           meeting effluent standards. This could be achieved by
 ment. A preliminary dynamic model developed by Carnahan           by-passing a portion of the  flow  around the carbon
 (52) semi-quantitatively simulates the chromatographic effects           contactors and blending it with the effluent from the
 that have been observed in the field. Carnahan assumed that           contactors  to produce  the  desired quality. These
 the  filtration and biodegradation effects were negligible. For           studies should be conducted on a full-scale level to
 many industrial wastewaters this assumption is realistic.                  establish the  operational cost benefits that  might be
                                                                    experienced through application of this control stra-
 Research Needs                                                     tegy.
     (1)  The dynamic model which incorporates the adsorp-
         tion, filtration, and biodegradation functions should   OVERALL CONTROL STRATEGIES
         be verified at either the pilot- or full-scale level.            Another  significant  research  need  is  the coupling of
     (2)  The dynamic model which accounts for chromatogra-   dynamic models for  individual  processes  into  an  overall
         phic effects should be verified at either the pilot- or   dynamic model for an  entire physicochemical wastewater
         full-scale levels and appropriately refined.              treatment facility. Such a model could subsequently be  used
     (3)  Control strategies using both of the dynamic mathe-   to explore control strategies for the entire  plant. As indicated
         matical models should be formulated.                  above, Kriegsman (46) coupled the primary clarification and
     (4)  Simple feedback control  strategies for adsorbers in   filtration processes as a preliminary effort in this direction.
                                                        62

-------
                                                                                                    PHYSICOCHEMICAL PROCESSES
REFERENCES
 1.  Bryant, J. O., Jr., "Continuous Time Simulation of the Conven-
     tional Activated  Sludge  Wastewater Renovation System," Ph.D.
     Dissertation, Clemson University, Clemson, SC (1972).
 2.  Pflanz, P.,  "Performance of (Activated Sludge) Secondary Sedi-
     mentation  Basins,"  Advances  in  Water  Pollution  Research
     (Jenkins, S. H., ed.), Pergamon Press, New York (1969).
 3.  Alkema, K. L., "The Effect of Settler Dynamics on the Activated
     Sludge Process," M. S. Thesis, University of Colorado, Boulder,
     CO (1971).
 4.  Tracy, K. D., "Mathematical Modeling of Unsteady-State Thick-
     ening of Compressible  Slurries,"  Ph.D. Dissertation, Clemson
     University, Clemson, SC (1973).
 5.  Tracy, K. D. and Keinath, T. M., "Dynamic Model for Clarifica-
     tion  and Thickening  of Activated Sludge," Amer. Inst. of Chem.
     Engrs. Symposium Series, Water 1973 (1973).
 6.  Dick, R. I., "Role of Activated Sludge Final Settling Tanks," Jour.
     San.  Engr. Div., Proc. Amer. Soc. Civil Engrs., 96, 423 (1970).
 7.  George, D. B., "Ditital S/L Interface  Settling Monitor,"  Special
     Development Report, Clemson University (1973).
 8.  Lesperance, T. W., "A Generalized Approach to Activated  Sludge:
     Part  III.  Process Variants, Why They Occur,"  Water Works and
     Wastes Eng., 2, No. 7, 34 (1965).
 9.  Logan, R. P. and Budd, W. E., "Effect of BOD Loadings on Acti-
     vated Sludge Plant Operation," Biological Treatment of  Sewage
     and  Industrial Wastes. Vol I Aerobic Oxidation, (McCabe, J. and
     Eckenfelder, W. W., eds.), Reinhold  Publ. Co., New York  (1956).
 10.  Stewart,  M. J., "Activated  Sludge Process Variations, The Com-
     plete Spectrum: Part  III.  Effluent Quality-Process  Loading
     Relationships," Water & Sew. Works, HI, R-260 (1964).
 11.  Bisogni,  J. J.,  Jr. and Lawrence, A. W., "Relationship Between
     SRT and Settling Characteristics of Activated Sludge," Water Res.,
     5, 753(1971).
 12.  Chao, A. C., "Destabilization  of Activated Sludge Suspensions
     with Ferric Chloride: Clarification  and Thickening," Ph.D. Dis-
     sertation, Clemson University, Clemson, SC (1973).
 13.  Jewell, W. J., Eckenfelder, W. W. and CavaHer, M. E., "The Use of
     Pure Oxygen for the Treatment of Brewery Wastes," Proc. 26th
     Ind.  Wastes Cow/., Purdue University (1971).
 14.  Okun, D. A., "System of Bio-Precipitation of Organic Matter from
     Sewage," Sew.  Works Jour., 21, 63 (1949).
 15.  Albertsson,  J. G., McWhirter, J. R., Robinson, E. K.  and  Vahl-
     dieck, N. P., "Investigation  of  the  Use of High Purity  Oxygen
     Aeration in the  Conventional Activated Sludge Process,"  FWQA,
     Water Poll. Control Res. Ser. 17050 DNW (1970).
 16.  Chudoba, J., Ottova', V. and Madera, V., "Control of Activated
     Sludge Filamentous  Bulking-I.  Effect of  Hydraulic  Regime or
     Degree  of  Mixing in an Aeration  Tank," Water Res.,  7,  1163
     (1973).
 17.  Chudoba, J.,  Grau,  P. and Ottova,  V., "Control of Activated
     Sludge Filamentous Bulking-II. Selection of Microorganisms by
     Means of a Selector," Water Res., 7,  1389(1973).
 18.  Chudoba, J.,  Blaha, J.  and Madera,  V., "Control of Activated
     Sludge  Filamentous  Bulking-III.  Effect  of  Sludge  Loading,"
     Water Res., 8,  231 (1974).
 19.  Bryant, J. O.,  Jr., Wilcox, L. C. and Andrews, J. F., "Continuous
     Time Simulation of Waste Water Treatment Plants," Presented at
     the AIChE National Meeting, Cincinnati (1971).
 20.  Naito, M., "Dynamic Model for Wastewater  Treatment Plant,"
     Presented  at  Fourth Joint AIChE/CSChE Meeting,  Vancouver
     (1973).
 21.  Johnson, A. I. and Yang, J. T., "Computer Simulation and Control
     Studies  for  a Sewage  Plant,"   Presented   at  Fourth  Joint
     AIChE/CSChE Meeting, Vancouver (1973).
 22.  Camp, T. R.,  "Sedimentation and the Design of Settling Tanks,"
     Trans. Amer. Soc. Civil Engrs., Ill,  895 (1946).
 23. Wills, R. F. and Davis, C., "Flow Patterns in a Rectangular Sewage
     Sedimentation Tank," Advances in Water Pollution Research, 2,
      335 (1962).
 24. Murphy, K. E.,  "Tracer Studies in Circular Sedimentation Basins,
     Proc. 18th Ind.  Wastes Con/.. Purdue University (1963).
 25.  Fitch, E. B. and Lutz, W. P.,  "Feedwells for Density Stabiliza-
      tion," Jour. Water Poll.  Control Fed., 32, 147(1960).
 26.  Takamatsu, T. and Naito, M., "Effects of Flow Condition on the
    Efficiency of a Sedimentation Vessel," Water Res., 1, 433 (1967).
27. Hazen,  A., "On Sedimentation,"  Trans.  Amer. Soc. Civil Engrs.,
    53, 63 (1904).
28. Silveston,  P. L., "Computer  Simulation  of  Waste  Treatment
    Plants," Presented  at  the  ACS/CIC Symposium on Pollution
    Problems of Our Environment, Toronto (1970).
29. Silveston, P. L., "Simulation at  the Mean Performance of Munici-
    pal Waste Treatment Plants," Water Res., 6, 1101 (1972).
30. Singh, D. P., Bryson, A. W., Jr. and Silveston, P. L., "A Stochastic
    Model for Primary Settlers," Proc.  5th  Annual Symp. on  Water
    Poll. Res., Waterloo, Ontario, Canada (1970).
31. Sakata, N. and Silveston, P. L.,  "Exponential Approximation for
    Settling Rate," Water Res. (In Press).
32. Sakata, N., Cordoba-Molina, J. F. and Silveston, P. L., "Studies on
    the Application of Power Spectrum Analysis to Primary Clarifier
    Dynamics," Presented at the 9th Canadian Symposium on Water
    Pollution Research (1974).
33. Smith, R., "Preliminary Design and Simulation of Conventional
    Wastewater Renovation Systems  Using the Digital  Computer,"
    Water Poll. Control Res. Ser. Publ. WP-20-9 FWPCA, Cincinnati,
     OH (1968).
34. Convery, J. J.,  Roesler, J. R. and Wise, R. H., "Automation and
     Control  of  Physical-Chemical  Treatment  for  Municipal  Waste-
     water," in Applications of New Concepts of Physical-Chemical
     Wastewater Treatment, (Eckenfelder, W. W. and Cecil,  L. K., eds.),
     Pergamon Press, New York (1972).
35.  Zurn  Environmental Engineers, "Westerly Advanced  Wastewater
     Treatment  Facility:  Process Development and Preliminary En-
     gineering Design," Zurn Environmental  Engineers and Battelle
     Memorial Institute (1972).
36.  Ison,  C.  R. and Ives,  K. J.,  "Removal Mechanisms in Deep-Bed
     Filtration," Chem. Engr. Sci., 24, 717 (1969).
37.  Ives, K. J., "Rational Design of  Filters," Proc. Inst.  Civil Engrs.,
     16, 189 (1960).
38.  Ives,  K. J.,  "Simulation of Filtration  on an Electronic  Digital
     Computer," Jour. Amer. Water Works Assn., 52, 933 (1960).
 39.  Ives,  K. J.,  "Simplified  Rational  Analysis of Filter Behavior,"
     Proc. Inst. Civil Engrs., 25, 345 (1963).
40.  Ives,  K. J.  and  Sholji, I.,  "Research on  Variables Affecting
     Filtration,"  Jour. San. Engr. Div., Proc. Amer. Soc. Civil  Engrs.,
     91, 1 (1965).
 41.  Ives, K. J., "The Use of Models in Filter Design," Effl. & Water
     Trt. Jour., 6, 522 (1966).
 42.  Ives,  K. J.,  "Advances in Deep Bed Filtration," Symposium  on
     Advances in Filtration, Inst. of Chem. Engrs., Manchester (1968).
 43.  Ives,  K. J.,  "Theory  of Filtration,"  Special  Subject  No.   7,
     International Water Supply Congress, Vienna (1969).
 44.  Diaper,  W. J.  and Ives, K. J., "Filtration Through  Size-Graded
     Media," Jour.  San. Engr. Div., Proc. Amer. Soc.  Civil Engrs.,  91,
     89 (1965).
 45.  Deb, A. K., "Theory of Sand Filtration," Jour.  San. Engr. Div.,
     Proc. Amer. Soc. Civil Engrs., 95, 399 (1969).
 46.  Kriegsman,  G. R., "Simulation of the  Dynamics of Clarification
     and Filtration," M. S. Special Problem Report, Clemson  Univer-
     sity, Clemson, SC (1973).
 47. Keinath, T. M. and Weber, W. J., Jr., "A Predictive Model for the
     Design of Fluid-Bed  Adsorbers," Jour.  Water Poll. Control Fed.,
     40, 741  (1968).
 48.  Payatakes,  A. C., Turian, R. M. and  Tien,  C.,  "Integration  of
      Filtration  Equations  and Parameter Optimization  Techniques,"
      Research Report 70-4, Syracuse University, Syracuse, NY (1970).
 49.  Mehter, A. A., Turian, R. M. and  Tien, C., "Filtration in Deep
      Beds  of Granular  Activated Carbon," Research Report  70-3,
      Syracuse University, Syracuse, NY (1970).
 50.  Hsieh, J. S., Turian, R. M. and Tien, C., "Experimental Investiga-
      tion  of the  Adsorption of Organic Contaminants in Waste Water
      on Granular  Activated Carbon," Research Report 69-1, Syracuse
      University, Syracuse, NY (1969).
 51.  Hsieh, J. S.,  Turian,  R. M.  and Tien, C.,  "Batch  Adsorption
      Kinetics with Chemical Reaction," Research Report 69-2, Syra-
      cuse University, Syracuse, NY  (1969).
 52.  Carnahan, R. P., "Mathematical Modeling of Heterogeneous Sorp-
      tion in Continuous Contactors for Wastewater Decontamination,"
      Ph.D. Dissertation, Clemson University, Clemson, SC (1973).
                                                                   63

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                                DISCUSSION
Robert A. Ryder:
   Would Mr. Schuk please comment on the type of ammonia
measurement and sensing system utilized, and its reliability?
   In connection with breakpoint chlorination, was there any
attempt  to control  overdosage  of  chlorine,  to minimize
wastage as well as the effect of excess residual on the sorption
capacity of the granular activated carbon filters?

Robert H. Wise:
   We in the municipal wastewater treatment field  tend to
believe that most of the many novel sensors and instruments
needed  to  implement WWTP automation  will almost auto-
matically be forthcoming from instrumentation suppliers once
a need for such devices has been conclusively shown. I have
discussed  this  "philosophy"  with  representatives of  many
instrumentation  manufacturers, all of whom assured me that
this  prevailing   concept  is  false.  Mr. Russell Babcock has
already commented on this topic.
   Once  the need for a particular sensor  or instrument has
been conclusively shown,  development of a reliable "device"
to  fill  this need  in the  hostile environment of a  sewage
treatment  plant requires  years  of  intensive  research  and
development. Therefore, if we postpone development of such
devices (using the rationale that such  equipment might not be
really needed at some future  date),  we thereby throw away
lead time corresponding to the developmental time required to
reduce sensor concept (or instrument concept) to practice.
Typically, such lead time is 4  to 8 years*. Can we afford this?

'Example Lead Times for Development of Existing Devices
1)   On-line TOC analyzer: about 8 years
2)   On-line polarographic DO probe: about 10 years
3)   On-line glass electrode: about  10-15 years
4)   On-line NH3  probe: after 3 years, still not acceptable for on-line
     use.

L. A. Schafer:
   In the automation of continuous processes, the development
of new instruments usually occurs in three  interacting phases,
performed  by three separate  and distinct  groups. The first
phase (innovation) carries the basic idea through bench-scale
testing.  The  second  phase  (development)  transforms the
concept  into marketable hardware.  The  third phase (applica-
tion) adapts  the  instrument to  its  proper function  by
modification and  the  addition of sub-systems. A  prime
example is the development of the gas chromatograph, where
the  three phases  were worked out  by chemists and other
researchers, instrument manufacturers, and  industrial users,
respectively. No one group  was  capable of producing  a
working instrument;  all three performed their function and
interacted to develop an effective machine.
   In the municipal wastewater field, the first two groups are
generally available,  but there is no one  with  the ability,
motivation, and funding to perform the third function. As has
been pointed out, instrument manufacturers have little incen-
tive  to  put  their  funds into application. In the treatment
facilities themselves, there are no funds for such work, the
operating staff  has no  motivation,  and  the one  possible
innovation group,  the  analysts or technicians,  are trained
toward  invention  and  laboratory  work  and  are  therefore
conditioned  toward the innovative phase, rather  than the
necessary and different application effort.
   Application, like innovation  and development, is almost
never achieved by any one single group, no matter what the
capability or motivation. Technology transfer is effective only
when it involves the user, or the person who has to live with
the  results. Instrument application by salesmen, design en-
gineers  and consultants has, all  too often, led  to dismal and
expensive failures, with the taxpayer taking the loss.
   It would seem, therefore, that to develop instruments such
as nutrient meters (TOC, BOD, respirometers, and the like) or
suspended solids meters, a group or groups must somehow be
established to promote  attempts at application, evaluate the
results,  disseminate and  promote the findings, and encourage
the selection and adoption of proven instrument systems.
   (The word instrument refers to many classes of devices. In
this case I am talking about process-type instruments, suitable
for continuing service for automatic control. I am particularly
excluding laboratory instruments.)

John F. Andrews:
   It would  appear to me that many  of  our problems with
sensors  could be reduced by using simple, on-line recalibra-
tions and performing tteese more frequently. Has anyone had
any experience with techniques suitable for on-line calibration
of large  flow meters?

Robert H. Wise:
   On-line (in-place) calibration  of  flow meters, regardless of
size, can  be  accomplished relatively easily with  fluorimetric
tracer techniques.  Equipment and/or procedures for making
such measurements  can be obtained from Turner Designs, Palo
Alto, Calif. Mr. Ron Doty (one of the  participants in this
workshop) has had  first-hand experience with this calibration
technique and with some of Turner Designs' flow-calibrating
equipment; his comments on both were favorable.

P. M. Berthouex:
   Mr. Schuk spoke of many problems with on-line sensors. I
agree that to demonstrate the feasibility of on-line control you
need (a) working hardware, (b) a usable control algorithm, and
(c)  performance data.  To  learn  about   process dynamics,
however, we need  performance data,  preferably data taken
                                                          64

-------
                                                                                         PHYSICOCHEMICAL PROCESSES
during controlled experiments. We need more of this kind of
data and we need to properly analyze such data. Data may be
collected manually  or by  any other convenient means, and
methods are available to fit such data in an attempt to derive
the dynamic model, to identify the control  problem, and to
design the control algorithm. This approach,  to which George
E. P.  Box  at  the University of Wisconsin  has contributed
substantially, has the  advantage of minimizing  the need for
assumptions about  mechanism, although known mechanisms
can and  should be  incorporated. Basically it lets the process
tell its own story.
   Aside from the data analysis for which alternatives exist, I
should like to encourage more effort toward manual collection
of data since heavily instrumented plants are not  yet common.
We can  learn some  things  about  automation  and control
without having automated  plants, if we have  the energy to do
the data collection necessary.

Harry Fertik:
   By  developing and using scientific models one can learn
how  various  physical and   chemical  phenomena produce
process behavior, but  scientific models are of limited useful-
ness in developing and testing real-time control strategies. For
example,  in  developing  computer control  algorithms  for
cement kilns more  than a decade ago, the initial effort  was
placed  in  writing  a  scientific description  of  the process
dynamics. This work was dropped when it became clear that
the model, although complex and of high order, still would
not exhibit the behavior familiar to plant operators (material
buildup into "rings" in the kiln producing a  typical tempera-
ture behavior, etc.). The control strategy was developed from
input-output analysis  using  process measurements obtained
from dynamic testing of the process.  Low-order dynamic
models  that  incorporate  process stochastic  behavior  and
nonlinear characteristics were the basis for this  development.
The control strategy was  refined and initially tested on these
models, with final testing on the process.
   Optimization  techniques  have been  applied  to various
classes of problems with varying degrees of success. They can
be  applied to design  problems, the determination of steady
state operating conditions of a plant and to the  calculation of
the dynamic response following a process  disturbance or a
setpoint change. Design optimization is an accepted, successful
technique, and steady-state optimizers are working routinely
in various industries to calculate setpoints, but control design
by dynamic optimization in process industries  has not in
general  resulted  in control  system  performance with  any
significant improvement over conventionally  designed control
systems.

Richard I. Dick:
   Much of the discussion at this conference relates to the
conflict  between complex theoretical models  and practical
control procedures. A middle ground in some cases would be
to modify  the  design and/or operation  of processes to make
them conform to  the  theoretical  model. This sounds  like
idealism, but in some cases it may be realistic. An example is
in the design of sedimentation basins. The theoretical analysis
of steady-state thickening in basins such as the final settling
tank in  the activated sludge process is well established, and the
performance of closely  controlled laboratory units conforms
well to  the theoretical predictions. Parallel measurements with
full-scale tanks indicate  that they  often  perform far less
satisfactorily than is theoretically possible. Professor Keinath
has  correctly  suggested a  need  for modifying theoretical
predictions in comparison with the actual performance of real
tanks. However, an alternative and more  satisfying approach
would be  to  change the design  of sedimentation  basins to
make their  performance approach that  which is theoretically
achievable.  It  would  seem that sedimentation equipment
manufacturers  and design engineers might  give some attention
to  the  differences  between the predicted and actual  per-
formance of their installations.

Harold  D. Oilman:
   The  discussions  so far appear to me to have focused too
strongly on the computer application in process mathematical
modeling. We should not neglect the basic computer capabil-
ities of  logging, reporting, data management, alarming, and its
response to the  information needs of operators, engineers,
management, and maintenance.
   Research  should be  undertaken on how to take advantage
of the new dimensions  of information processing and display,
notably with  the cathode ray tube,  graphics and color—just
start with improved and tailored information for operators.

Poul Sorensen:
   In considering the needs for research, I think a point should
be made  about  optimizing the chemical  post-precipitation
process.
   In Scandinavia, alum is normally used in this process.  Some
provisional experiments carried out in  Denmark have shown
that  a  pH adjustment  with sulfuric acid has improved the
process in  lowering the P content in the effluent from about 1
ppm to 0.2 ppm.  The reason  for  this could be the  high
alkalinity in the  wastewater, but the process requires further
investigation.

John F. Andrews:
   Physicochemical processes offer  the  advantage, in compari-
son with biological processes, of removing a greater percentage
of pollutants  from wastewaters. They should also be more
amenable  to  variable-efficiency operation. However,   their
operating  costs are usually higher than  those for biological
processes.  These  characteristics point to a potential research
need,   this being  selection of  the  proper combination  of
biological  and physicochemical processes  for  operating at a
variable efficiency and interfacing with  the  time-dependent
                                                          65

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS


requirements of the receiving stream.
   For example, treatment by a biological process  may be
adequate for high flows in the receiving stream but inadequate
during low  flow periods. A proper time-dependent basis for
plant  operation  might  therefore  be a  biological process
followed by a physicochemical process, with the proportion of
the total flow treated by the physicochemical process being
controlled in accordance with the flow  rate in the receiving
stream.  An  extension  of  this would be  to  regulate the
proportion of flow going to the physicochemical process in
accordance with  the pollution load in the stream above the
plant  discharge.  The  ultimate application  of  this concept
would be to use  feedforward control of the flow proportion-
ing, based on real-time computer simulation using a dynamic
model of the stream below the point of discharge.
                                     Report of Working  Party
                                                      on
                         RESEARCH  NEEDS  FOR  AUTOMATION
                           OF  PHYSICOCHEMICAL   PROCESSES
                                           Wesley W. Eckenfelder, Jr.
                  Professor, Environmental and Resources Engineering, Vanderbilt University,
                                              Nashville, TN 37235
                                                John Stamberg
                     Office of Research and Monitoring, Environmental Protection Agency.
                                            Washington, DC 20460
 INTRODUCTION
    An extensive Advanced Waste Treatment research program
 sponsored by the Federal Water Pollution Control Administra-
 tion  during the 1960's delineated a variety of physical and
 chemical  wastewater treatment processes that are capable of
 producing very high  quality product  waters.  Subsequent
 demonstrations  at  pilot  plants  and small  full-scale  plants
 showed the utility of these processes, defined their respective
 operational parameters and provided economic data.
    Automation of these physicochemical processes has for the
 most part been limited to studies conducted on a pilot scale
 under contract to  the  Environmental  Protection Agency.
 Although these studies have  defined several control strategies
 that apparently are applicable to full-scale systems, they have
 been evaluated only on a limited scale at small wastewater
 treatment plants.   Because  the process of  automating any
 treatment system is an iterative procedure whereby strategies
 or algorithms are proposed and then repetitively evaluated and
 refined,   it is   clear that automation  of  physicochemical
 processes is yet in its infancy.
    It is noteworthy, furthermore, that most physicochemical
 processes are not yet fully understood with respect to process
 variables  and operational parameters. This is due to a lack of
 long-term operating experience at large  wastewater treatment
 plants. Until such  information becomes available, strategies
 directed toward optimal control cannot be formulated and the
 full potential of physicochemical processes for reliably produc-
 ing high-quality product waters cannot be attained. According-
 ly, it is apparent that research needs in the area of automation,
 instrumentation, and process control often cannot be divorced
 from process research needs.
   Physicochemical processes considered by this working party
 include chemical clarification, filtration, adsorption, and nitro-
 gen  removal. Another extremely important application of a
 physical process, clarification  and thickening of biological
 slurries, was considered as a primary subject by  several other
 working parties and, therefore,  was  not included  in  the
 discussions of this group.  Let it suffice to point out that great
 potential payout exists for automating activated sludge bio-
 logical systems so as to optimize the performance of secondary
 clarifiers.

 PROBLEM AREAS AND IDENTIFIED RESEARCH NEEDS
   Relative to the  automation  of physicochemical treatment
 processes the following problem areas were delineated:
     1.  Although existing data on lime clarification would in-
        dicate that  pH is the only parameter required to prop-
        erly   control  the  process,  it  is  questionable
        whether such a control strategy is universally transfer-
        able to all  geographic locations. For example, in areas
        where the  carriage water is  relatively soft, it  may be
        necessary to add a polymeric flocculant to enhance
        solids-liquid separation. In  such a case a simple  pH
        control algorithm would not suffice  and  a more
         complex control strategy would be required.
      2. Chemical clarification using either Al(III) or Fe(III)
         requires, at  a  minimum, control of both  pH and
                                                        66

-------
                                                                                     PHYSICOCHEMICAL PROCESSES
    dosage of the chemical in response to the concentra-
    tion  of soluble  orthophosphate,  turbidity  or sus-
    pended solids, and alkalinity. Even though the litera-
    ture is  replete with appropriate parametric relation-
    ships, the task of formulating, evaluating and refining
    a suitable control algorithm remains.
 3. With  respect to chemical clarification,  furthermore.
    each  of the control schemes evaluated heretofore
    have provided only for control of the chemical dosing
    function in response to a readily measured parameter
    such as pH. No scheme has been provided to control a
    specific effluent-quality parameter, e.g., suspended
    solids or soluble orthophosphate  concentration. To
    accomplish this, the chemical precipitation and chem-
    ical  coagulation/flocculation  processes   would  first
    have to be  mathematically  modeled. The resultant
    models would then have to be coupled with a suitable
    model for the primary clarifier. Until such an overall
    model  is available,  optimal  control of the chemical
    clarification process cannot be established.
 4. Automation  of the deep-bed  filtration process has not
    advanced materially beyond what  has been common
    practice in the  water treatment  industry for  several
    decades.  Control schemes  that  are  currently  em-
    ployed in practice generally relate  to two  factions:
    (1) now splitting; and (2) backwash initiation. While
    these  are  very  useful,  they are  not  amenable  to
    establishing optimal control algorithms. If descriptive
    dynamic models of the filtration process were avail-
    able, they  could  be employed  in a  feedforward
    manner to determine optimal control strategies so as
    to  maximize  filter  run times, thereby  minimizing
    backwash water requirements and recycle  flows. Such
    a  dynamic model  would  have  to consider  filter
    loading,  bed  penetration,  head  loss, and effluent
    turbidity. Moreover, any filtration  model would have
    to provide for prediction of the effects of adding a
    filter aid and be  capable  of controlling the  dosing
    function.
 5.  Operation of a series of deep-bed filters that are part
    of  wastewater   treatment   systems  frequently  is
    plagued either by cascading or  effluent deterioration
    effects  during  periods  of high  flow.  Conversely.
    during periods of low flow,  backwash water storage
    requirements increase due to the  lack of water which
    can  be  obtained  from  the  main  flow.  Suitable
    backwash supervisory  (executive control) strategies
    can materially enhance operation  of deep-bed  filtra-
    tion  systems  by either preventing or minimizing the
    problems mentioned.
6.  Although  a  dynamic  mathematical model  which
   describes the  adsorption  of organics  in municipal
   wastewaters onto granular activated carbon was devel-
   oped  under  an  EPA  contract,  it has  not  been
         employed  to  evaluate the feasibility of alternative
         control strategies. For example, a treatment objective
         might  be to minimize  carbon loading subject to the
         constraint  of  meeting  effluent standards. Using the
         existing model, this as well as other control alterna-
         tives could  be evaluated.
      7. Treatment  of industrial  wastewaters using granular
         activated carbon is commonly beset with problems of
         chromatographic displacement of adsorbed organics.
         Solute displacement  can be  prevented through the
         application of feedforward control using a descriptive
         dynamic model. A need exists,  therefore, for the
         development of a dynamic model and for formulation
         of an appropriate feedforward control strategy.
      8. Because adsorption contactors are generally designed
         and  operated  in  a fashion analogous to  deep-bed
         filters, considerations  regarding cascading and  efflu-
         ent  deterioration  effects also apply.  Accordingly,
         executive control strategies  must  be developed for
         adsorption systems as well,
      9. Because of the lack of a complete understanding of
         process kinetics, removal of  nitrogen  from  waste-
         waters either by air stripping, selective ion exchange or
         breakpoint  chlorination cannot yet be reliably accom-
         plished over the long term. This shortcoming can be
         resolved by a  complete  parametric description  of
         these  processes, by dynamically modeling  the  pro-
         cesses and then by  developing and evaluating control
         strategies.
     10. Process monitoring and control can only be accom-
         plished  if  reliable sensors  are available.  Although
         many sensors required  to automate physicochemical
         processes are  available, the need  exists to improve
         their  reliability  and  to  reduce   the  maintenance
         requirements. This  is particularly important for the
         measurement of flow, pH, phosphorus, ammonia, and
         organic carbon (TOC, TOD, COD, etc.)

PRIORITIZED RESEARCH NEEDS
1. Develop  and/or  refine structured dynamic mathematical
   process models for:
   a.  Lime clarification with or without addition  of polymeric
      flocculants.
   b.  Chemical clarification using either Al(III) or Fe(III).
   c.  Deep-bed  filtration  of  wastewaters  (single   or   dual
      media).
   d.  Adsorption of organics from municipal  wastewaters.
   e.  Adsorption of organics from complex industrial waste-
      waters.
   f.  Removal  of nitrogen from waste  streams by either air
      stripping, selective ion exchange, or breakpoint chlorina-
      tion.
2.  Develop  practical  control strategies for process operation
   for each of the processes listed in 1. above.
                                                     67

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS


3. Develop executive or supervisory control strategies for:           organic carbon sensors and reduce their respective mainte-
   a.  Backwashing of a series of deep-bed filters.                   nance requirements.
   b.  Backwashing of a series of adsorption contactors.           6. Couple dynamic models  for individual processes into an
4. Evaluate process models and candidate control strategies on       overall  dynamic model for an entire independent physico-
   a pilot scale (Blue Plains) and subsequently at a full-scale       chemical  wastewater treatment facility and subsequently
   wastewater treatment plant.                                   explore and evaluate cost-optimal control  strategies for the
5. Improve reliability of flow, pH, phosphorus, ammonia and       entire plant.
                                                         68

-------
                                  AUTOMATION
                                             OF
SLUDGE PROCESSING, TRANSPORT & DISPOSAL
                                  Workshop on Research Needs
                          Automation of Wastewater Treatment Systems
                      69

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                      AUTOMATION  OF  SLUDGE  PROCESSING,
                                  TRANSPORT  AND  DISPOSAL

                           Bart T. Lynam, Raymond R. Rimkus and Stephen P. Graef
                             The Metropolitan Sanitary District of Greater Chicago,
                                        100 East Erie, Chicago, IL 60611
 INTRODUCTION
   The  Workshop mission of defining  research needs and
 establishing priorities in the area of wastewater treatment
 systems automation is a positive step toward improving facility
 operation. The benefits  are  especially germane  to large
 municipal utilities such  as the Metropolitan Sanitary District
 of Greater Chicago.  In recent years it has been the District's
 experience that process automation, where feasible, yields the
 following benefits:
   1. Performs routine tasks
   2. Provides process information
   3. Enacts control  action decisions based on process status
   4. Provides  information for  a planned maintenance  pro-
      gram
 These benefits in effect
   •  Enable the operator to cover larger service areas
   •  Improve process quality and stability
   •  Reduce capital  and M & O costs
   The  District has either implemented or experimented with
 automation in  all three  areas of sludge processing, transport
 and disposal. The District's experience suggests that the area of
 sludge processing has more potential  application for automa-
 tion  than does sludge transport  and  disposal.  Transport and
 disposal  operations  are usually carried out  by multiple,
 single-unit  pieces  of equipment  mobilized  over  a broad
 territory and are thus more difficult to regulate automatically.
 The areas of sludge  transport and disposal do have a serious
 need for basic development work of a  general nature: however,
 this is beyond the scope of this workshop.

 SLUDGE PROCESSING
   The objective of sludge processing is to convert raw sludges,
 which are produced in wastewater  treatment, into a form
 which can be disposed of in an environmentally compatible
 fashion. Raw sewage sludge  production fluctuates somewhat
 proportionally to  the changes in raw sewage flow rate. The
amplitude is damped and out of phase compared to  that of
raw sewage flow rate variations, because  of  the dynamics of
the settling and biological oxidation processes. Unfortunately,
the discontinuous  schedule  for transferring  sludges,  which
many plants follow, causes abrupt changes in volumetric and
mass loading rates to sludge processing units. Equipment is
available, such as variable speed pumps and surge tanks, which
can smooth the volumetric loading rates to sludge processing
units.
   Fluctuations in mass loading, on the other hand, are more
difficult to regulate. First of all the dynamics of few, if any. of
the sludge handling processes have been experimentally tested
on a full scale  basis in the field. This is in spite of the fact that
numerous dynamic  studies on sludge  processing have been
performed  on  a  pilot  basis  or by  computer  simulation
techniques. These  field studies  are very important for imple-
menting automatic control and should be undertaken.  Sec-
ondly,  solids concentration and/or sludge density have not
been reliably measured  on-line  under  long-term field condi-
tions. Entrained gas,  stringy materials,  inadequate velocity
gradients  and grease  coatings  are several  problems which
invalidate the accuracy of measured solids concentration. Field
testing and development is needed to make the sensors work.
For  example  sensor  redundancy, sensor orientation,  and
preconditioning of the sludge before meeting; e.g., degasifica-
tion and  grinding  should  all  be explored as  means for
improving the solids concentration measurements.
   Research needs for implementing automation in the various
types of solids handling processes are  presented below. The
needs outlined are primarily oriented toward automation of
field units  which process municipal sludges on an continuous
basis.

Concentration
   Sludge concentration  is usually accomplished by two types
                                                        70

-------
                                                                         SLUDGE PROCESSING, TRANSPORT AND DISPOSAL
 of thickeners, (1) gravity and (2) flotation. The research needs
 for these units are as follows:
    Gravity Concentration
      1. Perfect devices for automatic monitoring and control
         of underflow sludge concentration.
      2. Develop  a  field-verified  dynamic  model  for  the
         gravity thickening process.
    Flotation Thickening
      1. Develop  feedforward control  of flocculant  addition
         based,  for example, on  measured influent  sludge
         concentration.
      2. Perfect feedback trimming of polymer  addition rate
         and pressurized  recycle  flow rate based on a process
         variable; e.g., subnatant turbidity.
      3. Develop  a  field-verified  dynamic  model  for  the
         flotation thickening process.

 De watering
    Municipal  sludges  may   be  dewatered  mechanically  via
 centrifuges,  vacuum filters  and  filter  presses, or  by heat
 treatment. Control actions  for mechanical dewatering involve
 either  changes  in mechanical  settings  and  parameters   or
 variation in  the rate of sludge conditioner chemical addition.
 In most instances, chemical addition control can be  adjusted
 more  readily  than  mechanical  control settings. Potential
 control  actions for heat treatment include  reactor residence
 time and reactor temperature. The research needs for mechani-
 cal dewatering and dewatering by heat treatment include:
    Mechanical Dewatering
      1. Develop feedforward and/or feedback control schemes
         for  addition of chemical conditioners.  The  District.
         for  example,  controls  ferric  chloride addition  to
         waste  activated sludge  based  on  the conditioned
         sludge  pH. Other potential  feedback control  signals
         and sensors may be on-line viscometers monitoring
         the  conditioned sludge stream  and on-line  turbidi-
         meters monitoring   the  filtrate  or  centrate  streams
         from these units.
      2. Develop a reliable field measurement system for esti-
         mating solids concentration of sludge cakes.

   Heat Conditioning
      1. Refine  a  feedforward/feedback  control  system   to
         regulate reactor temperature  and residence  time based
         upon influent total and volatile solids concentration.
         Several manufacturers have  steady-state control sys-
         tems and these could  be  augmented  to  provide
         dynamic  control.  Since  the  settling rate  of heat
         treated solids is rapid, a settlometer  could potentially
         be modified to obtain an effective feedback signal.

Combustion
   Manufacturers of most wet and dry combustion processes
have incorporated steady state control into their combustion
systems. These  systems  may include feedforward control  of
thermal  energy  input, air input rates and reactor  pressure.
based on total and volatile solids loading rate measurements.
Stack gas composition  and temperature, ash content  of the
combustion residue and  organic content  of the recycle liquor
measurements have been used for feedback trimming control.
In general, equipment and technology for automatic regulation
of combustion processes are more  developed than for other
sludge processing methods.

Heat Drying
   Like  other  mechanical equipment  for processing sludge,
automatic regulation has also been incorporated into proprie-
tary  heat drying systems. Research needs for improving upon
present  techniques for the heat drying process are as follows:
   1.  Develop means for determining an  on-line water balance
      around the reactor.
   2.  Develop an automatic control scheme  for blending dry
      solids with feed sludge.
   3.  Formulate  a  dynamic control strategy  for regulating
      energy  utilization,  stack gas quality and volatile content
      of the dried sludge.

Digestion
   Plant  scale automatic control of  biological  processes  is
practically non  existent. There have been, however,  several
dozen pilot  plant and computer simulation studies in  which
automatic control strategies have been formulated for these
processes. These studies have indicated that both aerobic and
anaerobic digestion appear suited   for  automatic control
actions  and the  research needed to achieve implementation
include the following:
   Aerobic Digestion
      1.  Perfect  sensor installations which yield reliable mea-
         surements  for  obtaining organic carbon or  similar
         balances on a full scale reactor in the field.
      2.  Develop aeration  or  oxygen  supply  rate control
         techniques for plant scale digesters. Potential feedback
         control signals include off-gas composition from and
         dissolved oxygen  concentration  in  the  reactor. An
         aerobic digester could  potentially  be  operated as a
         continuous  flow respirometer with respiration rate
       -  governing the rate of aeration.
     3. Develop a  field-verified  dynamic control model for
        the aerobic digestion process.

   Anaerobic Digestion
      1. Develop equipment  and  instrumentation  systems
        which would automate accurate  digester sludge feed-
        ing and withdrawal on a continuous basis. The system
        should be reliable when handling the heterogeneous
        materials in municipal sludges.
     2. Automate digester stability evaluation based on mea-
        surements such  as gas  composition and flow, in-line
                                                          71

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
        pH and short  chain fatty acid concentration in the
        liquid phase.
      3. Develop a feedforward off-line fermenter for screen-
        ing the raw sludge for inhibitory characteristics.
      4. Develop a field-verified dynamic control model for
        the anaerobic digester.

 Composting and Air Drying
   Sludge  handling processes  such  as composting and air
 drying are not especially suited for automation. Both processes
 incorporate  natural forces, have extended reaction times, and
 utilize mobilized solids handling machinery which are operated
 by one or more persons.

 SLUDGE TRANSPORT
   At the  present state of the art, automation in the area of
 sludge transport may be applicable in  only a few situations.
 The  best suited areas for control implementation seem to be
 pipeline transport of sludge slurries and conveyor transport of
 sludge cake solids and dry sludge  solids.  Other transport
 methods such as barge,  train and truck are not well suited for
 automation  since  control  of such equipment requires direct
 human intervention.  However,  in moderate to large  sludge
 transport operations which utilize barges, trains and/or trucks,
 a  small computer or time-shared computer terminal can be
 very beneficial for
    (a) Monitoring sludge shipment and storage
    (b) Developing operations research models for simulating
        and evaluating sludge routing strategies
    (c) Establishing   a  sludge management  information
        system.

 Pipeline Transport
   The District has experimented with automatic  control of
 pipeline transport  both  inside and  outside  of  the  plant.
 Automatic regulation of fluid flow can be very effective. The
 research needs  to  implement automatic control of sludge
 transport include the following:
    1. Develop reliable apparatus and techniques  for starting
      and stopping pumps, opening, modulating  and closing
      valves and routing sludge  through a network  of piping
      based upon volume  transferred and instantaneous flow
      rate.
    2. Develop automatic control of sludge viscosity reduction
      techniques such as (a) high energy mixers for increasing
      the rate of shear and (b) addition of chemicals which
      reduce viscosity. On-line viscometers  could serve as a
      feedback control sensor for regulating both mixer speeds
      and rate of chemical addition.

 Conveyor Transport
    Conveyors  have  been used extensively in  the automatic
 mode  for feeding  dry chemicals to assorted unit processes.
 Feedback  control  of belt  speed  based  upon  belt scale
 measurements insures a  desired  mass feed  rate.  Additional
 work is  needed  to perfect sensing devices  which estimate the
 moisture content of both sludge cake and  dried sludge solids.
 Such a  sensor could provide a feedback control signal for
 regulating the rate of sludge delivery on a dried solids  basis.
 SLUDGE DISPOSAL
   Sludge disposal by land spreading, landfill or ocean disposal
 is not especially  suited for implementing automatic control.
 Individual pieces  of machinery and equipment used in sludge
 disposal may be equipped with some automatically controlled
 mechanisms but total automatic  control has not been imple-
 mented. Although considerable research is needed in the broad
 area of sludge  disposal  few avenues for applying automatic
 control techniques seem available.
                            CONTROL OF SLUDGE HANDLING:
                            SOME  SUCCESSES  AND  PROBLEMS

                                                   J. B. Parrel!
                     Chief, Ultimate Disposal Section, Advanced Waste Treatment Research
                  Laboratory, National Environmental Research Center, Cincinnati, OH 45268
INTRODUCTION
   Steady advances, dating back to the late 1930's, have been
made in process control-advances in instruments for measur-
ing process variables as well as in control systems.
   Instrument  development  has progressed beyond the mea-
surement of primary  variables such as pressure, temperature,
and  density, and instruments can rapidly measure complex
properties such as flow rate  of  heterogeneous mixtures or
composition of gas or liquid mixtures. Instrument develop-
ment sometimes seems slow  or inadequate because there are
almost  as  many  needs  as  there  are  processes.  Complex
properties must be measured: the quality of a bread dough, or
the plastic and thixotropic behavior of a  paint,  or the
filterability of a sludge. As the needs become more specialized,
the economic  motivation to instrument developers becomes
                                                       72

-------
less because the numbers of potential installations are fewer.
Nevertheless,  the  need  exists,  and improved  control  of
complex processes depends on such developments.
   The  chemical  and  petroleum  industries  offer  excellent
examples of successful application of process instrumentation
and  control.  Several factors   have  played a  part in this
development.  Very often the  materials that  are processed
are simple fluids. Almost always the feedstock is quite uniform in
composition, the flow rate is held constant  or is controlled at
will, and the chemical reactions are straightforward. As under-
standing of the process  steps  diminishes or  the  feed  or
intermediate streams become complex fluids, control dimin-
ishes as well.

CONTROL IN WASTEWATER PROCESSING
   Wastewater processing combines  a number of factors that
work  against satisfactory control. Flows vary widely and un-
predictably. Feedstock composition is poorly defined and fluc-
tuates  greatly. Process steps are poorly  understood. Residue
streams (sludges) that result are complex  in physical and chem-
ical characteristics. It is no surprise that wastewater processing.
particularly  sludge processing,  has not seen  extensive applica-
tion of instrumentation and control systems. Some reasons are
related to the nature of the "products", the scale of operation.
and the interest of top management. The "products" of waste-
water treatment do not return a profit, so there is little incent-
ive  to produce a superior product. Most plants are too small to
justify instrumentation for optimum performance. Top man-
                                                                        SLUDGE PROCESSING, TRANSPORT AND DISPOSAL
 agement  (the municipal authorities)  is understandably  more
 likely to  be concerned about creative innovation in areas  other
 than wastewater treatment.

 SOME SUCCESSES
    The most extensive and successful use of sophisticated con-
 trol techniques  in  sludge processing is in sludge incineration.
 There have been very good reasons for this development.  In-
 cinerators work  poorly  and use excessive quantities of fuel
 when they are controlled manually, and manufacturers would
 have seen their  business disappear under  the combined on-
 slaughts of a populace interested in clean air and strict air pol-
 lution codes. A  second reason is that  significant process  varia-
 bles  can  be  measured accurately, and relationships  between
 variables are well understood.
    The manner in which excess air is controlled in a multiple-
 hearth furnace (MHF)* is illustrated in Figure 1. Control  is ex-
 ercised by pressure at the outlet of the furnace or by oxygen
 content  of the  cooled exhaust gas. In one mode, pressure is
 sensed at the furnace outlet and transmitted  to  a two-mode
 controller, which positions a damper in the exhaust line. When
 conditions have  stabilized, control  is manually transferred to
 the second mode of the  controller.  Oxygen content of the ex-
 haust gas is measured and a proportional signal is transmitted
"Information on control systems has been obtained from the Enviio-
 tech Corporation. Mention of products or  use of such information
 does not indicate EPA endorsement.
   FEED
                                                                                                  TO  STACK
                                                                         SCRUBBER
Fgure 1. Draft Control for MHF.
                                                         73

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS


to the controller, which controls the damper (and the exhaust
gas pressure) to give the desired air flow and exhaust gas oxy-
gen content. This is an example of cascade control.
   Temperature on hearths is controlled by a conventional
 feedback circuit (Figure  2). Temperature is sensed, and if it is
 off the setpoint, the air flow rate is changed in the proper di-
 rection. A ratio controller changes fuel flow in proportion to
 air flow to give a constant percent excess air.
   Variations in flow rate of solids can cause large fluctuations
 in the fuel requirements on the different hearths. For example,
 an increase in flow rate can cause  the burning zone to travel
 down the furnace and start at a lower hearth; this would sub-
 stantially  change the fuel demand on  the  hearth. A feed-
 forward system has been  devised (Figure 3) that increases the
 rabble arm speed as feed rate increases. The increased agitation
 of the rabble teeth exposes more surface, which tends to make
 the sludge start  burning sooner. The two tendencies counter-
 balance each other, and the burning zone  remains fixed.
    Successful  applications of instrumentation are not limited
 to incineration.  Magnetic flow meters  and sludge  density
 gauges  are used in  open-loop control of sludge streams. Con-
 stant-displacement pumps and conventional pressure, tempera-
 ture, and flow controls are used to achieve closed-loop control
 of the heat treatment process.

 A DIFFICULT CONTROL PROBLEM
    Some sludge  handling processes and operations could  be
 profitably controlled, but so far, satisfactory feed-forward or
 feedback control techniques have not  been devised. One ex-
 ample is vacuum filtration. The operation is costly, so there is
 incentive to reduce costs. The major expense is  the cost of
 chemical conditioner, which conceivably could be reduced by
 proper control procedures. The scale of operation is import-
 ant. Some filters are  too small to automate; they  do not con-
 sume enough conditioning agent to warrant the cost of a con-
 trol system. However, a 500-square-foot filter, which is a com-
 monly used  size, can process over 4,000 dry tons of sludge per
 year and can consume $40,000 worth  of conditioning poly-
 mer. If a control system could reduce this cost by 10 percent,
 it would be a sound investment.
     A vacuum filtration installation is illustrated in  Figure 4.
 Best operation is achieved when each filter has its own sludge
                              FEED
EXHAUST
                                                                       	(TRd	
                                                                         	TRQ—
                               AIR
         \
       ASH
 Figure 2. Temperature Control on MHF Hearths.
                                                         74

-------
conditioning system. In most plants, control is manual. Sludge
flow rate is controlled by varying the speed of a constant dis-
placement pump. Conditioning agent is  fed in proportion to
the sludge feed rate by changing the adjustments on a chemical
proportioning  pump. The  proportion between polymer and
sludge is established by bench-top filtration tests.
  If the filter capacity is adequate to handle the plant capa-
city in the allotted operating period, a suitable control object-
ive would be  to  minimize  polymer  consumption. A simple
automatic control procedure that is used in some plants, parti-
cularly those in the chemical industry, is to control sludge feed
by the level in the sludge feed pan and match the polymer
dose to sludge flow by means of a ratio controller. This proce-
                                                                   SLUDGE PROCESSING. TRANSPORT AND DISPOSAL
dure does not account for changes that occur in the solids con-
tent of the liquid sludge and in the filterability of the sludge.
An alternative procedure, which is under test by a filter manu-
facturer, is to add to this  control system a continuous mea-
surement of solids content of the sludge by means of a nuclear
density gauge. The polymer dose is matched by the ratio con-
troller to the product of volumetric flow rate and sludge solids
content; i.e.,  the mass flow rate of dry solids. The control pro-
cedure now accounts for changes in sludge solids  content but
still does not account  for  changes in the filterability of the
sludge.
  Swanwick (1)  describes  the use of a rapid  filterability test
to control dose of chemical conditioning agent to a sludge that
              BELT  SCALE
                             BELT CONVEYOR
                                                                FEED
                    I WEIGHT TRANSMITTER
                    *
                        RATIO CONTROLLER
                                                                    FURNACE
                      SPEED CONTROLLER/"        A
                                     TRANSMITTER     T
                                                                                     ASH
                                                                              SHAFT DRIVE
Figure 3. Control of MHF Rabble Arm Speed.
                                                      75

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                            CONDITIONING
                                            TANK
                       FILTER
         POLYMER
                                                                                             CAKE
                                                                                  TO VACUUM
           SLUDGE
 Figure 4. Vacuum Filtration of Sludge.

 was to  be pressure  filtered. The test, which combined hand-
 operated mixing operations with an automated version of the
 capillary suction test (CST), took 4 minutes to complete and
 produced a recorded output. The sludge was dosed with suffi-
 cient lime to give a CST value below an empirically determined
 critical  value. This method eliminates the need for a measure-
 ment of the solids content of the sludge. If sludges are blended
 before  filtration so  that changes in solids content and filter-
 ability are slow, the time required for the number of CST tests
 needed to establish optimum  conditions should not  seriously
 limit the  ability to  use these  measurements to automatically
 control conditioning chemical dosage.
FUTURE DEVELOPMENTS
   It seems certain that the pace at which instrumentation and
control systems are introduced into sludge handling processes
will accelerate,  particularly in plants that handle large waste-
water flows (for example, in excess of 30 million gallons per
day).  The primary problems to be overcome are inadequate
understanding of complex processes, such as digestion, and the
lack  of  suitable instrumentation  for measuring  complex
phenomena.

REFERENCES
1.   Swanwick. J. D.. Water Poll. Control. 72. 78-86 (1973).
                                              DISCUSSION
 Richard I. Dick:
    Is there  any potential for improving and automating  the
 performance of land systems for sludge disposal by use of in
 situ sensors for constituents such as soil moisture, ORP, oxy-
 gen concentration,  ammonium, and nitrates? It would seem
 that sludge  application schedules could be optimized, ground
 water contamination could be minimized, and possibilities for
 reduction of the nitrogen  load by controlled nitrification-
 denitrification  could be realized if  such  a  system were
 implemented.
Ted Lejeune:
   At the R. M. Clayton plant in Atlanta it is planned to pro-
vide  polymer dosing control based on centrate density. So far
we have no operating experience with this method. As an alter-
 nate, a ratio proportioning meter is available.
   We have experienced continual problems in  trying to use
 sludge density and flow meters, due  to blockages and materials
 such as sticks in the sludge. Because  of these problems, the At-
 lanta computer will be used mainly for  logging, rather than
 control purposes.
                                                        76

-------
                                                                         SLUDGE PROCESSING, TRANSPORT AND DISPOSAL
   The multiple-hearth furnace at Atlanta is operated on an in-
termittent basis, and we feel that control of the final burn-out
is best done manually. This also points out the need for better
training of operators for automated plants.

Carmen F. Guarino:
   In  contrast to the experience  in  Atlanta, I  want to point
out that there has also been good experience in operating
sludge density gauges and flow meters in Philadelphia and Los
Angeles.

Richard I. Dick:
   Traditionally, problems  of sludge treatment and disposal
have been addressed independent of the  problems related to
the wastewater treatment process which generates the sludges.
This  is regrettable, for overall  optimization  would  require
simultaneous consideration of all aspects of the entire treat-
ment  process.
   It  would be unfortunate to extend this  traditional approach
to automation. Shouldn't attempts  to automate sludge handl-
ing extend beyond the sludge treatment and disposal area and
go  back to the  wastewater treatment  process  itself? For ex-
ample,  are there opportunities for  controlling the quantities
and physical, chemical, and biological characteristics of sludges
by feedback  control  of  the processes which generate the
sludges?

John  F. Andrews:
    I agree that one should guard against looking at individual
processes in isolation. Work is in progress on the development
of an overall plant model,  which will enable us to explore in-
teractions between units, and start  to  think about optimiza-
tion of the total operation.

P. .M. Berthouex:
    There are  two  ideas I invite  Steve Graef to comment on.
First, we have talked mostly about automation to monitor or
control traditional processes. Can't we use control problems to
direct our thinking toward  design  changes and the develop-
ment of new processes. One use for theoretical models is to
explore  process and operational changes. Promising ideas are
identified for pilot plant or full-scale trials. This is true even if
the model is never implemented for control, or, perhaps, even
if the model  is later found to have inadequacies.
    Secondly, the operator himself may be a good sensor. Some
operators can adjust to color and  sludge appearance. This is
not optimal  control, but it is control. As an example I cite the
use of closed-circuit television at the Blackbirds Sewage  Treat-
ment Works, England. Sludge  flows over a circular wire  and
 the operator can  estimate sludge density by appearance. It is
 simple and apparently helpful. It is also automation in a sense.
 Clever design ideas such as  this may be an alternate-perhaps
 an intermediate solution-to more sophisticated automation
 systems.
John F. Andrews:
   A sensor which could be of significant value in anaerobic di-
gestion is an on-line calorimeter for continuous monitoring of
the caloric value of digester gas. The discusser has observed the
use of such an instrument at the San Jose, California, waste-
water treatment plant in which the caloric value  of the digester
gas was measured and this signal used to control  the amount of
natural gas blended with the digester gas. The result was a gas
with a constant caloric value for improving operation of the
gas engines used in this plant.
   The discusser would like to  suggest that the  rate of energy
production, as calculated from the calorimeter and digester gas
flow rate measurements, would be a valuable indicator of the
condition of the anaerobic digestion process. This is analogous
to the rate of methane production which has been proposed as
an  indicator  of digester condition by Graef and Andrews (1).
Moreover,  this measurement, when used in combination with
COD measurements on the  feed to the digester, could be used
to  compute  the  efficiency of the digester with  respect to
energy removal from the feed sludge, since one  pound of COD
is approximately equivalent to 6,300 BTU of energy (2).

REFERENCES:
 1.   Graef, S. P. and Andrews. J. F., "Stability and Control of Anaero-
     bic Digestion," Jour. Water Pott. Control Fed., 46, 666, (1974).
2.   Andrews. J. F. and Kambhu, K., "Thermophilic Aerobic Digestion
     of Organic Solid Wastes," Final Report  to Office of Solid Wastes,
     U. S.  Public Health Service, Environmental Systems Engineering
     Department, Clemson University, Clemson, SC (1971).
 CLOSURE
 Stephen P. Graef:
    There is a potential for improving land application systems
 via  soil moisture, ORP, oxygen concentration, ammonia, ni-
 trate, etc., surveillance. Moreover, the benefits recognized by
 Dr. Dick can and have been realized. Automation, however,
 would not enhance a land application operation. Most land ap-
 plication variables change  quite slowly; Le., on  the order of
 weeks and months rather than hours and days. Although com-
 plete  and accurate surveillance is essential for gauging the con-
 dition and performance of a land application program, the
 monitoring does not have to be  continuous or on-line. Periodic
 sampling and analysis both in situ  and at the laboratory can
 provide the information needed to effectively manage a land
 application program.
    Automation and continuous on-line  surveillance can  be a
 distinct advantage in  controlled small-scale pilot studies. In a
 research environment as opposed to a large-scale production
 effort, agricultural instrumentation  and  control apparatus can
 be  serviced efficiently  and can provide the large volume of
 data needed for investigative studies.
    In his second question, Dr. Dick states an important con-
 cept. Efforts  to  automate  sludge  handling as well as other
 wastewater treatment processes should not be  confined to the
 process itself. The control strategy should take into considera-
                                                            77

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
tion the interactions between the process under consideration
and others throughout the plant.
   Such a system control strategy is developed on two levels.
First a strategy must be formulated for each component of the
system. The component control scheme must enable the pro-
cess to meet its objectives, given the range of inputs which can
be expected from the system. A more encompassing system
strategy should then be developed which balances the inter-
actions and tradeoffs among processes to meet the objectives
of the system.
   As an illustration consider a system, two components of
which are the  activated sludge and anaerobic digestion pro-
cesses. Assume that under severe conditions the digester may be
required  to process 0.45 Ibs. of volatile solids/ft.3 day for a
short period. A strategy of pH, temperature, and gas recircula-
tion  regulation must be formulated if the activated sludge pro-
cess does indeed provide such large masses of volatile sludge to
be processed from time to  time. Under normal  conditions a
system strategy could control the sludge produced by the acti-
vated sludge process within  0.25 to 0.3 Ib VSS/ft.3  day. By
regulating SRT and sludge blanket depth in the final clarifier,
the organism loading to the digester will also be regulated. A
system strategy should  be formulated which  will control and
balance the interactions so that the total output of the system,
e.g. digested sludge and clarified secondary effluent, meet or
exceed the quality criteria.
    In response to Dr. Berthouex's first question, math models
and computer simulations are powerful tools for evaluating de-
sign  changes and development of improved process trains, if
not improved processes as well. It is probably the only way a
 designer can answer dozens of questions such as, "What would
happen if...", without an enormous hand calculation effort.
Moreover, some questions of this sort could not be answered
 at all without computer simulations. As indicated, several par-
ticipants  in this workshop have successfully utilized theoretical
models  to  formulate   productive  process  and  operational
 changes.
    In response  to the second question, there  is no reason for
 intentionally making automation and control  techniques com-
 plicated.  Process output criteria or objectives  should be speci-
 fied  for the designer. He should  in turn provide the control
 scheme to meet the requirements. The scheme should be  reli-
 able, easy to use by operations personnel and easy to maintain.
J. B. Fairell:
    In response to Mr. Lejeune's comment, the use of centrate
 density (i.e., solids content) is an interesting way to control
polymer dose for centrifugation. Unfortunately, gamma ray
density  gauges are not very sensitive indicators of sludge solids
content in the range of 0 to 1% solids. Mr. Lejeune might want
to consider using a turbidity meter to indicate solids content.
If the centrate is too  opaque, the centrate stream could be di-
luted in a  fixed proportion  with clear  water to get an on-
stream reading.
   Mr. Lejeune  observes that the multiple-hearth furnace  at
Atlanta will  be operated on an intermittent basis,  and that
burnout (in preparation  to shutdown) is best done manually.
He is probably right that during burnout, when sludge feed is
being shut off, control should be turned over to manual opera-
tion. I suggest that Mr. Lejeune should do all that he can  to re-
duce the number of shutdown-startup cycles for the incin-
erator. Operation of a multiple-hearth incinerator for one shift
a day with shutdown for two shifts consumes excessive quanti-
ties  of  supplementary fuel and disproportionately  increases
maintenance  cost. We should really be thinking of continuous
operation of multiple-hearth incinerators with  shutdown  at
most once a week rather than once a day. Perhaps we should
design to run the dewatering room during the first part of the
week, provide intermediate sludge cake storage, and run the in-
cinerator continuously with the staff of the dewatering  room
during the second half of the week. Alternatively, the dewater-
ing  and incinerating  equipment could be  operated continu-
ously only during the first few days  of the week, but this pro-
cedure presents awesome personnel scheduling difficulties.
   In response  to Professor Dick's question, I believe that the
design of the water purification processes  of a sewage  treat-
ment plant and the sludge treatment and  disposal processes
should  be  considered simultaneously, with  the intention  of
minimizing   cost  and  optimizing  performance.  Processes
selected and conditions  at which  the processes are  operated
have a strong effect on sludge quantity and properties. As an
illustration, it is well known  that the quantity of biological
sludge produced in the activated sludge process depends on the
operating conditions of the process. However, these conditions
are fixed within relatively narrow limits by the initial design. It
is difficult to picture a  control scheme, for example, which
would feed back into the  water  purification processes  when
the sludge  concentration at the sludge thickener outlet  starts
to decrease or when the sludge production rate exceeds a de-
sired figure. The water purification processes should primarily
be controlled to produce high quality water. Control schemes
for sludge handling and disposal should probably not  feedback
into  the water purification processes.
                                                           78

-------
                                                                    SLUDGE PROCESSING, TRANSPORT AND DISPOSAL
                                    Report of Working  Party
                                                      on
                        RESEARCH  NEEDS FOR  AUTOMATION
           OF  SLUDGE  PROCESSING, TRANSPORT  AND  DISPOSAL

                                                 Richard I. Dick
            Professor, Department of Civil Engineering, University of Delaware, Newark, DE 19711
                                                  John R. Trax
             Office of Water Programs, Environmental Protection Agency, Washington, DC 20460
   In spite of the fact that sludge treatment and disposal has
been a problem since  the first wastewater treatment plant was
built, there is probably a greater backlog of unsolved problems
related to automation in the area of sludge treatment and dis-
posal than in other areas. Development of technology in sludge
treatment and disposal has largely been ignored in the past and
even comparatively new processes (such as the chemical-physi-
cal processes developed in recent years) are probably closer to
being capable  of precise  automation than  are processes for
sludge treatment and disposal.
   Because of the larger number of unsolved problems  relating
to sludge treatment and disposal, and because of the  accom-
panying difficulties in accomplishing effective automation of
sludge treatment and  disposal processes, the area deserves high
priority in competition for research funds. There is appreciable
potential for significant returns on research money spent in
this area.

STATEMENT OF PROBLEM
   Attempts to effectively and extensively automate sludge
treatment and  disposal schemes might be frustrated by a num-
ber of problems. Some of these are indicated below.
   1. Automation  of sludge treatment and  disposal facilities
     inevitably involves sensing of sludge properties and con-
     trol of sludge flows. To accomplish this, it is necessary
     that sludges  be  free of gross solids, rags, sticks, stones.
     etc. This problem is not  expanded in the later list  of re-
     search  needs  because it  is not  felt that research is wai-
     ranted on the  subject. However, in the considerations
     which  follow, it is assumed that complications caused by
     such materials  in sludges have been eliminated. Ques-
     tions remain as to whether the solids can best be elimi-
     nated from the  raw waste flow or from the sludge itself.
   2. To automate sludge treatment and disposal  schemes it
     will be necessary  to sense appropriate properties of the
     sludge. In the  case  of sludges, some  of the important
     properties are less well  defined than  with wastewater.
     For example, automated procedures for sensing physical
     parameters such as settleability and dewaterability may
     be difficult to develop.
   3. Attempts to effectively  automate sludge treatment and
     disposal processes may be expected to be hampered by a
    heritage of ineffective integration of the various steps in-
    volved in sludge treatment and disposal with each other
    and with the wastewater treatment processes which gen-
    erate  the sludges.  The relationships between various pro-
    cesses of sludge treatment and disposal are less well un-
    derstood than the relationships between various stages in
    wastewater treatment.
  4. Temporal  variations  in  sludge  characteristics  may
    hamper the performance of many sludge treatment  and
    disposal schemes.  An example is the anaerobic digestion
    process which, particularly in  cities of small and moder-
    ate size, often does not operate under stable conditions.
  5. Although  the performance of all processes of sludge
    treatment and disposal is related to the suspended solids
    content of  the sludge, it  is difficult to automatically
    measure those concentrations.
  6. Because processes for sludge treatment and disposal in
    general have not  been developed to as high a degree as
    those  for wastewater treatment,  appreciable  work is
    needed to develop predictive models describing the  per-
    formance of these processes.

RESEARCH NEEDS
   The following specific research needs are noted.
1.  Investigation  of the interaction  between various  processes
   used in sludge treatment and disposal must be undertaken
   so that  the overall process of sludge management can be op-
   timally  integrated.  In  the area of automation of sludge
   treatment and disposal, this is considered to be a major re-
   search need.  Additionally, benefits are to be derived from
   better integration of sludge treatment and disposal facilities
   with the processes which generate the sludges.
2.  Methods for sensing suspended solids concentration over a
   wide range of sludge consistency are needed. Three ranges
   of  suspended solids concentrations might arbitrarily be
   identified:
  •a. Dilute concentrations (filtrates,  centrates, etc.): turbidi-
     meters offer possibilities for sensing in this region.
   b. Intermediate  range  (concentrations typical of gravity
     thickener feed and underflow):  some need  for improve-
     ment of available  instrumentation in this area is seen.
   c. High concentrations  (filter cakes, centrifuge cakes, etc.).
                                                       79

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
      In spite of the importance of these measurements, means
      for sensing them in real time are not available.
3. There is a need for sensors to continuously monitor the
   physical properties of sludge which are important to various
   sludge treatment and transportation methods. Specifically:
   a.  On-line monitoring of the rheology of sludges.
   b.  On-line measurement of the settleability of sludges.
   c.  Continuous   measurement   of the  dewaterability  of
      sludges.
   d.  Continuous  measurement  of  the   calorific  value  of
      sludges.
4. Existing  steady-state  models  and developing non-steady
   state  models  of the performance of gravity thickeners
   would seem ready  for implementation. The final step of
   field testing of  these models to demonstrate the control
   strategy could seem to be important for future progress in
   sludge  thickening  and  automation  of sludge thickening
   processes.
5. The following research needs exist in the area of biological
   stabilization of sludges.
   a.  Improve the stability of the anaerobic digestion process
      for stabilization of sludges-Representatives of larger cities
      did not feel that this need deserved high priority because
      their  sludges  tend to be  more uniform in quantity and
      quality.  The need was considered to exist for moderate
      and small-sized cities, however.
   b.  Field scale evaluation of available models of anaerobic
      digestion performance  should be undertaken. Sensors
      such  as  those for gas composition, gas production rate,
      volatile  acid  concentration,  alkalinity and suspended
      solids concentration would be useful in undertaking such
      field  scale evaluation and refinement of models of anaer-
      obic  digester  performance  to develop a  demonstrated
      control strategy.
   c.  Similarly, existing models of the performance of aerobic
      digesters also should  be evaluated in field scale studies.
   d.  The effects of aerobic and anaerobic digestion  on solids
      separation and dewaterability need  to be evaluated and
      developed into performance models.
6. Sludge conditioning processes are not fully understood and
   are very difficult at present to automate. Two phases of re-
   search are required in this area:
   a.  Stage one would involve empirical, short-term evalua-
      tions of sludge conditioning  processes.  Control pro-
      cedures  would then be  developed on an empirical basis
      to regulate machine variables such as submergence or
      drum speed and control coagulant doses.
   b. Long-term research should be conducted to improve fun-
      damental understanding of conditioning and dewatering
      processes. This  research should lead to development of
      models  based on fundamental mechanisms involved in
      conditioning and dewatering.
7. In the area of incineration, the prospect for avoiding short-
   term transients by.storage and blending of dewatered sludge
   prior to introduction into incinerators  to assure a constant
   supply of sludge of uniform quality should be undertaken.
   Existing incinerator monitoring techniques for automation
   procedures are considered to be relatively well developed
   except that additional work is needed on suspended solids
   monitoring  devices, and devices for sensing the calorific
   value of sludges need to be utilized.
8. Currently,  little research  for automation of ultimate  dis-
   posal schemes such as land and ocean disposal is considered
   to be of high priority. Much additional  work, however, is
   needed in developing the ultimate disposal processes them-
   selves. In the area  of automation, there is need for a means
   of sensing the numbers of pathogenic organisms and viruses
   in sludge.

Priority of Research Needs
   As argued in the Introduction, all research needs in the area
of sludge treatment and disposal have a comparatively high pri-
ority because  technology in this area has lagged behind  that of
other wastewater treatment processes. Following is a ranking
of priorities within this high priority area:
  1.  The interrelations between the various processes used for
     sludge treatment and disposal need to be explored to al-
     low integration of the processes  with aid of automation
     procedures in an optimal way.
 2.  Develop  reliable  suspended solids concentration sensors
     for the entire range of solids encountered in sludge treat-
    ment  and disposal. Sensors for the low range of suspended
    solids concentrations (for flows such as filtrates and cen-
     trates), for the intermediate range of concentrations (typi-
     cal of liquid  sludges), and for the high range of concentra-
     tion (corresponding to dewatered sludges) will operate on
     different principles.
 3.  Develop  a "settleability meter." Such a device might be
    based on actual measurement of settling properties or by
    sensing a related physical property.
 4.  Develop a "dewaterability meter." Such a device is essen-
    tial for  automation of sludge dewatering  processes and
    monitoring of sludge conditioning processes.
 5.  Carry  out field-scale testing of gravity thickening models.
    Such models  currently are available but require extensive
    full-scale testing and possible modification.
 6.  Develop control  strategies for improving the stability of
    anaerobic  digesters. This is not a high  priority item  for
    large installations, but it is important for smaller installa-
    tions receiving occasional toxic discharges.
 7.  Develop an empirical model for control of conditioning
    and dewatering facilities.  This basis for automatic control
    of conditioning and dewatering processes is needed until
    more fundamental models can be developed.
 8.  Develop and apply sensors for sludge rheology. Such sen-
    sors should be useful for controlling pumping and piping
    operations and, potentially, other operations whose per-
    formance is related to the physical properties of sludges.
                                                          80

-------
                                                                         SLUDGE PROCESSING, TRANSPORT AND DISPOSAL
 9.  Carry out  field-scale evaluation of anaerobic digestion
    models. Such models currently  are available, but lack the
    extensive field-scale verification and modification required
    to achieve a demonstrated control strategy. Effects of an-
    aerobic digesters on  solids separation and dewaterability
    should be included.
10.  Develop rational models for  control of conditioning and
    dewatering facilities.  This is a more long-term solution to
    the same problem addressed in  high priority item number
    7 above.
11.  Develop a sensor for pathogenic organisms and viruses in
    sludge.
12.  Conduct field-scale demonstrations of controlled "steady-
    state" incineration accomplished by storage and blending
    of dewatered sludge prior to incineration.
13.  Evaluate the transferability of industrial solids-handling
    technology to municipal and industrial sludge  treatment
    and disposal.
14.  Conduct field-scale evaluations of models for aerobic di-
    gester performance. Effects of aerobic digestion on solids
    separation and dewaterability should be included.
15.  Develop and test on-line sensors for the calorific value of
    sludges.
                                                           81

-------
COMPUTER APPLICATIONS
          IN AUTOMATION
              Workshop on Research Needs
     Automation of Wastewater Treatment Systems
83

-------
   AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                   CURRENT PRACTICE  IN  INSTRUMENTATION  AND
   COMPUTER  APPLICATION  AT  THE  COUNTY SANITATION  DISTRICTS
                                   OF  LOS  ANGELES COUNTY

                                   Walter E. Garrison, Kip Payne and Tim Haug
                  County Sanitation Districts of Los Angeles County, 1955 Workman Mill Road
                                                Whittier,CA 90601.
  INTRODUCTION
    The  Sanitation Districts of  Los Angeles County plan, de-
  sign,  construct, and operate a sewerage system serving almost
  4,000,000  people residing in all or  part of 71 incorporated
  cities and adjacent County territory.  The basic operations are
  supported by a substantial research and development effort di-
  rected toward  improving existing design and operational tech-
  niques and  to pilot testing of advanced treatment processes for
  future water quality upgrading. District  management is dedi-
  cated to a  well-financed program to  improve plant reliabilily
 and performance through development and application of any
 advanced technology which will  aid in achieving this goal.
    The purpose of this paper is to present a manager's  over-
 view of the  present "State-of-ihe-Art" technology of the appli-
 cation of instrumentation  and  computers in  the  Districts1
 wasiewater  treatment plants and collection system. The text
 has been  prepared to briefly summarize salient features of the
 existing system, to explain the purpose and function of auto-
 mation in routine operation al the plants, to describe the cur-
 rent practice in computer application and, finally, to explore
 potential  uses  for future application  of instrumentation and
 computers.

 DESCRIPTION OF EXISTING SYSTEM
   All  service areas of the Districts are served by separate  sew-
 erage systems with minimum storm water  infiltration. Rainfall
 generally occurs for about three  months during the year. Thus
 inland plants discharge to dry water courses most of the year
 or to reclamation projects whenever available. Public contact di-
 rectly with plant discharges is common, requiring the mainte-
 nance  of  high-quality effluent,  adequately disinfected at all
 limes.
   In the principal Los Angeles drainage basin, approximately
420 million  gallons  per day of sewage  is  processed  by six
 wastewater treatment  plants. Five of the plants utilize the acti-
vated  sludge process  and are  strategically located inland to
 withdraw  raw sewage from existing trunk sewers which are
 tributary to a large primary plant near the ocean. The inland
 plants are unique in  that all solids removed in the treatment
 process are relumed  to the exJsting trunk sewers for central-
 ized  processing al the primary treatment  plant. Table I  sum-
 marizes  the flow parameters of the treatment plants on this
 system.
    Six other District  treatment plants ranging in size from 0.2
 MOD to 5 MOD serve other inland drainage basins. Solids pro-
 cessing at these planls is handled conventionally by use of two-
 stage anaerobic digestion followed  by onsite digested sludge
 drying  beds or  truck  transport  to  off site  drying and/or
 disposal.
    Al the large primary plant (JWPCP) 500 tons per day of
 raw sludge is stabilized in a highly automated single-stage,  high
 loading  anaerobic digestion process followed by a (wo-siage di-
 gested sludge dcwatering station. The  dewalering station  con-
 sists of a firsl stage, continuous-flow solid-bowl centrifuge fol-
 lowed by a second slage, baich-type basket centrifuge, the lat-
 ter currently under construction. Centrate from the dewater-
 ing station, at about  1000  mg/l,  will be recycled to the raw
 sewage channel, with the cake'discharged lo an open-bed com-
 posting  operation. Mosl of the dried solids will be sold as a fer-
 tilizer, with the balance hauled to landfill.
   Expansionof the above described system, based upon curreni
 federal  legislation, will consist  of construction of secondary
 treatment at  the existing primary plant. Pilot-plan! testing is
 currently in  progress for (his 300 million dollar expansion. An
 interim  short-term improvement in effluent quality will be
 achieved  by chemical  treatment using polymers,  lo achieve
 80% removal  of suspended solids, followed by  fine screening
 with a 20 mesh travelling water screen to remove remaining
 floatables prior to ocean discharge.
   At  the inland plants, planning for upgrading is concentraied
on advanced treatment techniques to reduce coliform bacteria
                                                       84

-------
                                                                          COMPUTER APPLICATIONS IN AUTOMATION
                                                     TABLE 1
                               LACSD TREATMENT PLANTS IN LOS ANGELES BASIN
   TREATMENT PLANT

   Primary Plant
    Joint Disposal
    Plant (JWPCP)

   Activated Sludge Plants
    Whittier Narrows WRP
    San Jose Creek WRP
    Pomona WRP
   *Los Coyotes WRP
    Long Beach WRP
DESIGN FLOW
     MGD
    450
      12.5
      38
      14
      38
      13
    "Includes 25 MGD currently under construction.
AVERAGE DAILY FLOW
          MGD
   (AS OF JULY 1973)
         350
           12
           31
            8.5
            8.5
            9.1
  RATIO OF PEAK
TO MINIMUM FLOW
        2.6
        1.5
        3.1
        4.3
        1.9
        4.3
to less than a median value of 2.2 per 100 ml and to the re-
moval of virus below levels of detection. Such sophisticated
treatment processes will not only assure greater public health
protection where human contact with undiluted effluent oc-
curs, but also will upgrade water quality to a level where reuse
projects will gain  much wider public support. Currently, the
Districts' staff is designing a carbon-adsorption advanced treat-
ment  module for the Pomona WRP which will  provide 10
MGD of water meeting the above criteria. Award  of contract
for construction is scheduled for December 1974.

AUTOMATION IN ROUTINE OPERATION
   The following discussion is intended to provide  a  general
summary of those instrumentation systems which are in rou-
tine use to provide operational control of District  plants. A
considerably more detailed report is  currently being prepared
under an EPA  research  grant titled. "State-of-the-Art Tech-
nology for Semi-Automatic Control of Activated Sludge Treat-
ment Plants". This detailed report is scheduled for completion
by March 1,1975.
   The management policy of the Districts encourages maxi-
mum  utilization of advanced control systems where it can be
demonstrated that improved reliability  and performance will
result. A continuing effort is being made to field-test new and,
hopefully, improved  control equipment  to clearly establish
capital,  operation and maintenance costs. Backup equipment
and/or manual override of all automatic controls  is obviously
                      necessary to assure continuous compliance with requirements
                      of regulatory agencies. Equipment described in this paper is
                      satisfactorily controlling plant operations, but improvements
                      are being sought which will reduce operational manpower re-
                      quirements and improve effluent quality at a reasonable capi-
                      tal cost and without excessive maintenance costs.

                      Inlet Pumping Controls
                        All of the influent  pumping plants in the LACSD system
                      are designed with variable speed pump and liquid level control.
                      In general, water surface elevations in the incoming sewers are
                      maintained  near  normal depth by varying pump speed with
                      depth of flow. Although this  proportional  relationship be-
                      tween depth  and pump speed does not exactly produce normal
                      depth for all flows, it approximates it closely enough to avoid
                      adverse effects caused by excessively high or  low velocities in
                      the sewer. Utilization of the potential storage capacity in in-
                      coming trunk sewers to reduce peak flow has not been aug-
                      mented since the warm climate and correspondingly high tem-
                      perature of the sewage would cause odor problems and higher
                      hydrogen sulfide  levels  than already  exist. Variable-speed
                      pumping  to  match  normal  water evaluations  eliminates
                      "breathing" of wet wells with attendant discharge of odorous
                      air to the neighborhood.
                        In a typical recent design, smaller pump units consist of a
                      vertical 10-inch  centrifugal pump, flexible shafting to a close
                      coupled magnetic drive, and electric motor. Larger pump units
                                                        85

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
are similar except for the addition of a right-angle gearbox
speed-reduction unit. Major components of the control system
consist of a liquid level sensor and transmitter, liquid level con-
trol  unit, and the motor/variable speed drive units used to
power the pumps.
   The level control system controls the influent pumps so that
the rate  of discharge from  the pumping station is approxi-
mately equal to the varying rate of sewage inflow. Level con-
trol  in the wet well is initiated by means of proper pump se-
quencing and by varying the pump speed. The system is capa-
ble of sequencing pump operation and varying the speed of all
pumps as necessary to pump a variable station flow without
storage in the wet well.
   Primary intelligence for the level control system is obtained
from an electronic differential pressure transmitter or  trans-
ducer. This unit is used to sense rising and falling liquid level in
a stilling well which is directly  connected to  the influent wet
well. Pressure sensed in the stilling well is converted to a 4-20
milliampere direct-current control signal, proportional to the
stilling well water level.
   The liquid level control unit receives the control signal gen-
erated by the differential pressure transmitter. This signal is
used as a pilot signal  for operation of the control equipment.
These operations include indication of the liquid level  in the
wet  well, initiation of start and stop functions for  all pumps.
modulation of pump speed, and actuation of alarms for high
and low water levels in the wet well.
   In a special situation where available flow in a trunk sewer
greatly exceeded treatment  plant capacity and the plant was
being operated for reclamation purposes, pump controls were
set to vary pump speed to maintain a constant water surface
elevation in  a  channel feeding the  treatment plant. Thus, a
constant flow could be maintained through the plant. General-
ly, however, existing inlet pump controls function well and no
special need exists for more sophisticated control. Later in this
paper  the  potential  for  flow equalization in existing CSD
plants will be discussed and, in  that instance, more compli-
cated controls would be warranted.

Process Air Controls for Activated Sludge Plants
   Basically the quantity of oxygen (air) required to satisfac-
torily operate an activated  sludge plant will vary  depending
upon the organic load  in the primary effluent, the cell resi-
dence time at which the plant is operated, and whether nitrifi-
cation does or does not occur. Oxygen requirements to stabi-
lize  a wastewater can be determined within reasonably narrow
limits depending primarily upon influent COD, NH3-N concen-
trations  and mean cell residence time.  However, air quantities
required to satisfy oxygen demand are more variable depend-
ing upon the type of oxygen transfer equipment used and the
efficiencies at which they operate. In all but completely  mixed
systems, rates  of oxygen consumption will vary along the
length of the aeration tank. Therefore, process air control is
critical to the stable operation  of the activated sludge system.
Not  only must total air quantities be controlled, but the rate
at which it is supplied along the aeration tank must be varied
to provide for stable and efficient treatment.
   Process air control systems used at  CSD activated  sludge
plants utilize either a  control signal to throttle a centrifugal
compressor  or preset timers to control the number of on-line
positive  displacement  compressors.  In  the  former case, dis-
solved oxygen probes, cam programmers, and plant flow rate
are currently used to provide the command signal. The control
method used at various CSD activated sludge plants is  shown
in Table 2.
   None  of the  above  control systems have operated without
some disadvantages, discussion of which is too lengthy to be
included in this short paper. From a theoretical standpoint
control by use of D.O. should be the most desirable since such
a system can respond to  any diurnal fluctuation in oxygen de-
mand as well as shock loads, whether they are caused by flow
rate  or waste strength.  Until recently, a reliable D.O. probe has
not been available which was capable of holding calibration to
the extent that air compressor programming could be depend-
ably set by  the sensor. Recent developments in this field now
make D.O. sensing a viable control scheme.
   Experience with D.O. probe  control systems has revealed
several operational characteristics which must be considered in
any  application. D.O. levels below 1  or 2 mg/1 will affect nitri-
fication kinetics by imposing a rate limitation. Based upon ex-
perience in Southern California, if the mean cell residence time
is above 5 to 6  days (dependent upon temperature) and the
D.O. set point is below about 0.5  mg/1, nitrification may not
occur because of the D.O. rate limitation. Conversely, if the
D.O. set point is increased much above 0.5 mg/1 the rate limit-
ation is removed and nitrification will proceed. This will pro-
duce a sharp increase in  oxygen demand and process air flow
will increase dramatically to try and match the demand. If ni-
trification is not desired, the best operational  procedure is to
keep the  MCRT below about 4 days. However, this may not
produce  the best quality effluent in terms of suspended solids.
   In the D.O. probe system currently used at the CSD plants,
only one D.O. probe can be  selected at any time to control
process air flow. The D.O. probe system does not regulate the
distribution of air  flow along the aeration system, but only
regulates total air flow to the  aeration tank in order to main-
tain  a dissolved oxygen  concentration at the site of the con-
trolling probe.
   Another  operational consideration with regard to the D.O.
probe control system is the added maintenance required, com-
pared with other control techniques in use at the CSD plants.
Probes in current use require weekly routine maintenance, ad-
justments and calibration by instrument repair personnel and
daily comparative checking by the plant operator. If reliable
response is  obtained,  this level of  maintenance is not con-
sidered to be excessive.
   If the full potential of the  D.O. probe system is to  be rea-
lized, total air flow and distribution of air throughout the aera-
                                                          86

-------
                                                                            COMPUTER APPLICATIONS IN AUTOMATION
                                                      TABLE 2

                                   ACTIVATED SLUDGE PROCESS AIR CONTROL
     Treatment Plant

     Whittier Narrows WRP
     San Jose Creek WRP
     Pomona WRP
     Los Coyotes WRP
     Long Beach WRP
     Dist. 26 WRP
     Dist. 32 WRP
 Average
Flow MOD

   12
   31
    8.5
    8.5
    9.1
    3.3
    1.3
  Air to
Flow Rate
    X
    X
D.O.
Probe
                     X
                     X
   Cam
Programmer
Timers
                                                           X
                                                           X
tion system must be controlled while still preventing closure of
the inlet guide vanes beyond the point that would cause com-
pressor surge. This may well represent a potential application
for a mini-computer.

Waste Activated Sludge Control
   Control  of mean cell residence  time  requires  wasting of a
predetermined mass of activated sludge each day. Two alter-
nate  control schemes haye been proposed to accomplish solids
wasting, either wasting directly from the mixed liquor or wast-
ing from the return sludge line coming from the secondary sedi-
mentation tanks.  The latter method has been in dependable
use at CSD plants. Propeller  meters are used for waste  acti-
vated sludge measurement. The  propeller meter has proven to
be cheaper than other types of  flow measurement devices and
is relatively maintenance-free, provided that effective primary
sedimentation tanks precede the secondary system and that no
solids recycle from aerobic or anaerobic digestion  is discharged
into  the  secondary system.  Dependable 24-hour composite
sampling is also mandatory, but such sampling should be  pro-
vided regardless of the wasting system.
   In the CSD design, a pulse signal generated by  the propeller
meter is sent to a pulse-to-current converter. Converter output
is a 4 to  20 ma signal, proportional to flow rate.  This signal is
compared with a set point and used to control the position of
a motorized throttling valve. The signal  also actuates  a re-
corder, flow totalizer and a  no-flow alarm. Thus, if something
                             should accidentally stop the propeller, an alarm is immediately
                             sounded, alerting the operator.
                                The technique of wasting a constant flow rate from the re-
                             turn sludge line  as set by the operator each day has been used
                             in LACSD installations for many years.  The control system
                             functions with a minimum of maintenance. However, the main
                             test  of any sludge wasting technique is whether it  actually
                             wastes the  correct  mass of cells and can  maintain  desired
                             MCRT values. Years of operational experience with  daily in-
                             ventories of treatment plant solids indicates that the wasting
                             technique   and  associated   control  equipment  function
                             properly.

                             Primary Sludge Pump Control
                                When primary sludge from he  Districts' inland renovation
                             plants is wasted to the existing trunk sewer, precise control is
                             unnecessary. The only critical observation is to assure that no
                             sludge blanket  accumulates in the tank. Conversely,  in the
                             downstream primary plant (JWPCP) where solids must be pro-
                             cessed, the maximum attainable sludge concentration is desir-
                             able to minimize the volume of sludge to be treated.
                                The Districts utilize radioactive density meters to  monitor
                             and control raw sludge pumping, followed by a control system
                             which distributes incremental amounts of sludge to each di-
                             gester until a preset  maximum amount is fed to each operating
                              cell. Seed sludge can be circulated to any digester if required,
                             but routinely, a completely mixed cell does not require seed
                                                         87

-------
 AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
 sludge. All controls are automatic with an operator decision re-
 quired at  the end of each 24-hour period as to the quantity to
 be fed to each of the 31 active digesters currently in operation.
 Loading decision is  based upon volatile acid concentration,
 pH, gas production  and previous loading history with some
 minor consideration  given to alkalinity and gas composition.
 Sludge concentration and quantity of flow are critical mea-
 surements. Flow measurement is made by two Venturi-type in-
 struments in series, one of which is redundant and used as a
 check on meter calibration.
    Further automation of this process awaits instrumentation
 capable  of  continuous measurement of the critical control
 parameters.

 Chlorination Control
    Control systems in current use at LACSD secondary plants
 provide  for rate control of chlorine  feed, paced by plant flow
 and fine-tuned by a chlorine residual feedback. Either free or
 combined residual  chlorine can be used as the feedback signal.
 Liquid chlorine is stored on site and fed through evaporators with
 the resultant solution discharged through mixers to the final ef-
 fluent. These systems are dependable and easily maintained.
    At the  primary plant (JWPCP), where chlorination is not
 routinely required  for deep ocean discharge, liquid chlorine is
 mixed with  slaked lime, Ca(OH)2, to produce a hypochlorite
 solution. When  chlorination is required, dosage of about 25 to
 30 ppm into an average flow of 350 MGD with peaks up to
 550 MGD requires massive amounts of chlorine. The hypo-
 chlorite  solution is fed into the suction side of effluent pumps
 to  provide  rapid  mixing in the  pump  and  downstream
 manifold.
   Control by chlorine residual is not particularly effective at
 present because of the  return  of  centrate from the sludge pro-
 cessing plant. Following completion of current construction of
 additional  solids processing equipment, the centrate will be re-
 moved from the plant effluent. This, together with polymer
 addition to the primaries, will result in an improved effluent
 quality. Control will then be more reliable.

 Return Shidge Pump Controls
   Control of sludge return rates from  the final clarifiers is
 normally only critical when a  treatment plant approaches de-
 sign capacity. Under these conditions, excessively high return
 rates can reduce mixed-liquor aeration time below desirable
 levels  and  decrease detention  time in the final sedimentation
 tanks. Also, during periods of sludge bulking, return rate must
 be adjustable to prevent solids accumulation in the final clarifi-
 ers with consequent overflow into the plant effluent.
   The most effective  design utilized  by the LACSD is a
 blanket detector unit which controls either direct pumping
 rate from the sludge hopper or the modulation of a valve con-
 trolling withdrawal  to a sludge sump. Relatively simple con-
 trols can then be used to vary  pumping rate from the sump to
maintain a relatively constant  water surface elevation in the
sump.
    A blanket detector control system has the operating advant-
 age of increasing return rate as mixed liquor flow increases and
 decreasing the return rate as mixed liquor and plant  flow de-
 crease.
    Manual override of a return sludge system can also be useful
 if return rate is to be reduced during periods of low plant flow
 to allow maximum compaction of sludge for wasting purposes.
 This mode has not been utilized at LACSD plants,  however.
 since  return  of activated sludge to the aeration system as
 promptly as possible has been a prime goal.

 EXISTING USAGE OF COMPUTERS AT CSD PLANTS
    Computer usage by  the Districts can be divided  into two
 categories. The first is  the  classic  data processing function,
 which includes normal accounting activities, scientific process-
 ing and  data base systems. The second category of computer
 usage by the Districts is the  use of computers to respond in a
 predefined manner to "real time" signals from remote sensing
 devices.  These computers are commonly called process control
 computers because, in many cases,  they actually control  the
 processing function in a manufacturing plant.
    The classical data processing activities are carried out on an
 IBM System 370  Model 125 located at the Joint Administra-
 tion Office's computer facilities. The IBM 370/125  is a gen-
 eral-purpose business and scientific computer. It  has multi-
 processing capabilities for the running of five jobs  concur-
 rently and for spooling all input and output for unit record de-
 vices to  direct-access disk storage. The  370/125 has the  capa-
 bility, although it is not implemented, for communication and
 status inquiry, and control of a MODCOMP mini-computer at
 the Districts' primary plant.
   There are five major areas of data processing activity at  the
 computer facilities in the Joint Administration Office.
   The first area of activity is batch  scientific data processing.
 To meet the needs of the Districts'  engineering departments,
 facilities  for  the batch submission and  execution of scientific
 programs have been provided. The programs written either in
 the  FORTRAN or PL/1  languages are submitted on punched
 cards at  the data center in  the Joint  Administrative Office.
 Typically, applications consist of problems such as the  solu-
 tion of  traverses,  large-scale  statistical analyses, and simula-
 tions for  planning purposes.
   A second area of activity is the processing of data from the
 batch accounting and personnel departments. Information sub-
 mitted by the Accounting Department, Personnel Department,
 District landfill sites, etc., is keypunched and submitted at the
 data center in the Joint Administration Office. Applications
typically  consist of projects such as the Refuse Disposal Billing
System which handles billing for a system handling in excess
of 15,000 tons per day of solid waste.
   In addition to the batch processing, there are three areas of
tele-processing activity at the data center.
   The Districts' major  teleprocessing  project is the Water
Renovation Plant Data Base system. The seven larger  second-
                                                        88

-------
                                                                              COMPUTER APPLICATIONS IN AUTOMATION
ary sewage treatment facilities located throughout Los Angeles
County dial into the data center once daily using teletypes.
Each plant enters about  100 operational parameters into the
370/125. A  series of interactive programs, using the para-
meters supplied by the  operator, then sends  a detailed re-
sponse to the plant indicating parameters to be used for the
next  day's plant operation. This series of programs also pro-
duces daily information  as to whether the plant is exceeding
any of the standards of the Water Quality Control Board. If
the plant exceeds these standards, the plant is informed by the
program so that immediate action may be taken. Additionally,
the operational status (total flows, etc.) and  Water Quality
Control Board compliance information for all of the plants can
be retrieved from terminals at the Joint Administration Office.
In completing water quality compliance and operational para-
meters, the interactive programs use plant  readings for a 30-
day period which are stored  by on-line direct-access disk de-
vices. Once monthly, these stored records are used as input for
batch programs which compute a Monthly Operational Status
Report showing long-term trends. A typical report is shown in
Appendix 1.
   A second teleprocessing activity  at  the Districts is the In-
dustrial Waste Project. On-line data inquiry and  updating of in-
dustrial waste records for industrial waste surcharge processing
is done by a  transaction-oriented data communication/data
base  management system (CIVS/VS). The 90 million character
data  base consists of sewage flow data, surcharge information,
county  assessment figures, and  internal  accounting informa-
tion  for 60.000 industrial properties in Los Angeles County.
Companies with significant waste discharges are charged based
on discharge characteristics and amounts in compliance with
state  and  federal  law. CRT terminals used in  this data base/
data  communication system are located in the Industrial Waste
Department and the data center at the  JAO.
   The final major area of activity under development at the
JAO data center is the On-Line Personnel System.  On-line in-
quiry and updating  of  personnel information is  done from
CRT's at  the  JAO Personnel Department.  Hard-copy audit
trails of these  transactions are made concurrently at the data
center.
   The second category of computer usage, the  use of comput-
ers to respond  to signals from remote sensing  devices,  is
limited to an alarm monitoring and logging systemat the Joint
Water  Pollution Control  Plant.  The Districts have not yet
found it feasible  to install process  control computers  in sec-
ondary  treatment plants. However,  the Water  Renovation
Plant Data  Management System,  as previously  described,
should allow the Districts to have  the process control func-
tions of a secondary treatment plant well defined before instal-
lation of a computerized system.
   The Districts have installed a process control computer,  a
MODCOMP  11/25, at the Joint Water Pollution Control Plant
(JWPCP). The existing system is primarily an alarm  monitoring
and  display system. It consists of a central alarm panel and  a
data logging system for both remote pumping plant  alarm
status and inplant (JWPCP) alarm status.
   There are 34 remote pumping plants located at various sites
around the Los Angeles County  area. All but  one  plant are
connected by  leased  telephone  lines to the  JWPCP  alarm
facility. The most remote pumping plant is periodically con-
nected to the alarm facility by use of an automatic dialer. The
conditions presently being monitored at the pumping plant are
(1) high water level in wet well, (2)  high water  level in dry
well, (3)  power failure, (4) pump control failures, (5)  com-
munication failure, (6) A/C power failure to  telemetry device.
   The  inplant  alarm  system can be  broken down into two
types of subsystems. The first is  the Methane Detection Sub-
system, which continuously monitors  the level of methane gas
at 6 points within the enclosed pipe galleries. The meters are
directly wired to display devices on the central alarm  panel.
An alarm is activated if a certain level is exceeded.
   The second subsystem is the in-plant telemetry system and is
responsible for  the majority of the alarm data transmission.
There  are 654  contact-closure  type alarm points scattered
throughout the JWPCP. All of these alarms are assigned to and
wired into an existing local annunciator panel  at some local-
ized area,  i.e., Chlorination Station,  Sludge Dewatering Sta-
tion, etc.  There are some 14 local annunciator panels through-
out the JWPCP.
   The  Central Alarm Facility can be best described by tracing
an  alarm  as it  travels  through the Facility.  Suppose, for ex-
ample,  a  contact  closure occurs at the chlorination facility.
The signal is transmitted to the local annunciator panel where
a window  is permanently devoted to displaying the  condition
of its alarm point. A hardwired system (in reality a combina-
tion of hardware and a read-only memory computer) polls for
alarms. This polling system passes the alarm to the MODCOMP
computer which in turn lights the proper display  window and
logs the alarm on a printer. If the MODCOMP computer is in-
operative,  the window will  still be lighted by  the hardwired
system, but the individual alarm will not be  logged. The alarm
windows at the  Central Alarm Facility respond to  the local an-
nunciator  panels  within the  JWPCP and not to  individual
alarms.
   The  MODCOMP system also has provisions for the employ-
ment of analog monitoring and control functions, depending
on future requirements.

POTENTIAL  APPLICATION  FOR  INSTRUMENTATION
AND COMPUTERS
   The  full potential of the mini-computer at  the JWPCP has
not been  explored since the initial decision to purchase the
unit was  based upon  the  unit  being less expensive than an
equivalent  electronic  system  using  conventional relays and
control wiring, for primary usage  as a central alarm station. In-
creased familiarity with the unit  should result in  reduced fre-
quency of attendance  by pumping plant operators to each in-
stallation  and the  consolidation of surveillance duties by plant
                                                          89

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
operators. Improved procedures should be possible for control
of the chlorination  system, inlet and effluent pumping, sludge
transfer  pumps and the  control  of the solids  processing
centrifuges.
   The IBM 370/125 was initially installed at the Joint Admin-
istration Office in the first week in August 1974 to replace a
Univac 9300 plus a number of  time-sharing terminals in the
Engineering Department. As promptly as advisable, all separate
functions  will  be  transferred  to  the  new unit. The  IBM
370/125 is capable  of storing and manipulating vast quantities
of data  which  would be impossible to  collate without the
machine.
   As newer applications are tested and proven, CSD wiD have
the hardware and personnel needed to  make the necessary
changes. The industrial  waste surcharge ordinance will vastly
increase the work involved in computing, billing, and recording
collection of the surcharges from industries using the system.
Coincidental to this effort to charge industries individually for
services rendered,  data  logging of information from a much
more exacting  source control program  must be initiated, to
provide information on potential sources of toxic discharges
into the system. Both the Federal Government and the  State
of California have  instituted limitations on the quantity of
toxic  waste material discharged into the sewerage system.
Sludge discharges of industrial waste which upset the treat-
ment processes  must be controlled or the Districts will be sub-
ject to fines for non-compliance.
   With  the  advent of unionization  of District  employees,
ready access must be provided for personnel records. Union
negotiations on wages and fringe benefits result in extremely
complicated settlements which make it more difficult  to arrive
at valid comparisons of prevailing wage levels. Computer data
logging for quick information retrieval is urgently needed to
evaluate  the status  of over 230 class  specifications for the
multiplicity of District jobs.
   A desk-top study recently completed indicates that serious
evaluation should be given to provision of flow equalization at
the Districts' secondary treatment plants. Certainly, every indi-
cation points towards more complicated, expensive treatment
requirements with far less tolerance of deviation from normal
conditions. The  question then must be posed as to whether the
cost of storage  and pumping of peak flows is less expensive
than the provision of adequate process equipment and tankage
to handle normal peak flows. Tertiary treatment processes are
almost as expensive  as secondary  treatment in many cases, and
if consideration to storage of secondary effluent shows it to be
cost-effective, it would be even more economical to store pri-
mary effluent to provide for constant hydraulic flow through
the secondary and tertiary process.
   The engineering problems involved in storing primary efflu-
ent are not  difficult to resolve and very possibly computer
control of storage  capacity versus inflow could assure maxi-
mum utilization of facilities while providing constant hydrau-
lic loading and, to some extent, equalization of organic load.
   Primary sedimentation prior to storage could preclude  any
solids handling problem. Covered storage with air under the
covers going to aeration blowers could resolve any odor prob-
lems. Modest mixing and aeration could prevent septicity from
increasing air requirements. Computer control  of gates  and
pumping rates could minimize operator time demands and re-
sult in very effective use of facilities.
   A saving in capital cost of 10% to 12% appears to be a con-
servative estimate of cost benefit without regard to rather ob-
vious operator advantages in elimination of peak flow through
CSD treatment  plants. The time  has probably arrived for im-
plementation of this  idea. District experience at the Whittier
Narrows  WRP  with  a reasonably  constant hydraulic flow
through the  plant has already indicated that a plant originally
designed for 10 MGD can handle 12 MGD even while fully ni-
trifying. Without nitrification the same plant can handle 14 to
 18 MGD,  depending on sewage temperatures.
CONCLUSION
   Increased automation and application of computer tech-
nology in  District  treatment plants will be implemented by
District management whenever an improvement in  operating
efficiency can be demonstrated.
   Application of  computer  technology  to  management of
new and complex industrial  waste regulations is necessary to
implement source control of  toxic industrial wastes and to ad-
minister the Districts' new industrial waste surcharge.
   Exacting requirements for improved effluent quality and
the high cost of capital improvements now makes considera-
tion of flow equalization in secondary plants a viable alternate
to standard designs which must provide adequate capacity for
peak discharges. Storage  of primary effluent to equalize hy-
draulic loading on secondary and tertiary processes could be
controlled by a mini-computer so that effective use of storage
facilities is assured.
                                                         90

-------
APPENDIX 1. MONTHLY OPERATIONAL STATUS REPORT.
                                        WHITTIER NARROWS WATER RECLAMATION PLANT
                                               COUNTY SANITATION DISTRICTS
                                              OF LOS ANGELES COUNTY, CALIF.
                                  JOHN D. PARKHURST - CHIEF ENGINEER 6 -GENERAL MANAGrR
                                              *** SUMMARY OF OPERATIONS ***
                                                        JUNE 1974
DATE







I
i
J
H
3
o
7
0
V
10
11
U
13
It
1,-j
16
17
Id
Iv
20
21
22
2j
2*
23
2b
27
26
2->
30
MLn.'J
PLANT r.HARCTERISTICS
FLOW

EFFLUENT


MGD
1
11.98
12.02
12. Ob
11.61
11.81
12.40
12.05
11.83
12.07
11.95
11.79
11.73
11.63
11.99
12. CO
11.99
11. bt
12.27
12.15
11.70
11.74
11. b4
11.63
11.43
10.41
12.02
12.22
12.03
12.CE
11.39
11.66
TOTAL
RETURN
ACT
SLUDGE
MGD
2
6.08
6.03
6.05
6.10
5.96
6.00
6.08
5.94
5.85
5.87
5.97
5.93
5.91
5.83
5.70
5.73
5.«3
5.86
5.71
5.55
5.55
5.40
5.36
5.44
5.76
5.51
5.46
5.41
5.53
5.33
5.76

WASTE
ACT
SLUDGE
i-iGD
3
.171
.185
.162
.190
.171
.161
.144
.139
.133
.111
.122
.111
.132
.149
.145
.144
.152
.168
.189
.1V3
.167
.172
.172
.151
.150
.14?
.143
.143
.143
.138
.153

AIR


VCF/DAY
4
29.9
29.9
29.8
29.8
29.1
29.0
29.1
29.1
29.2
29.2
29.2
2 B. 6
2H.4
28.7
28.8
29.1
2<;.i
29.4
2V. 4
2e.6
30.2
29.9
29.6.
29. e
29.5
29.7
29.1
29.5
29.5
29.5
29. ?<•
SUSPENDED SOLIDS

RAW
SEWAGE

MG/L
5
408
362

414
364
274
434
418
302
366
414
472
428
M4
396
380
412
394
5':>6
453
476
400
394
434
424
450
466
302
5f-3
403
414.2

PRIMARY
EFFLUENT

MG/L
6
98
74

106
118
102
104
7B
79
106
106
90
86
94
86
88
130
116
148
222
14
96
78
98
96
96
114
114
96
f!2
102.9

RETURN
SLUDGE

MG/L
7
6396
6161

5323
5267
5627
5582
5523
5240
5277
5664
5883
5714
5901
6020
59feO
6126
6434
6798
7372
6871
6690
6355
4691
6579
6218
6383
6322
6206
6236
6036

SEC.
EFFLUENT

MG/L
8
4
4
5
4
7
7
7-
4
5
7
4
6
6
5
5
5
8
5
7
6
6
4
4
5
4
4
4
6
6
4
5.2
; .
CHLOR'Nr
CONTACT
EFFLUENT
MG/L
9































COD

RAW
SEWAGr
TOTAL
MG/L
10

473

716

442


553

646

556


642

568

973


580



549


699
616

PRIMARY
EFFLUENT
TOTAL
MG/L
: i

21'.-

269

239


230

254

224


725

351

414


21-r



234


201
255.8
SECONDARY
EFFLUENT


TOTAL
MG/L
12

22

27

41


29

27

31


26

26

30


27



28


23
28.1
SOL.
MG/L
13

20

23

31


22

25

22


20

23

25


22



24


17
22.8

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
         APPENDIX 1. MONTHLY OPERATIONAL STATUS REPORT-(CONTINUED)
                         WHITTIER NARROWS WATER RECLAMATION PLANT
                                COUNTY SANITATION DISTRICTS
                               OF LOS ANGELES COUNTY* CALIF.
                    JOHN 0. PARKHURST - CHIEF ENGINEER 6 GENERAL MANAGrR
                               *** SUMMARY OF OPERATIONS ***
                                         JUNE 1974
DATE






1
2
3
A
5
6
7
6
9
10
11
12
13
1-
15
16
17
18
19
20
21
22
23
24
25
26
•27
26
29
30
MEAN
PLANT CHARCTERISTICS
NITROGEN
PRIVARY
EFFLUENT
CRG.
• MG/L
14































NH3
P.G/L .
15
20
20
19
22
21
21
22
22
20
21
21
22
22
21
19
21
22
22
22
25
18
20
19
20
20
20
20
20
20
20
20.7
SECONDARY
EFFLUENT
NH3
MG/L
16
.1
.2
.2
1.2
.2
. 3
1.1
1.0
.2
.3
.4
.3
.6
.4
.6
.2
1.3
* A
1.0
2.4
.6
.2
.1
. 1
. 1
.3
.6
.4
.'1
. 2
. 5
N02
MG/L
•7
.02
.Cl
.01
.04
.02
.03
.06
.02
.02
.02
.04
.04
.04
.05
.04
.04
.09
.04
.07
.06
.04
.02
.02
.02
.03
.03
.03
.03
.02
.02
.03
N03
KG/L
18
18.00
13.50
21.00
19.00
17.50
20.00
20.00
18.5O
19.00
18.00
17.00
20.00
19.00
17.00
16.50
16.50
15.50
10.50
9.50
e.50
13.00
12.00
17.50
16.00
16.00
15.00
12.50
15.00
15.50
15.30
16.24
TOS
ELEC.
CCND. '•
jui » /*onwu^
Hi V.KU" rl
/CK
19
































M/" /\
• nvj/ L

20
683
58'

650
645
593
606
627
577
605
627
567
594
600
633
551
637
566
617
588
651
638
588
654
527
66*
666
703
618
562
6.4.9
                                            92

-------
     APPENDIX 1. MONTHLY OPERATIONAL STATUS REPORT-(CONTINUED)
                                             WHITTIER NARROWS WATER RECLAMATION PLANT
                                                    COUNTY SANITATION DISTRICTS
                                                   OF LOS ANGELES COUNTY, CALIF.
                                        JOHN 0. PARKHURST - CHIEF ENGINEER 6 GENERAL MANAGER
                                                   *** SUMMARY CF OPERATIONS ***
                                                                  1974
DATE







1
<:
3
4
3
6
7
6
y
lo
It
U .
ii
It
la
lb
17
U
it
2o
21
& £.
2s
2t
23
2o
27
2a
2*
3o
.••••trtN
PLANT CHARCTLRISTICS
EFFLUENT CHAFACTER I ST I CS

SECChI
01SC
DEPTH
FEET
21
9.0
9.0
9.5
9.5
9.0
8.0
6.0
6.0
7.0
7.5
7.0
6.5
7.5
6.0
6.5
7.0
6.5
6.0
7.5
7.5
7.5
7.5
9.0
9.5
9.0
10.0
9.C
e.o
7.0
9.C
7. a

EFFLUENT
Tt>.f».

FAREM.
22
73.0
73.0
74.0

74. C
74.0
74.0
74.0
75.0
74.0
74.0
74.0
74. C
74.0
75.0
74. &
74.0
74. u
74.0
75.0
74.0
74.0
75. C
75.0
75.0
75.0
76.0
77.0
77.0
75.0
74.5
CL2
RESIDUAL
EFFLUcNT
GSAu
KG/L
23


.90

.72
.18
.54


2.40
I.d5
1.90
1.56



0.
.45
.24
.42
.45


2.
-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
APPENDIX 1. MONTHLY OPERATIONAL STATUS REPORT-(CONTINUED)

                         WHITTIER NARROWS WATER RECLAMATION PLANT
                                COUNTY SANITATION DISTRICTS
                               OF LOS ANGELES  COUNTY, CALIF.
                    JOHN 0. PARKHURST - CHIEF ENGINEER 6 GENERAL MANAGER
                               *** SUMMARY OF  OPERATIONS ***
                                         JUNE  1974
            Date

           20, 21

             18


             All
DATE

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
13
19
?0
21
22
23
24
25
76
27
28
29
30
«rAN
MISCELLANEOUS


34

































35

































36

































37
C.
0.
0.
C.
0.
.01
0.
0.
0.
• 0.
0.
0.
0.
.02
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.

.001
    Column

     6, 15

 8, 12, 13, 16,
17, 18, 26, 37

      37
NOTES

            Remarks

 Primary effluent sampler malfunction

 Secondary effluent sampler malfunction
 Hexavalent chromium, secondary
 effluent, 24-hr composite (mg/L)
                                           94

-------
                                                            COMPUTER APPLICATIONS IN AUTOMATION
APPENDIX 1. MONTHLY OPERATIONAL STATUS REPORT-(CONTINUED)
                         WHITTIER NARROWS WATER REC1 AV.ATION PLANT
                                COUNTY SANITATION DISTRICTS
                               OF LOS ANGELES COUNTY, CALIF.
                   JO"N 0. PARKHURST - CHIEF ENGINEER 6 GENERAL MANAGER
                               *** SUMMARY OF OPERATIONS ***
                                         JUNE 1974
DATt







1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
23
29
30
MEAN
AERATION SYSTEM NO. 1
SUSPENDED SOLIDS

TANK
1

MG/L
38
2626
2530

2443
2296
2461
2360
2299
2021
2532
2163
2429
2401
2416
2313
2212
2397
2273
2606
2605
2R3?
2530
2C75
2225
2251
2C9b
23^2
2106
25C3
2278
2373

TANK
2

• MG/L
39
1943
1860

1B27
1647
1732
1737
1699
1648
1710
1603
1870
1871
1730
1694
1770
1834
2054
2182
1937
Ib92
1P37
1742
1748
169C
1"04
1844
1926
15)83
1780
ieie

TANK
3

MG/L .
40
1830
1788

1671
1664
1501
1659
1543
1560
1572
1500
1762
1912
1622
1570
1690
1734
1860
2027
1664
1912
1858
1668
1667
1738
1 7 1 :l
1712
1846
1827
1732
1711

TANK
^

Mo/L
41
































VOLATILE
SOLIDS

*
•V2
75.0
74.0
74.0
75.0
75.0
77.0
7^.o
76.0
7A.O
76.0
77.0
73.0
75.0
74.0
74.0
7^.0
77.0
77.0
76.0
77. C
78.0
77.0
75.0
77.0
77.0
78.0
77.0
77.0
77.0
76.0
7t=..8

RETURN
ACT.
SLUDGE

MGD
43
6.1
6.0
6.1
6.1
6.0
6.0
6.1
5.9.
5.9
5.9
6.0
5.9
5.9
5.8
5.7
5.7
5.9
5.9
5.7
5.6
5.6
5.5
5.4
5.4
5.8
5.5
5.5
5.4
5.5
5.3
5.76
M^XED
L I-^UOP
D.O.
MG/L
MAX
44































WIN
45































RETURN
SLUDGE
AERATION
VOLUME
MG
4(
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.550
.55J
.550
.550
.550
.550
.550
.550
.550
                                            95

-------
APPENDIX 1. MONTHLY OPERATIONAL STATUS REPORT-(CONTINUED)
|
                                     WHITTIER NARROWS WATER  RECLAMATION
                                            COUNTY SANITATION DISTRICTS
                                           OF LOS ANGELES  COUNTY. CALIF.
                               JOHN D. PARKHURST - CHIEF ENGINEER 6 GENERAL MANAGPR
                                           *** SUMMAKY OF  OPERATIONS ***
                                                     JUNE  1974
o
o
•n
S
DATt







1
2
3
4
5
b
7
6
9
10
11
12 .
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
26
29
30
?-;EAN
AERATION SYSTEM MO. 1
LOADING PATTERN

TANK
t Ml"i\
I

%
47
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
66
6b
66
65.7

TANK
2

ft
48
34
34
34
34
34
3
-------
    APPENDIX I. MONTHLY OPERATIONAL STATUS REPORT-(CONTINUED)
                                              WHITTIE* NARROWS WATER RECLAMATION  PLANT
                                                     COUNTY SANITATION DISTRICTS
                                                    OF LOS ANGCLES CCUNfYt  CALIF.
                                        JOHN D. PARKHURST - CHIEF ENGINEER  6  GENERAL  MANAGER
                                                    *** SUMMARY CF OPERATIONS ***
                                                              JUNE 1974
OATt








1
2
3
4
5
fa
7
6
9
10
11 '
12
13
1*
15
16
17
13
19
20
21
22
23
24
25
26
27
28
29
30
tiEArJ
KINETIC PARAMETEkS
AIK PATFS


CUdIC
FEET/
CjMLLCr.
EFFLUENT

9(3
2.50
2.49
2.47
2.57
2.4o
2.34
2.41
2.46
2.4,;
2.44
2.4o
2.44
2.40
2.40
2.4Q
2.43
2.4o
2.39
2.42
2.46
2.57
2.53
2.56
2.CJ
2.6 c. i* i w v c. LV
99

1537

1251

1349


1394

1296

1423


1419

675

758


1600



1359


1689
1329
COD



PENOVAL


«
100

90.7

91.-+

87.0


• 90. 4

90.2

90.2


91.1

93.4

94.0


89.7



89.7


91,5
90. a

AERATION
SYSTEM
LOAD
(LBS.)
101

21453

26047

24716


23153

24975

22100


22499

35918

40397


20757.



23848


19094
25413
RETURN SLUDGE
AERATION TIMES

AERATION
SYSTEM
1
(H,%S.)
102
2.17
2.19
2.18
2.16
2.21
2.20
2.17
2.22
2.26
2.25
2.21
2.23
2.23
2.26
2.32
2.30
2.24
2.25
2.31
2.38
2.33
2.40
2.46
2.43
2.2S
2.40
2.42
2.44
2.39
2.48
2.29
AERATION
SYSTEM
2
(HRS. )
103































AERATION
SYSTEM
3
(MRS.)
104































MIXED LIQUOR
AERATION TI'-'ES

AEPATION
SYSTEM
r
(MRS.) •
103
3.26
3.26
3.24
3.32
3.31
3.20
3.24
3.31
3.28
3.30
3.31
3.33
3.31
3.30
3.32
3.32
3.31
3.24
3.29
3.41
3.40
3.39
3.46
3.48
3.63
3.35
3.33
3.37
3.34
3.52
3.34
AERATION
SYSTEM
2
(HRS. )
106































AERATION
SYSTEM
3
(HRS.)
107































TOTAL AERATION
SOLIDS

AERATION
SYSTEM
1
(LBS.)
103
53384
•31691

49548
48430
47488
48005
46254
43610
484K9
43V18
50549
51575
48122
4651:
47304
49748
51600
56837
53426
55344
51916
45912
470T8
47363
46ti71
49023
49023
51816
48289
49279
AERATION
SYSTEM
2
(LBS.)
109































AERATION
SYSTEM
3
(LBS.)
110































co

-------
APPENDIX 1. MONTHLY OPERATIONAL STATUS REPORT-(CONTINUED)
                                       H'HITTIER NARROWS  WATER  RECLAMATION PLANT
                                              COUNTY SANITATION  DISTRICTS
                                             OF LOS ANGELES  COUNTY, CALIF.
                                 JOHN 0. PARKHURST - CHIEF ENGINEER 6 GENERAL MANAGER
                                             *** SUMMARY OF  OPERATIONS ***
                                                      JUNE  1974
5
5
o
DATE








1
2
3
4
5
6
7
0
9
10
11
12
13
1*
lf>
16
17
IB
1*
20
£1
22
23
24
25
26
27
23
29
30
MEAN
KINETIC PARAMETERS
MIXED LICUOR
SL'SPCMGFtj SOLIDS
wwJ~ta!\U/t,L/ J wl» 1 I/ w
H *• • r
AERATION
SYSTEM
1
LBS.
*»•• •» «
111
41447
4C199

38451
33001
36309
37285
35«10
34430
36V87 •
34C93
3VS15
40669
37138
36005
37257
3efcfcO
4l01
AERATION
SYSTEM
2
LIS.
112































AERATION
SYSTEM
3
LBS.
113































COP LOADING

L?S.COD/
LtiS.
TPVSf.X
DAY

114

.42

.53

.52


.54

.56

.42


.44

.?8

.76


.45



.43


.39
.519
LES.COD/
LBS-
MLVS5/
DAY

115

.72

.90

.88


.91

.95

.72


.62

1.13

1.29


,76



.80


.66
.88
DAILY
ru i i
V.C.LL
Rt:S.
TH'.E


(DAYS)
1J.6
7.4
6.9

7.4
7.8
7.5
8.6
8.9
9.2
11.3
9.4
11.0
10.2
Q.I
7.0
8.2
7.7
7.2
6.6
5.5
7.2
6.8
6.5
9.6
7.4
8.1
e.i
8.2
8.6
r. .5
e.i
SOLIDS BALAMCE

TOTAL
PtANT
SUSPEND.
SOLIDS
(IPS.)
117
70509
66422

65184
64C01
61534
63529
60739
58208
63199
57955
67037
69466
63300
61203
63119
65974
69004
75805
68997
73^36
69303
61707
62637
63626
62947
65043
66296
68V13
64496
65358
WASTTD
SOL ITS


(LBS.)
118
91?2
9506

6435
7511
7556
6703
6403
5612
4(3^5
5763
5446
62°0
73?3
72PO
71P2
7766
9015
107)5
116C-.6
9570
9SP4
9116
5908
8230
7364
7612
75^-0
74C1
7177
7738
SE".
EFFL"ENT
SUSPEND.
SOLTOS

-------
                                                                          COMPUTER APPLICATIONS IN AUTOMATION
                  COMPUTER APPLICATIONS IN  AUTOMATION

                                               William E. Dobbins
                 President, Teetor-Dobbins, P.C., 515 Johnson Avenue, Bohemia, NY 11716
INTRODUCTION
  The computer, which has been widely utilized and accepted
in power plants and industrial process control, is now being ap-
plied to wastewater treatment. Most of the initial applications
have been for data logging systems, in which the plant variables
are monitored and the data manipulated so as to present it in
forms and units of greatest use to the operator. The data can
also be stored on magnetic tapes from which it  can be re-
trieved for future research. This is a great improvement  over
the  old systems in which limited historical data  can be re-
trieved, but only by being laboriously dug out of numerous re-
corder charts.
  While data  logging and processing constitutes an important
step forward, the most important advance will be the applica-
tion of the computer to plant  operation.  The computerized
system has an important advantage over the system of numer-
ous independent analog control loops; control actions can be
based on algorithms which can  utilize any or all of the varia-
bles. Thus, it can utilize calculated variables in arriving at deci-
sions regarding process adjustments. The control capability of
the computer goes far  beyond that of the human operator,
who, surrounded by dozens of meter indicators and recorders,
cannot possibly react in  time to prevent  process upsets. The al-
gorithms can  be  changed and tuned as fundamental  process
knowledge and knowledge of the peculiarities of the particular
plant become available.
   Two examples of computer algorithms for  plant operation
are  presented  in this paper. Both were developed at the two
highly automated activated sludge plants in Bridgeport, Con-
necticut. One describes a program for the operation of the
main raw sewage pumps and the other the results of a program
which attempts to measure the respiration rates in the aeration
tanks and the BOD of the sewage delivered to them.

PUMP CONTROL PROGRAM
   At Bridgeport, two existing primary  treatment plants  were
modernized and expanded by the provision of secondary treat-
ment by the step-aeration version of the activated sludge pro-
cess. At each plant the sewage is delivered through a large in-
terceptor with its invert about 20 feet  below the ground sur-
face. It flows through a bar screen and a trapezoidal-shaped
grit chamber to a wetwell, from which it is pumped to the pri-
mary treatment tanks. The pumping is done by three constant-
speed and two 3-step-speed pumps. Both collection  systems
are combined sanitary and storm systems.  At the West Side
Plant the flow varies from about  12 mgd to 60 mgd;the East
Side Plant flows are generally about one-half as much. Except
for  the magnitudes of the flows, the following discussion ap-
plies to both plants.
  No control section was provided for the grit chamber, with
the   result that the depth of flow is determined by the level in
the  wetwell. In the  old plant, the pump selections were acti-
vated by  a float-operated  control system by which the pump-
ing  rate at any given wetwell level was selected so as to provide
an acceptable grit chamber velocity. However, a great deal of
hunting took place, particularly for some of the individual
pumps. Figure 1 shows the operating record in a typical day of
one of the pumps. In such a control system, the combination
of pumps running at any  particular wetwell level is always the
same, and the combination is changed when the level changes
by some adjustable increment of about one foot. The basic rea-
son for the  excessive hunting is that the system cannot take
the  time  frame into  account. At some times it may require an
hour for  the incremental  level change to occur; at other times
it may  take  only five minutes. The computer program resolves
this by continuously calculating the inflow rate and selecting
pump combinations to pace this calculated rate. The algorithm
is very simple:
  Inflow rate = Outflow rate + Rate of change of storage
The computer calculates this rate every  two minutes, from its
knowledge of the present pumping rate and  the relationship
between  wetwell storage  and water level. From the tables of
flow rates delivered  by various  pump combinations, which are
stored  in its memory, it selects the combinations which most
closely produce the calculated inflow rates. Of course, it is not
quite as  simple as this. The program also maintains, within a
designated band, the relationship between water level and inflow
rate  which  results  in an  acceptable  velocity  in the  grit
chamber. One of the benefits of this program is that it auto-
matically forces the inceptor to back up during storm flows,
and thus make use of the storage in the collection system to
moderate the inflow rates. That the program  produces a very
smooth flow pattern is demonstrated by Figure 2. This is a
typical chart for the Parshall flume meter which measures the
total pumping rate.

RESPIRATION RATE PROGRAM
   A knowledge of the strength of the sewage being fed to the
secondary system, and the rate of oxygen utilization in the
aeration  tanks, are essential to  the implementation of various
control strategies which are currently being studied by numer-
ous investigators. Much of the attention has been on the devel-
opment of on-line instruments for the measurement of TOC,
                                                        99

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
Figure 1. Pump Chart Showing Considerable Hunting
            (Single Pump)
                                                        100

-------
                                                                             COMPUTER APPLICATIONS IN AUTOMATION
                                                        NOON
Figure 2. Chart of ParshaH Flume Meter Showing Smooth Flow Pattern
                                                         101

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS


COD and other parameters which might correlate with BOD.
An attempt was made at the Bridgeport West Side Plant to de-
velop a computer program which would utilize the operating
condition of the aeration tank itself to evaluate the  sewage
strength and the current respiration rate.

   The installation consists of three aeration tanks, each equip-
ped with its own final settling tank, sludge return and sludge
wasting facilities. Thus, it is possible to operate each system in-
dependently. A diagram of the Aeration Tanks is shown in Fig-
ure 3. When the program was developed the mode of operation
was to return the sludge to pass No. 1 and to divide the sewage
flow equally between the  east and west ends. Thus, each of
passes 2 and 4 received about one-quarter of the total sewage
and pass 3  about one-half of the flow. The return rate was
about 50% of the average sewage flow.
                                                                 FROM   PRIMARY  TANKS
      TO  FINAL  TANK N° I
      TO  FINAL  TANK  N° 2
      TO  FINAL  TANK  N°3
                        Propeller
                        Flowmeter
 Figure 3. Layout of the Aeration Tanks
                                                    102

-------
                                                                            COMPUTER APPLICATIONS IN AUTOMATION
Program Logic
  The program logic is based on the equations expressing the
mass balance of oxygen in the aeration tank. It was actually
applied to the  three passes which received the sewage. The
basic equations are :

A = Rate in by flow  =  Qs • DOI + Qr • DOR         (1)
B = Rate in by Air   =  Ka Qa (DOS - DOT)          (2)
C = Rate out by flow =  (Qs + Qr) DOE                (3)
D = Rate of change within tank = ADOT • V / At         (4)
R = Rate of respiration within tank
In these equations:
Qs   =  Total flow of sewage,  MGD
Qr   =  Flow of return sludge,  MGD
Qa   =  Flow of air, cfm
DOI =  Dissolved oxygen in influent sewage ,     Ib/MG
DOE =  Dissolved oxygen in tank effluent,        Ib/MG
DOR =  Dissolved oxygen in return sludge.        Ib/MG
DOT =  Dissolved oxygen in tank (average),       Ib/MG
DOS =  Dissolved oxygen saturation value.        Ib/MG
V    =  Volume of the 3 passes, MG
At   =  Time increment, days
Kj   =  Coefficient relating to oxygenation efficiency
The mass balance relationship is:
R=A + B-C-D    Ibs/day                            (5)

   All of the factors in this equation, except R, DOS and Ka,
are  fixed quantities or measurable variables. DOS is computed
from the temperature of the water and  the known depth of
submergence of the  air diffusers. Therefore, by  determining
either Kg, or R, the other would be determined. The value of
K,, for the particular operating conditions could really only be
estimated and there would be great  uncertainty  about it.  R
could be subject to measurements on samples of the mixed liq-
uor. Once reliable values were established, the respiration rate
could be calculated periodically.
 Field Measurements
   During September, 1973. various attempts were made  to
 measure the respiration  rates  on samples of mixed liquor by
 taking  them to the laboratory, aerating them and measuring
 the DO values with a portable DO probe. As expected, the
 rates varied  considerably from place to place in the tank and
 from time to time at each place. It was decided that it would
 require a great deal of work to accurately  determine the pro-
 files of respiration values that would be necessary for the eval-
 uation  of Ka.  Therefore, a different approach was taken. The
 respiration rate for the  tank as a whole might be expressed as:
                                                           and effluent sewages, Re is the endogenous respiration and Kb
                                                           a  coefficient. This equation was proposed some years ago by
                                                           Eckenfelder and O'Connor (1). The detention time of the sew-
                                                           age in the three passes varied from 2 to 4 hours and averaged
                                                           about 3 hours. Therefore, the total weight of added BOD, con-
                                                           tributed by the sewage present in the tank at any one time,
                                                           could be  estimated as some factor times the weight added dur-
                                                           ing the previous interval plus a lesser factor times the amount
                                                           added during the second previous interval, etc. The use of this
                                                           equation  introduces two new factors, Kb and Re, to be deter-
                                                           mined along with K.,.
                                                              During a 24-hour period in July, 1973 an attempt was made
                                                           to calibrate the system by the use of equations (5) and (6). DO
                                                           probes were  placed  at  the  influent  sewage   distribution
                                                           chamber, at the exit end  of pass  4 and at  four other points
                                                           within the tank. Every hour, on the hour, a sample of the in-
                                                           fluent sewage  was  taken and three dilutions were set up for
                                                           subsequent   determinations.   Unfortunately,  no   effluent
                                                           samples were taken, because of the lack of sufficient BOD bot-
                                                           tles, as well as sufficient time. The computer was programmed
                                                           to log out  the values of the sewage flows, air flows and DO
                                                           readings every  12 minutes. Study of the data showed that a
                                                           good value for At was one hour. This  eliminated  most of the
                                                           noise in  the readings and showed smooth patterns of change.
                                                           Equations (5) and (6) were combined in the form
                                                              BODI   =
                                                                         A+B-C-D+E
                                                     (7)
R = Kfe (BODI - BODE) Qs +
                                                      (6)
 where  BODI and BODE  are the 5-day BOD's of the influent
in which the term E represented the magnitude of Kb . BODE
.  Qs - Re. When  the lab BOD data were turned out (5 days
later) all of the data were then available to fit to the equations.
The  values of  BODI and  Qs used in the fitting were the
weighted values over the previous  three hours and the other
values were the averages for the previous hour. This provided
21 sets of data which could be fitted to the equations, although
there were only 3 unknowns. The fit was made by the method
of least squares, with the following results:
   K,, = 0.0901. Kb  = 0.314. E = 55.0
Figure 4 shows the computed BOD values compared to the
measured ones, and Figure  5 shows the computed  respiration
rates. It is noted that the BOD values are the 3-hour composite
values, whereas the respiration rate is the average  for the previ-
ous hour. Although the effluent BOD's were not determined, the
plant effluent values were  generally around 25 mg/1 during
that month. Also, the average mixed-liquor volatile suspended
solids were about 800 mg/1. Assuming that these prevailed dur-
ing the test period,  and  from the value of E = 55 Ibs/day, the
following is derived:
   R = Oxygen used up  = 0.314 x BOD5 removed + 0.0357 x
lbsofMLVSS(lb/day)
Although they seem to be somewhat low, the  coefficients are
in the range of reported values.
                                                        103

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
   O)
   o.

   d>
   E
   in
   i
   o
   o
   CD
   V)
   O
   Q.


   O
   O
   ro
        120
100
           1600       2000

              7/12/73
                          2400       0400        0800

                                          7/13/73
1200
                             Time   Of  Day
                                                              Legend

                                                         -e	 computer  predicted

                                                          A     lab.  results
 Figure 4. Comparison of Computer Calculated BOD With Lab Measured BOD



                                          104

-------
                                                                            COMPUTER APPLICATIONS IN AUTOMATION
                  1600          2000
                        7/12/73
2400
                                            Time
          0400
    7/13/73
Of   Day
0800
1200
Figure 5. Calculated Respiration Rates
DISCUSSION
  The computer was programmed using the values previously
derived for the coefficients  and it printed out the computed
BOD values on the hourly log. It was done individually for
each of the three aeration tanks, to provide a  check on the
values.  They  all  fell within the general range  of expected
values. The values for Tanks 1 and 3 agreed quite well and
those for Tank 2  were consistently low, a condition which can
be attributed  to faulty meter readings.
  The values of  the  three coefficients are  influenced by the
temperature of the sewage and will, therefore,  be subject to
change. The temperature of the sewage at Bridgeport does not
usually vary  greatly within the day. but is subject to a wide
seasonal variation. A program was written whereby the coeffi-
cients can be  updated every three days. The previous equations
can be summed to represent the composite BOD for a 24-hour
period. The resulting  equations for this calculated composite
contain the three unknown coefficients. The actual composite
BOD for  the  same 24-hour  period is determined in the lab.
When  three sets  of data  are determined, the lab  results are
typed  into the computer and  the computer  then solves the
          three simultaneous equations to up-date the three coefficients.
          They will obviously have a built-in lag of 5 days. However, be-
          cause  the seasonal temperature changes are quite gradual, the
          errors should not be  substantial.  This  program has not yet
          been implemented.
            It is believed  that the results are quite  promising and that
          this program can be  very useful. It is also  believed that it can
          be improved in several ways. For example, the effluent BOD
          should be  incorporated into the program.  Also, the  air flow
          readings should be adjusted to standard conditions from the
          readings of  air  temperature and  pressure. The endogenous
          respiration rate might be expressed as a factor times the mea-
          sured  concentration  of mixed liquor suspended solids. Addi-
          tional work should result in the development of a reliable pro-
          gram  for  the  evaluation  of this  very  important  calculated
          variable.
          REFERENCES
          1.   Eckenfelder,  W. W., Jr.  and O'Connor,  D. J., "Biological Waste
              Treatment," Pergamon Press. New York, NY (1961).
                                                        105

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                                DISCUSSION
 Joseph f. Roesler:
   The installation  of a computer can be considered as upgrad-
 ing  a  wastewater treatment plant. Or  computers can be con-
 sidered as O&M tools, especially if they are used for data ac-
 quisition and handling. I would appreciate comments from the
 various EPA regions in regards to the region's policy towards
 awarding construction grant money for computers. Secondly, I
 would appreciate  any  one else's comments on what  EPA's
 policy should be and why?

 Michael  K. Stenstrom:
   From a manager's point of view, how much calibration of
 Dissolved Oxygen probes can you afford?

 Walter W. Schuk:
   To provide some guidelines for instrument manufacturers it
 would be helpful if treatment plant management would define
 the  meaning of reasonable maintenance in more detail. Please
 give your  definition, including both the frequency and man-
 hours that would be considered acceptable for maintenance of
 an on-line analyzer.

 Allen E. Molvar:
   Would  Mr. Garrison comment on the expected benefits of
 flow equalization,  particularly in terms of potential process
 improvement?

 Heinrich O. Buhr:
   The  development of a  method for calculating the respira-
 tion rate in  an activated sludge system, as in Equation (5) of
 Dr. Dobbins' paper, is an important step towards the goal of
 controlling the respiration rate  per unit mass of sludge, or
 specific  oxygen utilization rate (SCOUR). It may be shown,
 from  the classical substrate and sludge mass balance equations.
 that a constant value for  SCOUR also means a constant spe-
 cific growth rate, constant F/M ratio and a constant value for
 soluble  BOD in the effluent. Keeping these parameters con-
 stant  further holds out the promise of more consistent sludge
 settling  characteristics, and it seems likely that SCOUR control
 will become one of the most useful concepts in automatic con-
 trol of activated sludge plants.
   Once a value for R, Equation (S), is available on  a regular
 basis, the  practical implementation of SCOUR control would
 be to calculate the sludge  concentration required in the aerat-
 ors, and then  to control the sludge return rate appropriately.
 For example,  if primary flow rate,  Qs, RAS return rate Qr,
 and, say, RAS concentration, X,, can be measured, then the
 RAS flow rate required will be
            Qr = Qs- ^required / C*r - X^,,,,;,.^),
 if sludge growth in the aerators is ignored  for practical pur-
 poses. As Dr. Dobbins pointed out, one of the advantages of a
computer-based system is that it facilitates the implementation
of such multi-input control strategies.
   In a system which is subject to strong diurnal load fluctua-
tions, the maintenance of a constant SCOUR will dictate equi-
valent variations in  the sludge content of the aeration tanks.
This infers that sludge storage  capability will be  required, well
beyond  what could normally  be  provided by the  secondary
settlers. Apart from the use of  storage tanks, one way of tackl-
ing the  problem is  to employ the step feed process,  where a
portion  of the aeration basin at the head of the plant is  used
essentially for  sludge  storage  during  periods   of  low flow.
Another method, which  reduces  the  variation  in  sludge re-
quirements, would be  to reduce incoming load fluctuations by
providing equalization capacity, as proposed by  Mr. Garrison.
Front-end load balancing should certainly be recognized as a
valid control technique, and. in general, an appropriate balance
should be struck between more equalization capacity, on the
one hand, and more complex in-plant control, on the other.

Peter C. Young:
(Editor's note: Dr. Young has combined his comments on the
individual workshop topics into a single document, which  is
given under this section because of the particular applicability
of many of his remarks to computer-based systems.)
   This Workshop has raised many important questions about
research needs for  the automation of wastewater treatment
systems. But the term "automation"  has, I feel,  been inter-
preted in a  rather narrow sense, with  much of the discussion
centering on the automation  of existing manually operated
processes without considering, in  anything but fairly super-
ficial detail, the possibility of improving these processes by the
utilization of advanced control systems analysis  and design
methodology.
   We have, for example, talked at great length about  the limi-
tations  of existing sensors and the need for research  into the
development of improved or entirely new sensing devices. But
automatic control is  not only concerned with the measure-
ment of variables; important as they are. sensors are only one
aspect of control system design. And the modern approach to
systems analysis and control systems synthesis requires an inte-
grated approach  to the  design problem,  with the sensor re-
quirement studies proceeding  in parallel with equally import-
ant  investigations such as  system modelling, simulation and
overall control system design.
   It might  well  be, for instance, that a better understanding
of the dynamic process, obtained by thorough modelling exer-
cises, could  lead to  the design of new types of control schemes
whose sensor requirements might not  necessarily be the  same
as existing designs. In other words, there is a need  for con-
tinual cross-fertilization  of ideas  from different  component
                                                          106

-------
                                                                             COMPUTER APPLICATIONS IN AUTOMATION
areas of research if the final automated system is to attain its
full potential. To give but one example: there has been a great
deal of discussion  on BOD measurement and the need for ob-
taining estimates of BOD more rapidly  than at present. And
yet it might be that future research into control systems design
will negate the requirement for BOD measurement, either be-
cause variables other than BOD are found to be more useful
for control  purposes or. alternatively,  because good statistical
estimates can be obtained in real time from other more easily
obtainable measurements.
   My plea  is simply that we do not close our eyes to the pos-
sibility of radical innovation at this early stage  in the study of
wastewater  treatment automation, but rather that we keep all
options open, bearing in mind the kind of developments which
may be possible if an adequate program of research in systems
analysis and control systems design is undertaken. At least it is
incumbent on those who authorize and finance research in the
area to promote studies which encompass all important aspects
of the  problem: they should not. it  seems to me. allow them-
selves to  be unduly influenced by groups who represent only
certain special,  albeit  important aspects of  the problem and
who may not be  cognizant of developments in areas outside
those covered by their own expertise and interests.
   Of  the  more  detailed  topics we have considered  at the
Workshop,  I feel  that there is need for  some clarification on
the adequacy of dynamic models. When describing a dynamic
system in mathematical  terms,  it should be emphasized that
there is not one, all-encompassing dynamic model-some, as it
were,  universal panacea  to all  modelling problems.  Rather
there is a whole family of models having different degrees of
complexity  depending upon  the  nature of the problem for
which  they are formulated; indeed, it could be argued that one
of the major arts  of systems analysis is the choice of a mathe-
matical model which suits the nature  of the problem at hand.
   In  this  latter  sense, the more  complicated  and  detailed
models, such as those considered by Dr. Andrews, are simula-
tion models aimed first at deriving  a  better understanding
of the physico-chemical and biological nature of the  process,
and second at providing a vehicle for assessing the possible effi-
cacy of  different  control  strategies before  these are imple-
mented in  practice. Provided such  models are developed and
used with care  and, in particular, provided  the user does not
succumb to the inherent danger of believing that his model is
the system  and not  merely a mathematical  model, then such
modelling exercises can be extremely useful, providing as they
do a natural prelude to systems analysis and control system de-
sign.
   But it should be realized that complicated simulation mod-
els are, more often than not, unsuitable  bases for analytical
control system  design. For these tasks, the systems analyst re-
quires much simpler and more analytically tractable models
 which reflect the dominant dynamic modal behaviour (Le., the
 behaviour of the  dominant modes that characterize the system
 response to excitation), without including the extraneous de-
tails (i.e., extraneous to  control system design requirements)
that tend to characterize a thorough micro-analysis of the pro-
cess. For example, it has been established recently (1,2) that a
simple second-order differential equation model  can explain
the dominant daily variations of DO and BOD in a non-tidal
river system, even though we know that the detailed behaviour
is much more complex. In many examples it would seem that
the multitude of higher-order, nonlinear and stochastic effects
that naturally characterize  the  detailed operation of the sys-
tem tend to combine in  their overall effects and can often be
represented in gross or macro terms as additive stochastic dis-
turbances with  fairly simple statistical properties: it  is as if
there is a Law of large  systems, somewhat analogous  to the
well known Law of Large Numbers,  so that, while the system
is exceedingly complex, its overall small-perturbation response
characteristics can often be quite simple  in form, apparently
dominated by a clearly defined and, more often than not,
linear mode of operation.
   It seems to me that the importance of the  above points in
relation to the automation of wastewater treatment systems is
that, while excellent work is being carried out in the simula-
tion modelling area and  a growing understanding of the prob-
lem is emerging, there appears to be no determined research ef-
fort in the other modelling areas. And this is  despite the fact
that analytical techniques  such as model identification, para-
meter estimation and evaluation are available and have been
applied  with reasonable success to  other  related problems,
such as  the analysis of  water resource systems and chemical
processes (3,4).
    To consider one example where  available  analytical tech-
niques may be  of use. the question of rainfall prediction has
been raised in connection with wastewater collection systems.
Such problems are closely related to  the flow forecasting prob-
lems encountered by hydrologists and recent work on flow
forecasting which uses sophisticated but relatively  simple  re-
cursive  methods of time-series analysis (3,5) may well have po-
tential in the analysis of wastewater collection systems; cer-
tainly initial reference  to  these techniques could provide im-
portant a priori information for research workers dealing with
such problems.
    Another  analytical  technique developed  in the   control
systems area which appears to have reasonable potential for
application to wastewater treatment problems is the numerical
filtering of noisy data from dynamic systems-or, as it is often
called in the technical literature, state variable estimation or
Kalman filtering (6). This approach to data processing provides
a method for using the  measured variables (which will almost
 certainly be corrupted by measurement noise  or subject to un-
 certainty of some type) to reconstruct, on a firm statistical
 basis, estimates of any unobserved or non-measurable variables
 that characterize the process dynamic behaviour while, at the
 same time, helping to filter off the measurement noise effects
 from the measured variables.
    A heuristic (Le., intuitive, not based on rigorous analysis or
                                                          107

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
 derivation) and largely deterministic analog of such a proce-
 dure is the estimation of BOD as suggested by Dr. Dobbins. In
 effect, the available analytical techniques are able to put such
 heuristic procedures on a firm statistical base, so making them
 more  systematic to design and potentially more reliable. Of
 course the analytical techniques, like their heuristic relatives,
 require a simple dynamic model of the process which explains
 the  dominant model behaviour; thus the provision of an ade-
 quate  mathematical modal is once again seen to be an inherent
 requirement of system design.
   Another extremely important point raised in the Workshop
 discussion is the use of computers in automation. Although
 many  of the advantages arising from the use of computers have
 been mentioned, I  feel there has been insufficient emphasis on
 the  fact that one of the major attractions of having a computer
 as part of the operation is that it enables the designer to con-
 sider  the integrated  control of the whole system: no longer
 does he only have to consider  the local  control of unit pro-
 cesses, but  he can also consider the  interaction of such pro-
 cesses and the design of control schemes that allow for such in-
 teraction. And  because the computer is such a powerful tool,
 the designer can, at the unit process level, also consider the de-
 sign of a control system that acknowledges the multivariable
 fie., multi-input, multi-output) nature of many unit processes
 (7,8)  and takes account, for  example, of factors such as  the
 stochastic nature of the process and the overall objectives.
    Of course, the  sophisticated use of computers in  control
 system design holds  many problems, most of which still have
 to be  solved. But there is no doubt that an important research
 need  is for people in the wastewater treatment field to become
 acquainted with recent developments in advanced computer
 control so  that at least they are  able to judge the claims of
 over-enthusiastic control system design consultants! There is
 no  doubt that in the next few years at least some of the tech-
 niques which are at present only gleams in the eyes of the re-
 searcher will come  to fruition.
    Finally, lest it  be thought  that I am over-emphasizing  the
 need  for research  on advanced systems analysis, it should be
 stressed that I am recommending here the use of techniques
 which, for the most part, have proven practical potential in
 other areas of application. Indeed, I am a firm opponent of in-
 novation for its own sake and would urge great caution in con-
 sidering  the use of  some of the  more advanced concepts in
 systems design such as self-adaption and optimality.
    Such  concepts  are, of course, extremely tempting because
 they  seem to offer solutions which, in the one case are able to
 counteract the effects of possible changes in the controlled
 process,  and in the  other are "best" in some sense. But it is
 only  fair to point  out that both approaches to control system
 design have not been particularly successful when applied to
 real (in contrast to simulated) systems. Even in the aerospace
 field  for example, where the need for adaptive adjustment of
control gains is emphasized by the large and rapid changes in
vehicle  dynamic characteristics that occur  because of  the
rapidly changing environment, the only widely used method of
adaption is the schedule adaptive system, in which the control
gains are pre-programmed to  vary as functions of "air data"
variables such as dynamic pressure,  altitude. Mach number.
etc.;  true self-adaptive control, in  which the system adjusts
itself without considerable a priori information, has been used
on  relatively few occasions and then  mainly for  advanced
"one-off" research aircraft.

   In the case of dynamic optimal control the picture is much
worse. First  of all, it should be realized that the system so
designed is only optimal  in the  sense  that  it  achieves  the
extremum  (max. or  min.)  of some  chosen  performance
objective or cost function; there is no guarantee that this cost
function is necessarily the best available such function, or even
that a suitable analytical statement of the desired performance
objective can be established at all.
    An example  of the inappropriate use of optimal control is
 the popularity of the "Linear-Quadratic"  (L-Q)  approach to
 optimal control system  design  which   received   so  much
 attention in the period 1960-1970(9). A considerable amount
 of time and money was expended on the design of optimal
 systems to this kind of specification even though, in its basic
 form,  the resulting  control  law  does not  include integral
 action, so  that the resulting controlled system can have rather
 undesirable steady-state  performance characteristics. (In  par-
 ticular,  the  resulting controlled  system  can be  extremely
 sensitive to the unavoidable uncertainty in the model parame-
 ters  and can  be characterized by  considerable steady-state
 error to setpoint or  constant input commands).  While this is
 not the only undesirable aspect of L-Q optimal designs (10) it
 is, perhaps, one of the principal reasons for the notable lack of
 success of such design procedures during the 1960's; certainly
 the need for inherent zero steady-state  offset to  set point
 inputs, as  provided by integral action, is one of the principal
 requirements of most practical control schemes, and the  fact
 that  a considerable period  of time passed before this disadvan-
 tage  was diagnosed and corrected (see for example Young and
 Willems (10))  illustrates   the  dangers  introduced by   the
 uncritical use of theoretical design procedures.
    It is because of such mistakes as these in the past that I urge
 the support  of a strong program of research into  advanced
 control systems analysis and design  for wastewater treatment
 processes; I am sure that if an enlightened program of research
 is initiated now, such aberrations are less likely to occur. A
 healthy body of control systems specialists will become firmly
 established; specialists who will be able to judge the utility of
 new innovations within  the context of wastewater treatment
 and so, in the words of the preamble to this Workshop, help to
 realize the many potential benefits offered by automation.
                                                          108

-------
                                                                            COMPUTER APPLICATIONS IN AUTOMATION
REFERENCES:
 1.  Young, P. C. and  Beck, M. B-,  "The  Modelling and Control of
    Water Quality in a River System," Automatica (1974) (in press).
 2.  Beck, M. B. and Young, P. C-, "A Dynamic Model for DO-BOD
    Relationships in a Non-tidal stream," Water Res. (to be published).
 3.  Young, P. C., "Recursive Approaches to Time-series  Analysis,"
    Bull. Inst. of Math, and its Applications (I.M.A.), 10.  No's 5/6,
    209-224(1974).
 4.  Astrom, K. J. and Eykhoff, P., "System Identification: A Survey,"
    Automatica, 7.123-162(1971).
 5.  Young, P. C., Shellswell, S. H. and Neethling, C. G., "A Recursive
    Approach   to  Time-Series  Analysis,"  Tech.  Report  No.
    CUED/B-Control/TR16 (1971),  Engineering Department, Univer-
    sity of Cambridge, England (1971).
 6.  Kalman, R. E., "A  New  Approach to  Linear  Filtering and
    Prediction  Theory," Trans, ASME, J. Basic Eng., 82-D,  35-45
    (1960).
 7.  Macfarlane, A. G. J., "A Survey of Some Recent Results in Linear
    Multivariable Feedback Theory," Automatica, 8, 455-492 (1972).
 8.  Macfarlane, A. G- J., "Relationship Between Recent Development
    in Linear Control Theory and Classical Design Techniques," Third
    IFAC Symp. on Multivariable Technological Systems, Manchester,
    16-19 Sept. 1974.
 9.  Special Issue on Linear, Quadratic Gaussian Problems, IEEE Trans.
    on Automatic Control, AC-16, No. 6 (1971).
 10.  Rosenbrock, H. H. and McMorran, P. D-, "Good, Bad or Optimal,"
    IEEE Trans, on Automatic Control. AC-16. No. 6, 552 (1971).
 11.  Young, P. C. and Willems, J. C., "An Approach to Multivariable
    Servo-mechanism  Problems," Intl. Jour. Control, 15, 961-979
    (1972).
CLOSURE
  (Editor's note:  In  preference to  submitting  a closing
statement, Messrs. Garrison, Payne and Haug have  modified
their paper in  the  light  of the points raised  during the
discussion.)
W. E. Dobbins:
   The question has been raised as to what is an acceptable
level of maintenance requirement for various water quality
sensors. The experience  in  Bridgeport,  as  reported by  the
Instrument Engineer, is as follows:
D. O. Probes: The three step-aeration tanks have a total of 18
probes—general maintenance, including  cleaning, membrane
and  electrolyte replacement  and calibration, requires about 5
man-hours per week.
Suspended  Solids  Meters:  Six meters of the  optical  type
require about 2 hours  per week, principally for  lens cleaning
and  calibration. This  moderate  requirement was  after  the
installation  of a simple automatic flushing system utilizing a
tuner and two small electrically controlled valves (see article in
WPCF "Deeds and Data" October, 1974).
Water  Quality Analyzer: An  on-line water quality analyzer has
probes for  pH,  conductivity  and  temperature,  and  wet-
chemistry analyzers for chromate, cyanide and copper. A total
of about  9 hours per week  are required to  keep the system
operating satisfactorily. The  work for the most part is cleaning
the probes, and replacing filters in the pretreatment system.
   The total effort  consists of about 16  man-hours  per week.
The  Instrument  Engineer  estimates that  one adequately
trained technician  can  keep the  systems  in  both of the
treatment  plants in generally satisfactory operating condition.
In the writer's opinion,  this is an acceptable requirement for
these  plants in which  the total operating personnel numbers
about 80.
                                      Report  of Working  Party
                                                        on
                                            RESEARCH  NEEDS
                                                       for
                    COMPUTER APPLICATIONS  IN  AUTOMATION
                                                Carmen F. Guarino
                                 Commissioner, Philadelphia Water Department,
                            1160 Municipal Services Building, Philadelphia, PA  19107
                                                  John M. Smith
                Chief, Municipal Treatment and Reuse,' National Environmental Research Center,
                            Environmental Protection Agency, Cincinnati, OH 45268
 INTRODUCTION
   Computers are being widely  used to expedite work in the
 chemical process industry and to economize the endeavors of
 man. To date, computer  application  in  the  wastewater
 treatment field  has been minimal. The use of computers will
 solve many  problems  that presently confront the effective
 treatment of domestic and industrial waste.
   Without the  use of computers, it will be impossible  to meet
 the  requirements of P.L. 92-500  and  other related laws
 concerning the  control and treatment of pollution. Presently,
 there is little control of the treatment process and this is more
obvious in the smaller plants than the large. The only way that
we can adequately control the treatment system is through the
development of instrumentation and automation.
   We must  have some method whereby the  process is
monitored continuously, not once every  24 hours or once a
week which is the case in  many  small plants.  Remembering
that there are far more small plants than large plants, this does
have  a tremendous impact  on national pollution control and
treatment. The  development of model systems  that can be
applied to both small and  large plants will result in effective
                                                        109

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
treatment and in a reduced cost of treating wastewater.
   There is very little regulation of energy used in treatment
plants today because we do not have adequate control. There
should be  no  doubt in our  minds that there  would  be  a
tremendous saving of energy as  well as materials if we only
applied them as necessary, and not on a routine basis as is now
being done. A check of most activated sludge plants will reveal
that they usually use the same number of blowers to supply air
at 3 o'clock in the  morning as  they  do at 3 o'clock in the
afternoon.  Yet, conditions usually vary greatly at these hours.
Over a period of time,  the prudent use of power nationally
would amount to a great saving.
   Instrumentation   and automation  will  furnish  the best
treatment at the least cost.

PROBLEMS
1. The  application  of  computers to  wastewater treatment
   systems is new. Consequently, there is a primary need to
   assemble all  information from both the national and inter-
   national viewpoint.  The first problem is  the lack of  a
   complete computer library as applied to wastewater treat-
   ment systems.
2. The  unavailability  of suitable  models  to  simulate  the
   treatment process.
3. The  absence of  reliable  sensors  to monitor  important
   process  variables. This  limits the  use of  models  and
   computers at this point in time.
4. For the  most part, the wastewater  treatment system  is
   treated separately from the collection system. For effective
   treatment  and the most effective automation, the system
   must  be  considered and  made  one.  The  absence  of
   sufficient information as to the dynamics of sewer collec-
   tion as well as the impact of rainfall, infiltration, and means
   of controlling flows in  the wastewater treatment system
   limits the most effective use of the computer.
5. There are many  specialists in the  wastewater treatment
   systems field and there are many specialists in the use of
   the computer, but there are very few people that under-
   stand both. Consequently, this limits the full application of
   the computer in  solving the  problems of the  wastewater
   treatment  process. There is  a need for greater education to
   produce personnel who understand  both facets of this
   work.

Summary
   Successful computer application will  require educated and
trained personnel, reliable sensors, and  time-tested and avail-
able mathematical models.


RESEARCH NEEDS
  1. Education—complete knowledge of the wastewater treat-
    ment  process  and complete knowledge of  computer
    applications.
        It  was obvious at the Workshop that the treatment
   specialists  and  computer  specialists  were not  able  to
   completely communicate. Before the treatment process
   will be successfully computerized, we must have a clear
   understanding of  the  treatment  process as well as the
   application of the computer.
 2. Development of reliable sensors which can, on a real-time
   basis, monitor the treatment process. These instruments
   would include:
   a.   B.O.D.
   b.   C.O.D.
   c.   Suspended Solids—mixed liquor suspended solids
   d.   Sludge density
   e.   Sludge blanket level detectors
   f.   Ammonia and nitrate
   g.   Biological activity indication
   h.   Flow measurement in the collection system
 3. Preparation of basic models which can be made available
   to those who are attempting to automate the treatment
   processes.
        There should be two types of models; one that can be
   used effectively  with  the  present  knowledge  of  the
   treatment process as  well as  the  present knowledge of
   sensors; and  another to make allowances for the sensors
   which are being developed.
        Models worthy of development are:
    a.   Activated sludge process and all its modifications
    b.   Chemical treatment
    c.   Aerobic  sludge digestion
    d.   Anaerobic sludge digestion
    e.   Collector system control
        All of the  above, of course, have  many variations
    depending on the particular  process chosen; for example.
    under 'Chemical Treatment',  adsorption  could be  in-
   cluded, etc.
 4. Test and evaluate models on a pilot plant basis. There is a
   great need  for a pilot  plant which can be  used  to both
   formulate and test models. This would  be the  greatest
   contribution  made towards  exploring the use of com-
   puters in controlling the treatment process. Proper pilot
   plants and  proper personnel could develop and de-bug
   models which could then be made available for use.
 5. Establish some unit that would serve as  an information
   center for all agencies working towards the automation of
   the  treatment process.  This central  office would serve
   many purposes,  from information dissemination  to con-
   tinuously updating and improving models and sensors.
6. Integration  of treatment system with collection  system.
   The collection system, through the use of computers,can be
   the first step in the treatment process. Control of the flow
   to the  treatment  plant, through effective use  of  the
   collector system, has many advantages from flow equaliza-
   tion  to being able to treat more flow, particularly during
   rainstorms.
7. Inter-relationships in the treatment  process  should be
                                                         110

-------
                                                                            COMPUTER APPLICATIONS IN AUTOMATION
clearly defined.  In order to computerize the system, a
program  must be written. A program cannot be written
until the  process can be clearly defined.
A complete technological assessment of wastewater treat-
ment  via computer applications is required.  The assess-
ment should include, but not be limited to. the following:
a)   Social and psychological implications
b)   Economics
c)   Risks, liability, etc.
d)   Man/machine interfacing
e)   Political and regulatory implications
Determine  the role of the computer  in wastewater treat-
ment with  respect to matching the  size  and the type of
computer (full-size, mini, micro-processor or analog con-
trol) to the plant size and the planned function (control.
data acquisition,  scheduling,  accounting,  etc.) of  the
computer.
10.  Determine specifications or guidelines in the selection of a
    computer for  a  particular facility. Preferably, the guide-
    lines  should  be presented so that the function of the
    computer is described and related to the size and type of
    computer required.
11.  Large-sized wastewater treatment systems display a poten-
    tial for using a centralized computer for data acquisition,
    monitoring by use of  the central  computer and finally
    using the central computer to assist in the  preparation of
    the monthly state  report. Programs should  be written for
    this application  and the concept should be  demonstrated.

12.  Using a centralized computer and several satellite plants,
    each having a terminal, demonstrate the effectiveness of
    using a man-in-the-loop  to control each plant by commun-
    icating with the computer.
                                                       111

-------
                           EVALUATION
                                     OF
THE EFFECTIVENESS OF AUTOMATION
                           Workshop on Research Needs
                  Automation of Wastewater Treatment Systems
             113

-------
 AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
          EVALUATION OF THE EFFECTIVENESS  OF AUTOMATION

                                                Joseph F. Roesler
               Pilot & Field Evaluation Section, Advanced Waste Treatment Research Laboratory,
                         National Environmental Research Center, Cincinnati, OH 45268.
   The relative value of automation in the field of wastewater
treatment can only be determined by accurately evaluating the
effectiveness  of  the various  automatic control  strategies.
Careful planning  of  an  evaluation is  such an important task
that failure can almost  be guaranteed if adequate planning is
neglected.  Two methods that can be used for evaluating the
effectiveness of automating a wastewater treatment plant are
field evaluation and desk-top computer simulation. Important
and interrelated parameters in any such evaluation are cost-
effectiveness of the  control strategy, performance improve-
ments, and energy  requirements. This  paper describes  the
criteria that  should  be  considered in planning an  evaluation
and presents field data from several evaluations sponsored by
the U. S. Environmental Protection Agency (EPA). Problems
encountered  and  some of the  final conclusions are  also
included.

EVALUATION TECHNIQUES
   The usual  technique employed  in field evaluation is to
compare  the performance of a manually operated plant with
an  automated plant; however, the  standards  for  manual
operation vary according to the idiosyncrasy of each plant.
Plant layout, piping,  type of sewage, etc., may affect the ease
with which the plant  is manually operated.  There are  no
baseline data available that show the performance of a well-run
typical wastewater   treatment  plant. Therefore each per-
formance  evaluation  is a comparison of automatic  versus
manual operation  at  only one plant, and each performance
evaluation requires  an   exacting  definition  of the  manual
operating procedures.
   Before beginning an  individual  plant evaluation to assess
automation performance, the operator-manager attitude must
be carefully evaluated. In some cases, operators believe that
the security of their jobs is threatened by automation. In such
cases, the instruments and automatic control  loops  will  be
placed under the most severe test conditions possible and may
 thus "fail." The cause of the  failure may be neglect or even
 sabotage of the equipment  by the operators. On the other
 hand, operators  and  managers  may welcome automation;
 however, because  of their inexperience with  the equipment
 and lack of proper  training, the equipment may still  "fail"
 because of unintentional neglect or faulty maintenance.
    Negative attitudes must be  corrected.  Assurances must be
 given  that  the operator's job security  is not  threatened.
 Adequate  instructions must  be given to operators and man-
 agers. In plants installing a digital computer, instruction must
 be given to the managers so that they comprehend the needs
 and requirements of such an installation.
    Many plants are symmetrically constructed, giving rise to
 duplicate systems. In such cases, the tendency is to compare
 the automatic control  of  one  stream  with a manually
 controlled  duplicate stream. In such a comparative evaluation,
 the "Hawthorne" effect (1)  must  be considered. The classic
 studies at Hawthorne, Illinois, were conducted by the Western
 Electric  Company in the late 1920's. The results of  this
 research  indicated that the employees' behavior and attitudes
 toward their jobs and toward  management are conditioned to
 a considerable extent by the values, standards, and expecta-
 tions of the work groups to which they  belong. Thus there is
 an  excellent possibility  that  operators working on the man-
 ually operated portion of the plant will work harder and more
 diligently  to demonstrate that man is  superior to machine,
 thus invalidating the comparative test results.

 COMPARATIVE FIELD EVALUATIONS SPONSORED  BY
 EPA
   Once the group behavior  problems are eliminated, simul-
 taneous  comparative  evaluations  will  yield  excellent  and
 reliable results. An  example  of  such  an evaluation (Le.,
without the group behavior problems) is the dissolved oxygen
(DO) study recently carried out at the U. S. EPA's Pilot Plant
at Blue Plains. The objective of the study was to determine  the
                                                       114

-------
                                                                  EVALUATION OF THE EFFECTIVENESS OF AUTOMATION
effect of DO level on  filamentous growth by comparing the
performance of two automated control strategies against one
another. Both systems  were operated at a level of 1 mg/1 for a
period of five sludge retention times (SRT's) without incurring
sludge bulking problems. One  system (D system)  was  then
operated at a DO level of 3 mg/1, while the second system (E
system) was maintained at  a DO level of 1  mg/1. After five
SRT's,  the sludge  volume  index  (SVI)  of the D  system
increased to 200-300, while the SVI of the E system remained
at a level of about 100. The DO level of each system was  then
set at 1 mg/1 for a period of five SRT's, thus causing the SVI
of the D system  to decrease to 50.  This simple comparative
test makes the dramatic point that DO set point does have an
influence on the sludge handling characteristics.
  Two other  DO studies in which the comparisons (manual
versus  automatic)  were  sequentially  performed  are   next
described. One of the studies, at Rent on, Washington (2) tends
to confirm the above observations, but no such clear confirma-
tion  resulted from the second study, conducted at Palo Alto,
California  (3). Though it  is  too early  to come  to  any
conclusions, it appears that the  nature of the comparative  tests
(simultaneous  versus  sequential) does have an effect on the
results. Obviously,  the  main  risk of conducting sequential
comparative tests (Le.,  manual  versus automatic operation) is
that  raw sewage entering the plant may differ in nature from
                            one  time frame to the  next. This possibility is an inherent
                            disadvantage that,  at best,  can be  compensated  for  only
                            partially.
                               The  Renton plant was operated for  about a year (March
                            1970 to April 1971) under manual control while an automatic
                            DO control system was being installed in the new aerator. The
                            following  year, the  plant was successfully operating  with
                            automatic DO control. Data were  collected for comparative
                            purposes during the  months  of  October, November and
                            December for years 1970 and 1971. The operators and  plant
                            management had  an excellent attitude toward  automation.
                            Also, the manual control policy was well defined and expertly
                            carried  out.
                               The  obvious question, however, is whether the sewage was
                            identical for both time frames. One partial answer is that the
                            BOD loading to  the plant increased about  50% during the
                            automatic  control   period.  In  spite of  this  increase, the
                            performance of the  plant did  improve.  The effluent  BOD
                            decreased  from a geometric  mean of  11.1 ppm, obtained
                            during manual operation, to a mean of 3.9 ppm for automatic
                            operation. Figure 1  shows the  effluent  BOD data plotted on
                            logarithmic  probability  paper  to obtain  a more normal
                            distribution of measurements.
                               Further  analysis indicated  that  the  sludge characteristics
                            may also have  been affected by automatic  DO control. The
                  100
                  90
                  80
                  70

                  60
                                           95  90    BO  70 60 50 40  30  20    10   5    21  0.5  0.2 0.1 005 0.01
               a
               o  »
               a  to
                  50
                  3P
                                        I     I
                                                                    AUTOMATIC CONTROL
                   IP
                   0.01
i  I   I   I _ '
            I _ I
                                                     1 - L
                       00501 02
0512    5   10    20  30  4O  50  60  70   80    90   95   98 99 995 99.8999

  PERCENT OF OBSERVATIONS EQUAL TO OR LESS THAN STATED CLASS MEAN
 Figure 1. Frequency Distribution of BOD in the Effluent
                                                        115

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
 frequency  distribution  of the  sludge volume index (SVI) is
 shown in  Figure 2. The arithmetic mean for the SVI with
 manual  control  was  332. This  was reduced  to  86 with
 automatic  control.  Since BOD is a broad analytical term and
 since this  was  a sequential comparative test,  the  data  are
 promising  but inconclusive. This  study, therefore, requires
 confirmation with further research. The U. S. EPA Pilot Plant
 at Blue  Plains has  already begun such research, as indicated
 previously.
   At  the  Palo Alto Sewage  Treatment  Plant,  four control
 strategies were evaluated and compared to manual operation.
 The control schemes were DO, air/return sludge (DO/RAS),
 respirometry, and mixed liquor suspended solids (MLSS). Each
 control strategy was evaluated  for about  30 days. There were
 three  well-defined  manual operations to accommodate  the
 operating requirements associated with wet- and dry-weather
 flows.
   As  compared  to manual operation, DO control showed a
 13%  performance  improvement  in terms of TOC  measure-
 ments and an 11%  reduction in air use. For MLSS control,  the
 solids level in the aerators were maintained close to set point,
 although a low limit on RAS pumping resulted in uncontrolled
 increases in MLSS during periods of low plant flow.
   During  the evaluation of the MLSS control scheme, proper
 sludge management became necessary.  A separate aeration
                                          basin was converted to a sludge storage tank to accommodate
                                          high loadings to the plant. Perhaps because an aerator basin
                                          was being used for sludge storage, or perhaps because of sludge
                                          age,  a scum appeared in the final clarifier. This scum would
                                          not settle and, as a result, caused a poor quality effluent. When
                                          sludge storage was eliminated, the scum disappeared and the
                                          effluent quality improved.
                                             One technique for complete control  of the aeration basin
                                          consists  of regulation of food-to-microorganism  ratio (F/M).
                                          The  technique  for measuring the  food  differs: It  may  be
                                          accomplished by measuring TOC, COD,  or  oxygen uptake.
                                          Two different types of suspended solids meters were used to
                                          indicate  the  microorganism  concentration. The  readings of
                                          both   meters  correlated  well  with  each  other and  with
                                          laboratory suspended solids values.
                                             Suitable  automatic TOC and  COD  analyzers were  not
                                          available for on-line control during the Palo Alto experiments.
                                          Therefore, only two  F/M  control strategies were evaluated.
                                          These  were  feedback  respirometry  control  and  DO/RAS
                                          control.  For DO/RAS control, the air flow was used to infer
                                          food, and then  the return sludge was adjusted accordingly. In
                                          other  words, the entire  aeration tank was used as a respir-
                                          ometer to set the return sludge flow.  The  results of these
                                          experiments were inconclusive, probably for three reasons: (1)
                                          more time was needed for collecting  data than was allotted.
               IOOO
               9OO
               •00
               7OO
               6OO

               5OO
             E
             _-  no
                        999998 995 99 98   95  9O   »0  70 60 50 4O  30   20    10   5    21  05 02 O.I 0.05  001
	    I
                                                                     AUTOMATIC CONTROL
                       I	I
                                                               1	I
                  O-OI 00501 0.2  O.S  I   J    5   10    2O   3O  40  SO  6O 70  80   9O  95   9«  99 99.5 99.899.9

                                PERCENT OF OBSERVATIONS EQUAL TO OR LESS THAN STATED CLASS MEAN
Figure 2. Frequency Distribution of SVI
                                                         116

-------
                                                                 EVALUATION OF THE EFFECTIVENESS OF AUTOMATION
Table 1. Costs for Various Control Schemes




Control scheme
Mass-proportional
Flow-proportional
Operator control
No control
for


Ibs Alum
per day
16,290
26,850
28.260
35,700
Alum Addition at a 10-mgd

Capital
cost
S
$112,000
128,000
134,000
145,000
Plant (August 1971)

Cost of
alum
c/Kgal
1.66
2.74
2.88
3.64

Total
cost of
plant,
c/K gal
2.22
3.33
3.48
4.27

Annual
cost
Thousands/
year
81
121
127
156
(2) difficulties were encountered in separate sludge storage and
in the ability  to  supply the  proper amounts of RAS at low
flows, and (3) assurance  was lacking that  the manual tests
could be accurately correlated with their adjacent automatic
tests.

COMPUTER SIMULATIONS
  An alternative to field evaluation  is computer simulation.
But for effective studies, such simulations must use mathe-
matical models that consider the entire system and have been
demonstrated  to  be  accurate  by  field  evaluations.  Most
mathematical models emphasize treatment of liquid waste; as a
result, there is little work on determination of sludge quality,
separation, and  subsequent  handling. There is evidence that
DO control does affect sludge characteristics, but the mathe-
matical models that exist today do not simulate this. Conse-
quently, accurate cost and performance data must be  obtained
from field evaluations.

COST-EFFECTIVE ANALYSIS AND PERFORMANCE IM-
PROVEMENTS
  The  relationship of  cost  to  performance  is  sometimes
overlooked in  evaluations. To obtain an accurate cost-effective
analysis, all data  must be normalized. The cost of improving
the performance of a plant removing about 90% BOD  has been
shown  (4)  to  follow  an  exponential  function.  Manually
operated plants typically remove 85 to 90% BOD; if automa-
tion were to  improve this performance, then such  improve-
ments must be considered in computing cost savings  to avoid
significant errors in cost-effective determinations.
  However, there  are  certain processes in  which  the per-
formance appears to be negligible for cost-effective  analyses.
Such processes are normally  dependent on a chemical addition
in which  a  minimum  amount  of  chemical achieves the
performance desired. Further addition of a chemical  is simply
wasted. An example  of such a process is alum addition for
phosphorus removal.
  Costs for alum  addition at a 10-mgd plant for various
control schemes were calculated by Convery et al. (5) and are
shown in Table 1.
   Periodic operator control provided a saving of 529,000 per
year. This saving could be increased by more frequent operator
attention. Similarly, if flow-proportional control were em-
ployed, an additional saving of 55,500  per year would result.
Feedforward mass-proportional  control  provided  an  annual
savings  of $40,500 compared to flow-proportional control
alone. This amount does not include the  significant savings in
sludge-handling costs  that would also accrue,  yet  it  is more
than adequate to justify the purchase of a phosphate analyzer.
Similar  cost  comparisons  have  been  made  for methanol
addition for denitrification (6)  and for breakpoint chlorination
(7).

ENERGY REQUIREMENTS
   The  final parameter  in  evaluating  automation is  energy
utilization.  The Palo  Alto study  showed  that  simple  DO
control can  provide a power saving of 11% and consequently
an economic savings of 55,500 per year.  Since air blowers are
one of the largest energy consumers within a treatment plant,
this energy saving is very significant.
   If  the  entire U. S. population were  served by activated
sludge treatment plants, it has been estimated that the average
power consumption would be 0.113 kwh per capita per day;
yet the energy available in sludge (8) is 0.154 kwh per capita
per  day.  We  must begin  investigating  the  possibility  of
automating sludge-handling so  as to achieve cheaper energy for
plant operation. Data obtained from the  U. S.  EPA's contract
with Raytheon  Company indicate  that  instrumentation and
control schemes for automated sludge processes (those for pH,
temperature, and  methane  production, for  example) have a
pay-back  period of 0.94 years for a 10-mgd plant, and 0.65
years for a 100-mgd plant. This is only a beginning.
   The  problem is serious.  Consider, for example, the Min-
neapolis-St.  Paul Metropolitan Sewerage District,  which has
received an  ultimatum  from  the  local utility;  this utility
threatens  to deny the District  any natural gas by 1978. In the
intervening years, the District's gas supply will be progressively
cut back.  The District is now seriously  looking at all forms of
sludge-handling  to  optimize energy production and  reduce
energy use. With the energy crisis that we face today, there are
                                                         117

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
no assurances that other plants will not eventually suffer the
same fate. Automation could play a vital role in this area.

PRESENT NEEDS
   In spite of the work that has been done in automation at
the  National Environmental Research  Center, there are  still
many research needs that must be answered. For example, we
still need accurate mathematical models,  especially models
describing  sludge characteristics  and the performance  of
sludge-handling and removal  processes. New  or  improved
models for sludge handling, removal, and disposal should ease
the  task  of formulating  and evaluating the corresponding
control strategies  for such sludge processes.
   The  mechanics  of implementing  automatic control are
reasonably well established, and the Renton, Washington study
demonstrated  the benefits of DO control. This technology
must be  documented and  promulgated  to encourage future
plants  to install  automatic DO  control and to  encourage
current plants to  upgrade  their  operation  by adding  DO
control. Meanwhile, research should  be conducted to investi-
gate the  effects of DO setpoint on  effluent quality, sludge-
handling characteristics, optimum  DO  setpoint,  proper loca-
tion, and immersion depth for the DO probe, etc.
   We  must utilize instruments  and automation for  energy
conservation and  generation. At  the Raymond Clayton Plant
in Atlanta there is a methane monitor for manual control of a
digester. The feasibility of automating a control loop to utilize
such a monitor for better and  more reliable production of
methane should also be investigated because sludge digesters
that function properly are a valuable energy source. We must
develop monitors and alarms that  alert the operator to toxic
substances entering either the plant or the anaerobic digester.
Alternative modes of treatment must also be considered in the
event that toxic  materials still manage to enter  the plant in
spite of any precautionary measures taken to prevent this from
occurring.
   Further energy savings can be made by instrumenting the
dewatering of sludge before incineration.  Instrumentation to
measure sludge  characteristics  is essential to  developing  an
automatic control loop for polymer addition before vacuum
filtration or centrifugation.
   Finally, we must help  engineering firms and small munici-
palities by offering typical instrumentation specifications that
plant designers can readily incorporate in their process designs
and for which the city planners could easily solicit bids.
   Design manuals for automated  control loops  should  be
published to enhance  the state-of-art  and encourage imple-
mentation.
   The U. S. EPA  is doing work in all these  areas, but  our
resources  are meager in comparison to the job that we know
must be done. It is our hope that this workshop will provide
the correct priorities, generate new ideas, and serve as a vehicle
of communication.
 REFERENCES
 1.   Roethlisberger, F. J., and Dickson. W. J., "Management and the
     Worker." Harvard University Press. Cambridge, MA (1939).
 2.   Roesler, J. F., "Plant Performance Using Dissolved Oxygen Con-
     trol," Jour. Environ. Eng. Div.. 100. 1069 (1974).
 3.   Petersack, J. F. and Smith. R. G., "Full-Scale Demonstration of
     Advanced Automatic Control Strategies for the Activated  Sludge
     Process," Final  Report  submitted to the U. S. Environmental
     Protection Agency, July 1974.
 4.   Smith, R., "Wastewater  Treatment Plant Control," Presented at
     the Joint Automatic Control Conference, Washington University.
     St. Louis, MO (1971).
 5.   Convery, J.J., Roesler. J.F. and Wise, R. H., "Automation and
     Control of Physical-Chemical Treatment for Municipal  Waste-
     water," Proc. Conf. Applications of New Concepts of Physical-
     Chemical Wastewater Treatment. Vanderbilt University (1972).
 6.   Roesler, J. F., "Factors to Consider in the Selection of a Control
     Strategy," Proc. Second U. S.-Japan Conference on Sewage Treat-
     ment Technology, Cincinnati. OH and Washington, DC (1972).
 7.   Roesler, J. F., Internal Memorandum, November 1971.
 8.   Anon., "Shortages:  Water Pollution  Control  Facilities Face the
     Crisis," Jour. Water Poll. Control Fed., 46,621  (1974).
                     FIELD  EVALUATION  OF  THE  EFFECTIVENESS
                                           OF AUTOMATION
                                                  Allen E. Molvar
                   Environmental R&D Director, Raytheon Company, Portsmouth, RI 02871
INTRODUCTION
   From  successful  experiences  with instruments and au-
tomatic  control  devices,  the  automation of both wet- and
dry-weather wastewater  treatment  plants offers these  well-
known and widely published potential advantages:
   Improved process performance
   Reduced energy consumption
   Reduced chemical consumption
   Reduced sludge production
   Reduced equipment size
   Reduced operating labor
   Reduced maintenance requirements
   Increased equipment life.
   Most of the available literature, however, fails to point out
the potential costs and  maintenance requirements that are
characteristic of the  proposed instrumentation. Instrument
loops usually include measuring or  sensing  elements, signal
                                                         118

-------
                                                                   EVALUATION OF THE EFFECTIVENESS OF AUTOMATION
transmitting  devices, display  elements, controllers,  and final
control elements. Accordingly,  this equipment  must be pur-
chased,  installed,  checked out,  and  properly maintained.
Although some  on-line instrumentation may be essential for
process operation or mandated  by regulatory agencies, most
on-line instruments  and all automatic control devices used in
wastewater treatment projects are optional; they are installed
to effect savings. In order to select intelligently the instrumen-
tation and automatic control devices that should  be installed, a
single touchstone must be met: the potential  benefits must
offset the added costs  and maintenance  burdens.  Figure  1
displays pictorially  the  factors involved  in  a cost-benefit
analysis that is used as a decision-making aid.
   To assess  the effectiveness  of  currently available instru-
ments and automatic control devices under field conditions
and  to accumulate the  baseline  information  necessary for
pragmatic  cost-benefit analyses, the United States Environ-
mental Protection Agency sponsored a comprehensive study of
the state-of-the-art. These analyses were intended to determine
the type and degree of  automation that are  best used in
wastewater treatment facilities. As  part of this project, a team
of engineers surveyed 50  selected municipal and  industrial
wastewater treatment facilities, located throughout the United
States (as shown in Tables  1  and 2). These plants practiced a
wide  array  of  processes.  Although  the  majority  of those
surveyed were  dry-weather or combined-treatment facilities,
some storm-water treatment  plants and control centers were
also examined.
      TABLE 1. TYPES OF FACILITIES SURVEYED
  Type of Facility

  Primary treatment plants
  Secondary treatment plants
  AWT treatment plants
  Storm-water detention facilities
  Computer data centers
  Industrial waste treatment
  Pilot plants
Number Visited

       9
      25
       3
       3
       5
       2
       3
    REDUCED EQUIPMENT SIZE
1
\
PERFORMANCE
niUCIIAflDTiniVI
LJIMouiyir 1 IUIM
CONSUMPTION
iODUCTION
TSIZE
G LABOR
iNCE REQUIREMENTS
NT LIFE







>„
^
~
(
, 1
/
EQUIPMEP
INSTALLA
STARTUP
ENGINEEF
MAINTEN;
LIFE EXPE
INTEREST


\
H-rr+
NET SAVINGS
COST/BENEFIT ANALYSIS
VIA TOTAL ANNUALIZED COSTS
 Figure 1. Cost/Benefit Analysis
                                                         119

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
TABLE 2.
REGIONAL LOCATIONS
EPA Region
1
2
3
4
5
6
8
9
10

OF PLANTS SURVEYED
Number Visited
2
4
4
5
16
2
1
10
6
 METHODS
   Prior  to the  on-site  inspections,  the survey  engineers
 attended a  2-day orientation session during which the type of
 measuring and sensing devices that might be encountered, as
well as the standardization of all  survey  reports, question-
naires, and drawings, was discussed. Extensive questionnaires
(Figures 2,  3,  and 4),  which detailed pertinent  background
information, instrument performance and  cost,  and control
loop experiences, were prepared in advance.
   At the start of each facility visit, the survey engineer met
with  plant  management and  those persons responsible  for
instrumentation. Plant histories, design flow rates, and opera-
tional characteristics  were  discussed at these meetings, with
special emphasis placed  on  the overall benefits or liabilities of
the  installed  instrumentation. This  information was then
documented on the General Survey Questionnaire (Figure 2).
A plant tour (with the facility's instrument  engineer usually
functioning  as  guide)  permitted  the  survey  engineer to
examine the operating instruments and control loops item-by-
item. During this tour, measuring devices were inspected, and
pertinent  data  (including the manufacturer, model  number,
maintenance characteristics, accuracy, and application) were
recorded  on the Instrument Survey Form (Figure 3). More-

if
i!
i








BC






0


















3
|

i <
w °
\\
S£
«
£














i

>-
|
1
s
l
t
i ?
i I
s &

| |§
1 | |
i ^
; £ 1
If! i
: I s =
z ; £ 1
I 111! 1 1!














s
^
1

i
|





a.
l»,l
fill





1

4





s
i



I
j


i


j j
i
1






I
i

I

8
5 i-
§ I
"j
|
-C

j
1
i 1
u
1 1
1
i
II 1 i















i
£
3




1
i

Konp
lollwin DtacHpllon
Camptuat COM:






















j
I

Central Control
SupMTtwry Control:
Alarm and Sataty Syiiamr
Automatic Enwrfincy Program!













4
^c
1 ?
i!








Malnlananca and Callbtallon
Spacial Eqvlpmanl:
Spaclal Oparalor Tralnlnl
Tolal In-Flanl Man Howi/Yaar
Tolal Coal of OulUda 5ar-all Icnallu ol Ini



























Figure 2. General Survey Questionnaire (Sample Form)
                                                          120

-------
                                                                                                 INSTRUMENT SURVEY FORM
Instrument
Parameter







Manufacturer







Model Number







Equipment Cost







Operating Experience
In-Plam Maintenance
(mh/yr)







Maintenance Frequency
(no./mo.)







Special Training







Service by Contract
($ or mh/yr)







On-Demand Service
(S or mh/yr)







6
_E
|
f,
I
V
u.







Total Downtime







Downtime Frequency
(no./mo.)







Problems*







n
3
o
o







Equipment
Auxiliary Devices**







Recording Devices***







Comments







 "B
t
i
I
VS
c
 S1
                                                                                                                                                                                                                          5
                                                                                                                                                                                                                          >
                                                                                                                                                                                                                          o
                                                                                                                                                                                                                          I
                                                                                                                                                                                                                          m
                                                                                                                                                                                                                          m
                                                                                                                                                                                                                          §
                                                                                                                                                                                                                          m
                                                                                                                                                                                                                          m
                                                                                                                                                                                                                          O
                                                                                                                                                                                                                          H
                                                                                                                                                                                                                          O
*    Corrosion, fouling, etc.
**   Limkers, alarms, ratio relays.
***  Local and central.

-------
                                                                                         LOOP AND raOCESS CONTROL SURVEY FORM
4

I1
s!







Control Techniques
!
]







!
•s
1







•







Type of Controller**







1







*
*
•
u
1







Estimited Rnfunc
TimediM.)







Remflu
Annul COM Si«lngi
S
j







UlBily (kW-lu/yr)







s







Pmctti Impfovtmtfil
Increase Removal (?)







Parameter Variance
min/max (mf/M







Operating Experience
Maintenance & Calibration
by In-flant Personnel
(Sor mh/yr)







y.
ii
1!







!
M







Service by Contract
(S or mh/yrl







On Demand Service *
< Sor mh/yr)







Downtime (hrs/yr)







Downtime Frequency
(no./mo.)







Comments







    '    Conlrol mode:  relay, proporlionil, propoillonil plui rcxl, tic.
    ••   Typei of conlrolleri:  iniloi (pnc., hyd. or clec. medli); computer (iiiperviiory, direct dlgllil or «l iniloi).
    •••  Finilconlrol element: pne. nluei, virlihli ipeed pump, etc.

-------
                                                                   EVALUATION OF THE EFFECTIVENESS OF AUTOMATION
over, the survey engineer examined the control techniques,
costs, benefits derived, and operating experiences. His observa-
tions were recorded on the Loop and Process Control Survey
Form  (Figure  4).  In order to  coordinate  the accumulated
information with respect to in-plant applications, applicable
instrument  diagrams were constructed  using  standard ISA
symbols.

RESULTS AND CONCLUSIONS

Measuring Devices
   Unreliable  sensors accounted for most  of the difficulties
experienced  with  automatic  measurement  and  control  in
wastewater projects. The  accumulated  instrument operating
experiences, summarized in Table 3, clearly show that waste-
water instruments require more maintenance than their indus-
trial counterparts. Since most measuring devices in wastewater
service interface directly with raw  sewage, mixed liquors, or
thickened  sludge,  these  devices  are  subject  to continued
fouling  from solids deposition,  slime buildup,  and precipita-
tion.  Accordingly, they  need  more frequent cleaning and
calibration.  Poor mechanical  reliability, interferences, and  a
lack of established measuring principles are also responsible for
the unsatisfactory state of some analytical sensors.
   The distribution of measuring devices (Figure 5)  indicates
that flow  and  level  devices  account for  nearly  half  the
instrumentation employed in treatment facilities. Analytical
instruments represent approximately a  quarter of the instru-
ments  observed,  and  position, speed,  weight,  and  other
mechanical-type measurements add up to about  15%.
   The  following  measuring  instruments possess sufficient
reliability for on-line use in wastewater treatment facilities and
are commercially   available:  level,  flow rate, temperature,
pressure, speed, weight, position, conductivity, rainfall, turbid-


VARIABLE

LEVEL



FLOW




DENSITY

ANALYSIS







MISC.






CONTROL







* Estimated

INSTRUMENT

Bubbler
d/p Trans.
Float & 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 & 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
AH
Engines, etc.
Sludge or C12
Sloice Gates
Liquid Streams
Storm Waters
Wells & Basins
Al! 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
O.G
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* 4
4 4
- -
60* —
30* —
20 -
50* -








SKILL

1
3
1
2
3
3
4
4
4
(3)
3
4
4
4
4
3
4
3
5
3
3
4
4
3
2
3
3
3
3
3
3

4


RELIAB.
(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.
.2-1 yr.
. 1-1 mo.
.5-2 yrs.
.1-5 yrs.
.6-5 yrs.
.6-2 yrs.
.1-1 yr.
.1-1 yr.
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.
1 yr.
4 yrs.
12 yrs.

NOTE :
STP = TREATMENT PLANT

IND = INDUSTRIAL.

SEE TEXT

** d/p Converter only
  TiHe 3. Instrument Performance
                                                          123

-------
 AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
 Figure 5. Distribution of Measuring Instruments Observed During
         User Survey
 ity, pH, residual chlorine, free chlorine gas, and fire hazardous
 (flammable) gas. This is based on actual field experiences in
 the surveyed facilities (summarized in Figure 6).
    Sludge  density  meters,  sludge  blanket   level  detectors,
 on-line respirometers,  dissolved oxygen probes, and  many
 automatic sampling systems use well-established principles that
 are suitable for wastewater monitoring and  control activities
 but  require  so much   maintenance that  many  users are
 dissatisfied  with  them.  These instruments  need  improved
 maintenance characteristics before they will become widely
 used.
    In spite of  the many successful flow-measuring devices
 observed  in the treatment plants, accurate and reliable flow
 rate monitoring for storm water does pose special problems.
 Highly transient flows, large operating ranges, high suspended
 solids, and frequent collisions with large debris are only some
 of the obstacles that  an  acceptable storm-water  in-sewer
 flowmeter must overcome. Consequently, a  suitable storm-
 water flowmeter needs to be developed that will produce the
 accurate flow rate data required for sewer regulation.
 Automatic Control  Loops
   As shown  in  the summary of automatic  control devices
 (Figure  7),  most  facilities  successfully practice automatic
 liquid level, liquid  flow  rate, and  air flow rate control since
 fluid  regulation  is  important  for  proper  operation, and
 satisfactory flowmeters are readily available. The flow control
 systems that are presently available use established designs that
 are entirely adequate for wastewater treatment activities.
   Process  control,  however,  is used only  occasionally  in
 wastewater treatment. The nationwide survey (summarized  in
 Figure 7) found  that flow-ratio chemical addition, feedback
residual chlorine, and digester temperature  control systems
 worked well and caused no difficulties. Most plant managers
 considered these automatic control systems cost-effective since
 they  save  both  energy  and  chemicals, and improve  plant
 operation.  Automatic  feedback  dissolved   oxygen  control
 systems effectively  reduced oxygenation power consumption,
 but some users complained that  these systems require  con-
 siderable probe maintenance.  The turbidity and pH control
 systems observed during this survey were inadequate and gave
 unsatisfactory performance  because of faulty system design
 and installation. Some of the process control parameters (such
 as substrate concentration. MLVSS. and food/microorganism
 ratios) that may be the most potentially useful have not been
 successfully monitored and controlled in wastewater treatment
 plants.
    Although  a  collection  of detailed capital costs, operating
 improvements,  and maintenance data  was one of the prime
 objectives,  fewer than 30% of  the surveyed  plants  had this
 information.  In all cases, the necessity and cost savings for
 automatic flow rate, liquid level, and temperature  control
 systems were readily apparent. The cost benefits of automatic-
 residual chlorine, dissolved oxygen, chemical addition via flow
 pacing,  and  pH,  were somewhat more  difficult  to assess
 because of  the lack of  data. Based  on the survey team's
 judgement and  experiences,the range of potential cost benefits
 was estimated  for these automatic control  systems. Table 4
 highlights the control strategies, confidence, advantages, limi-
 tations, and  recommendations.  Most  plant  superintendents
 reported that any operating labor savings due to automatic
 control  were offset  by  the  added  maintenance  burdens.
 Accordingly,  most automatic process control systems should
 be evaluated on their chemical and utility savings capabilities.

 Central Control
    Central control organizes the  plant operation so that all
 important  events, alarms, and  treatment  information  are
 displayed,  indicated, and  recorded in  a centralized location
 (usually referred to as the Control Room). In addition, most
 central facilities practice automatic or  remote manual actua-
 tion of final control elements. The success of central control is
 ensured by the commercial availability of reliable transmitters,
 displays, and indicating and recording equipment. Virtually all
 the facilities  surveyed successfully utilized a high degree of
 centralized  control.  Since centralized control  reduces  the
 number of men  required to operate a large treatment plant, it
 is  one of the few forms of  instrumentation that are  readily
justifiable on a labor savings basis.

 Computers
   Modern data-logging systems  accumulate, format,  record,
 and display large quantities of data effectively; consequently,
 most  new plants have automatic data acquisition systems.
 Approximately 20% of the visited facilities used  data-logging
 computers, and  90% of these users were satisfied with their
 automatic data acquisition systems.
                                                         124

-------
                                                        EVALUATION OF THE EFFECTIVENESS OF AUTOMATION
                             1C
                                       .:
                                                 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
              [""] UNSATISFACTORY
                                        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

                                                 RESIDUAL CHLORINE
                                        OTHER ANALYTICAL SENSORS

                                         ^]      GAS MONITORS

                                                 SAMPLING SYSTEMS
FAIR
n
SATISFACTORY
Figure 6. Performance Summary for Measuring Devices in Wastewater Treatment Service
                                                125

-------
          AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                           50 FACILITIES
       1972-3
                                              10
   45
20
25
                 NO. OF CASES
30
35
                                                       LIQUID LEVEL CONTROL
                                  LIQUID FLOW RATE CONTROL
                                                   SLUDGE PUMPING
                                       AIR FLOW RATE
                                 CHEMICAL ADDITION
                              RESIDUAL CHLORINE
                        L3
DISSOLVED OXYGEN

PH

TURBIDITY

AUTOMATIC SCUM REMOVAL
                                              J        AUTOMATIC DATA,

                                                        SUPERVISORY COMPUTER CONTROL

                                                        DIRECT PROCESS CONTROL BY DIGITAL COMPUTER
n
                                     UNSATISFACTORY
                FAIR
          | _ |
    SATISFACTORY
           Figure 7. Performance Summary for Automatic Control Loops in Wastewater
                  Treatment Service
             Although process and supervisory control computers have
          demonstrated their merits in many  industries, they are not
          well established in dry-weather treatment plants. Only two of
          the surveyed facilities had process control computers, whereas
          three  storm-water control centers used  computerized  super-
          visory  control. All of these computer systems worked well.
          Computerization of dry-weather  facilities is still in its infancy.
          and not enough operating experience has been accumulated to
          assess its desirability in wastewater treatment projects.
             Computerized   supervisory control  of large storm  and
          combined sewer  systems is cost-effective because  the  vast
          number  of variables  and  control  points  exceeds  human
          computational and decision-making  ability  within corrective
          time limits.
          Maintenance
             In spite of the  low instrument usage rates, wastewater
          treatment  personnel exhibited a good attitude toward instru-
          mentation, as  measured  by  their  willingness to use  and
          maintain the installed  instruments.  The survey team  found
          that the treatment plants supplied approximately 90rf of the
          maintenance resources  needed. Small abandonment rates also
          attest  to  the  favorable  attitude  of  wastewater  treatment
          personnel. Individual plant  managers'  disposition  toward in-
          strumentation, however, ranged from poor to excellent.
             As a group, satellite storm-water treatment facilities  sup-
          plied less  than adequate maintenance.  Possibly because of its
          newness,  storm-water  instrument maintenance is not  well
          understood. Since none of the satellite facilities started up or
                                                               126

-------
                                                                  EVALUATION OF THE EFFECTIVENESS OF AUTOMATION
Table 4. Automatic Control Benefits
Description
feedback compound
residual chlorine
control
Feedback dissolved
oxygen control
pH
Chemical addition via
flow pacing
Settled sludge removal
from clarifiers based
on feedback density
measurements
Potential Savings
25 to 50% savings in Cl
10 to 40% savings in
aeration power
consumption
Essential for pH adjust-
ment processes
15 to 30% chemical
savings
20 to 50% reduction of
sludge volume
pumped
Confidence
High
Moderate
Fair
High
Fair
Basis
Field observations and
engineering judgement
Field observations

Engineering judgement
Engineering judgement
Advantages
Produces good residual
chlorine control;
responds rapidly to flow
changes; highly stable
Energy -effective control
system; responds to
actual oxygen demand
Keeps pH with a reason-
able range of desired
point
Simple, reliable, cost-
effective control sys-
tem; responds to flow
rate variations; stable
control strategy
Produces a dense sludge
with minimum amount
of wate r
Limitation
Residual chlorine ana-
lyzer requires con-
siderable maintenance
DO probes require a
moderate amount of
maintenance; may be
difficult to apply to
some plug flow systems
pH probes tend to foul
in wustewiiter service
Does not respond to
strength variations
Density analyzers
require frequent repair
and maintenance
Recommendations
Suitable for most medium
and larger sized plants
Suitable for most plants,
especially completely
mixed aeration systems
Advisable to use ultra-
sonic cleaning
Suitable for service
wherever demand per
unit volume remains
constant
Useful in large plants
where downstream
sludge processes are
sensitive to water
content
 shut down automatically, it would behoove individuals con-
 cerned with storm-water treatment  facilities  to  direct more
 attention to instruments and automatic devices and to their
 maintenance. On the other hand, storm-water control centers.
 which  typically receive storm-water  and  combined  sewer
 network  information, were  well maintained  and  operated
 satisfactorily.
   Although most plants have reasonably well-qualified instru-
 ment maintenance staffs, any plans that call for  the addition
 of sophisticated instruments and automatic control devices
 must provide for upgrading the staff's qualifications.

 Instrument Budgets
   The nationwide  survey  of 50 wastewater  facilities  found
 that most of the treatment plants used fewer instruments and
 automatic  control  devices than the closely related  water
 supply and chemical processing plants. The  amassed cost data
 shows that the average secondary wastewater treatment plant
 spends  about  3%  of  its  construction  costs  for installed
 instruments, whereas water supply  and chemical  processing
 plants allocate about 6%  and 8% respectively,  for installed
 instruments. Remote satellite wet-weather  treatment plants,
 which in theory should operate unattended or with just a
 minimal  amount of operating manpower, budgeted only about
 2% for instrumentation and automation.
    In all fairness,  it must be  pointed out  that  the larger
 instrument budgets of the chemical and water supply indus-
 tries are due principally to the higher usage rates of physical-
 type measurement and control systems (that  is,  temperature,
 tevel, pressure, and flow). Most of the analytical sensors that
 the wastewater  treatment  field  needs  must  detect  trace
 amounts of  a specific  substance  from   a  multi-substance
 mixture. These analyzers are  challenged by a difficult  assign-
 ment and should be viewed in some respects as being similar to
 the early stages of the on-line gas chromatographs, which took
10 years and hundreds of thousands of dollars to develop.

RECOMMENDATIONS
   In summary, the field survey team assessed the instrument
utilization rates  and performances, and estimated the special
manpower skills, training, and equipment necessary to operate
and  maintain  instruments and automatic  control devices in
wastewater  treatment  service. Wherever available, the total
control system costs, as well as the economic and performance
benefits obtained, were also  noted. With  this information a
cost-benefit  analysis can be performed to  decide  whether or
not  to use  a  particular instrumentation  or control system.
Moreover, the commercially  available on-line analyzers and
automatic control devices were classified, on  the basis of the
field  experiences, as  being:  1) unacceptable,  2) fair, and
3) satisfactory.
   The small number of automatic loops observed in the plant
survey attests  to the low level of automation that is charac-
teristic of  most  wastewater  treatment plants. The survey's
observations   indicate  that  a lack of  sufficiently reliable
analytical sensors for automatic control has impeded process
control efforts. Other  commercially available process control
components (such as transmitters, display devices, controllers,
and final control elements) have proved their ability to provide
reliable service in wastewater treatment plants. To help assess
an automatic  control loop's desirability, a uniform and easily
practiced record keeping system is badly needed. Also, much
misunderstanding and  confusion can be avoided in the future
by using standard instrument symbols and drawings.
   An intensive  application  of elaborate  and  novel  logic
schemes, computers, displays, and recorders will not improve
wastewater treatment effectiveness. Instead, well-documented
field  evaluation  programs  are  needed  to  help  ferret out
desirable control systems from the numerous potentially viable
ones.
                                                          127

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                              DISCUSSION
Robert A. Ryder:
   DO control has been  utilized at the Reno WWTP since
1966, and at the present time is set to maintain DO = 0.5 mg/1
at the end of the aeration tanks. This has been found to be
highly effective in maintaining a low SVI, preventing bulking,
and producing a good  effluent. There  is implicit indication
that this is  due to ORP suppression of the sludge, inhibiting
certain strict-aerobic filamentous organisms.
   Recent experience with a pilot plant of Phostrip with 10-14
hours of quiescent sludge settling, further indicates  great
stability of the sludge, low SVI's and excellent BOD removal.
   Scum has been noted at many activated sludge plants, and
seems to be a function of very long sludge age or SRT. More
research is needed on this problem.

N. J. Biscan:
   Mr. Roesler mentioned the need for  monitors and alarms
which alert  the operator to  toxic substances entering  the
treatment facility. I'd like to point out  that we at Dow have
developed such an instrument under EPA Grant No. S800 766,
"Optimizing  a  Petrochemical Waste  Bio-oxidation  System
Through Automation."
   The instrument operates on a one-hour time cycle and is
based on a measure of the oxygen  consumed in the oxidation
of the organic  substrate in the samples added. Briefly,  the
instrument  operates by first  determining  the  area under an
oxygen-uptake curve of  a sample  of standard biodegradable
substrate, then a feed  sample, and finally the standard sample
again. The ratio of the standard areas gives the indication of
toxicity. Questions that remain to  be answered, however, are
the following:
 1. Is there in fact a significant problem with toxicity in feeds
   to municipal  activated  sludge  plants,  and what is  the
   potential demand for a toxicity indicator?
2. How can  instrument  companies market an instrument,
   which was developed under an EPA grant, without  the
   company obtaining patents or licensing rights?

Edmond P. Lomasney:
   I should like to comment on the  subject of utilizing gaseous
products that are generated in the anaerobic process.
   Recently  completed  research  has  indicated  that  the
dynamics of the anaerobic process in sewage treatment have
been  incorrectly interpreted. The  basic concepts of the  be-
havior of the bacteria  involved  indicates that those  bacteria
which participate also manufacture combustible gases. These
gases, methane and hydrogen, are  products generated in this
process, and the efficiency of the process is indirectly related
to the ratios of these gases  (excess hydrogen  acting as an
inhibitor to the performance  of the methanogenic bacteria).
As  a consequence,  a  gaseous monitoring  procedure  for
detecting and measuring these gases is a necessity  for  the
continuous  efficient   and  effective  performance   of  the
anaerobic process.

Ronald N. Doty:
   Much work has been done to evaluate available instruments
and  to  determine what is needed. However,  we seem to be
experiencing a dichotomy which will negate much ot the R &
D and evaluation work which has, and  is continuing to be
done: the construction grants division  of EPA. and the States.
force us to write  open specifications, thus precluding  us from
specifying that equipment  which has been proven superior.

L. A. Schafer:
   Although setting standards for wastewater instrumentation
can be restrictive, some standards (similar, for instance, to fire
or electrical regulations) could well be applied, subject to the
following rules:
 1. The  standard-setting body should include both theoretical
   and practical people.
2. Standards should be firm, but with an autonomous applica-
   tion   assessing  procedure for  evaluation  of results and
   provisions for exceptions.
3. Standards should be proposed  and discussed by  a cross-
   section of users before adoption.
Some standards  which  might profitably  be  adopted  are:
standard symbols, standard  classification  of facility plans by
disciplines (Civil, Structural, Architectural. Mechanical,  Elec-
trical, Instrument  (or Control)  and General), a  standard
selection of  instrument signal ranges (4-20 ma. 10-50 ma.
3—15 psi, 20—80 inches vac., etc.).  a requirement  that  all
facilities above a certain size  contain a signal interface area (Le.
terminal strips)  suitable  for connection  of temporary  com-
puters or data loggers, etc.

Richard R. Keppler:
   In view  of  the  large  amount of  discussion that  has
concerned itself with the proper maintenance of monitoring
equipment utilized  for  wastewater  treatment  systems, my
experience  in the  petroleum refining  industry  might  be
helpful, since a similar problem existed in this industry about
 15 or 20 years ago. It was the custom at  that time to utilize,
for  maintenance  purposes, separate  groups,  such  as the
insrument department, and to have these  various maintenance
groups  respond  to  requests from the  operators  for  main-
tenance  required. This was  ineffective in that the insrument
people,  for  example, would  consider  that  some  of the
problems which  resulted in breakdown  of their equipment
were due  to poor  operator  performance and  likewise the
operators said that an upset on the equipment was a  result of
poor maintenance on the  part of the  instrument people.
Therefore, there  was  a continuous  discussion between the
                                                         128

-------
                                                                   EVALUATION OF THE EFFECTIVENESS OF AUTOMATION
maintenance people and the operating people concerning the
responsibility for the  proper  operation of the  equipment.
Later, in an effort to overcome this problem, it  became the
industry's  pattern to  contract to outside organizations for
maintenance of  equipment  in the operation  of a refinery.
Although  possibly  more expert  and skilled  maintenance
became  available by  this means,  it  did  not  eliminate the
fundamental problem of one man  blaming another for poor
operation.
   In the past several  years there has developed a so-called
"refinery technician"—this man begins as an  operator and
learns additional  mechanical  skills, receiving an increase in pay
for  each additional  skill  that  he  masters.  In  this way the
operator is  also a machinist  and instrument repair man, and
therefore takes great  interest  in proper maintenance  of the
equipment which he is required to operate. It is my suggestion
that this kind of philosophy  be adopted by the municipalities
operating wastewater  treatment  systems,  in  an effort  to
improve the level of maintenance and performance of those
systems. This combined "refinery technician" has resulted in a
significant improvement in efficiency  of refinery operations.

J. F. Andrews:
   The discusser would like to suggest  that in order to obtain
maximum benefit from the automation of wastewater treat-
ment plants it  will be necessary to  significantly upgrade the
quality of personnel in plant operations. One method of doing
this  would  be to involve more  professional engineers  directly
in plant operations. Unfortunately, with the exception of our
large  cities,  the major  portion of environmental engineering
profession has  frequently been  guilty of considering plant
operations as a subprofessional  area not worthy of significant
engineering attention.
   Most  environmental engineering  programs in  universities
have considered wastewater treatment plant operations to be a
low-level task which is not suited for education or research in
universities.  Those few  universities which  have become in-
volved in plant  operations  have  usually  done  so   at  the
technician or  "operator  training"  level by offering  "short
courses" or  correspondence  courses.  These  courses  have
provided a valuable public  service; however, they may also be
partially  responsible  for  the   lack of recognition  of  the
importance of the engineer in treatment plant operations. The
major portion of  the material taught in these courses is at the
technician  and not  the engineering level and would best be
taught at the vocational  high school or two-year  technical
school level.  There is a strong trend in  this direction at the
present time and it is expected that this trend will continue.
   What, then, would be  the major differences between the
traditional  environmental  engineering  curriculum  and a cur-
riculum emphasizing plant operations? There would, perhaps,
be  as  many  different opinions  on this point  as there are
environmental engineering professors; however, in  the discus-
ser's opinion  the  major  differences would be  an increased
emphasis on  the dynamic behavior  of plants and techniques,
such  as  the  application of control systems, to  improve
dynamic  behavior. Most current design criteria, both as taught
in universities and practised  in the field,  are based  on average
inputs or, at best, maximum and minimum inputs  and have
not directly  considered that  inputs to  the plant are highly
variable with respect to time. These variations are the primary
reason why control is needed and they must be considered in
design as well as in operation. It should be noted that a course
in Process Dynamics and Control  is commonly found in most
chemical engineering curricula and,  in the discusser's opinion.
we would be well advised to include a course in  "Dynamics
and Control  of Wastewater  Treatment Systems"  in environ-
mental engineering curricula.
   In  addition  to  providing  educational programs for  plant
operations  engineers, universities should also become  more
involved  in research on plant  operations. Most environmental
engineering  programs have  concentrated on  design-oriented
research  and  have  neglected operations-oriented  research,
perhaps because of the assumption  that this is not worthy of
Ph.D.-level  studies  with  which most  university  research  is
associated.  However, the  discusser's experience has been that
this type of research is more difficult, from both  theoretical
and experimental points of  view, than the usual scientific or
design-oriented research. The discusser has had a total of nine
Ph.D. students who have conducted  their research in either the
dynamics  or  control  of wastewater  treatment   processes.
Although much is still to be learned, since the theory put forth
in these Ph.D. dissertations  has for  the most  part yet to be
tested  in  the field, a beginning has been made toward a
quantitative description of the dynamic behavior of treatment
processes,  and  control  systems  for  improvement  of  this
behavior.
                                                         129

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS
                                   Report  of Working Party
                                                     On
                                         RESEARCH NEEDS
                                                     FOR
         EVALUATION  OF THE  EFFECTIVENESS  OF  AUTOMATION
                                               Walter G. Gilbert
                                      Chief, Municipal Operations Branch,
                           Environmental Protection Agency, Washington, DC 20460
                                               James  A. Mueller
                      Associate Professor, Environmental Engineering & Science Program,
                                     Manhattan College, Bronx, NY 10471
   The success of utilizing instrumentation and automation in
the  operation of  wastewater  treatment  facilities  will  be
measured to a large degree by the capability to produce better
treatment  in a  more  effective and  efficient manner.  It is
important  to note  that the discussions of this  workshop
recognized the fact that automation can mean many different
things, from the simple and reliable instrumentation needed for
the smallest plant to the complex, computer-assisted operation
of our  largest facilities. There will  be many variations in
between, depending upon  the  size,  complexity, and many
other variables of the  facility involved. Measuring  the effec-
tiveness of  automation, therefore, involves  many  different
factors.
PROBLEMS
   In attempts to evaluate the effectiveness of automation for
wastewater treatment, problems  are  encountered  in  three
broad areas as follows: (1) effluent quality improvement, (2)
cost-effectiveness, and (3) human-machine interface. Specific
statements and relative priorities of these problem areas are
discussed more fully below.
1. Determine the  extent of  effluent  quality improvement,
   lower contaminant concentrations and reduced variability
   attainable through the use of instrumentation and automa-
   tion.
     This problem area should receive the highest priority in
   future automation research in keeping with the goals of the
   Federal  Water Quality  Act Amendment  of  1972 (PL
   92-500) for improvement  of our Nation's water quality.
   Regardless of other constraints, if increased automation of
   wastewater treatment  systems reduces effluent concentra-
   tions and variability, its  effectiveness  will  have  been
   demonstrated and future utilization insured.
     In applying  automation to attain the  above goals, the
   interaction of the collection system and treatment plant
   must be considered. Improvement of existing systems by
   increased utilization of automation with  appropriate con-
   trol strategies and required system  modifications must be
   evaluated. New system development incorporating process
   and system theory with required  automation  should be
   encouraged.
2.  Evaluate the cost-effectiveness of incorporating automation
   into wastewater systems which are  required to meet a
   specified performance objective.
     Existing and future  wastewater systems typically  are
   required to meet performance specifications due to either
   administrative or legislative policy or local allocation based
   on receiving water quality objectives. To meet this specified
   performance objective the type  and degree of automation
   that should be employed in a wastewater system to yield
   reduced  system costs must be evaluated. Unless this is
   accomplished, a manager will rarely consider employing
   automation  in his  system  if it is already meeting  per-
   formance objectives with an existing degree of manual
   operation.
     To accurately evaluate the cost-effectiveness of increased
   automation, a uniform set  of  criteria  incorporating all
   economical factors must be  developed. Systems must also
   be evaluated to determine the most cost-effective location
   of instrumentation as well as type.
3.  Evaluate the effectiveness of the human-machine interface.
     If treatment  plant automation is  to  have the desired
   degree of  effectiveness, interactions  of the  wastewater
   system operators, designers, and managers must  be con-
   sidered. Due to  its importance,  this problem area has  the
   same priority as the above cost-effectiveness area. If a plant
   operator lacks the training or desire to effectively utilize and
   maintain instrumentation, the automation system will fail.
   If  a  designer  does  not obtain operator feedback with
   respect  to automation  effectiveness, or long-term data
   feedback in a readily utilizable fashion, the effectiveness of
   automation  will  not be properly evaluated and future
   process and automation improvements will not result. If a
   manager does  not allocate sufficient funds nor personnel to
   adequately maintain and utilize  the automated facilities, a
   significant portion of the capital outlay for the plant will be
   wasted and it will not meet treatment objectives.
     Therefore,  to insure the effectiveness of automation,
   adequate  operator  training  with  data feedback  to both
   designer and managers must be employed.
                                                      130

-------
                                                                   EVALUATION OF THE EFFECTIVENESS OF AUTOMATION
RESEARCH NEEDS
   To respond to the problem areas discussed above, research
needs were identified which must be satisfied if we are to be
able to  fully evaluate the effectiveness of automated systems
for plant operations.  These  needs are  addressed  below in
priority order.
  1. Develop  control strategies to utilize  instrumentation and
    automation in achieving maximum process reliability.
         Control strategies are needed that .can fully utilize the
    information and response capabilities derived  from vary-
    ing degrees  of instrumentation and automation to provide
    the  process control data  necessary for good plant  opera-
    tion. Such strategies must utilize the  short response times
    that would be available through such automation to react to
    the widely  varying extremes  in influent quality, so as to
    produce  a more consistent effluent quality. The develop-
    ment of such strategies must also  recognize  the need for
    varying  degrees of automation for  different  sizes and
    types of plants. Achieving process reliability is an inherent
    part of this  need.
  2. Investigate  the full utilization of surge capacity or  flow
    equalization in maximizing plant  performance.  Consider
    the  use  of in-line  or  off-line storage or control,  with
    determination  of  the optimal  location of such control.
    Determine to what extent treatment facilities can  accept
    varying  loads  and produce  acceptable effluent quality.
    Conduct field studies required for verification.
         It is essential that  feasible  means of  control  over
    influent  quantity and quality  be evaluated,  along  with
    related impacts on plant  operations.  Determination must
    be  made on the degree of control needed to achieve the
    desired  end result. Means of  automating  such control
    systems,  which provide  feedback to  treatment process
    control,  are needed.
  3. Establish a data  collection  system  with  appropriate
    feedback mechanisms to design engineers for continued
    evaluation  of  plant  performance and use  in  improving
    future plant design. The system should be assembled with
     adequate correlation of related data elements for ease of
     retrieval.
         Although  this need  satisfies essentially a long-range
     objective, it is necessary to start it now so as to  fully
     utilize much of the historical data now  being collected.
  4. Develop uniform criteria for evaluation of the cost-effec-
     tiveness of varying degrees and types  of facility automa-
     tion.
         Cost-effectiveness  is  determined  on a  comparative
     basis between alternative systems. Uniformity of  evalua-
     tion is essential to allow  ease of comparison and allow for
     final selection  of  instrumentation and control systems.
     Inherent to this  research need is the identification and
     quantification in economic terms of  the benefits which
     accrue from use  of instrumentation  or automation in
     wastewater collection, treatment, and disposal systems.
5. Identify areas of potential cost saving in plant operations
   to  guide  priorities  in developing  instrumentation  and
   control systems. Determine and  investigate in detail the
   most cost-effective points of  application of instrumenta-
   tion in wastewater treatment processes.
        Such  efforts are  needed  to  provide a  guide in
   determining  the  most  effective  course  of  action in
   developing instrumentation and automation for  any given
   treatment facility.
6. Evaluate the effectiveness of  existing automated systems
   based on operator feedback.
        The operator must be recognized as a key element of
   the total control system.  Such systems must be  evaluated
   with this in mind.  Optimization of  automated systems
   must consider the role of  the operator in the control loop.
7. Investigate upstream monitoring for predictive purposes
   with adequate lead times for process control.
        As well  as  providing information for  immediate
   reaction to influent quality  variations, such  monitoring
   would provide historical data  needed to develop effective
   process  control strategies needed  to respond  to such
   variations. Efforts should also include a determination of
   the  types of monitoring  required, including both quality
   and quantity parameters.
 8. Identify and evaluate available sensors and instruments as
   applied in  wastewater treatment processes in  terms of
   their performance and reliability.
        This relates directly to the need for standard tests and
   specifications identified in earlier workshop groups.
9. Develop  the  research  input  to a  training program to
   permit full utilization of instrumentation and automation.
        A good training program is essential to  develop the
   proper operator  attitude  and  encourage operator  ac-
   ceptance  of automated systems. Such training  must also
   address the technical  details  of maintenance  and calibra-
   tion needs of automation as well as the use of the control
   system to improve performance.
10. Develop new and innovative treatment processes utilizing
   process  theory  in  conjunction  with the capabilities of
   instrumentation and automation.
        This  is conceived as  a need to develop new  treatment
   processes and  methods  based  on  a "no  holds  barred"
   approach.  It  is  essentially  pure   research free  of  the
   constraints of "traditional" design  of treatment facilities.
11. Develop  improved  models  to  more  closely  simulate
    physical, chemical, and biological processes using measura-
   ble parameters.  Utilize these  models for process develop-
    ment, process control, and research requirements.
        Satisfaction  of this  need  will help  accelerate  the
    process  of  evaluating  the  effectiveness  of  alternative
    automation strategies.
PRIORITIES
   Priorities  for  the research needs  identified  above  were
assigned primarily  on the basis  of  the need for immediate,
                                                          131

-------
AUTOMATION OF WASTEWATER TREATMENT SYSTEMS


 short-term answers to certain pressing questions regarding the    can proceed in a more logical sequence. Lower priority was
 role of instrumentation and automation in plant operations. The    assigned to  needs  with longer-range benefits  or where  the
 framework for procedures to evaluate the effectiveness of such    beneficial effects of such efforts may be largely unknown at
 systems must be  established now so that development work    this time.
                                                        132

-------
                                           LIST  OF  PARTICIPANTS
ANDERSON, James J., President
Watermation, Inc.
1404 East 9th Street
Cleveland, OH 44114

ANDREWS, John F., Professor
Environmental Systems Engineering
Clemson University
(present address:
Dept. of Civil Engineering
University of Houston
Houston, TX 7 7004)

AUSTIN, John H.. Department Head
Environmental Systems Engineering
Clemson University
Clemson, SC 29631

BABCOCK, Russell  H., Chief Engineer
C. E. Maguire Inc.
60 First Avenue
Walt ham, MA 02154

BALLOTT1, Elmer F., Partner
Greeley and Hansen Engineers
222 South Riverside Plaza
Chicago, 1L 60606

BARNES, George D.
Assistant Dirctor of Public Works
303 City Hall
Atlanta, GA 30303

BERTHOUEX, Paul Mac, Associate Professor
Civil and Environmental Engineering
University of Wisconsin
Madison, WI 53706

BERTRAM, William H.
Supervisory Sanitary Engineer
Environmental Protection Agency
 1860 Lincoln Street
Denver, CO 80203

 BISCAN, N. J., Research Specialist
 Dow Chemical U.S.A.
 A-l 127 Building
 Freeport,TX 77541

 BLAKELY, Christopher P., Market Manager
 Water Management
 Honeywell, Inc.
 1100 Virginia Drive
 Ft. Washington, PA 19034

 BREIDENBACH, Andrew W., Director
 National Environmental Research Center
 Environmental Protection Agency
 Cincinnati, OH 45 268

 BRUBAKER, Jay H., Section Leader
 Union Carbide Corp.
 P.O. Box 8361
 South Charleston, WV 25303
BRYANT, James O.
Municipal Operations Branch
Environmental Protection Agency
Springfield. VA 22153

BUHR. Heinrich O.
Visiting Associate Professor
Environmental Systems Engineering
Clemson University
(on leave from:
Dept. of Chemical Engineering
University of Cape Town
Rondebosch. CP 7700
Republic of South Africa)

CARKEEK, John G.? Associate
Consoer Townsend & Associates
360 East Grand Avenue
Chicago, IL 60611

CARROLL, Leo J., Manager
Environmental Division
Fischer & Porter Co.
Warminstei. PA 18974

COHEN, Al, President
Astro Ecology Corp.
801 Link Road
League City. TX 77573

CONVERY. John, Director
Advanced Waste Treatment
Research Laboratory
National Environmental Research Center
Environmental Protection Agency
Cincinnati, OH 45268

CROW, Merle
 Stevens T. Mason Building
 Lansing, MI 48926
 DAY, Robert E., Project Manager
 Black, Crow & Eidsness, Inc.
 Penn Towers-Suite 434
 1819 John F. Kennedy Blvd.
 Philadelphia, PA 19103
 DENIT, Jeffery D., Chief
 Impact Analysis Section
 Environmental Protection Agency
 Washington, DC 20402

 DICK, Richard I., Professor
 Department of Civil Engineering
 University of Delaware
 Newark, DE 19711

 DOBBINS, William E., President
 Teetor-Dobbins, P.C.
 515 Johnson Avenue
 Bohemia,  NY 11716
DOTY, Ronald N.
Supt. Water Quality Control
250 Hamilton Street
Palo Alto, C'A 94301

DU CROS, Michael. Marketing Manager
Technicon
Benedict Avenue
Tanytown, NY 10591

ECKENFELDER, Wesley W.. Jr.. Professor
Environmental & Resources Engr.
Vanderbilt University
Nashville, TN 37235

ELSAHRAGTY, Mohammed
Quirk. Lawler & Matusky Engineers
415 Route 303
Tappan. NY 10983

FARRELL, J.B., Chief
Ultimate Disposal Section
Advanced Waste Treatment Research Laboratory
National Environmental Research Center
Environmental Protection Agency
Cincinnati, OH 45268

FERTIK, Harry A., Senior  Scientist
Leeds & Northrup Company
Technical Center
North Wales, PA 19454

FETCH, John J.
Capital Controls Co.
Div. Dart Ind.
Advance Lane
Colmar, PA 18915

FLANAGAN, Michael J.
Brown & Caldwell Consulting Engineers
66 Mint Street
San Francisco. CA 94103

FREEMAN, Mark  P., Senior Scientist
Dorr-Oliver Incorporated
77 Havemeyer Lane
Stamford, CT 06904

GARRISON, Walter E., Assistant Chief Engineer
Sanitation Districts of L. A. County
 1955 Workman Mill Road
Whittier, CA 90601


GILBERT, Walter G., Chief
Municipal Operations Branch
 Environmental Protection  Agency
Washington, DC 20460

OILMAN, Harold D.
Greeley and  Hansen Engineers
 Six Penn Center Plaza
Philadelphia, PA 19103
                                                              133

-------
 GLENN, Don
 Space Division
 General Electric Company
 P. O. Box 8555
 Philadelphia, PA 19101

 GRAEF, Stephen P.,
 Principal Sanitary Engineer
 Metropolitan Sanitary District of
 Greater Chicago
 100 East Erie
 Chicago, IL 60611

 GUARINO, Carmen F., Commissioner
 Philadelphia Water Department
 1160 Municipal Services Building
 Philadelphia, PA 19107

 GULEVICH, Whdimir, Chief
 Water Quality Engineering Division
 U. S. Army Environmental Hygiene Agency
 Aberdeen Proving Ground, MD 21010

 HAYES, Charles H.
 Assistant General Manager
 Gulf Coast Waste Disposal Authority
 910 Bay Area Blvd.
 Houston, TX 77058

 HUMPHREY, Marshall  F.
 Jet Propulsion Laboratory
 4800 Oak Grove Drive
 Pasadena, CA 91103

 JACOBSON, Kenneth J.
 Dept. of Chemical Engineering
 University of Pennsylvania
 Philadelphia, PA 19104

 JOYCE, R. J., Production Manager
 Dohrmann-Envirotech
 3240 Scott Blvd.
 Santa Clara, CA 95050

 KEINATH, Thomas M., Associate Professor
 Environmental Systems Engineering
 Clemson University
 Ciemson, SC 29631

 KEPPLER, Richard R.
 Research and Monitoring
 Environmental Protection Agency
 Boston, MA 02203

KEY, Phi
 FMC Corporation
Denver, CO 80210

 KUGELMAN, Irwin J.,  Chief
 Pilot Plant Field Investigation Program
 National Environmental Research Center
 Environmental Protection Agency
 Cincinnati, OH 45268

KURLAND, Paul
Delta Scientific Corporation
 1172 Route 109
 Lindenhurst, NY 11757

 LAGER, John A., Vice President
 MetcalfA Eddy, Inc.
 1029 Corporation Way
Palo Alto, CA 94303
 LEISER, Curtis P.
 Manager of Computer Services
 Seattle Metro
 410 West Harrison Street
 Seattle, WA 98119

 LEJEUNE, T., Manager
 R. M. Clayton WPC Plant
 2240 Bolton Road, NW
 Atlanta, GA 30318

 LOMASNEY, Edmond P.
 R & D Program Director
 Environmental Protection Agency
 1421  Peachtree Street, NE
 Atlanta, GA 30309

 MANKES, Bob
 Vice President, Sales
 Delta Scientific Corporation
 1172 Rt. 109
 Lindenhurst, NY 11757

 MASTERS, Hugh E.
 Storm & Combined Sewer Section
 Environmental Protection Agency
 Woodbridge Avenue
 Edison,  NJ 08817

 MATHEWS, Michael B.
 Sanitation Superintendent
 City of Ventura
 P. O. Box 99
 Ventura, CA 93001

 MATSON, Jack, Assistant Professor
 Civil Engineering Department
 University of Houston
 Houston, TX 77004

 McKNIGHT, M. Dolan, Plant Superintendent
 Fort Worth Water Dept.
 P. O. Box 870
 Fort Worth, TX 76101

 MEYER, John M.
 Vice President, Marketing
 Systems Control, Inc.
 1810 Page Mifl Road
 Pato Alto, CA 94304


 MILLER, G. Wake, Project Director
 Public Technology. Inc.
 1140 Connecticut Ave., NW
 Suite 804
 Washington, DC 20036


 MOLVAR, Alten  E.
 Environmental R & D Director
 Rat neon Company
 P. O. Box 360
West Main Road
 Portsmouth, RI 02871


 MONTAGUE, Albert, Director
Office of Research & Development
 Region III
 Environmental Protection Agency
 6th and Walnut Streets
 Philadelphia, PA 19106
MOSS, J. E.
Union Carbide
P.O. Box 8361
South Charleston, WV 25 303

MUELLER, James A., Associate Professor
Environmental Engineering & Science Program
Manhattan College
Bronx, NY 10471

NELSON, John K.
Assistant Director Planning Operations
Metro Denver
3100 East 60th Avenue
Commerce City, CO 80022

NORKIS, Charles M., Sanitary Engineer
Philadelphia Water Department
Broad and Arch Streets
Philadelphia, PA 19106

OLSSON, Gustaf, Assistant Professor
Division of Automatic Control
Lund University
Lund 7,  Sweden

OPATKEN, Edward
Municipal Pollution Control Division
Environmental Protection Agency
Washington, DC 20402

PALLESEN, Lars
Dept. of Statistics
University of Wisconsin
Madison, WI 5 3706

PEIL, Kefly M.
CPT/Engineering Branch Chief
USA Med Bioengineering R & D Laboratory
Ft. Detrick
Frederick, MD 21701

PETERSEN, W.Carl
Associate-Instrumentation
Consoer Townsend & Associates
360 Grand Avenue
Chicago, IL 60611

POLTA, Robert C., Research Engineer
Metropolitan Sewer Board
350 Metro Square Building
7th and Robert Streets
St. Paul, MN 55101

QUIGLEY.John
Assistant Professor of Engineering
University of Wisconsin
Madison, WI 5 3706

RADEMACHER, John
Department of Natural Resources
State of Maryland
Annapolis, MD  21401

RICHARD, Dan
Technology Applications Programs
NASA Headquarters
Washington, DC 20007

RISLEY, Clifford
Region V
Environmental Protection Agency
Chicago, IL 60606
                                                             134

-------
ROESLER, Joseph F.
Advanced Waste Treatment Research Laboratory
National Environmental Research Center
Environmental Protection Agency
Cincinnati, OH 45 268

ROSENKRANZ, William, A., Director
Municipal Pollution Control Division
Environmental Protection Agency
Washington, DC 20460

ROTH, John A., Chairman
Division of Chemical, Fluid and
Thermal Sciences
Vanderbilt University
Nashville, TN 37235

RYCKMAN, D. W., President
Ryckman, Edgerley, Tomlinson & Associates
12161 Lackland Road
St. Louis, MO 63141

RYDER, Robert A., Vice President
Kennedy Engineers, Inc.
657 Howard Street
San Francisco, CA 94105

SAKO, Frank F., Staff Engineer
 FMC Corporation
 1185 Coleman Avenue
 Santa Clara, CA 9505 2

 SALLOUM, J. Duane, Director
Wastewater Technology Centre
 Environment Canada
 Ottawa, Ontario KIA OH3
 CANADA

 SCHAFER, Lawrence A.
 Principal Engineer
 C. E. Maguire, Inc.
 60 First Avenue
 Waltham, MA02154

 SCHUK, Walter W.
 Pilot Plant
 Environmental Protection Agency
 5000 Overlook Avenue
 Washington, DC 20032

 SMITH, John M., Chief
 Municipal Treatment and Reuse
 National Environmental Research Center
  Environmental Protection Agency
  Cincinnati, OH 45268

  SMITH, Wayne
  Process Control Branch
  National Field Investigations Center
  Environmental Protection Agency
  Denver, CO 80225

  SORENSEN, Poul Erik, Chemical Engineer
  Water Quality Research Institute
  Academy of Technical Sciences
  Soborg, Denmark
   STAMBERG.J.
   Office of Research & Monitoring
   Environmental Protection Agency
   Washington, DC 20460
STENSTROM, Michael K.
Environmental Systems Engineering
Clemson University
Clemson, SC 29631

STEPHENS, Ed, Associate
Clinton Bogert Associates
2125 Center Avenue
Fort Lee, NJ 07024

SWEENEY, Robert F.
Associate Professor
Dept. of Chemical Engineering
Villanova University
Villanova, PA 19085

TAYLOR, Reuben
Water & Wastewater Systems
Urban Systems Project Office
NASA, Johnson Space Center
Houston, TX 7705 8

TOBIN, Patrick
Municipal Pollution Control Division
Environmental Protection Agency
Washington, DC 20460

TOMCZYK, Harry
 Senior Electrical Engineer
 Metropolitan Sanitary District of Chicago
 100 East Erie
 Chicago, IL60611

 TORNO, Harry
 Office of Research & Monitoring
 Environmental Protection Agency
 Washington, DC 20460

 TRAX, John
 Office of Water Programs
 Environmental Protection Agency
 Washington, DC 20460

 TREUPEL.HansW.
 Senior Physicist
 IBM Corporation
 18100 Frederick Pike
 Gaithersburg, MD  20760
  WEST, A. W.
  Office of Enforcement and General Counsel
  National Field Investigations Center
  Environmental Protection Agency
  Cincinnati, OH 45 268

  WILKINS, Judd R., Microbiologist
  NASA
  Langley Research Center
  Hampton, VA 23665

  WISE, Robert H., Research Chemist
  National Environmental Research Center
  Environmental Protection Agency
  Cincinnati, OH 45 268

  WOODRUFF, Paul H., President
  Roy F. Weston, Inc.
  WestonWay
  West Chester, PA 19380
WRIGHT, Darwin R.
Municipal Pollution Control Division
Environmental Protection Agency
Washington, DC 20460

YOUNG, Pet erC.
Centre for Resource and Environmental Studies
Australian National University
Canberra, Australia

Z1CKEFOOSE, Charles S.
Operation Consultant
Stevens Thompson & Runyan, Inc.
5505 S. E. Milwaukie
Portland, OR  97202
                                                                135

-------