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 )
serial   ( sobekcm )
periodical   ( marcgt )


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

che m ia engineering education

PROCTER & GAMBLE is looking for

in R&D/Product

This organization is responsible for the
creation and improvement of new and
existing products, together with developing
the associated technology advances and
solving technical problems.
While this organization encompasses the full
range of scientific and engineering
backgrounds, the primary need at the BS/MS
level is for Chemical Engineers and
MBAs with a chemical or engineering
undergraduate degree.
Your initial responsibilities in the organization
would be primarily technical, with varying
degrees of interactions with P&G's
Engineering, Manufacturing and Marketing
divisions. As you advance, your career could
evolve along technical and/or management
routes. This evolution will include progressive
assignments, exposure to other divisions, and
in many cases a transfer to another
R&D/Product Development division, or
where appropriate to an Engineering,
Manufacturing or Marketing division.
The R&D/Product Development organization
is headquartered in Cincinnati, consists of
over 20 divisions, focuses on U.S. consumer
and industrial products, conducts P&G's basic
research, and provides technical support for
our international operations and technical
centers. (This technical support includes
international travel by certain of our
U.S.-based division personnel.)

If you are Interested In this area, please send
a resume to:
The Procter & Gamble Company
R&D BS/MS Recruiting Coordination Office
Ivorydale Technical Center
Spring Grove and June Avenues
Cincinnati, Ohio 45217


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




14 ,4a-~rd ._ectte
A New Look at an Old Fossil: Kinetics of
Coal Processing, D. D. Perlmutter

2 Departments of Chemical Engineering
ChE at Notre Dame, Faculty
8 The Educator
Richard M. Felder, R. W. Rousseau

20 ChE Lecture
Infinite Possibilities for the Finite Element,
Bruce A. Finlayson

26 ChE Laboratory
A Simple Tubular Reactor Experiment,
Robert R. Hudgins, Bertrand Cayrol

30 ChE Curriculum
Impressions of Process Control Education
and Research in the U.S.,
Kurt V. Waller

12 Stirred Pots
Sodales Princetonienses, Rutherford Aris

36 ChE News
Chemical Engineering Symposium at
Carnegie-Mellon, Michael Locke

38 Problems for Teachers
Solution: Prarie Dog Appendix,
R. L. Kabel

40 Classroom
Teaching Market Analysis, J. T. Ryan,
Bret Haugrud

25 Letter to the Editor

37 Positions Available

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. O. 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 1981 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 ChE department is located in the new Fitzpatrick Hall of Engineering, completed in July 1979.

[ department


University of Notre Dame
Notre Dame, IN 46556

Dame began in 1909. The 1908-9 Bulletin
published in July 1909 noted that "chemical manu-
facture has developed so rapidly and grown so
exacting that there has arisen a demand for men
who not only can create and improve chemical
processes strictly so called, but who can deal with
the problems of construction and maintenance as
far as they are related to chemical industries. To
prepare young men for such work, the course in
Chemical Engineering has been designed." There
were no specific courses in chemical engineering
per se; of the total 167 semester credit hours in
Copyright ChE Division, ASEE, 1981

the curriculum there were various courses in
chemistry (58 cr hrs), mechanical engineering
(30 cr hrs), drawing (15 cr hrs) and shopwork
(14 cr hrs). By today's standards, the university
was small then, with all students numbering 995,
and only 42 faculty of which two were in chemis-
try. (Current total enrollment is 8500 and faculty
number about 700). The first undergraduate "de-
gree of chemical engineer" was awarded in June
1911 at the 67th annual commencement to
Guillermo Patterson, Jr. of New York City. (Inci-
dentally, ours is the only Chemical Engineering
Department in the world which can boast of an
alumnus named Jacques Rousseau. He did not
minor in philosophy.)
It was in the 1920-21 academic year that the
university was reorganized in a form that exists to
this date, with four undergraduate colleges in Arts


The early chemical engineers at Notre Dame were taught chemistry by the
brilliant chemist J. A. Nieuwiand, CSC-who with H. H. Carothers of du Pont is credited
with inventing the first synthetic rubber "neoprene" . -and by Knute Rockne, who was on the chemistry
faculty during 1916-22 before moving on fulltime to some other activity.

and Letters, Science, Engineering, and Business
Administration (then Commerce)-each with its
own Dean-and a separate Committee on Gradu-
ate Study, which preceded formal establishment of
the Graduate School in 1932. A department of
Law, now the Law School, was established in 1869.
The heads of various departments were also ap-
pointed in 1920-21, and the first Head of Chemical
Engineering (1920-41) was Henry B. Froning.
Froning was a distinguished chemist who joined
Notre Dame in 1920, and also served as Head of
Chemistry (1920-41) before becoming Dean of
Science (1940-43). The course "Principles of
Chemical Engineering" based on the pioneering
book of the same name by Walker, Lewis and
McAdams (1923) was added to the curriculum in
1921, replacing its predecessor course in Industrial
Chemistry. The early chemical engineers at Notre
Dame were taught chemistry by the brilliant
chemist Rev. Julius A. Nieuwland, CSC-who with
W. H. Carothers of du Pont is credited with in-
venting the first synthetic rubber "neoprene" in
the late 20's-and by Knute Rockne, who was on
the chemistry faculty during 1916-22 before
moving on fulltime to some other activity.
By 1920, there were only 11 graduates of the
department. The present undergraduate degree
"bachelor of science in chemical engineering" was
instituted in 1925. The number of graduates rose
steadily; 59 degrees in the 20's and 156 in the
30's. The headship of the department passed to
Ronald E. Rich in 1941 when Froning left to

Henry B. Froning Ronald E. Rich
1920-41 1941-59

devote fulltime as Dean of Science. Rich had
joined the faculty in 1933, and was a Notre Dame
alumnus (B.S. Ch.E. '28, M.S. Chem. '36); he
guided the department until 1959 when Julius T.
Banchero arrived from Michigan as Head. Other
Chemical Engineering faculty during this period
were P. J. Byrne, Jr. (1920-22), H. Wenzke (1922-
39), E. G. Mahin (1926-33), H. Hinton (1929-38),
A. Boyle (1929-39), G. Hennion (1933-40) and
E. J. Wilhelm (1938-71). The Chemical Engineer-
ing Department also moved to a then new building
in 1941.

P RIOR TO 1945, THERE WERE NO graduate courses
offered within the Chemical Engineering De-
partment. Several graduate courses including lec-
tures and a laboratory in Applied Electrochemis-
try, Advanced Thermodynamics, Advanced Unit
Operations lectures and laboratory, and Advanced
Plant and Equipment Design were initiated in
1945-46 to accommodate the returning WWII
veterans who were anxious to renew their engi-
neering background. Allen S. Smith (Ph.D. Michi-
gan, '40), who joined the Chemical Engineering
faculty in 1946 and remained here till his death
in 1967, was a major factor in this development.
He introduced additional graduate level courses in
1946 on analysis of Distillation and Extraction,
and Applied Chemical Kinetics; these along with a
Heat Transmission course by Rich and Corrosion
of Metals and Alloys offered by Wilhelm, plus

Julius T. Banchero Roger A. Schmitz
1959-79 1979-


those initiated the previous year provided sufficient
offerings to begin a formal master's program
with required thesis research. Other engineering
departments were also beginning their master's
programs at this time. The very first master of
science in engineering was awarded to a mechani-
cal engineer in June 1947; the first wave of ten
masters of science in Chemical Engineering gradu-
ated in August 1947.
Other than Rich, it was Wilhelm and Smith,
F. L. Benton (1940-56), M. T. Howerton (1949-
56), J. Treacy (1950-56), G. Parravano (1956-
59) and J. P. Kohn (1955-present) who taught
Chemical Engineering at Notre Dame during the
40's and 50's. A total of 212 bachelors degrees in
Chemical Engineering were awarded during the
40's, and 336 in the 50's. By 1959, 77 master's
degrees had also been awarded.
Due to Rich's failing health, J. T. Banchero
was acquired as Head in 1959. He came with
assurance from the administration that a doctoral
program in Chemical Engineering could be
started. By this time, G. S. John (1958-63) had
already joined the faculty, but the major addition
was that of E. W. Thiele-well known for his
pioneering contributions to two major areas of
Chemical Engineering science; the McCabe-
Thiele method for graphical design of distillation
columns for binary separations and the Thiele
modulus in the problem of simultaneous diffusion
and reaction in porous catalyst pellets. Thiele had
just retired after a long and productive research
career with American Oil Company in 1960, and
was persuaded to take on a second career at Notre
Dame. Thiele retired from active teaching in 1970
at age 75, and was awarded an honorary doctor-

Kinetics and reactor studies are conducted in the
Catalysis and Reaction Engineering laboratory.

ate in 1971 in recognition of his contributions to
Chemical Engineering and Notre Dame.
The next addition to the faculty was J. J. Car-
berry in 1961, a former Notre Dame undergradu-
ate, who had been with du Pont after receiving
his Ph.D. at Yale in 1957. Along with Banchero,
Kohn, Smith, Thiele and Wilhelm already present
then, his addition provided a major thrust towards
the growth of the graduate program. A total of
55 Ph.D. degrees in Chemical Engineering have
been awarded since the first one in June 1963 to
Joosup Shim who did his thesis research with Jim
During the sixties and much of the seventies,
undergraduate enrollments were about 30 in each
class, and fulltime graduate students in residence
averaged 22 per year. The graduate program
was considerably strengthened during this period.
Several other faculty members, A. H. P. Skelland
(1963-69), E. D. Crandall (1965-69), T. G. Smith
(1969-75), N. D. Sylvester (1969-73), F. H. Ver-
hoff (1969-74) and K. D. Luks (1967-79) played
a major role in this development. Various new
graduate courses were added to the program.
Three other present faculty members, W. Strieder
(1969), A. Varma (1975) and E. E. Wolf (1975)
also joined during this period.
The department moved to its present location
in the new Fitzpatrick Hall of Engineering in July
1979. Our classroom and laboratory facilities are
now among the best and most modern available

SITH THE RETIREMENT and elevation to emeri-
tus status of Julius Banchero in 1979, the de-
partment was very fortunate to attract Roger A.
Schmitz as Keating-Crawford Professor and
Chairman. Schmitz had taught at the University
of Illinois-Urbana since 1962, and is widely known
for his pioneering experimental research on dy-
namics and control of chemical reactors.
As with other Chemical Engineering depart-
ments, we are also experiencing a large growth in
our undergraduate enrollments. We had 30
bachelor's degree graduates in 1978-the average
per year for the preceding twenty years-and 50
in 1979. Our current senior class is 70, and we
have 87 sophomores! Female students were first
admitted to Notre Dame at the undergraduate level
in 1972; our first batch of female bachelor's degree
holders graduated in 1975. Today 22%o of our


undergraduate students are female, which almost
equals the university figures and is significantly
higher than the rest of the College of Engineering.
Three new tenure-track faculty, all fresh
Ph.Ds, have been hired this year; C. F. Ivory and
J. C. Kantor from Princeton, and M. A. McHugh
from Delaware. We have 31 fulltime graduate
students in residence this year-which with the
present faculty strength permits excellent student-
faculty interaction.

T HE B.S. DEGREE IN CHEMICAL Engineering is a
four year program requiring 128 semester
hours. The 24 hours of liberal arts required, are
well in excess of ECPD requirements. We hope to
increase that number as Notre Dame deems a

Professor James J. Carberry providing valuable
tips on reactor design.
liberal education to be of paramount importance
whatever be the students professional calling. In
addition to the usual Chemical Engineering and
allied chemical and physical sciences, there is an
engineering core including a course in the Fresh-
man year, Introduction to Engineering in which
engineering methodology is taught along with
computer training, including use of FORTRAN.
The remaining core engineering courses are in
probability, engineering mechanics, electrical engi-
neering science, thermodynamics, materials
science, and applied mathematics. Chemical Engi-
neering courses include mass and energy balances,
stagewise operations, transport phenomena, chemi-
cal reaction engineering, thermodynamics, process
design, process modeling and control, transport
processes laboratory and unit operations labora-
The graduate curriculum includes courses in

advanced thermodynamics, statistical thermo-
dynamics, mathematical methods (2 courses),
transport phenomena (2 courses), advanced re-
action engineering, heterogeneous catalysis, poly-
mer engineering, numerical methods, process
control and modeling, dynamics of reaction pro-
cesses (2 courses), heterogeneous phase equilibria,
and equilibrium state operations. In addition to
the formal courses, there is a Chemical Efigineer-
ing seminar, advanced topics in Chemical Engi-
neering, advanced studies projects, and research
projects leading to M.S. and Ph.D. degrees. There
is a research M.S. degree which requires 24 hours
of course work and a thesis. However most
students elect to pursue the non-research M.S.
degree which requires 30 hours of course work.
Graduate students are encouraged to take gradu-
ate level courses from outside the department,
particularly in chemistry, physics, mathematics,
or engineering. First year graduate students are
required to take a 3 credit special topics course in
which they choose an advisor and devote them-
selves to solution of an original problem.
The Ph.D. degree requires the course work for
the M.S. degree plus additional courses in Chemi-
cal Engineering and science deemed important to
the candidate's research area. The progress of each
Ph.D. student is reviewed periodically by his or
her research committee. Normally the M.S. degree
can be obtained in one calendar year of fulltime
study, while the Ph.D. degree ordinarily takes
from three to four years of fulltime study beyond
the undergraduate degree.

A casual discussion of important issues among
faculty members, from left: Arvind Varma, James Kohn
and William Strieder.


Graduate research in the Thermodynamics and Phase
Equilibria laboratory.

J. T. Banchero, Professor Emeritus, Ph.D.
University of Michigan, 1950. Co-author of text-
book "Introduction to Chemical Engineering" with
W. L. Badger, McGraw-Hill Company (1955), and
"Unit Operations" with G. G. Brown et al., J. Wiley
Company (1950). Dr. Banchero has interests in
liquid-phase epoxide reactions, design of chemical
reactors, and thermodynamics of solutions.
J. J. Carberry, Professor, Dr. Eng., Yale Uni-
versity, 1957. Author of textbook "Chemical and
Catalytic Reaction Engineering," McGraw-Hill
Book Company (1976), and co-editor of the journal
"Catalysis Reviews." Dr. Carberry has con-
centrated his research interests in chemical re-
action engineering and heterogeneous catalysis,
and in 1976 was recipient of the Wilhelm Award
in Chemical Reaction Engineering (AIChE).
C. F. Ivory, Assistant Professor, Ph.D., Prince-
ton University, 1980. Dr. Ivory has wide-based
interests in transport phenomena and biosepara,
J. C. Kantor, Assistant Professor, Ph.D. Prince-
ton University, 1980. Dr. Kantor is interested in
process analysis, dynamics and control, and applied
J. P. Kohn, Professor, Ph.D. University of
Kansas, 1955. Dr. Kohn has interests in applied
thermodynamics, heterogeneous phase equilibria,

transport phenomena, and solar energy.
M. A. McHugh, Assistant Professor, Ph.D.
University of Delaware, 1980. Dr. McHugh has
interests in high pressure phase equilibria, super-
critical solvent extraction, and application to coal
R. A. Schmitz, Keating-Crawford Professor
and Chairman, Ph.D. University of Minnesota,
1962. Dr. Schmitz has interests in dynamics and
control of chemical reactors; instabilities and
oscillatory phenomena in chemically reacting
systems and in kinetics of gas-liquid reactions. He
won the Colburn Award of AIChE in 1970, and the
Westinghouse Award of ASEE in 1977.
W. C. Strieder, Associate Professor, Ph.D.,
Case Institute of Technology, 1963. Co-author
with Rutherford Aris of the monograph "Varia-
tional Methods Applied to Problems of Diffusion
and Reaction," Springer-Verlag, (1973). Dr.
Strieder has interests in diffusion in porous media,
kinetic theory, surface phenomena, and molecular
theory of transport processes.
A. Varma, Professor, Ph.D., University of
Minnesota, 1972. Dr. Varma has interests in chemi-
cal and catalytic reaction engineering, modeling
and simulation and applied mathematics. Varma,
with Aris, recently co-edited "Mathematical
Understanding of Chemical Engineering Systems
-Selected Papers of Neal R. Amundson," Perga-
mon Press (1980).
E. E. Wolf, Associate Professor, Ph.D., Uni-
versity of California, Berkeley, 1975. Dr. Wolf
has interests in heterogeneous catalysis, chroma-
tographic separations, and chemical reactor engi-
neering. Wolf recently translated Carberry's text
into Spanish.
In sum, the Chemical Engineering program at
Notre Dame has evolved through three stages:

An exclusively undergraduate program, 1909-1946.
An undergraduate and master's program, 1946-1960.
A B.S., M.S., and Ph.D. program since 1960.
With the appointment of Roger Schmitz as
Chairman in 1979, and the addition of three as-
sistant professors in 1980, a fourth stage of ad-
vancement is manifest insofar as our research
strength has been broadened: chemical and
catalytic reaction engineering, heterogeneous
catalysis, applied mathematics, thermodynamics
and phase equilibria, transport phenomena, and
process dynamics and control. We are happily
charged by our administration in the spirit of
Verdi's Falstaff: "Continua!". D



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
join us in the 1980's.

:. NIO

an equal opportunity employer

We invite you to encourage qualified students
to see our representatives on campus-
or write to:
Coordinator, Professional Placement
Union Carbide Corporation
270 Park Avenue
New York, N.Y. 10017


Rilahd M. qelAde

North Carolina State University
Raleigh, NC 27650

RICH FELDER JOINED the Department of Chemi-
cal Engineering at North Carolina State Uni-
versity in July 1969. He holds the rank of Pro-
fessor and has established himself as a leader in
chemical engineering education. His contributions
in teaching, research and administration reflect
commitment to the profession and engender ap-
preciation and admiration from his colleagues.

action Kinetics and Reactor Design, Thermo-
dynamics, Chemical Process Systems and Chemi-
cal Process Principles, which is the first course in
chemical engineering. In addition he teaches
graduate courses in Chemical Kinetics, Reactor
Design, Optimization, Process Modeling and
Special Topics in Coal Gasification.
His ability to establish rapport with students
makes him one of the most popular faculty mem-
bers in the Department. Undergraduate and gradu-
ate students have consistently evaluated his
courses as being among the best they have taken.
In recognition of his teaching performance he has
received a School of Engineering Outstanding
Teacher Award and has been named to the NCSU
Academy of Outstanding Teachers.
Students particularly compliment the clarity
with which Rich can present a lecture. Course
evaluation forms describe him as "always pre-
pared," "clear and easy to follow" and a "great
teacher." More than any other faculty member,

he is responsible for glowing compliments con-
sistently used by students to describe the intro-
ductory course in chemical engineering. The de-
manding nature of this course makes such atti-
tudes remarkable.
Rich has a classroom style centered about what
might be called a present-and-probe approach.
The "present" part always consists of a routine
in which the class is told first what is going to
be covered and why it is important. The concept
is subsequently presented clearly and concisely
and, finally, an example is used to illustrate its
application to a practical problem. The "probe"
part actually occurs during his presentation; he
sprinkles an oversupply of "okay"? and other
rhetorical questions throughout the lecture, some-
how sensing which points have been grasped by
students and which need additional coverage.

His ability to establish rapport with students makes him one of the most popular faculty
members in the Department . In recognition of his teaching performance he has received a School
of Engineering Outstanding Teacher Award and has been named to the
NCSU Academy of Outstanding Teachers.

Copyright ChE Division, ASEE, 1981
Copyr-ight ChE Division, ASEE, 1981


The reorganization of the
Ph.D. qualifying exam is one of Rich's
most significant administrative accomplishments.

His work with students outside the classroom
is also exemplary, both in his official capacity as
Graduate Administrator and in his unofficial role
of confidant and consultant. His concern for the
development of the complete person fosters state-
ments on student evaluations like "he is always
willing to discuss a problem, even if it is a personal
In recent years Rich has given numerous in-
dustrial short courses in Basic Principles of
Chemical Engineering, Process Maximization,
Polymer Reactor Technology and Separation Pro-
cess Technology. In participant critiques of these
courses he draws raves for his clarity, style and
quantity of material covered.
Finally, no description of Rich's teaching
would be complete without mentioning his work
with countless elementary and junior high school
students. It is not unusual to see him charging
out of his class on chemical kinetics and into a
meeting with a third grade class from one of
Raleigh's magnet schools for extraordinarily
gifted and talented kids. He will have volunteered
to discuss something like coal gasification with
these children, and will do so with the same clarity
and enthusiasm he has for his senior class in re-
actor design. It is believed that he could, given a
little notice, walk into a class on almost any subject
and present a lecture that would result in students
learning the material he presented.


all graduate student activities in the Depart-
ment, including applications for admission, ad-
mission of new students, selection of a research
advisor and graduate committee, selection of a
minor, administration of the Ph.D. qualifying
examination and scheduling of final oral examina-
tions. Graduate students recognize him as their
most important contact, outside of their thesis
supervisor, during their stay at North Carolina
State University. The reorganization of the Ph.D.
qualifying exam is one of Rich's most significant
administrative accomplishments. It has been given
in its present format for almost 10 years without
significant complaint.

RICH'S RESEARCH IS BASED on his interests in
modeling chemical processes and the effects
of chemical processing on the environment. These
broad concerns have led to a variety of research
programs, including radioisotope applications to
process analysis, photochemical reactions, use of
polymeric interfaces for stack sampling, modeling
of electrostatic precipitators and, most recently,
environmental effects associated with coal gasifica-
tion. He has published forty-five articles in
refereed journals as a result of research in these
fields and he was recently awarded a patent for
the novel use of polymeric interfaces in stack gas
The Environmental Protection Agency sup-
ports his research on the development of a collec-
tion tube for use as an interface in stack sampling,
modeling the performance of electrostatic precipi-
tators and the environmental effects of coal gasi-
fication. His research on the evaluation of trace
metals and sulfur gases from coal is supported
by the Department of Energy. The NCSU Faculty

Professor Felder giving impromptu lecture on
Research and Professional Development Fund has
supported the construction of a batch photoreactor,
measurement of beta spectra emitted by a slab
source, and dynamic simulation of a pulp chlorina-
tion tower.
The research efforts described above have been
marked by the same thoroughness and quality as
Rich's teaching. He was given the 1974 NCSU
Sigma Xi Research Award for superior ac-
complishments in the field of scientific research.
Rich has unselfishly contributed to other re-
search programs in the Department. His library
of computer programs is extensive and he makes


Felder with colleagues, Hal Hopfenberg (Depart-
ment Head) and Ron Rousseau.
them available to colleagues and students. Further-
more, he can generally be counted on to have the
latest information on analytical instrumentation.
When the coal gasification research project got
underway, there was a need for someone to learn
capabilities of various instruments available for
analyses of coal and char, waste water and gases
found in the gasification/gas cleaning pilot plant.
Rich accepted the responsibility and maintains an
expertise in the use of an array of instruments,
including atomic absorption spectrometer, ion
chromatograph, gas chromatographs, sulfur ana-
lyzer, nitrogen analyzer, etc.
Despite the enthusiasm Rich has for these re-
search efforts, he never loses sight of the proper
relationship between university research and
graduate students. He is concerned with their total
professional and personal development, as well as
their accomplishment of short range goals like ac-
cumulating data points and, ultimately, a degree.
None of Rich's students has the slightest re-
luctance to take their "problem of the day" to
him. His mortality and temper have been ob-
served, however, when a student has just broken
the third calibrated rotameter in a single day.


complete without mentioning his enjoyment of

the written word. (That's not to say he doesn't
like the spoken one, too.)
Although the 45 publications mentioned earlier
and the textbook Elementary Principles of Chemi-
cal Processes may be indicative of the quality of
Rich's writings, they are not true indicators of the
quantity. He is prolific. It is common knowledge
among graduate students that any question may
result in being given a handout specially prepared
for such occasions. A look around his office will
reveal stacks of documents he has written on
everything from "What is Chemical Engineering"
to "Radiotracer Applicators in System Analysis."
Writing a paper or a book with him can be both
excrutiatingly painful and exhilirating. It can be
painful because Rich's search for just the right
word or phrase can result in numerous rewritings.
None of his students or colleagues can escape his
trusty scissors, tape and stapler. His saving grace
in this regard is that he is just as likely as not to
rewrite one of his own drafts. And the exhilirating
part is that the document always sounds and reads
It seems only fitting that I say a few words
about Rich's unique contribution to our joint
authorship of Elementary Principles of Chemical
Processes. This textbook was published in 1978
and has been adopted for use at approximately 80
Departments of Chemical Engineering in the
United States, several European universities, and
will be translated into Spanish and Chinese. Two
characteristics of the book which have been a
factor in its widespread adoption are its pedogogi-
cal approach and style of writing. Quoting from
a review of the book that appears in the AIChE
Journal, 25, 382 (1979), "... the style and level of
presentation of the content is excellent, and the
subject matter represents the ideal body of
knowledge which should be imparted to students
in the first year of a chemical engineering curricu-
lum." These features are primarily the result of
Rich Felder's insight into what students need to
facilitate learning and his very special gift with
the written word. In addition, the exploits of
Sebastian Goniff, Johann Sebastian Farblunget,

It is not unusual to see him charging out of his class on chemical kinetics
and into a meeting with a third grade class from one of Raleigh's magnet schools for
extraordinarily gifted and talented kids. He will have volunteered to discuss something
like coal gasification with these children and will do so with the same clarity
and enthusiasm he has for his senior class in reactor design.


I -

Edd Seddera and other seedy but colorful charac-
ters sprinkled throughout the text, are illustrations
of his sense of humor and his ability to couple wit
with the illustration of chemical engineering


A AFTER A CHILDHOOD IN Manhattan, Queens,
Buffalo and North Miami, Rich settled down
to make good grades and enter the City College of
New York to study chemical engineering. Why
Chemical Engineering? (a) Everyone who knew
anything in 1957 was going into engineering and
(b) he was intrigued by the idea of mixing color-
less liquids to produce a bright orange (his favor-
ite color and the reason for his subsequent

Felder posing (i.e. acting like he's doing something)
in front of coal gasification/gas cleaning pilot plant
control panel.

matriculation at Princeton) fluid. Why CCNY?
He didn't like the letter MIT wrote to him, and
think of all the carfare he saved.
He graduated from City College No. 2 in engi-
neering and entered graduate school at Princeton.
His interests became chemical physics, Barbara
Cowl (whom he married), and hopeless liberal
causes, not necessarily in that order. He was
Morton Kostin's first graduate student, but he
claims his major achievement was a catch in a
softball game for which he is still revered by old
timers on the Princeton faculty like Ernie Johnson
and Bob Axtmann. He also tried Electrical Engi-
neering at Princeton but, on falling through the
ceiling at 224B Halsey Street while installing air
conditioner wiring, he decided his future remained
in Chemical Engineering.

He enjoys music ranging
from Mozart to McCartney and is
especially fond of playing classical guitar,
but he likes to play in groups so that
his mistakes are less obvious ... he also
memorizes Gilbert and Sullivan patter
songs, believing that someday
they will come in handy.

Following his Ph.D., he spent a year at Har-
well, England as a NATO postdoctoral Fellow,
and two years at Brookhaven National Laboratory.
His research interests had now shifted to photo-
reactor design and analysis, and mixing effects
in reactors. Letting it become known that he was
interested in an academic career, some enlightened
soul recommended he interview for a vacant posi-
tion at North Carolina State University. With his
interest in United States geography confined to the
blighted land mass east of the Hudson River and
the hedonistic environment surrounding the San
Francisco Bay, he is reported to have uttered the
famous quote for which many New Yorkers later
claimed credit: "North Carolina??? Get serious,
wouldja!!". Nevertheless, being basically curious,
he came, he saw, he stayed. Why? He fell in love
with the school, faculty, and (you guessed it)
North Carolina. He claims never to have regretted
the decision, except when he finds himself almost
enjoying eating grits with red eye gravy.
He maintains his passion for hopeless liberal
causes, bialys with whipped cream cheese and
Chivers Olde English Marmalade. He enjoys music
ranging from Mozart to McCartney and is es-
pecially fond of playing classical guitar, but he
likes to play in groups so that his mistakes are less
obvious. He is often awed by our friend and
colleague Jim Ferrell, who makes and plays classi-
cal guitars. You're not going to believe this, but
he also memorizes Gilbert and Sullivan patter
songs, believing that some day they will come in
Rich is also the devoted father of Kenneth,
Elena and Gary, ages 14, 12 and 10. Despite his
tireless professional efforts, he always seems to
have enough energy left over to spend time with
his children. One of his passions that he indulges
in frequently is being beaten in chess by these
budding Bobby Fishers. As a father, he exhibits
a degree of patience and concern which reflects his
feelings about children and, in fact, all people. EI


stirred pots


EDITOR'S NOTE: We are grateful to Prof. John Seinfeld
(California Institute of Technology) for bringing our
attention to the following poem, written by Prof. Ruther-
ford Aris of the University of Minnesota to congratulate
Princeton University's Chemical Engineering Department
on their fiftieth anniversary.

When you're taking a whack,
in your orange & black,
at a problem defying solution
In kinetics or flow,
or the ways that bugs grow,
or just how to get rid of pollution.
When you're wracking your brain
& you cannot obtain
a result that fulfills your intentions
Remember that Chem.E.s
do nothing by semis,*
and always get rid of dimensions
Recollect that p'raps Schmidt's
is the number that fits
and that Peclet is Reynolds times Prandtl.
Just remember your D's,
it'll come out with ease
when it's wrapped up in Damkdhler's mantel.
If you wish you could borrow
relief from your sorrow,
the Ohnesorge number we'll mention,
Into Reynolds you shoot
the good Weber's square root-
it accounts for the surface's tension.
For tubes in the boiler,
use t'number of Euler
and for drag use the factor of Fanning,
Which in matters of fiction
was used for the friction
twixtt Newman and Cardinal Manning.
But it's really quite rude
to rhyme Froude with St. Jude
or to mess up a grave patronymic,

To dress up in gingham
the number of Bingham,
or regard Boussinesq's as a gimmick,
While to think that Biot
has to do with the glow
that both heat and hard work often foster,
Or that Deborah's song
has the permanent pong
of an overly ripe double Gloucester.
To madden a Hatta
or exchange for the latter
the group that's been named Svant Arrhenius,
To mock at the balding
of Sherwood or Spalding;
these are all of 'em crimes hyperhein (i)ous.
If none of these fit,
just reflect for a bit,
devise one of your own-
it's a sine qua non-
for there's many a name
that has brought Princeton fame.
For size of reactor,
use great Wilhelm's factor;
for hard computation,
'tis Leon's equation;
and if that don't fit well
the number of Whitwell.
If you've pulled a slight boner,
be absolved by Dick Toner;
your problem's all garbled?
Why, Dean Elgin's not marbled;
While, if biking's your goal,
use the Johnson Control.
You've got Schowalter (Bill),
and there's Dudley Saville,
Andres, Russel and Hall,
Prud'homme, Ollis et al.**,
for from such you can scoop
a dimensionless group
and you'll out of your fix be
with some help from Ms. Bixby.
But away cerebration.
Now 'tis your celebration
and we all think it's nifty
that you flourish at 50.
So we wish you good cheer
for full many a year.


*Except Centenaries, of course.
**With due apologies to those not mentioned.



Chevron Oil Field

Research Company

PhD Chemical Engineers e
For Research And Development
In Enhanced Oil Recovery

Chevron's laboratory in La Habra, California is
engaged in research directed toward Increased
recovery of oil and gas from known subsurface
reservoirs. Chemical engineering technology is
extremely important in the very complicated
business of recovering petroleum from known
reservoirs-reservoirs of oil and gas already
discovered and in quantities large enough to
make a real difference in the United States'
domestic energy supply. That is, if we can find
more effective processes for breaking it free
from the rocks and bringing it to the surface.
The research, the development and the field
trials of new ideas for recovering oil carry
high risks and high costs. But the stakes are
high, too! When you realize that typically,
twice as much oil is left behind as is
produced by conventional methods, it is
easy to understand how large these
stakes really are and the energy resources
that will be available if we can find the
unlocking processes.
Our chemical engineers are also
working on the problems of in situ
recovery of heavy oils and oil from tar
sands and shale.
If you want to learn more about
research in the more complex
applications of chemical engineering,
send your resume to:

J.C. Benjamin
Chevron Oil Field Research Company
P.O. Box 446
La Habra, CA. 90631


- -


~~g "' 0




;1P, ZJ0 C/7

4wawd 2ecW e



The 1979 ASEE ChE Division Lecturer was
Dr. Daniel Perlmutter of the University of
Pennsylvania. The 3M Company supports this
annual award.
Dan Perlmutter earned his Bachelor's Degree
(Magna Cum Laude) from New York Univ. and
his Doctorate in ChE from Yale Univ. From 1955
to 1958 he worked for the Exxon (then Esso)
Standard Oil Company. His academic career
began in 1958 at the U. of Illinois, Urbana, and
moved to the U. of Pennsylvania in 1964. Follow-
ing a reorganization of the College of Engineer-
ing and Applied Science, he became the first Chair-
man of the department of Chemical and Biochemi-
cal Engineering, a position he held from 1972 to
1977. He has been a visiting Professor at Harvard
U., the U. of Manchester, the U. of Zagreb, and
the Hebrew U. of Jerusalem, in the capacity of a
Fulbright Professor in England and Yugoslavia
and as a Guggenheim Fellow in Cambridge.
He has developed new course materials in
chemical reactor control, optimization, and
stability problems. His textbook, Introduction to
Chemical Process Control, was one of the first
available in its field, and was widely adopted in
the U.S. and abroad. His monograph on Stability
of Chemical Reactors provided a unified view of
a wide range of questions by combining some
original work with an integrated survey of ma-
terial only available in scattered journal articles.
His most recent research has been on the kinetics
of gas-solid reactions, prompted especially by
their connections with energy-related problems in
coal drying, oxidation, and gasification, as well
as reversible storage in the form of heat of re-

University of Pennsylvania
Philadelphia, PA 19104

T HE WIDELY PROCLAIMED energy crisis that so
profoundly affects world affairs should be
understood to be caused not by a lack of fossil fuel
resources, as much as by a shortage of fuel in the
highly desirable gaseous and liquid forms. It is
accepted in all quarters that proved coal reserves
are ample to supply anticipated world energy
needs for at least several hundred years and
further that no geographic monopoly is possible,
since the deposits are very widely distributed over
virtually every continent. It is evident in this
context that the technology of the so-called gasi-
fication and liquefaction processes are central to
handling the crisis: techniques for coal conver-
sion are being sought that will supply transporta-
tion, space heating, chemical feedstocks, and in-
dustrial needs at prices comparable to petroleum
sources. The objective moves closer as OPEC
prices continue to rise.
It is convenient to think of gasification as com-
posed of three steps: (i) coal pretreatment, (ii)
pyrolysis, and (iii) char reaction, even though a
particular process may circumvent one or more
of these, or may be designed to accomplish several
objectives in a single reactor. The step-wise view-

Copyright ChE Division, ASEE, 1981


point leads most smoothly into an examination of
the appropriate chemical and physical properties
of coal, and of the changes that need to be sought
and monitored.

T HE PRIMARY MOTIVATION for a study of coal
pretreatment is a practical processing diffi-
culty. When common bituminous coals are sub-
jected to high temperature gasification conditions
they tend to soften and swell, and can plug re-
actors and transport lines before ultimately re-
solidifying as a porous char or coke. This process
is commonly referred to as caking; it must be con-
trolled if the solid fuel is to be successfully gasi-
fied. A mild oxidative pretreatment is commonly
used to reduce or eliminate this coal caking pro-
pensity, providing two steps in one: drying to re-


360 400 440 480 020 640
FIGURE 1. Caking test results for feed coal samples.

move naturally occurring water on the coal
surfaces, and the chemical oxidation.
Many laboratory test methods have been used
to study the plastic properties of coals, all em-
pirical in nature and devised to characterize the
plastic behavior of coal by means of numerical
indices. Comprehensive summaries and reviews of
the methods have been given by Brewer (1945)
and Loison et al. (1963). One common test is the
Gas Flow test (Foxwell, 1924; Coffman and
Layng, 1927, 1928), chosen for its simplicity in
apparatus design and operation and for its close
resemblance to the caking of coal in a packed bed

... techniques for coal
conversion are being sought that
will supply transportation, space heating,
chemical feedstocks, and industrial
needs at prices comparable
to petroleum sources.

reactor. In this test, a sample of the coal is heated
at a constant rate of temperature rise (2C/min)
inside a tube with a constant flow of nitrogen
through the coal sample. The differential pressure
drop across the sample is monitored, as the sample
goes through its plastic range. Pressure drop vs.
temperature readings are taken as data, as shown
in Figure 1. After the oxidizing pretreatment the
same test shows a sharp reduction in pressure
drop as in Figure 2, reflecting a lesser degree of
softening, swelling, and "caking."
Changes recorded as mechanical properties
need to be matched to chemical effects, if one is
to understand the details of the preoxidation. For
this purpose, a HVA bituminous coal was oxidized
in a packed-bed reactor with a once-through mix-
ture of N2 and 0,, monitored by analyzing the
composition of the feed and product gases with a
dual-column Gas Chromatograph.
Typical results on overall reaction rates are
shown in Figures 3 to 7, emphasizing the effects of
flow rate, particle size, oxygen concentration, pres-
sure, and temperature, respectively. In addition,
the chromatographic analysis run on each sample
of reactor effluent provided a record of carbon
dioxide and carbon monoxide production to match

s0 o400 440 480 520 540
FIGURE 2. Caking test results for -18 to +50 mesh
oxidized coal samples.


the rates of oxidation. Cross plots are presented
as Figures 8 and 9 to illustrate the linearity and
to demonstrate that the correlations for the
various coals do indeed persist as flow rate,
particle size, and conversion vary. Evidently, the
two carbonic gases are formed in a constant ratio,
presumably by the same mechanism.
Correlation of all these findings by means of a
comprehensive argument requires formulation of
a model that includes several steps. The two car-
bonic gases appear to form together by a direct
burnoff reaction. Simultaneously, oxygen is ad-
sorbed by the coal to form any of a series of oxy-

0 2 4 6

FIGURE 3. Effect of feed gas flow rate

on 0, reaction

functional groups, and hydrogen is removed in a
direct water formation reaction. Superimposed on
this complex chemistry are possible transport
limitations outside and inside the coal particles.
Quantitative tests have been used to demonstrate
consistency between the data and such models
(Kam et al., 1976; Karsner and Perlmutter,
Turning to more immediately practical results,
it is of interest to relate oxidation rates to the
geologic history of various coals. In Figure 10
comparable rates are presented as a function of


S 0 22 -18 +50
S/ A 21 -14 +18

o 16'8 E]23 -6 +14


c 14 -




0 2 4 6 a
FIGURE 4. Effect of particle size on 02 reaction rate.
T 225 C
0 47 286
40 -
< 44 20-2
S45 13.9

S[] 46 9-7
2 30 C)


o o


0 2 4 6 8
FIGURE 5. Effect of feed 02 concentration on oxygen
reaction rate.


FIGURE 6. Effect of pressure on oxgen reaction rate.

S 2 4 6 8
FIGURE 7. Effect of temperature on oxygen reaction

carbon content, a commonly used index of a coal's
rank. As shown, by the duplicate runs, the cross-
hatched band is primarily a reflection of vari-
ability among coals and only in minor part caused
by experimental scatter. A balanced evaluation of
pretreatment must also consider the economic
losses that accompany any reduction in heating
values. The data show a strong correlation with
the carbon content of a sample. Regardless of
particle size, the heating value of a material
changed upon oxidation primarily as the carbon
changed for a wide range of coals of different
types. There is also a suggestion in the data that

Regardless of particle size, the
heating value of a material changed
upon oxidation primarily as the carbon
changed for a wide range of
coals of different types.


3.0 -



Run No. Mesh Size
o 10A -6+16
o 11 A -16+18
, 12 A -18+50
* 23W -6 +16
* 22 W -16+18

1.0 /


CO Production Rate, g/KG Coal/Hour
FIGURE 8. Relative carbon dioxide and carbon
monoxide production rates. Coal: MV
bituminous, PSOC 135.


2.0 r-

generalizations regarding the fixation of oxygen
that occurred for all the various bituminous coals
do not apply to the lignite (PSOC 87) and the
anthracite (PSOC 80) coals.


T URNING ATTENTION TO the downstream side of
a coal gasification process, models are needed
to describe kinetics of the several gas-solid re-

20 -

10 l-

CO Production Rate, g/KG Coal/Hour
FIGURE 9. Relative carbon dioxide and carbon
monoxide production rates. Coal: HVC
bituminous, PSOC 190.

actions involving porous chars. The experimental
reports of Hashimoto et al. (1979) provide a good
point of departure, since any model to be de-
veloped must be consistent with the features of re-
action such as are presented in Figure 11.
Above all a viable model must permit the de-
velopment of a maximum in rate (or reactive
surface) as a function of conversion. A promising


0 8W 4
A12W 197
V27W 127
024W 135
K ---
V 9A 127
*12A 135

0 I I I I I
65 70 75 80 85 90
Carbon Content, Wt. :' (ultimate)
FIGURE 10. Effect of coal carbon content on initial
oxidation rate for -18 + 50 mesh particles
at 2250C, 100 SCCM.

candidate in this direction has been developed by
Bhatia and Perlmutter (1980), who considered
the isothermal chemical reaction of particles of
the solid B with a fluid A according to the stoi-

aA(g) + bB(s) -> pP(g) + qQ(s) (1)

The reaction is initiated on the surfaces of pores
in the solid B. As further reaction occurs, a layer
of product Q is formed around each pore, which
separates the growing reaction surface of the
solid B from the fluid reactant A within the pores.

Continued on page 49.

6001 CS 850 t
1:0 c. \ "C
,-, f tx4260mesh
E 40 -0 14.16mesh
S 8.9 2h \ \
300 *


O0 02 0' 06 01 10
Burnof. Xc [-3
FIGURE 11. Effect of char conversion on surface area,
after Hashimoto et al. (1979).


I i

Monsanto Drive.
It takes you a very long way.

VK'I h 40 q

This sign marks the
into our International
St. Louis.

road that leads
Headquarters in

These words, "Monsanto Drive"
have another and more significant mean-
ing at Monsanto. It's a way of expressing
the special qualities of Monsanto people,
who have the will to meet challenges
head-on-to accomplish and succeed.
We offer bright and energetic people
with this drive the opportunity to help
solve some of the world's major problems
concerning food, energy, the environment
and others.
Challenging assignments exist for
engineers, scientists, accountants and

marketing majors at locations throughout
the U.S.
We offer you opportunities, training
and career paths that are geared for
upward mobility. If you are a person
who has set high goals and has an
achievement record, and who wants to
advance and succeed, be sure to talk
with the Monsanto representative when
he visits your campus or write to:
Buck Fetters, University Relations and
Professional Employment Director,
Monsanto Company, 800 North Lind-
bergh, St. Louis, MO 63166.

An equal opportunity employer


Aq~r, 7W,



University of Washington
Seattle, WA 98195

F INITE ELEMENT METHODS are being used in-
creasingly for engineering studies of struc-
tures, heat transfer, fluid flows, design of dams
and flow of water in aquifers. Because of this
widespread use the well-educated engineer should
have some experience applying the finite element
Finite elements are most effective for solving
two-dimensional problems on irregular domains,
but their use is sometimes warranted in one-
dimensional problems. Chemical engineers have
used finite element methods in 1-D, without calling
it a finite element method. Sometimes finite ele-
ments are not needed, and an analyst should know
when to use finite elements and when not. This

Bruce Finlayson has extensive experience modelling chemical
engineering systems. He is well known for developing the orthogonal
collocation method as an efficient computational tool for modelling
chemical reactors with both radial and axial dispersion and monolith,
wall catalyzed reactors. More recently he is known for work in finite
element methods solving heat transfer and flow problems, particularly
viscoelastic polymeric fluids and flow in porous media. He is the
author of two books: The Method of Weighted Residuals and Varia-
tional Principles, Academic Press, 1972, and Non-Linear Analysis in
Chemical Engineering, McGraw-Hill, 1980.

*This paper was presented at the ASEE meeting in
Baton Rouge, LA, on June 24, 1979.

J, x / J
'7/7,.., r1771/1, 1/ 77 / 77/71'//.'/ / V- 7/-

FIGURE 1. Flow Past a Flat Plate.

distinction is less important for the practicing
engineer who may use an available program.


P PROBABLY THE OLDEST USE of finite element
methods is for flow past a flat plate (see
Figure 1). In the integral method (1, p. 142) we
divide the region of space into two elements: a
small element near the plate and a larger element
away from the plate. We represent the velocity in
these two regions by two functions:
outer element: u = U.
inner element: u = U~. (71),
S= y / 8(x)
The boundary layer thickness is 8(x). It is this
division of the region into two, with expansions of
velocity in each region, that makes this a finite
element method. We must still determine 4 (7)
and 8(x).
These expansions are substituted into the
boundary-layer equations (see 1, p.142; 2, 4.2 for
the details) and the result is called a residual. We
would like the residual to be zero since then the
differential equation is satisfied. That may not be
possible, particularly if the expansions for 4)())
or 8(x) are simple. In the integral method we
integrate the differential equation over the
domain-in this case over the inner element-and
set the result to zero, thereby making the
differential equation zero "on the average." The
result is an equation for the boundary layer
Copyright ChE Division, ASEE. 1981


(B A) 8 -- C (1)
dx y.
The numbers A, B, and C can be calculated once
)(7) is known, and Eq. (1) is easily solved.
The expansion for 4 (q) is made as simple as
possible: a polynomial. There are certain condi-
tions it must satisfy: The velocity should be zero
at y = 0 (on the wall) or at ) = 0: ((0) = 0.
The velocity should match at the node between
the inner and outer element, which occurs at y =
8(x) or -q = 1: 0 (1) = 1. Likewise it is con-
venient to have the slopes match at i- = 1: (1)
= 0. Applying these conditions to a quadratic
function, 4 = a + bq + ci2, gives the expansion
for ()7).
4 = 2r7 q2
Ref. (1) uses

2 2
S- 2. 2 -<(2)

which is derived by using a cubic polynomial and
adding the condition )" (0) = 0, which is obtained
by making the differential equation be satisfied at
the point 71 = 0 on the wall. Eq. (2) gives a
slightly better result. With this function we have
a representation for the velocity field in the whole
domain. By satisfying the equation "on the aver-
age" we are using the integral method, and by
representing the velocity by piecewise polynomials
in different regions we are using finite elements.
The finite elements are necessary because of the
sharp changes of velocity near the wall.
For the next example we turn to diffusion and
reaction in a spherical catalyst pellet. Consider the
reaction of carbon monoxide with oxygen in an
alumina catalyst coated with small amounts of
platinum. The rate of reaction has been measured
to be
Rate =
(1 + Kac*)2
where k and Ka depend on temperature (see 3,
p. 461). The equations to be solved are

1 d dc
r2 dr dr

R2 k(T)c
D, (1 + ac)2

Bim(c-l) atr=l

where we have written the non-dimensional r =
r' / R, with r' the dimensional radial position in
the catalyst and R the radius. c = c* '/ co, where
co is the external concentration of CO, taken here

Finite elements are most
effective for solving two-dimensional
problems on irregular domains, but their
use is sometimes warranted in
one-dimensional problems.

as 4%, and a = Kaco = 5 at 58C. For purposes
of illustration we solve this for Bim = 20 and a
variety of R2k (T) / De = 42, the Thiele modulus
squared. A quantity of interest is the average rate
of reaction, compared to the rate of reaction if
there were no diffusion limitation. The effective-
ness factor is
1 1
S Rate (c(r))r2 dr / Rate (c = 1)r dr
0 0
Now we know the concentration profile depends
strongly on 02, which is the ratio of diffusion
time to reaction time. If the diffusion time is very
small (i.e. diffusion is very rapid) then the con-
centration profile is as indicated in Figure 2a.
Under these conditions the concentration can be
easily represented by a polynomial in r, and finite
elements are not indicated. By contrast, if 02 is
large then the reactants are rapidly depleted in
the pellet, and concentration profiles such as Fig.
(2b) are possible. Then finite elements are indi-
cated-express the solution as one function in the
inner core and another in the outer boundary
For low 0 we can use the orthogonal collocation
method to solve the problem. The concentration is
expanded in a quadratic polynomial.
c(r) = c2+ a(l-r2)
Higher order polynomials are possible but are not
needed except for high accuracy. We find it more
convenient to solve in terms of c2 = c(1) and c, =
c (rj), where r, is the collocation point = 0.65465.
The orthogonal collocation equations are given
here (for details see 2, 5.3 and 4).

sk,141 d

FIGURE 2. Concentration in Catalyst Pellet.


Bilc + B12c2 = R2k(T)cl /1 De(1 + ac,

-(A21 1 + A22c2) = Bi(C2-1)
7 = 3(wR (cl) + wR(c2)) / R(c = 1)
Bil = -B12 = -10.5, A21 = -A22 = -3.5
wt = 0.2333, w2 = 0.1, rl = 0.65465
T, = 1 + 8 + p(1 8)c, 3c
8 = Bi / Bi = kke / (hpDe)
/ = (-AHa) CoDe / keTo)
The equation for Ti is exact. (4, pp. 96-97). 1
we solve for / = 0 isothermall case) and Bi,
20. This solution is not valid for any )2, but i
a good approximation provided c(r = 0)
(see Figure 2a). Since

ce = c2 + a(1-r 2)

a = C C2
a -
1 r2

the approximation is valid for
c2 + a> 0
or c1 r12c2 > 0.
The - 0 curve is shown in Figure 3 and is
distinguishable from the exact solution w
b < 10. This curve is easily calculated, witl
iteration, simply by choosing a c2, finding c, f
Eq. (4), T1 from (5) (if non-isothermal),
R2k(T,) / ci from (3), with a evaluated at'
For large 0 we must use the Patern
Cresswell (5) approach using finite elements
the inner element of Figure 2b we take c = 0
we let the inner element have length b < 1. In
outer element we use the trial function

FIGURE 3. Effectiveness Factor Thiele Modulus for
Carbon Monoxide Reaction, Bim = 20,
a = 5. ---exact, - approximate.

(3) The advantage of the finite
(4) element method is that the elements
can easily be deformed, small elements
can be used in important regions, and
irregular domains are easily handled.

(5) T 100
h 3 h=3
To= To= O

lere =3 h -s
To=O To = 0
n =
ST= 100
- lh =3

h 3 h = q=0
T= 100
FIGURE 4. Heat Transfer Problem.

c = u23,,
u= (r-b) / (-b)
which satisfies c = 0 and dc/du = 0 at the node
between elements. The collocation equations are
now (see 4), for planar geometry,
1-b2 (B21 + B22C2 + B23c3)

= R k(T,)c2 / (De(1 + aC2)2) (6)
1- b (Ai1ci + A32CZ + A33c3)

= Bi (c 1). (7)
Now c = c(u = 0), c2 = c(u = 1/2), at the mid-
point of the outer element, and c3 = c(u = 1)
= c(r=1). We have c3 = 0 and c2 = c3 / 4, so we
can solve (7) for c,.

(1-b) Bim
c 2 + (1 b) Bi ()
For various b from 0 to 1 it is easy to see the
influence of external diffusion resistance. When
cs = 1 there is no concentration drop across the
boundary layer surrounding the pellet, and this
condition depends on (1 b)Bim. For large Bim,
we have cs = 1. These equations are solved with-
out iteration by choosing b, finding c. by (8),
C2 = C3 / 4, and then using (6) to get 0,2 =
Rk (T2) / De. Choosing b = 0 gives the dividing


changing, Figure 2b, whereas non-finite element
methods are suitable for smooth solutions; Figure


A MORE IMPORTANT USE of finite element methods
is for two-dimensional problems. Consider
heat transfer in the region shown in Figure 4.
The boundaries are maintained at temperature =
100 and the interior is cooled with a heat transfer
coefficient h = 3, To = 0. We wish to solve kV2T
= Q in strut, T = 100 on top and bottom boundary,
-kn-VT = h(T To) on side boundaries. Space
does not permit a complete discussion of all
aspects of the finite element method but the high-
lights can be given.
We first discretize the space, as shown in
Figure 5. We can divide the region into triangles,

FIGURE 5. Mesh for Heat Transfer Problem.
line between the orthogonal collocation solution
and the finite element solution. Here we have
solved the finite element solution for planar
geometry (it is easier than spherical geometry),
and we use

A =-
to get the results for spherical geometry. The
7) 0 curve is shown in Figure 3. Over the whole
range of 0 we get a reasonable solution; compared
to much more difficult numerical solutions.
Certainly the solution exhibits all the important
phenomena, usually within experimental accuracy.
This example clearly shows the importance of
using finite elements when the solution is sharply

fl F_

FIGURE 6. Element Deformation.

FIGURE 7. Trial Function, Linear on Triangle.
rectangles or quadrilaterals. Because of the way
finite element programs are written it is very easy
to use deformed elements, as shown in Figure 6.
The actual calculations are made on the square
element, with appropriate account taken of the
transformation properties. Within each element
we expand the unknown function in a polynomial
-either linear or quadratic, usually (see Figure
7). Having decided on the trial function (here
linear functions on triangles) we number the
nodes and elements (see Figure 5). Now in two
adjacent elements the equations are similar when
written in terms of the local coordinates (see
Figure 8). The appropriate equations must be
assembled properly. The equations are solved with

Since finite methods are being used increasingly by industry
for design of structures, heat transfer, fluid flow, design of nuclear reactors,
etc., it is important that the modern student be exposed to them. It is not necessary that
the engineer be familiar with the details ... but (he) should know the general idea.

WINTER 1981:

an LU-decomposition or a Gaussian elimination
(6). These techniques are beyond the scope of
this article.
The user then specifies the thermal conductivity
and heat generation rate in each element. Here we
ke = 1, Qe = 0.
Finally the boundary conditions are specified as
Ti = 100
nodes i = 7, 14, 17, 20, 23
hi = 3, To, = 0
nodes i = 8, 9, 10, 11, 12, 15, 18, 21, 22, 23.
For a boundary node at which temperatures are
not specified by the user the boundary condition
is automatically the natural boundary condition
ZT / an = 0
nodes i = 1, 2, 3, 4, 5, 6, 7, 8
with n the outward pointing normal.
This problem is solved using the program in
Huebner's book (7). The solution is shown in




I 2


I 2

FIGURE t. Finite Element Numbering System.

Figure 9. Naturally more elements are needed
for good accuracy, but the essential elements are
clear in Figure 9.
The advantage of the finite element method is
that the elements can easily be deformed, small
elements can be used in important regions, and
irregular domains are easily handled. The reason
for these features is that the computer programs
are written on an element-by-element basis, since
each element is similar. The user, then, has
complete freedom in how the elements are to be

FIGURE 9. Solution to Heat Transfer Problem. Contours
are for each 10 degree increment.


between finite element and finite difference
methods. For the heat conduction problem

32T +aT
+ -0
Ex2 ry2
a finite difference grid is shown in Figure 10a.
The finite difference method uses the following
equation representing the differential equation at
node 5.

V2T = (T + T T + T 4T- ) / h2.

The same equation results from application of the
finite element methods, provided linear trial
functions are used on the triangles shown in
Figure 10b. (Other types of terms in an equation
may be different, however). If we use quadratic

7_ _8 9 B

4 5 6 4 5 6

I 2 3 1 1 .2 3
FIGURE 10. Comparison of Finite Difference and Finite
Element Methods.


polynomials in the finite element method, then we
have more nodes and the equation at node 5 in-
volves terms at each node represented by in
Figure 10c. Because the equations are derived in
different ways the finite element method is, how-
ever, easy to apply with the irregular geometries
shown in earlier figures. In addition, certain types
of boundary conditions are easily handled in finite
element methods, particularly boundary conditions
involving derivatives and/or free surfaces, whose
location is to be determined.

SINCE FINITE ELEMENT methods are being used
increasingly by industry for design of struc-
tures, heat transfer, fluid flow, design of nuclear
reactors, etc., it is important that modern students
be exposed to them. It is not necessary that the
engineer be familiar with the details of the method,
but the engineer should know the general idea and
be able to apply the method. The author has found
students are quick to learn how to use finite ele-
ment programs, and once experienced will always
know what someone means when they say Finite
Element Methods.

COMPARISONS OF finite element, collocation, and
finite difference methods are given in Ref. (8).
One-dimensional cases are emphasized since that
allows the easiest description of the methods and
details of applications. Two-dimensional problems
are treated there as well as books by Huebner (7)
and Chung (9). Huebner's book contains a simple
finite element program for heat transfer problems.
More elaborate programs are available (10). Ap-
plications to engineering design are widespread,
and one concentrated source is the International
Journal of Numerical Methods in Engineering.
Probably there are published accounts of research
applications of the finite element method in the
journals related to your area of interest. O
1. Bird, R. B., W. E. Stewart and E. N. Lightfoot,
"Transport Phenomena," Wiley (1960).
2. Finlayson, B. A., "The Method of Weighted Residuals
and Variational Principles," Academic Press (1972).
3. Carberry, J. J., "Chemical and Catalytic Reaction
Engineering," McGraw-Hill (1976).
4. Finlayson, B. A., "Orthogonal Collocating in Chemical
Reaction Engineering," Cat. Rev.-Sci. Eng. 10, 69-138
5. Paterson, W. R. and D. L. Cresswell, "A simple method
for the calculation of effectiveness factors," Chem.
Eng. Sci. 26, 605-616 (1971).

6. Hood, P., "Frontal Solution Program for Unsym-
metric Matrices," Int. J. Num. Methods. Eng. 10, 379-
399 (1976); 11, 1055, 1202 (1977).
7. Huebner, K. H., "The Finite Element Method for
Engineers," Wiley (1975).
8. Finlayson, B. A., "Nonlinear Analysis in Chemical
Engineering," McGraw-Hill (1980).
9. Chung, T. J., "Finite Element Analysis in Fluid
Dynamics," McGraw-Hill (1978).
10. Program DOT available from Professor E. L. Wilson,
Department of Civil Engineering, University of Cali-
fornia, Berkeley, California 94720.


Dear Sir:
We read with interest about the successful Study-
Travel Program at Virginia Tech (Summer 1980 issue).
I would like to briefly mention two intensive foreign
study programs which are open to selected Chemical
Engineering undergraduates at Case Western Reserve.
They may also serve as examples for other Departments
who wish to initiate such programs.
Each year since 1978, three of our undergraduates
have spent their junior year at the University of Edin-
burgh. There they are regular full time students in the
third year Chemical Engineering program. The students
normally live in university student housing and can par-
ticipate in the usual range of student activities. Full
academic transfer credit for a years work is granted upon
successful completion of the third year course at Edin-
burgh. This arrangement has been extremely successful,
primarily due to the excellent and continued cooperation
of the Edinburgh faculty. The experiences of the students
have been uniformly good and there have been minimal
academic re-entry problems after returning to Case for
their senior year.
We also participate, with Iowa State University and
Georgia Tech, in a summer laboratory course at Uni-
versity College London. This very well run program
lasts for approximately one month and, in addition to the
intensive laboratory course, includes a one week bus tour
of various British chemical industries. Credit for our
Unit Operations Lab is given upon completion of the
course. Part of the reason for the success of the pro-
gram is the dedicated work of the faculty representatives
from Iowa State and Georgia Tech that accompany the
Overseas study has been an area in which we in engi-
neering education have lagged behind our colleagues in
the liberal arts. Part of the reason has been the necessity
of meshing requirements from two highly structured
curricula. Despite these difficulties, the remarkable benefits
to the students involved make the effort worthwhile. More
programs of this type should surely be offered.
Sincerely yours,
John C. Angus
Case Western Reserve University




University of Waterloo
Waterloo, Ontario, Canada

University de Sherbrooke
Sherbrooke, Qu6bec, Canada

USING AS A REACTION system the hydrolysis of
acetic anhydride, Anderson [1] developed a
laboratory demonstration of tubular reactor be-
haviour. In this article, another such demonstra-
tion is presented with two novel additions: (i) a
color change is added as a visual reinforcement
of the measured results and (ii) the temperature
constraint is removed in order to provide an ex-
periment operable at room temperature. The

Robert R. (Bob) Hudgins, born in Toronto, Canada, obtained all
his degrees in Chemical Engineering; B.A.Sc. and M.A.Sc. at the Uni-
versity of Toronto and the Ph.D. degree at Princeton University. He
came to the University of Waterloo in 1964 where he has remained
except for study leaves at Polysar, Sarnia, Ontario, Universite de
Sherbrooke, Sherbrooke, Quebec, and most recently (1979) at the
Swiss Federal Institute of Technology (EPF), Lausanne. His research
studies have focused upon the influence of inert diluent gases in
heterogeneous catalytic reactions, and on the behavior of chemical
reactors under forced cycling. (L)
Bertrand Cayrol received his Ph.D. from McGill University, Canada,
in 1972. He has been a Research Visitor at Chalmers University of
Technology, Sweden, and a Research Assistant at the Universite
d'Orsay, France. Presently he is Assistant Professor at the Universite
de Sherbrooke, Canada, with special interest and work on the visco-
elastic properties of polymers synthesized by micro-organisms. (R)

latter innovation reduces the complexity and thus
the cost of the apparatus.
Corsaro [2] described the hydrolysis of crystal
violet dye by sodium hydroxide. Reaction is first
order in the concentrations of each of the reacting
species. If the base is in great excess, the kinetics
of reaction become pseudo-first order,

-rdye = k' [dye]
where k' = k[NaOH]. The dye concentrations
needed for this experiment are of the order of
10-5 mol/L. Thus, a 0.01 mol/L solution of NaOH
is in 1000-fold excess and in a concentration
readily achieved at low cost. Also, NaOH con-
centrations can be varied considerably in order to
achieve a desired pseudo-first order rate constant
k' at room temperature. This feature of the experi-
ment eliminates the expensive temperature control
equipment of the sort used by Anderson [1].
Another advantage of this system over that
using acetic anhydride hydrolysis is that the
system remains isothermal during reaction. The
fact that the reagents are also very dilute means
that there is little safety hazard if the reagents
are spilled, though safety glasses are recom-
The results of this experiment may be ap-
proximated by idealized models, such as the plug
flow tubular reactor (PFTR) and laminar flow
tubular reactor (LFTR).

The PFTR model is widely used, (see Leven-
spiel [3] for example) and is related to the inlet
and outlet dye concentrations as follows.
V 1 dye]i __ 1
S Vo k' dye]. k x)
The LFTR model has been developed by Cle-
land and Wilhelm [4]. The radial mean conversion
at the exit from the tube is given by the expres-

Copyright ChE Division, ASEE. 1981



i-= l- 2 dy
= 1-v2Ei(v) +e-' (v-1)
where v = k'ro, y = t/To
and ro = L7 r2/(2vo)
Ei is the exponential integral, available in mathe-
matical handbooks.

The principal components required for this ex-
periment are listed below; code letters refer to
the schematic in Figure 1.
Tanks: (T1) 200-L polyethylene tank bottle with 3/4-in
spigot for draw-off. (T2) 10-L Nalgene aspirator
bottle with spigot.
Pumps: (P1) Century type SPS (1/4 HP); available from
Flotec Inc., Norwalk, California. (P2) Magnetic-
drive variable speed micropump; Cole-Parmer Cat.
No. 7004-92.
Mixer: (M) Graphite impeller pump (whose impellers have
been removed for use as a mixer); Eastern Pumps,
LFE Fluids Control Div., Hamden, Conn.
Rotameters: (R1) Brooks type R-6-15-A. (R2) Brooks
type R-2-15-B.
Reactor: 40 m of 3/8-in I.D. Tygon tubing, wound on a
spool 28 cm in diameter and 55 cm in length.
Valves: (V) any adjustable valve to adjust flow in the
recycle lines.
Spectrometer: Spectronic 20 (Bausch & Lomb) fitted with
a "Flow-Thru Accessory" for rapid sampling.


7 | RI i R2

.I Q.0.9JQ L 0 I
FIGURE 1: Schematic Diagram of Apparatus. T1, T2 -
tanks for NaOH solution, dye solution re-
spectively; P1, P2 pumps; R1, R2 -
rotameters; V adjustable valves; M -
mixer; Sl, S2 sample points.

A 200-L reservoir (T1) is used for the caustic
solution and a 10-L reservoir (T2) for the crystal
violet dye solution. These are metered into a mixer
(M) at the entrance to the tubular reactor.
Samples may be withdrawn from the entrance
(S1) to the reactor and from the exit (S2) from
it at tee-junctions. These tees are fashioned by
welding short lengths of 1/8 in O.D. stainless
tubing to 3/8 in O.D. stainless tubes used as
connectors at the entrance to and exit from the
reactor tube.
For convenience, the reactor tube itself takes
the form of a helical coil, although there is no
reason in principle why it could not be straight
or folded (a comparison between the conversions
using helical coils vs. straight tubes might make
an interesting variation of this experiment, to in-
vestigate the importance of secondary flow in the
helical reactor). During operation, a strong change
in color may be observed along the reactor tube
between inlet and outlet.
The analysis of crystal dye at both the inlet to
and outlet from the reactor is readily done with
a spectrometer, and the method is given by
Corsaro [2]. The Flow-Thru Accessory of the
Spectronic 20 consists of a vacuum system to draw
small samples through a fine tube into cuvette,
and then to empty into a waste jar. Since the
sampling time is brief, and reproducible, errors in
measuring the concentration of the reacting
sample are kept to a minimum. The concentration
is readily obtained by calibration with known
concentrations of the dye. Occasional sampling of
methanol or ethanol will prevent the build-up of
adsorbed crystal violet dye on the walls of the


T HE MAXIMUM FLOWRATES of crystal violet dye
solution and caustic soda solution are 50 mL/
min and 1000 mL/min respectively. Thus, in the
reactor volume, the holding time is just a few
minutes, and since plug flow is approximated, little
more than a single filling of the reactor is needed
to obtain steady state conditions after a change is
made in the input flows and/or concentrations.
For purposes of illustration of the behavior of
this reactor, several points may be obtained at
different space velocities. This is done by deciding
on the range of flowrates available and then keep-
ing the two feed pumps in the same ratio. Con-
centrations are measured at a sample point just at


1.0-- -- -----------------



07 6 0
3.0 40 50 60
r (min )
FIGURE 2: Comparison of Experimental Conversions
with those Predicted from PFTR and LFTR

the entry to the reactor, and just at the exit. Some
student data are presented in Table 1, for the
experiment at Universit6 de Sherbrooke. Note that
the volume of the reactor was 2780 mL, and the
concentration of NaOH (from the titration tests)
was 0.04 mol/L. In addition, batch tests of dye
with caustic soda solution of this concentration
provided a pseudo-first order rate constant k' =
0.54 min-1.
Using this information in the reactor models
provided, the model conversions of Table 1 were
calculated and the graphs in Figure 2 prepared
for both PFTR and LFTR. The conversion data
from Table 1 are also plotted in Figure 2. It ap-
pears that the behavior of the reactor lies between
the LFTR and the PFTR. No particular signifi-
cance is attached to the fact that the conversion
curve crosses the LFTR curve at short holding
times; the conversion data of most student groups
appear to lie between the two theoretical curves.
A wider range of flows might be used in order to

discern trends in the behavior of the real reactor
with holding time.
Finally, we offer a comment or two on student
reaction to the experiment. The change in color of
the reactants across the reactor is quite a novelty
in a chemical engineering laboratory experiment.
This visual effect helps to reinforce what is learned
from both the measurements and the theory. The
measurements are readily made, and the theoreti-
cal models used to bracket the expected per-
formance of the reactor are inherently interesting.
This happy combination of factors has meant that
the experiment has always been well received
by students working on it. El

1. J. B. Anderson, "A Chemical Reactor Laboratory for
Undergraduate Instruction," Princeton University,
2. G. Corsaro, Chem. Educ., 41, 48 (1964).
3. 0. Levenspiel, "Chemical Reaction Engineering," 2nd.
ed., Wiley New York, 1972.
4. F. A. Cleland, and R. H. Wilhelm, AIChE J 2, 489
5. H. Kramers and K. R. Westerterp, "Elements of
Chemical Reactor Design and Operation," Nether-
lands University Press, 1963, p. 93.

e,i = subscript symbols for exit, inlet
k = rate constant L/molmin
k' = k[NaOH]; pseudo-first order rate
constant (min-1)
L = length of reactor tube (m)
-r = reaction rate mol/L.min
t = time (min)
Vo = volumetric flow rate (L/min)
V = volume (m3)
X = conversion
y = t/7o
v = k'ro
S = V/vo (min-1)
To = Lr / (2vo) (min-1)

Experimental Results

Flow of Flow of Dye Concentration Holding Fractional Conversion
dye NaOH Inlet Outlet Time PFTR LFTR
Run No. (mL/min) (mL/min) (mol/L x 105) (min) Exp'tl Model Model

1 15 450 1.1 0.08 5.98 0.93 0.96 0.90
2 20 600 1.05 0.15 4.48 0.86 0.91 0.84
3 25 750 1.05 0.22 3.59 0.79 0.86 0.77
4 30 900 1.05 0.30 2.99 0.71 0.80 0.72






One of the biggest engineering challenges facing the world today is
the future of transportation. And at Michelin, we're doing our best to
contribute some answers.
We've been anticipating the transportation needs of an ever-
changing world since 1889, when Michelin invented the detach-
able bicycle tire, and then again in 1948 when we introduced the
radial tire.
Now, as the world's leader in radial tire technology, Michelin has
openings for chemists and chemical engineers. Dedicated to main-
taining its leadership role through innovation, Michelin is expanding
its United States manufacturing and research capabilities to meet the
fast growing demand for gas-saving radial tires.
To take advantage of these opportunities with Michelin, a
background in organic chemistry, polymers and textile fibers is very
desirable. Michelin relies heavily on research and development to
keep the Michelin radial tire the best in the world. Facilities are new
and fully equipped. Living conditions are ideal. Find out how you can
be a part of this professional challenge by sending your resume, with
salary history, to Michelin Tire Corporation, Manufacturing Division,
Department SPM-JH-91, Box 2846, Greenville, SC 29602, or to
Michelin Americas Research & Development Corporation, Dept. SP,
Box 1987, Greenville, SC 29602.

An equal opportunity employer M/F






Abo Akademi
SF-20500 Abo Finland

T HE FIELD OF CHEMICAL process control was
heavily criticized in 1973. Foss [1] wrote a
much cited critique of the theory research. Athans'
critique[2] pointed in a different direction when
he stated: "It is indeed unfortunate that the
process control engineers have disassociated them-
selves from the mainstream of modern control
theory. . In this manner, the tremendous ad-
vances during the past three years in system
identification, adaptive control, sensor location,
and accuracy tradeoffs have probably gone un-
noticed by the very profession that stands the
benefit most."
Athans was aware of the difficulties: "I do
believe that far greater care and finesse is re-
quired to apply modern control theory to process
control problems than missile autopilotes," but
his opinion about what action should be taken is
clear: "Everybody knows that chemical process
control systems are inherently less understood in
their dynamic behaviour than Newtonian systems.
But it seems to me that this should be viewed as
a challenge, rather than an outcry that the theory
is no good."
It is easy to criticize. This is especially true
of process control research, which is in the

In design of control systems
there are two basic approaches. The
first, in this paper referred to as the forward
approach ... The second here called
the backward approach, starts
from the process operator.

*This is a condensed version of Prof. Waller's original
paper published in the Finnish journal Kemia-Kemi, Vol, 7
(1980), page 85.

Kurt V. Waller was born in Mariehamn, Finland, in 1940. After
studying in Finland and West Germany he got the degrees M.Sc., Tech.
Lic. and Tech.Dr from Abo Akademi, Finland, in 1966, 1970, and
1972. A member of the faculty of Chemical Engineering at Abo
Akademi since 1970, Dr. Waller was appointed full Professor in 1975.
Professor Waller has written about 50 scientific papers, mainly on
process dynamics, optimization, and control.

difficult position that it [3] "must compete for
funds with such endeavors as process de-bottle-
necking, process improvements, new plant con-
struction, and other undertakings. Returns from
these latter efforts are usually more easily proven
or at least more readily believed."
Still, has the critique from 1973 been seriously
considered by the chemical engineering profes-
sion? Is the field of chemical process control in
USA today a vital field, or does it lack enthusiasm
and perspectives?

A good picture of how established process
control has become in chemical engineering educa-
tion in the United States and Canada is obtained
from Seborg's recent survey [4]. It shows that the
basics of process control is now taught at practi-

0 Copyright ChE Division. ASEE, 1981



cally all chemical engineering departments. The
main part of the teaching is, however, concentrat-
ing on very basic facts only and using surprisingly
old textbooks, the clearly dominating one (by
Coughanowr and Koppel [5]) written over 15 years
The vast majority of departments have only
one faculty member in process control. There are
some departments, however, which put more
emphasis on control and which can offer quite
advanced programs. One example of a two-person
process control program was recently described
in [6].

Research and Development
In design of control systems there are two basic
approaches. The first, in this paper referred to as
the forward approach, is to take a theory with
potential for the specific process in question and
apply it after a possible adaptation to the class
of problems at hand.
The second, here called the backward approach,
starts from the process operator. In this approach
the goal is more or less to imitate a skillful and
experienced operator by the automatic control
system. In many respects the automatic system
can perform better than the manual, since it is
not influenced by the many sources for human
errors, it won't get tired, etc. It is natural to im-
prove the system continuously by extending it to
include more and-more new and abnormal situa-
tions, making the reasons for the operator to
switch to manual less frequent.
It seems natural that a sound design includes
elements from both the forward and the backward
approach. The impression obtained is that applica-
tions in the U.S. industry have a stronger emphasis
on the backward approach than is the case in
Finland (and Scandinavia).
One explanation for such a difference is easily
found. The connections between the universities
and industry seem to be much closer in Finland
than in U.S.A., as also indicated by Foss and
Denn [7] in summing up the 1976 Asilomar
Conference on Process Control: "The ties between
industry and the university appear closer in
Europe than in North America, and the closeness
seems to have afforded European university re-
searchers opportunities to implement and test their
control methods on industrially significant pro-
cesses." There are, however, a few university
groups in the U.S. sponsored by and working in

The connections between
the universities and industry seem
to be much closer in Finland than the U.S.A.

close cooperation with industry. One is the group
at the Case Western Reserve University directed
by I. Lefkowitz.
If there is a gap between industry and the
university on the one hand, there also seems to be
quite a gap between the control researchers in
U.S. electrical and chemical engineering depart-
ments. The electrical engineers are often very
theoretical while their chemical colleagues are
quite oriented towards applications, often even to
that extent that they rather work on specific
problems than on concepts. Thus there seems to
be a certain vacuum in the study of the applic-
ability of the more advanced theory, seemingly
the most natural process control field for activity
in chemical engineering departments. However,
there seems to exist a strong controversy concern-
ing this latter statement.
There are some workers in the field though.
At the University of California, to provide just
one example, Foss at the Berkeley campus has
for many years studied a packed bed chemical
reactor, the studies including state space modeling
by orthogonal collocation [8] and recently [9] [10]
the applicability of LQG-theory to the reactor,
which exhibits inverse response. At the Santa
Barbara campus Seborg and Mellichamp are work-
ing both on reactor and distillation control, two
of the applications in chemical process control
which seem to have a large potential for modern
multivariable estimation and control theory.
There is also quite an interest in exploring the
potential of the frequency response methods for
multivariable systems developed in Great Britain
by Rosenbrock, MacFarlane and coworkers.
Indeed, such an interest was already expressed by
Fisher [11] in 1973. The largest advantage is
generally considered to be the insight into and feel-
ing for the system obtained by the methods. Recent
applications treat Foss' reactor [12] and Wood's
distillation column [13] in Alberta, Canada.
Among new approaches being studied can be
mentioned "Inferential Control," studied by
Brosilow and coworkers [14]. The goal is to infer
unmeasurable product qualities from secondary
measurements, and here e.g. standard least squares
estimates are used. The novelty is how the second-
ary measurements are selected so as to minimize


the number of such measurements required to
obtain an accurate estimate which is insensitive
to modeling errors. A compensator, which is a
process model, compensates for the effect of the
control effort on the secondary measurements. This
leads to the appealing feature of the approach
that the control effort is not fed back in the
system, only the disturbances and the model mis-
matching. This fact is said to give the system
excellent stability properties. One interesting task
would be to exploit the approach systematically
for the model mismatching problem, which is of
such a central importance in process control. A
first step in this direction is taken in [15].
At present there does not seem to be much
active interest among chemical engineers to try
to apply and explore the self-tuning regulator
(STR) of Astrom and coworkers [16] [17] or
Zadeh's fuzzy control concepts [18] [19]. The

Another use of the interaction analysis is
for design of decoupling schemes.

common attitude is to wait and see until convincing
application studies are presented in which these
approaches have turned out to result in something
that more familiar and commonly accepted
methods have not. Previous studies in North
America of the STR and of Fuzzy Control have
mainly been done on the north side of the Canadian
One of the subjects of great research activity
in chemical process control in the U.S. today is
interaction analysis, in which coupling between
inputs and outputs in multi-input multi-output
systems is studied. Interestingly enough the topic
is studied by consults and industrial researchers
as well as by university people.
The analysis starts in many cases from
Bristol's Relative Gain Array [20] [21] or similar
interaction indices suggested by Rijnsdorp [22] or
Nisenfelt and Schultz [23]. One of the main uses
for the technique is to pair variables, resulting in
(interacting) single-input single-output control
systems. It should be emphasized that one is not
restricted to basic manipulators but can use
various combinations of manipulative variables.
Interesting examples are given by McAvoy [24]
in exploring two-composition control in distilla-
tion. Conventional control, in which reflux and
boilup are manipulated, is compared to Shinskey's
material balance control (with either reflux and

bottoms flow rate or boilup and distillate flow rate
as manipulators), and also to Rijnsdorp's sug-
gestion to use the ratio reflux to boilup as one
control variable in addition to boilup. The least
amount of interaction is found for a hybrid
between Shinskey's and Rijnsdorp's suggestions.
While the interaction analysis usually uses
only steady-state data, extensions to include
dynamics in the interaction analysis in a rational
way have been treated in several papers. Witcher
and McAvoy [25] as well as Tung and Edgar [26]
show by including dynamics in the analysis that
it can lead to wrong pairing to take only steady-
state properties of the process into account.
While Witcher and McAvoy, Tung and Edgar,
and recently Gagnepain and Seborg [27] work in
the time domain, Kominek and Smith [28] work
with polar plots in the frequency domain. The
screening of a dual composition control scheme
from 12 potential candidates for an ethylene
column is used by Kominek and Smith to illustrate
the use of the theory. Use of only steady-state
interaction index is shown to result in systems un-
suitable for implementation.
Another use of the interaction analysis is for
design of decoupling schemes. The discussion is
largely concerned with decoupling in two-composi-
tion control in distillation. At present the dis-
cussion is highly concerned with the question of
whether one-way (also called partial) decoupling,
advocated e.g. by Shinskey [29], or two-way (com-
plete) decoupling, which is the approach pre-
viously investigated, is to be preferred. Strongly
related to this question in distillation is the
question of degeneracy in decoupling, recently
studied by McAvoy and coworkers [30] [31].
Additional papers on the topics of interaction
and decoupling analysis are [32] to [34].
Some further research objects of present and/
or planned activity are mentioned below in the
discussion of trends.

Since no major changes concerning process
control can be foreseen in chemical engineering
departments in U.S.A. in the near future, most
departments will continue to have one faculty
member who teaches the basics of process control
and systems engineering. Advanced teaching pro-
grams in process control will probably be given
only in the very few departments (the order of 5


Since no major changes concerning process control can be foreseen in ChE
departments in the U.S.A. in the near future, most departments will continue to have one
faculty member who teaches the basics of process control and systems engineering.

schools in the whole U.S.A.) where there is
more than one faculty member in the control field.
This scarcity of more advanced programs in
process control seems to be very unfortunate, be-
cause the correct interpretation and intelligent ap-
plication of modern control concepts is a task that
takes quite a lot of understanding of and famili-
arity with the theory as well as it demands engi-
neering skill. There are several examples where
the application has failed solely because of unsuit-
able interpretation of the theory. A perspicious
example of the importance of a correct interpreta-
tion of the theory is given in [35] and [36] concern-
ing the way of treating the integrals in optimal
PI-control of systems with time delay.
What will be taught? There probably won't
be much enthusiasm in the teaching of only the
basic process control courses, so for that matter
the text by Coughanowr and Koppel could be used
for another 15 years. It seems likely, however,
that it will be replaced by a text which has a much
stronger emphasis on computers and digital
systems. The text may very well start the treat-
ment directly in discrete time and skip much of
the old material used for continuous time treat-
ments. The same goes for the future teaching of
process control, which is likely to put less
emphasis on equations and more on algorithms
than has been done in the past.
What will be taught in the more advanced
courses will usually reflect the research.

Research and Development
Largely the industry is likely to continue to
go its own way. Then the "backward approach,"
i.e. the one in which the operator is translated into
the computer system, may be one of the main
approaches for design of industrial control
Undoubtedly the approach has its merits. One
rationale for it is that things are almost never
"normal" in process operation: "the normal state
is the abnormal one." Therefore one should start
from the process at hand, (and why not) as viewed
by the experienced operator.
Important university research in this area
will be concerned with plant diagnosis, fault detec-

tion, control under failure, and system reliability.
Also, the emphasis will shift towards treating
more difficult processes and to be more "devilish"
in testing the control systems. The important
subject of man-machine communication, however,
is probably best studied in industrial environment.
There is, however, also in industry an aware-
ness of the serious drawbacks inherent in the back-
ward approach. It conserves bad habits of the
operators and retards the progress. A catalytic
reformer is a good example: The operator tends
to recycle much hydrogen because the operation is
then very stable and safe. However, less hydrogen
recycled generally means less energy consumed and
more economic operation. Another example [37]:
In distillation it is common practice for operators
to increase the reflux rate above the design value.
Then disturbances in the feed composition seldom
cause the top product to fall below specifications,
but this advantage is balanced by the excess steam
supply needed to overreflux the column.
Today we are faced with a lot of double con-
servatism in the process industries. Firstly, pro-
cesses and plants are designed to be conservative
meaning a lot of capacity for decoupling and dis-
turbance attenuation. And secondly, the processes
are operated in such a conservative way so that
most of the disturbances left by the design dis-
appear into the conservative operation. And this
usually means a waste in energy and equipment.
However, it seems quite clear, that industry
has to cut down the waste of resources by means
of more sophisticated technology. Processes will
have to be designed as well as operated in a sig-
nificantly more integrated way, closer to con-
straints, than today. It probably won't be possible,
and certainly not economic, to operate such plants
manually. Or as seen by Evans [38]: ". . new
processes are bigger, more integrated and more
highly automated. Operators are at the same time
becoming further divorced from the process, re-
moved from the operation of the process, so when
it comes time to start up or shutdown, or
emergency situations occur, the operator does not
have the intimate familiarity with the process
which his counterpart in the less automated
facility once had."
Closer ties between industry and universities


would, of course, decrease the gap. The industry
would get competent help to apply more advanced
control concepts and the university people could
more easily ask the right questions and attack the
relevant problems.
The gap can be illustrated by some views on
process modeling [39].
Much of the university research in modeling
for process control has been concerned with the
small perturbation, linear approach. The tendency
has been to include more and more smaller and
smaller effects into the model. An industrial re-
sponse is: These complicated models can never
take everything into account anyway, the operator
does not understand them since they are so compli-
cated, so they are not used.
Or to put it another way: University people
usually suppose a "normal" state for the process.
An industrial conception of what is "normal" was
expressed as: Things are never normal-or per-
haps 2% of the time.
Models will have to become simpler and more
robust. (Signs of steps in this direction are already
visible [40]). In many cases a simple model struc-
ture only will be determined from first principles.
In operation there will be a strong shift towards
experimental modeling, i.e. towards on-line
identification and estimation, with a subsequent
reduction in importance of modeling from first
It seems likely that the concept Fuzzy Control
will find increased use in the backward approach,
since the most important property of a system
designed by the backward approach is that it is
kept on automatic and not switched over to
manual. System performance comes only second
in importance. The natural language flavor of
fuzzy control seems ideal to prevent the operator
from feeling unfamiliar with the concepts used
in the computer system. Another rationale for
the use of fuzzy logic is that much of the control
work going on in industry is done, and will
continue to be done, almost without any control
theory in the sense of "hard" systems theory. So
far, however, use of fuzzy logic has been reported
mainly from Europe [41] [42].
For a long time it has been recognized that
there is a large potential in an integration between
process and plant design and control. Indeed [43],
"it is important to consider process control in
general-the development and the application of
its theory-as an integrated part of the plant
design and the process operation." There is quite

a lot of interest in the subject among chemical
engineers in the U.S.A. today, and increased re-
search activity can be expected. In this integration
the potential of the chemical engineer can be
utilized at its very best.
The increased integration of processes and
complexity of plants, will make the control aspects
more critical. There will be a greater need for,
and emphasis on, multivariable control methods.
Analysis of plants will to an increased extent rely
on efficient simulation methods for large systems.
More work in these areas is needed and is likely
to be done [44].
What about adaptive control methods? Will, for
example, the self-tuner see an increased number of
applications? The answer is most likely yes. One
of the reasons was formulated, somewhat
cynically, by an industrial control system design-
er: These methods will be used more because
university people like to play around with them.
Starting from the single process unit rather
than from the concept point of view, no significant
changes are visible. Thus the large interest in
distillation control of the last decades continues.
Today research in distillation control in U.S.A.
is concerned with such questions as pressure
control, coupling and decoupling in two-composi-
tion control, multivariable control in general, and
nonlinear control of high-purity columns.
The other "large" process will probably
continue to be the chemical reactor, which offers
a wealth of challenging design, optimization,
identification, estimation, and control problems.
It is also felt by several workers, both in uni-
versities and industry, that these two processes,
the chemical reactor and the distillation process,
have the greatest potential for use of modern
control theory.

H ow HAS THE CHEMICAL process control com-
munity responded to the critique expressed in
Athans' opinion 5 years later is not less critical,
when he comments on the-in his opinion sad-
state of applications of advanced control in the
process industries. He states [45] that the chemi-
cal engineers are so conservative that they have
only got what they deserve. Shinnar [46] (1977)
laconically states: "the state of process control is
rather sad."
Continued on page 51.




A world of energy...
Conoco's development of Petroleum,
Chemicals, Coal, Uranium and Minerals
Shas taken us around the globe. This year we
are spending two billion dollars on energy. Our chal-
lenges are great and our need for talented, innovative
individuals is growing.
We have many interesting and rewarding positions for
chemical engineers throughout our operations. If you would
like more information, contact your campus placement office
Or write: Dr. M.L. Sharrah
Sr. Vice President
Research & Engineering
Conoco Inc.
High Ridge Park
Stamford, Connecticut 06904

An equal opportunity employer, m/f.





Carnegie-Mellon University
Pittsburgh, PA 15217
ON NOVEMBER 15th and 16th, 1979, Carnegie-
Mellon University held its first annual Chemi-
cal Engineering Symposium. It was a chance for
graduate students in the department to present
their work before an audience of interested pro-
fessors and students. The symposium was intended
to serve several purposes. It would give students
experience in speaking before an audience. This
experience would be helpful in the future, making
it easier to speak at technical meetings, and, for
those who desire to teach, in the classroom. The
symposium was also a good way for research
groups in the department and in other depart-
ments in the university to examine each others'
work. Here was a chance for the left hand to at
last find out what the right hand is doing. It was
hoped that the symposium would lead to co-opera-
tion and an exchange of ideas between research
groups. Also, the symposium was a good time for
new graduate students to check out the work
that the professors in the department are doing.
This exposure helped the new students to decide

Mike Locke received his B.S.Ch.E. from the University of Massa-
chusetts in 1976, his M.S.Ch.E. from Carnegie-Mellon in 1978 and is
presently working on his Ph.D. at Carnegie-Mellon. His areas of interest
include Optimization, Computer Aided Design, and Equation Solving.
He presently serves as a consultant to NTP Corporation of Pittsburgh.

Westerman-Clark presenting his winning paper . .

which faculty members they would choose as their
The idea for a symposium was first brought
up in September during a meeting between depart-
ment chairman Tomlinson Fort and the officers of
Carnegie-Mellon's Chemical Engineering Graduate
Student Association (ChEGSA). Arrangements
for the symposium began with a call for papers.
Prospective speakers were required to turn in the
titles of their talks. It was at this point that the
organizers realized that they were on to something.
The response from students and faculty was over-
whelming. Twenty-seven students responded
positively, with 29 papers to present. This was
twice the number that had been anticipated.
One month before the symposium, speakers
submitted abstracts of their talks. The abstracts
were collected in a booklet, along with the agenda
of talks. Copies of the booklet were distributed to
all faculty members and graduate students in CMU
Chemical Engineering, and also to the heads of the
departments in the School of Engineering. Each
student was given 20 minutes to talk, with 5 addi-
tional minutes for questions. Topics ranged from
separation of polymer wastes to synthesis of heat
exchange networks. Other interesting topics were
Copyright ChE Division, ASEE, 1981


trace metal levels in Pittsburgh air (there's more
than just a trace), and the velocity of blood flow
in a rabbit's ear. Interesting discussions started
during the symposium, and are continuing.
As an incentive for the students, a prize was
awarded for the best paper. It consisted of an
expense paid trip to the next AIChE convention, to
deliver the winning paper, and a check for $100
from Tomlinson Fort. The judging committee was
made up of 5 faculty members, headed by Dr.
Robert Rothfus, the senior member of the depart-
ment. Speakers were judged on content, presenta-
tion, and a written copy of the talk.
After much deliberation, the judging commit-
tee chose Jerry Westermann-Clark as the prize
winner. The title of his talk was "Coion Exclusion
Potential in Charged Membranes." Jerry entered

S. .and receiving his award and the congratulations
of Professor Fort.

CMU in September, 1976, with a B.S. from the
University of Pennsylvania. He has a Masters
Degree from Carnegie-Mellon and his Ph.D. in
September, 1980. His work was supervised
by Dr. John Anderson. Jerry is now an Assistant
Professor at the University of Florida.
The judges also named three runners up. They
were Greg Townsend, for his talk "Pharmaco-
kinetics of Adriamycin in Normal and Neoplastic
Tissues" with Dr. Rakesh Jain supervising; Lewis
Grimes, with a talk entitled "The Synthesis and
Evaluation of Networks of Heat Exchangers that
Feature the Minimum Number of Units," super-
vised by Dr. Arthur Westerberg; and Michael
Reilly for his talk "Ambient Trace Metal Levels
in Pittsburgh Air," supervised by Dr. Eric Suu-
berg. O

Use CEE's reasonable rates to advertise. Minimum rate
% page $50; each additional column inch $20.

Chemical Engineering. Two Assistant or Associate Pro-
fessor Positions. These are tenure-track positions and will
be approximately half-time teaching and half-time re-
search. We will help successful candidates establish re-
search by providing initiation funds, co-investigation op-
portunities with senior faculty, and proposal preparation-
processing assistance from our Office of Engineering Re-
search. Candidates must possess an earned Ph.D. degree
from an accredited Department or School of Chemical
Engineering or have a Ph.D. degree in related areas and
have strongly related qualifications. We welcome applica-
tions from candidates with competencies and interests in
any field of chemical engineering, but especially seek those
with strengths in material sciences or controls. The posi-
tions are available as early as July, 1981. Applications
will be received thru June, 1981. Salary and rank are
commensurate with qualifications and experience. Please
send your resume and list of three references to: Dr.
Billy L. Crynes, Head, School of Chemical Engineering,
423 Engineering North, Oklahoma State University, Still-
water, Oklahoma 74078, 405-624-5280. (Calls for addi-
tional information invited.) OSU is an equal opportunity/
affirmative action employer.


OF THE '80's

on microfilm


For complete details and specifications,
write or call collect:
Research Publications, Inc.
12 Lunar Drive
Woodbridge, Connecticut 06525
(203) 397-2600


problems for teachers

Pennsylvania State University
University Park, PA 16802

We see that the wind velocity increases as a
logarithmic function of the height above the
earth's surface. Because of the higher velocity at
2.5 m than at 0.5 m there will be a lower pressure
at the top of the tube than at the bottom. This
pressure difference will induce an upward flow in
the tube.
A Reynolds Number can be calculated for flow
through the tube:

Re = Dp
0.01m(lms-1) (1.2 kg m-3)
1.8 (10-5 kg m-1 s-1)
which indicates laminar flow.
Thus the Hagen-Poiseuille equation can be used
to find the pressure drop. Eq. 2.3-19 Bird, et. al.
(or the Prairie Dog problem) gives

r Ap R4

or Ap Q/L

Q = A = 7rR2
= 1ms-1 (r) (0.0052 m2) = 7.85 (10-5 m3 s-1)
= 1.8 (10-5 kg m-1 s-1)
L = 2m
R = 0.005 m
Substituting, we obtain
Ap = 11.52 kg m- s-1
This pressure drop can be related to horizontal
wind velocity by the Bernoulli equation:
1 V3> ^A 1 A +
1 + b) + I-dp + W + E = 0
Neglecting friction and work, E, and W are zero.
If the air can be assumed to be incompressible
under these conditions, p is constant and

*The problem statement was presented in CEE Vol.
14, No. 4 (Fall 1980).

z = 3m
z= 2 m

z= 1m
z = oM

U 0 ms
d = 0.01 m
z = 2.5 m

= 1 ms-

z= 0.5 m
7 7 / 77 / / / / 7/ / 7 /7


fdp P2 P1
jP P

The potential energy difference A 4 between
points is negligible and for turbulent flow of air,
/ a 2, and
S 2 (p, p2)
2- 2 =
Taking point (1) at the bottom and point (2)
at the top, both terms are positive. Thus we have
one equation and two unknowns. The logarithmic
velocity profile
U(z) 1 z U. z
U(z) = I In or U(z) = 0.4n
U. k zo 0.4 01.04
provides two more equations but only one more
unknown, U.. Since U in the log velocity profile
and in the Bernoulli equation are the same
thing (i.e. the horizontal wind velocity) we can
combine these equations.

<> n [.5 ]= 10.34 U.
<.4 0.0 4

_U In f 0.5 = 6.31 U.
S-0.4 [L 0.04 J

(10.34 U.)2 (6.31U.)2 = 2(pl -P2)
2 (11.52 kg m-' s-2)
67.10 U," -
1.2 kg m-3
= 19.20 m2 -2

U. / 19.20 = 0.535 ms-1
S= 67.10
This value of U. can now be used in the velocity
profile to get velocity at 3 m.
U. z 0.535 3
U (z) =- In = In
k 0.04 0.4 0.04
= 5.77 ms-1


Create, construct and control
your career with MILLIKEN

We provide an expanding range of opportunities
to accelerate your managerial and professional
career development.
Milliken's commitment to technological leader-
ship in providing new products, processes and
equipment is exemplified by the Millitron ma-
chine, a computer controlled, jet-injection dyeing
system created and developed by Milliken engi-
neers and scientists for the production of multi-
colored carpet, rugs and upholstery products.
The men and women we need have open
minds, open enough to let us expose you to the
details of the Milliken challenge. When we select
each other, horizons broaden for both of us.
Call or write:
A/// h. College Relations Dept.
S Post Office Box 1926
MI LLIKEN Spartanburg, SC 29304
Phone (803) 573-2508

An equal opportunity employer.



University of Alberta
Alberta, Edmonton, Canada

Sherritt Gordon Mines
Fort Saskatchewan, Alberta, Canada

which is almost completely ignored in a
chemical engineer's university education. This
omission is unfortunate since marketing is usually
the crucial constraint on the commercial viability
of many projects. In spite of its obvious im-
portance, market analysis is not taught to under-
graduates; the undergraduate is not trained to


Capital Costs(1)

Capacity 1
Direct Fixed Cost 6
(excluding working
Debt(2) 2
Equity 4
Interest rate on debt (

Term on debt 1
Yearly debt charge(4) $
Interest rate on equity (

Term on equity 1
Yearly charge: equity
recovery, return on invest-
ment (5) 5
Total Yearly Capital
Charges on D.F.C. 7
Unit Capital Charges
on D.F.C. I

.000 T/D
'20 mm lb/yr
i6 mm $Can

2000 T/D
1440 mm lb/yr
80 mm $US

!0.5 mm $Can 25 mm $US
45.5 mm $Can 55 mm $US
Canadian prime US prime
t I/Vz% (3) + 1%
5 years 15 years
;2.2 mm $Can $3.1 mm $US
Canadian prime US prime
- 4% + 3%

5 years

15 years

.6 mm $Can 7.65 mm $US

'.8 mm $Can

.09 tCan/lb

(1)Unless stated otherwise, costs are
currency of the home country.
(2)Debt-equity ratio of 0.45.

10.75 mm $US

0.747 SUS/lb
expressed in the

(3)Canadian prime interest rate = 8%%, U.S. prime in-
terest rate = 8%.
(4)Both debt and equity charges taken as uniform payment
(5)Working capital and salvage neglected.

Copyright ChE Division, ASEE, 1981

Brett Haugrud graduated from the University of Alberta with a
B.Sc. in Chemical Engineering in 1973 and a MSc in Chemical Engi-
neering from the same school in 1978. He worked for several years
as an engineer in the Research Department of a Canadian Potash
Company before accepting his present position as process engineer
for Sherritt Gordon Mines, Fort Saskatchewan, Alberta. (L)
Jim Ryan received his Ph.D. from the University of Missouri
(Columbia) in 1966. He has taught at the University of Alberta since
that time. Primary teaching duties include the senior design project,
undergraduate fluid mechanics and introductory thermodynamics. He
serves as a consultant to the petroleum industry, mainly in the area
of pipeline design and explosions. Occasionally serves as a writer
and researcher for CBC-TV and Radio. (R)

analyze a market problem even in a rudimentary
fashion. The neglect of the marketing process by
the students' professors is understandable since
their training emphasizes the scientific aspects of
the chemical production process, such as research,
development and design. From an academic's view,
the estimation of a market size and its location is
regarded as a black art best left to someone else.
This paper presents a technique which hope-
fully transforms market estimation from a black
art into a dark gray one. Its objective is a quanti-
tative means of estimating a potential market area
at a given time for an undifferentiated chemical.
An undifferentiated product is one which does not
have any unique property that can be attributed
to its manufacturer. The manufacturer cannot
claim that his product is better than his competi-
tor's since both products are identical within
normal engineering specifications. Methanol,
fertilizer-grade ammonia and polymer-grade ethy-
lene are examples of undifferentiated products,
while almost anything that is advertised as unique


or better is a differentiated product. Consequently,
the major factor which will govern the sale of this
type of chemical is its price, not its quality.
The potential market area for an undifferenti-
ated chemical, produced by a given plant, is defined
as that region where the delivered cost of the
chemical is equal to or less than other competing
producers. The delivered cost of a chemical com-
modity to a customer is the sum of the produc-
tion costs, transportation charges, tariffs and
currency exchange. The last two expenses are ap-
plicable only if the commodity is traded inter-
nationally. Rather than analyzing the position of
all potential competitors for a market, the tech-
nique of estimating a market area can be il-
lustrated more simply by considering only two
As an example, assume that two hypothetical
methanol producers, one located in southern Al-
berta and the other on the Texas Gulf Coast, are

in competition with each other. In recognition of
historical precedence, the capacity of the Canadian
plant is taken initially as one half that of the
Gulf Coast plant. Table I gives the illustrative
capital costs and financial charges for each of the
two plants. The financial arrangements are simpli-
fied in order to make the presentation of the results
concise. Unit prices of inputs and associated unit
manufacturing costs are shown in Table II. The
cost figures in both tables are given in the
currency of the home country.
Until recently, the plant gate cost of Gulf
Coast methanol was significantly below that of
methanol produced in Alberta. In fact, at one
time the cost of methanol delivered from the Gulf
Coast to Alberta was less than the plant gate cost
of Alberta methanol. That situation was due
primarily to the economics of scale typical of the
American plants and the generally low level of
natural gas prices. As is obvious from Table II,



Direct Fixed Capital
Working Capital

Manufacturing Costs

720 mm lb/yr
1000 T/D
54 mm $Can
4.9 mm $Can

Natural Gas as Feedstock and Fuel (32 mm Btu/ton)
Water, misc.

Labour & Admin.-80 people
Maintenance & Local Taxes

$1.25/mm Btu

5% of DFC
2% of Plant
._Gate Cost

Unit debt on D.F.C.
Unit equity on DFC
Interest on W.C.


Variable Costs
Fixed Costs
Capital Costs


Unit Cost





1440 mm lb/yr
2000 T/D
80 mm $US
5.8 mm $US

$1.95/mm Btu


Unit Cost






Process Engineers

Let's Build For The Future Together.
At Bechtel, we believe in personal and professional growth. We've grown into a world
leader in the field of Engineering and Construction during the past eighty years, and
we continue to grow through leadership in new developments.
We are looking for career engineers with a commitment to quality and innovation to
work on conceptual studies and detailed process designs leading to major, first-of-a-
kind projects in the following areas:

Positions are available in SAN FRANCISCO, HOUSTON, and other locations. Take
the opportunity to build your future with a leader For immediate consideration, send
your resume or application to:
0 Rufus Williams Gall Owens
Bechtel National, Inc. Bechtel
Employment Dept. 24-6B-80 Employment Dept. 24-6B-80
P.O. Box 3965 P.O. Box 2166
San Francisco, CA 94119 Houston, TX 77001

Bechtel and People.

We Grow Together.


The potential market area for
an undifferentiated chemical, produced by
a given plant, is defined as that region where the
delivered cost of the chemical is equal to or
less than other competing producers.

the situation is now reversed. The plant gate cost
of Alberta methanol is below that of the Gulf Coast
even though there is a considerable size advantage
in favour of the American plant. The reason for
the reversal is that the Alberta plant now has a
large advantage in natural gas prices. It should
be noted these are new gas prices (as of January,
1978) and do not apply to older plants which have
long term contracts at lower natural gas prices
or to plants that can blend the old and new gas
To estimate the delivered cost at any point, the
transportation costs, tariffs and currency ex-
changes must be added to the average plant gate
production costs. The effect of these factors can
be illustrated in a simple example. Suppose the
two plants attempt to sell methanol in the San
Francisco area. The Gulf Coast producer can
transport methanol by ship through the Panama
Canal while the Canadian producer is restricted
to more expensive rail transportation. In addition
the American producer will not have to pay a tariff
while the Alberta plant must. All of these factors
favour the American plant and tend to erode the
plant gate cost advantage enjoyed by the Canadian
plant. On the other hand, the devaluation of the
Canadian dollar relative to the U.S. dollar in
effect lowers the plant gate cost of Alberta
methanol by about 10%. Table III summarizes all
of these effects and shows that the U. S. producer
is able to deliver methanol into San Francisco
more cheaply than his Canadian competitor. As a
consequence, San Francisco is regarded as a po-
tential market area for the U. S. producer.
Cost of Methanol Delivered to San Francisco

Plant Gate Cost

Rail to San Francisco
1,324 miles at 20/ton-mile
Ship to San Francisco
4000 at .5/ton-mile
Tariff-7.6 O/US gal
Delivered cost

4.05 Can/lb
(3.65) CUS/lb
1.32 OUS/lb

1.15 OUS/lb
6.12 OUS/lb


5.74 OUS/lb

The technique used to calculate the delivered
costs of the two competitors in the San Francisco
area can be generalized to any point of sale. The
results of this generalization are best presented on
a map as a series of curves of constant delivered
costs. These lines will be circles with their centers
located at each plant. When a commodity shipment
encounters an international border, the "iso-cost"
lines will have a discontinuity as a result of tariffs
and exchange rates. In general, the magnitude of
the discontinuities will not be the same for both
A further complication arises where there is
a possibility of change in mode of transportation.
For instance, to market methanol in the U.S. Mid-
west, the Gulf Coast producer might ship his
product via barge to Chicago, then distribute from
Chicago either by rail or by truck. The base cost
in Chicago would then be the Gulf Coast plant gate
cost plus the barge charges to Chicago. Local Mid-
west costs would be the Chicago cost plus rail or

FIGURE 1. Iso-Cost Lines and Market Line for Base
truck transportation charges to the market area.
Consequently, the iso-cost lines in the Midwest are
again circles, but their origin is at Chicago instead
of the plant site. In addition to Chicago, San Fran-
cisco and Montreal were chosen as distribution
The technical exercise is to construct the lines
of constant cost for each distribution center. A
set of these curves is shown in Figure 1. Each of
them is labelled with the delivered cost in the
currency of the consuming country. These circles
are idealizations since the railroads do not run in
straight lines. The actual iso-cost lines would be
severely distorted, particularly in the west. At this
stage of the analysis, this is viewed as a second
order effect and is neglected. The potential market


What does
engineering mean
total energy



It means many things. The development of better and
more efficient equipment to sink a drill bit into the ground.
The design, development and operation of giant offshore
drilling devices to find and produce natural resources
from beneath the oceans. Improved methods of coal and
uranium mining. More efficient operation of chemical
plants, gas processing plants and petroleum refineries.
In short, chemical, civil, electrical, mechanical, metal-
lurgical, mining and petroleum engineers are vital to and
an integral part of our organization.
For information concerning employment opportunities,
write to:
E. D. Marvin, Jr., Mgr.
Professional & Technical Recruiting
Kerr-McGee Corporation
P.O. Box 25861
Oklahoma City, Oklahoma 73125


An Equal Opportunity Employer M/F/H/V

FIGURE 2. Approximate Yearly Consumption and
Growth Rate of Methanol by Region.

area for each competitor is found by drawing a line
connecting the intersections of the lines of equal
constant costs. This, the market line, is also shown
in Figure 1. The market area for Alberta methanol
is north of the line and the Gulf Coast market is
to the south.
As can be seen from the figure, Alberta's
market area is all of Canada west of Quebec City
and a large portion of the northwest and west
central United States. The Gulf Coast producer
has a lower delivered cost everywhere else. Once
the geographical areas which represent potential
markets for the two competitors are established,
the objective is to locate customers within those
areas. Figure 2 shows the approximate demand for
methanol by region. The upper number is the 1977
capacity of chemical plants which use methanol as
a chemical feedstock, expressed in mm lb/yr.
Below this number is one in parentheses which is
the six year net growth in demand for feedstock
methanol. The data used to construct this figure
were taken from the Chemical Economics Hand-
book [2] and include only specific plants listed in
that publication. These values are somewhat low
because they do not take into account demand for
methanol as a solvent, gas line antifreeze, etc.
Since a large fraction of methanol is used in manu-
facturing resins for plywood and other wood
products, the markets are concentrated in those
regions which have a significant harvestable
forest. The pacific northwest and southeastern
states are examples of this concentration.
By superimposing the market line on the con-
sumption patterns, it is possible to compare the
production of the two plants with the require-
ments of customers potentially available to each.
It is clear from Table 2 that the price paid for

natural gas is the most significant component of
methanol manufacturing costs. The base case
natural gas prices were taken as the prevailing
new gas prices (1978). Actual prices could be
significantly different if old gas was under
contract. A number of cases were studied to show
the effect of gas prices on the market line. The
cases examined are designated on Figure 3 as one
through four with line 2 being the base case. When
the Alberta producer has a price advantage of
$1.00/mm Btu, the change in the market size is
fairly small. The only new market penetrated is
San Francisco where about 100 mm lb/yr are con-
sumed. Case 3 occurs when both competitors have
the same natural gas costs. The Gulf Coast pro-
ducer can now capture most of the Pacific north-
west market as well as the eastern Canadian
market. This situation would force the Alberta
producer to operate at less than full capacity,
further improving the position of the U. S. com-
petitor. Since a small change in any economic or
technical variable could cause a large change in
market size, competition in the PNW would be
intense. Price discounting by both parties would
be a definite possibility. Finally, if the Gulf Coast
operator enjoys old prices and the Alberta pro-
ducer pays new gas prices, the Texas plant would
capture all of the PNW. The Alberta producer
would be left with only the three western Cana-
dian provinces, whose total market is less than
his breakeven production rate. The problem is

FIGURE 3. Effect of Natural Gas Price on Market Area.

solved since he must then shut down if actually
operating or decide not to build a new plant.
Operating under the base case conditions, the
Gulf Coast producer has access to a market of
approximately 6000 mm lb/yr, while the Alberta
manufacturer has a maximum market of 970 mm





Weston M. Stacey, Jr.
This volume deals with the physics of magnetically-confined fusion
plasmas. Plasma physics is treated as an element in the development of
fusion power, with important technological constraints and interactions
included in the analysis. The development of the material starts from first
principles and is carried through to an engineering physics formulation
that can be applied to the analysis of fusion reactor plasmas.
approx. 384 pp. (1-08095-0) Jan. 1981 $32.95
Samuel Strelzoff
Supplies all the necessary information and data needed by qualified
engineers for choosing an optimum system and designing large-
capacity production facilities for ammonia synthesis. Designed as a
working tool that enables engineers to save time by avoiding extensive
searches for pertinent information. Divided into four sections-Ammonia
Synthesis, Syngas Generation, Carbon Dioxide Removal, and Final
Purification-each containing descriptions with data in tables and charts.
approx. 220 pp. (1-02722-7) April 1981 $22.00 (tent.)
Daniel I.C. Wang, Charles L. Cooney, Arnold L. Demain, Peter
Dunnill, Arthur E. Humphrey, & Malcolm D. Lilly
A practical, up-to-date introduction to fermentation and enzyme
technology, outlining the fundamental microbiological, biochemical,
genetic, and engineering aspects of fermentation and presenting
advanced methods of fermentation and control. Covers the isolation of
enzymes, especially those found in intracellular contents of
microorganisms. Discusses enzyme immobilization and the factors
influencing the use of enzymes in reactors.
374 pp. (1-91945-4) 1979 $28.95
William E. Biles & James J. Swain
Unifies theory and methodology of optimization statistical and
experimental methods. Discusses single and multiple objective
optimization problems, with and without constraints. Introduces the
fundamentals of statistical experimentation. Encompasses development
of regression estimation, experimental design, optimization techniques,
and a synthesis of these methods as a single methodology.
368 pp. (1-04244-7) 1980 $35.00
Aksel L. Lydersen
Provides a concentrated and unified treatment of mechanical unit
operations and heat transfer while filling the need for a short, refresher
text for practicing engineers. All theory and empirical equations
presented are related to practical applications demonstrated through
worked examples from industry and laboratory work.
357 pp. (1-99697-1) 1979 $53.95 Cloth
(1-99696-3) $22.50 Paper
Karla Longrie
Longr6e presents up-to-date information on the major reservoirs of
various contaminants, including microorganisms, causing foodborne
illnesses. Topics include the conditions leading to contamination of
ingredients and menu items while being prepared, stored, and served;
the conditions favoring multiplication of contaminating bacteria; and
appropriate measures of control.
456 pp. (1-06424-6) 1980 $27.50

Revised 2nd Ed.
The M.W. Kellogg Company
CONTENTS: Strength and Failure of Materials. Design Assumptions,
Stress Evaluation, and Design Limits. Local Components. Simplified
Method for Flexibility Analysis. Flexibility Analysis by the General
Analytical Method. Flexibility Analysis by Model Test. Approaches for
Reducing Expansion Effects: Expansion Joints. Supporting, Restraining,
and Bracing the Piping System. Vibration: Prevention and Control.
385 pp. (1-46795-2) 1964 $37.50

D.C. Stewart
A broad overview of all of the many factors involved in the safe handling
of radioactive materials on the bench scale. Administrators of nuclear-
oriented programs, architect-engineers, nuclear facility managers and
those responsible for training of personnel will find many practical
matters of interest. The more highly technical areas (dosimetry, shielding,
nuclear criticality) are surveyed and summaries presented in
straightforward, non-mathematical language.
approx. 336 pp. (1-04557-8) Jan. 1981 $35.00

James J. Duderstadt & William R. Martin
Presents a general unified theory, surveying methods used to analyze
the transport of such microscopic particles as neutrons, photons,
electrons, and molecules through matter. Covers numerous fields
including nuclear reactor physics, astrophysics, gas and plasma
dynamics, and statistical mechanics. Includes frequent references to
leading review articles.
613 pp. (1-04492-X) 1979 $39.95
H. Hugo Buchter
Demonstrates how to identify and solve all kinds of sealing problems,
consolidating data previously scattered throughout the literature.
Discusses basic principles of how seals work, what different kinds of
seals can accomplish, and what range of environmental and mechanical
specifications each seal type is suited for.
441 pp. (1-03184-4) 1979 $34.50
Environmental Movement of Chemicals in
Air, Water, and Soil
Louis J. Thibodeaux
This advanced text covers the general topic of movement within the
exchange of man-made chemicals between the three phases of the
environment-air, soil, and water. The topic of chemodynamics in the
exterior environment is organized and developed to provide students
with a unified approach.
501 pp. (1-04720-1) 1979 $36.00
Examination Review
Donald G. Newnan & Bruce E. Larock
A systematic, comprehensive review of basic engineering principles,
together with a practical guide to successful preparation for state
Engineer-in-Training (EIT) examinations. Covers all important concepts
in nine major areas: mathematics, statics, dynamics, mechanics of
materials, fluid mechanics, thermodynamics, chemistry, electricity, and
economic analysis.
503 pp. (1-01900-3) 1978 $27.50
Chemistry and Chemical Technology, 3rd Ed.,
Vols. 1, 2, & 3
James P. Casey
Here's an in-depth look at the chemistry and chemical technology
involved in the manufacture of pulp and paper, the properties of paper,
and the uses for paper. The new edition contains contributions by forty
recognized authorities in the field. Emphasizing the underlying science
and technology, this edition reviews, in detail, chemical and engineering
Vol. 1: 820 pp. (1-03175-5) 1980 $55.00
Vol. 2: 625 pp. (1-03176-3) 1980 $50.00
Vol. 3: approx. 700 pp. (1-03177-1) April 1981 $48.50 (tent.)

Books under consideration as classroom texts are available for a 60-day
free examination. Write to Jules Kazimir.
Order through your bookstore or write to Nat Bodian Dept. 7436
TO ORDER BY PHONE call toll tree 800-526-5368.
In New Jersey, call collect (201) 797-7809.
a division of John Wiley & Sons, Inc.
605 Third Avenue
New York, N.Y. 10158
In Canada: 22 Worcester Road, Rexdale, Ontario
Prices subject to change without notice, 092 1-7436

lb/yr. With the indicated growth rates, these
markets will have expanded to 7050 and 1220 mm
lbs/yr respectively during the three years taken
for planning and construction. To operate at full
capacity, the new American producer must cap-
ture 20% of the market available to him, whereas
the equivalent number for the Alberta plant is
60%. The potential market available to the Al-
berta producer at start-up is about twice the
capacity specified in the base case, i.e. 1000 T/D.
As a first step, he might consider doubling the
plant capacity to 2000 T/D in order to capture
the entire base case market area plus any other
market that becomes available as a result of the
economics of scale. Figure 4 shows the market

FIGURE 4. Effect of Plant Size and Duty Draw Back on
Market Area.
lines for the 1000 and 2000 T/D plants. All other
factors were held constant. The interesting
feature of the increase in plant size is that, in spite
of the reduced unit cost of methanol, not one more
major customer was added to the Alberta market.
The only effect of the increase was to add sufficient
capacity to satisfy the demand of the customers
that a 1000 T/D plant could not. It is obvious that
the next plant size that should be tried is about
1200 T/D. However, this is too precise a refinement
in view of the potential effects of changes in other
On occasion, a commodity is imported into a
country, processed, then exported in a different
form. When this occurs, the duty or tariff origin-
ally imposed on the imported good is forgiven.
This is called a duty drawback. In the PNW up
to one-third of imported methanol is exported in
the form of resin. In effect, the cost of the Alberta
produced methanol is reduced by an amount equal
to the duty drawback, which improves the com-
petitive position of Alberta methanol in a con-

tested area. The sensitivity of the market line to a
change in duty drawback is also shown on Figure
4 for the case of equal natural gas prices of $1.25/
mm Btu. With no duty drawback, the Alberta pro-
ducer should be able to capture virtually all of
Washington and half of the Idaho market area.
With one-third of the tariff forgiven, he will be
able to penetrate one-half of the Oregon market in
addition to the states of Idaho and Washington.
The incremental increase in geographical area due
to the change is relatively small; however, the
amount of increased sales should be significant
because of the concentration of users located in
the area. It should be noted that the line repre-
senting the one-third duty drawback is mislead-
ing. The line is for drawback based on the average,
whereas in reality the drawback is specific for only
those plants which exported the processed
So far it has been assumed that the determina-
tion of the market size was undertaken so that
both potential competitors can make the capital in-
vestment decision of whether to build a plant or
not. A more interesting situation occurs when one
of the plants is already in production, while the
decision to construct the other has yet to be made.
For the sake of argument, suppose the Texas plant
exists and that the Alberta plant has yet to be
built. The Alberta plant must use the average cost
of production which includes return on equity to
determine its market. On the other hand, the
Gulf producer makes his decisions based on the
marginal cost of production.
In the short term, the existing plant can sell
its production at a cost equal to the variable pro-
duction cost or breakeven. In the midterm, the
producer can sell at a price which includes all
manufacturing and debt repayment costs, but ex-
cludes equity capital recovery or return on equity.
Operating for an extended period under this
condition will have adverse effects on either the
value of the company's stock or the plant liquida-
tion value. Figure 5 shows the market lines for
the Texas plant selling its product under several
pricing options. The Alberta producer is assumed
to have full equity recovery and return. On this
figure, four market lines are shown. The three
most southern lines are parametric in the fraction
of original anticipated return and recapture of
equity. The line designated as 100% corresponds
to the original base case and that line marked
as 0 % is the case where only production and debt
costs are met. Finally, the market area outlined


FIGURE 5. Effect on Percent Capital Recovery for Gulf
Coast Plant on Western Market Area.
by the breakeven line is the situation where only
the variable costs of production are achieved. In
the last case there is no allowance made for equity
or debt repayment. By selling its product at the
marginal cost of production, the northernmost
line, the Gulf Coast producer is able to capture
most of the PNW market. However, this situation
can only be a short-lived phenomenon, since this
market is equal to about one-third of the Texas
plant's production. Usually, selling at the margin
is resorted to when a small fraction of the plant's
capacity is under-utilized. In the long run, the
Alberta producer does not need to worry about
this single competitor capturing a large share of
the PNW methanol market. This is certainly not
the case if a large number of U. S. producers are
operating at less than their capacity. The market
line in Eastern Canada was shown only for the
base case. The Gulf Coast producer selling into
Eastern Canada at breakeven or at the margin
would be in violation of Canadian dumping laws,
so these lines were not drawn. Dumping occurs
when a company sells internationally at a price
below its domestic price.
Thus far, the analysis has been used to find the
respective market areas for two competitors under
a variety of conditions using the concept of de-
livered cost as the sole criterion. By interpreting
the technique in a slightly different manner, the
two competitor price can be estimated. Here, a
two competitor price at a given point is defined as
the higher of the two delivered costs at that point.
The concept of this price is best illustrated by a
simple example. Suppose the two methanol pro-
ducers are attempting to sell methanol to hypo-
thetical customers in Pendleton, Washington and
Vancouver, B.C. As shown on Figure 6, the cost
of Gulf Coast produced methanol in Pendleton is

6.5 U.S. //lb, whereas the Alberta plant cost de-
livered at the same point is 5.34 U.S. 0/lb. Given
that there are no other suppliers, the Alberta
producer can charge slightly less than 6.5 U.S.
/lb and still capture the Pendleton market.
By subtracting transportation, tariff and
currency conversions from the Pendleton price,
the FOB plant gate price of methanol in Calgary
is approximately 5.2 Can. 0/lb. This is shown on
the figure as a dashed line. Since the plant gate
cost was only 4.05 Can. /lb, the sales into Pendle-
ton will yield a profit of 1.15 Can. /lb in addi-
tion to the required profit embedded in the cost
calculation. On the other hand, the American plant
can set the price in San Francisco at the Alberta
delivered cost of about 6.1 U.S. 0/lb. The basic
principle is that in a given market area, the price
is set equal to the competitor's delivered cost at
any point in that area. The situation can be more
complicated, as illustrated by an analysis of the
Vancouver market. The Gulf Coast plant could
transport methanol via ship from San Francisco
and land it in Vancouver at 6.8 Can. $/lb. This
delivered cost would fix the FOB plant gate price
available to the Alberta producer at 6.2 Can. I/lb.
\ \a

e= "..... '. \
.... ,, 5.6 |
,I 52.


I .e I 4.0
San Base Case Pendleton INTL Calgary
Francisco Market Line Border
FIGURE 6. Two Competitor Pricing in PNW.
However, the Canadian plant might then be violat-
ing American anti-dumping laws by selling to the
U.S. at 5.2 Can. /lb FOB plant gate and in the
Canadian market at 6.2 Can. //lb. Because the
PNW market size is much larger than that in
Vancouver, the delivered price of methanol in Van-
couver should be set at 5.8 Can. /lb instead of
the 6.8 Can. /lb calculated in the absence of anti-
dumping laws. Oddly enough, the U.S. anti-dump-
ing law has had the effect of lowering the price to


the Canadian methanol consumer while benefitting
neither the American methanol producer nor
American consumers.
A word of caution is necessary to warn the
reader that the actual price will probably not be
equal to the one calculated above. Obviously the
two competitor model is much too simple to be
applicable in the commercial world. The extension
of the technique to a multi-competitor one is
straightforward in principle and a natural appli-
cation for a computer simulation. However, even
the more complicated model would probably not
arrive at a realistic price because of the un-
availability of accurate input data. The model
would be useful in the commercial world in the
hands of someone familiar with the industry. The
technique has proven to be a very useful tool in
showing students how the technical, economic and
regulatory factors influence the potential markets
for a chemical commodity. E

1. Various industry sources.
2. Chemical Economics Handbook, Standford Research
Institute, Menlo Park, California, Vol. II, 1977.
3. Brett Haugrud, M.Sc. Thesis, University of Alberta,

Kinetics of Coal Processing
Continued from page 18.
Reactant A diffuses through the product layer to
the reaction surface where chemical change occurs.
As reaction progresses, each pore of the particle
has associated with it a growing reaction surface
which initially corresponds to the inner surface
of the pore. As the various reaction surfaces in
the particle grow, it is inevitable that neighboring
surfaces will intersect one another as the solid B
separating them is consumed and replaced by the
product Q.
The growth of the total reaction surface may be
followed in terms of the radial growth of a given
set of overlapping cylinders, as shown in Figure
12. If the rate of reaction on the actual surface is
proportional to the total surface area

I -
/ 1-/ln 1X
( i-


and the conversion is

X=1- (1- ) exp [-r(1+ )] (3)

The surface development predicted by Equation
(2) is shown in Figure 13, exhibiting the antici-
pated maximum, but only for pore structures that




FIGURE 12. Overlapping of cylindrical surfaces. The
hatched area shows the overlapped
portion. The blackened area represents un-
reacted solid B. The reaction surface is the
interface between the nonoverlapped
portion of the cylindrical surface and the
unreacted solid. The product layer that is
deposited as the reaction surface moves is
not shown in this figure. (After Avrami).

provide 9 > 2. In this regard the model is more
flexible than those based on an order of reaction
or a grain model. The Petersen model (1957) can
also describe either kind of behavior; but it makes
no provision for further pore wall intersections
and neglects the distribution of pore sizes. As an
example of the use of the above technique, it is
worthwhile to return to the data of Hashimoto et
al. (1979) on the surface areas produced by steam
activation of chars. As a- -- oo, Equation (2) pre-
dicts a linear semilog relationship between the
group [S/(1 X)]2 and (1 X). This expectation
is supported by the data within the limits of ex-
perimental error, as shown in Figure 14.
That the subject of coal processing is a com-
plex matter with multiple branches is a common-
place. The emphasis of this paper has been
to demonstrate that the newer techniques of ex-
perimentation and modeling can shed new light
on the old fossil, if the broad problems of practice
are directed into manageable parts, each to be
digested in its turn. O


S 1-X
S. 1 T





U .

FIGURE 13. Development of the reaction surface with
conversion according to the random pore
model, compared with grain model for
m = 2/3 and Petersen model for co =
0.26, Lo = 3.14 x 106 cm/cm3, So = 2,425


stoichiometric coefficients
concentration of gaseous reactant A
rate constant for surface reaction
length of overlapped system
reaction order with respect to gas A
stoichiometric coefficients
initial particle radius
reaction surface area per unit
Sat t = 0


E = porosity
Eo = initial value of e
r = 4,Lo (1 o) /So2, structural
o- = R.So/ (1-E.), particle size parameter
T = k,CnSot/ (1 o), dimensionless time




0 02 01 06 08 10 1.2 1.4

FIGURE 14. Correlation of the data of Hashimoto et al.
(1979) with random pore model.


Bhatia, S. K. and D. D. Perlmutter, "A Random Pore Model
For Fluid-Solid Reactions," AIChE Journal, 26, 385
Brewer, R. E., "Chemistry of Coal Utilization," Vol. 1,
H. H. Lowry, Ed., p. 160, Wiley, New York, N.Y., 1945.
Coffman, A. W., Layng, T. E., Ind. Eng. Chem., 19 (8)
924-925 (1927).
Coffman, A. W., Layng, T. E., Ind. Eng. Chem., 20 (2),
165-170 (1928).
Foxwell, G. E., Fuel, 3, 122-128 (1924).
Hashimoto, K., K. Miura, F. Yoshikawa, and I. Imai,
"Change in Pore Structure of Carbonaceous Materials
During Activation and Adsorption Performance of Ac-
tivated Carbon," Ind. Eng. Chem. Process Design De-
velop., 18, 73 (1979).
Kam, A. Y., A. N. Hixson, and D. D. Perlmutter, "The
Oxidation of Bituminous Coal, Part 1: Development
of a Mathematical Model," Chem. Eng. Sci., 81, 815
Kam, A. Y., A. N. Hixson, and D. D. Perlmutter, "The
Oxidation of Bituminous Coal, Part 2: Experimental
Kinetics and Interpretation," Chem. Eng. Sci., 31, 821
Kam, A. Y., A. N. Hixson, and D. D. Perlmutter, "The
Oxidation of Bituminous Coal, Part 3: Effect on Caking
Properties," Ind. Eng. Chem. Process Design Develop.,
15, No. 3, 416 (1976c).
Karsner, G. G. and D. D. Perlmutter, AIChE Journal
Loison, R., Peytavy, A., Boyer, A. F., Grillot, R., "Chemis-
try of Coal Utilization," Suppl. Vol., H. H. Lowry, Ed.,
p. 150, Wiley, New York, New York, 1963.


a, b
p, q


Continued from page 34.
Then, if the field has not been able to respond
in the way the critics had hoped, what explana-
tion can be found?
Process control seems to have been unable
to successfully compete for research and develop-
ment funding with other fields, which have been
able to be more convincing concerning return of
investment, especially short term. Indeed, Shinnar
[46] states that "the total expense of any major oil
company on research in process control in any
given year is less than for one major television
Changes can be expected, however. The in-
creased emphasis on energy utilization will, as
mentioned above, put increased demands on a less
conservative and more tightly and carefully
controlled operation of existing processes. One
of the most efficient means to meet the increased
cost of energy is and will be a more efficient
energy utilization. Actually, one of the reasons
for the receptivity of European industry to more
sophisticated control approaches has been the high
cost of energy in Europe. Now when the U.S.A. is
experiencing higher energy costs-which will un-
doubtedly continue to climb-much more in-
dustrial interest in advanced control can be an-
The rise of the energy costs will force process
design in the same direction as process operation.
The processes and plants will become more inte-
grated, more sensitive, and more difficult to
control. Thus, both from the design and the
operational point of view more emphasis will
have to be put on advanced control in the future.
Another circumstance which might have re-
tarded the progress in process control research,
is the structure of U.S. faculties. The American
professor, be he assistant, associate or full, often
has no other research assistance than his (gradu-
ate) students. And when students have learned
enough to be fast and efficient workers they are
ready to leave. This system is not very efficient
in an applied field like process control, in which
such a large amount of knowledge and technics
from various fields have to combined in every
single non-trivial application. As expressed by
Wallman et al. [10] in a LQG-application to the
Foss chemical reactor: "It is undoubtedly evident
from the account given here that the approach
taken exacts a high price in engineering effort

and expertise. To recount, effort is required in
process modeling, parameter estimation, variable
reconstruction, and control system design. On top
of all this rests an overhead in approximation and
numerical methods."
The hierarchical European faculty system, in
which research groups form naturally, almost by
themselves, seems to be more efficient in this field.
The difference may become larger in the future as
the systems and problems to be treated are be-
coming more complex.
This organizational drawback in U.S. faculties
is pronounced by the fact that most faculties have.
not gone in for process control whole-heartedly
enough. One faculty member (the usual situa-
tion) is below the critical mass and no chain re-
action, so important for creative work, is
During the keynote address at the 1973 Joint
Automatic Control Conference Richard Bellman,
recently recipient of the IEEE Medal of Honor,
the highest honor that can be given by the world's
largest professional society, predicted that control
science will be the most influential discipline for
solving pressing national problems for the next
two to three decades. Athans has agreed [2], "with
no reservations whatsoever."
Even if Bellman and Athans turn out to be ex-
cessively optimistic, it is easy to agree with Athans
when he continues: "Modern control science will
continue to flourish with or without the support
of the process control community."
It seems to me that the question of whether the
chemical engineering community will be the one
that bears the burden and wins the victories in
chemical process control is still open. The potential
is there, but so are the obstacles. El


The material providing the background for the
present review was gathered during a 9-month
stay as a Senior Fulbright Fellow 1978-79 at the
Chem. Eng. Department at the University of Cali-
fornia at Berkeley. The stay was made possible by
grants from the United States Educational
Foundation in Finland and the Foundation of
Neste Oy. Visits to a number of other universities
and to an AIChE meeting were made possible by
a special travelers grant from the Research
Institute of the Abo Akademi Foundation. These
grants are acknowledged with sincere gratitude.


Most stimulating and valuable discussions with
a number of scientists and designers concerning
the topic of this paper as well as comments on and
criticism of the first draft of the paper are greatly
appreciated. Above all Alan Foss, so perfect a host
in Berkeley, as well as M. Athans, E. H. Bristol,
C. B. Brosilow, P. S. Buckley, J. M. Douglas, T. F.
Edgar, L. A. Gould, W. L. Luyben, T. J. McAvoy,
D. E. Seborg, F. G. Shinskey, R. M. Tong, R. K.
Wood, and L. A. Zadeh.


1. Foss, A. S., AIChEJ 19 (1973) 209.
2. Athans, M. in Weaver, R.E.C. (Coordinator), Pro-
ceedings of the Workshop on priorities in process
control research, Tulane-Univ. 1973.
3. Smith, C. L., Ibid.
4. Seborg, D. E., A survey of process control education
in the United States and Canada, Paper presented at
the 71st Annual AIChE Meeting, Miami, Nov. 1978.
5. Coughanowr, D. R., Koppel, L. B., Process Systems
Analysis and Control, McGraw-Hill 1965.
6. Morari, M., Ray, W. H., The Integration of Real-Time
Computing into Process Control Teaching at Wis-
consin, Paper presented at the 71 Annual AIChE
Meeting, Miami, November 1978. The first part of
the paper has been published in Chem. Eng. Education
13 No. 4 (1979) 160. The second part is in Chem. Eng.
Education 14, No. 1 (1980) 32.
7. Foss, A. S., Denn, M. M. (Eds), Chemical Process
Control, AIChE Symposium Series, 159 Vol. 72, 1976.
8. Vakil, H. B., Michelsen, M. L., Foss, A. S., Ind. Eng.
Chem. Fund. 12 (1973) 323.
9. Silva, J. M., Wallman, P. H., Foss, A. S., Ibid. 18
(1979) 383.
10. Wallman, P. H., Silva, J. M., Foss, A. S., Ibid. 18
(1979) 392.
11. Fisher, D. G. in Weaver, R.E.C. (Coordinator), Pro-
ceedings of the Workshop on priorities in process
control research, Tulane Univ. 1973.
12. Foss, A. S., Edmunds, J. M., Kouvaritakis, B., Ind.
Eng. Chem. Fund. 19 (1980) 109.
13. Bilec, R., Wood, R. K., Multivariable Frequency
Domain Controller Design for a Binary Distillation
Column, Paper presented at the 86th National AIChE
Meeting, Houston, April 1979.
14. Joseph, B., Brosilow, C. B., Tong, M., AIChEJ 24
(1978) 485.
15. Brosilow, C. B., The structure and design of Smith
predictors from the viewpoint of Inferential control,
Paper presented at JACC 1979.
16. Astrim, K. J., Wittenmark, B., Automatica 9 (1973)
17. Astr6m, K. J., Borison, U., Ljung, L., Wittenmark, B.,
Ibid. 18 (1977) 457.
18. Zadeh, L. A., Trans. ASME, J. Dyn. Syst. Meas.
Control 94 Ser. G (1972) 3.
19. Zadeh, L. A., IEEE Trans. Systems, Man, Cyber-
netics SMC-3 (1973) 28.

20. Bristol, E. H., IEEE Trans. Auto. Contr. AC-11
(1966) 133.
21. Shinskey, F. G., Process-Control Systems, McGraw-
Hill 1967, Second Edition 1979.
22. Rijnsdorp, J. E., Automatica 1 (1965) pp. 15, 29.
23. Nisenfelt, A. E., Schultz, H. M., Instr. Technology 18
No. 4 April (1971) 52.
24. McAvoy, T. J., ISA Trans. 16 (1977) No. 4, p. 83.
25. Witcher, M. F., McAvoy, T. J., Ibid. 16 (1977) No. 3,
p. 35.
26. Tung, L. S., Edgar, T. F., Analysis of Control-Output
Interaction in Dynamic Systems, Paper presented at
the 71 Annual AIChE Meeting, Miami, Nov. 1978.
27. Gagnepain, J. P., Seborg, D. E., An analysis of process
interactions with applications to multiloop control
system design, Paper presented at the 72nd Annual
AIChE Meeting, San Francisco, November 1979.
28. Kominek, K. W., Smith, C. L., Analysis of System
Interaction, Paper presented at the 86th National
AIChE Meeting, Houston, April 1979.
29. Shinskey, F. G., The Stability of Interacting Control
Loops with and without Decoupling, Proc. IFAC
Multivariable Technological Systems Conf. 4th Inter-
national Symposium, Univ. of New Brunswick (1977)
p. 21.
30. Jafarey, A., McAvoy, T. J., Ind. Eng. Chem. Proc. Des.
Dev. 17 (1978) 485.
31. Jafarey, A., McAvoy, T. J., Douglas, J. M., Ibid. 49
(1980) 114.
32. McAvoy, T. J., Ind. Eng. Chem. Fund. 18 (1979) 269.
33. Jafarey, A., McAvoy, T. J., Douglas, J. M., Ibid. 18
(1979) 269.
34. Bristol, E. H., Recent Results in Interaction in Multi-
variable Process Control, Paper presented at the 71st
Annual AIChE Meeting, Miami, Nov. 1978.
35. Donoghue, J. F., IEEE Trans. Ind. Elect. Contr.
Instr. IECI-24 (1977) 109.
36. Hammarstr6m, L. G., Waller, K. V., IEEE Trans.
Ind. Elect. Contr. Instr. IECI-27 (1980) 301.
37. Jafarey, A., Douglas, J. M., McAvoy, T. J., Ind. Eng.
Chem. Proc. Des. Develop. 18 (1979) 197.
38. Evans, L. B. in Weaver, R. E. C. (Coordinator), Pro-
ceedings of the Workshop on priorities in process
control research, Tulane Univ. 1973.
39. Waller, K. V., Paper and Timber, 62 (1980) 128.
40. Chen, F. T., Douglas, J. M., Georgakis, C., Ind. Eng.
Chem. Fund. 18 (1979) 321.
41. Mamdani, E. H., Applications of Fuzzy Set Theory
to Control Systems: A Survey, in Gupta, M. M.,
Saridis, G. N., Gaines, B. R. (Eds), Fuzzy Automata
and Decision Processes, Elsevier North-Holland 1977.
42. Tong, R. M., Automatica 13 (1977) 559.
43. Asbjornsen, 0. A., Survey Paper, Symposium on use
of process computers, European Federation of Chemi-
cal Engineering, 166th Event, Firenze, Italy, Sept.
44. Ballard, D., Brosilow, C. B., Dynamic Simulation of
Multicomponent Distillation Columns, Paper pre-
sented at the 71st Annual AIChE Meeting, Miami,
45. Athans, M., Discussion remark in round table dis-
cussion on "The next decade in control theory and
applications," IFAC VII Triennial World Congress,
Helsinki, June 1978.
46. Shinnar, R., Chem. Eng. Education 11 (1977) 150.



Departmental Sponsors: The following 140 departments contributed
to the support of CHEMICAL ENGINEERING EDUCATION in 1981 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
Cleve:and 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
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
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 Mississippi
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
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
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 Florida
University of Southern California
Stanford University
Stevens Institute of Technology
Syracuse University
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
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 CHEMICAL ENGINEERING EDUCATION, c/o
Chemical Engineering Department, University of Florida, Gainesville, Florida

"Can a new plastic

be produced efficiently? It's

up to me to decide."
Mark Carlson BS, Chemical Engineering

"This kind of decision-
making is a great responsibility.
At Du Pont I have the freedom to
do whatever testing is necessary
to make accurate judgements.
"Working with plastics was
my chief interest at the South
Dakota School of Mines and
Technology. I interviewed with
Du Pont because its strength in
the field matched my interests.
"I started work at the
Parkersburg, West Virginia, site
on process development for
DELRIN. Then I worked on the
engineering of a color develop-
ment facility. My present assign-
ment is compounding glass,
mineral and rubber reinforced
"All this calls for initiative
and gets me into design activities
with the marketing people.That's
the great thing about Du Pont.
You use a lot more than just your
If you want to develop all
your talents-whether you're a
ChE, ME or EE-see the Du Pont
Representative when he's on cam-
pus. Or write Du Pont Company,
Room 37798, Wilmington,
DE 19898.
At DuPont...there's a
world of things you can
do something about.

An Equal Opportunity Employer, M F.

Full Text