MEMORANDUM TO D/DCI/NIPE FROM (Sanitized)
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP68R00530A000200040021-9
Release Decision:
RIPPUB
Original Classification:
S
Document Page Count:
15
Document Creation Date:
December 14, 2016
Document Release Date:
October 9, 2002
Sequence Number:
21
Case Number:
Publication Date:
December 2, 1965
Content Type:
MF
File:
Attachment | Size |
---|---|
![]() | 1.34 MB |
Body:
v;LJiii;1
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
DD/S&T 5630-65
2 December 1965
[v~EMQRANDUiwt OR D/DCI/NYPE
The attached monograph briefly describes some of the
potential opportunities and problems for the Intelligence Community
of the future.
25X1A
Attachment:
EXCIUGcB tram aute,natl
do~rnradln; and
Approved For Release 2 CIA-RDP68R'.OA00 0020 040021-9
Approved For Release 2002/10/29 CIA-RDP68R0053OA000200040021-9
A DESCRIPTION OF SOME OPPORTUNITIES
AND PROBLEMS IN INTELLIGENCE
Opportunities
1. Exploratory research and development can now be profitably
and effectively pursued toward the objective of establishing multi-
sensor collection capabilities on a world-wide basis. Such a
capability would allow most overt activities to be monitored to
whatever degree may be desired. There is now significant activity
in this direction, and it is clear that satellites and versions of the
MOL, etc. , could provide platforms to carry sensors or to corn
municate with sensors on or near earth. Activities which could
readily be monitored include transportation movement, weather,
various types of production, movements of groups or individuals,
etc.
2. In addition'to being able to collect and monitor, capabilities
can be developed for the selection and assessment of data in near
real time. The information explosion phenomena. should properly be
regarded as an opportunity since it permits' selectivity of the data
which will be considered, whereas in the past it often has been
necessary to do a great deal of work with marginal information.
Obviously, an exercise of judgment is required, and if this judgment
is to be optimum appropriate tests for the importance of data must
be devised.
3. Concurrent with the development of the appropriate sensor
systems and their related communications and the development of
criteria for the assessment and selection of data in real time, various
economic models can be developed with which predictions of
capabilities which will set feasible action limits can be developed.
Such modeling can be extended into areas other than economic and,
in fact, some criteria for such model's based on observable actions
and events are explicit in descriptions of escalation and 25X`1A
other national relationships. . .
1. I feel the greatest obstacle to the exploitation of our
opportunities is the attitude of people; inertia to change is real and
cannot be overlooked. To some extent we have motor vehicle
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
0P. in :r
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
capabilities, but require that the motor vehicle be drawn by horses,
simply because we are used to thinking in terms of horses, and may
also be organized along these lines. There are limits in our mobility
both in developing concepts and in assessing those concepts in terms
of their potential value. Perhaps the best and most effective method
for changing our attitudes and perspectives is through the clear
delineation of goals, and the resulting development of clear under-
standing among the important parties concerned.
2. In order to realize the opportunites increased lead time,
particularly in the identification and selection of national goals, is
required. Although the necessity for lead time in the engineering
aspects of complex equipment is understood, there does not appear
to be an equivalent appreciation of the necessity for systems studies
and the development of logic as to how complex ensembles of equip-
ment can be most effectively designed and used. It is important that*
this preliminary, vital, work be carried on continuously. Such
studies will indicate needs for self-discipline in stating requirements
and identify the requirements which are most significant. Because
of the long lead time required, this work is often, neglected, and
sometimes, if it is started, the, end goal is forgotten before the work
is finished; the problem then includes continuity in the program.
3. The attached paper "The Automatic Control of Electric Power
in the United States" identifies and exemplifies many of the opportunities
and many of the problems which may be anticipated in the application of
technology to the intelligence process. While the paper is of some
interest because of the recent power failure in the Northeast, it pro-
vides an excellent illustration of technical capabilities for handling an
extremely complex situation. Note in particular the necessity in con-
trolling the generation of electric power for "feed forward". "Feed
forward" is analogous to prediction in the intelligence process, and
"feed back" also serves to modify the control process. Through a
back and forth flow of these two processes an optimum control situation
can be achieved. In the area of human factors and attitudes, it is
interesting to note that operators still manually close some switches
in the start up'process. The control room console also shows a
number of display devices by which the operator may monitor the
operations. These displays and the analog processes described are,
I believe, reflections of the reluctance of people to adopt techniques
and equipment which is strange and foreign to their experience. There
is a considerable spread in the degree of progress which has been
Approved For Release 7002/10/29 : CIA-RDP68R0053OA000200040021-9
~'+F7~yJ' y,
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
achieved in various power generating. station equipment. This
analogy probably fits the Intelligence Community also. Considered
internationally, however, the disastrous consequences of being left
technologically behind are apparent.
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
The automatic control of
electric power in the United States
Automation of the electric power industry is well under way.,
Comprehensive coordinated control systems; derived from theory
and practice; have been developed to solve complex multivariable
control problems of the modern steam plant. Controls are primarily
analog in nature but direct digital control is being investigated
Nathan Cohn Leeds & Northrup Company
In reviewing the state of the automatic control art in a
given field of application, it seems appropriate first to
take note of progress in the field itself. Changes in process
techniques or objectives define new requirements and.
opportunities for advances in control concepts and
equipment. Accordingly, this survey article will endeavor
to "close the gap" between theory and practice by doing
three things: (1) review in brief the growth and progress
of the power industry itself; (2) make some comments on
recent significant trends and innovations in its processes
insofar as they relate to major areas of automatic control
application; (3) discuss the state of the art in these ap-
plications.
The electric power industry
A considerable portion of the economic growth and
industrial strength in the United States relates to the
growth of its electric power industry, which is by far the
largest of the country's industries' (Fig. 1). In recent
decades the electric power industry in the United States
has been growing at a rate of about 7 percent per year,
more than doubling its generating capability every ten
years2 (Fig. 2). At the beginning of last year, generating
capacity totaled over 228 million kW.
Annual output of generating plants has been steadily
growing over the years (Fig. 3), and is now of the order of
1000 billion kWh, or-using more modern termi-
nology-1000 TWh (terawatthours). An interesting fig-
ure is that with only 6 percent of the world's popula-
tion, the United States generates 37 percent of the world's
power.
Power processes
Electric power processes are concerned with energy
conversion and the delivery of energy to users in useful
form, when and where and for as long as wanted, as
economically and dependably as possible.
To help clarify these processes and their control prob-
lems, and to provide a basis for an orderly review of recent
progress and present state of the control art in this field,
two general areas in this review shall be considered,
namely:
1. Energy conversion plants. The objective is to
operate each plant at optimum efficiency at the power
generation level assigned to it.
2. Interconnected networks of such plants. Here the
objective is to assign and maintain plant generation
levels and tie-line power flows that will yield optimum
overall system economy consistent with continuity of
system operation.
Generating plants
As can be noted from Fig. 3, over 81 percent of the
present output of the industry is derived from fuel, and
almost all of this is from fossil-fuel-burning steam-
electric plants. In discussing progress in plant processes
and automatic control applications within plants, this
article will accordingly confine itself to such fuel-burning
plants. That is not to say that other forms of energy con-
version are not important, but time and space limitations
will not permit their inclusion in this discussion beyond
these brief comments :
Pumped storage. In a typical area, over a typical day,
power demand will vary over a range of about two to one
or more between the periods of highest and lowest de-
mand. Power demand in the valley period is likely to be
below the total capability of the most efficient available
units, some of which would therefore have to be idle for
parts of the day. There has been increasing interest in
recent years in taking advantage of the variation in
power demand by utilizing "pumped storage" for "peak-
ing" purposes.3
Under some conditions it is advantageous, during
off-peak hours, to use low-cost steam plant energy to
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
Approved For Release 2002/10/29: CIA-RDP
Electric Oil Railroads Communi- Metals
power refining cations
Fig. 1. Gross capital assets (in billions of dollars) of the
largest United States industries in 1962.
1920 1925 1930 1935 1940 1945 1950 1955 1960 1963
Year
800
0
600
1925 1930 1935 1940 1945 1950 1955 1960 1965
Year
Fig. 4. Largest turbine-generator set, 1946-1964. Progressive
growth in size of individual units.
pump water up to high-elevation reservoirs, and then run
the pumps as generators during high-demand periods.
About 2000 MW of pumped storage capacity is currently
in operation or under construction. Several thousand
megawatts of additional capacity are under consideration.
Control specialists will recognize that there are in-
teresting optimizing challenges in the operation of such
facilities.
Nuclear power. Nuclear plants at present account for
only a small fraction of a percent of the nation's power
capability. It has been estimated, however, that by 1980
nuclear power installations will aggregate about 70 000
MW, or about 13 percent of the total capability expected
to exist at that time. Such plants will engender a par-
ticular awareness of economy and safety, and will provide
many automatic control opportunities.
Fossil-fuel plants. Let us turn now to fossil-fuel-burn- _
ing plants. Here a first major point of interest is the
progressively increasing sizes of individual units.
The trend of increasing sizes is shown in Fig. 4. A
typical size in 1950 was 175 MW (compared with 100
MW in 1940). A number of units in the 500-700-MW
range are now in operation or under construction. Units
of 900 and 1000 MW will start commercial operation
within the next two years. A 1200-MW unit is already
under contract.
Increased size has generally meant increased complica-
tion, with greater demands from, and greater dependence
on, automatic control.
Paralleling the increased sizes, and reflecting continuing
progress and improvements in the energy conversion
process, has been improved efficiency. This is illustrated
for the years 1930-62 in Fig. 5.
Best plant heat rate, which was 15 000 Btu per kilo-
watthour back in 1925 (not shown in Fig. 5), had im-
proved to approximately 8600 Btu per kilowatthour by
1962. For all plants in operation the drop has been from
25 000 Btu per kilowatthour in 1925 to approximately
10 500 Btu per kilowatthour in 1962.
Such steady and significant improvements in plant ef-
ficiency have counterbalanced the continuing increases in
the price of fuel over the years, permitting a fairly con-
stant cost of fuel per kilowatthour despite the fuel price
increases. These factors are illustrated in Fig. 6. Auto-
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
Fig. 3. United States electric power production.
77 U" U:)J JAUU~_ I
Fuel pri.ce/,--
Fuel use per kWh
L Cost of fuel per kWh
COL 0.3
0.1
1945
Year
Fig. 6. Fuel prices and efficiencies, 1926-1963. Improved
efficiencies counterbalance increased fuel prices.
matic control has played an important role in the achieve-
ment of these improved efficiencies.
The supercritical pressure unit. As part of the con-
tinuous search for improved operating efficiencies, design
pressures and temperatures of steam-generating units
have been steadily increased over the years. As a most
noteworthy advance in this field, several units have been
built and placed into operation within the past five years
which operate in the supercritical range. Most units in
the 500-MW and above size are or will be in this category.
When operating above critical pressure-3206 psi-
steam and water do not exist as a mixture. A steam gen-
erator of this type is therefore of the "once-through"
design. It has no steam drum with its storage effects, and
drum level is therefore not available as a control reference
as in conventional boilers. Its response characteristics
and degree of self-regulation differ markedly from con-
ventional units. It includes a very large number of ma-
nipulated variables, many of which have major interacting
effects on output parameters. Its operating complexities
are manifold. It has created the need for markedly new
concepts and executions in an automatic control system
that will provide stable coordinated regulation over the
full range of "light-off" to rated load, under both steady-
state and varying load conditions.
The solutions developed for such boilers are repre-
sentative of the most advanced state of the art in this
facet of power plant control.
In turning to more detailed comments on in-plant con-
trols, I think it will be helpful to consider two aspects of
such applications separately, even though they are to
some degree interrelated, namely: (1) continuous oper-
ating functions, including on-line automatic control,
performance computation, and safety monitoring; and (2)
automatic start-up and shutdown functions.
Continuous plant controls
Use of control theory. In the development of present-
day control systems, such as those now applied to once-
through boilers, modern control theory and simulation
techniques have been used, to a degree, to replace or
supplement the essentially empirical approaches of earlier
years. Simulation in particular, using both analog and
digital computers, has been helpful in dynamic modeling,
a.
o E
r- o
E
Analog computer simulation
Field data
4 6 8 10 12 14 16 18 20 22
Time, minutes
Fig. 7. Responses to step change in fuel and air at full load.
Comparison of L&N simulation with field data.
Fig. S. Responses to ramp change in feedwater. Compari-
son of L&N simulation with field data.
a
E~
a
c E
d.2
o
-,"Ramp input.
tL
~a
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
- Analog computer simulation
Field data
10 12 14 16 18 20 22.
Time, minutes
Approved For Release 2002/10/29 : CIA-D
~p8R00~530gOgo. 0qQ~4o a t 0~?d1 a9 process
in the development of advanced control concepts, an a e to inpu noes
in the synthesis of multivariable control systems."-?
Several companies feel that here, at least, a gap has
been effectively bridged and that there have been pro-
ductive exchanges between engineers with long-time
practical power plant field experience and scientists in
computer laboratories.
Simulations from theoretical design data of complex
once-through units have proved to be reasonable when
checked against subsequently available field data. Typical
comparisons of predicted variations in output parameters
for step function changes in input variables, based on a
theoretical simulation of a supercritical once-through
boiler from its design data, with subsequently available
field data are shown in Figs. 7 and 8.
These simulations were conducted on an analog com-
puter having 170 operational amplifiers, and, together
with similar simulations, were helpful in carrying out
significant studies and in reaching decisions on the rela-
tive merits of proposed alternative control arrange-
ments. This manufacturer is studying supercritical unit
responses more extensively on a large digital computer,
and subsequently performing the control studies on an
analog computer. With this approach, it has been pos-
sible to study all of the significant process inputs and
outputs to represent all of the major control loops.
In addition to simulation, use is being made in modern
control systems of noninteracting concepts, of feed-
forward techniques, and of adaptive arrangements to
adjust controller settings for nonlinear control responses
at varying loads and with different operating combina-
tions.
Where direct digital control techniques have been
undertaken, available knowledge of sampled data theory
has been used to establish sampling rates properly re-
In general, manufacturers feel they are making com-
mendable progress in the development of control con-
cepts as coordinated noninteracting control systems to
meet the exacting requirements of complex present-day
plants. They might well be the first to grant, however,
that not too much of the extensive control theories being
developed, particularly in academic circles, finds use in
present-day installations. Much that is done remains
empirical or experimental.
Empiricism and experimentation. In the context of
comparisons between theory and practice, it seems appro-
priate at this point to say a few words about empiricism,
its practitioners, and its relation to control applications.
Empiricism should not necessarily be downgraded. In-
deed, when it takes place in the most advanced of ex-
perimental laboratories it is usually called "science." In
the development and application of control systems,
empiricism is unscientific only when the attitude is obtuse
and narrow, and when available applicable contemporary
knowledge is not fully used.
Theorists should recognize that it is the empiricist, the
"practical" engineer, who is largely responsible for the
control systems that are installed and operating so well in
complex generating plants and on widespread inter-
connected systems today. Typically, he has gathered
much of his information and understanding from field
observations and experiences, from adjustment of con-
trol systems in customer plants, and from cut-and-try
techniques for improving their performance. He has
sought to absorb and utilize all the applicable theory he
can understand. In addition, the most effective em-
piricists have listened with respect to the more theoreti-
cally minded individuals at their plant headquarters who
Fig. 9. Simplified schematic of a coordinated control system for once-through
boilers. Major coordinating controllers are marked "C."
Boiler
Desired
generation
Desired
pressure
Desired
temperature
Desired 02
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
have been able to d S%,W fl@r it n /29traf ,~? IMP ~ 1~1 ~~~4~t {~A g~$a nerating
and solving field problems and better ways of executing actuator feedback signals.
recognized control principles. - There have been important improvements in the ac-
But there is one view from which the practical engineer curacy and reliability of flow and pressure transmitters,
cannot depart. Today's control problems need solu- but even greater performance capabilities are being re-
tions today. Customer commitments must be satis- quested by users.
factorily resolved with contemporarily available tools. Digital techniques. Power companies have been tradi-
Tomorrow's knowledge will not be today's until tomor- tionally alert to the need for collecting data for operator
row, and will not become fully useful until then. guidance and for evaluation and improvement of plant
The practical engineer recognizes, however, the im- performance. Where a company's fuel bill runs into many
portance of innovation, and he encourages all activity millions of dollars per year, big savings can be achieved
that will improve the state of the art. But he also recog- by even modest improvements in plant efficiency.
nizes that newness does not necessarily equate with Several years ago, conventional measuring and re-
progress. He welcomes innovation, but requires that cording instruments began to be supplemented, and in
what is new be proved, in actual operation, to have ap- some instances replaced, by digital data gathering as-
propriate qualities of reliability and improved per- semblies. Some of the systems included modest computing
formance before it qualifies for general acceptance as a capability. Pertinent information was in this way pre-
contribution to the art. sented to operators in a more centralized and coordinated
I am certain that all manufacturers are on the alert fashion, hopefully resulting in more rapid corrective
for, and ready to welcome, all theoretical contributions, steps when such were required.
from within and outside their organizations, that will With the advent of the digital computer, centralized
help them do a better practical applications job. data gathering systems were expanded to include addi-
A coordinated control concept. Let us turn to the syn- tional functions.9
thesis of a coordinated control system for a complex One objective was to provide the operator with con-
multivariable application. One control now in use and current computations of cycle and plant efficiencies in-
being supplied for new projects, derived from both stead of the previously available historical analysis.
theoretical considerations and field experimentation,4 is Another was to provide more extensive monitoring and
shown in highly simplified form in Fig. 9. alarm functions, which would aid in preventing both
The large-sized once-through units to which this sys- minor and major shutdowns. Efforts to extend the com-
tem is currently being applied have as many as 90 or more puter to plant control functions soon followed.
manipulated variables to be placed under coordinated In addition to these essentially continuous plant func-
automatic control. The diagram of Fig. 9 does not by any tions, computers were also installed to fully automate
means begin to show the extent of the control problem, plant start-up and shutdown without human intervention,
with its multiple fuel, air, and feedwater inputs, its re- as will be additionally referred to in a section that follows.
heat steam cycle, and its complicated, almost endless de- The pioneer on-line solid-state digital computer in-
tails. The diagram is intended only to show how the major stallation was made in 1958. Its functions were logging
input variables are acted on in coordinated fashion by of approximately 100 key variables, alarm scanning, per-
pertinent measured, set, and computed parameters, in formance calculations, and closed-loop direct digital con-
some cases in the same sense, in others in the opposite trol of two auxiliary temperatures.
sense. Since then, it is estimated that about 65 or 70 com-
Fecdforward from desired output, itself manually set puters have been installed in steam plants throughout the
at the station or automatically set by a remote dis- country and additional ones are in the process of in-
patching computer, is utilized. Appropriate limits for stallation.
ranges and rates of response are provided, as are auto- As experience has been gained, and new generations of
matic runbacks on loss of major auxiliaries. Each in- computers have become available, functions have been
dividual control action is conditioned to give weight to extended to larger numbers of points scanned, alarmed,
its influence on process output parameters, to varying and logged, to more meaningful and useful performance
response characteristics, and to varying time lags. These computations, to logical sequence control functions
sophistications, essential to coordinated nonhunting and and-in a few cases-to more extensive direct digital
safe regulation, are not shown, but perhaps this brief control.
reference to them will help to establish the dimensions of There is by no means agreement in the industry at this
the control problem. point, however, on the functions that can and should be
Analog executions. Virtually all major plant control assigned to a plant computer.
systems currently in operation or in the process of instal- One example of current practice is shown in Fig. 10.
lation are analog in nature. A recent trend has been to This is an artist's sketch of the control room for a large
all-electric and electronic executions, though a number plant now under construction. For each of two 900-MW
of the new large plants still favor pneumatic or elec- units, control will be of the analog electric-electronic type,
tropic-pneumatic assemblies. Electric-electronic systems and a digital computer will perform monitoring, per-
are felt by many to have the advantages of speed and formance computation, alarm, and logging functions.
flexibility, and they coordinate well with digital data and Start-up and shutdown switching functions will be
computing systems. manually executed, but the computer will provide
The newest analog systems make extensive use of sequence and safety monitoring instructions and check-
solid-state technology. Increasingly, SCR power switches back for operator guidance. The screen for projecting.
are replacing electromechanical contactors for operation monitoring messages to the operator can be seen in the
of large electric actuators. Characterizable contact-free center of the figure.
Approved For Release 2002/10/29 : CIA-RDP68R00530A000200040021-9
AnotheiApX9MfISLfar ,ql i n?OQZtJQI'gP i'09IA-I PFr6AW?gP4 WPQ ltN%ing way the full
operation this year, and exemplary of the present state
of the applications art for steam-plant computers, will
include the following functions: off-normal alarming,
performance computation, logging, trend recording, on-
demand display of stored history of plant variables,
trip sequence monitoring, some start-stop functions,
digital data control for several temperature and level
loops that completely replace analog loops for these func-
tions, and digital override of major analog loops for con-
trol of feedwater, combustion, and steam temperature.
Design, installation, and operating experiences with
computer installations have been well documented. 10-14
(References 10 and 14 are particularly helpful in portray-
ing the present state of this art.)
In general, I think it is clear from available docu-
mentation that there has been much to learn in applying
on-line computers to power plants. In many cases, the
experience has been time consuming and expensive, for
manufacturer and user alike.
One manufacturer, in providing information for this
survey, has cited a number of typical problem areas en-
countered with field installations. For most problems, he
points out, solutions have been provided. For others, he
notes, solutions still require verification.
The problem areas cited are: input signal noise, system
reliability (now felt to be of a high order), expansion
capabilities, computer speed (solved with presently
available units), transducer failures and errors, high cost
and excessive time required to define, code, and check out
programs (emphasized as a continuing real and major
problem), and poor operator/computer communication.
Start-up and shutdown functions. An understandable
objective in the operation of steam plants has been to
extend the use of the digital computer beyond con-
tinuous monitoring, computation, and control functions
to extensive start-stop functions, thereby achieving full
plant automation.
One of the pioneering utilities
Fig. 10. Centralized control and monitoring, large
supercritical unit.
degree of automation that it hoped to achieve: "The
boiler and turbine-generator should be capable of being
safely and reliably started, operated automatically at
optimum efficiency, and automatically shut down with-
out benefit of manual operator assistance."
References 11 and 13 document this company's and
its consultant's experiences, and detail the extensive com-
plications and problems encountered in undertaking so
comprehensive an objective. While indicating that con-
crete evidence had not yet been developed that a digital
computer system is the most economical method of auto-
mating a steam plant, they do call attention to benefits
gained from experiences with efforts to achieve such full
plant automation.
Prevailing views
There is not, at this point, a unanimity of views among
users and their consultants concerning the place of digi-
tal computers in steam plants. Performance of present in-
stallations is being appraised. Technical and economic
considerations with respect to future plants are being
evaluated. Here are excerpts from comments obtained
from major consultants and operating companies for
this survey article.
Consulting engineers. One consultant says: "Analog
systems will continue to be used for automatic boiler
control. There does not appear to be justification from the
standpoint of initial cost or improved efficiency to en-
courage development of digital techniques in this area.
"It is for on-line up-to-date performance computation
and safety monitoring that the computer initially derives
its justification. Eventually, performance calculations will
be standardized on a national basis, after which it is ex-
pected that such computers will be considered standard
equipment for new projects.
"Due to the high. equipment cost and complexity of
programming for intermittent start-up and shutdown
functions, and based upon reports of operating experience
in this area, we cannot find justification for providing
computer equipment for this purpose on an overall unit
basis. We do recommend that consideration be given to
computer start-up and shutdown in some areas, such as
the turbine-generator combination, where programs have
been successfully applied.
"Solid-state logic circuitry lends itself to start-up-
shutdown functions as exemplified by progress in the
burner light-off area.
"Overall start-up and shutdown appear to be a job for
solid-state switching circuitry working in parallel with a
computer."
Another consultant writes: "A direct economic justifi-
cation by savings in efficiency and safety is almost im-
possible to prove. Another approach is that of demon-
strating how small an increment of efficiency is sufficient
to cover the cost of a computer installation, and assuming
that the potential for such savings is inherent in a modern
on-line installation.
"Once this concession is made, it would appear reason-
able to justify an on-line computer for a large station by
the many protective possibilities it presents in addition to
performance calculation.
"While computer automated stations have yet to
prove their worth, there is little doubt that this is the way
the wind blows. A great deal of work remains to make
Approved For Release 2002/10/29 : CIA-RDP68R00530A000200040021-9
Approved For Release 2002/10/29: CIA-RDP68R00530A000200040021-9
more reliable, and to develop systematic methods tical operating experience in computer control before
for flow charting operations."
Still another consultant writes: "Any increase in the
extent of automation must contribute to greater safety to
both personnel and equipment, improvement of operating
and operator efficiency, increase in station availability,
and reduction of maintenance.
"All of the automation equipment must be justified
by such criteria. When we consider highly sophisticated
digital computing systems, justification tends to be based
on more of the intangible advantages. It is extremely dif-
ficult to assign tangible values to such things as preventing
catastrophic failures.
"Complete automation will be an integrated overall
approach where some systems are independent and others
are supervised by a stored program computer.
"Wide acceptance of advanced automation must de-
velop slowly, and must be based on sound economic
justification."
One consultant sees real potential worth of expected
benefits from full automation, on a basis of fuel economy,
minimization of manpower needs, reduction in main-
tenance, and reduction in both major and minor mis-
haps. "Expected savings," he points out, "can vary con-
siderably with staffing practices, unit characteristics, and
care in operation. For the same size units savings may
vary by a factor of 3. Individual study is therefore es-
sential."
He feels that direct digital control has promise, and
cites the possibility of digital programs being used pro-
tectively to catch analog control failures before great up-
sets occur.
"Efforts to compute efficiency and heat rate are re-
stricted by lack of solid understanding of heat storage
dynamics. We need a full range nonlinear simulation. We
live with imperfect sensors and position switches. A
major problem is getting people with the right tempera-
ment, knowledge, and interest to work at bridging the
gaps. We're still in R&D for the next couple of jobs."
Users. One of the users writes: "There are three draw-
backs to computer control at present. They are the high
cost of equipment, the high cost and effort required for
programming, and lack of reliable sensing devices for
many applications.
"We have not seriously considered digital computers
for continuous control functions. We have felt that the
most promising application is the continued use of analog
loops with the digital computer used to control the time
sequence of major operating steps and to make certain
logical decisions."
Another user furnished a copy of reference 14, in which
he comments about the future this way: "It is hoped that
the two-year slippage of computer-controlled start-up at
[Plant P] will be over in mid-1965. It is also hoped that
experience will be achieved earlier at [Plant B]; however,
the general inability of boiler, turbogenerator, and other
contractors to produce on time the necessary information
for analysis and flow-charting has already delayed com-
pletion of logic diagrams by 17 months. Programming
for control has not yet begun [February 1, 1965], and the
12 to 13 months estimated for this work cannot now be
completed by the commercial operation date. These
comments are an acknowledgment of underestimating
the job and overestimating ability in varying degrees by
all participants. The worst effect is not getting any prac-
having to proceed with another unit installation. In short,
we have so far been unable to prove monetary savings
equal to or exceeding cost of the systems. Information on
costs however seems to accumulate steadily. Automation
in various forms and degrees will continue to be applied
until further experience and operating data dictate the
requirements in a more adequate manner."
Another user, who has pioneered in plant automation,
writes: "With regard to our present policy concerning
automation, we are only making provisions for the future
addition of computers. Conventional control systems are
used. Automation features are included, but as wired logic
subloops.
"We have retrogressed somewhat from our initial auto-
mation philosophy which was fully automatic start-up
and shutdown programs with absolutely no manual in-
terventions, to a more comfortable position which allows
operators to perform most of the on-off type functions.
"It is true our new unit design lacks the degree of checks
and balances afforded by complex computer program-
ming, but with the automation we are providing in new
units, the operator has at his disposal more control ap-
paratus with which to circumvent plant hardware or con-
trol equipment troubles. Where set computer control
programs provide a start-up procedure, increased remote
manual control provides a start-up flexibility that has
been lacking in all computer control programs to date."
Finally, still another user has this view: "More reliable
primary sensors are needed to eliminate paralleling for
redundancy.
"A great deal of effort should be expended toward ap-
plication of digital techniques to present analog control
functions. Other power plant processes which are at
present inadequately sensed and controlled should be
studied for digital control application."
It is clear from the foregoing summary of viewpoints
that widespread differences of opinion currently prevail
as to the degree of steam plant automation that is justifi-
able or desirable, and the extent to which it should be
analog or digital or both.
Probably the best way to appraise the state of this facet
of the art is simply to say that it is in a state of flux.
Views have certainly not yet hardened. Utilities and
their consultants, to their credit, have encouraged ex-
perimentation. Not all the approaches or innovations
have been the same, and not all the results have been
satisfactory. Such uncertainty and differences of view
are as good an indication as any that progress is being
made. Work, in many plants, involving many manu-
facturers and consultants, continues. Inevitably, better
clarity will emerge in the years ahead, and more uni-
versally shared views will doubtless develop.
Interconnected systems
Increasingly, over the past four decades, adjacent power
companies have interconnected with one another for
parallel operation. By this means, generation and re-
serves can be shared. Advantage is taken of load diversity
and of time-zone differences to transfer generation over
interconnecting tie lines from an area of low demand to
one of high demand. Larger, more efficient units can be
purchased and their outputs shared, and rotating reserves
in a given area reduced. Overall operating economies are
correspondingly achieved.
Approved For Release 2002/10/29 : CIA-RDP68R00530A000200040021-9
tively limited areas. As operating problems were analyzed
and resolved, and parallel operation technologies were
developed, interconnections have been steadily expanded.
Today, five operating interconnections account for the
entire country. The largest interconnection extends to the
cast from the Rocky Mountains, and includes the Mid-
west, the Gulf Coast, the Eastern Seaboard, and eastern
Canada. Constituent groups within this single intercon-
nection are the Interconnected Systems Group (ISG) of 115
operating utilities, private and public, the Pennsylvania-
New Jersey-Maryland pool (PJM) of 12 operating com-
panies, and the Canadian-Eastern United States group
(CANUSE), which has 31 operating utilities.
The more than 150 utilities of this interconnection,
having a total peak load greater than 130 million kW,
operate continuously in parallel, smoothly and coopera-
tively. Automatic control makes that possible.
Similar parallel operation is achieved and automatic
control is similarly utilized in the four other intercon-
nections.
In earlierAyrPv~it~r~o4i~i~d[fb~St~~~/dJ'c~i^el~,CIA-R~~8og,3R~Oglggg,~gt4~n~-ave undertaken
Fig. 11. Control areas of the country's largest interconnec-
tion (ISG, PJM, and CANUSE).
Fig. 12. Varying sizes and ages of available generators.
Distributing of thermal generating units.
Bands,
megawatts
Less than 10 years
Over 10 years
2[I F] n
M in r. rn O
( S 0
short-term test periods of parallel operation with each
other. It is.generally anticipated that by about the end of
this decade requisite new tie lines will have been built and
closed, permitting all five interconnections to operate as a
single interconnection covering the entire United States
and portions of neighboring nations.
The control problem
Coordinated control is essential to successful parallel
operation. Concepts for system regulation and optimiza-
tion have been well developed over the years, and are in
widespread use within the limits of present-day equip-
ment and technologies. II
Control requirements are twofold:
Area regulation. Total generation within an operating
area must be adjusted to follow the moment-to-moment
load changes within that area, in suitable coordination
with generation and load changes in all other operating
areas, so that scheduled tie-line interchanges with ad-
jacent areas, system frequency, and system synchronous
time are all properly maintained. This function is re-
ferred to as "area regulation."
Economic dispatch. Optimal assignment of the total
generation required at any moment from an area should
be made among the many plants and units within that
area to achieve optimum economy consistent with safe
operation. This objective is identified as "economic dis-
patch."
Let us explore briefly the nature of these two functions,
and the present state of the art in achieving them.
Area regulation
The word "area" probably needs to be explained. An
"area" can be a part of a company, a whole company, or
a group of adjacent companies, which operate, from the
viewpoint of interconnection, as a single entity. It
schedules and maintains levels of tie-line interchange with -
its neighbors, but permits tie lines within the area to be
free flowing; i.e., no effort is made to maintain them at
any scheduled levels. All load changes within the area,
regardless of where they occur, are treated alike insofar as
automatic control is concerned.
A first objective of area regulation is to automatically
adjust total generation to match total load changes within
the area. A second objective is to adjust area generation
when required to assist on a preprogrammed basis any
other area of the interconnection that may be in trouble,
and which cannot at the moment fulfill its own area
regulation obligations.
The universally accepted technique by which coordi-
nated neighborly area regulation fulfilling both of these
objectives is achieved is known as "net interchange tie-
line bias control."
The country's largest interconnection (ISG, PJM, and
CANUSE) has some 88 control areas for its 115 operating
utilities. These areas, with their major tie-line inter-
connections, are shown in Fig. 11. It will be understood
that each circle is a control area, and may include, as
several of them do, a number of companies "pooled"
together to operate as a single entity from the viewpoint
of area regulation and economic dispatch. Also, each line
between two areas may represent many tie lines, and not
just a single tie.
This interconnection extends into Quebec in the North-
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
R
on eleap~ 2exas 002/0/29
east, Florida in the outne st, F porti
Southwest, and Montana and the Dakotas in the North-
west.
All 88 of the operating areas shown in Fig. 11 are
equipped for net interchange tie-line bias control.
The tie-line telemetering required for the automatic
computation of an area's net interchange with the inter-
connected system is largely analog, although in recent
times digital techniques have also been used. Telemetering
transmission has increasingly been by microwave.
Almost all of the area controllers, including most of
those that have been recently installed, are of the analog
type, the newest using solid-state components. One or two
on this interconnection, and a similar number on the
Pacific Southwest-New Mexico interconnection, use
digital control techniques for area control function; these
will be referred to further in discussing the use of digital
computers for the economic dispatch function.
Area regulation performance. Area regulation perform-
ance on the interconnection of Fig. 11, though by no
means perfect, has been very good. A commendable job
is being done, permitting the areas of this widespread
network to achieve the benefits of parallel operation and
to stay together in operating synchronism, even during
periods of large system disturbances.
It is no mean task to assure that 88 controllers, spread
out over thousands of square miles of area-each in an
independent privately or publicly owned utility, each
depending on widespread telemetering networks, each
requiring frequency and tie-line schedule settings co-
ordinated on a system-wide basis, each requiring appro-
priate "bias" settings, each to be backed up by adequate
and responsive generating capacity, each to be adjusted
so that it corrects errors and does not create them, and all
operating simultaneously on a single integrated net-
work-will effectively fulfill their individual objectives
and obligations and at the same time contribute to the
common overall network objective of sustained, stable
parallel operation.
That these complex control objectives are achieved
as well as they are is a credit not only to the state of
the applicable control art, but also to the operating people
around the interconnection who are charged with making
the system and its equipment work.
For many years there have been informal, voluntary
operating and test committees on the various pools and
interconnections who have appraised performance,
analyzed problems, and established operating guides,
thereby contributing immeasurably to the results cur-
rently being achieved. More recently, an informal,
voluntary nationwide group, the North American Power
Systems Interconnection Committee (NAPSIC) has been
formed, with representation from all operating regions of
the country, to deal on a national basis with the coordina-
tion problems of massive networks. This committee will
make important contributions to improved system opera-
tions.
With regard to area regulation control performance,
the refinements still to be achieved are: better and more
rapid responses in some areas to changes in demand
within the area; minimizing the regulating assistance re-
quired from other areas; fuller coordination of tie-line
schedules; better telemetering channels for more sus-
tained communications between tie lines and control, and
between control and regulated generators; better co-
orCIA DP68R0g0530A00tQ2000o40021 -9
rror cor-
rection.
Improvement in all of these factors will decrease present
levels of "inadvertent" interchange-i.e., deviations from
scheduled interchanges between areas-and will assure
equitable distribution of system regulating burdens.
Economic dispatch
In the execution of area regulation, generation is auto-
matically adjusted within the area to match area load
changes. Clearly, it would be advantageous to assign each
required generation change to sources within the area
that can most economically absorb it. If we take into
consideration the different ages and sizes of units (Fig.
12), the resultant differences in their efficiencies, and their
different locations and consequent differences in trans-
mission loss factors to load centers, there is opportunity
for substantial economies by loading them optimally
with respect to one another. This is economic dis-
patch. It is achieved when generating sources within
the area are loaded to equal incremental costs of de-
livered power. For fuel-burning plants the well-known
coordination equation for such optimization is
df t r,,
dP,z
1-CPL
oPn
X is the incremental cost of power delivered for the area
dF,z is the incremental cost of power generated at source n
dPn
dPL is the incremental transmission loss for source n
c6P"
dHn is the incremental heat rate for source n
dP,
f, is the cost of incremental fuel for source n, adjusted to
include other varying costs at source n
There has been an interesting evolution in the auto-
matic control equipment used to achieve economic dis-
patch. A brief summary of some of its highlights and
comments on the present state of the art follows. Those
interested in fuller details of early steps in this evolution
and in information concerning the derivation and applica-
tions background of Eq. (1) are referred to papers listed
in the bibliography of Ref. 15.
Flexible loading consoles. The first areawide automatic
economic dispatch systems date back to the early 1950s.
Kilowatt loading schedules for each controlled source of
the area as a function of total area generation were
computed in advance, and were manually programmed
into a centralized computer-control console. As area
demand varied, the centralized console and its auxiliary
equipment would compute and maintain the proper load-
ing level for each generator of the area, and in the process
would fulfill the area regulating requirements.
A unique and important feature of these control as-
semblies was the use of a feedforward signal from area
control error, which, when combined with feedback from
prevailing area generation, provided a reference for pre-
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
dictively eAr~ lro etd For Release 20m 021/0/29 : CIA- pe6 0A~5 (~gA(~g~OSeg0or the late 1950s eac
source. This assignment was then independent of the
rate at which other participating sources responded to
their control assignments.
Another important feature involved arrangements for
overriding economic dispatch when the area regulating
requirement would not be fulfilled rapidly enough by
the sources responding to economic dispatch.
A large number of these early consoles are still in use,
and a number of new ones have been installed in quite
recent times, some for predominately hydro areas16
where the flexibility of wide-range schedule setters is very
useful, and others for coordinated use with digital com-
puters, as will be noted later. The newer assemblies
utilize solid-state circuitry in place of the electromechani-
cal elements of earlier units.
Analog desired-generation computers. An advance from
Fig. 13. Operating companies and ties of the circle marked
"P" in Fig. 11. Pennsylvania-New Jersey-Maryland pool.
Fig. 14. Major operating companies and ties of the circle
marked "A" in Fig. 11. American Electric Power System.
Public
Service of Ind
P. & L. Cincinnati
G. & E.
Carolina
P. & L.
with the introduction of analog-computer-control assem-
blies based on the coordination relationships of Eq. (1).
Here desired generation for each available source was
computed on a continuous basis from the following:
1. Stored information of its incremental heat rate vs.
output relationship.
2. Incremental transmission losses for its station,
computed dynamically from continuous measurement of
prevailing area generations and power flows and from
stored constants related to the area configuration.
3. A continuously computed lambda value that would
yield an economic dispatch whose total generation would
satisfy the prevailing area needs.
Some two dozen or so large centralized analog com-
puter controls of this general type are currently in opera-
tion, or in the process of installation. A combination of
feedforward and feedback, as with the flexible loading
consoles, provides a predictive lambda computation and
contributes to nonhunting simultaneous control of area
sources.
Digitally directed analog control. A relatively recent
development has been the inclusion of digital computers
in centralized economic-dispatch control systems. One
technique is to link the digital computer to an analog
console of the general type discussed earlier.17-11 With
this arrangement, the digital computer at appropriate
intervals executes the economy dispatch computations,
and transmits the resultant desired generation levels for
the controlled sources to setting devices in the analog
console. The latter executes the control assignment,
achieving the economic dispatch while simultaneously ful-
filling the area regulation obligation.
The digital computer can advantageously be pro-
grammed for other on-line functions, such as determining
the appropriate time to bring generators on the line or take
time off as peak load goes through its daily peak and _
valley cycles, checking reserves and imposing security
restraints for various parts of the area, evaluating pos-
sible advantageous interchanges for neighboring areas,
optimizing pumped hydro operation, checking area volt-
ages, and logging pertinent operations and measurements
data.
When the digital computer is not available for eco-
nomic dispatch because of off-line use, or because of main-
tenance or malfunctioning, the analog console provides
continuity of multiple-unit area regulation while execut-
ing manually programmed economic dispatch.
Several digitally directed analog systems have re-
cently been placed in operation or are presently being in-
stalled.
Direct digital control. An alternative technique when
utilizing a digital computer is to apply it directly for area
regulation and economic dispatch, without the use of an
intermediate analog control console. At the present state
of the art, a suitable stand-by analog control would
probably be retained, or made available, for area regula-
tion.
Initial installations of this type are now in opera-
tion,"," and are reported to be performing satisfactorily.
Typical installations
It may be of interest at this point to look briefly at
two major recent installations, which will illustrate a
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9
number of the PApprevttdiFci 2eilea se)tQD02d14dt~/29 :T01AT bP@ 0~ 2110M Ql1tsJ9 Automatic
connected systems controls. Control rt in tic Electric ower Industry of the United States,"
h' h , bl' 11 'd' 1 d, 1 196"
P
A
Pennsylvania-New Jersey-Maryland pool. The control
area marked P in Fig. 11 is the Pennsylvania-New Jersey-
Maryland pool. Companies of this group were pioneers
of the "pooling" concept, in which member com-
panies operate as a single control area with free-flowing
ties and a common economic dispatch.
The 12 independently owned companies of this pool,
and their intrapool and external ties to neighboring
utilities, are shown in Fig. 13.
Control execution for this pool is hierarchical. At the
pool headquarters in Philadelphia a net interchange tie-
line bias controller acts with an analog computer unit to
establish a pool lambda that will satisfy area generation
requirements, and will simultaneously maintain scheduled
interchanges over the 16 northern, Southern, and western
tie points to the utilities with which the pool inter-
connects.
The computer control at the central headquarters com-
municates this lambda value to computer-control as-
semblies located at the dispatch centers of the respective
member companies, which in turn compute and execute
corresponding economic dispatch assignments for their
generating units.
American Electric Power System. The circle marked
A in Fig. 11 represents the American Electric Power Sys-
tem. A closer look at this system and its major tie points
with its principal neighbors is provided by Fig. 14.
This system, the country's largest investor-owned elec-
tric-energy producer, has pioneered in interconnected
operation, as is apparent from its many ties with other
utilities. Here a group of adjacent companies having
common ownership operate as a single control area.
Tntrasystem ties are free flowing. Schedules for ad-
vantageous interchange are established and maintained
for the 40 major tie points with 19 other utilities, and
economic dispatch for the 38 principal generators of the
system is automatically computed and maintained, all
from one central location.
A new digitally directed analog system22,23 to fulfill
these functions was placed into operation late in 1964
at the company's new power control center in Canton,
Ohio. Control commands are routed directly from Can-
ton over microwave to the participating generators of the
operating companies.
This installation, with its solid-state analog console, its
digital computer, its individual unit approach, its use of
antihunting concepts, its display arrangements for unit
conditions, its arrangements for computing advantageous
interchanges with its many neighbors, its tie-in with a
large billing computer, and its extensive use of micro-
wave telemetering, reflects well the present start of the art
in interconnected power system controls.
I should like to acknowledge my indebtedness to a number of
my associates at the Leeds & Northrup Company and to indi-
viduals of the following organizations for information helpful in the
preparation of this paper: American Electric Power Service Corp.,
Bailey Meter Co., Control Data Corp., Ebasco Services, Inc.,
General Electric Co., Gilbert Associates, Hagan Controls Corp.,
IBM Corp., Metropolitan Edison Co., North American Power
Systems Interconnection Committee, Sargent & Lundy, Southern
California Edison Co., Stone & Webster Engineering Corp.,
Tennessee Valley Authority, and Westinghouse Electric Corp.
Figs. 1, 4, 5, and 12 are derived from reference 1; Figs. 2, 3, and
6 from reference 2; Fig. 11 is reproduced, with permission, from
material of the North American Systems Interconnection Com-
mittee.
w tc was pu is c m t UL c
rocee rags of t to J it;,
ltto.natrc
Control Conference.
REFERENCES
The Power Industry
1. National Power Survey, Federal Power Commission, 1964.
2. Statistical Year Book, Edison Electric institute, New York,
N. Y., 1963.
3. Friedlander, G. D., "Pumped storage-an answer to peaking
power," IEEE Spectrum, pp. 58-75; Oct. 1964.
Simulation and Plant Control
4. Argersinger, J. I., Laubli, F., Voegeli, E. F., and Scutt, E. D.,
"The development of an advanced control system for superer; ;cal
pressure units," paper presented at Nat'l Power Conf., Cincinnati,
Ohio, Sept. 22-26, 1963.
5. Stephens, W. M., deMello, F. P., and Ewart, D. N., "Simula-
tion as a design tool for plant Jack McDonough boiler controls,"
paper presented at 7th Ann. Power Instrumentation Symp., ISA,
Denver, Colo., May 1964.
6. Chen, T. S., and Schwartzenberg, J. W., "Cascade and model
control methods for superheater temperature control," ISA Trans.,
vol. 3, no. 4, 1964.
7. Adapts, J., Clark, D. R., Louis, J. R., and Spanbauer, J. P.,
"Mathematical modeling of once-thru boiler dynamics," IEEE
Traits. on Power Apparatus and Systems, vol. PAS-84, no. 2, pp.
146-156; Feb. 1965.
Digital Techniques
8. Gupta, S. C., and Ross, C. W., "Simulation evaluation of digi-
tal control system," ISA Trans., vol. 3, no. 3, 1964.
9. Summers, W. A., "Central station control," paper presented at
1st Ann. Power Instrumentation Symp., ISA, May 21-23, 1958.
10. Garney, R. J., Rankin, R. A., and Lloyd, A. G., "Experience
with direct digital control at the Little Gypsy Steam Electric Sta-
tion," paper presented at 19th Annual ISA Conf., Oct. 12-15,
1964.
11. Norry, R. A., Quist, L. R., and Emerson, L. R., "Engineering
aspects of a fully automated power plant," Paper 10.1, presented
at WESCON, Los Angeles, Calif., Aug. 25-28, 1964.
12. Livingston, R. G., "Computer system aspects of Etiwanda
automation," Paper 10.3, WESCON, 1964.
13. Ward, A. A., and Knapp, R. V., "Basic approach and experi-
ence with Etiwanda automation," Paper 10.4, WESCON, 1964.
14. Williamson, M. M., "TVA progress report on power plant
data logging and control," paper presented at Power Ind. Com-
puter Application Conf., Clearwater, Fla., May 1965.
Interconnected Systems Controls
15. Colin, Nathan, "Control of interconnected power systems,"
chapt. 17, in the Handbook of Automation, Computation and Con-
trol, vol. 3. New York: Wiley, 1961.
16. Benson, A. R., Johannson, D. E., and McNair, H. D., "Cen-
tralized load-frequency control for the U.S. Columbia River
power system," Paper CP63-230, presented at IEEE Winter Power
Meeting, New York, N.Y., Jan. 27-Feb. 1, 1963.
17. Blodgett, D. G., Hisscy, T. W., Falk, A. K., and Schultz,
W. B., "Application of an on-line digital computer for dispatch
and control of the Detroit Edison System," Paper CP62-247,
presented at AIEE Winter General Meeting, New York, N.Y.,
Jan. 28-Feb. 2, 1962.
18. Lydick, H. W., and Sutherland, J. F., "ADDAPU-Auto-
matic digital dispatch and processing unit," paper presented at
Power Ind. Computer Appl. Conf., Phoenix, Ariz., Apr. 1964.
19. Beyer, W. G., Chamberlain, H. H., Fiedler, H. J., and Simonds,
W. B., "Hybrid dispatch system at Florida Power Corporation,"
paper presented at Power Ind. Computer Appl. Conf., Clearwater,
Fla., May 1965.
20. Cunski, R. C., Harkness, M. A., Adibi, M. M., and Glimm,
A. F., "A digital dispatch system," paper presented at Power Ind.
Computer Appl. Conf., 1964.
21. Baker, A. D., Giras, T. C., and Nelms, W. B., "Breaking the
all-digital' barrier in systems operation computers," paper pre-
sented at Power Ind. Computer Appl. Conf., 1965.
22. Kinghorn, J. H., McDaniel, G. H., and Zimmermann, C. P.,
"Development of coordination and control of generation and
power flow on the AEP System," paper presented at American
Power Conf., Chicago, Ill., April 27-29, 1965.
23. Morgan, W. S., Willennar, A. H., Cohn, Nathan, and Nichols,
Clark, "Facilities for the AEP System power production and con-
trol center," paper presented at American Power Conf,, 1965.
Approved For Release 2002/10/29 : CIA-RDP68R0053OA000200040021-9