Chemical engineering education ( Journal Site )

Material Information

Chemical engineering education
Alternate Title:
Abbreviated Title:
Chem. eng. educ.
Physical Description:
v. : ill. ; 22-28 cm.
American Society for Engineering Education -- Chemical Engineering Division
Chemical Engineering Division, American Society for Engineering Education
Place of Publication:
Storrs, Conn
Publication Date:
annual[ former 1960-1961]


Subjects / Keywords:
Chemical engineering -- Study and teaching -- Periodicals   ( lcsh )
periodical   ( marcgt )
serial   ( sobekcm )


Chemical abstracts
Additional Physical Form:
Also issued online.
Dates or Sequential Designation:
1960-June 1964 ; v. 1, no. 1 (Oct. 1965)-
Numbering Peculiarities:
Publication suspended briefly: issue designated v. 1, no. 4 (June 1966) published Nov. 1967.
General Note:
Title from cover.
General Note:
Place of publication varies: Rochester, N.Y., 1965-1967; Gainesville, Fla., 1968-

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 01151209
lccn - 70013732
issn - 0009-2479
lcc - TP165 .C18
ddc - 660/.2/071
System ID:

Full Text

ch m ai eg aiern e Sato






Braun & Co
is a world leader
in the engineering and
construction industry. For
more than 70 years, we have
provided a wide range of services to the
process and power industries.
Our principal fields of activity are
chemical and petrochemical plants, oil
refineries, ore processing plants, coal
gasification facilities, and power


processes never before employed
on a commercial scale. We also have been
involved in the emerging synfuels
Our rapid growth has opened up many
challenging opportunities and
assignments for professional growth.
Positions are available at our engineering
headquarters in Alhambra, California, and
at our eastern engineering center in
Murray Hill, New Jersey.




I _

Department of Chemical Engineering
University of Florida
Gainesville, Florida 32611
Editor: Ray Fahien (904) 392-0857
Associate Editor: Mack Tyner
Editorial & Business Assistant:
Carole C. Yocum (904) 392-0861
Publications Board and Regional
Advertising Representatives:
Lee C. Eagleton
Pennsylvania State University
Past Chairman:
Klaus D. Timmerhaus
University of Colorado
Homer F. Johnson
University of Tennessee
Ralph W. Pike
Louisiana State University
James Fair
University of Texas
Gary Poehlezn
Georgia Tech
Darsh T. Wasan
Illinois Institute of Technology
J. J. Martin
University of Michigan
Lowell B. Koppel
Purdue University
William H. Corcoran
California Institute of Technology
William B. Krantz
University of Colorado
C. Judson King
University of California Berkeley
Angelo J. Perna
New Jersey Institute of Technology
Stuart W. Churchill
University of Pennsylvania
Raymond Baddour
A. W. Westerberg
Carnegie-Mellon University
Charles Sleicher
University of Washington
Leslie W. Shemilt
McMaster University
Thomas W. Weber
State University of New York

Chemical Engineering Education




Department of Chemical Engineering
50 University of Minnesota,
Rutherford Aris

The Educator
56 J. D. Seader of Utah,
A. D. Baer, Vickie Jones

4aw~ J2ect/es
62 Design Research: Both Theory and
Strategy (Part 2), A. W. Westerberg

ChE Lecture
68 Chemical Process Synthesis, J. J. Siirola
72 Chemical Process Design: An Integrated
Teaching Approach,
Kenneth A. Debelak, John A. Roth
76 Simulation of the Manufacture of a Chemi-
cal Product in a Competitive Environ-
ment, T. W. F. Russell,
D. F. Brestovansky, R. L. McCullough
82 On the Tensorial Nature of Fluxes in a
Continuous Media,
Vijay Kumar Stokes, Doraiswami
88 The Integration of Energy Conservation
Principles into a Course on Staged
Operations, Thomas J. McAvoy

Class and Home Problems g
94 Ice Rink Problem, Thomas J. McAvoy
67 ChE Division Activites
93, 95 Book Reviews

CHEMICAL ENGINEERING EDUCATION is published quarterly by Chemical
Engineering Division, American Society for Engineering Education. The publication
is edited at the Chemical Engineering Department, University of Florida. Second-class
postage is paid at Gainesville, Florida, and at DeLeon Springs, Florida. Correspondence
regarding editorial matter, circulation and changes of address should be addressed
to the Editor at Gainesville, Florida 32611. Advertising rates and information are
available from the advertising representatives. Plates and other advertising material
may be sent directly to the printer: E. O. Painter Printing Co., P. 0. Box 877,
DeLeon Springs, Florida 32028. Subscription rate U.S., Canada, and Mexico is $15 per
year, $10 per year mailed to members of AIChE and of the ChE Division of ASEE.
Bulk subscription rates to ChE faculty on request. Write for prices on individual
back copies. Copyright 1982 Chemical Engineering Division of American Society
for Engineering Education. The statements and opinions expressed in this periodical
are those of the writers and not necessarily those of the ChE Division of the ASEE
which body assumes no responsibility for them. Defective copies replaced if notified
within 120 days.
The International Organization for Standardization has assigned the code US ISSN
0009-2479 for the identification of this periodical.



The older students provided a clue-sheet for the new

LN. department

students at the Fall faculty-graduate student party.


U. Minnesota
Minneapolis, MN 55455

AS EARLY AS 1902, G. B. Frankforter of the De-
partment of Chemistry wrote to Cyrus North-
rop the President of the University of Minnesota
to point out
"the importance of offering .... a course of study in
which both chemistry and engineering are represented.
Such a source is now offered in all of the larger institu-
tions of Germany and in nearly all of the larger Uni-
versities and technical institutions of this country.
"What I would respectfully recommend then is, that
a course of study be offered which will meet the urgent
demands of the present time. I will state that I have
consulted the various members of the engineering
faculty concerned and the plan meets with their most
hearty approval. I will also say that in offering this
new course, there need not be a single dollar of ad-
ditional expense to the University.
I would also recommend that the School of Chemis-
try be separated from the College of Science, Litera-
ture and the Arts, in as much as the work is entirely
foreign to that College. We graduate four men this
year and there are some twenty members of the
present Freshman class. With this new course which

is sure to be popular although it is an exceedingly
difficult one, we shall be quite on a par with the School
of Mines.
Very respectfully submitted,
(Signed) G. B. Frankforter."

The School of Mines has since passed away
but the "new course" continues to be both "popu-
lar" and "exceedingly difficult", and what presi-
dent could fail to approve a program urged with
such enthusiasm and costing "not a single dollar
of additional expense."!
It was not until 1919 however that the cur-
riculum was formalized as the "Chemical Engi-
neering Course" and what was effectively a de-
partment arose as a division of the School of
Chemistry. Charles A. Mann was promoted to
the rank of professor with the title of acting
head of the Division of Chemical Engineering. For
the time being very inadequate quarters had to be
used, and no satisfactory development of the ex-
perimental phases of chemical engineering could
be put into operation until the completion of the
fourth wing of the Chemistry Building. In the
meantime, however, every effort was made to de-
velop a satisfactory course in chemical engineer-


Copyright ChE Division, ASEE, 1982


Twenty years later the department was still
in the bowels of the chemistry building. Montillon
and Montonna had joined Mann and were the
three professors, with Rogers and Stoppel as
assistant professors, Grove and Piret as in-
structors, and Armstrong, Chamberlain, Clegg,
Eldredge among the teaching assistants. B. F.
Ruth had been on the faculty; Amundson, Piercy
and a host of others had graduated and the cur-
riculum now contained Unit Operations, ChE
Thermodynamics, Plant Design and Economics as
well as courses on Explosives, Dyestuffs, Cellulose,
Gas Manufacture and Sanitary Chemistry.
Thirty years later, just after Mann's death,
Ceaglske and Piret were professors, Amundson
and Stoppel associate professors and Stevenson,
Preckshot and Madden assistant professors. The
search outside the department for a new head was
unsuccessful and Amundson, who had been made
Acting Head in 1949, became the Head in 1951
and the big move from the basement of Chemis-
try to a new building was about to begin. It was
a building of some 65,000 sq. ft. designed to handle
80 seniors and 100 graduate students. Apart from
some dissatisfaction with certain of the materials
(walls somewhat permeable to sound, for
example) the building proved satisfactory but is
once again stretched at its seams with the
numbers of students. Many of the details of the
transfer to the new building fell to some of the
new faculty who had joined the department in the
late 40's, notably to George Preckshot. There have
been several episodes of remodeling, including the
latest one to meet OSHA's requirements but in
the fiscal climate of the moment there seems little
prospect of the radical reorganization that is
really necessary.
In the early 50's there were about eight
members of the department, Ceaglske and Piret
being the professors until joined by Amundson in
'51 and Stoppel in '52. Isbin had come in 1950 and
soon became an associate professor, being joined
in that rank by Madden and Preckshot around the
middle of the decade. Tsuchiya came in 1956 to
bring the total number, including Hap Earle, to
The pace of research increased in the 1950's
also with such outstanding graduate students as
Acrivos and Lapidus. Piret attracted a number of
French students. Outstanding among them was
Olegh Bilous who did his Ph.D. work with Piret

but also, on the side, wrote with Amundson some
of the first papers on chemical reactors as
dynamical systems.
During his sabbatical at Cambridge in 1954-5,
Amundson not only wrote up much of this re-
search but saw something of what could be done in
a department when colleagues of diverse back-
grounds are brought together and realized the
potential of team teaching for developing really
good cooperation. Toward the end of the decade
he was able to put these ideas into effect as both
numbers and research activity took off again.
With Stoppel's death, but Isbin's promotion and
Ranz's appointment, the tale of professors stood

... the department used to
have a reputation for being dominantly
theoretical. This was never really accurate,
for there has always been a predominance
of experimental programs.

at five, with three associate professors and an in-
creasing number of assistants. Dahler, Fredrick-
son and Aris came in 1958 and were joined by
Scriven in 1959. Edgar Piret left to be the Scien-
tific Attache in Paris. Davis came in 1963 and
Keller in 1964 being followed in 1965, by Schmidt
and Carr. These years also saw some of our out-
standing graduates, such as Ramkrishna and Luss
(and later Rhee and Varma), remaining for a
short while as assistant professors. In 1969 there
were nine professors (we avoid the neologism
"full professors" as having too dyspeptic an
image), three associates and three on the tenure
track among a total of seven assistant professors.
Of course it is not only the faculty that matter
and graduates, of the department probably re-
member best the outstanding figures of the staff:
Ruth Nelson, who had everything at her finger tips
or in a little 3x5 file box; Verne Nelsen, to whom
the mysteries of stock and budget were as plain as
day, and John Antolak without whom no experi-
mental project got off the ground and but for
whom the lab would have given up the ghost
long before OSHA boxed it about with regulations.
In 1970 the Dean, disbanding the School of
Mines and Metallurgy, allowed some of the more
scientific types to come to chemical engineering
and it became a department of Chemical Engi-
neering and Materials Science, almost half as
large again as before. The development of the de-
partment in the 1970's was governed by two con-
siderations; namely, the need to integrate into it


the newly acquired aspect of materials science
and to maintain and strengthen the standing
that, by this time, the department had acquired.
The area of polymers was a natural bridge be-
tween the chemical and the material ends of the
operation and Macosko came in 1970, starting a
very active program in polymer processing and
rheology. Thomas was with us from '73 to '77
when he returned to Massachusetts, changing
places with Tirrell, whose powerful progress has
been recognized both internally and externally, by
a Dreyfus Award and a Sloan Fellowship. Then
Wellinghoff, a polymer morphologist, came to
strengthen the bridge from the materials science
side in 1978. Meanwhile, Gerberich had come in as
a metallurgical materials scientist and Hutchin-
son, an electron microscopist (who had been part
of the transfer of 1970) had left for the Uni-
versity of Washington. George Stephanopoulos
came from the University of Florida to bring a
great strength in the theory of design and control
and was with us from 1974 to 1980, when, to our
regret, he returned to Greece.
When, in 1974 (after 25 years of headship)
Amundson asked to be relieved, his colleagues,
much as they deplored the suggestion, could
scarcely refuse. Aris was Head for four years
and was succeeded by Keller in 1978. The depart-
ment's magnanimity was further tested in 1980
when Keller was appointed the Academic Vice-
President of the University of Minnesota and,
bowing to the greater good of the greater number,
we forged ahead under Davis' dynamic leadership.
This brought in, in 1980, four new faculty
members; two, Cussler and Evans, at a senior level
and two, Jensen and Griffin, at the junior.
The continuing strength and harmony of the
faculty is the final tribute to Amundson's years
of building, for it survived his departure in 1976.
On May 17, 1979, a date not inappropriate to one
of Norweigan descent, we were able to incorporate
Amundson's achievement into the permanent
fabric of the University by renaming the chemi-
cal engineering building after him. As was com-
mented at the time, there is no one in the engi-
neering profession who should be in any doubt
as to the subject taught in an Amundson Hall.
The strength of the department lies not merely
in the achievement of its individual members,
which has attracted the usual round of awards, but
in the easy spirit of cooperation which marks it. At

7Tis building has been named

in honor of


for twently-fiveLyears : cad of dti 'Department
off (lencl-nal tngqieerinq aind nlaterials Science-
and 'Reqents'Profiess1or of Chemical 'Bitinecnri
who. -bl his eximplew and insistence- on the-
highleifracadetmic Standards. has built, not
merely a building. but a departuenr and has
left his mar1dboa od n Othll eniversiti l of
Minnesota and on die- whole profession n f
chemical enqineerint.
No one in the engineering profession would have any
doubt as to the discipline taught in an Amundson Hall.

the graduate level this is manifested in the co-
operative supervision of students, not in any regi-
mented or organized fashion, but as the oppor-
tunity evolves naturally. We have been fortunate
over the years in attracting some excellent
students and this, of course, makes such coopera-
tion easy. This team spirit is fostered at the under-
graduate level by the system of team teaching.
Graduate students spend the first quarter getting
acquainted with the work of the different members
of the department and organizing their prefer-
ences. This results in the assignment to an advisor
at the beginning of the second quarter of their
graduate experience and allows them to get moving
on their research during the first summer. Though
course work continues into the second year, the
student presents a dossier showing the beginning
of his or her research by the seventh quarter and
faces, at this time, an oral preliminary examina-
tion on graduate course work. Beyond this the
student is free to spend full time on research. The
different groups of graduate students jointly or
severally supervised by the faculty have different
modes of operation; some meet regularly for
group seminars but others work rather more in-
dividually. The general attitude is that it is the
subject matter that counts and the way in which
it can be developed in the intellectual context of
the university.
Over the years the department has been


fortunate in its visitors and has entertained for
sabbatical or other visits such persons as K. G.
Denbigh, H. Kramer, E. Wicke, L. Waldmann, W.
Schowalter, H. Brenner, W. Resnick, E. Rotstein,
R. W. Fahien, J. Sinfelt and R. Rosensweig to
name but a few. We also benefit from the adjunct
professorships of not a few engineers in the Twin
Cities area who come in and teach a course or take
a lab with great enthusiasm and devotion, and who
immensely enrich the student's experience by their
industrial experience. K. J. Valentas, now the
Vice-President for Engineering at Pillsbury, has
taught a food engineering course for years: H.
Kramer of General Mills, before his sudden death
last year, had made a major contribution to the
design course. Among the others, W. C. Johnson of
3M and B. Koepke of Honeywell have taught
regularly as have J. Johnson, R. Minday, L. White
and many more.
To summarize the scope and interests of the
faculty is a difficult task and it is best to proceed
alphabetically, letting the grouping and over-
lapping of research areas become evident as we go.

ARIS is interested in any problem that yields a
significant mathematical structure and more generally
in the whole question of how physical systems may be
modeled intelligently. With CARR he shares an interest
in continuous chromatographic reactors of rotating or
counter-current design. Of course, Carr's interests in
kinetics go far beyond this area of joint investigation
reaching to the kinetics of reactions among species
having nonequilibrium energy distributions and to
multiphoton infrared photochemical reaction engineer-
ing. Atomic and radical reactions, particularly those
occurring in the atmosphere and in combustion, are also
one of Carr's interests and he has investigated large
kinetic systems using sensitivity analysis.
CUSSLER brings an interest in the mass transfer
processes of biological systems and this includes work
on corrosion of ionic materials as well as problems in
membrane diffusion. With EVANS he explores the
kinetics of detergency, especially of the bile which is
the human body's detergent. His interests even extend
into the psychology of perception, for, in food develop-
ment, the understanding of the impact of food texture
can only be understood by a combination of engineer-
ing and psychology. Cussler is one of the keenest
runners in the University and has actually participated
in the Boston marathon; with Evans and Griffin, Tirrell
and Macosko we can field a formidable team.
DAHLER, who also has an appointment in Chemis-
try, has research interests that encompass a broad
spectrum of problems in nonequilibrium statistical
mechanics and in the theory of atomic and molecular
collisions. DAVIS also comes from a background in
chemical physics and has applied this in several engi-
neering contexts as well as having continued his work
in the theory of liquids. Thus, with Scriven and others

he has looked into some questions of interfaces,
micelles, thin films, microdispersions and emulsions
and their applications to enhanced oil recovery. Perco-
lation theory has been one of his tools and this has
recently led to an analysis of some problems in surface
diffusion connected with catalysis. With Evans and
Tirrell he has an interest in the study of transport
processes and the statistical mechanics of diffusion re-
laxation and flow. Evans studies micelle formation
under unusual conditions ranging from those that
form in low melting fused salts to those in aqueous
solutions up to 2000C. Using a Taylor dispersion tech-
nique he has also done much work in determining the
diffusion coefficients of small solutes and works with
Davis to link these determinations up with the basic
theory. In collaboration with Cussler he is trying to
understand why organosilicon films are so effective in
inhibiting corrosion.
FREDRICKSON has the reputation of being of that
select band of teachers who can make thermodynamics
both interesting and understandable. It is to be hoped
that he will have time to get his text out soon but as
Director of Undergraduate Studies and Editor of
Chemical Engineering Science he has his hands more
than full. Besides all this he has a very active research
program in the dynamics of bacterial populations and,
in particular, of mixed cultures and the dynamics of
competition and predation.
GERBERICH is an expert in the fields of flow and
fracture, studying hydrogen embrittlement, polymer
fracture, elastic-plastic fracture mechanics and fatigue
crack growth. GRIFFIN joined the department in 1980
after a post-doctoral stay at the Surface Science Di-
vision of NBS and his doctorate at Princeton. His re-
search interests are in the fields of heterogeneous
catalysis and surface chemistry with particular empha-
sis on the reactions that involve oxide catalysts. He
uses in-situ measurement techniques such as infrared
spectroscopy to examine the catalyst under reaction
conditions. His interests link most closely with Schmidt's
and he is one of a group, with Jensen and Aris, which,
under Schmidt's leadership, looks at various aspects of
catalysis and its applications in chemical engineering.
ISBIN, the doyen of the department in virtue of his
more than 30 years of service, is a well-known nuclear
and chemical engineer and, for many years, a member
of the Advisory Committee on Reactor Safeguards. His
expertise on two-phase flow and transport has spread
now to the problems involved in the design of solar
boilers while his experience in reactor safeguards had
led him to a more general concern with risk analysis.
When JENSEN came from Wisconsin in 1980 he
brought with him an interest both in chemical reaction
engineering and in process control. In the first area his
projects concern the fundamental reaction and trans-
port phenomena of catalytic coal gasification and the
chemical deposition of microelectronic materials. He
is also investigating the problem of multiple steady
states, oscillations and pattern formation in combustion
and heterogeneous catalysis. In the control area, he is
our expert on development of algorithms and real time
experiments with the department's PDP 11/60 mini-


Before his elevation to the academic vice-presidency,
KELLER, besides heading the department, conducted
a research program in various biomedical areas, par-
ticularly those involving blood flow and its relation to
artificial organ design, thrombogenesis and arterio-
sclerosis. MACOSKO, the first monomeric unit of the
polymer bridge, is interested in all aspects of polymer
processing. This includes models for processes that
involve network polymerization and relating the re-
action kinetics to the development of network struc-
tures and to the physical properties that result from
these. He is also bringing his theological interests to
bear on one of Scriven's main concerns, that of coating
flows with non-Newtonian fluids. In ORIANI we are
fortunate in having a man of extensive industrial back-
ground who took early retirement from U.S. Steel to
become the first Director of the Corrosion Research
Center. RANZ has applied his understanding of fluid
mechanics and his great skill and feeling for the order
of magnitude of different physical effects to a number
of rather difficult fluid mechanical and mass transfer
problems. His particular emphasis is now concentrated
on new models of laminar and turbulent mixing with
chemical reaction in one and two phase systems, but
aerosol technology continues to be an active subject for
him. As one of the leaders, with Amundson, in the de-
velopment of team teaching 20 years ago his influence
and example are still most valuable.
SCHMIDT, one of the five members of the depart-
ment whose primary training was in chemistry, came
from the University of Chicago in 1965. He leads a
very active group of students in research on many
aspects of surface science and catalysis. He uses tech-
niques, such as Auger and photoelectron spectroscopy,
to characterize reactions on single and polycrystalline
surfaces of transition metals. The dynamics of both
natural and forced oscillations of catalytic reactions
is another of the things he has looked into both ex-
perimentally and theoretically and on which a natural
collaboration with Aris has developed. Schmidt's is the
initiative behind an informal center for catalysis that
has grown up to use the mutually interlocking interests
of Griffin, Jensen and Aris with his own.
Of SCRIVEN's interests a complete article could
be written in itself. His work on porous media has
covered both the application to enhanced recovery and
to the properties of catalysts pellets, packed beds and
foodstuffs; some of this is being done in conjunction
with Davis. A second area covers micro-structured
fluids either in thin films or at interfaces or in dis-
persions. The need for computational results has led
him to pioneer the application of finite element methods
and to develop a range of computer-aided analytical
techniques which allow the generation of numbers to
assist, rather than obscure or dominate, the insight that
can be obtained through mathematical analysis.
SIVERTSEN's current research interest has centered
on the relationship between the structure and various
physical properties of bulk and magnetic thin film
solids, especially the oxide analogs of the alkali halides.
TIRRELL, who also serves as the Director of Gradu-
ate Studies in the Department, is very active in many
areas of polymer science. These include the control and

design of polymerization reactors where experimental
studies are coupled with appropriate mathematical
models. Of particular interest is the copolymerization
of many components where temperature and mass
transfer, reaction kinetics and reactor design all in-
fluence the composition and molecular weight distribu-
tions and properties of the product. He also studies
high conversion free radical polymerization both
theoretically and experimentally and is using dynamic
scattering techniques to develop and apply modern
theories of polymer diffusion in concentrated systems.
Another area of interest is the influence of hydro-
dynamics on the kinetics processes involving macro-
molecules. These influences are exerted through in-
duced conformational changes which in turn affect
some of kinetic processes. TSUCHIYA came to the de-
partment in 1956 as the forefront of the development
of interest in biological and bacteriological matters.
He still is very interested in the role of metabolic
products in the growth of microorganisms and in the
development of continuous processes. On the materials
science side, WALLACE uses positron probing to in-
vestigate the microstructures and substrates of metals
and alloys. He has a considerable knowledge of the
history of metallurgy and has taught an Honors Semi-
nar in that area. WELLINGHOFF is interested,
amongst other things, in polymers which normally
crystallize very slowly in solution by chainfolding but
can exhibit a fine spinodal phase separation along with
rapid formation of fringed micelle crystals to provide
the crosslinks for an extended elastic network. The
interaction of flowfields upon polymer solvents is an-
other part of his research, as is the kinetics and
morphology of polymer phase separation in blends of
polymers and the exciting new area of conducting
It would be futile to try and indicate the range
of the papers published by members of the de-
partment for at the last count that the Dean re-
quired of us there were over a hundred papers
published last year and almost as many the year
before. It may, however, be germane to point out
that several earlier contributions to this journal
have reflected the approach of the department to
the enjoyment of chemical engineering and have
discussed various aspects of its program.* For
some strange reason the department used to have
a reputation for being dominantly theoretical.
This was never really accurate, for there has al-
ways been a predominance of experimental pro-
grams. What we have tried to cultivate is the
lively interaction of theory and experiment that
is the hallmark of fruitful research. How well we
have attained this balance is for others to judge.

*See II, 36-39; III, 48-52; VII, 19-40; X, 2-5; X, 114-
124; XI, 68-73; XII, 148-151; Profiles of faculty are at V,
104-106 1971; XI, 50-52 177; XIII, 8-12 1979. Light-
hearted ephemera appear at IX, 118-119 1975 and XV, 12


Read any best-sellers


I 1-

These DOW CAREERS booklets can help college graduates

At Dow, we're firmly committed to
attracting the best college graduates
every year. And our DOW CAREERS
booklets have been our "best-sellers"
-convincing a lot of people to choose
careers with us.
Each booklet outlines a specific
career at Dow. You'll read about people
who have those jobs now and see how
they like them. And you'll find out about

find uncommon career opportunities.
some typical job challenges-and how
Dow people solve them.
And every booklet explains impor-
tant facts about Dow-like how we're
one of the largest chemical companies
in the world, and yet decentralized so
you don't feel lost in some corporate
Dow gives new employees actual
"hands-on" experience beginning with
the first day on the job. A chance to
develop, to grow, and show us what
they can do. Plus the recognition they
So find out more about the Com-
mon Sense/Uncommon Chemistry

People in our DOW CAREERS book-
lets-and learn about our uncommonly
good careers.
If you know of qualified graduates
in engineering or the sciences, or with
an interest in marketing, finance or
computer science, we hope you will
encourage them to get copies of the
DOW CAREERS booklets by writing
us: Recruiting and College Relations,
P.O. Box 1713-CE, Midland, Michigan
48640. Dow is an equal opportunity
employer- male/female.
Trademark of The Dow Chemical Company 4 0 .
01981. The Dow Chemical Company


f. .. Seade4

University of Utah
Salt Lake City, UT 84112

three months after Lindbergh's flight across
the Atlantic. Being raised in the Bay Area in the
30's and 40's was a plus for any young man since
it was then possible for a bright student of limited
means to aspire to and receive an outstanding
technical education. The Lindbergh flight signaled
a public awareness of our developing, high tech-
nology, society and Bob has been both fascinated
by and a contributor to the development of that
After graduating from high school and spend-
ing one semester at City College of San Francisco,
Bob (at 17 years of age) enlisted in the Navy a
few months before VJ Day. When the war was
over, the Navy reneged on its promise to train
Bob as an electronic technician and put him in
charge of a demobilization interviewing team at
Treasure Island in San Francisco Bay. Once Bob
learned how the system worked, however, he ar-
ranged his own discharge in time to return to
college in the fall of 1946. His C.O. and the Navy
are still trying to understand his ploy, and the
story illustrates why it can be undesirable to have
enlisted men in the Navy who are too bright.
Bob's father wanted his son to be an entertain-
er (and perhaps enter politics) and in his pre-teen
years Bob performed as a tap-dancer, an actor,

... he had the experience of
seeing his designs built and operated.
His most satisfying project was a 100-plate,
215-foot tall, 10 foot diameter deisobutanizer ...
which was the largest distillation column
built by Chevron at that time.

Copyright ChE Division, ASEE, 1982

Bob and a dancing partner, 1935

and a boy soprano in such prestigious and diverse
places as the War Memorial Opera House in San
Francisco and at San Quentin Prison in Marin
County. When his voice changed, however, he de-
cided that it would be best to confine his singing to
the shower and instead began to pursue a life-
long interest in athletics. He had lettered in
basketball in San Francisco but later decided that
desire and a competitive spirit were not enough
for a successful career in sports. Thus, the father's
and son's separate aspirations were now com-
promised to consideration of a career in science
or engineering.
A word of explanation is required to explain
how the nickname "Bob" comes from the initials
J. D. By use of the peculiar thinking that only a
parent naming a son can understand, Bob's father
selected the name "Junior" since he didn't like his
own first name (George) but did want to name a
son after himself. However, Junior is not a name
to be known by on the streets of San Francisco.
The middle name DeVere likely comes from his


mother's Mormon heritage, and is a name to be
used only in Utah. So, with the help of some of his
peers in high school, the nickname Bob was se-
lected out of desperation, and it stuck.
After an additional year at City College of San
Francisco, Bob completed his sophomore year and
transferred to the chemical engineering depart-
ment at Berkeley, graduating at the top of his
class in 1949. He spent an additional year at U.C.
and received the M.S. degree in chemical engineer-
ing while working as the first graduate student
for Charles Tobias. In the fall of 1950, Bob and
his good friend Robert "Bob" Brodkey (now a
professor at Ohio State) both left Berkeley and
went to the University of Wisconsin in pursuit of
the PhD degree. At Wisconsin, Bob worked for
W. R. Marshall as one of Marshall's 16 PhD
candidates. Seader had seldom seen snow in Cali-
fornia, and the winters of 1950-51 and 1951-52
along Lake Mendota motivated him to complete
his dissertation on drying in as short a time as
possible so that he could return to sunny Cali-
fornia. He maintains that he considered jobs with
only two companies after completing his work at
Wisconsin. The companies, both in the San Fran-
cisco Bay area, were Shell Development (then at
Emeryville) and Chevron Research at Richmond.
Since Chevron offered $1 more a month than the
competition, Bob reported to Richmond in Septem-
ber of 1952 and spent seven years at that location,
involved in research and process design. In 1959,
the lure of the unknown and the adventure of our
manned space program attracted him to the aero-
space industry. He joined the Rocketyne Division
of North American Aviation, Inc., at Canoga Park
in Southern California to work on problems associ-
ated with the development of very large liquid-
fueled rockets.
His friends in Northern California couldn't
understand how he could leave the culture of that
area for the "playground" of Southern California.
However, while at Canoga Park Bob met his
future wife, Sylvia, and they were married in
1961. They have four children; Suzanne, age 19,
Robby, age 16, Kathleen, age 13, and Jennifer, age
7. Bob also has four older children, Steven, Clay-
ton, Gregory, and Donald (two sets of twin boys),
from a previous marriage.
When he left Wisconsin, Bob's original in-
tention was to work for seven years in industry
and then become a professor; but 13 years later
he found himself still working in industry.
Characteristically, he decided that it was time to

get back to his original plan and seek a position
at a university. He decided to look outside of Cali-
fornia (but not outside the Western States) and he
subsequently accepted a position at the University
of Idaho. In September of 1965, Bob arrived at
the University of Idaho with the intention of
spending the rest of his days there. But fate, in
the form of a fortune cookie, intervened. At a
Chinese restaurant in Moscow, Bob read that he
was to receive an important telephone call in three
days; and, sure enough, three days later, E. B.
Christiansen (who had been Chairman of the
Chemical Engineering Department at the Uni-
versity of Utah since 1947) called with an offer
for a Professorship. Chris had heard from Bob
Marshall that Seader was looking around, and he

"Where is the computer?"

finally tracked him down in Moscow. Bob was
reluctant to leave his new friends at Idaho; but
being a fatalist and not wishing to alienate the
Chinese, he agreed to join the department at the
University of Utah in the fall of 1966. Even
though Salt Lake City is the headquarters of the
LDS (Mormon) Church, he maintains that was
not a factor in his decision. He is an active member
of the LDS Church and has served in several im-
portant positions in the lay clergy of the church.
Seader was then the sixth member in the de-
partment at Utah, and his many talents and varied
background really strengthened the faculty. In
1975, he became Departmental Chairman, succeed-
ing Christiansen, who had been in that position
for 28 years. Bob served as chairman for a single
three-year appointment and then stepped down
with the comment that it was too easy to become
corrupt in that position, although it is hard to see


what could be gained in an academic department
even if one were corrupt. In fact, Bob really
wanted to spend all his time with his first loves,
research and teaching. His most significant ac-
complishment as chairman was to succeed in
adding three new members to the departmental
staff with a budget increase for only one new
member. This accomplishment still baffles the Uni-
versity administration, and it remained for later
chairmen to reconcile the budget; but through this
maneuver, the department was well staffed at the
time of the recent enrollment surge.
Bob Seader's work has been at the forefront
of the major thrusts in chemical engineering for
the past 30 years. It was planned that way, and
he is fond of setting objectives and then making
sure that the objectives are achieved. A good il-
lustration of his approach and use of objectives
occurred when he was informed that he had been
selected to coach his then-12-year-old son's basket-
ball team. He accepted the challenge but quickly
recognized that there were no budding Julius

While at Richmond, he helped
develop and publish the "Chao-Seader"
correlation for distillation K values. He
is "that" Seader.

Irvings in the group. However, he could teach his
four- to five-foot tall charges something about
man-to-man defense; and the objective was set to
prevent opposing teams from scoring. His success
was almost fatal for the league. In the 11-game
season his team completely shut out the opposi-
tion three times (this was basketball, not base-
ball!) ; and during the first half of one game, the
opponents never got the ball out of their back
court. The league rules were changed the follow-
ing week to disallow the backcourt press.
One of Bob's favorite hobbies is to harrass
Jerry Pimm, the coach of the University's
"Running Utes" basketball team. He firmly be-
lieves that if the coaches had listened to him, the
University of Utah would have been national
champions in 1981, when they started two senior
All-Americans. Pimm good naturedly answers
Seader's letters and, if pressed, might concede
that Bob could be right. In spite of Seader's
interests in athletics, none of the several skiiers
in the department have been able to get Bob on
skis. He is indifferent to all sports having to do
with ice and snow.

"Don't let the other team score!"

At the time Bob worked for Chevron, the
petrochemical industry was in a period of very
rapid growth, and he was involved in promoting
that growth. During one year he worked with
Wayne Edmister in developing the "Standard of
California" data books. Many of the projects that
originated in his process design group resulted in
completed plants, and he had the experience of
seeing his designs built and operated. His most
satisfying project was a 100-plate, 215-foot tall,
10-foot diameter deisobutanizer for the El Se-
gundo refinery, which was the largest distillation
column built by Chevron at that time. A picture
of this unit is shown in the frontispiece of his
book (with E. J. Henley) entitled Equilibrium-
Stage Separation Operations in Chemical Engi-
neering. Bob reads all accounts of L.A. earth-
quakes closely to make sure that the unit still
stands. While at Richmond, he helped develop and
publish the "Chao-Seader" correlation for distilla-
tion K values. He is "that" Seader. Bob suspects
that his reputation is occasionally enhanced by
being in a profession influenced by another Seider
(Warren D.) with whom he has written books and
given short courses.
He also learned something of the pitfalls in
solids handling when he designed a system con-
sisting of a high-speed vacuum rotary filter feed-
ing a rotary steam-tube dryer, which in pilot tests
performed very well when fed with a filter cake.
How was he to know that, during plant startup,
lack of attention to operation of the filters would
permit them to feed a slurry instead of a cake
to the dryer? The slurry set up like cement in
the dryer and plugged it so badly that the tubes


His colleagues consider Bob as a "super-star." Some of the secretaries have referred
to him as "Super-Seader" or "Dr. Dynamo." He does everything well, and it is a humbling experience to see
him work. He appears to be able to master tough technical material after only one reading ...
he requests six to eight papers from the weekly Current Contents in widely varying fields.

had to be completely replaced. His experience at
Chevron is unique among current-day engineer-
ing professors and, even when his research be-
comes very mathematical and abstract, he is able
to draw on that background to relate his results
to the real world. In the current vernacular, "he
has been there."
While with Standard of California, Bob
championed the then-unreliable digital computer
for design and control applications. In 1958, he
directed the development of Chevron's first pro-
cess simulation and distillation computer pro-
grams. His "love affair" with the computer has
never diminished. He was a charter member of
the CACHE committee and is currently the Execu-
tuve Officer and Treasurer for that group. With
that committee, he was instrumental in making
the arrangements for chemical engineering de-
partments to use the FLOWTRAN process simu-
lation program written by Monsanto.
The attraction of Rocketdyne and the aero-
space industry in the 1960's was simply that "that
was where the action was." The aerospace in-
dustry was an ideal place for high technology
work. Bob enjoyed the excitement of the goals of
that time and recalls with wonder how he and his
colleagues would attend night classes or lectures
at UCLA where the materials at the frontiers of
science and applied mathematics were taught and
then try to apply the information the next day to
solve real problems. There was always enough
money to do the job right, and teams of dedicated
and outstanding engineers and scientists were
assembled to accomplish the impossible. Seader
claims that he lost a lot of his "chemical engineer-
ing provincialism" while working with so many
competent people from so many other disciplines.
His own work included fundamental studies of
heat transfer to cryogenic fluids (for which he re-
ceived a NASA Tech Brief Award in 1971), ap-
plications of computers for data acquisition,
control and real-time analysis, and a wide range
of heat, mass, and momentum transport and fluid
dynamics problems. He became a recognized
authority in the area of ablative materials and of
theories of charring-ablation heat transfer.

His research interests at the University of
Utah have been of broad scope and, as usual, have
been in the mainstream of chemical engineering
work. In the latter half of the 1960's, he con-
centrated on kinetics and transport phenomena
in curved tubes. In the early 1970's, when the
vogue was on relevance of research, he served
as co-director of and supervised research in the
Flammability Research Center at the University
of Utah. His primary efforts dealt with characteri-
zation of the smoldering properties of synthetic
and natural polymers and the smoke generated by
pyrolysis and combustion in catastrophic fires. In
1975 he teamed with Alex Oblad and Wendell
Wiser of the College of Mines and Mineral In-
dustries at the University in support of efforts

CACHE Committee receiving grant from Monsanto,

to produce liquid coal and to develop processes for
recovery of oil from the unique tar sands forma-
tions in the state of Utah. He is co-inventor of a
promising patent for a tar sands thermal recovery
process that represents a novel application of heat
pipes. As usual, he continues to find new applica-
tions for the digital computer. A major interest
of his is computer-aided synthesis of separation
sequences using artificial intelligence. Currently
he has a student developing methods for design
of coupled systems of separators. He is not only
conversant with computers but also communicates


with mathematicians. Bob quickly recognizes and
uses applied mathematical achievements that have
engineering applications.
While at the University of Utah, Bob has been
a very effective industrial consultant, and he is
popular as a lecturer. Since 1974, with Dick
Hughes and Warren Seider, he has conducted
AIChE Continuing Education Courses in com-
puter-aided design and in mathematical modeling.
He has lectured in "Spanglish" in Puerto Rico,
Mexico, Brazil (in broken Portugese), Chile, and
Peru. Often, after a lecture tour, he is followed

Seader claims that he lost a lot
of his "chemical engineering provincialism"
while working with so many competent people from
so many other disciplines.

from South of the Border by exceptionally cap-
able graduate students. In an unguarded moment,
most of these students will admit that Bob's
mathematics is easier to understand than his
His colleagues consider Bob as a "super-star."
Some of the secretaries have referred to him as
"Super Seader" or "Dr. Dynamo." He does every-
thing well, and it is a humbling experience to see
him work. He appears to be able to master tough
technical material after only one reading.
Typically, he requests six to eight papers from
the weekly Current Contents in widely varying
fields and when the articles are received, he reads
each one carefully. Anyone answering his request
for a reprint can be sure that the paper will be
read; and if there are errors, one can expect to
hear from him. He has a reputation as an out-
standing reviewer of technical material, and
several standard texts have benefited from his
comments. He swears that he will review no more
texts, but he intends to write more books. His
latest book, to be published by the MIT Press,
deals with the thermodynamic efficiency of chemi-
cal processes.
Bob is one of the most effective and likely the
best-liked teacher in the department and in the
Engineering College. He received the University
of Utah Distinguished Teaching Award in 1975
and has always been very popular with students.
His lectures are concise, well organized, and al-
most totally free from distractive, trivial mis-
takes. His presentations always include many
references to recent practical applications of the

material in areas of immediate interest to the
students. He is a master of bringing material
from his research activities into his courses at
all levels. In addition to carrying a regular teach-
ing assignment in the department, he has de-
veloped a unique and popular second-year gradu-
ate course in advanced concepts of reactor design,
which is based on his experience at Chevron and
at Wisconsin. His senior-level course in computer-
aided design (FLOWTRAN based) draws
students from many disciplines on campus. He
works pretty much in a one-on-one basis with his
graduate students and has a weekly conference
with each student in which it is well understood
that some progress in the last week had better be
reported. Somehow he gets an enormous amount
of work out of all students, but no one ever com-
plains. There are no threats, no posturing; and he
is neither a hard nor an easy grader. All students
simply like to work hard for him.
The opinion held by people who have worked
with Bob for many years is best expressed by a
comment whose source has been forgotten, "It
is almost too bad that Bob is such a nice guy; if
he were real nasty, I could at least feel superior
to him in something." 0

Persons with a degree or job-experience in various
fields to teach adults for one year should write:
Personnel Director
International Education Services
Shin Taiso Bldg.
10-7, Dogenzaka 2-chome
Tokyo 150 Japan
Instructors employed by I.E.S. will teach Japan-
ese businessmen and engineers English as a second
language and the terminology in their own field of
Japanese language and teaching experience are
not required. Teacher training is given in Tokyo.
Further information can be obtained by providing
I.E.S. with a detailed resume and a letter indicating
an interest in the position.
Interviews will be held in Seattle, San Francisco
and Los Angeles in June, 1982 and selected ap-
plicants are expected in Tokyo July through Oc-
tober, 1982.



I /

Sometimes it's not all it's
cracked upto be.

However, at Union Carbide innovation continues to improve peoples' lives.
Union Carbide pioneered the petrochemicals industry. Today the Corporation's many hun-
dreds of chemicals are used in everything from automobile bumpers to shampoos. A leader in
the field of industrial gases, our cryogenic technology led to the development of the Oxygen
Walker System, which allows mobility for patients with respiratory diseases. Union Carbiders
are working on the frontiers of energy research-from fission to geothermal-at the world
renowned Oak Ridge National Laboratory in Tennessee. Our revolutionary Unipol process
produces polyethylene, the world's most widely used plastic, at one half the cost and one
quarter the energy of standard converting processes.
From sausage casings to miniature power cells, the Union Carbide tradition of innovation
extends beyond research and development activities to our engineering groups, manufactur-
ing operations, and sales forces.
Continued innovation will largely spring from the talents of the engineers and scientists who
jofn us in the 1980's.

We invite you to encourage qualified students
to see our representatives on campus-
or write to:

' :

an equal opportunity employer

Manager, Professional Placement
Union Carbide Corporation
Old Ridgebury Road
Danbury, Conn. 06817

4caMd .w ectae


Both Theory and Strategy

Carnegie-Mellon University
Pittsburgh, PA 15213

EDITOR'S NOTES This is the second and con-
cluding installment of Professor Westerberg's
1981 ASEE Award Lecture. The first installment
appeared in the Winter 1982 issue of Chemical
Engineering Education (Vol. 16, No. 1, page 12).

ASCEND-II: An Analysis Aid for Arbitrarily
Configured Processes

We shall move off on an entirely new tack at
this point and describe briefly the ASCEND-II
flowsheeting system (Locke, et al (1980)) that
we are developing in my research group at
Carnegie-Mellon University. The persons involved
are Michael Locke (Locke (1981)), Selahattin
Kuru (Kuru (1981)), Peter Clark (Clark
(1980)), Dean Benjamin and Andrew Hrymak.
The messages to be conveyed by this example are
two: the breadth of research activities which
support this project and a description of the use
of this system to develop a working analysis model
for a process in a manner which is consistent with
the design strategy that has been the main theme
of this paper.
To examine ASCEND-II we need first to es-

Arthur W. Westerberg received his degrees in chemical engineer-
ing at Minnesota, Princeton, and Imperial College, London. He then
joined Control Data Corporation in their process control division for
two years. In 1967 he joined the University of Florida where he
remained for nine years. In 1976 he joined the faculty at Carnegie-
Mellon University. He was Director of the Design Research Center from
1978 to 1980 and just became Head of Chemical Engineering this

tablish what we mean by analysis. We include
the following types of analysis for a given but
arbitrarily configured process flowsheet.
1) Simulation. The inputs to the process, the tempera-
ture and pressure levels at which to operate and the
equipment sizes are fixed. The calculation is to dis-
cover how the equipment performs, a rating calcula-
2) Design. Some outputs from the process and some
intermediate stream variable values may be specified
in exchange for calculating an equal number of the
inputs, levels of operation and/or equipment sizes.
3) Dynamics. The dynamic behavior of a process may be
4) Optimization. We may wish to optimize the process
over the set of continuous variables that describe
equipment sizes and process operating levels.
Fig. 4 illustrates the breadth of questions
which one can address in the area of design re-
search. Many persons identify design research

FIGURE 4. Aspects of Design Research.

Copyright ChE Division, ASEE, 1982


Unlike conventional flowsheeting systems, each unit with a flowsheet can
be tested by itself in ASCEND-II, permitting a bottom up solving of the units at any
time. In this mode and using the simple model as the base case design, much testing can be done
to see where to add complexity, and where perhaps to remove complexity.

with only the two aspects highlighted with a dark
line: 1) Identify Abstract Problem and 2) De-
velop Relevant Mathematical Theory. We have
been arguing all along about the importance of
developing a correct design strategy. Support
techniques are often shrugged off as not funda-
mental enough, but, if not done correctly, the
implementation of the theory will likely prove too
complex to be practical. Finally one should not
overlook the problem of placing sophisticated tools
into the hands of unsophisticated users. There is
research lurking there too.
The abstract problem for developing
ASCEND-II is how to solve large sets of simul-
taneous nonlinear, sparse algebraic, ordinary and
partial differential equations, perhaps subject to
inequality constraints and perhaps containing
discrete variables. There is certainly enough of a
problem here to require considerable effort.
Relevant theory includes convergence proofs,
analysis techniques to take advantage of structure
and Lagrange theory. We have already discussed
strategy ideas at length. The supporting tech-
niques include consideration of data structures,
problem decompositions (see Westerberg and
Berna (1978), Berna, et al (1980), Clark (1980)),
data bases and use of network computing. Finally
the ideas involved in placing the tools into the
hands of the practitioner include language design,
level of interaction, online documentation system
design and use of graphics. We are making con-
siderable progress at dealing with the above ideas
and others in the development of ASCEND-II.
Fig. 5 illustrates the underlying evolutionary
aspect of ASCEND-II.
ASCEND-II is intended to help a process
engineer "design" a computer model for his pro-
cess, using the available building blocks provided
within the program. "Design" here refers to find-
ing and solving a model of the needed complexity
to answer the questions being asked of the pro-
cess, where the engineer is learning both about
the questions he should ask and about the model as
he proceeds. We could broaden the meaning of
design to that of designing the process for which
the model is being developed, a task for which
ASCEND-II is also well suited, but we want to

limit ourselves here to the narrower model design
The axes in Fig. 5 are axes along which the
model design can evolve. Model complexity can
evolve from simple to complex, where simple
models consist of only a few units and the use of
the simplest of physical property models; e.g., a
flash unit using constant relative volatilities.
With each model a range of analysis types can
be performed, starting with simulation, moving
to design and finally (when ASCEND-II is
further developed) to optimization. Simulation
is intuitively the easiest mode to use for the
engineer. In that mode he can usually establish a



LG ________- MODEL

FIGURE 5. The ASCEND-II Flowsheeting System.

set of specifications which will lead to a solution
for the remaining variables. For example, one
has some confidence that, if he fixes the feed
stream to a flash unit, fixes the fraction of the
feed which will vaporize and the flash pressure,
then the flash unit will have to operate and so will
the corresponding calculation. Why not allow the
user to start then with this "comfortable" calcu-
lation? Once he can simulate the flash unit
ASCEND-II allows him to alter the set of vari-
ables to be specified. For example, he could re-
quire that the recovery fraction of one of the
components be specified and that the pressure be
calculated. If the trade is illegal, he will be warned
Running through a few design calculations
will acquaint him with the shape of the solution
space and when he gets near to a good solution,


he can switch into doing an optimization calcu-
Once this sequence is solved using simple
algebraic models, he can selectively add more
complexity to the model by adding more units
and/or more sophisticated physical property
calculations and continue.
A type of complexity which can be added is
to broaden the type of equations which are used
to model portions or all of the process, i.e., by
allowing models involving ODE's (Kuru (1981))
and PDE's to be introduced. With ODE's and
PDE's one can consider doing dynamic studies.
The last axis is that reflecting the degree of
interaction ASCEND-II will have with the user.
In ASCEND-II a standard command file can be
created which will attempt to solve any model once
it is set up. Invoking this "standard" file is like
running the problem in batch mode on a computer.
At the other extreme, the commands can be exe-
cuted interactively one at a time in fairly arbitrary
order. (The computer is a DEC-20 which pro-
vides a very friendly interactive environment.)
Examples of the types of commands available

The heat exchanger network
synthesis problem epitomizes the
effective use of approximate criteria
to locate excellent final network designs.
Using thermodynamic arguments, one can
predict a priori the least amount
and kind of utilities needed.

are 1) to input some more structure to the flow-
sheet, 2) to delete some of the existing structure,
3) to save and retrieve variable values, 4) to
initialize variable values (selectively), 5) to
change the set of variables whose values are to
remain fixed, 6) to cause variables and equations
to be rescaled to reflect current variable values, 7)
to do one or more Newton-Raphson iterations, 8)
to determine the constrained derivative of one
variable with respect to another, 9) to display
variables selectively, and 10) to display equation
errors selectively.
With this structure for ASCEND-II, the user
can "drive" his computation around computational
obstacles much as he drives a car and can become
very effective at getting solutions quickly, even for
stubborn problems.
We have set the stage now to argue that
ASCEND-II allows the model for a given process
to be designed using our earlier guidelines.

Clearly the first guideline is dealt with: evolving
from simple to complex. The depth first strategy
can be followed by developing first a simple model
for the entire process.
Unlike conventional flowsheeting systems, each
unit within a flowsheet can be tested by itself in
ASCEND-II, permitting a bottom up solving of
the units at any time. In this mode and using the
simple model as the base case design, much test-
ing can be done to see where to add complexity,
and where perhaps to remove complexity.
Answers can be obtained to a simpler version of
the problem to use as a starting solution point for
the more complex versions, a strategy often needed
when solving highly nonlinear equations. The
notion of developing and using approximate
criteria is also possible. One usually gets a solu-
tion to the equations, perhaps far from the desired
solution point. This solution may be from doing
a simulation rather than the desired design calcu-
lation. Not unlike the idea behind a continuation
method, the calculation can be converted to the
desired design calculation in terms of which vari-
ables are specified. Then one can move to the solu-
tion point desired through a series of small steps,
converging to the solution at each step.
While it is obvious that much of the power of
a program like ASCEND-II comes from its being
interactive, it is equally as obvious when using
it that the ability to find a base case solution and
then to move from that solution in almost any
manner desired (top down/bottom up, simula-
tion/design, etc.) is the heart of the rest of its
power. It is the learning that can occur which
helps to decide the nature of the next calcu-
lation, to see its impact and to alter one's path as a
consequence, that makes ASCEND-II so useful.
Traditional flowsheeting systems (and for that
matter, traditional equation solving packages) do
not offer the flexibility provided by ASCEND-II
for this approach.
ASCEND-II has been designed under the as-
sumption that calculations will often fail until one
learns about the problem. Diagnostic tools are
thus provided to allow the user a chance to detect
where the failures are occurring. As mentioned
only briefly before, these include interactive access
at any time to every variable in the problem by
a convenient name and similarly to every equation
error. This latter access allows one to note, for
example, that the phase equilibrium equations on
stage 3 of the diethyl ether column are not con-
verging. The variables around that stage can


"Design" here refers to finding and solving a model of the needed complexity to answer
the questions being asked of the process, where the engineer is learning both about the questions he
should ask and about the mode las he proceeds. We could broaden the meaning of design to that of
designing the process for which the model is being developed, a task for which ASCEND-II is
also well suited, but we want to limit ourselves here to the narrower model design problem.

then be examined to see if one is perhaps too large
or worse yet, negative. Having located the
problem, the user can then start to work on
correcting it.
Interestingly, the current version of this
system is a third generation version. We designed
ASCEND-II following our guidelines by proto-
typing it twice, at each step improving the design
based on the previous version. This was and re-
mains a deliberate policy for creating ASCEND-II.

Heat and Power Integration of a Process

T HE LAST PROCESS PROBLEM to be considered is
to integrate the heat and power requirements



I '
0 W



FIGURE 6. Minimum Utility Calculation. Upper right
illustrates how to extract "work" efficiently from

for a process for which one has just set tempera-
ture and pressure levels for each of the units
and has solved the process heat and material
balances (using ASCEND-II for example). Great
progress has been made for solving this problem.
The heat integration portion is usually called a
heat exchange network synthesis problem. See
Nishida et al (1981) for an extensive review of
the heat exchanger network synthesis problem.

The heat exchanger network synthesis problem
epitomizes the effective use of approximate
criteria to locate excellent final network designs.
Using thermodynamic arguments, one can predict
a priori the least amount and kind of utilities
needed to solve this problem. Also using graph
theoretic ideas one can guess the fewest number
of heat exchanger units likely to be needed. Ex-
perience has shown that the better designs meet
these goals, or come very close to meeting them.
Finally, effective design techniques exist to aid
one to find such designs.
In preanalyzing the heat integration problem,
one discovers for most problems a bottleneck will
occur to further heat integration in the form of a
temperature pinch. Fig. 6 illustrates the way Hoh-
mann (1971) located this pinch. He merged all hot
streams into a single "super" hot stream and all
cold into a single "super" cold stream. Placing
them as illustrated on a temperature versus total
enthalpy diagram reflects the opposing hot and
cold stream temperature profiles one would see if
these super streams met in a single counter
current heat exchanger. The pinch is the point
which precludes further integration. Cerda (1980)
from our group and, in parallel, Mason and Linn-
hoff at ICI recently developed an approach which
generalizes this minimum utility calculation.
Umeda et al (1979) have exploited the pinch
to aid in locating where in a process the re-
establishing of temperature and pressure levels
will permit more heat integration. Very recently
Linnhoff and coworkers (Townsend and Linnhoff
(1981), Dunford and Linnhoff (1981)) have
shown how to exploit large temperature differences
between these super streams which occur either
entirely above or entirely below the pinch. They
show how to convert heat entirely to mechanical
work or obtain some "free" separation work
within a process. For example, the upper right
part of Figure 6 shows how one can place a turbine
to get 100% of the thermal energy which must
be added into the process converted to the desired
mechanical work. The cost is the degrading of the
thermal energy which enters and is later rejected
by the turbine. If that energy can be degraded


and still be rejected at a temperature where it is
useful as heat input to the process and if that heat
can be extracted and rejected entirely above or
entirely below the pinch, then 100% of the extra
energy added to drive the turbine is converted to
The design strategy is to establish first a pro-
cess design not yet heat integrated. Then by
examining the process, one finds the pinch
temperature and predicts the minimum utility
costs associated with the process. Next one can
modify the process near the pinch if further heat
integration is desired. Finally one can place some
turbines if possible so they degrade thermal
energy either entirely above or below the pinch.
The design can then be reassessed and improved
from this thermally integrated base case.


How would one "prove" that the design guide-
lines are basically sound? That problem is itself a
design problem and should be (if we are correct)
solved using a strategy consistent with the guide-
lines themselves. The concept should be recursive.
In our case it is leading to the design and testing of
the ASCEND-I system. We are only at present
proving we are right by demonstrating how
rapidly one can put together a working computer
model for a process using this system. In one
example a model was constructed using a .con-
ventional flowsheeting system and the exercise
took two full time days. Using ASCEND-II, it
took two hours.
Since teaching these guidelines to our students
in the undergraduate design course, we see a
noticeable reduction in the time needed to get
realistic designs.

The guidelines suggested to aid one to do
design more efficiently have been illustrated on
several diverse problem types. Only qualitative
"proof" exists as to their correctness. If correct
a principal use can be to examine a proposed or
existing design tool (or design effort) to see if it
abides by them. Where it fails should suggest
modifications to the tool which could significantly
change its effectiveness. Designers have to make
a conscious effort to stick to the guidelines as

they do not always coincide with the most natural
approach. They can be taught; we try to do so in
the undergraduate design class. O

Berna, T. J., M. H. Locke and A. W. Westerberg, "A New
Approach to Optimization of Chemical Processes,"
AlChE J., Se(1), pp 37-4 (1980).
Cerda, J., Transportation Models for the Optima Synthesis
of Heat Exhanger Networks, Ph.D. Dissertation,
Carnegie-Mellon Univ., Pittsburgh, PA 15213 (1980).
Clark, P. A., An Implementation of a Decomposition
Scheme Suitable for Process Design Calculations, MS.
Thesis, Carnegie-Mellon Univ., Pittsburgh, PA 15213
Douglas, J. L, Manuscript for Process Design Text
Dunford,H. A. and B. Linnhoff, "Energy Savings by Ap-
propriate Integration of Distillation Columns into
Overall Processes, Paper No. 10, Cost Savings in
Distillation Symposium, Leeds, July 9-10, 1981.
Hohmann, E. C. Optimmm Netwoorks for Heat Exchange,
Ph.D. Dissertation, Univ. of So. Calif. (1971).
Kuru, S., Dynamic Simlation with an Equation Based
Flowsheeting Systm, Ph.D. Dissertation, Caregie-
Mellon niv., Pittsburgh, PA 15213 (1981).
Locke, IL HL, S. Kuru, P. A- Clark and A. W. Westerberg,
"ASCEND-H: An Advanced System for Chemical
Engineering Design," 11th Annual Pittsburgh Confer-
ence on Modeling and Simulation, Univ. of Pittsburgh,
May 1-2, 1980.
Locke, I. EL, Cotpter-Aided Design Tools Which Ac-
commodate an Evolutionary Strategy inp Eieeriug
Design, Ph.D. Dissertation, Carnegie-Mellon Ulniv,
Pittsburgh, PA 15213 (1981).
Motard, R. L. and A. W. Westerberg, Process Sythesis,
AIChE Advanced Seminar Lecture Notes, AIChE,
New York, NY (1978).
Nishida, N. G. Stephanopoilos and A. W. Westerberg,
"A Review of Process Synthesi AChE J., 27(3),
pp 321-351 (1981).
Seader, J. D. and A. W. Westerberg, "A CaDmbined Heur-
istic and Evolutionary Strategy for Synthesis of
Simple Separation Sequences," AlChE J, 23(6), pp
951-954 (1977).
Simon, H. A., The Sciences of the Artiicial MIT Press,
Cambridge, XA (1969).
Townsend, D. W. and Heat and Power Net-
works in Process Design-Part 1, Criteria for Place-
ment of Heat Engines and Heat Pmps in Process
Networks." sbmitte for publication (1981).
Unmedaa T., K. Nida and K. Shiroko, "A Thenanoynami
Approach to Heat Integration in Distilation Systems"
AIChE J, 25(3), pp 423-29 (1979).
Westerberg, A. W. and T. J. Berma, "Deoumposition of
very LargeScale Newton-Raphson Based Flowsheet
ing Problems," Computer and Chenical Engg. ., 2,
pp 61-3 (1978).





The Sunnor Sehool for ChemAiesl Enginoring
Faculty is organized by the Chemista Engineering
Division of the Afferiesn Seiety for Engineering
Eduation and is held every fve years. Thoe 10 O
Sfumier Shool, ninth in a series begun in 181,
will he held August -1= at tho Univmrsity of Cali=-
fornia in Santa Barbara and has been generously
upprted by indu1stri nal eoB ItS purpose is to
quaint faulty with new developments in hefmi-
esl engineering ednaetion and to provide for per=
sonal intoreation with other faculty and irepre
sentativds from industrial fris onorrned with
the eduestional process, The 19 Summir Slhwool

i Meing oo=mhaired by T: W: FPash rissll amd
Stanfley 1: Sandler of the Unfiversity of Dolawaef:
In February of 1o program and egistration
information was maidd to ohlimiai 0ngimerifg
department hairffnn in the Unitd States, CaR&-
da and Mexioo: Att4ndafno at the smmer hlool
must he limited due to Bonstraintf on both elass=-
room sps and living a@em dations and the
Chemist Mnigineering Division has oteod to lhave
ea@h hafirperson f t an attend@@ from their d&-
SummarFy information oneorinfig the topis to
hoe ovgerd at the Summer & Woi is shows in
Figure 1 beoow,

NounS I
P iif Efi6iWiT& &WsIW i #S &M 6Sv W hW L1W wI6W iIAW Pii I/ MW 1IiL
P&60 ONw AIMAI W #k Rt &M# ffika'w# i'6EWMcYi'MW 6f Mw I IW' I#M
&II 'ITII "A AV m lWIL 6# W AV -IIALL I TL ;1byW R IIIT -= 4L-
S 1T\ :.- ~.S fk: f,, -: $& #'4 : & ~ ep LL : f-*eer
MIWOW ll iirg i II I, 1F il~ -tii Ir I [ &WssIII TAWT iiI& Aws r !Ye I &fr hf'f -
A DrAY piiii&k 9w'wI &m 9wPbi PIiTrs g~fMlliiisy 1, 9w 0 Ni &i WOO i
WftIW il'6Iae ft~M fiiir ## a 6isy i Ii ii6q ii tit 9 WiW &RA f sW y'sf IfFF

Yisyll~W6 69 'l-#66W M s-ii iir fl-a5t 9 ~r 9t 'ii f rssII iLc4ii A Ws6iWi1Tssf w1is II ',' 6 i s
AInwsiW,1 'Wos6W~iBi ii' c ^ stk f 6 Eis w Msfi II't siW 6tsw1si' IdXMP6W 6 _ iWiiW _ ___ i'i'__f .

LWifi W Pst6e 9lyiiW 6sisisws ruhrwstfi f iir 9rwi 6 EI iiwi Ai' LWi'ri iiiij 9w e*s f |j
(#IIIr(4 F sP sriI 64e &W*its iip IIJA s Ms l l if'f tj II !i4w9y 49gm. s ~ vi L i Ffis '
fW TllFj il l fTm iwi w Ig= f i &lv ii& s A ff '4 lt lr l & IW f&ir 1r M W i r 1',s ii
6i',itW IPse54w~i~li ^6y6Wws' iw'i'j 'Msfi' p 6W L(i59i',lWI 4Wii'sW7 #SWAS 6W66W9#19iiwYSW 6 roEu fevwfiuIl
"M1WiLerwommyi i&wm & piWa &s 6v fw t'i9 6SIH'T6 &Eo
Abl"IrK &P'9d9il9 0Ift1~Ww 6W1 I' L6Y611w6~16 1W !wIIItt 4Ir & Ws6

amsNlg l

Lena lecture


Eastman Kodak Company
Kingsport, TN 37662

chemical and physical operations and the se-
lection and interconnection of equipment to imple-
ment these operations to effect desired chemical
processing transformations. Synthesis is the first
of an iterative set of process design activities
which also includes analysis, evaluation, and op-
timization. Significant progress has been made,
particularly since the advent of the digital com-
puter, in the development of a more scientific and
less empirical framework for modeling, simulat-
ing, and improving equipment and operating
parameters for a given design. However the
necessarily prerequisite invention of the design,
the generation of the processing structure repre-
sented as a flowsheet, remains largely a creative
art. Economic competitiveness has demonstrated
the importance of correct structural choices in
meeting process objectives. This realization has
led to present efforts to investigate the possibility
of formalizing the synthesis activity. The ultimate
goal is not just to invent technically feasible de-
signs, but to produce structural configurations
that when analyzed, evaluated, and optimized will
prove by design objective criteria to be superior
to (possibly all) other structural arrangements.
Inventing flowsheets is, of course, not new and
some guidance embodied in design rules-of-thumb
based on experience developed through trial and
error has existed for many decades. However,
formal consideration of structural generation

The ultimate goal is not just to
invent technically feasible designs, but to
produce structural configurations that when analyzed,
evaluated, and optimized will prove by design
objective criteria to be superior to ...
other structural arrangements.

lC Copyright ChE Division, ASEE, 1982

J. J. Siirola received his B.S. from the University of Utah and his
Ph.D. from the University of Wisconsin-Madison in 1970 where he
developed the AIDES process synthesis system and coauthored the
introductory text, Process Synthesis. Besides a continuing involve-
ment in synthesis technique development, implementation, and ap-
plication, his research interests also include simulation and optimiza-
tion aspects of computer-aided design, non-numeric programming,
artificial intelligence, and technology assessment. He is currently a
Research Associate in the Eastman Chemicals Division of Eastman Kodak

dates only from the late 1960's. Some investi-
gators studied the general overall invention of
complete flowsheets, while others concentrated on
specific process synthesis subproblems such as
reaction paths, reactor network configurations,
separation train systems, energy recovery net-
works, fault trees, and safety and control systems.
Several excellent reviews with extensive bibli-
ographies [(Hendry, Rudd, and Seader (1973),
Hlavacek (1978), Westerberg (1980), and
Nishida, Stephanopoulos, and Westerberg (1981)]
chronicle the developments and progress made
in these areas to the present time. Here we will
consider some of the underlying paradigms which
have been suggested for continuous process syn-

Process synthesis has been defined as the in-
vention of the process structure. This definition
can be extended to include invention of associ-



ated control and safety systems, development of
steady state and transient (start-up and shut-
down) operating procedures, and other similar
design specification activities. Synthesis should be
differentiated from 'flowsheeting,' a term which
refers to the simulation, often by computer, of an
existing structure.
The number of tasks that must be performed
to convert available raw materials into desired
products and environmentally acceptable by-
products may be quite large. Even given the con-
cept of unit operations, the wide range of existing
technologies capable of accomplishing these tasks
and the multitude of ways in which these tech-
nologies can be interconnected to achieve the de-
sign objectives with varying efficiency leads to
synthesis problems characterized by discrete de-
cisions and combinatorial difficulties. Further-
more, although the existence of some feasible pro-
cessing structure is usually not in question, the
multivariable and sometimes subjective nature of
the design objective criteria (safety, reliability,
operability, environmental impact, feedstock se-
curity, quality, resilience, familiarity, experience,
as well as capital investment, operating expense,
etc.) makes it difficult to quantify the concept of
'optimum.' Often it is desirable to generate a
series of structural alternatives judged to be
nearly optimal by some economic criteria for
further subjective evaluation.
Despite the fundamental importance of syn-
thesis in chemical process design, few existing
design texts give more than passing reference to
the necessity for first establishing a workable
manufacturing process for producing the desired
product, preferring instead to concentrate on the
more deductive and quantitative aspects of
analysis, evaluation, and (parameter) optimiza-
tion. When mentioned at all, more is noted about
such attributes as experience, self confidence, con-
structive discontent, open mindedness, and demon-
strated intuition and common sense of engineers
who appear to be successful at synthesis than
about invention techniques themselves. However,
some qualitative concept discovery aids do seem
to be used by successful designers. The most basic
of these is historical information from such
sources as engineering textbooks, handbooks, cata-
logs, proprietary company reports, commercially
available comparative process analyses, and open
scientific and patent literature, experience, and
rules-of-thumb. Design concepts have also been
discovered through individual or group creative

Often more than merely
the invention of a feasible design,
process synthesis involved an optimization
over structure.

efforts. These include functional and morphologi-
cal analyses (mechanical schemes for forcing
associations among desired design goals), imita-
tion of nature, analysis of models of other designs,
and brainstorming. These methods and aids rely
heavily on analogies with or adaptations of past
working designs and none have explicit provisions
for generating optimal structural arrangements.
Are there attributes specific to the synthesis
problem that should be exploited and developed
into solution procedures not necessarily re-
sembling techniques now employed by successful
designers? Are there aspects of 'artificial intel-
ligence' research, a relatively new branch of com-
puter science concerned with understanding and
simulating intelligent problem solving, that might
contribute to a formalism for process invention?
Early in systematic process synthesis research, op-
timization, evolution, and decomposition emerged
as potential invention approaches. These remain
the dominant framework for synthesis develop-
ment today.

Often more than merely the invention of a
feasible design, process synthesis involves an op-
timization over structure. A natural approach,
and in fact one of the earliest proposed, is to
combine structural optimization with equipment
and operating parameter optimizations now
routine because of recent advances in computer-
ized numerical methods. In fact, the synthesis
problem has been formally defined as a nonlinear
mixed integer and continuous variable optimiza-
tion problem. Equipment selection, interconnec-
tion, and other discrete decisions are represented
by integer variables (often zero or one) and
constraints among these variables. The difficulty is
that no code exists to solve this generally exceed-
ingly large mixed optimization problem.
One approach is to invent a superstructure in
which all technologies under consideration and all
their possible interconnections are embedded. As-
sociated with each branch of an interconnection
leading from one item of equipment to several


others is a 'structural parameter,' a continuous
variable which represents the fraction of flow to
be split to each of the alternate destinations, As
fractions, structural parameters are constrained
to be non-negative and to sum to unity for all
branches from a junction. These structural pa-
rameters are included with other equipment and
operating parameters in an overall design opti-
mization, Should a structural parameter be
optimized to zero, the corresponding interconnee-
tion is deleted from the superstructure, and should
all flows associated with an item of equipment
be zero, it too is deleted from the structure. In
effeet, the discrete decision making aspects of

Considerable savings may
result f pairs of complementary heating
and cooling tasks can be made to drive each
other rather than require the use
of auxiliary utilities.

process synthesis are replaced by a continuous
nonlinear constrained optimization for which
solution techniques exist,
Unfortunately, the method does not provide
guidance for the initial generation of the super.
structure. The larger the superstructure, the
greater will be the number of structural pa-
rameters and the size of the corresponding op-
timization. The problem of assuring that the
optimal structure has been included within a
given solution formulation is one which arises re-
peatedly in process synthesis research, In the
present case, human experience, rules-of-thumb,
and other design aids are used to select the
equipment and interconnections to be included in
the superstructure.
Because of the size of the optimizations re-
quired, the structural parameter method has been
applied primarily to smaller synthesis sub.
problems (generally fewer than ten interconnec-
tions) such as reactor and distillation networks,
Several solution techniques have been investigated
including simultaneous and sequential optimiza.
tion of the structural parameters and other de-
sign variables by a variety of codes. The feasible
region is often multimodal and search methods
starting from different points may be required to
find a global optimum. The convergence rate of
nonlinear constrained optimization is often slow,
and transformations have been suggested to
eliminate constraints on the structural pa.

rameters, However, inequality constraints on de-
sign variables may result in a discontinuous
feasible region boundary necessitating discrete
decisions the structural parameter method was
developed to circumvent, Inequality constraints
associated with equipment deleted from the super-
structure during structural parameter optimiza-
tion can force constrained behavior in the remain-
ing structure. For many problems a completely
continuous formulation may not be possible.
The optimal assignment formalism from
operation research is another optimization ap-
proach that has been applied to the synthesis of
heat integration networks. This synthesis sub-
problem concerns the heating and cooling tasks
which must be performed within a process. The
problem statement includes a complete description
of the streams to be heated or cooled, descriptions
of auxiliary utilities capable of performing the re-
quired tasks, and descriptions and design equa-
tions for equipment (furnaces, reboilers, con-
densers, and other heat exchangers, etc.) useful
for implementing the tasks, The object is to
synthesize a scheme to accomplish these tasks in
some optimal manner, for example at minimum
net present cost. Considerable savings may result
if pairs of complementary heating and cooling
tasks can be made to drive each other rather than
require the use of auxiliary utilities.
The discrete process of selecting the sequence
of energy interchanges among streams and utili-
ties to be heated and cooled is accomplished by
dividing the heat loads of these streams into suit-
ably small, equal-size elements. Subject to thermo-
dynamic temperature constraints, a cost can be
associated for the 'allocation' (effected by a tiny
exchanger) of heat from any element of a stream
or utility to be cooled to a suitable element to be
heated. If it is assumed that these costs are inde-
pendent of each other, then linear programming
can be employed to perform an optimal allocation,
Later, if multiple adjacent elements for two
streams are found to be paired, the tiny ex-
changers are merged into a larger countercurrent
unit, In this manner, the amount of heat to be
transferred in each exchanger in the structure
has also been made discrete.
Different kinds of structures can be synthe-
sized depending on whether the original streams
are divided into elements sequentially, or partly
in parallel resulting in stream splitting, and also
on how exchangers are merged. In fact, networks
that involve minimum total exchanger area can be


synthesized graphically without linear pro-
gramming by recognizing that maximum thermo-
dynamic efficiency for the total network results
when the temperature driving force is mini-
mized in individual exchangers.
In addition to problems related to choosing the
size and division of the heat load elements, the
principal difficulty with the optimal assignment
approach is that the synthesis problem is not
linear and the match costs are not independent of
each other. Large merged exchangers cost less
per unit area than small ones. Near-minimum
area networks generated by this method are
characterized by a large number of small ex-
changers, whereas minimum cost networks tend
to have much fewer exchangers. Furthermore,
some versions of the method automatically ex-
clude the warmest portions of the streams to be
heated and the coolest parts of the streams to be
cooled from integration, assigning them im-
mediately to an auxiliary heating or cooling utility
which, however, might not be the optimal policy
where multiple utilities of varying cost are avail-
Recently, it has been suggested that process
synthesis problems be solved by mixed-integer
linear programming techniques. Discrete vari-
ables not only indicate the existence of specific
equipment and interconnections but also make
possible very good approximations of many non-
linearities inherent in real systems such as fixed
plus variable operating costs. Mixed-integer linear
programming codes, which generally use branch
and bound algorithms to handle the discrete vari-
ables, exist which can handle quite large problems
of this type. The question remains whether it is
more efficient to let such a generalized algorithm
blindly find appropriate values for numerous dis-
crete variables or whether specific features of the
synthesis problem and other external procedures
should be utilized to set combinations of variables
known in advance to be feasible and expected to
be near-optimal.

Another class of process synthesis approaches
attempts to systematize the methods and aids that
many successful engineers are thought to use to
improve upon initial design concepts. These ap-
proaches assume the existence of a feasible design
and examine it for evolutionary improvement op-
portunities. The initial design may be a com-

mercial, laboratory, or literature scheme for the
same or similar product or process, or it may have
been obtained by the application of a series of
design rules or by some other synthesis technique.
Some methods use cost and functional analysis or
thermodynamic analysis to detect weaknesses and
suggest structural modifications, whereas others
concentrate on systematic reordering of design
rules, or processing tasks. Two-level Lagrangian
analysis can confirm the advantage of an evo-
lutionary change, sometimes without requiring a
complete optimization of the revised design.
Evolutionary methods for structural improve-
ment are quite appealing, possibly because they
resemble much historical process development.
They have been applied to almost all synthesis
subproblems including heat integration, separa-
tion systems, and even entire flowsheets. These
methods, like the embedded superstructure op-
timization approaches, are concerned with the
structural optimization aspects of process syn-
thesis. Their success depends critically, as do
most gradient-type searches in highly irregular

Evolutionary methods for structural
improvement are quite appealing, possibly
because they resemble much historical
process development.

domains, on the nature of the starting point.
Given a poor initial design, they will rarely evolve
to a truly novel globally optimal flowsheet.

To find new, less prejudiced, or more creative
solutions for synthesis problems, it may not be
desirable to use existing designs singly as the
starting point for evolutionary approaches or in
combination for embedded optimization ap-
proaches. Yet efforts to invent new designs face
enormous combinatorial difficulties arising from
the large number of processing tasks that may be
required and technologies available to perform
them. Recognizing that it is generally not possible
to consider all possible equipment arrangements
simultaneously, a synthesis approach based on a
decomposition was proposed whereby any design
problem for which a solution is not known is
broken into smaller subproblems. Such a de-
Continued on page 96.


Mna design


An Integrated Teaching Approach

Vanderbilt University
Nashville, Tennessee 37235

T HE EFFECTIVE TEACHING OF design to chemical
engineering seniors often presents a challenge
for many chemical engineering faculties. Philo-
sophically, the senior design project should inte-
grate flow sheet generation, chemical engineering
sciences, transport and reaction equipment design
skills, and engineering economics. The problem
should be structured to emphasize group problem
solving, project organization and planning, and
effective communications. Further, the problem
should be an "as realistic as possible" industrial
design problem requiring synthesis of a final
process. One way to create an appropriate en-
vironment is by combining the design and labora-
tory courses, thereby relating the design project
to real facts, real chemicals, real alternatives, and
involving industrial practitioners.
For the past several years we have conducted
our senior plant design course as a combined one
semester five hour design/laboratory course.
Typical design projects have included:

polymerization of DMT to polyester [1]
manufacture of acetic anhydride from acetic acid
design of a yeast plant
design of a coking wastewater treatment plant

These projects have been carried out with the
aid of engineers in industry.
In the early stages of the design course/
laboratory development, real problems were en-
countered with obtaining the necessary commit-
ment from industry. The proprietary information
hurdles were almost insurmountable. Many
projects involved hazardous 'chemicals and
hazardous processes. They also quickly involved

*A preliminary version of this paper was presented at the
73rd Annual AIChE meeting, November, 1980.

John A. Roth is a Professor of Chemical and Environmental Engi-
neering at Vanderbilt University. He received his education at the
University of Louisville (B.ChE., 1956; M.Ch.E., 1957; Ph.D., 1961). He
has had experience in industry and government, both full-time and
as a consultant in the environmental area. His research interests are
in Chemical reactor kinetics and physical chemical waste water
treatment. (L)
Kenneth A. Debelak is an Assistant Professor of Chemical Engi-
neering at Vanderbilt University. He received his B.S. (1969) in
Chemical Engineering from the University of Dayton and M.S. (1973)
and Ph.D. (1977) from the University of Kentucky. He has had in-
dustrial experience with General Motors and governmental experience
with the Environmental Protection Agency. His research interests are
in coal kinetics, process control and synthesis, and physical-chemical
processes for industrial waste water treatment. (R)

considerable expense both for chemicals and ex-
perimental equipment. We have found that the
use of an energy or environmental design problem
provides a solution to many of these difficulties.

during the first week of class. They are as-
signed to individually synthesize a process to
achieve the design objective. The class as a group
then decides the process alternatives it wishes
to consider in detail. Interaction and limited di-
rection from the course instructors, other faculty,

'O Copyright ChE Division, ASEE, 1982


and industrial consultants are essential through-
out this process.
The class is organized as a company. A Project
Manager and Research Director are nominated by
the class and selected by the course instructors.
The class is then divided into research/design
groups with one member assigned as group leader.
These groups develop design methods and cost
curves; they also design, execute and conduct ex-
periments to supplement the data and informa-
tion in the literature. Since some students take
only the three hour design course (non-chemical
engineering graduate students and selected senior
project students), they participate only in de-
velopment of the design methods and cost in-
formation but not the laboratory.
A portion of the classroom time is devoted to
a series of formal lectures on

cost estimation and engineering economics
related topics (i.e., corrosion, materials selection,
patents, the technical library, statistics)
supporting technical topics pertaining to the major
design problem.
Additional homework assignments are made,
primarily using Peters and Timmerhaus [2] as a
text. One individual short design problem is as-
signed. Problems are also assigned covering cost
estimations, economics, and statistics. One day
a week is reserved for discussion of the major
project and dissimination of data. Outside speak-
ers are used extensively for the appropriate areas
of expertise. The Project Manager, Research Di-
rector, and group leaders meet weekly with the
course instructors to report on the week's activi-
ties and to plan for the next week. A major design
report is required from each design group (3-4
persons). Weekly laboratory reports and a final
report are required of the laboratory groups.


T HE SELECTION OF AN appropriate problem is
most crucial. It must be challenging but must
also have good chances for successful completion.
It must be common enough so that there is litera-
ture available and also require experimental work
to supplement existing data. To actively involve
local industries it must not include proprietary in-
formation. In the past this has been a serious
problem in enlisting industrial support. Two
types of problems seem to overcome many of
these difficulties-those concerned with energy
or environment. These are topical problems and

The choice of an environmental process
design requires that the student become familiar
with a new area, one which has its own terminology
and language but which is based on
chemical engineering principles.

have been enthusiastically supported by both in-
dustry and student involvement.
For the past two years we have focused on the
design of wastewater treatment facilities for coal
conversion processes. Industrial assistance from
the Chemical Technology Division of Oak Ridge
National Laboratory (ORNL) and Alabama By-
Products Corporation (ABC) has been obtained.
The group at ORNL has been responsible for the
design of wastewater treatment facilities for coal
gasification processes, while Alabama By-Products
Company is a coking operation which has had to
meet increasingly stringent discharge require-
ments from the EPA.
The choice of an environmental process design
requires that the student become familiar with a
new area, one which has its own terminology and
language but which is based on chemical engineer-
ing principles. We feel that this is a common ex-
perience for graduating chemical engineers since
they will be required to learn the specific
terminology as they take their places in the specific


T HE FOLLOWING IS THE statement of the design
of a coking wastewater treatment plant:

TO: ChE Project Staff
FROM: Design/Laboratory Instructors
Our firm has been retained as consulting engineers to
design a wastewater treatment facility for the Tennessee
Coking Company, Cumberland Plant. The attachment
shows the coking process and characterization of the raw
waste load (influent). You are asked to complete a detailed
estimate design for the wastewater treatment facility.
Data not available in the literature or uncertain will be
determined in the laboratory. Effluent limitations are to
be based on Tennessee standards, assuming that this plant
will be constructed on the Cumberland River.
The final formal report is due This
report will include cost estimates as well as process design.
Attach as an appendix your calculations and the source


of your data. You will be supplied the basis for cost

The chemical engineering faculty were available
as consultants. The industrial participants were
also available to answer questions. The State En-
vironmental Agency was contacted for technical
information on waste treatment processes and for
discharge requirements the plant would have to
meet. Class lectures were presented as requested
by the students on the following: biological waste
treatment, liquid extraction, sludge dewatering
and disposal, carbon adsorption, equilization,
clarifier design. A number of textbooks, EPA
documents and manuals, and training manuals
were made available from the library or from
individual faculty.
Biological waste treatment emerged as the
cornerstone of the treatment scheme. One group
was assigned to investigate this process. For am-
monia removal, air stripping and steam stripping
appeared as viable processes. Groups were
assigned to each of these processes. The remain-
ing laboratory groups were organized for this
project as shown in Fig. 1. Each group was re-


FIGURE 1. Student Organization.

sponsible for establishing the appropriate an-
alytical tests. Members from other groups were
instructed on the tests by the group that es-
tablished the procedure.
Before entering the laboratory each group
submitted a pre-experimental plan outlining their
experiments and analytical tests. Also included
in the plan was a time-table for completion of the
work. These plans were reviewed by the Research
Director, Project Director, and instructors. The
final laboratory report presented design data and
methods and a solved example of a design using
design data and equations. A recommendation as
to the feasibility of each process was made. Each
group also presented an oral report to the entire
class. The final reports along with any other in-
formation were made available to all class


A SAMPLE OF PROCESS wastewater and sludge
was obtained from ABC. Typical laboratory
data needs for this project included:
rate constants for decomposition of phenol by acti-
vated sludge
sludge settling data for clarifier design
neutralization curves
chemical oxidation rates for phenol
carbon adsorption data.
Bench scale studies were conducted on the acti-
vated sludge process in a completely mixed stirred
tank reactor (CSTR). Two activated sludge re-
actors were operated. After flow rates were es-
tablished for the influent, the process was allowed
to reach steady state. Samples were collected over
a period of several hours and stored at 0C until
the analyses were made. Kinetics for the removal
of Chemical Oxygen Demand (COD), phenol, and
cyanide were determined by varying the retention
time for each run. Once the reaction rate constants
were determined for each pollutant, the retention
time for the limiting pollutant could be calculated
and the size of the aeration basin determined. The
group also determined the amount of excess sludge
produced, and made recommendations for oxygen
and nutrient requirements. An example of their
laboratory data is shown in Fig. 2. This group
established analytical procedures for COD, phenol
and total cyanides in accordance with Standard
Methods. Other groups were instructed in these
The process wastewater obtained from ABC
had already been treated to remove ammonia. It
was necessary to make a synthetic waste for
these experiments. The removal of ammonia by
air stripping consisted of the following steps: pH
adjustment, air stripping, ammonia adsorption of
off gas in dilute sulfuric acid. The last step was
then extended to recover the resulting ammonium
sulfate, a by-product used for fertilizer. The group
determined kinetic data to size the pH adjustment
vessel. Mass transfer coefficients were determined
for both stripping and absorbing the ammonia.
Not all of the lab groups were totally success-
ful in completion of their assigned tasks. There
are numerous reasons for the wide variation in
performance among the groups. Some groups had
a better understanding of their goals and ob-
jectives and better understanding of the processes
they were investigating. Some groups just had
people who were more motivated and enthusiastic
about the project and worked more diligently to


achieve their goals. Some of the experimental
problems were much more difficult, either because
of inherent complexity, lack of equipment, or time
constraints when supplies were not readily avail-

T HE END PRODUCT OF THE design course is the
final report. We have required both individual
final reports and 8-4 person group reports. This
report is a formal report using the format as
recommended by Peters and Timmerhaus [2]. To
maintain some measure of uniformity, common
economic parameters such as amortization rates,
percentage of administrative costs, and other
pertinent fixed-capital investment and production
costs are specified. Common equipment/process
cost curves are used. These are developed by the
appropriate laboratory/design groups. There are
nevertheless a wide number of options open to
the design groups. We have also found that
presentation of results by the groups to a joint
faculty-industrial audience enhanced interest.
For grading, we heavily weighed peer evalua-
tions for the group laboratory reports and the
group design reports.

F OR THE PAST EIGHT YEARS we have taught our
senior design course as an integrated design
laboratory course. The students have responded
very well to this approach. They perform their
work in a mature and professional manner and
can see the results of their efforts. This experience
forces the students to organize their efforts; make
engineering judgments along the way; obtain,
find and evaluate data; and organize an experi-
mental program. We feel we have been successful
in achieving an atmosphere similar to an industrial
situation in which the interaction of a number of
groups is necessary for the completion of a
We do not wish to gloss over the pitfalls of this
approach, for there are several. Advanced
planning is necessary. The faculty must develop
some feel for the problem since there are an ex-
tensive number of student questions and problems
which arise. It is important that equipment,
reagents, and supplies be on hand before the be-
ginning of the semester. During the first several
weeks there is considerable chaos. Students are
often unsure of what is expected of them. It is






glop. o kO.ed HR'


100 200
FIGURE 2. Phenol Blodegradability.
imperative that a timetable be established to keep
the project moving forward. We found that a good
approach to this problem was to encourage the
Project Manager, Research Director, and group
leaders to apply group pressure to their peers.
Most students accept the challenge and are re-
sponsible. It is occasionally possible for a student
to slide by with minimum effort. To counteract
this, the peer evaluations are quite effective. For
the most part, we find the students evaluate each
other honestly.
This approach to design in the chemical engi-
neering curriculum has been quite effective. We
believe it comes close to providing a realistic ex-
perience for the student and a professional
challenge for the student and the faculty. We
heartily endorse this approach for a more effective
design experience. 3
We would like to thank Jerry Klein, Oak
Ridge National Laboratory; John Koon, AWARE,
Inc.; and Moyer Edwards and John Shores, Ala-
bama By-Products Corporation, for their time and
1. Overholser, K. A., C. C. Woltz, and T. M. Godbold,
Chemical Engineering Education, 9:16 (1975).
2. Peters, M. S., and K. D. Timmerhaus, Plant Design
and Economics for Chemical Engineers, McGraw-Hill,
Second Edition, New York (1908).

SPatRN 1982





with the assistance of C. K. WANG
University of Delaware
Newark, DE 19711

L cators is the infusion of notions of compe-
tition, uncertainty, and risk-taking into the de-
terministic methods traditionally used in the in-
struction of engineering students. An educational
tool was described in a previous article [1] which
utilized computer simulation techniques to provide
an evaluation of a process design in a more real-
istic fashion than the usual procedure of
"grading" a design report. In this article, we de-
scribe an advanced simulation which illustrates
how basic engineering skills can be combined with
economic considerations and risk-taking judg-
ments in order to effectively design and operate a
chemical processing unit in a competitive environ-
For this advanced simulation, data are pro-
vided to establish a design concept, capital cost,
and operating cost. The completion of the design
requires the specification of the production ca-
pacity (or "size") of the plant. This crucial de-
cision is left to the participants and is based on
their judgments concerning the uncertainty of
market projections as well as their appraisal of
the behavior of the competition. This decision is
further confounded by the the availability of other
options which compete for the limited funds avail-
able ($9 million). Research and development pro-
grams are proposed which may have potentially
high returns. However, the commitment of funds
to these high risk projects reduces the funds avail-
able to build the processing plant and leaves open
the possibility of not being competitive in an es-
tablished market place. Information concerning
the estimated risks involved in the research and

The completion of the design
requires the specification of the production
capacity (or "size") of the plant. This
crucial decision is left to
the participants ..

development programs is provided so that rational
qualitative judgments can be made concerning the
disposition of limited funds. These elements of
risk associated with the pay-off from the research
and development projects along with possible
fluctuations in the market and the uncertainty as-
sociated with the actions of competitors are in-
tended to illustrate the difficult conditions under
which many important engineering decisions are
This simulation has been used as one of five
projects in Technical Project Management, a
course of the Chemical Engineering Department
of the University of Delaware. This course is open
to advanced seniors, graduate students, and
practicing engineers. Their enthusiastic response
to the simulation project has provided the motiva-
tion for this article.

The participants are grouped into four com-
peting teams with 4 to 5 members per team. Each
team is provided with technical, marketing, and
economic information in the form of memoranda
in order to familiarize the participants with this
important form of communication. The central
theme of the project is the design and operation
of a plant for the manufacture of a primary
product "D". A by-product "C" is produced that
can be a potential source of profit. Each group is
allotted an initial sum of $9 million to finance the
The first simple task involves the determina-
Stion of the rate constant from concentration-time


SCopyright ChE Division, ASEE, 1982

data for the catalytic decomposition of raw ma-
terial "A" into the product "D" and by-product
"C" via the reaction:

A -> C + D

A simple process with an isothermal con-
tinuous flow stirred tank reactor and distillation
column can be used to manufacture "D" and "C".
The process flow diagram is shown in Fig. 1 and
described elsewhere [2]. Relationships that can be
used to estimate the capital cost, operating cost,
and materials cost are given in the memoranda.
The information contained in these various
memoranda can be easily formulated into an
economic model for the process unit cost as a
function of plant capacity by procedures de-
scribed in engineering textbooks [3].
Estimates for the total demand and associ-
ated selling price for the primary product "D"
are provided for a ten year period with the
caution that these estimates may be in error by
approximately 15 %. The price-quantity data given
are thus in the form of a standard demand curve
used by the microeconomist. The marketing data
for "D" provide a basis for judgments concern-
ing the specification of plant capacity (reactor

T. W. Fraser Russell is the Allan
P. Colburn Professor of Chemical
Engineering and the Director of
the Institute of Energy Conversion,
a laboratory devoted to photo-
voltaic cell research at the Uni-
versity of Delaware. He received
his MSc and MSc degrees in chemi-
cal engineering from the University
of Alberta and his PhD in chemi-
cal engineering from the University
of Delaware. His industrial ex-
perience includes work as a design
engineer with Union Carbide Can-
ada where he designed the first
multipurpose continuous chemical -
processing plant built in Canada, and as a research engineer for the
Research Council of Alberta, where he did some initial development
work on the Athabasca tar sands. (L)
Dennis F. Brestovansky is an Instructor of Chemical Engineering and
Research Associate in the Institute of Energy Conversion's Photovoltaic
Unit Operations Laboratory at the University of Delaware. He received
his BS and MS degrees from the University of Delaware. Mr. Bresto-
vansky's industrial experience includes employment as a research and
development engineer with the Buckeye Cellulose Corporation in
Memphis, Tenn. At the Institute, he is involved with research into
the development and design of a continuous process for the manu-
facture of low cost, thin-film solar cells. (C)
R. L. McCullough, Professor of Chemical Engineering and As-






FIGURE 1. Process flow diagram

size, distillation column capacity). This crucial de-
cision requires judgments concerning (i) the
market share that can be obtained in relationship
to the possible strategies and behavior of com-
peting teams, and (ii) the uncertainty of the
market projections.
The processing unit size and capital cost de-
cision is made more difficult by the availability of
options, all of which require funds which could
be used for plant construction. As one option, data
concerning improvements in process efficiency and
the related cost are given which can be used as a
basis for a "cost-benefit" analysis. These data are
constructed to illustrate the option for an optimum
investment in process related research. In a second
option, a research and development program is
proposed which could lead to the potential com-

sociate Director of the Center for Composite Materials at the Uni-
versity of Delaware, received his undergraduate chemistry training
at Baylor University, and was awarded a Ph.D. in Chemistry by the
University of New Mexico, while employed by the Los Alamos
Scientific Laboratories. Prior to his appointment at the University of
Delaware in 1971, he was associated with the Boeing Company as a
Senior Research Scientist and Acting Director of Materials Science
Laboratory; and the Chemstrand Research Center, Monsanto Company,
as Group Leader and Acting Manager of the Fiber Science Section.
Dr. McCullough has published technical papers and organized symposia
in the areas of polymer structure, molecular mechanics, structure-
property relations, and composite materials and is the author of several
books. (R)


Upon specification of the production capacity (I.e, the anticipated share of the markets),
some simple calculus can be used to obtain an optimal conversion of reactant, The distillation
column is constrained by a maximum design flow rate slnce a flow in excess of the design
limit will result in entrainment and loss of proper product specifications,

mercialization of by-product "C" and attendant
new sources of income. This R&D program is a
"high-risk" venture with no guarantee of success
and requires an initial commitment of $2 million
for pilot plant capital and then additional support
each year for a research staff The allocation of
the support funds is described in terms of
"normal", "crash", and "minimal" programs.
"Case studies" are given to illustrate the proba-
bilities of success under these various levels of
funding. For example, under the "crash" program
it is projected that a 50% chance exists for
marketing "C" in Year 2 and there is a 95%
chance that "C" can be sold in Year 3. For a
"normal" schedule, there is a 50% chance that
"C" can be sold in Year 4 and a 95% chance that
"C" will be on the market by Year 6. Other case
studies are given to illustrate the effects of erratic
funding patterns. Again, demand and price pro-
jections for "C" are given with the caution that
these projections could be in error by approxi-
mately 20%. The diversion of funds to these
alternate projects reduces the capital available to
build the plant and therefore reduces the capacity
for producing product "D",
The development of a strategy is further
complicated by a provision which allows for an ex-
pansion of production facilities in the fifth year
of the simulation. In order to accomplish this ex-
pansion, the team must have generated sufaient
funds from the sale of "D" (and possibly "C") in
the prior years to cover the capital eost of the ex-
pansion. A team with a large R&D effort (and a
small capacity plant) could accumulate sufmMcent
funds from the early sale of "C" for a large scale
expansion in Year 5. Ths could allow them to
dominate the future market fr or both "C" and
"I'", Alternately, a team which foregoes research
could build a prodtn facility capable of
generating saflent funds, solely from the sale
of "D", to nane an expansion that could
maintain their domination of the "U" market.
Upon hpedfiionti of the preduetion Capa ty
(i.e. the anticipated share of the markets), some
simple calculas can be used to obtain sa optimal
conversion of reseatant The distillation column is
costrained by a maximum design fow rate state

a flow in excess of the design limit will result in
entrainment and loss of proper product specifiea-
tions. The teams may elect to "oversize" the distil.
nation column (relative to the size dictated by the
optimum conversion) to gain flexibility in d@-
veloping a production strategy,
Each team submits proprietary reports for the
initial design specifications, operating conditions,
and research strategies (Fig, 2). The data for all
teams are submitted to the simulator, Internal
checks are conducted to identify trivial errors
such as those resulting from unit conversions, etc,
If design errors have been made, the value leading
to a lower production rate i selected Errors in
operating conditions are adjusted to meet the de
sign constraints, such as the flow limit on the
distillation column. If the capital or research
funds are allotted in excess of the current eash
balance, the funds will be reduced to give a current
cash balance of zero, Funds for plant construction
are given priority over reea re funds,
In each period a random number generator is
used to establish deviations away from the pro-
jected demand. The sampling procedures are de-

iW]tfi fsg it

_A__ffl_ flPa_
&W&'' .9 A- .tti09 S6 APUe f ffA
*99091% 09? 090S 09 9c^_^ ^ 0
(asf~ls 'ifi s'999 iit099 $if .0 -------

099i5 fM> f #afis IP ^^ .^___ _____
Iflf9M 2. byspft eSmfren hUbMaf roPa


signed so that random fluctuations of 15% (or
less) and 20% (or less) may occur from the
yearly market projections for "D" and "C" re-
spectively. The selling price is established from
the actual values of the market sizes through the
"demand-curve" relationships. These fluctuations
can have a significant effect on the teams' cash
flows. The assignment of market share is biased
in favor of the teams with large capacity and large
inventories. The specific relationship used to es-
tablish this bias is not made available to the par-
Progress toward the R&D breakthrough is de-
termined by a Monte Carlo procedure which
operates by sampling at random, but with the
frequency of a "research probability function".
The "spread" of the research probability function
is established by the funding levels and patterns
as reflected by a "Case Studies Summary". Al-
though each team can influence the "odds" for a
breakthrough by modifying research strategies,
the breakthrough time remains an unpredictable
(stochastic) event within the limits described in
the "Case Studies Summary".
The inventory of the primary product, "D",
is controlled by the simulator. Any unsold "D"
will be stored automatically and placed on the
market (along with past inventories and current
production) in the following year. No provisions
are available for dumping "'" so that large in-
ventories and high inventory cost can accumulate.
Deposition of the secondary product, "C", is
controlled by the team, and "C" can be dumped
or retained in inventory. The product "C" can
only be sold from inventory. If a new product
research "breakthrough" occurs and no "C" is in
inventory, no revenue will be generated from the
sale of "C". The amount of "C" to be placed on
the market in a given year must be specified by
the team.
At the end of each period of operation per-
formance summaries are returned to each group.
An income statement is provided which gives in-
formation on revenue from sales of "D" and "C",
cost of material sold, gross profit on sales, operat-
ing expenses (including plant depreciation ,
general expenses (including inventory charges
and R&D expenditures), interest from spent
funds, and net profit. A balance sheet lists the
company's current assets, plant and equipment,
and liabilities and equity In a marketng report
the market fractions of "D" and "C" captured,
the product selling prices, product costs, and the

amount of product produced in inventory and
sold are reported. Operating conditions for the
year are listed in a plant operations report and
an R&D status report is given. This includes a
projection of the year in which the sale of "C"
will begin (at the 95% confidence level) if a pro-
gram has been initiated, and the plant is operating
efficiently. These reports are treated as proprietary
for each company. In addition to supplying
essential information, these reports provide ex-
perience in analyzing accounting statements.
Production facilities may be expanded at the
begiinning of Year 5 if sufficient funds have been
generated from the sale of the primary product
"D" and possibly the sale of "C" resulting from a
research breakthrough. Design specifications for

Progress toward the R&D breakhrugh
is determined by a Moane Carto procedure which
operates by sampling at random, but with
the frequency of a "research
probability function."

any expansion (consistent with the available
funds) are submitted for the beginning of Year 5.
Increased production will begin in Year 5 so that
increased sales can contribute to the revenues re-
ported for the fifth year. The simulation is re-
peated through Year 10. The performance of each
group is evaluated on the basis of the return
generated from the initial 9 million investment.

The simulation has been performed at various
stages of evolution six times in the course CHE
691, Technical Project Management, offered in
the Department of Chemical Engineering at the
University of Delaware. The current form has
been used for three cases of CHE 691 and has
been tested in three additional exercises with
volunteer groups. Student time to complete the
game varies from 25 to 35 hours ad some 3 to 5
hours are required for class discussion and
presentations. The results summarized in Tables
1 through 3 and Fig. 3 are typical eults from a
simulation carried out in the fall of 1978.
The plant capacities specified by each team
before and after expansim are summarized in
Table 1. In this particular simulation two teas
(II and IV) elected to pursue the R&D programs
for the commem ization of the by-product "C"
while teams I and H built large production faci-

SPRIG 1982

1 2 3 4 5 6 7 8 8 10
FIGURE 3. Trends in cumulative profit over the ten-
year game.

ties in an effort to capture the established market
for the primary product "D". The initial pro-
duction capacities of the teams conducting re-
search were about 40% smaller than the teams
that declined the research program.
Team I elected to build a large capacity distil-
lation column in order to insure flexibility of pro-

duction rates. Due to the high cost of the column
relative to the reactor, they were forced to con-
struct a smaller reactor to stay under the $9
million limit. Team II also elected to construct an
oversized distillation column. Team III constructed
a more modest column which could handle a flow
20% in excess of the optimum flow rate. Team
IV tightly designed the distillation column to
match the optimum flow rate from the reactor.
The teams that elected not to support the R&D
program designed their processing units to pro-
vide about 40% of the projected demand for
product "D" in Year 3. Teams III and IV built
the maximum capacity facilities possible with the
funds remaining after the commitment to the
R&D program.
The sensitivity of the outcome shown in Fig.
3 to competitive interactions is illustrated by the
results summarized in Tables 1 through 3. Teams
I and II built large production facilities by fore-
going research. As a result, they captured a large
percentage of the market for product "D" and
acquired relatively high cumulative profits during
the first three years while Teams III and IV
carried out product "C" research. When Teams III
and IV made product "C" research breakthroughs
in Year 3, their profits began to swing upwards
from the increased revenues. Team IV went on
to accumulate the most profit by a great margin

Summary of Facilities


Reactor Size, in m3 45.1 50.0 50.3 50.5
Reactor Cost, in $ x 106 1.74 1.91 1.92 1.93
Distillation Column Design Limit, in Kg-"D"/yr x 105 6.7 5.6 4.2 3.8
Distillation Column Cost, in $ x 106 6.79 5.89 4.94 4.64
Initial Capital Investment (year 1), in $ x 106 8.53 7.80 6.86 6.57
R&D Investment (year 1), in $ x 106 0.00 0.00 2.00 2.00
Maximum Capacity, in Kg-"D"/yr x 105 6.7 5.6 4.2 3.8
Reactor Size, in ma 45.1 67.5 30.2 54.6
Reactor Cost, in $ x 106 1.74 2.50 1.22 2.07
Distillation Column Design Limit, in Kg-"D"/yr x 106 6.7 6.6 2.2 6.8
Distillation Column Cost, in $ x 106 6.79 6.34 3.50 6.64
Total Capital Investment in Expansion (year 5), in 8.53 8.84 4.72 8.71
$ x 106
Overall Capital Investment in Equipment, in $ x 106 17.06 16.64 11.58 15.28
R&D Investment (year 1), in $ x 106 0.00 0.00 2.00 2.00
Total Maximum Capacity, in Kg-"D"/yr x 105 13.4 12.2 6.4 10.6


Distribution of Market Share (%)

"D" "C"
Period Team Team

1 31 30 19 20 -
2 35 30 14 21 -
3 37 30 12 21 15 85
4 35 29 16 20 53 47
5 34 30 16 20 45 55
6 28 31 17 24 37 63
7 29 30 15 26 37 63
8 31 29 15 25 37 63
9 32 28 15 25 37 63
10 32 28 15 25 37 63

by the end of the simulation. This one-sided out-
come was due in part to mistakes in judgment
made by the other teams.
Team III did not initiate a "crash" research
program until the second year; nonetheless, they
managed to achieve a "breakthrough" in the
same period (Year 3) as Team IV (Table 2). Un-
fortunately, Team III did not anticipate such an
early breakthrough and had not built up sufficient
inventories of "C" by storing product made in
prior periods. Consequently, Team IV captured
85% of the "C" market in Year 3 simply because
they had the product available for sale. Team III
responded by increasing production to generate
more "C" (Table 3). At this point both Team III
and IV began to view product "C" as their primary
product with "D" relegated to the role of a by-
product. Table 2 shows the results of this strategy.
Team III's fear of high inventory cost is re-
flected in their production strategies for the
primary product "D". The mandatory accumula-
tion of inventories of "D", that resulted from the
small market share (19%), contributed to a deficit
of $1.6 million for Team III in period 1. In order
to avoid additional inventory costs, they
drastically reduced production in Year 2 (see
Table 3). This caused a further decline in the
market share for "D" that led to a cumulative
deficit of $1.9 million. As shown in Table 3, Team
III increased production slightly in Year 3. Shored
up by the breakthrough into the "C" market,
Team III became more aggressive in Year 4. Un-
fortunately for Team III, the past production
strategies left them with inadequate funds for a
significant expansion in Year 5 so that Team IV
gained and maintained control of the "C" market

for the remainder of the simulation.
Team I elected to allocate all the available
funds to the construction of facilities for the pro-
duction of the primary product "D". The decision
to build an oversized distillation column to pro-
vide flexibility in operation introduced a signifi-
cant penalty by reducing the funds available for
the construction of the reactor. As a consequence,
Team I had the largest capability for producing
"D" but was penalized by having the smallest re-
actor (see Table 1). The small reactor led Team
I to operate at a conversion of 71%-well below
the optimum conversion of 85%. Team II also
emphasized flexibility in operating conditions by
building an oversized column which they con-
sistently operated at an off-optimum conversion
of 78%. Neither Team I nor II used the flexibility
in production rates that had been so dearly pur-
chased. Both teams tended to operate at the non-
optimal maximum production rate throughout
the simulation. This operating strategy led to a
reduced return on the capital investments. In
order to meet demands, all teams ran their facili-
ties at the maximum production rate in the latter
periods of the simulation.
Team IV operated close to the optimum con-
version level consistently and stored a large

Summary of Operating Strategies

Production of "D"
Period Conversion (%) (kg/yr x 105)

1 75 78 86 85 5.8 5.5 3.5 3.8
2 71 93 6.7 1.8 "
3 71 91 6.7 2.3 "
4 71 84 6.7 4.2 "
5 71 84 6.7 4.2 "
6 80 84 4.7 4.2 "
7 71 84 6.7 4.2 "
8 71 84 6.7 4.2 "
9 71 84 6.7 4.2 "
10 71 78 84 85 6.7 5.5 4.2 3.8

amount of by-product "C" in inventory in anticipa-
tion of a research breakthrough. This strategy
gave them the highest cumulative profit at the
end of the simulation. However, had Teams I and
II operated their facilities closer to optimal and
had Teams III and IV been in closer competition
Continued on page 87.


S classroom



General Electric Company
Schenectady, NY 12301
Purdue University
West Lafayette, IN 47907

in fluid mechanics or transport phenomena in-
vest special care to establish the tensorial in-
tegrity of the stress system in a fluid, quite often
the status of a vector is nonchalantly offered to
the mass and energy fluxes. The usual scenario
begins with the instructor identifying a point in
the fluid continuum, isolating a direction which
orients an infinitesimal area dA normal to it and
asking for the traction vector representing the
net surface force per unit area of dA. The in-
finite multiplicity of directions at a point leads
to the customary despair about its implications
to the characterization of surface forces until
their redemption through the gift of a second
order tensor by the joint effects of the momentum
principle and the continuum postulates.
Indeed the foregoing exercise is a healthy one,
for the simplicity of description of fluid stress by
a second order tensor is not one to be taken for
granted. The motivation for this article is the
extension of the same considerations to the mass
and energy fluxes in a fluid: their claims to being
vectors is a matter to be established by argument.
More generally, the primitive instruments of

The infinite multiplicity of directions
at a point leads to the customary despair about
its implications on the characterization of surface
forces until their redemption through the gift
of a second order tensor by the joint effects
of the momentum principle and
the continuum postulates.

Copyright ChE Division, ASEE, 1982

transport phenomena are orientation-dependent
variables, which are an extraordinary en-
cumbrance to mathematical treatment. The
central issue here is the replacement of such
primitive variables by substitutes that have no
orientation-dependence for the price of a unit
increase in tensorial order. This is a considerable
bargain, in the realization of which the conserva-
tion laws are intimately involved. Of course, it is
not that the dependence of the said primitive
variables on orientation has been entirely elimi-
nated but rather that it has been reduced to one
of homogeneous linearity. The entire procedure
as applied to the traction vector has been called
Cauchy's fundamental theorem.t
If it is admitted that the velocity v is a vector,
then it would instantly follow that the mass flux
pv is a vector and no further 'proof' would be
called for. On the other hand, an independent
definition of point velocity in a deforming
continuum does not build on a limiting procedure
of a collapsing mass, a requirement basic to proper
definitions of "point" quantities. The alternative is
then to seek a suitable definition of something
related to velocity (such as mass flux) from which
the velocity may be obtained as a derived quantity.
The choice falls somewhat naturally on mass flux
which, as mass flow rate per unit area, could
depend on the orientation of the area. This com-
plex situation is alleviated by the fact that the
mass flux turns out to be a vector. Similar con-
siderations also hold for the energy flux. Before
demonstrating these results for mass and energy
fluxes, it is instructive to review the derivation of
Cauchy's fundamental theorem for stress, which
is related to the modeling of mechanical inter-
actions within a material continuum.

tSee for example, "The Elements of Continuum Me-
chanics," by C. Truesdell, Springer Verlag, New York, Inc.,


Vijay Kumar Stokes is a mechanical engineer. He received his
BScEngg. (HONS) Mech. degree (1961) from Banaras Hindu University
and his MSE (1962) and PhD (1963) degrees from Princeton University.
He taught at the Indian Institute of Technology, Kanpur from 1964
until 1970. During 1970-71 he was a Visiting Unidel Associate Pro-
fessor in the Department of Chemical Engineering at the University
of Delaware. After working for eighteen months at Foster-Miller
Associates he returned to India to teach at I. 1. T. Kanpur (1973-78)
where he was the Head of the Mechanical Engineering Department
(1974-77) and the Convener of the Interdisciplinary Program in Nu-
clear Engineering and Technology (1977-78). He has been with
General Electric Corporate Research and Development since 1978. He
is interested in continuum mechanics and applied mathematics, and
his research includes heat and momentum transfer between rotating
disks, the effects of couple stresses in solids and fluids, wave propa-
gation in springs, plastic wave propagation, and the design and
development of oil-water separators. He holds several U. S. patents.
Doraiswami Ramkrishna received his B(Chem)Eng. degree (1960)
from the Bombay University Department of Chemical Technology and
his Ph.D. (1965) from the University of Minnesota. After teaching for
two years at Minnesota, he returned to India in 1967 and taught at
the Indian Institute of Technology, Kanpur until 1974. He was a.
Visiting Professor at Wisconsin (1974-75) and at Minnesota (1975-76)
and is now Professor of Chemical Engineering at Purdue University.
Hb is a consultant to General Mills, Inc., Minneapolis. His research
includes dispersed phase systems, stochastic modeling and applications,
bioengineering and problems of general applied math interests.
The authors recall the many pleasant years of their interaction
during their early academic careers at the I. 1. T., Kanpur, where the
academic environment was extraordinary. The present article grew
out of an evening's discussion many years ago. (R)


At a point x = [x,] in a continuous body con-
sider a small area dA whose orientation is deter-
mined by its unit normal n = [ni]. Then, the me-
chanical action of one part of the body on its
neighborhood across dA can be represented by a
force density t such that t dA gives the force
exerted on the area dA. Because of the nature of
force t is a vector, called the traction vector, which
has the dimensions of stress-force per unit area.
Besides being a function of the position x and

time t, the traction vector t = t(x,t;n) is also a
function of the orientation n of the area dA. As
a result, the traction vector is not suitable for
purposes of analysis, even though it is the primi-
tive variable of interest. Cauchy's fundamental
theorem guarantees the existence of a second
order tensor T(x,t) = [Tij], called the stress
tensor, such that

t(x,t;n) = n.T (x,t)
or (1)
ti(x,t;n) = njTji (x,t) summation over j

The proof of this result is in two steps. First
we show that t (x,t;-n) = -t (x,t;n). This follows
from an application of Euler's first law of motion
(the rate of change in linear momentum of a body
is equal to the total force applied to the body) to
a thin cylindrical disk of thickness dl and cross-
sectional area dA, as shown in Figure 1, which
results in the equation of motion

p(dldA)a = dA[t(x,t;n) + t(x,t;-n)]
+ dlJt(x,t;s)ds + p(dldA)f(x,t)

where a is the acceleration of the material disk,
p is the density of the material, the line integral
along the contour c gives the contribution to the
force due to the stress vector t (x,t;s) acting on the



t (x,t;-n)

X -

FIGURE 1. Traction vectors acting on a thin disk.


cylindrical surface of the disk, and the last term
is the contribution due to the body force. Taking
the limit dl ->0 in Eq. (1) and dividing the result-
ing equation by dA yields

t (x,t;-n) = t (x,t;n) (2)

Next we consider the motion of the small
material tetrahedron, PABC, at the point x, as
shown in Fig. 2, whose edges are along the
orthogonal unit vectors e1, e3 and es, and where
face ABC has an area dA and normal sn = [ni].
The areas of the faces with normals -er are then
given by dAr = nrdA, i = 1,2,3. Let the distance
of P from the face ABC be ds. Then, an applica-
tion of Euler's first law of motion to the material
tetrahedron results in the equation

(1 dsdA)a = dA[t(x,t;n) + nrt(x,t;-er)]
3 r=1

+ p( --dsdA)f (x,t)

Division of Eq. (3) by dA, followed by taking the
limit as ds -> 0, and the use of Eq. (1) then gives

t (x,t;n) = X nt (x,t;er). (4)

This equation shows that if the traction vector t
is known at a point on three mutually orthogonal

The same procedure could be
used for mass and energy fluxes,
both of which arise from orientation
dependent scalars.

planes (here e,, e, and es), then it can be deter-
mined on any plane n at that point. We can reduce
this result further: Let the components of
t (x,t;er) along the axes el, e2 and e, be given by
t(x,t;er) = Trs es.
Then Eq. (4) becomes
t(x,t;n) = S S n,T,,es
r=l s=
3 3
= n Y Tr, r e
r=l s=l
t (x,t;n) = n T (x,t), (5)



FIGURE 2. Traction vectors acting on a infinitesimal

where T is the dyadic

T = : Trs e es.
r=1 s=1

If Cartesian tensors are used, with summation im-
plied over a repeated index, then this equation
can be written alternatively as

ti(x,t;n) = n, Tri(x,t)

Let us summarize what we have so far. We
started with a primitive variable of interest-the
traction vector which is orientation dependent.
By using a physical principle, here Euler's first
law of motion, we were able to. establish the
existence of a second order tensor-the stress
dyadic T (x,t), which is a point function and which
factors out the dependence of t on n.
The same procedure could be used for mass
and energy fluxes, both of which arise from
orientation dependent scalars. Application of
physical laws-here conservation of mass and the
first law of thermodynamics, respectively, would
result in suitable point functions that are vectors
which factor out the dependence of the primitive
scalars on the orientation n. In each case the de-
sired result would be obtained because, in the
equation which results as a consequence of the
application of the physical principle, the volume
dependent terms vanish faster than the surface
dependent terms in the limit when one of the
dimensions of the body approaches zero-just as


in the derivation of Eqs. (2) and (5). Instead of
going through this exercise individually for each
case, it is useful to formalize this property in
the form of a theorem that is discussed in the
following section. Since the result applies to
tensors of all orders, it is more descriptive to use
Cartesian tensors rather than Gibbsian dyadics
and polyadics. However, each result will also be
presented in its Gibbsian polyadic form.

Consider a cartesian tensior field of order p
(the case of p = 0 represents a scalar and that
for p = 1 represents a vector), b = [bii2... ],
whose components at time t depend not only on
the spatial coordinates x = [xi] of the point, say
P at which they are considered but also on a
specified orientation n = [ni] at P. The foregoing
dependence on spatial coordinates and orientation
is presumed to be continuous and we write

biLi...ip = b112... i (x,t;n)
b = b(x,t;n)

Before we state the theorem of interest, some
preparations are essential. At any point P in the
continuum we identify a volume V bounded by an
area A enclosing P. The surface integral

A biti2... X(x,t;n)dA
A b b(x,t;n)dA

where n is everywhere normal to area A and point-
ing outwards, is defined for every volume enclosing
P and in fact, through the mean value theorem,
equals the value of the integrand somewhere on
the area A. It is now possible to take the limit of
this surface mean by contracting the volume to
zero around P.
The theorem concerned states that there
exists a tensor of order p + 1, denoted by
B = [Bji, i,...i (x,t] such that

= njBjil2..."p(x,t), summation on j, (6)

which is b (x,t;n) = n. B (x,t) in Gibbsian notation,
if and only if

lim 1
VlimO a1ibiliZib(x ,t;n)dA = 0

lim 1
V_>A0bb(x,t;n)dA = 0

Note in particular that the new tensor B(x,t)
= [Bjii2...ip] is not a function of the vector n.
The proof of this theorem is straightforward.

Thus the conservation principles
are themselves responsible for "fluxes" being
vectors or tensors, a result of tremendous
significance in the investigation of
the mechanics of continue.

To prove the 'only if' part we assume that Eq. (6)
is true. Thus we establish Eq. (7) by
lim 1
V.0 A jbii2... (Xx,t;n) dA
lim 1 I
= V-oA jj [Bji,2]...]dV = 0
where we have used (i) the divergence theorem,
(ii) the summation convention on index j and (iii)
the assumption that the divergence (aBj.../3xj) is
continuous and bounded. For the converse, we
start from Eq. (7). By taking V to be a right
cylinder of infinitesimal cross-section dA with its
normal along n, as shown in Figure 1, and contract-
ing the cylinder along its axis to a section passing
through x, we conclude that
b... (x,t;-n) = -b... (x,t;n)
or (8)
b(x,t;-n) = -b(x,t;n)
where for simplicity the indices have been re-
placed by dots. Next, at any point x (position
vector referred to some origin) and orientation
n, we apply the result of Eq. (7) to a sequence of
tetrahedra such that each of them has a vertex
at x and the face opposite is no#ral to n, which
points out of the tetrahedron (see Figure 2). The
other faces are formed such that the edges passing
through x lie along three mutually perpendicular
lines chosen to be a set of cartesian coordinate
directions. The normals pointing out of these
faces are the unit vectors -e, -e and -e3 re-


spectively. The application of Eq. (7) to this
collapsing sequence together with Eq. (8) gives
b...(x,t;n) = nj b...(x,t;ej), summation over j,

b(x,t;n) = S nj b(x,t;ej)

=n. S ejb(x,t;ej)
Since b...(x,t;n) and nj are tensors of orders p and
1, respectively, it follows from the quotient law
that 2 ejb(x,t;ej) is a tensor of order (p+l).
Thus the theorem is proved with
B,...(x,t) = b...(x,t;ej)

B(x,t) = S e bi(x,t;ei)

At any point x in the deforming continuum,
given a direction n it is possible to define a mass
flow rate per unit area (a continuum postulate!)
across an area normal to n. Denoting this by
m(x,t;n), we obtain the mass flow rate through
dA by m(x,t;n)dA. Conservation of mass for a
volume V bounded by A gives
dt \\ pdV + (m(x,t;n)dA = 0 (10)
where p is the fluid density. If Eq. (10) is divided
by A and we let V -> 0, the continuum postulate
that ap/at be continuous on V and A leads to the

lim 1
Vlim0 1 m(x,t;n)dA = 0
which is precisely the condition given by Eq. (7)
so that the theorem just discussed leads to the
existence of a mass flux vector M(x,t) such that

m(x,t;n) = njMj(x,t)
m(x,t;n) = n. M(x,t)

Thus the mass flow rate through area dA = ndA
is given by dA M. The point velocity vector is
now readily defined by v = (1/p)M. In an n-com-
ponent mixture, the mass flux of the kth com-

ponent of mass fraction wk is defined M(k) =
Mok, from which its velocity is defined by v(k) =

Although the stress tensor has been covered
earlier, the repetition here will serve this tutorial
well by reinforcing the result as a fallout from
the generalized theorem above.
A momentum balance for the volume of fluid
V bounded by A must account for the traffic of
momentum across A because of mass flow, body
forces throughout V and surface forces along the
entire surface A. The traction (force per unit
area) on an area dA, located at x on A, with normal
n (directed out of V) is denoted by [ti (x,t;n]. For
the ith component one has

dt (pvidV + ( viMjdAj

= ppfidV + I t,(x,t;n)dA


where [fi] represent components of the body force.
Dividing Eq. (11) by A and letting V->0, the
volume integrals are readily seen to vanish. Since
we have shown that M is a vector, [Mjvj] trans-
forms as a vector for each fixed i and the theorem
is applicable to the second term on the left hand
side of Eq. (11). Thus we obtain
lim 1
Vm>O X ti(x,t;n)dA = 0

so that the theorem is again applicable, and yields
second order tensor [Tj i (x,t)] such that
ti(x,t;n) = nTjii(x,t)
t (x,t;n) = n T (x,t)

The surface force on an area d)A ndA has for
its ith component dAjTji.

By energy flux here we mean that which occurs
by molecular conduction.t Of course energy is
transported by fluid motion but since this is associ-
ated with the mass flux vector the resulting energy
flux is already known to be a vector. Thus our con-

tRadiative transport through weakly absorbing media
would require a slightly more elaborate treatment.


cern here belongs to the energy transported by
conduction per unit area across an area located
at x and normal to n, denoted by q(x,t;n). The
first law of thermodynamics applied to volume V
bounded by A, gives
d 1 A 1
dt I [ + -2-v] dV + h [h + vl]MdAj
= ( dAjT,,v, + (I / pfividV- q(x,t;n)dA
The interpretations of the various terms in Eq.
(12) are available in any standard textbook. In
view of the pattern already set before we are able
to conclude that
lim 1
V-_0 A- q(x,t;n)dA 0
so that there exists an energy flux vector
[Qj (x,t)] such that
q(x,t;n) = njQpi(x,t)
q(x,t;n) = n. Q(x,t)
Thus the conservation principles are them-
selves responsible for "fluxes" being vectors or
tensors, a result of tremendous significance in the
investigation of the mechanics of continue. It is
a fact that deserves mention in courses on fluid
mechanics and transport phenomena.

Personal discussions have borne out that the
issues raised in this paper are a routine matter
to many. The authors would like to specially
acknowledge Professors L. E. Scriven at the Uni-
versity of Minnesota and Stephen Whitaker at
the University of California, Davis. In particular,
Professor Whitaker was kind enough to provide
evidence that the contents of this paper are not
common knowledge and represent useful informa-
tion. O

Continued from page 81.
for product "C", Team II could have gained the
largest profits.
Upon the completion of 10 operating periods,
a year by year summary of the performance of all
teams is made available to the participants. Each

team is asked to prepare a report to analyze their
performance and identify the important decisions
and actions that led to their relative position in
the competition. Although the simple "return on
investment" is suggested as a possible economic
evaluator of performance, the teams are at liberty
to select alternate evaluators; e.g., "internal rate
of return" (the interest rate that makes the
cumulative discounted cash flow equal to zero) or
"borrowing power". As would be expected, those
groups that make significant profits in the early
periods of the simulation tend to base comparisons
on the "internal rate of return". The use of these
alternate measures of performance can cause re-
versals in the relative positions established by
comparisons based on "net profit" or the simple
"return on investment" criteria.

Student response to this project has been en-
thusiastic. The immediate consequence of their
decisions provides a sense of realism for the inter-
play between technical, marketing, and economic
factors. The computer simulator, the package of
memoranda, and instruction manual are available
at a total cost of $275.00 from Engineering Edu-
cational Materials, 805 Baylor Drive, Newark,
Delaware 19711. O

We would like to thank all those who have
helped us in the development of this educational
tool. Mr. R. N. Pratt, Engineering Computation
Specialist in the College of Engineering at the
University of Delaware, has been most helpful
in setting up and modifying the computer pro-
gram for the simulation. His efforts in running
the simulation are also greatly appreciated.
Student feedback from experiences in playing the
game was essential to proper game development
and their input is gratefully acknowledged.

1. T. W. F. Russell and D. S. Frankel, Chemical Engi-
neering Education, "Teaching the Basic Elements of
Process Design with a Business Game," Winter, 18-
23 (1978).
2. T. W. F. Russell and M. M. Denn, Introduction to
Chemical Engineering Analysis, John Wiley & Sons,
Inc., New York (1972).
3. J. Wei, T. W. F. Russell, and M. W. Swartzlander,
The Structure of the Chemical Process Industries,
McGraw-Hill Book Co., New York (1979).


I classroom




University of Maryland
College Park, MD 20742

M Y INTERESTS IN ENERGY conservation began
with a one credit special topics course which
involved studying our local ice rink to suggest
energy conservation measures [8]. The course was
very successful and student enthusiasm was high.
The motivating factors were that we were work-
ing on a real system, we had to define the
problem, and that energy conservation was
topical. After the course the Department of
Energy awarded a contract to develop the class'
results into a technology transfer manual which
was recently published [1].
The Department of Energy's technology trans-
fer series covers a wide range of topics such as
energy conservation in distillation, evaporation,
and the use of computers for energy conserva-
tion. Staged operations in general and distillation
in particular are prime areas for energy con-
servation. It has been reported [9] that 3 % of the

Thomas McAvoy received his B.S. in Chemical Engineering from
Brooklyn Polytechnic Institute in 1961. His M.A. (1963) and Ph.D.
(1964) degrees in chemical engineering were obtained from Prince-
ton. He has taught for 17 years and recently joined the University
of Maryland. He has carried out research in process dynamics control
and more recently in applied pharmaco-kinetics. He has authored over
40 technical papers and one audio short course.



Course Outline-Staged Operations
Vapor Liquid Equilibrium Review
Enthalpy Concentration Diagrams
One Stage Flash Calculations
Ponchon Savarit Method
McCabe-Thiele Method
Design Considerations-Flooding, Weeping, Tray
Shortcut Techniques
Multicomponent Column Calculations
Energy Conservation

energy used in the United States goes into distilla-
tion. When I taught our senior course on staged
operations I decided to integrate energy conserva-
tion into the course. About 25% to 30% of the
course involved energy conservation. Again
student interest and enthusiasm were very high.
There are two other benefits to including the
energy conservation material. First, evaluating
energy conservation measures forces one to focus
on economics. Thus, the staged operations course
reinforced the students' experience in their pro-
cess design and evaluation courses. Secondly,
energy conservation generally involves good,
sound engineering. Invariably it only takes an
undergraduate training to appreciate the con-
servation techniques. What students learn is that
energy conservation measures already exist and
the question of their use depends primarily on
The purpose of this paper is to discuss how
energy conservation can be integrated into a
course on staged operations. Table 1 gives an out-
line of the material covered in the course.
Topics I to V were taught in the usual way with-
out consideration of energy conservation. Some
energy conservation material was introduced in
covering topics VI and VII. For the somewhat
dry subject matter such as flooding, tray

Copyright ChE Division, ASEE, 1982


There are two other benefits to including the energy conservation material. First,
evaluating energy conservation measures force one to focus on economics. Thus, the staged operations
course reinforced the students' experience in their process design and evaluation courses.
Secondly, energy conservation generally involves good sound engineering.

efficiencies, etc., the energy conservation im-
plications of these topics helped greatly to moti-
vate students. During the three week period when
the students worked on a multicomponent com-
puter project, lectures on energy conservation
were given.

Essentially all of the energy conservation
material that is required can be gotten from the
literature. The two primary sources that were
used are the Department of Energy (DOE)
Technology Transfer Manual [10] and Shin-
skey's text on distillation control [11]. The DOE
manual lists over 50 journal references. The Mix
[9] article cited earlier is also useful and it pre-
sents a number of rules of thumb.
In the DOE manual and Mix article energy in-
formation is given on the 29 most important
towers in the petroleum industry as well as for
the 131 key towers in the chemical industry. This
information includes the energy required/lb
product, the number of trays, reboiler tempera-
ture and condenser temperature. The DOE
manual divides energy conservation measures up
into three broad categories: as shown in Table II.

Energy Conservation Measures In Distillation
1. Lower reflux operation
2. Lower pressure operation
3. Changing feed plate location
4. Proper maintenance
5. Reducing heat exchanger fouling
(up to $50,000)
1. Insulation particularly for valves and flanges
2. Waste heat recovery
3. Retraying for higher efficiency/lower pressure drop
(over $50,000)
1. Advanced instrumentation and control
2. Heat pumping
3. Intermediate condensers and reboilers
4. Two stage condensation
5. Multiple tower operation

At this point it is useful to discuss three typical
examples that were used in the course. Other

examples are given in the DOE manual and home-
work problems can be developed from them.

Example 1: Lower pressure operation.
Fig. 1, taken from Shinskey's text [11], shows
a typical operating window for a butane splitter
with the various column constraints labeled. It is
necessary for a column to operate within these
constraints, or window as they are sometimes
called. The dashed curves, labeled contours of
constant separation, are for a fixed feed (F) and
a fixed product split (x,, x,). It is possible to use
Fig. 1 to show that minimum energy consumption
corresponds to minimum pressure operation. Since
coolant temperature sets column pressure, when
favorable cooling conditions arise one should
take advantage of them and minimize column pres-
sure. This minimum pressure can be determined
from the intersection of the condenser constraint
and the appropriate contour of constant separa-
The calculation of the constraints and the
contours of constant separation is straightforward
and an excellent homework assignment. For

20 40 60 80
Column pressure, psio

100 120

FIGURE 1. Typical operating windows for a butane
splitter, illustrating that production may be maximized
at the condenser constraint. (From "Distillation Control,"
by F. G. Shinskey, 1977. Used with permission of
McGraw-Hill Book Company.)


example, the flooding curve is a simple rearrange-
ment of the standard flooding correlation given by
Van Winkle [14]

0.2 ___ L l pv
V= A v).vVpLPV) f wV_ )
Similarly the weeping constraint can be calcu-
lated from standard correlations. The reboiler
and condenser constraints are heat transfer con-
straints. For the condenser assume a constant
coolant temperature, To, constant heat transfer
coefficient, U, and fixed area, A. Changes in
column pressure will change the temperature of
the overhead product. For nearly pure products
this temperature pressure relationship is given
by the Clausius Clapeyron equation

P = Poexp T ) (2)

As P increases the temperature at the top of the
tower increases and the amount of heat that can

As can be seen, lower pressure operation
gives substantial energy savings. Such savings can
occur at night, during the winter and during a
rain storm for air cooled condensers.

be transferred increases. If this heat is assumed
to be only latent heat then V and P can be related
along the condenser constraint as

V= UA 1 To
X 1 R n P
To AH Po In mean
The reboiler constraint can be similarly calcu-
lated. The steam condensate capacity constraint
and the vessel pressure constraint are self ex-
The contours of constant separation can be
calculated using any of several shortcut tech-
niques. The equation of Jafarey et al. [5] will be
x l-xw\
In xD Xw I

N 1 -a (4)
1n 1 1/2
(1+ RXF)

For a fixed column separating a fixed feed into
fixed products, Eq. (4) shows that the relative

volatility, a, and the reflux ratio, R, are related

a Constant (5)
1 + RxF

The reflux ratio and the vapor boilup are related

V = (R + 1) F x (6)

Assuming that the reboiler and condenser heat
loads are equal and that only latent heat effects
are important, the energy to run the tower is
given as

Q = XV

As tower pressure is lowered due to the avail-
ability of lower cooling conditions, several effects
occur. Tray efficiency decreases and a and X in-
crease. Table III shows the results of using Eqs.
(5) (7) on Shinskey's butane splitter. The
constant in Eq. (5) was calculated from Shin-
skey's values for a 120F condensate tempera-
ture. The effects of tray efficiency are not included
in Table III. As can be seen, lower pressure ope-
ration gives substantial energy savings. Such
savings can occur at night, during the winter and
during a rain storm for air cooled condensers. All
of these cases result in a lower To in Eq. (3) and
the condenser constraint is shifted to the left.
One of the interesting aspects of the calcu-
lations given in Table III is that relatively small
increases in a are translated into relatively large
decreases in energy consumption. Kister and Doig
[6] [7] have published papers on the effect of pres-
sure on column performance and these papers are
useful supplements to Shinskey's text.

Energy Required to Separate Butanes
at Constant Feed Conditions
(x, = 0.5; 97-3 Split)



S(BTU/b) V/F
a (BTU/lb) V/F


120 95.2 1.32 129 5.00 645 76
100 71.9 1.35 135 4.09 552 61
80 53.1 1.38 140 3.48 487 45
60 38.1 1.41 145 3.03 439 28


Another important aspect of Fig. 1 is that it
can be used to discuss the concept of maximizing
throughput. For the butane splitter maximum
throughput occurs at the intersection of the con-
denser and flooding constraints. For other systems
maximum throughput can occur at the intersection
of the reboiler and flooding constraints [11], [6],
[7]. The slope of the flooding curve relative to the
slope of the contours of constant separation deter-
mines the maximum throughput point. In going
through either the minimum pressure analysis or
maximum throughput analysis students see the
importance of such mundane considerations as

Example 2: Retraying for higher efficiency.
In the DOE manual retrofitting old towers
with more efficient trays is discussed. The example
chosen for illustration is a naphtha debutanizer
shown in Fig. 2. In calculating energy conserva-
tion results, the Eduljee [2] fit of Gilliland's [4]
correlation is used

N Nmi 0.75 1 (R-Rmin)o j (8)
N+ 1 R + 1]

The Underwood equation [12] [13] can be used to
calculate Rm and Nm can be calculated from
Fenske's equation [3]. By increasing tray
efficiency, the number of theoretical plates N is
increased and the reflux ratio R can be decreased.
To estimate energy savings assume that AN
is the increase in N and that this produces a AL
reduction in reflux. From Eq. (8) and the column
material balance AL can be calculated. The result-
ing energy savings are approximately

AQ = AL X (9)
The results presented in the DOE manual are
given in Table IV. The cost of new trays was
estimated to be $11,300 and the installation

Retrofitting Trays for Naphtha Debutanizer

Case 1 Case 2 Case 3

Increase Theoretical Trays by 2% 5% 10%
R (Gilliland) 2.12 2.04 1.94
Energy Savings, (106 BTU/hr) 0.12 0.32 0.59
Annual Savings @ $2.50/106 BTU $2376 $6369 $11,682
Payout time (yrs) for $16,950 7.1 2.7 1.5
Payout time (yrs) for $22,600 9.5 3.5 1.9

FIGURE 2. Naphtha Debutanizer
Manual (1980)].



[Taken from DOE

charges were estimated to be between 50% and
100% of tray costs.
The DOE manual estimates that a 3% in-
crease in N is necessary to achieve a maximum
reasonable payout time of 5 years. However, this
estimate is based upon relatively cheap energy
($2.50/106 BTU) and therefore it should drop
with increasing energy costs.

Example 3: Heat pumped towers.
In the DOE manual a propane/propylene
splitter was chosen to illustrate heat pump eco-
nomics. This system involves reboiler flashing,
shown in Fig. 3. With reboiler flashing the bottoms
product is used as the working fluid in the re-
frigeration loop. The economics for heat pumping
the propane/propylene splitter are given in Table
V. As can be seen, substantial energy and dollar
savings can be realized by going to a heat pump
design in this case.
The DOE manual gives calculation procedures
for determining energy and dollar savings for all
the examples discussed. Using these procedures
the results in Table V can be calculated. In ad-
dition the manual gives detailed guidelines for
application of each of the energy, conservation
measures. The most important guideline for heat
pumping is that the AT between the condenser and
reboiler should be less than 650F. An interesting


homework problem can be given to illustrate this
guideline. Suppose that two engineers decide to
spend the same amount of money to reboil a tower.
One buys low pressure steam at $2.50/10 BTU
and uses this steam directly in the reboiler. The
other buys high pressure steam at $3.50/106 BTU,
and uses this steam to run a compressor in a heat
pump arrangement. If the condenser temperature
is 1200F at what AT between the reboiler and con-
denser will the heat pumped tower begin to be
competitive with an ordinary tower? Assume
reasonable efficiencies and neglect the added
equipment costs for the heat pumped system. For
80% efficient compressor and a 33% efficiency for
generating work from steam the answer can be
calculated from the Carnot cycle equation as

AT = ($2.5/$3.5) (0.33) (0.8) (580R) (10)
= 109F

The difference between 109F and the DOE
recommendation of 65F results from the added
capital and maintenance costs for the heat pumped

The three examples which have been presented
are typical of those which can be incorporated
into a standard staged operations course. More
than enough material is available from literature
sources to easily cover the subject of energy con-
servation. Because of the topical nature of the
material student interest and motivation are high.
Energy conservation involves good, sound engi-
neering and it forces students to consider the
economic aspects of their designs. Lastly, students

Economics for the Application of a Heat Pump
to a Propane/Propylene Splitter*

Con- Reboiler Reboiler
ventional flashing flashing
Operating and tower as a on a new
economic factors design retrofit column

Column pressure, psia 275 275 275
Overhead temperature, OF 115 115 115
Bottoms temperature, F 135 135 135
Steam consumption,
106 BTU/hr 101t 12+ 12+
Cooling water,
106 BTU/hr 101 6 6
Utility costs, 106 $/yr 3.06 1.10 1.10
Increased maintenance,
106 $/yr 0.28 0.28
Savings, 106 $/yr 1.68 1.68
Additional capital cost,
106 $/yr 3.44 2.04
Simple payout, yrs 2.1 1.2
After-tax ROI, % 24.4 41.2

*All costs are in 1978 dollars.
tLow-pressure steam (valued at $2.50/103 lb).
+High-pressure steam (valued at $3.50/103 lb).

begin to learn about column operation. For all
these reasons it is felt that energy conservation
is a useful addition to any course, but in particular
to one on staged operations. E


f =
H =
L =
N =

Heat capacity
Feed flow
Reflux flow
number of trays

Greek Letters
a = Relative volatility
A = Difference
p = Density
C = Coolant
D = Distillate
F = Feed
L = Liquid

P = Pressure
Q = Heat duty
R = Reflux ratio
T = Temperature
U = Heat transfer
V = Vapor flow
x = Mole fraction

X = Latent heat of
= Surface tension


= Minimum
= Base case
= Vapor
= Bottoms


FIGURE 3. Reboiler Flashing [Taken from DOE Manual

1. Dietrich, B., and T. J. McAvoy, "Energy Conserva-
tion in Ice Skating Rinks," published by Department


of Energy, Oak Ridge, Tenn., 1980.
2. Eduljee, H. E., "Equations Replace Gilliland Plot,"
Hydrocarbon Processing, 120-122 (Sept. 1975).
3. Fenske, M. R., "Fractionation of Straight-Run
Pennsylvania Gasoline," Ind. Eng. Chem., 24, 482-485
4. Gilliland, E. R., "Multicomponent Rectification.
Estimation of the Number of Theoretical Plates as a
Function of the Reflux Ratio," Ind. Eng. Chem., 32,
1220-1223 (1940).
5. Jafarey, A., J. M. Douglas, and T. J. McAvoy, "Short-
Cut Techniques for Distillation Column Design and
Control. I. Column Design," I & EC Process Design
& Develop., 18, 197-202 (1979).
6. Kister, H. Z. and I. D. Doig, "Distillation Pressure
Ups Thruput," Hydrocarbon Processing, 54, 132-136
(July 1977).
7. Kister, H. Z. and I. D. Doig, "Computational Analysis
of the Effect of Pressure on Distillation Column Feed
Capacity," Trans. I. Chem. E., 57, 43-48 (1979).
8. McAvoy, T. J., "Energy Conservation in Ice Rinks. A
Unique Open-Ended Practice Course," AIChE Sym-
posium Series 183, Vol. 75, 81-86 (1979).
9. Mix, T. J., J. S. Dweck, M. Weinberg, and R. C.
Armstrong, "Energy Conservation in Distillation,"
CEP, 74, 49-55 (April 1978).
10. Radian Corporation, "Energy Conservation in Distilla-
tion," published by Dept. of Energy, Oak Ridge, Tenn.,
11. Shinskey, F. G., Distillation Control. Chapter 1,
McGraw-Hill, New York, 1977.
12. Underwood, A. J. V., "Fractional Distillation of
Multicomponent Mixtures," Chem. Eng. Prog., 44,
603-614 (1948).
13. Underwood, A. J. V., "Fractional Distillation of
Multicomponent Mixtures," 45, 609-618 (1949).
14. Van Winkel, M., Distillation, 525-526, McGraw-Hill,
New York, 1967.

book reviews

By G. S. Laddha and T. E. Degaleesan
McGraw-Hill, 1978, 485 pages
Reviewed by N. L. Ricker
University of Washington
From the title, one might expect this book to
be confined to the study of theoretical and experi-
mental developments in the field of transport phe-
nomena. While this is the authors' main emphasis,
they also give an overview of other important
facets of the practice of liquid extraction. The
general orientation is very similar to the well
known book by Treybal (1963) *, and it seems ap-
propriate to use Treybal's work as a frame of
reference for this review.
*Liquid Extraction, by R. E. Treybal, McGraw-Hill, 1963.

ChE Division ASEE
Annual Meeting 0 June 1983 0 Rochester NY
Papers are invited on any aspect of ChE educational innova-
tion but particularly relating to mass transfer
Abstract deadline Aug. 1, 1982
Send to: E. Dendy Sloan, Chem & Pet. Ref. Engg. Dept.,
Colorado School of Mines, Golden, CO 80401

Laddha and Degaleesan begin with a brief dis-
cussion of common industrial applications of liquid
extraction, followed by chapters devoted to the
fundamentals: phase-equilibrium thermodynamics,
theories of diffusion and interphase mass trans-
port, and calculational methods for stagewise ex-
traction and countercurrent differential extraction.
The material follows a logical sequence, and
practicing engineers will be comfortable with the
format, which emphasizes conventional graphical
methods, overall NTU's and HTU's, etc. Treybal,
however, covers much of the same material in
more depth. Also, Laddha and Degaleesan fail to
cite the more recent theories for the prediction
of liquid-liquid equilibria, and they do not discuss
the use of modern calculational methods for ex-
tractor design and simulation. There is an in-
correct statement, repeated in several places, that
the distribution coefficient is given by slope of the
distribution curve, which is not true in general.
The next two chapters deal with the behavior
of single drops and multiple interacting drops dis-
persed in a continuous phase, with an emphasis
on the fluid dynamics and mass transfer character-
istics of such systems. There is also a qualitative
discussion of the important Marangoni effects. The
material is presented in a unified form, whereas in
Treybal it is much more scattered.
The next major section of the book begins with
a description of the different types of extraction
devices used in practice and gives a brief summary
of the factors that might influence the selection of
a device for a specific application. Following this,
six common types of contractors: spray, packed,
perforated-plate, rotary-agitated, and pulse-
agitated columns, and mixer-setter extractors, are
treated in individual chapters. Each chapter
contains performance correlations and design cri-
teria that can be used for the given contactor.
These chapters comprise about one half of the
book, and are perhaps its best feature.
The final chapter reviews the special problems
that arise when extraction is accompanied by
Continued on page 96.


I class and home problems

The object of this column is to enhance our readers' collection of interesting and novel problems in
Chemical Engineering. Problems of the type that can be used to motivate the student by presenting a
particular principle in class or in a new light or that can be assigned as a novel home problem are re-
quested as well as those that are more traditional in nature that elucidate difficult concepts. Please sub-
mit them to Professor H. Scot Fogler, ChE Department, University of Michigan, Ann Arbor, MI 48109.


University of Maryland
College Park, MD 20742

In newer ice rinks a direct expansion re-
frigeration system is used to keep the ice frozen.
A schematic of such a system is shown in Fig. 1.
In effect the ice and the pipe under it serve
as the evaporator in the refrigeration loop.
Several different mechanisms contribute to
the heat, Q, which has to be removed from the ice.
First, the air in the building tends to heat up
the ice and second, water vapor in the air tends
to condense on the ice surface. Suppose that you
are a rink owner and a salesman shows up claim-
ing that a third mechanism is important as well.
He claims that radiational heat transfer from
the ceiling to the ice accounts for a significant
fraction of Q. To solve the problem he will sell
you a non-radiating ceiling which has an
aluminized face. The non-radiating ceiling will
be hung from the present ceiling if you decide to
buy it. The installed cost of the ceiling is $30,000.
Using the data below estimate if the investment
is worthwhile. Give any assumption that you make


w ... VALVE




in calculating your answer.

Condenser temperature
Evaporator temperature
Ice temperature at surface
Ceiling temperature
Ceiling height
Ice length
Ice width

55,000 kwh/month
40 ft.
200 ft.
85 ft.

The ceiling area can be taken equal to the ice
area. Use $ .05/kwh as a cost for electricity. The
rink operates for 8 months during the year.

Assume that the ceiling and ice are black
bodies and that the side walls are adiabatic. Eq.
(14.4-15) from Bird, Stewart, and Lightfoot [1]
gives the radiant heat transfer from the ceiling
to the ice as

QR = A F (Tc'- Ti4) (1)

a = 0.1712 x 10-8 BTU/hr ft20R"
A = 85 ft x 200 ft = 17000 ft2
Te = 525R
Ti = 4870R
Since the length to width ratio of the ceiling is
approximately 2.0, a value of F can be estimated
from curve 7 given in Figure 14.4-4 of Bird,
Stewart and Lightfoot. The estimated F is

F = 0.75
The hourly rate of heat transfer can be calculated


C Copyright ChE Division, ASEE, 1982


Each year CEE publishes a special Fall issue devoted to graduate education. This issue consists of
articles on graduate courses and/or graduate research, written by professors, and of announcements
placed by ChE departments describing their graduate programs. If you are interested in contributing to
the editorial content of this issue by submitting a paper on either a graduate course or on graduate re-
search, please send the Editor a letter describing the paper's content and possible date of submission.
Address correspondence to Ray Fahien, Editor, CEE, ChE Dept., U. of Fla., Gainesville, FL 32611.

from Eqn. (1) as
QR = (0.1712 x 10-s) (17000) (0.75)
(525 -4874) = 4.31 x 10 hrU

If it is assumed that the new ceiling completely
eliminates radiation to the ice then over an 8
month period the reduction in heat input to the
ice can be calculated as

QR = (4.31 x 105) (365 x ) (24)
= 2.53 x 109 BTU
Next the Carnot cycle equation can be used
for the refrigeration system to calculate the work
that is saved. This work is

W = QR Tond Tea, (2)
Tevap 71
7) = comp. efficiency (assume 0.80)
Teond = cond. temp. (5500R)
Tevap = evap. temp. (476R)
Substituting for QR, Teond, and Tevap, gives the work
savings for the 8 month operating period as

W = (2.53 x BTU) 550-476)

(2.931 x 10-' kwh
0.8) BTU
= 1.43 x 105 kw hr
The dollar savings can be calculated as

$ Savings = 1.43 x 105 kw hr x $ .05 $7150
kw hr
The estimated payout time for the non-radiating
ceiling is 4.2 years which is reasonable. Given the
fact that energy costs are bound to rise, the ceil-
ing is probably a good investment. O

1. "Transport Phenomena," R. B. Bird, W. E. Stewart,
and E. N. Lightfoot, John Wiley and Sons, New York,

I Ibook reviews

By Dean S. Shupe
Marcel Dekker, Inc., 1980. 186 pages
Reviewed by William G. Sullivan
University of Tennessee
This short book provides a concise treatment
of many principles of engineering economic evalu-
ation. It is full of good example problems to il-
lustrate these principles, and it can be read and
easily understood by the practicing engineer in a
few hours.
Several topics are covered that one might not
expect to encounter in a book of this length. For
example, several examples deal with inflation and
there are numerous solved problems related to
solar energy applications. The subject of debt
versus equity financing and how to handle it is
also included in a separate chapter. Furthermore,
several relatively advanced federal income tax
provisions are illustrated very clearly in another
The book is ideal for review of engineering
economy topics that appear on professional engi-
neering examinations. In addition, it could serve
well as a textbook for a shortcourse (1 or 2 days)
on this subject. Because of its brevity in con-
ceptual development regarding why various evalu-
ation methods work the way they do, the book is
not suitable as a text for ,a college-level course.
No homework problems are included at the end of
any of the eight chapters.
In Summary, What Every Engineer Should
Know About Economic Decision Analysis should
help engineers with no formal training in engi-
neering economics to more fully appreciate the
"how to," but not the "whys," of conducting
studies of engineering alternatives. The book
could make a valuable addition to the practition-
er's bookshelf. -


Continued from page 71.
composition of the total flowsheet generation
problem leads to the identification of many of the
synthesis subproblems mentioned before such as
heat integration, separations sequencing, etc. in
which much progress has been made. The de-
composition principle might also be applied re-
cursively to the solution of each of these sub-
problems. A special case arises if the decomposi-
tion is done in such a manner that at least one of
the resulting subproblems is recognized im-
mediately as solvable by some available technology.
If this process is repeatedly and successfully
applied until no unsolved subproblem remains,
the result is equivalent to the systematic specifica-
tion of a finite sequence of technologies which to-
gether form a feasible design for the original pro-
cessing problem.
The heat integration synthesis problem de-
scribed previously is amenable to such a de-
composition strategy. At each stage, a stream to
be heated and another to be cooled (either two
process streams or a process stream and a utility)
for which required tasks remain are selected and,
if thermodynamic or other constraints are met,
simultaneously solved with immediately recog-
nizable technology (an exchanger, furnace, etc.).
Should this heat exchange match fail to perform
the required task completely for a process stream
(reach the desired final temperature or phase
state), the remaining task, or 'residual', is simply
included among the other as yet unsolved process
heat transfer tasks for consideration in successive
stages. With appropriate available utilities, this
systematic generation scheme will produce a
feasible design in a finite number of steps. Differ-
ent designs result from alternative streams or
possible portions of streams selected for each
match and from how strictly the various task
specifications and other constraints are met. The
number of such designs for heat integration
among M streams is pn the order of ((M/2) !) 2.
Specification of separation sequences for
multicomponent mixtures is another synthesis
subproblem for which systematic generation ap-
proaches have been proposed. The problem arises
in reactor feed preparation, product purification,
by-product recovery, waste treatment and other
situations where the tasks of increasing concentra-
tion or component isolation are specified. In a
simplified form, a mixture is to be separated into
each of its components using a sequence of

technologies which sharply split a single multi-
component feed stream into two outlet streams,
one containing some subset of the feed com-
ponents and the other containing the remaining
components. Differences in various physical
properties of the components such as volatility or
solubility are exploited by various technologies to
effect the desired separations. If S different
separation technologies are available (such as
simple, azeotropic and extractive distillation, ex-
traction, fractional crystallization, etc.), the
number of different design sequences for the com-
plete separation of an N-component mixture is
on the order of S-1 (2N-2) !/ (N! (N-l) !). O

EDITOR'S NOTEs The concluding section of Dr.
Siirola's lecture will be published in the next issue
of CEE (Summer, 1982).

Continued from page 93.
chemical reaction. This subject is only given a
very brief treatment by Treybal. The theoretical
discussion in the present work draws heavily from
results obtained in studies of chemical absorption.
Although there are a few data for liquid-liquid
systems, there should be more. The presentation
generally falls short of the state-of-the-art in this
Topics covered by Treybal but not discussed in
the present work include calculational methods for
multicomponent extraction, methods for labora-
tory and small-scale extractions, economics of ex-
traction processes, and the competing factors in-
volved in the selection of a solvent. In general,
Treybal's book has the same theoretical basis, but
a more practical, process-oriented flavor than the
present work.
The subject material is probably too specialized
for the book to find much use in the undergraduate
chemical engineering curriculum. It seems better
suited as a reference for students and industrial
practitioners with a special interest in liquid ex-
traction. It would have been more valuable in this
regard had the authors included more references
to the recent literature. Only 2% are from the
period 1971-1978, with the most recent of these
being from 1974.
A final minor criticism: the printing quality in
the review copy was noticeably inferior to that
found in most technical books printed in the U.S.
The type was uneven and generally too light, and
pages were often slanted from the verticle. L



Departmental Sponsors: The following 146 departments contributed
to the support of CHEMICAL ENGINEERING EDUCATION in 1982 with bulk subscriptions.

University of Akron
University of Alabama
University of Alberta
Arizona State University
University of Arizona
University of Arkansas
Auburn University
Brigham Young University
University of British Columbia
Bucknell University
University of Calgary
California State Polytechnic
California Institute of Technology
University of California (Berkeley)
University of California (Davis)
University of California (Santa Barbara)
Carnegie-Mellon University
Case-Western Reserve University
University of Cincinnati
Clarkson College of Technology
Clemson University
Cleveland State University
University of Colorado
Colorado School of Mines
Columbia University
University of Connecticut
Cooper Union
Cornell University
University of Dayton
University of Delaware
U. of Detroit
Drexel University
Ecole Polytechnique (Canada)
University of Florida
Florida Institute of Technology
Georgia Technical Institute
University of Houston
Howard University
University of Idaho
University of Illinois (Urbana)
Illinois Institute of Technology
Institute of Gas Technology
Institute of Paper Chemistry
University of Iowa
John Hopkins University
Iowa State University
Kansas State University
University of Kentucky
Lafayette College

Lamar University
Lehigh University
Loughborough University
Louisiana State University
Louisiana Tech. University
University of Louisville
University of Maine
University of Maryland
University of Massachusetts
Massachusetts Institute of Technology
McMaster University
McNeese State University
University of Michigan
Michigan State University
Michigan Tech. University
University of Minnesota
University of Missouri (Columbia)
University of Missouri (Rolla)
Monash University
Montana State University
University of Nebraska
University of New Brunswick
New Jersey Inst. of Tech.
University of New Hampshire
New Mexico State University
University of New Mexico
City University of New York
Polytechnic Institute of New York
State University of N.Y. at Buffalo
North Carolina State University
University of North Dakota
Northeastern University
Northwestern University
University of Notre Dame
Nova Scotia Tech. College
Ohio State University
Ohio University
University of Oklahoma
Oklahoma State University
Oregon State University
University of Ottawa
University of Pennsylvania
Pennsylvania State University
University of Pittsburgh
Princeton University
University of Puerto Rico
Purdue University
University of Queensland

Queen's University
Rensselaer Polytechnic Institute
University of Rhode Island
Rice University
University of Rochester
Rose-Hulman Institute
Rutgers U.
University of South Carolina
University of Saskatchewan
South Dakota School of Mines
University of South Alabama
University of South Florida
University of Southern California
Stanford University
Stevens Institute of Technology
Syracuse University
Teeside Polytechnic Institute
Tennessee Technological University
University of Tennessee
Texas A&M University
Texas A&I University
University of Texas at Austin
Texas Technological University
University of Toledo
University of Toronto
Tri-State University
Tufts University
Tulane University
University of Tulsa
University of Utah
Vanderbilt University
Villanova University
Virginia Polytechnic Institute
University of Virginia
Washington State University
University of Washington
Washington University
University of Waterloo
Wayne State University
West Virginia Inst. Technology
West Virginia University
University of Western Ontario
Widener College
University of Windsor
University of Wisconsin (Madison)
Worcester Polytechnic Institute
University of Wyoming
Yale University
Youngstown State University

TO OUR READERS: If your department is not a contributor, please ask your department chairman to write CHEMI-
CAL ENGINEERING EDUCATION, c/o Chemical Engineering Department, University of Florida, Gainesville, Florida


With Applications to Phase Equilibria
Hendrick C. Van Ness and Michael M. Abbott,
both of Rensselaer Polytechnic Institute
1982, 482 pages
Solutions Manual
This unique text-the only textbook
devoted entirely to the subject-includes
numerous worked examples and problems for
solution. The book gives a systematic and
.comprehensive exposition of classical
thermodynamics of nonelectrolyte solutions,
including applications to the formulation and
solution of problems in fluid-phase equilibria. In
a thorough discussion of fluid properties and
their interrelations, the authors include state-of-
the-art methods for representing the properties
of pure fluids and mixtures.

Ronald F. Probstein, Massachusetts Institute of
Technology, and Water Purification Associates; and
R. Edwin Hicks, Water Purification Associates
1982, 496 pages
Solutions Manual
The first complete textbook covering a new
and important field, presented here as a
unified engineering subject. The text includes
relevant chemical and physical fundamentals
and develops the basic principles of each

Arthur P. Fraas
1982, 720 pages (tent.)
A comprehensive study of all major energy
systems, including-for each system-
historical background, underlying theory, and a
critical review of actual test experience that
defines .the practical limitations of the system.
Early chapters are devoted to the
characteristics of basic components and
problem areas common to most types of
energy systems. Includes summaries for all
chapters and for most sections, and over 300
figures and tables.

Ferdinand Rodriguez, Cornell University
1982, 576 pages
Solutions Manual
This up-to-date second edition continues
to teach students how to describe polymers in
quantitative terms and how to manipulate
quantitative data to predict and correlate the
behavior of real polymer systems. Adopting a
practical approach, this edition includes a new
example of drag reduction based on the most
current research, and an examination of torsion
pendulum that gives students a test method for
measuring mechanical properties at small
deformations. It treats chain transfer and
emulsion polymerization in more quantitative
terms, and it includes new sections on
feedstocks for monomer production and on
new polymer production methods such as gas
phase polymerization of ethylene and olefin
metathesis of norbornene.

Carroll O. Bennett, University of Connecticut; and
J.E. Myers, University of California, Santa Barbara
1982, 848 pages (tent.)
Solutions Manual
This text is designed for chemical
engineering students in the transport staged
operations sequence. Changes for the third
edition include: a general updating throughout;
50% SI units; chapter summaries; and
additional problems.

Nicholas Tsoulfanidis,
University of Missouri-Rolla
1982, 656 pages (tent.)
Covers all aspects of radiation
measurement, using worked examples and
problems to support concepts. The text
includes the most current radiation detectors
and techniques, teaching.students to select the
proper detector given the energy and type of
particle to be counted and the purpose of the

1221 Avenue of the Americas New York, N.Y. 10020

Full Text