Citation
Chemical engineering education

Material Information

Title:
Chemical engineering education
Alternate Title:
CEE
Abbreviated Title:
Chem. eng. educ.
Creator:
American Society for Engineering Education -- Chemical Engineering Division
Place of Publication:
Storrs, Conn
Publisher:
Chemical Engineering Division, American Society for Engineering Education
Publication Date:
Frequency:
Quarterly[1962-]
Annual[ FORMER 1960-1961]
quarterly
regular
Language:
English
Physical Description:
v. : ill. ; 22-28 cm.

Subjects

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

Notes

Citation/Reference:
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:
01151209 ( OCLC )
70013732 ( LCCN )
0009-2479 ( ISSN )
Classification:
TP165 .C18 ( lcc )
660/.2/071 ( ddc )

UFDC Membership

Aggregations:
Chemical Engineering Documents

Downloads

This item has the following downloads:


Full Text














cheial engie g edu





GC C
achowleddesa and tanko....





THE
FLUOR FOUNDATION







CHEMICAL ENGINEERING EDUCATION
wimi a doowdiot oj /a#nds.









EDITORIAL AND BUSINESS ADDRESS
Department of Chemical Engineering
University of Florida
Gainesville, Florida 32611
Editor: Ray Fahien
Associate Editor: Mack Tyner
Editorial & Business Assistant:
Carole C. Yocum (904) 392-0861
Publications Board and Regional
Advertising Representatives:
Chairman:
Klaus D. Timmerhaus
University of Colorado
Vice Chairman:
Lee C. Eagleton
Pennsylvania State University
SOUTH:
Homer F. Johnson
University of Tennessee
Vincent W. Uhl
University of Virginia
CENTRAL: Leslie E. Lahti
University of Toledo
Camden A. Coberly
University of Wisconsin
Darsh T. Wasan
Illinois Institute of Technology
WEST: R. W. Tock
Texas Tech University
William H. Corcoran
California Institute of Technology
William B. Krantz
University of Colorado
C. Judson King
University of California Berkeley
NORTH: J. J. Martin
University of Michigan
Edward B. Stuart
University of Pittsburgh
NORTHEAST: Angelo J. Perna
New Jersey Institute of Technology
NORTHWEST: Charles Sleicher
University of Washington
Charles E. Wicks
Oregon State University
LIBRARY REPRESENTATIVE
Thomas W. Weber
State University of New York
PUBLISHERS REPRESENTATIVE
D. R. Coughanowr
Drexel University
UNIVERSITY REPRESENTATIVE
Stuart W. Churchill
University of Pennsylvania
James H. Hand
University of Michigan


Chemical Engineering Education
VOLUME XIV NUMBER 1 WINTER 1980


2 Departments of Chemical Engineering
Georgia Tech, Gary Poehlein

8 The Educator
Art Humphrey and Biochemical
Engineering at the University of
Pennsylvania, Alan E. Myers

IS Classroom
A Full-Year Course Sequence in Real-
Time Computing, D. A. Mellichamp
32 The Integration of Real-Time Computing
Into Process Control Teaching: Part II,
The Undergraduate Course,
M. Morari, W. H. Ray

26 Laboratory
Advanced Process Control Experiments,
Pradeep B. Deshpande, W. L. S. Laukhuf,
Nandkishor G. Patke
38 Process Control Experiment: The Toilet
Tank, Thomas J. Ward
42 Curriculum
A Survey of Process Control Education in
the U.S. and Canada, Dale E. Seborg

Class and Home Problems
45 Solution: The Mirror Fog Problem,
R. L. Kabel
46 In the "Heat" of The Night, R. L. Gordon
47 Two-Dimensional Heat Transport, A. Basio
14 Lecture
The Rate of Reaction: A Definition or the
Result of a Conservation Equation?
Alberto Cassano

12, 24 Book Reviews
44 ChE News


CHEMICAL ENGINEERING EDUCATION is published quarterly by the Chemical
Engineering Division, American Society for Engineering Education. The publication
is edited at the Chemical Engineering Department, University of Florida. Second-class
postage is paid at Gainesville, Florida, and at DeLeon Springs, Florida. Correspondence
regarding editorial matter, circulation and changes of address should be addressed
to the Editor at Gainesville, Florida 32611. Advertising rates and information are
available from the advertising representatives. Plates and other advertising material
may be sent directly to the printer: E. O. Painter Printing Co., P. 0. Box 87,.
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 1980 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.


WINTER 1980































Atlanta's skyline is shown behind the Tower of Tech's Administration Building.

department


CHE AT GEORGIA TECH

4 Pe"od oj T'ia*ditei


GARY W. POEHLEIN
Georgia Institute of Technology
Atlanta, GA 30332

LAST WINTER WHILE browsing in the campus
bookstore I found a stack of T-shirts imprinted
with the message "MIT-THE GEORGIA TECH
OF THE NORTH." I returned two days later to
purchase one for my friend Professor James Wei
(MIT Department Head and a Georgia Tech
graduate) only to find they had all been sold. In
their place were T-shirts which bore the message
"NORTH AVENUE TRADE SCHOOL" (North
Avenue is the southern boundary of the campus).
I decided not to buy one of the replacement shirts.
It probably would not have been received with

Copyright ChE Division, ASEE, 1980


much enthusiasm by Professor Wei.
Both T-shirt messages contained elements of
truth and fiction. Georgia Tech, like MIT, has a
long-standing reputation for quality engineering
and scientific education. Unlike MIT, Georgia Tech
is a State Institution, being part of the University
System of Georgia. Four colleges (Engineering,
Architecture, Sciences and Liberal Studies, and
Industrial Management) offer undergraduate and
graduate degrees in areas that could be described,
in a broad sense, by the term "Technology." De-
grees are not offered in areas such as music,
English, art, history, etc. This last fact is re-
sponsible for the "trade school" label on the second
stack of T-shirts. Of course, those of us associated
with Tech know that this label does not reflect


CHEMICAL ENGINEERING EDUCATION








reality. Our liberal studies departments are staffed
with high-quality faculty who offer a very wide
range of courses to help our students obtain a
broad education.
The Georgia Tech campus is located on 280
acres about 11/2 miles from the center of Atlanta.
With a population of about 1,750,000, Atlanta is
the center of commerce in the rapidly growing
Southeast. It is a dynamic, beautiful and exciting
city with a very diverse population. Atlanta is an
educational center with 29 degree-granting
colleges, junior colleges, and universities; a center
for the arts with a symphony orchestra, a ballet
company, numerous art exhibits, local and im-
ported theater groups, a wide variety of special
festivals; and a center for sports, with teams in
all major professional leagues.
The School of Chemical Engineering at Georgia
Tech is comprised of a Chemical Engineering Di-
vision, a Metallurgy Division, and the Fracture
and Fatigue Research Laboratory, an interdis-
ciplinary research organization. Faculty and other
scientific staff are listed in Table 1. The Chemical
Engineering Division offers B.S., M.S. and Ph.D.
degrees. The Metallurgy Division does not offer
a designated undergraduate degree but does have
a very active graduate program leading to M.S.
and Ph.D. degrees.
The School of Chemical Engineering has been
changing rapidly during the past few years. Dr.
Waldemar Ziegler retired at the end of the 1977-
78 academic year, several faculty have left for
other positions and, unfortunately, Dr. Leon
Bridger and Dr. Homer Grubb died suddenly
during the past academic year. Thus much of our
efforts during the past fifteen months have been
involved with recruiting new faculty. These efforts
have been very successful, with ten outstanding
individuals accepting offers to join our faculty.
Dr. Edvin Underwood was a Senior Research
Scientist here prior to accepting a faculty posi-
tion. Eight of the remaining nine have moved to
Atlanta since April, 1979. Dr. Amyn Teja, pres-
ently at Loughborough University, will join us in
September, 1980. The faculty members and their
areas of interest are identified in Table 1.


GEORGIA TECH-A BRIEF HISTORY*
The Georgia Institute of Technology began
as the Georgia School of Technology in 1888. The
School of Chemical Engineering evolved from the
chemistry curriculum. An Engineering Chemistry
Program was published in the 1900-1901 catalog.
In addition to chemistry and chemical engineer-
ing topics, areas such as metallurgy, dyeing, me-


Chemical Engineering is housed in the Bunger-Henry
Building.

chanics, electricity, minerals, and industrial pro-
cesses were included in this early program. The
1929 catalog listed a B.S. in Engineering Chemis-
try, and the 1930 catalog indicated that a B.S.
in ChE could be obtained under the Department
of Chemistry.
McLaren White, whose father, Alfred H.
White, was an MIT graduate and Head, of ChE
at Michigan, was the first chemical engineer to
come to Tech. This happened during the 1920s,
and Professor White was part of the chemistry de-
partment faculty. Changes toward chemical engi-
neering were too slow to suit Professor White, so
he left. The first-prize for the AIChE contest
problem is named in honor of McLaren White.
The first Chief of the Chemical Engineering
Division within the Department of Chemistry was

*The information presented here is condensed from a
paper, "An Early History of Chemical Engineering at
Georgia Tech," by Marcella M. Lusby, Unpublished (No-
vember, 1977).


The School of Chemical Engineering currently enrolls about 950
undergraduates. This figure includes freshmen who declare a major at Tech
and co-op students who are on work assignments. Cooperative education dates back
to 1915 at Tech. Today nearly 30% of undergraduate ChE's are in the Coop Program.


WINTER 1980 a








Dr. Harold Bunger, who started at Tech around
1929. The name of the academic unit was changed
to the Department of Chemistry and Chemical
Engineering in the 1930s. This department was

TA

Staff: ChE Division

Pradeep K. Agrawal; Assistant Professor; Ph.D. 1979,
University of Delaware; Heterogeneous Catalysis
Charles J. Aloisio; Lecturer; Ph.D. 1970, Purdue Uni-
versity; Polymer Engineering and Science
William R. Ernst; Associate Professor; Ph.D. 1974,
University of Delaware; Heterogeneous Catalysis;
Assistant Director for Ch.E. Undergraduate Pro-
grams
Larry J. Forney; Associate Professor, Ch.E./C.E.; Ph.D.
1974, Harvard University; Aerosol and Particle
.' Technology
George A. Fowles; Adjunct Professor; Retired Vice
President, Marketing, B.F. Goodrich Co.; Plastics
Pioneer
Charles W. Gorton; Professor; Ph.D. 1953, Purdue Uni-
versity; Transport Phenomena, Fluidization
,Edwin M. Hartley; Associate Professor; Ph.D. 1973,
.Georgia Tech; Pulp and Paper Engineering
SH. Clay Lewis; Professor; Sc.D. 1943, Carnegie Insti-
tute of Technology; Chemical Process Design
Albert A. Liabastre; Research Scientist; Ph.D. 1974,
Georgia Tech; Surface Science
Michael J. Matteson; Professor; D. Eng. 1967, Technical
University Clausthal (Germany); Aerosols, Particle
Technology, Air Pollution Control; Assistant Di-
rector for Ch.E. Graduate Programs
John D. Muzzy; Professor; Ph.D. 1970, Rensselaer Poly-
technic Institute; Polymer Engineering, Energy
Conservation, Economics
Allan S. Myerson; Assistant Professor; Ph.D. 1977,
University of Virginia; Thermodynamics, Crystalli-
zation, Biochemical Reactions
Clyde Orr, Jr.; Regents' Professor; Ph.D. 1953, Georgia
Tech; Instrumentation and Particle Technology
Gary W. Poehlein; Professor; Ph.D. 1966, Purdue Uni-
versity; Emulsion Polymerization, Latex Tech-
nology; Director of the School
Ronnie S. Roberts; Assistant Professor; Ph.D. 1976,
University of Tennessee; Biochemical Engineering,
Mass Transfer, Reactor Design
Robert J. Samuels; Professor; Ph.D. 1960, University of
Akron; Polymer Science and Engineering
A. H. Peter Skelland; Professor; Ph.D. 1952, University
of Birmingham (England); Non-Newtonian Fluids,
Mixing and Fluid Dynamics, Heat and Mass Transfer
Jude T. Sommerfeld; Professor; Ph.D. 1963, University
of Michigan; Computer Applications
D. William Tedder; Assistant Professor; Ph.D. 1975,
University of Wisconsin; Process Synthesis, Optimi-
zation and Waste Management
Amyn S. Teja; Associate Professor, Ph.D. 1972, Im-
perial College (London), Phase Equilibria, Thermo-
dynamics
Henderson C. Ward; Professor; Ph.D. 1953, Georgia


headed by Dr. Boggs, with Dr. Bunger continuing
as Chief of the ChE Division. Both Boggs and
Bunger died in 1941 and the two divisions were
split, forming separate departments. Professor


LBLE 1
Tech; Transport Phenomena, Process Design, Co-
Siting
Mark G. White; Assistant Professor; Ph.D. 1978, Rice
University; Heterogeneous Catalysis
Jack Winnick; Professor; Ph.D. 1963, University of
Oklahoma; Thermodynamics, Electrochemical Engi-
neering, Air Pollution Control
Ajit P. Yoganathan; Assistant Professor; Ph.D. 1978,
California Institute of Technology; Biomedical Engi-
neering, Polymer Rheology
Alex Zhavoronkov; Visiting Scientist; Ph.D. 1975,
Moscow Technical Institute; Aerosols
Staff: Metallurgy Division
Helen E. Grenga; Professor; Ph.D. 1967, University of
Virginia; Catalysis, Corrosion, Extractive Metal-
lurgy
Robert F. Hochman; Professor; Ph.D. 1959, University
of Notre Dame; Phys-Chem. of Metals, Corrosion,
Biomaterials; Associate Director for Metallurgy
John E. Husted; Professor; Ph.D. 1970, Florida State
University; Mineral Engineering
Miroslav Marek; Associate Professor; Ph.D. 1970,
Georgia Tech; Corrosion, Dental Materials
Pieter Muije; Associate Professor; Ph.D. 1971, Wash-
ington State University; Metallurgy, Mineral
Processing
Stephen Spooner; Professor; Sc.D. 1965, Massachusetts
Institute of Technology; Physical Metallurgy, Metal
Physics
Staff: Fracture and Fatigue Research Lab.
Saghana B. Chakrabortty; Research Scientist; Ph.D.
1974, Georgia Tech; Mechanical Metallurgy, Electron
Microscopy
Albrecht Gysler; Visiting Research Scientist; Ph.D.
1965, University of Stuttgart (Germany); Micro-
structure-Properties Relationships
Ludmilla Konopasek; Research Engineer; M.S. 1975,
Manchester University (England); Fracture and
Fatigue of Materials
Fu-Shiong Lin; Research Scientist; Ph.D. 1978, Georgia
Tech; Corrosion, Fatigue, and Ti and Al Alloys
T. H. B. Sanders; Research Scientist; Ph.D. 1974,
Georgia Tech; Aluminum Alloy Development, Micro-
structure and Fatigue
Bhaskar Sarkar; Postdoctoral Fellow; Ph.D. 1979,
Georgia Tech; Stress Corrosion Cracking and
Fatigue
Edgar A. Starke, Jr. Professor; Ph.D. 1964, University
of Florida; Fracture and Fatigue; Director of the
Fracture and Fatigue Research Laboratory
Edvin E. Underwood; Professor; Sc.D. 1954, Massa-
chusetts Institute of Technology; Physical Metal-
lurgy, High Temperature Deformation and Stere-
ology


CHEMICAL ENGINEERING EDUCATION








Jesse Mason became Director of the Chemical
Engineering Department; and Dr. Paul Weber, a
chemical engineering faculty member, was named
Assistant Director of the Engineering Experi-
ment Station.
A reorganization took place in 1948 with Pro-
fessor Mason becoming Dean of the College of
Engineering and Dr. Weber the new Director of
the School of Chemical Engineering. Dr. Weber
held this position until 1955 when he became Dean
of the Faculty (equivalent to the present position
of Vice President for Academic Affairs), and Dr.
Robert Raudebaugh from the Metallurgy Division
was appointed Acting Director of the School of
Chemical Engineering for the year. Dr. W. M.
Newton assumed the acting director position and


The "Ramblin Wreck" Parade, a real demonstration of
student creativity-weird things that move!

served for about three years until Dr. Homer
Grubb was named Acting Director. The name
change from "Department" to "School" was made
to identify ChE as a degree-granting component
of the Institute; the School of Chemical Engineer-
ing offers degrees, the Departments of English,
History, etc., do not. In addition, schools are
generally more autonomous than departments.


One of the reasons for
the dramatic, almost three-fold,
increase in ChE undergraduate enrollment at
Tech during the past five years has
been the number of women and
minorities entering the School.


Dr. Grubb was later named Director of the
School, a position he held until 1965 when Dr. Leon
Bridger became the Director. During this period
the Metallurgy Division developed a significant
graduate program which operated partially inde-
pendently of the Chemical Engineering Division.
Dr. Robert Hochman came to Tech in 1959 and
is currently the Associate Director responsible
for the Metallurgy Division.
Dr. Bridger returned to a full-time faculty
position in ChE during the summer of 1978; and
Dr. Gary Poehlein was appointed Director after
moving to Tech from Lehigh University. The
Fracture and Fatigue Research Laboratory,
headed by Dr. Edgar Starke, Jr., was also es-
tablished in 1978.

UNDERGRADUATE PROFILE

A PPROXIMATELY 8500 undergraduates were en-
rolled at Georgia Tech to start the Fall Quarter,
1979. About one-half of these students are resi-
dents of the State of Georgia. The others include
representatives from every other state in the U.S.
and many foreign countries. A selective a'dmis-
sions policy continues to produce a high-quality
undergraduate student body. The 1978 freshman
class had an average SAT score of 1161 comprised
of an average verbal score of 533 and an average
math score of 628. Georgia Tech ranks seventh in
the nation in attracting National Merit Scholars
and second to Harvard-Radcliffe in the' aintu of
National Achievement Scholars enrolled. :
The School of Chemical Engineering currently
enrolls about 950 undergraduates. This figure in-
cludes freshmen who declare a major at Tech and
coop students who are on work assignments. Co-
operative education dates back to 1915 at Tech.
Chemical Engineering became involved later, and
Professor Emeritus Waldemar Ziegler was the
first ChE Coop graduate receiving his degree in
1932. Today nearly 30 percent of undergraduate
ChEs are in the Coop Program.
One of the reasons for the dramatic, almost
three-fold, increase in undergraduate ChE enroll-


WINTER 1980









The undergraduate program in chemical engineering is quite
rigorous with a good balance between theory and practice. Required courses
include two quarters of transport phenomena, three quarters of unit operations, three quarters
of design, as well as courses in stoichiometry, reaction kinetics, process control, etc.
Engineering drawing and physical education remain as required courses.
i _!J! "


ment at Tech during the past five years has been
the number of women and minorities entering the
School. These numbers have increased from near
zero in 1974 to about 30 percent women and 8
percent minorities in 1979. A second significant
reason for our increased enrollment has been more
participation in dual-degree programs with other
universities.
The undergraduate program in chemical
engineering is quite rigorous with a good balance
between theory and practice. Required courses in-
clude two quarters of transport phenomena, three
quarters of unit operations, three quarters of de-
sign, as well as courses in stoichiometry, reaction
kinetics, process control, etc. Engineering draw-
ing and physical education remain as required
courses.
Students come to Georgia Tech expecting to
work hard, and most are willing to make the


Jack Childs checks the assembly of Polymer Fabric Ex-
truder.

necessary commitment to succeed. The vast ma-
jority leave Tech with a very high opinion of the
Institute and a fond feeling for Atlanta. This
loyalty is clearly manifested in the fact that
Georgia Tech almost always ranks first among
the nation's public institutions in terms of support
by alumni and alumnae.


GRADUATE PROGRAM

T HE GRADUATE PROGRAM is in a period of rapid
growth. The addition of ten new faculty mem-
bers should provide us with the manpower neces-
sary to improve an undergraduate program that
is already quite good and, at the same time, to
build a graduate program of equivalent stature.
Enrollment of full-time graduate students in the
Fall of 1978 was about 55. These students were
about equally divided between the ChE and Metal-
lurgy Divisions. Present graduate enrollment in-
cludes 61 chemical engineering students and 40
metallurgy students.
New graduate students entering the School of
Chemical Engineering in September, 1979, con-
sisted of 33 U.S. citizens and 9 from other
countries. Thirty-six of these students are in the
ChE Division and 6 in the Metallurgy Division. A
continuation of this sort of success in recruiting
graduate students will insure the proper develop-
ment of our graduate program. We would like to
achieve a steady-state graduate enrollment of
about 100 ChEs and 40 Mets. with an 80/20
balance between citizens and non-citizens.

THE FUTURE

T HE NUMBER OF NEW faces at our first faculty
meeting in September prompted a request for
round-the-table introductions. Of course, faculty
on our staff last year had an opportunity to meet
the new faculty during the campus visits. In fact,
a number of significant working relationships
have already been developed between new faculty
and those continuing to serve on our staff. During
a year when more than fifty interviews occurred,
however, names and faces can become confused.
This introduction exercise clearly illustrates
that the School of Chemical Engineering at
Georgia Tech is indeed in "A Period of Rapid
Transition." Prediction of the future in such an
environment is surely an uncertain endeavor. Our
faculty and students are very optimistic. We look
forward to an exciting period in the life of an out-
standing institution. O


CHEMICAL ENGINEERING EDUCATION








You know about

Atlantic Richfield Company.


Now start reading

between the lines.

ARCO Oil and Gas Company ",
Division of AtlanticRichfieldCompany
ARCO Petroleum Products Company ->
Division of AtlanticRichfieldCompany
ANACONDA Industries 4
Division of The ANACONDA Company
ARCO Transportation Company 0
Division of AtlanticRichfieldCompany
ARCO Chemical Company 1,
Division of AtlanticRichfieldCompany
ANACONDA Copper Company 4
Division of The ANACONDA Company
ARCO International Oil and Gas Company 'I
Division of AtlanticRichfieldCompany
ARCO Coal Company <,
Division of AtlanticRichfieldCompany
With all the companies you should be talking to, why not meet eight with
one interview? For more information regarding career opportunities
with us, see your Placement Office.


AtlanticRichfieldCompany A,
IF


An equal,opportunity employer, m/f









educator


441t ec444HfMJe"

and Biochemical Engineering
at the University of Pennsylvania

ALAN L. MYERS
University of Pennsylvania
Philadelphia, PA 19104

T HE PAST ACADEMIC year served as a milestone
for the field of biochemical engineering and
the University of Pennsylvania. Not only did the
year mark the 25th anniversary of the biochemi-
cal engineering program at the University, it was
also the 25th anniversary of the leadership of the
program by its founder at Penn, Arthur E.
Humphrey.
There are older and larger biochemical engi-
neering programs in existence, but none can rival
the record of Penn's program in terms of its dy-
namic growth and production of graduates who
have gone on to become leaders in the field. To
mark these achievements, the University hosted a
special 25th anniversary symposium to which all
the graduates of the program were invited. The
"alumni" who attended participated in panel dis-
cussions on the future of biochemical engineering
in the areas of food production (discussion led by
Stanley Barnett), energy production (led by
Charles Cooney), production of chemicals (led by
Daniel Wang), and preservation of the environ-
ment (led by Larry Erickson), with a roundtable
discussion on the future of biochemical engineer-
ing education led by Richard Mateles.

ORIGINS OF THE PROGRAM
IN 1953 ART ARRIVED at Penn, fresh from his PhD
studies in chemical engineering at Columbia
University. While working on his doctorate, he

As a scholar of biochemical
engineering in general, and fermentation
technology in particular, he has published
over 150 technical papers and co-authored
two textbooks which are regarded
as the bible in their fields.

Copyright ChE Division, ASEE, 1980


Art inspects the fermenter laboratory.


developed an interest in fermentation technology
and subsequently took additional training in food
technology at M.I.T. At Columbia, Art, along
with Ernest Henley (now at the University of
Houston), was the first doctoral student of Pro-
fessor Elmer Gaden who was later to be cited as
the "father of biochemical engineering" (in a
cover story in the May 31, 1971 issue of Chemical
Engineering News). It was through Elmer
Gaden's influence that Art's interest in fermenta-
tion systems matured. Elmer had written his own
thesis on "Mass Transfer in Fermentation
Systems under the supervision of Arthur W.
("Pop") Hixson at Columbia, working in conjunc-
tion with a biochemical engineering team (Karow,
Bartholomew, and Sfat) at the Merck Company.
(It is of some interest to note that reporting of
the 1945 McGraw-Hill Process Development
Award, presented to Merck for its "Biochemical
Engineering Development of the Penicillin Pro-
cess," is apparently the first mention of "bio-
chemical engineering" in the chemical engineer-
ing literature.)
Art joined Penn's chemical engineering faculty


CHEMICAL ENGINEERING EDUCATION









as an assistant professor and in the fall term of
1953 offered a new course in biochemical engineer-
ing. Since 1953 the biochemical engineering pro-
gram has grown steadily, and in 1972 it was
officially recognized when the Department of
Chemical Engineering formally changed its name
to the Department of Chemical and Biochemical
Engineering.

THE MAN BEHIND THE PROGRAM
AS OVERWORKED AS THE WORD "dynamic" is in
today's usage, it is nevertheless the appropri-
ate term to characterize the founder of Penn's
biochemical engineering program-Art Humph-
rey. Whether at work in his laboratory, or lead-
ing a faculty debate, or scaling a mountain in
Guatemala, Art is possessed of energy that leaves
others much younger in a race to catch up, and he
inspires his contemporaries to set equally chal-
lenging goals for themselves.
In his 25 years at Pennsylvania, Art has es-
tablished professional credentials that are indis-
putable. As a scholar of biochemical engineering
in general, and fermentation technology in par-
ticular, he has published over 150 technical
papers and co-authored two textbooks which are
regarded as the bible in their fields, Biochemical
Engineering (Academic Press, 2nd ed., 1976) and
Fermentation & Enzyme Technology (John Wiley,
1979). He holds three U.S. patents and has actively
consulted for more than 20 chemical companies
during his career. He has served on and chaired
numerous AIChE committees including the FBP


Art seems to approve of his students' efforts at office
"redecoration."


Over the years he has formed
a special bond with his students, based
on mutual respect, willingness to devote intense
effort to a project, and in no small
measure on the ability to dish
out and take a practical joke.


Division, and between 1975 and 1978 was a di-
rector of the organization. He has also been a
past chairman of the MC&T Division of ACS and
of the Working Group on the Production of Sub-
stances by Microbial Means of the US/USSR
Committee on Cooperation in Science and Tech-
nology. In addition, he served from 1971-1973 as
a member of the NSF Advisory Committee for
Engineering. In 1973 he was elected to the Na-
tional Academy of Engineering.
Such a summary of professional activities,
however, doesn't begin to capture Art Humphrey,
the man, who by virtue of his enthusiasm and
exuberant love for life succeeds in making people
want to work together toward a common goal. Art
always has a full complement of five doctoral
students to work with him as advises or, more
appropriate to the man, associates. Over the years
he has formed a special bond with his students
based on mutual respect, willingness to devote
intense effort to a project, and in no small measure
on the ability to dish out and take a practical joke.
For pranks he's played on his students, Art was
rewarded on one occasion with the careful "re-
decoration" of his office (as shown in the ac-
companying photograph) and on another with the
clamping of a ball and chain on his leg only
minutes before he was to make a presentation to
the University Trustees. In Houdini-like fashion
he arrived before the Trustees on time sans ball
and chain.
Art is an avid outdoorsman who cheerfully
ignores middle age (and some say common sense)
as he pursues his interests. Having hiked the trails
of much of this country, he frequently substitutes
an opportunity to climb a mountain in a foreign
country in place of the honorarium he would re-
ceive for lecturing there. Such negotiations have
enabled him to climb Mount Fuji in Japan, Pop6-
catepetl in Mexico, and Augua in Guatemala,
among others. Last year when taking up skiing
for the first time, Art found himself by mistake in
an advanced intermediate class and, not having
the good sense to get out of the class, "mostly fell


WINTER 1980










Collectively, this group conducts a research program amounting to
$500,000 a year, and this past year they published the program's 150th
paper on biochemical problems.


down the hill" and received a National Standard
Medal for downhill racing as a result. Each year
Art leads a departmental canoe trip down the
Delaware River, an activity he organized 16 years
ago (when Bob Bird was a member of that group).
The annual event now involves between 60 and
70 people, including faculty members, students,
alumni, and participants from other schools. This
year an eight-mile run was included as part of
the trip, and he typically outpaced the younger
generation-without having trained for the run.
The honesty and openness with which Art ap-
proaches every aspect of his life generates an un-
swerving loyalty among his associates. Never
afraid to admit when he's wrong or simply doesn't
know an answer, he leads people to trust that
they'll "hear it straight," good or bad, in dealing
with him. Add to these qualities Art's rare talent
of being a good listener, and the result is an in-
dividual who is extremely effective in accomplish-
ing the work he sets out to do. His spirit has
inspired and guided the development of Penn's
biochemical engineering program over 25 years,
and it has set the tone with which the biochemical
engineering faculty now approaches the next 25
years.

BIOCHEMICAL ENGINEERING AT PENN TODAY
ART HUMPHREY STILL heads the biochemical
engineering program within the University's
Department of Chemical and Biochemical Engi-
neering, despite his heavy administrative load as
Dean of the School of Engineering and Applied
Science. (From 1962 until he was named Dean in
1972, he served as Chairman of the Department
of Chemical Engineering.) Art and three others
of the Department's 12-member faculty have the
focus of their teaching and research activities in
the area of biochemical engineering. At least three
others conduct a significant portion of their re-
search within the field. Collectively, this group
conducts a research program amounting to
$500,000 a year, and this past year they published
the program's 150th paper on biochemical engi-
neering problems (out of a total of more than 350
scientific articles published by these individuals).
Six national AIChE, ACS, and ASEE awards have


been won by members of the group, and two mem-
bers have been elected to the National Academy
of Engineering.
Those members of the faculty of the Depart-
ment of Chemical and Biochemical Engineering
who participate in the biochemical engineering
program are:


David J. Graves:
Arthur E. Humphrey:
Douglas A. Lauffenburger:
Mitchell D. Litt:
Daniel D. Perlmutter:
E. Kendall Pye:

John A. Quinn:


enzyme kinetics
fermentation technology
cell population dynamics
biorheology
enzyme reactor dynamics
enzyme behavior and
purification
bound membrane systems


The facilities which are used in both the re-
search and teaching functions of the biochemical
engineering program consist of four primary
laboratories: the fermenter laboratory, a wet
chemistry laboratory for enzyme analysis, a mem-
brane laboratory, and a reactor laboratory.
The fermenter laboratory (pictured in an ac-
companying photograph) centers about a 70-liter
highly instrumented fermenter that is coupled to
a PDP 11/34 computer with a 96K core capacity
and three discs (two fixed), a Calcomp plotter
and a video screen for data acquisition and
control. The facility is supported by a Nuclide
Mass Spectrometer and other appropriate analysis
equipment. In addition, the laboratory has a
number of smaller fermenter units, including a
20-1, 14-1, two 1-1, and five 500-ml systems-most
having temperature, pH, foam, and dissolved
oxygen control.

ACADEMIC PROGRAMS

U NDER THE LEADERSHIP of Art Humphrey, the
biochemical engineering faculty has always in-
sisted that the academic program remain a part of
the basic program in chemical engineering, allow-
ing the "biochemical" aspects of the program to
emerge from an emphasis on biological processes.
Because of the emerging significance of bioconver-
sion processes in the production of energy, food,
and chemical feedstocks and as a means for con-
trolling the environment, students are eager to in-
vestigate these problems. The faculty believes the
students' eventual careers are better served by


CHEMICAL ENGINEERING EDWUATION







having them pursue these interests from a solid
foundation in chemical engineering, rather than
by focussing exclusively on a subspecialty.
Thus, the undergraduate biochemical engineer-
ing student takes the standard chemical engineer-
ing curriculum, but biochemistry is substituted for
one. of the courses in organic chemistry, as is a
biology course for the course in nuclear physics.
In addition, the student will take two or three
courses in microbiology, biological processes, utili-
zation of wastes, biochemical engineering, or food
engineering as his senior technical electives; and
his senior research project will focus on a bio-
logical process.
Out of a senior chemical engineering class of
45 students, about 7 of them will be taking courses
with a focus in biochemical engineering. Most of
these students will either continue their biochemi-
cal engineering studies at the graduate level or
will enter medical school.
At the graduate level the student planning a
focus in biochemical engineering is also expected
to take the core courses in chemical engineering-
in applied mathematics, transport processes, re-
actor design, and thermodynamics. The student is
then expected to take courses in advanced bio-
chemistry, molecular biology, and genetics, in addi-
tion to the four basic graduate courses in bio-
chemical engineering.
The basic graduate level biochemical engineer-
ing courses include:
Biochemical Engineering: fermenter kinetics, design
and operation


Biological Processes:

Enzyme Technology:


At any given time, about 15 students in the
chemical and biochemical engineering graduate
program (which numbers approximately 50 full-
time students) will be focussing their studies in
the direction of biochemical engineering. Upon re-
ceiving their doctoral degrees, these individuals
generally seek employment in the food production,
pharmaceutical, waste treatment, and chemical
process industries, or in academia.
When questioned, most of the alumni of Penn's
biochemical engineering program say they con-
sider themselves chemical engineers who have an
interest in biological processes, for such is the
slant of their curriculum. This may in part ex-
plain why Penn graduates have never encountered
problems on seeking employment. Indeed, a recent
graduate of the program looking for a position in
the biochemical field received more than ten offers
from firms, with several offers in excess of $31,000
a year.

FUTURE OF BIOCHEMICAL ENGINEERING:
ART HUMPHREY'S VIEW

T HE FACULTY OF THE biochemical engineering
program at Pennsylvania looks forward to con-
tinued growth of the program. The field is now


physics and chemistry of
biological processes
behavior and utilization of
enzymes


Utilization of Wastes: waste utilization
and treatment
These courses are taught by the faculty on a
rotating basis, i.e., once every other year. They
are frequently team-taught with the help of visit-
ing professors and members of the University's
Department of Biochemistry and Biophysics.
Students in the graduate program can develop
and shape their programs to serve their own par-
ticular career emphases by selecting additional
courses in areas ranging from nutrition to micro-
biology. They are free to select these courses from
throughout the University's graduate and profes-
sional programs, including its Medical School and
School of Veterinary Medicine.


Over 60 faculty members and students now participate
in Art's annual canoe trip down the Delaware River.
coming into its own and the opportunities are
unlimited. Perhaps the enthusiasm of those active
in Penn's biochemical engineering program comes
across best in the words of the man responsible
for its flourishing here-Art Humphrey.
"Never has the future for biochemical engi-
neering looked so bright. This is due largely to
the energy crisis and the attendant emphasis on
the use of renewable resources, meaning materials
of biological origin. It seems fairly evident that
many of these materials will be processed by bio-


WINTER 1980








logical means, with the use of enzymes, in order
to achieve low temperature, energy saving pro-
cesses. The environmental crisis and the increas-
ingly strict limits placed on the use of nitrates,
phosphates, and other surface water contaminants
mean that more efficient and more complicated
waste treatment systems will have to be evolved.
Also, wastes will be viewed in the future as valu-
able resources which can be treated to yield useful
materials.
"Perhaps the most significant development
affecting the future of biochemical engineering is
the explosion of knowledge concerning genetic
engineering techniques. Not only do we now
possess the ability to cultivate both animal and
plant tissue cells in large-scale reactor systems,
but we can transfer their genetic information for
making various biologically active molecules such
as, insulin into more easily cultivated bacterial
cells by gene splicing techniques. Soon the
scientist will be able to create cells with virtually
any desired metabolic activity. When this comes
to pass, the biochemical engineer will become
active in efficiently simulating and optimizing
many of nature's special reactions in stainless steel
fermenters. In many ways biomass can be re-
garded as the crude oil of the future. Just as crude
oil now serves as the feedstock of the petrochemi-
pal industry, from 'barrels of biomass' will come a
number of the chemical feedstocks of the future. It
Would not surprise me to see biomass refineries
emerging within the next decade.
"I for one will welcome the change. I believe
the chemical engineering textbooks of the future
will reflect this change and will include examples
of biomass problems along with those from the
petroleum ,industry. Chemical engineering is a
truly broad-based discipline, and I believe it is
already demonstrating its concern not just with
physical and chemical changes, but with biological
changes as well." FE


RMNbook reviews


CONTACT CATALYSIS, VOLS. 1 and 2
Edited by Z. G. Szabo, Elsevier Scientific, 1976
Reviewed by John B. Butt, Northwestern U.

This monumental two volume set is an essay
of the Catalysis Club of the Hungarian Academy
of Sciences with individual chapters contributed


r POSITIONS AVAILABLE
Use CEE's reasonable rates to advertise. Minimum rate
% page $50; each additional column inch $20.
(Please mention CEE when responding to ad.)
RENSSELAER POLYTECHNIC INSTITUTE
CHAIRPERSON. The Department of Chemical and
Environmental Engineering at Rensselaer Polytechnic
Institute invites applications for the position of Chair-
person. The candidate must hold the Ph.D. degree, have
made significant contributions to the literature of chemical
and/or environmental engineering and be nationally
recognized. The candidate must have administrative and
leadership abilities and be able to represent the depart-
ment externally. The department is balanced, has an ex-
cellent undergraduate program, numerous active research
programs and seeks "first rank" status over the next
decade. In the past five years, the Engineering School
as a whole has expanded its graduate programs sig-
nificantly and currently ranks sixth nationally in total
external funding (over $8 million). The salary is open.
The deadline for applications is April 1, 1980. Submit re-
sume and four letters of recommendation to Prof. H. Litt-
man, Chairman, Search Committee, 123 Ricketts Bldg.,
RPI, Troy, NY 12181. RPI is an Affirmative Action/Equal
Opportunity Employer.


by thirteen different authors. In many ways the
work is reminiscent of the series "Catalysis"
edited by Prof. P. H. Emmett in the 1950's, and
it promises to be as useful. True to the title, the
entire field of contact catalysis is treated, starting
with the fundamentals of solid state science,
chemisorption and kinetics in the first volume,
with applications concerning preparation, charac-
terization, and catalytic reaction engineering in
the second volume. The topics included are treated
in quite some detail and in many instances
represent current state of the art in both catalysis
research and applications.
With so many different aspects of the field
treated in such detail, it is difficult in a review of
reasonable length to do other than cite certain
parts that are of particular use to the reviewer.
In this respect, there are particularly fine treat-
ments of adsorption on solid surfaces and physical
characterization methods which eclipse much
existing literature. For example, the characteriza-
tion methods discussed include x-ray diffraction,
electron-optical methods, magnetic properties,
electrical properties, adsorption, infrared and
EPR spectroscopy.
Another very useful chapter deals with the
preparation of catalysts. This is particularly
timely now, since we have attained sufficient
abilities in characterization that the long-time
Continued on page 44.


CHEMICAL ENGINEERING EDUCATION






Monsanto Drive.
It takes you a very long way.


This sign marks the road that leads
into our International Headquarters in
St. Louis.
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: Ray
Nobel, Monsanto Company, Professional
Employment Dept., Bldg. A3SB, 800
North Lindbergh, St. Louis, Mo. 63166.

Monsanto
An equal opportunity employer


m .'.4









JIM lecture


THE RATE OF REACTION: A DEFINITION OR THE

RESULT OF A CONSERVATION EQUATION?*


ALBERTO E. CASSANO
Institute de DesarroUo Tecnol6gico para la Ind.
Quimica
3000-Santa Fe, Argentina

A quick survey of the classical literature on
chemical kinetics and reactor analysis reveals
different criteria concerning the definition of the
rate of reaction. Relying on historical rather than
rational bases, the following "definition" has been
used:
dC-
ri = (1)
dt
Eq. (1) was probably derived from the first
physicochemical studies on rates of reaction de-
veloped in constant volume batch systems.
This "definition" has been used by Glasstone
[1], Benson [2], Daniels [3], Laidler [4], Frost and
Pearson [5], and Johnston [6], among others.
On the other hand, chemical engineers mainly
engaged in design problems, distrusted the validity
of the "definition," taking into account its in-
applicability to reacting systems of variable
volume. The alternative "definition" was based on
the number of moles instead of concentration. The
difference lies in defining beforehand an extensive
rate of reaction which can then be turned into an
intensive property dividing by the reaction volume.
dNi
r' d (2)
dt

ri d (3)
V, dt
The greater generality of expression (3) as
compared to equation (1) is easily demonstratable,
in either conceptual or mathematical terms.

*This paper is the result of mutually beneficial dis-
cussions on the subject with my former classmates
Charles Allen and Sieghard Wanke. Any merit to this
paper is theirs. Any errors in it are mine.


The "definitions" belonging to the kind pro-
vided by equation (3) can be found in books such
as those written by Hougen and Watson [7],
Smith [8], Levenspiel [9], Walas [10], Boudart [11],
Pannetier and Souchay [12]; and with explicitly
stated limitations by Kramers and Westerterp
[13], by Aris [14] and Denbigh [15] as well. Many
other cases could be quoted.
A good example for analysis is the "Continuous
Flow Stirred Tank Reactor" (CFSTR). In this
particular case the product is continuously re-
moved at the system outlet and it becomes evident
that at the steady state dNi/dt is zero (the product
being species i) but the rate of reaction is utterly
different from zero. Here, as Denbigh remarks, ex-
pression (3) is not valid, either. This author tries
to narrow the applicability of the "definition" (3)
to processes where the only change in reactant i is
due to a chemical reaction. This would exclude any
other form of physical phenomena causing changes
in reactant concentration. The restriction seems
valid for the CFSTR and for all forms of diffusive
flows. Nevertheless, it does not seem so clear for
the "Piston or Plug Flow Reactor" (PFR), in
which if, on the limit, dt is taken as the reacting
mass average residence-time in the elementary re-
action volume of length dz, equation (3) may be
properly applied with some substitutions.*
This work shows the futility of arguing about
the accurate "definition" of the reaction rate and

*It should be noticed that now vz may be a function of
z if changes in the number of moles take place.


This work shows the
futility of arguing about the
"definition" of the reaction rate and
the convenience of dealing with the subject
from a different viewpoint.

( Copyright ChE Division, ASEE, 1980


CHEMICAL ENGINEERING EDUCATION







the convenience of dealing with the subject from
a different viewpoint, that is, drawing the
necessary equations from a more fundamental
principle, as the general mass conservation equa-
tion for multicomponent systems would be.

REDEFINITION OF THE PROBLEM FOR
HOMOGENEOUS REACTORS

S FAR, IT IS obvious that the existing "defini-
tions" seem to depend either on the author's
personal likings or on the reactor to which the
equations will be applied.
It seems logical to assume that any reaction
rate should be a function not of the system in
which it has been determined but only of tempera-
ture, pressure and concentration of the species
participating in the reaction, just to mention the
most commonly encountered variables affecting
the rate. The reaction rate could be naturally in-
fluenced by the type of reactor-continuous or
batchwise-but only up to the extent that it may
affect concentrations, temperatures or pressures.
It is necessary to reconcile an expression de-
rived from the ontological concept of the rate of
reaction with a mathematical equation expressing
exactly the same, involving the variables suscepti-
ble of experimental measurement.
Ontologically (and considered as an intensive
property), "The rate of reaction is the change in
the number of moles which takes place in unit time
and unit reaction volume, due to a transformation
of reactants into products."
The mathematical expression of the above de-
finition will be set equal to the kinetic equation of
the chemical system under consideration; i.e. if
such a definition is adequately represented by a
mathematical proposition which will be called r,
it is evident that:
r = ) (C,, C2 .... etc. T, P, etc.) (4)
where 4 depends on the complexity of the reacting
system (order, molecularity, activation energy,
etc.). It is the concern of this work to determine
the adequate formulation for the left-hand side of
equation (4).
The problem at issue has been partially, even
though quite accurately, discussed by Petersen
[16] and also singularly viewed by Amdur and
Hammes [17], particularly for batch reactors. The
reader may resort to the quoted references.
In many cases, experimental measurements will
register not only the change in concentration but
also the resultant of the chemical process of re-


Alberto E. Cassano is the founder and Chairman of INTEC; Pro-
fessor at the Universidad Nacional del Litoral and Member of the
Scientific Research Staff of the National Council for Scientific and
Technological Research of Argentina. He received his Chemical Engi-
neer's degree from the Facultad de Ingenieria Quimica of U.N.L.
(Santa Fe, Argentina) and his Ph.D. degree from the University of
California, Davis. His research interests are in Photochemical Re-
actors and Gas-Liquid Reactions Catalyzed by Solids. At present he
is also responsible for a consulting contract (undertaken by INTEC)
for the Argentinian Atomic Energy Commission engaged in the de-
velopment of a Heavy Water Experimental Plant.

action and the physical processes of diffusion, con-
vection and volume changes. From this point of
view, all the "definitions" previously analyzed are
erroneous and inaccurate, and only applicable, at
best, to systems universally used but no less par-
ticular (e.g. the constant volume stirred batch re-
actor, etc.).
Petersen clearly points out that the left-hand
side of equation (4) is a direct resultant of the ex-
perimental system employed in kinetic determina-
tions, and that, therefore, there is no single de-
finition which could be widely used, still less if
the reacting system were non-isothermal. This
statement can be effectively generalized and is the
subject of the next paragraph.

SOLVING METHODS
F THE RATE OF reaction is to be defined, the
problem is usually reduced to finding the rate
of change of the reacting species with respect to
some independent variable (time in a batch reactor
and position in many continuous reactors). The
rate of change of the reacting species is generally
measured considering the change in the number
of moles of the said reactant or the rate of change
in its concentration.
An apparent way of solving the problem would
be to consider the reaction within a species ma-
terial volume, instead of the generally used fixed


WINTER 1980








control volume. But it is difficult to think of a
species material volume since they are not pre-
served [18]. If the "particles" of the material
volume are elements, the difficulty is overcome at
the expense of a greater complexity. Therefore,
this strategy is quite troublesome.
A more adequate way of looking into the
problem for an isothermal system is to consider
the various mechanisms which can bring about
concentration changes of a reactant within a
system and, then, determine which part of these
total changes is due to the chemical reaction, sub-
tracting all contributions other than the reaction.
This means to state in detail a mass conserva-
tion balance for a multicomponent system. Gener-
ally speaking, there are two categories of phe-
nomena through which a species concentration
may vary in a fixed volume in space: (1) The
species may appear or disappear by chemical re-
action and (2) There is a net flow of this species
through the area of this volume element. This flow
mechanism, be it diffusion, forced convection or
any other means of mass transport, needs not be
detailed here. The description of this flow will ex-
clusively depend on the system at stake.
The general conservation equation may be thus
written [19]:
C + V Ni = ri (5)
at
or in a more general way: if a, is the stoichio-
metric coefficient of species i:
ri = r ai (6)
Then:

-Ci + V 'Ni = ar (7)
at
This well known conservation equation is the only
general formulation for homogeneous isothermal
rates of reaction that is independent of the re-
action system being employed. It clearly shows
that the rate of reaction is the "source" or "sink"
term'in the mass inventory; therefore all the re-
maining non-zero terms resulting from taking r
out of expression (7) must be substituted into the
right-hand side of equation (4).

So far, it is obvious
that the existing "definitions" seem
to depend either on the author's personal
likings or on the reactor to which
the equations will be applied.


APPLICATION TO COMMON REACTING SYSTEMS

EQUATION (7) WILL now be used to obtain the
adequate expressions for some classical react-
ing systems. The following homogeneous iso-
thermal systems will be considered:
1) Isothermal constant volume batch reactor
2) Isothermal variable volume batch reactor
3) Steady state, isothermal continuous plug flow re-
actor
4) Steady state, isothermal continuous flow stirred
tank reactor
In every case, the adequate assumptions will
be made in order to allow an analytical description
of the flux Ni so as to simplify the rate expression.

ISOTHERMAL CONSTANT VOLUME BATCH REACTOR
F WE ASSUME THAT properties are constant in
the whole volume of the reactor (especially con-
centration and temperature), then the divergence
of the flux becomes zero since Ni will be inde-
pendent of position. Therefore, if V Ni=0, equa-
tion (7) is reduced to:
1 dC1
r = (8)
ai dt
where the derivative is now total insofar as con-
centration will be a function of time only.

ISOTHERMAL VARIABLE VOLUME BATCH REACTOR
TF WE ASSUME THAT the expansion of volume is
slow so that, as in the previous case, such proper-
ties as concentration, pressure and temperature
are independent of their position within the re-
actor, no diffusional effects whatsoever will result,
and the flow will be caused only by expansion.
In this case, by definition:
Ni = Cr.vi*
but, as all species expand at the same rate:
VI = Vj = V "=V*
the velocity of all species coincides with the global
velocity of the system.
With the previous relationships, equation (7)
results:

S+ V (C v) = a r
at
But, since it has been assumed that concentration
is independent of position V Ci = 0. Then:

i+ Ci (V v) =air (9)
Continued on page 48.
Continued on page 48.


CHEMICAL ENGINEERING EDUCATION











UNION


'4~1 "f


THE MOST VERSATILE
RAW MATERIAL ON EARTH.
Modern technology-much of it
developed by Union Carbide-
transforms oil and natural gas
into thousands of vital petrochem-
ical products: synthetic fabrics,
plastics, even medicines.
Yet, sadly, most oil and nat-
ural gas are burned up, as fuel.
Onlyfourpercent goes into
petrochemicals. At Union Car-
bide, we're working on ways to
make thatfour percent go as far
as we can, to build a better future
for more of the world's people.
_ JA C'


JUST WHAT THE DOCTOR
PRESCRIBED.
The purity of the medicines
you depend on may well begin
with Union Carbide ingredients.
We provide basic materials for
everything from aspirin to
sedatives and antibiotics.


PLASTICS THAT SAVE GAS.
FABRICS THAT SAVE BOTHER.
Increasingly, we're using oil to
save gas- by turning it into
strong, light plastics that trim
cars' weight and boost their
mileage. Union Carbide also
supplies automakers with petro-
chemicals for everything from
resilient urethane bumpers to
polyethylene electrical insula-
tion. We even make Prestone II
anti-freeze.
Your carefree synthetic
fabrics are petrochemicals, too.
Union Carbide makes one of
the basic ingredients in polyester,
the most popular
synthetic of all. l


PESTICIDES TO HELP FEED
THE WORLD.
Union Carbide pesticides are
registered for control of a wide
variety of crop-destroying
pests. Around the world,
they're helping increase
yields of rice, soybeans
and other basic foods.


BRIGHT FUTURES.
Millions of pound&


of Union


resins and
latexes help
modern paints
protect and
Sbeautify just
about every-
thing under
the sun. And
the rain.

An equal opportunity employer.


UNION
WORKING WITH NATURE TODAY,
FOR THE RESOURCES WE'LL NEED TOMORROW.
Union Carbide Corporation, 270 Park Avenue, NewYork, N.Y. 10017


P1 ";i Lr-F-p









Classroom


A FULL-YEAR COURSE SEQUENCE

IN REAL-TIME COMPUTING


D. A. MELLICHAMP
University of California
Santa Barbara, CA 93106

T HE CHEMICAL ENGINEERING program at U.C.
Santa Barbara began in the mid-1960's with
an intended emphasis on the process control part
of the curriculum. The use of digital computers
for data acquisition and control, what often is re-
ferred to as real-time computing, at that time was
an object of relatively intense interest in industry.
However, except for one or two schools with the
financial resources to acquire an industrial-scale
computer (e.g. IBM 1800), there was little op-
portunity for universities to give students hands-
on experience with data acquisition and control
computers.
The appearance of the minicomputer in the late
1960's made it possible for virtually any depart-
ment to acquire or build a real-time computing
facility which could be used in teaching and in re-
search. The more recent introduction of the micro-


Duncan A. Mellichamp is Professor of ChE at the University of
California, Santa Barbara, where he was initially responsible for de-
velopment of the process dynamics and control program. An early
interest in computers led subsequently to activities with the CACHE
Real-Time Task Force (Chairman 1974-76), to the CACHE Corp. Board of
Trustees (President 1977-78), to the Editorship of the CACHE Mono-
graph series in Real-Time Computing, and to a long succession of
aggravations associated with building and maintaining a real-time
laboratory facility. He would like to forswear all future activities in-
volving computers, but cannot.


processor (and microcomputer) has merely ac-
celerated the trend of chemical engineering depart-
ments to install real-time facilities and to introduce
elements of real-time computing into the curricu-
lum.
In the past few years, the early fascination
with hardware and software among real-time

TABLE 1
Topics Covered in Undergraduate Control Courses
(2 hr. lecture and 3 hr. lab per week for two quarters)
Derivation of process dynamic models
Transfer function models
Open- and closed-loop systems
Frequency-response methods for controller design
Process applications
Overview of advanced control methods

users has decreased to some extent, and the more
important educational questions of what, where,
and how to teach this new subject area are re-
ceiving more attention. Since real-time computing
was and is tied to the subject of process control,
it seems reasonable to explore some ideas concern-
ing the teaching of real-time computing within
that context. The chemical engineering depart-
ment at U.C. Santa Barbara has developed what
must be one of the most extensive teaching and
research programs in real-time computing in this
country (at least among chemical engineering de-
partments) ; and I would like to describe it, to
discuss some of the background behind its de-
velopment, and to note how it is changing. To
begin with, however, it might be useful to discuss
the process dynamics and control program which
now coordinates with the more recently developed
real-time computing program.

THE PROCESS CONTROL PROGRAM
SN DESCRIBING THE PROCESS control program at
UCSB, it probably is fair to say that it is a tra-
ditional one. Two one-quarter undergraduate
courses are required of all seniors and the usual

Copyright ChE Division, ASEE, 1980


CHEMICAL ENGINEERING EDUCATION










If there is any non-traditional aspect of the undergraduate program
it would be the emphasis on "practical" experiments. The associated laboratory contains
several experimental units for the study of liquid level and stirred tank heating dynamics
and control, pneumatic and electronic control systems, and simulation facilities ....


range of topics is covered. Depending on who is
teaching the course, there may be more or less
emphasis on process dynamics and on "advanced
control" topics; nevertheless the core areas listed
in Table 1 are fairly rigorously and extensively
covered.
If there is any non-traditional aspect of the
undergraduate program it would be the emphasis
on "practical" experiments. The associated labora-
tory contains several experimental units for the
study of liquid level and stirred tank heating dy-
namics and control, pneumatic and electronic
control systems, and simulation facilities for other,
more complex processes. The experimental units
are all bench scale and designed to have time
constants on the order of one to two minutes.
Since individual dynamics and control experiments
can be carried out in 15 to 30 minutes, a whole
range of experiments (summarized in Table 2)
can be run during the two quarters. Students write
up their results in brief (i.e. memo) form;
although theory underlies all experimentation,
emphasis is on the use of theory to evaluate practi-
cal consequences such as "Will the surge tank
system overflow?", "Is the catalyst mixing system
adequately stirred?", etc. By the end of two
quarters students will have designed and tested
TABLE 2
Undergraduate Process Dynamics
and Control Laboratory Experiments
FIRST COURSE
Liquid level system:
1. Step response
2. Pulse response
Stirred tank heating system:
3. Step response
4. Transportation lags
Stirred tank reactor (simulated):
5. Model parameter fitting
6. Steady-state optimization
SECOND COURSE
Controllers:
1. Dynamic characteristics of three-mode controllers
2. Control of 1st- and 2nd-order systems
Stirred tank heating system:
3. Frequency response
4. Closed-loop control system design
Liquid level system:
5. Closed-loop control system design


TABLE 3
Topics Covered in Graduate Control Courses
(3 hr. lecture each week, per course)
FIRST COURSE
Derivation of models for multivariable systems
Formulation of state space models
Solution of multivariable system models (matrix
methods)
Modal analysis
Design of controllers using modal theory
Simulation and computer-aided controller design
Discrete systems analysis
SECOND COURSE
Sampled data systems
Sampled data controller design (or digital control
algorithms)
Decoupling control systems
Controllability, observability, etc.
Optimal control: quadratic and time-optimal
Analysis of stochastic systems
Observers, filters and state estimators
control systems for each of the laboratory bench
scale processes using: (1) theory only, (2) em-
pirically-determined process models, and (3) on-
line (loop-tuning) methods.
At the graduate level we presently offer two
courses in process dynamics and control. In recent
years the first course has covered both time and
frequency-domain methods with emphasis on state
space techniques used in conjunction with the com-
puter for analysis and design. The course deals
substantially with multivariable systems and, at
least in part, parallels the undergraduate courses
at an advanced level (Table 3).
The second graduate course covers advanced
control topics exclusively (also summarized in
Table 3). Although the graduate offerings are
traditional in nature (there are no laboratory ex-
periments at the graduate level), the rigorous yet
extensive coverage of advanced materials reflects
the same approach as at the undergraduate level.

THE REAL-TIME COMPUTING PROGRAM

O NE OF THE MAIN PROBLEMS in bringing new
material into the curriculum is that (usually)
an equivalent amount of old material will have to
come out. In developing the real-time teaching
program we began with four basic tenets:
1. Real-time computing instruction will be offered to both


WINTER 1980









undergraduate and graduate students on an elective
basis.
2. Real-time computing course work will supplement, not
replace, existing process control course work.
3. Real-time computing will be taught from a fundamental
point of view. Students will be expected to understand
basic hardware and software structures and how they
are used.
4. Lectures in real-time computing will be paralleled by a
"hands-on" laboratory with appropriately-designed ex-
periments.
We have followed these principles substantially
down to the present day; hence a few words of dis-
cussion might be appropriate: Tenet 1 arose out
of an early realization that many chemical engi-
neers, not just a few, will be involved with on-line


.... the ratio of "outsiders" to
chemical engineers was about 30/70 in
this second year; through the last academic
year it has been more like 70/30, with the
number of chemical engineering students
relatively constant at 12-15.


process computing as part of their professional
careers. Tenet 2 was based on a natural reluctance
to tamper with established control courses, in par-
ticular to remove some significant amount of ma-
terial so as to introduce lectures on real-time com-
puting. Tenet 3 might well be open to argument
but has its parallel in the controls area: twenty
years ago many people felt that control theory was
not chemical engineering; probably no one today
would argue in favor of a control course based
totally on an empirical approach. Our experience
with chemical engineering students is that they
do not, in general, like to spend time on computer
fundamentals; nevertheless those who do find that
practical applications are much easier to under-
stand and are able to transfer their knowledge to
other real-time systems much easier. Tenet 4 will
need no explanation.
Historically, we began our real-time course
offerings with a one-quarter graduate seminar in
Spring 1972 and followed it immediately in the
Fall of that year with a senior-level elective course
open to graduate students. Development of the
course and the associated laboratory were under-
written by the National Science Foundation
through two grants totalling almost $100,000
ver .a three and one-half year period. In the Fall
of i973 the course was elected by about twice as
many chemical engineering students (ten).
Several electrical: engineering students also took


the course; in fact the ratio of "outsiders" to
chemical engineers was about 30/70 in this second
year; through the last academic year it has been
more like 70/30 with the number of chemical
engineering students relatively constant at 12-15.
In the past three years we have had to restrict
enrollment because of the limitation in our real-
time laboratory facilities.
One of the interesting developments that came
out of the real-time computing course resulted
from the large number of student requests for se-
quel courses in the same area, but covering more
advanced topics. Many of the requests came from
electrical engineering or computer science students
who claimed that there were no equivalent appli-
cations courses within their own departments.
Additional requests came from some of our own
students, both undergraduate and graduate, who
planned to work in the process computer control
applications areas. The real-time field naturally
divides into three applications areas: (1) single
process/single computer, e.g. the topics covered
in our first course, (2) multiple processes/single
computer (multitasking or multiprogramming ap-
plications), (3) multiple processes/multiple com-
puters (multiprocessing or networking). So far as
the author can determine, it generally is the case
that the real-time instruction offered by most com-
puter science departments is (a) non-existent, (b)
concerned only with on-line systems, e.g. airline
reservation systems, (c) theoretically- rather than
practically-oriented, e.g. concerned with hypo-
thetical job scheduling problems in a multipro-
cessor environment. Condition (c) holds on our
campus; hence in order to accommodate student
requests we decided to add two additional courses
to our offerings to cover substantially multitask
programming and operating systems for real-time
applications, and networking and digital computer
control systems.
Several points are worth noting here concern-
ing the decision to expand the real-time computing
course to a full-year sequence:

* Some of the specialized computer-oriented material we
now teach is outside the area of expertise of most
chemical engineering faculty even though the applica-
tions-oriented material is not. We have avoided po-
tential problems somewhat by using a Teaching Associ-
ate, a Ph.D. candidate in computer engineering, to share
teaching responsibilities and to supervise the labora-
tory. In the four years we have offered the full se-
quence, several students working on joint research
projects involving the real-time laboratory facilities
have been supported financially in this way and have


CHEMICAL ENGINEERING EDUCATION









contributed significantly to the development of our
teaching and research program.
Attenuation of students enrolled in the sequence
historically was relatively high, running 40-50% per
quarter. Hence by the third quarter the enrollment
might have dropped from approximately 40 to about
10; most students continuing through the entire se-
quence have been our own graduate students, chemical
engineering undergraduates who have accepted jobs
involving a process control starting assignment, com-
puter science undergraduates, and undergraduate or
graduate electrical engineering students with an interest
in computer applications.
Chemical engineering students who have taken the real-
time sequence along with the required courses in dy-
namics and control have relatively little difficulty finding
employment in process-control-related areas. Several
process-oriented companies now recruit process control
engineers actively at Santa Barbara and, if statements
from recruiters can be believed, would have hired about
twice as many students for control work last year if
they had been available.
The mixing of chemical engineering students who have
relatively little computer background (in general only
experience in programming a higher-level language, i.e.
FORTRAN) together with computer science students
who have little or no experience with physical equipment
never was totally satisfactory. The distribution of abili-
ties in any particular prerequisite subject area is in-
variably bimodal: e.g. ChE students will have relatively
little background in binary arithmetic and logic (com-
puter science students will feel they have mastered the
subject); the reverse situation is true in the area of
physical measurements and measurement errors. This
situation led to major problems in the introductory
course where so much of the lecture material must
cover topics which will appear to be elementary to a
computer science major.

This year, for the first time, we have not per-
mitted computer science students to take the intro-
ductory course. The entire sequence has been re-
arranged somewhat to reflect these new develop-
ments. These actions represented an attempt to re-
turn the first course to what it originally was-
an introduction for chemical engineers. At the
same time we hoped to retain a reasonable enroll-
ment of "outside" students in the two following
courses. This hope did not grow out of any purely
altruistic motivations; rather the presence of
outside students furnished the department with a
claim on the additional teaching staff resources
necessary for us to offer such an extensive pro-
gram. Also, in a rapidly changing field such as
real-time computing, the presence of relatively
advanced computer science students in our courses
has kept the discussions lively and the lectures
more nearly "state-of-the-art." The success of
these changes is now apparent as will be noted in
the sequel. In any case, this rather lengthy descrip-


TABLE 4
A First Course in Real-Time Computing
(3 hr. lecture and 2 hr. lab per week)

Introduction to BASIC and to real-time BASIC
Structure of real-time systems
Measurements, transducers, and signal handling
Number systems and computer arithmetic
Introduction to computer architecture and
hardware
Input/output systems: ADCs and DACs
ISA FORTRAN
Device controllers and device drivers


tion of the development of the real-time sequence
is intended to motivate the description of the
present courses which follows immediately.

Real-Time Computing Courses. There is no
"traditional" first course in real-time computing;
we have, after much experimentation, settled on
coverage of the topics listed in Table 4. From the
table it can be seen that we spend considerable
time on computer fundamentals; number systems
and digital arithmetic, digital logic and hardware,
computer architecture, interfacing, assembly
language programming, interrupt handling, etc.
We also spend time on some topics which have
long since been dropped from most process control
courses; measurements and measurement errors,
transduction, signal transmission, etc. In a first
course of this sort the emphasis is on single
process/single computer systems and the coverage
must, unfortunately, be light. Our purpose is to
develop a basic understanding of all the elements
in a real-time system, how these interact, .and
how they comprise the whole.
Our purpose in teaching this course has not
been to treat real-time computing as an isolated
subject area but to teach it so that the material
can be integrated into the control courses, at least
into the undergraduate process dynamics and
control laboratory. Since the introductory real-
time course is taught in Fall quarter and precedes
the two-quarter sequence in dynamics and control,
students normally are in a position to make im-
mediate application. Those students who have
elected to take the real-time course are "permitted"
to run all of their dynamics (data logging) and
control experiments using one of the real-time
computers. Although this normally requires more
work of the student-outside reading, pro-
gramming, debugging programs, etc.-our ex-
perience shows that they often take this oppor-


WINTER 1980










FALL WINTER SPRING


INTRODUCTION REAL-TIME a REAL-TIME
TO REAL-TIME COMPUTING COMPUTING
COMPUTING I II
(NON EE/CS)
o LABORATORY LABORATORY LABORATORY

PROCESS DYNAMICS PROCESS DYNAMICS
AND CONTROL AND CONTROL













graduate process control program, also how the
three different groups of students-undergradu-
ate and graduate chemical engineers and "out-of-
department" (C.S., E.E., etc.) -can be accommo-
dated.
The remaining two courses in the sequence
(labeled "Real-Time Computing I and II" in
Figure 1) cover the major areas of multitask andRED
LABORATORYLABORATORY

ADVANCED ADVANCED
PROCESS DYNAMICS PROCESS DYNAMICS
3 AND CONTROL AND CONTROL









FIGURE 1. Process Dynamications, respectively, with an
emphasis on appliaomputions of computers, eitherferings.








singly or in networks, for control purposes. Tables
t5 and 6 furnish a brief description how the real-time course
contend for each of these courses; thergraduate necessarily
gradmust be procea small degree of repetition to how bring
three different groups of students-undergradu-
ate and graduate chemical engineers and "out-of-
department" (C.S., E.E., etc.)--can be accommo-
dated.





entering students up remaining to operates in the sequence
(labeled "Real-Time Computing Laboratory. The present
real-time 1) laboratory (shown in Figure 2 wmultitah one
multiprocessor applications, respectively, with an
emphasis on applications of computers, either
singly or in networks, for control purposes. Tables
5 and 6 furnish a brief description of the course
content for each of these courses; there necessarily
must be a small degree of repetition to bring
entering students up to operating speed.

Real-Time Computing Laboratory. The present
real-time laboratory (shown in Figure 2 with one


of the undergraduate process dynamics and
control experiments visible in the background)
contains three minicomputers, two of which are
configured for real-time operations. These facili-
ties, built up over the past ten years, will be sub-
stantially replaced in early 1980 by the single-
computer system shown schematically in Figure
3. This multiprogrammed system will accommo-
date up to six real-time user programs in main
memory simultaneously and, potentially, can be ex-
panded to handle many more if program swapping
using the fast disk can be tolerated. Features of
the new system will include a link with the main
campus computer, full graphics capabilities, a dial-
up facility for one remote user, and one or more
terminals in remote study rooms and laboratories.

TABLE 5
Second Course in Real-Time Computing
(2 hr. lecture and 2 hr. lab per week)
Overview of real-time computing
Introduction to real-time FORTRAN
Analog and digital input/output
Operating systems and schedulers
Introduction to multitask programming
Multitask program design
File handling and bulk storage
Assembly language device driver routines
Multitasking applications

As part of the instructional laboratory we have
constructed several interesting auxiliary units:

* a set of input and display panels for experiments in-
volving input, output and conversion of analog and
digital quantities;
an air pressure experiment with binary inputs (solenoid-
operated valves) and outlays (pressure-operated) relays
for instruction in digital 1/0 and simple on-off control;
* a metal bar heated at one end, with eight temperature









FIGURE 2.
The Real-Time
Computing Laboratory
(Foreground).


.... .. ,.-' .. .


CHEMICAL ENGINEERING EDUCATION








TABLE 6
Third Course in Real-Time Computing
(2 hr. lecture and 2 hr. lab per week)
Real-time computers in process control
Controller and filtering algorithms
Controller design and applications
Overview of computer networks
Network architectures
Interprocessor communications
Distributed processing
Networks in process control
sensing elements located along the unheated section for
multipoint data logging studies;
a fully interfaced model railroad designed to demonstrate
the control of multiple, largely-random processes.
A list of experiments which typically would be
performed as part of the introductory real-time
course is given in Table 7. The model railroad is
used as the basis of a sequence of five experiments
in the second course. In the third course, several
of our present computers as well as the new system
will form the basis for the networking portions
of the course. The stirred-tank heating systems are
used for the process control portions. Students
completing the laboratory sequence will, as a
final project, put together a two-computer real-
time system (one computer for data acquisition
and control, the other for process operator com-
munications and report generation) with inter-
processor communications carried out over an
existing multiprocessor bus.

SUMMARY AND CONCLUSIONS
CHEMICAL ENGINEERS WHO plan to work closely
with digital computer control systems need a
much more fundamental exposure to real-time
systems principles than can be obtained through
a brief exposure as part of a senior-level control
course. Even a single-quarter course in real-time
computing cannot cover important advanced
topics in the field such as real-time operating
systems, multitasking, multiprogramming, and
TABLE 7
First Course in Real-Time Computing:
Laboratory Experiments
* Calibration of a resistance thermometer for a stirred
tank heating system
* Estimation of dynamic measurement error in the
stirred tank temperature transducer
* Automated number conversions
* Digital input/output: "super pong"
* Analog input/output: simulation of a staircase ADC
* Data logging of the heated bar temperature profile
a Data logging and control of the pressure tank


... in a rapidly changing field such
as real-time computing, the presence of
relatively advanced computer science students
in our courses has kept the discussions
lively and the lectures more
nearly "state of the art."


FIGURE 3. The UCSB Real-Time Computing System.

networking. At Santa Barbara we have expanded
our offerings to a three-quarter sequence with
heavy emphasis on laboratory exercises and ex-
periments. As a service course, the sequence
attracts enough outside students to warrant aug-
mented teaching support staff from the college.
Also, chemical engineering students appear to
benefit from the experience of working with com-
puter specialists; still it is clear that mixing them
at too early a stage is not optimum. The move we
have made to restrict enrollment to chemical engi-
neers in the introductory class has eliminated most
of the problems arising from mismatches in basic
skills. This year, considerably more of our own
students elected to take the first course than have
in the past, and more of them are continuing in
the sequence. Additionally, the teaching loads over
the entire academic year have been considerably
smoothed out by closing the first course to EE and
CS students.


WINTER 1980








TEXT MATERIALS
Before 1977-78 there was available virtually no single
source of information which could be used as a text in a
course on real-time computing. In 1977 the first volumes
of the CACHE Monograph Series in Real-Time Computing,
edited by this author, appeared. The monograph series, an
attempt to produce a definitive treatment of each major
area in the field, will consist of eight volumes initially. All
of them will be available in 1979. A listing of the titles in
the present series:
I. An Introduction to Real-Time Computing
II. Processes, Measurements, and Signal Processing
III. Introduction to Digital Arithmetic and Hardware
IV. Real-Time Digital Systems Architecture
V. Real-Time Systems Software
VI. Real-Time Applications Software
VII. Management of Real-Time Computing Facilities
VIII. Process Analysis, Data Acquisition, and Control
Algorithms
In 1978 a good book dealing with industrial applica-
tions of real-time computers was published: Minicomputers
in Industrial Control, T. J. Harrison (Editor), Instrument
Society of America, Pittsburgh (1978).
With respect to the laboratory facilities, the author
earlier documented the three real-time laboratory experi-
ments developed at Santa Barbara in reports, a few copies
of which are still available:
D. A. Mellichamp and F. Kayihan, The Tank Pressure Ex-
periment, UCSB Department of Chemical and Nuclear
Engineering Report C-74-1, 79 pp, (August 1974).
D. A. Mellichamp and G. P. Engelberg, The Digital Com-
puter Controlled Model Railroad, UCSB Department of
Chemical and Nuclear Engineering Report C-74-3, 119 pp,
(October 1974).
D. A. Mellichamp and T. W. Moore, The Heated Bar Ex-
periment, UCSB Department of Chemical and Nuclear
Engineering Report C-76-1, 96 pp, (March, 1976). O


book reviews

CHEMICAL REACTOR DESIGN FOR PROCESS
PLANTS; VOLUME I, PRINCIPLES AND
TECHNIQUES; VOLUME I, CASE STUDIES
AND DESIGN DATA
By Howard F. Rase
Wiley-Interscience, New York, 1977
Reviewed by Charles H. Ware, Jr.,
Commercialization Insights, Poughkeepsie, N.Y.

The author has written this book, as the pre-
face states, "for the professional engineer who
either daily or periodically must deal with design
or operation of chemical reactors. But in addition
to serving as a reference in the personal libraries
of professionals, it should also be useful as a text-
book for advanced design courses, including
courses taught in continuing education." It will
serve all of these purposes very well.


Volume I (772 pages) is divided into four
parts: basic data and principles of design; general
aspects of reactor design; single-phase reactors;
and design of reactors for multiphase processes.
Volume II (242) pages consists of 14 case studies
including three oxidation reactions, two polymeri-
zations, and two hydrogenations.
Part 1 is devoted to reaction rate theory and
applications; chemical and physical aspects of
catalysis and catalysts; idealized models of re-
action rates and reactor performance; and ex-
perimental methods and equipment for developing
design data. Two chapters devoted to catalysis
and catalysts provide a good summary of them
with attention to both theoretical foundations and
practical considerations. Experimental methods
and equipment to obtain chemical reaction data
free of transport effects are emphasized.
Part 2 is concerned with selection of reactor
type and mode of operation based upon yield and
safety, as well as general design considerations
such as mixing of reactants, flow distribution,
residence-time distribution within reactors, and
briefly, vessel design.
Part 3, which comprises almost half of the
text in Volume I, covers the design of CFSTRs,
tubular, batch, semi-batch, fixed-bed catalytic,
fluid-bed catalytic, and many special reactors. In
addition to the various design equations, there are
numerous drawings of actual reactors and con-
siderable attention is given to flow and heat effects,
feed systems, pressure drop, scale-up, start-up and
shutdown procedures.
Part 4 consists of an excellent chapter on gas-
liquid reactors plus a short account of liquid-liquid
reactors. In the former, stirred tanks, sparged
vessels, plate and packed columns, trickle beds, and
pipeline contractors are considered. Many theo-
retical and practical aspects are discussed: scale-
up, heat transfer, power consumption, pressure
drop, design models and procedures, hold-up, mass
transfer, dispersion, liquid distribution, and many
others. Only the agitated reactor is treated in the
last chapter.
The case studies of Volume II have been se-
lected to illustrate various types of design
problems. They are indicated in each case, often
accompanied by a comment on the principal
weakness of the design. The data that are needed
are presented, along with intermediate results,
alternatives, and bases for decisions.

Continued on page 47.


CHEMICAL ENGINEERING EDUCATION





"At Du Pont you don't get lost


in a big company atmosphere.


It's very personal:


-George D. Peterson


BS, Chemical Engineering

I


"Du Pont is a big com-
pany but it's broken down into
satellites. So you don't get lost
in a big-company atmosphere.
It's very personal, and I think the
people are top-notch.
"I started in technical
here at the Belle Plant in West
Virginia. Now I'm a production
supervisor. Production is solv-
ing problems on a day-to-day
basis. I like working under that
kind of pressure. When things


work out, it's very rewarding. So
is working with people. I'm
responsible for helping 22 peo-
ple do their jobs:'
George was recruited by
Du Pont from the Michigan
Technological University
campus in 1973. He interviewed
about 25 companies.
George's story is typical
of many Chemical, Mechanical
and Electrical Engineers who've
chosen careers at Du Pont.


We place no limits on
the progress our engineers can
make. And we place no limits
on the contribution they can
make-to themselves, the
Company or to society.
If this sounds like your
kind of company, do what
George Peterson did. Talk to the
Du Pont representative who
visits your campus. Or write:
Du Pont Company, Room
35972, Wilmington, DE 19898.


At Du Pont...there's a world of things YOU can do something about.



RE US PrTaR OFF
An Equal Opportunity Employer. M/F










-M 3 laboratory


ADVANCED PROCESS CONTROL EXPERIMENTS


PRADEEP B. DESHPANDE,
W. L. S. LAUKHUF and
NANDKISHOR G. PATKE
University of Louisville
Louisville, KY 40208


N A COURSE ON ADVANCED process control a sub-
stantial portion of the time is spent in discussing
the fundamentals, design, and implementation of
advanced control concepts. When the course was
first offered several years ago, computer simula-
tions were used to demonstrate the concepts. While
simulation is certainly a very valuable tool in the
analysis and design of control systems, the
students felt that it would have been much more
satisfying if they had a physical process to work
with. In subsequent years, several laboratory ex-
periments were developed to eliminate this de-
ficiency.
The equipment for the process, around which
the present experiments were developed, was con-
structed from data provided by Exxon Oil
Company (then Humble Oil & Refining Co.) on an
identical setup at their Bayway refinery. [1] The
rig was used by Exxon to train their instrument
and process personnel. It has four of the most


FIGURE 1. Schematic of Process Control System (Arrows
indicate Signal transmission between the process and
the computer).


It has four of the most
commonly encountered control
loops in process industry, i.e., liquid-level,
flow, pressure, and temperature.


commonly encountered control loops in process
industry, i.e., liquid-level, flow, pressure and
temperature. The equipment was used to demon-
strate feedback control concepts for these loops.
Additional instrumentation has been added to the
apparatus at the University of Louisville in order
to demonstrate advanced control concepts.

EQUIPMENT & INSTRUMENTATION
A schematic of the thermal process unit is
shown in Figure 1. The process involves heating
of a continuous stream of water by steam. A
vertical cylindrical tank approximately one foot
in diameter is located in the center of the unit.
The tank contains a steam pipe in the form of a
vertical U tube. Water flows continuously in and
out of the tank where it is heated by steam.
As shown in Figure 1, the process is instru-
mented with conventional controllers as well as
with computer control hardware. Three variables
can be controlled in this process: the flow rate of
water into the tank, the level of water in the tank,
and the temperature of water in the tank.

FLOW CONTROL LOOP
This loop regulates the flow of cold water into
the tank. Supply water passes through a 7/16-
inch diameter orifice mounted in a 1/2-inch pipe,
then through a 1/2-inch control valve made by
Uniflow Valve Corporation, and into the top of the
tank. The differential pressure across the orifice
is transmitted to a mercury manometer (FI in
Figure 1) and to a Honeywell flow indicating
transmitter (FT). The pneumatic 3-15 psig out-
put of the transmitter is fed to a flow recording
controller (FRC). It is also fed to an AMTEK
pneumatic to voltage (P/E) transducer. The
electrical output of the P/E transducer is con-


CHEMICAL ENGINEERING EDUCATION


Copyright ChE Division, ASEE, 1980









nected to one of the analog-to-digital (A/D) con-
verter channels of the control computer. As indi-
cated in Figure 1, the position of a Foxboro air-
switch determines whether the transmitter output
is fed to the conventional controller or to the
control computer.
The conventional flow recording controller is
a Honeywell, proportional + reset type, controller
which sends a signal to an "air-to-open" control
valve on the process unit. Alternately, the signal
to the control valve may also come from one of the
digital-to-analog converter channels on the control
computer, via a Fisher E/P transducer. Again,
the position of an air switch determines whether
the signal to the valve comes from the conventional
controller or from the control computer.

LIQUID-LEVEL CONTROL LOOP

Water level in the tank is controlled by
manipulating the flow of water out of the tank.
The level sensor infers the liquid level by measur-
ing the pressure required to cause air bubbles to
form slowly at the bottom of the tank. This pres-
sure signal is fed into the high pressure side of a

Pradeep B. Deshpande is associate professor of ChE at the Uni-
versity of Louisville. Prior to coming to Louisville, he was with
Bechtel, Inc., at San Francisco, CA. He has served as a consultant to
Mobil Exploration Norway, Inc., and to other companies in the areas of
control systems design and simulation. He has a Ph.D. in ChE from
the University of Arkansas and is a registered control systems engi-
neer in California. (L)
Walden L. S. Laukhuf is an associate professor of ChE at the Uni-
versity of Louisville. Prior to coming to the University, he was a
Captain in the Air Force at the Air Force Materials Laboratory in
Dayton, OH. He has a BChE, MSChE and a PhD in ChE from the
University of Louisville and is a registered professional engineer in
Kentucky. (C)
Nandkishor G. Patke is presently working on a Ph.D. in ChE at
the University of Louisville. His research interests are process model-
ing, simulation and control. He has a B.Tech. in ChE from the Indian
Institute of Technology, Kanpur, India. (R)


Load, L(s)
load transfer
steam GI'5l' function
I- -vove process
poantt --[- lt[).--- Go C temperature
L- -_ .. J__



measuring
element
FIGURE 2. Block Diagram of the Temperature Control
System

Foxboro differential pressure transmitter (LLT),
set at a range of 40 inches of water. The low
pressure side of the transmitter is vented to the
atmosphere as is the surface of the liquid in the
tank. Thus, the differential pressure transmitter
output is proportional to the liquid level. The
signal from the transmitter is fed to a Honeywell
proportional + reset controller (LLRC) and to
one of the A/D converter channels via an air
switch which determines whether the loop will be
on conventional control or on computer control.
The outputs of the controller and a D/A con-
verter channel are fed to an air switch and then
to a 3/4-inch, "air-to-close" control valve made by
the Uniflow Valve Corporation, installed in the
drain line from the tank. The air switch selects
computer control or conventional control. A 1/4-
hp, Barray pump in this line insures sufficient fluid
pressure on the upstream side of the valve.

TEMPERATURE CONTROL LOOP

The temperature of water near the bottom of
the tank is measured by an iron-constantan
thermocouple immersed in an oil-filled well ex-
tending into the bottom portion of the tank. The
voltage produced by the thermocouple is converted
by a Honeywell electropneumatic transducer (TT)


. Ll
'0


WINTER 1980








101


S93

a
S85


77
77


output



10
-input--


0 4.8 8.8 12.8 16.8


TIME(MIN)
FIGURE 3. Input/Output Records From Pulse Test

into a pneumatic signal. This air signal is fed to a
Foxboro three-mode controller (TRC). When con-
ducting cascade control experiments, this con-
troller serves as the master controller. The trans-
mitter output is also fed to an A/D converter
channel via an air switch. The output of the three-
mode controller is fed to the set-point input of the
Honeywell, proportional + reset, pressure record-
ing controller (PRC). In cascade control experi-
ments PRC serves as the slave controller.
The coil-side steam pressure is fed to a Honey-
well transmitter. The output of the transmitter
is fed to the pressure controller. The output of the
controller operates a 1/2-in. Uniflow pressure
control valve. The signal to the valve may alter-
nately come from a D/A converter channel of the
control computer.
In conventional control experiments, cascade
control is achieved if both, master and slave con-
trollers, are placed in automatic. If the pressure
(slave) controller is switched to manual, the in-
put to the control valve comes from the tempera-
ture controller. The temperature control loop is
then a simple closed-loop rather than a cascade
system.
The control computer used in some of the ex-


FIGURE 4. Frequency Response Diagram
Frequency co, Radians/min.


periments is a PDP 1103 microcomputer system
manufactured by the Digital Equipment Corpora-
tion. It has 24K words of memory and is equipped
with an 8-channel D/A converter and a 16-channel
A/D converter. The computer comes with a dual
disk drive. A floppy disk, containing systems pro-
grams (e.g. Fortran support programs, real-time
subroutines) resides in one of the drives while a
second floppy disk, containing user-developed
control programs, resides in the other drive. Com-
munication with the computer is via a teletype-
writer (LA 36 DECWRITER).

EXPERIMENT 1: Process Identification
This experiment is concerned with dynamic
identification of an open-loop process by pulse test-
ing. The resulting information is used to find (1)
suitable tuning constants for a feedback controller
or (2) to develop an approximate process model
which is useful in designing advanced control
strategies. The pulse testing technique [2, 3] has
been applied to the temperature control loop whose
block diagram is shown in Figure 2.
The input and output records from the pulse
test are shown in Figure 3. Numerous data points
from these records are entered into a computer
program [3, 4] which generates frequency response
data as shown in Figure 4.
From Figure 4, the ultimate AR, which refers
to the amplitude ratio for which the phase lag
equals 180 degrees, is 0.0615. Also, the crossover
frequency, which is the frequency corresponding
to the phase lag of 180 degrees, is 1.1 radians
per minute. Therefore, the ultimate gain, Ku, and
the ultimate period, Pu, are


1
Ku =
0.0615


16.23 psi/F


27r
Pu = = 5.71 min.
1.1
Since the gain of the transmitter, K,, is 0.06
psi/F, the Ziegler-Nichols tuning constants for a
PI controller are

Gain, Ke = 0.45 Ku/KT = (0.45) (16.23)/0.06
= 121 psi/psi
Integral Time, 7T = Pu/1.2 = 5.71/1.2
= 4.75 min. (2)

To assess the adequacy of the controller settings
found in this section, a closed-loop control experi-
ment was conducted. The response of the system
to a step change in set point and load is shown in


CHEMICAL ENGINEERING EDUCATION








Figure 5. This plot shows that the tuning
constants found through pulse testing are ade-
quate.

EXPERIMENT 2: Multivariable Control
Most large processes have many controlled
variables and many manipulated variables. Ideally,
a change in a given manipulated variable should
affect only its own controlled variables and no
others. Unfortunately, in many cases, this is not
the case. The interaction among different loops can
lead to poor control and even instability.
Since interaction can be a problem in multi-
variable control systems, it is important to know
the extent of interaction and to be able to develop
criteria for proper pairing of manipulated and
controlled variables.
A measure of the extent of interaction in multi-
variable control is obtained by Bristol's method
[5]. The method is based on steady-state input-out-
put relationships for the process. It yields a
measure of steady-state gain between a given
input-output pairing. By using the most sensitive
input-output connections, interaction is mini-
mized.
Since Bristol's method does not take systems




-104



S1 -
96


0 4 8 12 16
TIME MINUTES
FIGURE 5. Transient Closed-Loop Response to (a) Set
Point Change (b) Load Change

dynamics into account, it would be very useful to
evolve an experiment which assesses the beneficial
effects of proper pairing upon the dynamic re-
sponse of the multivariable system. The present
experiment [6] is designed to accomplish this ob-
jective.
The hardware for this experiment is essentially
that shown in Figure 1 with the exception that
the steam line is replaced by a pipe which intro-
duces hot water into the tank. The air switches
must be in the computer control position for this
experiment.


Since interaction can be a
problem in multivariable control systems,
it is important to know the extent of interaction
and to be able to develop criteria for
proper pairing of manipulated
and controlled variables.

The process objective is to control the level (in
effect, total flow) and temperature of water in
the tank. There are two inputs to the process,
namely, the flow of cold water and the flow of hot
water into the tank. So, the controlled variables
are temperature and total flow and the manipu-
lated variables are cold water flow rate and hot
water flow rate. The question is, should the
temperature be controlled by manipulating hot
water flow and level (i.e. total flow) by cold water
flow or vice versa? Bristol's method provides the
answer.

Bristol's Relative Gains Analysis
The functional steady-state relationship be-
tween temperature, total flow and the flow
streams is


T = f(m, m,) = (meT + mhTh) /mt

mt = f(me, mh) = m + in


Around some steady-state operating
relationships can be expressed as

AT = Am + TAm,
am- Ambh
= K, Am, + K2lAm,


point, these


Amt A-m e + Am a Amn

= K21 Ame + K2,2Amh
The K's are the open-loop steady-state gains which
quantitatively describe how the m's affect T and
mt. They can be determined from a mathematical
model of the process or by experimental step or
pulse-testing on the plant. To evaluate K1l and
K21 for example, a small change in the flow of cold
water is made, while the process is operating
under steady state conditions (under manual
control with the flow of hot water maintained
constant). When the temperature and level reach
their new steady-state values, K11 and K21 can be
evaluated by
SAT
Ki1 Am mh constant (5)


WINTER 1980










K21 = l( -T
K (6)
\ Ari ) h = constant (
The gain K,:, then, determines the change in
temperature, T, due to a change in me when mh
is held constant. Now, suppose instead of holding
m, constant, while a small change in me is being
made, mh is manipulated so as to bring mt back to
the original value it had before the change in m,
was made. Then, another gain Anl can be defined
as

An -A
AI A(me )mt = constant (7)

An, is a measure of how me affects temperature T,
if level were under closed-loop control (i.e. held
constant). The ratio of Ku, to An is called the
relative gain X,,. Thus,
K11 ('AT/Ame) h = constant
S All (AT/Ame)mt = constant
By comparing the relative gains for each manipu-
lated variable, it is possible to assess which m
has the most effect on a given controlled variable


The equipment for the process,
around which the present experiments
were developed, was constructed from data
provided by Exxon Oil Company... on an
identical setup at their Bayway refinery.


and therefore how to pair the manipulated and
the controlled variables.
While K's can be determined easily, the experi-
mental determination of A's is not so easy. How-
ever, they can be evaluated from the K's as
follows:
By definition

A,, = (AT
\ Am, = constant


The open-loop relationships (Equation
come
Am = = = K21Am, + K22Amh
Thus,

Amh K -- Am,
K22
Also in view of Equation (4)
K,2K21
AT = KnAm, K Am,
KJ22


S- Incorrect pairing


time (minutes)
FIGURE 6. Transient Response of Level


Therefore,


and,


AT = KliK22 K1,K Ame
K22 m


SA t constant
HAmm = constant


(11)


KxiK22 K12K21
K22
(12)


The relative gain XL is then

KA, KlIK22
All ;KliK22 K12K21


(13)


Similar analysis yields the remaining relative
gains. Thus,


S K12K21
K12 K12K21 KlnK22
S K12K21
2 12K21 Kll K22
2__2 K11K22
-2 11K22 K1K


(14)

(15)

(16)


(4) be- To facilitate the pairing of manipulated and
controlled variables, it is convenient to present the
relative gains in a matrix form as shown in Equa-
9 tion (17).


min mh
T X, 112
mt A21 A22


(17)


For each controlled variable, the manipulated
variable selected is the one which has the largest
positive relative gain. Since a property of this


CHEMICAL ENGINEERING EDUCATION








matrix is that each row and column sums to one,
only one X need by explicitly computed in a 2 x 2
system.

Results

The relative gains matrix for the current pro-
cess is shown in Equation (18).


me mh
T me m__
mnmt mt
It mh me
Smt nTt


me mi

T 0.172 '0.828 (18)


mt 0.828 0.172


This equation shows that: T should be controlled
by manipulating mh and mt by manipulating me.
Both loops use a proportional + integral control
algorithm on the digital computer as the control
element. The algorithm was tuned by trial and


interaction ("fighting loops") would have been a
problem, particularly if the response times of the
two loops were comparable. Severe cross-coupling
can drive the multivariable system to instability.
In such cases decoupling will be required. Inter-
ested readers may consult reference 7 to obtain
further information on the various techniques
currently available for decoupling a multivariable
control system. O

NOMENCLATURE


AR
A's
G,(s)
K,
Ke

Kp
K's
Ku
me
mh
mt
Pu
T
To
Th

Greek


amplitude ratio
closed-loop gains
process transfer function
temperature transmitter gain, psi/F
proportional gain on temperature controller,
psi/psi
steady-state gain of process, F/psi
open-loop gains
ultimate gain, psi/F
cold water flow, lb/hr
hot water flow, lb/hr
total flow m, + mn, lb/hr
ultimate period, min.
temperature of the mixture, F
temperature of cold water, F
temperature of hot water, oF


phase angle
process dead-time, minutes
time constant, minutes
integral time, minutes
relative gain
frequency, radians/minute


10 20 30 40
time (minutes)
FIGURE 7. Transient Response of Temperature

error. The steady-state operating conditions were:
level set point, 50 % (which corresponded to total
outlet flow of 11.6 lit/min) ; temperature set
point, 24.4C; cold water flow, 9.61 lit/min; hot
water flow, 1.99 lit/min. The process was operated
with correct pairing as well as with incorrect
pairing. The benefits of proper pairing are clearly
evident in the set-point responses shown in
Figures 6 and 7. These results show that Bristol's
approach is a simple and powerful tool in the
control systems design of multivariable processes.
If the relative gains in Equation (18) had
turned out to be numerically close to each other,


REFERENCES CITED
1. Alper, William S., A Laboratory Demonstration of
Closed-loop Automatic Process Control, Master of
Chemical Engineering Thesis, University of Louisville,
1963.
2. Hougen, Joel 0., Experiences and Experiments with
Process Dynamics, Chemical Engineering Progress
Monograph Series, Vol. 60, No. 4, 1964.
3. Luyben, W. L., Process Modeling, Simulation and
Control for Chemical Engineers, McGraw-Hill Book
Co., New York, NY, 1973.
4. Dynamic Response Testing of Process Control Instru-
mentation, ISA-S26 Standard, Instrument Society of
America, 400 Stanwix Street, Pittsburgh, PA 15222,
October 1968.
5. Bristol, E. H., On a New Measure of Interaction for
Multivariable Process Control, Trans. IEEE, 1966.
6. Knabel, E. A., Multi Variable Control System for
Fluidized Bed Coal-gasification Units, Master of Engi-
neering Thesis, University of Louisville, 1978.
7. Foss, A. S., Denn, M. M. Ed., Chemical Process Control,
A.I.Ch.E. Symposium Series, No. 159, Vol. 72, 1976.


WINTER 1980










Classroom


THE INTEGRATION OF REAL-TIME COMPUTING INTO


PROCESS CONTROL TEACHING


PART II: THE UNDERGRADUATE COURSE*


M. MORARI and W. H. RAY
University of Wisconsin
Madison, Wisconsin 53706


P ROCESS CONTROL FOR undergraduates is offered
each semester and has enrollments of 30 to 50
students. The course has two lecture periods, one
recitation period, and one 3 hour laboratory period
each week. The curriculum (shown in Tables 1
and 2) has been selected to give the student a
balanced mixture of useful theory and hands-on
practical experience in process dynamics, measure-
ment, and control.

COURSE DESCRIPTION

T HE INSTALLATION OF A PDP11/55 minicom-
puter system described in Part I of this article
has allowed a complete restructuring of the course
material. A large library of computer programs
for the design of control systems is available and
is still growing. The use of these programs makes
it possible to design control systems for meaning-
ful practical processes without the drudgery of
laborious hand calculations. Thus, course time is
freed and it becomes possible to cover topics
generally neglected in undergraduate courses,
TABLE 1
Undergraduate Process Control Lecture Topics
1. Review of Laplace transforms and matrix algebra
2. Principles of real-time computation and data acquisi-
tion
3. Transient and frequency response of linear systems
4. Feedback control of linear systems
5. Stability of linear systems
6. Control system design for linear systems
7. Nonlinear systems
8. Case studies

*Part I dealing with graduate education in process
control appeared in the Fall 1979 issue of CEE.


TABLE 2
Undergraduate Process Control Laboratory Experiments
1. Techniques of analog simulation
2. Techniques of digital simulation
3. Dynamics of interconnected water tanks
4. Computer aided data acquisition
5. Frequency response and process identification through
pulse testing
6. Calibration and dynamic response of PID controllers
7. Feed forward, feedback, and cascade control
8. Multivariable control of a gas distribution system
9. Multivariable control of a multi-sidestream distilla-
tion column
10. Tuning of a level controller with strong system non-
linearities
e.g. multivariable control. The lecture material is
listed in Table 1 and includes considerations of
how to choose loop pairings, how to minimize
interactions between control loops and how to tune
multivariable systems. In the latter part'pf the se-
mester a number of graphical interactive com-
puter aided design programs are used for detailed
case studies. The present library includes routines
for the generation of Bode plots, root loci and




DESIGN 00EL
I) G S)=CtiPltS'I 'P2*tS+l)-T3tS+I)(T4*S+)(XT5tS+I)
OR ,
-2) G(S)[>= I )+A tS+ ....A(N)tSI*-H- > I ) B( I )B(2)S+.....B(M)tS*St(M-I)
GC,.S = t(l TOSS 1/TItS)
WHICH WOULD YOU LIKEEI OR 23?1 CONTROLLER
INPUT C = 1 1) P 2) PD 3) PI 4) PID
INPUT PI =8 TYPECE TO 43?1
INPUT P2 =1
INPUT T3 =2
INPUT T4 =-
INPUT T5 =5
INPUT TINE DELAY..B
UP TO WHNT UALUES OF K MOULD YOU LIKE?1B
MOULD YOU LIKE THE ROOTS LISTED ON THE LINE PRINTER?[Y/N3N
FIGURE 1. Input Data for Root Locus Program: Process
(s + 1)
G(s) = is under Propor-
(2s + 1).(8s + 1))(5s + 1) is under rop
tional Feedback Control.
Copyright ChE Division, ASEE, 1980


CHEMICAL ENGINEERING EDUCATION









Nyquist diagrams for open loop, closed loop and
cascade control systems. In addition, a number of
programs for the design of multivariable systems
are presently available. Output from these pro-
grams usually appears as plots drawn on the
screens of graphic terminals, and paper copies (to
be included in student reports) are obtained by
the student at the push of a button. Example: if
a root locus diagram is to be drawn the user
would be asked to specify the transfer function
parameters and other variables as shown in Fig.
1. Upon completion of the questions a diagram (as
in Fig. 2) will appear which can be subsequently
enlarged or otherwise modified.
The laboratory, which is designed to comple-
ment the lecture material, is comprised of some





.1=




."



4.



Manfred Morari was born in Graz, Austria on May 13, 1951. He
obtained his undergraduate education in chemical engineering at the
Swiss Federal Institute of Technology (ETH), Zurich. After his diploma
he started graduate school at the University of Minnesota in 1975.
Upon completion of his doctorate he joined the ChE faculty at the
University of Wisconsin in 1977 where he is currently assistant pro-
fessor. Last summer he worked for Exxon Research and Engineering
Company. His research interests include a variety of topics from the
areas of process synthesis and process control: synthesis of separa-
tion sequences, optimal measurement selection and inferential control,
optimizing control and the dynamics and control of large integrated
processing systems. (L)

W. Harmon Ray was born in Washington, D.C., on April 4, 1940.
He received the B.A. and B.S. Ch.E. degrees from Rice University,
Houston, Texas, in 1962 and 1963 respectively, and the Ph.D. degree
in ChE from the University of Minnesota in 1966. He has been on
the faculty of the University of Waterloo in Canada (1966-70), the
State University of New York at Buffalo (1970-76), and the Uni-
versity of Wisconsin, Madison, where he is presently Professor of
ChE. During the 1973-74 academic year, he was on sabbatical leave as
a Guggenheim Fellow in Belgium and Germany. His research in-
terests include chemical reactor engineering and process modelling,
optimization, and control. His publications include an edited volume
"Distributed Parameter Systems" (Dekker, 1977), and two monographs
"Process Optimization" (Wiley, 1973) and "Advanced Process Control"
to be published by McGraw-Hill in 1980. (R)


The use of these programs makes
it possible to design control systems for
meaningful practical processes without the drudgery
of laborious hand calculations. Thus, course
time is freed and it becomes possible to
cover topics generally neglected
in undergraduate courses


ten experiments (cf Table 2). Most of these ex-
periments are carried out by each laboratory
group in the course of the semester. Many experi-
ments involve real time computation and are se-
lected to familiarize the students with the modern
methods of implementing control algorithms.
These presently include:
Data acquisition (noise suppression, signal amplifica-
tion, A/D conversion, sensor calibration, etc.)
Pulse testing (data acquisition, input pulse selection,
Fourier transformation of data, frequency spectrum
analysis, frequency response parameter determination,
etc.)
Multivariable feedback control of interconnected gas
storage tanks (process modelling, data acquisition,
single loop PI control, supervisory computer control,
direct digital control, etc.)
Multivariable feedback control of a multi-side-stream
distillation column (process modelling, data acquisition,
single loop control, supervisory computer control, and
direct digital control)
Let us discuss two of these experiments in more
detail.

Pulse Testing

T HIS EXPERIMENT CONSISTS of putting a
measured pulse of hot water into a stirred
mixing tank having continuous inflow and out-
flow. Input and output temperatures are measured
under computer control and the resulting data
analyzed to provide frequency response informa-
-6eee6eee6e6e6eee66 es6e6asea ROOT LOCUS eeePeet eepeleleetett
P= 6 6O P2= 1. T3i 2 68 TINE DELATY 0.00
T4= 666 T5= 5668 C- 16, 0.
TDO= 0.@ I/TI= 8.6e K= 8.06 0 446E-06
CONTROLLER D
K ( P )
A= e.41
8= 2 45
C' 4 59
6 72
E= 9.92


OC a A


-e.665E+90


-8 333E.09


FIGURE 2. Root Locus Diagram for the
Control Loop in Figure 1.


e ee6E+*e


a8.earEBB -. 446E+B9


Feedback


WINTER 1980










tion and parameters for a process model (i.e., a
gain and time constant for this simple first order
process). Typical results, taken from a student
lab report, are shown in Figures 3 and 4. The
measured input and output temperatures are
shown in Figures 3 while the resulting Bode plot
showing the frequency response may be seen in
Figure 4. The process gain of 1.0 is readily found
from the low frequency asymptote of the ampli-
tude ratio (AR). By using both amplitude ratio
(AR) and phase angle (0) to estimate the corner
frequency, w two separate estimates of the tank
time constant are found. Usually these are in
reasonable agreement with the "theoretical" value
determined from the mean residence time of the
tank.

Multivariable Control of Interacting Gas Storage Tanks

One of the most sophisticated experiments
carried out by the students is the modelling and
multivariable feedback control of a pair of inter-

'T l'r r r T n Tr T n PULSE RESPONSE *t *t*S***lrtutt
I : IPUT TEMIP
2 OUTPUT TEMP


0 12E+03 02
e. 1 +03
Temp. tF)
0.-0E+ T lnt




a 6WME+02






O.E8MS+-3 0.531E+G3
0


*tT lttW* *tt itttrttltl* BODE PLOT


i~s:



2~'J


. 346Ea a
a.ea8Eae0


-0.SE=+02 -9o-
8.52 E-82
Do$


.15E-1) .C 8 o.53 1os
u(sec~~ ) 05


FIGURE 4. Frequency Response Bode Plot for Mixing
Tank.

fice, a rotameter, and is vented. The equipment is
fully instrumented with two pressure gauges, pres-
sure transducers, PI analog controllers and is also
interfaced with the minicomputer allowing data
acquisition, supervisory and direct digital control.
The first task given the students is to develop
a mathematical model for the tank system. An
unsteady state mass balance for each of the two
tanks yields

dp1 RT
dt V= [CdlAif(Popi) -Cd2A21(P1,P2)] (1)

dpC RT
dpt- = T_[Cd2A Z(PP~1P) -CdsA, (P2,P3)] (2)
dt V.M


where


0.873+22
seconds


0.174E+03
160


R = universal gas constant
T absolute temperature
Vi = volume of tanki
M = mean molecular mass of air stream
Cdi discharge coefficient for orifice i
Ai area of orifice i


FIGURE 3. Pulse Test Data for Mixing Tank.

acting gas storage tanks. This experiment requires
three laboratory periods plus some lecture prepa-
ration. The purpose of this experiment is to
demonstrate to the student the effects of inter-
actions in multivariable systems and to give him
or her the possibility of testing different multi-
variable control schemes on a real system.
A simplified version of the system flow sheet
is given in Figure 5. Air enters the system at a
constant pressure of 60 psig, flows through a
control valve into the first tank and from there
through another control valve into the second
tank. Finally, the air passes through a fixed ori-


K, (Pi(Pi- Pi+))1/2 P+, /PI > .5 (3)
0(p,Pi+l)
Kpi pi+,/pi < .5 (4)


KI,K2 = material constants

All parameters of the model are available to


FIGURE 5. An Interacting Gas Storage System.

CHEMICAL ENGINEERING EDUCATION








the student in tabulated form except the discharge
coefficients which have to be determined through
steady state experiments.
Computer control experiments are carried out
with the goal of achieving a given gas production
rate while meeting certain pressure constraints in
the two gas storage tanks. Both Supervisory
Control and Direct Digital Control algorithms are
tested by the student. Different control objectives
can be selected by the student but they all result
in the regulation of the two pressures through
changes in the two control valves. Let us briefly
indicate some of the choices available to the
student.

A. Supervisory control:
Possible control objectives:
1) specified gas flow rate from tank 2 and
2) p1/P2 fixed or p, minimized or p, maximized
In supervisory mode, the valves are under
local analog control. After a control option is


FIGURE 6. Control Structure for DDC Control with
Single Loop PI Controllers and with Added Steady State
Decoupling (dashed lines).

entered by the student via the computer terminal,
the set points are computed and transmitted by
the computer to the local controllers. The response
to changes in objectives is observed for different
flow regimes. (Critical or subcritical flow through
the valves.) Set point compensation is attempted
to yield a smoother servo behavior.

B. Direct digital control (DDC):
The possible control objectives are identical to
those listed in part A. Different multivariable
control algorithms are developed by the students
and supplied to the main control program in the
form of Fortran subroutines. As an example,
steady state decoupling is implemented and com-
pared with the usual single loop PI control. The
controller structure is seen in Figure 6. The re-
sults of one laboratory group are shown in Figures


PI COHTPCL, CYI=C'K. 5,. TII-TI121 0
PRORF = 90.8, PPA9 I0 = 1.1
I=PI 2=2, 3=CVI, D 2 r5FPODR
AY-AXI IS LINEARi -AXIS IS LINEAR
1*. .-,-r*T .' ]
2 O ,:,.; -' *.


e! le" e-._ ..






e eE
a eeE.$1 8 gmll*1


I 171**13


FIGURE 7. Process Response Under DDC With Single
Input-Single Output PI Control.


7, 8. With simple PI control, significant oscilla-
tions in the pressure response were found (cf.
Figure 7). However, with the addition of steady
state decoupling, the response was much improved
(Figure 8).

CONCLUSIONS

T HE NEW MINICOMPUTER has become an integral
part of the undergraduate control course at
Wisconsin. Aspects of digital computer control are
demonstrated to the students and they have the
opportunity to gain some practical experience with
the implementation and application of modern
control algorithms. Computer aided control


I CONTROL uITH SS DECOUPLIING. CKlCK2- 05. TI1.TI2-1 J
FROCE = 0 00a. PRATIO 1.1
*r-:i' IS LINEAR X-AXIS IS LINEAR





1 ',.'.'PO E .,:







0.18E+0e1 e.141E*3 1 28*II8
FIGURE 8. Process Response Under DDC With Steady
State Decoupling and PI Control.


WINTER 1980








system design methods utilizing interactive
graphics have replaced classic pencil and paper
methods and have thus made time available to
include new theoretical material in the curricu-
lum. O

ACKNOWLEDGMENTS
Our progress in bringing real time computing into the
process control curriculum at Wisconsin is due to the co-
operative efforts of many individuals:
1. Emeritus Professor R. J. Altpeter, who originally
established process control as a discipline at Wisconsin.
He laid the foundation upon which the present cur-
riculum is built.
2. Visiting Professor Ram Lavie, who shared his ex-
perience in the development of laboratory experiments.
3. The students who contributed their time and talents
to the development of new experiments and computer
aided design programs-these include Dennis Arnon,
Dean Berceau, John Bolling, John Greiner, Tim Heisel,
Sunny Lo, Bob Lojek, Diana Meseck, David Roark,
John Seymour, and Pat Vilbrandt.
4. The technical staff and faculty of the department of
Chemical Engineering who should be recognized for
the support they have given this endeavor. In par-
ticular, the efforts of Mike Lynch, Todd Ninman, Jim
Wenz and Don Zentner have been invaluable.
5. The more than 200 undergraduate and graduate
students who have "consumer tested" the changes in
curriculum and provided useful feedback.











Ever been in the midst of research and realized that
easy access to U.S. Patents would be the answer to
your problems? U.S. Patents on microfilm from
Research Publications, Inc. offers to attorneys,
researchers, inventors, and librarians the largest
single body of scientific and technical information
in existence. RPI offers a complete retrospective
file of U.S. Patents and current subscriptions, as well
as the Official Gazette and the CDR File. Access is
immediate, and information retrieval is quick and
easy. Delivery of current subscriptions is weekly,
within four weeks of date of issue. Find the answers
to your questions with U.S. Patents on microfilm
from:


research publications, inc.
12 Lunar Drive, Woodbridge, Ct 06525 (203) 397-2600


THE INSTITUTE OF


PAPER CHEMISTRY


ASSOCIATE


DEAN

Nominations and applications are invited
for the position of Associate Dean at The
Institute of Paper Chemistry.

The Associate Dean is Director of Ad-
missions and chairs the Admissions Com-
mittee. This includes responsibility for the
recruitment and evaluation of prospective
graduate students.
The Associate Dean also works closely
with the Dean in administering the gradu-
ate programs at the Institute, assists in the
development and implementation of aca-
demic policy and shares responsibility for
the management of the Dean's Office.
The Associate Dean reports directly to
the Vice President-Academic Affairs/Dean of
the Institute. The Dean's Office has re-
sponsibility for faculty planning, student re-
search, admissions, financial aid, student
records, advisement, placement, continuing
education, alumni affairs and the computer
center.
The position includes an appointment on
the faculty which affords graduate teach-
ing and research opportunities.


Letters of application, with credentials,
should be sent by April 1, 1980 to:

Harry T. Cullinan, Jr.
Vice President-Academic Affairs & Dean
Institute of Paper Chemistry
P. O. Box 1039
Appleton, WI 54912

The Institute is an equal opportunity, affirmative
action employer.


CHEMICAL ENGINEERING EDUCATION







The Leading Authorities...

and the Leading Texts
aU 0


Hill




COLLEGE DIVISION
McGRAW-HILL
BOOK COMPANY
1221 Avenue of the Americas
New York, N.Y. 10020


Separation Processes
SECOND EDITION
C. Judson King, University of California at
Berkeley
1979, 880 pages (tent.), $27.95 (tent.)
Here is a book featuring complete, general
coverage of separation processes, with material
organized by concept. The common features and
key differences between processes are
emphasized, and many worked examples,
problems, and case studies are included. Solutions
Manual.
Available Now

Mass Transfer
Operations
THIRD EDITION
The late Robert E. Treybal
1980, 800 pages (tent.), $26.50 (tent.)
A thorough, highly acclaimed study of
.mass-transfer operations in chemical
" engineering, this book covers relevant
principles of equipment design, data
concerning equipment design, and
applications of modern theory. Solutions
Manual.
Available Now


brought to

you by

McGraw6


SI m i i I ---------------
McGraw-Hill Please send me the following books) for a free 10-day
Book Company examination. If not completely satisfied, I will return the
Book Compay books) within the trial period; otherwise I will remit
P.O. Box 400 payment, plus postage, handling, and local sales tax
Hightstown, N.J. 08520 (McGraw-Hill pays postage and handling if payment is
included with order).
O King: SEP PROCESSES, 2/e Name
(034612-7)
I Peters-T: PLANT DES & ECO, 3/e Address
(049582-3)
E Satterfield: HET CATALYSIS, City
(054875-7)
O Treybal: MASS TRANS OPS, 3/e State Zip
(065176-0)
U-070-0107-1 Prices Subject to change. Rev72/CE Ed
am m m m mmnmmmmm n m m m mmm mm mm


in cnemical

Engineering


Heterogeneous
Catalysis in Practice
Charles N. Satterfield, Massachusetts
Institute of Technology
1980, 464 pages (tent.), $25.50 (tent.)
This book is primarily devoted to a study of
catalysts and reactions that have industrial
significance for large-scale operations. It includes
suggestions for designing experiments and
analyzing data, and a guide to relevant literature.
Solutions Manual.
Available May, 1980

Plant Design and
Economics for
Chemical Engineers
THIRD EDITION
Max S. Peters and
Klaus D. Timmerhaus, both of the
University of Colorado at Boulder
1980, 992 pages (tent.), $28.50 (tent.)
In this book, economic principles are applied to
the study of chemical engineering, showing the
relevance of economics to industrial operations
and plant design. Solutions Manual.
Available February, 1980









laboratory


PROCESS CONTROL EXPERIMENT: THE TOILET TANK

THOMAS J. WARD
Clarkson College of Technology
Potsdam, N.Y. 13676
THE LEVEL CONTROL mechanism in the home
toilet tank is a nonlinear, proportional control
system that illustrates various process control
concepts. It can also serve as an introduction to
data acquisition, process analysis, and model de-
velopment. This simple experiment can be de-
veloped as an example problem, a classroom
demonstration, or a laboratory exercise.


DESCRIPTION

A TYPICAL TOILET TANK is shown in Figure 1.
The level control system regulates the tank
water level C at the desired steady state C, by
manipulating the inlet water rate M to compen-
sate for the disturbance (the flushing rate U).
The control logic is given in the flow diagram of
Figure 2A.
An actual toilet tank can be used for the labora-
tory exercise if a flowmeter is installed in the feed
line and a measuring scale is fastened to the in-
side of the tank wall. For classroom demonstra-
tions, the level mechanism can be installed in a
clear plastic tank (0.5 by 0.15 by 0.4 meters high).
A quick-opening valve can provide the necessary
flush for the plastic tank.


Tom Ward has been on the chemical engineering faculty at Clarkson
College of Technology for many years. He received his M.S. from
the University of Texas and his Ph.D. from RPI. His research interests
are principally in the area of process control.


FIGURE 1. A typical tank.

Typical exercises involve the development of
analytical models, determination of model pa-
rameters, and measurement of the closed-loop re-
sponse to a flushing disturbance. Two suitable
process models are presented in the following
sections.

NONLINEAR MODEL
THE EQUATIONS FOR EACH of the elements of
Figure 2A can be obtained as follows:
Tank: A mass balance on the tank gives
dC
A = M- U (1)
dt


0 ~ E



B.
FIGURE 2. Information flow diagrams: A. Nonlinear
logic, B. Linear block diagram.


CHEMICAL ENGINEERING EDUCATION


Copyright ChE Division, ASEE, 1980








where A is the tank cross-sectional area. A typical
value for A is suggested by the plastic tank dimen-
sions given above so that
Tank: A mass balance on the tank gives
A = (0.5) (0.15) = 0.0752 (2)
Float: The float center height B is assumed to
correspond to the water level so that the equation
for the measuring element is simply
B = C (3)
Alternately, it could be assumed that the float is
characterized by a first or second order model.
While this would lead to an interesting higher-
order process model, it might be difficult for
students to estimate the parameters in a higher-
order model.
Valve: It is assumed that the valve flow rate
M can be related to the valve position M' by the
equation
M = a (M')o.5 (4)
Many toilet valves have an adjustable ring so that
the valve coefficient can be changed. A typical
value for the coefficient a is 0.00212.5/s. The
experimental determination of a suitable valve
equation and coefficient can provide an interest-
ing short study.
Lever: A reasonable controller equation relat-
ing the valve position M' to the error (C, B)
can be written as
M'= K (C B) (5)
where the controller gain K is given by the lever
ratio. A typical value for the desired steady-state
level C. is assumed to be 0.3 meters. If the lever
is assumed to be 0.40 meters long (to the float
center) and the pivot is 0.04 meters from the
valve end, then the controller gain is given as
K = 0.04/0.40 = 0.1 (6)
These can be combined to give the nonlinear model


dC
A dt a[K(C -C)]0.-5-U
dtA-=


dC ^ ^
Tank: A = M U
dt
A A
Float: E = -B = -6
A A
Lever: M' = KE


(10)


The nonlinear valve equation can be linearized
around the steady state by the truncated Taylor
series approach to give


A IdM ^ A A
MM ( 1dM ,' iM'


(11)


where p should be the slope of the valve curve at
the origin. As can be seen in Figure 3, this would


LU.UUUV i


M 0.0002
(


0


/S'
A-


S0.01 SLOPE 0.02 0.3m


0 0.1 0.02 0.03


FIGURE 3. Linearization of valve response.


give a value of p that would be much too high
away from the origin. A line with slope p =
0.0159 is arbitrarily selected to approxi-
mate the value behavior over the region of
interest. This illustrates how experimental data
can be used to improve on the classical steady-
state linearization.
The linear model corresponds to the block
diagram of Figure 2B, where G, = K, G, = P,
H = 1, Gp = 1/AD, and D = d/dt. These can be
combined to give the linear model as


(TD + 1)C = -aU
where T = A/P/K and y = 1/pK.


(12)


LINEARIZED MODEL
A S NOTED ABOVE, the steady state is selected as
the filled tank (a nonflow condition). For any
variable X, the perturbation from the steady
A
state is defined as X = X X,. In terms of such
perturbation variables, the equations for three of
the elements are


DISTURBANCE INPUT
T IS ASSUMED THAT the flush, given by SUdt,
empties the tank. Then JUdt = (0.5) (0.15)
(0.3) = 0.0225 3. Various approximations
can be suggested for the disturbance U. A suit-
able function might be the displaced cosinusoid
with a period of 10 seconds


WINTER 1980









S= -0.00225 cos 0.27rt + 0.00225

0 t>10 sec (13
Since the flush occurs quickly compared to the
filling time, it might be reasonable to approximate
the disturbance as an impulse of 0.0225
occurring at t = 0.

MODEL RESPONSES

T HE RESPONSE FOR VARIOUS model and dis-
turbance forms can be compared with experi-
mental responses to show the significance of the
approximations made. Three model responses will
be given here:
a. Nonlinear Model With Cosinusoid Dis-
turbance. The response for this case was obtained
numerically for the parameter values assumed
above. This is given by the curves NL-C in Figure
4.
b. Nonlinear Model With Impulse Dis-
turbance. If the flush is an impulse at t = 0,
then the solution for t>0 can be obtained by
assuming that U = 0 and C = 0 at the instant
t = 0+. Equation 7 can then be solved by the sepa-
ration of variables technique to give
(C,- C)--5/2A)t + C1 (14)
For the assumed parameter values, the integra-
tion constant C1 = (0.3)0.5. The response curves
for this case are given as NL-I in Figure 4.
c. Linear Model With Impulse Disturbance.
A A
If it is assumed that U(0+) = 0 and C(0+) =
-0.3, then the solution to Equation 12 is


C = -0.3 e-
C -= 0.3 e-t/T


(15)


The responses for this case are shown in Figure
4 as the curves LI.

OTHER FEATURES

T HE TOILET TANK CAN be used to illustrate other
process control concepts. Some of these are:
a. Offset. The leaking toilet provides one of
the simplest demonstrations of offset. If the out-
flow at steady-state is not zero, then the inflow at
steady state is not zero. Since the flow into the
tank is a fixed function of C, then the steady-state
level must decrease if the toilet tank is leaking.
b. Measurement Error. The water-logged
float implies that the float center-line does not


0.
LEVEL C



FLOW M


TIME (s)
FIGURE 4. Level and flow responses (LI = linear model
with impulse disturbance, NL-I = non-
linear model with impulse disturbance,
NL-C = nonlinear model with cosinusoid
disturbance).

correspond to the level. At steady state, the level
must increase in order to close the feed valve.
c. Setpoint Changes. The float end of the
level can be bent to represent setpoint changes.
d. Controller Gain. Most toilet tank controls
have some provision for changing the loop gain.
In some cases, this is accomplished by changing
the effective lever ratio. In others, an adjustable
ring is used to change the valve gain.

CONCLUSION
The toilet tank, as either a classroom problem
or experiment, provides a simple introduction to
process control. D


CHEMICAL ENGINEERING EDUCATION









rec.og.niL tion\, rek-ig-'nish-an,
-og-\ n 1 : the action of recognizing; the state of being
recognized; as a : ACKNOWLEDGMENT 2 : special notice
or attention.




rec.og.ni.tion\ as we see it\
1 : the primary motivation to do creative work for an out-
standing company 2 : ACKNOWLEDGMENT of the quality of
that work; as a : self-satisfaction and pride b : respect
from peers and associates c : opportunity for advancement
3 : to recognize the challenge of the world today 4 : to be
recognized for doing something to meet those challenges
tomorrow.



If you know of qualified graduates in engineering or the
sciences, or with an interest in marketing, finance or computer
science, we hope you will encourage them to write us: Re-
cruiting and College Relations, P.O. Box 1713-CE, Midland,
Michigan 48640. Dow is an equal opportunity employer--
male/female. DOW CHEMICAL U.S.A.
1 *Trademark of The Dow Chemical Company


WINTER 1980 ,,










Curriculum


A SURVEY OF PROCESS CONTROL EDUCATION


IN THE UNITED STATES AND CANADA*


DALE E. SEBORG
University of California
Santa Barbara, CA 93106


N FEBRUARY, 1978 a questionnaire on process
control education was distributed to the 158
chemical engineering departments in the United
States and Canada. Completed questionnaires were
returned by 143 schools, or 90% of the 158 depart-
ments. This response compares quite favorably
with the 59-101 replies that were received in re-
cent AIChE surveys of undergraduate curricula
[1, 2]. The large number of replies is probably due
to two factors: 1) the questionnaire was kept very
brief, and 2) copies of the final report were
promised to those departments which submitted
completed questionnaires.

SURVEY RESULTS

T HE SURVEY RESULTS indicate that process
control is firmly established in the undergradu-
ate curriculum since only 7 of the 143 respondents
(5%) do not offer undergraduate courses. By
contrast, 108 schools (75%) have required courses
and an additional 28 schools (20%) offer elective
courses. Interestingly enough, 4 of the 7 schools
which do not offer undergraduate courses in pro-
cess control do offer graduate courses. Thus only
3 schools of the 143 respondents do not offer any
process control courses.
Process control courses are also firmly es-
tablished at the graduate level. Seventy-two
schools (50%) offer graduate courses while an


The survey results-indicate that
process control is firmly established in
the undergraduate curriculum since only
7 of the 143 respondents (5%) do not
offer undergraduate courses.


*A preliminary version of this paper was presented at
the Miami Beach AIChE Meeting.


TABLE 1
Textbook Selection for Undergraduate and Under-
graduate/Graduate Courses


TEXT


NUMBER OF DEPTS.


Coughanowr and Koppel
Luyben
Weber
Harriott
Douglas
Perlmutter
Smith, Cecil
Ogata
Others (one each)


2
2
17
Total 135


additional 15 schools (10%) offer courses which
are open to both graduate students and advanced
undergraduate students. Tables 1 and 2 list the
process control textbooks which have been adopted
for undergraduate and graduate courses, re-
spectively.
The most striking result here is the con-
tinuing popularity of the book by Coughanowr and
Koppel which has been selected as an undergradu-
ate text by 69 departments and as a graduate text
by 7 departments. The dominant position of this
15 year old text is quite remarkable in view of the
significant developments which have occurred
since 1965 in both computer control hardware and

TABLE 2
Textbook Selection for Graduate Courses


TEXT
Smith, Cecil
Coughanowr and Koppel
Luyben
Douglas
CACHE Monographs
Lapidus and Luus
Other (one or two each)


NUMBER OF DEPTS.
10
7
4
2
2
2
17
Total 44


Copyright ChE Division, ASEE, 1980


CHEMICAL ENGINEERING EDUCATION









TABLE 3
Laboratory Control Experiments in Undergraduate
Courses
NUMBER OF
EXPERIMENTS/COURSE NUMBER OF DEPTS.
0 43
1-2 30
3-4 26
5+ 37
Some experiments
(number not available) 7
Total 143

in control theory. The results in Table 1 agree
quite well with a 1975 survey on undergraduate
process control courses [1]. It should be noted
that the numbers in Tables 1 and 2 are reported
on the basis of individual departments rather
than on the basis of courses offered. For example,
if a particular department offers two undergradu-
ate process control courses which use the same
textbook, this was counted only once in Table 1
rather than twice. By contrast, if two textbooks
were required for a particular course, they both
were included. Many of the 17 textbooks included
in the "Other" category in Tables 1 and 2 were
written for mechanical or electrical engineers and
are used in classes taken by both chemical engi-
neering students and other engineering students.
One hundred departments (70% of the re-
spondents) indicated that their curriculum in-


Dale E. Seborg received a B.S. degree from the University of
Wisconsin in 1964 and a Ph.D. degree from Princeton University in
1969. Both degrees were awarded in chemical engineering. He
taught at the University of Alberta for nine years before joining the
University of California, Santa Barbara in 1977. He is presently Pro-
fessor and Chairman of the Department of Chemical and Nuclear
Engineering at UCSB. Dr. Seborg is co-author of the book, Multi-
variable Computer Control: A Case Study. He is also the past chair-
man of Area 15B (Systems and Process Control) of the AIChE Na-
tional Program Committee.


The most striking result here is
the continuing popularity of the book by
Coughanowr and Koppel which has been selected
as an undergraduate text by 69 departments
and a graduate text by 7 departments.


cludes one or more laboratory experiments in pro-
cess control. Table 3 shows that 63 departments
offer courses that contain at least three control
experiments and 30 departments have one or two
experiments, usually as part of a unit operations
laboratory.
In compiling these statistics, each department
was included in only a single category. Thus if a
particular department offers two process control

TABLE 4
Use of Real-Time Computers or Micro-Processors in
Control Experiments
NUMBER OF DEPTS.
Currently have a real-time system 48
Equipment on order or being installed 19
No equipment (but have tentative plans to
add equipment) 22
No equipment (and no plans for future
equipment) 53
Total 143

courses which include three and five experiments,
respectively, this department was included in the
tally for the "5+" category in Table 3.
During the past decade there has been con-
siderable interest in "real-time computing," that
is, in digital computers which are used for data
acquisition and control. Both industrial and aca-
demic personnel in the process control field have
maintained an active interest in the field of real-
time computing for the following reasons:
The widespread availability of inexpensive minicom-
puters and microprocessors;
Changing process control objectives in industrial
plants due to energy and environmental considera-
tions;
The realization that the application of most advanced
control strategies will inevitably require an on-line
digital computer.
The results in Table 4 indicate that 48 depart-
ments (34%) currently have control experiments
which involve a real-time computer while an ad-
ditional 19 departments (13%) have computer
systems on order or being installed.
Table 4 includes only those departments which
use real-time computers in conjunction with under-
graduate control experiments. It does not include


WINTER 1980








TABLE 5
Real-Time Computers or Microprocessors
Currently Installed (53) or On Order (19)


EQUIPMENT AND VENDOR
Minicomputers
Digital Equipment Corp.
Data General Corp.
Hewlett Packard
IBM
Texas Instruments
Foxboro
Interdata
Miscellaneous (one each)
Not specified
Microprocessors


NUMBER OF DEPTS.

24
8
5
5
3
2
2
6
7
10
Total 72


other departments which use minicomputers or
microprocessors exclusively in research labora-
tories. The 22 departments in the third category
in Table 4 typically are in a preliminary planning
stage or are seeking funds to purchase a real-time
system. Thus the results of this survey indicate a
continuing trend for incorporating a real-time
computer system in the undergraduate curricu-
lum.
Table 5 presents a summary of the 62 mini-
computers and 10 microprocessors which are
currently operating in chemical engineering de-
partments or on order. The numbers in Table 5
do not correspond directly to those in Table 4
since several chemical engineering departments
use more than one real-time computer in the under-
graduate curriculum.

CONCLUSIONS

T HE RESULTS OF THIS survey indicate that the
topic of process control has become firmly es-
tablished in the chemical engineering curriculum.
Only 3 of the 143 departments surveyed do not
teach any courses in process control. One hundred
and eight schools (75% of the respondents) have
required undergraduate courses while 87 schools
(61%) teach graduate level courses in process
control. Laboratory experiments in process control
are now available at 100 schools (70%). There is
a continuing trend toward providing students with
exposure to real-time computer systems in con-
junction with process control experiments; 67 de-
partments currently have such a system operating
or on order while an additional 22 departments
have tentative plans for such a system.
Fifteen years ago, process control was
generally regarded as a new, specialized topic


which was not part of mainstream chemical engi-
neering. The present survey demonstrates that this
situation no longer exists. Process control has
joined the more traditional topics such as trans-
port phenomena, thermodynamics and reactor
analysis in playing a central role in the chemical
engineering curriculum. E

REFERENCES
1. Eisen, E. O., "Teaching of Undergraduate Process Dy-
namics and Control," paper presented in a mini-session
at the 68th Annual AIChE Meeting, Los Angeles (No-
vember, 1975).
2. Barker, D. H., "Undergraduate Curriculum 1976,"
Chem. Eng. Educ., Vol. XI, No. 2, (Spring, 1977).

BOOK REVIEW: Contact Catalysis
Continued from page 12.
hope of being able to reproduce catalysts of a given
type in different laboratories is rapidly becoming
a reality.
As one might infer from the variety of topics
and extent of treatment, these volumes are not
exactly for the beginner. One might have wished
some discussion of homogeneous catalysis, at least
in terms of analogs to heterogeneous systems, and
a more general inclusion of the concepts of co-
ordination chemistry as they relate to catalysis.
In all, however, some balance must be struck be-
tween coverage and length and the editor has
done an admirable job. The English translation
of the original Hungarian edition of 1966 is ex-
cellent and the text has been updated. The dust
jacket states that "the book will be useful to
workers studying catalysis in industrial and uni-
versity laboratories." The present reviewer feels
this is a correct statement and commendable for
its modesty. O


m news

ART HUMPHREY HONORED
Arthur E. Humphrey, dean of Penn's School
of Engineering and Applied Science, became the
eighth honoree to receive the James M. Van Lanen
Distinguished Service Award for his "life long
dedication and service to fermentation science and
the fermentation industry.
The award is named for a pioneer in fermenta-
tion technology and was established in 1976 as the
foremost award and citation of the ACS Division
of Microbial and Biochemical Technology.


CHEMICAL ENGINEERING EDUCATION











class and home problems


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


SOLUTION: MIRROR FOG PROBLEM

R. L. KABEL
Pennsylvania State University
University Park, PA 16802

Editor's Note: Professor Kabel presented the
"Mirror Fog Problem" in the Fall 1979 issue of
GEE. We extended an invitation for student solu-
tions to this problem at the time of publication and
would like to congratulate Mauricio Fuentes of
Ecole Polytechnique, Montreal, Canada, who sub-
mitted the winning entry and by so doing has won
a year's subscription to GEE. Professor Kabel
graded the responses and, in his words, Mr.
Fuentes' entry was both "correct and excellently
done." The following is Professor Kabel's solution
to the problem.

Derivation of equations:
Use a microscopic model because momentum and
energy equations are not required due to isothermality
and no bulk flow.
Mass balance equation:
DCk 3c, DCA DCA
S+ v + VY + V
at Dx DY ?z
DAB D2C + D2CA + A]+Ca
DAB 2 2 + RA

Since vx = Vy = v = 0, CA =t f (x,z) and there is no
generation in the vapor space this equation becomes.
DCA 2 DCA
at DY'
which shows that the concentration at any point changes
with time because of diffusion in the y-direction.
Initial condition: At t = 0, CA = CA,sat at all y
Boundary conditions: At y = 0, CA = CA,room at all time
At y = Y, CA = CA,sat at all time
where Y is the location of front edge of the remaining
fog on the mirror. Note however that Y varies with time
going from 0 when t = 0 to 0.3 m when t = tf.
If an analytical solution of the equation is to be sought
this second boundary condition should be respecified. If the


solution is to be numerical, then one merely needs to keep
track of Y(t). The end of the calculation is t = t, when
Y = Ym, = 0.3 m. Y(t) can be obtained by equating
the total amount evaporated to the integrated mass flux
into the room neglecting the slight accumulation of water
vapor in the enlarging vapor space.
Let MAo be the initial total mass of water condensed, then
Y Y
Amount evaporated = MAo = MA-
0.3 YXma
ft p yf0
Amount transferred to room = DA CA S dt
J y y= 0


where S = ZX

Yflax t "A
Y =M- ZX DAB,
.oJ DY


XZYmaxDAB
Y = MA,


DCA dt
DY y = 0


The above is an adequate answer to the exam question. A
very simple analytical solution can be obtained as follows.
We can say that the flux at y = Y is equal to the amount
of moisture evaporated there per unit time. Then the
amount of moisture can be related to the rate at which
the boundary moves. Thus, if R = thickness of liquid
film,

dCA g H20 evap.
Bdy y Y area time

PHno ZR dY g H0 evap/time
ZX dt area of transfer

S dCA PHo R dY
B dy y= X dt

If we assume that a steady state concentration profile is
established rapidly and maintained (shown by dynamic
analysis to be an excellent assumption) we get


dCA
dy
y-Y


CA,sat CA,room
Y


WINTER 1980








A further simplification is obtained by taking CA,room = 0
Let CA,sat = CA,.

SCAS H2o R dY
DAB X
Y X dt


tf
SDAB CSA X dt

o oR


Ymax

B dt = Y dY

o o


Ymax
B tf = 0.5Y2 I = 0.5 Y2ma
0
d tf = 0.5 Y2max/B
for the mirror fog problem


PHo2
R
DAB
CAs
Ymax
X
tf


106 g.m-3
1.3 x 10-5 m (obtained from experiment)
2.47 x 10-5 m2 s-1
17.3 g-m-3
0.3 m
4 x 10-3 m
3.5 x 106 s = 96 m = 4 days


This result appears high by about a factor of 4. There
are several explanations and we have calculated for
different assumptions. Probably the experimental circum-
stances (e.g. leakage around edges, etc.) do not meet
the idealizations of the model. [


IN THE "HEAT" OF THE NIGHT

R. J. GORDON
University of Florida
Gainesville, FL 32611

Y OU ARE SPENDING the evening in a small town on
your way home for the holidays. At about
11:00 p.m. the local sheriff calls you and asks for
your help. He knows from the desk clerk that you
are a chemical engineer, and naturally assumes
you have some knowledge of forensic chemistry.
It seems that the body of John Lurie, a local
car dealer, had been found somewhat earlier in a
wooded area just outside of town. The local
coroner had gone fishing and there was no one
else to estimate the time of death. John Lurie had
been known to deal in "hot" cars and was thought
to be going to the police to confess and name his
four accomplices, Gus Nusselt, Bill Gurney, Ed
Reynolds, and Bob Prandtl. Nusselt had been
known to be out of town until 11:00 A.M. that
morning, Gurney had a solid alibi from 1:00 p.m.
on, Reynolds was with his girlfriend until about


6:00 A.M., when he left to go fishing, and Prandtl
was in jail the night before for drunkenness, and
was not released until about 8:00 A.M.
When you finally get to the body it is about
12:00 p.m. (midnight). You measure a rectal
temperature of 80F, and an air temperature of
70F. The air temperature has been about 700F all
day.
Luckily, you brought your Perry's along.
Recognizing that the human body is mostly water,
1) calculate the latest possible time the murder
could have occurred and 2) state the possible
suspect.
NOTE: For practical purposes, John Lurie
can be assumed to be shaped like a rectangular
slab. He is 10 inches thick from his back to his
breastbone. Body temperature is 98.60F. Rectal
temperature is equivalent to core or centerline
temperature.
For comparison, a pathology formula some-
times used to estimate the time of death is*
of h. s d 98.6 rectal temperature
No. of hrs. since death -1
1.5

ONE-DIMENSIONAL SOLUTION
We will use the Gurney-Lurie charts, Perry's
4th Ed., p. 10-6, 10-7. To calculate the latest time
the murder could have occurred, assume maximum
rate of cooling, or in other words that the surface
of the body is at 70F (same as saying h = oo or
m = 0). We also neglect radiative losses since the
body was found in a "heavily wooded" area.
Then, if we assume infinite width and depth,

S T, -T 70-80 0.35
T To 70-98.6
From graph, for n = 0
at
X 0.54
(x12)
k 0.36
a pC 62.4 x .0 0.0058 ft2/hr

5 inches
X1 5 incs 0.42 ft.
12
0.54 x (0.42) 2
t 0.0058 = 16.4 hours or
16 hours, 24 minutes

1) Murder had to occur before 7:40 A.M.

*"Medical Jurisprudence and Toxicology," Glaister and
Rentoul, 12th Ed., Livingston Ltd., Edinburgh, 1966, p. 110.


CHEMICAL ENGINEERING EDUCATION


Copyright ChE Division, ASEE, 1980







2) Possible Suspects: Gurney, Reynolds
From the pathology, formula:
98.6 80
98.6-80 = 12.4 hours
1.5
(or murder occurred at 12 noon)
This formula, however, takes no account of
changes in room temperature, or body thickness,
and in fact is known to underpredict the time of
death except for the first few hours. From our
superior knowledge of heat transfer, we have
eliminated Prandtl and Nusselt as suspects. OE

ACKNOWLEDGMENT:
Helpful comments were provided by Professor
J. H. Hand, University of Michigan.
Editor's Note: Professor Gordon's purpose in his
solution to the foregoing problem, "In The Heat
of the Night," was to illustrate the use of the
Gurney-Lurie charts assuming a simple one-dimen-
sional model. Professor Fogler, CEE Problem
Section Editor, asked his student, Alan Basio, to
comment on this simplified solution. Mr. Basio's
reply follows.

TWO-DIMENSIONAL HEAT TRANSPORT
ALAN BASIO
University of Michigan
Ann Arbor, MI 48109
It was previously assumed that Lurie, the dead
man, is an infinite slab. From this assumption, the
time is 16.4 hours since Lurie was killed.
I used Newman's Rule and assumed Lurie is
an infinitely long slab with a. finite width and
depth. Newman's Rule in this situation is the
following:

Y = Y, =T- T =0.350 (1)
-T T
T,8 T.
Let Lurie be 10" deep, as previously specified,
and 1.3 feet wide. Use the same values as before
for Y and a. There are now two values of X to be
found on the Gurney-Lurie Charts:
Xx = at/(5/12)2 and X, = at/(1.35/2)2. The
time must be the same in both Xx and Y,, and the
product YY, = 0.350.
Criteria for solution: (1) YxY, = 0.35
(2) X.(x) X,(y) t
a a
Results: By trial and error the times are found


to be within 2.7% of each other.
Yx = 0.420 Y, = 0.833
X. = 0.45 X, = 0.18
YY, = (0.42) (0.833) = 0.350

t (0.45) (5/12)2 13.47 hrs.
0.0058
S 0.18 (0.65)2 = 13.10 hrs.
t-
0.0058
13.47 13.10
13.47 x 100 2.7% difference
13.47
If the width of Lurie is 1.3 ft., he died 13.3
hrs. ago, not 16.4 hrs.
The width of Lurie is important. If Lurie is
2.6 ft. wide, for example, he dies 16.3 hours earlier.
In other words, the infinite slab assumption im-
proves when Lurie is assumed over 2.0 feet wide,
approaching an answer of t = 16.4 hrs. O

BOOK REVIEW: Reactor Design
Continued from page 24.
The book is an excellent work. The author has
covered a very large area of relatively difficult
material in a highly readable fashion and has pro-
vided enough detail so that the reader is able to
come to grips with the realities of chemical re-
actor design. It is accurate and relatively com-
plete. There is a considerable amount of specialized
knowledge, based upon over 1000 references, aug-
mented by the author's own considerable ex-
perience. In many areas, it stands at the edge of
chemical reactor design knowledge that is in the
public domain. As such it will continue to be a
valuable reference work for many years to come.
Its only major shortcoming is insufficient il-
lustrations and a lack of exercises or problems
for the student. The fourteen case studies of
Volume II serve to illustrate design principles but
only cover a fraction of the material in Volume I.
In order to serve as a text for a graduate course in
chemical reactor design, it would have to be sup-
plemented by problems developed to reinforce
specific points and others which would require the
student to integrate these ideas into a chemical
reactor design. The latter would be an under-
taking of the order of a term paper.
These two volumes are a major contribution to
the chemical engineering literature. They belong
in the library of every chemical engineer who is
concerned with research, development, design, or,
in many cases, operation of chemical reactors or
conversion processes. O


WINTER 1980








RATE OF REACTION
Continued from page 16.
The evaluate V v the continuity equation is
used (see ref. 19) :


+ +V (pv) =0
at


gaseous phase, equation (13) is not so useful for
v, will become a function of z and it will not be
possible to take it out of the differential. If the
number of moles stays constant, equation (13) is
reduced to:


(10)


But for this system V p = 0. Besides p = m/V and
am/at = 0 because the total mass of the system
is constant in time. Substituting these relation-
ships into equation (10) and operating:
1 DV
Sv (11)
V at
Substituting Ci by Ni (Ci = Ni/V) and per-
forming the differentiation, the equation for r,
in the case of an isothermal variable volume batch
reactor, results:
dNi
r (12)
a V(t) dt
The partial differential is now a total differen-
tial insofar as Ni is a function of time only. It
must be noticed, however, that we should know
the relationship existing between the change of
volume and the reaction extent, i.e., if the extent
of the reaction is expressed in terms of conver-
sion, there should be at hand a relationship of the
following type: V (t) = V (Vo,x). In most gaseous
systems a linear variation of volume with conver-
sion is often assumed (9).

STEADY STATE ISOTHERMAL CONTINUOUS
PLUG FLOW REACTOR

U NDER STEADY CONDITIONS, general equation
(7) is reduced to:
V Ni = ai r
If, according to the model it is supposed that
there are no diffusion or dispersion effects, the
mass flux is only due to the global convective flow.
Hence:
Ni = Ci v
from where we finally have:
r d(Ci v) (13)
a, dz
since in this reactor the flow is uni-directional.
(Direction z has been chosen as representative of
the model).
If the total number of moles is not preserved
and, furthermore, if the reaction takes place in a


vr dCi
ar- dz
ai dz


(14)


which, obviously, also includes the assumption that
pressure changes along the reactor are small (due
to losses by friction, for example) because other-
wise v, would not be independent of z, either.
Equation (13) can be adequately modified to
become useful even in those cases in which the
number of moles is not constant.
To do so, we simply transform the equation
and work in terms of mass fractions (w). Re-
calling that:

C p wi
Mi

In this equation wi is the mass fraction of com-
ponent i and Mi its molecular weight. Substituting
in equation (13):


d LMpw vZ =air
~M1 I


(15)


and since the mass flow rate Go is constant:
p v. = Go = constant; hence, the rate of reaction
can be written as:


SG dwi
ai Mi dz
ai Mi dz


(16)


STEADY STATE, ISOTHERMAL CONTINUOUS FLOW
STIRRED TANK REACTOR
The SSICFSTR is an ideal type of system, in
which concentration of reactants does not depend
upon time or position within the reactor. The
mathematical description of the flow within the
reactor is exceedingly difficult, as we are dealing
with a highly idealized case which, consequently,
cannot be fully achieved in practice. Nevertheless,
in many cases the deviations are almost negligible
and the system has been successfully modelled.
The exact mathematical description cannot be ac-
complished because the model implies a transport
of mass, instantaneously, over finite distances.
But the difficulty may be overcome if we do
away with the necessity of describing the internal
pattern of the flow inside the reactor. To do so,
the general equation (7) is integrated over the


CHEMICAL ENGINEERING EDUCATION








This statement sums up the whole problem:
The rate of reaction expression is the "sink" or "source" term in the
continuity equation for multicomponent systems which will take into account the creation
or destruction of the said species by chemical reaction.


volume of the reactor so as to obtain a macroscopic
balance.

C' dV=- (VV Ni) dV + Fai rdV (17)
V V V
Under steady state conditions the left-hand
side is zero and, as the assumptions of the model
grant that all properties within the reactor are
constant, r is not a function of the space co-
ordinates and it may be taken out of the integral.
With this, the previous equation is reduced to:

airVR= (V Ni) dV (18)
V
Applying the divergence theorem to the right-
hand side of equation (18), we obtain:

air VR= (Ni n) dA (19)

The integration performed over the whole
surface of the reactor may be evaluated because
the flux is non-zero only at the inlets and outlets.

S(Ni n) dA = (Ni n) dA + (Ni n) dA
A In. Out.
(20)
On the right-hand side, the first term repre-
sents the inlet flow of component i and the second
one, the outlet flow. Taking into account the di-
rections of unit normal vectors n, the final equa-
tion is:
r = (Fi outlet Fi, inlet) (21)
ai VR
which is the expression usually written for the re-
action rate in a continuous stirred tank reactor.
The integration of the differential equation (7),
so accomplished, brings about a macroscopic
balance for species i, which is usually the start-
ing point of the derivations in the books on applied
kinetics. In this work we have followed this ap-
proach in order to demonstrate the absolute
generality of equation (7).
In the following section, this analysis will be
extended to two more complex experimental
systems, the recycle reactors. These are often


useful for obtaining kinetic data.

APPLICATION TO ISOTHERMAL RECYCLE REACTORS
R ECYCLE REACTORS HAVE been widely used since
the publication of the original papers by
Hougen [20], Perkins and Race [21], Biskis and
Smith [22], Korbach and Stewart [23], and
Cassano, Matsuura and Smith [24] as a means of
retaining the differential operation of the reactor
and, at the same time, eliminating the restrictions
of inaccuracy in the analysis of the small composi-
tion changes.
Moreover, control of the flow rates in the re-
cycle allows the reduction of diffusional resistances
and the elimination of temperature gradients. Diffi-
culties may be centered around the effects of re-
action by-products, the considerably longer time
usually needed to obtain the steady state condition
in the recycling section of the apparatus and the
difficulty in operating under pre-fixed concentra-
tion conditions. (The last two are especially im-
portant for the continuous type).
Once more, "general definitions" will be useless
and we shall have to resort to the general mass in-
ventory. In order to simplify the matter, let us
consider the case when the total number of moles
is constant.
ISOTHERMAL CONTINUOUS RECYCLE REACTOR
FIGURE 1 (A) SHOWS the system under con-
sideration. Operating conditions must be ad-
justed in order to fulfill the following assump-
tions:
1) The operation in Va is differential, i.e.,
the outlet concentration Ci,f is very close
to the inlet concentration of the reactor
Ci,i.
2) Differences in concentration between Ci,o
and Ci,f are accurately measured.
3) High recycling flow rate (Q).
When Q co the whole reactor, analyzed in
the control volume (2), is an excellent approxi-
mation to a continuous flow stirred tank reactor,
working at differential conversions as the react-
ing mixture goes through the control volume (1).
The General equation (7) will be applied to each
of these systems.


WINTER 1980








F ------------.




t Ci,Q+q Cf





fCo,q Cf,q


FIGURE 1

Under steady state conditions, the concentra-
tion in the reactor is independent of time. The
differential equation is reduced to:

V Ni = r, ai (22)
On the other hand, the operational characteristics
allow us to neglect any form of dispersion or diffu-
sion effects. Consequently:
Ni = Ci v (23)
If the differential equation (22) is integrated
over the control volume, the result is:


S(V -Ni) dV =f rai dV
V V
On the left-hand side we shall apply the di-
vergence theorem and on the other side we shall
take into account the fact that the reactor is
differential. Substituting expression (23) yields:

f(n Ci v) dA = r ai dV
A V
and finally:
(q + Q) (Ci,f Ci,i) = ai ri Va (24)
On the other hand, if the whole system (2) is
treated as a CFSTR the application of equation
(21) will yield:


q (Ci,, Ci,o) = r2 ai Va


(25)


In equation (25) we have VR only on the right-
hand side, because it is the only part of the total
volume where r is different from zero. But the
global velocity r, has to be equal to that produced
in the reactor itself, that is to say r,. If it is shown
that r, accurately represents a differential rate of


I c
i V I


L------

Analysis
(B)


r = q (C, C,o)
ai VR


(26)


Notice that the condition of a high recycling
flow, necessary for the differential operation in
VR, coincides with the requirements for the opera-
tion of the global system as a continuous stirred
tank reactor. The extension to a system with
variable number of moles only complicates the
algebra.

ISOTHERMAL BATCH RECYCLE REACTOR

F IGURE 1 (B) illustrates the system under con-
sideration. For extremely slow reactions and
for cases where one wishes to avoid serious limita-
tions in the size of the samples for analysis, this
is an adequate experimental device. Within this
system there are no disadvantages such as those
pointed out for a continuous recycle reactor, but
an experimental problem may arise: the existence
of a recycling device (with movable parts in most
cases) could introduce contamination of the react-
ing mixture from the outside. For a batch system,
the impurity level will grow with time. This is
not so severe in continuous systems.
The operating conditions are as follows:
1) The operation in VR is differential.
2) For reasonable intervals of time, concen-
tration differences are accurately
measured in V.
3) High recirculating flow, and adequate
mixing in V.
We treat the whole system as a batch reactor:
if the recirculating flow is high, the concentration
will be uniform. Hence, V N =- 0 and results:


CHEMICAL ENGINEERING EDUCATION


reaction, then the values which may be obtained
through the application of equation (25) will
portray the exact rate of reaction and not an aver-
age value. This will be true if Ci,i is very close
to Cif as has been initially assumed. The only re-
maining doubt would be to know how these condi-
tions could be accomplished. As r, = r2, we have:

(q + Q) (Ci, Ci,,) = q (Ci,f Ci,o)

C Q Ci,f q C,o
q+Q

If, as was first assumed, Q is sufficiently high,
Q >> q and therefore Cii 0 Ci,(.
For isothermal continuous recycle reactors,
working under the conditions stated above, an
adequate expression for the rate of reaction
results:


(AI








r r
S dC1 dV = ai rdV
JV dt JV
Since according to previous assumptions, con-
centration is not a function of position, the total
derivative has been employed and it may be taken
out of the volume integral. On the: right-hand
side, the integral will be different from zero only
in those portions of the volume where r is different
from zero. Due to the differential performance of
VR, one then obtains:
dC (V + VR) =ai rVR
dt
and the rate of reaction, when there is no varia-
tion in the number of moles, is finally:
= (V + Va) dC (27)
r = --Vdt (27)
VR dt
This equation must be corrected for additional
changes of concentration, in case the sample
volumes for analysis were significant. A great re-
lationship of V/VR reduces this problem, but may
largely prolong the necessary reaction time to
attain accurately measured conversions.
Abalance in the reactor itself shows the con-
ditions for the differential operation of Va. At
each cycle the reactor behaves as a steady state
isothermal continuous plug flow reactor.
V *Ni = ai r

1 (V Ni) d = r dV
JV fV
Assuming, for the time being, an average
value of r, the result will be:
Q (C,, Ci,i) = ai ravg V
where Ci,t and Ci,i are the outlet and inlet con-
centrations of reactor Vn at each cycle. If the rate
of reaction has a finite value and, as previously as-
sumed, Q is high, the difference between inlet and
outlet concentrations will be small. By increasing
Q the difference could be made small enough to
turn rav9 into a differential rate of reaction. Notice
that this condition was also assumed as necessary
to propose a uniform composition model. It must
also be noticed that an adequate experimental
device should minimize the volumes of the connect-
ing lines between the reactor Va and the tank V.

HETEROGENEOUS SYSTEMS OF REACTION
N THIS WORK WE have emphasized the analysis
of the different isothermal homogeneous react-


ing systems. The adequate treatment for hetero-
geneous reactions (catalyzed and non-catalyzed)
is outside its scope.
It is not far-fetched, however, to notice that,
in most cases, the flaws of the so-called definitions
are even more evident within these systems. The
problem may be summarized as follows:
1) At the microscopic level all heterogeneous re-
actions take place at interphases.
2) Even at the macroscopic level, many heterogeneous
reactions take place at interphases.
3) In many cases the rate of reaction takes place not
only inside the control volume but on its boundaries
as well.
4) Hence, accurately speaking, the rate of reaction
will, in many cases, be a boundary condition of
the general mass conservation equation.
5) When this happens at the microscopic level (in the
case of a catalyst, for example), an "effective"
rate has been used due to the difficulty in solving
the conservation equations with complicated geo-
metrics for the boundary conditions.
6) At the macroscopic level (for example, the case of
free radical termination reactions on the walls of
a reactor) when the rates of reaction are boundary
conditions, the conservation equations are compli-
cated, since diffusional terms cannot be neglected.
See for example reference [25].
7) Even ontologically, the rates of reaction in hetero-
geneous systems undergo a change, insofar as their
intensive character is attained by means of an ex-
pression referring to the area of this boundary
(real or ideal).
All this makes even more evident the futility
of trying to establish a general "definition" of a
reaction rate.

CONCLUSIONS
N THE PREVIOUS sections, six different expres-
sions have been attained for the rate of reaction:
equations (8), (12), (21), (26) and (27). Many
others could be found for other physical systems
using equation (7) as a starting point. Quoting
Petersen [16]: "to argue that any of these (rate
of reaction expressions) is more correct than all
of the others as its defining equation, is to confuse
a conservation equation with a definition." This
statement sums up the whole problem. The rate
of reaction expression is the "sink" or "source"
term in the continuity equation for multicompon-
ent systems which will take into account the
creation or destruction of the said species by
chemical reaction.
The rates of reaction thus attained will be in-
dependent of the system used to measure them
(provided it is a homogeneous reaction). This
means that the kinetic expression of the rate of


WINTER 1980









reaction r = ( (T, Ci, P, etc .....), if determined
using equation (7) as a starting point, will be in-
dependent of the physical device employed to
obtain the kinetic data.
There may exist a sound definition of an ex-
tensive rate of reaction based on the concept of
the "extent of reaction," both introduced by De
Donder [26], which may be useful for thermo-
dynamic calculations but the idea becomes devoid
of general validity when one needs an intensive
property. For the purposes of reaction engineer-
ing this is legitimate but it is of little help.

ACKNOWLEDGMENT
To Prof. Elsa I. Grimaldi to whom I am greatly in-
debted not only for her invaluable contribution to the
writing of the manuscript but also for her everlasting
patience in doing it over and over again until it finally
achieved its final form.
NOMENCLATURE
A : Area (cm2)
a :Stoichiometric coefficient
C :Concentration (mole-cm-3)
F :Molar flow rate (mole*s-1)
G : Mass flow rate per unit area (gr.s-l.cm-2)
M Molecular weight (gr-mole-1)
m Mass (gr)
N Number of moles
N Molar flux (mole.cm-2.s-1)
n : Outwardly directed unit normal vector
P :Pressure (Kgf.cm-2)
Q :Volumetric recycle flow rate (cm8.s-1)
q :Volumetric input flow rate (cms.s-1)
r Rate of reaction (intensive) (mole'cm-38s-1)
T :Temperature (oK)
t :Time (s)
V :Volume (cm3)
v :Linear velocity (cm*s-1)
V, : Volume of reactor (cm3)
w Mass fraction
x :Conversion
p : Density (gr.cm-s)
S: Kinetic expression for rate of reaction
(mole.cm-8.s-1)
V :Vector operator
Subindices and Supraindexes
avg: Average value
i Indicates species or initial condition, according
to context
f :Final condition
o :Inlet condition
R :Indicates reactor
z Indicates direction
: Indicates vector
Indicates molar average velocity

REFERENCES
1) GLASSTONE, S., "Textbook of Physical Chemistry",
D. Van Nostrand Co., Princeton, N.J., (1946).


2) BENSON, S.W., "The Foundations of Chemical
Kinetics", Me Graw Hill Book Co., New York,
(1960).
3) DANIELS, F., "Chemical Kinetics", Cornell Uni-
versity Press, (1938).
4) LAIDLER, K. J., "Chemical Kinetics", McGraw Hill
Book Co., New York, 2nd. ed., (1965).
5) FROST, A. A., and PEARSON, E. G., "Kinetics and
Mechanism", J. Wiley & Sons, New York, 2nd. ed.,
(1961).
6) JOHNSTON, H. S., "Gas Phase Reaction Rate
Theory", The Ronald Press Co., New York, (1966).
7) HOUGEN, O. A., and WATSON, K. M., "Chemical
Process Principles", Part III, J. Wiley & Sons, New
York, (1947).
8) SMITH, J. M., "Chemical Engineering Kinetics", Me
Graw Hill Book Co., New York, (1956).
9) LEVENSPIEL, 0., "Chemical Reaction Engineer-
ing", J. Wiley & Sons, New York, (1962).
10) WALAS, S. W., "Reaction Kinetics for Chemical
Engineers", McGraw Hill Book Co., New York,
(1959).
11) BOUDART, M., "Kinetics of Chemical Processes",
Prentice Hall Englewood Cliffs, N. J., (1968).
12) PANNETIER, G., and SOUCHAY, P., "Chemical
Kinetics", (Translation by H.D. Gesser and Emond),
Elsevier Publishing Co., New York, (1959).
13) KRAMERS, H. and WESTERTERP, K. R., "Ele-
ments of Chemical Reactor Design and Operation",
Academic Press, New York, (1963).
14) ARIS, R., "Introduction to the Analysis of Chemical
Reactors", Prentice Hall, Englewood Cliffs, N. J.,
(1965).
15) DENBIGH, K. G., "Chemical Reactor Theory", Cam-
bridge University Press, Cambridge, (1965).
16) PETERSEN, E. E., "Chemical Reaction Analysis",
Prentice Hall, Englewood Cliffs, N. J., (1965).
17) AMDUR, I. and HAMMES, G. G., "Chemical
Kinetics", McGraw Hill Book Co., New York, (1966).
18) ARIS, R., "Vectors, Tensors, and the Basic Equa-
tions of Fluid Mechanics". Prentice Hall, Englewood
Cliffs, N. J., (1962).
19) BIRD, R. B., STEWART, W. F., and LIGHTFOOT,
E. N., "Transport Phenomena", J. Wiley & Sons,
(1960).
20) HOUGEN, O. A., "Reaction Kinetics in Chemical
Engineering", Chem. Eng. Progr. Monograph Ser.
47, No. 1, (1951).
21) PERKINS, T. K., and RASE, H. F., A.I.Ch.E.
Journal, 4, 351, (1958).
22) BISKIS, E. G. and SMITH, J. M., A.I.Ch.E. Journal,
9, 677, (1963).
23) KORBACH, P. F. and STEWART, W. E., I.E.C. Fund.
3,24, (1968).
24) CASSANO, A. E., MATSUURA, T., and SMITH,
J. M., I.E.C. Fund. 7, 655, (1968).
25) CASSANO, A. E., SILVESTON, P. L., and SMITH,
J. M., I.E.C. 59, 18, (1967).
26) DE DONDER, Th., "Legons de Thermodynamique et
de Chimie Physique", Paris, (1920); quoted from
Prigogine I. and Defay, R., "Chemical Thermody-
namics", Longmans, Green and Co. Ltd., London,
(1954).


CHEMICAL ENGINEERING EDUCATION














ACKNOWLEDGMENTS


Departmental Sponsors: The following 137 departments contributed
to the support of CHEMICAL ENGINEERING EDUCATION in 1979.


University of Akron
University of Alabama
University of Alberta
Arizona State University
University of Arizona
University of Arkansas
Auburn University
Brigham Young University
University of British Columbia
Bucknell University
University of Calgary
California State Polytechnic
California Institute of Technology
University of California (Berkeley)
University of California (Davis)
University of California (Santa Barbara)
Carnegie-Mellon University
Case-Western Reserve University
University of Cincinnati
Clarkson College of Technology
Clemson University
Cleveland State University
University of Colorado
Colorado School of Mines
Columbia University
University of Connecticut
Cornell University
University of Dayton
University of Delaware
U. of Detroit
Drexel University
University College Dublin
Ecole Polytechnique (Canada)
University of Florida
Georgia Tech
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
Manhattan College
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)
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
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 University
University of Western Ontario
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
32611.






PROCTER & GAMBLE is looking for


in R&D/Product
Development


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.)


RESPONSIBILITY NOW!
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


PROCTER & GAMBLE
AN EQUAL OPPORTUNITY EMPLOYER




Full Text

PAGE 2

THE FLUOR FOUNDATION ... jtu CHEMICAL ENGINEERING EDUCAT I ON

PAGE 3

EDITORIAL AND BUSINESS ADDRESS Department of Chemical Engineering University of Florida Gainesville, Florida 32611 Editor: Ray Fahien Associate Editor: Mack Tyner Editorial & Business Assistant: Carole C. Yocum (904) 392-0861 Publications Board and Regional Advertising Representatives: Chairman: Klaus D. Timmerhaus University of ColOl'ado Vice Chairman: Lee C. Eagleton Pennsylvania State University SOUTH: Homer F. Johnson University of Tennessee Vincent W. Uhl University of Virginia CENTRAL: Leslie E. Lahti University of Toledo Camden A. Coberly University of Wisconsin Darsh T. Wasan Illinois Institute of Technology WEST: R. W. Tock Texas Tech University William H. Corcoran Califorlllia Institute of Technology William B. Krrintz University of Colorado C. Judson King University of California Berkeley NORTH: J J. Martin University of Michigan Edward B. Stuart University of Pittsburgh NORTHEAST: Angelo J. Perna New Jersey Institute of Technology NORTHWEST:
PAGE 4

Atlanta's skyline is shown behind the Tower of Tech's Administration Building. ti)" a department I CHE AT GEORGIA TECH GARY W. POEHLEIN Georgia Institute of Technology Atlanta, GA 30332 LAST WINTER WHILE browsing in the campus bookstore I found a stack of T-shirts imprinted with the message "MIT-THE GEORGIA TECH OF THE NORTH." I returned two days later to purchase one for my friend Professor James Wei (MIT Department Head and a Georgia Tech graduate) only to find they had all been sold. In their place were T-shirts which bore the message "NORTH AVENUE TRADE SCHOOL" (North Avenue is the southern boundary of the campus). I decided not to buy one of the replacement shirts. It probably would not have been received with Copyr i ght ChE Divis i on, ASEE, 1980 2 much enthusiasm by Professor Wei. Both T-shirt messages contained elements of truth and fiction. Georgia Tech, like MIT, has a long-standing reputation for quality engineering and scientific education. Unlike MIT, Georgia Tech is a State Institution, being part of the University System of Georgia Four colleges (Engineering, Architecture, Sciences and Liberal Studies, and Industrial Management) offer undergraduate and graduate degrees in areas that could be described, in a broad sense, by the term "Technology." De grees are not offered in areas such as music, English, art, history, etc. This last fact is re sponsible for the "trade school" label on the second stack of T-shirts. Of course, those of us associated with Tech know that this label does not reflect CHEMICAL ENGINEERING EDUCATION

PAGE 5

reality. Our liberal studies departments are staffed with high-quality faculty who offer a very wide range of courses to help our students obtain a broad education. The Georgia Tech campus is located on 280 acres about 1 miles from the center of Atlanta. With a population of about 1,750,000, Atlanta is the center of commerce in the rapidly growing Southeast. It is a dynamic, beautiful and exciting city with a very diverse population. Atlanta is an educational center with 29 degree-granting colleges, junior colleges, and universities; a center for the arts with a symphony orchestra, a ballet company, numerous art exhibits, local and im ported theater groups, a wide variety of special festivals; and a center for sports, with teams in all major professional leagues. The School of Chemical Engineering at Georgia Tech is comprised of a Chemical Engineering Di vision, a Metallurgy Division, and the Fracture and Fatigue Research Laboratory, an interdis ciplinary research organization. Faculty and other scientific staff are listed in Table 1. The Chemical Engineering Division offers B.S., M.S and Ph.D. degrees. The Metallurgy Division does not off er a designated undergraduate degree but does have a very active graduate program leading to M.S. and Ph.D. degrees. The School of Chemical Engineering has been changing rapidly during the past few years. Dr. Waldemar Ziegler retired at the end of the 197778 academic year, several faculty have left for other positions and, unfortunately, Dr. Leon Bridger and Dr. Homer Grubb died suddenly during the past academic year. Thus much of our efforts during the past fifteen months have been involved with recruiting new faculty. These efforts have been very successful, with ten outstanding individuals accepting offers to join our faculty. Dr. Edvin Underwood was a Senior Research Scientist here prior to accepting a faculty posi tion. Eight of the remaining nine have moved to Atlanta since April, 1979. Dr. Amyn Teja, pres ently at Loughborough University, will join us in September, 1980. The faculty members and their areas of interest are identified in Table 1. GEORGIA TECH-A BRIEF HISTORY* The Georgia Institute of Technology began as the Georgia School of Technology in 1888. The School of Chemical Engineering evolved from the chemistry curriculum. An Engineering Chemistry Program was published in the 1900-1901 catalog. In addition to chemistry and chemical engineer ing topics, areas such as metallurgy, dyeing, meChemical Engineering is housed in the BungerHinry Building. chanics, electricity, minerals, and industrial pro cesses were included in this early program. The 1929 catalog listed a B.S. in Engineering Chemis try, and the 1930 catalog indicated that a B.S. in ChE could be obtained under the Department of Chemistry. McLaren White, whose father, Alfred H. White, was an MIT graduate and Head of ChE at Michigan, was the first chemical engineer to come to Tech. This happened during the 1920s, and Professor White was part of the chemistry de partment faculty. Changes toward chemical engi neering were too slow to suit Professor White, so he left. The first-prize for the AIChE contest problem is named in honor of McLaren White. The first Chief of the Chemical Engineering Division within the Department of Chemistry was ,: ,The information presented here is condensed from a paper, "An Early History of Chemical Engineering at Georgia Tech," by Marcella M. Lusby, Unpublished (No vember, 1977). The School of Chemical Engineering currently enrolls about 950 undergraduates. This figure includes freshmen who declare a major at Tech and co-op students who are on work assignments. Cooperative education : dates back to 1915 at Tech. Today nearly 30% of undergraduate ChE's are in the Coop Program. WINTER 1980 3

PAGE 6

Dr. Harold Bunger, who started at Tech around 1929. The name of the academic unit was changed to the Department of Chemistry and Chemical Engineering in the 1930s. This department was headed by Dr. Boggs, with Dr. Bunger continuing as Chief of the ChE Division. Both Boggs and Bunger died in 1941 and the two divisions were split, forming separate departments Professor TABLE 1 Staff: ChE Division Pradeep K. Agrawal; Assista.nt Professor; Ph.D. 1979, University of Delaware; Heterogeneous Catalysis Charles J. Aloisio; Lecturer; Ph.D. 1970, Purdue Uni versity; Polymer Engineering and Science William R. Ernst; Associate Professor; Ph.D. 1974, University of Delaware; Heterogeneous Catalysis; Assistant Director for Ch.E. Undergraduate Pro .-.? gr ~ ms :t arry J. Forney; Associate Professor, Ch.E./C.E.; Ph.D. 1974, Harvard University; Aerosol and Particle ,\J f Technology: 1 George A. Fowle~; Adjunct Professor; Retired Vice President, Marketing, B.F. Goodrich Co.; Plastics Pioneer Charles W. Gorton; Professor; Ph.D. 1953, Purdue Uni versiti; Transport Phenomena, Fluidization ;Edwin M. Hartley; Associate Professor; Ph.D. 1973, Georgia Tech; Pulp and Paper Engineering 1 H. Clay Lewis; Professor; Sc.D. 1943, Carnegie Insti tute of Technology ; Chemical Process Design Albert A. Liabastre; Research Scientist; Ph.D. 1974, Georgia Tech; Surface Science Michael J. Matteson; Professor; D. Eng 1967, Technical University Clausthal (Germany); Aerosols, Particle Technology, Air Pollution Control; Assistant Di rector for Ch.E. Graduate Programs John D. Muzzy; Professor; Ph.D. 1970, Rensselaer Poly technic Institute; Polymer Engineering, Energy Conservation, Economics .. Allan S. Myerson; Assistant Professor; Ph.D. 1977, University of Virginia; Thermodynamics, Crystalli zation, Biochemical Reactions Clyde Orr, Jr.; Regents' Professor; Ph.D. 1953, Georgia Tech; Instrumentation and Particle Technology : Gary W. P~ehlein; Professor; Ph.D. 1966, Purdue Uni versity; Emulsion Polymerization, Latex Tech nology; Director of the School Ronnie S. Roberts; Assistant Professor; Ph.D. 1976, University of Tennessee; Biochemical Engineering, Mass T'ran sfer, Reactor Design Robert J; Samuels; Professor; Ph.D. 1960, University of Akron; Polymer Science and Engineering A. H. Peter Skelland; Professor; Ph.D. 1952, University of Birmingham (England); Non-Newtonian Fluids, Mixing and Fluid Dynamics, Heat and Mass Transfer Jude T. Sommerfeld; Professor; Ph.D. 1963, University of Michigan; Computer Applications D. William Tedder; Assistant Professor; Ph.D. 1975, University of Wisconsin; Process Synthesis, Optimi zation and Waste Management Amyn S. Teja; Associate Professor, Ph.D. 1972, Im perial College (London), Phase Equilibria, Thermo dynamics Hend~rson C. Ward; Professor; Ph.D. 1953, Georgia Tech; Transport Phenomena, Process Design, Co Siting Mark G. White; Assistant Professor; Ph.D. 1978, Rice University; Heterogen~ous Catalysis Jack Winnick; Professor; Ph.D. 1963, University of Oklahoma; Thermodynamics, Electrochemical Engi neering, Air Pollution Control Ajit P. Yoganathan; Assistant Professor; Ph.D. 1978, California Institute of Technology; Biomedical Engi neering, Polymer Rheology Alex Zhavoronkov; Visiting Scientist; Ph.D. 1975, Moscow Technical Institute; Aerosols Staff: Metallurgy Division Helen E. Grenga; Professor; Ph.D. 1967, University of Virginia; Catalysis, Corrosion, Extractive Metal lurgy Robert F. Hochman; Professor; Ph.D. 1959, University of Notre Dame; Phys-Chem. of Metals, Corrosion, Biomaterials; Associate Director for Metallurgy John E. Husted; Professor; Ph.D. 1970, Florida State University; Mineral Engineering Miroslav Marek; Associate Professor; Ph.D. 1970, Georgia Tech; Corrosion, Dental Materials Pieter Muije; Associate Professor; Ph D. 1971, Wash ington State University; Metallurgy, Mineral Processing Stephen Spooner; Professor; Sc.D. 1965, Massachusetts Institute of Technology; Physical Metallurgy, Metal Physics Staff: Fracture and Fatigue Research Lab Saghana B. Chakrabortty; Research Scientist; Ph.D. 1974, Georgia Tech; Mechanical Metallurgy, Electron Microscopy Albrecht Gysler; Visiting Research Scientist; Ph.D. 1965, University of Stuttgart (Germany); Micro structure-Properties Relationships Ludmilla Konopasek; Research Engineer; M.S. 1975, Manchester University (England); Fracture and Fatigue of Materials Fu-Shiong Lin; Research Scientist; Ph.D. 1978, Georgia Tech ; Corrosion, Fatigue, and Ti and Al Alloys T. H. B. Sanders; Research Scientist; Ph.D. 1974, Georgia Tech; Aluminum Alloy Development, Micro structure and Fatigue Bhaskar Sarkar; Postdoctoral Fellow; Ph D. 1979, Georgia Tech; Stress Corrosion Cracking and Fatigue Edgar A. Starke, Jr. Professor; Ph.D. 1964, University of Florida; Fracture and Fatigue; Director of the Fracture and Fatigue Research Laboratory Edvin E. Underwood; Professor; Sc.D. 1954, Massa chusetts Institute of Technology; Physical Metal lurgy, High Temperature Deformation and Stere ology CHEMICAL ENGINEERING EDUCATION

PAGE 7

Jesse Mason became Director of the Chemical Engineering Department; and Dr. Paul Weber, a chemical engineering faculty member, was named Assistant Director of the Engineering Experi ment Station. A reorganization took place in 1948 with Pro fessor Mason becoming Dean of the College of Engineering and Dr. Weber the new Director of the School of Chemical Engineering. Dr. Weber held this position until 1955 when he became Dean of the Faculty ( equivalent to the present position of Vice President for Academic Affairs), and Dr. Robert Raudebaugh from the Metallurgy Division was appointed Acting Director of the School of Chemical Engineering for the year. Dr. W. M. Newton assumed the acting director position and The "Ramblin Wreck" Parade, a real demonstration of student creativity-weird things that move! served for about three years until Dr. Homer Grubb was named Acting Director. The name change from "Department" to "School" was made to identify ChE as a degree-granting component of the Institute; the School of Chemical Engineer ing offers degrees, the Departments of English, History, etc., do not. In addition, schools are generally more autonomous than departments. WINTER 1980 One of the reasons for the dramatic, almost three-fold, increase in ChE undergraduate enrollment at Tech during the past five years has been the number of women and minorities entering the School. Dr Grubb was later named Director of the School, a position he held until 1965 when Dr. Leon Bridger became the Director. During this period the Metallurgy Division developed a significant graduate program which operated partially inde pendently of the Chemical Engineering Division. Dr. Robert Hochman came to Tech in 1959 and is currently the Associate Director responsible for the Metallurgy Division. Dr. Bridger returned to a full-time faculty position in ChE during the summer of 1978; and Dr. Gary Poehlein was appointed Director after moving to Tech from Lehigh University. The Fracture and Fatigue Research Laboratory, headed by Dr Edgar Starke, Jr., was also es tablished in 1978. UNDERGRADUATE PROFILE APPROXIMATELY 8500 undergraduates were enrolled at Georgia Tech to start the Fall Quarter, 1979. About one-half of these students are resi dents of the State of Georgia, The others include representatives from every other state hi the : u.s : .,,. and many foreign countries. A s~lective idmis sions policy continues to produce a high-quality undergraduate student body. The 1978 freslitjlaii. class had an average SAT score of 1161 comprised of an average verbal score of 533 and an average math score of 628. Georgia Tech ranks sev enth in the nation in attracting National Merit Sch.olars '),. .,,, and second to Harvard-Radc liffe,in the'-ll.il.:rttJ: } N of National Achievement Scholars enrolled. '/ The School of Chemical Engineering curre ntly enrolls about 950 undergraduates. This figure in cludes freshmen who declare a major at Tech and coop students who are on work assignments. Co operative education dates back to 1915 at Tech. Chemical Engineering became involved later, and Professor Emeritus Waldemar Ziegler was the first ChE Coop graduate receiving his degree in 1932. Today nearly 30 percent of undergraduate ChEs are in the Coop Program. One of the reasons for the dramatic, almost three-fold, increase in undergraduate ChE enroll5

PAGE 8

The undergraduate program in chemical engineering is quite rigorous with a good balance between theory and practice. Required courses include two quarters of transport phenomena, three quarters of unit operations, three quarters ef design, as well as courses in stoichiometry, reaction kinetics, process control, etc. Engineering drawing and physical education remain as required courses. ment at Tech during the past five years has been the number of women and minorities entering the School. These numbers have increased from near zero in 1974 to about 30 percent women and 8 percent minorities in 1979. A second significant reason for our increased enrollment has been more participation in dual-degree programs with other universities. The undergraduate program in chemical engineering is quite rigorous with a good balance between theory and practice. Required courses in clude two quarters of transport phenomena, three quarters of unit operations three quarters of de sign, as well as courses in stoichiometry, reaction kinetics, process control, etc. Engineering draw ing and physical education remain as required courses. Students come to Georgia Tech expecting to work hard, and most are willing to make the Jack Childs checks the assembly of Polymer Fabric Ex truder. necessary commitment to succeed. The vast ma jority leave Tech with a very high opinion of the Institute and a fond feeling for Atlanta. This loyalty is clearly manifested in the fact that Georgia Tech almost always ranks first among the nation's public institutions in terms of support by alumni and alumnae. 6 GRADUATE PROGRAM THE GRADUATE PROGRAM is in a period of rapid growth. The addition of ten new faculty mem bers should provide us with the manpower neces sary to improve an undergraduate program that is already quite good and, at the same time, to build a graduate program of equivalent stature. Enrollment of full-time graduate students in the Fall of 1978 was about 55. These students were about equally divided between the ChE and Metal lurgy Divisions. Present graduate enrollment in cludes 61 chemical engineering students and 40 metallurgy students. New graduate students entering the School of Chemical Engineering in September, 1979, con sisted of 33 U.S. citizens and 9 from other countries. Thirty-six of these students are in the ChE Division and 6 in the Metallurgy Division. A continuation of this sort of success in recruiting graduate students will insure the proper develop ment of our graduate program. We would like to achieve a steady-state graduate enrollment of about 100 ChEs and 40 Mets. with an 80/20 balance between citizens and non-citi'zens. THE FUTURE THE NUMBER OF NEW faces at our first faculty meeting in September prompted a request for round-the-table introductions. Of course, faculty on our staff last year had an opportunity to meet the new faculty during the campus visits. In fact, a number of significant working relationships have already been developed between new faculty and those continuing to serve on our staff. During a year when more than fifty interviews occurred, however names and faces can become confused. This introduction exercise clearly illustrates that the School of Chemical Engineering at Georgia Tech is indeed in "A Period of Rapid Transition." Prediction of the future in such an environment is surely an uncertain endeavor. Our faculty and students are very optimistic. We look forward to an exciting period in the life of an out standing institution. CHEMICAL ENGINEERING EDUCATION

PAGE 9

You know aboot Atlantic Richfield Company. Now start reading between the lines. ARCO Oil and Gas Company <> Division of Atlant1cRich f ieldCompany ARCO Petroleum Products Company <> Division of AtlanticR1chfieldCo m pany ANACONDA Industries 4 Division of The A N ACONDA Company ARCO Transportation Company <> D1v1s1 o n of AtlanticRichfieldCompan y ARCO Chemical Company <> Division of AtlanticRichfie l dCompany ANACONDA Copper Company 4 D iv i sion of The A NACON D A Co m pa n y ARCO International Oil and Gas Company <> D ivision of Atlantic Ri c hf ie ld Com p a n y ARCO Coal Company <> D i vi sio n of Atl a nti c Ri c hf ie ld Co mp a n y With all the companies you should be talking to why not meet eight with one interview? For more information regarding career opportunities with us, see your Placement Office. AtlanticRichfieldCompany <> An equal opportunity employer m/f

PAGE 10

r, Na educator and Biochemical Engineering at the University of Pennsylvania ALAN L. MYERS Uni v ersity of Pennsyl v ania Philadelphia, PA 19104 THE PAST ACADEMIC year served as a milestone for the field of biochemical engineering and the University of Pennsylvania. Not only did the year mark the 25th anniversary of the biochemi cal engineering program at the University, it was also the 25th anniversary of the leadership of the program by its founder at Penn, Arthur E Humphrey. There are older and larger biochemical engi neering programs in existence, but none can rival the record of Penn's program in terms of its dy namic growth and production of graduates who have gone on to become leaders in the field. To mark these achievements, the University hosted a special 25th anniversary symposium to which all the graduates of the program were invited. The "alumni" who attended participated in panel dis cussions on the future of biochemical engineering in the areas of food production ( discussion led by Stanley Barnett), energy production (led by Charles Cooney) production of chemicals (led by Daniel Wang), and preservation of the environ ment (led by Larry Erickson), with a roundtable discussion on the future of biochemical engineer ing education led by Richard Matele s ORIGINS OF THE PROGRAM JN 1953 ART ARRIVED at Penn, fresh from his PhD studies in chemical engineering at Columbia University. While working on his doctorate, he As a scholar of biochemical engineering in general, and fermentation technology in particular, he has published over 150 technical papers and co-authored two textbooks which are regarded as the bible in their fields. C o p y rig ht ChE D iv i si on, A SEE, 1 980 8 Art inspects the fermenter laboratory. developed an interest in fermentation technology and subsequently took additional training in food technology at M.I.T. At Columbia, Art, along with Ernest Henley (now at the University of Houston), was the first doctoral student of Pro fessor Elmer Gaden who was later to be cited as the "father of biochemical engineering" (in a cover story in the May 31, 1971 issue of Chemical Engin e ering N ew s). It was through Elmer Gaden's influence that Art's interest in fermenta tion systems matured. Elmer had written his own thesis on "Mass Transfer in Fermentation Systems under the supervision of Arthur W. ("Pop") Hixson at Columbia, working in conjunc tion with a biochemical engineering team (Karow, Bartholomew, and Sfat) at the Merck Company. (It is of some interest to note that reporting of the 1945 McGraw-Hill Process Development Award, presented to Merck for its "Biochemical Engineering Development of the Penicillin Pro cess," is apparently the first mention of "bio chemical engineering" in the chemical engineer ing literature.) Art joined Penn's chemical engineering faculty CHEMICAL ENGINEERING E DUOA'l'IQN

PAGE 11

as an assistant professor and in the fall term of 1953 offered a new course in biochemical engineer ing. Since 1953 the biochemical engineering pro gram has grown steadily, and in 1972 it was officially recognized when the Department of Chemical Engineering formally changed its name to the Department of Chemical and Biochemical Engineering. THE MAN BEHIND THE PROGRAM As OVERWORKED AS THE WORD "dynamic" is in today's usage, it is nevertheless the appropri ate term to characterize the founder of Penn's biochemical engineering program-Art Humph rey. Whether at work in his laboratory, or lead ing a faculty debate, or scaling a mountain in Guatemala, Art is possessed of energy that leaves others much younger in a race to catch up, and he inspires his contemporaries to set equally chal lenging goals for themselves. In his 25 years at Pennsylvania, Art has es tablished professional credentials that are indis putable. As a scholar of biochemical engineering in general, and fermentation technology in par ticular, he has published over 150 technical papers and co-authored two textbooks which are regarded as the bible in their fields, Biochemical Engineering (Academic Press, 2nd ed., 1976) and Fermentation & Enzyme Technology (John Wiley, 1979). He holds three U.S. patents and has actively consulted for more than 20 chemical companies during his career. He has served on and chaired numerous AIChE committees including the FBP Art seems to approve of his students' efforts at office "redecoration." WINTER 1980 Over the years he has formed a special bond with his studen,s; based on mutual respect, willingness to devote intense effort to a project, and in no small measure on the ability to dish out and take a practical joke. Division, and between 1975 and 1978 was a di rector of the organization. He has also been a past chairman of the MC&T Division of ACS and of the Working Group on the Production of Sub stances by Microbial Means of the US/USSR Committee on Cooperation in Science and Tech nology. In addition, he served from 1971-1973 as a member of the NSF Advisory Committee for Engineering. In 1973 he was elected to the Na tional Academy of Engineering. Such a summary of professional activities, however, doesn't begin to capture Art Humphrey, the man, who by virtue of his enthusiasm and exuberant love for life succeeds in making people want to work together toward a common goal. Art always has a full complement of five doctoral students to work with him as advisees or, more appropriate to the man, associates. Over the years he has formed a special bond with his students based on mutual respect, willingness to devote intense effort to a project, and in no small measure on the ability to dish out and take a practical joke. For pranks he's played on his students, Art was rewarded on one occasion with the careful "re decoration" of his office (as shown in the ac companying photograph) and on another with.!J:ie clamping of a ball and chain on his leg only minutes before he was to make a presentation to the University Trustees. In Houdini-like fashion he arrived before the Trustees on time sans ball and chain. Art is an avid outdoorsman who cheerfully ignores middle age (and some say common sense') as he pursues his interests. Having hiked the trails of much of this country, he frequently substitutes an opportunity to climb a mountain in a foreign country in place of the honorarium he would re ceive for lecturing there. Such negotiations have enabled him to climb Mount Fuji in Japan, Popo catepetl in Mexico, and Augua in Guatemala, among others. Last year when taking up skiing for the first time, Art found himself by mistake in an advanced intermediate class and, not having the good sense to get out of the class, "mostly fell ( 9

PAGE 12

Collectively, this group conducts a research program amounting to $500,000 a year, and this past year they published the program's 150th paper on biochemical problems. down the hill" and received a National Standard Medal for downhill racing as a result. Each year Art leads a departmental canoe trip down the Delaware River, an activity he organized 16 years ago (when Bob Bird was a member of that group). The annual event now involves between 60 and 70 people, including faculty members, students, alumni, and participants from other schools. This year an eight-mile run was included as part of the trip, and he typically outpaced the younger generation-without having trained for the run. The honesty and openness with which Art ap. proaches. ever.y aspect of his life generates an un swerving loyalty among his associates. Never afraid to admit when he's wrong or simply doesn't know an .' answer, he leads people to trust that they'll "hear it straight," good or bad, in dealing with him. Add to these qualities Art's rare talent of being a good listener, and the result is an in dividual who is extremely effective in accomplish ing the work he sets out to do. His spirit has inspired and guided the development of Penn's biochemical engineering program over 25 years, and it has set the tone with which the biochemical engineering faculty now approaches the next 25 years. BIOCHEMICAL ENGINEERING AT PENN TODAY A RT HUMPHREY STILL heads the biochemical engineering program within the University's Department of Chemical and Biochemical Engi neering, despite his heavy administrative load as Dean of the School of Engineering and Applied Science. (From 1962 until he was named Dean in 1972, he served as Chairman of the Department of Chemical Engineering.) Art and three others of the Department's 12-member faculty have the focus of their teaching and research activities in the area of biochemical engineering. At least three others conduct a significant portion of their re search within the field. Collectively, this group conducts a research program amounting to $500,000 a year, and this past year they published the program's 150th paper on biochemical engi neering problems ( out of a total of more than 350 scientific articles published by these individuals). Six national AIChE, ACS, and ASEE awards have 10 been won by members of the group, and two mem bers have been elected to the National Academy of Engineering. Those members of the faculty of the Depart ment of Chemical and Biochemical Engineering who participate in the biochemical engineering program are: David J. Graves: enzyme kinetics Arthur E. Humphrey: fermentation technology Douglas A. Lauffenburger: cell population dynamics Mitchell D. Litt: biorheology Daniel D. Perlmutter: enzyme reactor dynamics E. Kendall Pye: enzyme behavior and purification John A. Quinn: bound membrane systems The facilities which are used in both the re search and teaching functions of the biochemical engineering program consist of four primary laboratories: the fermenter laboratory, a wet chemistry laboratory for enzyme analysis, a mem brane laboratory, and a reactor laboratory. The fermenter laboratory (pictured in an ac companying photograph) centers about a 70-liter highly instrumented fermenter that is coupled to a PDP 11 / 34 computer with a 96K core capacity and three discs (two fixed), a Calcomp plotter and a video screen for data acquisition and control. The facility is supported by a Nuclide Mass Spectrometer and other appropriate analysis equipment. In addition, the laboratory has a number of smaller fermenter units, including a 20-1, 14-1, two 1-1, and five 500-ml systems-most having temperature, pH, foam, and dissolved oxygen control. ACADEMIC PROGRAMS UNDER THE LEADERSHIP of Art Humphrey, the biochemical engineering faculty has always in sisted that the academic program remain a part of the basic program in chemical engineering, allow ing the "biochemical" aspects of the program to emerge from an emphasis on biological processes. Because of the emerging significance of bioconver sion processes in the production of energy, food, and chemical feedstocks and as a means for con trolling the environment, students are eager to in vestigate these problems. The faculty believes the students' eventual careers are better served by CHEMICAL ENGINEERING EDVCATION

PAGE 13

having them pursue these interests from a solid foundation in chemical engineering, rather than by focussing exclusively on a subspecialty. Thus, the undergraduate b i ochemical engineer ing student takes the standard chemical engineer ing curriculum, but biochemistry is substituted for one of the courses in organic chemistry, as is a biology course for the course in nuclear physics. In addition, the student will take two or three courses in microbiology, biological processes, utili zation of wastes, biochemical engineering, or food engineering as his senior technical electives; and his senior research project will focus on a bio logical process. Out of a senior chemical engineering class of 45 students, about 7 of them will be taking courses with a focus in biochemical engineering. Most of these students will either continue their biochemi cal engineering studies at the graduate level or will enter medical school. At the graduate level the student planning a focus in biochemical engineering is also expected to take the core courses in chemical engineering in applied mathematics, transport processes, re actor design, and thermodynamics. The student is then expected to take courses in advanced bio chemistry, molecular biology, and genetics, in addi tion to the four basic graduate courses in bio chemical engine~ring. The basic graduate level biochemical engineer ing courses include: Biochemical Engineering: fermenter kinetics, design and operation Biological Processes: Enzyme Technology: Utilization of Wastes: physics and chemistry of biological processes behavior and utilization of enzymes waste utilization and treatment These courses are taught by the faculty on a rotating basis, i.e., once every other year. They are frequently team-taught with the help of visit ing professors and members of the University's Department of Biochemistry and Biophysics. Students in the graduate program can develop and shape their programs to serve their own par ticular career emphases by selecting additional courses in areas ranging from nutrition to micro biology. They are free to select these courses from throughout the University's graduate and profes sional programs, including its Medical School and School of Veterinary Medicine. WINTER 1980 At any given time, about 15 students in the chemical and biochemical engineering graduate program (which numbers approximately 50 fuU time students) will be focussing their studies in the direction of biochemical engineering. Upon re ceiving their doctoral degrees, these individuals generally seek employment in the food production, pharmaceutical, waste treatment, and chemical process industries, or in academia. When questioned, most of the alumni of Penn's biochemical engineering program say they con sider themselves chemical engineers who have an interest in biological processes, for such is the slant of their curriculum. This may in part ex plain why Penn graduates have never encountered problems on seeking employment. Indeed, a recent graduate of the program looking for a position in the biochemical field received more than ten offers from firms, with several offers in excess of $31,000 a year. FUTURE OF BIOCHEMICAL ENGINEERING: ART HUMPHREY'S VIEW THE FACULTY OF THE biochemical engineering program at Pennsylvania looks forwardto con tinued growth of the program. The field is now Over 60 faculty members and students now participate in ArYs annual canoe trip down the Delaware River. coming into its own and the opportunities are unlimited. Perhaps the enthusiasm of those active in Penn's biochemical engineering program comes across best in the words of the man responsible for its flourishing here-Art Humphrey. "Never has the future for biochemical engi neering looked so bright. This ti; due largely to the energy crisis and the attendant emphasis on the use of renewable r esources, meaning materials of biological origin. It seems fairly eviden C that many of these materials will be wocessed by bio11

PAGE 14

logical means, with the use of enzymes, in order to achieve low temperature, energy saving pro cesses. The environmental crisis and the increas ingly strict limits placed on the use of nitrates, phosphates, and other surface water contaminants mean that more efficient and more complicated waste treatment systems will have to be evolved. Also, wastes will be viewed in the future as valu able resources which can be treated to yield useful materials. "Perhaps the most significant development affecting the future of biochemical engineering is the explosion of knowledge concerning genetic engineering techniques. Not only do we now possess the ability to cultivate both animal alld plant tissue cells in large-scale reactor systems, but we can transfer their genetic information for making various biologically active molecules such as insulin into more easily cultivated bacterial cells by gene splicing techniques. Soon the scientist will be able to create cells with virtually any desired metabolic activity. When this comes to pass, the biochemical engineer will become active in efficiently simulating and optimizing many of nature's special reactions in stainless steel fermenters. In many ways biomass can be re garded as the crude oil of the future. Just as crude oil now serves as the feedstock of the petrochemi 'tal indu~t'ry, from 'barrels of biomass' will come a ~umber of the chemical feedstocks of the future. It would not surprise me to see biomass refineries ~tgthg wjthin the next decade. "I ior one wi11 welcome the change. I believe the chemical engineering textbooks of the future wiTl reflect this change aiid will include examples ctlb iJ.aSs problems along with those from the pe t:r6leum .industry Chemical engineering is a truly broad-based discipline, and I believe it is already demonstrating its concern not just with physical and chemical changes, but with biological changes as well." D tin pl book reviews CONTAC'l' CATALYSIS, VOLS. 1 and 2 Edited by Z. G. Szabo, Elsevier Scientific, 1976 Reviewed by John B. Butt, Northwestern U. This monumental two volume set is an essay of the Catalysis Club of the Hungarian Academy of Sciences with individual chapters contributed 12 POSITIONS AVAILABLE Use CEE's reasonable rates to advertise. Minimum rate page $50; each additional column lilch $20. ( Please mention CEE when responding to ad.) RENSSELAER POLYTECHNIC INSTITUTE CHAIRPERSON. The Department of Chemical and Environmental Engineering at Rensselaer Polytechnic Institute invites applications for the position of Chair person. The candidate must hold the Ph.D. degree, have made significant contributions to the literature of chemical and / or environmental engineering and be nationally recognized. The candidate must have administrative and leadership abilities and be able to represent the depart ment externally. The department is balanced, has an ex cellent undergraduate program, numerous active research programs and seeks "first rank" status over the next decade. In the past five years, the Engineering School as a whole has expanded its graduate programs sig nificantly and currently ranks sixth nationally in total external funding ( over $8 million). The salary is open. The deadline for applications is April 1, 1980. Submit re sume and four letters of recommendation to Prof. H. Litt man, Chairman, Search Committee, 123 Ricketts Bldg RPI, Troy, NY 12181. RPI is an Affirmative Action / Equal Opportunity Employer. by thirteen different authors. In many ways the work is reminiscent of the series "Catalysis" edited by Prof. P. H. Emmett in the 1950's, and it promises to be as useful. True to the title, the entire field of contact catalysis is treated, starting with the fundamentals of solid state science, chemisorption and kinetics in the first volume, with applications concerning preparation, charac terization, and catalytic reaction engineering in the second volume The topics included are treated in quite some detail and in many instances represent current state of the art in both catalysis research and applications. With so many different aspects of the field treated in such detail, it is difficult in a review of reasonable length fo do other than cite certain parts that are of particular use to the reviewer. In this respect, there are particularly fine treat ments of adsorption on solid surfaces and physical characterization methods which eclipse much existing literature. For example, the characteriza tion methods discussed include x-ray diffraction, electron-optical methods, magnetic properties, electrical properties, adsorption, infrared and EPR spectroscopy. Another very useful chapter deals with the preparation of catalysts. This is particularly timely now, since we have attained sufficient abilities in characterization that the long-time Continued on page 44. CHEMICAL ENGINEERING EDUCATION

PAGE 15

Monsanto Drive. It takes you a very long way. This sign marks the road that leads into our International Headquarters in St. Louis .. 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 t .; marketing majors at locations thrqtigl).out 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: Ray Nobel, Monsanto Company, Professional Employm~nt Dept., Bldg. A3SB, 800 North Lindbergh, St. Louis, Mo. 63166. Monsanto An eq u a l o pp o rt u nit y e mp l o ye r

PAGE 16

[j;jl 1ecture THE RA TE OF REACTION: A DEFINITION OR THE RESULT OF A CONSERVATION EQUATION?* ALBERTO E. CASSANO lnstituto de Desarrollo Tecnologico para la Ind. Quimica 3000-Santa Fe, Argentina A quick survey of the classical literature on chemical kinetics and reactor analysis reveals different criteria concerning the definition of the rate of reaction. Relying on historical rather than rational bases, the following "definition" has been used: (1) Eq. (1) was probably derived from the first physicochemical studies on rates of reaction de veloped in constant volume batch systems. This "definition" has been used by Glasstone [1 ], Benson [2], Daniels [3], Laidler [4], Frost and Pearson [5], and Johnston [6], among others. On the other hand, chemical engineers mainly engaged in design problems, distrusted the validity of the "definition," taking into account its in applicability to reacting systems of variable volume. The alternative "definition" was based on the number of moles instead of concentration. The difference lies in defining beforehand an extensive rate of reaction which can then be turned into an intensive property dividing by the reaction volume. r' 1 = d! 1 (2) 1 dN1 r i = Vr dt (3) The greater generality of expression (3) as compared to equation (1) is easily demonstratable, in either conceptual or mathematical terms. This paper is the result of mutually beneficial dis cussions on the subject with my former classmates Charles Allen and Sieg hard Wanke. Any merit to this paper is theirs. Any errors in it are mine. 14 The "definitions" belonging to the kind pro vided by equation (3) can be found in books such as those written by Hougen and Watson [7], Smith [8], Levenspiel [9], Walas [10], Boudart [11], Pannetier and Souchay [12] ; and with explicitly stated limitations by Kramers and Westerterp [13], by Aris [14] and Denbigh [15] as well. Many other cases could be quoted. A good example for analysis is the "Continuous Flow Stirred Tank Reactor" (CFSTR). In this particular case the product is continuously re moved at the system outlet and it becomes evident that at the steady state dN 1 / dt is zero (the product being species i) but the rate of reaction is utterly different from zero. Here, as Denbigh remarks, ex pression (3) is not valid, either. This author tries to narrow the applicability of the "definition" (3) to processes where the only change in reactant i is due to a chemical reaction. This would exclude any other form of physical phenomena causing changes in reactant concentration. The restriction seems valid for the CFSTR and for all forms of diffusive flows. Nevertheless, it does not seem so clear for the "Piston or Plug Flow Reactor" (PFR), in which if, on the limit, dt is taken as the reacting mass average residence-time in the elementary re action volume of length dz, equation (3) may be properly applied with some substitutions This work shows the futility of arguing about the accurate "definition" of the reaction rate and It should be noticed that now v z may be a function of z if changes in the number of moles take place. This work shows the futility of arguing about the "definition" of the reaction rate and the convenience of dealing with the subject from a different viewpoint. Co pyright ChE D i v isi on, A SEE, 1980 CHEMICAL ENGINEERING EDUCATION

PAGE 17

the convenience of dealing with the subject from a different viewpoint, that is, drawing the necessary equations from a more fundamental principle, as the general mass conservation equa tion for multicomponent systems would be. REDEFINITION OF THE PROBLEM FOR HOMOGENEOUS REACTORS S o FAR, IT IS obvious that the existing "defini tions" seem to depend either on the author's personal likings or on the reactor to which the equations will be applied. It seems logical to assume that any reaction rate should be a function not of the system in which it has been determined but only of tempera ture, pressure and concentration of the species participating in the reaction, just to mention the most commonly encountered variables affecting the rate. The reaction rate could be naturally in fluenced by the type of reactor-continuous or batchwise-but only up to the extent that it may affect concentrations, temperatures or pressures. It is necessary to reconcile an expression de rived from the ontological concept of the rate of reaction with a mathematical equation expressing exactly the same, involving the variables suscepti ble of experimental measurement. Ontologically (and considered as an intensive property), "The rate of reaction is the change in the number of moles which takes place in unit time and unit reaction volume, due to a transformation of reactants into products." The mathematical expression of the above de finition will be set equal to the kinetic equation of the chemical system under consideration; i.e. if such a definition is adequately represented by a mathematical proposition which will be called r, it is evident that: r = cf, (C1, C 2 .... etc. T, P, etc.) (4) where cf, depends on the complexity of the reacting system (order, molecularity, activation energy, etc.). It is the concern of this work to determine the adequate formulation for the left-hand side of equation ( 4). The problem at issue has been partially, even though quite accurately, discussed by Petersen [16] and also singularly viewed by Amdur and Hammes [17], particularly for batch reactors. The reader may resort to the quoted references. In many cases, experimental measurements will register not only the change in concentration but also the resultant of the chemical process of reWINTER 1980 Alberto E Cassano is the founder and Chairman of INTEC; Pro fessor at the Universidad Nacional del Literal and Member of the Scientific Research Staff of the National Council for Scientific and Technological Research of Argentina. He received his Chemical Engi neer's degree from the Facultad de lngenieria Quimica of U.N L. (Santa Fe, Argentina) and his Ph.D. degree from the University of California, Davis. His research interests are in Photochemical Re actors and Gas-Liquid Reactions Catalyzed by Solids At present he is also responsible for a consulting contract (undertaken by INTEC) for the Argentinian Atomic Energy Commission engaged in the de velopment of a Heavy Water Experimental Plant. action and the physical processes of diffusion, con vection and volume changes. From this point of view, all the "definitions" previously analyzed are erroneous and inaccurate, and only applicable, at best, to systems universally used but no less par ticular (e.g. the constant volume stirred batch re actor, etc.) Petersen clearly points out that the left-hand side of equation ( 4) is a direct resultant of the ex perimental system employed in kinetic determina tions, and that, therefore, there is no single de finition which could be widely used, still less if the reacting system were non-isothermal. This statement can be effectively generalized and is the subject of the next paragraph. SOLVING METHODS JF THE RATE OF reaction is to be defined, the problem is usually reduced to finding th,e rate of change of the reacting species with respect to some independent variable ( time in a batch reactor and position in many continuous reactors). The rate of change of the reacting species is generally measured considering the change in the number of moles of the said reactant or the rate of change in its concentration. An apparent way of solving the problem would be to consider the reaction within a species ma terial volume, instead of the generally used fixed 15

PAGE 18

control volume. But it is difficult to think of a species material volume since they are not pre served [18]. If the "particles" of the material volume are elements, the difficulty is overcome at the expense of a greater complexity. Therefore, this strategy is quite troublesome. A more adequate way of looking into the problem for an isothermal system is to consider the various mechanisms which can bring about concentration changes of a reactant within a system and, then, determine which part of these total changes is due to the chemical reaction, sub tracting all contributions other than the reaction. This means to state in detail a mass conserva tion balance for a multicomponent system. Gener ally speaking, there are two categories of phe nomena through which a species concentration may vary in a fixed volume in space: (1) The species may appear or disappear by chemical re action and (2) There is a net flow of this species through the area of this volume element. This flow mechanism, be it diffusion, forced convection or any other means of mass transport, needs not be detailed here. The description of this flow will ex clusively depend on the system at stake. The general conservation equation may be thus written (19] : (5) or in a more general way: if a1 is the stoichio metric coefficient of species i: (6) Then: aC 1 + n N 1 = a r at V (7) This well known conservation equation is the only general formulation for homogeneous isothermal rates of reaction that is independent of the re action system being employed. It clearly shows that the rate of reaction is the "source" or "sink" term in the mass inventory; therefore all the re maining non-zero terms resultin'g from taking r oi,_it of expression (7) must be substituted into the right-hand side of equation (4) So far, it is obvious that the existing "definitions" seem to depend either on the author's personal likings or on the reactor to which the equations will be applied. T6 APPLICATION TO COMMON REACTING SYSTEMS EQUATION (7) WILL now be used to obtain the adequate expressions for some classical react ing systems. The following homogeneous iso thermal systems will be considered : 1) Isothermal constant volume batch reactor 2) Isothermal variable volume batch reactor 3) Steady state, isothermal continuous plug flow re actor 4) Steady state, isothermal continuous flow stirred tank reactor In every case, the adequate assumptions will be made in order to allow an analytical description of the flux Ni so as to simplify the rate expression. ISOTHERMAL CONSTANT VOLUME BATCH REACTOR IF WE ASSUME THAT properties are constant in the whole volume of the reactor (especially con centration and temperature), then the divergence of the flux becomes zero since N1 will be inde pendent of position. Therefore, if 'v N 1 = 0, equa tion (7) is reduced to: (8) where the derivative is now total insofar as con centration will be a function of time only. ISOTHERMAL VARIABLE VOLUME BATCH REACTOR IF WE ASSUME THAT the expansion of volume is slow so that, as in the previous case, such proper ties as concentration, pressure and temperature are independent of their position within the re actor, no diffusional effects whatsoever will result, and the flow will be caused only by expansion. In this case, by definition: N1 = C r V1* but, as all species expand at the same rate: the velocity of all species coincides with the global velocity of the system. With the previous relationships, equation (7) results: aC1 at + '\J (C1 v) = al r But, since it has been assumed that concentration is independent of position 'v C 1 = 0. Then: a 0 ~ 1 + C1 ( 'v v) = a ; r (9) Continued on page 48. CHEMICAL ENGINEERING EDUCATION

PAGE 19

THE MOST VERSATILE RAW MATERIAL ON EARTH. Modern technology-much of it developed by Union Carbide transforms oil and natural gas into thousands of vital petrochem ical products: synthetic fabrics, plastics, even medicines. Yet, sadly, most oil and nat ural gas are burned up, as fuel. Only four percent goes into petrochemicals At Union Car bide, we're working on ways to make that four percent go as far as we can, to build a better future for more of the world's people. ~~~ .;:..,' PESTICIDES TO HELP FEED THE WORLD. Union Carbide pesticides are registered for control of a wide variety of crop-destroying pests. Around the world, they're helping increase yields of rice, soybeans and other basic foods. BRIGHT FUTURES. Millions of pounds of Union Carbide solvents, resins and latexes help modern paints protect and beautify just about every thing under the sun. And the rain. An equal opportunity employer JUST WHAT THE DOCTOR PRESCRIBED. The purity of the medicines you depend on may well begin with Union Carbide ingredients. We provide basic materials for everything from aspirin to sedatives and antibiotics. PLASTICS THAT SAVE GAS. FABRICS THAT SAVE BOTHER. Increasingly, we re using oil to save ga s by turning it into strong, light plastics that trim cars' weight and boost their mileage. Union Carbide also supplies automakers with petro chemicals for everything from resilient urethane bumpers to polyethylene electrical insula tion. We even make Prestone II anti-freeze. Your carefree synthetic fabrics are petrochemicals, too. Union Carbide makes one of the basic ingredients in polyester, the most popular synthetic of all. UNION CARBIDE WORKING WITH NATURE TODAY, FOR THE RESOURCES WE'LL NEED TOMORROW. Union Carbide Co rp oration, 270 Park Avenue, NewYo rk, N.Y. 10017

PAGE 20

[eJ n a classroom A FULL-YEAR COURSE SEQUENCE IN REALTIME COMPUTING D. A. MELLICHAMP University of California Santa Barbara, CA 93106 THE CHEMICAL ENGINEERING program at U.C. Santa Barbara began in the mid-1960's with an intended emphasis on the process control part of the curriculum. The use of digital computers for data acquisition and control, what often is re ferred to as real-time computing, at that time was an object of relatively intense interest in industry. However, except for one or two schools with the financial resources to acquire an industrial-scale computer (e.g. IBM 1800), there was little op portunity for universities to give students hands on experience with data acquisition and control computers. The appearance of the minicomputer in the late 1960's made it possible for virtually any depart ment to acquire or build a real-time computing facility which could be used in teaching and in re search. The more recent introduction of the microDuncan A. Mellichamp is Professor of ChE at the University of California, Santa Barbara, where he was initially responsible for de velopment of the process dynamics and control program. An early interest in computers led subsequently to activities with the CACHE Real Time Task Force (Chairman 1974-76), to the CACHE Corp. Board of Trustees (President 1977-78) to the Editorship of the CACHE Mono graph series in Real-Time Computing, and to a long succession of aggravations associated with building and maintaining a real-time laboratory facility He would like to forswear all future activities in volving computers, but cannot. 18 processor (and microcomputer) has merely ac celerated the trend of chemical engineering depart ments to install real-time facilities and to introduce elements of real-time computing into the curricu lum. In the past few years, the early fascination with hardware and software among real-time TABLE 1 Topics Covered in Undergraduate Control Courses (2 hr. lecture and 3 hr. lab per week for two quarters) Derivation of process dynamic models Transfer function models Openand closed-loop systems Frequency-response methods for controller design Process applications Overview of advanced control methods users has decreased to some extent, and the more important educational questions of what, where, and how to teach this new subject area are re ceiving more attention. Since real-time computing was and is tied to the subject of process control, it seems reasonable to explore some ideas concern ing the teaching of real-time computing within that context. The chemical engineering depart ment at U.C. Santa Barbara has developed what must be one of the most extensive teaching and research programs in real-time computing in this country (at least among chemical engineering de partments) ; and I would like to describe it, to discuss some of the background behind its de velopment, and to note how it is changing. To begin with, however, it might be useful to discuss the process dynamics and control program which now coordinates with the more recently developed real-time computing program. THE PROCESS CONTROL PROGRAM JN DESCRIBING THE PROCESS control program at UCSB, it probably is fair to say that it is a tra ditional one. Two one-quarter undergraduate courses are required of all seniors and the usual C o pyri ght ChE D i v isi on, A SEE, 1980 CHEMICAL ENGINEERING EDUCATION

PAGE 21

If there is any non-traditional aspect of the undergraduate program it would be the emphasis on "practical" experiments. The associated laboratory contains several experimental units for the study of liquid level and stirred tank heating dynamics and control, pneumatic and electronic control systems, and simulation facilities .... range of topics is covered. Depending on who is teaching the course, there may be more or less emphasis on process dynamics and on "advanced control" topics ; nevertheless the core areas listed in Table 1 are fairly rigorously and extensively covered. If there is any non-traditional aspect of the undergraduate program it would be the emphasis on "practical" experiments. The associated labora tory contains several experimental units for the study of liquid level and stirred tank heating dy namics and control, pneumatic and electronic control systems, and simulation facilities for other, more complex processes. The experimental units are all bench scale and designed to have time constants on the order of one to two minutes. Since individual dynamics and control experiments can be carried out in 15 to 30 minutes, a whole range of experiments (summarized in Table 2) can be run during the two quarters. Students write up their results in brief (i.e. memo) form; although theory underlies all experimentation, emphasis is on the use of theory to evaluate practi cal consequences such as "Will the surge tank system overflow?", "Is the catalyst mixing system adequately stirred?", etc. By the end of two quarters students will have designed and tested TABLE 2 Undergraduate Process Dynamics and Control Laboratory Experiments FIRST COURSE Liquid level system: 1. Step response 2. Pulse response Stirred tank heating system: 3. Step response 4. Transportation lags Stirred tank reactor (simulated) : 5. Model parameter fitting 6. Steady-state optimization SECOND COURSE Controllers: 1. Dynamic characteristics of three-mode controllers 2. Control of 1stand 2nd-order systems Stirred tank heating system: 3. Frequency response 4. Closed-loop control system design Liquid level system: 5. Closed-loop control system design WINTER 1980 TABLE 3 Topics Covered in Graduate Control Courses (3 hr. lecture each week, per course) FIRST COURSE Derivation of models for multivariable systems Formulation of state space models Solution of multivariable system models (matrix methods) Modal analysis Design of controllers using modal theory Simulation and computer-aided controller design Discrete systems analysis SECOND COURSE Sampled data systems Sampled data controller design (or digital control algorithms) Decoupling control systems Controllability, observability, etc. Optimal control: quadratic and time-optimal Analysis of stochastic systems Observers, filters and state estimators control systems for each of the laboratory bench scale processes using: (1) theory only, (2) em pirically-determined process models, and (3) on line (loop-tuning) methods. At the graduate level we presently offer two courses in process dynamics and control. In recent years the first course has covered both time and frequency-domain methods with emphasis on state space techniques used in conjunction with the com puter for analysis and design. The course deals substantially with multivariable systems and, at least in part, parallels the undergraduate courses at an advanced level (Table 3). The second graduate course covers advanced control topics exclusively (also summarized in Table 3). Although the graduate offerings are traditional in nature (there are no laboratory ex periments at the graduate level), the rigorous yet extensive coverage of advanced materials reflects the same approach as at the undergraduate level. THE REAL-TIME COMPUTING PROGRAM o NE OF THE MAIN PROBLEMS in bringing new material into the curriculum is that (usually) an equivalent amount of old material will have to come out. In developing the real-time teaching program we began with four basic tenets: 1. Real-time computing instruction will be offered to both 19

PAGE 22

undergraduate and graduate students on an elective basis. 2. Real-time computing course wor k will supplement, not replace, existing process control course work. 3. Real-time computing will be taught from a fundamental point of view. Students will be expected to understand basic hardware and software structures and how they are used. 4. Lectures in real-time computing will be paralleled by a "hands-on" laboratory with appropriately-designed ex periments. We have followed these principles substantially down to the present day; hence a few words of dis cussion might be appropriate: Tenet 1 arose out of an early realization that many chemical engi neers, not just a few, will be involved with on-line .... the ratio of "outsiders" to chemical engineers was about 30 / 70 in this second year; through the last academic ye ar it has : been more like 70/30, with the number of chemical engineering students relatively constant at 12-15. process computing as part of their professional careers. Tenet 2 was based on a natural reluctance to tamper with established control courses, in par ticular to remove some significant amount of ma terial so as to introduce lectures on real-time com puting. Tenet 3 might well be open to argument but has its parallel in the controls area: twenty years ago many people felt that control theory was not chemical engineering; probably no one today would argue in favor of a control course based totally oh an empirical approach. Our experience with chemical engineering students is that they do not, in general, like to spend time on computer fundamentals; nevertheless those who do find that practical applications are much easier to under stand and are able to transfer their knowledge to othe r. real-time systems much easier. Tenet 4 will need no explanation. Historically, we began our real-time course offerings with a one-quarter graduate seminar in Spring 1972 and followed it immediately in the :fall of that year with a senior-level elective course open to graduate students. Development of the course and the associated laboratory were under written by the National Science Foundation thrQugh two grants totalling almost $100,000 Q ~~!i \ a three and one-half year period. In the Fall ~ fo l,973 the course was elected by about twice as many chem,ical engineering students (ten). $~ve:r;,al electrical engineering students also took 20 the course; in fact the ratio of "outsiders" to chemical engineers was about 30 / 70 in this second year; through the last academic year it has been more like 70 / 30 with the number of chemical engineering students relatively constant at 12-15. In the past three years we have had to restrict enrollment because of the limitation in our real time laboratory :facilities. One of the interesting developments that came out of the real-time computing course resulted from the large number of student requests for se quel courses in the same area, but covering more advanced topics. Many of the requests came from electrical engineering or computer science students who claimed that there were no equivalent appli cations courses within their own departments. Additional requests came from some of our own students, both undergraduate and graduate, who planned to work in the process computer control applications areas. The real-time field naturally divides into three applications areas: ( 1) single process / single computer, e.g. the topics covered in our first course, (2) multiple processes / single computer (multitasking or multiprogramming ap plications), (3) multiple processes / multiple com puters (multiprocessing or networking). So far as the author can determine, it generally is the case that the real-time instruction offered by most com puter science departments is (a) non-existent, (b) concerned only with on-line systems, e.g. airline reservation systems, (c) theoreticallyrather than practically-oriented, e.g. concerned with hypo thetical job scheduling problems in a multipro cessor environment. Condition (c) holds on our campus; hence in order to accommodate student requests we decided to add two additional courses to our offerings to cover substantially multitask programming and operating systems for real-time applications, and networking and digital computer control systems. Several points are worth noting here concern ing the decision to expand the real -time computing course to a full-year sequence: Some of the specialized computer-oriented material we now teach is outside the area of expertise of most chemical engineering faculty even though the applica tions-oriented material is not. We have avoided po tential problems somewhat by using a Teaching Associ ate, a Ph.D. candidate in computer engineering, to share teaching responsibilities and to supervise the labora tory. In the four years we have offered the full se quence, several students working on joint research projects involving the real-time laboratory facilities have been supported financially in this way and have CHEMICAL ENGINEERING EDUCATION

PAGE 23

co n tributed significantly to the development of our teaching and research program. Attenuation of students enrolled in the sequence historically wa s relatively high, running 40-50% per quarter. Hence by the third quarter the enrollment might have dropped from approximately 40 to about 10; most students continuing through the entire se quence have been our own graduate students, chemical engineering undergraduates who have accepted jobs involving a process control starting assignment, com puter science undergraduates and undergraduate or graduate electrical engineering students with an interest in computer applications. Chemical engineering students who have taken the real time sequence along with the required courses in dy namics and control have relatively little difficulty finding employment in process-control-related areas. Several process oriented companies now recruit process control engineers actively at Santa Barbara and, if statements from recruiters can be believed, would have hired about twice as many students for control work last year if they had been available. G The mixing of chemical engineering students who have relatively little computer background (in general only experience in programming a higher-level language, i.e. FORTRAN) together with computer science students who have little or no experience with physical equipment never was totally satisfactory. The distribution of abili ties in any particular prerequisite subject area is in variably bimodal: e.g. ChE students will have relatively little background in binary arithmetic and logic (com puter science students will feel they have mastered the subject) ; the reverse situation is true in the area of physical measurements and measurement errors. This situation led to major problems in the introductory course where so much of the lecture material must cover topics which will appear to be e l ementary to a computer science major. This year, for the first time, we have not per mitted computer science students to take the intro ductory course. The entire sequence has been re arranged somewhat to reflect these new develop ments. These actions represented an attempt to re turn the first course to what it originally was an introduction for chemical engineers. At the same time we hoped to retain a reasonable enroll ment of "outside" students in the two following courses. This hope did not grow out of any purely altruistic motivations; rather the presence of outside students furnished the department with a claim on the additional teaching staff resources necessary for us to offer such an extensive pro gram. Also, in a rapidly changing field such as real-time computing, the presence of relatively advanced computer science students in our courses has kept the discussions lively and the lectures more nearly "state-of-the-art." The success of these changes is now apparent as will be noted in the sequel. In any case, this rather lengthy descripWINTER 1980 TABLE 4 A First Course in Real-Time Computing (3 hr. lecture and 2 hr. lab per week) Introduction to B A SIC and to real-time BASIC Structure of real time systems Measurements, transducers, and signal handling Number systems and computer arithmetic Introduction to computer architecture and hardware Input/output systems: ADCs and DACs 1 ISA FORTRAN Device controllers and device drivers tion of the development of the real-time sequence is intended to motivate the description of the present courses which follows immediately. Real-Time Computing Courses. There is no "traditional" first course in real-time computing; we have, after much experimentation, settled on coverage of the topics listed in Table 4. From the table it can be seen that we spend considerable time on computer fundamentals; number systems and digital arithmetic, digital logic and hardware, compute r architecture, interfacing, assembly language programming, interrupt handling, etc. We also spend time on some topics whi'ch have long since been dropped from most process control courses; measu r ements and measurement errors, transduction, signal transmission, etc. In a first course of this sort the emphasis is on single process / single computer systems and the coverage must, unfortunately, be light. Our purpose is to develop a basic understanding of all the elements in a real-time system, how these interact, and how they comprise the whole. Our purpose in teaching this course has not been to treat real-time computing as an isolated subject area but to teach it so that the material can be integrated into the control courses, .at least into the undergraduate process -. dynamics and control laboratory. Since the introductory real time course is taught in Fall quarter and precedes the two-quarter sequence in dypamics and control, students normally are in a position to make im mediate application. Those students who have elected to take the real-time cou 'i se are "permitted '. to run all of their dynamics (data logging) and control experiments using one of the real-time computers. Although this normally requires more work of the student-Qutside reading, .. pro gramming, debugging programs, etc.---0tir ex perience shows that they often take this oppor21

PAGE 24

C> FALL INTRODUCTION TO REAL-TIME COM PUTIN G (NON EE /CS) LABORATORY WINTER REAL TIME CO MPUT I N G I LAB O RATOR Y PROCE S S DYNAMICS ANO CO NTR OL I { REQUIRED ) LABORATORY w ..1 ADVANCED 5 _...------' "--41 PROCESS D Y NAMI CS 5 i AND CO NTROL $5~ I SPRI N G REAL-TI ME CO MPUTIN G 11 LAB ORATORY PROCESS DYNAM ICS ANO CO NTR OL 11 {REQUIRED) LABORATORY ADVANCED PROCESS DYNAMI CS AND CO NTROL II FIGURE 1. Process Dynamics and Control and Real-Time Computing Offerings. tunity. Figure 1 indicates how the real-time se quence fits together with the undergraduate and graduate process control program, also how the three different groups of students-undergradu ate and graduate chemical engineers and "out-of department" (C.S., E.E., etc.)-can be accommo dated. The remaining two courses in the sequence (labeled "Real-Time Computing I and II" in Figure 1) cover the major areas of multitask and multiprocessor applications, respectively, with an emphasis on applications of computers, either singly or in networks, for control purposes. Tables 5 and 6 furnish a brief description 'Of the course content for each of these courses; there necessarily must be a small degree of repetition to bring entering students up to operating speed. Real-Time Computing Laboratory. The present real-time laboratory (shown in Figure 2 with one 22 of the undergraduate process dynamics and control experiments visible in the background) contains three minicomputers, two of which are configured for real-time operations. These facili ties, built up over the past ten years, will be sub stantially replaced in early 1980 by the single computer system shown schematically in Figure 3. This multiprogrammed system will accommo date up to six real -time user programs in main memory simultaneously and, potentially, can be ex panded to handle many more if program swapping using the fast disk can be tolerated. Features of the new system will include a link with the main campus computer, full graphics capabilities, a dial up facility for one remote user, and one or more terminals in remote study rooms and laboratories. TABLE 5 Second Course in Real-Time Computing (2 hr. lecture and 2 hr. lab per week) Overview of real-time computing Introduction to real-time FORTRAN Analog and digital input/output Operating systems and schedulers Introduction to multitask programming Multitask program design File handling and bulk storage Assembly language device driver routines Multitasking applications As part of the instructional laboratory we have constructed several interesting auxiliary units: a set of input and display panels for experiments in volving input, output and conversion of analog and digital quantities; an air pressure experiment with binary inputs (solenoid operated valves) and outlays (pressure-operated) relays for instruction in digital 1/0 and si mple on-off control; a metal bar heated at one end, with eight temperature FIGURE 2. The Real-Time Computing Laboratory (Foreground). CHEMICAL ENGINEERING EDUCATION

PAGE 25

TABLE 6 Third Course in Real-Time Computing (2 hr. lecture and 2 hr. lab per week) Real-time computers in process control Controller and filtering algorithms Controller design and applications Overview of computer networks Network architectures Interprocessor communications Distributed processing Networks in process control sensing elements located along the unheated section for multipoint data logging studies; a fully interfaced model railroad designed to demonstrate the control of multiple, largely-random processes. A list of experiments which typically would be performed as part of the introductory real-time course is given in Table 7. The model railroad is used as the basis of a sequence of five experiments in the second course. In the third course, several of our present computers as well as the new system will form the basis for the networking portions of the course. The stirred-tank heating systems are used for the process control portions. Students completing the laboratory sequence will, as a final project, put together a two-computer real time system (one computer for data acquisition and control, the other for process operator com munications and report generation) with inter processor communications carried out over an existing multiprocessor bus. SUMMARY .6.ND CONCLUSIONS C HEMICAL ENGINEERS WHO plan to work closely with digital computer control systems need a much more fundamental' exposure to real-time systems principles than can be obtained through a brief exposure as part of a senior-level control course. Even a single-quarter course in real-time computing cannot cover important advanced topics in the field such as real-time operating systems, multitasking, multiprogramming, and TABLE 7 First Course in Real-Time Computing: Laboratory Experiments Calibration of a resistance thermometer for a stirred tank heating system Estimation of dynamic measurement error in the stirred tank temperature transducer Automated number conversions Digital input/output: "super pong" Analog input/output: simulation of a staircase ADC Data logging of the heated bar temperature profile Data logging and control of the pressure tank WINTER 1980 ... in a rapidly changing field such as real-time computing, the presence of relatively advanced computer science students in our courses has kept the discussions lively and the lectures more nearly "state of the art." CAMPUS COMPUTER CENTER SENSOR 1/0 4-16 BIT IN { 16 BIT OUT } 4 2 PULSE OUT 32-A/D 20-D/A EXPERIMENTAL PROCESSES REMOTE I GRAPHICS TERMINAL PLOTTER FIGURE 3. The UCSB Real-Time Computing System. networking. At Santa Barbara we have expanded our offerings to a three-quarter sequence with heavy emphasis on laboratory exercises and ex periments. As a service course, the sequence attracts enough outside stud@ts to warrant au mented teaching support staff from the college. Also, chemical engineering students appear to benefit from the experience of working with com puter specialists; still it is clear that mixing them at too early a stage is not optiinum. The move we have made to restrict enrollment to chemical engi neers in the introductory class has eliminated most of the problems arising from mismatches in basic skills. This year, considerably more of our own students elected to take the first course than have in the past, and more of them are continuing in the sequence. Additionally, the teaching loads over the entire academic year have been considerably smoothed out by closing the first course to EE and CS students. 23

PAGE 26

TEXT MATERIALS Before 1977-78 there was available virtually no single source of information which could be used as a text in a cou rse on real-time computing. In 1977 the first volumes of the CACHE Monogra ph Series in Real-Time Computing, edited by this author, appeared The monograph series, an attempt to produce a definitive treatment of each major area in the field, will consist of eight volumes initially. All of them will be available in 1979. A listing of the titles in the present series: I. An Introduction to Real-Time Computing II. Processes, Measurements and Signal Processing III. Introduction to Digital Arithmetic and Hardware IV. Real-Time Digital Systems Architecture V. Real-Time Systems Software VI. Real-Time Applications Software VII. Management of Real-Time Computing Facilities VIII. Process Analysis, Data Acquisition, and Control Algorithms In 1978 a good book dealing with industrial applica tions of real-time computers was published: Minicomputers in Industrial Control, T. J. Harrison (Editor), Instrument Society of America, Pittsburgh (1978). With r e spect to the laboratory facilities, the author earlier documented the three real-time laboratory experi ments developed at Santa Barbara in reports, a few copies of which are still available: D. A. Mellichamp and F. Kayihan, Th e Tank Pressure Ex periment, UCSB Department of Chemical and Nuclear Engineering Report C-74-1, 79 pp, (August 1974). D. A. Mellichamp and G. P. Engelberg, The Digital Com puter Controlled Model Railroad, UCSB Department of Chemical and Nuclear Engin ee ring Report C-74-3, 119 pp, (October 1974). D. A. Mellichamp and T. W. Moore, The Heated Bar E periment, UCSB Department of Chemical and Nuclear Engineering Report C-76-1, 96 pp, (March, 1976). D ( eJ ;j :I book reviews CHEMICAL REACTOR DESIGN FOR PROCESS PLANTS; VOLUME I, PRINCIPLES AND TECHNIQUES; VOLUME II, CASE STUDIES AND DESIGN DATA By Howard F. Rase Wiley-lnterscience, New York, 1977 Reviewed by Charles H. Ware,. Jr., Commercialization Insights, Poughkeepsie, N.Y. .. The author has written this book, as the pre :fa~e states, "for the professional engineer who either daily or periodically must deal with design or op~ration of chemical reactors. But in addition to serving as a reference in the personal -libraries of professionals, it should also be useful as a text book for advanced design courses, including cou~ses fa~ught in continuing education." It will serve all of these purposes very well. 24 Volume I (772 pages) is divided into four parts: basic data and principles of design; general aspects of reactor design; single-phase reactors; and design of reactors for multiphase processes. Volume II (242) pages consists of 14 case studies including three oxidation reactions, two polymeri zations, and two hydrogenations. Part 1 is devoted to reaction rate theory and applications; chemical and physical aspects of catalysis and catalysts ; idealized models of re action rates and reactor performance; and ex perimental methods and equipment for developing design data. Two chapters devoted to catalysis and catalysts provide a good summary of them with attention to both theoretical foundations and practical considerations. Experimental methods and equipment to obtain chemical reaction data free of transport effects are emphasized. Part 2 is concerned with selection of reactor type and mode of operation based upon yield and safety, as well as general design considerations such as mixing of reactants, flow distribution, residence-time distribution within reactors, and briefly, vessel design. Part 3, which comprises almost half of the text in Volume I, covers the design of CFSTRs, tubular, batch, semi-batch, fixed-bed catalytic, fluid-bed catalytic, and many special reactors. In addition to the various design equations, there are numerous drawings of actual reactors and con siderable attention is given to flow and heat effects, feed systems, pressure drop, scale-up, start-up and shutdown procedures. Part 4 consists of an excellent chapter on gas liquid reactors plus a short account of liquid-liquid reactors. In the former, stirred tanks, sparged vessels, plate and packed columns, trickle beds, and pipeline contractors are considered. Many theo retical and practical aspects are discussed: scale up, heat transfer, power consumption, pressure drop, design models and procedures, hold-up, mass transfer, dispersion, liquid distribution, and many others. Only the agitated reactor is treated in the last chapter. The case studies of Volume II have been se lected to illustrate various types of design problems. They are indicated in each case, often accompanied by a comment on the principal weakness of the design. The data that are needed are presented, along with intermediate results, alternatives, and bases for decisions. Continued on page 47. CHEMICAL ENGINEERING EDUCATION

PAGE 27

1 6At DuPont you dont get lost in a big company atmosphere. It's very personal:' 1ll "Du Pont is a big com pany but it's broken down into satellites. So you don t get lost in a big-company atmosphere. It's very personal and I think the people are top-notch. I started in technical here at the Belle Plant in West Virginia Now I'm a production supervisor. Production is solv ing problems on a day-to-day basis. I like working under that kind of pressure. When things -George D. Peterson BS Chemical Engineering work out it's very rewarding. So is working with people. I'm responsible for helping 22 peo ple do their jobs : George was recruited by Du Pont from the Michigan Technological University campus in 1973. He interviewed about 25 companies George s story is typical of many Chemical Mechanical and Electrical Engineers who ve chosen careers at Du Pont. We place no limits on the progress our engineers can make A 11 d we place no limits on the contribu t ion they can make-to themselves the Company or to society. If this sounds like your kind of company do what George Peterson did Talk to the Du Pomt representative who visits your campus Or write : Du Pont Company Room 35972, Wilmington DE 19898 At Du Pont ... there's a world of things YOU can do something about ~Ci I.J S AII.T. 8 T M !)Ff An Eq u a l Oppo rtuni ty Emp l oyer M / F

PAGE 28

r---,--------------------[i) ;j :I laboratory ADVANCED PROCESS CONTROL EXPERIMENTS PRADEEP B. DESHPANDE, W. L. S. LAUKHUF and NANDKISHOR G. PATKE University of Louisville Louisville, KY 40208 JN A COURSE ON ADVANCED process control a substantial portion of the time is spent in discussing the fundamentals, design, and implementation of advanced control concepts. When the course was first offered several years ago, computer simula tions were used to demonstrate the concepts. While simulation is certainly a very valuable tool in the analysis and design of control systems, the students felt that it would have been much more satisfying if they had a physical process to work with. In subsequent years, several laboratory ex periments were developed to eliminate this de ficiency. The equipment for the process, around which the present experiments were developed, was con structed from data provided by Exxon Oil Company (then Humble Oil & Refining Co.) on an identical setup at their Bayway refinery. [1] The rig was used by Exxon to train their instrument and process personnel. It has four of the most FIGURE 1. Schematic of Process Control System (Arrows indicate Signal transmission between the process and the computer). Copyripht ChE Div isi on, ASEE, 1980 26 It has four of the most commonly encountered control loops in process industry, i.e., liquid-level, flow, pressure, and temperature. commmonly encountered control loops in process industry, i.e., liquid-level, flow, pressure and temperature. The equipment was used to demon strate feedback control concepts for these loops. Additional instrumentation has been added to the apparatus at the University of Louisville in order to demonstrate advanced control concepts. EQUIPMENT & INSTRUMENTATION A schematic of the thermal process unit is shown in Figure 1. The process involves heating of a continuous stream of water by steam. A vertical cylindrical tank approximately one foot in diameter is located in the center of the unit. The tank contains a steam pipe in the form of a vertical U tube. Water flows continuously in and out of the tank where it is heated by steam. As shown in Figure 1, the process is instru mented with conventional controllers as well as with computer control hardware. Three variables can be controlled in this process: the flow rate of water into the tank, the level of water in the tank, and the temperature of water in the tank. FLOW CONTROL LOOP This loop regulates the flow of cold water into the tank. Supply water passes through a 7 /16inch diameter orifice mounted in a 1/2-inch pipe, then through a 1 / 2-inch control valve made by Uniflow Valve Corporation, and into the top of the tank. The differential pressure across the orifice is transmitted to a mercury manometer (FI in Figure 1) and to a Honeywell flow indicating transmitter (FT). The pneumatic 3-15 psig out put of the transmitter is fed to a flow recording controller (FRC) It is also fed to an AMTEK pneumatic to voltage (P / E) transducer. The electrical output of the P / E transducer is conCHEMICAL ENGINEERING EDUCATION

PAGE 29

nected to one of the analog-to-digital (A/D) con verter channels of the control computer. As indi cated in Figure 1, the position of a Foxboro air switch determines whether the transmitter output is fed to the conventional controller or to the control computer. The conventional flow recording controller is a Honeywell, proportional + reset type, controller which sends a signal to an "air-to-open" control valve on the process unit. Alternately, the signal to the control valve may also come from one of the digital-to-analog converter channels on the control computer, via a Fisher E/P transducer. Again, the position of an air switch determines whether the signal to the valve comes from the conventional controller or from the control computer. LIQUID-LEVEL CONTROL LOOP Water level in the tank is controlled by manipulating the flow of water out of the tank. The level sensor infers the liquid level by measur ing the pressure required to cause air bubbles to form slowly at the bottom of the tank. This pres sure signal is fed into the high pressure side of a Pradeep B. Deshpande is associate professor of ChE at the Uni versity of Louisville. Prior to coming to Louisville, he was with Bechtel Inc., at San Francisco, CA. He has served as a consultant to Mobil Exploration Norway, Inc., and to other companies in the areas of control systems design and simulation. He has a Ph.D. in ChE from the University of Arkansas and is a registered control systems engi neer in California. (L} Walden L. S. Laukhuf is an associate professor of ChE at the Uni versity of Louisville. Prior to coming to the University, he was a Captain in the Air Force at the Air Force Materials Laboratory in Dayton, OH. He has a BChE, MSChE and a PhD in ChE from the University of Louisville and is a registered professional engineer in Kentucky. (C} Nandkishor G. Patke is presently working on a Ph.D. in ChE at the University of Louisville. His research interests are process model ing, simulation and control. He has a B.Tech. in ChE from the Indian Institute of Technology, Kanpur, India. (R} WINTER 1980 steam c ontroller contro l 1 -------7 valve Sff + _IG(sl7 i X(Sl ~GJCS),___ __ point I L_ ______ I measuring element Lood, L(s) FIGURE 2. Block Diagram of the Temperature Control System Foxboro differential pressure transmitter (LLT), set at a range of 40 inches of water. The low pressure side of the transmitter is vented to the atmosphere as is the surface of the liquid in the tank. Thus, the differential pressure transmitter output is proportional to the liquid level. The signal from the transmitter is fed to a Honeywell proportional + reset controller (LLRC) and to one of the A / D converter channels via an air switch which determines whether the loop will be on conventional control or on computer control. The outputs of the controller and a D / A con verter channel are fed to an air switch and then to a 3 / 4-inch, "air-to-close" control valve made by the Uniflow Valve Corporation, installed in the drain line from the tank. The air switch selects computer control or conventional control. A 1/4hp, Barray pump in this line insures sufficient fluid pressure on the upstream side of the valve. TEMPERATURE CONTROL LOOP The temperature of water near the bottom of the tank is measured by an iron-constantan thermocouple immersed in an oil-filled well ex tending into the bottom portion of the tank. The voltage produced by the thermocouple is converted by a Honeywell electropneumatic transducer (TT) 27

PAGE 30

101 !1-gl 93 0:: w a.. ::;; 85 w ... 77 I I I -output\ > V 0 4.8 I"'------in putI 8.8 TIME(MIN) 12 8 16.8 l> ;ij cl 1l 15 1l ;o 10 Gl (/) C ;o !'1 1l (/) c;; FIGURE: 3. Input / Output Records From Pulse Test into a pneumatic signal. This air signal is fed to a Foxboro three-mode controller (TRC). When con ducting cascade control experiments, this con troller serves as the master controller. The trans mitter output is also fed to an A / D converter channel via an air switch. The output of the three mode controller is fed to the set-point input of the Honeywell, proportional + reset, pressure record ing controller (PRC). In cascade control experi ments PRC serves as the slave controller. The coil-side steam pressure is fed to a Honey well transmitter. The output of the transmitter is fed to the pressure controller. The output of the controller operates a 1 / 2-in. Unifi.ow pressure control valve. The signal to the valve may alter nately come from a D / A converter channel of the control computer. In conventional control experiments, cascade control is achieved if both, master and slave con trollers, are placed in automatic. If the pressure (slave) controller is switched to manual, the in put to the control valve comes from the tempera ture controller. The temperature control loop is then a simple closed-loop rather than a cascade system. The control computer used in some of the ex0 0 -----LOG MODUL US-,. __ PHASE ANGLE --15 --' r--C) 0 ...J 0 N ;;. 3 0 AT -,eo, LOG MODULUS "' 24 ,2 ULTIMATE AR 0 .061 5 ", 'i\ ::, ...J ::, 0 0 4 5 0 ...J -60 0.01 \ 0 .0 4 0.10 0.20 0.40 FIGURE 4. Frequency Response Diagram Frequency w, Radians / min. 28 I I I I I I 1. 0 1) I I> -50 Vl "' 0 ,; ..., "' -100 2 0 "' J -150 C> ;o -200 2.0 "' "' V> periments is a PDP 1103 microcomputer system manufactured by the Digital Equipment Corpora: tion. It has 24K words of memory and is equipped with an 8-channel D / A converter and a 16-channel A / D converter. The computer comes with a dual disk drive. A floppy disk, containing systems pro: grams ( e.g. Fortran support programs, real-time subroutines) resides in one of the drives while a second floppy disk, containing user-developed control programs, resides in the other drive. Com munication with the computer is via a teletype writer (LA 36 DECWRITER). EXPERIMENT 1 : Process Identification This experiment is concerned with dynamic identification of an open-loop process by pulse test ing. The resulting information is used to find (1) suitable tuning constants for a feedback controller or (2) to develop an approximate process model which is useful in designing advanced control strategies. The pulse testing technique [2, 3] h a s been applied to the temperature control loop whose block diagram is shown in Figure 2. The input and output records from the pulse test are shown in Figure 3. Numerous data points from these records are entered into a computer program [3, 4] which generates frequency response data as shown in Figure 4. From Figure 4, the ultimate AR, which refers to .the amplitude ratio for which the phase lag equals 180 degrees, is 0.0615. Also, the crossover frequenc y which is the frequency corresponding to the phase lag of 180 degrees, is 1.1 radians per minute. Therefore, the ultimate gain, Ku, and the ultimate period, Pu, are Ku Pu 1 0.0615 21T 1.1 16.23 psi F (1) 5.71 min Since the gain of the transmitter, KT, is 0.06 psi F, the Ziegler-Nichols tuning constants for a PI controller are Gain, K c = 0.45 Ku / K T = (0.45) (16.23) / 0.06 = 121 psi / psi Integral Time, r1 = Pu / 1.2 = 5.71 / 1.2 = 4.75 min. (2) To assess the adequacy of the controller settings found in this section, a closed-loop control experi ment was conducted. The response of the system to a step change in set point and load is shown in CHEMICAL ENGINEERING EDUCATION

PAGE 31

Figure 5. This plot shows that the tuning constants found through pulse testing are ade quate. EXPERIMENT 2: Multivariable Control Most large processes have many controlled variables and many manipulated variables. Ideally, a change in a given manipulated variable should affect only its own controlled variables and no others. Unfortunately, in many cases, this is not the case. The interaction among different lo9ps can lead to poor control and even instability. Since interaction can be a problem in multi variable control systems, it is important to know the extent of interaction and to be able to develop criteria for proper pairing of manipulated and controlled variables. A measure of the extent of interaction in multi variable control is obtained by Bristol's method [5]. The method is based on steady-state input-out put relationships for the process. It yields a measure of steady-state gain between a given input-output pairing. By using the most sensitive input-output connections, interaction is mini mized. Since Eristol's method does not take systems 112 ,-------------, ~104 ._ w a: ::, I/(a) '-(b) ,!is I 88 '---' 1 --' 1 ----'1 ----'1 ---L l ....1. l ....1. I _Ll_...,_I~ 0 4 8 12 16 TIME MINUTES FIGURE 5. Transient Closed-Loop Response to (a) Set Point Change (b) load Change dynamics into account, it would be very useful to evolve an experiment which assesses the beneficial effects of proper pairing upon the dynamic re sponse of the multivariable system. The present experiment [6] is designed to accomplish this oh j ective. The hardware for this experiment is essentially that shown in Figure 1 with the exception that the steam line is replaced by a pipe which intro duces hot water into the tank. The air switches must be in the computer control position for this exJi)erimerit. WINTER 1980 Since interaction can be a problem in multivariable control systems, it is important to know the extent of interaction and to be able to develop criteria for proper pairing of manipulated and controlled variables. The process objective is to control the level (in effect, total flow) and temperature of water in the tank. There are two inputs to the process, namely, the flow of cold water and the flow of hot water into the tank. So, the controlled variables are temperature and total flow and the manipu lated variables are cold water flow rate and hot w a ter flow rate. The question is, should the temperature ,be controlled by manipulating hot water flow and level (i.e. total flow) by cold water flo w or vice versa? Bristol's method provides the answer. Bristol's Relative Gains Analysis The functional steady-state relationship be twee n tempe r ature, total flow and the flow streams is mt = f(m c mh) =m e + m h Around some steady-state operating point, these relationships can be expressed as AT = aT a T -Am e + -Am h am c amh = K n Am e + K 12 Am b and .. (4) Am t am t am t = ~m e + -Am11 m e amh = K 2 1 Am e + K 22 Am 11 The K's are the open-loop steady-state gains which quantitatively describe how the m's affect T and m 1 They can be det e rmined from a mathematical model of the process or by experimental step or pulse-testing on tlie pl a nt. To evaluate K 11 and K 2 1 for example, a small change in the fl.ow of cold water is made, while the process is operating under steady state conditions (under manual control with the flow of hot water maintained constant). When the temperature and level reach their new steady-state values, K u and K 21 can be evaluated by ( AT) K 11 = Am e mh; consta n t (5) 29

PAGE 32

~----------------------------and K 2 1 = (AmT) Llmc mh = co n st an t (6) The gain Ku, then, determines the change in temperature, T, due to a change in m e when mh is held constant. Now, suppose instead of holding mh constant, while a small change in m e is being made, mh is manipulated so as to bring mt back to the original value it had before the change in m e was made. Then, another gain Au can be defined as Au= -( ~T ) Ame mt = consta nt (7) Au is a measure of how me affects temperature T, if level were under closed-loop control (i.e. held constant). The ratio of K u to A u is called the relative gain A 11 Thus, Kn ('~T/t. .m c) mh = constant Au= --------(8) Au ('u T / A1I1c) mt = constant By comparing the relative gains for each manipu lated variable, it is possible to assess which m has the most effect on a given controlled variable The equipment for the process around which the present experiments were developed, was constructed from data provided by Exxon Oil Company ... on an identical setup at their Bayway refinery. and therefore how to pair the manipulated and the controlled variables. While K's can be determined easily, the experi mental determination of A's is not so easy. How ever, they can be evaluated from the K's as follows: B y definition Au = ( :~J m t = c on s tant The open-loop relationships (Equation ( 4) be come Thus, Also in view of Equation ( 4) 30 K A K 12 K 21 A AT = uumc K um c 2 2 (10) 100 incorrec t pa iring 50 .. 0 C i! I 'li ;t. 0 0 1 0 20 30 40 lim (minUIH) FIGURE 6. Transient Response of Level Therefore, AT (11) and, Au = ( :~J m t = co n s t a nt KnK 22 K12K21 K2 2 (12) The relative gain An is then Kn K11K2 2 Au = -= -=:----,=-----==-=Au ;K11K 22 K12:K.:i1 (13) Similar analysis yields the remaining relative gains. Thus, A12 K12K21 (14) K1 2 K 21 K11K 22 A 21 = K12K21 (15) iK12K 2 1 K11K2 2 A 22 = KnK22 (16) KnK 22 K 12 K 21 To facilitate the pairing of manipulated and controlled variables, it is convenient to present the relative gains in a matrix form as shown in Equa tion (17). (17) For each controlled variable, the manipulated variable selected is the one which has the largest positive relative gain. Since a property of this CHEMICAL ENGINEERING EDUCATION

PAGE 33

matrix is that each row and column sums to one, only one >.. need by explicitly computed in a 2 x 2 system. Results The relative gains matrix for the current pro cess is shown in Equation (18). me mh me mh T me mh T 0.172 '0.828 (18) mt mt mt mh me 0.828 0.172 mt mt mt This equation shows that: T should be controlled by manipulating mh and mt by manipulating m e Both loops use a proportional + integral control algorithm on the digital computer as the control element. The algorithm was tuned by trial and 40 lime (minutes) FIGURE 7. Transient Response of Temperature error. The steady-state operating conditions were: level set point, 50 % (which corresponded to total outlet flow of 11.6 lit/min) ; temperature set point, 24.4 C; cold water flow, 9.61 lit/min; hot water flow, 1.99 lit/min. The process was operated with correct pairing as well as with incorrect pairing. The benefits of proper pairing are clearly evident in the set-point responses shown in Figures 6 and 7. These results show that Bristol's approach is a simple and powerful tool in the control systems design of multivariable processes. If the relative gains in Equation (18) had turned out to be numerically close to each other, WINTER 1980 -----------~-----~ interaction ("fighting loops") would have been a problem, particularly if the response times of the two loops were comparable. Severe cross-coupling can drive the multivariable system to instability. In such cases decoupling will be required. Inter ested readers may consult reference 7 to obtain further information on the various techniques currently available for decoupling a multivariable control system. NOMENCLATURE Kp K's Ku Greek amplitude ratio closed-loop gains process transfer function temperature transmitter gain, psi/ F proportional gain on temperature controller, psi/psi steady-state gain of process, F/psi open-loop gains ultimate gain, psi/ F cold water flow, lb/hr hot water flow, lb/hr total flow m e + mh, lb/hr ultimate period, min. temperature of the mixture, F temperature of cold water, F temperature of hot water, F phase angle process dead-time, minutes time constant, minutes integral time, minutes relative gain frequency, radians/minute REFERENCES CiTED 1. Alper, William S., A Laboratory Demonstration of Closed-loop Automatic Process Control, Master of Chemical Engineering Thesis, University of Louisville, 1963. 2. Hougen, Joel 0., Exp e rienc es and Experim e nts with Proc ess Dynamics, Chemical Engineering Progress Monograph Series, Vol. 60, No. 4, 1964. 3. Luyben, W. L., Process Modeling, Simulation and Control for Chemical Engine e rs, McGraw-Hill Book Co., New York, NY, 1973. 4. Dynamic Response Testing of Process Control Instru mentation, I SA S26 Standard, Instrument Society of America, 400 Stanwix Street, Pittsburgh, PA 15222, October 1968. 5 Bristol, E. H., O n a New Measure of Interaction for Multivariable P rocess Control, Trans. IEEE, 1966. 6. Knabel, E. A., Multi .Variable Control System for Fluidi ze d Bed Coal-gasification Units, Master of Engi neering Thesis, University of Louisville, 1978. 7. Foss, A. S., Denn M. M. Ed., Chemical Process Control, A.I.Ch.E. Symposium Series, No. 159, Vol. 72, 1976. 31

PAGE 34

[eJ;) ;>I classroom THE INTEGRATION OF REALTIME COMPUTING INTO PROCESS CONTROL TEACHING PART II: THE UNDERGRADUATE COURSE* M. MORAR! and W. H. RAY University of Wisconsin Madison, Wisconsin 53706 p ROCESS CONTROL FOR undergraduates is offered each semester and has enrollments of 30 to 50 students. The course has two lecture periods, one recitation period, and one 3 hour laboratory period each week. The curriculum (shown in Tables 1 and 2) has been selected to give the student a balanced mixture of useful theory and hands-on practical experience in process dynamics, measure ment, and control. COURSE DESCRIPTION T HE INSTALLATION OF A PDPll/55 minicom puter system described in Part I of this article has allowed a complete restructuring of the course material. A large library of computer programs for the design of control systems is available and is still growing. The use of these programs makes it possible to design control systems for meaning ful practical processes without the drudgery of laborious hand calculations. Thus, course time is freed and it becomes possible to cover topics generally neglected in undergraduate courses, TABLE 1 Undergraduate Process Control lecture Topics 1. Review of Laplace transforms and matrix algebra 2. Principles of real-time computation and data acquisition 3. Transient and frequency response of linear systems 4. Feedback control of linear systems 5. Stability of linear systems 6. Control system design for linear systems 7. Nonlinear systems 8 Case stu dies *Part I dealing with graduate education in process control appeared in the Fall 1979 issue of C.EE. 32 TABLE 2 Undergraduate Process Control labo~atory Experime~ts 1.. Techniques of analog simulation 2. Techniques of digital s imulation 3. Dynamics of interconnected water tanks 4. Computer aided data acquisition 5. Frequency response and process identification through pulse testing 6. Calibration an d dynamic response of PID controllers 7. Feed forward, feedback, and cascade control 8 Mult ivariable control of a gas distribution systeJ,U 9. Multivariable control of a multi-sidestream distilla tion column 10. Tuning of a le ve l controller with stro ng systertl. nonlinearities e.g. multivariable control. The lecture material is listed in Table 1 and includes cortsideratio'ns of how to choose loop pairings, how to minimize interactions between cop.trol loopi;; a:dd .Q,9W to tune multi variable systems: In the latt~r part'\>f the se mester a number of g\raphical interacti1Ve com puter aided design prog ~ f1,mS are Jsed for detailed case studies. The present Hbrary iricludes r'outines for the generation of Bdde plots,' root loei and ---.--~ GC ( 5 > t----:11 -i G ( $ > ~1 OE SI GII ~OOEL 1 > G < 5 > = C I ( PU~ 1 )( P2*S+l ) / (T3* S + 1 )( T4*S+l )( TSfS+l ) OP. 2 ) G ( S > =C A 1 ) +A ( 2 H S+ .. A < H HSU ( N-1 l l t CB< 1 ) +8 ( 2 l*S+ .. B < N >*SU( N I ) l GC 1 S l = ~*(1 + TOfS + J;TUS> WHICH WOULD \' O l l Ll K ECl OP. 21 ? CONTROLLER INPUT C = I I) P 2) PD 3 ) Pl 4 ) PIO INPUT Pl =0 TYPE[! TO 41?1 IHPUT P2 I INPUT T3 =2 INPUT T4 S INPUT TS S INPUT Tl NE DELAV UP TO WHAT VALUES OF K WOULD Y OU L!KE?18 WOULD YOU LI KE TH E ROOTS LI STEO OH THE LI HE PR! HTER>[Y tNlH FIGURE 1. Input Data for Root Locus Program: Process (s + 1) G~s} = (2s + 1) {8s + 1} (5s + 1) is under Proportional Feedback Control. Copyright ChE Div i sion, ASEE, 1980 CHEMICAL ENGINEERING EDUCATION

PAGE 35

------,-Nyquist diagrams for open loop, closed loop and cascade control systems. In addition, a number of programs for the design of multivariable systems are presently available. Output from these pro grams usually appears as plots drawn on the screens of graphic terminals, and paper copies (to be included in student reports) are obtained by the student at the push of a button. Example: if a root locus diagram is to be drawn the user would be asked to specify the transfer function parameters and other variables as shown in Fig. 1. Upon completion of the questions a diagram (as in Fig. 2) will appear which can be subsequently enlarged or otherwise modified. The laboratory, which is designed to comple ment the lecture material, is comprised of some Manfred Morari was born in Graz Austria on May 13, 1951. He obtained his undergraduate education in chemical engineering at the Swiss Federal Institute of Technology (ETH ) Zurich After his diploma he started graduate school at the University of Minnesota in 1975. Upon completion of his doctorate he joined the ChE faculty at the University of Wisconsin in 1977 where he is currently assistant pro fessor. Last summer he worked for Exxon Research and Engineering Company. His r~search interests include a variety of topics from the areas of process synthesis and process control: synthesis of separa tion sequences, optimal measurement selection and inferential control, optimizing control and the dynamics and control of large integrated processing systems. (L) W Harmon Ray was born in Washington, D.C., on April 4, 1940. He received the B.A. and B S. Ch.E. degrees from Rice University Houston, Texas, in 1962 and 1963 respectively, and the Ph.D. degree in ChE from the University of Minnesota in 1966. He has been on the faculty of tlae University of Waterloo in Canada (19 66 -70), the State University of New York at Buffalo (1970 76) and the Uni versity of Wisconsin Madison, where he is presently Professor of ChE During the 1973 74 academic year, he was on sabbatical leave as a ~uggenheim Fellow in Belgium and Germany. His research in terests include chemical reactor engineering and process modelling optimization, and control. His publications include an edited volume "Dist r ibuted Parameter Systems" (Dekker, 1977), and two monographs "Pro ces s Opt imizati on (Wiley, 1973) and "Advanced Process Control" to be published by McGraw Hill in 19B0. (R) WINTER 1980 The use of these programs makes it possible to design control systems for meaningful practical processes without the drudgery of laborious hand calculations. Thus, course time is freed and it becomes possible to cover topics generally neglected in undergraduate courses ten experiments (cf Table 2). Most of these ex periments are carried out by each laboratory group in the course of the semester. Many experi ments involve real time computation and are se lected to familiarize the students with the modern methods of implementing control algorithms. These presently include: Data acquisition (noise suppression, signal amplifica tion, A/D conversion, sensor calibration, etc.) Pulse testing (data acquisition, input pulse selection, Fourier transformation of data, frequency spectrum analysis, frequency response parameter determination, etc_) Multivariable feedback control of interconnected gas storage tanks (process modelling, data acquisition, single loop PI control, supervisory computer control, direct digital control, etc_) Multivariable feedback control of a multi-side-stream distillation column (process modelling, data acquisiti!)n, single loop control, supervisory computer control, and direct digital controJ) Let us discuss two of these exper~ments in more detail. Pulse Testing T HIS EXPERIMENT CONSISTS of putting a measured pulse of hot water into a stirred mixing tank having continuous inflow and out flow. Input and output temperatures are measured under computer control and the resulting data analyzed to provide frequency response informae ~e @@@@@H@@HH @H@HHHH ROOT LO C US UHHHH HH@HHHHtt Pl 8 80 P2 1.88 T 3 2 88 T4 8 00 T5 5 00 C 1 00 TO 0 0 0 1 TJ: 0 00 K 0 80 CO NTROLLER K ( p > A 0 41 8 2 45 C 4 5 9 O= 6 72 E 9 92 0 C B A (1 I I I)( -8 J33E+08 T !NE DELA Y 8 88 8 6E+88 8 008E+88 0 888E+08 -8 446E+O8 FIGURE 2. Root Locus Diagram for the Feedback Control Loop in Figure 1.

PAGE 36

tion and parameters for a process model (i.e., a gain and time constant for this simple first order process). Typical results, taken from a student lab report, are shown in Figures 3 and 4. The measured input and output temperatures are shown in Figures 3 while the resulting Bode plot showing the frequency response may be seen in Figure 4. The process gain of 1.0 is readily found from the low frequency asymptote of the ampli tude ratio (AR). By using both amplitude ratio (AR) and phase angle (cp) to estimate the corner frequency, w e two separate estimates of the tank time constant are found. Usually these are in reasonable agreement with the "theoretical" value determined from the mean residence time of the tank. Multivariable Control of Interacting Gas Storage Tanks One of the most sophisticated experiments carried out by the students is the modelling and multivariable feedback control of a pair of intern n tn n rt, l"l:'t'I' tt*'l''l"t'I'* PULSE RESPONSE ******* *~ **t*; tnuut:t 0 120E+03 1 20 0 1 2(1 +0 3 Tp. ( F) {T Ta111bientJ I IH? UT TE /1 ? 2 OUTPUT TEMP a em~2 a 17 4 E->03 90 ti r: e 1 8 0 ae c orids FIGURE 3. Pulse Test Data for Mixing Tank. acting gas storage tanks. This experiment requires three laboratory periods plus some lecture prepa ration. The purpose of this experiment is to demonstrate to the student the effects of inter actions in multivariable systems and to give him or her the possibility of testing different multi variable control schemes on a real system. A simplified version of the system flow sheet is given in Figure 5. Air enters the system at a constant pressure of 60 psig, flows through a control valve into the first tank and from there through another control valve into the second tank. Finally, the air passes through a fixed ori34 1 .-. ,F. 2 =;j BOOE PLOT 0 ':"? l) E+ QO 1.0 ,---J'-----1 .. 0. ]5 0 346.,ai;; 0 .J--~~---------0 ee&+w ,,, -0 ~+e2 4-...,_~___. ___ ........,-+-~--'---...._-'---1 8 52ZE-02 e 1SSE-01 w e t "' co 0 sa~-e1 005 w l ec l) .OS FIGURE 4. Frequency Response Bode Plot for Mixing Tank. flee, a rotameter, and is vented. The equipment is fully instrumented with two pressure gauges, pres sure transducers, PI analog controllers and is also interfaced with the minicomputer allowing data acquisition, supervisory and direct digital control. The first task given the students is to develop a mathematical model for the tank system. An unsteady state mass balance for each of the two tanks yields dpl RT = V M [Cd 1 A i, (P o ,P 1 ) cd 2 A 2 (p1'p2)] 1 dp 2 RT = y M [Cd 2 A 2 (P 1 ,P 2 ) Cd a A a (P 2 ,P 3 )] 2 where universal gas constant absolute temperature volume of tank i (1) (2) mean molecular mass of air stream discharge coefficient for orifice i area of orifice i ~,K 2 = material constants (3) (4) All parameters of the model are available to 60 P S I G A IR T O VENT FIGURE 5. An Interacting Gas Storage System. CHEMICAL ENGINEERING EDUCATION

PAGE 37

-------------------------------~ the student in tabulated form except the discharge coefficients which have to be determined through steady state experiments. Computer control experiments are carried out with the goal of achieving a given gas production rate while meeting certain pressure constraints in the two gas storage tanks. Both Supervisory Control and Direct Digital Control algorithms are tested by the student. Different control objectives can be selected by the student but they all result in the regulation of the two pressures through changes in the two control valves. Let us briefly indicate some of the choices available to the student. A. Supervisory control: Possible control objectives: 1) specified gas flow rate from tank 2 and 2) p 1 / p 2 fixed or p 1 minimized or p 1 maximized In supervisory mode, the valves are under local analog control. After a control option is S TE ADY P l ST A TE CONTROL DECOUPLING PR O CE SS .,. ., r--, ., .,, I ,,, I ,, ,, I I I G I I I I I ,, ,, 2 2 I I L __ _. FIGURE 6. Control Structure for DDC Control with Single Loop Pl Controllers and with Added Steady State Decoupling {dashed lines). entered by the student via the computer terminal, the set points are computed and transmitted by the computer to the local controllers. The response to changes in objectives is observed for different flow regimes. ( Critical or subcritical flow through the valves.) Set point compensation is attempted to yield a smoother servo behavior. B. Direct digital control (DDC): The possible control objectives are identical to those listed in part A. Different multivariable control algorithms are developed by the students and supplied to the main control program in the form of Fortran subroutines. As an example, steady state decoupling is implemented and com pared with the usual single loop PI control. The controller structure is seen in Figure 6. The re sults of one laboratory group are shown in Figures WINTER 1980 FIGURE 7. Process Response Under DDC With Single Input-Single Output Pl Control. 7, 8. With simple PI control, significant oscilla tions in the pressure response were found (cf. Figure 7). However, with the addition of steady state decoupling the response was much improved (Figure 8) CONCLUSIONS THE NEW MINICOMPUTER has become an integral part of the undergraduate control course at Wisconsin. Aspects of digital computer control are demonstrated to the students and they have the opportunity to gain s ome practical experience with the implementation and application of modern control algorithms. Computer aided control f: [ljNT ROL W ITH SS 0EC0U~LIIIC C ~ i O2 05 111! !"1 F~ C, CF : = 0 oe as, ~R~T JO I. I ,a : n; IS Llll E AR X AX I S JS LINEt.P l l) l; 0(1 E +02 .!. ;:1 C,t)( 1 + (12 1: (1 J (1~1 E+G2: 2 = 0 3(10 E +1 J 2 1 0 .0 00 E+ 00 ~e eouE+oo 0 100E+81 8 1 !+13 I 281l(+tJ FIGURE 8. Process Response Under DDC With Steady State Decoupling and Pl Control. 35

PAGE 38

system design methods utilizing interactive graphics have replaced classic pencil and paper methods and have thus made time available to include new theoretical material in the curricu lum. ACKNOWLEDGMENTS Our progress in bringing real time computing into the process control curriculum at Wisconsin is due to the co operative efforts of many individuals: 1. Emeritus Professor R. J. Altpeter, who originally established process control as a discipline at Wisconsin. He laid the foundation upon which the present cur riculum is built. 2. Visiting Professor Ram Lavie, who shared his ex perience in the development of laboratory experiments. 3. The students who contributed their time and talents to the development of new experiments and computer aided design programs-these include Dennis Arnon, Dean Berceau, John Bolling, John Greiner, Tim Heisel, Sunny Lo, Bob Lojek, Diana Meseck, David Roark, John Seymour, and Pat Vilbrandt. 4. The technical staff and faculty of the department of Chemical Engineering who should be recognized for the support they have given this endeavor. In par ticular, the efforts of Mike Lynch, Todd Ninman, Jim Wenz and Don Zentner have been invaluable. 5. The more than 200 undergraduate and graduate students who have "consumer tested" the changes in curriculum and provided useful feedback. 36 NEED QUICK~ ACCESS TO.,~ PATENTS Ever been in the midst of research and realized that easy access to U.S. Patents would be the answer to your problems? U.S. Patents on microfilm from Research Publications, Inc. offers to attorneys, researchers, inventors, and librarians the largest single body of scientific and technical information in existence. RPI offers a complete retrospective file of U.S. Patents and current subscriptions, as well as the Official Gazette and the CDR File. Access is immediate, and information retrieval is quick and easy. Delivery of current subscriptions is weekly, within four weeks of date of issue. Find the answers to your questions with U.S. Patents on microfilm from: research publications, in!P 12 Lunar Drive, Woodbridge, Ct 06S2S (203) 397 2600 THE INSTITUTE OF PAPER CHEMISTRY ASSOCIATE DEAN Nominations and applications are invited for the position of Associate Dean at The Institute of Paper Chemistry. The Associate Dean is Director of Ad missions and chairs the Admissions Com mittee. This includes responsibility for the recruitment and evaluation of prospective graduate students. The Associate Dean also works closely with the Dean in administering the gradu ate programs at the Institute, assists in the development and implementation of aca demic policy and shares responsibility for the management of the Dean's Office. The Associate Dean reports directly to the Vice President-Academic Affairs/Dean of the Institute. The Dean's Office has re sponsibility for faculty planning, student re search, admissions, financial aid, student records, advisement, placement, continuing education, alumni affairs and the computer center. The position includes an appointment on the faculty which affords graduate teach ing and research opportunities. Letters of application ; with credentials, should be sent by April 1, 1980 to: Harry T. Cullinan, Jr. Vice President-Academic Affairs & Dean Institute of Paper Chemistry P. 0. Box 1039 Appleton, WI 54912 The Institute is an equal opportunity affirmative action employer. CHEMICAL ENGINEERING EDUCATION

PAGE 39

The Leading Authorities ... and the Leadi~ Texts Separation Processes in Chemical SECOND EDITION fe;ktl~ydson King, University of Californ i a at Eng1neer1ng 1979 880 pages (tent.), $27 95 (te n t.) Here is a book featuring complete general coverage of separation processes with material organized by concept. The common features and key differences between processes are emphasized, and many worked examples problems and case studies are included. Solutions Manual. Available Now Mass Transfer Operations THIRD EDITION The late Robert E. Treybal 1980 800 pages ( tent.) $26.50 ( tent. ) A thorough highly acclaimed study of mass transfer operations in chemical ~J engineering this book covers relevant principles of equipment design data concerning equipment design and applications of modern theory Solutions Manual. Available Now r htto you l>y Heteroaeneous Catalysis in Practice Charles N. Satterfield, Massachusetts In stitute of Technology 1980, 464 pag es ( tent. ), $25.50 ( tent. ) This book is primarily devoted to a study of catalysts and reactions that have industrial significance for large-scale operations It includes suggestions for designing e x periments and analyzing data and a guide to relevant literature Solution s Manual. Available May 1980 Plant Design and Economics for Chemical Engineers THIRD EDITION Max s. Peters and Klaus D. Timmerhaus, both of the University of Colorado at Boulde r 1980 992 pages ( tent. ) $28.50 ( tent. ) In this book economic princ i ples are applied to the st udy of chemical engineering showing the relevance of economics to industrial operations and plant design Solutions Manual. Available February 1980 M Hill r-------------------------I I I ~,q~: 1tMri : COLLEGE DIVISION McGRAW-HILL BOOK COMPANY 1 22 1 Avenue of the Americas New York N.Y. 10020 I I I I I I McGraw-Hill Book Company P O. Bo x 400 Hight s town N J 08520 Please send me the following book(s) for a free 10 -day examination. I f not comp l etely satisfied I wi l l re turn the book(s) within the trial period; otherwise I will remit payment, plu s postage, handling and local sa le s tax (McGrawHi l l pays postage and handling if payment is included with order). 0 King : SEP PROCESSES 2/e (034612-7) 0 Peters-T : PLANT DES & ECO 3/e (049582 3) D Satterf ield: HET CATALYSIS (054875 7) D Treybal : MASS TRANS OPS 3/e (065176 0) U 070 0107-1 Pri ces S ubj ect to c hang e. Name _________ Address ___ ______ City __________ State ______ Zip __ Rev72 / CE Ed I I I I I I I I I I I I I ~----------------------------------------

PAGE 40

-----------------------(in #I laboratory PROCESS CONTROL EXPERIMENT: THE TOILET TANK THOMAS J. WARD Cl,arkson College of Technology Potsdam, N.Y. 13676 THE LEVEL CONTROL mechanism in the home toilet tank is a nonlinear, proportional control system that illustrates various process control concepts. It can also serve as an introduction to data acquisition, process analysis, and model de velopment. This simple experiment can be de veloped as an example problem, a classroom demonstration, or a laboratory exercise. DESCRIPTION A TYPICAL TOILET TANK is shown in Figure 1. The level control system regulates the tank water level C at the desired steady state C. by manipulating the inlet water rate M to compen sate for the disturbance (the flushing rate U). The control logic is given in the flow diagram of Figure 2A. An actual toilet tank can be used for the labora tory exercise if a flowmeter is installed in the feed line and a measuring scale is fastened to the in side of the tank wall. For classroom demonstra tions, the level mechanism can be installed in a clear plastic tank (0.5 by 0.15 by 0.4 meters high). A quick-opening valve can provide the necessary flush for the plastic tank. Tom Ward has been on the chemical engineering faculty at Clarkson College of Technology for many years. He received his M.S. from the University of Texas and his Ph D. from RPI His research interests are principally i n the area of process control. Copyright ChE Divisicm, ASEE, 1980 38 C B I FIGURE 1. A typical tank. Typical exercises involve the development of analytical models, determination of model pa rameters, and measurement of the closed-loop re sponse to a flushing disturbance. Two suitable process models are presented in the following sections. NONLINEAR MODEL THE EQUATIONS FOR EACH of the elements of Figure 2A can be obtained as follows : Tank: A mass balance on the tank gives A_j_Q_ = M U dt A 0 B DISTURBANCE U (1) FIGURE 2. Information flow diagrams: A. Nonlinear logic, B. Linear block diagram. CHEMICAL ENGINEERING EDUCATION

PAGE 41

where A is the tank cross-sectional area. A typical value for A is suggested, by the plastic tank dimen sions given above so that Tank: A mass balance on the tank gives A= (0.5) (0.15) = 0.075 G! (2) Float: The float center height B is assumed to correspond to the water level so that the equation for the measuring element is simply B : = C (3) Alternately, it could be assumed that the float is characterized by a first or second order model. While this would lead to an interesting higher order process model, it might be difficult for students to estimate the parameters in a higher order model. Valve: It is assumed that the valve flow rate M can be related to the valve position M' by the equation M = a (M') 0 5 (4) Many toilet valves have an adjustable ring so that the valve coefficient can be changed. A typical value for the coefficient a is 0.0021 2 5 /s. The experimental determination of a suitable valve equation and coefficient can provide an interest ing short study. Lever: A reasonable controller equation relat ing the valve position M' to the error (C. B) can be written as M' = K (C -B) (5) where the controller gain K is given by the lever ratio. A typical value for the desired steady-state level c is assumed to be 0.3 meters. If the lever is assumed to be 0.40 meters long (to the float center) and the pivot is 0.04 meters from the valve end, then the controller gain is given as K = 0.04/0.40 = 0.1 (6) These can be combined to give the nonlinear model as dC A---ar-= a[K(C. C)] 0 5 -U (7) LINEARIZED MODEL As NOTED ABOVE, the steady state is selected as the filled tank (a nonflow condition). For any variable X, the perturbation from the steady I\ state is defined as X = X x In terms of such perturbation variables, the equations for three of the elements are WINTER 1980 Tank: I\ dC 11 /\ A---ar-= M U (8) I\ I\ = -6 Float: E : = -B (9) Lever: I\ I\ M' = KE (10) The nonlinear valve equation can be linearized around the steady state by the truncated Taylor series approach to give M ( !!, ) M' = /3M' where f3 should be the slope of the valve curve at the origin. As can be seen in Figure 3, this would 0 000 4 ------------M 0 0 002 ( /s ) / I/ ~OPE/3 = 0. 0159/s o ...__ ___ .._. ___ j_. ___ .J 0 0 01 0,02 0 03 M' is arbitrarily selected to approxi mate the value behavior over the region of interest. This illustrates how experimental data can be used to improve on the classical steady state linearization. The linear model corresponds to the block diagram of Figure 2B, where G e = K, Gv = /3, H = 1, Gp = 1/ AD, and D = d / dt. These can be combined to give the linear model as I\ I\ (TD + 1) C = ---aU (12) where T = A / {3K and y = 1 / ,BK. DISTURBANCE INPUT IT IS ASSUMED THAT the flush, given by 5Udt, empties the tank. Then 5Udt = (0.5) (0.15) (0.3) = 0.0225 3 Various approximations can be suggested for the disturbance U. A suit able function might be the displaced cosinusoid with a period of 10 seconds 39

PAGE 42

U i -0.00:25 cos 0.21rt + 0.00225 0 s tsl0 sec l t>lO sec I (l3 Since the flush occurs quickly compared to the filling time, it might be reasonable to approximate the disturbance as an impulse of 0.0225 occurring at t = 0. MODEL RESPONSES THE RESPONSE FOR VARIOUS model and disturbance forms can be compared with experi mental responses to show the significance of the approximations made. Three model responses will be given here: a. Nonlinear Model With Cosinusoid Dis turbance. The response for this case was obtained numerically for the parameter values assumed above. This is given by the curves NL-C in Figure 4. b. Nonlinear Model With Impulse Dis turbance. If the flush is an impulse at t = 0, then the solution for t>0 can be obtained by assuming that U = 0 and C = 0 at the instant t = o + Equation 7 can then be solved by the sepa ration of variables technique to give (C -C) 0 5 /2A)t + C1 (14) For the assumed parameter values, the integra tion constant C 1 = (0.3) 0 5 The response curves for this case are given as NL-I in Figure 4. c. Linear Model With Impulse Disturbance. I\ I\ If it is assumed that U (0 + ) = 0 and C (0 + ) = -0.3, then the solution to Equation 12 is I\ C = -0.3 e-t / T (15) The responses for this case are shown in Figure 4 as the curves LI. OTHER FEATURES THE TOILET TANK CAN be used to illustrate other process control concepts. Some of these are: a. Offset. The leaking toilet provides one of the simplest demonstrations of offset. If the out flow at steady-state is not :zero, then the inflow at steady state is not zero. Since the flow into the tank is a fixed function of C, then the steady-state level must decrease if the toilet tank is leaking. b. Measurement Error. The water-logged float implies that the float center.,line does not 40 0.2 LEVEL C < m> 0.1 o...._ ___ _._ ___ __. ___ ..--J FLOW M < m3r s 0 0002 120 180 TIME (sl FIGURE 4. Level and flow responses (LI = linear model with impulse disturbance, NL-I = non linear model with impulse disturbance, NL-C = nonlinear model with cosinusoid disturbance} correspond to the level. At steady state, the level must increase in order to close the feed valve. c. Setpoint Changes. The float end of the level can be bent to represent setpoint changes d. Controller Gain. Most toilet tank controls have some provision for changing the loop gain. In some cases, this is accomplished by changing the effective lever ratio. In others, an adjustable ring is used to change the valve gain. CONCLUSION The toilet tank, as either a classroom problem or experiment, provides a simple introduction to process control. CHEMICAL ENGINEERING EDUCATION

PAGE 43

rec-og-ni-tion\ I rek-ig1 nish-Jn, -~g-\ n 1 : the action of recognizing; the state of being recognized; as a : ACKNOWLEDGM EN T 2 : spec i al notice or attention. rec-og-ni-tion\ as we see it\ 1 : the primary motivation to do creative work for an out standing company 2: ACKNOWL E DGM E NT of the quality of that work; as a : self-satisfaction and pride b : respect from peers and associates c : opportunity for advancement 3 : to recognize the challenge o f the world today 4 : to be recognized for doing something to meet those challenges tomorrow If you know of qualified graduates in engineering or the sciences or with 9 n interest in marketing finance or computer science we hope you will encourage them to write us : Re cruiting and College Relations P O Box 1713-CE Midland Michigan 48640 Dow is an equal opportunity employer male/female DOW CHEMICAL U.S.A T ra d emark o f The D ow C hemical Co mpa n y WINTER 1980 ; ,, 41

PAGE 44

(e) ;j I curriculum A SURVEY OF PROCESS CONTROL EDUCATION IN THE UNITED STATES AND CANADA* DALE E. SEBORG University of California Santa Barbara, CA 93106 JN FEBRUARY, 1978 a questionnaire on process control education was distributed to the 158 chemical engineering departments in the United States and Canada. Completed questionnaires were returned by 143 schools, or 90 % of the 158 depart ments. This response compares quite favorably with the 59-101 replies that were received in re cent AIChE surveys of undergraduate curricula [1, 2]. The large number of replies is probably due to two factors: 1) the questionnaire was kept very brief, and 2) copies of the final report were promised to those departments which submitted completed questionnaires. SURVEY RESULTS THE SURVEY RESULTS indicate that process control is firmly established in the undergradu ate curriculum since only 7 of the 143 respondents ( 5 % ) do not off er undergraduate courses. By contrast, 108 schools (75 % ) have required courses and an additional 28 schools (20 % ) offer elective courses. Interestingly enough, 4 of the 7 schools which do not offer undergraduate courses in pro cess control do offer graduate courses. Thus only 3 schools of the 143 respondents do not offer any process control courses. Process control courses are also firmly es tablished at the graduate level. Seventy-two schools ( 50 % ) offer graduate courses while an The survey results-indicate that process control is firmly established in the undergraduate curriculum since only 7 of the 143 respondents (5%) do not offer undergraduate courses. A preliminary version of this paper was presented at the Miami Beach AIChE Meeting. 42 TABLE 1 Textbook Selection for Undergraduate and Under graduate / Graduate Courses TEXT Coughanowr and Koppel Luyben Weber Harriott Douglas Perlmutter Smith, Cecil Ogata Others (one each) NUMBER OF DEPTS. 69 21 9 6 6 3 2 2 17 Total 135 additional 15 schools (10 % ) offer courses which are open to both graduate students and advanced undergraduate students. Tables 1 and 2 list the process control textbooks which have been adopted for undergraduate and graduate co~rses, re spectively. The most striking result here is the con tinuing popularity of the book by Coughanowr and Koppel which has been selected as an undergradu ate text by 69 departments and as a graduate text by 7 departments. The dominant position of this 15 year old text is quite remarkable in view of the significant developments which have occurred since 1965 in both computer control hardware and TABLE 2 Textbook Selection for Graduate Courses TEXT Smith, Cecil Coughanowr and Koppel Luyben Douglas CACHE Monographs Lapidus and Luus Other (one or two each) NUMBER OF DEPTS. 10 7 4 2 2 2 17 Total 44 Copyright ChE D ivision ASEE, 1980 CHEMICAL ENGINEERING EDUCATION

PAGE 45

--------------------~ TABLE 3 Laboratory Control Experiments in Undergraduate Courses NUMBER OF EXPERIMENTS/COURSE NUMBER OF DEPTS. 0 43 1-2 30 3-4 26 s+ 37 Some experiments (number not available) 7 Total 143 in control theory. The results in Table 1 agree quite well with a 1975 survey on undergraduate process control 1 courses [1]. It should be noted that the numbers in Tables 1 and 2 are reported on the basis of individual departments rather than on the basis of courses offered. For example, if a particular department offers two undergradu ate process control courses which use the same textbook, this was counted only once in Table 1 rather than twice. By contrast, if two textbooks were required for a particular course, they both were included. Many of the 17 textbooks included in the "Other" category in Tables 1 and 2 were written for mechanical or electrical engineers and are used in classes taken by both chemical engi neering students and other engineering students. One hundred departments (70 % of the re spondents) indicated that their curriculum inDale E. Seborg received a B.S degree from the University of Wisconsin in 1964 and a Ph.D degree from Princeton University in 1969 Both degrees were awarded in chemical engineering He taught at the University of Alberta for nine years before joining the University of California, Santa Barbara in 1977 He is presently Pro fessor and Chairman of the Department of Chemical and Nuclear Engineering at UCSB. Dr. Seborg is co-author of the book, Multivariable Computer Control: A Case Study He is also the past chair man of Area 15B (Systems and Process Control) of the AIChE Na tional Program Committee. WINTER 1980 The most striking result here is the continuing popularity of the book by Coughanowr and Koppel which has been selected as an undergraduate text by 69 departments and a graduate text by 7 departments. eludes one or more laboratory experiments in pro cess control. Table 3 shows that 63 departments offer courses that contain at least three control experiments and 30 departments have one or two experiments, usually as part of a unit operations laboratory. In compiling these statistics, each department was included in only a single category. Thus if a particular department offers two process control TABLE 4 Use of Real-Time Computers or Micro-Processors in Control Experiments NUMBER OF DEPTS. Currently have a real-time system 48 Equipment on order or being installed 19 No equipment (but have tentative plans to add equipment) 22 No equipment (and no plans for future equipment) 53 Total 143 courses which include three and five experiments, respectively, this department was included in the tally for the "5 +" category in Table 3. During the past decade there has been con siderable interest in "real -time computing," that is, in digital computers which are used for data acquisition and control. Both industrial and aca demic personnel in the process control field have maintained an active interest in the field of real time computing for the following reasons : The widespread availability of inexpensive minicom puters and microprocessors; Changing process control objectives in industrial plants due to energy and environmental considera tions; The realization that the application of most advanced control strategies will inevitably require an on-line digital computer. The results in Table 4 indicate that 48 depart ments (34 % ) currently have control experiments which involve a real-time computer while an ad ditional 19 departments ( 13 % ) have computer systems on order or being installed. Table 4 includes only those departments which use real-time computers in conjunction with under graduate control experiments. It does not include 43

PAGE 46

TABLE 5 Real-Time Computers or Microprocessors Currently Installed (53) or On Order (19) EQUIPMENT AND VENDOR Minicomputers Digital Equipment Corp. Data General Corp. Hewlett Packard IBM Texas Instruments Foxboro Interdata Miscellaneous (one each) Not specified Microprocessors NUMBER OF DEPTS. 24 8 5 5 3 2 2 6 7 10 Total 72 other departments which use minicomputers or microprocessors exclusively in research labora tories. The 22 departments in the third category in Table 4 typically are in a preliminary planning stage or are seeking funds to purchase a real-time system. Thus the results of this survey indicate a continuing trend for incorporating a real-time computer system in the undergraduate curricu lum. Table 5 presents a summary of the 62 mini computers and 10 microprocessors which are currently operating in chemical engineering de partments or on order. The numbers in Table 5 do not correspond directly to those in Table 4 since several chemical engineering departments use more than one real-time computer in the under graduate curriculum. CONCLUSIONS THE RESULTS OF THIS survey indicate that the topic of process control has become firmly es tablished in the chemical engineering curriculum. Only 3 of the 143 departments surveyed do not teach any courses in process control. One hundred and eight schools (75 % of the respondents) have required undergraduate courses while 87 schools (61 % ) teach graduate level courses in process control. Laboratory experiments in process control are now available at 100 schools (70 % ). There is a continuing trend toward providing students with exposure to real-time computer systems in con junction with process control experiments; 67 de partments currently have such a system operating or on order while an additional 22 departments have tentative plans for such a system. Fifteen years ago, process control was generally regarded as a new, specialized topic 44 which was not part of mainstream chemical engi neering. The present survey demonstrates that this situation no longer exists. Process control has joined the more traditional topics such as trans port phenomena, thermodynamics and reactor analysis in playing a central role in the chemical engineering curriculum. REFERENCES 1. Eisen, E. 0., "Teaching of Undergraduate Process Dy namics and Control," paper presented in a mini-session at the 68th Annual AIChE Meeting, Los Angeles (No vember, 1975). 2. Barker, D. H., "Undergraduate Curriculum 1976," Ch e m. Eng. Educ., Vol. XI, No. 2, (Spring, 1977). BOOK REVIEW: Contact Catalysis Continued from page 12. hope of being able to reproduce catalysts of a given type in different laboratories is rapidly becoming a reality. As one might infer from the variety of topics and extent of treatment, these volumes are not exactly for the beginner. One might have wished some discussion of homogeneous catalysis, at least in terms of analogs to heterogeneous systems, and a more general inclusion of the concepts of co ordination chemistry as they relate to catalysis. In all, however, some balance must be struck be tween coverage and length and the editor has done an admirable job. The English translation of the original Hungarian edition of 1966 is ex cellent and the text has been updated. The dust jacket states that "the book will be useful to workers studying catalysis in industrial and uni versity laboratories." The present reviewer feels this is a correct statement and commendable for its modesty. kiN?I news ART HUMPHREY HONORED Arthur E. Humphrey, dean of Penn's School of Engineering and Applied Science, became the eighth honoree to receive the James M. Van Lanen Distinguished Service A ward for his "life long dedication and service to fermentation science and the fermentation industry. The award is named for a pioneer in fermenta tion technology and was established in 1976 as the foremost award and citation of the ACS Division of Microbial and Biochemical Technology. CHEMICAL ENGINEERING EDUCATION

PAGE 47

liJ;jlclass and home problems The object of this column is to enhance our readers' collection of interesting and novel problems in Chemical Engineering. Problems of the type than can be used to motivate the student by presenting a particular principle in class or in a new light or that can be assigned as a novel home problem are re quested as well as those that are more traditional in nature that elucidate difficult concepts. Please sub mit them to Professor H. Scott Fogler, ChE Department, Uni ver sity of Michigan, Ann Arbor, MI 48109. SOLUTION: MIRROR FOG PROBLEM R. L. KABEL Pennsylvania State University University Park, PA 16802 Editor's Note: Professor Kabel presented the "Mirror Fog Problem" in the Fall 1979 issue of CEE. We extended an invitation for student solu tions to this problem at the time of publication and would like to congratulate Mauricio Fuentes of Ecole Polytechnique, Montreal, Canada, who sub mitted the winning entry and by so doing has won a year's subscription to CEE. Professor Kabel graded the responses and, in his words, Mr. Fuentes' entry was both "correct and excellently done." The following is Professor Kabel's solution to the problem. Derivation of equations: Use a microscopic model because momentum and energy equations are not required due to isothermality and no bulk flow. Mass balance equation: aCA + v acA + v acA + v aCA at x ax y aY oz = DAu [a 2 CA + a 2 CA + a 2 CA] + R A Ox 2 0Y 2 az 2 Since v x = v Y = v. = 0, CA =I= f (x,z) and there is no generation in the vapor space this equation becomes aCA D a 2 CA AB OY 2 which shows that the concentration at any point changes with time because of diffusion in the y-direction. Initial condition: At t = 0, CA = CA ,snt at all Y Boundary conditions: At y = 0, CA = CA room at all time At y = Y, CA = CA ,snt at all time where Y is the location of front edge of the remaining fog on the mirror. Note however that Y varies with time going from 0 when t = 0 to 0.3 m when t = t r. If an analytical solution of the equation is to be sought this second boundary condition should be respecified. If the WINTER 1980 solution is to be numerical, then one merely needs to keep track of Y (t). The end of the calculation is t = tr when Y = Yma x = 0.3 m. Y(t) can be obtained by equating the total amount evaporated to the integrated mass flux into the room neglecting the slight accumulation of water vapor in the enlarging vapor space. Let MA o be the initial total mass of water condensed, then y y Amount evaporated = MAo -= MA 0 0.3 Ymax Amount transferred to room = J t DAB oCA Is dt oY y = o 0 wheres= ZX y Y= Ym nx J t oC I zx DAB __ A dt Ao oY y = 0 0 XZY maxD AB J t MAo 0 acAJ at oY y = o The above is an adequate answer to the exam question. A very simple analytical solution can be obtained as follows. We can say that the flux at y = Y is equal to the amount of moisture evaporated there per unit time. Then the amount of moisture can be related to the rate at which the boundary moves. Thus, if R = thickness of liquid film, D dCA g H20 evap. AB~ area time y=Y PH 2 o ZR dY g H 2 0 evap/time ZX dt area of transfer dY dt If we assume that a steady state concentration profile is established rapidly and maintained (shown by dynamic analysis to be an excellent assumption) we get c.A,F.iat CA,room y 45

PAGE 48

A further simplification is obtained by taki n g C.A ,r o om = 0. Let c .A,s at = C A S" dY cit Y max dt = f tr B dt = f Y dY 0 0 I Ym ax B t r = 0 5Y 2 0 = 0.5 Y 2 max and t r = 0 5 Y 2 max/ B for the mirror fog problem pa 2 0 10 g m 3 R 1.3 x 10 -s m (obtained from experiment) DAB = 2.47 x 10 5 m 2 s 1 CAs 17.3 g m -s Y ma x 0.3 m X 4 x l03 m t r 3.5 x 10 5 s = 96 m = 4 days This result appears high by about a factor of 4. There are several explanatio n s and we have calculate d for different assumptions Probably the experimental circum stances (e.g. leakage around edges, etc ) do n o t meet the idealizations of the model. D IN THE "HEAT" OF THE NIGHT R. J. GORDON Un iv ersity of Florida Gaines v ille FL 32611 y ou ARE SPENDING the evening in a small town on your way home for the holidays At about 11 :00 p.m. the local sheriff calls you and asks for your help. He knows from the desk clerk that you are a chemical engineer, and naturally assumes you have some knowledge of forensic chemistry. It seems that the body of John Lurie, a local car dealer, had been found somewhat earlier in a wooded area just outside of town The local coroner had gone fishing and there was no one else to estimate the time of death. John Lurie had been known to deal in "hot" cars and was thought to be going to the police to confess and name his four accomplices, Gus Nusselt, Bill Gurney, Ed Reynolds, and Bob Prandtl. Nusselt had been known to be out of town until 11 :00 A.M that morning, Gurney had a solid alibi from 1 :00 p.m. on, Reynolds was with his girlfriend until about Co py right C h E D ivisi on A SEE 1 98 0 46 6 :00 A M when he left to go fishing, and Prandtl was in jail the night before for drunkenness, and was not released until about 8 : 00 A.M. When you finally get to the body it is about 12 :00 p.m. (midnight). You measure a rectal temperature of 80 F, and an air temperature of 70 F. The air temperature has been about 70 F all day. Luckily, you brought your Perry's along. Recognizing that the human body is mostly water, 1) calculate the latest possible time the murder could have occurred and 2) state the possible suspect. NOTE : For practical purposes, John Lurie can be assumed to be shaped like a rectangular slab. He is 10 inches thick from his back to his breastbone Body temperature is 98 6 F. Rectal temperature is equivalent to core or centerline temperature For comparison, a pathology formula some times used to estimate the time of death is N f h d th 98.6 rectal temperature o. o rs. smce ea 1. 5 ONE-DIMENSIONAL S O LUTION We will use the Gurney-Lurie charts, Perry's 4th Ed., p. 10-6, 10 7. To calculate the latest time the murder could have occurred, assume maximum rate of cooling, or in other words that the surface of the body is at 70 F (same as saying h = oo or m = 0). We also neglect radiative losses since the body was found in a "heavily wooded" area. Then, if we assume infinite width and depth, y = T s T 70 80 = 0 35 T .T o 70-98.6 From graph, for n = 0 X at 0.54 = (x i2 ) k 0.36 0.0058 ft 2 / hr a = pCp 62.4 X 1.0 5 inches 0.42 ft. X 1 = 12 t 0.54 X (0.42) 2 0.0058 16.4 hours or 16 hours, 24 minutes 1) Murder had to occur before 7 :40 A.M "Medical Jurisprudence and Toxicology," Glaister and Rentoul 12th Ed., Livingston Ltd., Edinburgh, 1966, p. 110. CHEMICAL ENGINEERING EDUCATION

PAGE 49

2) Possible Suspects: Gurney, Reynolds From the pathology 1 formula: 98.680 1. 5 = 12.4 hours (or murder occurred at 12 noon) This formula, however, takes no account of changes in room temperature, or body thickness, and in fact is known to underpredict the time of death except for the first few hours. From our superior knowledge of heat transfer, we have eliminated Prandtl and Nusselt as suspects. D ACKNOWLEDGMENT: Helpful comments were provided by Professor J. H. Hand, University of Michigan. Editor's Note: Professor Gordon's purpose in his solution to the foregoing problem, "In The Heat of the Night," was to illustrate the use of the Gurney-Lurie charts assuming a simple one-dimen sional model. Professor Fogler, GEE Problem Section Editor, asked his student, Alan Basio, to comment on this simplified solution. Mr. Basio's reply follows. TWODIMENSIONAL HEAT TRANSPORT ALAN BASIO University of Michigan Ann Arbor, MI 48109 It was previously assumed that Lurie, the dead man, is an infinite slab. From this assumption, the time is 16.4 hours since Lurie was killed. I used Newman's Rule and assumed Lurie is an infinitely long slab with a finite width and depth. Newman's Rule in this situation is the following: T -T Y = Y x Y y = T -To 0.350 (1) Let Lurie be 10" deep, as previously specified, and 1.3 feet wide. Use the same values as before for Yanda. There are now two values of X to be found on the Gurney-Lurie Charts: Xx= at/(5 / 12) 2 and X y = at / (1.35 / 2) 2 The time must be the same in both X x and Yn and the product YxY y = 0.350. Criteria for solution: (1) YxY y = 0.35 ( 2 ) Xx(x) = X y (y) = t a a Results: By trial and error the times are found WINTER 1980 to be within 2.7 % of each other. Y x = 0.420 Y y = 0.833 Xx = 0.45 X y = 0.18 t t YxY y = (0.42) (0.833) = 0.350 (0.45) (5 ,f 12) 2 0.0058 0.18 (0.65) 2 0 ( 0058 13.47 hrs. 13.10 hrs. 13 .4 7 l 3 .10 100 2 7 01 d'ff 13 .4 7 x = 10 1 erence If the width of Lurie is 1.3 ft., he died 13.3 hrs. ago, not 16.4 hrs. The width of Lurie is important. If Lurie is 2.6 ft. wide, for example, he dies 16.3 hours earlier. In other words, the infinite slab assumption im proves when Lurie is assumed over 2.0 feet wide, approaching an answer oft = 16.4 hrs. BOOK REVIEW: Reactor Design Continued from page 24 The book is an excellent work. The author has covered a very large area of relatively difficult material in a highly readable fashion and has pro vided enough detail so that the reader is able to come to grips With the realities of chemical re actor design. It is accurate and relatively com plete. There is a considerable amount of specialized knowledge, based upon over 1000 references, aug mented by the author's own considerable ex perience. In many areas, it stands at the edge of chemical reactor design knowledge that is in the public domain. As such it will continue to be a valuable reference work for many years to come. Its only major shortcoming is insufficient il lustrations and a lack of exercises or problems for the student. The fourteen case studies of Volume II serve to illustrate design principles but only cover a fraction of the material in Volume I. In order to serve as a text for a graduate course in chemical reactor design, it would have to be sup plemented by problems developed to reinforce specific points and others which would require the student to integrate these ideas into a chemical reactor design. The latter would be an. under taking of the order of a term paper. These two volumes are a major contribution to the chemical engineering literature. They belong in the library of every chemical engineer who is concerned with research, development, design, or, in many cases, operation of chemical reactors or conversion processes. 47

PAGE 50

RATE OF REACTION Continued from page 16. The evaluate V v the continuity equation is used (see ref. 19) : op + V (p v) = o at (10) But for this system V p = 0. Besides p = m/V and 0 m/ 0 t = 0 because the total mass of the system is constant in time. Substituting these relation ships into equation (10) and operating: 1 aV V v= V at (11) Substituting Ci by Ni (Ci = Ni/V) and per forming the differentiation, the equation for r, in the case of an isothermal variable volume batch reactor, results: dN, r= ---ai V(t) dt (12) The partial differential is now a total differen tial insofar as Ni is a function of time only. It must be noticed, however, that we should know the relationship existing between the change of volume and the reaction extent, i.e., if the extent of the reaction is expressed in terms of conver sion, there should be at hand a relationship of the following type: V (t) = V (V 0 ,x). In most gaseous systems a linear variation of volume with conver sion is often assumed (9). STEADY STATE ISOTHERMAL CONTINUOUS PLUG FLOW REACTOR UNDER STEADY CONDITIONS, general equation (7) is reduced to: V 1 = a, r If, according to the model it is supposed that there are no diffusion or dispersion effects, the mass flux is only due to the global convective flow. Hence: N, = Civ from where we finally have: d(C, v ) r = ----''---=--=-'-a 1 dz (13) since in this reactor the flow is uni-directional. (Direction z has been chosen as representative of the model). If the total number of moles is not preserved and, furthermore, if the reaction takes place in a 48 gaseous phase, equation (13) is not so useful for v will become a function of z and it will not be possible to take it out of the differential. If the number of moles stays constant, equation (13) is reduced to: v dCi r= -a i dz (14) which, obviously, also includes the assumption that pressure changes along the reactor are small ( due to losses by friction, for example) because other wise v would not be independent of z, either. Equation (13) can be adequately modified to become useful even in those cases in which the number of moles is not constant. To do so, we simply transform the equation and work in terms of mass fractions (w). Re calling that: C, = pw 1 M i In this equation wi is the mass fraction of com ponent i and Mi its molecular weight. Substituting in equation (13): (15) and since the mass flow rate G 0 is constant: p v = G o = constant; hence, the rate of reaction can be written as: (16) STEADY STATE, ISOTHERMAL CONTINUOUS FLOW STIRRED TANK REACTOR The SSICFSTR is an ideal type of system, in which concentration of reactants does not depend upon time or position within the reactor. The mathematical description of the flow within the reactor is exceedingly difficult, as we are dealing with a highly idealized case which, consequently, cannot be fully achieved in practice. Nevertheless, in many cases the deviations are almost negligible and the system has been successfully modelled. The exact mathematical description cannot be ac complished because the model implies a transport of mass, instantaneously, over finite distances. But the difficulty may be overcome if we do away with the necessity of describing the internal pattern of the flow inside the reactor. To do so, the general equation (7) is integrated over the CHEMICAL ENGINEERING EDUCATION

PAGE 51

---,--This statement sums up the whole problem: The rate of reaction expression is the "sink" or "source" term in the continuity equation for multicomponent systems which will take into account the creation or destruction of the said species by chemical reaction. volume of the reactor so as to obtain a macroscopic balance. I dV = J ( \7 1 ) dV + r a 1 rdV (17) V V JV Under steady state conditions the left-hand side is zero and, as the assumptions of the model grant that all properties within the reactor are constant, r is not a function of the space co ordinates and it may be taken out of the integral. With this, the previous equation is reduced to: a1 r VR = J ( \7 N 1 ) dV (18) V Applying the divergence theorem to the right hand side of equation (18), we obtain: a1rVR= J (N1n)dA (19) The integration performed over the whole surface of the reactor may be evaluated because the flux is non-zero only at the inlets and outlets J (N1 n) dA = J (Ni n) dA + J (N 1 n) dA A In. Out. (20) On the right-hand side, the first term repre sents the inlet flow of component i and the second one, the outlet flow. Taking into account the di rections of unit normal vectors n, the final equa tion is: 1 r = V (F 1 outl et F !, i nl e t) (21) a1 R which is the expression usually written for the re action rate in a continuous stirred tank reactor. The integration of the differential equation (7), so accomplished, brings about a macroscopic balance for species i, which is usually the start ing point of the derivations in the books on applied kinetics. In this work we have followed this ap proach in order to demonstrate the absolute generality of equation (7). In the following section, this analysis will be extended to two more complex experimental systems, the recycle reactors. These are often WINTER 1980 useful for obtaining kinetic data. APPLICATION TO ISOTHERMAL RECYCLE REACTORS RECYCLE REACTORS HA VE been widely used since the publication of the original papers by Hougen [20], Perkins and Race [21 ], Biskis and Smith [22], Korbach and Stewart [23], and Cassano, Matsuura and Smith [24] as a means of retaining the differential operation of the reactor and, at the same time, eliminating the restrictions of inaccuracy in the analysis of the small composi tion changes. Moreover, control of the flow rates in the re cycle allows the reduction of diffusional resistances and the elimination of temperature gradients. Diffi culties may be centered around the effects of re action by-products, the considerably longer time usually needed to obtain the steady state condition in the recycling section of the apparatus and the difficulty in operating under pre-fixed concentra tion conditions. (The last two are especially im portant for the continuous type). Once more, "general definitions" will be useless and we shall have to resort to the general mass in ventory. In order to simplify the matter, let us consider the case when the total number of moles is constant. ISOTHERMAL CONTINUOUS RECYCLE REACTOR FIGURE 1 (A) SHOWS the system under consideration. Operating conditions must be ad justed in order to fulfill the following assump tions: 1) The operation in V R is differential, i.e., the outlet concentration C 1 ,t is very close to the inlet concentration of the reactor C 1, 12) Differences in concentration between C 1, o and Ci ,f are accurately measured. 3) High recycling flow rate ( Q) When Q oo the whole reactor, analyzed in the control volume (2), is an excellent approxi mation to a continuous flow stirred tank reactor, working at differential conversions as the react ing mixture goes through the control volume (1). The General equation (7) will be applied to each of these systems. 49

PAGE 52

t Co,q (Al FIGURE 1 I I I I I I I I --------------~ Analysis !Bl Under steady state conditions, the concentra tion in the reactor is independent of_ time. The differential equation is reduced to: (22) On the other hand, the operational characteristics allow us to neglect any form of dispersion or diffu sion effects. Consequently: (23) If the differential equation (22) is integrated over the control volume, the result is: J ( V N 1 ) dV = J !'Ii a1 dV V V On the left-hand side we shall apply the di vergence theorem and on the other side we shall take into account the fact that the reactor is differential. Substituting expression (23) yields: J (n C1 v) dA = r1 a1 J dV A V and finally : (q + Q) (C1 t C1,1) = a1 r1 VR (24) On the other hand, if the whole system (2) is treated as a CFSTR the application of equation (21) will yield: q (C 1,t C1,o) = r 2 a1 VR (25) In equation (25) we have V R only on the right hand side, because it is the only part of the total volume where r is different from zero. But the global velocity r 2 has to be equal to that produced in the reactor itself, that is to say r 1 If it is shown that r 1 accurately represents a differential rate of 50 reaction, then the values which may be obtained through the application of equation (25) will portray the exact rate of reaction and not an aver age value. This will be true if C1,1 is very close to C1 t as has been initially assumed. The only re maining doubt would be to know how these condi tions could be accomplished. As r1 = r 2 we have: (q + Q) (C 1,t C 1,1 ,) = q (C1, 1 C1,o) C Q C1 ,r q C1,o i I q + Q If, as was first assumed, Q is sufficiently high, Q > > q and therefore G 1,1 :::::: C 1,r For isothermal continuous recycle reactors, working under the conditions stated above, an adequate expression for the rate of reaction results: r = q (C 1 r C1,o) a1 VR (26) Notice that the condition of a high recycling flow, necessary for the differential operation in V n, coincides with the requirements for the opera tion of the global system as a continuous stirred tank reactor. The extension to a system with variable number of moles only complicates the algebra. ISOTHERMAL BATCH RECYCLE REACTOR FIGURE 1 (B) illustrates the system under consideration. For extremely slow reactions and for cases where one wishes to avoid serious limita tions in the size of the samples for analysis, this is an adequate experimental device. Within this system there are no disadvantages such as those pointed out for a continuous recycle reactor, but an experimental problem may arise: the existence of a recycling device (with movable parts in most cases) could introduce contamination of the react ing mixture from the outside. F'or a batch system, the impurity level will grow with time. This is not so severe in continuous systems. The operating conditions are as follows: 1) The operation in V R is differential. 2) For reasonable intervals of time, concen tration differences are accurately measured in V. _, 3) High recirculating flow, and adequate mixing in V. We treat the whole system as a batch reactor: if the recirculating flow is high, the concentration will be uniform. Hence, V N 1 = 0 and resuits: CHEMICAL ENGINEERING EDUCATION

PAGE 53

---------------------------------, f JV dC1 dV = dt r J air dV V Since according to previous assumptions, con centration is not a function of position, the total derivative has been employed and it may be taken out of the volume integral. On the : right-hand side, the integral will be different from zero only in those portions of the volume where r is different from zero. Due to the differential performance of V n one then obtains: (V + Vn) = ai rV n and the rate of reaction, when there is no varia tion in the number of moles, is finally: (27) This equation must be corrected for additional changes of concentration, in case the sample volumes for analysis were significant. A great re lationship of V /V n reduces this problem, but may largely prolong the necessary reaction time to attain accurately measured conversions. Abalance in the reactor itself shows the con ditions for the differential operation of V R At each cycle the reactor behaves as a steady state isothermal continuous plug flow reactor. V i = a 1 r J ('v Ni) dV = J a;i r dV V V Assuming, for the time being, an average value of r, the result will be: Q (C1 r C 1 ,1) = a1 ravg V R where C 1,r and C 1 ,1 are the outlet and inlet con centrations of reactor V R at each cycle. If the rate of reaction has a finite value and, as previously as sumed, Q is high, the difference between inlet and outlet concentrations will be small. By increasing Q the difference could be made small enough to turn ra vg into a differential rate of reaction. Notice that this condition was also assumed as necessary to propose a uniform composition model. It must also be noticed that an adequate experimental device should minimize the volumes of the connect ing lines between the reactor VR and the tank V. HETEROGENEOUS SYSTEMS OF REACTION I N THIS WORK WE have emphasized the analysis of the different isothermal homogeneous reactWINTER rnso ing systems. The adequate treatment for hetero geneous reactions (catalyzed and non-catalyzed) is outside its scope. It is not far-fetched, however, to notice that, in most cases, the flaws of the so-called definitions are even more evident within these systems. The problem may be summarized as follows : 1) At the microscopic level all heterogeneous re actions take place at interphases. 2) Even at the macroscopic level, many heterogeneous reactions take place at interphases. 3) In many cases the rate of reaction takes place not only inside the control volume but on its boundaries as well. 4) Hence, accurately speaking, the rate of reaction will, in many cases, be a boundary condition of the general mass conservation equation. 5 ) When this happens at the microscopic level (in the case of a catalyst, for example), an "effective" rate has been used due to the difficulty in solving the conservation equations with complicated geo metrics for the boundary conditions. 6) At the macroscopic level (for example, the case of free radical termination reactions on the walls of a reactor) when the rates of reaction are boundary c onditions, the conservation equations are compli cated, since diffusional terms cannot be neglected. See for example reference [25]. 7) Even ontologically, the rates of reaction in hetero geneous systems undergo a change, insofar as their intensive character is attained by means of an ex pression referring to the area of this boundary (real or ideal). All this makes even more evident the futility of trying to establish a general "definition" of a reaction rate. CONCLUSIONS I N THE PREVIOUS sections, six different expres sions have been attained for the rate of reaction: equations (8), (12), (21), (26) and (27). Many others could be found for other physical systems using equation (7) as a starting point. Quoting Petersen [16]: "to argue that any of these (rate of reaction expressions) is more correct than all of the others as its defining equation, is to confuse a conservation equation with a definition." This statement sums up the whole problem The rate of reaction e x pression is the "sink" or "source" term i n the continuity equation for multicompon ent systems wh i ch will take into account the creation or destruction of the said species by chemical reaction. The rates of reaction thus attained will be in dependent of the system used to measure them (provided it is a homogeneous reaction) This means that the kinetic expression of the rate of 5J.

PAGE 54

reaction r = cf> (T, Ci, P, etc ..... ) if determined using equation (7) as a starting point, will be in dependent of the physical device employed to obtain the kinetic data. There may exist a sound definition of an ex tensive rate of reaction based on the concept of the "extent of reaction," both introduced by De Donder [26], which may be useful for thermo dynamic calculations but the idea becomes devoid of general validity when one needs an intensive property. For the purposes of reaction engineer ing this is legitimate but it is of little help. ACKNOWLEDGMENT To Prof. Elsa I. Grimaldi to whom I am greatly in debted not only for her invaluable contribution to the writing of the manuscript but also for her everlasting patience in doing it over and over again until it finally achieved its final form. NOMENCLATURE A Area (cm 2 ) a Stoichiometric coefficient C Concentration (molecm-3) F Molar flow rate (moles1 ) G 0 Mass flow rate per unit area (grs1 cm2 ) M Molecular weight (grmole 1 ) m Mass (gr) N Number of moles N Molar flux (mole cm 2 -1 ) n Outwardly directed unit normal vector P Pressure (Kgfcm2 ) Q Volumetric recycle flow rate (cm 8 1 ) q Volumetric input flow rate ( cm s 1 ) r Rate of reaction (intensive) (molecm-ss-1) T Temperature (K) t Time (s) V Volume (cm B ) v Linear velocity ( cms 1 ) VR Volume of reactor (cm 3 ) w Mass fraction x Conversion p Density (grcm-s) cf> Kinetic expression for rate of reaction (molecm3 s1 ) V Vector operator Subindices and Supraindexes avg: Average value i Indicates species or initial condition, according to context f Final condition o Inlet condition R Indicates reactor z Indicates direction Indicates vector Indicates molar average velocity REFERENCES 1) GLASSTONE, S., "Textbook of Physical Chemistry", D. Van Nostrand Co., Princeton, N.J., (1946). 52 2) BENSON, S W "The Foundations of Chemical Kinetics" Mc Graw Hill Book Co., New York, (1960). 3) DANIELS, F., "Chemical Kinetics", Cornell Uni versity Press, (1938). 4) LAIDLER, K. J., "Chemical Kinetics", McGraw Hill Book Co ., New York, 2nd. ed., (1965). 5) FROST, A. A., and PEARSON, E. G., "Kinetics and Mechanism", J. Wiley & Sons, New York, 2nd. ed., (1961). 6) JOHNSTON, H. S., "Gas Phase Reaction Rate Theory", The Ronald Press Co., New York, (1966). 7) HOUGEN, 0. A. and WATSON, K. M., "Chemical Process Principles", Part III, J. Wiley & Sons, New York, (1947). 8) SMITH, J. M., "Chemical Engineering Kinetics", Mc Graw Hill Book Co ., New York, (1956). 9) LEVENSPIEL, 0 "Chemical Reaction Engineer ing", J. Wiley & Sons, New York, (1962). 10) WALAS, S. W., "Reaction Kinetics for Chemical Engineers", McGraw Hill Book Co., New York, (1959). 11) BOUDART, M., "Kinetics of Chemical Processes", Prentice Hall Englewood Cliffs, N. J., (1968). 12) PANNETIER, G., and SOUCHAY, P., "Chemical Kinetics", (Translation by H.D. Gesser and Emond), Elsevier Publi shing Co., New York, (1959). 13) KRAMERS H. and WESTERTERP, K. R., "Ele ments of Chemical Reactor Design and Operation" Academic Press, New York, (1963). 14) ARIS, R., "Introduction to the Analysis of Chemical Reactors", Prentice Hall, Englewood Cliffs, N. J., (1965). 15) DENBIGH, K. G., "Chemical Reactor Theory", Cam bridge University Press, Cambridge, ( 1965). 16) PETERSEN, E. E., "Chemical Reaction Analysis", Prentice Hall, Englewood Cliffs, N. J., (1965). 17) AMDUR, I. and HAMMES, G. G., "Chemical Kinetics", McGraw Hill Book Co., New York, (1966). 18) ARIS, R., "Vectors, Tensors, and the Basic Equa tions of Fluid Mechanics". Prentice Hall, Englewood Cliffs, N. J., (1962). 19) BIRD, R. B., STEWART, W. F., and LIGHTFOOT, E. N., "Transport Phenomena", J. Wiley & Sons, (1960). 20) HOUGEN, 0. A., "Reaction Kinetics in Chemical Engineering", Chem. Eng. Progr. Monograph Ser. 47, No. 1, (1951). 21) PERKINS T. K., and RASE, H. F., A.I.Ch.E. Journal, 4, 351, (1958). 22) BISKIS, E. G and SMITH, J. M., A.I.Ch.E. Journal, 9, 677, (1963) 23) KORBACH, P. F. and STEWART, W. E., I.E.C. Fund. 3, 24, (1968). 24) CASSANO, A. E., MATSUURA, T., and SMITH, J. M., I.E.C. Fund. 7, 655, (1968). 25) CASSANO, A. E., SILVESTON, P. L., and SMITH, J. M., I.E.C. 59, 18, (1967). 26) DE DONDER, Th., "Le~ons de Thermodynamique et de Chimie Physique", Paris, (1920); quoted from Prigogine I. and Defay, R., "Chemical Thermody namics'', Longmans, Green and Co. Ltd., London, (1954). l CHEMICAL ENGINEERING EDUCATION

PAGE 55

ACl(NOWLEDGMENTS Departmental Sponsors: The following 137 departments contributed to the support of CHEMICAL ENGINEE RING EDUCATION in 1979. University of Akron University of Alabama University of Alberta Arizona State University University of Arizona University of Arkansas Auburn University Brigham Young University University of British Columbia Bucknell University University of Calgary California State Polytechnic California Institute of Technology University of California (Berkeley) University of California (Davis) University of California (Santa Barbara) Carnegie-Mellon University Case-Western Reserve University University of Cincinnati Clarkson College of Technology Clemson University Cleveland State University University of Colorado Colorado School of Mines Columbia University University of Connecticut Cornell University University of Dayton University of Delaware U. of Detroit Drexel University University College Dublin Ecole Polytechnique (Canada) University of Florida Georgia Tech 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 Manhattan College 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) 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 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 Washmgton University University of Waterloo Wayne State University West Virginia University University of Western Ontario Uni v ersity 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 32611.

PAGE 56

PROCTER & GAMBLE is looking for in R&D / Product Development 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.) RESPONSIBILITY NOW! If you are Interested In this area, please send a resume to: The Procter & Gamble Company R&D BS/ MS Recruiting Coordination Office lvorydale Technical Center Spring Grove and June Avenues Cincinnati, Ohio 45217 PROCTER & GAMBLE AN EQUAL OPPORTUNITY EMPLOYER


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZ5UP6V7D_61VWQ7 INGEST_TIME 2011-09-26T20:43:45Z PACKAGE AA00000383_00065
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES