Chemical engineering education ( Journal Site )

Material Information

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


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


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

Record Information

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

Full Text



Chemical Engineering Education
Department of Chemical Engineering
University of Florida
Gainesville, FL 32611
PHONE and FAX: 904-392-0861
Ray W. Fahien
T. J. Anderson
Mack Tyner
Carole Yocum
James O. Wilkes and Mark A. Burns
University 1,. Ih,.ri
William J. Koros
University of Texas, Austin


E. Dendy Sloan, Jr.
Colorado School of Mines

Gary Poehlein
Georgia Institute of Technology
Klaus Timmerhaus
University of Colorado

Anthony T. DiBenedetto
University of Connecticut
Thomas F. Edgar
University of Texas at Austin
Richard M. Felder
North Carolina State University
Bruce A. Finlayson
University of Washington
H. Scott Fogler
University -r Ih. i,...,
J. David Hellums
Rice University
Angelo J. Perna
New Jersey Institute of Technology
Stanley I Sandler
University of Delaware
Richard C. Seagrave
Iowa State University
M. Sami Selim
Colorado School of Mines
James E. Stice
University. of Texas at Austin
Phillip C. Wankat
Purdue University
Donald R. Woods
McMaster University

Chemical Engineering Education

Volume 29

Number 1

Winter 1995

2 The University of Tennessee, Fred E. Weber, John P. Prados
8 Stan Sandler, of the University of Delaware, by His Colleagues

12 Teach 'Em Particle Technology, Ralph D. Nelson, Jr., Reg Davies, Karl

18 The Phillips Petroleum Company Industrial Experience Program,
R. Bruce Eldridge

22 Correlation and Overcorrelation of Heterogeneous Reaction Rate Data,
Mordechai Shacham, Neima Brauner

26 Simulating the Air Products Cryogenic Hydrogen Reactive Cooling
Process, S. Jayakumar, R.G. Squires, G.V. Reklaitis, K.S. Grassi

46 Role and Impact of Computers in Engineering Education,
Richard S.H. Mah, David M. Himmelblau

56 Chemical Engineering with Maple, Ross Taylor, Katherine Atherley

62 Students Journals: Are They Beneficial in Lecture Courses?
Douglas E. Hirt

50 Computers in Undergraduate Chemical Engineering Education: A
Perspective on Training and Application,
James F. Davis, Gary E. Blau, G. V. Reklaitis

32 We Never Said It Would Be Easy, Richard M. Felder

34 Laboratory Projects: Should Students Do Them or Design Them?
Anton P. J. Middelberg

40 Putting Commercial Relevance Into the Unit Operations Laboratory,
W. A. Davies, T.A.G. Langrish

6 Letter to the Editor
6 Response to Letter to the Editor
7 Positions Available
16,17,39 Book Review
39 Callfor Papers

CHEMICAL ENGINEERING EDUCATION (ISSN 0009-2479) is published quarterly by the Chemical Engineering
Division, American Society for Engineering Education, and is edited at the University of Florida. Correspondence
regarding editorial matter, circulation, and changes of address should be sent to CEE, Chemical Engineering Department,
University of Florida, Gainesville, FL 32611-2022. Copyright 1995 by the Chemical Engineering Division, 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, ASEE, which body assumes no responsibility for them. Defective copies replaced
if notified within 120 days of publication. Write for information on subscription costs and for back copy costs and
availability. POSTMASTER: Send address changes to CEE, Chemical Engineering Department., University of Florida,
Gainesville, FL 32611.

Winter 1995


Chemical Engineering at...




The University of Tennessee Knoxville,

TN 37996-2200

he University of Tennes-
see is located in Knox-
ville near the headwaters
of the Tennessee River and
roughly in the center of the
Great Valley of East Tennes-
see. The main range of the Ap-
palachian Mountains lies forty
miles to the southeast, with the
Cumberland Plateau about the
same distance to the northwest.
Within an hour's drive are six
Tennessee Valley Authority
lakes and the Great Smoky
Mountains National Park. The
Knoxville metropolitan area has

Aerial view of the University of Tennessee campus.

a population over 600,000, but enjoys a pleasant, generally
uncrowded atmosphere, consistently ranking among the
nation's top ten metropolitan areas in surveys on quality of
life. East Tennessee has a four-season climate, ranging from
summer temperatures in the 90s to winter temperatures cold
enough for snow skiing in nearby mountain resorts.

Founded in 1794 as Blount College, the first non-sectarian
college west of the Appalachians, The University of Tennes-
see today is the state's largest university and Land-Grant
institution with about 17,000 undergraduates, 7,500 gradu-
ate and professional students, and a faculty of more than
1,600. Bachelor's degrees are offered in over 150 fields,

master's degrees in 85, and doc-
toral degrees in 52.
Although a program called
"chemical engineering" ap-
peared in the university catalog
as early as 1905, true chemical
engineering courses were not of-
fered until 1934, first in the De-
partment of Chemistry and two
years later in a separate Depart-
ment of Chemical Engineering.
The first chemical engineering
faculty member was Robert M.
Boarts, a PhD graduate of the
University of Michigan and stu-

dent of the legendary W.L. Bad-
ger. The Master's program was begun in 1935 and the PhD
program in 1949 as the first doctoral program in engineering
offered by any institution in Tennessee.
The undergraduate program in chemical engineering re-
ceived its initial accreditation from the Engineers' Council
for Professional Development (now known as ABET) in
1939, making it one of the first four chemical engineering
programs in the South to receive accreditation. (Programs at
Georgia Tech and Virginia Tech were accredited in 1938,
while those at Tennessee and Louisiana State were accred-
ited in 1939.) The program has been continuously accredited
since that time. The department has had only five heads in its
58-year history: Robert M. Boarts (1936-1960); Homer F.
Johnson (1960-1984); Joseph J. Perona (1984-1990); John
W. Prados (1990-1993); and Charles F. Moore (1993-present).
Chemical Engineering Education

Copyright ChE Division ofASEE 1995

The undergraduate program in chemical engineering received its initial accreditation from the
Engineers' Councilfor Professional Development (now known as ABET) in 1939, making it one of the first
four chemical engineering programs in the South to receive accreditation. (Programs at Georgia Tech
and Virginia Tech were accredited in 1938, while those at Tennessee and Louisiana State were
accredited in 1939.) The program has been continuously accredited since that time.

In accordance with the University's overall mission as a
Land-Grant institution, the department has the threefold mis-
sion to: (1) provide excellence in chemical engineering edu-
cation at both the undergraduate and graduate levels; (2)
provide a strong research program which fortifies the educa-
tional process and which advances engineering knowledge;
and (3) serve industries in Tennessee and the nation by
providing educational opportunities for practicing engineers,
research services, and technical assistance.

The chemical engineering faculty includes thirteen full-
time members and one regular, part-time member. Some
course offerings and graduate student research direction are
shared with the Materials Science and Engineering Depart-
ment, which was a part of chemical engineering prior to
1984. Adjunct faculty members from Martin Marietta En-
ergy Systems, the Eastman Chemical Company, and the
Texas Instruments (now Siemens) Company provide a rich
source of industrial expertise for teaching specialized gradu-
ate courses and directing graduate students' research. The
faculty, in turn, are supported by a well-qualified secretarial
and technical staff.
The full-time faculty include six "old timers" whose com-
bined service at the University of Tennessee totals 146 years:
Don Bogue, George Frazier, Joe Perona, John Prados, and
Carl Thomas. Duane Bruns, Robert Counce, and Charlie
Moore constitute a "mid-range" group of senior faculty,
while Marion Hansen, Tse-Wei Wang, Paul Bienkowski,
and Fred Weber are more recent arrivals. The newest addi-
tion is Peter Cummings, who joined the faculty in January of
1994 as a Distinguished Scientist and Professor with a joint
appointment in the UT Chemical Engineering Department
and the Chemical Technology Division of the Oak Ridge
National Laboratory (ORNL). There are about a dozen such
faculty currently at the University; their positions are sup-
ported by special funds provided by the State of Tennessee
and the US Department of Energy.
According to Peter Cummings, "These are exciting times
at the University of Tennessee. Construction is nearing
completion on a 200,000-ft2 science and engineering re-
search building that will house only research laboratories
and which will result in a dramatic increase in both the
quantity and quality of research space available to the sci-
ence and engineering departments. There is an increasingly
high level of cooperation between the University of Tennes-
Winter 1995

see and nearby Oak Ridge National Laboratory, the latter
being an extraordinary resource of theoretical and experi-
mental expertise with 4500 employees, half of whom are
PhD engineers and scientists. The Chemical Technology
Division, the division with which chemical engineering has
the most direct connection, is one of its largest, embracing a
wide range of basic and applied research in such areas as
chemical processing, fluid mechanics, separations processes,
molecular thermodynamics, and environmental engineering."
Some recent faculty activities that have attracted outside
recognition include:
Three Fellows of the AIChE: John Prados, Jack
Watson, Joseph Perona
A past president and Fellow of the ABET, along
with a recipient of ABET's L.E. Grinter Award:
John Prados
A recent NATO Postdoctoral Fellow: Marion
A recipient of a research grant for outstanding
young engineering faculty from E.I. du Pont de
Nemours and Company for work in pollution
prevention: Robert Counce
Recipient of a large, unrestricted grant by E.I. du
Pont de Nemours and Company to support
graduate education and research in chemical
process control: Charles Moore and Duane
> One UT/ORNL distinguished Scientist: Peter
> A former director of the AIChE: John Prados
A faculty member is on a one-year appointment
as Senior Education Associate with the NSF
Engineering Education and Centers Division:
John Prados
N Chair of the Nuclear Engineering Division of
AIChE: Joseph Perona
Perhaps more important than any of the preceding recog-
nition is that six of the chemical engineering faculty have
received multiple Outstanding Teacher awards.

At present, the Chemical Engineering Department enrolls
230 undergraduate students, 49 full-time graduate students,

At left is the architect's
rendering of the new research
building that will be ready for
occupancy in 1995,
below is the Dougherty
Engineering Building, home of
chemical engineering at

and 32 part-time graduate students who are em-
ployed full-time as engineers. After declining sig-
nificantly in the mid-1980s, undergraduate chemi-
cal engineering enrollments are again beginning to
increase, while graduate enrollments remain rela-
tively stable; both are consistent with national trends.
Approximately 32% of the undergraduate students
are women, 13% are minority, and 5% are interna-
tional. The corresponding percentages for the gradu-
ate students are 21% women, 2% minority, and
32% international. Over the past eight years the
department has granted 197 bachelor's degrees, 52
Master's degrees, and 27 doctorates.
Chemical engineering undergraduates regularly
fill leadership roles outside the department. Patty
Wiegand, a 1992 graduate, was an All-American
distance track star and captain of the varsity women's
track team. Last year, Jerry Johnson, a junior co-op
student, was president of the Tennessee Alpha Chapter of
Tau Beta Pi. (It is worth noting that the Tau Beta Pi National
Headquarters have been located at The University of Ten-
nessee since 1907, with offices currently in the Dougherty
Engineering Building along with chemical engineering.)

A generous grant from the Eastman Chemical Company
has been used to modernize the undergraduate chemical
engineering laboratory with new equipment and instrumen-
tation. Additional grants, primarily from the Siemens Com-
pany, are being used to modernize the chemical process
control laboratory to provide undergraduate and graduate
students with experience in a state-of-the-art industrial data-
acquisition and control system.
Graduate students' experimental research is conducted in
on-campus laboratories of the chemical engineering and ma-
terials science and engineering departments as well as at
research facilities operated by the University at its Pellissippi
Biotechnology Research Facility and by the Oak Ridge Na-
tional Laboratory. At present, six chemical engineering gradu-
ate students are conducting research at the Pellissippi facil-
ity, and seven full-time graduate students, as well as several

part-time students and six postdoctoral research scholars, all
in chemical engineering, are conducting experimental re-
search at the Oak Ridge National Laboratory.
On-campus opportunities for experimental research will
be enhanced significantly in 1995 with the completion of a
new, multidisciplinary Science and Engineering Research
Building under construction adjacent to the present chemical
engineering facilities in the Dougherty Engineering Build-
ing. Space and equipment in the new building will support
research in the areas of bioprocess engineering, environmen-
tal and waste minimization studies, polymers and compos-
ites, and process control.
Both undergraduate and graduate students have access to
mainframe computers operated by The University of Ten-
nessee Computing Center and, under special conditions, to
supercomputers accessed through the internet. The depart-
ment operates two microcomputer laboratories, one prima-
rily for undergraduates and one exclusively for graduate
student use, along with a SUN SPARC-10 workstation. These
laboratories presently contain a mixture of Apple Macintosh
and MS/DOS personal computers, connected through a local
area network to allow sharing of expensive peripherals and
infrequently used programs; each computer contains a hard

Chemical Engineering Education

disk loaded with frequently used software. Faculty comput-
ers and those of the departmental clerical and technical staff
are also connected through the network. A user-friendly
electronic mail system has been implemented to simplify
communication among faculty, students, and staff. The net-
work has an interbridge connection to the University's
ethernet, through which any personal computer can access
mainframe computers and the internet.
The combined assets of The University of Tennessee and
ORNL in massively parallel computers make it one of the
richest computational environments in the U.S. Between
them, the two institutions have two Intel Paragons, a Kendall
Square KSR1, an Intel IPSC/860, a Masspar, and a Thinking
Machines CM-5. The UTK/ORNL Joint Institute for Com-
putational Science provides instruction on the efficient use
of these machines.

In recent years the chemical engineering faculty have at-
tempted to focus their research in a few limited areas to
allow better mutual support and interaction with the
University's interdisciplinary research centers. Current re-
search is concentrated in the following four principal areas:
Bioprocess Engineering Four faculty members are cur-
rently working in this area, with much of the work con-
ducted in collaboration with the Department of Botany, the
Center for Environmental Biotechnology, and the Oak Ridge
National Laboratory. Active research areas include the de-
velopment of reactor systems for biodegradation of toxic
organic, modeling and analysis on in-situ remediation pro-
cesses, and the application of bioluminescent sensors to
measurement of reaction rates in bioreactors.
Polymers and Composites Four faculty members have
active research in this area, with much of the work con-
ducted in collaboration with the Department of Materials
Science and Engineering and the Center for Materials Pro-
cessing. Work is currently in progress in the rheology of
polymers, modeling of tubular film blowing, multivariable
control of a fiber spinning process, modeling of polymer
crystallization in pipe extrusion, and in-line monitoring of
polymeric processes.
Process Control Three faculty members are collaborat-
ing with the Measurement and Control Engineering Center
in this area. Research is currently being done in plant-wide
process control system design and analysis, the use of neural
networks for system process monitoring and control, sensor
development and data acquisition and control for polymer
processing operations, and the monitoring, modeling, and
control of bioprocesses such as biowastewater treatment
plants and fermentation processes.
Pollution Prevention and Separations Technology
Seven faculty members are currently active in this area, with

much of the work conducted in collaboration with the En-
ergy, Environment, and Resources Center, the Environmen-
tal Engineering Program, and the Oak Ridge National Labo-
ratory. Areas of research include the development and evalu-
ation of non-halogenated solvent cleaning technologies, the
development of tools to aid in performance of life cycle
assessments, supercritical extraction techniques, and air strip-
ping of volatile organic compounds from ground water.

All full-time chemical engineering graduate students con-
duct research under the direction of one or more faculty
members as a part of their MS or PhD degree requirements.
In addition, undergraduate students with strong academic
records may elect to conduct a senior research project with a

On-campus opportunities for experimental
research will be enhanced significantly in 1995
with the completion of a new, multidisciplinary Science
and Engineering Research Building under construction
adjacent to the present chemical engineering
facilities in the Dougherty Engineering Building.
Space and equipment in the new building will
support research in the areas of bioprocess engineering,
environmental and waste minimization studies,
polymers and composites, and process control.

faculty member for academic credit. Several such projects
are conducted each year, some as individual efforts and
some as small-group projects.
The University operates a long-standing Engineering Co-
operative (Co-op) Program in which students alternate se-
mesters of academic study with semesters of work related to
their career goals. The program is open to all students in the
College of Engineering who are making satisfactory progress
toward their degrees, and 25-30% of the chemical engineer-
ing undergraduates participate.
In addition to Co-op and undergraduate research experi-
ences, a limited number of seniors may be invited to partici-
pate in an industrial internship for academic credit. In this
internship, small groups of students under the direction of a
faculty member are introduced to an actual industrial prob-
lem and are given the opportunity to develop a proposed
solution. Most of the work is conducted on-campus, and at
the end of the semester the students make a formal presenta-
tion of their proposed solution to the engineers who posed
the original problem. In a number of instances, these solu-
tions have been implemented in the industrial setting or have
led to further studies that ultimately yielded a satisfactory
solution. Currently, these internships are conducted with
cooperation and support by the Eastman Chemical Company
in the area of chemical process control under the direction of
Charlie Moore, and by the du Pont Company in the area of

Winter 1995

industrial pollution prevention under the direction of Pete
The undergraduate curriculum allows some flexibility for
specialization through the inclusion of four technical elec-
tives. Three 3-hour technical elective courses may be se-
lected from a wide variety of advanced engineering, science,
mathematics, or business courses, and the fourth 3-hour
elective is taken in advanced chemistry or other advanced
science (e.g., microbiology, materials science). Possible ar-
eas of specialization include biotechnology, process control,
industrial pollution prevention, and polymer engineering; a
designated faculty member in each area can advise students
on proper elective selection. Other areas of emphasis can
also be developed in consultation with a faculty advisor.
The department participates actively in the Engineering
college program of videotaped, off-campus graduate instruc-
tion. Core and elective courses in the Master's program are
offered on a regular cycle to students at off-campus sites,
including Oak Ridge, The Kingsport University Center, and
other chemical plants in Tennessee. Additional videotaped
graduate courses will be offered as dictated by demand.
When student demand is sufficient, graduate courses are

also taught "live" at Oak Ridge and Kingsport. Students may
complete MS degrees at these locations (and the PhD at Oak
Ridge) without a period of full-time residence in Knoxville.

Over the years, The University of Tennessee Chemical
Engineering Department has enjoyed a happy combination
of assets: faculty members with a wholeheared commitment
to students in their roles as research directors, as graduate
and undergraduate teachers, and as academic and career
advisors; a highly trained technical support staff for both
instrumentation/computers and mechanical systems, who pro-
vide invaluable support for experimental research and in-
struction laboratory development; and last but not least, a
capable and caring secretarial staff whose motto is "how can
I help you?" and not "go away and don't bother me!" Our
chemical engineering graduates may soon grow hazy about
the finer points of transport phenomena or the Second Law
of Thermodynamics, but they will remember clearly how the
departmental secretaries, Sancy Hail and Betty Frazier, cre-
ated within a large and sometimes impersonal university a
warm atmosphere of genuine care and concern for their
welfare. O

B letter to the editor

Dear Sir:
This letter is motivated by the paper "Exothermic CSTRs-Just
How Stable are the Multiple Steady States?" by Shacham, Brauner,
and Cutlip, that appeared in the winter 1994 issue of CEE.
The authors argue that the upper steady state in an exothermic
CSTR can be unstable; this is wrong and, as a matter of fact, it
results from a misunderstanding of the mathematical stability con-
Their conclusion stems from a linearized analysis of the problem
which is not exact around that steady state; as they say in their
paper, "...when it is integrated for a long enough time, the basic
model will produce a limit cycle." As their basic model is, in
reality, the nonlinearized model, we can believe that this is the
correct answer to the problem; but, and this is the main mistake, a
limit cycle is a stable periodic solution, as any standard textbook
would teach them! Even if the oscillations were much larger than
the ones observed in their Figure 6, where the "terrible oscillations"
(which they incorrectly interpret as instabilities) are of the order of
0.10R! Finally, one can say that one thing is the stability of the
steady state, which can, in fact, be concluded from a plot with both
the heats generated and removed (their Figure 1) and another is the
nature of this stable steady state (limit cycle, in this case), which
cannot be established from that plot.
Let's then restate that when three steady states result from the
intersections of the lines of the heat generated and the heat re-
moved, the upper and lower are stable (irrespective of the nature of
these steady states), while the middle one is unstable.

Thank you very much for your attention.
Jose Miguel Loureiro, Associate Professor
Dept. de Engenharia Quimica *
Univ. do Porto
Rua dos Bragas *
4099 Porto Codex, Portugal

Response to Letter to the Editor

To the Editor:
The letter of Prof. Loureiro contains several misconceptions
which are very important to correct. These misconceptions are:
1. The results of a stability analysis for a linearized system are
not valid for a nonlinear system.
2. A stable periodic solution is equivalent to a stable steady
3. The stability of a system can be deduced from a small initial
response of the system to a disturbance.
These misconceptions will be discussed separately.
N Relationship between stability of the nonlinear and the
linearized system.
This relationship was established by the Liapunov theorem (see
for example, reference 1). This theorem deals with the stability of a
nonlinear system in the vicinity of a particular critical (steady state)
point, and it states that "If the linearized solution is unstable, then
the actual operation (as described by the nonlinear equations) will
be unstable...." The meaning of Liapunov's theorem is that instabil-
ity indicated by linear stability analysis is a sufficient condition for
Chemical Engineering Education

instability of the nonlinear system. Liapunov's theorem is the basis
for the linear control theory which has been used successfully for
decades in control system design, and it certainly cannot be dis-
missed in an offhand manner.
- Relationship between stable periodic solution and stable steady
In order to clarify the behavior of the system in the vicinity of the
upper steady state, we have prepared a figure which extends the
simulation time scale for the basic model of the original paper.
Figure 1 shows the three variables associated with the CSTR (C, -
outlet concentration; T temperature inside the CSTR; T, tem-
perature in the cooling jacket) after the model equations have been
integrated for up to 150 hrs (using a powerful numerical integration
package: DDASSL). Initially the variables begin to oscillate with
monotonically increasing amplitudes, and after about 100 hours the
major oscillations continue with constant amplitudes. These con-
stant amplitudes oscillations provide what is called "stable limit
cycle" in a phase plane diagram. This is indeed a stable periodic
solution of the dynamic equations, as Prof. Loureiro refers to it, but
it has nothing to do with a stable steady state.
Steady state is defined as a state where none of the variables
change with time. A stable steady state is indicated where the
system returns to
the original steady
0.1 state after some
time when a small
0.08 perturbation is ap-
-plied to the sys-
tem. Clearly a re-
1actor oscillating
.E between 634'R
004 and 6740R (as see
in Figure 1) can-
0020 1 not be considered
TIME, t [hours] as being at steady


S650 -



0 50 100 150
TIME, t [hours]


"r 660

650 -


0 50 100 150
TIME, t[hours]

Winter 1995

It should also
be mentioned that
the oscillations
are not symmetri-
cal around the
upper steady state.
The center of
the oscillations is
T 6540R where
at the upper steady
state it is T =
651 R. Thus there
is no direct con-

Figure 1.
results to
large time for
a disturbance
of the basic

Use CEE's reasonable rates to advertise.
Minimum rate, 1/8 page, $100;
Each additional column inch or portion thereof, $40.

Faculty Position in Chemical Engineering: Responsible for teach-
ing undergraduate and graduate courses, supervising graduate re-
search. Applicants must have a Ph.D. and be a U.S. citizen or have
permanent resident certification. Candidates should have a strong
commitment to teaching, research, and professional activity. Send
curriculum vitae, list of three references, transcripts, and statement of
teaching and research objectives to Dr. W. J. Koros, Chairman, De-
partment of Chemical Engineering, The University of Texas at Aus-
tin, Austin, Texas 78712-1062. Affirmative Action/Equal Opportu-
nity Employer.

nection between the limit cycle and the upper steady state.
1 Small magnitude initial response as an indication of stability.
Prof. Loureiro notes that the magnitude of the initial response of
the CSTR to disturbance is small (0.1 0R), and as such it is judged as
The disturbance that was introduced to the system in the paper
was just due to the numerical values used as initial conditions in the
numerical simulation. The upper steady state values of the tempera-
tures were input with 5 decimal digits accuracy, and the concentra-
tion with 4 digits. These initial conditions were enough to perturb
the system from the upper steady state. Thus this disturbance in the
temperature was of the order of 10-3OR and the immediate system
response was a hundred times larger. Obviously much bigger initial
responses can be obtained by increasing the disturbance in many
ways. For example, changing the steady state concentration from
0.0591 to 0.06 leads to an initial temperature oscillation with an
amplitude of 4R. The initial amplitude values are less important
than the dynamic trends which develop after a disturbance. Figure
6 in the published paper"12 indicates growing oscillations, and Fig-
ure 1 clearly shows that the oscillations continue to increase dra-
matically until a stable limit cycle is reached.
In conclusion, we have shown using both a stability analysis of
the linearized system and a numerical simulation of the nonlinear
system that the upper steady state of this particular CSTR is
unstable for the basic model. Prof. Loureiro's contentions
that upper steady state is stable and is at steady state are just not
correct, as we have demonstrated in this letter. An excellent sum-
mary of the mathematics of CSTR multiplicity and stability is
given by Uppal, et
Mordechai Shacham
Neima Brauner
Michael B. Cutlip
1. Coughanowr, D.R., Process Systems Analysis and Control,
2nd ed., McGraw-Hill, New York, NY (1991)
2. Shacham, M., N. Brauner, and M.B. Cutlip, Chem. Eng. Ed.,
28(1), 30 (1994)
3. Uppal, A., W.H. Ray, and A.B. Poore, Chem. Eng. Sci., 29,
967(1974) 0

R educator


of the University of Delaware

University of Delaware Newark, DE 19716
S tanley I. Sandler, a native
of New York City, earned
his BS at the City College
of New York and his PhD at the
University of Minnesota, both in
chemical engineering. After a
year as an NSF Postdoctoral Fel-
low at the Institute for Molecular
Physics at the University of Mary-
land, he joined Delaware's chemi-
cal engineering faculty in 1967.
At various times since then he
has been the department chair and
interim dean of the College of
Engineering at Delaware, in ad-
dition to being a visiting profes-
sor at universities in England,
Germany, Australia, and Argen-
tina. He is currently the H.B. du
Pont Professor of Chemical En-
gineering and Professor of Chemistry at the University of
Delaware, and Director of its Center for Molecular and
Engineering Thermodynamics.
Stan's areas of research include thermodynamics, statisti-
cal mechanics, separations processes, and phase behavior.
He is the author of Chemical and Engineering Thermody-
namics and approximately 190 refereed papers, and is the
editor of seven books on thermodynamics and chemical
engineering education. Among the awards he has received
are the Professional Progress Award from the American
Institute of Chemical Engineers, the 3M Lectureship Award
from the American Society for Engineering Education, and
the Distinguished U.S. Senior Scientist Award from the
Alexander von Humboldt Foundation (Germany).

1962. What a year! Stan graduated from the City College
of New York, got married (the same week he graduated),

Stan then (circa 1941) and now.

and went to Minnesota for graduate school
shortly thereafter. One of his senior design
partners at CCNY was Richard Felder, now
at NCSU.
In those days new graduate students at Min-
nesota had to take written exams before classes
started. That was part of the qualifying pro-
S cess-indeed the only written qualifying ex-
--j ams. There were a total of five exams, and
Stan passed all but the thermodynamics exam!
(We've hitherto kept this secret from his
thermo classes!) During that first summer at
Minnesota he had a job at General Mills and
spent considerable time studying thermodynamics. He sub-
sequently developed a structure for understanding thermo-
dynamics which enabled him to pass the qualifying exam
and, about fifteen years later, it became the basis for his
widely used textbook Chemical and Engineering Thermody-
namics. Clearly, the exam process had many unanticipated
Stan's graduate student colleagues at Minnesota who have
since gone into academia include Bruce Finlayson, Dan
Luss, Ramkrishna, Mort Denn, Harmon Ray, George Gavalas,
Ben McCoy, and others. During the last year of his graduate
work, his thesis advisor, John Dahler, took a sabbatical leave
at the University of California, Berkeley, and brought Ben
McCoy and Stan with him. John was in the physics depart-
ment, while Ben and Stan were housed in a "temporary"
building built during the first or second world war that was
still being used as graduate student offices in 1965.
Notable events while Stan was at Berkeley included the

Copyright ChE Division ofASEE 1995

Chemical Engineering Education

... he joined Delaware's chemical engineering faculty in 1967. At various times since then he has been
the department chair and interim dean of the College of Engineering at Delaware, in addition to being a
visiting professor at universities in England, Germany, Australia, and Argentina. He is currently the
H.B. du Pont Professor of Chemical Engineering and Professor of Chemistry at the University of
Delaware, and Director of its Center for Molecular and Engineering Thermodynamics.

birth of his first child (Cathy, now As-
sistant Vice President at MBNA), and
discovering the West Coast, the Sierra
Mountains, and the desert. It should be
noted that Stan grew up in midtown
Manhattan and that his parents never
had a car. Their philosophy was that
anyplace you could not get to via the
New York City subway system was not
worth going to! Consequently, going to
Minnesota, and still further to Califor-
nia, were great adventures for Stan. His
wife Judy claims he travels so much
because he has this unending urge to
discover what lies beyond the New York
City subway system.
During Stan's time in California, he
also interviewed for jobs for the first
time, and his first interview trip, to the
Los Alamos and Sandia Laboratories,
was memorable. Not only was he
amazed by the beauty of the South-
west, but it was also his first experience Stan and Judy d
in flying. He arranged to fly from the bicycling v
Oakland airport to Los Alamos by way
of San Francisco and Albuquerque. The flight from Oakland
to San Francisco was by Hoovercraft, then to Albuquerque
by jet prop, to Los Alamos by a twin-engine, eight-seater
plane, to Albuquerque by a single-engine, four-seater plane,
and from San Francisco to Oakland by helicopter. Conse-
quently, on his first airplane trip he experienced virtually
every type of air transportation then available.
Stan finished his thesis at Berkeley, and then, to defend,
"dropped in" at Minnesota on his way as a National Science
postdoc to the (then) Institute for Molecular Physics at the
University of Maryland, working with Ed Mason. His postdoc
work combined a continuation of things he had done at
Minnesota with some more applications of kinetic theory.
One application was on the kinetic theory of ionized gases,
which resulted in consulting (in 1970) for Martin Marietta to
help design the heat shield for a Project Galileo probe to
Jupiter that is only now on its way to that planet.
During the time Stan was a postdoc, Mort Denn invited
him to give a seminar at Delaware, and this visit subse-
quently led to an offer of a faculty position. Having grown
up in New York and living in Minneapolis, the San Fran-
Winter 1995


cisco Bay area, and the Washington,
DC area, his wife and he could not
imagine living in the small town of
S Newark. They decided to try it for
three years, however-and twenty-
seven years later they are still there.
They found Newark to be a delight-
ful town, close to Philadelphia, Bal-
timore, Washington, and New York.
Stan was welcomed to the depart-
ment by excellent, friendly, and ex-
tremely helpful colleagues. They
quickly recognized Stan's talents, and
he was promoted to associate pro-
fessor three years after he arrived!
His promotion to full professor came
six years after his arrival and several
weeks before his 33rd birthday.
Clearly, Stan was ready for a bril-
liant career at Delaware.

g a recent summer Stan's versatility was on display
ion in Europe. during his early years at Delaware,
when he taught a large variety of
courses. In those days, classes were held Monday through
Saturday noon, and during his first two or three years he was
regularly assigned to teach classes at 8 and 9 a.m. on Tues-
days, Thursdays, and Saturdays. Since he has been at Dela-
ware, he has taught virtually the entire curriculum at one
time or another, including (of course) several "flavors" of
His transition to applied thermodynamics was catalyzed
by his organizing an Engineering Foundation conference of
both academic and industrial engineering thermodynami-
cists. This conference, "Phase Equilibria and Fluid Proper-
ties in the Chemical Industry," was held in January of 1977
and probably marked his entry into, and acceptance by, the
applied thermodynamics community. The meeting was a
great success, helping to revitalize thermodynamics (includ-
ing sessions at AIChE meeting). It also resulted in several
long-term consulting contracts for him that have further
helped Stan develop the field. He co-organized the second
conference, held in Berlin in 1980, with Helmut Knapp of
the Technical University of Berlin. The conference now
continues, being held every three years and alternating be-

Stan in one of his laboratories.
tween North American and Europe.
Another conference he started (in 1982) is the so-called
"Mid-Atlantic Thermodynamics Conference." This meeting
rotates among the core schools of Delaware, Penn, Princeton,
and Johns Hopkins, and now also includes Rutgers,
Virginia, and Penn State, with occasional visitors from
farther afield. One characteristic of these symposia is that
only graduate students are allowed to make presentations. It
is meant to be an event during which graduate students
working in the broadly defined area of thermodynamics
can meet and interact with their colleagues at other (rela-
tively) nearby schools.
The thermodynamics conferences in California and Berlin
led to Stan's close association and friendship with Helmut
Knapp of the Technical University of Berlin. As a result, he
spent one sabbatical leave in 1981 (taking only his middle
child Joel with him since Judy was working at the time) and
part of another 1988-89 sabbatical leave at the Technical
University of Berlin. Both were supported by the Alexander
von Humboldt Foundation (Bonn, Germany). As a result of
this Berlin connection, approximately a dozen students from
Berlin, and also two from the University of Karlsruhe, have
done their studienarbeit or diplomarbeit in Stan's laborato-
ries at Delaware, and five Delaware students have spent
varying amounts of time at T.U. Berlin. Another especially
delightful part of Stan's Berlin sabbatical was the develop-
ment of a friendship with John and Susie Prausnitz, who
were also there.
Stan was also co-chairman of the ASEE Chemical Engi-
neering Faculty Summer School held in Santa Barbara, Cali-
fornia, in 1982, and organizer and chairman of the 1983
National Science Foundation workshop "Thermodynamic
Needs for the Decade Ahead: Theory and Experiment." He
has also served on the organizing committees for many other
meetings, including the two "Beijing Symposia on Thermo-


The Sandier family at home: Joel and Michael
(first row), Catherine, Stan, and Judith.
dynamics in Chemical Engineering and Industry," as well as
meetings conducted by the National Bureau of Standards
(now NIST) and the ASME. In addition, he has served as
meeting session chairman, on various committees, subcom-
mittees and panels of the National Science Foundation and
the AIChE, on the National Research Council/National Acad-
emy of Sciences Evaluation Panel for the Center for Chemi-
cal Engineering of the National Bureau of Standards, as a
member of the Board of Trustees of the CACHE corpora-
tion, and on editorial boards of Industrial and Engineering
Chemistry Fundamentals, the AIChE Journal, and the Uni-
versity of Delaware Press. At present he is Thermodynamics
Area Editor for Chemical Engineering Education, and is on
the editorial advisory boards of the IEC Research, the Jour-
nal of Chemical Engineering Data, Engineering Science and
Technology (Malaysia), Indian Chemical Engineer (Calcutta),
and the John Wiley & Sons Series in Chemical Engineering.
He is also on the external advisory board of the Department
of Chemical Engineering at Carnegie-Mellon University,
having previously served in a similar role at LSU.
Another of Stan's activities had a major impact of chemi-
cal engineering education in this country. Recognizing the
need to bring examples from new technologies and nontradi-
tional areas into chemical engineering courses, he organized
and chaired a 1988 Engineering Foundation Conference titled
"Chemical Engineering Education in a Changing Environ-
ment." One goal of the conference was to discuss, in the
context of examples and case studies, the use of fundamental
chemical engineering principles in such new technology ar-
eas as solid state processing and electronics, biochemical
and biomedical engineering, hazardous waste management
Chemical Engineering Education

and disposal, food processing, etc. Another goal was to
produce a proceedings volume of examples and case studies,
with sufficient data and detail that they could easily be
included into traditional chemical engineering courses. This
made it possible to rapidly bring new technology areas into
the chemical engineering educational program and class-
room without the need for major curricula revision or the
introduction of new courses. Stan and Bruce Finlayson ed-
ited these conference Proceedings, which were published by
the Engineering Foundation and distributed by the AIChE in
the fall of 1988. Stan and colleague Kenneth Bischoff, rec-
ognizing a need to introduce safety and environmental con-
siderations and risk and hazard analysis in an already crowded
chemical engineering curriculum, have been teaching a very
popular one-semester elective course titled "Risks, Safety,
Hazards, and the Environment." Perhaps Stan's next book
will arise from that course.
Stan's textbook, Chemical and Engineering Thermody-
namics, has made a major contribution to chemical engi-
neering education and went through thirteen printings. It
was a featured selection of the McGraw-Hill Book Club, has
been translated into Spanish and Chinese, and is reprinted in
Taiwan, Korea, and India. The second edition of the book is
the first chemical engineering textbook to contain, as an
integral part, a disk of calculational personal computer pro-
grams to be used for thermodynamic and phase equilibrium
calculations. These programs have also been used by stu-
dents in the stagewise operations and capstone design courses
to estimate phase equilibria easily and realistically. This new
edition has been well received, with numerous adoptions
and sales in excess of 9,000 copies in the U.S. and an
additional large number of paperback copies in Europe and
Asia since its introduction in 1989. It has already been
translated into Korean.
Stan has been a strong advocate of the use of computers in
chemical engineering courses. He introduced process simu-
lation into the senior design course, was instrumental in
establishing a microcomputer laboratory, and acquired a
VAX 785 computer for the department (in the early 1980s!).
Nationally, he has advocated chemical engineering compu-
tation as a member of the CACHE Corporation Board of
Trustees. He was also one of the first chemical engineers to
develop computer-assisted instructional materials for the
PLATO system and for the IBM-PC.

Stan's research program has grown from a very fundamen-
tal base to include applications-oriented components. For a
number of years after coming to Delaware, he continued to
do research in statistical mechanics, though moving away
from the kinetic theory of dilute, structured gases to various
aspects of equilibrium properties of liquids. Recognizing
that a lack of knowledge of the interactions between mol-
ecules was a difficulty in his work, Stan initiated research
Winter 1995

directed toward obtaining such information for nonspherical
molecules. Here he pioneered the method of comparing
the results of statistical mechanical theory and computer
simulations with X-ray diffraction data taken by his re-
search group in collaboration with scientists at the Oak Ridge
National Laboratory to obtain information on inter-
molecular forces and order. This work was a major step in
using theory, rather than geometric models, to interpret mo-
lecular diffraction data.
A significant accomplishment during this period of his
career was the development of a statistical mechanical per-
turbation theory based on the use of a nonspherical reference
potential. Before his work, statistical mechanical perturba-
tion theories had been based on the use of spherical mol-

Stan has been a strong advocate of the
use of computers in chemical engineering courses.
He introduced process simulation into the senior
design course, was instrumental in establishing a
microcomputer laboratory, and acquired a VAX 785
computer for the department (in the early 1980s!).

ecules (generally hard spheres) as the reference fluid. While
such perturbation theories were of some use for long-range
forces, such as polar and multipolar forces, they were poorly
convergent for the addition of shape and overlap forces.
Stan's work showed how these latter effects could be ac-
counted for accurately in perturbation theory.
Stan then moved into an extremely productive phase of
research in which he applied the lessons learned from his
statistical mechanical research to the area of applied thermo-
dynamics, which is of more immediate need to chemical
engineers. First, with Professor Amyn Teja (Georgia Insti-
tute of Technology), he developed a highly accurate corre-
sponding states principle using two reference fluids. Next he
initiated a successful and continuing research program on
the measurement of vapor-liquid equilibrium of systems that
are chemically reactive, associate, or are difficult systems in
other ways, to provide data for his theoretical research and
that of others, especially in the area of group contribution
methods. A part of this work has been an international coop-
erative effort involving researchers at the Technical Univer-
sity of Berlin and the University of Paris.
Coincidental with this work has been a continuing theo-
retical research program which has included the develop-
ment of a rigorous thermodynamics of continuous, but
bounded, mixtures, which was immediately adopted by the
Chevron Oil Field Research Company in its reservoir simu-
lations and crude characterizations and is being used by
others in designing crude oil distillation columns. Stan is
also using this work for the description of polymer solution
phase equilibrium and biochemical separations.
Other aspects of Stan's research have also had important
Continued on page 17.




E.I. du Pont de Nemours and Co., Inc. PO Box 80304 *

(This paper is based on presentations made at the ASEE
National Symposium on Chemical Engineering Education,
August 11, 1992, at Montana State University, Bozeman, MT.)

Particles are critical to the success of many products,

but the U.S. chemical industry lags behind its foreign
competitors in the understanding and application of
particle science to chemical process technology. This is due,
in part, to the neglect of particle science and technology in
the education of scientists and engineers. Although most
chemical processes involve particles, the typical U.S. chemi-
cal engineering curriculum devotes little time to particle
technology. Consequently, new particle processes have dis-
mal prospects for startup compared to liquid and gas pro-

Ralph Nelson is a Research Associate in the
DuPont Corporate Center for Particle Science
and Technology. He taught at Middlebury Col-
lege and West Virginia University for eight years
before joining DuPont. He spent eight years
providing technical support for pigment manu-
facture, followed by twelve years consulting in
particle technology, specializing in problems re-
lated to particle-liquid interactions.

Reg Davies is Principal Consultant and Re-
search Manager of the DuPont Corporate Cen-
ter for Particle Science and Technology. He
consulted in particle technology for ten years
with IIT, then twenty-six years with DuPont. He
was Technical Chairman of the International
Fine Particle Research Institute for 1979-91
and is currently Chairman of the AIChE's Inter-
national Particle Technology Forum.

Karl Jacob is Research Leader with the Dow
Chemical Company. He has devoted eleven years
to understanding particle processing. In 1988 he
founded the Solids Processing Lab, which pro-
vides support to Dow's engineering, research,
and manufacturing communities. He specializes
in solving problems related to drying, fluidization,
bins, hoppers, and powder flow.

* The Dow Chemical Company, Solids Processing Lab, Bldg.
1319, Midland, MI 48667

Wilmington, DE 19880-0304

cesses, and our engineers must struggle to optimize and
retrofit existing production facilities that process particles.
We propose that the present set of undergraduate courses
in chemical engineering be modified to include more prob-
lems illustrating the challenges posed by particle processes
and exploring practical routes to their resolution. Eventually
the curriculum should also include a separate course provid-
ing an integrated approach to engineering applications of
particle technology.

Table 1 outlines the technical areas of particle technology.
We have found that some areas of powder technology are
virtually absent from U.S. engineering curricula: powder
storage, Jenike's theory of hopper flow, cyclones, dilute
phase conveying, fluidization, and powder milling.
Particle technology is becoming increasingly important.
As the microcomponents of information retrieval, expert
systems, medical diagnostics, and robotic manufacturing and
surgery become more sophisticated, the need for carefully
controlled structures has created an enormous market for
particles smaller than a micrometer in diameter. Modern
agricultural, ceramic, pharmaceutical, and medical diagnos-
tic materials require sophisticated control of particle size and
agglomeration so that they can release ingredients on a well-
defined schedule. Conversion of what was once considered
waste into products useful to society often relies on sophisti-
cated particle-generation and control schemes.
Customer needs and specifications for more mundane com-
mercial products (pigments, plastics, pharmaceuticals, de-
tergents) have become more stringent as companies seek to
increase the effectiveness and to reduce the environmental
impact of their products. Companies that cannot adequately
control particle characteristics will lose market share to those
companies who technologists have a better grasp of particle
There will be many opportunities to apply particle tech-
nology to
Copyright ChE Division ofASEE 1995
Chemical Engineering Education


Introduce differentiated products at a premium
price based on carefully tailored particle proper-
ties or size distributions
Reduce scale-up time and cost by understanding
what basic data must be acquired and what full-
scale processes best implement the unit operations
developed at lab-scale
Reduce rework costs by having meaningful
specifications and process controls to make
particles "right" early in the process
Improve quality control and reduce environmental
problems by improved sampling and on-line
monitoring for properties critical to efficient
Increase capacity and reduce costs by tailoring
product properties to give higher solid-liquid
separation rates, shorter drying times, and lower
yield losses.
Textbooks designed for undergraduate courses on particle
technology include Principles of Powder Technology[lI (used
by the Institute of Chemical Engineering in Great Britain for
their post-graduate training course) and Bulk Solids Han-
dling.[2] Unfortunately, neither book contains any problems.

DuPont and Dow are the fifth and sixth largest chemical
manufacturers in the world (after BASF, Bayer, Hoechst,
and ICI). Together, Dow and DuPont spent about $2.4 bil-
lion on research and development in 1991-half the entire
total for the US chemical industry and ten times the amount
of federal support for university non-defense research and
development in chemistry and chemical engineering.[31 A
1985 survey found that about 60% (by value or volume) of
the products sold by DuPont are sold in particulate
form, while another 18% have particulate additives. Some
50% (by volume) of Dow's products are solids. The amount
of solids handled is three to four times the amount finally
sold when raw materials, intermediates, and minimal-value
coproducts are considered.
DuPont and Dow operate several thousand major unit
operations involving particulates, each one requiring super-
vision by a technologist who understands the relevant areas
of particle technology. In 1988 DuPont designated particle
technology as one of its core technologies (critically impor-
tant to success in global competition), and Dow constructed
a solids processing "technology well" (to focus technical
Because it is difficult to find US graduates adequately
trained in particle technology, we must spend considerable
time training people in particle technology and we recruit a
disproportionate number of such specialists abroad (com-
Winter 1995

We propose that the present set of undergraduate
courses in chemical engineering be modified
to include more problems illustrating the
challenges posed by particle processes
and exploring practical routes
to their resolution.

Technical Areas of Particle (Powder/Slurry/Emulsion)

Characterization Sampling methods and statistics; methods for
determining and characterizing particle size distribution, shape,
surface roughness, porosity; packing in heaps and sediments;
particle charge; adsorbed material; interfacial tension; granule
Powder Storage Angle of repose; hopper discharge; fines
Mechanical Transport belt/bucket/screw conveyors;
flowability; dynamic weighing; power consumption; selection
based on particle size/shape/strength of agglomeration
Convective Transport sedimentation rate and solidosity;
suspension in stirred tanks; rheology of slow-settling slurries;
cyclone separation; dilute and dense-phase flow in pipes; static
generation and dust explosions; flow-enhancement additives;
fluidized beds, fluid flow through packed beds; erosion; selection/
designing of valves/pipes/pumps; metering
Flocculation/Agglomeration Interparticle and interface
forces; polymer adsorption and bridging; wetting and capillarity;
floc strength and structure; granulation, briquetting, and tabletting
Deflocculation/Deagglomeration/Grinding Dispersion
stability; grinding aids; particle fracture strength and toughness
Foam/Emulsions/Fouling Interfacial tension (interface
energy); effects of surface curvature on solubility and vapor
pressure; adsorption isotherms; foam/antifoam additives; froth
flotation; emulsion formation and coalescence; decanter designs
Drying Capillary flow and compression; solubility and salt
transfer; moisture diffusion and vaporize-condense cycling effects;
use of convection to speed drying; heat transfer; induction heating;
freeze drying; displacement drying
Mixing and Blending Effects of component size, shape, and
density on blend time; gas release during wet-in; considerations in
mixer design; demixing during flow; achieving remixing or
resuspension; power consumption; rate of dissolution
Coating Alternatives to dry blending; caking, dusting, binders,
slurry flow and drying in film coaters; film surface defects caused
by particles
Solid-Fluid Separation Selecting solid-liquid separation
equipment; sedimentation; cyclones; electroseparation; cake
formation and structure; washing and dewatering the cake

pared to national vs. international sales dollars). Some 40%
of the engineers in the DuPont Center for Particle Science
and Technology were trained abroad, and Dow's Solids Pro-
cessing Center is located in Europe. Those US companies
which are not multi-national must rely on post-college train-
ing and the rather small flow of US graduates who are
trained in particle technology. All of us would benefit if US

chemical engineering curricula were stronger in particle tech-

Very definitely yes! The presence of particles introduces
considerable complexity (compared to particle-free gases
and liquids) into the selection of equipment and also into the
estimate of the complexity and cost of operation and mainte-
nance. Areas of particular concern are powder flow from
bins, pneumatic conveying, metering, valves and pumps for
solids and slurries, mixing and wetting-in, pipe plugging
under low flow or shut-down conditions, and quality control.
Particles can cause problems with commercial-scale opera-
tion that are not evident during laboratory or even pilot-scale
tests, so someone skilled in particle technology should be
involved with development of a new process from the earli-
est stages of research.
In a study of forty industrial startups, E.W. Merrow found
that processes involving particles took several times longer
to start up than those involving liquids or gases,141 even
though manufacturers anticipated problems and planned
lengthier startup times for those projects! Plants involving
several new unit operations were particularly hard to start
up, and some particle processing plants have been aban-
doned (without producing any significant amounts of prod-
uct) after tens of millions of dollars had been spent trying to
make them work.
Merrow suggested that much of the problem is due to a
lack of fundamental understanding of particulate phenom-
ena and to a lack of applied research in this area. Inventors
assume that engineers will take care of any awkward details
of powder technology during scaleup, and the engineers
assume that designers will solve the problems. In the end,
the production staff is left to cope with a marginally operable

Unfortunately, no. In our recruiting visits to many cam-
puses over the past four years we have found no U.S. univer-
sity with a full undergraduate program in particle technol-
ogy. Michigan Technological University offers an elective
course every other year, and a few other universities offer
graduate-level courses in agglomeration and fluidization. A
recent study found that the average engineering curriculum
provides less than twenty minutes of discussion on solid-
liquid separations.?5i The distribution was found to be quite
uneven-most schools had no discussion, while a few had
several hours.
The strongest programs in particle technology are found in
the departments of Mining and Mineral Science. But the
courses on size reduction, agglomeration, and particle-liquid
separation tend to ignore the molecular aspects of particle

technology since few mining engineers pursue careers in the
chemical industry. Chemical engineering courses on crystal-
lization, solid-liquid separation, drying, and bulk handling
focus on mass and energy balances and tend to ignore the
mechanical aspects of particle technology, population bal-
ance, pore-size distribution, agglomerate strength, and the
structure of packed beds.
Other nations recognized the importance of particle tech-
nology several decades ago and now have well-established
programs. Japan has courses in powder technology in 24 of

Introductory Course in Powder Technology

Outline of a typical one-year undergraduate course in powder
technology at a German technical university, comprising 90 hours of
lecture, 30 hours of discussion, and 12 half-day experiments.

When Fluid Is Not a Major Factor
n Description of Particulates
Single particle size/shape
Measuring/modeling size distribution
n Separation of Solid/Solid Mixtures
Mass balance
Separation parameter
n Mixing of Solids
Mixture types/statistics/sampling
H Size Reduction/Grinding
Surface energy/cracks/fracture strength distribution
Energy efficiency, equipment
n Size Enlargement/Agglomeration
Types of particle bonding
n Storage and Bulk Flow
Angle of repose, mass/funnel flow
Design of hoppers/silos

When Fluid Must Be Considered
n Two-Phase (Solid/Fluid) Flow
Pressure drop in fixed/fluidized beds
n Size Classification/Composition Sorting
Sifting (cross/counter-current)
Differential sedimentation/centrifugation
n Separation of Solids from Fluid
Spray scrubbers/electrostatic precipitation
n Fluidized Beds
Principles/nozzle patterns
Gas vs. liquid
Applications to catalysis/combustion
n Pneumatic Conveying
Dense/dilute phase
Pressure/vacuum systems
n Solid/Fluid Mixing
Impeller design/power use
Wetting, dispersion
Heat/mass transfer
Chemical Engineering Education

its 38 universities, and 16 government institutes have strate-
gic programs designed to aggressively focus on and capture
markets that are dependent on particle technology.161 There
are strong programs in Europe at Delft and Twente (The
Netherlands), Bradford, Surrey, and Loughborough (Great
Britain), Hamburg-Harburg, Clausthal, Karlsruhe, and
Braunschweig (Germany, where thirteen universities have
powder technology programs), and in Canada at McGill,
Laval, British Columbia, and Western Ontario. A typical
curriculum is given in Table 2.

Society's mastery of each area of technology goes through
several stages. At first, mastery increases slowly as research-
ers explore the basic phenomena (the emerging stage). Then
mastery increases rapidly as major quantitative theories are
utilized (the vital stage), and finally, mastery levels out as
the area is so well studied that new understandings are rare
(the mature stage).
The emerging stage covers science where the markets are
still uncertain, where there is a question of what variables
are important in controlling the process, and where the pro-
cesses for commercial production have not been designed.
There is no organized body of understanding for the area, so
there is little justification for including it in an undergradu-
ate engineering curriculum. New developments cover wide
areas and are patented to protect them from use by competi-
tors, and production volumes are low. Current examples
include genetic engineering, superconductor applications,
and supercritical extraction.
The vital stage covers processes that are widely used but
which often cause operating problems. Basic theories and
models exist, but applying them to commercial processes is

Examples of Industrial Support for
Nonproprietary Studies in Particle Technology
(Authors will provide addresses on request.)

-- International Fine Particle Research Institute
Association of Crystallization Technology
American Filtration Society
Separation Process Services
Particle Technology Forum of the American Institute of Chemical
-- Colloid and Interface Science Division of the American Chemical
Gordon Research Conferences on Colloid Science
Center for Microengineered Ceramics (University of New
-- Consortium for Advanced Materials Processing (Clarkson
-- Particulate Materials Center (University of Pennsylvania)
Winter 1995

a challenge. Capability is developed incrementally, and com-
petitive advantage is maintained by treating new develop-
ments as trade secrets. Current examples include slurry flow,
agglomeration, crystallization, filtration, and milling (espe-
cially the kinetics of breakage). We are beginning to build
quantitative models to predict the behavior of commercial-
scale particulate processes. Expert systems and parallel pro-
cessors can help chemical plant operators anticipate prob-
lems and decide on the best response, and some particulate
processes are now run under computer control.
The mature stage covers areas that have been reported
extensively in the technical literature and are widely and
confidently practiced to design commercial equipment that
operates trouble-free and well within the limits of past expe-
rience. Expertise and equipment are available from many
sources at a lower cost than developing and maintaining the
capability in-house (unless the company plans to specialize
in the area). Current examples include pumps for gases and
liquids, distillation towers, heat exchangers, ion exchange
columns, gas adsorption beds, refrigeration units, and elec-
tric motors.
For emerging technologies, teaching can be restricted to
the basic concepts. As areas move from the emerging to the
vital stage, they should receive increased attention in engi-
neering curricula since it is the application of vital technolo-
gies that provides industry with the greatest competitive
advantage. As areas move into the mature stage, their place
in the curriculum should be pruned back. The natural temp-
tation to "overteach" mature technologies must be resisted;
all that is needed is to make the fundamental principles clear
and to describe the range of applications for commercially
available technology. The elaborate design protocols for
mature technologies are best taught within the companies
that specialize in applying them.

The U.S. has become increasingly aware of the impor-
tance of particle technology. In 1973 a group of international
filtration experts strongly urged more research and educa-
tion in solid-fluid separation.'7' In 1988 the U.S. National
Research Council Committee on Separation Science and
Technology recommended focused research in six specific
areas of solid-liquid separations."5 Industry now provides a
great deal of financing for nonproprietary research in par-
ticle technology (see Table 3), and many industrial practitio-
ners (including the authors) regularly teach continuing edu-
cation courses in particle technology. It is now time for U.S.
universities to respond to the need by increasing the empha-
sis on particle technology in their curricula.
Although industry would like to achieve overnight what
our competitors have built up over decades, we realize that
gradual change is more likely. For example, classes in fluid
mechanics could introduce consideration of dilute-phase gas

and slurry transport. For the longer term, the National Sci-
ence Foundation is supporting a joint academic-industrial
effort to develop undergraduate courses in particle technol-
ogy. Four week-long workshops will be held in 1994 and
1995, each helping twenty-five engineering faculty build
elements of particle technology into their courses. The teams
are DuPont and Penn State University, 3M and the Univer-
sity of Minnesota, Dow and the University of Houston, and
Westinghouse and the University of Pittsburgh.

First, in the academic tradition, take this quiz:
Have you included examples of the relationships be-
tween particle size distribution, state of agglomeration,
and end-use properties in your lectures, in required read-
ing, in labs, and in homework?
Have your graduates learned enough about particulate
operations to participate on a startup team without
extensive additional training?
Can your graduates design a dust collector, a slurry
transport line, or a storage vessel for a cohesive
Can your graduates make a computer modelfor the
behavior and control of a process involving particles?
Are your graduates aware of the typical problems
encountered in systems involving particles and do they
know enough of the terminology to discuss the problem
with a consultant in powder technology?
Have your graduates learned the most meaningful
parameters to monitor and what options exist for
resolving typical problems?

Options for enriching the curriculum include

Permitting students to take (and count as electives
toward a degree) courses in other departments, such as
mining engineering, which already have strong courses
in particle technology.
Assigning the chemical engineering department to focus
exclusively on gas and liquid processes and starting a
new department devoted exclusively to particle technol-
ogy, as is done in Germany.
Developing continuing education courses in particle
technology to be taught on campus, at a local corpora-
tion site, at AIChE meetings, via satellite through the
National Technological Institute, or at a 'for-profit"
training center such as the Center for Professional
We hope that this paper will inspire you to act. If you can
educate engineers to function effectively in this area of criti-
cal need, the U.S. chemical industry will hold onto its sub-
stantial share of the world chemical market, enhance the
employment opportunities for technologists and plant work-

ers, and increase the support for academic research in par-
ticle technology.

1. Rhodes, M.J., ed., Principles of Powder Technology, John
Wiley & Sons, Ltd., Chichester, England (1990)
2. Woodcock, C.R., and J.S. Mason, Bulk Solids Handling: An
Introduction to the Practice and Technology, Chapman and
Hall, New York, NY (1987)
3. Brennan, M.B., and J.R. Long, "Facts & Figures for Chemi-
cal R&D," C&E News, 70(33), 38 (1992)
4. Merrow, E.W., "Estimating Startup Times for Solids-Pro-
cessing Plants," Chem. Eng., 95(18), 89 (1988)
5. Tiller, F.M., "Separation and Purification: Critical Needs
and Opportunities," Fluid/Particle Sepn. J., 1, S10 (1988)
6. Iinoya, K. Particle Technology in Japan, International Fine
Particle Research Institute (1980)
7. Alt, C., et al., "The Crisis in Solid-Fluid Separation Technol-
ogy," Filtration & Separation, 10(6), 670 (1973) 0

Book review

by Vladimir Kucera
Prentice Hall, 472 pages, $66 (1992)

Reviewed by
B. Wayne Bequette
Rensselaer Polytechnic Institute

The subject of this book is the analysis and design of control
systems for systems characterized by linear, constant coefficient,
discrete-time models. It is assumed that the systems are perfectly
modeled and unconstrained; the modernr control theory" approach
[ca. 1960s with some recent (early 1980s) results] is used.
The intended audience for the book, as stated by the author,
includes "the graduate student who intends to specialize in linear
control and the practicing engineer or applied scientist who is
interested in new perspectives of linear control theory. For the
specialist, the book is intended as a reference and, hopefully, as an
inspiration for further research." It is assumed that the reader has a
background in abstract and linear algebra, linear system theory, and
stochastic processes.
I liked the author's philosophy of placing bibliographical notes at
the end of each chapter rather than disrupting the presentation of
material with references. The text portion of each chapter is con-
cisely written, with the bulk of the material dominated by math-
ematical equations. Beginning students may have trouble using this
text for self-study since they will typically need more motivation
and justification for the derivations.
Another aspect of the book that I liked were the problems at the
end of each chapter. Short answers and more detailed solutions to
all of these problems are provided in appendices. I found it odd that
the author bothered to include the short answers section (eight
Continued on page 31.
Chemical Engineering Education

EDUCATOR: Stan Sandier
Continued from page 11.

industrial applications. For example, his research on
calculational methods and equations of state has been used
in the design many commercial processes and to achieve
major reductions in computation time in reservoir simula-
tion calculations.
Another important research effort has been his work on
the generalized van der Waals theory. This work has led to a
new understanding of the theoretical bases and assumptions
that underlie the equations of state and activity coefficient
models currently in use, as well as the development of com-
pletely new, theoretically based models. Perhaps the most
significant product of his recent research has been a new,
theoretically correct mixing rule that allows equations of
state and activity coefficient models to be combined in such
a way that equations of state can now be used to predict the
behavior of highly nonideal mixtures over wide ranges of
temperature and pressure. The manuscript describing this
work was so enthusiastically received by the reviewers that
it was published in the AIChE Journal within nine weeks of
its original submission. Other recent work that Stan is espe-
cially proud of includes developing thermodynamic bounds
for microbial growth processes, the use of quantum mechan-
ics and other computational chemistry methods in chemical
engineering thermodynamics, and (with Eric Kaler) on the
separation of biological materials. His research activities
also include phase behavior studies relevant to environmen-
tal protection which have included obtaining data for the
reformulation of gasoline, the behavior of organic pollutants
in aqueous solution, and (with Michael Paulaitis) on CFC
replacements. All this work has been enthusiastically re-
ceived by the international thermodynamics community, and
in recent years Stan has been invited to speak on his research
in France, Germany, Canada, Mexico, Denmark, China, South
Korea, Italy, the Czech Republic, Poland, Australia, Argen-
tina, and at many universities and companies in the United
States. Recently Stan has been the Phillips Lecturer at Okla-
homa State University, the Aston Hall Carey Lecturer at
Georgia Tech, and the ICI Distinguished Lecturer at the
University of Alberta.

Shortly after returning to Delaware from his 1981 sabbati-
cal leave, Stan was named the H.B. du Pont Professor of
Chemical Engineering and a short time later was asked to
become chair of the department (in which capacity he served
until 1986). His efforts as chair led to the successful nomina-
tion of five PYIs (Paulaitis, Dhurjati, Klein, Barteau, and
Lenhoff). He may still hold the record for the ChE depart-
ment chair, and perhaps any department chair, with the most
successful Presidential Young Investigator nominations! He
also rebuilt the chemical engineering faculty by hiring Antony
Winter 1995

Beris, Prasad Dhurjati, Hank Foley, Bramie Lenhoff, and
Andrew Zydney. He remains a mentor to many of these, and
other, faculty.
Stan has also given his time to the University. His service
activities have included several terms in the University Sen-
ate, chair of successful search committees for the chemistry
and electrical engineering department chairs, membership
on the Provost search committee (twice), head of major
University committees such as the Budget and Space Prior-
ity Committee and the Committee on Committees, and mem-
ber of many other committees.
The University of Delaware recognized Stan's special cam-
pus role by bestowing on him its highest faculty honor, the
Francis P. Alison Award. This nicely summarizes the high
esteem in which he is held by colleagues, who continue to see
him as the consummate scholar, educator, and mentor. O

S book review

Volume II, Part A, 4th Edition
Edited by G.D. Clayton and F.E. Clayton
John Wiley and Sons, Inc., New York, NY; 945 pages (1994)

Reviewed by
Klaus D. Timmerhaus
University of Colorado

This is the first of six books of a well-respected industrial toxi-
cology source book. Part A of this comprehensive compilation in
Volume II contains fifteen chapters from a total of forty-six chap-
ters covering thousands of chemicals. The first book of Volume 2
begins with an overview of industrial toxicology followed by the
general criteria for identifying and clarifying toxic properties and
recognizing occupational carcinogens of chemical substances. The
following twelve chapters analyze a number of well-known toxins
with their physical and chemical properties, industrial resources,
analytical determination, physiological response in terms of toxic-
ity and human effects, and exposure limits. The chemical sub-
stances covered in Part A of this revised series include:
Occupational carcinogens
Complex mixtures of tobacco smoke
Aldehydes and acetals
Epoxy compounds
Organic peroxides
Aliphatic nitro, nitrate, and nitrite compounds
Continued on page 38.

e M learning in industry

The Phillips Petroleum Company



Phillips Petroleum Company
Bartlesville, OK 74004

A significant number of PhD graduates currently en-
tering the teaching profession do not have industrial
experience. Over the past several years the trend
has been aggravated by the poor job market where PhD
graduates who would like to gain industrial experience prior
to starting an academic career find limited opportunities.
The shortage of funding in "traditional" chemical engi-
neering areas has moved chemical engineering research away
from the mass, heat, and momentum transfer unit operations.
While research in biotechnology, microelectronics, and bio-
medicine has made valuable contributions to society and has
significantly expanded the field of chemical engineering, it
does not prepare the graduate student to teach traditional
material covering pumps, heat exchangers, distillation col-
umns, etc. The professor also is not exposed to the world of
economically driven, short-deadline, industrial research.
The distribution of teaching assignments becomes diffi-
cult as older professors with industrial experience retire and
are replaced by professors with limited industrial experi-
ence. The current emphasis on incorporating design into all
stages of the chemical engineering curriculum can best be
accomplished if the instructors have been exposed to indus-
trial research and design processes.

The cornerstone of the curriculum, the senior design course,
is best taught by a professor with industrial experience.
Many departments are currently using adjunct professors to
meet this need. While this approach is viable, it does not
offer the benefits of having full-time faculty with industrial

In an effort to assist recent graduates in gaining industrial
experience, Phillips Petroleum's Corporate Technology or-
ganization initiated a program to expose professors to indus-
trial research. The program, The Industrial Experience Pro-
gram (IEP), is coordinated through our Bartlesville research
center where the professor spends up to twelve weeks during
the summer. While some of the work might be conducted at
the professor's home, residency in Bartlesville is required

R. Bruce Eldridge received his BS and MS from
the University of Arkansas and his PhD from the
University of Texas at Austin, all in chemical
engineering. He is a Senior Engineer for Phillips
Petroleum Co., specializing in separations re-
search, and is also an Adjunct Professor of Chemi-
cal Engineering at Oklahoma State University,
teaching a graduate-level separations course.

Chemical Engineering Education

This column provides examples of cases in which students or faculty have gained knowledge, insight,
and experience in the practice of chemical engineering while in an industrial setting. Summer interns and
coop assignments typify such experiences; however, reports of more unusual cases are also welcome.
Description of analytical tools used and the skills developed during the project should be emphasized.
These examples should stimulate innovative approaches to bring real world tools and experiences back to
campus for integration into the curriculum. Please submit manuscripts to Professor W. J Koros, Chemical
Engineering Department, University of Texas, Austin, Texas 78712.

Copyright ChE Divtsion ofASEE 1995

for the duration of the program. This requirement maximizes
the professor's informal interaction with our design and
research personnel and significantly enhances the benefits of
the program. Both the professor's research and teaching
interests are reviewed during the selection process to ensure
that a match exists between Phillips' ongoing research and
design efforts and the professor's interests and expertise.
Phillips is a worldwide leader in the production of com-
modity polymers, motor fuels, ethylene, and specialty chemi-
cals. A partial list of our current research areas is given in
Table 1. Visiting professors have access to state-of-the-art
experimental equipment, sophisticated computer-controlled
pilot plant facilities supported by a skilled operations staff,
and computing platforms ranging from high-speed PCs to
our Cray super computer. We use the latest in commercial
process simulation software in combination with in-house
and commercially available unit operations design software.
Most of our research is targeted at problems of interest to
our business units. The IEP is structured to give the profes-
sor an opportunity to visit a plant site during the summer. In
most cases the customer interested in the project is a plant
engineer and frequent communication is a necessity. This
interaction is a valuable component of the program and
gives the professor an opportunity to see how industrial
research is closely linked to business unit needs. The assign-
ments are structured to allow completion during the sum-
mer, and a report summarizing the project results is required.

We believe the program is a good start toward exposing

the professor to the world of industrial research and design.
A schematic of the program information flow is given in
Figure 1. While on the site, the participant can interact with
nationally recognized experts in polymer catalysis, thermo-
dynamics, reactor design, seismic imaging, heat transfer,
and separations technology. This interaction can significantly
enhance the background the professor brings to the
classroom. Exposure to the demands placed on newly hired
engineers provides a feedback mechanism that will help
ensure that the undergraduate curriculum remains relevant
to industrial needs. The direct communication with the
business units and the plant-site visits expose the participant
to commercial equipment operation and the constraints
of the "real world."
The program also connects the professor with industrial
researchers who are good sounding boards for research ideas.
Most funding organizations (NSF, DOE, etc) require that
proposals demonstrate industrial relevance, and in some cases
they require industrial participation. The understanding and
contacts obtained during the summer internship could be
very useful as proposal ideas are formulated. The program
could also initiate an on-going consulting relationship which
will benefit both the academic and industry. The ultimate
benefit is the broader perspective that the professor brings to
the classroom and to his research program.
Of course, the program is not without benefit to Phillips
Petroleum Company. We will be employing extremely bright,
energetic, chemical engineers who understand the latest tech-
nologies. Contact made at this stage of an academic's career
will benefit both parties for many years. Continued contact

Figure 1. IEP information flow paths

Winter 1995

Corporate Technology
Research Areas

* Advanced seismic mapping
* Polymer catalyst develop-
M Fixed bed catalysis
M Process simulation and
* Separations technology
* Polymer process develop-
* Advanced process control
* Specialty chemicals

with the academic community during this period of re-engi-
neering, restructuring, and realignment in the oil industry
will pay dividends as the industry recovers and begins large-
scale recruiting of university graduates.

Some of the program benefits are clearly pointed out in the
following example of an initial project in the IEP.

simulator. We complicated the picture significantly by
moving to a commercially available process simulator
that used slightly different reference conditions than those in
the in-house package. Dr. Gasem's assignment was to put
together a consistent data base of thermodynamic data,
with an emphasis on vapor-liquid and liquid-liquid interac-
tion parameters, for incorporation into the commercial simu-
lation package.

U Introduction U Project Execution
* Introduction

Dr. Kahlid Gasem (Associate Professor, Oklahoma State
University) worked in the Corporate Technology Petrochemi-
cal Division during the summer of 1994. While Dr. Gasem's
level of experience and industrial background is well above
that of an entering assistant professor, his project illustrates
the benefits to the IEP. A project which fully uses Dr. Gasem's
significant thermodynamic expertise was developed by our
Technology Licensing, Corporate Engineering, and Corpo-
rate Technology groups. In this case the customer was our
Technology Licensing business unit and plant visits were
not made. An effort was made, however, to ensure that Dr.
Gasem was exposed to the process that his thermodynamic
calculations would impart.

* Problem Description
The problem involved verification of physical properties
used in the design simulation of Phillips' methyl tertiary
butyl ether (MTBE), ethyl tertiary
butyl ether (ETBE), and tertiary
amyl methyl ether (TAME) pro- Mixed
cesses. An overview of the MTBE *torag
process is given in Figure 2. The s"
production of ethers for inclusionED
in gasoline has become increas-
ingly important as reformulated
fuels are mandated. Our Technol- "
ogy Licensing group has observed
a dramatic rise in requests for quo-
tations and we have recently
started several new units. ,.
Operating data from the new
units indicated that our design
procedures were meeting custo-
mer demands, but we wanted nar
to review our thermodynamic
data for consistency and to add
recently obtained data. Over a
period of several years, a variety
of thermodynamic data, corre-
lated by a variety of engineers,
had been used to build a physical
property deck for our in-house

One valuable aspect of university/industry cooperation is
the different perspective the two groups can bring to a
problem. In this particular case, we benefited significantly
from Dr. Gasem's background in the retrieval of basic
thermodynamic data. He also had significant expertise in
preparing data banks of thermodynamic information. Our in-
house work had evolved in a somewhat nonlinear fashion,
and the correlation of all the data into a rational/retriev-
able catalogue proved to be very beneficial. Literature
searches of a variety of data bases were conducted, some
familiar to Dr. Gasem and some familiar to our in-house
thermodynamics experts.
Both parties benefited from exposure to each other's meth-
ods of obtaining data. Frequent update meetings were held
with the customer groups. These meetings served to update
the funding organization on the project's progress as well as
providing Dr. Gasem with a view of the larger picture. The

Figure 2. MTBE Process

Chemical Engineering Education

Reactor 2


MTBE To Stmrage

ultimate goal of the project was to ensure accuracy in
TABLE 2 our process designs. Design factors that were impacted
Model Combinations Evaluated by the thermodynamic properties were reviewed in de-
tail to ensure that everyone understood the level of
Ideal gas-Ideal solution accuracy that was required. Input was received from a
Ideal gas-Wilson variety of groups inside the company. Each group
Ideal gas-Non-random two liquid brought specific concerns that impacted the project.
Redlich Kwong-Wilson
Redlich Kwong-Non-random two liquid N Results
Soave Redlich Kwong-Non-random two liquid Computer data files were generated for forty-two VLE
binaries involving twenty-five chemical species. Three
sets of pure-fluid physical properties were used in the
1.00- model evaluation: the Phillips' in-house data bank, a
mTBE .MET // commercial process simulator data bank, and the
Pr ~. 1.0s1 / DIPPR data base. The six model combinations evalu-
0.80- Ah.daL(m190) ated are given in Table 2. The Redlich Kwong/NRTL
Schid. ml rm model yielded adequate predictions, as did the
-- RX4NUL P. dMb
SSoave Redlich Kwong/NRTL model. Example plots of
0.eo the RK-NRTL prediction versus experimental data are
given in Figures 3 and 4. The prediction accuracy
is typical of the binaries reviewed. It should be noted
S0.40- that the use of the process simulator default NRTL
parameters produced an inaccurate estimation of the
equilibrium K value. Additional work is being under-
0.20 taken to evaluate the ability of the optimized physical
properties to predict the performance of a multicompo-
nent separations column.
0. 0 I In summary, the project was a complete success.
o.0o 0.20 0.40 0.60 0.0 1.o Phillips obtained an expanded, consistent, data base of
Liquid Mole Fraction
Liquid Mole Fraction physical properties that was inserted into the physical
Figure 3. Equilibrium phase composition for the property system. The project provided Dr. Gasem addi-
MTBE-methanol system tional exposure to the use of thermodynamic data in an
industrial setting. This exposure will enhance the infor-
1.00 mation Dr. Gasem can supply students in both the un-
TUrEa f dergraduate classroom and the research laboratory. We
Topr-SW.I I // hope this project is one of many successful interactions
o.e0- waid.v.(lnl J that develop from the IEP.
-- u1it4K4Klition
o.eo- Inquiries about the IEP can be sent to Dr. Mark
3Dreiling; Phillips Petroleum Co.; 147 CPL, PRC;
Bartlesville, OK 74004. Participation in the program
S0.40 requires the execution of a secrecy agreement and the
/ completion of all normal Phillips Petroleum Company
employment procedures. Phillips Petroleum Co. is an
0.2o equal opportunity employer.

0.00- Aim, K., and M. Ciprian, J. Chem. Eng. Data, 25, 100
0.00 0.20 0.40 0.0 0.80 1.00 (1980)
Liquid Mole Fraction
Liquid Mole Frat* Churkin, V.N., V.A. Gorshkov, and S. Yu Pavlov, Zh. Fiz.
Figure 4. Equilibrium phase compositions for the TBA- Khim., 52, 488 (1978)
ETBE system. Wilding, V., DIPPR 805 (Phillips Data Files) (1991) 0

Winter 1995






Ben-Gurion University of the Negev Beer-Sheva, 84105, Israel

Experimental data is hard to come by, and obtaining it
can often be expensive, difficult, and time-consum-
ing. But with the computational tools available to-
day, it is easy to manipulate available data in order to extract
every bit of information that exists. There are, however,
dangers in manipulating the data. If it is not done carefully,
the data can be altered without warning and unnoted, and it
will no longer provide the right information. Another danger
is overcorrelation, which is done by forcing a model when
there is insufficient or inaccurate data, trying to get informa-
tion that is not there.
How can we know what is the valid use of data and what
represents unjustified altering or overcorrelation? The an-
swer is: there is no need for complicated statistical analysis.
Common sense and judicious use of some basic statistical
and error-analysis tools provide the answer in most cases.
Most of today's students have access to programs which

Mordechai Shacham is Professor and Head of
the Chemical Engineering Department at the Ben
Gurion University of the Negev, Beer-Sheva,
Israel. He received his BSc and DSc from the
Technion, Israel Institute of Technology. His re-
search interests include applied numerical meth-
ods, computer-aided instruction, chemical process
simulation, design, and optimization, and expert

Neima Brauner received her BSc and MSc from
the Technion, Israel Institute of Technology, and
her PhD from the University of Tel-Aviv. She is
currently Associate Professor in the Fluid Mechan-
ics and Heat Transfer Department and has been
elected to President of the Israel Institute of Chemi-
cal Engineers. She teaches courses in Mass and
Heat Transfer and Process Control. Her main re-
search interests include two-phase flows and trans-
port phenomena in thin films
* Address: School of Engineering, Tel-Aviv University, Tel-Aviv,
69978, Israel

carry out linear, polynomial, and often even nonlinear re-
gression. While these programs fit the parameters to the
requested model, they will not provide guidance regarding
the ability of the model to represent the data. The critical
analysis of the model's adequacy to represent the data must
be done by the students, and we should provide them with
the tools needed to carry out this analysis.
In this paper, we will use an example from the area of
heterogeneous reaction rate data analysis to demonstrate
some potential pitfalls in data correlation. Churchill noted
that, "The published correlations for reactions catalyzed by
solids provide many examples of overcorrelation."11l The
example we will be using involves verification of a rate
expression for platinum catalyzed oxidation of sulfur diox-
ide. This example is from Smith's textbook[21 on chemical
engineering kinetics.

The data of reaction rate versus partial pressures of SO2,
SO3, and 02 in oxidation of sulfur dioxide using a packed
bed of platinum-on-alumina catalyst pellets at 4800C is shown
in Table 1. The data is from Example 9.2 in Smith. The data
shown in the table is actually part of a larger set of data
which was published by Olson, et al., in 1950.'31
Based on a postulated reaction mechanism, Smith devel-
oped the following rate expression for this reaction.

PS, p/2 (1/ K)pS0
r= 2 3
A + Bpso3

where K is the equilibrium coefficient (K = 73 at 4800C) and
A and B are constants to be calculated by regression of the
experimental data. Smith calculated the constants using lin-
Copyright ChE Division ofASEE 1995
Chemical Engineering Education

ear regression with the linearized form of Eq. (1)

Pso2 P2 (1/ K)pso j1
-2 = A + BpSO3

Since the unknown coefficients in Eq. (2)
in linear form, they can be calculated either
using linear regression or by plotting the expr
sion on the left-hand side of this equation ver
Pso, and fitting a straight line. The coeffici
values obtained by Smith using this proceed
are A = 0.176 and B = 12.9. The test for feasil
ity of a particular mechanism to represent
data, when a Langmuir-Hinshelwood type r
expression (such as Eq. 1) is used, is that
coefficients of the proposed model must be pc
tive. In this case, the coefficients (A,B) are
deed positive; hence, the proposed rate expr

Reaction Rate Data for
Sulfur Dioxide Oxidation at 4800C
(from Smith21)

r Partial pressure (atm) at catalyst surface
g mol
(h)(g catalyst) SO, SO, 02

0.02 0.0428 0.0255 0.186
0.04 0.0331 0.0352 0.190
0.06 0.0272 0.0409 0.193
0.08 0.0236 0.0443 0.195
0.10 0.0214 0.0464 0.196
0.12 0.0201 0.0476 0.197





We have recalculated the parameter values using nonlinear regression
and the original Eq. 1. The calculation was done using the POLYMATH'14
(2) package. The parameters obtained using nonlinear regression (including
confidence intervals) are
are A = 0.1017 0.0958
by B= 16.02 4.33
sus These values are also positive. A comparison of the experimental and
ent calculated reaction-rate values is shown in Figure 1.
ure It can be seen that the fit between the calculated and experimental values
bil- is very good. As an additional check, the sum of squares of errors, S2, can
the be calculated
ate 6
all S = (ri.exp- ricalc
si- i=l
es- For the parameters obtained using nonlinear regression, S' = 5.226 x 10 ,
a small number, indicating that indeed the fit between the experimental
data and the calculated values is acceptable.
When all the indicators show that the fit is acceptable, does it ensure that
Eq. (1) is the right rate expression to represent the reaction under the given
conditions? Churchill warns that, "One should be wary of accepting the
validity of a model merely because it is successful in correlating the
data..." So what else should be checked?

Is the data really experimental data?
Looking at the data in Table 1 may have led one to suspect that this data
does not represent the real measured values. The reaction rate, which is the
dependent variable, appears in nice round numbers. While the value of the
independent variable can often be set to rounded numbers, which are more
convenient to work with, the measured (or calculated) value of the depen-
dent variable will usually contain at least as many decimal digits as
provided by the measuring instruments.

2 Regression data
o Calculated value
(A=0.176, B=12.9)
A Calculated value
8 (A=0.1017, B=16.02) 9


0 1 2 I
0 1 2 3 4 5 6

I Point No.
Figure 1. Experimental and calculated reaction rate, using Eq. (1)
as rate expression
Winter 1995

There is an explanation for the round numbers
that represent the reaction rates in Table 1. Smith
noted that the data was "interpolated for even inter-
vals of reaction rates...." But when the original
source of the data is consulted (Olson, et al.), one
finds that this was not the only interpolation per-
formed. Due to difficulties in controlling and mea-
suring the temperature, none of the experiments
was actually carried out at 4800C. The data in Table
1 is probably based on four measurements made at
mean bed temperatures of 461 C, 467C, 4820C,
and 488C. The results at these temperatures were
again interpolated to 480'C. In 1950, interpolation
meant putting the experimental data points on a
graph paper and manually fitting the "best" curve.
Hence, the sophisticated regression analysis is not
carried out on the real measured data, but rather on
a curve that was drawn freehand.

1 4

sion is a feasible one.

While we must realize that this was perhaps the best that
could be done with the calculating tools that were available
forty years ago, the approach is totally unacceptable today.
Interpolation and extrapolation introduce changes in the er-
ror distribution of the data. Statistical analysis of regression
models is based on certain assumptions regarding the error
distribution. Therefore, with such smoothed and extrapo-
lated data, the application of statistical analysis for evaluat-
ing the validity of regression models is meaningless.

Can the rate data be represented better by a different model?
Since the equilibrium coefficient value is K = 73, it seems
that the reversibility of the reaction has very little signifi-
cance at this temperature. Actually, it can be seen that the
contribution of the reverse reaction (in the numerator of Eq.
1) is smaller by two orders of magnitude than the contribu-
tion of the forward reaction. This possibly suggests that the
expression related to the reversible reaction can be omitted
from the rate expression. Thus, we may try to find out the
coefficients for the following rate expression:

r = s2 2 (3)
(A+ Bpso )

The calculated coefficients using nonlinear regression are
A = 0.094 + 0.093
B = 16.53 + 4.20
S2 = 4.8 x 10-5
The coefficients are positive and the sum of squares of errors
is smaller than that obtained for Eq. (1). Thus, Eq. (3) is a
valid rate expression and represents the data even better than
Eq. (1). Is it the best rate expression?
Brauner and Shacham151 have shown that the parameters of
a power-law rate expression can be used to discriminate
between feasible and infeasible mechanism-based rate
expressions. For the reaction considered here, a power-
law model reads
r = kp so PC (4)
so3 PSo2 02

By denoting amin the lowest power of pso in the
expression obtained for 1/r in a particular feasible
mechanism-based model and amax its highest power
(according to the criteria developed by Brauner and
Shacham'51), the following inequality should hold:

amin <-a < amax (5)
Similar criteria are valid for the powers of pso, and
po The following results are obtained using nonlin-
ear regression for the power-law model:
k = 0.517 113.3

a = -1.98 7.02
b = -0.216 4.556
c = 6.078 124.7
S2 = 1.85 x10-

It can be seen that S2 is considerably reduced compared to
the values obtained in the two previously considered rate
expressions, indicating that the power-law correlates the
data much better than the previously suggested models.
We could proceed and postulate different feasible mecha-
nism-based models using Eq. (5) and similar inequalities for
the powers of the various partial pressures. But we should
first pay attention to the huge uncertainty in the constants of
the power-law model. The 95% confidence interval in k, for
example, is larger by more than two orders of magnitude
than the value of k itself. At this point, the statistical infor-
mation on confidence limits becomes very important. Too
much uncertainty in the parameter values may indicate that
one or more of the variables should not be included in the
Since the uncertainty is the largest in the parameter c (the
power of po,), we can try first to remove p,, from the
correlation by setting c = 0. The results obtained for this
case are
k = 1.217 x 10-5 + 1.268 x 103
a = -2.308 1.23
b = -0.052 1.87
S2 = 1.89X 10-5
The value of S2 is changed very little by removing po,
from the correlation, but the confidence limits still remain
too wide. This time the parameter b is associated with the
largest uncertainty, so we may try now to also remove pso,
from the correlation by setting b = 0. The results obtained
now are
k = (1.621.37) x10-
a = -2.270 0.222
S2 = 1.9 x105
The plot of the calculated curve and the experimental data

2 2.5

3 3.5 4 4.5
P So 102

Figure 2. Calculated curve and experimental data versus
PS03 with power-law expression
Chemical Engineering Education

versus pso are shown in Figure 2.
All the indicators (plot, S2, and confidence limits) show
that so far this equation is the best to represent the data in
Table 1. It should be emphasized, however, that this equa-
tion cannot be extrapolated outside the region where the
measurements were made; otherwise, absurd rate values
may result. For instance, at the start of the reaction when
products' concentration is zero, this equation predicts an
infinite reaction rate.
The poor asymptotic behavior of the power-law rate ex-
pression, obtained in this case, demonstrates very clearly the
disadvantage of empirical models compared to mechanism-
based models, which are derived based on physical insight.
The power-law parameters, however, can be used for postu-
lating mechanism-based rate expressions.

Summary of Fitted Parameter Values

Langmuir-Hinshelwood Type Model
Eq.# A B S2
1 0.102 0.0958 16.02 4.33 5.23 x 10-5
3 0.094 0.093 16.53 4.2 4.8 x 10-5
6 0.586 0.207 72.49 9.35 2.56 x 10i5

Power-Law Model
k a b c
0.517 +113.3 -1.98 7.02 -0.2164.556 6.078 124.7 1.8
(0.0121 1.27)x 10-3 -2.308 1.23 -0.052 1.87 0 1.8
(1.62 1.37)x 10- -2.27 0.222 0 0 1.

PS2- 102

Figure 3. Linear dependency between ps and ps
in the experiments. 3 2

Winter 1995

Inequality Eq. (5) with the power-law constants obtained
above yields amin <2.27 < amax Using this inequality, sev-
eral rate expression of the form of Eq. (3) can be postulated.
One possibility is

r -A + 3

A^+ BpSO

Nonlinear regression with Eq. (6) yields
A = 0.586 +0.207
B = 72.49 9.35
S2 = 2.56x 105

The plot of calculated curve and experimental data versus
partial pressure of SO3 for this correlation is practically
identical to Figure 2, and thus this model also represents the
data excellently.
The results for the different models are summa-
rized in Table 2. They strongly suggest that the only
variable to be included in the correlation of the data
is Pso,. This is definitely true for the numbers ap-
pearing in Table 1, but will it hold in general for the
sulfur dioxide oxidation reaction?

Why does the reaction rate depend only on Pso ?

Statistical analysis has shown that for the data of
Table 1 the reaction rate does not depend on either
Pso, or po,. But before generalizing this conclu-
S2 sioi, we have to seek the answer to two questions
5 x 10o5 concerning experimental design:
9 x 10-5 1. Were these variables altered during the
Sx 10. experiments to cause significant differences in the
reaction rate?
2. Were the variables altered independently so
that the data includes independent information
with regard to their effect on the reaction rate?

Inspection of the partial pressures in Table 1
reveals that the relative variation of p,0 is much
smaller than that of pso and Po The maximal
change in po, is less than 6%, while the changes in
the other variables are of the order of 100%. But
when considering the absolute change, it is the
same order of magnitude for all the independent
variables. Thus, the answer to the first question
above is inconclusive.

5.0 To check the assumption that the partial pressures
of the various components were altered indepen-
dently, plots of one independent variable versus the
others are prepared. The resulting correlation be-
tween pso, and so is shown in Figure 3.

Continued on page 45.



peO classroom





Purdue University West Lafayette, IN 47907-1283

his module, simulating a cryogenic hydrogen reac-
tive cooling process, is one of a series of computer
simulations initiated by a grant from NSF in 1988.
Computer simulations were chosen to supplement the chemi-
cal engineering laboratory facilities for a variety of reasons
Size, complexity, and safety of processes will not
preclude their use.
Realistic time and budget constraints can be built into
the simulation, giving the students a taste of "real
world" engineering problems.
Emphasis of the experiment can be shifted from the
details of operating a particular piece of laboratory
equipment to the more general considerations of
proper experimental design and data analysis.
The time duration of running simulations and obtain-

S. Jayakumar is a visiting instructor in the School of Chemical Engineer-
ing at Purdue University with a consulting appointment at Eli Lilly and
Company. He received his BTech from Indian Institute of Technology,
Madras, in 1985, and his MS and PhD degrees from Purdue University in
1988 and 1992, respectively. His research interests include process
design, simulation, optimization, plant layout, and educational applica-
tions of computer technology.
R.G. Squires is a professor of chemical engineering at Purdue Univer-
sity. He received his BS from Rensselaer Polytechnic Institute in 1957
and his MS and PhD from the University of Michigan in 1958 and 1963,
respectively. His current research interests center on the educational
applications of computer simulation.
G. V. Reklaitis is Head of the School of Chemical Engineering at Purdue
University. He earned his BS from Illinois Institute of Technology in 1965
and his MS and PhD from Stanford University in 1969. His research
interests include process systems engineering, process scheduling meth-
odology, and the design and analysis of batch processes.
Kimberly S. Grassi is a Senior Principal Process Engineer at Air Prod-
ucts and Chemicals, Inc. She received a BS from Purdue University in
1980 and a MS from Lehigh University in 1986. With Air Products since
1980, she has worked primarily in cryogenic processing of hydrocarbons
from natural gas and refinery streams, and in air separation.

*Address: Air Products and Chemicals, Inc., Allentown, PA

ing data is greatly reduced. (A batch reactor experi-
ment, which may take three hours in a laboratory, can
be done within minutes on a computer.)
Computer simulation is relatively inexpensive com-
pared to the cost of building, operating, and maintain-
ing complex experimental equipment.
Simulated experiments take up little laboratory space
and can serve large classes, the same computer can
run many different simulations, and several worksta-
tions can run the same simulation simultaneously.
Previously developed modules have been discussed in
several publications.11 Each module consists of the follow-
ing components:
1. Video Tape: A twenty-minute video of the actual process,
furnished by the company, is used to introduce the project.
This typically contains a brief discussion of the appropriate
theory and background, a "visit" to the bench scale and pilot
facilities, a "tour" of the plant, and an overview of how the
process fits into the overall company operations. The aim is
to reinforce the fact that the project is not merely a computer
game but is based on a real-life state-of-the-art process
2. Problem Statements: There are two separate problem
statements. The experimental section requires the students to
design and run experiments to determine the characteristic
model parameters. The application section involves using
these parameters in the simulation of a more complex
application such as process optimization, start-up of a new
plant, or scale-up. The problem statements are printed on
company stationary to add to the sense of an industrial
project. In addition, budget and time constraints are imposed.
3. Background Information: A description of the theory
behind the processes, and details about the experiments and
the analysis necessary to execute the project, are provided.
4. The Computer Module: This is a software package, written
Copyright ChE Division ofASEE 1995
Chemical Engineering Education

Although computer simulations may offer a number of advantages
over experiments, the experience (and frustrations!) of running traditional experiments
in the laboratory is an important component of chemical engineering education. For this reason, we
require that at least two of the three projects in our senior laboratory course be real experiments.

in the C language, that the students must use to run both the
experiments and the application problem. It uses some IMSL
routines for random number generation, solution of a system
of nonlinear algebraic equations, and numerical integration.
The module can run on any machine that supports the X
Window System. At Purdue it runs on Sun Sparc worksta-
tions with 12 MB of memory. Each module uses less than 10
MB of disk space.
Although computer simulations may offer a number of
advantages over experiments, the experience (and frustra-
tions!) of running traditional experiments in the laboratory is
an important component of chemical engineering education.
For this reason, we require that at least two of the three
projects in our senior laboratory course be real experiments.
In our curriculum, groups of three students are given eight
three-hour laboratory periods (four weeks) to complete the
project. The first two periods are used for planning the
experiments, followed by a conference with the professor.
Experiments and analysis are done during the next three
periods, followed by a fifteen-minute oral presentation which
is video-recorded for a later private discussion with a com-
munication specialist. The application problem is solved in
the last three periods, and a final report is due a week later.
The output values of students' data are randomly perturbed
to simulate experimental error, based on variation param-
eters set by the instructor. In addition, a separate computer
"simulation" of the process (without statistical fluctuations)
is available to the students for validating their parameter
estimates. To use this simulation, the students must first
enter the estimated parameters. The students need only run
this "simulation" at conditions identical to those used in
running the experiments. Results within experimental error
would indicate reasonable estimates of the parameters.
The Air Products
Hydrogen Reactive Cooling Process

* Background
At first glance, the liquefaction of hydrogen appears to be
similar to that for nitrogen, where the gas is compressed,
refrigerated, further cooled by heat exchange to a tempera-
ture near its boiling point, and finally partially condensed by
means of an expansion through a throttling valve (or an
expansion engine). The liquid product is transported to cus-
tomers or sent to storage.
But the hydrogen liquefaction process is complicated by
the fact that hydrogen exists in two forms: ortho and para. As
shown in Figure 1, the equilibrium mixture is about 25%
para-hydrogen at room temperature and nearly 100% para-
Winter 1995

hydrogen at liquid hydrogen temperatures.
Therefore, if we liquefy hydrogen as described above, the
composition of hydrogen leaving the process would be al-
most the same as that entering-25% para-hydrogen. In the
storage tank, this liquid hydrogen will tend to react to reach
100% para-hydrogen (the equilibrium composition at this
low temperature). These are shown by the dashed lines in
Figure 1. Since this ortho- to para-hydrogen reaction is exo-
thermic and the heat of reaction is of the same order of
magnitude as the heat of liquefaction, much of the stored
hydrogen will vaporize.
In order to avoid this boil-off of liquid hydrogen during
storage, it is necessary to accomplish conversion of ortho
hydrogen to the para form as it is being cooled down so that
the stored liquid hydrogen is already near chemical equilib-
rium, minimizing any further reaction.
Our reactive cooling process achieves this aim: the cool-
ing and chemical reaction take place in a single tower (see
Figure 2a). Of course, this places additional cooling require-
ments on the tower since both the sensible heat and the heat
of the exothermic reaction must be removed from the cool-
ing hydrogen stream.
Figure 2b shows the warm and cold streams in one section
of the tower, with different points in the streams numbered.
After the warm stream is cooled (2), it is passed through an
adiabatic converter in which the stream is heated to T, by the
exothermic heat of reaction, hence T, > T,. Optimal design
considerations require that this stream which exits the reac-
tor be returned to the heat exchanger (we call this the recycle


C 90

2 80
= 70

a 60
50 '

30 Cool
0 60 120 180
Temperature (K)

Figure 1. Equilibrium composition of para-hydrogen.

point) at the location at which T3 = T,. Then, since inlet
temperatures, flow rates, and heat transfer areas for the stream
between points 1 and 2 are identical to those between points
3 and 4, the temperatures of these two streams must be equal
at same elevations in the column. In particular, T2 = T4.
Note that (a) the enthalpy gained by the cold stream in the
adiabatic heat exchanger must equal the enthalpy lost by the
warm stream, and (b) the effective flow rate of the warm
stream is twice as much below the recycle point as it is
above the recycle point. This results in a temperature-en-
thalpy diagram for the reactive heat exchanger cooling tower
as shown in Figure 3.

* The Problem Statement
The students are told that the supplier of the compact heat
exchanger modules has developed a new welding technique
that will enable them to redesign the module to make it
more efficient. The supplier has furnished a prototype mod-
ule. Since the module must be redesigned, the students
are told that this is an appropriate time to also incorporate
the use of Air Products' newly developed ortho-para
hydrogen catalyst.
In order to properly design the tower, the students must
first plan and perform a set of laboratory experiments to
The heat transfer coefficients of the new heat exchanger
The rate constants of the reaction over the new catalyst as a
function of temperature.

* Heat Transfer Experiments

late to the much lower temperatures and higher flow rates
used in the plant.

* Reaction Experiments
Laboratory experiments must also be designed to deter-
mine the rate constants of the ortho-para hydrogen reaction
as a function of temperature. The kinetics are complicated
by mass transfer effects. Since we are interested in the cata-
lyzed reaction, we might expect the reaction rate to be corre-
lated in terms of the surface concentration of the adsorbed
species C, and a surface reaction rate constant k,. For ex-
ample, the rate of the forward reaction might be
r = ks Cs
But since it is very difficult to measure the surface concen-
tration of the adsorbed species, we use an empirical reaction
rate model based on the para-hydrogen concentration in the
gas phase, an apparent rate constant, k, for the forward
reaction and equilibrium compositions. The analysis assumes
The temperature difference between the bulkflowing
hydrogen stream and the surface of the catalyst particles is
negligibly small.
The temperature difference between the interior and
exterior of the catalyst particles is also negligible.
The concentration gradients between the bulk gas and
catalyst particles surface are considered negligible.
The concentration gradients between the surface and the
interior of the catalyst pellet are taken into account by
means of a pore diffusion effectiveness factor.
For a first-order reversible rate law, written in terms of the
bulk concentrations for an isothermal plug flow reactor, the
final integrated expression is

Neglecting heat transfer resistance through the metal par- F
titions, the overall heat transfer coefficient based on the cold W
side, U, for the heat exchanger can be correlated by


Ce n Ci -Ce
Co -C
C -Ce

1 1 1
=-- +-
UAc hcAc hwAw

For many different types of heat exchanger surfaces (e.g.,
circular tubes, plain plate-fin surfaces), the heat transfer
coefficients, h, have been correlated in the form

Nst(Npr)2/3 = a(NRe)b

h Cpu pvdH
NSt Np, = c NRe
pvCp k t

Nst(N,)2/3 is also known as the Colburn j jactor for heat
transfer. Students must design a series of experiments (using
the prototype module) in which the countercurrent flow rates
and temperatures (80 120 K) of hydrogen can be varied
within specified limits to determine the heat transfer coeffi-
cients. The above correlations can then be used to extrapo-

Figure 2. (a) Reactive cooling tower; (b) Expanded view of
one heat exchanger-reactor combination unit.
Chemical Engineering Education

Tot Twin

Heat Exchanger



Tcin Two t
(a) (b)

where k is the rate constant corresponding to the conversion
of ortho- to para-hydrogen, and the effectiveness factor is

3( 1 1 (coth )- 1
p tanh D (DD 2

D is the dimensionless Thiele modulus, which in turn is
given by

dp [ kppRT
2 MDeff P

Students are provided with an isothermal plug flow reactor
and are expected to design experiments in which the flow
rates and temperatures may be varied and the compositions
of the inlet and exit streams are measured. This data would
allow the estimation of the two parameters in the equation
k = ke-E/RT. An iterative calculation scheme may be used
to account for the effectiveness factor.

* Budgetary Limitations
Each student group is given a $35,000 budget to complete
the laboratory experiments. Liquid hydrogen and nitrogen
cost $10,000 and $5,000, respectively. In addition, $700 per
reaction experiment and $500 per heat exchanger experi-
ment are also charged against their budget. A consultation
fee of $500 per session is charged (after the initial free
introductory lecture). The data which is collected includes
statistical fluctuations. Hence, the students are expected to
perform replicate runs as necessary to report confidence
limits on the constants they estimate.
The students should complete this part of the project by


0. -

Ia- R
F- R

Unit 1 Unit 2 Unit 3


Figure 3. Reactive heat exchanger cooling curve for
three units in series.
Winter 1995

the fifth (of eight) laboratory periods. Then they must present
their results during a fifteen-minute oral presentation, after
which they use the estimated parameters to design the reac-
tive cooling tower (during the last three laboratory periods).

* Optimal Reactive Cooling Tower Design
The reactive cooling tower is a series of vertically ar-
ranged units, as shown in Figure 2a. Each unit is comprised
of a heat exchanger and a converter (Figure 2b). The design
variables of interest are the number of units, the length of
each heat exchanger, and the amount of catalyst in each
converter. The annualized cost of the tower as a function of
design variables and parameters is provided.
Referring to Figure 2b, after ortho-to-para conversion in
the reactor, the hydrogen stream is fed into the heat ex-
changer (at the recycle point) and is further cooled before
exiting from the heat exchanger. Thus, even if the reaction
were to reach equilibrium at the recycle point, the last sec-
tion of cooling (3 to 4) forces the stream temperature away
from the equilibrium temperature. Hence, one can see that
the longer the individual exchanger units, the further away
the final composition will be from equilibrium. This will
lead to greater H, boil-off during storage.
Using a larger number of shorter units may be an alterna-
tive. Every heat exchanger, however, has 1.1 feet of header
that distributes the flow properly throughout the unit, but in
which no heat transfer takes place. Thus, if a tower 21 feet
high is designed with 10 units, each 2.1 feet high, only one
foot of each unit (or a total of 10 feet of the tower) is
available for heat transfer. A large number of short heat
exchanger units will, therefore, result in closer approach to
equilibrium (and consequently less H, boil-off) but with
higher heat exchanger costs.
Since the total available heat transfer area in the column is
relatively less for this configuration, a cold stream inlet
temperature of 20 K may not be realizable in the simulation.
These competing trends will result in an optimal design. For
the optimal design, the students must determine the number
of units, length of each unit and the amount of catalyst in
each reactor, for fixed warm- and cold-stream flow rates and
inlet temperatures, 98% equilibrium in each reactor, and a
tower height limitation of 21 feet. A screen-dump of a typi-
cal reactive cooling simulation is shown in Figure 4.

Each section normally has twenty-four students working
in eight groups of three students. Each group is required to
complete three one-month-long projects, of which one, at
most, may be a computer simulation of the type described in
this paper. The three students in each group are assigned to
function as either group leader, experimental engineer, or
design engineer, and during the semester each student serves

........ Cold Side
- Warm Side
R Recycle Point

once in each capacity. A summary of our grading system is
given in Table 1.
Please note that our Communication Specialist, Dr. Frank
Oreovicz, assigns grades for report writing and prevention
style. These two grades, reflecting a student's communica-
tion skills, add up to 150 out of the leader's 550 points for the
project. If the leader does not pass in the communication
skills category, he or she fails the project regardless of the
quality of technical work.

The projects in this laboratory clearly meet the ABET
design criteria. In particular, the Air Products project is an
open-ended problem that requires students to make deci-
sions in designing their own laboratory experiments. Within
a carefully estimated limited budget, students must decide
on the type, number, and operating conditions of each ex-
periment. Safety, environmental, and other issues are cov-
ered in the opening lecture by the professor. The data col-
lected from the computer includes statistical fluctuations
which must be properly handled by the student. The project
requires the students to demonstrate their abilities to work in
a team and to communicate both orally and in writing.

The Air Products module was used for the first time in the
spring 1994 at Purdue University and is also undergoing a
test at Georgia Institute of Technology. To date, at least one
of the Purdue-Industry modules has been used at twenty-five
schools, and the response has been encouraging. Apart from
the undergraduate design laboratory, the modules have also
been used successfully in kinetics, reactor design, and pro-
cess control courses at Connecticut, Carnegie-Mellon, West
Virginia Institute of Technology, and Utah (see, for ex-
ample, reference 6).
The module is written in the C and FORTRAN languages,
with graphics based on the X Window System. At Purdue,
the software runs on Sun Sparc workstations. A project to
port these modules to other Unix platforms like HP, DEC,
IBM, and Silicon Graphics has recently been funded by the
National Science Foundtion. This work is expected to more
than double the current usage. Porting of the software onto
personal computers is also being considered.

Interactive multimedia is making inroads into the educa-
tional arena, [78 although the power of "true" multimedia has
not been well exploited. Work is currently progressing on
the development of an integrated multimedia application
involving a coffee decaffeination process. Some new fea-
tures include integration of animation, video, pictures, and
audio into the simulation in order to present realistic indus-
trial equipment operation, measurement, and safety details.

This three-year project is sponsored by Procter and Gamble.

The display of the commercial reality of the processes via
video technology, the ease of its presentation and use, its
applicability to both laboratory and lectures, and a true-
to-life simulated budget and time constraints, together
with other aspects, make projects of this kind very effective
in the chemical engineering curriculum. Although these
modules have been used in chemical engineering, the
ideas presented in this paper are equally applicable to all
engineering domains.

This work was supported by Air Products and Chemicals.

Entalpy (Milko a)n X,,,)te

Figure 4. A typical simulation of the reactive cooling unit
for hydrogen liquefaction. The simulation corresponds to
five heat exchangers, each 3.82 ft., and about 100 lbs. of
catalyst in each adiabatic converter. The cold stream inlet
temperature is 20K, which is a requirement of the validity
of the design.

Grading System

Project Experimental Design
Leader Engineer Einner
Planning conference 50 50 50

Written Report: (Technical)
Leader's Section 100 25 25
E.perimenial Engineer's Section 100 100 25
Design Enqgneer's Secnon 100 25 100
ToLal 300 150 150

Written Report: Writing 100 100 100
Oral Report: Technical 50 25 25
Oral Report: Presentation 50

Total Points 550 325 325

Chemical Engineering Education

a,b Colburn j factor parameters
A Heat exchange area
C Para-hydrogen mole fraction
c Heat capacity
dH Hydraulic diameter of one side of heat ex-
dp Catalyst particle diameter
Dff Pore diffusivity inside catalyst pellet
E Activation energy
E Pore diffusion effectiveness factor
F Flow rate
h Heat transfer (film) coefficient
k ortho-para-hydrogen rate constant
ko Preexponential (frequency) factor
k, Thermal conductivity
M Molecular weight of H,
Nst,N ,NRe Stanton, Prandtl, and Reynolds numbers
P Pressure
R Universal gas constant
T Temperature
U Overall heat transfer coefficient
v Linear flow velocity
W Mass of catalyst
r Density (unsubscripted -*Hydrogen)
g Viscosity
D Thiele modulus

Subscripts and Superscripts
c Cold side of heat exchanger
e Equilibrium
i Inlet
o Outlet
p Catalyst particle
s Surface of the catalyst
w Warm side of heat exchanger

1. Squires, R.G., P.K. Andersen, G.V. Reklaitis, S. Jayakumar,
and D.S. Carmichael, "Multi-Media Based Educational Ap-
plications of Computer Simulations of Chemical Engineer-
ing Processes," Comp. Appns. Eng. Ed., 1(1), 25 (1992)
2. Squires, R.G., G.V. Reklaitis, N.C. Yeh, J.F. Mosby, I.A.
Karimi, and P.K. Andersen, "Purdue-Industry Computer
Simulation Modules: The Amoco Resid Hydrotreater Pro-
cess," Chem. Eng. Ed., 25(2), 98 (1991)
3. Jayakumar, S., R.G. Squires, G.V. Reklaitis, P.K. Andersen,
and L.R. Partin, "Purdue-Industry Chemical Engineering
Laboratory Computer Module: 2. Eastman Chemicals Reac-
tive Distillation Process," Chem. Eng. Ed., 27(2), 136 (1993)
4. Jayakumar, S., R.G. Squires, G.V. Reklaitis, P.K. Andersen,
B.C. Choi, and K.R. Graziani, "The Use of Computer Simu-
lations in Engineering Capstone Courses: A Chemical Engi-
neering Example-The Mobil Catalytic Reforming Process
Simulation," Int. J. Eng. Ed., 9(3), 243 (1994)
5. Jayakumar, S., R.G. Squires, G.V. Reklaitis, P.K. Andersen,
and B.K. Dietrich, "The Purdue-Dow Styrene-Butadiene Po-
lymerization Simulation," J. Eng. Ed., in press (1995)
6. Cutlip, M.B., "Use of the Purdue-Industry/NSF Laboratory
Modules Within the Chemical Reaction Engineering Course,"
presentation at Annual AIChE Meeting, Miami, FL; No-
vember (1992)
Winter 1995

7. Bailey, H.J., and N.E. Thornton, "Interactive Video: Innova-
tive Episodes for Enhancing Eduction," Computer Appns.
Eng. Ed., 1(1), 97 (1992)
8. Meyer, D.G., "The Videojockey System: A Testbed for Cost-
Effective Multimedia Instructional Delivery," presentations
at 1992 Frontiers in Education Conference, Nashville, TN 0

REVIEW: Linear Control Systems
Continued from page 16.
pages) in addition to the detailed solutions (seventy-one pages).
Chapter 1 provides a concise review of the necessary mathemati-
cal background, including linear algebra, random sequences, matri-
ces over rings, and matrix equations over rings. Thirty-two prob-
lems illustrate these concepts.
Chapter 2 reviews discrete-time linear systems theory, including
converting differential equation models to discrete-time models.
Included are the standard concepts of reachability, controllability,
observability, constructability, stabilizability and detectability, in-
put/output (transfer function) models, and invertibility. A number
of the twenty-four problems provide nice illustrative examples of
the techniques. Control problems include cattle population, inven-
tory, ball and beam, antenna, aircraft, and stirred-tank concentra-
Chapter 3 introduces state feedback (including pole placement),
deadbeat control, and the linear quadratic regulator. Results for
both state-space and transfer-function models are presented. There
are nine illustrative examples. A number of the twenty-two prob-
lems are extensions to problems in Chapter 2.
Chapter 4 covers state estimation, including the linear quadratic
predictor (Kalman filter) and offers six illustrative examples and
eighteen problems. Chapter 5 presents output feedback (including
the linear quadratic compensator) in both state space and input/
output forms. There are eight illustrative examples and twenty
Chapters 3 through 5 cover control system design techniques
which take an initial state to a desired final state. Chapter 6 devel-
ops control designs which achieve desired closed-loop relation-
ships between setpoints or load disturbances and outputs. There are
thirteen illustrative examples and twenty problems.
Faculty who are teaching advanced process control courses may
want to use portions of this book for background material, particu-
larly for the concise review of discrete-time linear system theory
concepts. A number of the problems can serve as nice illustrations
of the theory, in addition to the homework and example problems. I
doubt, however, that any chemical engineering faculty will want to
adopt this as a textbook since most advanced process control courses
devote less than one quarter of a semester to the material covered in
this text. Indeed, most courses cover the continuous time version of
this material.
Graduate students taking optimal control courses may wish to
use this text for supplementary material. The book should be avail-
able in any well-stocked university library but will not be necessary
for most personal libraries. Despite the desires of the author, this
text will not meet the needs of the practicing process control engi-
neer. It will also be of limited use to faculty and students conduct-
ing process control research since important topics such as con-
straints and robustness are not covered at all. 0

Random Thoughts...



North Carolina State University
Raleigh, NC 27695-7905

K, here's the scenario. You go to a teaching work-
shop presented by Woods or Wales or Stice or Smith
or that joker from North Carolina who's always
ranting about this stuff. The presenter instructs you to im-
merse your students in real-world problems without rou-
tinely providing all the requisite facts and formulas. He also
tells you-repeatedly-to stop doing so much lecturing and
instead get the students to work in teams and teach each
other. Once they realize they can no longer count on you to
tell them all they need to know, they'll start to rely on
themselves to figure it out-which is to say, they will learn
to learn.
Whether the instructional approach being promoted in the
workshop is called guided design, problem-based learning,
cooperative learning, 4MAT, or whatever, it's based on the
reasonable premise that students learn more by doing things
than by watching lectures. The presenter cites hundreds of
studies showing that compared to traditional lecturing, ac-
tive/cooperative learning leads to deeper understanding, im-
proved attitudes toward the subject, and greater self-confi-
dence. It all sounds like just what you've been looking for to
counter the apathy and poor performance that have charac-
terized an uncomfortably high percentage of your students
lately. You leave the workshop fired up and ready to switch
to the new approach in your very next class.
You may be in for a rude shock. It's not that the methods
don't work-they do. I've had great success with some of
them, particularly cooperative learning, and I do my fair
share of missionary work on their behalf. The success is
neither immediate nor automatic, however, and the awk-
wardness and frustration and student resistance and hostility
you may experience before you get to the payoff can be
formidable. It's tempting to give up in the face of all that, and
many instructors unfortunately do.
The problem is that doing anything new and nontrivial
always involves a learning curve, and the curve may be
particularly steep for both you and your students when you
try an active learning approach for the first time. The stu-

dents, whose teachers have been telling them everything
they needed to know from the first grade on, don't appreciate
having this support suddenly withdrawn, and complaints
like "Meachley never teaches us anything-we have to do it
all ourselves" start echoing through the corridors. It's even
worse if you use cooperative (team-based) learning; students
then gripe loudly and bitterly about other team members not
pulling their weight or about being slowed down by having
to explain everything to that lemon they've been forced to
team with. Sometimes instructors who are effective lecturers
get lower student ratings when they start using active and
cooperative learning methods.
My goal here is to assure you that these initial glitches are
both common and natural and that they may be a cause for
concern but not for panic or discouragement. The trick is
knowing how the process works, taking a few precautionary
steps to smooth out the bumps, and waiting out the inevi-
table setbacks until the payoffs start emerging.
Consider the students. Woods[' observes that students
forced to take major responsibility for their own learning go
through some or all of the steps psychologists associate with
trauma and grief: (1) Shock: "I don't believe it-we have to
do homework in groups and she isn't going to lecture on the
chapter before the problems are due?" (2) Denial: "She
can't be serious about this-if I ignore it, it will go away."
(3) Strong emotion: "I can't do it-I'd better drop the course
and take it next semester" or "She can't do this to me-I'm
going to complain to the department head!" (4) Resistance
and withdrawal: "I'm not going to play her dumb games-I
don't care if she fails me." (5) Surrender and Acceptance:

Richard M. Felder is Hoechst Celanese Pro-
fessor of Chemical Engineering at North Caro-
lina State University. He received his BChE from
City College of CUNY and his PhD from
Princeton. He has presented courses on chemi-
cal engineering principles, reactor design, pro-
cess optimization, and effective teaching to vari-
ous American and foreign industries and institu-
tions. He is coauthor of the text Elementary
Principles of Chemical Processes (Wiley, 1986).

Copyright ChE Division ofASEE 1995
Chemical Engineering Education

"OK, I think it's stupid but I'm stuck with it and I might as
well give it a shot." (6) Struggle and exploration: "These
other guys seem to be getting this stuff-maybe I need to try
harder or do things differently to get it to work for me." (7)
Return of confidence: "Hey, this is really working. I don't
understand why I had so much trouble with it." (8) Integra-
tion and success.
Just as some people have an easier time than others in
getting through the grieving process, some students may
enthusiastically dive right into active learning and short-
circuit many of the eight steps, while others may have diffi-
culty getting past the negativity of Step 3. The point is to
remember that the resistance you encounter from some stu-
dents is a natural part of their journey from dependence to
intellectual autonomy, and if you provide some help along
the way, sooner or later most of them will make it.
So what can you do to help them and yourself get through
the process? Out of painful necessity,* I've developed an
arsenal of strategies. For whatever they may be worth, here
they are.

C Set the stage. When I plan to use active or cooperative
learning in a course, I explain on Day One exactly what I'm
going to do and why. I assure the class, for example, that I'll
be making them work in class not to make my life easier
(quite the contrary), but because research shows that stu-
dents learn by doing, not by just watching and listening. I
reinforce the point by citing some of the research; as always,
McKeachie121 and Wankat and Oreovicz131 provide good gen-
eral summaries, and Johnson, et al.,141 cite results specifi-
cally for cooperative learning.

C Provide coaching on the skills you want the stu-
dents to develop. When students complain (or make evident
in other ways) that they don't know how to set up problem
solutions or prepare for tests or work effectively in teams, I
try to offer some guidance during my office hours and occa-
sionally hold a miniclinic in class. Woods, Wankat and
Oreovicz, and Johnson, et al., are rich sources of methods
for facilitating development of learning and teamwork skills.

C Get feedback and try to be responsive to it. Espe-
cially when many students in a class seem to be spending a
great deal of their time hovering around Stages 3 and 4 of the
trauma scale (loss of confidence, anger, and withdrawal), I
grit my teeth and conduct a midsemester evaluation, asking
them to list things they like about the class, things they
dislike, and things that would improve the class for them.
The first list often surprises me: the complaints I've been

* Believe me, my observations about student resistance are
neither theoretical nor speculative.
Winter 1995

hearing tend to monopolize my attention, clouding my aware-
ness that what I'm doing is working well for many or most of
the students. The things they dislike are not exactly fun to
read, but I learn from them and the students seem to appreci-
ate the opportunity to vent. The suggested improvements
may include some that are unacceptable to me ("Stop assign-
ing problems that you haven't lectured on." "Cut out this
group garbage. ") but I may be able to act on others without
seriously disrupting my plans or compromising my prin-
ciples. When I respond positively to some of their sugges-
tions (like easing off on the length of the homework assign-
ments, or giving them the option of doing a few assignments
individually), it usually goes a long way toward getting them
to meet me halfway.

C Be patient. I expect many of my students (especially
those I haven't previously taught) to be frustrated and upset
in the first few weeks of my courses. I deal with it now better
than I used to, knowing from experience that most of them
will turn around by the final exam.

C Go back to the references periodically. When some
of my cooperative learning groups seem to be disintegrating
halfway through the semester, I look back at one of Karl
Smith's monographs (or, for that matter, at my own work-
shop notes). I'm usually reminded that I've been neglecting
one or another of the recommended CL practices, like hav-
ing the groups regularly assess their functioning and work
out what they need to do differently in the future.

C Don't expect to win them all. In the end, despite my
best efforts, some students fail and some who pass continue
to resent my putting so much of the burden of their learning
on their shoulders. A student once wrote in a course-end
evaluation, "Felder really makes us think!" It was on the list
of things he disliked. On the other hand, for all their com-
plaints about how hard I am on them, my students on the
average earn higher grades than they ever did when I just
lectured, and many more of them now tell me that after
getting through one of my courses they feel confident that
they can do anything. So I lose some, but I win a lot more. I
can cheerfully live with the trade-off.

1. Woods, D.R., Problem-Based Learning: How to Gain the
Most from PBL., Donald R. Woods, Publisher, Waterdown,
Ontario, Canada (1994)
2. McKeachie, W., Teaching Tips, 8th ed., Heath & Co., Lex-
ington, MA (1986)
3. Wankat, P., and F.S. Oreovicz, Teaching Engineering,
McGraw-Hill, New York, NY (1993)
4. Johnson, D.W., R.T. Johnson, and K.A. Smith, Cooperative
Learning: Increasing College Faculty Instructional Produc-
tivity, ASHE-ERIC Higher Education Report No. 4, George
Washington University (1991) J

, L1laboratory


Should Students Do Them or Design Them?

University of Adelaide Adelaide SA 5005, Australia

Laboratory exercises are an integral part of any chemi-
cal engineering curriculum. They perform several
crucial roles, including1 illustrating and reinforcing
chemical engineering theory and principles, providing
hands-on experience with commonly used process equip-
ment, and demonstrating experimental methodology and tech-
niques. But laboratory projects organized along strict tradi-
tional lines have some distinct disadvantages, which can
include minimal training in the communication aspects of
reporting laboratory results and the possibility of reports
being passed down from a group finishing the course to an
incoming group.
The communication problem has been widely recognized,
and steps are being taken in most engineering courses to
address it.[21 The second problem, however, is more serious.
One aim of an engineering course is to foster the develop-
ment of higher-level skills (e.g., analysis and synthesis), but
the passing of reports between groups allows students to
complete projects without acquiring such skills; the require-
ment for independent thought and problem solving is conve-
niently circumvented. While this allows time for a more
rigorous study of the traditional lecture-based material, the
understanding that would be obtained through careful study
of the associated practical work is lost. The high cost associ-
ated with changing laboratory apparatus makes it impracti-
cal to overcome this problem by altering the available projects.
These problems were identified as existing in the practical
component of the Level-3 chemical engineering course at
the University of Adelaide. Typically, at the start of the
academic year the students were given a series of practical

Copyright ChE Division ofASEE 1995

scripts and a handout listing the desirable sections that should
be included in a technical report, with the likely content of
each of those sections explained. The students were required
to undertake a total of twelve projects over the academic
year, with each project requiring six hours of laboratory
work and approximately twenty hours of data analysis,
calculations, and report writing. They worked in pairs
and each of them had to submit a total of six reports, which
were graded and returned to the student with comments and
corrections. While assessment was based solely on the
submitted reports, comments generally focused only on
technical aspects of the project. The grading system did
not reflect the importance which employers attach to com-
munication skills.131
Because of the decline in resources available to Australian
engineering departments in the past few years,[41 it has been
impossible to update expensive laboratory equipment.
Reports of previous-year projects are readily available to
students, and there is evidence that results, calculations,
and in some instances text, are copied from those earlier
reports. While software exists that can compare files for
similarities and which could therefore detect such duplica-
tion, it is not entirely useful in this case since the reports
are rendered in hard copy rather than on disk. Also, such
software is not suited to detecting copied results and is
unable to ascertain whether calculations were conducted
independently or by using sample calculations from previ-
ous reports as a template.
To address the above problems, I initiated a series of
major changes to the Level-3 laboratory course. The nature
of those changes, and the results, are the topic of this paper.

The new course structure consists of three assessable com-
1. Laboratory project work and reports (70% of the
overall grade)
2. Report writing and data analysis workshops (15% of
the overall grade)

Chemical Engineering Education

Anton Middelberg received both his BE and
PhD degrees from Adelaide University, where he
is currently a Lecturer and the Bioprocess Facil-
ity Manager. His expertise and industrial experi-
ence include the design, modeling, and optimiza-
tion of biochemical processes, particularly involv-
ing the production and recovery of recombinant
proteins from Escherichia coli.

The students were required to undertake a total of twelve projects over the academic year, with each
project requiring six hours of laboratory work and approximately twenty hours of data analysis,
calculations, and report writing. They worked in pairs and each of them had to submit a total of
six reports, which were graded and returned to the student with comments and corrections.

3. Laboratory project design exercise (15% of the
overall grade)
The relative weightings reflect the workload for each com-
ponent. The entire subject constitutes one-sixth of the Level-
3 course and has four times the credit points of the Level-3
design subject. Reducing the total number of projects has
allowed for introduction of the new components.
The first component is laboratory based and is similar to
the old format, although students now submit nine joint
reports and the method for report assessment has been
changed. Half of the report grade is now awarded for presen-
tation in order to emphasize its importance. A pro forma
marksheet is used to grade the reports, with half of the marks
being awarded for presentation and the other half for techni-
cal content. Initial use of the proforma led to a clustering of
report marks, and markers complained that the form did not
allow sufficient flexibility. A qualitative category called "Gen-
eral" was therefore introduced to compensate for this weak-
ness. A fraction (10%) of the grading for the first component
is based on laboratory performance and covers such aspects
as safe behavior, preparation, and experimental technique.
The "preparation" component has been central to reducing
the reliance of students on laboratory teaching assistants.
The second component examines formal report writing
and data analysis, while the third component requires that
students undertake a comprehensive design exercise involv-
ing communication with technical staff and outside organi-
zations, project planning, and budgeting. In combination,
these components address the problems of poor communica-
tion and the lack of higher-level skill development. A more
detailed discussion of these modifications follows.

A series of thirteen workshops that cover various aspects
of report writing and data analysis has been introduced.
Notes discussing report writing are distributed to the stu-
dents and are reinforced with both good and bad examples
from previous reports. A particularly useful technique is
placing earlier students' graphs and reports on overheads and
allowing the class to critique them. In this way students
understand the need to critically analyze their own work
from both a presentation and a technical point of view.
The workshops on data analysis cover such aspects as
randomization, linear regression, error analysis, and simple
comparative statistical tests. Although students have been
exposed to these topics in prior years, the connection to
Winter 1995

engineering is not always clear to them. The workshops aim
to reinforce the students' earlier exposure to statistics through
worked examples and the need to solve engineering-related
problems. Assessment for this component is based both on
"hand-in" problems submitted each week for the data analy-
sis section and on a final data analysis assignment, which is
also graded on presentation.

One method for overcoming the second problem, that of
copying reports, is to force students to adopt a problem-
solving approach. A research-type of experiment has been
described in the literature"51 in which students are told what
needs to be discovered and are then asked to plan a solution;
after presenting their plan, the students are required to con-
duct the necessary experiment (within the constraints of
available equipment) and to present their findings.
The third component of the modified course at the Univer-
sity of Adelaide, the laboratory project design exercise, is
based on a similar approach. Students are told to
4 Select a concept from their chemical engineering
course which is appropriate for a Level-3 under-
graduate experiment and which does not already exist
within the department as a chemical engineering
experiment. Existing resource constraints are made
clear to the students to ensure that their chosen
concepts are suitable.
4 Submit a plan outlining the concept and the tasks
which will be undertaken in order to submit a
satisfactory final report. This plan is examined by the
instructor and is either approved or disapproved (in
which case there is a request for an alternative
4 Design the laboratory project, including equipment
specifications and sources, after the concept has been
4 Submit a final report detailing the concept and how it
can be implemented.
The exact form of the final report is deliberately not speci-
fied, but the students are told it must include the following:
4 Technical and executive summaries
4 An outline of the concept, including background
4 A full PID diagram of the apparatus
4 An experimental script to be handed to students
undertaking the project
4 A budget, including direct costs and departmental

resources (e.g., the workshop)
4 A detailed plan for implementation
4 An examination of safety implications
Students are not required to build the apparatus and per-
form the experiment, unlike the research experiment de-
scribed earlier. This is a conceptual design exercise. The
following resources are made available to assist the students:
4 Technical staff to comment on the practicality of the
designs, to indicate workshop resource requirements,
and to assist students in sourcing key equipment items
4 Teaching assistants to comment on theory develop-
ment and to assist with chemical engineering calcula-
4 Engineering directories listing equipment suppliers
4 Telephone and fax machines
The activities undertaken to complete the exercise make it
different from traditional chemical engineering process de-
sign exercises elsewhere in our curriculum. First, instead of
being given a project by the instructor, students are required
to select a concept by themselves; second, they must com-
municate with a wide range of people including technical
staff, outside organizations, and senior engineers (i.e., aca-
demics); and finally, they must consider implementation
issues, including budgeting, resource constraints, and likely
workshop difficulties in equipment construction. While many
of these tasks are not included in traditional process design
courses, they are performed on a daily basis by practicing
engineers. It has been stated that "properly organized projects
which allow students to function as engineers and to receive
feedback are an excellent teaching method."'61 This end, in
addition to fostering communication and higher-level skills,
is the aim of the laboratory project design exercise.

Students submitted a range of project designs covering
diverse topics in chemical engineering. Some were based on
projects previously outlined
in this journal (e.g., electro-
chemical reduction in a
monolith reactor) while oth- Results of Stud
ers were based on students' Laborator
Class size: 27 students Mean
experiences and interests
and a desire to illustrate # Question Asked
basic principles with novel I. This subject demonstrated lin
apparatus (e.g., a hydraulic 2. The workshops %ere relevant
ram and the adsorption of 3. The workshops were valuable
colored impurities from raw 4. I prefer this style of teaching
sugar solutions using acti- 5. I benefited more from this sty
vated carbon). Other if it were a lecture-based appi
projects were designed to 6. M.\ ability to think critically
address perceived deficien- 7. MN ability to think independe
cies in the course. For ex-
ample, one group felt that

exposure to standards in the laboratory was inadequate
and so designed a project based on the Australian Standard
for measuring thermal conductivity. Another group, feeling
that the biochemical engineering content in labora-
tories needed to be increased, designed a protein ultra-
filtration experiment.
Most projects were well researched and had a strong
theoretical background, although some students simply at-
tempted to modify existing equipment or to investigate con-
cepts suitable for prior levels. Such projects were largely
detected at the concept-approval stage, however, and as a
result the final projects represented an extremely pleasing
cross-section of chemical engineering. Students seemed to
enjoy the activity, as was evidenced by the quality of the
reports received. Given the resources, several of the projects
would be worth implementing.

We conducted a student evaluation of teaching on the
class following completion of the exercise. The results are
shown in Table 1. All responses show a higher-than-average
agreement with the questions, indicating that students ben-
efited from the design exercise. In particular, students seemed
to gain instructional benefit from the workshops with techni-
cal and teaching staff (questions 2 and 3). This is supported
by additional comments from students who stated that the
chance to interact with workshop and technical staff gave
them a greater feel for reality and the problems associated
with implementing engineering designs. Similar positive feed-
back was obtained from workshop staff.
Students also believe that the subject allowed them to link
various parts of their chemical engineering course (question
1) and increased their ability to think independently and
critically (questions 6 and 7). Surprisingly, the lowest re-
sponse was obtained for the question regarding this approach's
benefit relative to a lecture-based course (question 5). The
spread of responses is highest for this question, and exami-
nation of the raw
data shows that
LBLE 1 only a handful of
valuation of Teaching for students shifted
ject Design Exercise from positive to
e: 7=strongly agree, 1=strongly disagree n tive o
negative responses.
Mean/7 St Dev Is it possible that
theirr subject areas. 6.2 0.8 these students have
aims of the subject. 6.3 0.8 a different concept
ly understanding of the subject. 6.1 1.0 of benefit to our
cture-based approach. 5.8 1.1 own? Have we im-
eaching than I would have posed an artificial
5.5 1.4 concept of benefit
en increased. 5.7 1.1 on some students
a, been increased 5.8 1.0 by having a pre-
dominantly lec-
ture-based curricu-
Chemical Engineering Education

ent E

ks to
to the
for m
to a le
le oft
iha beb
iady h

lum? I shall not even attempt to answer these questions, but
they certainly warrant further investigation.
An "open" question in the evaluation allowed students to
provide additional feedback. Only positive and constructive
comments were received, some of which were:
More direction as to what to do is required.
Implement our practicals-students in coming years should
be able to understand them better than we did with the
existing ones.
This course was relevant to what I perceive as the "real
engineering world." It makes it worth doing.
Application of the knowledge and skills developed in
lectures for all subjects to date is the best aspect-it gives a
better understanding as to what it is really all about.
Learned where and how to contact people for equipment.
Good opportunity to apply theory to a real design.
Learned how to initiate a design (and an idea) and then
implement it on my own. Gave me confidence that I would
be able to work in the real world.

Technical staff members were asked to provide written
comments on how the exercise might be improved in subse-
quent years. Specifically, they were asked to comment on
organizational problems, on how involvement in the exer-
cise affected their other duties, and on perceived student
All the staff felt that the exercise was useful and should be
maintained, particularly since it raised student awareness of
practical issues. Many useful comments were received re-
garding procedural and organizational matters, and they will
be incorporated into the course in subsequent years. The
time commitment from workshop staff was relatively small
(about two hours a week for five weeks) and did not detract
from their other duties to any large extent. Most staff felt that
more time with the students would be useful.
The following were identified as key student skill defi-
A lack of practical knowledge and an intuitive "feel" for
design parameters (e.g., flowrates and volumes).
Overreliance on technical staff (e.g., what sort of pump
should I use, how big should it be, and how thick should I
make the pipe insulation?).
A desire to cost apparatus down to the smallest item (i.e.,
an unwillingness to use budget estimates and approximate
realistic costs for small items).
Inexperience in preparing questions to technical staff.
Inability of students to communicate with outside organiza-
tions in an effective manner.
The first three points relate to the students' lack of practi-
cal knowledge and an apparent desire to defer chemical
engineering questions (e.g., pump sizing by characteristic
matching) to someone with greater practical experience when
Winter 1995

confronted with a real problem instead of a paper exercise.
This lack of practical knowledge may also explain the prob-
lems associated with the costing exercise. As indicated, some
students provided costings to the last cent, despite being told
that only key equipment items should be outsourced and
other costs approximated using the knowledge of workshop
staff. The last two problems relate to poor communication,
or a desire to seek information without a properly formed
question. Again, the exercise in costing proved problematic,
as many students approached outside organizations with in-
complete specifications for key equipment items.

Overall, the laboratory project design exercise proved popu-
lar with both students and staff. Students developed and used
a greater range of skills than with traditional laboratory and
design projects and they felt they were undertaking some-
thing more closely aligned to the role of a workplace engi-
neer. In the long term, exercises such as this might help
reduce the frequency of complaints from graduates that their
undergraduate education failed to adequately prepare them
for the engineering workforce.
Despite the positive outcome, several areas can be im-
proved. First, both students and workshop staff feel that
more time should be made available for consultation. Much
of that problem, however, may arise from the quality of
communication. Also, upon reflection, I believe many stu-
dents failed to make efficient use of staff time. Poorly pre-
pared questions often resulted in staff having to interrogate
students, rather than the desired case of students interrogat-
ing staff! To overcome this problem, an additional stage will
be incorporated into the exercise: following concept ap-
proval, students will be required to submit their apparatus
design and engineering calculations for grading. Workshop
staff will only be made available after this process is com-
plete. In this way we hope that students will be able to form
pertinent questions without deferring the design responsibil-
ity to staff. Clearly, initial designs will be modified in the
light of feedback from staff.
Second, some students requested more direction. Although
the course handouts will be modified in response to feed
back, I am loath to increase the level of direction. After all-
students function best when challenged, despite their appar-
ent desire to avoid using higher-level skills.
Finally, the problem of outside communication will be
addressed by requiring that all outgoing communication be
done either by fax (with approval by an academic or teach-
ing assistant prior to transmittal) or by using an approved
telephone communication plan. In this way, both the content
and appropriateness of the request for information can be
In addition to the above modifications, another course in
the chemical engineering curriculum, "Managing People and

Business" (introduced at the same time as this course modi-
fication), will be modified to address some of the concerns
arising out of this exercise. Specifically, a series of work-
shops on project management and communication will be
run at the start of this laboratory project design exercise.

The modifications introduced in the chemical engineering
laboratory subject have been well received. Students have
received additional guidance in writing reports and analyz-
ing data and have been required to undertake a laboratory
project design exercise. This has necessitated the develop-
ment and use of a range of skills not required in traditional
laboratory and design exercises.
Introducing these modifications has meant a reduction
(from twelve to nine) in the number of projects undertaken
by each group. But this number reduction has been accom-
panied by an increase in report quality and the need to
submit nine joint-author reports rather than six single-author
reports. The possibility of introducing research-type experi-
ments for the remaining laboratory projects, as described
elsewhere,[51 will be investigated to further improve the qual-
ity of the remaining laboratory time. Such an approach has
already been adopted for Level-4 students. A reduction in
laboratory time at Level 3 has also led to a net reduction in
subject resource requirements. The savings are being em-
ployed to offset the cost of improving existing experimental
rigs (e.g., by replacing chart recorders with data loggers).
The laboratory design project exercise has been particu-
larly useful in fostering development of higher-level skills
and reducing the reliance on reports handed down from
previous years. Clearly, it will be necessary to restrict
the choice of possible designs in future years to prevent
copying. Although it would be possible to hand out a list
of concepts, I feel it is more useful to allow students to
develop their own concept and then disallow it if it is too
similar to a previous concept. To this end I am retaining
copies of the initial concepts submitted by students (2-3
pages). Given their brevity and diversity, future concept
submissions can be easily compared to these filed copies at
the concept-approval stage. This is sufficiently early in the
exercise to identify possible plagiarists and invite them to
submit a new proposal.
Finally, an answer to the question posed in the title of this
article. I firmly believe that traditional laboratory courses
are an integral part of any curriculum and cannot be dis-
pensed with. But they do have some shortcomings that can
be partly overcome by providing formal communication train-
ing and by giving students a real engineering problem in
addition to laboratory work-design it, don't just do it!

I would like to thank all staff who took part in this exercise

for their valuable input and comments. I would also like to
thank the Department of Chemical Engineering at the Uni-
versity of Adelaide, and in particular Professor John Agnew,
for encouraging innovative approaches to teaching.

1. Jones, W.E., "Basic Chemical Engineering Experiments,"
Chem. Eng. Ed., 27(1), 53 (1993)
2. Pettit, K.R., and R.C. Alkire, "Integrating Communication
Training into Laboratory and Design Courses," Chem. Eng.
Ed., 27(3), 188 (1993)
3. National Board of Employment, Education, and Training
(Australia), "Skills Sought by Employers of Graduates," Com-
missioned Report No. 20, December (1992)
4. Department of Employment, Education, and Training (Aus-
tralia), Review of the Discipline of Engineering: Vol. 1. Re-
port and Recommendations, Australian Government Pub-
lishing Service, Canberra (1988)
5. Macias-Machin, A., G. Zhang, and 0. Levenspiel, "The Un-
structured Student-Designed Research Type of Laboratory
Experiment," Chem. Eng. Ed., 24(2), 78 (1990)
6. Wankat, P.C., "Learning Through Doing," Chem. Eng. Ed.,
27(4), 208 (1993) 0

REVIEW: Hygiene and Toxicology
Continued from page 17.

Organic phosphates
Alkane materials
Phosphorous, selenium, tellurium, and sulfur
Silicon and silicates, including asbestos
A considerable amount of information on toxicology is detailed
in part A of this series. Chapters 2 and 3 should be required reading
for not only those who will be involved in the manufacture of any
type of chemical but also for all who will be using chemicals in one
way or another during their daily life. These chapters outline the
care that should be exercised and the risks that could be encoun-
tered with various chemicals that have some toxic tendencies.
The usefulness of the twelve chapters dealing with various chemi-
cals identified as toxins varies to a certain degree. Evidently there
were no fixed rules provided by the editor to the authors of the
various chapters. Thus, a number of presentations begin with an
overall consideration of toxicity for the chemicals being reviewed,
while other presentations begin with an analysis of specific chemi-
cals and their toxicity properties. Some authors provide a tabular
presentation of the physical and chemical properties of all the
chemicals covered in the chapter. Other authors do this separately,
requiring the reader to go through the chapter to make comparison
of the properties. The former approach is more satisfactory since a
comparison of the properties could give some guide to the increas-
ing or decreasing toxicity level in a chemical family. Some authors
provided more uniform details on the toxicity studies of the chemi-
cals reviewed, while others summarized these all under the topic of
physiological responses.
The attention to details was overdone in the chapter on acetone
where details were included from the 535 references quoted through
1991. On the other hand, the chapter on alkaline materials appears
to require a further update since the most recent reference of the 89
Chemical Engineering Education


provided at the end of the chapter is 1979. Finally, the chapter on
the effects of tobacco smoke on the occupational environment
seemed out of place with the following chapters that analyzed
specific chemical families for their toxic effect. The location of the
chapter might have been more appropriate if it had been grouped
with other overall classes of materials (e.g., radioactive materials,
municipal wastes, etc.). Apparently the location of chapters, as
pointed out by the editors, was directly influenced by the availabil-
ity of the manuscripts prepared by the authors.
Despite these objections, the text provides an intensive evalua-
tion of the toxicity effects of a number of widely used chemicals.
Fortunately, the extensive subject index and chemical index of
fifty-four pages simplifies the task of obtaining toxicity informa-
tion on a specific chemical even in those chapters where the infor-
mation is provided in a different format. In most chapters except
the one noted above, the information provided contains data ob-
tained through 1991. O

Mo book review

by Stephen B. Vardeman
PWS Publishing Company, Boston, MA 02116-4324; 712
pages plus appendices (1994)

Reviewed by
Charles H. Barron
Clemson University

Most academic curriculum builders have, in recent years, con-
fronted the question of how to incorporate statistics into their
efforts. This has been called for by our industrial feedback systems,
both from industrial advisory boards and from alumni in industry;
in addition, there is a mandate in the current ABET criteria that
stipulates the need for evidence of proper statistical incorporation
within our engineering laboratory and design courses. This book is
an excellent choice as a textbook for a statistics course designed to
help meet these needs.
There have been several problems with traditional statistics course
offerings in the past. Frequently, the courses have been populated
by students who do not have the mathematics background of the
typical engineering student, and for this reason, many faculty advi-
sors have been reluctant to recommend statistics as an elective
course. The approach taken in this book by Professor Vardeman is
one oriented toward solving problems by using statistical methods
Winter 1995

and thought patterns. This is exactly the approach needed in order
for the course to appeal to most engineering students. No formal
background in statistics is expected, but the material presented
advances rapidly through the necessary techniques and methodolo-
After a brief review of variability and random error as they arise
in data-collection processes, the author introduces simple descrip-
tive statistics and some common experimental plans are described
without belaboring the theoretical basis for their design. Standard
numerical summary measures are discussed by using extensive
graphical presentations of distribution functions and their moments,
especially means and variances.
With this background the students moves on to introductory
concepts of regression and computing and using residuals. This
section of the presentation benefits from an extensive use of graphs.
After introducing the fitting of curves and surfaces, a full discus-
sion of the fitted effects of factorial data structures is presented. By
this time, about a third of the way into the book, the student is
becoming quite accustomed to the author's strategy of introducing
statistical tools as they are needed in the context of the problems.
It is only at this point that the background notions of probability
and the mathematics of randomness are raised. Chapter Five deals
with discrete and continuous random variables and goes much
further than previous coverage in the development of a number of
distribution functions. Several of these functions will be new to the
student, such as geometric, Poisson, exponential, Weibull, and the
beta distribution. Each of these distributions arises in the context of
an application, and the motivation to understand the distribution
arises naturally. Joint probability distributions are discussed after
this background work in completed. Chapter Six introduces the
concepts of statistical inference, confidence intervals, and signifi-
cance testing.
The next few chapters continue to expand the student's toolkit by
describing multisample studies, simultaneous confidence intervals,
analysis of variance, and variance decomposition. Control charts
arise in this section of the discussion. Inference for multisample
studies provides a platform for introducing the full factorial data
structure of such systems. With this basis, the full inference meth-
odology for general least squares curve- and surface-fitting is de-
veloped, and response surfaces are introduced. Finally, the book
concludes with the elaboration of some of the more advanced
concepts of experimental design.
The author has provided numerous examples and problems for
the student at every step along the way. This textbook is highly
recommended both to teachers of statistics courses for engineering
students and to engineering faculty who would like to sharpen their
own statistics skills in order to do a better job of using these tools in
their engineering laboratory courses. 0

Fall 1995 Graduate Education Issue of Chemical Engineering Education

Each year CEE publishes a special fall issue devoted to graduate education. It includes
articles on graduate courses and research as well as ads describing the university graduate programs.
Anyone interested in contributing to the editorial content of the 1995 fall issue should write to CEE, indicating
the subject of the contribution and the tentative date it will be submitted. Deadline is June 15, 1995.

a laboratory




University of Sydney Sydney, New South Wales, Australia 2006

Imagine that you are a junior engineer in a minerals
processing company. Your manager calls you in one day
and the following conversation ensues:
We have a new project to recover a valuable ore product.
Your part is the filtration of a slurry. We want to separate
50 tons per week of dry alumina. It will be worth $1,000 per
dry ton. Find out the most profitable way to do it. You will
have to do some experiments to get the rates of filtration.
We only have a small test rig, so you will have to scale the
whole thing up. I have forgotten all of the theory, so you
figure it out and give me a report in two weeks' time that I
can take to the next Board Meeting and tell them how much
we have to spend. It'll have to be good because I don't want
to make a fool of myself Off you go.
So begins a typical student pep talk for the unit operations
laboratory where we have attempted to introduce the flavor
of a real commercial enterprise. Until recently, all of our
laboratory experiments addressed only the "engineering sci-
ence" aspects of the work. Typically, students were asked
to get data from a rig, to do correlations, and to compare

Timothy A. G. Langrish received his BE degree
in chemical and process engineering in New
Zealand and his DPhil at the University of Oxford.
After working as a Research Fellow with the Sepa-
ration Processes Service at Harwell Laboratories,
he returned to the University of Canterbury in New
Zealand and subsequently took up a position at
the University of Sydney. His research interests
include drying technology and fluid mechanics.

Wayne Davies took his BSc and PhD at the
University of Sydney. He subsequently held a
position at Massachusetts General Hospital as a
Research Fellow in Medicine before returning to
Sydney to continue research in biotechnology
and biological process engineering. He finally
rejoined his alma mater and now works as a
private consultant. His interests are bio-process
engineering and advanced technologies for de-
stroying hazardous chemical wastes.

Copyright ChE Division ofASEE 1995

With little perceived relevance and
no significant goal for their work, [students]
would struggle to find something relevant to say.
Frequently the implied conclusion to their
investigations was a grumpy "so what!"

these with literature values and theories. With little
perceived relevance and no significant goal for their work,
they would struggle to find something relevant to say.
Frequently the implied conclusion to their investigations
was a grumpy "so what!"
Of the thirteen experiments available, we have now intro-
duced a measure of commercial relevance into two of them:
Filtration and Leaching. The first is a realistic problem of
process scale-up, and the second is an economic optimiza-
tion of an existing plant.

Students are asked to design a full-scale filtration process
for a hypothetical mineral company, Total Recovery and
Marketing Proprietary Limited (TRAMPL). Because they
have access only to a laboratory-scale filter for detailed
work, the problem is one of scale-up. Using constant pres-
sure operation on a laboratory scale filter, students must
produce a set of filtrate volume-versus-time data from which
they calculate two fundamental design parameters: 1) the
specific filter cake resistance, and 2) the filter medium resis-
tance. These parameters can then be applied to determine the
operation of a full-scale plant.
A major variable is the addition of a "filter aid." Addition
of filter aid (diatomaceous earth) should reduce the cake
resistance and consequently the filtration cycle times. The
resultant reduction in capital and labor costs should be offset
by greater running costs to pay for the filter aid. Students
find the economic optimum of their scaled-up process.
Certain design constraints are provided:
Chemical Engineering Education

The scale of the desired operation (how many
tons per week of alumina are to be recovered)
The nature and composition of the slurry which
will be separated
The capital costs of installing filtration equipment
The running costs in terms of labor and overheads
The costs of added reagents such as filter aid
which may prove useful in lowering costs by
speeding up production rates
The provision of these economic constraints is additional to
any physical constraints due to the size and layout of the
laboratory equipment.
Students write a report with the aim of conveying to
TRAMPL managers enough written information to convince
them that the experimental work and the conclusions drawn
can be relied upon. We emphasize that management should
be able to make a confident decision, based on the report,
that spending capital and employing staff to carry out the
process specified will bring a good return on investment for
the company. In the briefing sheets which accompany the
laboratory experiment, considerations in deciding the con-
tent of critical sections of the report, such as the Introduc-
tion, the Apparatus and Method sections, the Results and
Discussion sections, and especially the Summary, are out-
lined. This additional guidance is important, since reports to
management must be succinct as well as relevant.
In designing their experimental protocol, students are asked
to anticipate all the independent parameters that make an
impact on the outcome, such as the filter area of the labora-
tory scale rig and the filter area and frame volume of the full-
scale rig and filter-aid dosage. We emphasize that students
should design the experimen-
tal work considering dosages HIGH-
of filter aid which are relevant PRESSURE
both economically and techni- AIR ---
cally. This can be done by as- SUPPLY
suming some simple profitabil-
ity constraints and showing
that the possible range of filter
aid dosage is between zero and
a certain upper limit. At the
upper limit, the cost of the fil- WATER
ter aid begins to make the pro- T
cess uneconomic. Information
and the problem statement
given to the students are shown
in Table 1.
The equipment consists of a
laboratory plate-and-frame fil- TO
ter press (see Figure 1) that DRAIN
can be operated at constant
pressure but not at constant Figure 1. ScJ

Winter 1995

Fixed Costs
- Cost of 24-frame filter press
Installation (piping and solids
handling plant) ----------------------------
> Slurry feed pump(s) -----------------

> Overheads (independent of number
of filter presses) ---------------------------

Variable Costs
> Filter aid ------------------------------------
> Running costs -----------------------------

> Labor (all-inclusive costs to TRAMPL)
for day shift, 7:30 am to 3:30 pm -----
for night shifts and weekends ---------

$10,000/filter press

$40.000/filter press
Students specify flowrate
and get capital cost by
asking local suppliers


$100/24-hour day per filter

$24/hour per laborer
$36/hour per laborer

Financial workup
> Estimated project life --------------------- 5 years
> Value of solids recovered ---------- $1,000/dry ton

Assume 10% interest rate and 10% rate of inflation
Assume that equipment is totally written off after 5 years

hematic diagram of laboratory test filter plant.

Technical and Economic Information for the Filtration

Technical Design Parameters
Required rate of dry alumina production: 50 tons per week
Slurry concentration: 5% (weight/weight in water)



flowrate. Students carry out most of the experimental trials
with this equipment. There is limited access to a full-size
plate-and-frame filter press (see Figure 2) which can be
operated at constant flowrate for testing, although the full-
scale filter is to be designed for constant pressure using a
centrifugal slurry pump to deliver the feed. This situation
simulates the case where a company laboratory is modestly
furnished with small-scale equipment and there is limited
access to a working large-scale filter which is available, say,
at a neighboring company. It is possible to do a few quick
but not elaborate measurements on the large-scale gear such
as total frame volume and the maintenance time, e.g., the
time taken to dismantle the press, remove the solids, and
return the press to service.
We direct students to Volume 2 of Coulson, Richardson,
Backhurst, and Harkerl'" for filtration theory, and to Sinnott[2'
for the discounted cash flow analysis which is required for
the economic analysis. Students are encouraged to decide
for themselves which theories need to be used and to de-
scribe them sufficiently well so that the reader can follow the
arguments presented. Typically, they choose to perform four
runs at filter-aid dosages of 0, 1%, 2%, and 3% of the dry
solids concentration. For a uniformly formed cake of con-
stant specific resistance, a plot of t/V versus V (where t is
cumulative time and V is cumulative filtration volume) pro-
duces a straight line. The slope of this line is proportional to
the specific cake resistance and the intercept is proportional
to the medium resistance. The expression for specific cake
resistance contains the term 1/A2 (where A is the filter area),
and the expression for medium resistance contains the term
1/A. Finding the working resistances for the full-scale filter
is then easily done by factoring in the appropriate ratio of
areas of the small-to-large scale filters.
Typical experimental results show that filter aid reduces
the filter cake resistance significantly, but that adding filter
aid is not always economically optimal. The optimum profit
is achieved by a single shift of six to eight operators, work-
ing a forty-hour week, running one or two full-scale rigs and
using no filter aid. Students observe
from a sensitivity analysis that maxi- M.
mum profit corresponds to a bal-
ance between the increased through-
put due to using more operators and
the greater labor costs which this
entails. The experimentally mea-
sured cake resistances and the eco-
nomic analysis show that the use of FEED
filter aid for this type of porous TANK
filtercake (alumina) is usually un-
necessary because the added cost
of filter aid is not rewarded ad-
equately by lower capital and labor

Conceptual mistakes can lead to absurd results. The full-
scale filter press is designed to hold twenty-four frames for
normal operation. For experimental scale-up purposes, we
have it set up with only two frames. Students may forget that
the installed working area of this filter is twelve times the
filter area of two frames. If the rest of their analysis is
correct, they will discover that not one or two filter presses
but twelve or twenty-four are required!
To complete the exercise successfully, the technical and
economic analysis must be reported succinctly in a self-
contained "Summary." Managers need to be able to read a
summary and get an immediate idea of what the report is
about and of what use it is to their company. A frequent
oversight is to omit the aim and purpose of the laboratory
exercise and launch immediately into the numerical results.
Students know that the academics who mark their work are
totally familiar with the experiments, and some of them lose
sight of the need to write for a readership which may be
unfamiliar with those experiments. Although it is frequently
only one-half of a page of writing, we emphasize that the
summary is the most difficult part of any report to write well
and it deserves a significant effort.

Students are presented with a simulated commercial prob-
lem in the minerals-processing industry involving optimiza-
tion. There is a virtually unlimited supply of low-grade ore
containing a valuable solute, and a three-stage countercur-
rent leaching plant of fixed maximum capacity with which
to extract this solute. The design and performance of this
model plant have been previously described."3 Feed consists
of a slurry of inert material, usually 20% by wt. of PVC
granules in water, together with a dye fluoresceinn) at 50 to
80 ppm representing the valuable solute. This slurry is me-
tered into the first-stage mixer via a positive-displacement
pump where it contacts the overflow from stage two. The
combined slurry is then fed to the first-stage settler where
the solids are allowed to settle, leaving as underflow and

Figure 2. Schematic diagram of full-scale filter press plant.
Chemical Engineering Education

'' '' '~''

producing an overflow which is the concentrated solution
product stream. The underflow from stage one is the feed to
the stage-two mixer, which contacts the stage-three over-
flow, etc. Make-up water is fed into the third-stage mixer
and the well-washed inert solids leave as underflow from the
third-stage settler. Dye concentrations are determined spec-
trophotometrically and flowrates are determined by volu-
metric apparatus and stopwatch.
As in nearly all separation technology (of which this is just
one example) there is a compromise between "recovery" and
"quality" of the valuable product. At low make-up water
flowrates the product solution is obtained at relatively high
concentration, but its flowrate is small and total recovery of
the chemical is relatively low. As make-up water flowrate
increases, the product solution concentration falls-but the
total recovery increases asymptotically to 100%.
To make this laboratory experience more like the real
thing, economic criteria are given to the students describing
The value of the dry solid chemical
The cost of processing the product solution (by
evaporation, for example)
The cost of power, maintenance, and labor
The cost of treating unrecovered chemical, since it
may be an undesirable burden on the environment if
it is dumped as landfill, for instance.
The equations for income and costs have been arbitrarily
defined. Overall profit is simply income minus costs. Thus
I= Q (10 100/Cp)
I = income ($/hour)

Economics of Single-Stage Leaching Plant

Water Flowrate Dye Conc'n. Dye in UF Dye in OF Recovery Income
(liters/hour) (ppm) (g/h) (g/h) (%) ($/hour) ($

0 60.00 0.785 0.94 55 7.85
5 51.12 0.669 1.06 61 8.52
10 44.54 0.583 1.14 66 8.88
15 39.45 0.516 1.21 70 9.04
20 35.41 0.464 1.26 73 9.07
25 32.12 0.420 131 76 9.00
30 29.39 0.385 1.34 78 8.86
35 27.08 0.355 1.37 79 8.66
40 25.12 0.329 1.40 81 8.42
45 23.41 0.307 1.42 82 8.14
50 21.93 0.287 1.44 83 7.84
1000 1.68 0.022 1.71 99 -84.51

Winter 1995

Q = dye production rate (gram/hour)
Cp = dye concentration in product stream (ppm)
This equation gives a steadily falling income, even be-
coming negative, as the product stream becomes more di-
lute. This reflects the greater costs of subsequent processing.
There are penalty costs for disposal or treatment of
unrecovered dye and a flat-rate running cost which applies
regardless of dye recovery. These costs are represented by
C = 2 + 2 (C)(V,)
C = Total running costs ($/hour)
C, = concentration of dye in underflow (ppm)
V, = solution flowrate in underflow (m3/hour)

The constant $2/hour is the flat-rate running cost and the
coefficient 2 is the treatment cost at $2/gram of dye.
The object of the laboratory exercise is to operate
the leaching process in such a way as to maximize the
profit. Students must identify the key parameter to be varied
(most importantly, the water flowrate and the underflow
solids concentration) and the range of operation to find
the maximum profit most efficiently. Students need to
set a number of run conditions, operate the laboratory
plant to achieve steady state, and perform mass balances
to show that the data can be relied upon before determin-
ing peak profitability.
In order to estimate the optimal experimental range, we
ask students to write a mathematical model of the simpler
case of a perfectly mixed single-stage plant which is conve-
niently done using a spreadsheet calculation. In this case, the
dye-containing slurry is mixed with water and the mixture is
allowed to separate into an underflow
(UF) fraction containing all of the solids
together with some solution and an over-
flow (OF) fraction containing dye solu-
tion only. Table 2 shows a typical result
Costs Profit with make-up water flowrate varying be-
/hour) ($/hour) tween 0 and 50 liters/hour. As can be
seen, the addition of water steadily re-
3.57 4.28 duces the dye concentration such that the
3.34 5.18 dye lost in the UF decreases. Although
3.17 5.71 the dye concentration in the OF also de-
3.03 6.01 I creases, the total recovery of dye in-
2.93 6.15 creases. As an internal control of the cal-
2.84 6.16 culation, a very large value of make-up
2.77 6.09 water, e.g., 1000 liters/hour, shows that
71 595 recovery approaches 100% as expected.
2.671 5.95 Using our arbitrary formulae for incomes
6 56 and costs it can be seen that
2.61 5.53
2.57 5.26 Income is at a peak at 20 liters/hour
2.0 -of make-up water flow
2.04 -86.55y decrease
Costs steadily decrease

Profit is at a peak at 25 liters/hour of water flow
It now becomes interesting to compare the profits gener-
ated by the three-stage experimental rig to those generated
by the single-stage theoretical prediction. The commercial
justification of using a more complex and expensive three-
stage rig is that it should create a more concentrated product
dye solution and result in less dye lost in the underflow. We
should expect therefore that the three-stage rig should return
a greater profit and use less make-up water to do so, com-
pared to the single-stage prediction. Results from laboratory
experiments (see Figure 3) show that the three-stage rig
returned a maximum profit of 14.1 $/hour at a water flowrate
of 23 liters/hour. The single-stage prediction returned corre-
sponding values of 6.2 $/hour and 25 liters/hour.
The overall mass balance around the process indicates the
reliability of the measurements as well as the level of under-
standing students have for steady state operation. With care-
ful operation and analytical measurement, discrepancies in
the mass balances for both dye and solids should be less than
10%. Discrepancies of this magnitude are within the 95%
confidence limits of the uncertainties in the experimental
measurement of concentration (the largest contribution to
error) and flowrate. From these mass balances students are
encouraged to argue a case that their data indicate steady
state operation with the flow of all streams accounted for,
before they discuss the effects of experimental variables
such as feed or water flowrate.
Students work in groups of two, taking turns to act as
Group Leader. We assess students on their ability to design
a set of experiments to achieve peak profitability most
efficiently and convincingly. Emphasis is placed on the
ability of the Group Leader to organize the work to be done
on the day in question, and on the ability of both students
to communicate the critical operation parameters in a con-
cise form in their reports which are due two weeks after the
date of the experiment.
The Unit Operations Laboratory is a core element of the
undergraduate curriculum at the University of Sydney. At
this stage of their studies, the third year of a four-year
curriculum, they have had initial exposure to economic theory,
but are not expected to perform elaborate economic analysis.
Despite this, the introduction of economic relevance into the
Unit Operations Laboratory has been accepted and even
welcomed by the students.

Putting a commercial flavor into the Unit Operations Labo-
ratory has increased the perceived relevance to students as
well as enhancing the engineering-science aspects of the
work. By making the laboratory results "do something," we
find that all students, whether proficient or of average abil-
ity, must understand what the relevant equations actually

10 20 30
Water (solvent) flowrate, I/hr
Predicted profit Actual profit
single-stage apparatus three-stage apparatus
---__ a

40 50

Figure 3. Predicted profit from a single-stage math-
ematical model and actual profit from the
three-stage plant.

mean and not just how to substitute blindly into them. With
this emphasis there is a greater point in interpreting results
critically. In particular, the analysis of experimental error,
which has been hitherto a vague concept at best, is now
given tangible importance as dollar figures depend on the
outcome. Students perceive that their decisions in the design
of the process may lead to unrealistic operating costs and
that these decisions will be challenged by management. In
this case they are encouraged to take the initiative and to
seek better alternatives.
We use the commercial aspects of the laboratory to pro-
mote a more responsible attitude toward experimental work
and report writing. In the past we have been aware of the
students' tendency not to analyze data critically. Lines of
best fit might be drawn through a random scatter of data
points, and data which do not fit some pet theory might be
conveniently ignored. Values of profit, equipment sizes, trans-
fer coefficients, etc., might be reported with four or five
significant figures when the primary data have two.
Just as "real life" managers might react, we are especially
critical of reports which disown responsibility for students'
work, with statements such as "...there were leaks in the rig
which did not allow precise analysis of results..." or "...ow-
ing to lack of agreement between theory and experiment,
additional work will be required..." or "...there were many
errors in this work, the major one being human error and to a
lesser extent experimental error...". We tell students that a
professionally written report presents a case to the reader
that the results are reliable within certain confidence limits.
Including a list of perceived sources of error, as shown in the
last example, with no explanation and no reassurance that
the errors do not contradict the conclusions, spells instant
self-disqualification. Our simulated management teams are
not interested in excuses. This message is getting through;
Chemical Engineering Education


/ \
... . . . .

students have increasingly accepted the challenge of being
responsible for their work and they now feel more produc-
tive and stimulated. As a result, the addition of an economic
flavor in the Unit Operations Laboratory has been well re-
ceived by students who have appreciated its relevance to
education leading to a commercial career.

By providing a balance between an innovative commer-
cial aspect and the traditional engineering-science aspect of
the Unit Operations Laboratory, we have introduced ele-
ments of "real-life" into laboratory work. Solving problems
which have the flavor of industry makes laboratory work
more challenging and interesting and, we think, more rel-
evant for the students.
In the future we intend to extend the approach beyond the
two experiments discussed here and involve the majority of
the Unit Operations Laboratories. Most experiments can be
augmented readily, according to the same principles used
with our leaching and filtration experiments, by 1) defining
profitability equations for an existing plant or process and
seeking an economic optimum, or 2) defining a scale-up
problem and using laboratory data to predict the economics
of a full-scale plant and process.

1. Coulson, J.M., J.F. Richardson, J.R. Backhurst, and J.H.
Harker, Chemical Engineering. Volume 2: Unit Operations,
3rd ed., Chapter 9 (Filtration) and Chapter 10 (Leaching);
Pergamon Press, Oxford, England (1978)
2. Sinnott, R.K., Chemical Engineering. Volume 6: An Intro-
duction to Chemical Engineering Design, Chapter 6 (Cost-
ing and Project Evaluation), Pergamon Press, Oxford, En-
gland (1983)
3. Davies, W.A., "A Three-Stage Counter Current Leaching
Rig for the Senior Laboratory," Chem. Eng. Ed., 23(2), 96
(1989) O

Heterogeneous Reaction Rate Data
Continued from page 25.
Figure 3 shows that pso is a linear function of pso, im-
plying that these variables were actually not changed inde-
pendently during the experiments. Thus, there is no way to
separate the information in the reaction rate data related to
these two variables. Plotting po versus pso gives
similar results, indicating that there is also a linear depen-
dency between these two variables.

Initial results have shown that the reaction rate data of
Table 1 can be represented well by the rate expression of Eq.
(1). It is very tempting to jump to the conclusion that the
experimental data verifies the mechanism postulated in Eq.
(1), but using numerical and statistical analysis of the data,
Winter 1995

we have proven that such a conclusion is completely ground-
less because

1. The data itself is not experimental, but extrapolated,
the accuracy of which is impossible to assess.
2. Because of the large value of the equilibrium coeffi-
cient and limited accuracy of the reaction rate values,
no effect of the reversibility can be detected in the
reaction rate values. As a result, assuming irreversible
reaction yields a more accurate correlation over that
obtained with the reversible model.
3. The partial pressures of SO,, SO,, and 02 were not
varied independently during the experiments; there is
linear dependency between the partial pressures of
SO,, SO,, and 0-. As a result, it is impossible to
discriminate between the effects of pso, Ps,, and po,
on the reaction rate.
4. The equation

r = 1.62 x 10-5 (-2.27)

has been found to best represent the data in Table 1,
but because of the limitations of the data that were
mentioned earlier and the empirical nature of the
power-law rate expression, there is absolutely no
certainty that this rate expression is applicable for
other combinations of partial pressures.

We have used one particular example to demonstrate
several potential pitfalls in correlation of experimental data.
It can be expected that in most practical cases not all
these pitfalls will show up; but these four points can serve
as guidelines in assessing the quality of the data, the accu-
racy of the correlation, and the adequacy of a model to
represent the data.
In conclusion, it is appropriate to quote Churchill, who
noted that "...if the observed behavior of the process re-
quires the use of a more complex model than the data
justifies, resolution lies in the laboratory rather than in
further analysis."

1. Churchill, S.W., The Interpretation and Use of Rate Data:
The Rate Concept, revised printing, Hemisphere Publishing
Co., Washington, DC (1979)
2. Smith, J.M., Chemical Engineering Kinetics, 3rd ed.,
McGraw-Hill, New York, NY (1981)
3. Olson, R.W., R.W. Schuler, and J.M. Smith, "Catalytic Oxi-
dation of Sulfur Dioxide: Effect of Diffusion," Chem. Eng.
Prog., 46(12), 614 (1950)
4. Shacham, M., and M.B. Cutlip, POLYMATH 3.0 Users'
Manual, CACHE Corporation, Austin, TX (1993)
5. Brauner, N., and M. Shacham, "Model Discrimination and
Error Analysis in Heterogeneous Rate Data Regression for
Irreversible Reactions," submitted for publication (1993) O




Northwestern University Evanston, IL 60208-3120

After three-and-one-half decades of development, the
computing environment is now highly intercon-
nected. Networks proliferate between computers,
laboratories, buildings, campuses, and across continents.
Computer use is integrated into many chemical engineering
courses to aid teaching, learning, and communication. Many
pioneers' dreams have now become a reality.

The Computing Environment When the university is in
session, the chances are that lights are on in the Computer
Teaching Laboratory, its computers are running, and stu-
dents are using them in various ways, some of which their
older brothers or sisters could not have done just a few years
ago. These days, the Laboratory is easily the most-used
facility in the chemical engineering department. System crash
is now a rare event. The Laboratory's opening hours are
dictated only by security and maintenance considerations
and there is no full-time staff associated with the facility. It
is user-serviced with a half-time teaching assistant acting as
a Laboratory manager, with policy guidance and planning
provided by a faculty director. For fourteen hours a day
during the week and eight hours on Saturdays and Sundays
the micros slave tirelessly at the friendly commands of users.
The micros are connected in a local area network (LAN)
served by a file server, printers, and other peripherals. The
LAN is linked to the campus fiber optic backbone, and
through it to the Internet worldwide.2
Access to the information superhighway is the most sig-
nificant step forward in the empowerment of faculty and
students and has already taken place on many campuses.
Give or take a few details such as the type of hardware and
software or the physical dimensions of the laboratory and
the size of the student population, the environment described
above reflects the computing and information processing
environment currently existing in many universities and, in

'Address: University of Texas, Austin, TX 78712-1062
2 In this simplified description we have omitted a few hardware
and software details.

particular, in the chemical engineering departments.
Impact of Computers on ChE Education This paper is
concerned with the present status of engineering education,
with specifics taken primarily from chemical engineering.
How have computers affected the learning and the teaching
of engineering? The use of computers has now been inte-
grated into most of our undergraduate courses, beginning
with material balances and stoichiometry (analysis of chemi-
cal process systems), thermodynamics, equilibrium separa-
tions, continuing with process dynamics and control, pro-
cess design, process optimization, and chemical engineering
laboratory, and ending with electives such as statistics in
process modeling. Significant changes have already taken
place in the content, learning, and teaching of these subjects.
For example, linearization and Laplace transformations play
a ubiquitous role in classical process control. In the days
before computers, we spent much time on inverse transfor-
mations and in the preparation of Bode and Nyquist dia-
grams in stability analysis. Now, with Program CC, we
simply input the appropriate polynomials in the numerator
and denominator of the transfer function in the Laplace
domain, and then let the computer do the tedious work.
Parametric studies are easy to carry out. Understanding and
insight, which used to take a long time to develop, are now

Richard S.H. Mah is Professor of Chemical Engi-
neering at Northwestern University. He received
his BSc from the University of Birmingham, his
DIC and PhD from Imperial College, and his DSc
(Engineering) from the University of London. As a
researcher, analyst, supervisor, and professor, he
has used computers for more than thirty-five years.

David M. Himmelblau is the Paul D. and Betty
Robertson Meek and American Petrofina Founda-
tion Centennial Professor of Chemicat Engineer-
ing at the University of Texas, Austin. He received
his BS from the Massachusetts Institute of Tech-
nology, and his MS and PhD from Washington
University. He is Executive Officer of the CACHE
Copyright ChE Division ofASEE 1995
Chemical Engineering Education

acquired rapidly and enthusiastically. Similarly, TK-Solver
and Lotus 1-2-3 take a lot of drudgery and mystery out of
balances and stoichiometry. With flowsheet simulators and
property libraries, the dual role of thermodynamics in pro-
cess analysis and in property estimation becomes very much
easier to teach and to explain. In statistics, by using Monte
Carlo simulation, the instructor can readily demonstrate and
verify, for instance, the Central Limit Theorem and can
display plots in vivid color graphics in dimensions which
"will cross a rabbi's eyes" (Fiddler On the Roof).
The result of all this is that by using computers, one can
cover more territory and tackle more realistic problems in
less time. Because of the availability of these new tools and
techniques, it is possible to begin experimenting with new
pedagogy1'"31 which, in time, may profoundly change the
ways students learn and instructors teach these subjects.
This is particularly true with subjects involving many ele-
ments, complex structures, and closely knit relationships
(such as systems engineering) that would be difficult to
demonstrate experimentally. With computer simulation we
can now reproduce precisely controlled "misbehavior" to
study its impact on every aspect of the system.
Communication and Productivity Tools Equally remark-
able are advances that have taken place in communication
and personal productivity tools. With only a modicum of
formal instruction, students can acquire serviceable skills in
word processing, graphics, desktop publishing, database, and
E-mail, and with spell checkers there is no excuse for mis-
spelled words. By making it "fun" to prepare texts and
illustrations, not only do the reports and illustrations look
more professional, but in due course they also improve in
substance and style. With universal access to computer net-
works, anyone can send a message or be reached by E-mail
without having to play phone-tag.* Through remote access,
instructors can just as easily review class records and assign
homework problems as they can conduct an electronic dia-
log with colleagues at other locations-all without leaving
the physical environment of home or office. Last but not
least, by greatly simplifying the protocol, distribution, and
delivery, E-mail lowers the threshold of communication and
shrinks the physical and psychological distances of an orga-
nization, be it a corporation, a government, or an university.
To appreciate the profound and pervasive changes that are
taking place in information technology in general, and com-
puters in particular, we need only look back at the path of
progress that has led us to the present state of development.

*As we were revising this manuscript in October of 1994, an all-
electronic conference was taking place on Internet, breaking new
ground in communication. As participants of the first Interna-
tional Chemometrics Internet Conference (InCINC'94), the
authors could ask and answer questions and conduct technical
discussions on papers in five simultaneous sessions, while
proceeding with our daily business in the background mode. We
were quite conscious of being witnesses to an historic event.
Winter 1995

This paper is concerned with the present
status of engineering education, with specifics
taken primarily from chemical engineering. How
have computers affected the learning and
the teaching of engineering?

By most reckoning, we are in the fourth decade of com-
puter applications, even though there may not be an exact
point of origin. The first two decades were dominated by
mainframes and minis. In chemical engineering, much of the
initial programming efforts were directed at replacing repeti-
tive calculations. Taking 1958 as our reference point, the
establishment of FORTRAN as the universal high-level pro-
gramming language for quantitative computation must rank
among the foremost achievements of that first decade.
By the second decade, LP (Linear Programming) and,
more specifically, codes based on the Simplex Method, had
become the single largest user of computer time in the pro-
cess industries. Time sharing, on line terminals, and
flowsheeting programs were some of the other notable de-
velopments of that decade.
The year 1978 heralded the introduction of the first com-
mercial-scale microcomputer, the Apple II, followed three
years later by the IBM PC and the mass marketing of soft-
ware (word processors, spreadsheets, and databases) that
fueled the revolution of microcomputers.
By 1990 personal computers (PCs, clones, Apples, and
Macintoshes) became the second most common tool of com-
munication-second only to the telephone. It is notable that
E-mail and networking did not gain popularity until well
into the third decade.
One of the most remarkable characteristics of the com-
puter industry is the continual improvement in performance
in relation to price which has been sustained for over three
decades. Figures 1 and 2[4] (next page) show that the price of
computing has dropped by one-half every two to three years
ever since computers were marketed commercially. A present-
day $3,000 PC is comparable in computing speed to a mil-
lion-dollar mainframe a decade ago. If progress in the rest of
the economy had matched progress in the computer sector, a
Cadillac would cost $4.98, while ten minutes' work of labor
would buy a year's worth of groceries![41
What Has Already Happened, or Is Happening One
important impact of the changing price/performance ratio is
improved user-friendliness. In the early 1980s, a word pro-
cessor ran on a 64 KB memory microcomputer, by the mid-
1980s it required 640 KB of memory, and now, in 1995, no
respectable application software requires less than several
megabytes of memory. But with this extravagance in memory
requirement came a much more fault-tolerant and user-

friendly interface, and the same "look and feel" (under Win-
dows, for example), that makes the task of assimilation
much less formidable for lay users. In fact, previous empha-
sis on learning to program in FORTRAN and similar high-
level languages has diminished since Matlab, MathCad,
Polymath, Mathematica and their ilk have enhanced their
capabilities and relieved users of the need to program-
raising anew the question whether it is necessary to teach
programming to students other than computer engineering
and computer science majors.
With ever-improving computing capabilities and methods
of solution, the bottleneck in process analysis is once again
the quality and fidelity of the models suitable for different
applications. Some attention is already being directed to
applications typically ignored by educators, such as model-
ing less structured, fuzzy problems and applications involv-
ing noisy and correlated data.
Cheap information storage and improved means of trans-
mission and distribution have already changed the modus
operandi of traditional institutions such as libraries and pub-
lishing houses. Journal abstracts and even articles are sold
on compact disks that can be searched at will at nominal
cost, and with an E-mail address and access to databases a
student can download information just as easily as he can
send an electronic file to a friend.
Textbook have changed substantially. Readers are expected
to have access to a computer to solve exercise problems.
Disks containing pertinent software are commonly found
inserts in the backs of books. The technology exists today to
customize, assemble, and electronically deliver textbooks
for each student, but it may take time to resolve all the
copyright issues and to provide suitable marketing mecha-
nisms. How to provide teaching material for engineering
courses will almost certainly be a major issue in the next
decade, and the opportunities for innovation will be limited
only by our imagination.
One important impact of computers on engineering educa-
tion has been to broaden the access to teaching and learning
styles. In a few instances, computer-aided learning has com-
pletely replaced the lecture-recitation format for learning,
but in most universities changes have occurred in a more
limited way over a period of many years as the role of
computers in education became better appreciated by the
faculty. Such changes are often caused more by the influx of
young faculty members who have hands-on knowledge in
using computers than by the action of accreditation or uni-
versity guidelines. So retooling of tenured faculty may well
be one limiting factor in introducing new information tech-
nology in our pedagogy. Nonetheless, the rate of technologi-
cal innovation will continue to be rapid, and equipment will
become technologically obsolete when it is still in good
mechanical condition. Short life cycles in computing tech-
nology will continue to be a fact of life. To stay in the

competition, schools must have plans and funding to rejuve-
nate programs and facilities. Those with foresight to antici-
pate will have a competitive advantage.
Curriculum revamping will surely be needed at some point,
since we cannot go on adding new material to the existing
courses without deleting other topics. This will create oppor-
tunities for experimenting with new pedagogy, which may
in turn make our profession more accessible to a wider range
of candidates, thereby contributing to retooling of the na-
tional workforce.
Historically, the path to progress is strewn with expensive
wreckage. A megabuck investment does not ensure that a
project will succeed, and today's success is no guarantee for
tomorrow. An example of innovative educational software is
the PLATO system, which reportedly cost CDC hundreds of

-o-- Computers
--Producers' durable
10,000- equipment

0 1,000


10 I I
1955 1965 1975 1985 1995

Figure 1. The cost of computing has declined substantially
relative to other capital purchases (based on data from U.S.
Department of Commerce, Survey of Current Business, 1990).
Figure reprinted with permission from Comm. ACM, 36(12), 67, Dec (1993)

1 0
o Microprocessors
SMemory 1M 586
.a 10 486
: 486
S256K 'i860

105- 64K /286

S 416K/ /'8086
104 4K 8048
/ M8080
103 004,
1970 1980 1990 2000
Figure 2. Microchip performance has shown uninterrupted
exponential growth (data provided by Intel.) Figure reprinted with
permission from Comm ACM, 36(12), 67, Dec (1993)
Chemical Engineering Education

millions of dollars in the 1970s but which has left no lasting
imprint on engineering education today. The IBM
multitasking operating system for PCs, OS/2, has been re-
ported to cost one billion dollars to develop and has pro-
duced perhaps one hundred million dollars in revenue. With
respect to hardware, in the 1970s IBM crushed RCA, Xerox,
Honeywell, and GE, and compelled those companies to take
large write-offs. In the parallel computing field, more re-
cently both Kendall Square and Thinking Machines have
declared bankruptcy.
However, we did learn some valuable lessons. Most po-
tential users cannot visualize how to use unfamiliar technol-
ogy in large mental steps. If the context of the new technol-
ogy is sufficiently dissimilar to the current context, rejection
is likely. Thus, quantum leaps often fail where incremental
changes may succeed. Another lesson for developers of new
computing technology is to focus on the relevance to educa-
tional needs and not be carried away by the clever, exciting,
or imaginative technology. Changing curriculum solely to
take advantage of computing technology is usually a waste
of resources.

What might we predict about the future? On the hardware
side, commercial technology already available includes
1280x1024 pixel graphics, 21 inch color monitors, 90 MHz
clocks, 150 MIPS computing speed, 10 GB hard drives, and
1200 dpi printers. In the near future we expect Intel to
continue maintaining its remarkable 44-month product cycle.
High-speed transmission and vast databases are expected to
make greater impact on education. Subject to usage monitor-
ing and accounting, software at one site may be exploited by
students at another site at little cost, making it unnecessary
and unattractive to reinvent the wheel. Further integration of
software and textbooks will eventually lead to multimedia
teaching material.
An education built on sound fundamentals and in-depth
understanding is still the best strategy to allow one's
knowledge base to evolve and grow with changing times.

Networking 21
Word processing 6
Statistical 2
Spreadsheet/database 5
Programming 13
Graphics 3
Desktop publishing 11
Scanning II
DOS environmental 12
0 10 20 30 40 50 60
Percent of users
Percentages total more than 100% due to multiple responses

Figure 3. Types of Software and use in computer labs. [Graphic by Steve
Alspach; reprinted with permission from NU Information Systems and Technology, North-
western University, Evanston, IL, News and Views, 2(3), Spring (1993)]
Winter 1995

While hands-on practical experience is indispensable to en-
gineers, one must avoid overspecialization. Paraphrased in
another way,5s' kilobit education is dangerous in a world of
gigabyte technology.
On the other hand, history also shows that the momentum
generated by a real winner can carry development a long
way. FORTRAN, LP, word processor, and E-mail are
some examples. Word processing was probably the single
largest application that spearheaded the commercializa-
tion of personal computers. Figure 3 shows that it continues
to be the dominant application of campus microcomputer
users even today.'61
Compared with the earlier decades when IBM accounted
for three of every four computers sold, we now have a global
market for buyers and sellers of information technology in
addition to vast capital and financial institutions, trained
manpower, and many potential winners. A list of promising
developments includes networks, optical and parallel com-
puters, CD ROMs, satellite broadcasting and reception, per-
sonalized portable phones and pagers, notebook computers,
and high-definition television. The potential for information
technology mergers which will further enhance the use of
computers in engineering education is very large and very
likely.* How engineering education can continue to make
use of these rapid changes remains a challenge. It is our
responsibility to take advantage of these changes in engi-
neering education to shape engineers who shape the future.
We are almost at the dawn of the 21st century. Looking
back along the pathway leading to the present makes us
realize how far we have traveled in a journey that was
propelled by just a few key inventions. How many more
wonders lie ahead of us to be discovered, invented, and
applied to engineering education in the coming decades?
The possibilities are truly exhilarating and exciting.

1. Felder, R.M., and L.K. Silverman, "Learning and Teaching
Styles in Engineering Education," Eng. Ed., 674, April (1988)
2. Schank, Roger, "How Students Learn: Educational Software
and the Future of Education," Sponsored by Searle Center
for Teaching Excellence, April (1994)
3. Stice, J.E., "Using Kolb's Learning Cycles to Im-
prove Student Learning," Eng. Ed., 291, February
2 (1987)
4. Brynjolfsson, E., "The Productivity Paradox of In-
formation Technology," Comm. ACM., 36(12), 67,
December (1993)
5. Augustine, N.R., "Socioengineering Age," ASEE
Prism, 24, February (1994)
6. Alspach, S., NU Information Systems and Tech-
I nology, Northwestern University, Evanston, IL.,
70 News and Views, 2(3), Spring (1993) 0

Mergers are taking place even as we write. Novell and
Word Perfect haOe just announced a new alliance to

compete with Microsoft which is teaming up with
McCaw to form Teledesic, a global communication
network linked by 840 satellites.

SSurvey Commissioned by CACHE




A Perspective on Training and Application

Ohio State University Columbus, OH 43210-1180

Over the past decade, computing has had an unprec-
edented impact on the chemical process industry in
terms of use and widespread acceptance of the tech-
nology. The impact has accelerated over the past five years
as computing has rapidly become a pervasive tool used for a
variety of purposes including numerical computation, analy-
sis, text processing, graphics, communication, and accessing
information. In these times of streamlined engineering staffs
and extensive outsourcing of engineering tasks, the com-
puter is recognized as a critical tool in conducting business.
It is no longer viewed as a stand-alone box only for numeri-
cal computation; it has become an extension of how prob-
lems are solved and a medium for processing information.
As such, it has become an integral part of virtually all as-
pects of the chemical processing industry.
In response to the rapidly changing technology and with
limited opportunity and time for dialog with industry, uni-
versities are forging ahead with curriculum changes to re-
flect new instructional objectives relating to computing tech-
nologies. These changes are proceeding in response to gen-
eral industrial expectations for increased levels of computer
literacy but without a clear and detailed perspective of com-
puting in industry. In addition, industry and academia are
simultaneously, but independently, trying to understand the
current and future impact of computing on engineering with
the result that expectations and objectives may not be articu-
lated clearly. As a result, there are many unanswered ques-
tions regarding industrial and academic transitions into this
technology. How is computing helping the process industry?
Has computing changed the way we do engineering? Are we
better engineers as a result? What skills are required to enter
the profession? Are universities meeting the challenge in
training future engineers?
To provide some insight into these questions from both

"[ DowElanco, Indianapolis, IN
' Purdue University, West Lafayette, IN 47907

academic and industrial perspectives, the CACHE Corpora-
tion Curriculum Committee commissioned a series of indus-
try and academic surveys on computing. Specifically, the
surveys targeted
Engineering management-to get a broad and current view of
computing in the chemical process industry
New BS chemical engineers with only a few years of profes-
sional service-to compare their professional computing
requirements with their recent college training
Chemical engineering faculty-to compare the academic view
with the industrial view.
This article summarizes the results of these surveys. Three
primary topics are addressed in separate sections: Computer
Use in Industry; Content of Training; Computing in the
Chemical Engineering Curriculum.

Recent BS Chemical Engineering Graduates
379 questionnaires 152 responses
The respondents were from four major companies reflect-
ing the chemical, petroleum, pharmaceutical, and consumer
products industries. Engineering professionals who have
graduated within the past three years made up 45% of the
respondents; another 30% have been in industry between
three and five years; and about 25% of them have been in
industry more than five years. As indicated by this distribu-
tion of experience, the survey by and large concentrates on
computing within the past five years. 83% of all respondents
were involved in technical work.
The recent graduates represented the following distribu-
tion of job descriptions:
*38% in-process/plant support
*29% research and development
*20% design and analysis
*7% process control
*15% other
Copyright ChE Division ofASEE 1995
Chemical Engineering Education

Industrial Management
205 questionnaires 156 responses
The management sample involved 156 total respondents
from a wide variety of companies-chemical, control, com-
puter, pharmaceutical, aerospace, petroleum, consumer prod-
ucts, government, food, and technology companies were all
represented in the sample. Of the respondents, 16% were
managers with ten to fifteen years in industry, and over 75%
had more than fifteen years experience. Clearly, the manage-
rial sample reflected an experienced viewpoint on the impact
of computing and the expectations of new engineers. Re-
spondents with job descriptions including technical manage-
ment made up 76% of the sample, while 12% were doing
technical work.
154 questionnaires 65 responses
The questionnaire was sent to each U.S. chemical engi-
neering department. Of the 65 respondents, about half have
been in academia more than fifteen years, The other half
have between five and fifteen years academic experience,
but over 60% of that number have had less than five years in
industry. The survey did not ascertain how many had no
industrial experience.

The management response reveals that engineers now spend
a substantial amount of time at the computer. Well over half
of the engineers average between 20-40% of their time at the
computer, while another 30% spend 40-60%. Interestingly,
academics substantially underestimated how much comput-
ers are used-the academic perspective estimated that about
70% of engineers are in front of the computer less than 20%
of their time, while more intensive users were estimated to
spend 25-50% of their day with the computer.
The results of the new graduate survey, shown in Table 1,
provide a perspective on what kind of computing is being
performed. This breakdown is very revealing. First, virtually
everyone is making some use of spreadsheets. It shows that a
large percentage of engineers (74%) are frequent users, with
the remaining people being occasional users. When asked
what the primary uses of the spreadsheets programs are,
there was strong concurrence by the recent graduates and
management. The greatest use is for data analysis-but quite
significant use is directed toward material balances, eco-
nomic studies, and numerical analysis, in decreasing order
of importance.
The table also reveals that very few engineers are pro-
gramming in FORTRAN, and that a majority (64%) never
do. Those few who do program in FORTRAN are only
occasional users. By and large, engineers do not program in
other languages either, but it is evident from those who do
that other programming languages are being used as much as
FORTRAN. It is revealing that statistical packages, numeri-
Winter 1995

Computer Use in Industry
Never Seldom Frequent
Spreadsheets 2% 23% 74%
FORTRAN programming 64% 28% 8%
Language other than FORTRAN 56% 28% 15%
Statistical 46% 40% 14%
Numerical method libraries 85% 11% 3%
Mathematical packages 86% 13% 1%

cal methods libraries, and mathematical packages are sel-
dom or never used except implicitly in application packages.
Clearly, programming and more specialized packages are
not in widespread use even though a fair amount of attention
is devoted to at least some of these in most academic institu-
tions. One possible explanation is that industry tends to
develop specialized users of scientific computing packages
and that they, in turn, serve the needs of other engineers
within the company.
Virtually everyone is using the computer for communica-
tions (e-mail, word processing, etc.), and a large percentage
of engineers are using graphics software for technical report-
ing, presentations, and visualizations. Database systems are
also being used with high frequency. Over 70% of the re-
spondents are heavily using DBMS applications for organiz-
ing project information, accessing general engineering data,
and processing information.
Generally speaking, management and recent graduates
agreed on the level and kind of computing they do. Given the
variety of companies represented, the fact that there is this
agreement supports the generalization of the survey results.
It is noteworthy that both management and academia con-
curred with the rank order of computer uses as reported by
the engineers. There were wide differences, however, in the
perceptions of managers and academics on the amount of



Communications heavily underestimated heavily underestimated
Graphics heavily underestimated heavily underestimated
Database systems underestimated heavily underestimated

Spreadsheets match

heavily underestimated

One question asked of all three groups was how much time
is needed to learn the computer skills required for the job.
There was strong concurrence on this question: over 80% of
recent graduates claimed less than three months, while man-
agement and academia both estimated three months or less
for 75% of new engineers. Two notable differences, how-
ever, did arise: the majority of engineers claimed that they
required less than one month for training, whereas manage-
ment estimated three months. Furthermore, management

claimed a significant number (19%) of engineers required as
much as three to six months of training. This agreed with the
responses of the recent graduates but not with the perspec-
tive of academia, which projected that very few graduates
would require this extensive amount of time.
Recent graduates overwhelmingly considered computing
to be an integral part of the undergraduate program, but 10%
thought that computing should not be included. While this is
a relatively low percentage, it is striking that there is this
percentage of respondents disclaiming the importance of
computing in education, given its wide spread use in indus-
try. With respect to undergraduate training in computing, the
recent graduates provide an important perspective:
13% training is more than enough
62% training is about right
25% training was not nearly enough
We can conclude from the above that academia is doing an
adequate job of preparing new engineers, but there is appar-
ently considerable room for improvement since a significant
percentage claimed that they had not had enough training.
Confirmation is provided by the response to a related ques-
tion showing that 34% of recent graduates felt they were not
adequately prepared in computing.
Regarding the content of their academic education, a ma-
jority (57%) of recent graduates felt their preparation con-
centrated too heavily on programming over applications,
while 38% felt that the mix was about right. Virtually no one

thought training overemphasized applications. To the ques-
tion of how much programming should remain part of
an undergraduate program, 40% still thought it should be
more strongly emphasized and essentially none thought it
should be eliminated. A majority (60%) recommended
some exposure. It is clear that recent graduates recognize the
importance of programming in learning how a computer
works, even though they may not do much program-
ming themselves. The exposure is seen as important to
understanding computing.
On this same question about programming, management
and academic viewpoints were in agreement with the gen-
eral feelings reported by recent graduates. But 67% of the
managers wanted to see stronger emphasis on applications,
while only 50% of the academics wanted to strengthen the
emphasis. A strong contingent of both academics and man-
agers (about one-third of each) advocated equal time to
applications and programming. With respect to the choice of
programming language, academics had strong opinions in
favor of FORTRAN: about three-fourths of the respondents
wished to continue with FORTRAN programming, but a
significant number (21%) did not. Managers were substan-
tially less adamant on this issue, with 25% having no opin-
ion. A majority still favored FORTRAN while a significant
number of respondents were not in favor of it at all. On the
usefulness of a second language, there was generally a mixed
opinion by management, but it leaned toward 'no.' Academ-
ics were relatively unenthusiastic about a second language,

N-Never S-Seldom F-Frequently Y-Yes N-No

1. Years since receiving BS degree:
<3 3-5 >5
2. Primary type of work:
administration technical management technical sales/marketing
3. Time required to learn the computer skills for current job:
< Month 1-3 months 3-6 months >6 months
4. Description that best fits your job:
process design/analysis research and development process control *
plant/process support other
5. Do you use the computer for communication (e-mail, word
processing, calendars, and access to on-line data)? Y N
6. Do you run spreadsheet programs? N S F
7. What are the primary uses of spreadsheet programs?
economic studies data analysis numerical analysis material balances
8. Do you program in FORTRAN? N S F
9. Do you program in language other than FORTRAN? N S F
10. Do you use graphics software?
never technical reporting presentations visualizations
11. Do you use statistical packages such as SAS, RS/1, etc.? N S F
12. Do you use numerical methods libraries such as IMSL, NAG,
etc.? N S F
13. Do you use symbolic and mathematical manipulation packages
such as Mathematica or Matlab? N S F
14. Do you use database management systems?
never project information general engineering data
process information
15. Do you use high level software packages? N S F

16. Do you feel you have had sufficient undergraduate training in
computing to prepare you for your work environment?
more than enough about right not nearly enough
17. Do you feel that your undergraduate computer training had the
proper mix of programming versus applications?
too much programming over applications
too much applications over programming about the right mix
18. To what extent do you feel computer programming should
remain a part of the undergraduate program?
strongly emphasized some exposure eliminated
19. Did undergraduate training expose you to more than one
operating system (e.g., DOS, UNIX, VMS, etc.)?
no two systems more than two systems
20. If not, would you have benefitted from exposure to multiple
operating systems?
yes no no answer
21. Were you sufficiently trained to understand and use flowsheeting
systems and physical property systems?
yes no unimportant
22. Do you feel you had sufficient exposure to computer tools to solve
non-trivial problems?
more than enough about the right amount not adequately prepared
23. Should computer programming remain part of the undergradu-
ate program? Y N
24. Do you feel there is a relationship between computer skills and
problem-solving skills?
yes, strong positive correlation yes, strong negative correlation
some correlation no correlation

2 Chemical Engineering Education

with 50% responding "no."
Multiple operating systems apparently is an unresolved
issue with recent graduates. Half of the respondents felt that
exposure to more than one operating system is important and
half did not. When asked if they would have benefited from
exposure to multiple operating systems, 33% answered yes
and 19% answered no. Almost half had no opinion.

Particularly noteworthy is the result that, while 57% of the
recent graduates thought they had sufficient exposure to
the computer to solve non-trivial problems, nearly a third
thought they were not adequately prepared. On a more spe-
cific question, only half of the respondents felt they were
sufficiently trained to understand flowsheeting systems
and physical property estimation systems. 44% felt they
were not adequately prepared, but 10% thought flowsheeting
was unimportant.
Academics felt considerably more strongly than either
graduates or management that graduates do not have enough
exposure to computing skills. Both management and aca-
demics, however, overwhelmingly considered computing to
be an enhancement to problem-solving. A significant nega-
tive impression was still apparent, though, in that 15% of
managers and 11% of academics considered computing to
have no effect or to be a hindrance to problem-solving. On a
related question about computing skills and the ability to

formulate or define problems, again 14% of the managers
considered computing to be a hindrance and another 42%
felt there was no effect. Academics marginally considered
computing to be more of an enhancement than did the man-
agers, but nearly half felt there was no effect or that comput-
ing was a hindrance. While the computer has come into
widespread use in industry and is generally considered to be
a positive element, there remains a significant contingent of
engineers who do not believe that computing has much of an
effect on how problems are conceptualized and defined.
An overriding issue with respect to computer education is
the effect that computing has on problem-solving skills. A
majority of recent engineers felt there is a correlation be-
tween computer skills and problem-solving skills, and nearly
a third more thought there was a strong correlation-but
11% felt there was no correlation. The academic perception
closely matched that of recent graduates, but managers were
somewhat less convinced, with 21% claiming no correla-
tion. Another significant question asked whether or not
recent graduates were bringing a systems analysis ap-
proach to problem solving. It is disappointing that a majority
(70%) of academics and management reported that engi-
neers are not adopting more of a systems viewpoint when
solving problems. On the other hand, there is general agree-
ment that computers are resulting in differences in how
engineering in conducted. Apparently, there is substantial


1. Years since receiving BS degree:
<5 5-10 10-15 >15
2. Primary type of work:
administration technical management technical sales/marketing
3. Type of positions filled by BS chemical engineers in your
process design/analysis research and development process control
administrative plant/process support systems other
4. Percent of the day a typical BS chemical engineer in your
department spends at the computer.
5. What percentage of BS engineers use the computer for communi-
cation (e-mail, word processing, calendars, and access to on-line
data)? _
6. What percentage of BS engineers run spreadsheet programs?
7. What are the primary uses of spreadsheet programs?
economic studies data analysis numerical analysis material balances
8. What percentage of BS engineers use graphics software (technical
reporting, presentations, and visualizations)? _
9. What percentage of BS engineers use statistical packages such as
SAS, RS/1, etc.?
10. What percentage of BS engineers use numerical methods libraries
such as IMSL, NAG, etc.? _
11. What percentage of BS engineers use symbolic and mathematical
manipulation packages such as Mathematica or Matlab?
12. What percentage of BS engineers use database management
systems for project information, general engineering data,
process information, etc.? _
13. How much time is required to train engineers to learn the
computer skills for their job function?
< 1 month 1-3 months 3-6 months > 6 months
14. Do you feel the new graduates have had sufficient exposure to
computer tools to solve non-trivial problems?

not enough about right more than enough
15. Do you feel the students' exposure to computer technology has
been an enhancement or hindrance in engineering problem
hindrance no effect enhancement
16. Has the exposure to computer skills enhanced or hindered the
ability of the graduate to formulate or define problems conceptu-
ally or mathematically?
hindrance no effect enhancement
17. Do you feel there is a relationship between computer skills and
problem-solving skills?
yes, strong positive correlation yes, strong negative correlation
some correlation no correlation
18. Do you believe that undergraduate training should emphasize:
programming over applications applications over programming
devote equal time
19. Do you believe undergraduate training should include exposure
to more than one operating system (e.g., DOS, UNIX, VMS, etc.)?
highly desirable not necessary unimportant
20. Should computer programming in FORTRAN be part of the
undergraduate curriculum for chemical engineering?
yes no no opinion
21. Should computer programming in any general purpose language
be part of the ChE undergraduate curriculum?
highly desirable not necessary no opinion
22. Are new graduates bringing a systems analysis approach to
process unit operations?
yes no, still doing things by conventional means unaware of a difference
23. Do you believe we are doing things differently (rather than faster/
more efficiently) with computers (e.g., design/analysis area)?
yes, significant innovations no unaware of difference

Winter 1995 5

recognition that computing allows engineers to do more
things faster but that it does not have a fundamental impact
on how we do engineering.


A final topic addressed specifically by the academic sur-
vey pertains to changes in curricula to accommodate com-
puting. One of the important questions considered the possi-
bility that computing education itself contributed to length-

ening the time to graduation for undergraduates. The survey
confirmed that a significant number of students are taking
longer than four years to complete their undergraduate pro-
grams. The respondents reported that a full 31% (on aver-
age) of undergraduates take an additional semester or quar-
ter and 25% take even longer. Of the 65 universities re-
sponding, 89% claimed that computing had no effect on the
length of time a student takes to graduate. In fact, 20% of the
departments have fully integrated computing into their cur-


1. Years you have been in an academic position:
<5 5-10 10-15 >15
2. Years of industrial experience:
<5 5-10 10-15 >15
3. Type of positions filled by BS chemical engineers. Please rank
process design/analysis research and development
process control _administrative
S_ plant/process support systems
S other don't know
4. Percent of the day a typical BS chemical engineer in your
department spends at the computer.
5. What percentage of BS engineers use the computer for communi-
cation (e-mail, word processing, calendars, and access to on-line
data)? _
6. What percentage of BS engineers run spreadsheet programs?
7. What are the primary uses of spreadsheet programs?
economic studies data analysis numerical analysis
material balances don't know
8. What percentage of BS engineers use graphics software (for
technical reporting, presentations, and visualizations)? _
9. What percentage of BS engineers use statistical packages such as
SAS, RS/1, etc.?
10. What percentage of BS engineers use numerical methods libraries
such as IMSL, NAG, etc.? _
11. What percentage of BS engineers use symbolic and mathematical
manipulation packages such as Mathematica or Matlab?
12. What percentage of BS engineers use database management
systems for project information, general engineering data, process
information, etc.?
13. How much time is required to train engineers to learn the
computer skills for their job function?
< 1 month 1-3 months 3-6 months > 6 months
14. Do you feel the new graduates have had sufficient exposure to
computer tools to solve non-trivial problems?
not enough about right more than enough don't know
15. Do you feel students' exposure to computer technology has been
an enhancement or hindrance in engineering problem solving?
hindrance no effect enhancement don't know
16. Has the exposure to computer skills enhanced or hindered the
ability of the graduate to formulate or define problems conceptu-
ally or mathematically?
hindrance no effect enhancement don't know
17. Do you feel there is a relationship between computer skills and
problem-solving skills?
yes, strong positive correlation yes, strong negative correlation
some correlation no correlation don't know
18. Do you believe that undergraduate training should emphasize:
programming over use of applications applications over use of
programming devote equal time don't know
19. Do you believe undergraduate training should include exposure to
more than one operating system (e.g., DOS, UNIX, VMS, etc.)?
highly desirable not necessary unimportant

20. Should computer programming in FORTRAN be part of the
undergraduate curriculum for chemical engineers?
yes no no opinion
21. Should computer programming in an additional language be
part of the undergraduate curriculum for chemical engineers?
yes no no opinion
22. Are new graduates bringing a systems analysis approach to
process unit operations?
yes no, still doing things by conventional means unaware of a
23. Do you believe we are doing things differently (rather than
faster/more efficiently) with computers (e.g., design/analysis
yes, significant innovations no unaware of difference
24. In courses using computing, have computing assignments tended
to be added on to previously existing course material or have
they been integrated into the course by changing/removing
previously used materials?
full integrated partially integrated added on
25. What percentage of the undergraduate students in your
program take longer than four years?
% taking an additional semester/quarter
% taking an additional two semesters/quarters
26. Has increased use of computers in the curriculum contributed to
students taking longer to graduate from your program?
significantly to some extent not at all
27. Is computing helping students to better learn chemical engineer-
ing principles?
yes unaware of a difference no
28. How would you currently rank order the value of teaching
students skills in the following? (Please number, with 1 being the
most important.)
object-oriented languages spreadsheets
statistical packages database systems
numerical methods libraries communications
symbolic and mathematical packages other
29. Please rank the following in the order they are emphasized in
your department's program (with 1 being the most important).
object-oriented languages spreadsheets
statistical packages database systems
numerical methods libraries communications
S symbolic and mathematical packages other
30. Looking five years into the future, how would you rank order
the value of teaching students skills in the following. (Please
number, with 1 being the most important.)
object-oriented languages spreadsheets
statistical packages database systems
numerical methods libraries communications
symbolic and mathematical packages other

4 Chemical Engineering Education

ricula and another 75% claim that it has been partially inte-
grated. This is important data in that it conveys the fact that
departments are indeed recognizing the role of computing in
engineer training and are willing to include it at the expense
of other topics. It is not simply an addition to their normal
course of study.
A critical question posed to academics was whether com-
puting helped students better learn chemical engineering
principles. About half of the respondents claimed that com-
puting did indeed help, but the other half felt computing
either had no effect or did not help. The large neutral-to-
negative response seems to indicate that computing is not
necessarily resulting in better chemical engineering educa-
tion. This again corresponds with the earlier responses on
the effect of computing on problem solving.
To provide both current and future academic perspective
on computing in education, the survey asked respondents to
rank order the value of developing student skills in a number
of computing areas. Surprisingly, the current and the five-
year perspectives on these skills were essentially the same.
Below is the rank-ordered list of skills:
FORTRAN programming
C programming
Object-oriented languages
Statistical packages
Numerical methods libraries
Symbolic and mathematical packages
Database systems
The interpretation of this list is, as indicated in the previ-
ous section, that academics see programming as an impor-
tant element in a chemical engineer's education. FORTRAN
programming will likely continue to be the most im-
portant language, but recent trends in C and object-oriented
programming have been noted by academics and their
importance to the curriculum is recognized. Spreadsheets
and graphics are also strongly mentioned. Skills in using
specialized packages are not valued nearly as much as
programming skills.

Computers are used extensively in industry by virtually
all engineers. Academics need to adjust upward their per-
ceptions of the amount of computing used by their students
in industry.
The primary uses of computing in industry are by far
for communications, spreadsheets, graphics for reporting
and presentations, and database systems. There is an educa-
tional benefit for engineering programs to expose students to
these computing applications.
>- A relatively small number of engineers do technical
Winter 1995

computing or programming (beyond spreadsheets). This does
not mean that industry is not doing much technical comput-
ing. Rather, it appears that industry tends to develop special-
ists in scientific computing packages who then serve the
needs of other engineers within the company. There appears,
therefore, to be little need to teach highly specialized com-
puting packages in depth. It may be more important to give
students a broad exposure to a variety of computing pack-
ages to develop a general appreciation of how computers can
be used.
> Relating to the above, the survey results convey a
strong message that training in the use of specific packages
is not as important to industry as is a general engineering
computing skill set. Time needed to train new employees in
specific computing skills used by a company is not that
significant. Universities should focus attention on this gen-
eral skill set, and industry will train employees in specific
skills. The general skill set apparently includes program-
ming experience to understand how a computer works and
experience with application packages to understand the is-
sues of interpreting computer-generated results and to be
able to relate them to real-world problems.
There is a disparity between the views of industry and
academia on the relative value of programming versus expe-
rience in using application packages. But even though engi-
neers do very little programming, most respondents recog-
nize the importance of programming experience in develop-
ing an understanding of how computers work. The general
sense of the survey is that the curriculum time commitment
to each should be about equal.
The survey results indicate that universities are gener-
ally doing a good job of graduating engineers with the neces-
sary computer skills for the profession. But it is noteworthy
that a significant number of recent graduates reported that
they were not adequately prepared and that their computer
training was not enough. It is important for departments to
continue to assess the computing component in their cur-
ricula and to continue the transition to full integration of
computing into all engineering courses.
An important observation from the survey concerns the
fundamental impact of computing on engineering. While
computing is generally considered an enhancement to engi-
neering problem-solving, this enhancement apparently re-
lates only to speed and efficiency of doing tasks. Significant
percentages of the respondents did not feel that computing
helped in better defining and solving problems. In fact, a
small but significant number felt it was a hindrance. Further-
more, computing has had little impact in reinforcing or pro-
moting a systems-analysis approach to the solution of engi-
neering problems. There appears to be considerable room in
both academia and industry for understanding, and then
teaching, new viewpoints for analyzing and solving prob-
lems more effectively. 0

S classroom



Clarkson University
Potsdam, NY 13699-5705
Chemical engineering students are required (by ac-
creditation agencies) to make appropriate use of com-
puters throughout their program. Appropriate use is
defined as including most of the following: programming in
a high-level language; use of software packages for analysis
and design; use of appropriate utilities; and simulation of
engineering problems.
Maple is a powerful and flexible computing tool that has
the potential of becoming the software package of choice for
much scientific and engineering work, perhaps replacing, at
least in part, other computer-based methods such as tradi-
tional programming languages and special purpose analysis
and design programs. In this paper we provide a brief de-
scription of Maple and discuss some of the ways it can be
used in the chemical engineering curriculum.

Maple[21 is a computer algebra system (CAS). Computer
algebra is defined as follows:[31
Computer algebra (sometimes called algebraic manipulation,
or symbolic computation) can be defined to be computation with
variables and constants according to the rules of algebra,
analysis and other branches of mathematics, or formula
manipulation involving symbols, unknowns, and formal
operations rather than with conventional computer data of
numbers and character strings.
There are several computer algebra systems in use today.
Macsyma, Reduce, Derive, Mathematica, Maple, and
Scratchpad (now known as Axiom) are some of the better-
known ones (Gonnet and Gruntz131 give brief histories). This
article focuses on Maple because it happens to be the CAS
we use, although much of what follows would also be true of
other systems.
The emphasis in the above definition on symbolic manipu-
lation should not be taken to imply that Maple is unsuitable

SThis paper is based on a presentation given at the 1994 Maple
Summer Workshop. An abbreviated version appears in the
proceedings of that event1'
2 Address: InContext Corp., Toronto, Ontario, Canada

Ross Taylor is a professor of Chemical Engi-
neering at Clarkson University. His interests are
in the areas of mass transfer and separation pro-
cesses, and he is the coauthor (with Professor R.
Krishna, University of Amsterdam) of Mutticom-
ponent Mass Transfer (Wiley, 1993) and (with
Harry Kooijman) of ChemSep, a software pack-
age for separation process simulation used in
universities in several countries.

Katherine Atherley has a degree in pure math-
ematics from the University of Waterloo where
she first encountered Maple. She worked with the
Maple research group before joining WMS, and
has been involved with many aspects of Maple,
from programming and writing documentation to
providing user support and teaching Maple
courses. She is currently Product Manager at
InContext Corporation.

for the numerical calculations that dominate engineering
computing today. Maple combines symbolic mathematical
capabilities (integration and differentiation, for example)
with numerical capabilities (integration of ODE systems and
sparse linear equation solving) and sophisticated graphics
(including three-dimensional plots of surfaces and
spacecurves, sparse matrix plots, and much more) that allow
new approaches to the teaching of traditional materials.
Maple provides a command-line-in-a-window style of in-
terface. Its commands can be entered at the prompt and
executed immediately. This allows more immediate and easier
experimentation and exploration in what are called
worksheets. Worksheets can be scrolled forward and back-
ward in order to review prior results; in fact, it is possible to
go back, change and re-execute just one or a few Maple
instructions without having to re-execute the entire worksheet.
Maple runs on a wide variety of computer platforms. The
worksheets created by one version of Maple are saved as text
files and can be imported with little difficulty. The interfaces
to different versions of Maple are not the same, however. In
some cases this is more or less inevitable (as, for example,
between the MS DOS and MS Windows versions). In other
cases (such as between the AIX, MS Windows, and Macintosh
versions), it is less easy to understand the differences. The
interfaces to all of the above mentioned versions possess
interesting features that should be available in all versions.
Copyright ChE Division ofASEE 1995
Chemical Engineering Education

A Personal Perspective

My first exposure to Maple came some
years ago when I considered using it for
solving the examples in a textbook I was
writing. At that time Maple was not able to
do many of the things we needed. More re-
cently I took another look at Maple when
the engineering school at Clarkson was con-
sidering adopting it (see box Maple at
Clarkson). To teach myself Maple, I did not
go through the tutorials; instead, I picked
problems from the literature in chemical en-
gineering or which were somehow related
to my research interests and tried to see if I
could get anywhere with Maple. Finding the
critical constants to a generic cubic equa-
tion of state with the help of Mike Monagan
was the first problem I looked at.1'6
I have done (and continue to do) a great
deal of programming in Fortran, and at first
I found it difficult to get used to Maple. I
believe that part of my difficulties with
Maple had to do with the fact that I had to
learn a new way of thinking when writing
Maple. Despite (or perhaps because of) these
difficulties, I became addicted to Maple in
a very short time, and now it is the tool I
turn to first.
By manipulating equations with Maple I
have gained new insights (at least for me)
into methods of solving certain problems.
Now that I am past the initial hurdles, I am
able to solve real engineering problems with
much less Maple code (and in much less
time) than would have been the case had I
used Fortran. Moreover, I can do things with
Maple that are not even possible with For-
tran. For example, after I had solved a bi-
nary distillation problem with Maple, it took
me only fifteen minutes to program Maple
to construct the McCabe-Thiele diagram
shown in Figure 4.
I expect (hope) that students who do not
have any prior experience programming in
Fortran (or other traditional language) will
not face the same degree of difficulty that I
I use the MS Windows and AIX versions
of Maple in my own work and have shared
files with others using Maple on Sun and
Next workstations and on Macintoshes (and
possibly others for all I know).
Ross Taylor

Winter 1995

In the next few sections we focus on a few ways in which Maple can be used in
selected chemical engineering courses.

Computer Programming
Undergraduate engineering students are required to do some programming in a
high-level language. Often the programming course is the first course to expose
students to elements of engineering problem solving. Most students learn FOR-
TRAN, and C or Basic are included in some curricula. With more and more
students learning to use Maple in the calculus classes (and fewer and fewer of
them using any kind of traditional programming language after they graduate), it
makes sense to consider adopting Maple as the programming language for use in
their engineering courses (see the box Maple at Clarkson).
Maple is built around a programming language that is custom-designed for
symbolic mathematical calculations and manipulations. Unlike FORTRAN, the
Maple language supports standard mathematical structures as data-types and
can work with them in sensible mathematical ways. We feel that the language
is as suitable as any of the traditional languages for emphasizing those
problem-solving skills that are acquired by learning a programming language.
One might even make the case that Maple is better suited; Maple code is more
natural for mathematical work and produces much shorter, easier to understand,
programs than FORTRAN.

Material and Energy Balances
Maple allows one to solve elementary material balance problems in a sys-
tematic way that makes it almost impossible to get the problem formulation
wrong. Just a few lines of Maple code can set up the material balances and
mole fraction summation equations for any process unit regardless of the number
of components and input and output streams. Systems with chemical reactions
also can be modeled.
Maple's ability to handle symbolic indices makes it possible to identify compo-
nents with a number, name, chemical formula, or any other convenient label. The
number of unknown variables and independent equations can be quickly counted
and, hence, the number of degrees of freedom determined. Specification equa-
tions can be added to the model equations and Maple asked to solve the entire set
of equations in one go.
Problems often can be solved symbolically in terms of an unspecified parameter
(reactor conversion, say). This is useful if it is desirable to evaluate the solution at
several parameter values. It is unnecessary even to choose a basis as the actual
specification of interest can be included among the set of equations.

Calculation of the critical constants for a cubic equation of state is a classic
problem in thermodynamics, one that is covered in most thermodynamics text-
books.4'51 Textbook examples usually include finding the critical constants for the
simplest cubic equation of state, that of Van der Waals or those of the Redlich-
Kwong family. With Maple, however, it is possible to obtain explicit expressions
for the critical constants for a generic cubic equation of state.[6] Constants for
particular equations of state can be obtained as special cases of the general result.
Maple is a useful tool for the visualization of thermodynamic functions. Figure
1 shows the roots of the compressibility polynomial for the Soave-Redlich-
Kwong equation as a function of reduced temperature at a reduced pressure of
0.75. The well-known fact that a cubic equation can have complex roots at certain
parameter values is illustrated here. This figure was obtained with about thirty

lines of Maple code, which included the calculation and
ordering of the roots themselves. Figure 2 provides a three-
dimensional view of a compressibility diagram; only one
line of code (over and above that used for Figure 1) was
needed to create Figure 2. Three-dimensional diagrams in
Maple can also be rotated and viewed from other angles.
A Maple procedure to obtain expressions for the activity
coefficients from any model of the excess Gibbs energy and
for a specified number of components can be written in
about ten lines of Maple code. Students can investigate
different models of the Gibbs energy function without run-
ning the risk of getting the derivation incorrect.
Simple phase equilibrium calculations and the creation of
phase diagrams for binary systems using Maple have been
discussed by Taylor.171

Separation Processes
Multicomponent distillation simulations require the nu-
merical solution of a large set of equations: material bal-
ances, energy balances, and equilibrium (thermodynamics)
equations. These equations are sparse, nonlinear, and can
easily number in the hundreds and sometimes in the thou-
sands. The literature on distillation contains scores of papers
discussing methods of solving these equations (see Seader181).
Issues such as what form the equations should take, what
variables should be used, in what way should the equations
and variables be ordered, what variables should be com-
puted from what equations, and what numerical methods
should be used to solve (each subset of) the equations have
been thoroughly explored. Figure 3 shows an incidence ma-
trix (sparsity pattern) for a column distilling a nonideal bi-
nary mixture in ten stages. It is pertinent to point out that

Maple at Clarkson
In common with many other schools, undergraduate engineering
students at Clarkson University are required to take a programming
course. Until this year. the language taught in this course was Fortran
(although some students received instruction in Basic). No longer will
Clarkson students take courses in Fortran (unless it be by choice); the
introductory computing course has been completely revised, and
Clarkson students will now be programming using the computer alge-
bra system Maple.
The decision to abandon Fortran in fa' or of Maple %wa, not reached
lightly or without considerable debate within the engineering school
at Clarkson. Factors in favor of Maple included the fact that our
students have been using Maple in their calculus classes for the past
few years. It was deemed to be unw ise to introduce first-3 ear ,tudenis
to a second majoi software package such as MATLAB. In addition.
our site license permits us to install Maple on nearly all machines
owned by Clarkson, including the machines that all of our students
(regardless of discipline) are issued when they arrive on campus.
Clarkson s PC program has been in place for over a decade, and in the
fall 1993 semester new students \ere issued a computer with Maple
already installed on the hard drive.

although with Maple it is simple to explore different compu-
tational strategies, Maple also makes topics like this largely
irrelevant (even if they are interesting).
A Maple session to obtain the flows inside a distillation
column under the assumptions of constant molar overflow
is shown in the Appendix. It is a simple matter to investi-
gate alternative operating strategies such as what happens to
the flows when the feed is a saturated vapor or partially
vaporized liquid.
Four pages of Maple code is all that is required to obtain
numerical solutions to many multicomponent distillation
problems (including the derivation of all the equations).
This compares to the many hundreds (or even thousands) of
lines that would be needed to solve the same problem using
FORTRAN. Interlinked columns and nonstandard specifica-
tions also are simple to deal with (and only slightly more
difficult to solve). Figure 4 shows the McCabe-Thiele dia-
gram plotted from the results obtained by numerically solv-
ing the equations that gave us Figure 3.

Chemical Reaction Engineering
Chemical reaction engineering problems often require so-
lution of systems of coupled differential equations. Text-
book problems sometimes are specially simplified so that
the equations can be solved analytically. While such solu-
tions can also be obtained with Maple, it is no longer neces-
sary to simplify problems in this way. Software packages
that possess numerical methods for solving ODEs can be
used to solve more realistic problems. This point of view has
already been expressed by Fogler,t19 who uses Mathematica
and Polymath for solving reaction engineering problems.
The advantages of a CAS over a purely numerical method of

The fact that Maple can exchange files across platforms is a great
advantage for a computationally diverse campus like Clarkson; in
addition lo the huge number of PCs (over three thousand), there are
nearly two hundred IBM RS/6000 workstations (most of which are
grouped in instructional laboratories) and more than a few Suns and
The 1994 graduating class was the last one to have had no formal
instruction in Maple. Nevertheless, several chemical engineering stu-
dents had found opportunities to use Maple on occasions and more than
a few of them used Maple to solve problems in our Design II course
where, as it happened, the plant could not be simulated using a more
conentional process, flovheeting program.
The reaction of the first-year students to the first course on using
Maple as a programming language was more positive than any time in
the past several years when Fortran was introduced to them. Faculty
members in the engineering departments also are learning Maple with
a view to incorporating aspects of problem solving with Maple into
selected engineering courses. It will be interesting to follow the progress
of both students and faculty as Maple finds increased use in the engi-
neering school.

Chemical Engineering Education

Figure 1. Compressibility as a function
of reduced temperature (Tr) at a reduced
pressure of 0. 75 computed using the
Soave-Redlich-Kwong cubic equation of
state. The vertical axis is the real part of
the compressibility; the horizontal axis at
the back of the figure is the imaginary
part of the compressibility. All axes are
dimensionless. The region of three real
roots is shown between a region
possessing a real liquid-like root and
complex vapor-like roots and another
region with a real vapor-like root and
two complex liquid-like roots.

Figure 2. Compressibility surface as a
function of reduced temperature and
pressure. The vapor-like root from
Figure I is used in the construction
of this figure.

Figure 3. Incidence matrix for a small
distillation problem (2 components, 10
stages). The equations and variables are
grouped by stage leading to the familiar
block tridiagonal pattern shown in the
figure; however, it only takes two lines of
Maple to reorder the equations so they
are grouped by type rather than by stage.
The pattern is "upside-down" because
Maple plots cannot (yet) go from high to
low on any axis.

Figure 4. McCabe-Thiele diagram
constructed from the solution obtained
with Maple to a binary distillation
problem for a column with 10 stages
(including a total condenser). About 10
lines of Maple code were used to create
this figure. The various parts of the
diagram are created separately (in
different colors if so desired) and then
combined into a single figure using a
Maple command designed for this

203 3

Winter 1995

solution include the fact that the reaction system can be
analyzed symbolically and the material and energy
balance equations also may be derived.
Figure 5 shows multiple steady states in a
nonisothermal continuous stirred tank reactor (CSTR).
This plot was created using the parameters given by
Shacham, et al/.t10 who also considered the CSTR
dynamics and noted that the stability of each steady
state could be determined by computing the eigenval-
ues of the state matrix using Polymath. All of these
things are possible with Maple; however, Maple is
capable of evaluating the eigenvalues of the state ma-
trix symbolically as well!

Process Design
We do not think it requires too great a leap of imagi-
nation to expect students to set up and solve entire
process flowsheeting problems using Maple. The tech-
niques that are useful for material and energy balances
around simple units are readily applied to process
flowsheets with any number of units and their inter-
connections. Flowsheets containing recycle streams are
easy to handle.
While Maple will not replace (nor should it) special-
ized programs designed for large-scale plant simula-
tion, it is a useful tool for teaching students how
flowsheeting simulations work. The Maple program-
ming language encourages problems to be formulated
in a way that is reminiscent of equation-oriented
flowsheeting,E" but it is not too difficult to instruct
Maple to solve flowsheet problems using tearing or
simultaneous modular strategies. Maple's open inter-
face and powerful language make it possible for engi-
neers to create their own unit models in the Maple
programming language; others would then be able to
use them as plug-in modules in their own problems if
they were made available.

Figure 5. Multiple steady states in a nonisothermal continuous stirred tank
reactor (CSTR). The vertical axis is the heat duty (Btu/hrx 10lO), the x-axis
is the reactor temperature (K). The straight line represents the heat lost due
to cooling the reactor. The S-curve is the heat generated by the reaction.

More Maple
There are, of course, a number of ways that Maple needs
to be improved. It needs the ability to read and write binary
direct-access files (as is possible in FORTRAN and C). This
would make it possible to access databanks of physical prop-
erty data for direct use in engineering calculations.
Maple also needs improved symbolic capabilities. There
are many engineering formulas where it is necessary to
differentiate arbitrary sums and products (e.g., a sum or
product of indexed variables where the index range is non-
numeric, such as i = 1..c). The fact that Maple cannot do this
is a serious impediment to using it for certain important
problems (such as the derivation of thermodynamic proper-
ties of mixtures). We also need to be able to (elegantly)
exclude selected elements from sums and products of in-
dexed variables.
A great many problems in chemical engineering require
finding numerical solutions to large (or small) systems of
(sparse) nonlinear equations. This is not one of Maple's
strengths. The floating point solver built into Maple lacks
some of the features that would be useful. In particular, it is
not possible to provide the initial estimates or to control the
iteration history. On the other hand, it is possible to program
Newton's method in Maple so that all the user must provide
is a set of equations, a list of unknown variables, and a
starting point; Maple can compute the Jacobian symboli-
cally, thereby removing one of the major chores that must be
faced when using the method as a part of a FORTRAN
program. In fact, Maple is used for precisely this purpose in
some companies that write software for process engineering
simulations. Differential arc-length homotopy continuation
methods, recommended by Seader181 for solving difficult
nonlinear problems, may also be easily (and elegantly) pro-
grammed in the Maple programming language.
Unfortunately, Maple is currently many times slower than
compiled FORTRAN when carrying out large-scale numeri-
cal computations. It needs better (i.e., faster) routines for
purely numerical computing. A fast sparse linear equation
solver, for example, would go a long way to making large-
scale flowsheeting problems and multicomponent distilla-
tion problems a practical proposition. For now, a problem
can be set up with Maple and then translated into FOR-
TRAN (or C) so that the application can be compiled.
Many models in chemical engineering consist of large sets
of (stiff) ordinary differential equations (ODEs), mixed sys-
tems of differential and algebraic equations (DAEs), or par-
tial differential equations (PDEs). Fast numerical methods
for stiff ODEs, DAE systems, and for solving PDEs by, for
example, the method of lines would be very welcome.
Maple's graphics capabilities, although quite good, could
be improved by adding more basic plot types that are en-
countered often in (chemical) engineering (such as triangu-

lar diagrams and their three-dimensional counterparts).

In this article we have highlighted only a few ways in
which Maple can be used in chemical engineering educa-
tion. Additional applications are listed in Table 1. Some of
these worksheets are included in the Maple share library
(which is provided as part of Release 3 of Maple V), and
others are available from the first author. We have also
identified a few areas where Maple needs improved capa-
The fact is that Maple can have a significant impact in
almost all areas of chemical engineering education, but there
are some problems associated with using Maple in existing
Maple (in common with all other computer algebra sys-
tems) contains far more bugs than does even the worst FOR-
TRAN compiler. This is largely a reflection of the relatively
recent development of computer algebra and the fact that
some mathematics are hard to do with a computer.
Using Maple to derive expressions that are standard fare in
current engineering textbooks will rapidly demonstrate that
it is not always easy to get from Maple an answer that you
recognize. It may be quite trivial for Maple to solve your

Maple Worksheets in Chemical Engineering

1. Chemical Process Calculations
Material balances on single and multiple process units
2. Thermodynamics
Critical constants for cubic equations of statet'6
Phase equilibrium calculations and phase diagrams for ideal
Activity coefficients in binary and multicomponent systems
Gibbs free energy surfaces
Phase equilibrium calculations for nonideal systems
Flash calculations for ideal systems
Advanced flash calculations
Thermodynamic property relations and the Maxwell equations"41
3. Reactor Engineering
Material balances in tubular reactors
Isothermal tubular reactor (multiple reactions, numerical
Nonisothermal tubular reactor
Multiple steady states in a CSTR
CSTR dynamics
Fitting reaction rate coefficients to rate data
4. Equilibrium Stage Separations
Constant molar overflow in distillation
Multicomponent distillation-stage-to-stage calculations
Multicomponent distillation-simultaneous solution
McCabe-Thiele diagrams
5. Numerical Methods
Newton's method for systems of equations
Homotopy-continuation for systems of equations

Chemical Engineering Education

problem correctly, but it can require considerable skill in
expression manipulation in order to get a familiar result.
While this may not be important in solving original prob-
lems, it can make matching the results in established text-
books a frustrating experience. Perhaps we will have to get
used to new ways of looking at old results. The problem of
simplifying the chore of obtaining recognizable results re-
mains as a challenge for the computer-algebra community.
The exercises and examples in many standard textbooks,
(e.g., Felder and Rousseaul 21 and Reklaitist131) were designed
to be solved by hand. Many (if not most) of these problems
are far too simple if Maple is on hand to assist with the
problem solving; when you have solved one, you have solved
them all.
This brings us to some important questions: Do we want
students to use Maple for solving engineering problems?
Can the use of a CAS prevent students from mastering
essential skills that are better assimilated when solving prob-
lems by hand? Computer algebra is finding increasing use in
the teaching of calculus at many schools. It is impossible to
turn back the clock and abandon the use of computer algebra
in mathematics courses, thereby making its use in engineer-
ing a nonissue. It will not be possible to prevent students
from using tools they have learned once it has become clear
that they are useful. It will be up to us as educators to find the
proper time and place in our courses to introduce students to
engineering problem solving with Maple.

1. Taylor, R., and K. Atherley, "Chemical Engineering with
Maple," in Maple V: Mathematics and Its Application, R.
Lopez (Ed.), Birkhauser, Boston, MA (1994)
2. Char, B.W., K.O. Geddes, G.H. Gonnet, B.L. Leong, M.B.
Monagan, and S.M. Watt, Maple V Language Reference
Manual, Springer-Verlag (1991)
3. Gonnet, G.H., and D.W. Gruntz, Algebraic Manipulation
Systems, in Encyclopedia of Computer Science and Engi-
neering, 3rd Ed., Van Nostrand Reinhold (1991)
4. Walas, S.M., Phase Equilibria in Chemical Engineering,
Butterworths, Stoneham, MA (1985)
5. Sandler, S.I., Chemical and Engineering Thermodynamics,
2nd ed., Wiley, New York, NY (1989)
6. Taylor, R., and M.B. Monagan, "Thermodynamics with
Maple. I. Equations of State," Maple Tech, 10, 50 (1993)
7. Taylor, R., "Thermodynamics with Maple V. II. Phase Equi-
libria in Binary Systems," Maple Tech, 1(1), 83 (1994)
8. Seader, J.D., "The BC and AD of Equilibrium-Stage Opera-
tions," Chem. Eng. Ed., 19(2), 88 (1985)
9. Fogler, H.S., Elements of Chemical Reaction Engineering,
2nd ed., Prentice-Hall, Englewood Cliffs, NJ (1993)
10. Shacham, M., N. Brauner, and M.B. Cutlip, "Exothermic
CSTRs: Just How Stable are Those Steady States," Chem.
Eng. Ed., 28(1), 30 (1994)
11. Westerberg, A.W., H.P. Hutchison, R.L. Motard, and P.
Winter, Process Flowsheeting, Cambridge University Press,
Cambridge, UK (1979)
12. Felder, R.M., and R.W. Rousseau, Elementary Principles of
Chemical Processes, 2nd ed., Wiley, New York, NY (1986)
13. Reklaitis, G.V., Introduction to Material and Energy Bal-
ances, McGraw-Hill, New York, NY (1983)
Winter 1995

14. Adams, S., and R. Taylor, "Thermodynamics with Maple.
III. Thermodynamic Property Relations and the Maxwell
Equations," Maple Tech, In press (1995)

Maple session to compute flows inside a distilation column
Note that Maple input can be terminated by a ; or by :. In the latter case the result
is not echoed to the screen
>ns:=10: # Define the number of stages (1 is the condenser, ns the reboiler)
Material balances
>forj from 2 to ns-1 do TMB[j]:=V[j]+L[j]-V[j+l]-L[j-l]=0; od:
>TMB[1]:=D+L[1]-V[2]=0; #Material balance for the condenser
TMB,:= D + L, V,= 0
>L[ns]:=B: # The liquid flow leaving the reboiler is given the symbol B
>TMB[ns]:=V[ns]+L[ns]-L[ns-l]=0; # Material balance for the reboiler
TMBo:=V,, + B L,= 0
Energy balances
>forj from 2 to ns- do EB[j]:=Vj]*H[Vj+LUj]*H[Lj]-Vj+l]*H[Vj+l]
-Llj-1]*H[Lj-1]=0; od:
>FeedStage:=5: # Put a feed on stage 5
j:=FeedStage: # Modify the balances for the feed stage
TMBs:=V,+L,- V,- L4-F=0
EB,:=VH, + LHL, V6,,, L4HL FH = 0
>forj from I to ns do H[Vj]:=H[V]; H[Lj]:=H[L]; od: # Make enthalpies of
each phase the same
>H[F]:=H[L]: # Assume feed is saturated liquid
The overall material balance for the column can be obtained by summing the
material balances for all stages
>j:='j': sum(TMB[j]j=l..ns): TCB:=op(solve({")},F}));
>RRdef:=R=L[1]/D: RRdef; #Define the reflux ratio
Now we solve the material and energy balances for the flows. We create a set of
equations which includes the material balances for stages 2 to ns, the energy
balances for stages 2 to ns-I and the definition of the reflux ratio, R. Then we
create a set of unknown variables that we wish to compute (the flows) and invoke
Maple's solve command.
>result:=solve(Eqns,Vars): #output hidden to save space

>subs(TCB,result): # Eliminate F using the overall column balance
>collect(",D): # Tidy up by collecting terms in D
>Flows :=linalg[matrix](ns,2): # Create an array to hold the flows
>V[1]:=": #Hide V[1] since there is no vapor flow from the condenser
>forj from 1 to ns do Flows[,l] := V[j]; Flows[j,2] := LUj] od:

(I + R)D
(I + R)D
(I + R)D
(I +R)D
(I+ R)D
(I + R)D
(I + R)D
(I +R)D
(I + RID

(I+ R)D+B
(I+ R)D+B
( + R)D+B

The vapor flows are in the left hand column, the liquid flows in the right

S classroom


Are They Beneficial in Lecture Courses?

Clemson University
Clemson, SC 29634-0909

ne of the most difficult classroom tasks that a fac-
ulty member faces is getting students to become
active in the learning process. Several methodolo-
gies can be used to promote active learning, such as asking
questions and waiting (sometimes at length) for a response,
dividing the class into groups and asking them to address a
particular issue, having students work problems on the board,
or even having them give the lectures. But it is often difficult
to implement these techniques, particularly for classes with
a large number of students or technical courses that are
typically lecture-driven and which involve numerous equa-
tions and definitions. Additionally, the methodologies de-
scribed above consume class time-a precious commodity
when there is pressure to cover the topical content outlined
in the course catalog.
In an attempt to involve students in the learning process, I
have used some of the techniques described above in several
engineering courses I have taught. While the various activi-
ties seem to work well, they take up class time that is needed
to convey important technical information to the students.
To overcome this difficulty, I have used journal writing to
encourage students to become more active in the learning
process without sacrificing class time. The journals are used
for informal writing, giving the students an opportunity to
express their thoughts and questions both about course-re-
lated issues and about issues outside the scope of the course.
This activity is part of a university-wide "Communication
Across the Curriculum" effort to integrate writing (formal

Douglas E. Hirt is DuPont Assistant Professor
of Chemical Engineering at Clemson Univer-
sity. He received his BS and MS degrees from
Virginia Tech and his PhD from Princeton Uni-
versity. He was a NAS/NRC Research Associ-
ate at NASA Langley Research Center from
1989-90. His research interests include interfa-
cial phenomena, fiber-reinforced composites,
and thin polymer films.

Copyright ChE Division of ASEE 1994

and informal), speaking, and collaborative learning into uni-
versity curricula.
Journal writing is certainly not a new concept. It has been
used frequently in liberal arts courses, and it has also been
used for many years in science and engineering courses,
although the body of literature describing journal writing in
science and engineering is somewhat limited (for examples,
see Selfe and Arbabi[' and Schulz and Ludlow[21). Selfe and
Arbabi used journal writing in a civil engineering lecture
course, requiring students to write at least one page a week
about the course, including documentation of their progress
on design projects. The pages were submitted for review
three times during the ten-week quarter. I used a similar
approach except that the journal writing was not required.
Those who elected to write in a journal, however, turned in
multiple pages each week.
Schulz and Ludlow used student journals in a different
way in a graduate chemical engineering course. The students
were required to write journal entries (comments and ques-
tions) about assigned reading before attending each class,
and each submission was evaluated on a scale of 1 to 5. In
the work described in this paper, however, the journal en-
tries were made after class and the submitted pages were not
graded. Obviously, there are many ways to integrate infor-
mal writing into a lecture course (an excellent book on the
use of journals has been edited by Fulwiler[31; also see
Marwinel41). The purpose of this article is to provide the
framework that I used to incorporate journal writing into
chemical engineering courses and to use examples of student
writing to illustrate the benefits of informal writing.

The courses in which I have experimented with journal
writing were on the junior and senior levels, with enroll-
ments from thirty to sixty students. The students were not
required to submit journal pages-they did so on a volun-
tary basis. I provided instructions at the beginning of the
semester, and the weekly assignment was always the same.
At the end of the week, the students could submit a journal
Chemical Engineering Education

page from each previous class period, describing the most
important things) that they learned from those lectures. The
objective was to encourage the students to go back into the
class notes, review the material, extract the important infor-
mation, and transfer their thoughts about the subject matter
onto paper. The students were then free to write on any other
The journal pages from the previous classes were handed
in each Friday (or Thursday for a T-Th class) and were
checked off. I told the students that the pages would not be
graded for grammar, punctuation, or political correctness. In
addition, I stressed that the information that I read would be
held strictly confidential. The incentive for submitting all
journal pages was two bonus points at the end of the semes-
ter, making it possible for a student to have a final course
grade of 102. It should be emphasized again that the students
were not required to submit journal pages, but the two bonus
points were incentive enough so that over ninety percent of
the students in each course participated in this activity.
I collected the pages at the end of each week and read
them over the weekend. The students often asked questions
about the lecture material or about homework problems. It
took approximately one hour per fifteen students to read the
pages and to give written feedback. The feedback ranged
from short answers to lengthy derivations to "words of wis-
dom." I debated whether to edit the writing by making spell-
ing and grammar corrections, but I decided it would be too
time consuming and that it was not necessary for this type of
informal writing (although in a few glaring instances, I did
point out to students that they needed to improve their spell-

As I mentioned earlier, the goal of journal writing is to
encourage students to be more active in the learning process.
For this to be successful, the students must make an attempt
to truly understand the material and to ask questions when
something is not clear. The value of journal writing is illus-
trated below, using examples of student writing:

"You said that Newtonian fluids do not support a normal
stress. What about buoyancy?"
This is an excellent example of a student who takes
what he is learning now and relates it to concepts
learned in previous courses. I responded to this ques-
tion by describing the difference between a buoyant
(static) force and the normal stress that is generated by
fluid motion. It is gratifying to see a large number of
students relating the lecture material to material from
other courses, to their everyday experiences and to
their industrial work experiences.

* "Could you explain why the second homework problem

Winter 1995

... I have used journal writing to encourage
students to become more active in the learning
process without sacrificing class time. The journals
are used for informal writing, giving the students
an opportunity to express their thoughts ...

only has pressure from gravitational forces and the first
problem has to account for both pressures (I'm not sure I
understand the difference)."
This illustrates what is probably the most valuable as-
pect of journal writing. Students are often apprehensive
about asking questions in class or visiting the instructor
in person-journal writing provides a non-threatening
environment where they can ask their questions and
receive feedback. It is also valuable for the instructor
since the common questions that are raised in journal
pages can then be addressed in class the following week.
Journal writing thereby provides weekly feedback from

"I could follow the example you worked until you chose
the velocity scale. I don't understand why you chose the
maximum velocity ... Ah, yes, but the plot of v vs. x would be
a little different, right? I think I answered my own question."
This is the type of entry all instructors would like to see.
The student posed the problem and then came up with
the solution on his own (an example of "writing to
learn"). Unfortunately, this type of entry is not encoun-
tered often enough.

The following example illustrates student insight...
"I have finally begun to notice an interesting trend this
year among my classes. Finally, all the little bits and pieces
of information and skills we have been taught in the past 2.5
years are coming together. For example, we are using Laplace
transforms from math class to solve problems in process
dynamics. We can use numerical methods to solve problems
in all other classes. Also, in many instances, pchem and
thermo go hand-in-hand. I find this trend very encouraging
and helpful-courses reinforcing each other."
... and creativity
"I have discovered that I am a Bingham tanner. Yes,
once our design project was turned in, I suddenly found a
wee little bit offree time. So I used it to lie in the sun for 1.5
hours one day last week to rid myself of my ghastly color.
Even though it was a very sunny day, with a temperature in
the 70s (F), and even though I (unwisely) wore no sun-
screen, you would not have been able to tell (on the follow-
ing day) that I had ever gone outside. I figure that I must
reach some minimum value of sun exposure before I tan."

Students were also encouraged to write about other things,
particularly items related to engineering (newspaper or maga-


zine articles, TV shows, seminars, discussions with friends,
etc.) and about personal experiences if they so desired. Most
of the students stuck to course-related issues, but many
students wrote about unrelated topics such as "significant
others," interviewing, politics, other classes, tests, lack of
sleep, and (usually toward the end of the semester) the
philosophy of life. I did not discourage writing on these
subjects because I wanted to give the students freedom to
express their thoughts and to solicit my opinion/feedback on
a variety of topics.

In the spring semester 1993, Clemson University's Pearce
Center for Professional Communication administered a sur-
vey to my junior-level chemical engineering class to obtain
student feedback on journal writing. The students were over-
whelmingly positive about the experience (a few of the
students said that it took too much time and a few had not
written any journal pages). These juniors (and the seniors in
a subsequent class) said that journal writing
Helped identify key concepts
Helped them retain information more easily
Provided an opportunity to get clarification from the
Was a place to speculate about the practical applica-
tions of the course theory
Gave them a chance to know the professor more
The last item is particularly important. Journal writing not
only gave the students a chance to know me on a more
personal level, but it also gave me the opportunity to know
them better. This seemed to create a more relaxed atmo-
sphere in class, and the students seemed less afraid to ask
questions. In essence, journal writing indirectly led to an
active-learning environment in the classroom.
Initially, I was met with some resistance when I proposed
the idea of journal writing. The students were skeptical
("Honestly, I thought the idea of a journal in my engineering
class was stupid at first, but the journal proved useful.") and
frankly, I did not know what to expect. After several semes-
ters, however, the process has evolved into a valuable tool
for the students.
The following suggestions are provided to help make jour-
nal writing a more positive experience, both in terms of
student learning and faculty time.
In the beginning, too many students strayed from
the original intent of the journals and wrote about
anything and everything except the course. The
instructor must emphasize that the journal pages
will not count unless they focus on at least one
major concept from each lecture.
The other extreme occurs when students simply

transcribe class notes onto the journal pages. The
professor should emphasize that the students must
summarize the lecture by selecting, say, one to three
key points.
If faced with large classes or severe time con-
straints, ask the students to summarize the important
points from each week of lectures on one page-
they still identify key points, but they tend to write
less when they are limited to one page. As an
alternative, collect journal pages every few weeks
and selectively read some of them.l' This has the
disadvantage, however, of not obtaining and
providing weekly feedback.
Offer an incentive. It is amazing how many
students will participate when offered two bonus
points for submitting journal pages throughout the
semester (". the two bonus points for my average
were highly motivating").
As the instructor, I found student journal writing to be
quite beneficial. It gave me an opportunity to answer stu-
dents' questions directly with written feedback and also al-
lowed me to identify common questions that then could be
addressed during subsequent class periods. The journals also
provided a forum for students to express their opinions on a
variety of topics. Journal writing takes no class time and is a
valuable means of increasing communication between stu-
dent and instructor. More importantly, it motivates students
to review the lecture material and to write about it on a
regular basis, a practice we should all follow to document
our thoughts and ideas.
I close with a student comment on journal writing after the
semester was completed.

"Although at times it was a pain, it allowed me to get
things out. Strange, but I have continued writing jour-
nal pages. ..

I would like to thank Professors Art Young and Chris
Benson from the Pearce Center for Professional Communi-
cation at Clemson for their helpful advice and interaction.

1. Selfe, C.L., and F. Arbabi, "Writing to Learn: Engineering
Student Journals," Eng. Ed., 74, 86 (1983)
2. Schulz, K.H., and D.K. Ludlow, "Using Writing-To-Learn
Assignments in Chemical Engineering Courses," Proceed-
ings of the ASEE Annual Conference, Urbana-Champaign,
IL, pp. 776, June 20-24 (1993)
3. Fulwiler, T., The Journal Book, Boyton/Cook Publishers,
Portsmouth, NH (1987)
4. Marwine, A., "Reflections on the Uses of Informal Writing,"
in Writing to Learn Mathematics and Science, P. Connolly
and T. Vilardi, eds., Teachers College Press, New York, NY
p. 56 (1989) O
Chemical Engineering Education

Full Text