Chemical engineering education ( Journal Site )

Material Information

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


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


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

Record Information

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

Full Text

*.. .. .. 0 i
chemical engne ~ dcto

ac0tswdees amwd t1hmi4....


atdA a doawti4o o juwda.


Department of Chemical Engineering
University of Florida
Gainesville, Florida 32611

Editor: Ray Fahien (904) 392-0857

Consulting Editor: Mack Tyner

Managing Editor:
Carole C. Yocum (904) 392-0861

Publications Board and Regional
Advertising Representatives:

Gary Poehlein
Georgia Institute of Technology

Past Chairman:
Klaus D. Timmerhaus
University of Colorado
Lee C. Eagleton
Pennsylvania State University

Richard Felder
North Carolina State University
Jack R. Hopper
Lamar University
Donald R. Paul
University of Texas
James Fair
University of Texas
J. S. Dranoff
Northwestern University
Frederick H. Shair
California Institute of Technology
Alexis T. Bell
University of California, Berkeley
Angelo J. Perna
New Jersey Institute of Technology
Stuart W. Churchill
University of Pennsylvania
Raymond Baddour
Charles Sleicher
University of Washington
Leslie W. Shemilt
McMaster University
Thomas W. Weber
State University of New York

Chemical Engineering Education

60 Purdue University,
W. Nicholas Delgass, Nicholas A. Peppas

66 Robert A. Greenkorn of Purdue, K. C. Chao, D. P. Kessler

Process Control
70 Use of IBM's Advanced Control System in Undergraduate
Process Control Education,
Lowell B. Koppel, Gerald R. Sullivan
74 Undergraduate Process Control, Dennis C. Williams,
A. Ray Tarrer
78 Discrete Processes in Undergraduate Process Control
Courses, Richard H. Luecke, Hsin-Ying Lin
82 A Useful Formula in Process Control, W. J. Chen

Views and Opinions
58 Accreditation: Plus or Minus, Robert R. Furgason
69 Teach Corrosion: If You Dare, Lyndon Hadley-Coates
92 Bridging the Gap: An Integrated Approach to Curriculum,
Ali H. Mansour, Michael S. Lane, John L. Harpell

84 Flow Curve Determination for Non-Newtonian Fluids: A
Sequel, Mahari Tjahjadi, Santosh K. Gupta

88 The Representation of Highly Non-Ideal Phase Equilibria
Using Computer Graphics,
Georgios N. Charos, Paulette Clancy, Keith E. Gubbins
94 Generation of a Ternary Phase Diagram for VLLE,
Ray E. Desrosiers

102 Learning Through Reading Scientific Papers,
Jose O. Valderrama
77 Letter to the Editor
77 Stirred Pots
77, 83, 101 Book Reviews
100 Editorial

CHEMICAL ENGINEERING EDUCATION is published quarterly by Chemical Engineering Division,
American Society for Engineering Education. The publication is edited at the Chemical Engineering Depart-
ment, 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 adver-
tising representatives. Plates and other advertising material may be sent directly to the printer: E. O.
Painter Printing Co., P. O. Box 877, DeLeon Springs, Florida 32028. Subscription rate U.S., Canada, and
Mexico is $20 per year, $15 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 O 1986
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
-identifatien-oef-this teriodieal.


[jInefl views and opinions


Plus or Minus

University of Nebraska-Lincoln
Lincoln, NE 68588-0420

IN A RECENT "Views and Opinions" article in
Chemical Engineering Education,* Professor
James Wei asks, "Are We Participants or Victims?"
The underlying issue is accreditation-the criteria,
the process, and the application. The basic question is
whether the current accreditation system is serving
the profession well through the education offered stu-
dents in chemical engineering programs. This accredi-
tation system is jointly administered by the Accredita-
tion Board for Engineering and Technology (ABET)
and AICHE-it is a joint effort. The person who quip-
ped, "But these are our organizations that we depend
upon. If we can't fix them, who else would do it?", is
right. AICHE has several direct influences on the ac-
creditation process.
Perhaps I should review the organizational struc-
ture of the accreditation process. AICHE, as well as
other engineering societies, has representatives on
ABET-the policy-establishing organization for ac-
creditation procedures. ABET has established three
commissions which handle the actual accreditation
evaluations and actions. These commissions are the
Engineering Accreditation Commission (EAC), the
Technology Accreditation Commission (TAC), and
Related Accreditation Commission (RAC). The EAC
is the body that acts on accreditation for our chemical
engineering programs. EAC derives its membership
from AICHE and the other participating engineering
societies. We have four standing members of EAC
and four ABET board members. The AICHE repre-
sentatives are chosen from participants in chemical
engineering accreditation activities both as a program

If the criteria themselves are
in contention, they can be changed. But
those who want them changed must trumpet their
cause-not in the Exxon suite but with AICHE's
representatives to the EAC and ABET, or
even as a representative ...

*Vol. XIX No. 3 (Summer 1985), page 120

evaluator and membership on the Education and Ac-
creditation Committee (E&A) of AICHE. Your cur-
rent representatives to ABET and the EAC are
ABET Board of Directors
Bryce Anderson, Southeastern Massachusetts
Robert Greenkorn, Purdue University
James Knudsen, Oregon State University
William Manogue, E. I. du Pont
Don Anderson, Michigan State University
Dee Barker, Brigham Young University
David Camp, Dow Chemical Company
Lee Eagleton, Pennsylvania State University
Robert Furgason, University of Nebraska-
Vice Chair-Operations
John Prados, University of Tennessee
Past Chair
So much for the gruesome organizational details.
Apparently at question are the ABET/EAC ac-
creditation criteria and their application to specific
programs. A substantive argument against having
criteria is not likely-the schools need criteria for
guidance as to what is expected, and those doing the
evaluation need a standard against which judgments
can be made. Uniform application, stringency, "bean
counting," interpretation, and final judgments become
the issues. The criteria are dynamic, not static ...
constantly changing to reflect changes in the profes-
sion. If the criteria themselves are in contention, they
can be changed. But those who want them changed
must trumpet their cause-not in the Exxon suite but
with AICHE's representatives to the EAC and
ABET, or even as a representative (but be prepared
to do a lot of work!).
The judgmental process can not and should not be
written out of the system. It has been my experience
that AICHE's representatives have constantly re-
sisted attempts toward over-specification of criteria
("bean counting"). And yet evaluators are torn be-
tween consistency and fairness to all institutions re-
0 Copyright ChE Division ASEE 1986



Robert R. Furgason received his BS and MS in chemical engineering
from the University of Idaho (in 1956 and 1958, respectively) and his
PhD in chemical engineering from Northwestern University in 1961.
He is presently Vice Chancellor for Academic Affairs and Professor of
Chemical Engineering at the University of Nebraska-Lincoln.

gardless of their perceived reputation, and a need to
promote, not inhibit, innovation and desirable change.
Perhaps a two-tiered or multi-tiered system could be
devised to address the mission and thrust of different
universities. I suspect that such an approach would be
very difficult to develop and that it would be a con-
stant source of contention as to how a particular in-
stitution would be classified. Remember, the accredi-
tation process applies exclusively to the under-
graduate program and has only an indirect relation-
ship to the research and graduate mission of the in-
stitution. AICHE has consistently opposed advanced
level accreditation of chemical engineering programs.
If the amount of design material in the curriculum is
the issue, nothing is necessarily sacred. The amount
can be changed, up or down, and interpretations re-
fined. But to condemn the system and cast doubts on
the motives of the people involved usually generates
resentment, not resolution. Any group action is, by
definition, political and will remain so even if new or-
ganizations and players are involved. Participation is
the name of the game and victimization not the object.
I wonder how many critics have read Section IV
of the Criteria for Accrediting Programs in Engineer-
ing in the United States, especially the Curricular Ob-
jectives and Content section. Since the design compo-
nent seems to be a major issue, let me extract the
exact wording from the Criteria document related to
engineering design.
"(3) Engineering Design.
(a) Engineering design is the process of devising a
system, component, or process to meet desired
needs. It is a decision-making process (often itera-
tive), in which the basic sciences, mathematics,
and engineering sciences are applied to convert
resources optimally to meet a stated objective.
Among the fundamental elements of the design

process are the establishment of objectives and
criteria, synthesis, analysis, construction, test-
ing, and evaluation. The engineering design com-
ponent of a curriculum must include at least some
of the following features: development of student
creativity, use of open-ended problems, develop-
ment and use of design methodology, formulation
of design problem statements and specifications,
consideration of alternative solutions, feasibility
considerations, and detailed system descriptions.
Further, it is desirable to include a variety of
realistic constraints such as economic factors,
safety, reliability, aesthetics, ethics, and social
(b) Courses that contain engineering design normally
are taught at the upper-division level of the en-
gineering program. Some portion of this require-
ment must be satisfied by at least one course
which is primarily design, preferably at the senior
level, and predicated on the accumulated back-
ground of the curricular components.
(c) Coursework devoted to developing drafting skills
may not be used to satisfy the engineering design
Obviously this language leaves a lot of room for
interpretation. It does not say, for example, that a
course in chemical kinetics and reactor design is, or
isn't, necessarily all design. The criteria spell out what
is expected in a course if it is to be classified as design
and certainly provide for portions of a course to fit
into more than one category. But it would be difficult
to stretch the interpretation to include a basic FOR-
TRAN programming course as design, although more
than one institution has tried. I suspect some people
do not thoroughly read the criteria and do not have
these requirements well in mind when they prepare
for an accreditation visit.
I recommend reading the entire document as the
wording has been carefully selected and there is a
standing committee whose job is revising the criteria.
Significant changes have occurred. For example, sev-
eral years ago the use of beginning foreign language
courses was not permitted to satisfy the humanities
and social sciences requirements. Now the H-SS
criteria allow, ". . and foreign languages other than
a student's native languagess)" This is a meaningful
change and relates in part to the recognition that an
engineer's perspective must be international and that
ABET/EAC should encourage engineering students
to take foreign languages, not to impede this study as
part of their education. I was a vocal proponent of this
change and served on the Criteria Committee of EAC
when it was instituted.
Maybe accreditation is unimportant to some in-
stitutions. Nothing compels the institution to be sub-
jected to the process-EAC program evaluators par-
Continued on page 107.


Seventy-Jfrh iAmiversary #Q&Schwof i(Cfend

p ,' -l

[lIjO department



Purdue University
West Lafayette, IN 47907

As we celebrate 75 years of chemical engineering
at Purdue we look, as always, to the future but also
back to our roots. The origins of chemical engineering
education can be traced almost exclusively to the first
half of the 19th Century in Germany. During that
period important chemists such as Justus von Liebig,
August Kekul6, August von Hoffman, Robert Bunsen
and others established in three major universities-
University of Heidelberg, University of Gottingen
and University of Giessen-chemical laboratories
which nurtured many generations of theoretical and
applied chemists.
Industrial chemistry, the forerunner of chemical
engineering, originated from these laboratories and
became an important field of research at many univer-
sities during the last quarter of the 19th Century. The
first chemical engineering course was given at the
University of Manchester in 1887 by George E. Davis

Copyright ChE Division ASEE 1986

One-fifth of the more that 338
PhD graduates from Purdue have continued
in academia, many achieving prominent positions. A
recent survey .. identified a total of 164 alumni
that have become university faculty.

in the form of twelve lectures covering various aspects
of industrial chemical practice. At Purdue University,
which had been established as a land-grant university
in 1874, the first chemical engineering course was of-
fered in 1902.
Chemical engineering at Purdue University
started in the Chemistry Department. In 1900, the
pioneering head of chemistry, Percy N. Evans,
suggested that some industrial applications be incor-
porated into the course, "Technical Analysis." In 1902,
fascinated by the first edition of G. E. Davis' Hand-
book of Chemical Engineering, Evans offered a
course called "Industrial Organic Chemistry Lec-
tures" to a select group of undergraduate students in
chemistry. It immediately became very popular with
the students, and in 1904 Evans introduced three
more "chemical engineering" courses, which he shared
with Edward G. Mahin.
On April 16, 1907, President Winthrop E. Stone's
recommendation to the Board of Trustees was ap-


19I11-i986 Purdue University

proved, and a "chemical engineering curriculum" was
formed within the Department of Chemistry. The first
BS in chemical engineering was awarded to Benjamin
M. Ferguson (a Purdue quarterback!) in May 1909. A
second degree was awarded in 1910 and nine more in
Only four years after the introduction of the chem-
ical engineering curriculum there were 79 under-
graduate students enrolled in the chemical engineer-
ing program. Thus, in 1911 President Stone, Dean
Charles Benjamin and Professor Evans sought to es-
tablish an independent school of chemical engineering.
On June 14, 1911, the Board of Trustees approved
the recommendation of President Stone and Purdue's
School of Chemical Engineering became a reality. Its
first faculty member and head was Harry C. Peffer,
former director of research at the Aluminum Com-
pany of America (Alcoa) in East St. Louis. Peffer di-
rected the school from 1911 until 1934. Faculty mem-
bers of the pre-World War II period who served for
a significant time were John L. Bray, Harold L. Max-
well, Robert B. Leckie (who developed a gas en-
gineering option), Frederick L. Serviss (who intro-
duced an engineering geology option), Clifton L.
Lovell, Edward C. Miller and George W. Sherman,
Jr. In 1930, R. Norris Shreve, a 1907 graduate of Har-
vard University, was hired to establish an "organic
technology option."
A graduate program was established in 1916, and
the first MS degree in chemical engineering was
awarded in 1921 to Ernest H. Hartwig. Shreve (unit
processes and industrial chemistry), Bray (metal-
lurgy) and Lovell (unit operations) became the three
main researchers of the school. The first PhD was
awarded in 1935 to William N. Pritchard, Jr. A strong
graduate program with more than 50 graduate stu-
dents already existed in the late 1930's. At the same
time the undergraduate program had become the
largest in the country with 296 undergraduates in
1920, and 441 undergraduates in 1932.
Bray was named head of the school on Peffer's
death and served from 1935 until 1947. He was fol-
lowed by Shreve from 1947 to 1951, Edward W. Com-
ings from 1951 to 1959, and Brage Golding from 1959
to 1966.
In the post-war years, new additions to the faculty
added to the excellence of an already flourishing pro-
gram. Prominent among the new faculty were Joe M.
Smith (1945-57), Carroll O. Bennett (1949-59), John
E. Myers (1950-66), and H. C. Van Ness (1952-56).
These Purdue faculty members were authors of
pioneering textbooks in thermodynamics, kinetics and

transport phenomena that are still used today. In
1945, Shreve published his monumental Chemical
Process Industries, which has sold more than 180,000
copies. The research programs of the school were also
augmented by the work of J. Henry Rushton on mix-
ing and that of Comings on high pressure ther-


Ronald P. Andres at the controls of a secondary ion mass
spectrometer in one of the catalysis research

modynamics. In the early 1960's another best-selling
textbook was prepared at Purdue by D. R.
Coughanowr and L. B. Koppel, Process Systems
Analysis and Control, a book that has served at least
two generations of chemical engineers.

The 70's saw a growing emphasis in the school to-
ward fundamental and interdisciplinary research and
on engineering science. This shaping of the current
educational and research philosophy began with the
appointment of R. A. Greenkorn as head in 1967, was
accelerated by L. B. Koppel from 1973 to 1981 in a
dedicated effort to add to the excellence of the
graduate program, and is now in the hands of R. P.
Andres, who became head in 1981.
The faculty have continued the tradition of
textbook writing with Transfer Operations, R. A.
Greenkorn and D. P. Kessler (1972); Ther-
modynamics of Fluids, K. C. Chao and R. A. Green-
korn (1975); Introduction to Material and Energy
Balances, G. V. Reklaitis (1983); Engineering Optimi-
zation, G. V. Reklaitis, A. Ravindran and K.
Ragsdell, (1983); and Linear Operator Methods in
Chemical Engineering, D. Ramkrishna and N. R.
Amundson (1985). The faculty are also recognized for
excellence in research (recently the 1984 AIChE



Awards in Computing in Chemical Enginee:
Materials Engineering and Sciences and the
Parravano Award in 1985) and in teaching
the George Westinghouse [1984] and Weste
Fund (1984) awards of ASEE and Purdue's
gineering Teaching Award [1985]).


The faculty member with the longest tenure in t
the school is Alden H. Emery who joined Purdue in 19
his PhD at the University of Illinois. His research int
thermal diffusion, mass transfer, fluid mechanics, p
ogy, and most recently biochemical engineering, wit
modelling algae and plant cell growth. Lyle F. Albrigh
Purdue in 1955, is an authority on chemical process
catalytic research. Lyle, who is the author of books
and chemical processes and the editor of several A
volumes, has been particularly active in the AICh
served as a director in 1982-84.
Eight faculty joined the school in the 1960's. Robe
joined in 1962. He was the first faculty member to do

Prominent among the new faculty were
Joe M. Smith, Carroll O. Bennett, John E.
Myers, and H. C. Van Ness. (They) were aut
of pioneering textbooks in thermodynamics,
and transport phenomena . still used tod

research in catalysis, concentrating his efforts on in
troscopy of catalyst surfaces. Squires is a gifted edi
currently director of the school's Coop Program. Dav
and Roger E. Eckert came in 1964. David's research
in transport phenomena, applied mathematics and b
gineering. He has also contributed to Purdue as an a
previously as associate provost and presently as th
Interdisciplinary Engineering Division. Roger has
polymer rheology and transport phenomena. His cur
interests are in applied statistics and the design of e:
Robert A. Greenkorn came to Purdue in 1965
years in industry and academia. His research is in the
and transport phenomena, more specifically flow in
He is author or co-author of four books and is present
ident of Purdue. Henry C. Lim joined Purdue in 1966
years at Pfizer. His early research was on process
1970 he has concentrated his efforts on modelling, co
timization of biochemical reactors. He is co-author
environmental engineering. K. C. Chao also work
and academia before joining Purdue in 1968. He direct
program in theoretical and applied thermodynamics.
has concentrated on developing equations of state.
the book with Greenkorn on thermodynamics, he has
ACS volumes on thermodynamics and equations of st
In 1969 Robert E. Hannemann, a chemical engine
cian (pediatrician), was added to the faculty as a "per
ing professor. His research focuses on various biomed
ing problems including the causes and treatment c
membrane disease, and the spectroscopic detection
That same year R. Neal Houze came to Purdue from t
of Houston. His research interest is in two-phase flo

Seventy-iffth Anniversary tof 'eSchoolofChemi
ring and in sently the director of the university-wide Coop Program.
e Giuseppe Eleven of the twenty full-time faculty members were hired in
Gr e the 1970's and 1980's. Phillip C. Wankat and Gintaras V. "Rex"
g (recently Reklaitis came in 1970. Phil is doing research in various aspects of
rn Electric separation science, including chromatographic methods, membrane
Potter En- science, magnetic separations, and cyclic zone separation. A
heralded educator, Phil is active in ASEE and has an MS in counsel-
ing along with his degrees in chemical engineering. Over the past
fifteen years, Rex has developed computer-aided design into one of
the school's major strengths. He also has written two textbooks,
he history of one on optimization and one on energy and mass balances. In 1985
54 fresh from he was appointed assistant dean of engineering for Graduate
crests include Studies and Research.
polymer rheol- George T. Tsao and W. Nicholas Delgass joined Purdue in
th a focus on 1974. George was at Iowa State University and NSF before coming
t, who joined to Purdue. Here he has developed a large research program on
s and applied biochemical and genetic engineering. He is founder and director of
on polymers the Laboratory for Renewable Resources Engineering. Nick came
CS symposia from Yale University. His research centers on heterogeneous
E, where he catalysis. He is co-author of a book on spectroscopic techniques in
rt G. Squires Three faculty members were added in 1976 and 1977. Dorais-
fundamental wami "Ramki" Ramkrishna came to Purdue after several years
at I.I.T. Kanpur. He is well-known for his research in the applica-
tion of mathematics to chemical engineering problems. He is co-au-
thor of a book on linear operator methods. Nicholas A. Peppas also
joined the school in 1976. His research is on polymers and biomed-
hors ical engineering. James M. Caruthers came to Purdue from M.I.T.
kinetics in 1977 and has developed research programs in thermorheology
ay. and viscoelastic properties of polymers and in rheology and light
scattering of colloidal dispersions.
situ IR spec- Elias I. Franses joined Purdue in 1979. His research centers on
ucator, and is colloids and interfacial science with emphasis on the thermodynamic
id P. Kessler and transport properties of surfactants and the microstructure and
interests are stability of dispersions. Linda N. H. Wang, who joined the faculty
iomedical en- in 1980, is doing research in ion-exchange theory, separation sci-
tdministrator, ence, and applications of these to control the local environment of
e head of the enzymes used in biomedical and biochemical engineering. Christos
s worked on G. Takoudis joined Purdue in 1981. He has developed research pro-
rent research grams in reaction engineering, catalysis, and chemical vapor depo-
xperiments. sition. The present head of the school, Ronald P. Andres, also
after several joined Purdue in 1981, after 17 years on the faculty at Princeton
rmodynamics University. His research interests are in the areas of chemical
)orous media. physics, aerosol science, and process control.
ly a vice-pres-
i after several RESEARCH
control. Since
ntrol, and op- Current research at Purdue covers a wide spec-
of a book on trum of activities. There is not space to discuss all of
d in industry these activities here, but we will illustrate a few.
ets a research Biochemical Engineering Biochemical en-

Recentli he gineering at Purdue dates back to the work of Lovell
In addition to
s edited three on production of 2,3 butane-diol from corn in the early
;ate. 1940's. By 1970 a strong research program under the
eer and physi- early leadership of Alden Emery and Henry Lim was
manent" visit- in place. Among current studies are Emery's work on
ical engineer- improving the efficiency of algae and plant cell growth
of the hyaline
Sof jaundice. and Henry Lim's work on adaptive control of bioreac-
lhe University tors. Lim's work utilizes highly instrumented bioreac-
iw. He is pre- tors interfaced to a hierarchical system of minicomput-



S1911 -1986 Purdue i' university

Left: Ray W. Fahien, current editor of this journal, doing research toward his 1954 PhD
under the direction of J. M. Smith. Right: Kathleen M. Keville, a current graduate student,
at a cone and plate rheometer in one of the polymer research laboratories.

ers for on-line data acquisition, analysis, optimization
and control. The reactors can operate in the batch,
semi-batch, continuous, or cyclic mode. On-line con-
trol and optimization allow Lim and his students to
force bioprocesses to maximize production of biochem-
ical products, such as penicillin, while regulating cell
growth to maintain cell mass.
The Laboratory of Renewable Resources En-
gineering (LORRE) supports the bioprocessing re-
search of approximately fifty professionals and stu-
dents. Among the many projects that LORRE direc-
tor George Tsao is involved in is the conversion of
biomass to basic chemicals. This work involves all as-
pects of biochemical engineering including, for exam-
ple, the use of genetic engineering to transfer xylose
isomerase genes into yeast to create new cultures that
can yield ethanol from both xylose and glucose.
Mike Ladisch and Martin Okos (faculty members
in agricultural engineering with courtesy appoint-
ments in chemical engineering), Phil Wankat, and
their students are addressing the problem of recovery
of ethanol from the dilute aqueous solutions produced
by bioprocesses. They have devised more energy effi-
cient distillation methods, have found that extraction
with gasoline can produce gasohol directly, and have
developed new methods to produce pure ethanol by
absorbing water from the ethanol-water azeotrope.
This latter process, developed by Mike Ladisch, is
being licensed commercially.
The behavior of cells in multinutrient media is gov-
erned by hundreds of microscopic chemical processes.
Ramki Ramkrishna and his students have found that
the internal regulation in cells can be modelled by as-
suming that cells seek to optimize their growth rate
and respond to changing conditions accordingly. This
cybernetic model has predicted cell responses to

changing nutrients and provides a framework for con-
trol strategies for microbial fermentations.
Control of the local microenvironment in bioreac-
tors requires chemical modification of the reaction
medium. Linda Wang and her students have ac-
complished this control by immobilizing enzymes near
ion exchange sites. In urea hydrolysis, for example,
the local pH is maintained near the optimum for
urease by removing the product carbonate ions by ion
Systems Engineering and Computer Aided De-
sign Perhaps nowhere in chemical engineering does
the computer offer greater opportunities than in the
area of process design and optimization. The univer-
sity's Cyber 205 supercomputer, a CDC 6500, office
and lab terminals linked to a powerful network of
VAX 11/780 computers, as well as the recently funded
NSF Engineering Research Center in Intelligent
Manufacturing provide a most supportive environ-
ment for these studies at Purdue. In addition, the
school has two unique facilities for education and re-
search-the Shell Interactive Graphics Laboratory
consisting of VAX 11/780 computer and 8 MEGATEK
interactive graphics terminals and the IBM Com-
puter-Aided Process Engineering Facility consisting
of an IBM 4341 computer and 22 color graphics termi-
Rex Reklaitis and his students use these facilities
heavily in three separate areas of research. Their
work in batch processing seeks to develop, implement,
and demonstrate computer-aided methodology to sup-
port the design, scheduling and operation of batch sys-
tems. They pay particular attention to the role of in-
termediate storage in the operation of batch plants
and are exploring methodology for automated plan-
ning and scheduling of multiproduct plants. Develop-


Seventy-Tfiti iAtmversary if tfzaScfwofoofCflenz

ment of adaptive local approximation methods for
physical property calculations has been a key to the
effective simulation of the inherently dynamic proces-
ses encountered in batch systems. A second project is
development of a process engineer's work station, in-
tegrating data base management, interactive comput-
ing, and computer color graphics. The present pro-
totype allows interactive flow sheet drawing, material
and energy balancing, and preliminary sizing and cost-
ing of equipment. Ultimately, the work station will

The 1986 plan of study includes
a core curriculum and structured options
in biochemical engineering, computer applications,
engineering science, microelectronics, process
management, and polymers.

allow generation and transmission of reports as well
as the progressive development and evaluation of new
process designs. Rex's research also encompasses the
synthesis and operation of energy recovery systems.
One of its aims is to develop energy management
methodology for energy intensive processes and to
test it under simulated conditions using the IBM facil-
ity and ACS (IBM's process control software).
Materials Many of society's technological chal-
lenges past, present and future are closely linked to
the development of new materials. The chemical en-
gineer's role at this frontier is being increasingly rec-
ognized. Polymer materials and metal clusters are two
areas of current research emphasis at Purdue. Jim
Caruthers' research on the mechanical properties of
polymer solids focuses on predicting the yield be-
havior of amorphous polymers. By carefully consider-
ing volume changes during deformation and combining
a full three-dimensional linear viscoelastic constitutive
equation with a statistical mechanical equation of
state, he and his students are developing new models
for predicting yield stress quantitatively over a wide
range in strain rate. These models also identify
mechanisms for pre- and post-yield behavior in
polymer solids.
Another area of fundamental materials research is
the work of Elias Franses, Jim Caruthers and their
students on the properties of filled polymers. They
have recently succeeded in preparing monodispersed
spheroids of chosen aspect ratio. These unique mater-
ials, together with a new theory for interpreting light
scattering from oriented, shaped particles will permit
them to study particle orientation in suspensions
under shear. Recent theological measurements on
suspensions of the shaped particles show that the par-

ticles do not interact directly but rather are associated
by entanglements of polymer chains absorbed at the
particle surface. The ultimate goal of these studies is
a theological constitutive equation for heavily filled
Particles of another sort are the focus of the re-
search of Ron Andres and his students. Using a care-
fully arranged series of supersonic expansions they
are able to separate the nucleation and growth of
metal clusters. The result is the ability to produce
metal aerosols with controlled cluster size from 5 to
500 atoms and with a size distribution of only +10%.
This uniquely narrow distribution provides special op-
portunities for studying the effect of cluster size on
the catalytic and electronic properties of metals.
SThermodynamics In recent years, K. C. Chao,
H. M. Lin, R. A. Greenkorn, and their research group
have developed a new equation of state, the chain-of-
rotators equation, that explicitly accounts for the rota-
tional motion of polyatomic molecules. The cubic
chain-of-rotators equation, derived from the original,
represents vapor pressure and liquid density to within
2%. Chao and his students have turned their attention
to combining group contribution, equation of state,
and local composition concepts to describe polar fluids
and mixtures. Particularly important to their recent
success has been the effective use of Purdue's Cyber
205 supercomputer for the simulation of polar fluids
and mixtures by averaging millions of microstates
computed by Monte Carlo methods. Their experimen-
tal research of fluid phase equilibria at elevated tem-
peratures and pressures has produced fundamental
data of lasting value.
Other Research Other research activities at
Purdue span applied mathematics, biomedical en-
gineering, chemical processes, colloid and interface
science, kinetics and catalysis, reaction engineering,
semiconductor processing, separations, and transport
phenomena. Many of the experimental research pro-
grams are the direct beneficiaries of the expertise of
David G. Taylor, our instrumentation specialist.

The curriculum has changed substantially in recent
years in response to changing technology and the
broadened knowledge base underlying chemical en-
gineering. Nevertheless, in comparing today's cur-
riculum with that of the 1920's and 30's, one is struck
by the continuity.
Today's undergraduates have the option of doing
a semester research or design project in lieu of an


gineeriu g 1 911 1986 Purdue-' university

elective course or of participating in an honors pro-
gram which includes two semesters of independent
research and a BS thesis. This option had its origin in
an independent research-project course for under-
graduates established by Peffer and Bray back in
Today's undergraduates have hands-on experience
with a modern control computer (Honeywell TDC-
2000) in their senior laboratory course and make use
of real-time computer simulations supported on the
IBM-4341 computer in their required process control
course. A course in process control was established
back in 1919 by George Sherman, existing first as an
instrumentation course and since 1937 as a control
The undergraduate plan of study for 1936 included
a core curriculum required of all students and struc-
tured electives for gas technology, general, metal-
lurgy, and organic technology options. The 1986 plan
of study includes a core curriculum and structured op-
tions in biochemical engineering, computer applica-
tions, engineering science, microelectronics, process
management, and polymers. Throughout this period
the school has stressed development of strong com-
munications skills. This aspect of the program is now
in the able hands of Frank S. Oreovicz, our communi-
cations specialist.
The current graduate curriculum consists of a core
of four courses in applied mathematics, reaction en-
gineering, thermodynamics, and transport phe-
nomena. These courses follow general syllabi agreed
upon by the faculty and have a rigor and depth that
establishes the tone for graduate study. The remain-
der of a student's plan of study may be chosen from
seventeen graduate electives (each course taught in
alternate years), about a dozen graduate-under-
graduate electives, or graduate courses from other de-
partments. These courses are chosen in consultation
with the student's thesis committee both to promote
breadth and to support the student's research in-

The school has had an illustrious past-one that is
intertwined with the history of chemical engineering
education and the chemical engineering profession.
The impact of Purdue goes beyond its 6804 BS and
1208 advanced degrees. One-fifth of the more than
338 PhD graduates from Purdue have continued in
academia, many achieving prominent positions. A re-

The curriculum has changed
substantially in recent years in response to
changing technology and the broadened knowledge
base underlying chemical engineering.

cent survey by the school identified a total of 170
alumni that have become university faculty. The same
1984 survey identified over 300 alumni who are chief
executive officers, presidents, or vice presidents of
companies. The 1983 National Research Council Sur-
vey of graduate programs in engineering ranked the
school second in the country in the "overall influence
of scientific contributions of its faculty on chemical
What can we predict for the future? What else ex-
cept to repeat after Heraclitus, "Nothing is perma-
nent except change." The 75th Anniversary of Pur-
due's School of Chemical Engineering finds both the
chemical engineering profession and chemical en-
gineering education in a period of rapid change.
From the perspective of the changes that have
taken place at Purdue over the past 75 years the cur-
rent situation is seen to be evolutionary rather than
revolutionary. The National Research Council has re-
cently charged a committee headed by Neal R.
Amundson to look to the future of chemical engineer-
ing. The subcommittees for this undertaking:
biochemical and biomedical engineering; chemical en-
gineering aspects of advanced structural materials;
chemical engineering aspects of electronic, photonic,
and recording materials and devices; computer-as-
sisted process and control engineering; energy and
natural resources processing; environmental protec-
tion, safety and hazardous materials; and surface and
interfacial engineering, have titles that qualify as
headings for the research currently being undertaken
at Purdue. Even more important than changes in re-
search direction is maintenance of an environment in
which students recognize, participate in, and generate
excellence in research and a curriculum that is de-
signed to prepare graduates for the needs of today as
well as the continuing evolution of challenges that the
profession will offer in the future.
The year 1986 finds the school stronger than ever.
The presence of an ambitious and enthusiastic faculty
and a dedicated undergraduate and graduate body,
the support provided by a strong research university
and by a large concerned alumni body, and the guid-
ance provided by the history of the past 75 years, all
point to a future even brighter than the past. O


Seventy-iftih Anniversary of wt/eScfwhoo#ofChem

a Deducator

Roast 4. Q4"eenhut

of Purdue

Purdue University
West Lafayette, IN 47907

Do you know any chemical engineer who:

a) Designed a flying submarine?
b) Acknowledged in a text gratitude to a
snooker table?
c) Was shot down into the sea in combat?
d) Painted his garage door in the style of
All of the above describes Bob Greenkorn, profes-
sor of chemical engineering at Purdue University.
a) While teaching in the Department of Theoretical
and Applied Mechanics at Marquette, he became in-
volved in the design of a flying submarine. (Bob ada-
mantly maintains that it was a sinking airplane he
designed, and that his competitors were designing the
flying submarine. For those in the audience who are
now thinking that it shouldn't be too difficult to design
an airplane that will sink-the idea was that it would
sink and then fly, not just the reverse.) At any rate,
he still has a model of the resulting design. A pro-
totype was never built, but it would have been in-
teresting to see if he would have volunteered as a test
b) His interest in snooker is as long-standing as
Andy Capp's and it led both to an inscribed snooker
cue which was presented to him upon his move from
industrial to academic life, and to the above mentioned
vote of thanks to "Snooker Table Number 4 in the
Purdue Billiard Room-where the emotions and frust-
rations associated with writing the text [Transfer Op-
erations] found release."
c) Bob was a pilot in the U.S. Navy during the
Korean War, flying P2V Neptune anti-submarine air-

craft from carriers, and was shot down in the China
Sea. At the end of his Navy tour he held the DFC and
three Air Medals. His interest in flying has remained
undiminished, however, and to this day he is an en-
thusiastic private pilot, holding a commercial license
with multi-engine and instrument ratings.
d) It was during his residency in Tulsa that he
painted his garage door in the style of Mondrian. The
garage door remained a garage door and was never
appropriated by any recognized art gallery; however,
the design, rotated and made three-dimensional, later
appeared on the cover of Transfer Operations.
Bob has been associated with chemical engineering
at Purdue for twenty years, which is about one-third
of the seventy-five years of chemical engineering at
our institution. He started as an associate professor
(1965), became a professor in 1967, and was head from

Copyright ChE Division ASEE 1986


9ineermg l913111986 PurTdue' university

Bob inherited [his parents] health and is an all-around athlete. He developed his vigorous style
of life early . Besides his chemical engineering studies and research, he taught theory and operation of
computers in the Mathematics Department. He completed his monumental generalized correlation of
thermodynamic properties of fluids for his PhD thesis in two years before meeting the minimum
residency requirement. ... He is the author of over 120 publications and four books . .

1967 to 1973. He then held various other administra-
tive positions in the university while continuing to
teach and do research, and presently is Vice-President
and Associate Provost. He directed and influenced the
chemical engineering program at Purdue in several
major ways, shaping it into what it is today. It is
fitting to present a short sketch of him as part of the
celebration of the Diamond Jubilee Year of our school,
for by knowing him one gains much in knowing chem-
ical engineering at Purdue.
Bob was born in 1928 in Oshkosh, Wisconsin, the
second son and youngest child of Frederick and Sophie
Greenkorn. Both parents were of vigorous health and
lived long and full lives. Bob inherited their health
and is an all-round athlete. He probably developed his
vigorous style of life early, but we can only go back
to his graduate student days, when his style was al-
ready developed. Besides his chemical engineering
studies and research, he taught theory and operation
of computers in the Mathematics Department. He
completed his monumental generalized correlation of
thermodynamic properties of fluids for his PhD thesis
in two years, before meeting the minimum residency
requirement. This he did by teaching chemical en-
gineering as an instructor during the third year.
When Bob came to Purdue in 1965, he had already
had a successful career in industry at Jersey Produc-
tion Research in Tulsa and had taught at the Univer-
sity of Tulsa and Marquette University. The five
years at Jersey Production Research saw a decisive
change in Bob. He launched into transport and flow,
made major contributions in porous media flow, and
supervised large industrial experimental research. He
became firmly established in this area, where he con-
tinues to do major work.
Under his direction as head, the School of Chemi-
cal Engineering underwent a remarkable expansion
of the PhD program, supplementing the strong BS/MS
program already in place. Bob conducted a vigorous
recruitment program by approaching the best re-
search groups in the country in the areas where he
wanted faculty and asking for direct recommenda-
tions. In this way he obtained a number of individuals

around whom the greatly strengthened department
was built. He developed quality and emphasis on re-
search and graduate education, without sacrificing the
large undergraduate program already in place.
After his tenure as school head, he was asked to
join the staff of the dean of engineering in the area of
research coordination. His activity in the dean's office
saw a remarkable expansion in the quantity of re-
search dollars generated by the schools of engineering
(in fact, the research budget for the schools of en-
gineering went from $5 million to $22 million annually
during this period). His work also improved proce-
dures, making generation and submission of research
proposals far easier, and he introduced quantitative
models for research and education that were later
adopted by a number of other schools on campus.
He served as director of the Institute for Ad-
vanced Interdisciplinary Engineering Studies from
1972-1975 while he was also the assistant dean of En-
gineering for Research as well as associate director of
the Engineering Experiment Station. During this
period he was called upon to change his chemical en-
gineering orientation to serve as the acting head of
Aeronautical & Astronautical Engineering. He has
also been active in Purdue's environmental area as
director of the Environmental Engineering Center
from 1976-1978, following a one-year term as its acting
director. From 1976 to 1980 he was Associate Dean of
Engineering, Director of the Engineering Experiment
Station, and Director of the Coal Research Labora-
His outstanding performance in the dean's office
brought him to the attention of the upper administra-
tion at Purdue, and he was asked to become associate
provost and vice-president of the institution. Among
his management duties as vice-president is the pro-
grammatic responsibility for all funded research at
Purdue where current annual research expenditures
are over $80 million. He is the Purdue member of the
Indiana University-Purdue University at Indianapolis
management team. (About 40% of the teaching re-
sponsibility of the 23,000 students at IUPUI are in
the Purdue School of Engineering and Technology and


Seventyl-'ffiti .nniuversary )ff tfie-Scwoo ol Clemica 'LTE ineering 111 -1986 Purdue- universityy

Purdue School of Science.)
Another of his current responsibilities is participa-
tion in the business and industrial development of the
state. He was a member of the Executive Planning
Committee of the Indiana Economic Development
Council which wrote the Indiana Strategic Economic
Development Plan, "In Step With The Future."

Greenkorn in the lab.

He is currently a member of the board of directors
of the Indiana Institute for New Business Ventures
(IINBV). The IINBV promotes economic develop-
ment through formation of new business enterprises
in Indiana. He is also a member of the board of direc-
tors of Bemis Company, Inc., a major packaging com-
pany in Fortune's 500, and INventure, Inc., a new
technology enterprise development company.
Despite many years of administrative responsibil-
ity, Bob's research has continued undiminished. His
current research interests include: flow, mass trans-
port, and reaction in spatially variable dispersed
media-in porous media and in the atmosphere;
equilibrium thermodynamic properties of fluids, and
refining of coal. He is also involved in system model-
ling of the environmental flow of trace elements and
simulation of universities.
He is the author of over 120 technical publications
and four books, and holds several patents in the area
of petroleum production. Bob is a registered profes-
sional engineer, a Fellow of the American Institute of
Chemical Engineers, and a member of the American
Association for the Advancement of Science, Amer-
ican Chemical Society, American Geophysical Union,
American Society for Engineering Education, and So-

city of Petroleum Engineers of the AIME. He is also
a member of the Accreditation Board for Engineering
and Technology (ABET) and a Registered Profes-
sional Engineer in Oklahoma and Wisconsin.
He has acquitted himself admirably in teaching as
well as in research and administration, and has won
the Outstanding Teacher in School of Chemical En-
gineering Award, 1965-1966, and the Tau Beta Pi Out-
standing Teacher Award, 1969. He is also the holder
of the Shreve Prize for 1971 in the School of Chemical
Engineering. He was recognized last year with a Dis-
tinguished Service Citation from his alma mater, the
University of Wisconsin-Madison.
Bob has long been a sports enthusiast, both as a
participant and as a spectator. He was named one of
ten members of the Oshkosh High School Hall of Fame
in 1984. The mayor of Oshkosh has also presented Bob
with the key to the city on his specially named day.
From being a competitive swimmer/diver in high
school he progressed to track (discus and high-jump),
football (running back and kickoff) and basketball (all
West Coast Service Team) in the Navy at Pensacola,
Seattle, and San Diego.
Until about a year or so ago, he was a regular
member of a noon-time group at Lambert Gymnasium
that played what was characterized by its members
as "basketbrawl." A rag-tag group of assorted former
college and professional basketball and football
players, faculty members, lawyers, tavernkeepers, in-
surance salesmen, librarians, and various other
geriatric athletes, this organization played enthusias-
tic basketball each noon (except for the occasional
fistfight-there were no referees) for many years.
One group member played four years as a starting
forward in Big 10 basketball without an injury-then
he joined the basketbrawl group and was immediately
incapacitated. The demise of the group was brought
about, alas, not by old age or antagonisms, but by the
rededication of the space by the university to newer
and supposedly nobler purposes. Presently Bob is re-
duced to jogging for several miles each noon, but re-
mains a faithful attendee at Purdue football and (par-
ticularly) basketball games. For outside interests he
sings in a choir, plays the classical guitar, and is in-
terested in art, astronomy, and the study of nature.
This has been an article about Bob and his ac-
tivities, but we cannot close without mentioning the
rest of the Greenkorns. Bob and Rosemary Greenkorn
have been married for 32 years and have four children:
David, Eileen, Nancy, and Susan. Purdue University
and the West Lafayette community are fortunate to
count them as associates and friends. D


MPl .views and opinions


If you dare

University of Petroleum and Minerals
Dhahran 31261, Saudia Arabia

ONE OF THE side-effects of being in an academic
environment is that we are protected from the
day-to-day problems faced by a typical engineer in
industry. Not the least of these problems is corrosion,
and bearing in mind how much it costs these days, it
is rather surprising that this subject is not given more
emphasis at the university level, particularly in the
chemical engineering curricula.
For many instructors, being assigned Corrosion
and Corrosion Control ranks in the category of a
nightmare, especially after semesters of stage to stage
calculations, well-behaved chemical reactions, heat
transfer, and so on. It comes as a bit of a shock, but
in the end it is immensely rewarding.
Corrosion control, it seems, is very much like tak-
ing care of your health. It takes will-power and great
effort to keep in shape and its care is often left too
late to be really effective. This fact alone presents a
strong case for a dedicated corrosion team in the
plant. If it's a part-time job it will always be neg-
A couple of months into the course one realizes
how broadly based the subject material is. It em-
braces physical, organic and inorganic chemistry,
metallurgy and ceramics, strength of materials, ther-
modynamics, plant design, and economics. It means
digging deep into forgotten corners of the mind and
pulling down some of those texts that you thought had
been closed forever.
Saudi Arabia has an inherent corrosion problem.
The ground is very saline, the natural waters are
brackish, the seawater is heavy with dissolved solids,
temperatures are high, there is lots of sunshine,
humidity is often close to 100%, the gas and oil are
sour, and there are great distances to transport them.
These are enormous driving forces. The University of
Petroleum and Minerals, its associated Research In-
stitute, and ARAMCO (Arabian American Oil Com-
pany) are doing battle against this powerful enemy on
its own territory.
Unfortunately it is difficult to tempt students into
this field. We offer it as an elective and as such it

Lyndon Hadley-Coates was educated at the University of Leeds,
England. After a short period working as an engineer on oil rigs in
Abu-Dhabi and teaching in England, he joined the University of Pe-
troleum and Minerals, Dhahran, Saudi Arabia. His research interests
are in the fields of residual-oil technology, mass transfer, and more
recently, corrosion.

competes with refining, petrochemicals, surface oper-
ations, desalination, etc. The Kingdom is an oil-pro-
ducing nation and some of the alternatives do look
very tempting. The population is small and a shortage
of students affects all disciplines due to the high de-
mand for engineers. Most of the good chemical en-
gineering majors are attracted to the traditional
areas, leaving the fringe subjects almost untouched.
The students argue that promotion comes more easily
this way. Become a specialist and you cannot be re-
placed, they say. Perhaps there is a psychological bar-
rier too. An engineer's job is making things happen,
not stopping them from happening. It's a lot easier to
think positively than anti-negatively.
A change of attitude is in the air however. Aramco
has made it clear recently that they are saturated with
business management majors and have introduced an
excellent program designed to create a pool of Saudi
engineering specialists that will include corrosion en-
gineers. Most recently the university, in collaboration
with NACE, sponsored a two-day corrosion sym-
posium, which was supported at the highest level.
As for me, I'm a confirmed corrosion man. Re-
learning electrochemistry was the activated step,
but now I actually enjoy it. It is no longer a night-
mare. F
Copyright ChE Division ASEE 1986



m |

D curriculum




Purdue University
West Lafayette, IN 47907
University of Waterloo
Waterloo, Ontario, Canada N2L 3G1

PROCESS CONTROL IS A difficult subject to teach
because it ranges all the way from abstractions
such as calcualtions in the complex frequency domain,
to realities such as how a distillation tower will re-
spond to a change in the reflux rate. Textbooks and
lectures are not the best mechanisms for teaching stu-
dents how to deal with real-time changes occurring
over time spans of minutes, or even hours, in a real
plant. Looking at a textbook graph of a previously
generated plant response is a poor substitute for ob-
serving the real response. And, no textbook can pro-
vide a way for the student to test, tune, and revise a
control strategy, which is a central thrust of profes-
sional practice.
These difficulties were highlighted in a recent
trade article ("Putting College Back on Course,"
Chemical Engineering, Sept. 19, 1983, pp. 48-60), re-
porting a survey in which 4,759 engineers evaluated
their undergraduate education. Process control
courses were singled out with comments such as: "In
college, process control is all Laplace transforms and
transfer functions in the s-domain. In the real world,
I need to .. understand how control systems work."
The article summarized its findings with this recom-
mendation: "Courses in process control should have a
practical orientation. Graduates are not using Laplace
transforms and other control theory, but they do have

This article traces some of the
history behind the IBM/academia arrangement,
describes ACS and how it is currently integrated
into the introductory undergraduate process control
courses, discusses benefits to the students and
instructors, and summarizes future plans.

*Present address: SETPOINT, Inc., Houston, TX

Lowell Koppel received his BSChE and PhD from Northwestern Uni-
versity and joined the chemical engineering faculty at Purdue Univer-
sity in 1961. He has been a frequent contributor to the literature and
is co-author of a widely used introductory text on chemical process
control. Recent awards include the 1982 Chemical Engineering Lec-
tureship Award from ASEE and the 1984 "Innovations in Helping Stu-
dents Learn" award from Purdue University. He has recently joined
SETPOINT Inc. in Houston. (L)
Gerry Sullivan has a BASc degree in chemical engineering from
Waterloo and a PhD from Imperial College in London, England. He has
several years experience working for a major petrochemical company
in the design and implementation of microprocessor and computer-
based process control, optimization and scheduling systems. He is in-
volved in an active research program in process design, control and
plant-wide systems. (R)

a pressing need for information on how to design and
operate practical control systems." As academicians,
we can recognize this as a manifestation of the "never
mind the theory, just teach me how to solve real-world
problems" syndrome. Nevertheless, the perception
regarding process control education is probably
symptomatic of the problems mentioned above.
A typical physically-based process control educa-
tional laboratory partially addresses these problems.
However, the large numbers of students and the prob-
lems of maintaining equipment (particularly when
used intermittently) make it difficult for universities
to operate such facilities. Furthermore, laboratory
processes are usually considerably smaller in scale

Copyright ChE Division ASEE 1986


than commercial facilities, thus negating some of the
real-time feeling. Such laboratories usually have fixed
control strategies which the students can test and
tune, but not revise in favor of an improved strategy
which they have synthesized.
The general-purpose computing facilities increas-
ingly available in most universities can help overcome
some of these difficulties. The physical process can be
replaced with a simulation, and the standard control
strategies can be pre-programmed. Maintenance
problems are greatly reduced. However, interactive
operation in real-time, over periods of minutes, is
cumbersome and expensive on most computing
facilities available for student use. In addition, pro-
gramming the software needed to allow user-friendly
implementation of alternative control strategies is a
major effort, to say nothing of all the other software
needed for effective student-to-computer interfacing.
One solution to this problem is the use of indus-
trially developed software, on a computing installation
devoted in major part to process control education.
Purdue University and the University of Waterloo are
now doing this. IBM's Advanced Control System
(ACS) is a licensed program for implementing plant
automation and is currently in use in refineries and
plants around the world. Chemical engineering under-
graduates at Waterloo and Purdue are using ACS, on
color graphics terminals supported by an IBM 4341
mainframe, to study process control.
This article traces some of the history behind the
IBM/academia arrangement, describes ACS and how
it is currently integrated into the introductory under-
graduate process control courses, discusses benefits
to the students and instructors, and summarizes fu-
ture plans.
In February, 1982, chemical engineering academi-
cians from several institutions attended an IBM semi-
nar on ACS, held in Houston. The purpose was to
expose the academicians to the capabilities of ACS
and to explore the possibilities for using it in chemical
engineering education. Considerable enthusiasm was
expressed, and in September, 1982, one of the authors
(LBK) met again with IBM officials in Houston, this
time to formulate a concrete arrangement for (1) in-
stalling the hardware and software at Purdue, (2) de-
veloping study guides for students to use ACS in an
educational framework, and (3) making the facility
available to students at other institutions. Final
agreements between IBM and Purdue University
were completed early in 1983, and the hardware and
software were delivered, installed, and operating by
July, 1983.

As stated by IBM, the ACS
package is a licensed program providing
a framework for implementing plantwide process
control in medium- to large-scale plants.

At Waterloo, the ACS system was first installed
in March, 1982, as part of a contract with IBM to
develop a two-week intensive engineer's user course
for ACS. At that time, only the ACS software was
delivered in order to allow course development. The
ACS system shared hardware resources with 350
other university users on a network of three IBM 4341
computers. Obviously, response was less than desir-
able. In May, 1983, as a result of a partnership
agreement with IBM Canada Ltd., and the University
of Waterloo, a dedicated IBM 4341 was installed solely
for ACS and computer-aided design activities in the
chemical engineering department. This, along with an
IBM Series I Interface computer, has allowed real-
time application of ACS for control on some of the
pilot plants (e.g., cyclic feed reactor, single cell protein
In the fall semester of 1983, 168 chemical engineer-
ing students at Purdue and 75 at Waterloo used ACS
as part of their introductory process control course.
Currently, intermediate process control elective
courses are being taught at Purdue (with over 50 stu-
dents) and at Waterloo (with over 40 students), all
using ACS. The remainder of this paper will be con-
cerned with the development and delivery of the intro-
ductory process control course.

As stated by IBM, the ACS package is a licensed
program providing a framework for implementing
plantwide process control in medium- to large-scale
plants. It uses IBM general purpose and sensor-based
computers for execution of control strategies and for
single-source storage of information for plant and
operating management.
As seen by plant operating personnel (and stu-
dents), ACS is (among other information management
functions) a user-friendly, menu-driven software
package for implementation of computer process con-
trol. The operator sits before a console, typically made
up of two or more color-graphic terminals, and can
call up a variety of vivid and easy-to-use "live" dis-
plays from which to monitor and control the plant.
Among the most frequently used of these are
SProcess schematics (Figs. 1,2). These can show the pro-
cess flow and the control strategy. They are operating


displays with real time update and change capability.
Faceplate displays (Fig. 3). These mimic the faceplates
of older board-mounted controllers.
Multitag plot displays (Fig. 4). These display short-term
trend plots of up to four single tags, and can be converted
to historical trend plots at the press of a button.

Control strategies for ACS are defined on-screen
by fill-in-the-blanks forms. Built-in control algorithms
are supplied for P-I-D, lag/lead, ratio, multi-level
bumpless-transfer cascades, high-low (constraint) con-
trol, interactive decoupling, and nonlinear or error-
squared control. In addition, a General Algorithm
facility enables straightforward programming of cus-
tom strategies, such as energy balance control.


A key aspect of the educational use of ACS is that
the manufacturing process itself can be simulated and
simultaneously controlled on the host computer. A
variety of process simulations are already available,
including the furnace shown in Fig. 1. These can be
constructed either within an ACS General Algorithm
(as is the furnace), in the User Fortran Interface (a
slave partition provided with the ACS system), or in
dynamic simulation languages such as SPEED-UP.
The simulations give a real-time feeling by having

TS 900

FIGURE 1. Overall schematic for furnace process. The
black-and-white reproduction belies the vividness and
easy comprehensibility of this and the other color ACS
displays presented in this paper.

time constants on the order of minutes. The students,
when operating the simulated process from the ACS
displays, experience virtually the same events that
would confront operators of the real process, with the
exception of drastic events such as equipment failure.

Study Guides for ACS Process Control Units

I Basic mechanics of ACS, such as display callup and enter-
ing changes.
II Introduction to the furnace process. More basics of ACS.
How to operate furnace through manual changes in fuel
oil and air flow rates.
III Manual operation and representation by single-variable
block diagrams.
IV Manual operation and representation by multivariable
block diagrams.
V Disturbances and their effects. How to overcome by ra-
tional manual operation. Limitations of manual opera-
tion for multivariable processes. Motivation for atuoma-
VI Introduction to feedback control. Study of single-loop
proportional control and its implementation to regulate
the furnace. Effect of changing gain.
VII Study of integral feedback. Implementation to regulate
the furnace by single-loop proportional-integral control.
Effects of changing integral time. Advantages and disad-
vantages of integral feedback.
VIII Study of derivative feedback. Implementation of single-
loop proportional-derivative and proportional-integral-de-
rivative control to regulate the furnace. Effect of chang-
ing derivative time. Advantages and disadvantages of de-
rivative feedback.
IX Tuning control loops by process reaction curve method
and by closed-loop cycling method. Tuning of loops domi-

nated by delay time. Use of the furnace regulatory loops
to practice and evaluate these tuning methods.
X A study of the effects of integration between control loops.
Effects of reversing loop pairings to alter the furnace con-
trol strategy.
XI Open-loop frequency response. Students apply sine waves
to fuel oil valve, observe temperature response, and pre-
pare Bode diagram. Comparison of results with point ob-
tained by loop tuning in Unit IX.
XII Closed-loop frequency response. Students apply sine wave
distribution in furnace throughput, observe temperature
response, prepare Bode diagram of ratio of responses with
feedback loop closed and open. Verify that loop is vulner-
able near the resonant frequency.
XIII Feedforward energy balance control. Use of a general al-
gorithm to compute correct fuel oil flow from measured
disturbance and feedforward to the correction through a
lag/lead compensator. Tuning of the lag and lead time
constants. Advantages and disadvantages over feedback.
XIV Cascade control. Purposes and implementation. Bumpless
transfer and coupling. Implementation of combined feed-
forward and feedback. Tuning of the controllers in these
combined circumstances, compared with tuning when op-
erated singly. Grand control strategy for furnace, using
all previous concepts plus ratio and material balance.
Comparison of results to show advantages of rational
strategy based on process knowledge vs. "black box" feed-
back strategies.


A good analogy is a flight training simulator. This ad-
dresses many of the difficulties of process control edu-
cation mentioned in the introduction. Students can
quickly perform and see the immediate impact of open
loop response tests, control strategy design, and
changes in loop tuning parameters.
We have chosen a series of self-study guides as the
vehicle with which to use ACS to introduce students
to the practice of process control. These guides were
written at Purdue as a series of Units, and were
tested both at Purdue University and the University
of Waterloo during the 1983.fall semester. The units
are designed to form a complete support package for

2. 000


.000oo I
TS F 004 IMP 8%7.14 8066.7

TS 949

612.50 j
TS F N IM1P 725.7 SP 3724.4

IP 0e.e01 SP 6e.e1 OUT 51.271
-- bO.L- -

FIGURE 2. Schematic of a control scheme for furnace
process. Note setpoint change being entered at bottom
of display.

Ts 4w4


cs 053
a. .


F ee, T 0ee cs 051 F 03
SI a
I. I -




TS 94 PAGE 1/1

F 894 AL 05
l- -


*4 *** * 0x E.P
INPUT 100.0e 250se. 51.7 .k350.b 586.b a5.3 2.0314 b1e.25
SETPT 1e.00 250.00 51.775 324.2 2.144 77T2.7 1.9 he.01
OuT: bb.hB8 37.501 51.775 51.509 38.813 39.b42 21.144 51.e07


FIGURE 3. Faceplate display. Displays a user-defined
group of variables.

FIGURE 4. Multitag plot display. Short-term trend plots
of a user-defined group of up to four variables. Pressing
programmed-function key converts it to an historical
trend plot.

a traditional introductory undergraduate course in
process control. Titles and brief descriptions of the
introductory units are shown in Table 1. The units are
now fully integrated into the Purdue and Waterloo
Additional units dealing with other simulated pro-
cesses, and with more advanced control strategies,
are either available or being developed. In these ad-
vanced units, students are allowed to develop, imple-
ment, tune, and test their own control strategies, for
a variety of processes. These units are designed for
students interested in learning more about process
control than is covered in a typical introductory
course, presumably in the framework of an elective
follow-up course.
Figures 5 through 8 illustrate some of the hard-
copy responses the students obtain while perfornrring
the tests in the study guides, and offer some insight
into the material covered. On all response plots, adja-
cent points are 8 seconds apart; the entire time axis
covers approximately 7.5 minutes of real time.


At Purdue, the introductory process control course
is required for all chemical engineering students. To
incorporate ACS, the three-credit hour class has been
structured into two formal lecture sessions, given to
the entire class, and one recitation session, given to
the class in six groups of approximately twenty-eight
students each.
The lectures cover a traditional process control

Continued on page 106.


TS q44

M AiLs00

Control off
*turn -4 on
T I 7
* messages
* plot 11I
TS T 5eese



Auburn University
Auburn, AL 36849

PROCESS CONTROL HAS become a subject of more
importance in recent years because of two major
trends. The first trend is that increased energy con-
servation, more stringent environmental standards,
and tougher economic competition have forced indus-
try to use more complex processes than in the past.
These processes are more difficult to control, yet con-
trol requirements have also increased since processes
often operate near process limits. The second trend is
that rapid increases in computing power and rapid de-
creases in computing cost have made feasible more
complex control strategies. The result is the need for
more training in process control at the undergraduate
level for all engineers rather than only for those with
special interests or career goals in process control.
Recognizing this, the chemical engineering faculty at
Auburn University doubled the process control course
requirements during a recent study and revision of
the total undergraduate curriculum. The 1985 class
was the second class to graduate under the revised
curriculum. We would like to describe our experiences
in the process control sequence.
Some background information should be provided
first. Our undergraduate enrollment has followed na-
tional trends, with the graduating class being approx-
imately 25 ten years ago, 90 two years ago, and 70 in
1985. Typically, 85-90% of the graduates take indus-
trial positions. The entering freshman class in 1985
had an average composite ACT score of approximately
The process control sequence consists of two four-
credit hour lecture courses taken in the first two quar-
ters of the senior year and a two-credit hour labora-

The result is the need for more training
in process control at the undergraduate level
for all engineers rather than only for those with
special interests or career goals in process control.
Recognizing this, the . faculty .. doubled
the process control requirements.

Copyright ChE Division ASEE 1986

tory course which may be taken in either the second
or third quarter. The chemical engineering prerequi-
sites include courses in material and energy balances,
thermodynamics, fluid mechanics, heat and mass
transfer, stagewise processes, reactor engineering,
and unit operations laboratory.

The text is Chemical Process Control by
Stephanopoulos [1]. The first course covers Introduc-
tion to Process Control (2 lecture hours, Ch. 1-3):
motivation for control, control hardware; Mathemati-
cal Modelling (12, Ch. 4-9): art of modelling, lineariza-
tion, deviation variables, Laplace transforms, transfer
function models; Dynamics of Linear, Single Variable
Processes (6, Ch. 10-12): first order processes, second
order processes, higher order processes, dead time,
inverse response; and Analysis and Design of Feed-
back Control Systems (13, Ch. 13-18): PID control,
stability analysis, controller tuning, and frequency re-
sponse analysis. The second course covers Digital
Control (16, Ch. 26-30): hardware, samplers and
holds, z-transforms, pulse transfer functions, stability
analysis, position and velocity PID algorithms, design
of algorithms for specified response including dead-
beat control and Dahlin's algorithm; Advanced Topics
in Single Variable Control (6, Ch. 19-21): dead time
compensation, inverse response compensation, cas-
cade control, selective control, split-range control,
feedforward control, ratio control; Multivariable Sys-
tem Control (8, Ch. 23-24): selection of control inputs
and outputs, interaction analysis, decoupling control;
and Other Topics (5, not in text): control valve charac-
teristics and sizing, P&I diagrams. The text is fol-
lowed closely with occasional omission of a topic, e.g.,
root locus, and the occasional addition of a topic not
in the text, e.g., the Shannon sampling theorem and
aliasing of sampled periodic signals. The last topics
covered in the second course are the only major excep-
tions to following the text closely. The level of cover-
age of control valve characteristics and sizing is more
like that of Smith and Corripio [2], and manufacturer's
literature is used for some homework problems. P&I
diagrams are taught by examples donated by various
industries. A useful reference here is Warren [3,4].


Dennis C. Williams is assistant professor of chemical engineering
at Auburn University. He received his BS from Auburn University and
PhD from Princeton University. His research interests include process
modelling, nonlinear control, adaptive control, and modelling and con-
trol applications in the pulp and paper industry and the textile indus-
try. (L)
Arthur Ray Tarrer is professor of chemical engineering at Auburn
University. He received his BS from Auburn University and MS and
PhD from Purdue University. His research interests include catalysis,
coal liquefaction, demetallization kinetics, hazardous wastes, waste
oil reprocessing, and waste water treatment. (R)

The teaching style is traditional. The lecture over
the quarter averages to be half spent on lecturing on
new material and half spent on reviewing assigned
problems. One or two homework problems are as-
signed each class and collected at the next class. Some
computer programs have been written by the instruc-
tors for aiding on homework, e.g., calculation of fre-
quency response from a transfer function. Occasion-
ally, computer programs written by the instructors
have been used as the basis for a lengthier homework
assignment. This has been done more commonly in
the laboratory course and is described in more detail

In the laboratory course five or six experiments
are performed. The experiment sequence is model
identification, controller and instrument calibration,
and controller tuning, with two experiments in each
category being typical, although there is wide vari-
ation. The students work in groups of three or four.
For each experiment there is a conference prior to the
experiment where the experiment objectives, proce-
dures, and theory are explained. For each experiment
the group makes a written and an oral report. The
requirements for the written report normally are
specified and involve only specific calculations, e.g.,
determining a transfer function. The oral report
covers the group's interpretation and evaluation of the

results of the written report. The group must also
justify the computation methods and assumptions
made in preparing the written report. The most satis-
factory schedule has the pre-experiment conference
on Wednesday, the experiment on Thursday or Fri-
day, the written report due on Monday or Tuesday,
and the oral report on Wednesday. Unfortunately,
due to other schedule constraints we have not been
able to maintain this schedule every quarter.
Both digital control and pneumatic analog control
are used. The analog controlled processes include level
control of a single tank, level control of two tanks in
series, and temperature control of a tank heated by
steam injection. The digitally controlled processes in-
clude level control of two tanks in series either in-
teracting or non-interacting, temperature control of
an oil reprocessing facility, multivariable pressure and
flow control of two air storage tanks in series, and
multivariable temperature and level control of a
stirred tank. Two additional equipment items are
being added for next year: packed distillation column
control and temperature control of two stirred tanks
in series with a variable time delay.
The CRISP control system from Anaconda Ad-
vanced Technology of Columbus, Ohio, is used to con-
trol most of the digitally controlled processes. This is
an off-the-shelf system based on a PDP 11/23 and
featuring three high resolution, eight-color CRT
operator stations. This provides the students with ex-
perience on the type of system now commonly being
installed in industry. The oil reprocessing facility is
controlled by a Texas Instruments PM550 system.
The distillation column is interfaced with a MACSYM
2 from Analog Devices of Norwood, Massachusetts.
In addition to experiments performed in the labo-
ratory, computer experiments have also been as-
signed. As an example, a program was written to
simulate the digital PI concentration and level control
of a blending process. The students were asked to
tune the controllers with one loop open, close both
loops, and compare the control with control obtained
with a decoupling controller, which they also were
asked to tune.
As a result of two years of experience with the
sequence, we have made several observations. Some
of these are specific to the control courses, but others
are more general and apply to other courses as well.
The students are better prepared in process control
than under the previous curriculum, which had one
required lecture course and a one credit hour labora-
tory course. Whether they are sufficiently prepared is
open to question. Our observation is that the control


In the laboratory course five or six experiments are performed. The experiment
sequence is model identification, controller and instrument calibration, and controller tuning,
with two experiments in each category being typical, although there is wide variation.

of multivariable processes is not that well understood
by many students. The exposure to control equipment
is too limited, although this is an area more easily
remedied on the job. Some concepts such as adaptive
control, optimization, and filtering have not even been
mentioned. Perhaps these are unnecessary for the BS
chemical engineer, but as these concepts become more
widespread as in Dynamic Matrix Control or the new
expert systems from vendors, the engineer who does
not understand how the process is being controlled
will be at a disadvantage.
The course is extremely fast paced from the stu-
dents' perspective. Many of the students have com-
mented that for them these are the hardest courses in
the curriculum. This implies that the topics mentioned
above as being incompletely covered or omitted can
not be included unless the course is expanded or some
currently covered material is deleted.
The mathematics obscures the engineering for
many students. Our students take four quarters of
calculus and a three credit hour course in ordinary
differential equations. In the course, as it is now
taught, the students must learn Laplace transforms,
difference equations, and z-transforms. They must
make a large effort to become proficient with these
techniques, yet in doing so, they have learned nothing
about control. They have generally not had linear
algebra, statistics, or optimization. The lack of linear
algebra makes work on multivariable systems particu-
larly difficult.
In a course with such a fast pace the text or distri-
buted class notes must be followed. In 1984, the first
course text was Coughanowr and Koppel [5], and the
second course text was Palm [6]. The material in Palm
was not in the order taught and was intermingled with
material which was not covered. Since the material in
class was presented more quickly than most students
could follow, they had to rely on study of the text.
With the material scattered many students were un-
able to learn from the text. This implies that if the
course is changed and no suitable text can be found,
the instructor must prepare and distribute class notes
on the level of a text for those sections of the course
not covered by the text.
Students often do not apply their engineering
knowledge and rely on brute force to solve problems
based on computer simulation. In the blending prob-
lem described above, the students were told to use
any method they chose to tune the controllers. The

only hint given was that simply guessing controller
parameters until the simulation showed the desired
response was probably a poor way to solve the prob-
lem. Despite this hint, that method was chosen by all
students, even though the problem was essentially
identical with "pencil and paper" homework problems
worked in the first course. Since use of simulation
packages is becoming more common in industry, it is
important to train students to use such tools properly.
The beginning section of the first course on
mathematical modelling is very beneficial as an inte-
gration of the sophomore and junior courses. There
is some discussion on modelling in general, but
primarily this time is spent assigning and reviewing
problems similar to those worked in earlier courses.
The two key differences from earlier courses are that
nothing is steady state and that students can not rely
upon "this is the heat transfer course, so this must be
a heat transfer problem" thinking. It is initially dis-
couraging for students to find themselves unable to
work problems they felt they had previously mastered,
but this begins the process of unifying the previously
compartmentalized knowledge of courses into a whole
body of chemical engineering knowledge.
The laboratory helps the students' understanding
of the lecture material. Of course, this should be true
of any laboratory course, but it is particularly benefi-
cial here because the lecture courses seem so abstract.
Ideally, the laboratory should be split into two one-
credit hour courses accompanying the lecture courses,
but this is impossible for us with our current enroll-
ments and laboratory space.
The laboratory reporting method is popular with
the students and the instructors. The instructor does
all the grading and conducts all the oral reports. This
is very time intensive, but it provides excellent feed-
back on student understanding. The written report
requirements are kept to a minimum, but we have the
freedom to do this because the written report require-
ments in other required courses are extensive.

We feel the process control sequence is valuable as
a requirement for all undergraduates. Although the
courses are difficult, most students feel the material
is valuable. We have received only a few comments
from former students about the value of the courses,
but they have been favorable. With the rapid changes


in technology the contents of the course may need
revision, but the amount of time devoted to process
control will not decrease.

1. Stephanopoulos, George, Chemical Process Control, Prentice-
Hall, Englewood Cliffs, NJ, 1984
2. Smith, Carlos A., and Armando B. Corripio, Principles and
Practice of Automatic Process Control, John Wiley and Sons,
New York, 1985
3. Warren, Cliff W., "How to Read Instrument Flow Sheets. Part
1," Hydrocarbon Processing, 53, 163-165, (July, 1975)
4. Warren, Cliff W., "How to Read Instrument Flow Sheets. Part
2," Hydrocarbon Processing, 53, 191-193, (September, 1975)
5. Coughanowr, Donald R., and Lowell B. Koppel, Process Sys-
tems Analysis and Control, McGraw-Hill, New York, 1965
6. Palm, William J., III, Modeling, Analysis, and Control
of Dynamic Systems, John Wiley and Sons, New York,
1983 D

gi stirred pots

The tradition of honoring the recipient of the
Lacey Lectureship Award each year with a poem
tailored to that recipient's particular interests con-
tinues at CalTech. The latest tribute was written by
Professor R. A. Aris and recognizes in verse the
Nineteenth Annual William N. Laey Lecturer, Tom

(With apologies to George Gordon Lord Byron)
He talks of transport like a night
Of boisterous wind and stormy skies,
Of mass that's moved from left to right
By Djs and vis,
Enhanced by eddies, loose or tight,
Which laminated flow denies.
One whorl the more, one turb the less
Had half impaired the transfer rate
Which 'stricted eddies would repress
In space which walls delineate,
Where Navier and Stokes express
How swift, how soon they dissipate.
So from the Shell Distinguished Chair
To Lacey's campus near the sea
Rings out the challenge, fair & square:
What shall th'interpretation be
Of matters physical transfer?
Sincerely yours, Tom Hanrattee.

R. A. Aris


Dear Editor:
The articles of Professors Luss and Denn in one
issue (Winter 1986) indicate the width of knowledge
our society would wish to expect of chemical engineer-
ing graduates. While it is important that a chemical
engineer involved e.g. with reactors should be equally
conversant with spreadsheets, the FORTRAN/PAS-
CAL/C languages and with multiple steady states/
bifurcations, our current (and perhaps inordinate) en-
rapture with computing devices threatens to relegate
the latter i.e. analytical thinking and a clear under-
standing of fundamentals, to second place. We will
doubtless form one day a mature attitude towards this
fast and powerful machinery, but until then, we must
insist relentlessly that our students acquire a balanced
view of what computers can and cannot offer.
T. Z. Fahidy
University of Waterloo

book reviews

By J. S. Rowlinson and F. L. Swinton
Butterworths, London, 1982. $69.95
Reviewed by
Keith E. Gubbins
Cornell University
Thirteen years have elapsed between the appear-
ance of the second and this, the.third, edition of this
authoritative monograph on liquids and liquid mix-
tures. In that time much has happened in both the
experimental and theoretical aspects of the subject.
Many new and more accurate measurements of ther-
Continued on page 83.


special fall issue devoted to graduate education. This issue consists
1) of articles on graduate courses and research, written by profes-
sors at various universities, and 2) of announcements placed by
ChE departments describing their graduate programs. Anyone in-
terested in contributing to the editorial content of the fall 1986
issue should write the editor, indicating the subject of the contribu-
tion and the tentative date it can be submitted. Deadline is June


SP classroom



University of Missouri-Columbia
Columbia, MO 65201

ALONG HELD criticism of academic instruction
and research in chemical process control concerns
the gulf that is so often perceived between what is
discussed in the classroom and university laboratory
compared with what is actually used in the field. This
problem extends to undergraduate courses in basic
process control. In a recent survey [1], chemical en-
gineers rated the teaching of Laplace transforms, for
example, as being among the least useful of under-
graduate topics. Nonetheless, most instructors in
chemical process control would probably agree that
the use of the Laplace transform is essential for the
understanding of dynamic processes.
However a dilemma has arrived at the process con-
trol classroom door-sampled data for digital process
control. Clearly, it is no longer adequate to give the
undergraduate a course involving only systems with
analog controllers. Almost all new process control is
installed as sampled data processing; the topic must
be included.
Discrete system analysis has raised a new spectre
for students in undergraduate control classes. Just
as the typical student is becoming comfortable with
Laplace transforms, a new and equally mysterious op-
eration called the Z-transform is introduced to deal
with discrete processes. This situation is exacerbated
by the fact that this topic usually appears near the
end of the semester when insufficient "acclimatizing"
time remains. It doesn't help that the algebra and
arithmetic associated with Z-transforms is just as
complicated as that with Laplace transforms. And the
results are probably more opaque.
The final irony is that, in fact, not much chemical
process sampled data control system design is actually
performed via Z-transform methods. Particularly for
sampled data processes, even the control specialists
tend'to think and work in the time domain rather than
the transform domain. This is because in the time do-
main sampled data methods are simpler, easier, and
Copyright ChE Division ASEE 1986

are intuitively more understandable. These seem to
be good reasons why it might be better, in under-
graduate control courses, to develop sampled data
processes in the time domain, leaving Z-transforms to
advanced courses (if at all).
In the following, we summarize and compare time
domain and Z-transform methods for sampled data
process control design and analysis. We show that all
of the results and concepts usually considered in un-
dergraduate exercises can be developed more quickly,
clearly, and intuitively in the time domain. In fact
there are only a few special situations where the Z-
transform could be considered to have advantages
over the time domain.

Although most chemical processes exhibit non-
linear dynamic behavior, a linear model that approxi-

Dr. Luecke received a BSChE in 1953 from the University of Cincin-
nati, and worked for nine years for the DuPont Company. He returned
to graduate school and received a PhD in chemical engineering from
the University of Oklahoma in 1966. After two years with the Central
Research Division of the Monsanto Company in St. Louis, he joined the
faculty of chemical engineering at the University of Missouri-Colum-
bia, and teaches design, process control, process simulation, advanced
computing, and engineering mathematics. His research areas are pro-
cess control, simulation and optimization, and simulation of physiolog-
ical systems. (L)
Hsin-Ying Lin is a PhD candidate in chemical engineering at the
University of Missouri-Columbia. He received his BSChE from Chung-
Yuan University at Chung-Li, Taiwan and his MSEE from the University
of Missouri-Columbia. His main research interests are in the areas of
process control, simulation and optimization. (R)


mates the dynamics over a specific operating range is
valid in most cases. Linearization allows the use of
Laplace or Fourier transforms. Operational calculus
and transform techniques were adopted in the early
half of this century for analysis of system dynamics
and to assist in control system design.
Extensive research and development effort has
been devoted in the past two to three decades toward
a different approach-time domain analysis of dy-
namic systems. Despite this work, a survey of recent
texts indicates that methods which are taught in
standard texts for analysis of dynamic systems and
design of process control systems are still principally
concentrated on the transform domain. This is true
not only for continuous or analog control but also for
sampled-data processes. Popular textbooks for sam-
pled-data process control [2-6], and including a recent
chemical engineering text [7], all focus on transform
Design for sampled-data process control systems
may be based on continuous system methods and Z-
transforms used to evaluate the resulting sampled
data response; or the design may be developed in the
Z-transform domain. In either case, Z-transforms usu-
ally require extensive and tedious algebraic computa-
tion. While this may be facilitated by a good library
of well documented computer programs [5], the exten-
sive computation that is required suggests that a di-
rect computation or simulation approach might be
preferable to transforms for sampled data processes.
Because we are dealing with linear systems, the inte-
gration is inherently simple.
To compare the two methods, let us first consider
analysis of a closed loop control system using the Z-
transform approach. We begin with an observable and
controllable dynamic process that may be described
by a set of linear, ordinary, differential equations.
While the theory and methods to be discussed are by
no means limited to single-input-single-output pro-
cesses (SISO), we shall restrict the discussion to this
class of systems for clarity of presentation. Actually,
multivariable control systems rarely get much atten-
tion in undergraduate instruction.
dx/dt = Ax + Qm(t td) + Rd (1)
y = Cx
where x is the vector of state variables,
y is the observed output variable,
m is the control variable,
d is the disturbance,
td is the delay in the control input,
A, Q, R, C are matrices.
We assume that the input control variable m is

Extensive research and
development effort has been devoted
in the past two to three decades toward
a different approach-time domain
analysis of dynamic systems.

restricted to the commonly used zero-order hold func-
tions, i.e.
m(t) = constant
kT < t s (k + 1)T (2)

T is the sample period.
It can be shown that the stable discrete step re-
sponse model of the observed variable y is given in
the Z-transform space by [8]
N -
y(z) = z-T a aiz'(1 z )m(z) + Bd(z) (3)

where = /T

B = C[zI exp(AT)]-1R
ai = h

h. = C exp[(n-1)AT]r
S x )d
r = exp(At)dt Q = -

[I exp(AT)]A-1Q

In the classical approach for analysis in the Z-
transform domain, the process model is considered to
be available in terms of a single nth order differential
equation. The Laplace transform of this equation
gives the transfer function of the continuous plant

y(s) = G(s)m(s) + Gd(s)d(s)

The open loop Z-transform of the transfer function
is formed by multiplication of the Laplace transfer
function by the zero-order hold function and then by
use of the Z-transform tables. The Z-transform of the
open loop process response is the Z-transform of the
inputs multiplied by Z-transform of the open loop
transfer function.

y(z) = HG(z)m(z) + Bd(z)

The Z-transform, Bd(z), must be found for the product
Bd(s), and not by multiplying the respective Z-trans-
forms, B(z) by d(z) [7].
The time domain response is found by inversion of
the Z-transform from Eq. (5). This might be ac-
complished analytically using partial fractions and the


Z-transform tables. Alternately, a numerical solution
for the response at the sample intervals can be calcu-
lated by long division of the denominator of y(z) into
its numerator. The coefficients of the polynomial in z-k
that is obtained by the division would be the discrete
response sequence.
In the time domain, these same results may be
derived following a more direct and intuitive ap-
proach, as used for example, by Cutler and Ramaker
[9]. The coefficients, ai in Eq. (3) are the unit step
response model for the process. As illustrated in
Figure 1, the output or response at any sample inter-
val due to a step input of size m, at t = 0 is given by

y(nT) = yo + man (6)

Because of superposition in linear processes, the
response after a series ofn step inputs mi is given by
y(nT) = yo + mian-i (7)

The response after a series of n step inputs in both
the control variables mi and disturbances di is calcu-
lated explicitly.
y(nT) = y + V (miani + dibni) (8)

where br is the step response to a unit step input in
the disturbance.
This equation is equivalent to Eq. (3) except that
the dead time delay in Eq. (8) is carried in the ai vector
as depicted schematically in Figure 1. The dead time
need not be an integer multiple of the sample time.



0 0.6-
cc 0.4-





FIGURE 1. Unit step response of a second order process
with a dead time
G(S) = e-/[(S+1)(2S+1)

A restriction implied by Eq. (8) is that the distur-
bance occurs as a step input. In fact, many but not all
real disturbance variables do actually occur as steps.
For those real physical disturbances that do not occur
as steps, the resulting inaccuracy in the predicted out-
put is usually far smaller than error that can occur for
other reasons. The overwhelming simplicity of the for-
malism is adequate compensation for this minor prob-
It may be evident just from the above descriptions
that considerably less effort is required to derive the
time domain representation of the open loop response
than with Z-transforms. The actual computation of the
response using Eq. (8) is straightforward and may be
accomplished on a pocket calculator.
Of course the closed loop response is of greater
interest; from that we can determine if the controller
will perform adequately. The classical Z-transform
computation of closed loop response for a typical DDC
control system is given by

HGfG(z)Gc(z) Z{Gd(s)d(s)}
y(z) + HGfGG(z)G (z) + 1 + HGfGGm(z)Gc(z)

This equation implies that the Z-transforms be
found for HGfG, HGfGmG, Gc, and Gd(s)d(s). The out-
put which is the inverse of y(z) would then be found
by algebraic manipulations similar to those described
above for the open loop. The algebra required for
y(nT) via Eq. (9) is far from trivial even for a modest
The direct computation approach solves the closed
loop equations recursively rather than as a set of
simultaneous equations. Eq. (8) is unchanged for com-
putation of the response except that an equation is
added to compute the control input. For example, if
the controller is a PID, then the control or manipula-
tive variable may be computed as follows [7]

mn= Kc(En + (T/TI) I i + (TD/T)(n n-l)] + ms(10)

where n = ysetn Yn
ms = steady state value for input
Eq. (10) may be easily programmed and used along
with Eq. (8) to compute the response. Again pro-
grammable calculators or very modest computer
programs can develop the time response even for
rather complex processes. The equations are explicit
and the algorithm is simple:
(1) Knowing the initial conditions and inputs up to time
nT, y(nT) is computed with Eq. (8);
(2) The errors ,, are computed, including the current one;


-- Set Point
---- Control Variable
--- Response

3 6 9 12 15 18 21 24 27
Time/Sample Time

FIGURE 2a. Closed loop response of sampled data pro-
cess. Third order system with dead time of Eq. 20. Con-
trol is PI with K. = 0.60, 7T = 1.0

Set Point
-- Control Variable
--- Response



/ I \

/ r: r

i--, r- ,, .-

3 6 9 12 15 18 21 24 21
Time/Sample Time

but with K, =

FIGURE 2b. Same process as in Figure 2a

-- Set Point
---- Control Variable
--- Response

,.- /" -.

\ ,- \ I"

a 0.0- * S- .
-1.0-. *- '

3 6 9 12 15 18 21 24 27
Time/Sample Time

FIGURE 2c. Same process as in Figure 2a but with Kc =

(3) The value for the manipulated variable for the next in-
terval is computed from the control law Eq. (10); this
calculation is exactly the same as would be made by the
actual process control computer.
(4) The input is applied to the process response function,
iterating this procedure through step (1).

These ideas can be illustrated by considering the
examples and problems from the recent under-
graduate text by Stephanopoulos [7]. All of the prob-
lems which involve finding the response, evaluation of
stability, or design of controllers for sampled data sys-
tems can be worked using the simulation approach, In
almost every case there is a considerable saving of
time and labor over the Z-transform. A typical exam-
ple would be Problem VII.28 in which the student is
asked to compute the ultimate gain for a third order
process with dead time

G = 2 exp(-tds)/(0.1s + 1)(s + 1)(3s + 1)

For the Z-transform solution, the open-loop Z-trans-
form of the transfer function must first be found. The
above Laplace transform is first multiplied by the
zero-order hold, and then factored into partial frac-
tions. The partial fractions are used with Z-transform
tables and additional algebra to obtain the pulse trans-
fer function
HG(z) = 2 z-(l+t)(l z-1)

S1 1/2610 5/9 135/29
1 Z-1 1 e-10z-1 1 e-1z 1 e-1/3z-1

This result is a rational function in z where the de-
nominator is third order and the numerator is of order
four plus td.
To find the ultimate gain, we must find the roots
of the characteristic polynomial
1 + D(z) HGp(z) = 0 (13)

where D(z) is a proportional controller. Thus the
characteristic polynomial is also of order four plus td.
The proportional gain in D(z) is varied repetitively
until one of the polynomial roots is on the unit circle.
While the roots may be found with a computer pro-
gram, extensive algebra is required before that point
is reached. The reader should visualize the amount of
"dog work" involved in clearing fractions and expand-
ing the characteristic polynomial using Eqs. 12 and
13, and then finding the location of the roots as the
proportional gain changes. Even with the partial
simplification of ignoring the-shortest time constant,
Continued on page 108.


Q P classroom


Ohio University
Athens, OH 45701

ANON-INTERACTING THIRD order system plays an
important role in process control as a typical sys-
tem employed in the classroom and laboratory to illus-
trate some of the control concepts. It is the simplest
system that can be used to show instability in the
closed loop control with any of the three controllers
(proportional, proportional-integral, proportional-in-
tegral-derivative). Lower order systems do not
exhibit instability with proportional control. It is also
used to demonstrate some of the control methods,
such as root-locus, frequency-response, and loop tun-
ing. In all these methods, the parameters to be deter-
mined are the maximum overall gain, Kmax, and the
critical frequency, we. When the overall gain is
greater than Kmax, the control system is unstable.
Therefore Kax is probably the most important
parameter in control design. For this reason, a lot of
time and effort can be saved if a simple formula can
be devised for Kmax.
Harriot [1] developed a formula to calculate Kmax
and we, but it was not widely used. Routh's array [2]
was most widely used, but it was not simple. My stu-
dents and I have employed a very simple formula
which is based on the fundamental information of the
system, namely the time constants. If the three time
constants are T1, T2 and T3, the maximum gain can be
calculated directly by

Kx = + (wT2 + T2 + (W 2 1
max cc 2 c 3

Wen-Jia R. Chen obtained his BSChE from National Taiwan Univer-
sity, and his MS and PhD from Syracuse University. He worked at
Allied Chemical Corporation and the University of New Mexico before
joining Ohio University. His current research interests are in coal con-
version and process control. He is also planning to integrate interactive
graphics with ASPEN.

= /T1 + T+ T3)/(T1T 2T3)

The above equations can be derived from the Bode
stability criterion.

tan-(WcT1) + tanl1(wT2) + tan-l(wT3) = 180 (3)
1 1 1
K m 1 (4)
max +(WcT1) A1+(wjc2 +(wcT3)2

Take the tangent of Eq. (3) and expand

w T + w T + w T w T w T2w T = 0 (5)
c 1 c 2 c 3 clc 2 c 3



My students and I
have employed a very simple
formula which is based on the fundamental
information of the system, namely
the time constants.

0 Copyright ChE Division ASEE 1986

wc = /(T1 + T2+ T3)/(TT2T3) (2)

And from Eq. (4)

K = + (wT1)2 + (WT2)2 + (WT3)2 (1

Some interesting information can be deduced from
Eq. (1). First, the maximum gain depends only on the


relative magnitude of the time constants. If the dom-
inant time constant is TI, and the other two time con-
stants, T2 and T3, are expressed in terms of TI

T2 = X T,,

T3 = Y T1

then Eq. (1) becomes

K = +Y +X+Y I+X+Y (7)
max XY Y X

which depends only on the ratios of the time con-
stants. Second, the maximum gain is the same for two
systems with the time constants as the reciprocals of
each other. In other words, if 1, 1/X, 1/Y are substi-
tuted into Eq (1), Eq. (7) remains the same. Third,
the minimum Kmax happens when all three time con-
stants are identical. The corresponding value is 8. Fi-
nally, Kmax goes to infinity when any one of the ratios
goes to infinity which corresponds to a lower order
system, and hence is always stable.


1. Harriot, P., Process Control, McGraw-Hill, New York, 1964
2. Routh, F. J., Dynamics of a System of Rigid Bodies, 3rd ed.,
Macmillan & Co., Ltd., London, 1877 O

REVIEW: Liquids
Continued from page 77.
modynamic properties have been made, including new
methods for studying and understanding fluid be-
havior near critical and tricritical points. On the
theoretical side, the 70's and 80's have seen the rapid
development of computer simulation methods and per-
turbation theory, which have placed the interpreta-
tion of liquid properties on a much sounder footing. In
particular, the use of simulation data to test theories
has eliminated the need to guess the intermolecular
forces-a serious source of error in almost all of the
tests of theories prior to 1970.
The aim and general organization of the book are
the same as for the previous edition. The first two
chapters give an introduction to liquid properties and
their classical thermodynamics, and are similar to
those of the last edition. The next chapter covers fluid
behavior near the critical point, and has been substan-
tially updated to cover recent developments. The next
three chapters cover our experimental knowledge of
thermodynamic properties and phase equilibria for liq-
uid mixtures at both low and high pressures. The
coverage is similar to that in the last edition, but the
material has been reorganized and updated. Welcome
additions include the introduction of phase diagrams
plotted in terms of field variables (temperature T,

pressure p, and chemical potential R in place of T, p,
and x, where x is the mole fraction), a section on tri-
critical points, and the use of the classification scheme
of van Konynenburg and Scott in discussing the be-
havior of fluid mixtures at high pressure. The last two
chapters of the book cover the statistical ther-
modynamics of liquids and liquid mixtures. Much has
been added here, particularly on the recent develop-
ments on fluids of nonspherical molecules, perturba-
tion theory, and measurements of molecular correla-
tion functions. The chapter on intermolecular forces,
present in the last edition, has been removed, al-
though there is a brief discussion on this subject in
Chapter 7. The authors conclude that the statistical
thermodynamics of mixtures of rigid (i.e. non-flexible)
neutral molecules has made major advances and is
now on a wound footing. The main barrier to progress
remains our relative ignorance of intermolecular
forces for all but the simplest molecules.
This latest edition of the book will be indispensable
to those working on liquid properties. O

by R. F. Hout, Jr., W. J. Pietro, and W. J. Hehre
Published by Wiley-Interscience,
Somerset NJ 08873 (1984),
403 pages, $39.95
Reviewed by
Gar B. Hoflund
University of Florida
It is well known that molecular orbital theory and
relatively simple symmetry arguments can provide
important information about the electronic and
geometrical structure of molecules and their chemical
reactivity. In this approach, extended molecular orbi-
tals are constructed from hydrogen-like atomic arbi-
tal. While the visualization of atomic orbitals is quite
easy, the visualization of molecular orbitals can be
quite difficult even for relatively simple molecules.
This book provides a pictorial representation of high-
lying (valence) molecular orbitals for a large number
of polyatomic inorganic, organic, and organometallic
molecules. The molecular orbitals have been gener-
ated using the GAUSSIAN-83 computer program,
and in this book photographs taken of a high resolution
graphical display of the molecular orbitals are pre-
sented. In addition to the filled levels, the lowest-
unfilled molecular orbital (LUMO) is also shown for
each molecule. This book provides an important
source of information for those trying to understand
molecular interactions on a fundamental level. O


O laboratoryy



A Sequel

University of Notre Dame Occ
Notre Dame, IN 46556

the literature [1-3] on demonstrating or measur- Buret
ing non-Newtonian effects in the flow of polymer so- (AREA, A)
lutions. The importance of these experiments can
hardly be overemphasized in view of the fact that in
several programs, this may be the only exposure of
an undergraduate student to non-Newtonian flow as 50 cc
well as to polymeric systems. It is, therefore, neces-
sary to maximize the learning experience associated Tygon Capillary Tube
with such an experiment. This article describes our Tube -
efforts to design such an experimental program with L--x
the set-up of Walawender and Chen [3], which we Ring rr--__--
found to be most amenable for this purpose. Stand

We simplified the set-up of Walawender and Chen
and did not use their elaborate temperature-control
system. Our set-up thus consisted only of a 50 cc;
buret connected to a glass capillary through a Tygon
tube (see Fig. 1). By doing so, we could have enough
parallel set-ups available so that students could take
more data on systems during their laboratory session.
Several capillaries were made available, having radii
of 0.505 and 0.287 mm. (obtained by weighing known
lengths of mercury threads in them), and lengths of
20.0, 30.0, and 39.6 cm. The experimental technique
was identical to that described by these workers and

The importance of these experiments
can hardly be overemphasized in view of
the fact that in several programs, this may be
the only exposure of an undergraduate student
to non-Newtonian flow as well as
to polymeric systems.

*On leave from the Indian Institute of Technology, Kanpur, India

FIGURE 1. The experimental set-up.

consisted of recording the height, h(t), of the meniscus
in the buret as a function of time for flow of several
solutions through the capillary. This was analyzed to
give the apparent viscosity, Tia, as a function of the
shear rate, j.

The volume rate of flow, Q, of a Newtonian liquid
in a horizontal capillary tube under steady, fully de-
veloped and laminar conditions, is described by the
following equation [4,5]

Q R4 (-P) (1)

-P' = = constant = -
SCopyright CE Division ASEE 1986
0 Copynght ChE Divisio" ASEE 1986




where -q, AP, L and R are the viscosity, pressure drop
across the capillary, length and radius of the capillary,
respectively. The pressure drop across the capillary
tube in the set-up shown in Fig. 1 is also given by
AP = pgh(t) (3)

where p is the liquid density, and g is the acceleration
due to gravity. The change of h with respect to time
can be expressed as

dh(t)=_ Q_ R4[pgh(t)] (4)
dt A 8 LA n

where A is the cross sectional area of the buret. Inte-
gration of Eq. (4) gives

Rn [h(t)] = R4 C- t + CE mt + C (5)
8LAn n

m d n [h(t)] = B
dt n

log [h(t)] vs t plots for such liquids are thus expected
to be linear having negative slopes (which may be used
to estimate 91 for Newtonian liquids).
In the case of a non-Newtonian liquid, the shear
stress, 7, is not linearly related to the shear rate, ;,
and the 'apparent' viscosity is a function of the shear
rate. Weissenberg, Rabinowitsch and Mooney have
presented an extremely ingenious method of analyzing
experimental data on flow through capillaries under
these conditions. Their final equations are given by
the following two equations [6,7]

S=RAP (8)
w 2L


B -AR4g

1 Y + T w dee
:T e 4 drw
n^y) w w

Here, Tw and %, are the shear stress and the shear
rate at the capillary wall at any time t, and 4e is de-
fined by

4e = 8QL
Ie rR3'r~ W R4 (AP)


Mahari Tjahjadi will receive his bachelor's degree from the Univer-
sity of Notre Dame in 1986. He is currently a senior (1985-1986) and
an undergraduate research assistant in the chemical engineering de-
partment. As an international student from Indonesia, he enjoys the
pleasant years of study at Notre Dame, where the academic environ-
ment is rewarding. (L)
Santosh K. Gupta received his BTech degree (1968) from the Indian
Institute of Technology, Kanpur, India, and his PhD (1972) from the
University of Pennsylvania. After a year of post-doctoral work he re-
turned to teach at the Indian Institute of Technology, Kanpur. He spent
about 24 months during 1985-1986 as a visiting professor at the Uni-
versity of Notre Dame. His research interests include polymerization
reaction engineering, physical properties of polymer systems and fluid
mechanical operations. He has authored/co-authored four books, the
latest of which 'Reaction Engineering of Step-Growth Polymerization'
(Gupta and Kumar-Plenum, N.Y.) is scheduled to appear in 1986. (R).

Eq. (3) may still be used to estimate AP as a function
of time, and the flow rate Q can be easily determined
experimentally at different times, using

Q = A dh( (11)

Thus, both w7 and )e can be obtained as functions of
time from a single experiment on flow through the
capillary. Plots of 4e vs Tw can be made and both 9a
and jw can be obtained at any particular time. Thus,
the function %la(w) can be determined. This is identi-
cal to '9a(i), since the apparent viscosity is a material
Because of errors introduced in computing the
slopes in Eq. (9), Walawender and Chen [3] suggested
curve-fitting experimental data using
h(t) = ho exp{-kt + (a + bt)2} (12)

where ho is the height of the meniscus at time t= 0 and
k, a and b are constants. The corresponding form of
the Weissenberg, Rabinowitsch and Mooney Eq. (9)
is then


1 ;w m 1 (m)
B-- p + d- t i (13)
ra (w) 4w 2
m dn h = k + 2b (a + bt) (14a)
dm= 2b2 (14b)
and B is given by Eq. (6).
The method of analysis of experimental data, h(t),
on any one system, is to obtain k, a and b by a non-
linear curve-fit parameter estimation computer pack-
age [8-10], use Eqs. (8) and (3) to give w7 at any time
t, use Eqs. (13) and (14) to give 'a and then get the
corresponding ,w as 7w/,la. The values of T7, y and
'9a change with the time, t, and thus, a(ww) or %la*()
plot can be obtained from a single experimental run.
Our numerical method is a little sensitive to the initial
estimates of k, a and b fed into the computer [3] and
students gain some valuable experience with con-
vergence problems encountered in numerical proce-


Flow curve measurements were made on car-
boxymethylcellulose (sodium salt, widely known as
CMC) solutions in distilled water. This system is non-
toxic and easy to work with. Not only do these poly-
mers exhibit typical pseudoplastic behavior, but,
being a polyelectrolyte, they can also be exploited for
educational purposes to show the relationship be-
tween molecular interactions and viscometric proper-
ties [11,12]. Three different molecular weight samples
of this polymer, C8758 (low viscosity), C4888 (medium

S i I I

0.07 Weight percent Mov
o-7.5 x o05
a4- 2.5x 05
So 8.9 x 104
-5----.-0-o- -0.
010 --
I 0 =-. jO.-o o -.0 nco "ifl
0.07 ** & *
0 7--'-- s a

200 300 400 500 700 1000
y, sec-I

FIGURE 2. Flow curves for three different molecular
weight samples of CMC in distilled water (22.80C). Dia.
of capillary = 1.010 mm, Length = 20 cm.

viscosity) and C5013 (high viscosity), having average
molecular weights, Ma, of about 8.9 x 104, 2.5 x 105
and 7.5 x 105 respectively (obtained from Sigma Chem-
ical Co., St. Louis), were used to make solutions of
concentrations ranging from 0.7% to 1.5% (by weight)
for the 'low viscosity', from 0.07%-0.13% for the
'medium viscosity' and from 0.03-0.07% for the 'high
viscosity.' This choice of concentrations was made so
that inordinately high times of flow through the capil-
laries are not required. The solutions of the 'high vis-
cosity' polymer had to be stirred for about 45 minutes
to attain homogeneity. A good way of checking the
homogeneity of the solution was to view it against a
bright light and not find any regions having different
refractive indices. The experience of making polymer
solutions of even such low concentrations as 0.05% is
itself quite educational.
Fig. 2 shows a sample set of flow curves for three

15- a 39.6
0 30.0
10 4 Ax 20.0
SI0o ^^aoo.o

7- ** oo..
Sp I I I I N I 39 I
6- 0-300~o

60 80 100

200 300 400 600 800 1000 1500
S, sec-i

FIGURE 3. Flow curves for the 7.5 x 105 molecular weight
CMC sample in distilled water (22.8C) taken with capil-
laries of different lengths. Concentration = 0.07%.

concentrations for each of the three CMC samples
mentioned above. For the 'high viscosity' sample, the
viscosity is observed to fall by a factor of about two
over a shear rate range of 200-1500 sec-'. The 'low
viscosity' samples, in contrast, show almost Newto-
nian behavior, at least for the concentrations studied.
A comparison of curves a and b in Fig. 2 leads to the
significant conclusion that shear-thinning is not al-
ways present in polymeric systems, and that two
polymeric solutions having the same viscosity at any
one shear rate, need not have equal viscosities at a
different shear rate. Fig. 2 also illustrates the signifi-
cant effect of molecular weight on the viscosities at
any shear rate. For example, at a shear rate of about
500 sec-1, the viscosity of a 1% solution of a low
molecular weight sample of CMC is the same as that
of a 0.03% of a high molecular weight sample. In fact,
cross plots of 9 vs concentration (c) or average molecu-
lar weight, May, at any /, may be made to illustrate
these effects. We observed that at / = 500 sec, plots



50 70

100 200 300 500 700 1000
y, sec-

FIGURE 4. Flow curves for the 7.5 x 105 molecular weight
CMC sample in distilled water (22.80C) taken with capil-
laries of different lengths and diameters. Concentration
= 0.0726%.

of Y vs (cMa) had slopes about unity (though they did
not superpose, both because of free volume effects
present at such high dilutions [6,7] and because of the
configurational changes associated with different de-
grees of ionization of the polyelectrolyte in pure
water). Students could be advised to read some write-
ups on simple molecular theories for the non-Newto-
nian behavior of polymeric systems [6,7] in order to
rationalize spme of their data.
Fig. 3 shows data taken on a 0.07% CMC (May =
7.5 x 105) solution in distilled water using capillaries
of different lengths. It illustrates the superposition of
data, as well as the possibilities offered for extending
the range of shear rates which could be studied by use
of different capillaries. Fig. 4 shows how similar ex-

I 0
b Sucrose H2O -
-- a K
7 0.7


S- 0.3 -
o 0

mm cm 0.2
mm cm CMC (O.IM NoC) 2
H20-- 1.01 396
CMC 1.01 20.0
Sucrose- 1.01 20.0
I I A I I 01
200 300 500 700 1000 2000 3000 5000

FIGURE 5. Flow curves for water (use right scale for vis-
cosity), for CMC (M., = 7.5 x 105, 0.07%) in distilled
water and in 0.1M NaCI solution, and for a 50wt% su-
crose solution in water. Temperature = 22.80C.

mm cm
5 0- 1.010 20
a- 1.010 40
0- 0.574 10

3 , I i I I I I i I


tension of the range of 's results when capillaries of
different diameters are used. There is some experi-
mental error in the data reported (about 10%), and
there are some additional errors introduced (longest
capillary, ~ 400 sec-1) because of viscous effect in
the buret and the Tygon tubing. Lower diameter
capillaries of longer dimensions were not tried since
they led to very high times of flow and so were not
suited for use in the laboratory sessions. However,
general trends can easily be inferred from these two
figures. It may be mentioned that the low-shear New-
tonian region was not reached in our experiments for
the high molecular weight samples, and so we failed



(a) (b)

FIGURE 6. (a) A typical molecule of CMC salt in water
with 0.1M NaCI (b) in absence of added NaCI.

in our efforts to illustrate how polymer solutions be-
have as Newtonian fluids till some y, and then show
a drop in viscosity till the second Newtonian region.
Use of more sophisticated equipment (e.g., Weissen-
berg rheogoniometer) could give a more extended
range of y, but their availability is limited only to some
chemical engineering departments.
Fig. 5 shows data on distilled water and shows its
value to match those reported in the literature.
Higher length capillaries were necessary to take data
on water; otherwise, considerable experimental error
would have resulted. This figure also shows the signif-
icant lowering of the viscosity of CMC solutions by a
factor of 3 to 4 by the addition of NaCl to the solution.
This is an excellent illustration of the molecular as-
pects of polymer rheology. Because of the presence of
significant amounts of NaC1, the sodium ions on the
polymer molecules do not move out too far into the
solution, resulting in very little ionic charge on the
macromolecules in such solutions. This is quite differ-
ent from CMC solutions in water alone, in which the
sodium ions of the polymer wander out into the solu-
tion, and the polymer has negative charges on its
backbone (see Fig. 6). Because of the electrostatic re-

Continued on page 105.

S classroom




Cornell University
Ithaca, NY 14853

ing of thermodynamic phase equilbria achieved
through the use of computer graphics has recently
been shown by Naik et al [1] for the two simplest
classes of phase behavior, as classified by Scott and
Van Konynenburg [2,3]. In this paper we extend the
work reported in [1] to include the considerably more
non-ideal phase behavior shown by classes III, IV and
V. Thus, at Cornell, we are now able to allow all five
classes of phase equilibria to be displayed on Evans
and Sutherland multipicture system II vector refresh
workstations, with the full range of interactive ma-
nipulation of the display described in [1]. Currently
the programs are being used by 120 students in five
courses, with great success.


The three-dimensional phase diagrams,
involving pressure, temperature and composition
axes, are depicted as wire-frame objects enclosed
within a normalized cube. The liquid regions are
shown as continuous lines . the vapor regions
are depicted as dashed lines.

[4] modification of the Redlich-Kwong equation of
state. The non-linear equations involved in this ap-
proach were solved using a Marquardt [5] algorithm,
available in Argonne National Laboratories' MINPAK
software package. The critical lines originating from
these mixtures were calculated using a method pro-
posed by Heidemann and Khalil [6]. It should be noted
that the parameters used for the generation of the
different classes (III, IV and V) do not correspond to
real systems; they are simply sufficient to generate
the phase diagram typifying a particular class of be-
havior. These parameters are given in Table 1.


For these more complex systems, the representa- The three-dimensional phase diagrams, involving
tion of the phase diagram (the pressures, tempera- pressure, temperature and composition axes, are de-
tures and compositions) was obtained from the Soave picted as wire-frame objects enclosed within a nor-
malized cube. The liquid regions are shown as continu-
ous lines while the vapor regions are depicted by
TABLE 1. dashed lines. Solid lines are also used to present the
Parameters Used for the Generation of the Phase Dia- pure component vapor lines and the critical lines. The
w is P grams 0 a c critical points of the pure components are highlighted
w is Pitzer's acentric factor; 9 and t are combining
rule parameters for the potential energy and size [8]. as bright dots.
Class III Class III systems exhibit the most
T,, K Pc, atm a 0 5 complicated phase equilibria; there are several sub-di-
CLASS III visions of this class which include liquid-liquid and gas-
Component 1 190.6 45.4 0.008 0.013 0.01 .
Component 2 373.2 88.2 0.100 gas immiscibility. The most interesting sub-class is
CLASS IV the one in which the critical line originating at the
Component 1 282.4 49.7 0.085 -0.017 0.00 critical point of the less volatile component exhibits
Component 2 588.0 62.3 0.346 both a maximum and a minimum in pressure. Systems
CLASS V which are classified as being Class III also show a
Component 1 588.0 62.3 0.346 -0.042 0.00
Component 2 272.4 49.7 0.085 Copyright ChE Division ASEE 1986



second, short, critical line which originates at the crit-
ical point of the more volatile component.
Figure 1 gives a clear presentation of the sigmoidal
behavior of the critical line which originates at the
critical point of the less volatile component. Figure 2
shows the same phase diagram from a different angle
superimposed with a low-temperature isothermal cut.
In this figure the temperature of the cut is within the
three-phase region and thus a three-phase tie-line is
present. Isobaric and constant composition cuts can
also be superimposed on the three-dimensional phase
Any of the aforementioned cuts can be seen as a
two-dimensional plot instead of being superimposed
on the diagram. A series of such plots is shown in

Figure 3. The ability of the software to allow a con-
tinuous set of cuts to be taken can be used to good
effect to demonstrate the transition between different
phenomena. Figures 3 (a)-(d) show the transition of
the isobaric cuts in the vicinity of the minimum of the
critical lines which originates at the critical point of
the less volatile component. Figure 3(a) shows a Tx
cut at high pressure; Figure 3(b) shows another cut at
a pressure very close to the minimum; Figure 3(c)
represents a cut just below the minimum and Figure
3(d) completes the series with a cut at a still lower
Class IV Systems in this class are characterized
by three critical lines and two regions of liquid-liquid
immiscibility. The first critical line is short and ex-

Georgios N. Charos, of Larnuca,
Cyprus, graduated from the Uni-
versity of New Hampshire with a
BS in chemical engineering. He
was awarded his MS in chemical
engineering from Cornell Univer-
sity in 1984 and is currently study-
ing for a PhD in the area of process
design. (L)
Paulette Clancy is currently an
assistant professor in chemical en-
gineering and associate director of Af J-
the Manufacturing Engineering
Program at Cornell University. She
received her BS degree at the Uni-
versity of London and a DPhil degree at the University of Oxford. She
held fellowships at Cornell University and at London University before
joining the faculty at Cornell in 1984. (C)
Keith E. Gubbins is currently the Thomas R. Briggs Professor of
engineering and director of chemical engineering at Cornell University.
He received his BS and PhD degrees at the University of London and

was on the staff at the University of Florida from 1962-76, when he
moved to Cornell. He had held visiting appointments at Imperial Col-
lege, London, at Oxford University, and at the University of California
at Berkeley. He has co-authored two books, Applied Statistical
Mechanics(Reed and Gubbins) and Theory of Molecular Liquids(Gray
and Gubbins). (R)


tends from the critical point of the more volatile com-
ponent to an Upper Critical End Point. The second
critical line is longer; it originates at the critical point
of the less volatile component, goes through a
maximum in pressure and terminates at a Lower Crit-
ical End Point. The Lower Critical End Point and
Upper Critical End Point enclose one liquid-liquid im-
miscibility region. The third critical line appears at
the low-temperature, low-pressure region of the sys-
tem. The relationship between Classes IV and III can
be seen by comparing Figure 4 and Figure 1; in Class
IV the long critical line intersects the three-phase re-
gion, whereas in Class III it does not. Figure 5 shows
an expanded view of the two liquid-liquid immiscibility
regions. One can easily see the critical line intersect-
ing the liquid surface and emerging in a lower pres-

SYSEM clas g 25C

sure region. Some three-phase tie-lines are also
Class V Class V systems are similar to the ones
in Class IV, but they lack the low-temperature, low-
pressure critical line and liquid-liquid immiscibility.
Actually, Class V systems would emulate Class IV
behavior if the solid phase did not extend to suffi-
ciently high temperatures and pressure to intersect
with the fluid phase emergence of the critical line [7].
Figures 6 and 7 show the phase diagram for this class
from two different views. Figure 7 also shows a con-
stant composition cut superimposed on the three-di-
mensional phase diagram.
The extension of our previous work [1] to the more




complex Classes III, IV and V permits us to provide
complete coverage of all the most commonly encoun-
tered types of phase behavior, with all the benefits
that an interactive software package on a sophisti-
cated computer graphics system can provide. Student
and instructor response to their inclusion in both un-
dergraduate and graduate level courses has been
overwhelmingly positive. We feel that this approach
provides the key to a significant improvement in the
teaching of this difficult subject.


It is a pleasure to thank Missy Mink and the staff
of Cornell's Computer Aided Design Instructional
Facility for their assistance. This work was supported

by grants from the National Science Foundation
(Grant No. CPE-8209187) and the Gas Research Insti-

1. C. D. Naik, P. Clancy and K. E. Gubbins, Chem. Eng. Educ.,
20, 2 (1985).
2. R. L. Scott and P. H. Van Konynenburg, Disc. Faraday Soc.,
49, 87 (1970).
3. P. H. Van Konynenburg and R. L. Scott, Phil. Trans., A298,
495 (1980).
4. G. Soave, Chem. Eng. Sci., 27, 1197 (1972).
5. D. W. Marquardt, J. Soc. Ind. and Appl. Math., 11,431 (1963).
6. R. A. Heidemann and A. M. Khalil, AIChE J., 26, 769 (1980).
7. C. H. Twu and K. E. Gubbins, Chem. Eng. Sci., 33,863 (1978).
8. U. Deiters and G. M. Schneider, Ber. Buns. Phys. Chem., 80,
1316 (1976). [




rj' views and opinions


An Integrated Approach to Curriculum

West Virginia University
Morgantown, WV 26506-6025

ACULTY IN MANY institutions are trying to study
and analyze their respective curricula in order to
bridge the gap between theory and practice. One
method of gathering data for analysis is the survey
method. Recently, several articles [1-9] have ap-
peared dealing with curriculum changes that academic
institutions need to make in order to bridge the gap
between what is being taught in the classroom and
what is actually needed in the workplace. One such
article, "A CE Survey-Putting College Back on
Course," is particularly striking. Essentially, the arti-
cle, using the survey method, solicited responses from
working chemical engineers and chemical engineering
professors in order to find out if curriculum changes
are needed in the undergraduate discipline so that
theory and practice can be meaningfully brought to-
gether [3]. There is a serious gap in their research
design; industry's environment and role in regard to
bridging the gap was totally ignored. No information
was solicited regarding the engineer's job environ-
ment and his perceptions of the company's role in job
placement, training, and career development. Accord-
ingly, this paper suggests an integrated approach for
the study of curriculum changes in academic institu-
tions in order to bridge the gap between theory and
The study of chemical engineering curriculum used
the traditional or common approach to curriculum re-
design. The authors summarized the results of 4,759
responses to an education survey in the April 18, 1983,
issue of Chemical Engineering. The analysis and re-
sults focused on the role of the respondents as past
students who are now practicing chemical engineers.
The authors addressed two time periods in their pre-
sentation of the data: "in college" and "after college."
In the chemical engineering survey, several con-
clusions were offered as a result of the analysis of

Copyright ChE Division ASEE 1986

. industry can and should shoulder
some of the responsibility . Specifically,
industry should communicate to academic institutions
what a chemical engineer is expected to do in
process engineering, in research and
development, in sales, etc.

responses. Two of these we feel required in-depth
analysis of the job environment in addition to more
academic analysis. The first conclusion stated that
"many students graduate from a four-year course in
chemical engineering with a hazy or inadequate idea
of what chemical engineers do and what careers are
open to them." The authors of the article explicit put
the total blame on academic institutions and their re-
spective curricula. There is no doubt that academic
institutions and curriculum changes can help define
what chemical engineers do and what careers are open
to them. However, industry can and should shoulder
some of the responsibility in this regard. Specifically,
industry should communicate to academic institutions
what a chemical engineer is expected to do in process
engineering, in research and development, in sales,
etc. This information can be communicated via indus-
trial seminars, video tapes, or in the recruitment of
students. In addition, it is industry's responsibility to
match the interest, talent, and academic background
of the recruited engineer with a job that requires such
talent and background. If a mismatch occurs, the
working engineer is not going to be effective on the
job and most likely will be unhappy with the job.
Given this situation, it is quite possible that the en-
gineer, when asked about the curriculum he had dur-
ing his college education, might view it in a negative
way. This kind of response is then translated into cur-
riculum recommendations and changes. We feel that
questions about the engineer's job setting, whether
the engineer was happy with his job and whether he
was properly matched to it, should be asked. Only
then can we meaningfully analyze curriculum changes
to bridge the gap between theory and practice.
The second conclusion stated, "These engineers
are less well prepared for their long-term careers in




that they lack essential management skills (ten years
after graduation three-quarters of our sample have
become managers)." [3, p. 48] The authors then rec-
ommend that "more and better business and manage-
ment courses should be available to students, perhaps
as electives-even at the expense of some of the more
advanced scientific and engineering courses, such as
reactor design."
This conclusion fails to recognize 1) the organiza-
tion's role in the education process, and 2) the time
lag associated with the engineer's move to manage-
ment positions. Just as a college curriculum cannot be
all things to all people, course work designed for entry
into specialized fields of endeavor cannot be expected
to cover all aspects of a career that might possibly
span thirty years. The education process is dynamic
in nature. The analysis used a static model which does
not provide for changes over time. Respondents indi-
cated that after being on the job they recognized the
need for management/business courses. Surely, tak-
ing some business courses might help initially, but
after being on the job for a period of time, one would
expect the company to absorb some of the responsibil-
ity for specialized on-going training to prepare the en-
gineer, who has been on the job for a long time, for a
managerial role. Matching the type of courses or train-
ing to the time period when it is needed recognizes
the inherent time-lag associated with the need for
management skills for professional engineers. This
can be done in a myriad of ways-by offering non-
credit courses in the area of business and management
or perhaps by encouraging and paying the engineer's
way to a nearby institution that offers an MBA pro-

Ali H. Mansour is an associate
professor of management at West
Virginia University. He obtained
his bachelor degree in mechanical
engineering from Youngstown
State University in 1962 and
worked as an aerospace research
engineer with NASA through 1971.
In 1974 he completed his MBA
from YSU in management science,
and in 1978 he earned a PhD from
the University of Georgia. In 1985
he was awarded a Fulbright Fel-
lowship to design MBA courses and
to lecture at the School of Economics
at the University of Jordan. He has several publications in aerospace,
management, and information systems. (L)
Michael S. Lane is an assistant professor of management at West
Virginia University. He obtained his bachelor's and MBA degrees from
the University of Nebraska and his DBA from Memphis State Univer-
sity. He teaches MBA courses both on- and off-campus to a number of
engineering graduates. His publications have focused on strategic
planning, ethics, and mentorship. (C)

gram. In fact, there are many companies which pre-
pare their engineers for a managerial role in the com-
pany in this way [2].
By ignoring the organization's perspective, it is
evident that alternative recommendations and/or solu-
tions are arbitrarily eliminated from consideration. In
addition the life-long learning approach and its impact
on both the individual and the organization with re-
gard to their respective responsibilities is being ig-
nored. It should be evident from the discussion that
an integrated approach should be used when studying
curriculum changes.
The chemical engineering survey addresses seri-
ous concerns in the area of curriculum design/rede-
sign. The results warrant serious consideration. The
analysis, though, must be integrated; both curriculum
and industry recommendations should be addressed.
In addition to curriculum recommendations, the study
needs to make recommendations for industry. The two
are intimately related; data collection and analysis
must be done concurrently. For example, it is possible
that the respondents feeling they should have had cer-
tain other courses in college to help them on their
jobs, may be due to improper placement of their tal-
ents and skills rather than a deficiency in their college
education. In this case, the researcher will be able to
alert industry to do a better job in recruiting place-
ment, and perhaps in training.
This paper presented a critique of existing
methodology used in curriculum updates in academic
Continued on page 107.

John Harpell is an associate professor at West Virginia University.
He received his bachelor's degree in engineering at Northeastern Uni-
versity and his MBA and DBA at Georgia State University. He teaches
graduate and undergraduate students at West Virginia University and
in their off-campus programs. His primary teaching areas include op-
erations research, statistics, MIS, and business policy. He has written
several books on operations research, production and nonparametric
statistics. (R)





Texas Tech University
Lubbock, TX 79409

namic concepts is of invaluable assistance in an-
ticipating the effect on a process of changing condi-
tions of temperature and pressure. This is particularly
true in the phase behavior of multicomponent sys-
tems. Estimates of this behavior can usually be made
Knowing only the general features of such a diagram,
even if a quantitatively accurate plot is not available.
An arsenal of such figures for binary systems is usu-
ally imparted in a course on solution thermodynamics.
The generation and use of a ternary phase diagram
for a system exhibiting vapor-liquid-liquid equilbrium
is presented here.

Consider a system consisting of two hydrocarbons
and water at 50 kPa. Each hydrocarbon is immiscible
with water but miscible in all proportions with the
other hydrocarbon. At this low pressure, no serious
error will be introduced if the vapor phase is assumed

to behave ideally. We therefore set all vapor phase
fugacity coefficients to unity. At a sufficiently low
temperature (to be determined), there will be no
vapor phase. The system then consists of a hydrocar-
bon binary and pure water. We will neglect the small
mutual solubility of the organic and aqueous phases.
At higher temperatures, three phases are possible: a
vapor containing all three components, a liquid or-
ganic phase containing only a mixture of hydrocar-
bons, and liquid water. Let us further assume that
the liquid hydrocarbon binary behaves ideally; that is,
the fugacity of each hydrocarbon in the liquid phase
depends linearly on its composition. The principles of
mapping the phase boundaries are best illustrated
with the simple system described above. The exten-
sion to more realistic liquid phase behavior is dis-
cussed later.
We number the components as follows: (1)-HCI
(least volatile hydrocarbon), (2)-HC2, (3)-H20, and
define the following composition variables
X = mol fraction of HCI in the liquid organic
Y = y1/(y1 + y2)
y3 = mol fraction of H20 in the vapor phase
There is no need to consider a composition variable
for the water phase since it is pure. The liquid organic
phase is a binary and its composition is completely
specified by X. The vapor is a ternary phase and re-
quires two specifications. We could use two of the
three mol fractions, yi, but it will be more convenient
to locate points in the ternary diagram using the set
Y, y3. They are related by

Y = (1 y3)Y

Ray Desrosiers is an assistant professor in the chemical engineering
department at Texas Tech University. He obtained a BS in chemistry at
Rensselaer Polytechnic Institute in 1970 and his doctorate in 1975.
Following postdoctoral work at the Ames Lab (ERDA) and the University
of New Hampshire, he worked for three years at the Solar Energy
Research Institute in Colorado. His research interests are in phase
equilibrium and biomass gasification.

Y2 = (1 y3)(l Y)


The only property data required are expressions
for the temperature dependence of the component
vapor pressures, Pi". Defining
Ki(T,P) = PiS(T)/P

the only phase equilibrium relations which may apply

a Copyright ChE Division ASEE 1986


At higher temperatures, three
phases are possible: a vapor containing all
three components, a liquid organic phase containing
only a mixture of hydrocarbons, and liquid water.
S380 K
to the vapor liquid boundaries are the Raoultian ex-
(1 y3)Y = KiX (2) 360
(1 y3)(1 Y) = K2(1 X) (3) --
y = K3 (4)

The first two apply if a liquid organic phase is present
and the third if the vapor is saturated with respect to
liquid water.
The vapor pressures of ethylbenzene and benzene EBZ 32
at several temperatures were fit to the Riedel equa-
tion and the resulting expressions were used to illus-
trate the calculation of the phase boundaries. b 20
If compositions are indicated on a triangular plane
and temperature is plotted vertically, then the ter-
nary phase diagram will consist of surfaces within a
FIGURE 2. Bubble and dew point surfaces.

triangular prism. Points in the prism are located with
three coordinates (Y,y3,T) as illustrated in Figure 1.
The coordinate Y fixes the hydrocarbon ratio. A line
380 K of constant HC ratio extends to the H20 apex. The
coordinate y3 is measured along this line.

The Component Binary Diagrams

HC 1 360 Before exploring the interior of the prism, let us
consider the phase boundaries lying in the faces of the
prism. These will be bubble and dew point curves of
the component binaries. For the hydrocarbon face,
320 these are obtained by summing (2) and (3) with Y3 =
"20 0, and solving for X or Y.
X = (1 K2)/(K1 K2) (5)
Y = K1(1 K2)/(K1 K2) T2 < T < T1 (6)
0 \0 Ki(Ti) = 1 (7)
The HCI-HC2 binary bubble-point line is the set of
points (X,0,T) and is drawn as the dotted line ab in
FIGURE 1. A point in the prism is located by specifying Figure 2. The dew point line is given by (Y,0,T) and
three coordinates: the HC ratio (Y), the steam ol frac- is the solid line joining ab. The saturation tempera-
tion (3), and the temperature (T). tures of ethylbenzene and benzene at 50 kPa are 379.5


K and 332 K and are obtained by solving (7).
The HCi-H20 binary displays an invariant point
at Ei. At fixed pressure, a binary can exhibit three
phases at only one temperature. This eutectic temper-
ature TEi is determined by summing (2) or (3) with (4)
and setting X = Y = 1 (or zero for i = 2).

Ki(TE) + K3(TE) 1 = 0 = =1,2
1 "1

Solving (8) for temperature gives TE, = 345 K. The
same procedure applied to the HC2-H20 binary yields
TE2 = 324 K. The eutectic compositions are

y3Ei Ki(TEi)

The eutectic points are E1:(1, 0.675, 345) and E2:(0,
0.255, 324). Below these points the system consists of
two liquid phases only. Hence the lines de and fg are
edges of the bubble-point surface for the ternary.
The water-rich branches of the dew point curves
are obtained from (4) since only a liquid water phase
is present.

Eic: Y3 = K (T)

Ti < T < T

i = 1,2 (10)

The branches Eic are the loci of points (1,y3,T) and
(0,y3,T) for the indicated ranges of T. The left
branches Ela and E2b are similarly obtained from (2)
and (3) by setting both X and Y to unity or zero,

Y3 = 1 K(T)

TEi T< T.s
Ei 1

The Invariant Line
We have just seen how the eutectic points of the
HCi-HO0 binaries dominate their phase boundaries.
The interior of the prism is similarly dominated by an
invariant line EIE2. Again using the phase rule, a sys-
tem of three components appearing in three phases at
fixed pressure has but a single degree of freedom. We
will need two representations of this line
All three of the phase equilibrium relations apply.
Let the invariant line be parameterized by tempera-
ture TE. Then it is the locus of points (Y(TE), y3(TE),
TE). Substituting (4) in (2) and (3) and eliminating X,
we obtain

3E = K3(TE) (12)

K,(1 K, K3)
(1 K3)(K K2)

TE2 < TE

It will also be useful to have this line characterized

FIGURE 3. The 330 K isothermal plane cuts the phase
boundaries of the bubble and dew point surfaces. The
348 K plane intersects only the dew point surfaces.

by the hydrocarbon ratio; i.e., as the set of points (YE,
y3(YE), T(YE)). Again summing (2) and (3) and
eliminating X, we obtain
(1 K3)(Y/K1 + (1 Y)/K2) 1 = 0 (14)

0 < Y < 1

For given Y, (14) can be solved numerically for TE.
Then (12) is used to determine y3E.

The Dew Point Surfaces
Given a vapor of composition (z1,z2,z3) we wish to
determine the temperature at which a liquid phase
first forms and the composition of that phase. The HC
ratio is
Y = zl/(zZ + z2) (15)

3 > K3[T(Y)]

then a liquid water phase is formed and the dew point


TD is found using Eq. (4) in the form
K3(TD) 3 = 0

Tie lines then extend from (Y,z3,TD) on the surface
ElcE2 to (*,l,TD) at the water apex.
z3 < K3[TE(Y)]
a liquid HC phase is formed. An objective function for
TD is found by summing (2) and (3) using (1)

zl/KI(TD) + Z2/K2(TD) 1 = 0


Tie lines extend from (Y,z3,TD) on the surface ElabE2
of Figure 2 to (z,/KI,O,TD) on the hydrocarbon face.
These tie lines will be illustrated more clearly when
we discuss isotherms.

The Bubble-Point Surface
For this system, a liquid of overall composition
(z1,Z2,Z3) would in general be a two-phase mixture. If
the aqueous and organic phases were suitably dis-
persed, then at the bubble point, TB, a vapor phase
containing all three components will form. The HC
ratio is
X = z /(Z1 + Z2)

All three of the equilibrium relations (2)-(4) apply and,
when summed with elimination of Y, they yield

KlX + K2(1 X) + K3 1 = 0 (17)

This equation is easily solved for TB. Once this tem-


.5 / H20

4 / '

- .0 E2 T(K) =330

BZN A ""
BZP (kPa)=50

FIGURE 4. The 330 K isotherm. The dashed lines are tie

.1 A2 T(K)=343

P (kPa)=50
FIGURE 5. The 343 K isotherm is above the saturation
temperature of benzene.

perature is known

Y = K1X/(K1X + K2(1 X)

and y3 = K3. Tie lines extend from the HC face to the
invariant line and then to the water apex at TB.

(X,0,T ) (Y,y3,T ) : (*,1,TB)

These tie lines generate the surface fE1gdE2e. It
should be noted that all of the lines and surfaces in
Figure 2 have been generated using only phase rule,
Raoult's law, and pure component vapor pressures.

The specification of the curves generated by the
intersection of an isothermal plane cutting the phase
boundary surfaces is an instructive exercise to test
our understanding of the ternary phase diagram. The
330 and 348 K planes are illustrated in Figure 3.
The intersections of the 330 K plane with the phase
boundaries and several tie lines appear in the plane of
Figure 4. The dot-dashed line from E1 to E2 is the
projection of the invariant line onto this plane. The
point E is on the invariant line and is located by set-
ting TE = 330 K in (12) and (13). The line EB2 is a line
of constant y, = y3E since the vapor mol fraction of
water is fixed by temperature as long as a liquid water
phase is present. Tie lines in the triangular region
formed by E, B2 and the water apex extend from
(Y,y3,TE) to (*,1,TE) for 0 < Y < YE. In the quad-
rilateral region formed by E,A1, the benzene apex,


and A2, a liquid HC phase is in equilibrium with a
vapor phase. The line EA2 is the set of points
(Y,y3,TE) where
y3 =1 [Y/K1 + (1 Y)/K2-1 (18)
Tie lines extend from (X,O,TE) to (Y,YS,TE). The liq-
uid HC composition is

X = (1 3)Y/K1 = K2Y/[Kl(1 Y) + K2Y] (19)

While the 330 K plane cuts the prism between the
eutectic points, it was below all three saturation tem-
peratures. The 343 K plane, illustrated in Figure 5, is
above T"2, the saturation temperation of the light hy-
drocarbon, in this case benzene. Eqs. (18) and (19)
still apply with TE = 343 K but only over the range
A2 > y < yE where YA2 is obtained from (6) with
T = TE.


Following the progress of condensing the overhead
vapor from a steam distillation is a practical example

Z \360



.500 -

.167 .387

FIGURE 6. Cooling path DEB for a steam-rich vapor. TD
= 349.1, TE = 340.2, TB = 332.5 K at 50 kPa.




o2 '3
E H20

2 -

P (kPa) 50

FIGURE 7. Projection of the cooling path DEB for a vapor
of initial composition (.1,.1,.8). The dashed lines are tie
lines. Points D,B show the compositions of the vapor
phase at the dew and bubble points.

of the use of the ternary phase diagram. Given a vapor
of known composition at a fixed pressure, we wish to
calculate the compositions and amounts of the phases
in equilibrium as the system is cooled. The equations
already developed for the determination of the phase
compositions must be augmented by the appropriate
mass balances. Each point on the cooling curve repre-
sents a flash calculation. The composition of the enter-
ing vapor relative to the invariant line determines the
phenomena which occur on cooling. We must therefore
consider two cases: steam-rich vapors and HC-rich
vapors. As before, pressure will be fixed at 50 kPa for
this discussion.

Case I: Steam-rich vapor

A vapor of composition (zl,z2,z3) is defined as

Characteristic Temperatures of the Cooling Path
for a Steam-Rich Vapor


Objective Function

TD K3(TD 3 = 0

TE Y = 1zl(z1 + Z2)
LI K3(TE)[Y/K (TE) + (1 Y)/K2(TE)] 1 = 0

TB X = zl/(z + z2)
K (T ) X + K2(T )(1 X) + K3(T ) 1 = 0


steam-rich if

3 > y3E = K3[TE(Y)] (20)

where Y = zl/(z, + z2) and TE is found using (14). A
typical cooling path is represented by the heavy solid
line in Figure 6. The projection of this path in the
composition plane appears in Figure 7. (Note that this
is not an isotherm.) As the vapor is cooled, the path
encounters the vapor/water boundary surface at TD.
Since only water condenses, the cooling path must de-
scend on this surface along a line of constant Y until
the invariant line is reached at TE. At this point, HC
begins to condense. Since three phases are present,
the remainder of the cooling path coincides with the


FIGURE 8. Phase compositions and extents resulting
from cooling a steam-rich vapor of initial composition
(.1,.1,.8) at 50 kPa.

invariant line until TB is reached. Note in Figure 7
that the HC liquid in equilibrium with the final point
of the cooling path has the same HC ratio as the enter-
ing vapor. The objective functions required to deter-
mine the characteristic points of the path are sum-
marized in Table 1.
For one mol of entering vapor, let us designate the
phase extents by V, H, W. These are the fractions,
respectively, of vapor, liquid HC, and liquid water.
For temperatures above TE, a balance on water is
sufficient to determine the phase extents

23 = W + y3(1 W)

This equation is easily solved for W, then V = 1 W,

Compositions and Extents while Cooling a Steam-Rich

T < T < T
Phases V, H, W
Compo- X = (1 K2 K )/(K1 K2)
Y = K X/[K1X + K2(1 X)]

TE < T < TD
V, W
X =
Y = z/(z1 + z2)

Y3 = K3
Y = (1 y3) Y
Y2 = (1 3)(1 Y)

Extents z1(1 y3) y1(1 z3)
H x(1 y3) 1
V = (z1 HX)/yI
W = 1 H V


V = 1 W
V= (z /--
W = (z3-Y3)/(1-Y3)

and H = 0. Below TE, two component balances in
addition to the overall mol balance are required.
z = Vy + HX z3 = Vy3 + (1 H V) (21)

The relations in (21) can be solved for V, H. A sum-
mary of the calculations required to fix the phase com-
positions and extents is listed in Table 2 and typical
results for an entering vapor of composition (.1,.1,.8)
are plotted in Figure 8.

Case II. Hydrocarbon-rich vapor
If the inequality in (20) is reversed, the vapor is
considered HC-rich. The cooling path for a vapor of
Continued on page 104.



E H20


P (kPa) 50
FIGURE 9. Projection of the cooling path DEB for a HC-
rich vapor of initial composition (.4,.2,.4). TD = 353.9,
TE = 340.5, TB = 336.1 K.



In a previous editorial it was indicated that if a
department is to receive public support, its goal
should be to serve society and not merely to seek
higher ratings. It was pointed out that the department
whose faculty was aware of professional needs and
attempted to fulfill them should gain in prestige, and
that this increased prestige may be reflected in the
ratings. But it was also pointed out that the depart-
ment that zealously sought high ratings by trying to
emulate what it felt were the attitudes and policies of
the high-rated departments, may fail to attain its self-
serving goal.
Perhaps critics will say that it does not make any
difference what goals are sought-that no one will
notice the difference between a self-serving depart-
ment whose faculty seeks high ratings and a society-
serving department that tries to fulfill human needs.
For that reason the following table has been prepared
indicating how the two differing philosophies might
effect the decision-making process of a department.*

If the department goal is serv-
ice, it should:
1. Develop a program that is
balanced between theory and
practice, basic and more
applied research, teaching
and research, math model-
ling and experiment.
2. Prepare students for a suc-
cessful career both in indus-
try and in grad school.

3. Critically examine today's
situation to see what profes-
sional needs can be met in an
innovative way.

4. Examine its curriculum in
the light of society's current
and future needs, and make
changes of an innovative na-

If the department goal is to
seek high ratings, it might:
1. Follow a specialized path
that emphasizes whatever it
feels has led to high ratings
in the past.

2. Prepare students primarily
for grad school by teaching
at a high graduate level with
few applications.
3. Try to emulate the innova-
tive departments that have
high ratings by adopting
their programs even though
it may be too late to make
much impact.
4. Adopt a curriculum that
seems currently popular
with accrediting bodies and/
or with high-rated depart-

*The entries in the above columns are not intended to represent
the policies or practices of any actual department. The table is in-
tended to portray cases where conflict may exist between the two
goals. The editor feels that his own department has generally fol-
lowed a goal of service (column 1). He personally favors the goal of
service as both a departmental and an individual goal.

5. Recognize and support inde-
pendent innovative research
on the part of its faculty
members even though it is
not presently fundable.
6. Support faculty in profes-
sional activities such as pro-
fessional service as officers
of professional organizations,
service on national commit-
tees, etc.
7. Permit and reward diversity
of faculty abilities and in-
terests with some emphasiz-
ing research, others teaching
or service; some doing basic
research, others more ap-
plied research.
8. Give appointments, promo-
tion, or tenure and salary in-
creases on the basis of all
contributions of its faculty
and not just on research.

5. Encourage only research
that already is supportable
by outside agencies.

6. Give support only to faculty
doing research, discouraging
or not promoting member-
ship or service in ASEE.

7. Expect all faculty to conform
to the model that only re-
search counts; that teaching
and service are poor uses of
one's time. Discourage indi-
viduality or departure from
8. At decision times, insist on
high research paper produc-
tion for all faculty, ignoring
teaching or service ac-

As seen above, the ratings race not only effects the
policies of the department, but also has an important
impact on individual faculty members. If the depart-
ment is striving for a high rating it will tend to reward
activities that it believes will improve its ratings (such
as the publication of research papers) over activities
that are hard to measure on a universal scale (such as
teaching and service). As a result, a philosophy that
"research is all that counts" is followed. The idealistic
young professor, whose personal goal is also service,
will follow his conscience and do what he believes is
important and what is right. But his more pragmatic
colleagues may do what they think will get them ahead
(publish papers) and may survive while he may not.
Thus a department seeking high ratings can foster
conformity on the part of its faculty since all will be
selected, not as persons with diversified strengths,
but as prolific producers of papers. Therefore teaching
and service might be minimized as the competition to
publish more papers intensifies.
In a department whose goal is service, all activities
in which faculty serve society as chemical engineers
are acknowledged. Like the organs of the human
body-the heart, the lungs, the brain etc.-each has
its own function. Instead of destructive competition
with each other the organs cooperate and the whole
body benefits. So also if a department is to thrive,
each faculty member must appreciate the unique con-
tributions of his colleagues so that as a department,
they may serve society, not only through quality re-
search but also through inspirational teaching and im-
portant professional activities.
Ray Fahien


book reviews

Published by the National Academy Press,
Washington, D.C.,
1985. $13.95.
Reviewed by
George Burnet
Iowa State University
Faculty shortages, overcrowded classrooms, and
inadequate equipment are diminishing the effective-
ness of engineering education in the United States,
according to this report from the National Research
Council, a part of the National Academy of Sciences.
The two-year $900,000 study was conducted by the
NRC Committee on the Education and Utilization of
the Engineer made up of 26 members about equally
divided between education and practice. Funding
came from NSF, several other government agencies,
and a number of major corporations.
The study attempts to examine both sides of the
engineering equation: education and utilization. In
view of the present times, the tone of the report seems
overly objective and its conclusions on the optimistic
side. A brief discussion of the development of en-
gineering in the United States is followed by an
examination of the status of engineering today. In
seeking an understanding of engineering's infrastruc-
ture, the report focuses on "organizing principles" and
comprehensive flow diagrams.
Roughly one-fourth of the report's 123 pages are
devoted to a critical look at the strengths and weak-
nesses of current engineering education. It is cor-
rectly noted that "the most critical and concerned at-
tention directed at the engineering profession in re-
cent years has focused on engineering education." On
the other hand, the report fails to convey a sense of
urgency in addressing the crisis of quality that it
agrees exists in engineering education today.
The section of the report on education documents
today's problems in a convincing manner. Engineering
undergraduate enrollment has grown at least 80 per-
cent during the last decade while in the same period,
engineering faculty has grown only 10 percent. En-
gineering PhD degrees have decreased by 30 percent
and one-third of these have been to non-U.S. nationals
who make up over 40 percent of the graduate students
in engineering. At the same time the pressures on
engineering faculty have been aggravated by an in-
crease in sponsored research of 50 percent in constant
dollars. The pervasive problems of equipment obsoles-

cence and the aging of physical plants are well
The report describes a two-tiered system of en-
gineering education that has been exacerbated by the
current crisis. The first tier consists of institutions
that are the major recipients of government funds for
graduate education and research. They are seen as
enjoying a distinct advantage that influences both
graduate and undergraduate engineering education.
The second tier consists of those institutions that have
as their primary focus undergraduate education. Be-
cause both government and industry focus their fund-
ing on graduate study and research, these colleges
are forced to depend on other appreciably smaller
sources of funding. Approximately one-half of the B.S.
engineering degrees are estimated to be granted by
colleges in the second tier. The report concludes that
separation of the two-tier system will widen unless
both government and industry introduce innovative
programs accompanied by more than token support.
Turning to curriculum, the committee concludes
that undergraduate programs should provide consid-
erable breadth across the disciplines of engineering
and within each discipline. Actions called for include
greater emphasis on non-technical education (greater
exposure to the world of ideas), computer technology,
orientation to the realities of the work world, and per-
sonal career management.
The committee could not reach a consensus about
whether or not the four-year program leading to the
first professional degree is any longer sufficient. Con-
siderable support was found for a pre-engineering un-
dergraduate program followed by a professional
school program with the combination requiring more
time than four years to earn the first professional de-
The final section of the report offers a preview of
engineering and its environment in the year 2000. The
discussion is based on a careful analysis of the en-
gineering workforce as it exists today. For working
engineers, the report paints a rather rosy picture not-
ing that the estimated 1.6 million engineers are among
the "best paid of all non-self employed professionals,"
and that their unemployment rates are the lowest of
any profession (2 percent). Looking to the future, the
report states that "we are entering an era in which
engineering will play a more dominant role than ever."
The report offers a 17-page executive summary
but the reader settling for this will be poorly served.
The summary includes recommendations selected
from the report that are largely unrelated and of
widely differing degrees of importance. The main
thrust of the report is found in conclusions and recom-
Continued on page 103.





Universidad Del Norte
Antofagasta, Chile

For several years I have assigned to students in
my undergraduate classes the reading of an interna-
tional publication. This research and discussion work
is done as part of a formal course which also contains
homework, quizzes, labs, and exams.
This activity may be commonplace in American or
European departments, especially at the graduate
level. However, it is unusual in some developing coun-
tries because of restricted access to many journals
published nowadays, because language problems arise
(since most publications of interest are not written in
the students' native language), and because professors
sometimes do not understand the importance of publi-

Jos6 0. Valderrama teaches chemical engineering at the University
of Petroleum & Minerals in Saudi Arabia. He holds BS and MSc degrees
from the Universidad de Concepci6n in Chile and a PhD from the
University of Delaware. He has worked at the Universidad de Concep-
ci6n and Universidad del Norte in Chile and has recently moved to
Saudi Arabia. He is very concerned with educational and teaching
problems and is constantly looking for improvements in what he con-
siders his main goal as a teacher: that his students learn and become
outstanding professionals. His research profile is revealed in his many
international publications and congress presentations in the areas of
thermodynamics and chemical reactors on which he has developed
both theoretical and experimental works.

*Present address: University of Petroleum & Minerals, Dhahran
31261, Saudi Arabia

I am convinced that analysis of and conferences
about an international publication give students a
broader vision of the subject being studied, encourage
constructive criticism, promote discussion, and stimu-
late efforts toward better oral and written communi-
cation (in the native language of the students). At the
level I usually assign this kind of work, most students
have already had to write laboratory reports in
chemistry, physics, physical chemistry, and so on.
However, it seems that instructors give no emphasis
to the quality of the written material, but rather
stress the quality of the experiments. Grammar, spel-
ling, and style are in general mediocre, but no mea-
sures are taken to remedy these defects. Usually, stu-
dents and professors become aware of the problem
when the students submit their senior projects for ap-
proval . too late to be corrected.
The idea of assigning articles for reading, as a par-
tial assignment within a course, is often resisted by
some in the class, a situation that sometimes results
in a poor and deficient report. Therefore, I have had
to develop a convincing argument and use part of my
teaching time to talk about the importance of going
beyond what is given in my lectures and about the
importance of clearly transmitting ideas, either in oral
or written form. Together with the assignment, I
hand out a two-page note in which I present some
ideas on why I bother them about reading the articles,
and give them some general rules about the work.
The note reads, in part

To be able to read a publication critically, to understand and
to interpret equations, methods, theories, and general ideas pro-
posed in a publication, is an activity that an engineer cannot disre-
gard. The only way of keeping up-to-date nowadays, when scientific
publications have tremendously increased in number, is to read at
least some of the large volume of articles. To learn how to read a
publication is a task which takes time. Why should we wait longer?
We will have to do it eventually. We should leave this university
with not only a solid basis in chemical engineering, but also with
a broad background to solve any kind of problem which may face
us in our professional work. To be able to transmit to other people,
orally or in written form, exactly what we want them to hear and
understand is a task as difficult as, or more so, than the preceding
one. There exists more than one case of a student from the last

Copyright ChE Division ASEE 1986


The idea of assigning articles for reading,
as a partial assignment within a course, is often
resisted by some ..., a situation that sometimes results
in a poor and deficient report.

semester who cannot meet face-to-face with another person to dis-
cuss clearly his or her point of view on any problem. There also
exists more than one case of a student who is about to graduate but
who does not feel comfortable speaking in front of his own class-
mates because of the dread of making afool of himself. We have to
learn how to overcome that drawback . it is not too late.
The assignment consists of the same five main
parts for any course I lecture in, namely
1. To read and understand the paper thoroughly. This in-
volves knowing what theories, methods, equations, ap-
proximations, and applications are contained in the pub-
lication, and, of course consulting the literature when it
is necessary.
2. To present an original contribution related to the subject
of the article and to the topics studied in the course.
Modifications, applications, and extensions of the sub-
jects of the publication are appropriate for this part.
3. To present two numerical problems related to the topics
of the article. The problems must be presented with de-
tailed solutions. Originality and ingenuity are the main
factors considered in the evaluation of this part.
4. To submit a written report about the article. The report
must be original and should contain what the student
understood about the subject. It should also contain the
original contribution and the problems.
5. To prepare and give a talk in front of the class and the
professor in which the main ideas, results, and conclu-
sions on the subject are presented. There is a time limit
of fifteen minutes for each student.

Of these steps, the first one has always been the
most difficult for my students, although they recog-
nize that it is the most important part for developing
a good contribution (step 2), inventing original prob-
lems (step 3), writing a good report (step 4), and giv-
ing a clear talk on the subject (step 5).
With points (2) and (3) my idea is to promote orig-
inality and develop creativity. With a new contribu-
tion, students create new situations, prove some of
the proposed concepts, compare their ideas with
others found in the literature, and speculate about the
scope of the subjects. Inventing problems and then
solving them is an exceptional exercise in which stu-
dents test themselves on their comprehension of the
publication. To invent a good problem, the ideas about
the subject must be absolutely clear. Otherwise, the
problems may result in unclear, ambiguous, and con-
tradictory situations . in summary, a poor prob-
For the written report, I give my students several
rules concerning the form and style of the report, but

allow them maximum freedom on the subject itself in
order not to limit their creativity or working capacity.
Clarity, depth, and continuity in the presentation of
the subject, originality of the written material, pre-
sentation and organization of the report, and quality
of writing are the main factors considered in the evalu-
ation of the report.
The importance of clearly transmitting ideas is
self-evident. With a time limit of fifteen minutes for a
talk, the students are forced to present the essence of
the article, select the main equations, show the most
relevant figures and tables, and discuss the most im-
portant conclusions and scope of the paper. Emphasis
must be given by students to their own findings, their
original contributions, and their application problems.
Participation by the rest of the class is encouraged as
a form of promoting discussion, enriching the presen-
tation of the subject, and obtaining the maximum
benefit from the talk. Clarity and continuity of the
presentation, adjustment to the time limit, answers to
questions from the class, and participation in other
students' presentations are the main factors consi-
dered in the evaluation of this part.
The most important factor for making the idea
work is to get students interested in what they are
doing. To that end the professor has to adequately
reward good work and punish poor work. Giving writ-
ten reports and oral presentation an appreciable im-
portance, say between 20% and 30% of the final grade
of the course works well. The professor should also
use strong and weak works as examples of what
should and should not be done by the students. The
results I have obtained with this kind of work have
encouraged me to improve upon the system each
semester. O

REVIEW: Engineering Education
continued from page 101.

mendations stated concisely at the end of each section.
Overall, the Committee recommends a variety of spe-
cific changes largely oriented toward the educational
process. Representatives of industry, academe, gov-
ernment, and engineering professional societies are
urged to work together to develop the necessary in-
In summary, the report presents carefully
documented findings and useful recommendations in-
tended to guide and inform the reader. It should be
read by those concerned about engineering education
and practice in the United States including engineer-
ing educators and administrators, government policy-
makers, and industrial leaders. D


Continued from page 99.

initial composition (.4,.2,.4) is projected onto the com-
position plane in Figure 9. The path encounters the
HC-vapor boundary surface at TD, but unlike the pre-
vious case, the HC ratio changes as the path ap-
proaches the invariant line. The point at which the
cooling path meets this line is again designated as TE
but if

Y = zl/(zl + z2),

TE f T (Y)

Whereas the equilibrium relations and mass balances
could be solved consecutively in Case I, here they
must be solved simultaneously for that portion of the
path between TD and TE. The objective function for
TE is conveniently expressed as
W = 1 H V = 0 (22)

When the expressions for H,V in Table 2 are substi-
tuted in (22) and all quantities are expressed in terms
of Zi and Ki, the expression listed in Table 3 for TE is
obtained. The objective functions for TD and TB have
already been discussed.
TE < T < T

there is no water phase and the phase extents are
determined using component balances on the hydro-
z = y1V + X(1 V) = X[1 + V(K1 1)]

2 = y2V + (1 X)(1 V) = (1 X)[1 + V(K2 1)]

The vapor fraction, V, is found by eliminating X in (23)
Szi/[1 + V(Ki 1)] 1 = 0 (24)

Characteristic Temperatures of the Cooling Path
for a Hydrocarbon-Rich Vapor
Temperature Objective Function

TD /K1(TD) + z2/K2(TD) 1 = 0
TE zl(l K1)(K1 K2) + (1 K2 K3)[z3(1 K1) K3] = 0
TB Ki(TB) X + K2(T ()(I X) + K3(TB) 1 = 0
X = Zl/(zI + z2)

Once V is found using (24), H = 1 V and W = 0.
The liquid HC composition, X, can then be calculated
using (23) and

y, = KX, Y2 = K2(1 X),

3 = z3/V

Once the cooling path encounters the invariant line, it
is restricted to this line as discussed in Case I. Then
TB < T < T

the calculations of the phase compositions and extents
are as outlined in Table 2.

An understanding of the phase diagram for the ter-
nary system consisting of two hydrocarbons and water
is useful in anticipating the phenomena which occur
when a vapor of this type is cooled. These are rela-
tively complex even for this idealized system in which
the small mutual solubilities of the HC's with water
were ignored. It is also astonishing to recall that Fig-
ure 2 was generated using only Raoult's law and phase
If the mutual HCi-H20 solubilities are not ignored,
considerably more thermodynamic data are required.
This would be in the form of expressions for the excess
Gibbs energy

GE(T,x 1PXP,x x3P)

where p identifies a particular liquid phase. The neces-
sary calculations are described by Prausnitz [2]. The
pairs of points (c,e), (b,d), (f,c) and (a,g) of Figure 2
would coalesce and the edges of the bubble-point sur-
face would become smooth curves all terminating at
the pure component saturation temperatures. Exam-
ples of such figures are available [1].
A BASIC program to perform all the calculations
outlined in this article is available from the author.
The program can generate the two-dimensional
isotherms and cooling curves illustrated in Figures 3,
7, 8. Machine requirements are modest: 32K and a
graphics package.

1. Findlay, A., A. N. Campbell and N. 0. Smith, 1951. Phase
Rule. p. 309, ninth ed. Dover Publications Inc., New York, NY
2. Prausnitz, J. M., T. F. Anderson, E. A. Grens, C. A. Eckert,
R. Hsieh and J. P. O'Connell, Computer Calculations for Mul-
ticomponent Vapor-Liquid and Liquid-Liquid Equilibria.
Prentice-Hall, Inc. Englewood Cliffs, NJ. O


Continued from page 87.

pulsion between these charges on the flexible polymer
molecules, the latter expand and give rise to higher
viscosities as well as to non-Newtonian effects. The
students are encouraged to read up on more elaborate
descriptions [12] of the salt-effect in polyelectrolytes.
Fig. 5 also shows the flow curve for a 50 (wt.)%
sucrose solution in distilled water [13]. This solution
is Newtonian in the range of shear rates studied and
has approximately the same viscosity as the CMC so-
lution (concentration = 0.07%) at the rate of shear of
200 sec-'. The comparison of curves a and b in Fig. 5
brings out the point further and more strongly that
two solutions having the same viscosity can still differ
significantly in terms of their non-Newtonian charac-
teristics. Students can also interpret these curves in

t, min

- Low Vis MediumVis EZZZ High Vis

FIGURE 7. A typical plan for two 3-hour laboratory ses-
sions. Remaining time could be used to duplicate same
results. C,, C2 and C3 represent different concentrations
(see Figs. 2-5).

terms of pressure drops required for achieving a de-
sired flow rate in a pipe.
It has been reported that some CMC solutions are
also thixotropic [11], i.e, the viscosity depends on the
time elapsed after the shear is applied. This is because
the molecules of the polymer could become associated
into some type of a macrostructure where the crystal-
line regions of different molecules aggregate to give
higher effective molecular weights and so higher vis-
cosities. These macrostructures could be broken up
by various means, e.g., by the application of high
shear for some time, leading to thixotropy. We carried
out some experiments to illustrate this effect. How-
ever, we found that the solutions, after vigorous stir-

ring for one hour and then storing for a week (when
the macrostructures could be formed again), exhibited
the same vi vs ~ behavior as before storing, and so we
inferred that thixotropy was absent in our polymer
samples. The reason for lack of thixotropy is probably
because our samples had a good uniformity of sub-
stitution by sodium [11].

A set of experimental runs on non-Newtonian flow
of dilute polymer solutions using an extremely simple
set-up has been described, illustrating some important
differences between the flow behavior of polymeric
systems and common low molecular weight liquids.
The program described also enables students to get a
grasp of the difficulty in making polymer solutions, as
well as gives some experience with convergence prob-
lems associated with nonlinear parameter estimation
techniques. More results have been reported in this
paper than can be obtained in two 3-hour laboratory
sessions. The instructor can choose any set out of
these. A typical plan for two 3-hour sessions is shown
in Fig. 7, wherein the effects of length and radius are
omitted. In a single 3-hour session, one could select
experiments on three concentrations of the low viscos-
ity CMC, one concentration of the high viscosity
polymer, and experiments on sugar solutions and

1. Rodriguez, F., Chem. Eng. Educ., 5, 82 (1971).
2. Weinberger, C. B., Chem. Eng. Educ., 9, 80 (1975).
3. Walawender, W. P. and T. Y. Chen, Chem. Eng. Educ., 9, 10
4. Bird, R. B., W. E. Stewart and E. N. Lightfoot, Transport
Phenomena, -1st ed., Wiley, N.Y., 1960.
5. Gupta, V. and S. K. Gupta, Fluid Mechanics and its Applica-
tions, 1st ed., Wiley Eastern, New Delhi, India, 1984.
6. Ferry, J. D., Viscoelastic Properties of Polymers, 2nd ed.,
Wiley, N.Y. 1961.
7. Kumar, A. and S. K. Gupta, Fundamentals of Polymer Sci-
ence and Engineering, 1st ed., Tata McGraw-Hill, New Delhi,
India, 1978.
8. Eikenberry, R. S., "Analysis of the Angular Motion of Mis-
siles," Report SC-CR-70-6051, Sandia Laboratories, Al-
buquerque, New Mexico, 1970.
9. Nielsen, K. L., Methods in Numerical Analysis, 1st ed., Mac-
millan, N.Y., 1960.
10. Ralston, A. and H.S. Wilf, Mathematical Methods for Digital
Computers, Vol 2, 1st ed., Wiley, N.Y., 1967.
11. Klug, E. D. in Encyclopedia of Polymer Science and Technol-
ogy, Vol. 3, eds. H. F. Mark, N. G. Gaylord and N. M. Bikales,
Wiley, N.Y., 1965.
12. Armstrong, R. W. and U. P. Strauss, in Encyclopedia of
Polymer Science and Technology, Vol 10, eds. H. F. Mark,
N. G. Gaylord and N. M. Bikales, Wiley, N.Y., 1965.
13. The authors would like to thank Prof. M. J. McCready for
suggesting this to us. D


Continued from page 73.
outline, using the Coughanowr and Koppel text. A
few of the recitations are orientation and preparation
sessions, instructing the students in how to use the
ACS system. Although the units are self guiding and
can be used without this preliminary instruction, stu-
dents can accomplish more study of process control
after the preparation sessions. Other recitations use
the ACS system in the demonstration mode to illus-
trate and reinforce particular course concepts. In
either type of recitation, a video projector is used to
put the instructor's CRT on a large screen for class
For the "hands-on" use of the ACS system, stu-
dents divide into groups of three. Each group has re-
served for it a weekly three-hour time period in which
to perform the units. During this period, they have
exclusive use of an ACS console, made up of two inter-
communicating color graphics terminals, and exclu-
sive control over a simulated process. Within this time

TS TC 001 INF 600.20 SF 600.2) TS AC 001 INP 1.9955 SP i '?f99
6S0.O 0 I 0E ND PL O I

Ts" C 100 INMP 2.599 y .P .F Q TS' 0C 101 INP i4.j45 SFP i-
....... .........
S. ........... .. ............

',0 "'0",

FIGURE 5. Controller output response to set-change in
error. Students check validity of controller action by
measuring slope of ramp and height of step.

TS TC 001 IN' 599.61 SF '/'9 I TSi AC 001 INP .46327 SF 1

1S cs 100 INP) .95 5 -F -- TS CS 101 INP 5H.2450SP -'

'2 ]. . . .."""""


FIGURE 6. Response of furnace feedback control to step
change in load. Students tune controller to obtain desir-
able response and to note effects elsewhere in process.

frame, students are able to proceed at a nominal pace
of one unit per week with little, if any, supervision.
At Waterloo, the first undergraduate course in
process control is given at the 3B level (second half of
third year). The course text is Coughanowr and Kop-
pel. Originally, the course consisted of three hours
per week of lectures for a term consisting of thirteen
weeks. The first undergraduate course using ACS was
given starting in September, 1983. With ACS, the
course structure was altered to two hours per week
of lecture with three hours per week in the computing
lab using the coursework modules. The computing lab
setup and hardware is essentially identical to that at
Purdue. Over the course of the term, ACS unit mod-
ules I-XII were covered during the lab sessions.


Near the end of the courses at Purdue and Water-
loo, students were invited to comment on the educa-
tional value of ACS. At both locations, approximately
70% of the respondents rated the use of ACS as "effec-

I'Tt: 001 INP /.3 SP 1 S A 00 1 INF 0. SPF

..... .. .' ........, .. .


SI .. ..
PF^l HISrTR-------------

FIGURE 7. Response of control scheme using two single-
loop feedback controllers. Students verify relative gain
array prediction that no controller tuning can stabilize
this control scheme.

IS TC 001 IMP 600.46 SP g' 9 TS TC .001 INP 600.46 SP 5'9.9

.S CE 1S.. IN' .5.416 SF T9 CS 106.. IMF 99.8 3 SF--
,,: : '- ., ,

... .. ... ,,......... 'I"*........,,,.


I .

(iC~~~.~ . ... '

FIGURE 8. Closed-loop frequency response plot. Students
verify vulnerability of feedback control schemes to dis-
turbances near the process resonant frequency.


tive" or "very effective" in teaching process control.
The bulk of the remaining responses rated it "OK."
Virtually all students agreed that the sacrifice of lec-
ture time for the use of ACS was well worthwhile. We
are confident that their responses would differ from
those of the engineers who participated in the Chem-
ical Engineering magazine survey. At both Purdue
and Waterloo, other indicators of positive student
reaction were

Students worked at a faster pace than anticipated by the
instructor. Only ten units were originally planned for the
introductory course; the last four had to be added to keep
up with the students.
Students arrived as much as half an hour early, even for
sessions scheduled at 7:30 AM, explaining that they
wanted the additional time to experiment in greater
The original enrollment limit for the follow-up elective
course was set at 38 students. Due to demand, this had to
be increased, and even the larger figure was quickly over-
subscribed. This is particularly unusual for process con-
trol, which is regarded as one of the most difficult sub-
jects in the curriculum.
The ACS facility was visited by over a score of industrial
recruiters, on a "drop in" basis. They indicated their in-
terest was spurred by the comments of their student in-
terviewees, who almost invariably listed it as one of their
favorite educational experiences.

We are enthusiastic and excited over the use of
ACS in the undergraduate process control course. Our
future plans are focused on the development of a wider
base of process simulations for additional senior pro-
cess control courses, and making the
coursework modules available to any other univer-
sities desiring them.
Plans are already being implemented to make ACS
and the study guides available to the chemical en-
gineering department at Northwestern University
and to the pulp and paper technology department at
the University of Wisconsin-Stevens Point, both by
remote dial-up to the Purdue facility. Other ACS sites
now include Louisiana State University, Imperial Col-
lege (England) and Queensland University (Au-
The authors are grateful to the many people at
IBM and IBM Canada Ltd. whose steadfast and en-
thusiastic support has made this valuable tool avail-
able to the chemical engineering academic community.
Specific acknowledgment is given to Ross M. Aiken at
Purdue as well as to Jerry van de Hoef and Blair
Thompson at Waterloo for their dedicated system and
tutorial support. []

Continued from page 59

ticipate only by invitation. Some university adminis-
trators would be delighted to see accreditation disap-
pear. Who needs those interlopers putting more heat
on for scarce resources for their favorite discipline?
Who cares whether someone else likes the curriculum?
I suppose the answer is that our profession is collec-
tively interested in knowing what type of graduate is
being produced beyond a potluck process.
This brings me back to my initial comments-those
of us involved in the accreditation system represent
AICHE and the chemical engineering profession, a
profession involving educators and practitioners in in-
dustry and government. I conclude by confessing that
participating in the camaraderie of the Exxon suite
sure beats grinding through a mountain of accredita-
tion reports and sitting through literally days of meet-
ings. Anyone wish to trade places? O

Continued from page 93

institutions. It is suggested that an integrated ap-
proach is more realistic and meaningful to study and
to bridging the gap between academic curriculum and
industry's needs. Specifically, we recommend that
curriculum-related data and job-related data be
analyzed simultaneously. The authors feel that this
approach should give us better insight to the much
reported 'gap' between theory and practice.

1. Giauque, William C. and R. E. D. Woolsey, "A Totally New
Direction For Management Education: A Modest Proposal," In-
terface (August 1981): 30-34.
2. Landis, Fred, "Employers Responsible for Education, Too"
ASME NEWS (April 1984): 3.
3. Lipowicz, Mark A., "Putting College Back on Course," Chemi-
cal Engineering (September 1983): 48-60.
4. McClenahen, John S., "New Marching Orders for MBA's," In-
dustry Week (August 1983): 49-51.
5. Newell, R. B., P. L. Lee, I. S. Leung, "A Resource-Based
Approach to ChE Education," Chemical Engineering Educa-
tion, Winter, 1985, 36-50.
6. Pollock, John A., Jon R. Bartol, Bruce C. Sherony, George R.
Carnahan, "Executives' Perceptions of Future MBA Pro-
grams," Collegiate News and Views, Spring 1983, 22-25.
7. Prentice, Marjorie G., "An Empirical Search for a Relevant
Management Curriculum," Collegiate News and Views, Winter
1983-84, 25-29.
8. Richman, Louis S., "B-School Students' Favorite Profs," For-
tune (January 1982): 72-79,
9. Windsor, Duane and Francis D. Tuggle, "Redesigning the MBA
Curriculum," Interfaces (August 1982): 72-77, E


Continued from page 81.

and with a good polynomial rootfinder, this problem
will take hours to complete.
In the time domain, computations for this problem
are not only much simpler, but also direct and trans-
parent. The student can get a feeling of the system
dynamics as part of the solution. The third order re-
sponse can be found directly from the Laplace trans-
forms. The algebra is the same as leading to Eq. 12
for the Laplace transform which is easily inverted to

y(t + td) = (1 e-0t/2610 + e-t/1.8
135 e-1/3t/29)u(t) (14)

The step response coefficients, a,, for use in Eq. 8 are
calculated from this equation. Note that it doesn't
complicate the calculation if the dead time is not an
integer multiple of the sample interval. The closed
loop response is then calculated as described above
using Eqs. 8 and 10. The calculations can be carried
out by hand, but the students quickly learn to imple-
ment these equations in a simple program on a hand
calculator, or on a computer.
Instead of searching for a root on the unit circle,
the ultimate gain is found as one would find it experi-
mentally: the gain is varied until oscillations are sus-
tained, neither waxing nor waning. This is shown in
Figure 2. Figure 2a, the closed loop response is seen
to be underdamped but stable when the gain is 0.60.
In Figure 2b, unstable response results when the pro-
cess gain has been increased to 0.95. The control vari-
able is a more sensitive indicator of stability than is
the output variable itself.
The decay ratio is the measure of the relative
amplitude of successive waves. In (a) the decay ratio
is 1.47; in (b) it is 0.902. Linear interpolation of these
values to 1.00 gives the ultimate process gain of 0.84.
In Figure 2c, it is seen that with this gain, sustained
oscillations occur, neither growing nor damping.
For additional information, it is a simple matter at
this point to modify the program to include the I and
D sections of Eq. 10 to evaluate a Ziegler-Nichols con-
troller designed using the ultimate gain and the ulti-
mate period.
Menu driven interactive programs have been de-
veloped which allow students to enter process
parameters and control constraints for first or second
order linear systems with dead time. These automat-
ically produce plots using graphics on the DEC PRO-
350 microcomputer similar to Figure 2 [10]. These
may be displayed in high resolution and in multicolor

directly on the terminal screen, or they may be
printed, as shown here, on a dot-matrix printer.
We have also installed these programs on the main
frame computer at the University of Missouri. In that
case, the graphical output on the terminal screen or
on the line printer is of lower quality than that from
the microcomputer since the graphics are produced
using the modified line printer plot subroutine,
FPLOT. We can send listings of these programs upon

Discrete process control systems can be simulated
using the Dynamic Matrix approach with relatively
simple computer programs. Problems that are very
tedious using Z-transforms can be solved in a straight-
forward manner using iteration and direct computa-
tion. Simulation gives intuitively meaningful results
and is pedagogically far more useful for analysis of
sampled data systems than are the more elegant but
difficult transform methods. It is an additional benefit
for students in process control classes to see the vari-
ation in system responses as a function of change in
control parameters. Hopefully this approach will re-
ceive greater stress in future texts.

We gratefully acknowledge support from Shell Oil
Co. during this work.

1. Lipowicz, M. A. and R. V. Hughson, "Putting College Back
on Course," Chemical Engineering, 90, No. 19, Sept. 19,
2. Kuo, B. C., DIGITAL CONTROL SYSTEMS, Champaign,
Ill., SRL Pub. Co., 1977.
TROL, Intext Educational Publishers, Scranton, Pa., 1972.
4. Isermann, R., DIGITAL CONTROL SYSTEMS, Springer-
Verlag, New York, 1981.
5. Franklin, G. F. and D. J. Powell, DIGITAL CONTROL OF
DYNAMIC SYSTEMS, Addison-Wesley Pub., Reading,
Mass. 1980.
6. Deshpande, P. B. and R. H. Ash, ELEMENTS OF COM-
PUTER CONTROL, Research Park, N.C., ISA, 1981.
7. Stephanopoulos, G., CHEMICAL PROCESS CONTROL,
Prentice-Hall, Englewood Cliffs, N.J., 1984.
8. Garcia, C. E. and M. Morari, "Internal Model Control," Ind.
Eng. Chem. Proc. Des. Dev., 21, 308, 1982.
9. Cutler, C. R. and B. L. Ramaker, "Dynamic Matrix Control-
A Computer Control Algorithm." 86th AIChE meeting, Hous-
ton, Tex., 1979.
10. Pao, A. S., "An Interactive Simulator for Digital Process Con-
trol System," M. S. Thesis, University of Missouri-Columbia,
Dept. of Chemical Engineering, 1985.



Departmental Sponsors

The following 157 departments contributed to the support of CEE in 1986 with bulk subscriptions.

University of Akron
University of Alabama
University of Alberta
Arizona State University
University of Arizona
University of Arkansas
University of Aston in Birmingham
Auburn University
Brigham Young University
University of British Columbia
Brown University
Bucknell University
University of Calgary
California State University, Long Beach
California Institute of Technology
University of California (Berkeley)
University of California (Davis)
University of California (Los Angeles)
University of California (Santa Barbara)
University of California at San Diego
Carnegie-Mellon University
Case-Western Reserve University
University of Cincinnati
Clarkson University
Clemson University
Cleveland State University
University of Colorado
Colorado School of Mines
Colorado State University
Columbia University
University of Connecticut
Cornell University
Dartmouth College
University of Delaware
Drexel University
University of Florida
Florida State University
Florida Institute of Technology
Georgia Institute of Technology
University of Houston
Howard University
University of Idaho
University of Illinois (Chicago)
University of Illinois (Urbana)
Illinois Institute of Technology
Institute of Paper Chemistry
University of Iowa
Iowa State University
Johns Hopkins University
University of Kansas
Kansas State University
University of Kentucky

Lafayette College
Lakehead University
Lamar University
Laval University
Lehigh University
Loughborough University of Technology
Louisiana State University
Louisiana Tech. University
University of Louisville
University of Lowell
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, Duluth
University of Minnesota, Minneapolis
University of Missouri (Columbia)
University of Missouri (Rolla)
Monash University
Montana State University
University of Nebraska
University of New Hampshire
University of New Haven
University of New South Wales
New Jersey Institute of Tech.
University of New Mexico
New Mexico State University
City University of New York
Polytechnic Institute of New York
State University of N.Y. at Buffalo
North Carolina A&T State University
North Carolina State University
University of North Dakota
Northeastern University
Northwestern University
University of Notre Dame
Nova Scotia Technical 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

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

If your department is not a contributor, write to CHEMICAL ENGINEERING EDUCATION,
do Chemical Engineering Dept., University of Florida, Gainesville FL 32611, for information on bulk subscriptions.

Chemical Engineering -

A new range of
'hands-on' teaching
'* *." *
F^;^; .'^ *-..- iS ^^ '^ '^^


Yesterday, laboratory teaching rigs had to be large,
expensive and difficult to set up.
Today, smaller scale, easy-to-use, self-contained
cheaper apparatus is available with greater educational
Technovate offers a full range:
*Mass Transfer *Reaction Engineering *Unit Operations
Ask us for more details of our equipment

CEK Liquid Mixing
CEL Fixed & Fluidised Beds
CEM Liquid Phase Chemical
Stirred Tank Reactor
CEN Solids Handling Bench
CEP Dynamics of Stirred Tanks
CERa Gaseous Diffusion
CERb Liquid Diffusion
CES Wetted Wall Column

CET Tubular Reactor
CEV Multi-stage Mixer/Settler
UOP1 Climbing Film Evaporator
UOP2 Double Effect Evaporator
UOP4 Solid-Liquid Extraction
UOP5 Liquid-Liquid Extraction
UOP7 Gas Absorption Column
UOP8 Tray Drier
Write also for our full range in Fluid Mechanics
and Heat Transfer

910 SW 12TH AVENUE *

POMPANO BEACH FL 33060 USA (305) 946-4470



: ;zu;.- b'l
ix .T~r~; yl
:~ ~- '"



Full Text