Chemical engineering education

De^^^^^^^^^^j^ran?7^^^
DPnesian

Featues .I

.. gn Ih atth

1996Annual Conference and Exposition
Sheraton Washington Hotel
Washington, DC
June 23-26, 1996

Plan now to attend the Annual Conference!

Conference highlights include:

* stimulating conference program featuring more than 350 technical sessions

* only forum specifically designed for all disciplines of engineering education

* bustling exposition with the latest in products and services for the engineering educator

* Awards Banquet, featuring the renowned political satire group "The Capital Steps"

* special Wednesday "Expo-Open House" for engineering students

* ideal networking environment for educators, researchers, administrators and related industry
professionals

* self-contained conference all sessions, meal events and exposition in one place!

* close to the nation's most historic monuments and museums - great place to bring the family!

ASEE PRISM will update you with ongoing conference
details. Don't miss out on this exciting annual event!

For more information, feel free to contact ASEE Meetings and
Conferences Department at (202) 331-3530.

EDITORIAL AND BUSINESS ADDRESS:
Chemical Engineering Education
Department of Chemical Engineering
University of Florida
Gainesville, FL 32611
PHONE and FAX : 904-392-0861
e-mail: cee@che.ufl.edu

EDITOR
T. J. Anderson
ASSOCIATE EDITOR
Phillip C. Wankat
CONSULTING EDITOR
Mack Tyner
MANAGING EDITOR
Carole Yocum
PROBLEM EDITORS
James 0. Wilkes and Mark A. Burns
University of Michigan
LEARNING IN INDUSTRY EDITOR
William J. Koros
University of Texas, Austin
PUBLICATIONS BOARD --

CHAIRMAN *
E. Dendy Sloan, Jr.
Colorado School of Mines

PAST CHAIRMEN *
Gary Poehlein
Georgia Institute of Technology
Klaus Timmerhaus
University of Colorado

MEMBERS *
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 of Michigan
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
Donald R. Woods
McMaster University

Winter 1996

Chemical Engineering Education

Volume 30

Number 1

Winter 1996

> DEPARTMENT
2 University of South Carolina, Michael D. Amiridis, Francis A. Gadala-Maria

> EDUCATOR
8 Deran Hanesian of the New Jersey Institute of Technology,
Zabel Sarian, Angelo J. Perna

0 LABORATORY
14 Application of Pseudo-Steady-State Approximation in Solving Chemical Engineering
Problems, Jordan M. Kmit, Dhananjai B. Shah

> CLASSROOM
20 Computational Results: How Reliable are They? A Systematic Approach to Model
Validation, Neima Brauner, Mordechai Shacham, Michael B. Cutlip
26 Applications of Some Modem Management Tools in Education, Richard Pollard
30 Application of Quality Management Techniques to ChE Processes,
Mary Ann Pickner, Bahman Ghorashi, Anne M. Ghorashi
40 Teaching Transport Phenomena with Interactive Computers to the Nintendo
Generation, Juan Eduardo Wolf Eduardo E. Wolf
58 On Using a Boundary Perturbation to Linearize a System of Nonlinear PDEs,
N. W. Loney
62 Design of Separation Units Using Spreadsheets, Mark A. Burns, James C. Sung
70 A Large-Group Senior Design Experience: Teaching Responsibility and Life-Long
Learning, Joseph A. Shaeiwitz, Wallace B. Whiting, Darrell Velegol
76 Freshman Design Course for Chemical Engineers, Carol McConica

> RANDOM THOUGHTS
34 The Warm Winds of Change, Richard M. Felder

> CLASS AND HOME PROBLEMS
36 Changing Vapor-Liquid Traffic in a Distillation Column, W. E. Jones, J. A. Wilson

> LEARNING IN INDUSTRY
46 EPIC: The Engineering Program for International Careers,
S. S. Melsheimer, C.E.G. Przirembel

> LABORATORY
50 Low-Cost Experiments in Mass Transfer: Part 1, I. Nirdosh, M.H.I. Baird

> CURRICULUM
54 On Selecting Approriate Control Valves for Pipework Systems, John R. E. Christy

1 7 Positions Available
> 19,69 Book Reviews
O- 45 New Books

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 32611-6005. Copyright 1996 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 responsibilityfor 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 32611-6005.

Department

University of

cr ,T

0

CAROLINA

MICHAEL D. AMIRIDIS, FRANCIs A. GADALA-MARIA
University of South Carolina Columbia, SC 29208
he University of South Carolina is located in Colum-
bia, the state capital of South Carolina. Columbia is
strategically placed in the geographic center of the
state, has a metropolitan population of 470,000, and com-
bines both the benefits of a big city and the charm and
hospitality of a small town. The area's sunny and mild
climate, combined with its lakes and wooded parks, provide
plenty of opportunities for year-round outdoor recreation. In
addition, Columbia is only hours away from the Blue Ridge
Mountains, the Atlantic Coast, Charlotte, and Atlanta-cit-
ies that serve as Columbia's international gateways.
Columbia's rapidly growing, yet balanced, economic base
includes both labor-intensive and technologically advanced
industries and serves as a model of the growth in the South-
east. Multinational corporations such as Allied Signal,
Amoco, BMW, DuPont, Eastman Kodak, Hoechst Celanese,
Michelin, Milliken, NCR, Roche, Sony, United Technolo-
gies, Westinghouse, and Westvaco have either research or
manufacturing facilities in Columbia or within a short com-
Copyright ChE Divsion ofASEE 1996

vearingen Engineering Center, home of
chemical engineering at South Carolina

muting distance. Our department benefits through partner-
ship with some of these industries and continues to develop
new relationships with others.
DEPARTMENTAL HISTORY
Engineering subjects were taught at what was then called
South Carolina College as early as 1840, and engineering
degrees were first conferred in 1882. The Department of
Engineering was formally established in 1908, was changed
to the School of Engineering in 1909, and finally to the
College of Engineering in 1961. The first program in chemi-
cal engineering was established in the Department of Chem-
istry in the mid-1920s.
In 1946, Broughton Leonard Baker came to the University
of South Carolina to form a Department of Chemical Engi-
neering as part of the School of Engineering. The program
was accredited for the first time in 1956 and has remained
fully accredited since then. The graduate program was estab-
lished in 1957. The Department has had only three chairmen
in its entire history: Professors Baker (1946-1978), Gibbons
Chemical Engineering Education

(1978-1993), and White (1993-
present). The Departmei
The Department has undergone tremendous trans
a tremendous transformation in the ten years-from
last ten years-from a small de- striving to maint
apartment striving to maintain the its undergrad
excellence of its undergraduate medium-size
program to a medium-size depart- strong under
ment with strong undergraduate graduate
and graduate programs. Several
factors contributed to this evo-
lution, including the increase in
faculty size and an attitude of
self-reliance.
When the faculty reached a low
of four members in 1982, teach-
ing the required undergraduate and
graduate courses did not allow
much time for research. Teaching
loads were three course sections
per semester, even for faculty with
research projects and outside fund-
ing. New faculty hires were given
a "reduced" teaching load of two
course sections per semester. Fac-
ulty size gradually grew to its cur-
rent size of thirteen tenured and
tenure-track professors and three
research professors, and as their
numbers increased, it became pos-
sible to seek and obtain outside
funding for the research activities
necessary for a strong graduate
program. Recently, research fund-
ing grew from five figures per year
to seven figures per year. Research The crew of FRED in
expenditures in the 1994-95 fiscal right: Tim Terwillig
year reached $3.9 million, the larg- niors), Travis Deal
est of any chemical engineering Price (senior), Kathy
department in the Southeastern oper), Professor Vin
United States, both on a total and MEhad Podellock seniort), Ralph
a per-faculty basis.
In 1988, then-Chairman Joseph
Gibbons organized a Chemical Engineering Industrial Advi-
sory Board consisting mainly of executives from large chemi-
cal companies in South Carolina. One of the Board's recom-
mendations was that the faculty meet with a facilitator to
define a vision for the department as well as the path for
achieving that vision. During these sessions it was perceived
that the Department could no longer merely depend on bud-
get increases from the administration or state legislature to
fuel its growth-that the department would have to "pull
itself up by its own bootstraps" in order to achieve future

at has undergone a
formation in the last
a small department
ain the excellence of
uate program to a
department with
graduatee and
e programs.

front of the facility. Left to
er and Nils Rasmussen (se-
(assistant engineer), John
Borg-Todd (project devel-
ce Van Brunt, Jason Aull
Haggard (engineer), and
r).

excellence. We thus resolved to
work hard and obtain the neces-
sary resources from outside the
University in order to make the
Department's vision a reality. The
"long-term planning" meetings
with a facilitator continue to take
place twice a year and have been a
tremendous asset in this quest for
excellence.

THE DEPARTMENT TODAY
> Mission: We will develop
high-quality chemical engineers
by continuously improving our
undergraduate and graduate
programs. We will conduct
world-class research and
innovative teaching, providing
an environment for professional
development, and be an effective
resource for industry, govern-
ment, and academia.

> Vision: We will be a Depart-
ment of Chemical Engineering
internationally known for
excellent undergraduate and
graduate teaching and research.
When Westinghouse took over
the operation of the Savannah
River Site from DuPont in 1989, it
agreed to set up a position of
Westinghouse Distinguished Sci-
entist at each of the three research
universities in South Carolina
(Clemson University, the Medical
University of South Carolina, and
the University of South Carolina).
The position carried with it a size-
able salary and discretionary funds.
Chairman Gibbons at that time had
been appointed Associate Dean of
Engineering for Undergraduate
Studies and a nationwide search

for a new chairman culminated with the appointment of
Ralph White, a University of South Carolina alumnus, as
both Chairman of the Department and the University's
Westinghouse Distinguished Scientist.
Currently, the Department has thirteen tenured or tenure-
track faculty members, including Professor Gibbons (who
remains very active in departmental issues despite his heavy
workload as Associate Dean). It has one of the youngest
chemical engineering faculties in the country (with an aver-
age age under 42 years old) and the one with the most

Winter 1996

assistant professors (eight). Our faculty reflects the growth
that has taken place at South Carolina over the past five
years and is a sign of the Department's youth, vitality, and
desire to succeed.
The research interests of the faculty cover the most dy-
namic areas in the spectrum of the chemical engineering
discipline, as can be seen from the following:
Ralph White
(PhD, UC-
Berkeley,
1977) works
in the areas of
batteries,
electrodepo-
sition, corro-
sion, electro-
chemical re-
actor design,
and numeri-
cal methods.
Vincent Van
Brunt (PhD,
University of
Tennessee,
1974) has re-
search pro- The department's faculty. Front
grams in the search), Carolyn Bolton researchi
areas of sepa- Kosanovich, John Weidner, and Ha
rations mod- Mike Matthews, Michael Amiridis
eling and Maria. Not shown: Joe Gibbons, Toi
chemistry. Emil Hanzevack (research).
Francis Gadala-Maria (PhD, Stanford, 1979) has interest
in the areas of rheology, composite materials, and
polymer processing.
John Van Zee (PhD, Texas A&M, 1984) does research in
the area of electrochemical engineering and, in par-
ticular, in the development of models for the nickel
electrode impregnation, the characterization of lithium
and hydride electrodes fabricated from fullerene-based
materials, and the electrochemical reduction of nitrate
and nitrite in alkaline wastes.
Michael Amiridis (PhD, University of Wisconsin, 1991)
has research interests in the areas of heterogeneous
catalysis, kinetics, and reactor design, and specifi-
cally in the emerging field of environmental catalysis.
Andrew Farell (PhD, University of Tennessee, 1990) does
research focusing on process modeling and control.
Karlene Kosanovich (PhD, Notre Dame, 1986) is study-
ing the development of batch and continuous chemi-
cal process systems emphasizing the aspects of mod-
eling and control.

row
h), R
rry P
, Joh
m Stc

Michael Matthews (PhD, Texas A&M, 1986) does re-
search in the areas of phase equilibrium thermody-
namics and characterization of complex mixtures,
supercritical fluid science and extraction, and diffu-
sion and adsorption.
Harry Ploehn (PhD, Princeton, 1988) has research inter-
ests in polymers, colloidal materials, and interfacial
phenomena
James Ritter (PhD,
SUNY Buffalo,
1989) focuses on cy-
clic adsorption based
separation processes
and new sol-gel de-
rived porous materi-
als for separation pro-
cesses and electrode
systems.
Thomas Stanford
(PhD, University of
Michigan, 1977) has
interests in reaction
engineering and pro-
cess control.

John Weidner (PhD,
(left to right): Branko Popov (re- North Carolina State,
alph White (chairman), Karlene 1991) works on the
loehn. Back row: Vince Van Brunt, application of chemi-
n Van Zee, and Francis Gadala- cal engineering prin-
nford, Andy Farell, Jim Ritter, and ciples in the study of
problems that com-
bine electrochemical technology and materials.
The faculty also includes the following research professors
who, while heavily involved in research, also find time to
bring their experience to the classroom:
Branko Popov (PhD, University of Zagreb, 1972) is study-
ing several aspects of electrochemical engineering
including electrochemical deposition, new materials
for electroplating, corrosion, and cathodic protection
for pipelines.
Emil Hanzevack (PhD, Northwestern, 1974) works in the
areas of neural modeling and control.
Carolyn Bolton (PhD, Princeton, 1989) has research inter-
ests in the application of chemical engineering exper-
tise to the prevention of adverse environmental im-
pacts and focuses on the areas of design for pollution
prevention and waste minimization from both the in-
dustrial and the municipal sectors.
Finally, the Department is still getting valuable advice from
emeritus professor Milton Davis, Jr., who enjoys retirement
at Hilton Head Island.

Chemical Engineering Education

PhD student Amanda Elmore
uses a liquid chromatograph
for analysis of complex
mixtures.

Professor Branko Popov and
Dr. Guanghong Zheng
(postdoctoral associate)
discuss the results of
an electrochemical
experiment. I

UNDERGRADUATE PROGRAM

Our undergraduate curriculum does not differ sub-
stantially from most other chemical engineering de-
partments. Therefore, only some of its most character-
istic features will be noted here. All engineering stu-
dents have a common freshman year, enabling them to
easily switch disciplines within the College of Engi-
neering before their sophomore year. Our students take
two Introduction to Engineering courses in their fresh-
man year, allowing them to have contact with engineer-
ing faculty, to get a better idea of what engineering is,
and to learn some of the skills they will use later. The
courses emphasize both computational and personal
skills, such as communications and teamwork.
The first chemical engineering course, Chemical Pro-
cess Principles, is scheduled for the first semester of
the sophomore year, while most of the remaining chemi-
cal engineering courses are taken by our students in
their junior and senior years. We have a three-semester
sequence of transport phenomena courses: one each in
fluid mechanics, heat transfer, and mass transfer. Our
students take a sequence of two three-credit chemical
engineering lab courses starting in the second semester
of their junior year. We also have a required safety
course. The curriculum comes together in the senior
year in a two-semester design sequence.
In our most recent long-term planning meeting, we
discussed and agreed on the need to revise the existing
curriculum to allow our students more flexibility in the
Winter 1996

.4

Professor
Karlene
Kosanovich
teaches in
the APOGEE
studio

form of elective courses. Such a revision is currently underway. One
example of an interdisciplinary engineering elective course intended
for engineering, science, math, and business majors is the Environ-
mentally Conscious Manufacturing course recently developed by a
team of chemical and mechanical engineering faculty. Lecture topics
include design for the environment, life cycle analysis, environmen-
tal economics and global competitiveness, legal and regulatory af-
fairs, and management of technological change.
Our undergraduate classes are relatively small. In the past ten years
we have graduated an average of about twenty-five undergraduates
per year; future plans call for an increase to about thirty-five students
per year. We believe that one of the strengths of our program is
the accessibility that our students have to our faculty, and small
class size contributes to that. Although we are experiencing tre-
mendous growth in our graduate program, we are committed to
making sure that growth does not jeopardize the excellence of
our undergraduate program.
Our students are mostly from South Carolina and vary greatly in
background. Indeed, it is not at all unusual to have students who are
working toward a second degree or who are more mature than the
average student at most universities. An Honors College within the
University helps us attract top students, who have done very well in

obtaining Goldwater and NSF Graduate Fellowships. About
twenty percent of our graduates go on to professional schools.
Finally, we have a very active co-op program that allows
students to alternate between periods in school and periods
in industry and to integrate industrial experience with their
education. We have found that co-op education is quite
beneficial to the students, and we encourage them to
participate in the program, thus paving the way to their
success after graduation.

GRADUATE PROGRAM
There are currently fifty-nine full-time students enrolled
in our graduate program. Of these, fifty-two are pursuing a
PhD degree and seven are enrolled in the MS/ME program.
This is a dramatic turnaround from just a few years ago
when the Department's graduate population consisted al-
most entirely of MS/ME students. It reflects the philosophy
of the faculty: the overall research productivity of a PhD
student is higher than that of an MS student. As a result, the
Department financially supports only those students who are
enrolled in the PhD program at a level of $17,700 to $22,500
per year plus a full tuition waiver.
An aggressive and well-organized recruiting campaign
over the last two years has been successful and, as a result,
the most recent incoming graduate classes have exception-
ally high caliber and promise. As an example, this year's
incoming class of fourteen consists of eight students with BS
degrees from U.S. institutions with an average GPA of 3.75
and six exceptional international students. Two of our new
students were designated by their undergraduate institutions
as Southeastern Research Fellows (SERF) and three received
University-level fellowships from the University of South
Carolina. We are confident that this trend will continue as
the news about the "revolution" taking place at the Univer-
sity of South Carolina continues to spread.
In addition to the standard graduate courses offered by
most chemical engineering departments (such as graduate-
level transport phenomena, kinetics and reactor design, ther-
modynamics, and process analysis), we are committed to
offering elective courses that cover a broad range of research
interests, including a three-course sequence in electrochemi-
cal engineering and corrosion and another one in separa-
tions. In development is a three-course sequence in process
optimization and control and a two-course sequence in inter-
facial phenomena, surface science, and catalysis. Additional
graduate courses are also offered in numerical methods,
supercritical fluids, and safety and loss prevention.
APOGEE (A Program of Graduate Engineering Educa-
tion) is another integral part of our graduate program. It was
developed by the University of South Carolina College of
Engineering in 1969 with the goal of providing a graduate
engineering education to practicing engineers in all parts of
the state through a combination of videotaped and live classes
6

via closed-circuit television with a 'talk back' phone system.
The Department has participated and supported APOGEE
from its inception, and today we have twenty part-time APO-
GEE students enrolled in the ME program and three enrolled
in the PhD program. From the teaching standpoint, the APO-
GEE setup creates some special challenges since the course
material must include enough practical examples to meet the
needs of a practicing engineer without compromising the
requirement to be on the cutting edge of science in a
given field. Additionally, APOGEE provides a unique
opportunity for full-time graduate students and faculty to

Professor Michael Amiridis and part of the ca-
talysis group with their in-situ IR cell. From left
to right: Sundaram Krishnamoorthy (PhD stu-
dent), Dr. Tiejun Zhang (postdoctoral associate),
Prof. Amiridis, and Ken Roberts (PhD student).

interact with practicing engineers and to be introduced to
their "real world" problems.

FACILITIES
The Department occupies over one-third of the new (1987)
220,000 square foot Swearingen Engineering Center. The
construction of this handsome building was made possible in
part by the generous contributions of our Department's alum-
nus John E. Swearingen (Class of 1938) who served as
Chairman of the Board of Standard Oil Company (Indiana).
This state-of-the-art building is fully equipped with modem
laboratories, including walk-in hoods, wet counters, large air
baths, furnaces, and an outstanding physical plant. In addi-
tion, the building has excellent teaching facilities, including
three modern TV studios that are used in the recording of the
graduate APOGEE courses.
A variety of specialized equipment is available to support
the experimental work conducted in the different research
programs. The Department's laboratories are equipped with
various spectrometers (e.g., FTIR, Raman, UV-Vis, atomic
absorption, ICP-mass spectrometer), chromatographs (e.g.,
GC, IC), and structure characterization instruments (e.g., X-
ray diffractometer, differential scanning calorimeter, BET
analyzers). In addition, several state-of-the-art analyzers are
dedicated to individual experiments in the various labs (e.g.,
Chemical Engineering Education

gas chromatographs, IR and UV-Vis analyzers, potentiostats
and AC impedance systems, thermogravimetric analyzers,
and a quartz crystal nanobalance). At other locations on
campus the faculty and students also have access to a scan-
ning electron microscope (SEM) with energy dispersive ca-
pabilities for elemental analysis (EDAX), transmission elec-
tron (TEM), scanning tunneling (STM), and atomic force
(AFM) microscopes, and a solid-state nuclear magnetic reso-
nance (NMR) spectrometer.
Available computer resources for modeling and data analy-
sis include an Intel Paragon supercomputer and a SUN
Sparcstation. A PC-based LAN, a DEC alpha, and other
SUN and VAX workstations are also used in different de-
grees for both research and instructional purposes.

RESEARCH
We like to say that "there is an explosion taking place in
the Department of Chemical Engineering at the University
of South Carolina." Well, here are the numbers to back up
that claim: in the last seven years the faculty size has more
than doubled (from six in 1988-89 to thirteen in 1994-95)
while the research expenditures have increased by a fac-
tor of 50 (from $72,000 in 1988-89 to more than $3.9
million in 1994-95). This level of research effort places
the Department at the top of the Southeastern schools
(despite its relatively small faculty size) and among the
top in the nation.
Such a dramatic change required determination and hard
work. The first key step in the process was the faculty's
recognition that change was necessary, followed by the de-
velopment of a strategic plan to make the change happen.
Other key decisions were to focus and strengthen certain
research areas and to hire junior but relatively experienced
faculty. All five new assistant professors added to the faculty
in the last three years had three to seven years of experience
either in industry or at another academic institution. Today,
the Department's research activities focus in five major ar-
eas: electrochemical engineering, separations, process mod-
eling and control, colloidal and polymeric materials, and
heterogeneous catalysis. Common elements that create op-
portunities for overlap and collaboration in more than one of
these areas are the environmental aspects and the materials
research elements that are present in many existing projects.
Among the highlights of our recent research activities are
the construction of the Filtration Research Engineering Dem-
onstration facility (FRED) and the establishment of the Cen-
ter for Electrochemical Engineering. FRED is a crossflow
filtration pilot plant constructed under the supervision of
Professor Vince Van Brunt with funding from the Depart-
ment of Energy. Its size is unique for a facility located at a
university and it offers many opportunities for potential peda-
gogical uses in addition to the crossflow filter tests that are
conducted for DOE. We anticipate that FRED will be used
Winter 1996

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

THE UNIVERSITY OF TEXAS AT AUSTIN

Faculty Position in Chemical Engineering-The Uni-
versity of Texas at Austin. Applicants must have sincere
interest in teaching, research, and professional activity,
and a PhD or satisfactory progress toward completion of
requirements for a doctoral degree in Chemical Engineer-
ing. Duties will include teaching undergraduate and gradu-
ate courses, and supervising graduate research. Send cur-
riculum vitae, list of three references, transcripts, and
statement of teaching and research objectives to: Dr. W.J.
Koros, Chairman, Department of Chemical Engineering,
The University of Texas at Austin, Austin, Texas 78712-
1062. Affirmative Action Employer. Women and minori-
ties are especially invited to apply.

by our Department's faculty, as well as by colleagues
from other departments in the College of Engineering, to
demonstrate applications in separations, fluid mechanics
and hydraulics, signal analysis and control, mixing, sched-
uling and sequencing, safety analysis, and ISO 9000 and
14000 certification.
The Center for Electrochemical Engineering was estab-
lished in the summer of 1995 and is currently funded by the
State of South Carolina, the DOE's EPSCoR (Experimental
Program to Stimulate Competitive Research) program, and
the industry. The Center, under the direction of Professor
Ralph White, brings together researchers from both the
Chemical Engineering and the Chemistry Departments at
the University of South Carolina in addition to faculty
members from Clemson, North Carolina State, and the
University of Virginia. The focus of the work is on the
development of new electrochemical power sources.
Projects are directed in four "thrust areas," namely de-
sign, materials, separators, and sensors.

SUMMARY
The Department of Chemical Engineering at the Univer-
sity of South Carolina has traditionally been a small, high-
quality undergraduate program. While we plan to maintain
this aspect, in the last few years we have also made a signifi-
cant effort to establish a first-class graduate research pro-
gram. The signs of this change, including the productivity of
the faculty, the quality of the graduate students, and recogni-
tion by the chemical engineering community, indicate that
we are moving fast in the right direction. There is still much
to be done; these are exhilarating times in Columbia! O

r educator

DERAN HANESIAN

of the New Jersey Institute of Technology

ZABEL SARIAN, ANGELO J. PERNA1"
State University of New York Oswego, NY 13126

Deran Hanesian has taught chemical engineering and
chemistry for thirty-three years. Reflecting back, he
ponders, "How did my career evolve? There were
no established goals, no firm plans." Perhaps his life's work
and effort fulfilled the hopes and dreams of Armenian immi-
grant settlers in America who had survived a massacre.
Deran was born in Niagara Falls, New York. The city had
a heavy concentration of electrochemical and electrometal-
lurgical industries offering an abundance of opportunity for
unskilled workers that attracted a hopeful immigrant popula-
tion. Deran's parents, Vahan and Anna, had both survived
the 1915 genocide of the Armenian people by the govern-
ment of Ottoman Turkey. Vahan was the sole survivor of a
large family, and the victims of the massacre included his
young wife and son. He escaped to South America and
eventually worked his way north to the United States. Deran's
mother was orphaned by the massacre at the age of twelve,
with only a few members of her family surviving. Anna,
together with an older sister and her son and a younger sister,
eventually made it to Aleppo, Syria. (Twenty years later, a
younger brother, who was six years old at the time of the
genocide, was miraculously found on the Turkish-Syrian
border, and in 1973, following fifty-eight years of separa-
tion, he and Anna were reunited in Niagara Falls.)
After searching for his first wife and family for ten years.
Deran's father gave up hope that his family was alive, and on
a subsequent trip to Syria he met and married Anna, bringing
her to America as his bride. Eventually, Anna brought her
younger sister to Niagara Falls, where she also married and
raised her own family. Together, the two families raised
seven children in a five-room apartment.
In 1927, Deran was the firstborn child of Vahan and
Anna. After the Stock Market crash of 1929 and the Great
Depression of the 1930s, many industries in Niagara Falls
laid off workers, and Vahan's employer, the Aluminum Com-
pany of America, completely shut down, making life even

' Address: New Jersey Institute of Technology, Newark, NJ 07102

... when he asked Mr. Field [the
Boy's Club director] to serve as a
reference on his [college] application,
Mr. Field suggested that Deran become a
chemical engineer because "civil engineering
was too political." Deran immediately
crossed out the word "Civil" on the
application, substituted "Chemical"
above it, and mailed it to Cornell.
With this stroke of a pen,
a career began.

more difficult for the immigrant families. Only intermittent
jobs with the WPA were available to him until ALCOA
reopened in 1940 because of World War II. Consequently,
Anna never realized her wish to bring her older sister and
sister's son to America, and in 1946 they emigrated from
Beirut to settle in the Soviet Republic of Armenia.
Despite the poverty of the depression years, both families
maintained a hopeful atmosphere. All of the children began
working at an early age, contributing their earnings to the
family's purse. Armenian was the language spoken at home;
English was learned in kindergarten.
Although the children excelled in school, economic hard-
ship excluded any hope for a college education. Following
the suggestions of his junior high school guidance teacher,
Deran at first concentrated on commercial studies such as
bookkeeping, business arithmetic, shorthand, and typing,
with a factory office job as the occupational goal. Fortu-
nately, he enjoyed mathematics and loved history, and he
simultaneously studied these subjects. During the second
half of his eleventh year, however, he changed to a math and
science course of study, and in 1945 he graduated from
Niagara Falls High School with high honors.
World War II was raging during Deran's high school

Copyright ChE Division ofASEE 1996

Chemical Engineering Education

Deran against the backdrop of
Niagara Falls

to the present...
Deran the chemical engineering
educator and researcher

Winter 1996

years. To participate in the war effort, he
worked as an inspector at the Auto-Lite Bat-
tery Corporation after school during the school
year and in the summer months. He continued
working at the plant after he graduated from
high school, expecting to be drafted into mili-
tary service when he reached the age of eigh-
teen in September.
Late in July, however, Deran's mother sug-
gested he apply for admission to a college
despite the family's lack of financial resources.
Deran wanted to be an engineer, and the only
school he knew in close proximity to home
that offered engineering was Cornell. While
in high school he had built a transit for survey-
ing and was very interested in trigonometry,
so he completed Cornell's application for its
civil engineering program. Fortuitously, he be-
longed to the Niagara Falls Boys' Club where
activities after school included working, play-
ing billiards, basketball, and sandlot football.
He had become good friends with the club
director, Mr. Field, and when he asked Mr.
Field to serve as a reference on his applica-
tion, Mr. Field suggested that Deran be-
come a chemical engineer because "civil
engineering was too political." Deran im-
mediately crossed out the word "Civil" on
the application, substituted "Chemical"
above it, and mailed it to Cornell. With this
stroke of a pen, a career began.
Deran was accepted into Cornell's chemical
engineering program. Cornell was on a war-
time schedule, and the semester was to begin
in November. In late September, however,
Deran became eighteen, registered for the draft,
passed the physical, and was summoned to
active duty in the U.S. Army. He wrote to
Professor Fred H. "Dusty" Rhodes, the de-
partment director at Cornell, about his cir-
cumstances. Dusty told Deran to complete mili-
tary service, assuring him there would be a
place at Cornell when he returned. He also
told Deran not to worry about money, assur-
ing Deran there was financial support for prom-
ising students. Dusty's words proved to be
true, for in Deran's last two years at Cornell,
he was awarded a scholarship.
Deran served as Private First Class, first in
the Medical Corps at Fort Dix, New Jersey,
later in the Corps of Engineers at Fort Belvoir,
Virginia, and finally, at the Yuma, Arizona,
Engineer Test Center. He was discharged from

the Presidio in San Francisco.
After his discharge from the Army, Deran visited a cousin's
farm in Fresno, California, and the visit convinced him that
his family should move from Niagara Falls to Fresno and
become grape farmers. When he returned to Niagara Falls
and presented this "great" idea to his father, he was un-
equivocally told that he was to go to college. Education was
greatly valued by his parents. Deran's mother had been
completely deprived of any
schooling, and although his fa-
ther had only completed the
sixth grade, he was a poet and
an intellectual who loved to
read and write. Deran remem-
bers these dreams and won-
ders-what would his life have
been like as a grape farmer?

UNDERGRADUATE
SCHOOL
Deran entered Cornell armed
with the GI Bill of Rights and a Deran, with his senior
War Service Scholarship, but absorption experii
in his freshman year, his engineering
father's employer again shut
down, so Deran's GI stipend
had to help support the family. From the time of his arrival at
Cornell until he completed the five-year program, Deran
worked at various part-time jobs. Such a heavy workload
made undergraduate study extremely difficult. During the
freshman orientation for the approximately 135 new stu-
dents, Dusty stated, "All of you came to Cornell from the top
of your high school classes. Not one of you had less than a
ninety percent average in high school. However, look at the
two people on your right and the two on your left, shake their
hands, and tell them you probably won't see them at gradua-
tion." Dusty was right-only thirty-five graduated.
Deran held on despite the fact that in his third year his
father passed away and the family finances became ex-
tremely desperate. Although those years were difficult, Deran
was determined to become a chemical engineer. The five-
year program was completed on schedule.

THE EARLY DU PONT EXPERIENCE
At the time of Deran's graduation, the chemical industry
was in rapid growth. He accepted a hometown position at the
Electrochemical Department plant of E.I. du Pont de
Nemours, Inc., and became involved in production and de-
velopment work at the Adiponitrile plant. He was also occu-
pied at the Lysine manufacturing semi-works plant doing
research on a new product where he was the supervisor of
the first three steps of a seven-step process. Those years at
du Pont impressed upon Deran the importance of techni-

s, op
men
g la

cal ability coupled with good communication skills, both
written and oral. As a field engineer, he was always
thankful for the discipline to which the faculty at Cornell
had subjected him.
The du Pont years were challenging, but Deran found that
climbing the production management ladder was an unsatis-
fying pursuit. He needed something more. While problems
appeared and then disappeared in production work, it was
often not understood how the
problem developed or how it was
resolved. Delving into problems
more deeply and solving prob-
lems scientifically were pre-
cluded by the need to move on
to solve a prevailing new prob-
lem. This frustrating aspect of
production work was neverthe-
less offset by learning more
about the nature of chemical
equipment. Deran also developed
rewarding interpersonal relation-
erating the ammonia ships with plant operators and
tin the chemical maintenance workers who had a
boratory. great deal of practical learning.
In applying the theoretical
knowledge gained at Cornell to
real industrial problems, Deran's understanding of chemical
engineering was greatly expanded.

GRADUATE SCHOOL
Deran's temperament for research motivated his desire
for graduate study. In late 1956, he went to Cornell to
discuss the matter with Chuck Winding, the director of
chemical engineering. Deran remembers, "Chuck was
kind to me, he was sensitive to my problems as an under-
graduate and to my current desires, and he accepted me
into the graduate program."
The graduate years, which began in September of 1957,
were very different from Deran's undergraduate experience.
With five years of strong industrial experience behind him
and more adequate finances, graduate school proved to be
less difficult. He recalls, "I worked very hard the first
semester, having forgotten how to integrate and differen-
tiate. I had to relearn almost everything." Progressively,
however, things became easier, and he was soon ac-
cepted into the PhD program.
Deran's thesis for the PhD degree was on "Simultaneous
Heat and Mass Transfer in a Packed Liquid-Liquid Extrac-
tion Column." Temperature differences of 0.1 0C between
phases were measured by the newly developed thermistors.
His advisor, Professor Robert von Berg, had many years of
experience at du Pont and in the nuclear industry. He says he
is indebted for much of his graduate training to Professors

Chemical Engineering Education

von Berg and Peter Harriot, for whom he
was a teaching assistant in the chemical
engineering laboratory. Professors von
Berg and Thorpe also guided Deran in
teaching the sophomore stoichiometry
course. Deran completed all graduate stud-
ies for the PhD requirements in a brief
period of three years.

RESEARCH AT DU PONT
Following graduate studies, Deran de-
cided to return to a research position at du
Pont's Jackson Laboratory, Chamber
Works, Organic Chemicals Department,
in Deepwater, New Jersey. Given two re-
search projects from which to choose, his
preference was to work on the develop-
ment of dielectric gases for electrical trans-
formers based on chlorofluorocarbon
chemistry. The work was a combination
of fundamental research, laboratory ex- Deran as dep
perimentation, field work, and market de-
velopment. With his technicians, he designed and built a
new laboratory. Deran became involved with the various
technical societies concerned with the establishment of text
standards. Traveling nationally to numerous professional
meetings brought him into contact with scientists from com-
peting companies. It was extremely enjoyable and challeng-
ing work. "Those were great years in the Wilmington, Dela-
ware, area," Deran remembers, "I was in my mid-thirties and
for the first time in my life I felt confident and secure."

THE NJIT YEARS

Despite his success at du Pont, Deran felt he should try
teaching. "If I didn't like it, I could always return to industry
because I had received excellent performance evaluations at
du Pont." Deran preferred to locate in the New York or
Boston area. He remembered that his friend, Tom Weber,
had obtained an MS degree at Newark College of Engineer-
ing (NCE) in New Jersey while he worked at Exxon, so
Deran applied there, and after an interview with Charlie
Mantell, the department chairman, he was given an offer on
the spot. Deran asked Charlie if he could have some time to
think it over because other schools had also expressed an
interest. Charlie replied, "Sure, take a week"-those who
knew Charlie will understand his response. Deran reasoned
that a "bird in the hand is worth two in the bush," and
accepted the offer.
It was 1963, and the first step of his teaching career had
thus been taken. It is ironic that Charlie started Deran on his
teaching career-Charlie was a renowned consultant and
authority in electrochemical engineering, and one day he
and Deran were having an informal conversation when the
Winter 1996

artn

subject changed to Niagara Falls. Charlie
casually informed Deran that before
coming to NCE in 1948, he had served
as a consultant for ALCOA, and that on
two occasions, in 1929 and again in
1947, he had been involved in a recom-
mendation to shut down the Niagara
Falls plant. Deran had come face to face
with one of the men who had been part
of the fateful decision that resulted in so
much suffering for his family.

Teaching
Students at NCE were primarily mem-
bers of local working-class families. The
college was focused toward its under-
graduate school and teaching loads were
very heavy. Deran was assigned an eigh-
teen or nineteen contact hour load per
semester, with four to five preparations.
tent chairman. His assignment included teaching chem-
istry courses. Since NCE did not offer a
degree in chemistry, the chemists teaching higher level chem-
istry courses were a part of the chemical engineering depart-
ment. Deran states, "I quickly learned what teaching was all
about and how to teach. Each student had special needs
that must be quickly recognized and addressed in order
to reach them. I learned that if you couldn't reach them,
you couldn't teach them." He taught every course in the
chemical engineering curriculum at the undergraduate
level, taught chemistry courses, developed graduate
courses in chemical reaction engineering, and developed
undergraduate technical electives.
In 1966, a large step change occurred. NCE was told that
unless it reduced teaching loads to twelve contact hours, it
would lose its accreditation. Many faculty were hired, and
Deran, who had many years of teaching and industrial expe-
rience, was called upon to guide his new colleagues.
Deran had volunteered to develop the Process Dynamics
and Control Course, and in 1966 received a NSF grant to
build a related laboratory. In 1968, the department was noti-
fied that it would have a new building. Deran was a member
of the committee assigned to design the new building and to
construct a new chemical engineering laboratory. He and Dr.
Perna visited numerous universities, selected appropriate
ideas, and synthesized the concept of the current laboratory
and laboratory course. With success, substantial funding
was received from industrial, state, and federal sources to
implement these concepts.
The years 1967-75 were heavily involved with numerous
research projects, the development and teaching of under-
graduate courses, new graduate courses, and building the
new chemical engineering laboratories. More than thirty unit

operations experiments were built in the four summers from
1972 to 1975. The late Professor Vinnie Uhl, on a visit from
the University of Virginia, praised the laboratory, stating
"the only other laboratory of comparable
quality that I have seen was in Germany."
Faculty from other schools came to see the Teachi
facility and used it as a model for their own Deran,
laboratories. Deran's continuous involve-
ment and teaching experiences in the chemi- profession
cal engineering laboratory resulted in pub- flexibilil
location of an extensive Chemical Engineer- industrial
ing Laboratory Manual, followed a few
years later by a second edition.
In 1988, Deran became reinvolved in
teaching stoichiometry to sophomores and various E
transfer students. He also undertook the and
challenge of teaching a large lecture class opportui
in chemistry that was required for all as consu
freshmen students. as consu
industry
Deran and Angie Perna are presently par-
ticipating in a National Science Foundation over
consortium of ten universities and have de- Concu
veloped a Chemical Engineering Measure- industry
ments Laboratory for freshmen as part of experiez
the Fundamentals of Engineering Design
program. This course is also offered in the reservoir
Institute's Summer Academy for outstand- new col
ing high school students. During recent sum- NJIT c
mers, they have also been working with 9-
14 year old girls in the FEMME program,

teaching the young ladies the basic principles of chemical
engineering in the Unit Operations Laboratory. A grant has
also been received for the Advance Technology Center pro-
gram in pollution prevention.

Administration
A new era of significant changes in the department began
in 1975. The name of the institution, Newark College of
Engineering, was changed to New Jersey Institute of Tech-
nology (NJIT), and Deran was selected as chairman of the
combined departments of chemical engineering and chemis-
try. A few years later it became the Department of Chemical
Engineering, Chemistry, and Environmental Science, grant-
ing degrees in all of these fields. At its peak, it was among
the largest departments in the United States, with approxi-
mately forty faculty, twenty technical and support staff, and
numerous adjunct faculty.
One of the department's distinguished achievements dur-
ing Deran's tenure as chair was the growth of the under-
graduate scholarship, merit award, and graduate fellowship
program. The department's aid program grew rapidly and at
its peak in the mid- 1980s awarded seventy-five merit awards,

fifteen scholarships, and seventeen graduate fellowships for
a total of $57,100 in aid to students. Almost thirty percent of
all chemical engineering students were receiving some fi-
nancial aid at that time.

ing allowed
a registered
mal engineer,
ty to continue
1 research for
summers at
mnt and the
xxon affiliates
presented
cities to serve
itant to other
ial concerns
the years.
rrently, his
ial research
ice became a
that provided
urses for the
urriculum.

Research
The launching of the Soviet Union's Sputnik in 1957
prompted a surge of growth in NCE's research and graduate
programs, and Deran was involved from the start. In a single
year (1972), one of the three PhD degrees and eight of the
twenty-five MS degrees granted were given to his students.
In 1964, Deran became involved in an adhesives study
with one of his fellow organic chemists and an orthodontist.
The work concerned replacing metal appliances on teeth
with plastic pieces adhering to the tooth's surface. Concur-
rently, he began to develop a research program primarily in
areas of reaction kinetics, including the effects of ultrasound
on reaction rates and in fluidization.
In the summers of 1964-66, Deran returned to du Pont's
Jackson Laboratory and performed research on a new pro-
cess to produce chlorofluorocarbons. This effort led to the
construction of a plant in Texas. Many of the problems
encountered during the research later served as the basis of
student theses. Beginning with the summer of 1967, he
worked with Exxon's Bayway Refinery developing a kinetic
model for the ethane-propane pyrolysis furnace, and this

Chemical Engineering Education

Deran has always been active in profes-
sional societies and took special interest
in student societies such as AIChE, ACS,
Omega Chi Epsilon, and the Biochemical
Club. From 1971 to 1990 the NJIT AIChE
chapter was voted by National AIChE as
an outstanding chapter in the United States,
a national record of twenty consecutive
years that still holds.
The laboratory focus during Deran's ten-
ure as chairman turned to on-line, rapid
data reduction and analysis. As part of the
CACHE/NSF project on the Modular In-
struction Series, Deran was invited to sub-
mit two modules on reaction kinetics and
chemical reactor design. He coauthored
the two modules with faculty from India
and Venezuela that are part of the
AIChEMI Series.
In the thirteen years that Deran was de-
partment head, he focused on being an
effective administrator while continuing
to teach a half-time load of two courses.
The graduate courses that he developed
attracted students and served as a source
for his research program.

work eventually led to sev-
eral research projects in-
volving simulations and the
development of new under-
graduate electives.
After more than a decade
as department chair, in 1988
Deran accepted a summer
position at the Center for
Plastic Recycling Research
(CPRR) at Rutgers, The
State University of New
Jersey, and continued his
work there during the
1988-89 academic year. Deran, here receiving.
When the center's direc- Award for Teaching Exce
tor later resigned, Deran
was asked to serve as Acting Deputy Director because of
his administrative abilities and experience. He accepted,
and his relationship with CPRR continues to the present.
He has been involved in research in all areas of recycling
and in soil remediation

Service
The trust placed in Deran by the NJIT faculty was ex-
pressed by his election to the office of Faculty Council Vice
Chairman and Chairman. He also has served on the Institute
Promotion and Tenure Committee. He has served and held
leadership offices on numerous AIChE and ASEE commit-
tees over the years, and is a Fellow and life-member of
ASEE and a Fellow and Emeritus Member of AIChE. Mem-
bership in various other professional societies include Ameri-
can Chemical Society, Society of Plastics Engineers, Omega
Chi Epsilon, Alpha Chi Sigma, Fulbright Association, and
the American Association of University Professors.
Deran has also been active in the community and serves
both the Diocese of the Armenian Church and the Parish
Council of St. Sarkis Armenian Apostolic Church in
Niagara Falls.

INTERNATIONAL OUTLOOK
Deran has been a tireless teacher to a wide international
community of students. In 1978, he taught at the Algerian
Petroleum Institute, in 1981 he taught at the University of
Edinburgh in Scotland, and in the spring of 1982 he received
a Fulbright grant to teach at the Yerevan Polytechnic Insti-
tute in the Soviet Republic of Armenia., where he lectured in
Armenian on chemical engineering subjects. He was instru-
mental in establishing an exchange agreement between NJIT
and the Polytrechnic Institute.
When the Iron Curtain was lifted, allowing restricted travel
to the USSR, Deran was one of the first to travel to the

NJIT
llenE

former Soviet Republic of
Armenia in 1962. He
found his aging aunt who
had rescued and raised his
mother following the 1915
genocide by the Ottoman
Turks. Since that meeting,
Deran has returned fifteen
times to the tiny land that
is all that remains of his
decimated ancient heri-
tage, assisting in the de-
velopment of a democratic,
independent Armenia.
's Robert W. Van Houten
ce on graduation day, 1977.
The most cherished
honor of Deran's teaching career was receiving NJIT's Rob-
ert W. Van Houten Award for Teaching Excellence in 1977,
given annually to the professor chosen by vote of alumni
who have graduated within the closest five-year period.
Other awards followed: ASEE's Mid-Atlantic, AT&T
Foundation Award for Excellence in Instruction of Engi-
neering Students; ASEE's Centennial Certificate; the John
Fluke Award for Excellence in Laboratory Instruction; and
the first recipient of NJIT's award for "Outstanding Profes-
sional Development by a Tenured Faculty Member," given
to "tenured faculty members who have demonstrated signifi-
cant achievement in teaching effectiveness and innovation
over a substantial period of time...."
Teaching allowed Deran, a registered professional engi-
neer, flexibility to continue industrial research for seven
summers at du Pont and the various Exxon affiliates and
presented opportunities to serve as consultant to other indus-
trial concerns over the years. Concurrently, his industrial
research experience became a reservoir that provided new
courses for the NJIT curriculum.
The special insights Deran gained have been disseminated
through publication in professional journals and through
oral presentations at professional society meetings. This reci-
procity between teaching and industrial experience has been
a rich, synergistic relationship for Deran, leading to global
and humanitarian endeavors.
Deran looks back in amazement at the evolution of his
career over the years It was a cooperative effort, an invest-
ment in the future of one person by many others. Deran says,
"I was lucky-I had good family, good friends, good teach-
ers, close colleagues, and dedicated coworkers. Our students
have always felt a strong commitment and sense of belong-
ing to the department and the institution that was theirs. For
all this, I am grateful." 0

Winter 1996

r.M laboratory

APPLICATION OF

PSEUDO-STEADY-STATE

APPROXIMATION IN SOLVING

CHEMICAL ENGINEERING PROBLEMS

JORDAN M. KMIT, DHANANJAI B. SHAH
Cleveland State University Cleveland, OH 44115

he concept of pseudo-steady-state approximation
(PSSA) has been used quite extensively in solving
chemical engineering problems where different steps
in an overall process take place at different time scales.
Ideally, the process takes place in two consecutive steps,
and the PSSA is invoked to eliminate the unsteady-state
nature of one of the steps. The effect of invoking such an
approximation is to considerably simplify the solution
of the problem.
The PSSA solution is an approximate solution and needs
to be carefully examined to determine its accuracy and to
identify the physical conditions under which the solution is
valid. In cases where the rigorous solution is available, it can
be compared with the PSSA solution to determine the error
involved in invoking the PSSA concept. In cases where the
rigorous solution is not available, application of pertur-
bation method may be necessary to determine the accu-
racy of the PSSA solution.
In this paper, we have chosen a number of examples from
fluid mechanics, heat and mass transfer, and reaction kinet-
ics to illustrate the principle of the technique, when and how
it can be applied, the importance of checking the accuracy of
the PSSA solution, and the physical conditions under which
the method may be used.
The general procedure is to identify the steps in the overall
process and the corresponding time scales or time constants
(e.g., cl and X2). The problem may then be examined under
the limiting conditions of either
TI / T2 << 1 or 'T 1/2 >> 1
This procedure is illustrated with the following examples,
the majority of which involve moving boundary problems.
The examples chosen are such that their more rigorous ana-
lytical solutions are available, and in each case, the PSSA
Copyright ChE Division ofASEE 1995

solution is checked against the more rigorous solution to
determine the conditions under which the PSSA solution is
valid.

EXAMPLE 1
Draining of a Tank

Draining a tank filled with a liquid is an unsteady-state
problem often simplified by the application of the PSSA
concept. Here, a tank having height H and radius R is filled
with a liquid to a height h. At time t=0, the liquid in the
tank is allowed to drain through a hole of radius R0 in the
bottom of the tank. We want to determine the time it
takes to empty the tank.
There are two time constants: they are determined by V,,
the velocity of the liquid level in the tank, and V2, the efflux
velocity through the hole. The exact solution of this problem
involves the solution of an unsteady-state mass balance

Jordan Munn Kmit received her Bachelor's De-
gree in Industrial Engineering from Case Western
Reserve University in 1987. She worked for Argo
Tech Corporation four years prior to entering
Cleveland State University for graduate studies in
chemical engineering. She received her Master's
Degree in 1994 and currently works at the Cleve-
land Advanced Manufacturing Program as a
Waste Reduction Engineer.

D.B. Shah is Associate Professor of Chemical
Engineering at Cleveland State University. He
obtained his BChE degree from the University
of Bombay and his Master's and PhD degrees
in chemical engineering from Michigan State
University. His research interests are in adsorp-
tion and diffusion in zeolites, simulation and mod-
eling of adsorption column dynamics, and appli-
cations of adsorption in separation and purifica-
tion.
Chemical Engineering Education

coupled with an unsteady-state macroscopic mechanical en-
ergy balance. The equations are considerably simplified,
however, if one assumes that R,, < < R. Under this condition,
the rate at which the liquid is leaving the tank is much
smaller than the total amount of liquid in the tank. The liquid
level in the tank may move so slowly that it can be essen-
tially considered as stationary. The rate at which the liquid
level in the tank drops is much smaller than the velocity of
liquid leaving the tank through the hole (i.e., V, << V,). As a
result, steady-state macroscopic mechanical energy balance
can be used instead of the unsteady-state balance. The prob-
lem can now be modeled as an unsteady-state mass balance
coupled with a steady-state mechanical energy balance. The
problem has been solved"' and the rigorous solution is given by

t = 2R H ,

where PN is a correction factor defined as

ON =(N- 2) f(rI 1N- drl (2)
0
Here, N is defined as (R/Ro)4. Since V, and V, are inversely
proportional to R2 and Roe, the variable N is equivalent to
(V,/V-. The correction factor, O,N is a direct measure of the
accuracy of the PSSA solution. As ON begins to differ sig-
nificantly from one, the PSSA solution begins to deviate
from the exact solution. The value of ON is undefined at zero
and at one. Numerical integration of 0N has been performed
as a function of N between N=1 o"' and 0.99999; the results
are shown in Figure 1. For values ofN > 100 (i.e., when the
tank radius is only about 3.16 times the radius of the hole),
the PSSA solution is more than 99% accurate. At N=3, when
the tank radius is only 1.32 times the radius of the hole, the
accuracy of the PSSA solution falls to 90%. As the radius of
the hole increases in relation to the tank radius, the level in
the tank falls faster and the assumption of a steady-state
energy balance involves greater errors.

0.98

0.96

0.94
0
0

0.92

EXAMPLE 2
Measurement of Diffusion Coefficients
by Bulb Technique

The measurement of diffusion coefficients using a two-
bulb apparatus is another example where using the PSSA is
advantageous. Here, two bulbs are separated by a capil-
lary whose volume is negligible in comparison to the
volume of each of the two bulbs. An impermeable mem-
brane located equidistant between the two bulbs sepa-
rates solutions of different compositions. At time t=0, the
membrane is ruptured and the two solutions are allowed
to diffuse into each other.
Two different time scales are pertinent for this system.
One time scale defines how fast the concentrations in the
bulbs are changing with time, and the second time scale
defines how fast the concentration profile is established in
the capillary for a given concentration driving force between
the bulbs. Since the volume of the bulbs is much larger than
that of the capillary, concentration in the bulbs changes
much more slowly. In contrast, the capillary itself contains
relatively little material. Changes in its concentration profile
occur much more quickly. Even if this profile is initially
quite different from that at steady state, it will approach
steady state before the concentrations in the bulbs change
much. Therefore, the system is modeled as an unsteady-
state mass balance for the material in the bulbs combined
with a steady-state flux across the capillary, i.e.., the
PSSA is invoked.
The PSSA solution for the concentration difference be-
tween the two chambers as a function of time can be easily
found through the solution of Fick's equation. First, two
dimensionless variables are defined: dimensionless time and
dimensionless volume ratio

Dt
T = -
L-

AL
and N = AL
V

where
D diffusion coefficient
L distance separating the two bulbs
A cross-sectional area of the capillary
V volume of a single bulb.
It is assumed that the two bulbs are of equal volume. The
concentration difference between the two chambers at any
time under the condition of PSSA is given by121
Ax(,t) = exp(-2Nt) (4)

The more rigorous solution for the concentration differ-
ence between the two chambers by the application of pertur-
bation method has been shown to be"3'

Ax(T) = exp -2NT- 2 ) /2(5)
t- t~-2) ()

Figure 1. Effect of changing the ratio of tank diameter to
hole diameter on the accuracy of PSSA solution.
Winter 1996

The ratio of the PSSA solution to the more exact solution
is then equal to

exp -2NT1-I N

This factor describes the accuracy of the PSSA solution as a
function of the ratio of the volume of the capillary to the
volume of the bulbs (N) and dimensionless time T. Figure 2
shows this relationship. When N=0.01, the capillary has 1%
of the volume of the bulb, and the PSSA solution is equal to
the exact solution for all values of T. As the value of N
increases, the PSSA solution deviates more and more from
the "exact" solution. When N=1, the capillary has a volume
that is equal to the bulb volume. At this N, the exact
solution is considerably different from the PSSA solu-
tion at all values of T except for very small values of T.
The PSSA solution has been shown to be valid12' when
the time constant for the capillary is much smaller than
that for the compartments, i.e.,

L2 1
D- >> -D
D Dp

2A
where = -
VL

Marrero and Mason'4' have discussed the effect of factors
such as end effects and the presence of Knudsen diffusion in
the capillary tube.

EXAMPLE 3
Measurement of Diffusion Coefficients
Using Stefan Tube

One of the common methods of determining DAB, diffusivity
of A and B in a binary gas system, is through the use of a
Stefan tube. Liquid A is placed at the bottom in
a long, relatively small-diameter tube filled with
gas B. Gas B is blown across the top of the tube
so that any A diffusing to the top of the tube is c
swept away, maintaining essentially zero partial 0.8
pressure of A at the top of the tube. As A vapor- 0
izes and diffuses into B, the liquid level in the -
tube drops. Time required for the liquid level to 0.
drop from an initial level to a final level is 0
measured and is used to calculate the diffusivity. 0) 0.4
-.

As in the previous examples, there are two
time scales involved here: the speed with which
the concentration profiles are established in the
gaseous phase in the tube, and the rate at which
the liquid level drops. The length of the diffu-
sion path continually increases. Since the den-
sity of liquid is three orders of magnitude higher
than that of a gas, we would intuitively surmise
that the liquid level would drop quite slowly
and, in contrast, the concentration profiles within
the gas space would establish much more quickly.
16

0.

0.2
'.,

For all practical purposes, the concentration profile in the
tube can be approximated as that at steady state. We can,
therefore, combine the steady-state flux at the boundary with
the unsteady-state macroscopic mass balance. Equating the
steady-state flux with the rate of vaporization, we obtain

PM dy =NA = DABP (PAl -PA2) (6)
dt RTy PBM

Integrating between the limits of yo (diffusion path length at
t=0) and y, (diffusion path length at t=t) and rearranging, we
obtain

DAB RTpBMpM(y -y (7)
2 P(PAI PA2)t

An obvious error in the analysis is that the initial concen-
tration distribution in the tube may be quite different from
the steady-state profile. However, we can easily calculate
the time needed to establish the steady-state profile. The
ratio of flux at the boundary at time t to flux at steady-state is
given by"5

N -2e-122 + 2e-2 2 -2-3 +... (8)
N

where
DABt
2
Yo
The approach to steady-state profile is depicted in Figure 3.
The steady-state profile is established when T = 0.5. For
gaseous systems, with y, = 10 cm and DAB = 0.1 cm2/s, the
steady-state profile is approached in about ten minutes. When
compared with the experimental time of the order of hours,
the PSSA appears quite reasonable.

0.01 0.1 1 10 100
Tau, Dimensionless Time

-- N=0.01 --N= 0.05 --N=0.1 -.-N=0.5 N = 1

Figure 2. Accuracy of PSSA solution as a function of the ratio of
capillary volume to bulb volume.
Chemical Engineering Education

A second error arises from the fact that the observed drop
in the liquid level produces vapor which occupies the space
originally filled with liquid and may not diffuse out of the
tube. But this error will be small because the partial density
of A in the newly created space will be much smaller com-
pared to the density of the liquid.

EXAMPLE
Gas-Solid Non-Catalytic Reactions

Gas-solid non-catalytic reactions represent another class
of problems where the PSSA has been advantageously ex-
ploited. Consider a non-catalytic reaction between gas (A)
and solid (S). The reaction may be represented by
aA (fluid) + S (solid) -> fluid and / or solid products
The solid particle is in the form of a spherical particle. As the
reaction proceeds, the particle may shrink with time, ulti-
mately disappearing completely, or it may retain its size but
the unreacted core of the solid continuously shrinks in size
with the formation of an ash layer on the outer shell of the
particle. Both the reactant A and the boundary of the unreacted
core move inward. But since the density of solid differs from
that of gas by a factor of about 1000, it is reasonable to
assume that the unreacted core moves so slowly towards the
center that for all practical purposes it may be considered
stationary; i.e., the time constant associated with the move-
ment of unreacted core is so much larger than that for estab-
lishing the concentration gradient in the ash layer. The PSSA
solution is relatively easy to derive and is given in a number
of books.6"' The exact solution for a spherical particle is not
possible, however, and a number of approximate analytical
solutions have been derived7 0" to determine the adequacy of
the PSSA solution. To illustrate the accuracy of the method,
we follow the suggestion of Wen"" and consider the prob-
lem in the Cartesian coordinate system.
Consider the reaction taking place on a horizontal surface
that moves inward as the reaction takes place, leaving a
porous inert layer at the top. The model equations for such a
scenario are

DCA D CA
at DeA az2

The boundary conditions are
CA AOG

0< Z

at Z=0

and at the moving boundary Z = S(t)

CA =0

dCA dS
-DeA aCSO d (10)
aZ dt
The initial condition for the moving boundary is that it
occupies the position Z=0 at t=0. The PSSA solution can be
Winter 1996

0.1 0.2 0.3
Dimensionless Time

0.4 0.5

Figure 3. Approach of transient flux to steady-state
value as a function of time.

easily shown to be

S(t) 2 CAODeAt
aCso

CAaCs1
CA = CAO 2 CADeAt Z (11)

An analytical expression for the exact case can be ob-
tained as follows. The solution of the differential equation
(9) is given by the combination of variables method and has
been shown to be"'

erf Z
CA CAO 1- erf (

where = S/(4DeAt/E)"/2 and is the root of

S- ke' erf (13)
aCso
By expanding the error function in an infinite series, we
can show that

CAOE =2(+ 23 4
aCso I 3 15

If )< i, then

X= (CAe/I2aCs0)1/2

and S(t) (2CAODeAt/aCs)1/2

This is exactly the same expression for the travel of the
moving boundary obtained for the PSSA case. Thus, PSSA
is a good approximation if X is much less than 1. It is indeed
the case for gas-solid reactions. Molar density of A is about
three orders of magnitude smaller than that for a solid.
Hence, = 10-3. As the value of X increases, however, so
does the inaccuracy of the PSSA solution. This is depicted in

Figure 4. The ratio of solid reactant conver-
sion (proportional to the location of the mov-
ing front at S(t)) predicted under PSSA to
that under unsteady-state condition is plotted
against C A0/aCs. It is clear that the error
associated with the PSSA solution increases
as the value of EC AO /aC0 increases.

( EXAMPLE
Dissolution of a Sphere in Liquidl21

In Examples 3 and 4, the PSSA concept
appeared quite reasonable because the densi-
ties of the two phases involved (gas-liquid in
Example 3 and gas-solid in Example 4) were
different by about three orders of magnitude.
Let us now consider the case where the densi-
ties of the two phases are of the same order of
magnitude. The problem of dissolving a ben-
zoic acid sphere of 1 cm diameter in water
at 250C has been considered by Sherwood,
wish to determine the time required for comp
tion of the sphere.
Again, there are two time constants involve
cess: one characterizes the rate of movement of
boundary, and the other characterizes how qui
centration profiles are established in the aqueoi
as in Examples 3 and 4, it is tempting to ass
boundary recedes very slowly compared to the
for the steady-state concentration profile to
determine how quickly the steady-state conce
files in the fluid phase are established, we car
following analysis. Benzoic acid dissolves at
and radially diffuses into water. Assume that t
tion of the aqueous phase at the surface of
maintained at 0.0278 g moles/1. The flux at the s
ary at r = r,, can be shown to be

= DABCs + DAB Cs
(NA ) + s~ AB
r=r ro0 At

The transient flux differs from the steady-stat
when (ltDABt) is equal to 100 r0. For DB = 1.
this time is calculated as 72.2 million seconds.
sis, the diameter of the sphere has been assume
constant. But when we integrate the flux equati
we find that in 72.2 million seconds the tot
benzoic acid dissolved will be about twenty-s
contained in a one-cm sphere.
The use of a steady-state equation, in this ca
considerable error. The transient concentration
proaches steady state extremely slowly. First,
DAB is small and molar concentration of acid
relatively large, much larger than that of gas

0.8 -
0.001

0.01 0.1
sCAol(aCSo), Ratio of Densities

Figure 4. Accuracy of PSSA solution as a function of ratio of two densities.

et al.112" We space in the Stefan tube. The amount of solute contained in
lete dissolu- the shell originally occupied by solid is significant. In this
case, PSSA does not work well. The value of ? (as defined
d in the pro- in Example 5) in this case approaches one, and for this
the spherical condition, the PSSA solution should deviate significantly
ckly the con- from the exact solution.
us phase. Just
us ptha st CONCLUDING REMARKS
ume that the
time required The PSSA is a powerful concept that can be applied in
establish. To many transport phenomena problems where multiple time
ntration pro- scales are involved. This has been demonstrated with the use
n perform the of several examples. In every case, however, it is imperative
the interface to determine the accuracy of the PSSA solution by compar-
he concentra- ing it with either an exact solution or an approximate closed
the sphere is form analytical solution or numerical solution. In the first
phere bound- four examples, the PSSA solution is fairly accurate; but for
the last example, the concept of PSSA solution results in
significant errors.
(15) REFERENCES

1. Bird, R.B., W.E. Stewart, and E.N. Lightfoot, Transport Phenom-
e flux by 1% ena, John Wiley & Sons, New York, NY (1960)
1 x 10-5 cm2/s, 2. Cussler, E.L., Diffusion: Mass Transfer in Fluid Systems, Cam-
In this analy- bridge University Press, New York, NY, 26 (1984)
ied to remain 3. Paul, R., The Physics of Fluids, 3, 905 (1960)
4. Marrero, T.R., and E.A. Mason, J. Phys. Chem. Ref. Data, 1, 1
on with time, (1972)
al amount of 5. Lee, C.Y., and C.R. Wilke, Ind. Eng. Chem., 46, 2381 (1954)
six times that 6. Levenspiel, 0., Chemical Reaction Engineering, 2nd Ed., John
Wiley & Sons, New York, NY, 365 (1972)
7. Bischoff, K.B., Chem. Eng. Sci., 18, 711 (1963)
ase, results in 8. Bischoff, K.B., Chem. Eng. Sci., 20, 783 (1965)
n profile ap- 9. Luss, D., Can. J. Chem. Eng., 46, 154 (1968)
the value of 10. Theofanous, T.G., and H.C. Lim, Chem. Eng. Sci., 26, 1297 (1971)
ie11. Wen, C.Y., Ind. Eng. Chem., 60, 34 (1968)
in solution is 12. Sherwood, T.K., R.L. Pigford, and C.R. Wilke, Mass Transfer,
A in the gas McGraw Hill, New York, NY, 69 (1975) 0
Chemical Engineering Education

[ book review

Boundary Element Methods
in Transport Phenomena
by P. A. Ramachandran
Published by Computational Mechanics, Inc., 25 Bridge
Street, Billerica, MA 01821; 424 pages, $160 (1993)

Reviewed by
Bruce A. Finlayson
University of Washington

The boundary element method solves an integral form of a
differential equation. This method has the marvelous advan-
tage that for a linear problem the entire solution is deter-
mined from the boundary conditions. This changes a three-
dimensional problem into a two-dimensional problem, or a
two-dimensional problem into a one-dimensional problem;
both reductions in order provide significant computational
savings. The reduction of dimensionality is useful if one
wants only a few features of the solution, such as the value
of the function at a few points, or the integrated flux. If one
wants the solution everywhere (say, for contour plotting),
then the computational cost goes up, and this may be why no
contour plots are presented in the book. Multiple equations
are also limiting, since the equations must be handled sepa-
rately, leading to iterations that may not converge.
The disadvantage of the boundary element method is that
nonlinear problems still require that the solution be repre-
sented everywhere, usually using finite elements throughout
the whole domain. This feature destroys the chief advantage
of the method, the reduction of dimensionality. The author
argues that one does not need to introduce approximations
at an early stage, but approximations are finally introduced,
and if they affect the result it doesn't matter when they are
introduced.
The first chapter carefully develops the differential equa-
tions and boundary conditions that are needed to solve trans-
port problems, including fluid flow and mass and heat trans-
fer. This is carefully done, and the treatment is concise.
The second chapter shows how the use of a weighting
function plus one integration by parts can lead to a weighted
residual method (like Galerkin); with two integration by
parts the boundary element method is obtained. The weight-
ing function can be a Green's function, in which case the
method is the "Fundamental Solution Method," or the bound-
ary integral method.
The book (which has a total of eleven chapters) is very
clearly written, with lots of simple examples. Fundamental
Solution Methods and Green's Functions are clearly pre-
sented, with lots of details carefully attended to. The Laplacian

Winter 1996

operator is clearly important, but chapters deal with exten-
sions involving time derivatives, one-, two-, and three-di-
mensional problems, the Poisson Equation, heat and mass
transfer applications, and application to fluid flow. Most of
the problems are simple; more complicated solutions are
given in tables, which is not a clear way to present material
unless it is to be used for checking computer code. A computer
diskette is provided, but the reviewer did not review it since it
is not in Macintosh format (it also uses the old 5 1/4" format).
There are some minor errors: reference is made to the
"Burger" equation; the person's name was Burgers. On page
153 the claim is made that the "quadratic" method for inte-
grating ordinary differential equations does not lead to oscil-
lations regardless of the step size-that is false.
The author makes the claim that the boundary element
method will become the most widely used numerical method
for engineering analysis in the 21st century. This reviewer
believes that won't happen for nonlinear problems, and most
problems are nonlinear. In fact, the book is deficient in that
hard problems are not presented, even to whet the reader's
appetite. The book does not usually demonstrate convergence
with mesh refinement, which is a standard required of first-rate
numerical journals, and this author believes that is essential.
Detailed equations are given for constant, linear, or quadratic
elements, but little guidance is given when to use which ones.
Despite these minor defects, overall the book is excellent.
The careful, concise treatment of both Green's functions and
boundary element methods will make it useful for anyone
doing analysis of transport problems, or anyone who wants
to learn about Green's functions. This reviewer believes that
reviewing a book is only worthwhile if he can learn some-
thing. He did. 1

EM^a book review

HAZARDOUS WASTE MANAGEMENT, 2nd ed.
by Charles A. Wentz
Published by McGraw-Hill Book Company, NY (1995)

Reviewed by
Ralph H. Kummler
Wayne State University

At Wayne State University, we used the first edition of
Hazardous Waste Management by Charles A. Wentz in the
developmental stages of our program. We viewed it as an
excellent undergraduate and graduate introductory overview
to the hazardous waste management (HWM) field. Our civil
engineering curriculum also used the text for their environ-
mental engineering course on landfill disposal techniques.
As our HWM program developed, we began taking in
Continued on page 61.
19

7 Classroom

COMPUTATIONAL RESULTS -

HOW RELIABLE ARE THEY?

A Systematic Approach to Model Validation

NEIMA BRAUNER, MORDECHAI SHACHAM,"l MICHAEL B. CUTLIP[2]
Tel-Aviv University Tel-Aviv 69978, Israel

wo recent papers in this journal[1'2] have discussed the
ever-increasing role of computers in chemical engi-
neering education and practice. While computers are
heavily used for word processing and communication, their
most noticeable effect in engineering education is their role
as mathematical modeling and numerical computation tools.
The range of numerical computational tools available to
the student and the practicing engineer includes spreadsheets
for simple calculations, numerical computation packages
such as MATLAB, MATHEMATICS, MAPLE, and
POLYMATH, and powerful, sophisticated steady-state and
dynamic simulation programs such as ASPEN, HYSIM,
PROII, and SPEEDUP. These tools have considerably re-
duced the time and effort required for engineering calcula-
tions. They also make it possible to simulate operation of a
complete process or even a plant. The sophisticated compu-
tational tools have not reduced, however, the need to verify
and validate the results. Actually, there is probably more
need than ever for verification of the results because some of
the computational tools are used as a "black box" where the
applied mathematical model is invisible to the user.
Most commercial simulation programs use the black-box
approach where the user has to provide only a minimum
amount of input data to specify the process. The mathemati-
cal model, the solution algorithm, and the physical and
thermodynamic properties are provided by the program,
and the user usually receives only the final results. This
approach saves much of the user's time, but it makes it
impossible to use some of the traditional methods for model
validation and verification.
Himmelblau[31 quotes Finger and Naylor's 41 steps for model
validation as: validation of the logic, validation of model
behavior, and validation of model assumptions. Clearly, when
the model is invisible to the user, neither its logic nor its

1 Address: Ben-Gurion University of the Negev, Beer-Sheva,
84105, Israel
Address: University of Connecticut, Storrs, CT 06269

simplifying assumptions can be validated. The user can only
rely on the final results for validating the model.
Validating the model and verifying the results is more an
art than a science, as Himmelblau notes. The model can
never be completely validated because there are only finite
number of tests that can be carried out,E5" and passing a
certain number of tests does not ensure that the model is
correct. In order to minimize the chance for errors, a verifi-
cation process that uses the final results only as a diagnostic
tool should be devised; this process should be used consis-
tently, without taking anything for granted. The use of so-
phisticated computational tools can save a lot of time, but
some of this saved time must be used for validation and
verification of the results.

Nelma 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 serves as
the President of the Israel Institute of Chemical
Engineers. She teaches courses in Mass and Heat
Transfer and Process Control. Her main research
interests include two-phase flows and transport
phenomena in thin films.

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 pro-
cess simulation, design, and optimization, and
expert systems.

Michael B. Cutlip received his BChE and MS
from The Ohio State University and his PhD
from the University of Colorado. He has taught
at the University of Connecticut for the last twenty-
five years, serving as Department Head for nine
years. His research interests include catalytic
and electrochemical reaction engineering, and
he is coauthor of the POLYMATH numerical
analysis software.

Copyright ChE Division of ASEE 1996
Chemical Engineering Education

. . sophisticated computational tools have not reduced, however, the need to verify and validate the
results. Actually, there is probably more need than ever for verification of the results
because some of the computational tools are used as a "black box" where
the applied mathematical model is invisible to the user.

In this paper, a model validation and verification process,
based only on the final computational results, is presented
and its use is demonstrated using several examples from the
literature. We recommend that students be introduced to
model validation toward the last quarter of a modeling and
simulation course. The examples included in the paper can
be best given as homework assignments where the student
can use a numerical computation package (such as
POLYMATH, MATLAB, or MATHEMATICS) to solve
the problem and use the validation procedure to detect what
is wrong with the solution. Some of the examples involve
solution of stiff ordinary differential equations, and it is
important to ensure that the software used by the students is
capable of solving such equations.
Most of the readers have probably come across examples
(even in research work) where the lack of model validation
has led to embarrassing glitches. The examples we present
are fairly simple, so that they can be easily understood by
undergraduate students, do not require excessive amount of
time for preparation, and can be solved using widely avail-
able software packages.

MODEL VALIDATION
AND VERIFICATION PROCEDURE
The following procedure assumes that the only informa-
tion available about the model is the final result.
1. Solve a problem similar to the one you want to solve,
but where the results can be verified using: process
data, results from the literature, an analytical solution,
or limiting cases (a typical example would be checking
the steady-state solution for a dynamic problem).
Compare the solution obtained by using the model with
the results or data obtained independently.
2. Always investigate error messages and warnings that
your program issues.
3. Check the results obtained for physical feasibility.
4. Carry out a sensitivity analysis by introducing small
changes in the input data and user-selectable or
adjustable parameters of the computer program (such
as solution algorithm, error tolerance, plot interval,
etc.). Look out for any unreasonable changes in the
results caused by these parameter variations.
While complete verification of the results is practically
impossible, consistently carrying out the above four steps of
verification can prevent most of the common errors encoun-
tered in simulation and numerical computation. Some ex-
amples that demonstrate this procedure follow.
Winter 1996

Example 1
Transient Behavior of a Catalytic Fluidized Bed
Luss and Amundson161 studied a simplified model for the
dynamics of a catalytic fluidized bed in which an irrevers-
ible gas phase reaction A -> B is assumed to occur. The mass
and energy conservation equations for this system were

dP = P H(P P)
dT / \
T=Te T + H(T, T)+ Hw(Tw T)
dP
-Pp g [P- Pp,(1 + k)]

dTp HT
dT~ C[(T T) + FkP ]

where
k=0.0006 exp(20.7 15000/Tp)
T(R),P(atm)=temperature and partial pressure of the reactant
in the fluid
Tp(R),Pp(atm)= temperature and partial pressure of the reactant
at the catalyst
T dimensionlesss time
Hg,HT,F,A,C=dimensionless constants

and the subscript e indicates entrance conditions. The fol-
lowing numerical values were provided by Luss and
Amundson:
Pe = 0.1 atm C = 205.74 Hg = 320
Te= 6000R F=8000 HT=266.667 A=0.17142
Luss and Amundson noted that the system of ordinary
differential equations (ODE) representing the catalytic fluid-
ized bed is a stiff system. At that time, there were no estab-
lished methods for solving stiff systems of ODEs and they
derived a special technique to solve it.
Subsequently, Aiken and Lapidus[71 proposed a different
method for solving stiff ODEs. They used the system of Eq.
(1) as a test example, but rewrote the system of equations by
introducing the numerical values into system (1) and round-
ing some of the coefficient as follows:

dP 0.1 + 320Pp -321P
dlt
dT = 1752- 269 T + 267Tp
dPr
dP -=1.88 x 103[ Pp(l + k)]
dT
- = 1.3( T-Tp)+ 1.04 x 104 kPp
dT

One possible assignment for the students in this example
can be to verify that Eq. (1) and Eq. (2) yield the same
steady-state solutions.
Luss and Amundson have identified three steady-state
solutions for this problem. The values of P, Pp, T, and Tp at
the three steady states reported by Luss and Amundson are
shown in Table 1.
To find the steady-state solutions, the time derivatives in
the four equations of systems (1) and (2) are set equal to
zero. The systems can then be reformulated to give a single
implicit nonlinear equation, which should equal to zero,
while the rest of the variables can be calculated from explicit
expressions. Introducing the numerical values of the con-
stants into Eq. (1) and reformulating yields

f(T) = 1.296 (T Tp)+ 10369 kPp (3a)

Tp = (269.267 T 1752) / 266.667 (3b)
Pp= -0.1/{321[320 / 321 (1+ k)]} (3c)

(320 Pp +0.1)
321

System (2) can be written in a similar manner. Figure 1
displays plots of f(T) versus T in the region 500R using both the original and revised formulations. It can be
seen that the original formulation yields three steady states
at the points indicated by Luss and Amundson, whereas the
revised formulation gives only a single root at T = 1210.8.
Thus, when there is very little difference between the

TABLE 1
Steady-State Solutions
of the Catalytic Fluidized Bed161
Steady States
First Second Third
p(atm) 0.09352 0.06704 0.006822
Pp(atm) 0.09350 0.06694 0.006531
T(oR) 690.445 753.344 912.764
Tp(R) 690.607 759.167 915.094

4
f(T) revised
S2 formulation
.........-.-........ -. -- - ..--.....-- .- .......- .... .......- ..
original
-2- formulation

-4 -

-6 I
0.4 0.6 0.8 1.0 1.2 1.4
)---Txo10-3
Figure 1. Steady states of the catalytic fluidized bed
reactor using original and revised model formulations.
22

original (Eq. 1) and revised (Eq. 2) formulation, they actu-
ally do represent a much different problem. The discrepancy
between the original and revised formulation was detected
by Michelsent19 two years after the revised formulation was
published. In the meantime, this formulation was exten-
sively used for testing software (see, for example, reference
8) without noticing that it actually was a different problem.
To understand the reason for this difference, the expression
for Tp in Eq. (3b) can be introduced into Eq. (3a) to yield
f(T) = 8.5147 0.0126 T + 10369 kPp (4)
Carrying out the same substitution using the modified for-
mulation yields
f(T) = 8.53 0.00974 T + 1.04 x 104 kPp (5)

It can be seen that the coefficients of T in Eqs. (4) and (5)
are significantly different, thus rounding the numbers at the
third decimal digit in Eq. (2b) resulted in not even one
correct digit in the coefficient of T in Eq. (5).
This example demonstrates that small changes in the
model equations may sometimes cause unpredictably large
changes in the results. Model validation is needed to
detect such errors.

Example 2

A Chemistry Problemi101

This problem has been frequently used to test stiff ODE
solver programs, and it is cited very often in both the chemi-
cal engineering[81 and numerical analysis ll.p.734] literature.
The equations of this example, as they appear in reference
11 are
dyl = -0.013 y 1000 YiY3 (6a)
dt
dy2 = -2500 y2y3 (6b)
dt
dy3 = -0.013 yi -1000 Y1Y3 2500 Y2Y3 (6c)
dt
The initial conditions are yi(0)=1, y2(0)=l, and y3(0)=0.
These equations are usually integrated from to = 0 up to tf =
50. Assuming that y,, y2, and y3 represent concentration of
different species, the students should check the physical
feasibility of the solution.
The variation of y3 in the requested time interval is shown
as curve "A" in Figure 2. It can be seen that y3 descends very
rapidly from the initial values y3(0)=0 to y3=-3.7x106
and stays negative for the whole range of solution. Assum-
ing that y3 represents concentration (a very probable
assumption given the form of the model equations), it can-
not be negative.
The original reference by Gear'o10 shows that there was a
typographical error in Eq. (6c). The equal sign is missing
and there is a minus sign in front of the 0.013 y, term.
Chemical Engineering Education

results in physically infeasible solution, and that often the
plot interval must be changed in order to obtain complete
details of a solution.

Example 3
Chemical Equilibrium
The following system of algebraic equations describes
equilibrium in a constant volume, gas-phase batch reactor
for a complex system of reactions:

Figure 2. Variation of y, in the chemistry problem in a
large time scale
r ___________________

0 10 20 30 40 50
t.--1 03
Figure 3. Variation of y, in the chemistry problem in a
short time scale.
Apparently when the equation was copied by others, the
equal sign was added and the minus sign was retained. This
formulation of the problem gives a physically infeasible
solution of a negative concentration of y3. The general form
of Eq. (6) indicates that it most probably represents reaction
rates among three reacting species, so the 0.013 y, term in
Eq. (6c) must definitely be positive.
The integration when the first minus is replaced by a plus
in Eq. (6c) yields all positive values for y3 as shown in
Figure 2 (curve "B"). Figure 2 demonstrates an additional
potential problem in interpreting the results. Since the initial
change of y3 is very fast, it seems from the figure that the
initial value of y3 is y3(0)=3.27x10-6 (or -3.27x10-6) instead
of the correct value of y3(0)=0. To see the exact details of the
solution at initial t, the integration interval must be reduced
considerably. Figure 3 shows the initial changes of y3 when
the integration interval is reduced by a factor of 103.
This example demonstrates that error in the model often

TABLE 2
Multiple Solutions of the Chemical
Equilibrium Problem

Variable 1 2 3
CD 0.7053 0.0556 1.0702
Cx 0.1778 0.5972 -0.3225
Cz 0.3740 1.0821 1.1304
CA 0.4207 -0.3624 -0.7007
CB 0.2429 -0.2348 0.8080
Cc 0.1536 -1.6237 -0.3782
C' 0.5518 1.6793 0.2623

Winter 1996

fl(CDC X C CcCD -K1 = 0
CACB

f2(CD,Cx,Cz) CCY K2 = 0

f3(CD,CX,CZ) Cz K3 = 0
CACX
CA = CAO -CD Cz
CB =CBO -CD -CY
CcC CD -CY
Cy= CX -CZ

where CA, CB, C, CD, CD Cy, and Cz are concentrations of the
various species, and CAO, CBO, K,, K9, and K3 are constants.
The assignment is to solve the system for the following
values of the constants:

CAO = CBO = 1.5
KI = 1.06
K2 = 2.63
K3= 5
for three different sets of initial estimates
(CD, CX, Cz) =0,1,10
Most programs for solving nonlinear algebraic equations
will not be able to solve this system (7) as it is written. The
difficulty is caused by division by the unknowns in the first
three equations. The problem can be made much less nonlin-
ear and easier to solve by eliminating division by the un-
knowns. Indeed, f, can be multiplied by CACB to yield CcCo
- KiCACB = 0. Similar transformations can be applied to f2
and f3. Using the modified set of equations POLYMATH
converged to three different solutions (as shown in Table 2)
from the three initial guesses.
Checking for physical feasibility reveals that only the first
solution is acceptable. In solutions 2 and 3, some of the
concentrations are negative, and thus these solutions cannot
represent a valid physical situation.
Contrary to dynamic simulation, in solving steady-state
models, the algorithm may converge to infeasible solutions,
even when the model is correct and the initial estimate lies in
the feasible region. If an infeasible solution is reached, a
sensitivity analysis (by changing the initial guess) should be
carried out in an attempt to locate a feasible solution. In this
23

B, Correct formulation

SA, Erroneous formulation

case, convergence to infeasible solutions does not necessar-
ily indicate an erroneous model.

Example 4

Equilibrium Conversion in an Isothermal Tubular
Reactor
The following equations represent the conversion in a
tubular reactor (X) as a function of the catalyst weight (w):

kiP0(1-X) k2Po2X2
dX 1+X (1+X)2 (8)
dw FAO1 +7 P ( ]
(1 + X)

where
k, = 1.277 x 109 exp[-90000/(8.31 T)]
k2 = 1.29x101 exp[-135000/(8.31 T)]
FAO = 20 P0 /(0.082 x 450)
P0 = pressure at the inlet
T = temperature in the reactor

The assignment is to find the equilibrium conversion in
the reactor for P0 = 10 atm and T = 3130K.
To find the equilibrium conversion in the reactor, dX/dw =
f(X) is set to zero. Solving the resultant algebraic equation
using the POLYMATH 3.01121 program yields two solutions:

X= 0.984 where f(X)= 0.114 x 10-7
and
X = 1.02 where f(X) = -0.7 X 10-7

At both points the function value is very small, and thus both
can represent legitimate solutions. But conversion of 1.02 is
unacceptable because it is physically infeasible to obtain
conversion higher than 1.
Carrying out sensitivity analysis, by changing the initial
guess for the unknown X, causes the program to find differ-
ent values for the first solution. A plot of f(X) versus X
(shown in Figure 4) reveals the reason for the inability of the
program to locate the root precisely. Between X = 0 and X =
1 the function value is always below the 3 x 10-8 in absolute
value. There are two changes of the function value sign
around X = 1. One at the root (the precise value is X =
0.999985) and the other at X = 1.029, which is a point of
discontinuity for the function. With such small function
values throughout the entire interval of interest and the pres-
ence of point of discontinuity near the solution, most pro-
grams will have great difficulty in locating the precise root.
Once a solution is reached, the root must be verified. For
verification, the values of the unknown must be introduced
into the functions to yield values close to zero. In some of
the nonlinear equation-solver programs, the user must ex-
plicitly request calculation and display of the function values
24

10

6-

S 2
4: -2-

-6 -

WO 0.3 0.6 0.9 1.2 1.5
--]1,- x

Figure 4. Function shape for equilibrium conversion
calculation in a tubular flow reactor.

at the solution. This is essential for avoiding acceptance of
incorrect results, as may happen when the program uses
minimization algorithms and occasionally converges to a
local minimum. More strict verification of the root is pos-
sible by carrying out a sensitivity study to obtain changes of
the sign of the function values in the vicinity of the solution.

Example 5

Aerobic Microbial Growth Problem1131
The following equation represents the amount of substrate
(S), cells (x), and concentration of oxygen (Co,) in an
aerobic microbial growth system:

dx
dt
dS gx mx
dt Yx/s

d =- KLA (C2 C02o)- (9)

where
S
R = l max Ks +S

9max, Ks,YxisYx/o02,m,KLA,Co2,mo2 = constants.
The assignment is to explore the dynamics of this system
from t = 0 to t = 10 hrs using the constants and initial values
shown in Table 3.
Figure 5 shows the variation of the biomass with time. It
can be seen that the amount of biomass increases up to
around t = 0.65 and from this point on the amount decreases
(as indicated by the curve of the original model). Checking
the physical feasibility for the other variables reveals that
when x reaches its maximum, the substrate value is reduced
to zero. It continues to decrease and obtains negative values.
This is, of course, impossible. The reason for the negative
amounts of material, in this case, is that the model presented
in Eq. (9) is correct only if S > 0. In order to make the same
model applicable for the S = 0, the differential equation dS/
dt must be rewritten as
Chemical Engineering Education

PX
dS Yx /
dt
0

mx if S > 0

otherwise

Using the revised model x remains constant after reaching
its maximum, as shown in Figure 5. Thus the original model
was used outside the region of its validity, and proper model
validation procedure detects this problem.
If, during dynamic simulation, some of the variables become
infeasible at a particular point, sensitivity analysis (by chang-
ing tolerances or parameters of the numerical solution algo-
rithm) at the vicinity of this point can detect whether the
source of the problem lies in the numerical solution algo-
rithm or the model fails to represent correctly the physical
situation at this point.

DISCUSSION AND CONCLUSIONS
We have shown five examples where computational re-
sults can be incorrect. The following reasons for incorrect or

TABLE 3
Constants and Initial Values for the
Microbial Growth Problem[131

Constant Value Units
Itmax 0.6 hr-'
K 0.05 gr/liter
YxV 0.5 gr cells/gr glucose
Yx/o, 1 gr cells/gr 0,
m 0.08 gr glucose/(gr cells hr)
K LA 400 hr1
C0, 8 mgr/liter
mo, 0.1 gr O,/(gr cells hr)

x(t=0) 0.1 gr/liter
S(t=0) 10 gr/liter
Co, (t=0) 8 mgr/liter

revised
x 5 __/model__

4
Original
3 model

2-

0 2 4 6 8 10

Figure 5. Variation of the biomass in the microbial
growth problem.
Winter 1996

imprecise results were demonstrated: carelessly rounding
numbers in the model equations; error in the sign in a model
equation; multiple problem solutions; using a model outside
the domain of its validity; numerical difficulties in finding
the precise solution when working with very small numbers.
There can be many more reasons for obtaining incorrect
results. Correlation of experimental data when the model
equations are improperly linearized1141 or when experimental
design for obtaining the data is not satisfactory1151 can be
common sources of such errors. Low resolution in present-
ing the results can lead to misinterpretation of the results
even if the solution is correct.1161
In an era when hand calculations have been replaced by
computation, it is more important than ever to consistently
validate and verify the results. The examples provided in this
paper demonstrate very clearly the necessity of model vali-
dation. The suggested procedures can serve as a basis for
systematic approach for validating the results.

REFERENCES
1. Mah, R.S.H., and D.M. Himmelblau, "Role and Impact of Computers
in Engineering Education," Chem. Eng. Ed., 29(1), 46 (1995)
2. Davis, J.R., G.E. Blau, and G.V. Reklaitis, "Computers in Under-
graduate Chemical Engineering Education," Chem. Eng. Ed., 29(1),
46(1995)
3. Himmelblau, D.M., "Mathematical Modeling," p 35 in Bisio, A., and
R.L. Kabel, eds, Scaleup of Chemical Processes, John Wiley, New
York, NY (1985)
4. Finger, G.S., and T.H. Naylor, Mang. Sci., 14, 92 (1967)
5. Riggs, J.B., "A Systematic Approach to Modeling," Chem. Eng. Ed.,
22, 26 (1988)
6. Luss, D., and N.R. Amundson, "Stability of Batch Catalytic Fluid-
ized Beds,"AIChE J., 14(2), 211 (1968)
7. Aiken, R.C., and L. Lapidus, "An Effective Numerical Integration
Method for Typical Stiff Systems," AlChE J., 20(2), 368 (1974)
8. Enright, W.H., and T.E. Hull, "Comparing Numerical Methods for
the Solution of Stiff Systems of ODEs Arising in Chemistry," p 45 in
Lapidus, L., and W.E. Schiesser, eds, Numerical Methods for Differ-
ential Systems, Academic Press, Inc., New York, NY (1976)
9. Michelsen, M.L., "An Efficient General Purpose Method for the
Integration of Stiff Ordinary Differential Equations," AIChE J.,
22(3), 594 (1976)
10. Gear, C.W., "The Automatic Integration of Stiff Ordinary Differen-
tial Equations," Proc. of the IP68 Conf., North-Holland, Amsterdam
(1969)
11. Press, W.H., S.A. Teukolsky, W.T. Vetterling, and B.P. Flannery,
Numerical Recipes, 2nd ed., Cambridge Univ. Press, Cambridge
(1992)
12. Shacham, M., and M.B. Cutlip, POLYMATH 3.0 User's Manual,
CACHE Corporation, Austin, TX (1993)
13. Bajpai, R., personal communication (1995)
14. Shacham, M., J. Wisniak, and N. Brauner, "Error Analysis of Lin-
earization Methods in Regression of Data for the Van Laar and
Margules Equations," Ind. Eng. Chem. Res., 32, 2820 (1993)
15. Shacham, M., and N. Brauner, "Correlation and Overcorrelation of
Heterogeneous Reaction Rate Data," Chem. Eng. Ed., 29(1), 22 (1995)
16. Shacham, M., N. Brauner, and M. Pozin, "Pitfalls in Using General
Purpose Software for Interactive Solution of Ordinary Differential
Equations," presented at the ESCAPE5 Conference, Bled, Slovenia,
June 11-14 (1995) D

G classroom

APPLICATIONS OF

SOME MODERN MANAGEMENT TOOLS

IN EDUCATION

RICHARD POLLARD
University of Houston Houston, TX 77204-4792
Since total quality management (TQM) (also known as
the continuous improvement process) is used in in-
dustry,"11 it would be helpful for students to have an
appreciation of what it involves before they graduate. One
teaching strategy might be to give lectures on TQM theory,
but such presentations often come across as being rather
abstract and dull. Hence, it is difficult for the theory to
demonstrate the real benefits of TQM. Descriptions of case
studies also tend to be ineffective since the students are not
actively involved, and the solution to a problem, once given,
is often perceived as obvious.
An alternative approach is to introduce TQM indirectly by
having students use the various TQM tools to address prob-
lems and issues that concern them. These tools are simple,
quick, and fun to use, yet they reduce the time required to
plan activities and accomplish goals. In addition, application
of the tools promotes student creativity and participation,
helps break down student-faculty barriers, and provides a
mechanism for rapid feedback to the instructor (who acts as
a facilitator for implementing the tools).
A variety of tools have been developed to help apply the
underlying principles of TQM.[2'3] Some of these tools (e.g.,
histograms, scatter diagrams, and control charts) are used to
display and interpret numerical data and, hence, they can be
introduced quite naturally with material on statistics and
experimental design. Other management tools address
"softer" issues such as organizing ideas, building consensus,
and making decisions. In this paper, several examples of the

Richard Pollard has been Professor of Chemi-
cal Engineering since 1989. He received his BA
and MA from the University of Cambridge, En-
gland, and his PhD from the University of Califor-
nia, Berkeley. His research interests range from
reactor engineering and complex reaction net-
works to processing of electronic materials and
electrochemical systems.
Copyright ChE Division of ASEE 1996

TABLE 1
Summary of Management Tools Described in this Paper

Management Tool
Affinity Diagram
Relatons Diagram

Pnonty Mainr

Deployment Chart
Nonunal Group Technique

Primary Function(s)
* To gather and organize ideasiopinions
* To citablbh the links among related items and
identify the controlling factor(si
* To decide on the most critical tasks/issues and
plan the sequence of events
* To divide responsibilities among team members
* To rank preferences in a list of itemr

second type of tools are presented (see Table 1), with em-
phasis on how to put them into practice.

APPLICATIONS IN THE CLASSROOM
At the beginning of an elective course titled "Reaction
Kinetics for Industrial Processes," the instructor asked the
students "What are the goals of reactor modeling?" Specifi-
cally, they were asked to write, as quickly as possible (and
silently), short phrases summarizing any ideas they had for
the goals. The instructor encouraged student participation by
emphasizing that all ideas are good ideas.
The students then formed teams (a total of 6-7 teams is
optimal), grouped their ideas, and transferred each one to a
4x6 Post-ItT' note (in large lettering, using a felt-tip pen).
The notes were immediately placed in a random fashion on a
large sheet of butcher paper attached to the wall. The stu-
dents were permitted to read the notes as they were posted
since this often sparks additional thoughts.
After many ideas had been posted (usually within 10-15
minutes), several students (one from each team) were given
a few minutes to move the notes into groups that had com-
mon threads. This procedure, performed without discussion,
was then repeated by other team members until everyone
had been given a chance to sort the ideas. (Occasionally it
was necessary to ask for clarification of a note's meaning,
but there was never any critique of the ideas.) Next, the
instructor asked the students to give each grouping a name
Chemical Engineering Education

(header card) that captured the essence of the ideas in that
group. Some additional note movement occurred at this stage,
i.e., if someone said "that idea doesn't fit there," they were
told to go and move it. Anyone who disagreed was invited to
move it again! The result is called an affinity diagram (see
Figure 1).
The header cards were arranged to form a circle on the
wall (with the notes placed outside the circle, next to the
corresponding header card). Initially, a card was chosen (at
random) and compared with each of the others, one at a time.
For each comparison, the instructor posed the following
question: "If we improve this item, does it improve the other
item, or vice versa?" Then, an arrow was drawn from the
cause to the effect (driver to outcome). For example, in
considering the goals of reactor modeling, it was felt that

Affinity Diagramr
Goals for Reactor Modeling

S t

Teaching Environmental Optimization Proess Process Proc Employmen
i To teach Environmental To improve e Understand F ind cheap Feed coo. Keepaoam r

sands eces. made demand
(Todelano na ------e mIne p t T, x iotcosts
satrty on prd e Hither T
vcr and by. ^slecivandyJ (Eine ral- To produce ts
reactor To oet- reator design
Lproperly J mine p. T..

Fgure 1. Illustrative example of Toan affinity diagram. Theo o

items in rectangles are header cards; individual ideas are
in rounded boxes.

Figure 2. Relations diagram for "goals for reactor model-
ing." The header cards (in oval boxes) are from Figure 1.
The arrows and numbers are explained in the text.
Winter 1996

. . management tools address "softer" issues such
as organizing ideas, building consensus, and
making decisions ..... examples .. are
presented with emphasis on how
to put them into practice.

better "Process Control" would lead to better "Environmental,
Health, and Safety," rather than the reverse (see Figure 2).
Other arrows were established in a similar manner. In
some cases, the cause-and-effect relationship was not clear
and some discussion ensued. If a consensus could not be
reached quickly on which item was the major influence, no
arrow was drawn. (In this situation, the relationship between
the two items is usually not crucial.) In a few cases, no
relationship between two cards was apparent, and when this
occurred, again, no arrow was used. After the comparison
procedure was completed for all the header cards, the total
number of arrows out/in was written next to each one. The
cards with more arrows going out are causes (drivers),
whereas the cards with more arrows going in are effects
(outcomes). The result is called a relations diagram.
For the example in Figure 2, "Process Economics" and
"Environmental, Health, and Safety" had arrows entering
but not leaving; hence, improving these items was regarded
as the mission of modeling reactors. One heading (Process
Fundamentals) only had arrows leaving; hence, it was re-
garded as the primary driver (i.e., where effort should be
focused in order to accomplish the mission). Copies of the
two diagrams were distributed to the class and students were
given a chance to comment and to present additional ideas.
Generation of Figures 1 and 2 took only about one hour of
class time, yet it got the students involved in the material to
be covered during the semester and it motivated them to
learn that material. Furthermore, all the students (rather than
just the ones with expressive personalities) felt comfortable
making contributions because no one was "evaluated" for
his or her comments. This approach had far more impact
than a lecture because the students themselves came up with
all the ideas, i.e., it was is "their" diagram-the instructor
did nothing but keep the procedure on track. At the same
time, it showed them how two management tools could be
used to answer a question efficiently and painlessly.
The same course involved a team project (with four people
per team). The students indicated that they would like guid-
ance on how to be effective and how to avoid conflict.
Consequently, information was presented on

building an effective team (e.g., practicing team-building
roles, dealing constructively with diverse opinions,1' thinking
"win-win, "141 recognizing differences in people's personali-
ties and accounting for them,t51 etc.)
holding effective team meetings
giving effective presentations

4/0 E
Teaching

( Employment

* using management tools to establish priorities and to get
everyone to participate.

Here, we will focus on the tools. After approximately two
weeks had been allowed for initial technical reading on the
assigned problem, the procedure for the affinity diagram
was applied during a team meeting, using the question "What
action items are needed to complete the project?" One slight
difference from before was that the extent of grouping was
kept to a minimum, although some ideas were rewritten to
avoid duplication. Note that if students had prepared infor-
mation in advance, they were asked to ignore it and just
write down ideas "off the top of their head." (Otherwise
there is a danger that one person will dominate and stifle the
creativity of the team.) Next, a large 3x3 priority matrix of
impact versus time was made on the wall (see Figure 3), and
definitions were established (e.g., what is meant by short,
medium, and long term). The notes were placed randomly
on the matrix and then moved around until a consensus was
reached (i.e., no more movement).
Sometimes there would be apparent disagreement regard-
ing the position of an item, and when that occurred it was
helpful to ask for clarification (e.g., to see if the parties
involved were interpreting it differently); often the solution
was to split one action item into two or more items, which
then fitted into different parts of the matrix. Occasionally,
too many items were posted under High Impact. This situa-
tion was alleviated by asking "If we do this item, does it help
the other item, or the reverse?" or "If this item isn't done,
does it prevent us from making progress?"
Subsequently, tasks were allocated among team members
using a deployment chart (see Figure 4). Low-priority items
were disregarded or performed only after more critical items
were completed. While team members volunteered for items
they felt they could do well, the facilitator helped to ensure
1) that everyone got a blend of short-, medium- and long-
term action items to help maintain an even workload, 2) that
the high-impact items were not given to just one person, and
3) that all team members got some tasks with which they
were comfortable. Some items required several or all team
members and some teams agreed on specific deadlines for
one or two critical items. The results were written up and
circulated to the team (and to the instructor). Development
of the priority matrix and deployment chart took only one to
one and one-half hours per team, and it got everyone in-
volved in the project at an early stage. The session demon-
strated to both the undergraduate and graduate team mem-
bers that everyone could make significant contributions to
the project, and it helped motivate them to do so. As the
work progressed, each team continued to evaluate its progress
and, where necessary, prioritized in more detail (especially
when time was running out).
There are many other places in the curriculum where man-
agement tools could be demonstrated to the students. For
28

example, in an "Introduction to Chemical Engineering"
course, an affinity diagram could be used to poll views on
why students want to major in chemical engineering. Also, a
relations diagram could show students the impact of safety
issues in the plant or to help students didcide which elec-
tive courses would be most beneficial to them in meeting
their goals. Priority matrices could help students plan
laboratory exercises or focus attention on critical unit
operations in a design project.
In any of these applications, it is important to realize that
while the instructor provides input on technical issues, his or
her role as facilitator is solely to maintain focus, to keep the
process on schedule, and to ensure that participation is bal-
anced. In particular, the instructor should resist any tempta-
tion to direct the outcome since there is no "right answer."
The important issue is that the results obtained from apply-
ing the tools truly reflect the consensus of opinion within
that group. For example, writing the final report was re-
garded as a high-priority item by some teams but as a me-
dium-priority item by others. Nevertheless, in each case, the
team members believed in their result, and their report was
completed successfully and on time. There are invariably
several routes to success.

APPLICATIONS IN STUDENT ACTIVITIES
The management tools can also be applied effectively
outside the classroom. One example is with student societ-

Short Term Medium Term Long Term
Develop Disribute Finish calculations Content
no impac t rtb. . andt..estthemod.el I of report

Wria me D3.c Prroo ea roa Pe ate or -
(Critical) of- Asrto r _Sfe
npamental das ase tion C presto Jer

betedneg Re nr Develop coelatin bet-
Medium impact I rmt r ti m a mhetanuarclogand
Perorm cculations sDecide on length hamonuclear adsorption
Pbeoo Cculations d T andMdepth tr o port J
R using abon e memhanrsm a dept-.-reor
O ercde on
Spraeentaton a

Low Impact Prepare for reading Prepare or turning-in
(not essential) .sess onnlsJ5 grup m eeaina tools

Figure 3. Priority matrix for a team project titled "Ener-
getics of Adsorption Onto Metal Surfaces."

Task Kelly Peter Chatphoi Jerry
Getting Reference Matenal g
Postulate Mechanisms
Perform Calculations and Test Model __
Reading Session on Monday
Plan the Report______
Write the Report
Presentation

Figure 4. Deployment chart for the priority matrix shown
in Figure 3. A shaded rectangle denotes responsibility and
an oval rectangle denotes assistance.

Chemical Engineering Education

ies. In our (QXE Honor Society we had previously initiated a
lunchtime lecture series (together with an equivalent group
from mechanical engineering), but the attendance had been
poor and there had been some disagreement about topics
(e.g., more talks focusing on one discipline than the other).
Consequently, the eight chapter officers used Post-It notes to
come up with ideas for seminar topics, sorted them into n
groups (as done for the affinity diagram), and asked their
constituents (the undergraduates) to rank the ideas by voting
for the n/3 they wanted the most (see Table 2). This is
more expeditious than applying the full nominal group
technique (NGT),[2] which would require each person to
vote on every topic (giving n points to their top choice,
n-1 points to the next choice, etc.).
The officers also sought information on how many semi-
nars people would be interested in attending per semester.
This motivated them to arrange seminars for the top three
choices and it avoided any potential conflict over subject
matter. It turned out that the top choices were "how to get a
job," "what engineers do," and "computing." For these top-
ics, the chapter officers made suggestions and agreed on
both format and speakers. Then QXE held a preparative
meeting where members used Post-It notes to come up with
ideas for questions. The highest priority questions were also
established by a variation of NGT: each student placed an
adhesive color coding label next to the n/3 questions (out of
n in total) they were most interested in, and those getting the
largest number of labels were regarded as most important.
The high-priority questions were then sent to the speakers so
that they could present the most pertinent information.

TABLE 2
Possible Topics for a Seminar Series
The numbers are the percentage of total votes (506) recorded for each topic.
Each student had up to four votes.
% Topic and Examples
17.7 How to Find a Job Interviewing skills; job-hunting skills; working with
consulting firms; cooperative education/internships
14.2 What Engineers Do Frontier areas of engineering/future directions;
employee experiences from industry; consulting
11.9 Computing Use of computers in engineering; how to use the computer
laboratory; using your HP (beginning to advanced)
10.0 Specific Areas of Study Nontraditional areas (materials); lasers;
aeronautical/astronautical engineering; academic research topics
7.9 Writing Reports
7.5 Information on Graduate School
7 1 Design Projects
5 7 Teamwnork How engineers Kork together: total quality management
5.3 Feedback Question/answer sessions with seniors; senior honors thesis
4.7 Experimental Design/Statistics
3.8 Safety Safety and environmental concerns
2.4 Another Lecture by Prof. Lienhard
1.8 Ethics

Winter 1996

The outcome has been highly successful seminars with
attendance three to five times larger than before. The stu-
dents feel that this is "their" seminar series and the officers
make the seminar arrangements gladly. Moreover, at each
seminar, feedback forms are handed out to attendees so that
they can make comments, suggest improvements, etc. This
information in turn helps the students prepare future activi-
ties that, hopefully, will be even more successful.

APPLICATIONS IN THE DEPARTMENT
An additional way that students can see TQM in action is
for faculty and departments to practice it. Examples could
include getting feedback from advises and alumni (e.g., to
see if we are providing what the students really need) and
establishing directives for offerings in both electives and
continuing education.
At the departmental level, the management tools can be
applied to operations that do not involve the students di-
rectly, e.g., establishing vision-mission goals for the depart-
ment and, in turn, setting priorities for faculty hiring, allocat-
ing resources and service tasks, and developing criteria/
measurements of success. In addition, the tools can help
identify more effective methods for recruiting the top
graduate students and interfacing with (local) industry.
For the staff, one can set up "quality circles" where the
tools are used to pinpoint systemic problems and help
find ways to streamline them.[61
In most departments, there will be some "low hanging
fruit" that can provide relatively short-term successes and
serve to illustrate the benefits of the approach. However,
application of the tools at the departmental level will fail
unless there is a clear commitment from the administration
(e.g., the Dean and the Chairperson) to embrace the results
of the TQM process;t7" if people go through the process
only to have administrators manipulate the results, it is
even more detrimental to morale than not doing it at all.
Conversely, if the Chairperson and the facilitator are
trustworthy and do not have any vested interests, appli-
cation of the tools could make a significant contribution
to the success of a department.

REFERENCES
1. Deming, W.E., Out of the Crisis, MIT Center for Advanced
Engineering Study, Cambridge, MA (1986)
2. Brassard, M., Memory Jogger Plus+, Goal/QPC, Methuen,
MA (1989)
3. Scholtes, P.R., The Team Handbook, Joiner, Madison, WI
(1988)
4. Covey, S.R., The Seven Habits of Highly Effective People,
Simon and Schuster, New York, NY (1990)
5. Kroeger, 0., and J.M. Thuesen, Type Talk at Work, Delacorte
Press, New York, NY (1992)
6. Ingle, S., and N. Ingle, Quality Circles in Service Industries,
Prentice-Hall, Englewood Cliffs, NJ (1983)
7. Brown, M.G., D.E. Hitchcock, and M.L. Willard, Why TQM
Fails and What To Do About It, R.D. Irwin, New York, NY
(1994) 0

classroom

APPLICATION OF

QUALITY MANAGEMENT TECHNIQUES

TO ChE PROCESSES

MARY ANN PICKNER,* BAHMAN GHORASHI, ANNE M. GHORASHI**
Cleveland State University Cleveland, OH 44115

his paper reports on a study we performed on the
application of the Deming Management method to a
practical chemical engineering project. Our objec-
tive was to examine a practical application of the manage-
ment technique when applied to a chemical engineering
process. We will first describe some of the basic principles
which Deming's technique is based upon, such as under-
standing that any job consists of a group of consecutive tasks
(a process) and that each process has inputs (suppliers) and
outputs (customers). Also, certain tools and procedures, such
as forming teams, will be discussed. Finally, an application
to a hypothetical chemical engineering industrial project,
patterned after a real industrial case, will be presented.
Deming's work initially became popular in Japan in 1950,

- Mary Ann Pickner is currently a Process Devel-
opment Group Leader in the Process Develop-
ment Department of Lubrizol (Wickliffe, Ohio)
where she has been since 1987, three years as
a process development engineer and two years
as a technical engineer.

Bahman Ghorashi received his BS from Wayne
State University and his MS and PhD from The
Ohio State University. He joined the Chemical
Engineering Department at Cleveland State Uni-
versity in 1978 and is presently Professor and
Graduate Program Director.

Anne Ghorashi has over sixteen years experi-
ence working in industry. She has worked at AGA
Gas Inc., for the past five years as a project man-
ager, responsible for projects involving application
and system development for air gas industry.

Lubrizol Corporation
** AGA Gas Inc.
30

Our objective was to examine a practical
application of the [Deming] management
technique when applied to a chemical
engineering process.

and in 1980 it began to take hold in the United States.
Deming studied with Shewhart, learning his ideas of "statis-
tical control." This became the basis of Deming's work. He
is the author of Quality, Productivity, and the Competitive
Position,t11 and Out of the Crisis, as well as other technical
books and brochures on statistics and sampling together with
numerous scholarly studies.[2] Several other related and per-
tinent studies are cited in the references to this paper.t3-121
Some of Deming's basic concepts are captured in his
fourteen points:
Create constancy of purpose for the improvement of product
and service
Adopt the new philosophy
Cease dependence on mass inspection
End the practice of awarding business on price tag alone
Improve constantly and forever the system of production
and service
Institute training and retraining
Institute leadership
Drive out fear
Break down barriers between staff areas
Eliminate slogans, exhortations, and targets for the
workforce
Eliminate numerical quotas
Remove barriers to pride of workmanship
Institute a vigorous program of education and retraining
Take action to accomplish the transformation
Deming's management methods are different from those
that have been traditionally taught. A summary of some of
Copyright ChE Division ofASEE 1996
Chemical Engineering Education

the more contradictory ideas can be found in Table 1.

Application of Quality Management Techniques to a
Hypothetical ChE Industrial Project
In this hypothetical project, patterned after a real industrial
project, the goal was to increase the capacity of a system
without investing capital. Ultimately, the required increased
capacity was obtained by process improvement through re-
ducing control time cycles.
Formation of a Team A team was formed to study the
process and to determine whether a capacity increase could
be made by reducing the existing controlling time cycles
instead of investing capital. Each person on the team was
chosen to bring expertise from his/her respective area. (Form-
ing a team is in line with Deming's point to "break down
barriers between staff areas.") The team members and their
corresponding areas of expertise were:
Production Technologist processing; time cycles and
equipment knowledge
Quality Assurance Lab Coordinator analytical
specification; alternative analytical methods

TABLE 1
Comparison of Traditional Business Techniques and Deming's Mi

Traditional Approach
Quality is expensive.

Inspection is the key to quality.

Quality control experts and inspectors can
assure quality.

Defects are caused by workers.
The manufacturing process can be opti-
mized by outside experts. No change in
system afterward. No input from workers.

Use of work standards, quotas, and goals
can help productivity.

Fear and reward are proper ways to motivate.

Deming's Approach
Quality leads to lower cost.

Inspection is too late. If workers
produce defect-free goods, elimir
inspections.

Quality is made in the boardroom

Most defects are caused by the pi
Process is never optimized; it can
always be improved.

Elimination of all work standards
quotas is necessary

Fear leads to disaster.

Buy at lowest cost. Buy from vendors committed to

Figure 1. Process flow diagram.

Winter 1996

Process Development Engineer process variable
effects; laboratory capability
Customer Service Representative customer concerns
Marketing Representative projected future demands
Process Operator hands-on experience in running the
process; knowledge of process problems; potential
alternatives
Operations Manager authority to approve process
changes or to authorize facility capital investments

In this case, the leader of the team was chosen to be the
production technologist. The entire team did not attend all
meetings. The marketing, customer service, and plant man-
agement representatives attended the first meeting to set the
stage and participated later as deemed appropriate. The team
chose to use a structured strategic problem-solving approach.
Note that the case described here is patterned after a case
in a large chemical company. Nonetheless, these principles
can be easily modified so that they are applicable to opera-
tions much smaller in size and personnel.
Application of Strategic Problem-Solving Ap-
proach In this case, the system being studied
consisted of a series of four processes. The first
ethod process was a reactor, the second a neutralizer, the
third a stripper, and the fourth a filtration unit.
These four units are shown in Figure 1.

1. The first step was to define the problem state-
can ment.
nate
The goal of the project was to determine if it
was possible to reduce the controlling time cycle
from 15.1 to 7.5 hours in order to increase capac-
ity from 5000 MT/yr to 10,000 MT/yr. The limita-
ocess. tion on the process changes was that the product
quality should not be compromised as determined
by analysis or performance tests.
2. The next step was to define the reason for
and improvement.

The reason for improvement was to increase the
capacity to meet the required customer demand in
quality. two years without a $5 MM capital investment
required to obtain the shorter time cycle.

3. The next step was to define the current situa-
tion.
As described previously, four processes con-
tributed to the existing system. Since the goal of
the team was to reduce the controlling time cycle,
the first step in describing the current situation
was to determine the time cycles of each process
and ultimately define the controlling time cycle(s).
Therefore, the production technologist gathered

31

data from the last forty batches to identify the time cycle for
each process. The control charts from these batches for each
of the four consecutive process steps were then constructed.
From the time cycle control charts, it was clear that the
reaction and the stripping processes had relatively longer
controlling time cycles. The production technologist applied
statistical tools to the data from each process by calculating
the mean, standard deviation, and upper and lower control
limits. The statistical time cycles for these four processes are
shown in Table 2.
The time-cycle analysis illustrated clearly that any reduc-
tion in the controlling time cycle had to be achieved by
reducing the reaction and stripping time cycles. Therefore,
to better understand the causes of the pro-
cesses having the longer time cycles, the T
team further analyzed the reaction and strip- Slatisti
ping steps using control charts..
4. Analysis of the chemical reaction. Unit
The team noticed from the control charts Reactor Time
Neutralizer Tin
that the reaction time cycle was out of con-
trol. The average reaction time cycle was Stnpper Lme e
12.7 hours, with a large standard deviation Fitraton Time

of 4 giving an expected range
of 0.7 to 24.7 hours based on
plus or minus three standard
deviations. Not only was the
range large, but also there ex-
isted special causes in the sys-
tem. The data appeared to have
some nonrandom patterns that
were viewed as a clue to a
special cause. Also, a second
special cause pattern was ob-
served since more than 8
points were below the mean.

very short, the reaction pressure had dropped. This would
parallel the steaming out of the condenser, i.e., the con-
denser was steamed in four batches.
Since the operators do not record this information on
the batch sheet, the production technologist never no-
ticed the increased pressure or the fact that they were
steaming the overhead.
A second trend on the chart also indicated a special cause
where the batches with exceptionally long reaction time
were not part of the upward trend effect. Three batches
having a reaction time of 20, 18, and 20 appeared suspicious,
although these points were not outside of the control limits.
The team looked on the batch sheets for these three points

MATERIALS OPERATORS I
Solvent
Catalyst Charge Accuracy
Reactant A Undercharge
Reactant B Reactant B
Reactant C (by drum)
Supplier x
Supplier y
S/ Pump Failure
Reaction incomplete / Mixer Failure
/ / Condenser Plugage
EASURENTS EQUIPMENT

Figure 2. Cause-and-effect diagram.

Once special causes were identified, the team studied data
obtained by the operator from the previous batches in order
to identify the variation's cause. To assist in their search,
the team had a meeting where they brainstormed the
possible causes for the time cycle variation and used a
cause-and-effect diagram, shown in Figure 2, to identify
the most likely causes.
During the construction of the cause-and-effect diagram,
the operator explained that when the time cycle gets excep-
tionally long, the next batch would run quicker if the over-
head condenser of the reactor was steamed out prior to
starting the batch. The production technologist thought this
comment was rather interesting as he theorized that if the
reactor vent was plugged, the desired reflux condition could
no longer be obtained and perhaps this would result in an
extended reaction time cycle. In light of this, the team
looked more closely at the batch sheets and noticed that
each time the reaction time cycle dropped from long to
32

and noticed that the final pounds out of the
batch were higher by approximately 200
pounds, as indicated by the manometer read-
ing. The team determined that possible ex-
planations could be equipment failure or
overcharge of a raw material. Later, the
operator mentioned that sometimes a whole
drum of raw material B was charged in-
stead of weighing out 100 lbs because it
was believed that the accuracy
of this charge was not all that
critical since it was only about
one drum in a 5000-gallon re-
actor. The production tech-
nologist was not absolutely
Longer certain of the sensitivity of the
Time product to raw material B, but
thought that based on the data
it would be worth ensuring the
raw material charges were
more accurately measured
from batch to batch. To assist

with this situation, the team
worked through the Purchasing Department to get the mate-
rial packed in the amount that they were to charge to each
batch so that weighing was not necessary.
It is worth noting that the operator spoke freely about not
charging the batch correctly because he knew the team was
interested in achieving improvement and not in pointing
fingers. This is in line with Deming's point to "drive out
fear" in the workplace. Important information can be lost
when people are afraid to come forward.
5. Solutions to reaction bottleneck.
The team implemented what was thought to be the solu-
tion to removing the special cause variation to the process by
requesting steaming of the condenser between batches. This
would ensure that the vent would be clear. Also, packaging
the raw material B in preweighed containers would elimi-
nate weighing and would help ensure that overcharges of
material B would not occur.
Chemical Engineering Education

ABLE 2
al Time Cycles
Cycle Hours
C.cle 12.7 12.0
ne Cycle 5.43.9
C.cle 15 1 60
Cycle 4.5 t 4.2

6. Results/analysis from eliminating reaction special causes
Next, the team monitored the process for another forty
batches. Based on the results, it appeared that the upward
trend seen with steaming the condenser had disappeared and
that the spikes in the data were reduced. But several points
again appeared below the mean, indicating a special cause.
The team theorized that the factors causing the reaction
time to go faster on a series of batches such as this would
probably be related to a raw material stock. All their raw
materials were single sourced except for one-raw material
C. Since it is more likely to have variability between two
suppliers as opposed to two lots from one supplier, it is
preferable to have only one supplier. The team looked fur-
ther to see if lots were received from two different suppliers
for raw material C during this time period and found that
producer x supplied during the period that the points below
the mean were obtained, while producer y supplied during
the period for other points.

The team requested material from only T.
supplier x for the next forty batches and Improvem
found that the process was in control, with T
no indications of special causes present. In Process Step
addition, the time cycle was reduced to an Reaction
average of 6.0 with an Upper Control Limit Neutralization
(UCL) of 9.2 and a Lower Control Limit Stripping
(LCL) of 2.8, which was a great improve- Filtration
ment over the original process having a UCL
of 25 and a LCL of 0.7. Now the team was in a position to
discuss the possibility of additional process improvements
to further reduce the time cycle.
Since all other process steps were 6 hours or less at this
point, the stripping time had to be reduced next. As men-
tioned earlier, the average stripping time was 15.1 hours.
From the data, the stripping process appeared to be in con-
trol, as no indications of being out of control were apparent.
The only way to reduce the time cycle was, therefore, to
somehow change the system. Since the process was in con-
trol, changes to the system could be made and the effects
observed without concern of mislead by special causes.
The team discussed alternatives to reducing the stripping
time. Since the team had to avoid a capital expenditure, the
idea that seemed most promising was the use of steam strip-
ping to remove the solvent. Furthermore, since this was a
process change, the process development engineer had to
evaluate the ideas and provide technical support.
This idea was proven in the lab, and a plant trial was
successfully completed. The resulting process showed that
the new average time for stripping was only 6.5 hours,
allowing the whole system to have a controlled time
cycle of less than 7.5 hours and resulting in the desired
doubling in capacity.
Table 3 shows the improvements that were made in the
Winter 1996

process time cycles. In order to ensure that the reaction and
stripping process improvements were implemented in a man-
ner to ensure consistency, a new process batch sheet was
written that incorporated the process changes.
The original goals of the team were now met. Further
discussion revealed that perhaps additional work on convert-
ing this process into a continuous process, as opposed to a
batch process, might be beneficial since the demand had
continued to grow through the years. The team drafted a
memo to process development suggesting the long-term in-
vestigation of such an idea.

CONCLUSIONS
In this hypothetical study, the fundamental definitions and
tools introduced by Deming were used. The tools were ap-
plied to a hypothetical chemical engineering project, pat-
terned after an industrial project, that incorporated real pro-
cess industry occurrences. They were both
LE 3 technical (steaming of the condenser be-
in the Process tween batches and steam stripping to re-
Cycle move the solvent) as well as procedural
before After (working with the suppliers to pack the
712 6.03.3 raw material in preweighed containers
3.9 5.4 3.9 to eliminate weighing and overcharge
1 6.0 6.5 4.2 as well as using a single source for raw
4.2 4.5 4.2 material supply). In addition to the stan-
dard chemical engineering principles,
we hope that these fundamental tools will be used by
others to enhance the capabilities of the more traditional
technical problem solving methods.

REFERENCES
1. Deming, W. Edward, Quality, Productivity, and Competi-
tive Position, Massachusetts Institute of Technology, Cen-
ter for Advanced Engineering Study (1982)
2. Walton, Mary, The Deming Management Method, Dodd,
Mead & Company, Inc., (1986)
3. Aguayo, Rafael, Dr. Deming: The American Who Taught the
Japanese About Quality," First Carol Publishing (1990)
4. Burr, I. W., Statistical Quality Control Methods, Marcel
Dekker Inc., New York, NY (1976)
5. Grant, E.L., Statistical Quality Control, McGraw-Hill, New
York, NY (1980)
6. Ishikawa, K., What is Total Quality Control? The Japanese
Way, Prentice-Hall, Inc., Englewood Cliffs, NJ (1985)
7. Jessup, P.T., Continuing Process Control, Corporate Qual-
ity Education and Training Center, Ford Motor Company
(1987)
8. Gryna, Juran, and Frank M. Gryna, Quality Planning and
Analysis, McGraw-Hill Book Co., New York, NY (1980)
9. Schltess, Peter R., The Team Handbook, Joiner Associates,
Inc., Madison, WI (1988)
10. Shewhart, Walter A., Economic Control of Manufactured
Product," Van Nostrand (1931); republished ASAC (1980)
11. Walton, Mary, Deming Managements at Work, G.P. Putnam's
Sons Publishers (1990)
12. Wheeler, Donald J. Understanding Statistical Process Con-
trol, Statistical Process Controls, Inc (1986) J

AB]
ients
ime

12.7
5.4
15.
4.5

Random Thoughts...

THE WARM WINDS OF CHANGE

RICHARD M. FIELDER
North Carolina State University Raleigh, NC 27695-7905

It might surprise you to learn that some people find me a
bit pessimistic. Somewhat cynical, a few would add. A
perpetually grumpy curmudgeon whose patron saint is
Eeyore, one might mutter (but she's only my wife-what
does she know?). This image is reinforced by how I spend
much of my time these days, writing papers and giving
speeches about the woeful state of practically everything
involving education in this country.
My dark reputation notwithstanding, I am currently more
hopeful than I have ever been about the direction of higher
education in general and engineering education in particular.
To lift the spirits of those who share my disposition to
gloom-and to prove that Rebecca is all wrong about me-I
offer my reasons for this unaccustomed optimism, starting
with some that might at first appear negative and depressing.
(All right, so maybe Rebecca isn't completely wrong.)
Growing Pressures to Upgrade Undergraduate Engi-
neering Education U Engineering schools are going through
turbulent times these days. The pool of qualified applicants
is shrinking and the dropout rate is higher than ever, leading
to losses in tuition revenues and state funding. Significant
numbers of entering students need remedial work in math-
ematics, science, and English, severely stretching campus
teaching and advising resources. Industrial recruiters and
supervisors complain that most engineering graduates lack
the skills (teamwork, writing, speaking, etc.) they need to
succeed in the workplace. Legislators, trustees, faculty mem-
bers, and students have begun to question-sometimes un-
fairly, sometimes with good cause-the minimal teaching
loads and low status of teaching at most research universi-
ties, and chancellors and deans are feeling increasing pres-
sure to respond with more than rhetoric. Traditional sources
of research funding are drying up and the PhD job market is
anemic, providing still more incentive to upgrade under-
graduate education. On the positive side, external funding
for improving teaching and advising has been growing, led
by the NSF's impressive investment in engineering educa-
tion in the past decade through Division of Undergraduate
Education grants and the education coalitions.
To be sure, most of these developments are not exactly
cause for celebration. The financial crunches at most univer-

sities are real and severe, the survival of some academic
programs and positions may be in jeopardy, and the thought
of legislators and politically appointed trustees attempting to
dictate academic policy is truly frightening. But our profes-
sion has weathered financial and enrollment crises before,
and I have no doubt that we will get through this one too.
What is different about this crisis is its potential positive
impact on the quality of undergraduate education. The trends
just described-notably the rising chorus of complaints about
the status and quality of undergraduate education and the
availability of external support for improving teaching and
advising-have given rise to an eruption of curriculum re-
form initiatives and innovative teaching and advising pro-
grams. Faculty participation in these programs is increasing
rapidly, and even professors who are not active partici-
pants-including some who are heavily involved in re-
search-are starting to examine their own teaching and to
explore ways of doing it better. Consider some examples.
Innovative Programs, Methods, and Instructional Mate-
rials U Current reform efforts involve virtually every aspect
of engineering education. There is Purdue's proactive coun-
seling program for freshman engineering students; minority
education programs at Arizona State, California State at Los
Angeles, and Georgia Tech; instructional software develop-
ment at Cornell and other schools in the SYNTHESIS Coali-
tion and also at Michigan, Connecticut, Virginia Tech, and
Purdue-Kokomo; integrated freshman engineering curicula
at Texas A&M, Rose-Hulman, and other schools in the
FOUNDATION Coalition, and also at Drexel, North Caro-
lina State, and Colorado; freshman engineering design and
laboratory programs at Maryland, Florida, and other schools
in the ECSEL and SUCCEED Coalitions, and also at the
Colorado School of Mines, Pittsburgh, and Wisconsin; and
other programs designed to help students develop skills in
problem solving, computer applications, creative and critical
thinking, teamwork, and communication. Some of these pro-
grams are experimental, but more and more are becoming
institutionalized on a large scale.
Leadership. Reformers and innovators have been around
in engineering education for many years. In the 1960s and
Copyright ChE Division ofASEE 1996
Chemical Engineering Education

1970s, folks like Jim Stice, Don Woods, Charles Wales,
Helen Plants, John Lindenlaub, Billy Koen, Lee Harrisberger,
Larry Grayson, and Lois Greenfield were a congenial bunch
who did wonderful work and had some memorable times at
gatherings of the Educational Research and Methods Divi-
sion of the ASEE. For years their numbers did not grow,
however, and their calls for educational reform went largely
unheeded outside of their own dedicated community.
Many of those pioneers are still gratifyingly active, but
now their ranks are swelling as younger colleagues enter the
game with growing effectiveness. In the literature and on
campuses around the country you can see the influence of
creative educators like Karl Smith of Minnesota, Phil Wankat,
Bill LeBold, and Dan Budny of Purdue, Dick Culver of
SUNY-Binghamton, Ed Lumsdaine of Michigan Tech, Ray
Landis of Cal State-Los Angeles, Susan Montgomery and
Scott Fogler of Michigan, Steve LeBlanc of Toledo, Doug
Cooper of Connecticut, Cindy Atman and Larry Shuman of
Pittsburgh, Don Evans and Lynn Bellamy of Arizona State,
Tom Regan of Maryland, Charley Yokomoto of IUPUI, and
Ron Miller, Barbara Olds, Mike Pavelich, and Dendy Sloan
of the Colorado School of Mines.* Significantly, some of
the strongest participants in the reform movement are deans,
like Lyle Feisel at SUNY-Binghamton, Landis, and Shuman,
who are putting their talents, energy, and money behind the
usual administrative rhetoric about the supreme importance
of teaching on their campuses. Equally significantly, some
of the emerging leaders are untenured assistant professors,
whose deans and department heads are gambling that a few
new faculty members can be allowed to dedicate their ca-
reers to undergraduate education without causing the entire
system to collapse. This display of courage on the part of
both the administrators and the new professors is a particu-
larly hopeful sign.
Growing Faculty Interest in Educational Methods U
Some engineering professors-Smith, Stice, Wankat, Landis,
Woods, Fogler, and Felder, to name a few-regularly present
teaching workshops on campuses around the country. His-
torically, engineering professors have either been indiffer-
ent, skeptical, or disdainful toward teaching workshops, but
in recent years interest has skyrocketed. Some of us now get
more invitations than we can handle, and as many as 150
professors have shown up at a single workshop. Also, for the
past five years Jim Stice and I have codirected the National
Effective Teaching Institute at the Annual Meeting of the
ASEE. The NETI has reached over 250 professors so far and
is oversubscribed each year, to the point that Jim and I are
contemplating a second offering to accommodate the over-
flow. On many campuses, NETI participants have with our
encouragement used our workshop materials in their own

* These are just a few of the people whose innovative work I
admire. There are many more I would also have cited if I had
more space.
Winter 1996

faculty development programs.
In short, the growing pressures on universities to pay more
attention to the quality of their undergraduate education
programs, the availability of external funding to support
educational reform and innovation, the proliferation of pro-
grams to improve education on campuses around the coun-
try, the increasing faculty involvement in these programs,
and the increased willingness of professors to learn about
and try better ways to teach, all suggest that engineering
education is on the brink of a major renewal. Granted, the
same thing might have been said in other periods-most
recently in the early 1970s. Call me an incurable optimist if
you will, but I'm convinced that this time it's for real.

Epilogue: How can you get in on the action?

If you're a faculty member currently putting most of your
time and energy into disciplinary research, and you're doing
it successfully (as measured by, say, number of citations and
invited presentations, not just dollars and papers) and enjoy-
ing it, and you're also doing an adequate or better job of
teaching, you don't need to do anything differently. Aca-
demic research is a vital university function, and doing it at a
world-class level is a full-time pursuit. More power to you.
If, on the other hand, you have the inclination to improve
undergraduate education on your campus or just to improve
your own teaching, there are several ways to go about it.
Read McKeachie (Teaching Tips) and Wankat and Oreovicz
(Teaching Engineering). Join the ASEE, read Prism and the
Journal of Engineering Education (both of which come with
ASEE membership), and attend the annual ASEE confer-
ence or the Frontiers in Education conference to get ideas
and to avoid reinventing the wheel. If you hear or read about
new instructional software or a new approach to a course
you teach, think about giving it a test run. If a teaching
workshop is given on your campus or at a conference, invest
a few hours or days and take it, especially if you've heard
good things about it. Find out which of your campus col-
leagues are already involved in educational reform and see
what they're doing. If their work strikes you as important,
consider the possibility of participating. If you plan to try
something innovative to improve teaching in your depart-
ment or school, seek support for it (including release time
for you) from public and corporate funding agencies and
alumni-you might be surprised at how much is out there.
Finally, if your efforts to improve teaching quality
are successful, share your results at conferences and in jour-
nals, and make sure the administration, alumni, parents,
prospective students, trustees, legislature, and local
newspapers know about it. As with disciplinary research,
spreading the word about successes helps both the profes-
sion and your university's reputation. It won't do you any
harm either. 0

class and home problems

CHANGING VAPOR-LIQUID TRAFFIC

IN A DISTILLATION COLUMN

W. E. JONES, J. A. WILSON
University of Nottingham University Park Nottingham NG7 2RD England

hanging vapor-liquid traffic in a distillation column
is associated with the use of side-reboilers, side-
condensers, and pump-arounds. Incorporating some
of these features into a exercise tests the students' grasp of
the McCabe-Thiele construction and gives an elementary
insight into one aspect of heat integration.
Nowadays, most final designs for distillation columns are
prepared using a simulation package, so it is easy to dismiss
McCabe-Thiele construction as a routine piece of teaching.
But for more complex columns, such as those incorporating
a side-reboiler, for example, the ability to plan the design
roughly on a McCabe-Thiele diagram is a great help in
obtaining a swift convergence of the simulation program.
Hence, this exercise is of value in making students con-
sider the McCabe-Thiele construction as a flexible tool
rather than a rigid routine.
Side-reboilers, side-condensers, and pump-arounds are typi-
cally, but not very accurately, illustrated as shown in Figure
1."[121 Side-condensers and pump-arounds are associated with
column heat removal and are thus located above the feed.
Heat removal condenses vapor and the operating line is of a
shallower gradient above the point of heat removal than
below. Correspondingly, side-reboilers are located below
the feed and result in a steepening of the operating line
gradient below the side-reboiler.
The justification for changing the vapor-liquid traffic in a
distillation column is economic. Distillation columns are
major energy users, and efforts to reduce plant utility costs
can lead to energy integration requiring heat addition/re-
moval at locations other than the main reboiler/condenser.131
For example, lower temperatures found higher up the distil-
36

lation column mean that a cooler, and hence cheaper, heat-
ing medium (often heat recovered from within the plant) can
be used in the side-reboiler. To be set against this advantage
is the tendency of side-reboilers to narrow the driving force
between operating and equilibrium lines, resulting in the
separation requiring more theoretical stages. Similarly, side-
condensers and pump-arounds allow heat to be removed at a
higher, and hence more useful or cheaper (if refrigerated),
temperature level compared to the main condenser.
Side-reboilers and pump-arounds are the most commonly
encountered. Pump-arounds are generally preferred over side-
condensers because it is easier to engineer the liquid circuit
of the pump-around than vapor withdrawal to a side-con-

Tony Wilson holds BSc and PhD degrees in
chemical engineering from the University of
Nottingham. With industrial and consulting ex-
perience in process control and batch process
engineering, and with active research in both
fields, he coordinates the department's research
in computer-aided process engineering and is
responsible for process control teaching at the
undergraduate level

Copyright ChE Division ofASEE 1996
Chemical Engineering Education

The object of this column is to enhance our readers' collections of interesting and novel
problems in chemical engineering. Problems of the type that can be used to motivate the student
by presenting a particular principle in class, or in a new light, or that can be assigned as a novel
home problem, are requested, as well as those that are more traditional in nature and which
elucidate difficult concepts. Please submit them to Professor James 0. Wilkes (e-mail:
wilkes@engin.umich.edu) or Mark A. Bums (e-mail: mabums@engin.umich.edu), Chemical
Engineering Department, University of Michigan, Ann Arbor, MI 48109-2136.

Warren Jones holds BSc and PhD degrees in
chemical engineering from the University of
Nottingham and is a registered Chartered Engi-
neer. He has a wide-ranging interest in both front-
end processes and detailed plant design, devel-
oped initially through nine years of experience
with a major engineering and construction com-
pany. Teaching responsibilities include several
design courses, process economics, and engi-
neering thermodynamics.

denser. Further, side-reboilers and side-
condensers are the easiest to analyze rig-
orously on the McCabe-Thiele construc-
tion. Therefore, a side-reboiler case has
been chosen as the basis for the main ex-
ample presented here. At the end, guid-
ance is given on setting up a pump-
around example, based on a simplifying
assumption. The side-condenser case is
a simple variation on the main example
and is left to the reader.

Figure 1. Typical representation
of distillation column with multiple
heat additions/removals

Figure 2. Distillation column with
side-reboiler

PRACTICAL IMPLEMENTATION
A practical side-reboiler arrangement is shown schematically in Figure 2,
where x and y denote the mole fraction of the more volatile component in the
liquid and vapor phases. The design is based on total liquid trap-out to a once-
through thermosyphon reboiler, where the stream is partially vaporized and the
equilibrium two-phase mixture is returned to the column. The equilibrium vapor,
P, combines with the ascending vapor, V", and the equilibrium liquid, L",
descends to the main reboiler. Partial vaporization in the reboiler is important to
reduce fouling and to maintain good heat transfer. A maximum vaporization of
20-25% of the feed to the reboiler is often used. Kister has presented an analysis
based on partial liquid trap-out followed by total vaporization of the liquid.[41 This
analysis would show a deep notch on the McCabe-Thiele construction, which
effectively reverses some of the separation effected in the lower section of
the column. Kayihan's McCabe-Thiele construction"1 shows the notch, but
no analysis is presented.

PROBLEM STATEMENT
Sketch the operating lines for a binary distillation column incorporating a
thermosyphon side-reboiler. Pay particular attention to the operating line
end-points around the side-reboiler. You should make the usual simplify-
ing assumptions and use the nomenclature in Figure 2. Figure 2 assumes a
total condenser on the overheads and a recirculating thermosyphon as the
main reboiler.
SIf the relative volatility of the two components in the binary mixture is
denoted by a, show that x, and x' are related by

Xn =' ( Pa +
xn t- (c J+L"

Saturated liquid comprising 50 mol % A and 50 mol % B is fed to a distillation
column. The distillate is to contain 95 mol % A and the bottoms 95 mol % B.
The relative volatility of A with respect to B is 2.5. Estimate the number of
theoretical stages required for the separation, assuming:
Reflux ratio is 1.4 times the minimum reflux ratio
25% of the liquid fed to the side-reboiler is vaporized
Temperature level of the heat input to the side-reboiler is such that liquid
containing a minimum of 35 mol % A can be vaporized
@What proportion of the total heat input in (c) is made through the side-
reboiler? Compare the number of theoretical stages and total heat input
required for the side-reboiler case with that required for a simple distillation
column operating with a reflux ratio of 1.4 times the minimum reflux ratio.
If you attempt to add the side-reboiler heat at successively lower temperature
levels, what limitation do you reach? Illustrate your answer using the relevant
information from (c).
( For discussion: A preliminary design recommendation commonly quoted for
higher energy cost regions is to use an operating reflux ratio 1.2 to 1.3 times
the minimum reflux ratio. Why might it be appropriate to use a higher factor,
say 1.4 to 1.5, when considering a side-reboiler?

SOLUTION

0 a. The distillation column representation will incorporate three operating lines.
The operating line applying above the feed will conform to the normal McCabe-

Winter 1996

SProduct

Reflux Ratio
R = L/D

Main
Reboiler

Thiele construction. Below the feed there will be two operat-
ing lines: line A applying below the side-reboiler (and down
to the main reboiler), and the other (line B) applying above
the side-reboiler (and up to the feed plate). The important
point to note is that, although the two operating lines below
the feed have different gradients (L'/ V' and L" / V"), they
pass through the same point, xo on the 450 line, because no
side-product is taken. Figure 3 illustrates the construction
where a saturated liquid feed has been assumed.
The actual construction is straightforward for a given feed
and required separation, and knowing xD and selecting R
permits construction of the top operating line. The intersec-
tion of the top operating line with the q-line gives one end of
operating line B; the other end is xo on the 450 line.
Operating line A can be added because we know it passes
through xo and has gradient L"/V". Also, L" and V" are
easily calculated from L' and v' by the equations
L" = L'- P
V"= V'-P

and L', V' are found in turn from L, V, F, and the feed
condition.
The one outstanding problem concerns the transition from
the operating line A to B. Point x', y,,n- is located at the end
of operating line A. Point x', y' represents the equilibrium
mixture returning from the side-reboiler. The vapor en-
tering the section of column above the side-reboiler is a
blend of compositions y' and yn-i, so y" must lie between
these two values and the point xn, y" must lie at the lower
end of operating line B.
Note the above design avoids the deep notch previously
mentioned. But care is needed when drawing the theoretical
stages. Stages can be drawn in the normal manner, com-
mencing at XD on the 45' line and terminating at xn, y". A
discontinuity occurs between x,, y" and x',yn-,, and the stages
will be recommended at the latter point, terminating at xo.

We cannot simply draw steps over the transition region, as
suggested by Petterson and Wells.131
b. The equation, which is useful for specifying the transi-
tion between operating lines A and B, is easily derived using
a componential mass balance on the side-reboiler
L'x, = Py'+ L"x'

where y' is eliminated using the equilibrium relationship

1 +(a-1)x'

to give

x=x'( Pa L"
L' 1 + (a 1)x'
c. After drawing the equilibrium line, a minimum reflux
ratio of 1.1 is easily found from the gradient of the operating
line giving an infinite number of theoretical stages at a feed
composition, x, = 0.5. The reflux ratio to be used in opera-
tion is 1.4 x 1.1 = 1.54, and this implies an intercept of 0.677
on the q-line.
Below the feed, operating line B must have slope
L' / V' = (0.677 0.05) /(0.5 0.05) = 1.393
while below the side-reboiler, operating line A must have
slope
L" L' -P 0.75 x 1.393 x V' 603
V" V'-P V'- (0.25 x 1.393 xV')

This implies yn-, = 0.531 at x' = 0.35, completely defining
operating line A. All that remains is to establish xn on operat-
ing line B, and this is achieved using the equation derived in
(b)

0.35 0.25 x L' x 2.5
xn= L + (1.5x0.35) +(L'-0.25 L') 0.406

The completed construction is shown in Figure 4. Above
the side-reboiler, 7.7 theoretical stages are required and be-

Figure 3. Operating line construction
for distillation column with
side-reboiler

0.1 0.2 0.3 04 0.5 0.6 0.7 0.8 0.9 1.0 x
Figure 4. McCabe-Thiele construc-
tion for the problem.

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 x

Figure 5. Introduction of a pinch by a
side-reboiler.

Chemical Engineering Education

low 6.5 are needed, but this includes a recirculating
thermosyphon reboiler which is not one theoretical stage151
but may be taken as roughly 0.5 of a stage. Hence, the total
number of theoretical stages is 13.7.

d. The usual simplifying assumptions for McCabe-Thiele
construction require the components to have equal latent
heats of vaporization. Hence, to compare heat inputs, we
simply need to compare vapor flows.
Above the side-reboiler, the vapor flow is directly related
to the total heat input and we know V' = L'/1.393 = 0.718 L'.
In the side-reboiler, vaporization P = 0.25 L', and hence
percentage heat input through the side-reboiler = (0.25 x
100)/0.718 = 34.8%.
In the case of the simple distillation column, operating line
B (from the side-reboiler case) now applies at all points
below the feed. This immediately tells us the total heat input
must be the same for the two cases. To complete the com-
parison, we need the number of theoretical stages for the
simple distillation column, and this is easily obtained by
stepping off along operating line B and its extension to give
12.2 stages, after allowance for the reboiler. (This con-
struction is not shown in Figure 4.) Hence, we have a
trade-off in which 34.8% of the heat is saved at the
highest level in return for installation of an extra 1.5
theoretical stages plus side reboiler.

0 e. Successively lower temperature levels for heat addition
means that the increasing amounts of more volatile compo-
nent A must remain in the unvaporized side-reboiler return
liquid. This is equivalent to lengthening operating line A and
correspondingly shortening B. Ultimately, an alternative pinch
would be generated at 0.421, 0.645 on the equilibrium line
(illustrated in Figure 5).
Note that the transition
has become horizontal, as
can be expected.

N XD

V L
p+1
xp+l

p yx, --

v'= v+w
P- -- L = L+W
V L'

Figure 6. Distillation column
with a pump-around
Winter 1996

xp+l,y, -
xp,y -_
S C, slopeL/V

D, slope L'/V'

XD 1.0 x
Figure 7. Operating line construction
for distillation column with pump-
around.

f. Optimum reflux ratio is a balance between operating
costs and capital investment. Use of a lower-cost heating
medium for part of the heating effectively reduces the aver-
age heating-medium cost and this changes the balance, mov-
ing the optimum in a direction that reduces capital invest-
ment and permits slightly more energy consumption, e.g.,
increasing the factor from 1.2 1.3 to 1.4 1.5, say.

PUMP-AROUND
In theory, a pump-around can be viewed as operating over
one theoretical stage, as shown in Figure 6. Part of the liquid
exiting stage p is withdrawn and circulated through a cooler
before returning to the same stage. The circulation rate and
extent of cooling can be adjusted to control heat removal.
For analysis, we represent the heat removal as equivalent
to the latent heat released by "flow" W changing phase
from vapor to liquid.
Stage p+1 and above is represented as operating line C,
while stage p-1 and below is represented by operating line
D. Operating line C is of a shallower slope than D, and
noting y. and xp are in equilibrium (a consequence of the
theoretical stage assumption), then the relationship between
the end points of operating lines C and D adjacent to the
pump-around is as shown in Figure 7.
In summary, we have succeeded in representing the pump-
around as a theoretical stage "jumping" between operat-
ing lines. Operating line C can be drawn based on col-
umn reflux ratio, and D can be added by adjusting for
quantity changing phase, W.
Strictly, this representation is optimistic because cold liq-
uid returned to the column will not be heated to its bubble
point on one real tray; generally a few trays are needed. But
if the extra trays are added, then, by way of compensation,
the mass transfer will be better than indicated in Figure 7,
e.g., Xp and yp will no longer correspond to a point on the
equilibrium line. Provided these considerations are borne
in mind, this pump-around analysis makes a thought-
provoking exercise, leading to a clearer un-
derstanding of the topic.

ACKNOWLEDGMENT
We appreciate Carl Pulford's help with the
figures.

REFERENCES
1. Kayihan, F., AIChE J Sym. Ser. No. 192, 76, 1
(1980)
2. Lieberman, N.P., Troubleshooting Process Op-
erations, 2nd ed., PennWell, Tulsa, OK, p. 4 (1985)
3. Petterson, W.C., and T.A. Wells, Chem. Engg.,
84, 78; 26 September (1977)
4. Kister, H.Z., Chem. Engg., 92, 97; 21 January
(1985)
5. Jones, W.E., Chem. Eng. Ed., 27, 178 (1993) 0

Ml classroom

TEACHING TRANSPORT PHENOMENA

WITH INTERACTIVE COMPUTERS

TO THE

NINTENDO GENERATION

JUAN EDUARDO WOLF, EDUARDO E. WOLF
University of Notre Dame Notre Dame IN 46556

his paper is the result of collaborative work between
myself (E.E. Wolf) and my son (J.E. Wolf). It is
written from my perspective since it relates to my
accumulated years of teaching, while my son's contribution
relates to computer software development.
In my twenty years of teaching chemical engineering
courses, I have always been challenged by how best to
involve students in the specific subject being taught. I have
devised many strategies to reach as many students as I could,
and I am especially fond of teaching via the Socratic method.
But the evading and delaying tactics of students who do not
get involved in class have often led me to call on a student
who will most likely know the answer I am looking for.
Unfortunately, this process leads to a dialogue between a
select group of students and myself, to the delight of those
students who prefer to be left alone. As a result, the effec-
tiveness of the in-class learning process is significantly re-
duced, and whatever students learn to pass the exam is done
primarily outside class (usually from a textbook).
This process of selective teaching occurs especially in
large classes where it is obviously impractical to reach ev-
eryone. Many students are thus deprived of the benefits of
Eduardo E. Wolf is Professor of Chemical Engineering at The University
of Notre Dame. He obtained his BS in Civil and Chemical Engineering at
the University of Chile (Santiago), his MS at the University of California,
Davis, and his PhD at the University of California, Berkeley. While his
research interests are in the areas of catalysis and reaction engineering,
he has been teaching transport phenomena to undergraduates for a
decade, and he is interested in the application of computers to interactive
teaching.
Juan Eduardo Wolf obtained his BS degree in Chemical Engineering
and Art at the University of Notre Dame and is currently finishing an MS
degree in chemical engineering at Northwestern University.
Copyright ChE Division ofASEE 1996

the experiential learning process because they do not
participate in the inquiry that the classroom provides.
Our teaching methodology needs to be revised in order to
improve class participation.
Computers may be the media needed to achieve experien-
tial learning. It has always amazed me how fast young people
learn computer games compared to how long it takes me to
save Mario" from all the traps in his unforgiving virtual
world. The younger generation (the Nintendees?) that has
been exposed to these games from early childhood seems to
be able to learn new games even without the aid of a manual.
This learning process is mainly experiential through com-
puter interaction with the player. If we could only get
students to learn at a fraction of the pace with which they
learn these games, we could significantly improve our
teaching capacity.
Several learning studies"1l have shown that involving stu-
dents in the educational process is the key to better learning.
I recall results from a study showing that when the sensorial
perception of information is only auditory (the average lec-
ture), the retention rate is about twenty percent. In a setting
that includes both audio and visual aids, retention increases
to forty percent (transparencies always help!). When the
process also includes an interactive element, however,
wherein students participate (small classes/recitations where
students ask questions), retention rises to eighty percent. I
still remember some problems discussed long ago when I
was a graduate student in our process-design brainstorming
sessions. All of us, I believe, experience events that create
such an impression in our minds that they remain in our
memory for years. Special circumstances cause the brain to
activate the processes required for long-term memory. We

Chemical Engineering Education

do not yet completely understand this
process, so the next best thing is to create
classroom experiences that stimulate
similar responses.
I have been teaching transport phenom-
ena to chemical engineering juniors for
some time. This subject is well suited to
intense interactive teaching via question-
and-answer sessions because it rests
firmly on a few fundamental principles.
The conservation laws provide the axi-
oms from which the basic governing
equations can be derived for most engi-
neering problems. In the past, after ex-
plaining the basic principles, I developed
in-class examples of their applications
by asking the students questions about
the model that describes the problems,
about the assumptions and boundary con-
ditions involved, and finally, about the
method for solving the problem. After
receiving the answers from some selected
student, I would reveal my own answers
on a transparency. Initially, I would de-
velop the equations using shell balances
and, later, by using the simplified vector
forms of the general conservation equa-
tions. To avoid selecting the same stu-
dents every time, I would use randomly
picked numbers to select a student from
the class list. This step-by-step pro-
cess was slow, but it generally received

good reviews

from the students. I felt, however, that many students had
not really participated and I felt that I needed a more
effective teaching method.
Clearly, computers can get people involved in a particular
task. The main use of classroom computers involves home-
work assignments, problem solving, and improved visual
presentations via special animated simulations. Computer
networks also provide an opportunity to reach students out-
side the classroom with assignments and notes.
Notre Dame recently inaugurated a special teaching facil-
ity (DeBartolo Hall) in which each classroom is equipped
with state-of-the-art communications facilities. In particular,
there are two classrooms where there is a computer available
for every student. These computers are connected to a local
area network (LAN) that communicates with a server and
with a podium that is also equipped with a computer. This
setting presented the opportunity I was looking for-the
ability to simultaneously reach and involve every student. A
room in which every student has a computer at his command
means that each student could be asked to answer the same
question and could provide his or her own individual an-
Winter 1996

Several lea
have sh
involving st
educational
the
to better Je
Special cir
cause th
activate th
requii
long-ternz
We do not y4
understand
so the next
to create
experience
stimulate
respond
Clearly, co
get people

swer. The challenge was to create an
interactive teaching method to work
ring studies with these facilities.
own that Translating this idea into reality
udents in the required software. I thought that such
process. software would be available in the
marketplace, but alas, I found no pro-
key gram that seemed suitable, and the
faming .... project was put on hold for a year.
Fortunately, my son was available
cumstancesfor the summer, and with the help of
e brain to a grant from Notre Dame's Office
e processes of University Computing, I asked
him to translate my concept into
red for workable software. He is a chemi-
H memory, cal engineering and an art major,
et completely he had taken my transport phenom-
ena course, and he happens to be
this process, well versed in computers. So, a
best thing is rare father-son academic partner-
classroom ship emerged.

nces that THE PROGRAM
e similar Things developed quickly, al-
though it took more than the sum-
... .mer to complete the working ver-
mputers can sion (an extra month was required).
involved... In the previous semester, with the
help of our secretarial staff, I had
transferred my teaching notes to a
computer disk, which helped expe-
dite the development of the interactive software. After learn-
ing six different computer languages, I have become a de-
voted Macintosh user, so Hypercard was chosen to be the
development environment.
Table 1 (next page) shows the course outline. I first go
over the principles (denoted as LECTURES on the outline)
and then present the application of the principle as a problem
(shown as Lessons). The outline is a hybrid of the texts of
Bird, Stewart, and Lightfoot (BSL)'2] and Welty, Wicks, and
Wilson (WWW).m31 The reference column in the outline re-
fers to either a specific example from WWW or to notes
adapted from BSL (especially when dealing with macro-
scopic balances). The course has three credit hours and it is
the second semester of a year-long course. It covers funda-
mental heat and mass transport; fluid mechanics is covered
in the first semester.
The software was designed as a sequence of "cards" (small
windows of text and graphic information) that ask generic
questions applicable to most lessons. For cases in which this
format does not match (e.g., turbulent flow considerations) a
regular lecture format is used. The first card displayed is the
problem statement; it is displayed in the lower half of a

TABLE 1
Course Outline: Transport Phenomena II
(Heat and Mass Transport)

Session Topic

LECTURE 1 TRANSPORT TUTOR, CONSERVATION'S LAWS
LECTURE 1 CONTINUITY, MOMENTUM BALANCE
LECTURE 1 CONSERVATION OF ENERGY, HEAT CONDUCTION
Lesson 1 Heat conduction, plane wall
Lesson 2 Heat conduction, composite walls
Lesson 3 Heat conduction, cylinder
LECTURE 2 HEAT CONDUCTION IN SOLIDS, VECTOR APPROA(
Lesson 4 Heat conduction with constant source
Lesson 5 Heat conduction with variable source
Lesson 6 Heat transfer from extended surfaces
Lesson 7 Two-dimension heat conduction
Lesson 8 Unsteady heat conduction, semi-infinite wall
Lesson 9 Unsteady heat conduction, lumped systems
LECTURE 3 DIFFERENTIAL ENERGY BALANCE IN FLOW SYST
Lesson 10 Boundary layer analysis-laminar flow
LECTURE 4 TURBULENT FLOW CONSIDERATIONS
LECTURE 5 NATURAL CONVECTION
LECTURE 5 CONVECTIVE HEAT TRANSFER CORRELATIONS
LECTURE 6 MACROSCOPIC ENERGY BALANCE
Lesson 11 Heat transfer equipment design

FIRST MIDTERM EXAM

LECTURE 7 MASS TRANSFER MECHANISMS
LECTURE 8 POINT DIFFERENTIAL MASS BALANCE
Lesson 12 Diffusion in gases: stagnant gas film
Lesson 13 Diffusion is gases: equimolar counter-diffusion
Lesson 14 Diffusion in gases: surface reaction
Lesson 15 Diffusion in liquids: gas absorption without reaction
Lesson 16 Diffusion in liquids: gas absorption with reaction
Lesson 17 Unsteady diffusion in liquids
Lesson 18 Diffusion in solids, porous catalyst pellet
Lesson 19 Diffusion and convection
Lesson 20 Boundary layer analysis, laminar flow
LECTURE 9 TURBULENT FLOW CONSIDERATIONS
LECTURE 10 INTERPHASE MASS TRANSPORT

SECOND MIDTERM EXAM

LECTURE 11 CONVECTIVE MASS TRANSPORT CORRELATIONS
LECTURE 12 MACROSCOPIC MASS BALANCES
LECTURE 13 MASS TRANSFER EQUIPMENT DESIGN
Lesson 20 Design of a batch tank
Lesson 21 Design of a continuous contact tower
LECTURE 14 REVIEW

FINAL EXAM

IN SOLIDS

CH

EMS

NOTES
NOTES
NOTES
17.1
17.1
17.1
NOTES
17.2
17.2
17.3
17.4
18.1
18.1
NOTES
19.4
19.7
20
20
NOTES
22

24
25
26.1
26.1
26.2
26.1
26.2
27
Notes
26.4
28.4
28.6
29

30
NOTES
31
31.2
31.3
NOTES

page-length screen during the entire time
that the student works on the problem.
An example is shown in the bottom half
of Figure 1 corresponding to Lesson
Three on heat conduction in a cylin-
der. The problem statement card
comes equipped with a help button
that, when clicked, opens a window
that provides hints to the students
should they need them.

With the appearance of the Problem
Statement card at the bottom of the
screen, a First Questions card appears at
the top of the screen. It contains a series
of multiple-choice questions that stu-
dents answer by clicking on a box next
to the answer they have selected. The
box is then highlighted. The questions
asked are the ones all transport students
should ask themselves each time they
attempt to solve a problem; the answers
are the simplifying assumptions that ap-
ply to the problem. Once the student has
answered these questions, he or she can
move on to the next card by clicking on
the arrow at the bottom of the page.

So far, the software appears rather one-
sided; in fact, the program was designed
so that students could review their class-
room work outside class in one of the
computer clusters around campus. But
the software really becomes interactive
when it is used in conjunction with an-
other piece of commercially available
software (Screen Link or Timbuktu) that
allows the professor to view a student's
screen on the podium computer. And
the interaction does not stop there since
the professor's screen can be projected
on a large projection screen in front of
the entire class. After first checking with
the student to avoid embarrassing him
or her, the student's work can then be
seen by the entire class. This allows the
professor to go over various points in
the problem and to clarify possible mis-
takes, particularly regarding assump-
tions, etc., as they occur in a student's
thinking process. It also creates an op-
portunity for the class to ask questions
or for clarification on a particular issue.
The professor's version of the pro-
gram contains a class list from which a

Chemical Engineering Education

student is randomly selected each time (see
Figure 2). This undoubtedly creates an in-
centive to become involved during class. In
every session I emphasize that mistakes can
be corrected at this stage before they become
misconceptions that cost the student dearly
in their exams.
After completing the First Questions card,
the students move to the Problem Set-Up
card. At the top of this card (see Figure 3,
next page) is a simplified form of a conser-
vation law: the time rate of change term
[Acc] equals the rate of change due to trans-
fer through open surfaces by convection
[AFc] plus the rate of change due to transfer
through closed surfaces by diffusion (of heat
or mass) [AFd] and the rate of generation of
the quantity being conserved [Rg]. This form
of the equation, although not strictly rigor-
ous, represents the majority of situations en-
countered in transport problems. In the spe-
cial case of momentum, gravitational forces
are considered as a generation term.
The rest of the card consists of a scratch
board that is initially blank. In this space the
student is expected to apply the conservation
equation at the top of the page to the prob-
lem at hand, keeping in mind the assump-
tions made on the previous card. (The stu-
dents can always check by clicking on the
back arrow that takes them back to the previ-
ous card.) The student should develop a solv-
able differential equation. Using a standard
keyboard, typing mathematical notation is
often difficult, if not illegible, so this soft-
ware has a special menu for the most com-
monly used symbols. A symbol menu ap-
pears by clicking the button at the bottom of
the card, and clicking on any of the indi-
vidual menu characters inserts that character
at the last place the student typed. Figure 4
(next page) shows an example of what a
student could type in this space using the
shell balance approach for the example prob-
lem. Again, to move ahead, the student clicks
on the arrow at the bottom of the card.
The next card (see Figure 4), the Bound-
ary Conditions card, asks questions about
the type of equation that the student has de-
veloped and the corresponding boundary con-
ditions that are appropriate to solve the equa-
tion. Students must know the answers to these
questions in order to choose the correct
Winter 1996

)j _first (- questions

Which conserved quantities are we interested in ?
O Momentum [ Energy E] Mass
What geometry describes the situation?
0 General/Cartesian 0 Cylindrical 0 Spherical
Does the situation change with respect to time? o
0 Steady State 0 Unsteady State /
What types of transport are occurring? / i
0 Connection in One Dimension ] Conduction in One Dimension
O Connection in 2D [ Conduction in 2
[ Diffusion in One Dimension ] Diffusion in 20
Does generation/consumption of a quantity happen?
0 Sources exist 0 No sources l

Lesson Three: Heat ProbleStatement:
Conduction in a Cylinder I
Consider a hollow cylinder with length, L, C
inner radius, Ri, and outer radius, Ro. The
outer surface has a temperature of To, while
To the inner surface's is Ti. Assume the cylinder
has a constant density, heat capacity and
thermal conductivity. Plus it is long enough
that we can ignore end effects. Write an
expression for the heat transfer rate, q, via
the temperature profile.

Student Help

Figure 1. First Questions card and the Problem Statement card
(which remains open in the lower half of the screen
along with all other cards).

Professor's Corner

Selected Student
Mac24- Tim B.
/Allison B.

Pick Again

Huailable t of Students: 30

Create/Edit

Open Solution

Huailable Students
MacOl-AS
Mac02-JD
Mac03-DH
etc.

Too Bad, Class Over...

Figure 2. The screen in the professor's computer podium.

method of solving the equation (which they will attempt on
the next card). Questions about the equation are answered by
highlighting the circle next to each correct characteristic of
the equation, and the boundary conditions are typed in on
fence signs. When this card is completed, students can move
on to the last card, where they attempt to solve the equation
they have developed and characterized. They do this by
typing in a chalkboard space with the same menu available
to them as before.
These last two cards usually present the greatest difficulty
to students since most of them have forgotten their calculus
at this point in their college career. (Students somehow have
the impression that once they pass a calculus course, calcu-
lus is over and done with!) In the case of the example
problem, the technique needed is a simple integration. Fig-
ure 5 illustrates the solution in terms of two integration with
the corresponding integration constants, temperature pro-
files, and heat flux. At the bottom of the last card, the
students have the options of saving the lesson to disk, print-
ing it, and/or moving on to the next lesson. Most of the time
the students save the lesson-thus, in a sense, the traditional
notebook has been replaced by a disk.
As professor, several more options are available to me, as
can be seen in Figure 2. In addition to randomly selecting
students for review of their work as mentioned before, I can
easily select from the options at the bottom of my card to add
new problems for the students to work on as well as choose
from a library of pictures to illustrate them. Also, I can pull a
particularly involved solution up on the screen from my
own stack to help explain it. Since the program allows
for the creation of new cards, it can be used not only in
the transport phenomena course but also, by adding new
lessons, in similar courses.

SUMMARY OF A TWO-YEAR EXPERIENCE
The interactive features of the program permitted me to
gain a better insight into the learning process than I could get
in the traditional lecture. One of the benefits I gained from
using this program was learning how little students assimi-
late material from previous courses into their present ones.
My earlier comment about calculus is not a humorous one
but reflects a well-established feeling among students. I also
found that after solving so many differential balances it was
difficult for students to work with macroscopic balances, in
particular with the plug flow model, when using
macrodifferential balances.
Clearly, there are many ways to improve the software. For
example, the program could be used with standardized math-
ematical software such as Mathematica to graph the solu-
tions. Obviously, animations showing the physics involved
in a given problem can be developed using many of the
programs available from the computer centers supported by
federal programs. As software such as described here be-

l IProblem Set-Up: Acc = AFc AFd + Rg 0

0 = AFd

0 = q*2nrL r q*2nrL Ir +Ar|

Dividing by 2nLAr and taking the limit as Ar approaches 0,
0 = 5(r-q)

Substituting in Fourier's Law for a cylinder, q = -k (T
0 = -2(r (-k:,T))

Continue when you've developed your p
mbo general balance to the point of integration. N

Figure 3. Problem Set Up card, initially empty, showing
how to set up the problem using shell balances.

Figure 4. Boundary Conditions card and classification of
the type of equation to be solved.

Solution
First Integration: fO dr = -fd(r-q)
C1= r-q or q= Cl/r

Second Integration: -k fdT = kT = f (-Cl/r) dr + f0 dr
-kT = -Cl In (r) + C2

From BC1 & BC2,
C1 = -k(Ti-To)/ln (Ro/Ri)
C2 = k-Ti Cl-In (Ri)

Substitute in for temperature profile,
T(r) = Ti [ (Ti-To)/ln (Ro/Ri)]- In (r/Ri)

Heat flux, q =-kT|Ri= -CI/r]Rij= k- (Ti -To)/Ri-ln (Ro/Ri)

Previous Symbols Save This Lesson Prin My Sluff Bye byse

Figure 5. Solution card, initially blank, showing the
completed solution for the selected problem.
Chemical Engineering Education

comes commercially available, more interactive features can
be added (for example, a method for self-grading). Another
alternative to the interactive mode is to include a short test as
part of the software, with some sort of point system to
evaluate the answers. This, however, would eliminate the
direct interaction and make the presence of the professor less
relevant. I still believe that direct interaction is the best
experiential learning.
The first time I tried the program the class had fifty-five
students, and every class was conducted in the computer lab.
I was not able to cover all the material listed in Table 1 (in
particular, the applications of macroscopic balances). Some
of these examples are revisited in our Design I course, so I
covered only the fundamentals. The second time around, the
new class of seventy-two students could not be accommo-
dated in a single session in the computer lab, so I divided the
class into two sections. Each section attended the computer
lab once a week for 75 minutes in addition to a 75-minute
lecture for the entire class in a classroom where there was a
podium computer to display my lectures. The interaction
here was the traditional question-and-answer format. Sev-
eral of the lessons which had been worked out previously
during the tutorial were assigned as homework problems.
This required that I volunteer 75 minutes of my time to
teach the course.
Even though it takes more time to cover the material when
using interactive software, it was an interesting experience.
The interactive features motivated the class and provided an
incentive for the students to get involved. Above all, it
made teaching fun for me, and the majority of the class
enjoyed it as well. In the computer lab I found I had to
alert the students to the fact that class was over, instead
of listening to the impatient rattle I would usually hear as
the period drew to a close.
The first time the course was presented, most teaching
evaluations were very positive, but there was a small group
(about 4%) that did not like the software and strongly voiced
their preference for the regular lecture. I suspect that for
those unfamiliar with computers, the new format presented
an extra burden of gaining computer literacy and this pro-
duced the negative reaction. The second time around when I
combined the tutorial with the regular lectures in a 50-50
mix, there were no negative responses as to the use of
computers in the classroom. This time I also posted the
lectures ahead of time so the students would have copies
available during class. I am still struggling with the question
of whether or not to give students access to the solutions or
let the class work them out. While computers can be won-
derful tools when used conscientiously, they can also be
expeditious copying machines, which defeats the purpose of
experiential learning.
The method is not limited to chemical engineering but can
also be applied to many different disciplines. During a week-
Winter 1996

end when the parents visited the university, I hosted an open
house for them where I set up a riddle for them to solve.
Everybody seems to have enjoyed the experience-even
those who did not get the right answer. I also used videos
to illustrate physical phenomena, such as boundary layer
flow. This gave the students a visual experience that
equations do not impart.
The method is still limited to situations where there are
classrooms with networked computers. Undoubtedly, there
will be more of these in the future, and software such as
described in this paper will become commonplace. I envi-
sion a future when computers will be an active part of our
teaching technology. We should continue to introduce the
latest multimedia technology into the classroom to improve
what I believe are less-than-effective teaching methods.

REFERENCES
1. Wankat, P.C., and F.S. Oreovicz, Teaching Engineering,
McGraw-Hill, New York, NY, (1993)
2. Bird, R.B., W. E. Stewart, and E.N. Lightfoot, Transport
Phenomena, John Wiley & Sons, New York, NY (1960)
3. Welty, J.R., C.E. Wicks, and R.E. Wilson, Fundamentals of
Momentum, Heat, and Mass Transfer, 3rd ed., John Wiley
& Sons, New York, NY (1984) n

[LM1 new books )

The Physiology and Biochemistry of Prokaryotes, by White; Oxford Uni-
versity Press, 200 Madison Avenue, New York, NY 10016; 378 pages, $45
(1995)
Encyclopedia of Chemical Technology: Imaging Technology to Lanthanides,
Kirk-Othmer; Wiley, 605 Third Avenue, New York, NY 10158; 1115
pages, $295 (1995)
Principles of Ceramics Processing, 2nd ed., by Reed; Wiley, 605 Third
Avenue, New York, NY 10158; 658 pages, $69.95 (1995)
Introduction to Chemistry, 7th ed., by Dickson; Wiley, 605 Third Avenue,
New York, NY 10158; $20.95 paper (1995)
Patty's Industrial Hybiene and Toxicology: Biological Responses, 3rd ed.,
edited by Cralley, Cralley, and Bus; Wiley, 605 Third Avenue, New York,
NY 10158; $195 (1995)
Laser Techniques in Chemistry, by Myers and Rizzo; Wiley, 605 Third
Avenue, New York, NY 10158; 429 pages, $125 (1995)
Encyclopedia of Chemical Technology: Lasers to Mass Spectrometry, 4th
ed., Kirk-Othmer; Wiley, 605 Third Avenue, New York, NY 10158; 1094
pages, $295 (1995)
Advances in Photochemistry, Vol. 20, by Neckers, Volman, and Bunau;
Wiley, 605 Third Avenue, New York, NY 10158; 301 pages, $95 (1995)
Hydrocarbon Chemistry, by Olah and Molnar; Wiley, 605 Third Avenue,
New York, NY 10158; 632 pages, $69.95 (1995)
Ketenes, by Tidwell; Wiley, 605 Third Avenue, New York, NY 10158; 665
pages, $69.95 (1995)
Conformational Theory of Large Molecules: The Rotational Isomeric State
Model in Macromolecular Systems, by Mattice and Suter; Wiley, 605 Third
Avenue, New York, NY 10158; 448 pages, $54.95 (1994)
Interfacial Transport Processes and Rheology, by Edwards, Brenner and
Wasan; Butterworth's, (1991)

r l learning in industry

This column provides examples of cases in which students 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.

--EPIC-

The Engineering Program for International Careers

S. S. MELSHEIMER, C.E.G. PRZIREMBEL
Clemson University Clemson, SC 29634

Competition in the world marketplace demands that
the US educational system produce engineers who
possess not only first-rate technical skills, but who
are also capable of functioning effectively in a global engi-
neering/industrial environment. American engineering edu-
cation is world-class in the technical arena, but it has slighted
the preparation of its graduates to compete in a global engi-
neering arena. For many years the economic dominance of
the US in the world economy permitted us this luxury.
Development of the European and Pacific Rim economies,
however, has brought us into a new era. Now, for example,
three of the five largest chemical companies in the world are
German, and only the fifth largest is US-owned. Perhaps
more important, manufacturing and engineering companies
have become multinational in their operations. The design of
new vehicles or chemical plants involves teams spread across
multiple continents. Moreover, US engineering graduates
often compete with engineers educated abroad for positions
in this global engineering/industrial arena. European and
Asian engineers commonly speak multiple languages and
have broad international experience, while neither of these
attributes is typical of US engineering graduates.
The growing need for US engineering graduates who are
prepared to work in this international environment has not
gone unnoticed. For example, a workshop addressing inter-
national engineering education issues was held by the Fund
for Improvement of Post Secondary Education in 1990.[1] In
1992, the National Science Foundation sponsored work-
shops to address US-German interaction in engineering re-
search as well as in education.121 Representatives of German
@ Copyright ChE Division ofASEE 1996

industry emphasized the need for US engineering graduates
prepared to function professionally in Germany (or other
European countries). The primary issue is not placing US
engineers in permanent positions in Europe, but short-term
assignments and interactions between US engineers and co-
workers (or customers) from other cultures. Specific re-
quirements cited included foreign language proficiency and
experience in engineering work or study in a foreign culture.
U.S. schools have responded to this need in various ways.
For example, Michigan State has long offered a summer
course taught by MSU faculty at the Rheinisch-Westfalische
Technische Hochschule in Aachen, Germany, and they now
offer academic year exchange with RWTH. Rhode Island
offers a unique dual-degree program (BS Engr/BA German),

Christian E.G. Przirembel received his BS,
MS, and PhD degrees in mechanical and aero-
space engineering from Rutgers University. He
has served as a faculty member and associate
dean of academic affairs at Rutgers, and from
1981-1994 as Professor and Head of Mechani-
cal Engineering at Clemson. Currently he is
Associate Dean of Engineering and Science.
His research areas include subsonic and su-
personic flows and separated flows.

Chemical Engineering Education

Stephen S. Melsheimer received his BS in
chemical engineering from L.S.U. (1965) and
his PhD in chemical engineering from Tulane
University (1969). He is currently Professor of
Chemical Engineering and Acting Associate
Dean of Engineering and Science at Clemson
University, where he has been since 1969. His
research interests are primarily in the area of
automatic control.

and Cincinnati has an International Engineering Co-Op Pro-
gram with an internship abroad. Various engineering col-
leges (e.g., Rose-Hulman, Wisconsin, and Rhode Island)
offer exchange programs with specific foreign schools, and
study abroad through the International Student Exchange
Program is widely available. Still, engineering student par-
ticipation in study abroad remains small, influenced by cost,
course transfer, and foreign language proficiency issues.131
The American European Engineering Exchange Consortium
(AE3) of over thirty-five American and European institu-
tions has recently been established to promote international
education in engineering. AE3 will address transfer and other
issues and offer programs that include both study abroad and
internship opportunities.

PROGRAM CONCEPT

The need for this international dimension is certainly evi-
dent from Clemson's perspective. Consider that 160 foreign-
owned companies have US headquarters in South Carolina.
In 1993, there were 457 international facilities in the state,
70% of which are located within fifty miles of Clemson
University. Adding to this international flavor are numerous
US-owned global manufacturing and engineering concerns
in the area. In responding to the obvious need to enhance the
international component of its undergraduate programs the
College of Engineering had to address the following key
problems:
Although there is some recent improvement,'41few US second-
ary school graduates have more than a passing knowledge of
a foreign language; this is especially true with respect to some
languages important from an industrial viewpoint (e.g., Ger-
man, Japanese).
Many engineering students come from middle- to low-income
families. Thus, significant additional costs to the student would
limit the number of students electing the program.
ABET and institutional general educational requirements con-
strain curriculum content.
Many engineering students are unaware of the implications of
economic globalization.
At Clemson (and many other institutions), engineering stu-
dents select a specific major at the end of the first year. Thus,
program entry after the first year should be accommodated.

Considering these factors, the EPIC program was developed
based on the following principles:
The program cornerstone is a company-sponsored Interna-
tional Internship of significant duration in the foreign envi-
ronment. This approach minimizes costs to the student and
provides a particularly valuable form of international experi-
ence. As an option, the program offers a study-abroad period
to complement and extend the internship experience.
At least three years college-level study of the foreign lan-
guage, including a period of "immersion" language and cul-
tural preparation just prior to the overseas internship.
Winter 1996

Prior to the international internship, the program requires at
least one domestic internship to provide experience in the
engineering work environment.
The program provides continued exposure to the foreign lan-
guage following the overseas experience and includes courses
addressing global culture and economics.
Industrial guidance in development, evaluation, and opera-
tion of the program is provided through an Industrial Advi-
sory Board.
A Certificate provides a credential for students completing the
program.

PROGRAM DEVELOPMENT
Recruitment of Industrial Partners The most critical
early activity in the project was recruiting companies to
participate in the program and to provide the international
internships that are at the core of EPIC. Indeed, despite the
discussion at the NSF workshop and with other corporation
representatives, the question of the real level of interest by
the corporate community remained open-would there actu-
ally be sufficient corporate support to make the program
viable? Corporate response to EPIC definitely answered this
question in the affirmative. To date, fourteen companies
have joined the program and are represented on the EPIC
Advisory Board. Moreover, company representatives on the
EPIC Board are unanimous in the view that there is a real
and growing need for engineers prepared to function in an
international environment.
Program Structure/Curriculum The program structure
outlined above was developed with the close collaboration
and approval of the EPIC Board. Additional features are:
Entry into EPIC requires completion of all freshman require-
ments with a grade point average of at least 3.2, and registra-
tion in first-year (or higher) language courses.
EPIC internships are arranged by US operating companies
(either US or foreign owned).
The selection process includes interviews of student applicants
by EPIC companies.
The overseas internship is scheduled after students have
completed most of their junior-level engineering courses.
Language options include French, German, and Japanese.
This is based on the overseas internship opportunities with the
current sponsoring companies.
A "typical" EPIC program schedule incorporating these
features is shown in Table 1 (next page). In terms of academ-
ics, the primary change is incorporation of twenty foreign
language credits. Many of these credits can qualify as elec-
tives within the base engineering curriculum. Clemson cur-
ricula, for example, include several "free" elective credits
plus 16-18 credits of humanities and social sciences, allow-
ing up to 14 language credits to qualify as electives within
the curricula. However, EPIC does require about six "ex-
cess" credits, equivalent to the credits in the intensive lan-

guage course. As in the case of co-op payments, the inclu-
sion of two internship terms results in a total program dura-
tion of five years.
The EPIC entrance requirement is set reasonably high
(GPA > 3.2) to assure that students have a good probability
of meeting the academic demands of the program. While
academic credentials are significant in company assessments,
personal attributes such as self-reliance, resourcefulness, flex-
ibility (i.e., ability to cope with new situations), and willing-
ness to take risks are also very important in evaluating candi-
dates. In addition, factors such as vision/ambition, commu-
nications ability, leadership, and interpersonal skills are im-
portant as in any professional hiring situation. Judgments on
these factors are subjective, and thus the interview process is
highly important to the participating companies, with many
of them conducting follow-up interviews at plant sites after
an initial on-campus interview.
The program structure will be reviewed and revised as
experience is gained. For example, alternative schedules
with two or more domestic internships have been developed
to suit the desires of some EPIC companies.
Student Recruitment With the program structure estab-
lished, student recruitment began in March of 1993. Clemson
freshmen enrolled in engineering were given information on
EPIC, and follow-up contact was made via a survey de-
signed for use in program evaluation. The survey indicated
that less than 20% of freshmen engineering students at
Clemson have any interest in a career involving international
assignments, and even fewer expressed specific interest in
EPIC. Virtually identical results were obtained in a survey
conducted in the spring of 1994. These results clearly indi-
cate the need for greater emphasis on language, international
culture, and globalization in our curricula. Programs like
EPIC address this issue both directly (by increasing the
number of students having international experiences) and
indirectly (through the presence of more engineering stu-
dents on campus who have had international experiences).
Ten students with the requisite qualifications applied for
the first round of interviews held in November of 1993. The
small number reflects the survey results noted above, plus
the newness of the EPIC program. In addition, only four
EPIC companies were available to interview in the fall of
1993. Due to the small number of both students and compa-
nies, some companies had no applicants to consider in their
desired language/major combinations. Similarly, some stu-
dents found no companies interested in their language/ma-
jor. As a result, only two students were placed. These stu-
dents had domestic internships in 1994-95 and will have
overseas internship in fall 1996.
Augmenting this, BMW (whose U.S. plant was under
construction) selected two senior-level students for intern-
ships in Munich beginning in the summer of 1994. These
two BMW internships were invaluable in that they provided

immediate experience with sending students abroad. As it
happens, one of the students had co-op experience (but mini-
mal German), while the second was fluent in German (but
had no prior engineering work record). Both had fine experi-
ences (despite a few rough edges due to their "pioneer"
status), and BMW was pleased with the overall performance
of both students. However, BMW concluded that future
interns needed to have both prior work experience and good
German competency, thus confirming the need for both of
these components in the EPIC program design.
The second round of EPIC applications was held in the fall
of 1994, with twenty-one qualified students (including one
from the University of North Carolina at Charlotte, another
SUCCEED school) applying to join EPIC and nine compa-
nies participating in the interviews. This represented a sig-
nificant increase in both applicants and participating compa-
nies. Still, the unbalanced distribution of language/major
combinations again hampered placements. In March of 1995
an additional round of interviews was held in which fresh-
men were permitted to interview for the first time. The
purpose of this was to enable students considering co-op
opportunities to simultaneously consider EPIC (previously,
students were required to have completed the entire fresh-
man year). Five companies and ten students participated. In
addition, BMW selected a third senior student (a Chem E
with co-op experience and excellent German) to send di-
rectly to Germany. The student reports that he is having a
fine work experience and is enjoying Munich greatly, while
BMW is extremely pleased with the performance of the
student and his preparation for the assignment. EPIC student
participation now totals fifteen, with Fall 1995 selections
under way.
Operational Factors The financial terms for the intern-

TABLE 1
Typical EPIC Program Schedule
(Italics denote EPIC-specific components)

First Year Standard freshman sequence ~32 cr
Second Year Interviews with EPIC companies
Two semesters foreign language (first year) 8 cr
Balance of "normal" sophomore courses ~26 cr
Third Year
Fall Industrial internship in US
Spring Third semester of language 3 cr
Normal first-semester junior classes -15 cr
Summer Normal junior classes 6 cr
Intensive language institute 6 cr
Fourth Year
Fall International internship (4-6 months)
Spring International social science elective 3 cr
Normal second-semester junior classes ~12 cr
(optionally Spring semester abroad at foreign institution
Fifth Year Upper division language course 3 cr
International social science elective 3 cr
Balance of normal senior year sequence ~27 cr

Chemical Engineering Education

ships are of considerable practical importance to the stu-
dents. The agreement reached by the EPIC companies is that
they will compensate EPIC students at the same rate as their
normal practice for co-op students and summer interns. How-
ever, for international internships consideration will be given
to cost of living, travel, taxes, etc. In this regard, the EPIC
companies agreed to provide round-trip transportation to the
host country and to address the key cost of living issue by
assisting the students with housing arrangements. As a gen-
eral rule, the goal is for the participants to "break even"
during their international assignment.
Issues such as visas, work permits, residence permits, and
tax regulations for international workers need to be addressed
before arrival in the country. For instance, in Germany an
internship's duration has a major effect on the income tax
liability, with substantially higher tax liability for intern-
ships exceeding six months.
Evaluation Plan As with any project, it is important that
the EPIC program be assessed to determine its effectiveness,
to follow up on the question of whether it is meeting a real
need, and to provide information for use in improving the
program. Data gathered from successive entering classes,
and follow-up data from graduates, will enable assessment
of the impact of EPIC in altering the international perspec-
tive of engineering students. Data from employers, EPIC
students, and EPIC graduates will be used to evaluate and
improve the EPIC program. Obviously, gathering these data
will be a multi-year undertaking.
Study-Abroad Linkages Discussions regarding enroll-
ment of EPIC students in engineering courses in the semes-
ter following their internship have been held with German
and French schools. All indicated an interest in having US
students enroll and expressed a willingness to assist the
students with housing, exam arrangements, etc. Details of
course equivalencies will have to be worked out for each
major and each institution.
Development of the Language Institute The initial of-
fering of the institute was in July-August 1995. The institute
is tailored to students with technical backgrounds who have
completed basic study of the language. The institute is spe-
cially targeted at meeting the needs of engineering students
about to embark on an overseas internship, bringing them
from basic language understanding to a reasonable level of
conversation, including an introduction to appropriate tech-
nical vocabulary.

IMPLEMENTING EPIC AT OTHER INSTITUTIONS
We certainly encourage others to borrow any or all of the
EPIC scheme. The key steps to doing so are fairly simple:
* Identify a core of enthusiastic supporters from international
companies that have strong ties to your school, and invite them
to form the charter membership of an advisory board for your
international engineering program.
Winter 1996

* Lay out a "curriculum "for the international program that
shows students (and companies) how the necessary language
courses and internships fit into a program of study.
Identify a mechanism to provide an "immersion" language
program prior to the overseas experience.

The first item is clearly the most important if the EPIC
concept of providing international experience through in-
ternships is to be used. Enthusiastic supporters will help
"make things happen" within their own companies and can
help in recruiting additional industrial partners. The curricu-
lum plan is important as a way of conveying to the students
an orderly way through the program. Finally, it is our view
that an immersion program is an essential part of the lan-
guage preparation. Of course, it is not practical for every
university to offer such an immersion program. However,
existing programs in the US (such as the Clemson program) or
abroad (e.g., the Goethe Institute in Germany) can be used.

CONCLUSIONS
Corporate interest in EPIC confirms the need for an inter-
national dimension in US engineering programs. It has been
well received by students since it addresses a key issue (cost)
that prevents some students from pursuing other interna-
tional education programs. Our experience to date clearly
shows it to be important to have a critical mass of companies
and students so as to minimize supply/demand imbalances in
the various major/language combinations. It is pertinent to
note that the demand for chemical engineers has exceeded
the supply in both German and French, while the ChE sup-
ply/demand picture in Japan is "perfectly balanced" (no
applicants and no positions!).
Overall, our experience validates both the premises upon
which EPIC is based and its practicality in terms of meshing
with existing engineering curricula. Coupled with the strong
company and student interest, this indicates that EPIC is a
viable model for similar programs at other institutions.

ACKNOWLEDGMENTS
Funding for this work was provided by the National Sci-
ence Foundation through SUCCEED (Cooperative Agree-
ment No. EID-9109853). SUCCEED is a coalition of eight
schools and colleges working to enhance engineering educa-
tion for the twenty-first century.

REFERENCES
1. "Internationalizing Engineering Education: A National
Workshop," FIPSE, Washington, DC (1990)
2. Workshops on "US-German Cooperative Programs in Chemi-
cal and Mechanical Engineering," NSF, New Orleans, LA
(1992)
3. "Open Doors 1994-95," Institute for International Educa-
tion, New York, NY (1995)
4. Draper, J.P., "Foreign Language Enrollments in Public Sec-
ondary Schools, Fall 1989 and Fall 1990," Am. Council on
the Teaching of Foreign Languages, Yonkers, NY (1991) 1
49

110 laboratory

LOW-COST EXPERIMENTS IN MASS

TRANSFER

Part 1*

I. NIRDOSH, M.H.I. BAIRD**
Lakehead University Thunder Bay, Ontario, Canada P7B 5E1

Laboratory work is an important component in the

chemical engineering curricula. In a recent CEE ar-
ticle, Stubington"' appropriately defined the objec-
tives of a teaching laboratory as
... to develop skills in the acquisition and analysis of
engineering data; to develop the ability to communicate
experimental findings in written and oral forms; and to
reinforce in a practical way theoretical concepts taught
in lectures.
Laboratory instructors are all too familiar with the impasse
created by lack of funds for equipment, combined with in-
creased student enrollment. The challenge is to devise sig-
nificant experiments using existing laboratory supplies such
as pipes, fittings, or glassware. A simple and inexpensive
method for determining liquid-side mass transfer coeffi-
cients in an undergraduate chemical engineering labora-
tory is described here.

THEORY
Gas absorption can be a complex process with resistances
in either phase, with the additional problem of estimating the
interfacial area in flow across packing, etc. In this experi-
ment, a model system with a simple geometry has been
chosen so that the mass transfer coefficient can be measured
and compared with a theoretical prediction. Students should
be cautioned that the equations given here are not sufficient
to design a piece of practical equipment. They do, however,
provide a partial theoretical basis for design.
Pure carbon dioxide is absorbed from a chamber into a jet
of water. Because the gas is pure, the entire mass transfer
resistance lies in the liquid phase. The rate of mass transfer
(R) from the gas to the liquid phase is related to the driving
force (AC), the area for mass transfer (A), and the mass

*Part 2 of this paper will appear in the spring 1996 issue of CEE.
"Address: McMaster University, Hamilton, Ontario, Canada
L8S 4L7
Copyright ChE Division of ASEE 1996

Inder Nirdosh received his BSc and MSc in
chemical engineering from Panjab University (In-
dia) and his PhD from Birmingham University
(United Kingdom). He has industrial and teach-
ing experiences, and his research interests are in
the fields of mineral processing and electrochemi-
cal engineering.

Malcolm Baird received his PhD in chemical
engineering from Cambridge University in 1960.
After some industrial experience and a post-doc-
toral fellowship at the University of Edinburgh, he
o joined the McMaster University faculty in 1967.
His research interests are liquid-liquid extraction,
oscillatory fluid flows, and hydrodynamic model-
ing of metallurgical processes.

transfer coefficient (ke) by the equation
R = kAAC (1)
Generally speaking, each of the three terms on the right
can vary independently and, in a given case, the estimation
of (R) has to be done by using a combination of theoretical
models and empirical correlations. This experiment allows
the driving force ( AC) and the area (A) to be known exactly,
thus permitting one to find the mass transfer coefficient (k,)
by measuring the rate of mass transfer (R). The value of k,
so determined can then be compared with the results pre-
dicted by the popular Higbie's Penetration Theory for un-
steady diffusion.121
This theory indicates that the value of ke depends on the
molecular diffusion coefficient (D) and the contact time (c)
as described below:

k=2 D (2)

Here, r is the time for which the liquid surface is in contact
with gas and can be set by forming a jet of liquid passing
continuously through the gas at a uniform velocity u. Since
T can be calculated by dividing the jet length (L) with the
average velocity of the liquid (u), the above equation can be
written as

k = 2 DU (3)
CeiL
Chemical Engineering Education

The above equation can also be written in dimensionless
form, using a length-based Sherwood number (Sh = k,L / D)
and Reynolds number (Re = uLp / g):

Sh = 2 Rel/2 Sc1/2 (4)
I1t
The liquid properties (p,p) cancel out on the right-hand
side of Eq. (4). That is consistent with the assumption of an
ideal jet moving at a plug flow velocity, u. In practice, there
is a very small hydrodynamic boundary layer near the nozzle.
The shear stress at the gas-liquid interface is very small
because of the low viscosity of the gas phase. Therefore, the
simplifying assumption of liquid plug flow can be justified
as a reasonable approximation.
Combining Eqs. (1) and (3), we get

R = 2 L AAC (5)

It may be noted that the contact time between a short jet of
water and the surrounding gas (CO2) is so small that only an
insignificant quantity of CO2 is picked up by the water jet
and the bulk concentration of CO, in water at any time (C.)
remains negligible. Thus, the driving force (AC = C, Cw) is
essentially given by c*, the solubility of CO, in water at the
operating temperature, which can be found in the literature.131
The surface area ( idL) of the water jet exposed to the gas
is a function of d and L, the diameter and length of the jet,
respectively. It is assumed that the liquid velocity (u) is great
enough that d remains constant over the jet length. Substitut-
ing these into Eq. (5) and simplifying gives

R = 2 Cwd DiruL (6)
The velocity term in Eq. (6) can be replaced by the average
volumetric flow rate (Q = und2 / 4), giving

R = *4C DQL (7)
Cullen and Davidson"4' showed that this form of mass
transfer expression holds even for cases where gravitational
acceleration has a significant effect on jet velocity u.

APPARATUS
A schematic of the apparatus is shown in Figure 1. The gas
chamber may be a glass tube 5-to-6 cm diameter and 25-to-
30 cm long. Water is fed from an overhead reservoir to a jet
nozzle located in the chamber. The water jet is in contact
with the gas for a short length (L) that can be adjusted by
raising or lowering the nozzle. The jet leaves the chamber at
the draw-off tube, which is filled with kerosene to provide a
non-absorbing liquid seal. A small layer of kerosene is also
present at the base of the chamber to cover any water drops
that may accidently spill from the draw-off tube. On the gas
side, a supply of pure CO2 can be connected to the chamber,
and a soap film gas meter is also connected so that small
changes in volume of the gas in the chamber can be mea-
sured.
Winter 1996

PROCEDURE
1. Set up the apparatus with the addition of kerosene layers as
shown in Figure 1. The purpose of various kerosene layers
is to allow CO, absorption only in the water in the jet and
not in any other location, such as in the draw-off tube or in
spillage.
2. Purge the gas chamber thoroughly with CO2.
3. With the gas flow on, squeeze the soap solution bulb and
run a few films up the graduated gas burette to lubricate it.
4. Test the apparatus for leaks.
5. Start the water flow to obtain a uniform jet falling directly
into the 1-cm draw-off tube. (Note that at very small flow
rates, the jet would tend to bead, and this should be
avoided.)
6. Bring the kerosene level to the tip of the draw-off tube by
lowering or raising the outlet tube.
7. Measure, and maintain, the jet length as the distance
between the nozzle tip and the draw-off tube.
8. Reduce the CO, flow to a very small value and run a few
soap films up the graduated burette. (At large gas-flow
rates the soap films will either break or exit the burette too
quickly.)
9. When a few films reach the top of the burette, shut off the
gas supply completely. (The soap films will start descend-
ing at a rate depending on the rate of adsorption of CO,.)
10. Measure the rate of CO, absorption by monitoring the rate
of descent of the soap films.
11. Measure the water flow rate (Q) from the outlet tube with
the help of graduated cylinder and stop watch. (Because the

CO,2-
From
Cylinder

Graduated
- Gas
Burette
Gas
Film

Soap
Solution
Bulb

Figure 1. The apparatus

total measuring time for the film-descent and water flow rate
measurements is not more than a few minutes, the water level in the
overhead reservoir, and therefore the hydrostatic head and hence
(Q), can be assumed to remain unchanged during the measurement
period.)
12. Repeat steps 8 through 11 at four or five more water flow rates for
the same jet length.
13. Measure the diameter (d) of the nozzle (and hence the jet) either
with a traveling microscope or by the following experiment that
makes use of the law of conservation of energy:
A water jet is made to rise at an angle 6 (see Figure 2). For a
given volume flow rate, the average kinetic energy of a fluid
element of mass, m, in the jet as it leaves the nozzle with an
average velocity u is Y2 mu2. As it rises against gravity, it decel-
erates and eventually its velocity becomes zero at a certain height,
h, after which it starts accelerating downward. At the instant its
velocity is zero, its kinetic energy based on the vertical velocity
component, u sin 6, is changed entirely to its potential energy
and the two may be equated as

y/2 mu2 sin2 0 = mgh (8)

which on replacing u in terms of the volumetric flow rate
Q (= und2 / 4) and simplifying gives

d2 Qsin (9)
it gh
from which d can be obtained.

TYPICAL RESULTS AND DISCUSSION
Figure 3 shows typical measured mass transfer rates R, plotted
against Q for several different jet lengths. It can be seen that the
data are well represented by straight lines passing through the origin,
as predicted by Eq. (7). The slopes of the lines can be determined
by linear regression and are found to be proportional to vL,
again in agreement with Eq. (7). The solubility c* and molecu-
lar diffusivity D are constant, provided the experiments are all
done at the same temperature.
In Figure 4, the dependence of R upon J!L is shown directly for
some experiments carried out at constant water flow rate with differ-

Figure 2. Determination of jet diameter.

ent jet lengths. Again, regression analysis should be
carried out to determine the slope.
A complete comparison of data with the theory can
be made if values of Cw and D from the literature
sources are given, and this is illustrated in Table 1.
Students should be asked to discuss possible rea-

20
JET LENGTH
= 2.6 cm
A= 3.4 cm
= 4.2 cm
V 7.0 cm

0 -
-

: /o
/
_/ / \

5 / ,/ /

0 1.0 2.0 3.0
m. ( .)05

Figure 3. Dependence of gas absorption rate on
water flow rate (Q) through a jet of 0.0984 cm
diameter at several different lengths.

0 I [ I I I I I I I I I I I J
0 1.0 2.0 3.C
-FJT, (cm)"

Figure 4. Dependence of gas absorption rate on
length of a jet 0.0984 cm diameter, for water flow
rate Q = 1.96 mL/s.
Chemical Engineering Education

sons why the measured k, values are slightly different from
the calculated values. For example, the penetration theory
assumes that the liquid is moving in plug flow, but the jet
leaving the nozzle may have a curved velocity profile, or it
may be turbulent.

CONCLUSIONS
The apparatus is easy to build and needs hardware that is
usually available in any chemical engineering department.
The experiment introduces the students to an experimental
test of Higbie's Penetration Theory, liquid-phase mass-trans-
fer resistance, and conservation of energy. The entire class
can be subdivided into various groups and each group can be
asked to study the effect of liquid flow rate on CO, absorp-
tion for one fixed jet length (or jet diameter), the other
groups studying different jet lengths (or jet diameters).
This will make it an all-group lab because each group
will need data from other groups to compile a compre-
hensive lab report. Specifically, each group can be asked
to do the following:
1. Plot R vs. -FQ and determine if Eq. (7) holds. Such a plot
should be linear for a given jet length. Data for all jet lengths
should be plotted on the same graph for comparison (see
Figure 3 for typical experimental data).
2. Obtain data for R vs. L for a given flow rate from (1) above,
and plot R vs. -iL (see Figure 4 for typical experimental
data).
3. Calculate k, expt from Eq. (1) and compare kc pred obtained
from Eq. (3) (see Table 1).
In addition, the students can be asked to comment on the
effect of jet acceleration due to gravity resulting in tapering
the jet and thus decreasing its diameter (and hence the sur-

TABLE 1
Typical Experimental Results

Jet Diameter =0.1914 cm C* = 1.649x 103g/cm3 p = 0.99823 g/cm3
Temperature = 20C D = 1.77 x 10` cm2/s p = 0.01005 g/ms

Q u T
(cm'/s) (cm/s) (s)

Re x 10 kept
(g/s) Eq.l(cm/s)

k pred
Eq.3(cm/s)

2.6 1.463 50.852 0.051 6.005 0.02328 0.02098
2.6 1.631 56.691 0.046 6.005 0.02328 0.02098
2.6 2.12 73.688 0.0353 6.629 0.02569 0.02387

3.244 112.76
1.434 49.84
1.875 65.17

0.023 8.227 0.03189
0.084 8.502 0.02039
0.065 8.575 0.02056

0.0295
0.01517
0.01762

4.2 2.172 75.49 0.056 9.567 0.02294 0.0268

4.348 150.85 0.028 13.5
5.125 178.14 0.024 15.06

0.03237 0.02845
0.03612 0.02358

face area that has been taken as ndL). The instructor may
encourage the students to review a paper by Cullen and
Davidson141 for a discussion of the basic assumptions and to
ascertain that the effect of the taper is exactly offset by the
effect of acceleration, so that Eq. (7) is rigorously valid.
Unpredictable results may sometimes be observed due to
one or a combination of the following reasons:
1. Gas burette is not clean and gas films do not descend at a
uniform rate.
2. Soap films are too frothy and soap-solution consistency
needs adjustment.
3. Jet length is too small (less than 2.5 cm).
4. Water flow rate is too small, e.g., it corresponds to incipient
drop formation.
5. The apparatus leaks.

NOMENCLATURE
A area
C carbon dioxide concentration in water at any time
C* carbon dioxide solubility in water at operating temperature
d jet diameter
D diffusion coefficient of CO2 in water at operating tempera-

ture
g acceleration due to gravity
k, water-side mass transfer coeft
L jet length
m water mass flow rate
R gas absorption rate
Ref Reynolds number, uLp / g
Q water volume flow rate through
Sc Schmidt number, g / pD
Sh, Sherwood number, kfL/D
u water velocity through the jet
Greek Symbols
0 angle
T contact time
A change
g liquid viscosity
p liquid density

icient

gh the jet

ACKNOWLEDGMENTS
Financial support for this work was provided by the Natu-
ral Sciences and Engineering Research Council of Canada.
Thanks are due to Mr. T. Bainbridge and Mr. S. Connell for
collecting some of the experimental data.

REFERENCES
1. Stubington, J.F., "Quality in Teaching Laboratory," Chem.
Eng. Ed., 29(3), 186 (1995)
2. Higbie, R., "The Rate of Absorption of a Pure Gas into a Still
Liquid During Short Periods of Exposure," Trans AIChE,
31, 365 (1935)
3. Perry, R.H., and C.H. Chilton, eds., Chemical Engineer's
Handbook, 5, 3-96 (1973)
4. Cullen, E.J., and J.F. Davidson, "Absorption of Gases in
Liquid Jets," Trans. Faraday Soc., 53, 113 (1957) O

Winter 1996

B curriculum

ON SELECTING

APPROPRIATE CONTROL VALVES

FOR PIPEWORK SYSTEMS

JOHN R. E. CHRISTY
University of Edinburgh e Edinburgh EH9 3JL, Scotland

While the subject of selecting appropriate centrifu-
gal pumps for a given pipework system is treated
reasonably well in many textbooks,""'2 there is
usually little or no discussion of selection parameters for
control valves. Even specialized books on control'34' tend to
concentrate on control theory rather than on what some
would regard as a more mundane, but nevertheless neces-
sary, practical understanding of the interaction between con-
trol valve characteristics and the dynamics of the system into
which it is installed. Although Seborg, et al.,"'4 covers much
of this material, it is largely embedded within an example in
a short section on final control elements.
At Edinburgh University, in the second year of our four-
year undergraduate course, selection procedures for both
centrifugal pumps and control valves are taught within the
same "Plant Engineering" course. Two aims of the course
are to introduce practical aspects of chemical plant control
and to give the students a practical understanding of the
interactions between pumps, control valves, and the pipework
systems in which they are installed.

SELECTING CENTRIFUGAL PUMPS
For pipework systems (excluding control valves) a pump
with a suitable characteristic can be chosen by reference to
the system curve (plot of pressure drop through system,

John Christy, a graduate of both Cambridge
University(MA) and Edinburgh University (PHD),
is a senior lecturer in chemical engineering at
the University of Edinburgh. He teaches first-
year mass and energy balances and vapour liq-
uid equilibrium, second-year separation pro-
cesses and plant engineering, and final-year fluid
mechanics, along with running final-year research
and design projects.

expressed in terms of fluid head, h, versus flowrate, Q), as
shown in Figure 1. The pump should be chosen so that 1)
the operating point is close to the most efficient opera-
tion of the pump, and 2) the net positive suction head
(NPSH) required by the pump is less than the calculated
available net positive suction head.
The students are also taught about other design criteria for
the pump, such as materials of construction, types of impel-
ler, or the possibility of increasing throughput by mounting a
larger impeller in the pump (provided that the pump casing
is large enough). To ensure that the students have an under-
standing of the significance of the operating point on the
graph, they are asked to consider what will happen to the
operating point upon changing 1) the static delivery head, 2)
the resistance to flow in the system (obtained, for example,
by partially closing a valve), and 3) the pump characteristic.

SELECTING CONTROL VALVE SIZE AND TRIM
In the following discussion, numerical values quoted are
based on the assumption that a wide range of controllable
flowrates is required, with significant turndown ratios
(rangeability).
When designing a pipework system including a control
valve, the same graphical technique can be used as shown in
Figure 1. The pump should usually be chosen, however, so
that 1) the operating point without a control valve in place
corresponds to a flowrate about 40-50% greater than the
normal operating flowrate, and 2) there is a reasonable head
difference between the characteristic curve and the system
curve at the normal operating flowrate (usually about 5m
or more). This will allow both for efficient control around
the normal operating point and for moderate controlled
increase in the flow.
From the plot of system curve and the chosen pump char-

Copyright ChE Division ofASEE 1996

Chemical Engineering Education

... in the second year of our four-year undergraduate course, selection procedures for both centrifugal
pumps and control valves are taught within the same "Plant Engineering" course. Two aims of
the course are to introduce practical aspects of chemical plant control and to give the
students a practical understanding of the interactions between pumps, control
valves, and the pipework systems in which they are installed.

Volumetric Flowrate, Q

flowrate

Figure 1. Calculation of operating flowrate in a pumped
system.

Developed
or
Required

Volumetric Flowrate, Q

desired /
flowrate

a

Static /
Delivery
Head System Curve
with valve

Flowrate without
control valve /Volumetric Flowrate,

Ah,

System Curve
without valve
H = P 1 + z2-, +

(NB 72- + z2 must be negative)

Figure 2. Calculation of Ahvfor pumped and
gravity-fed systems.
Winter 1996

acteristic, it is then possible to determine the head loss
across the control valve, Ahv, at a full range of flowrates.
The same procedure can be used even if the flow is
gravity driven rather than being pumped (as shown in
Figure 2).
For control valves, we must first specify the maxi-
mum volumetric discharge coefficient or valve con-
stant, C,,,-. The operating volumetric discharge coeffi-
cient, C,, is a function of the fractional stem position, x,
such that

Cv = f(x)Cv,max

The value of x lies between 0 (fully closed) where f(x) =
0 and 1 (fully open) where f(x) =1. It should be noted
that the mathematical functions describing valve char-
acteristics do not always fit these limits. Since

Cv='

if Ah,v is independent of flowrate, then a plot of Q vs. x
is identical in form to f(x) vs. x-the inherent valve
characteristic.

By plotting head loss, Ahv, versus flowrate, an
initial choice of valve trim can be made; if Ahv
remains almost constant over the desired control-
lable range, a linear trim should be chosen, whereas
if Ah, drops as the flowrate increases, a valve trim
giving increasing sensitivity (such as a hyperbolic or
equal percentage trim) may be more suitable. De-
creasing sensitivity trims, such as the square root
trim, are usually reserved for situations where rapid
opening is required.
Several criteria can be used to estimate an appropriate
maximum volumetric discharge coefficient Cv,max. If Ah,
is independent of flowrate, then the variation of Q with
x will be linear and a linear trim, such that f(x) = x,
should be chosen, with Cv.m. set so that the normal
operating, C, is around 60-70% of Cv.m. For most sys-
tems, however, Ahv varies with Q-in which case it is
better to select an upper limit for the flowrate. This,
for example, could be approximately 40% greater
than the normal flowrate, providing that there is still
sufficient head available to accommodate the pres-
sure drop across the valve. We can then evaluate
Cv.ma directly as

Head
Required
H

!

Qmax

where Ah, is determined at Q.ma. These values of Cv.ax are
best chosen using typical manufacturers' data and should in
any case be assessed to ensure that they correspond to a
valve with a diameter less than that of the pipework.
Values of Cv can be evaluated at a range of flowrates
between zero and Qmax. Using the chosen value of Cvm,,
values of f(x) (=Cj/Cv,) can then be tabulated against Q. By
rearranging the expression for f(x) for each type of valve to
give x in terms of f(x), the values of x for each trim can be
tabulated against Q. An example using four types of trim is
given in Table 1. (Note that where the mathematical func-
tion, f(x), leads to values of x outside the range of 0 to 1, the
appropriate limit of x has been entered in the table.)
By plotting Q vs. x for each type of trim (see Figure 3),
the valve trim giving the most linear response over the
required controllable range can be chosen. For the example
given, the hyperbolic trim gives the greatest sensitivity with
an almost linear response over the widest range of flowrates
(5-25 m3h '). At this stage, the chosen value of Cv... may be
altered with the last stage in the procedure being repeated, if
necessary, until a good linear response is obtained. In the
example above (Table 1), it is clear that the maximum con-
trollable flowrate and hence Cv,ma chosen is too close to the
operating capability of the pump with no valve present. It
would be better here to reduce the maximum controllable
flowrate to, say, 20 m3h-' or to use a larger pump, which
would allow a valve with a considerably lower Cv,. (say, 22
m3h 'bar-1/2) and perhaps a linear trim to be used.
For a rapid choice of valve trim, inspection of the graphs
showing both the pump characteristic and the system curve
excluding the control valve will yield useful information.

Figure 4 shows the regions of flowrate for which linear and
equal percentage trims would be applicable in four systems
with quite different dynamics.

OTHER SPECIFICATIONS
Apart from the choice of materials for the valve parts,
including the gland and the valve seat, the students are

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Fractional Opening (x)

Figure 3. Installed valve characteristics.

TABLE 1
Sample Data for Control Valve Trim Selection
Cvm = 113 m'h 'bar-12; a= 4 (equal%); a = 50 (hyperbolic)

Linear

Ah, C.

m mh 'bar-1/2

56.5

0
1.98
4.25
7.30
12.9
113

x v
x max

0
0.018
0.037
0.065
0.114
1.0

Equal %
[n(Cv/Cvmax)
a

Hyperbolic
0- (Cvmax/Cv)
al-I

0
0.018
0.037
0.065
0.114
1.0

6 Chemical Engineering Education

m3/h

Square Root

(Cv,
(Cv,max

Figure 4. Choice between linear and increasing sensitivity valve trims.

taught why it is important to specify the position of the
valve on control failure and to be able to identify which
way a valve will fail by inspection. For high pressure
drops in the flow, the use of double ported valves or
valve positioners is also discussed.

ASSESSMENT
Due to the practical nature of this topic and the range of
design assumptions to be made, examination questions of
the length traditionally set in the second year of our course
would not adequately test the students' abilities. Instead, the
students are given a hand-in exercise to complete that counts
towards the degree assessment. This normally involves a
given pipework system and the pump characteristic, with the
students being asked to evaluate the available net positive
suction head and the operating flowrate without a control
valve in place. A selection of four or more control valves
having different valve constants and trims are then given
and the students are asked to select the most appropriate
valve for the given duty. Normally, two of the given
valves could be chosen, neither being perfect for the
duty. The students are thus encouraged to discuss the
Winter 1996

reasons for their choice of valve.

CONCLUSIONS
A procedure has been described for teaching students about
the selection criteria for both centrifugal pumps and control
valves for a given pipework system, along with comments
on the way this is taught and assessed at Edinburgh Univer-
sity. While the typical values given in the text assume a
straightforward case in which a wide controllable range is
required, the same procedure is suitable for more specialized
applications where the choice of C, -. and trim may be based
on alternative criteria.

REFERENCES
1. Coulson, J.M., and J.F. Richardson, Chemical Engineering,
Vol. 1, Pergamon Press (1977)
2. Perry, R.H., D.W. Green, and J.O. Maloney, Perry's Chemi-
cal Engineers'Handbook, 6th ed., McGraw-Hill (1984)
3. Stephanopoulos, G., Chemical Process Control: An Intro-
duction to Theory and Practice, Prentice-Hall (1984)
4. Seborg, D.E., T.F. Edgar, and D.A. Mellichamp, Process
Dynamics and Control, Wiley (1989) O

classroom

ON USING A

BOUNDARY PERTURBATION

TO LINEARIZE A SYSTEM

OF NONLINEAR PDEs

N. W. LONEY
New Jersey Institute of Technology Newark, NJ 07102

While a mathematical-methods course in graduate
chemical engineering programs is a much-needed
vehicle for training students in formal analysis, a
large part of the course time is expended in bringing all the
students up to a common mathematical-maturity level. Al-
though students successfully pass the course, they often
leave with the opinion that applications problems can only
be solved by using a computer. This may be true in most
cases, but some modified version of a difficult problem can
be analyzed using the linear theories discussed in class and
can provide valuable insight into the phenomena being stud-
ied. In order to convince students of the value of this ap-
proach, connections must be made with applications prob-
lems. One way to demonstrate how such connections can be
made is given below.

BACKGROUND
Low-pressure chemical vapor deposition (LPCVD) pro-
cesses are currently popular methods to produce thin solid
films in the microelectronics industry. Of the numerous types
of LPCVD processes, the multiwafer hot-wall process is one
of the most economical and is widely used in production and
developmental research facilities. Currently, the design and
operation of the hot-wall process is done mostly through
trial and error. For those who can afford the expense, large
numbers of experiments are conducted to develop empirical

Norman W. Loney is Assistant Professor of
Chemical Engineering at New Jersey Institute of
Technology. He has studied chemical engineer-
ing at NJIT and applied mathematics at Courant
Institute of Mathematical Science. In addition, Dr.
Loney has practical experience in process devel-
opment, process design, and inplant engineer-
ing.

models that are then used to optimize the process.[" The
following is a problem taken from the process of chemical
vapor deposition (CVD) in a hot-wall reactor.

PROBLEM STATEMENT
Consider a cylindrical tube that is uniformly heated from the
outside. Inside, a pyrolyzable gas flows at a given set of flow
conditions. The inside tube geometry consists of a set of
circular disks placed vertically and equally spaced along the
cylinder axis (see Figure 1). These disks are supported in
such a way that their axes coincide with that of the cylinder.
Gas flows around the disks (through the annulus formed by
the disks and cylinder) by convection, while material trans-
fer between the disks is dominated by diffusion. Simulta-
neous with the mass transfer are chemical reactions, prima-
rily on the surface of the circular disks, that form the desired
deposit. One is interested in determining a reasonable con-
centration profile, given that the surface reaction rate is
klCACB (gmol cm-2) (1)
l+k2CA + k3CB (gmolm
where CA and CB are concentration of reactant and carrier
gas, respectively. For example, if the expected overall reac-
tion is

SiH2C12 + -4NH3 -- 1Si3N4 + 2 HC1 + 2H2 (la)

then species A is SiH2Cl2 and B is NH3. Note that species B
can actively participate in the chemical reaction, but CB must
be considerably larger than CA for it to be qualified as carrier
gas.
Equation (1) is typical of the surface rate expressions used
in CVD'1-5] systems and can be reduced to

Rate = klCACtt (2)
K+K'CA

@ Copyright ChE Division ofASEE 1996

Chemical Engineering Education

Although students successfully pass the [mathematical methods] course, they often leave with
the opinion that applications problems can only be solved by using a computer. This may be true in most
cases, but some modified version of a difficult problem can be analyzed using the linear theories
discussed in class and can provide valuable insight into the phenomena being studied.

where

K = 1 + k3Cto

K'= k2 k3 (4)

and Cot is the total concentration of species A and B. Also,
due to the abundance of species B in the system and the
small amount of A that is consumed in a given run, Ctot
remains relatively constant.

Before continuing, it is important to acknowledge some of
the other simplifying assumptions that are made, such as:
Gas phase kinetics are neglected
Isothermal condition exists in the reactor
Steady state prevails
CA < Ctot
The wafers are located perfectly axisymmetric with
the cylinder
Deposits on the wafer support are neglected
In order to initiate a solution, connections are to be made
between the equation of continuity and the simplifying as-
sumptions. It should be pointed out to the students that as an
alternate procedure, the final differential equation can be
derived directly from material balance considerations for
this system. It should also be noted that fewer simplifying
assumptions are usually needed to obtain numerical results
(solutions obtained by programmed numerical analysis). But
while it is an attractive feature, this approach requires great
skill and experience both in numerical analysis and in the
particular applications area. Interestingly enough, one
way to develop such needed skill and experience is
through analytical solution of a modified version of the
problem under consideration.

SOLUTION
Following application of the assumptions to the continuity
equation and employing constant mass density and diffusion
coefficient, we get the differential equation

28

GAS (r

Figure 1. CVD reactor configuration
Winter 1996

1 a (rCA) + ( CA) (5)
r r ar r z az
with r and z being respectively the radial and axial coordi-
nates as shown in Figure 1. Equation (5) is subject to the
boundary conditions of

_CA 0
az
0CA 0

at z=0

at r=0

CA(R,,Z)=CAb,; 0

DAB A = kCACtot
az K+K'CA

at = 8

In Eq. (8), CAb is the bulk concentration at the i* interwafer
region. This quantity (CAb ) is a function of the axial coordi-
nate and decays as the reactor exit is approached. But at the
1th interwafer region its value is assumed to be relatively
constant since wafer spacings are typically small.[6]
It is now convenient to recast the problem in dimension-
less form by making use of the following:
Let

CAbi

r z
R" C 6

Then Eqs. (5-9) become

-1 2F 2 F 1IF
a2 O2 2- V +

0F
-=0
8

at = 0

at = 0

F(1, ) = 1
--DABCAb, F(, 1) = Rxn rate (heterogeneous)
8 a(

(11,11a)

respectively, where

8
a-

Equation (6) or (13) is derived from the assumption that
the coordinate system being used is symmetric with respect
to z. The condition that species CA is finite at r=0 is conve-
niently expressed by Eq. (7) or (14). Equation (8) or (15)
simply describes the concentration of species A at that
interwafer region. Equation (9) or (16) states that the flux of
species A to a wafer surface is balanced by the surface
reaction there. This latter boundary condition, Eq. (9) or
(16), allows for some novelty in CVD modeling-in our
case, a boundary perturbation.

MOTIVATION FOR PERTURBATION
Upon examining Eqs. (5) through (9) or (12) through (16),
we see that the major obstacle to application of separation of
variables directly occurs in Eq. (9) or (16). The right-hand
side of the latter equations is a rational function. This quan-
tity can be represented by a convergent power series in an
appropriate region. For example, two possibilities for Eq. (2)
are

Rate = k k0CA + k A 2-... (2a)

valid for
K'CA >1
K
with
k0 k1Ctot (2b)
k2 k3

or

Rate kCACtot I K'CA K'CA 2 ... (2c)

valid for
K'CA <
--"< 1
K
By recasting Eq. (2) in its power series form we are able to
render the system of differential equations linear. That is, we
seek a general power series representation of the reaction
rate equation and modify the system as demonstrated below.
Considering the heterogeneous reaction rate expression to
be

where

Rate= koCAb (1+eF+E2F2 F 3F3 +-3.)

F=F0 +F1 +e 2F2 +

Eq. (20) is derived through comparison with Eqs. (2a) and
(2c). Then, substituting Eqs. (18) and (19) into Eqs. (12)
(17) through (16) results in

S(a2FO + a2pl 2 2F2
a2 2 2 + 2

a2F0 2F1 2 a2F2 1 (F0
2 2 2

F0+ DF 2+E 2F2 0
a + a + a+

F, aF2 +2.
+e -+e2 ***i

(21:

at = 0 (22:

8F0 8F1 2 _
ea +E 2 ... 0 at =0 (23)

F0 + eF1 +2 F2+=1 at -=1 (24)

DABCAb, (F0 aF1 2 aF2
-___-+.-. +

=koCAb, [I(Fo +eF1 2F2 +..)+2 (Fo +EFl +2F2 +...)2 +o(e3
(25)

Equating like powers of e, we get

o 0 1 a2F0 a2F0 1 aF0
a 2 2 a 2 a _

aFo(0, )
F ) = 0

a;

Fo(,I)= 1
DABCAb F0 kCAb at
8 k--OC' at =1

1 a2F1 a2F1 1 aF1
a2 a2 2- ;+

3F (0,) 0

a(

F,(,1)= 0

(k2 k3)CAO <
1 + k3Ctot

DABCAb, 8F1
8 --k0CAbFO
6 O

at =1

Chemical Engineering Education

1 a2F2 2F2 I 3F2
a2 2 W2 a4
F2 (0, _) 0

aF2 ( o, 0)

F2 (, 1) = 0 (39)

DABCAb, F2
-DABCAb koCAb (F, + FI) (40)
8 a

The continuation of this process is possible up to any
desired power of E. Note that by this process, an otherwise
nonlinear system is reduced to a set of linear problems.
Solving each of the above linear systems, we obtain

S= _2ko Jo (xn)cosh(aan0)
aDAB 2 n sinh2aaJan)

Jo(an)= 0 (42)

8ko cosh(aon)- an sinh(aan)J0o(an,)cosh(an,,t)
F 28ko DAB
a2DAB n=0 T sinh2 (an )J (n)

(43)

F2 = QnJo(0n)cosh(aaan) (44)
n=o

with

-25k0 fFo2JO(an)d 4 2FJo(a)di
DAB f 0aDAB )
Q n 0 o__ 0 (45)
aaun sinh(aan)J2 (n) aan sinh(aan)J (a5n)

where J0 and J, are Bessel functions of the first kind of order
zero and one respectively.
The first order approximation F0 is sufficient to describe
the concentration profile of interest. Also note that Eq. (18)
expresses a general form of the two possible power series
expansions of Eq. (2).
Equation (41) was tested in the prediction of silicon nitride
deposition data with reasonable results.[6]

REFERENCES
1. Badgwell, T.A., T.F. Edgar, and I. Trachtenberg, J.
Electrochem. Soc., 139, 524 (1992)
2. Collingham. M.E., and R.L. Zollars, J. Electrochem. Soc.,
136, 787 (1989)
3. Hitchman, M.L., J. Kane, and A.E. Widmer, Thin Solid
Films, 59, 231 (1979)
Winter 1996

4. Roenigk, K.F., and K.F. Jensen, J. Electrochem. Soc., 134,
(36) 1777 (1987)
5. Kuiper, A.E.T., C.J.H. Van den Brekel, J. de Groot, and
G.W. Veltkamp, J. Electrochem. Soc., 129, 2288 (1982)
(37) 6. Loney, N.W., and C.R. Huang, Thin Solid Films, 226, 15
(1993) 0

REVIEW: Hazardous Waste Management
Continued from page 19.
100-200 new students in the introductory course each year,
and as a result we eventually introduced a full MS in HWM,
with over thirty required and elective courses. We developed
a separate, in-depth course in each of the Wentz chapters and
subtopics, outgrowing much of the text. Moreover, the text's
data and examples had become dated. So we switched texts,
despite the excellent material contained in many chapters
and the presentation from the chemical engineer's perspec-
tive. We therefore looked forward to the second edition.

Clearly, most students will take only a single course in this
field; hence, there is a need for a broad-based text covering
the entire subject. The second edition of Hazardous Waste
Management fits that need well. Wentz provides an overview
of all the major topics that should be addressed in HWM.
A major strength of the revised text is its comprehensive
coverage within a reasonable number of pages. Another
strength is its hard-hitting chemical engineering approach to
the technical treatment and remediation subject areas. A
major asset is the case studies of many of the major chemical
tragedies that have driven federal law development, allowing
the student to gain a sense of the field's history as well as its
technical requirements. The second edition also groups topics
in a more logical order, such as merging toxicology with risk
assessment, landfills with injection well disposal, while giving
other areas, such as biological treatment, a separate chapter.
The new material in environmental auditing and site as-
sessment and the enhanced technical material in site
remediation are major positive additions to the second edi-
tion. Likewise, the addition of problems to the technical
chapters improves the text significantly as a teaching tool.
Weaknesses of the text include the aged data in Chapter 1,
the brevity of the toxicology and environmental auditing
sections, and the absence of the OSHA perspective and
statistical considerations of data, none of which are serious
drawbacks for the student who will take several courses in
the field. The international environmental examples pro-
vided will increase the students' global awareness, but are
largely out of context as they follow discussions of US law.
In summary, the Wentz text led the field since it was first
published in 1989; the revised second edition is greatly
improved and is a sound choice for a course to round out the
background of most chemical engineers. It is also a good
choice for an introductory course for a wide range of disci-
plines with interest in the field. 0
61

,O1 classroom

DESIGN OF SEPARATION UNITS

USING SPREADSHEETS

MARK A. BURNS, JAMES C. SUNG
University of Michigan Ann Arbor, MI 48109-2136

Most design calculations in an undergraduate sepa-
rations course involve solving a combination of
mass balance, mass transfer, and equilibrium equa-
tions. The resulting system of equations is usually not solved
analytically for two reasons: 1) the system of equations is
almost always nonlinear, and 2) some of the information
(typically the equilibrium data) is not available in analytical
form. Before the advent of digital computers, the only option
available to teachers and students alike was to use graphical
solution strategies. In recent years, however, the speed of
computers has increased to such an extent that even lower-
end personal computers have sufficient power to solve these
design equations.
In choosing an application to solve the system of equa-
tions, one must decide between faster, task-specific "separa-
tions" programs (bought commercially, written in-house, or
a combination of the two)"1 and more general "equation
solving" programs (e.g., Mathematica, Maple, HiQ, etc.).12'
The task-specific programs have the advantage that both the
Mark A. Burns received his BS in chemical
engineering from the University of Notre Dame
and his MS and PhD in chemical and biochemical
engineering from the University of Pennsylvania.
After spending several years teaching at the Uni-
versity of Massachusetts, he joined the faculty at
the University of Michigan. His research interests
are in the area of bioseparations and include ad-
sorption, chromatography, and micromachined -
systems.

James Sung just completed his second year of
graduate work in chemical engineering at the Uni-
versity of Virginia. He received his BR from the
University of Michigan in 1993. He is currently
studying bacterial migration in porous media with
applications to bioremediation.

input (the problems you choose) and the output (how the
results are displayed) can be rather complex. The disadvan-
tage is that the equations and the solution strategy are com-
pletely transparent to the user. Equation-solving programs
alleviate this problem but require that students learn what
may be rather complex programs. Although each strategy
has its advantages, we have chosen to use a simple equation-
solving application to teach separation-unit design.
Although many programs can be used to solve systems of
equations, we have chosen simple spreadsheet programs"3'
that can perform complicated mathematical calculations[4'5]
and display the solutions in a variety of tabular and/or graphi-
cal forms. The matrix-like structure of the spreadsheet is
ideal for solving a system of coupled linear algebraic equa-
tions using Gauss-Jordan Elimination"6' or simple matrix in-
version. Other techniques can be used to solve more compli-
cated nonlinear equations on spreadsheets, including sys-
tems of both ordinary and partial differential equations."'7
The limited power and speed of most spreadsheet applica-
tions is more than compensated for by the ease of "program-
ming" and the immediate presentation of results.

AN EXAMPLE: DISTILLATION
McCabe-Thiele and Ponchon-Savarit diagrams are the most
common graphical solution strategies taught in a separations
course. Realizing that this distillation-column design is merely
the graphical solution of a series of sequential, nonlinear
equations, one can construct a McCabe-Thiele diagram us-
ing a spreadsheet. Assuming a total condenser is used, the
distillate composition is equal to the vapor composition leav-
ing the top tray. With the additional assumptions of constant
relative volatility and constant molar overflow, the liquid
composition leaving the top tray can be calculated using the
equilibrium relationship. A mass balance on that tray then
yields the vapor composition entering that tray from the tray

Copyright ChE Division ofASEE 1996

Chemical Engineering Education

Before the advent of digital computers, the only option available to teachers
and students alike was to use graphical solution strategies. In recent years, however, the
speed of computers has increased to such an extent that even lower-end personal
computers have sufficient power to solve these design equations.

below. Using the equi-
librium relationship
again, the liquid com-
position can be calcu-
lated on the next tray
and so on. Each of
these equations is en-
tered into a particular
cell on the spread-
sheet, and the result-
ing tray compositions
can be plotted on an
x-y diagram.
Figure 1 shows the
parameter entry and
graphical results sec-
tion of such a spread-
sheet. Note that al-
though we have cho-
sen Microsoft Excel

Distillation (Constant relative volatility, Liquid tray efficiencies)

by Mark Bums and James Sung
Department of Chemical Engineering at the University of Michigan
Example from Geankoplis, 1993, p. 656 and 660

D= 41.18 mol/hr
X> d= 0.95

F= 100 mol/hr R
F____ --- J alpha 2.38
Xf-= 045 -' q= 1.195
E(ML)= 1

W = 58.82 mol/hr
Xw= 01

1.0-
0.8
0.6
0.4
0.2
0.0
0.0

0.5 1.0
X

Number of Stages Required = 8
Minimum Reflux Ratio = 1.2355
Minimum Number of Stages = 6
(at infinite reflux and E(ML)=1)

Figure 1. Parameter entry section of a distillation spreadsheet. After entering
the design parameters (boxes), the spreadsheet calculates the distillate (D)
and bottoms (W) flow rates, the number of stages required, the minimum
reflux ratio, and the minimum number of stages. After a new parameter is
entered, the calculated values are immediately updated.

(version 4.0) for all the examples in this paper, one should
be able to incorporate the concepts shown into almost
any spreadsheet program.
The parameters, including the reflux ratio (R), the relative
volatility (a), the feed composition (Xf), flow rate (F), and
quality (q), the distillate (Xd) and bottoms (Xw) composi-
tions, and the tray efficiency (E(ML)), are entered into the
boxes next to the appropriate symbols. The spreadsheet pro-
gram then calculates (within seconds) the number of trays
required for the separation and plots a McCabe-Thiele dia-
gram. Changing the value in any of the boxes and pressing
"Enter" immediately updates the screen with the new de-
sign. Note that no macros (prerecorded series of com-
mands or manipulations) are required to perform any of
these design calculations.
There are many advantages to using spreadsheets such as
this one in a separations course. The most obvious advantage
is that, from a teaching point of view, the instructor can
spend a significant amount of time discussing the phenom-
ena of distillation as opposed to the technique of graphical
calculation. With a notebook computer, an LCD display, and
an overhead projector, the instructor can use the spreadsheet
during class to show the effect of different variables on the
separation design. For example, one can slowly decrease the
reflux ratio until a pinch-point occurs to visually calculate
the minimum reflux ratio, or increase the reflux ratio to
show that the operating lines collapse to the y=x line. By
gradually changing one variable at a time, 50 to 100 column
Winter 1996

designs can be done
in a fifty-minute lec-
ture! If these designs
are saved to disk, stu-
dents can then take
these same spread-
sheets home and use
them to solve calcu-
lation-intensive
homework problems.
The disadvantage of
this technique is that,
if one avoids the use
of macros, some struc-
tures that are relatively
easy to accomplish in
a computer program
must be given careful
thought in a spread-
sheet. For example, a

series of nested "IF" statements must be used in the mass
balance section of the distillation spreadsheet to insure that
the proper "operating line" is used (upper or lower). Also,
since the solution is recorded in a specific location on the
spreadsheet, there are restrictions as to the maximum num-
ber of stages that can be calculated. In the spreadsheet shown
in Figure 1, the equations were copied into 20 rows corre-
sponding to 20 trays, resulting in a 20-tray maximum for
that particular spreadsheet. We have found this limit suf-
ficient for instructional purposes, although there is no
reason why a spreadsheet with a limit of 100 or more
stages cannot be constructed.
There are many possible problems that the students can
solve with this spreadsheet or slight modifications of it.
Retrofits can be performed by varying the parameter of
interest until the spreadsheet returns the stage number of the
existing column. Multiple feeds, sidestreams, and mislocated
feeds can easily be handled by changing the "IF" statements
that control the choice of operating line. More extensive
modifications allow the inclusion of heat effects by adding
an energy balance to the previously described equations.
Also, actual tabulated data can be used in place of analytical
equilibrium and enthalpy expressions.

USEFUL MATHEMATICAL
TECHNIQUES ON SPREADSHEETS
Although distillation with constant relative volatility can
be solved using successive substitution, many real separa-

tion systems have complexities that make their design on
spreadsheets more complicated (see Table 1). Over the last
several years, we have found that a number of mathematical
techniques are useful in solving a wide variety of separation
problems. They are listed below with a description of their
implementation on a spreadsheet and an accompanying sepa-
ration unit design problem.

Linear Interpolation and Lookup Tables (Distillation
with Equilibrium Data)

tion was, say, y=0.355, linear interpolation between the
y=0.304 and the y=0.418 data points would yield a value of
x=0.078 for the liquid composition, or

x= below + (Xabove- Xbelow )(- Ybelow )/(Yabove Ybelow) (1)

This interpolation is performed in the spreadsheet by using
lookup commands that return the y values above and below
the actual y value in the "data table" and the associated x
values. The data points do not need to be evenly spaced to

The previous distillation spread
sheet assumed constant relative vol
utility (a) for the binary mixture. TI
vapor-liquid equilibrium of most t
nary mixtures, however, is not ide
and is instead best represented t
tabulated equilibrium data. The u
of tabulated data as opposed to a
suming constant a is irrelevant wh(
using graphical techniques but pr
sents a significant problem in mc
analytical solution strategies.
Using tabulated data in sprea
sheets is quite simple due to the tab
lar format of the spreadsheets. On
the equilibrium data is entered, i
terpolation between data points
necessary operation when calculz
ing equilibrium values) can easi
be accomplished using either cub
(or higher) spline fits or simple li
ear interpolation. We have found th
linear interpolation is preferred
almost all cases because (a) the ca
culations are extremely easy, (b)
increase in accuracy in one re-
gion from a cubic spline fit is
usually offset by a decrease in
accuracy from an incorrect fit
at another location, and (c) in
a region of high curvature, ex-
tra data points can be added to
increase the accuracy of the
linear fit.
As an example, consider the
equilibrium data entered into
the spreadsheet shown in Fig-
ure 2. In "stepping down" a
distillation column design, one
might know the vapor compo-
sition leaving a tray and want
to calculate the liquid concen-
tration in equilibrium with that
value. If the vapor composi-
64

Lu-
a- Distillation using the McCabe-Thiele method with equilibrium
data points
Sby Mark umns and Jamnes Sung
Department of Chemical Engineering at the University of Michigan 1 .0
Example from King, 1980. 0 a
al p 2. Pro.5-B 19x 3 rnln, 0 6
by
se F C mou 09625 O
s- x[ n =/ oE(Mo) O0 0.s 1 0
03n x
e-
e- X xW= S20.46 nmL0hr
XW m= 20. r Number of Stages Required = 16
Ist
EaullbruDataMiEni
d-
PoilnrlNio- x Y
u- oo. o.ooo
1 0.020 0134
ce 0.060 0304
S3 0.100 0418
4 0.200 0.579
n- 5 0300 0.665
6 0.400 0.729
(a 7 0.500 0.779
a 9 0.700 0 8270
it- 10 0.800 0.915
11 0.900 0.953
ly 12 000 0.979
13 1.000 1.000
ic 14 1.0 1.o0o
n-
at Figure 2. Distillation spreadsheet using tabulated equilbrium data. Note that the data points
in do not need to be evenly spaced or fill the entire table (an extra (1.0,1.0) is added at the end of
l- the table). Although spline fits can be used to interpolate the data, we have found that using
simple linear interpolation gives the most accurate results provided a sufficient number of data
an points is used.

TABLE 1
Mathematical Techniques Used in Separation Spreadsheets

Separation
System
General Countercurrent
Unit design, linear equilibrium
Unit design, nonlinear equilibrium
Unit design, equilibrium data
Retrofit, linear equilibrium
Retrofit, nonlinear equilibrium
Retrofit, equilibrium data

Specific Unit Design
Ideal distillation
Distillation with equilibrium data
Distillation with equil./enthalpy data
Absorption (real equilibrium data)
Liquid/Liquid extraction immisciblee)
Liquid/Liquid extraction misciblee)

Matrix Jacobian/
Manipulations Iteration

Successive Linear Quadratic Numerical
Substition Interpolation Fit Integration

V V

V V V V

V
V V
V v V

Chemical Engineering Education

perform these calculations and there is no set maximum
number of points that can be used in the table.

Figure 2 also shows the earlier distillation spreadsheet
modified to include equilibrium data and lookup commands.
Although 13 data points are shown in the example (the 14th
is a repeat), the table can be expanded to include 100 or more
points. For display purposes (and for most calculations), the
14 points produce a relatively smooth curve. Note that, for
low x-values where the slope of the equilibrium curve is
large, additional data points have been added. Also, several
columns between "x" and "y" data columns are hidden that
shift the x- and y-values up one row; this technique makes
interpolation with lookup commands much simpler.

Matrix Inversion (Countercurrent Distribution)

For the design of countercurrent separation systems, suc-
cessive substitution is needed because the number of stages
that will be used in the system is unknown. If the design
involves a known number of stages (e.g., retrofitting a sepa-
ration to an existing column), then matrix techniques can be
used instead. Most spreadsheets include basic matrix-ma-
nipulation commands such as matrix inversion and multipli-
cation. Although other techniques can be used to solve the

equations, the simplicity and speed of matrix manipulations
on spreadsheets makes matrix techniques particularly attrac-
tive.

As an example, one can calculate the separation that would
be obtained for a countercurrent distribution system (such as
immiscible liquid-liquid extraction). If the equilibrium dis-
tribution coefficient is constant (linear equilibrium relation-
ship) and the number of stages is fixed, the mass balance
equations reduce to a set of linear algebraic equations. Once
the flow rates and compositions of the entering streams are
specified along with the equilibrium constant (K=y/x), the
equations for a 5-stage system are

-(L + KV)X + (KV)X2

LX (L + KV)X, + (KV)X3

LX, (L + KV)X3 + (KV)X4

LX3 -(L + KV)X4 + (KV)Xoa,

LX4 (L + KV)Xout

=LXin (2)

S0 (3)

= 0 (4)

=0 (5)

-YinV (6)

where X and L are the composition and flow rate of one
liquid, and Y and V are the composition and flow rate of the

Countercurrent or Cocurrent Distribution (Linear Equilibrium)

by Mark Bums and James Sung
Department of Chemical Engineering at the University of Michigan

Counter- (1) or Co- (0) current operation ?

V 60
< ...

Xln- 0.001
.53 (y = Kx)

Yout Y2
0.0057 0.0076

Xln Xi
0.0010 0.0023

Y3
0.0087

X2
0.0030

Y4
0.0094

X3
0.0034

Y5
00098
4 <-----

X4
0.0037

Yin
0.01

Xout
0.0039

Figure 3 (a) (above) Data entry section for staged countercurrent or cocurrent
separation assuming linear equilibrium and immiscible phases (liquid-liquid
extraction, absorption, adsorption, etc.).
(b) (below) Matrix calculation section of spreadsheet.

other. Figure 3(a) shows the flow diagram for
the process.

The equations above can also be written in
matrix notation:

AX=b (7)

To solve for X, the matrix A is first inverted
using the matrix inversion command. The ma-
trix multiplication command is then used to find
the product of the inverted matrix and the vector
b or

X = A-b (8)

The composition of the other stream (Y) is found
by merely multiplying X by the equilibrium con-
stant, or

Y = KX (9)

Solution using matrix Inverlion
Eauations in Matrix Notation:
Coefficent Mat i (A) f (xln
-241.80 151.80 0 0 0 X -0.09
90.00 -241.80 151.80 0 0 X20.00
0 90.00 -241.80 151.80 0 X3 0.00
0 0 90.00 -241.80 151.80 X40.00
0 0 0 90.00 -241.80 X5
Inverted Coefficient Matrix (A Inveme)
0.006382 -0.006036 -0.005451 -0.004466 -0.002804 Solution Scheme
-0a003579 -0.009614 -0.008684 -0.007114 -0.004466 (1) Ax=b
-0.001916 -0.005148 -0.0106 -0.008684 -0.005451 (2) x = Alnvers*b
-0.000931 -0.002501 -0.005148 -0.009614 -0.006036 (3) y= Kx
.0.000346 -0.00093, -0.001918 -0.003579 -0.006382
Solution
Yl = 8.00571 X1 = 0.00226
Y2 = 0.00759 X2 = 0.00300
Y3 0.0071 X3 = 0.00344
Y4 0.00937 X4 = 0.00371
Y5 0.00977 X5= 0.00386

Winter 1996

The section of the spreadsheet that
performs these calculations is shown
in Figure 3(b). The coefficients of the
above equations are entered into the
matrix A. Calculations are performed
each time a new flow rate, equilib-
rium constant, or composition is en-
tered into the spreadsheet. For nonlin-
ear equilibrium relationships, a simi-
lar spreadsheet can be constructed ex-
cept that an iterative scheme is neces-
sary (Newton's method, multivariable:

0.01

> 0 0.005- 0J0

0 0.002 0.004
x

see Figure 4). The nonlinear component-mass-balance equa-
tions for each stage are entered into the matrix f. The deriva-
tive of each equation with respect to each variable is then
entered into the Jacobian matrix and used to direct the itera-
tion steps. The equation

Xk+1 = Xk Jk fk (10)

describes this process where k is the iteration step and J is
the Jacobian matrix. This iterative spreadsheet usually con-
verges within seconds.

Numerical Integration (Absorption)

For staged separation systems, such as distillation in tray
columns, stage-by-stage calculations need to be performed,
as shown earlier. Separations in packed columns, however,
require integration of the mass-transfer driving force before
an accurate design can be made. This integration Countercu
is one of the easiest calculations to perform on a by Mark Bums a
spreadsheet and is used to calculate the neces- Departmentof
sary column length provided the driving force is Note: At
known as a function of position or composition. ".

Packed column absorption is a typical separa-
tion system in which the driving force can be
calculated as a function of composition. Although
many different forms of the design equations
exist, we use the following form of the absorp-
tion equations: 18

Ymi
GL s (1- Y)*m
C Kya (1- y)2(y y*)

Xin =
Yin =
K1 =
K2=
Yout
0.39927

Xin
0.1

Yin

LC = 1( -1) *dy (13)
c f (l-y)2I n 1- y
Yourt ( n y

This integration can be performed on a spreadsheet using the
trapezoidal rule. The interval from you, to y,, is divided into n
intervals and the integral is calculated using the equation

Integral= f(yi)+f(yi+) (yi+l i) (14)

i=1
where f(y) is the integrand in Eq. (13). The accuracy of this
technique is a function of the number of intervals chosen (n)
with larger values of n increasing the accuracy.

A spreadsheet that performs these calculations for a con-
tinuous absorption column is shown in Figure 5. The spread-

rrent Distribution (Nonlinear Equilibrium) I

ind James Sung
chemical Engineering at the University of Michigan
Sthe start of each spreadsheeting session, initiate
e iterative solution scheme by going under the
options" menu, clicking "Calculation.", and then
king on the "Iteration" button.

07 Equilibrium:
4 y=KIx/(K2+x) V = 50
< .. ..

Y2 Y3
0.21170 0.06864

X1 X2
0.05311 0.01734

Y4
0.01667

X3
0.00435

Y5 Yin
0.00317 0
4 -<----- 5 <-----
X4 Xout
0.00098 0.00018

L = 200

where Le is the length of the column, Gs is t
solute-free gas flow rate, kv is the overall m
transfer coefficient, a is the surface area ava
able to mass transfer per column volume,
y,n and you, are the inlet and outlet gas con-
centrations, y* is the equilibrium gas con-
centration based on the liquid-phase con-
centration, and (1-y),. is defined as

(1 -y)m = (1 y *) (1 y) (12)
en I-Y*]

In Eq. (11), the term (y-y*) is the driving
force for mass transfer. This driving force
must be modified when dealing with con-
centration solutions to include variable
flow rates ((1-y)2) and diffusion-induced
convection (1/(1-y).m).

Substituting Eq. (12) into Eq. (11) and
rearranging, we get

66

the
ass
il-

Figure 4. (a) (above) Staged countercurrent or cocurrent separation assuming
nonlinear equilibrium and immiscible phases (liquid-liquid extraction, absorp-
tion, adsorption, etc.). This spreadsheet is similar to that shown in Figure 3
except that a nonlinear isotherm is used.
(b) (below) Matrix multiplication and iteration section of spreadsheet.

Solution using Newton's method for multivariable systems reset I 0
Initial guess (X(k))
X1 = 0.05311 X2= 0.01734 X3= 0.00435 X4= 0.00098 Xout= 0.00018
Equations in Matrix Notation (Taylor Series Approximation)
Jacobian Matrixs J) (AX)
361.5 -425.8 0 0 0 AXi 0.00000
200 -625.8 711.82 0 0 AX2 -2E-15
0 200 -911.8 833.83 0 AX3 = 6E-16
0 0 200 -1034 867.08 AX4 -6E-17
0 0 0 200 -1067 AXout 3E-17
Inverted Jacoblan (J-1)
0.0061 -0.006 -0.006 -0.006 -0.005
0.0029 -0.005 -0.005 -0.005 -0.004
0.0008 -0.001 -0.003 -0.003 -0.002
0.0002 -3E-04 -6E-04 -0.002 -0.001
3E-05 -6E-05 -IE-04 -3E-04 -0.001
New guess (X(k+l))
X1 = 0.05311 X2= 0.01734 X3= 0.00435 X4= 0.00098 Xout= 0.00018
Solution Solution Scheme
X1 = 0.05311 Yout = 0.39927 X(k+l)=X(k) (J-1)f
X2= 0.01734 Y2= 0.21170
X3= 0.00435 Y3 = 0.06864 (* Iterative solution scheme)
X4 = 0.00098 Y4 = 0.01867
Xout = 0.00018 Y5 = 0.00317

Chemical Engineering Education

0.50-
0.40- 0
0.30-
,I
0.20
0.10-
0.00
0.00 0.05 0.10
x

oil

0

sheet uses tabulated equilibrium data and calculates the driv-
ing force at 20 points along the column (usually sufficient
although more can be added). The trapezoidal rule is then
used to calculate the integral listed above, and the length of
column necessary to perform the separation in either a co- or
countercurrent configuration is displayed. Note that the in-
tegral calculations in the spreadsheet are concealed in-
side the diagram of the column (see Figure 6). Also, in
this spreadsheet, the outlet gas-phase concentration is
specified; a similar spreadsheet can be developed speci-
fying the outlet liquid concentration.

Quadratic Fits (Liquid-Liquid Extraction)

The solutions presented thus far have involved relatively
simple mathematical techniques. If these techniques were
applied to liquid-liquid extraction with partially miscible
phases or distillation with tabulated enthalpy data, difficul-
ties would arise. Specifically, stagewise calculations are dif-
ficult when the stepping procedure occurs on graphs with
multiple "tabulated data" lines because the endpoints of the

Absorption in Packed Columns (Countercurrent, gas exit specified)
by Mark Bums and James Sung
Department of Chemical Engineering at the University of Michigan
SI Flow rates entering: Column sett
Liquid In Gas out Gas= 8 mol/hr*ft2 Kya = 27
40.80 mol/hr*ft2 27.77 mol/hr'ft2 Liquid = 40.80 mol/hr't2 Counter(l or (
0 mol frac 0.003 mol frac 1

mol/hr'ft2

mol/hr*ft2

Compositions entering Composition leaving
Gas-= 0.083 |molfrac Gas = 10.003 molfrac
Liquid= 0.000 mol frac Liquid = 0.056 mol frac

0.10

>- 0.05

0. 00+ I I I I
0.00 0.02 0.04 0.06 0.08
x

Length =
11.12 ft
Example from Welty, et al,
1984, p. 693

Liquid out J Gas in
43.2 mol/hr*ft2 30.18 mol/hr*ft2
0.06 mol irac 0.083 mol frac
t t___j

interpolations are not easily found. Using equations instead
of data eliminates this problem but requires the user to
generate and enter the equations into the spreadsheet.

An easier technique involves programming the spread-
sheet to generate a least-squares quadratic polynomial to fit
the data and then use the equation to design the separation
system. The least-squares quadratic polynomial is defined as
that polynomial, p(x), that minimizes the sum of the squares
of the error between the polynomial and the data points, (x,,
f,). The function to be minimized can be represented by

Q(f, P) = [fi P(xi)]2 (15)

i=l

where the quadratic polynomial is

P(x) = ax2 +a2x+a3 (16)

The coefficients a, a,, and a, are then found by taking the
derivative of Q with respect to each coefficient and setting
these equations equal to zero, thus minimizing the function
Q. After considerable mathematical manipula-
tions,"9' the coefficients can be shown to be

a = ( TX)1 T -.f (17)
.60 molVhr*ft3
^m1 where

Figure 5. Continuous packed column absorption spreadsheet. After entering the
parameters, the spreadsheet calculates (using Eq. 13) the length of packed column
necessary to perform the desired separation. Note that, although we have specified
the exiting gas composition, a similar spreadsheet can be devleoped that uses a
specified exiting liquid composition.

Ls = 408
Gs= 27.69
Liquid In Position x X y* I 1/(... I Negative? I J I Y y Gas out
40.80 mol/hr*ft2 0 0.0000 0.0000 0.0000 334.84 FALSE 0.0030 0.0030 27.77
0 molfrac 0.05 0.0029 0.0029 0.0037 284.11 FALSE 1.33 0.0074 0.0073 0.003
0.1 0.0059 0.0059 00075 247.92 FALSE 2.47 0.0117 0.0116
0.15 0.0088 0.0088 0.0112 220.84 FALSE 3.46 0.0160 0.0158
0.2 0.0117 0.0118 0.0149 199.84 FALSE 4.34 0.0204 0.0200
Gs= I 0.25 0.0145 0.0147 0.0185 183.08 FALSE 5.13 00247 0.0241 A
27.69 mol/hr*ft2 I 0.3 0.0174 0.0177 0.0221 168.09 FALSE 586 0.0291 0.0283 I Length=
I 0.35 0.0202 0.0206 0.0256 154.37 FALSE 6.52 0.0334 0.0323 I 11.12
Ls= I 0.4 0.0230 0.0236 0.0291 143.09 FALSE 7.12 0.0378 0.0364 I
40.80 mol/hr*ft2 I 0.45 0.0259 0.0265 0.0323 129.08 FALSE 7.67 0.0421 0.0404 I Example fi
V 0.5 0.0286 0.0295 0.0351 11365 FALSE 8.15 0.0465 0.0444 I 1984, p. 6'
0.55 0.0314 0.0324 0.0380 101.78 FALSE 8.57 00508 0.0484
0.6 0.0342 0.0354 0.0408 92.38 FALSE 895 0.0552 0.0523
Liquid Out 0.8 00451 0.0472 0.0519 68.60 FALSE 1017 0.0725 0.0676 Gas in
43.2 mol/hr*ft2 0.85 0.0477 00501 0.0545 64.47 FALSE 10.42 0.0769 0.0714 30.18
0.06 mol frac 0.9 0.0504 0.0531 0.0572 60.89 FALSE 10.66 0.0812 0.0751 0.083
S0.95 0.0531 0.0560 0.0598 57.76 FALSE 10.88 0.0856 0.0788 0
1 0.0557 0.0590 0.0625 55.00 FALSE 11.08 0.0899 0.0825

Figure 6. View of the hidden columns in the absorption spreadsheet shown in Figure 5.
Winter 1996

a= a,
a,

f
n-

Thus, the coefficients (a) are
found from the data (f) using
the matrix x.

Figure 7 shows the implemen-
tation of this technique on a
spreadsheet that calculates the
number of stages necessary for
a particular liquid-liquid extrac-
tion system. The coefficients of
the calculated equation are
shown in Figure 8 and can be

used throughout the spreadsheet in place of the data to calcu-
late the required number of stages. If the fit obtained by this
method is unsatisfactory, additional data points can be added to
improve the fit or a higher order curve can be used. In our
work, we have found that quadratic fits do remarkably well for
both liquid-liquid extraction and enthalpy data in distillation.

CONCLUSIONS

What spreadsheets lack in power, they make up for in ease
of use. For a student, the results of changes to design vari-
ables in separation spreadsheets can be seen in a
few seconds in tabular or graphical form. In countercur
addition to changing design variables, the stu- by Mark Bums
Department of Ch
dent can also change the structure of the spread-
sheet to design more complicated separation units.
For instance, in binary distillation, multiple feeds, V1
one or more sidestreams, or mislocated feeds Watcdr
can all be added to the basic spreadsheet. The Isopropyl ether
techniques used to solve systems of equations in
these spreadsheets can also be applied to other
problems in other courses.

For an instructor, separation spreadsheets en-
Vn+1
able the lecturer to focus on the principles be- Acetic Acid
Water
hind the separation processes instead of the te- Isopropyl ether
dious solution procedures. But care must be taken
to ensure that the students do not use the spread- Figure 7.
sheets to solve homework problems that, with extraction
the aid of the spreadsheets, are merely "plug- ping" betn
using a qu
and-chug." During the separations course at ofstages f
Michigan, each spreadsheet is typically in-
troduced in lecture when that separation unit Detarmine Qu.dratic L.
is first discussed. Homework problems then Solvent/Extract Lav
include both simple, plug-and-chug type 1
problems that require a single use of the 1
X = 1
spreadsheet and more complicated, multiuse 1
problems (e.g., plot the number of trays 11
needed to perform this distillation as a func- 1
tion of reflux ratio). 1
XT = -0.990
Qualitative homework questions are also 0.980
common. More difficult problems that re- 0
quire the student to redesign the spreadsheet 39
are sometimes used, but are typically reserved Raffinate Layer
for group projects. Overall, the students have
had a very favorable reaction to the introduc-
tion of these spreadsheets in the separation X
course. As a final note, all the spreadsheets
shown here are available from the authors.'31

ACKNOWLEDGMENTS XT= .0

Many students were involved in develop- -oor0.0
ing the spreadsheets shown in this paper. as .8.16

James worKea on tne original versions ot Le
countercurrent distribution and distillation

68

spreadsheets as well as perfecting a number of the others.
Wilbur Woo investigated the use of cubic splines (which we
found were not necessary) and constructed the original "data"
spreadsheets. Mike Johns followed Professor Frey's paper
and did some simple chromatography solutions. Mike Vyvoda
spent considerable time debugging the final versions and
writing the all-important nomenclature tables. In addition to
these students, many other undergraduates at the University
of Michigan contributed to the spreadsheets through feed-
back in our separations course.

rent Liquid/Liquid Extraction
id James Sung
chemical Engineering at the University of Michigan

Extract

1818.52
0.1564
0.0403
0.8034

Solvent
1475.00
0.00000
0.0000
1.0000

Feed

1000. Lo
0.35001 Acetic Acid
0.65001 Water
0.0000 Isopropyl ether

Example from Wankat,
1988, p.595

Raffinate

656.48 Ln
0 Acetic Acid
0.8786 Water
0.0214 Isopropyl ether

1 6
S 1.4
1.2
S 1.0
0.8
S 0.6
0.4

-0.5 0.0 0.5 1.0
Xsolute,Ysolute

Number of Stages Required = 6
Note: Spreadsheet may generate incorrect
solutions when attempting a separation
i : :

Parameter entry and graphical output section of the liquid/liquid
spreadsheet. This spreadsheet is particularly complex because "step-
ween the equilibrium curves with linear interpolation is difficult. But
adratic fit to the data allows easy calculation of the required number
or any given system.

ast Sauares Fit for the Equllibrium Phase Data

er
-0.990 0.980
-0.971 0.943
.0.847 0.717
-0.715 0.511
-0.487 0.237
.0.165 0.027
-0.165 0.027
.0165 0.027
-0.165 0.027

0.000
0.004
0.019
0.)14
I= 0.216
0.362
0.464
0.464
0.464
0.464

I 1 1 I 1
-0.989 -0.971 -0.847 -0.715 -0.487
0.978 0.943 0.717 0.511 0.237

a3
a2
a1

-0.010 0.000
-0.012 0.000
-0.016 0.000
-0.023 0.001
-0.034 0.001
.0.106 0.011
-0.165 0.027
-0.165 0.027
-0.165 0.027
-0.165 0.027

0 1 0 0

-0.165 -0.165 -0.165 -0.165
0.027 0.027 0.027 0.027

0.000
0.007
0.029
0.133
If 0.255
0.443
0.464
0.464
0.464
0.464

-0.012 -0.010 -0.023 -0.034 -0.106
0.000 0.000 0.001 0.001 0.011

-0.165 -0.165 -0.165 -0.165
0.027 0.027 0.027 0.027

a6
a 5
a5

Figure 8. Calculation of the quadratic equations used in the Figure 7 spreadsheet.

Chemical Engineering Education

REFERENCES
1. Jolls, K.R., M. Nelson, and D. Lumba, "Teaching Staged-
Process Design Through Interactive Computer Graphics,"
Chem. Eng. Ed., 28(2), 110 (1994)
2. Taylor, R., and K. Atherley, "Chemical Engineering with
Maple," Chem. Eng. Ed., 29(1), 56 (1995)
3. All the spreadsheets shown in this paper are available from
the author. For copies of the spreadsheets, send a Macintosh
or IBM compatible disk (state which one) and a stamped,
self-addressed envelope to Mark A. Burns, Department of
Chemical Engineering, University of Michigan, Ann Arbor,
MI 48109-2136
4. Arganbright, D., Mathematical Applications of Electronic
Spreadsheets, McGraw-Hill Book Co., New York, NY (1985)
5. Rosen, E.M., and R.N. Adams, "A Review of Spreadsheet
Usage in Chemical Engineering Applications," Comput.
Chem. Eng., 11(6), 723 (1987)
6. Burns, M.A., "Mass and Energy Balance on Microbial Pro-
cesses," in Chemical Engineering Problems in Biotechnol-
ogy, M.L. Shuler, Ed., AIChE, New York, NY (1989)
7. Frey, D.D., "Numerical Simulation of Multicomponent Chro-
matography Using Spreadsheets," Chem. Eng. Ed., 24(4),
204 (1990)
8. Geankoplis, C.J., Transport Processes and Unit Operations,
3rd ed., Prentice Hall, Englewood, NJ (1983). Equation simi-
lar to 10.6-16
9. Yakowitz, S., and F. Szidarovszky, Introduction to Numeri-
cal Computations, Macmillan Publishing Company, New
York, NY (1989) 1

book review

CHEMICAL THERMODYNAMICS: BASIC THEORY
AND METHODS, 5th ed.
by Irving M. Klotz, Robert M. Rosenberg
Published by John Wiley and Sons, Inc., NY; 533 pages,
$54.95 (hard cover) (1994)

Reviewed by
Pablo G. Debenedetti
Princeton University

The fifth edition of Klotz and Rosenberg's Chemical Ther-
modynamics is similar in spirit to its four predecessors. It is a
text on classical thermodynamics and its applications to
mixtures, chemical reactions, and other situations of interest
to chemists. It can be used both for undergraduate and gradu-
ate instruction and requires no previous knowledge of ther-
modynamics. The simple mathematical tools needed to un-
derstand the material are explained in the book.
The twenty-three chapters cover a wide range of topics.
Following two introductory chapters on the history and ob-
jectives of classical thermodynamics and on mathematical
prolegomena, the First Law and its applications to chemical
reactions and to the behavior of gases is discussed. Three
chapters are also devoted to the Second Law, its consequences
(reversibility, spontaneity, free energy functions) and its appli-
cation to simple cases of phase equilibria (e.g, the Clapeyron
equation, temperature dependence of enthalpies of transition).
Winter 1996

Other chapters discuss the Third Law, reaction equilibria,
systems of variable composition, gas mixtures, the phase
rule, ideal solutions, dilute solutions, activities in non-elec-
trolyte solutions, the calculation of partial molar quantities
from experimental data, the determination of activities of
non-electrolytes, electrolyte solutions, free energy changes
in solutions, gravitational fields, and the estimation of ther-
modynamic quantities. The fifth edition also contains a new
chapter on simple analytical and numerical methods (least
squares regression, numerical and graphical differentiation
and integration). The above subjects are of obvious interest
to chemical engineers, but important topics such as open
systems and phase equilibria are not discussed with the
depth needed in many engineering applications.
A useful feature of the book is the presence of several
examples and problems dealing with biological systems.
Specific topics include the calorimetric study of conforma-
tional transitions in proteins, free energy and useful work in
biological systems, the dissociation of DNA, the solubility
of proteins in aqueous solution, osmotic work in biological
systems, and protein centrifugation. Several geological ex-
amples are also given, especially in the chapter on the phase
rule. These biological and geological illustrations, the ma-
jority of which can also be found in the fourth edition, add
significantly to the book's value and originality.
The book aims at training students in the use of thermody-
namics for solving practical problems. This is accomplished
very well indeed. Each chapter contains illustrative examples,
as well as a good number of problems (typically between ten
and twenty). More rigorous and satisfying discussions of the
logical structure of thermodynamics are available (e.g.,
Denbigh's Principles of Chemical Equilibrium). In Chapter
3, for example, the authors define adiabatic systems by in-
voking the notion of thermal equilibrium; however, neither
temperature nor equilibrium have been discussed at that
point. Similarly, the definition of an ideal gas as one satisfy-
ing PV=RT and, in addition, having a temperature-indepen-
dent energy is redundant. The latter condition follows from
the former, but this can only be proved by invoking entropy,
which the authors have not defined at that stage (Chapter 5).
On balance, however, the book's virtues outweigh its limi-
tations. Few texts provide the student of chemical thermody-
namics with a wider selection of exercises and examples to
assist in the development of problem-solving skills. Because
of this, Klotz and Rosenberg's book is useful not only for
chemists, but also for biologists, engineers, and geologists.
The back cover of the copy of the book that I received
from CEE for review, and that of a second copy subse-
quently sent to me by the publishers, says that this fifth
edition contains new chapters on the thermodynamics of the
electrochemical cell and on pH diagrams. This is not correct;
the book does not include such chapters. I have been assured
by the publisher that this matter will be corrected. 17

r@ classroom

A LARGE-GROUP

SENIOR DESIGN EXPERIENCE

Teaching Responsibility and Life-Long Learning

JOSEPH A. SHAEIWITZ, WALLACE B. WHITING, DARRELL VELEGOL1
West Virginia University Morgantown, WV26506-6102

he subject of this paper is a unique, two-semester,
senior design experience in which students learn to
be responsible in a team environment and to work in
an organizational structure. Under the direction of a student
chief engineer, the class works on one design project for the
entire academic year, beginning with a feasibility analysis
and ending with a detailed, preliminary design. The project
emphasizes team effort and teaches lifelong learning skills.
This year-long design project is only one facet of the
undergraduate design experience at West Virginia Univer-
sity. In the freshman year, there is an introduction to design
through a cross-disciplinary course in guided design and
decision making.'" The initial chemical engineering design
experience is the integrated sophomore and junior design
project in which students work on progressively more com-
plex versions of the same process as they advance through
the curriculum, so that they appreciate the practical applica-
tions of what they are learning."23' Another portion of the
design experience consists of individual design projects in
which students work on three design projects of increasing
complexity during the senior year (in parallel with the year-
long design project described here), in order to enhance their
individual problem-solving skills.'41
These experiences ensure that each student acquires tradi-
tional equipment design and process synthesis skills inde-
pendent of the large-group project described here. The verti-
Joseph A. Shaeiwitz received his degrees in chemical engineering from
the University of Delaware (BS in 1974) and Carnegie Mellon University
(MS in 1976, PhD in 1978). His interests are in design, design education,
and outcomes assessment in higher education.
Wallace B. Whiting is Professor of Chemical Engineering at West
Virginia University, where he has taught since 1982. He is active in
ASEE and AIChE, and his research and teaching interests range from
thermodynamics to process safety and process design.
Darrell Velegol is a National Science Foundation PhD Fellow at Camegie
Mellon University. He received his BS in chemical engineering from
West Virginia University in 1992, having served as chief engineer for the
class of 1992. His PhD research focuses on the electrophoresis of
colloidal aggregates.
' Address: Carnegie Mellon University, Pittsburgh, PA 15213
70

cal integration of design through the curriculum, the exten-
sive oral and written communication assignments accompa-
nying these experiences, and the communication assign-
ments required in the senior laboratory, create the backbone
of what we call "The Holistic Curriculum."'5' The role of
specific required and elective courses is seen as supportive
of this backbone, adding strength and breadth.
Professor Harold P. Simons introduced the year-long project
into our curriculum in 1941, and many of its details have
been presented previously.'6 It has survived the inevitable
turnovers in faculty, continually evolving and incorporating
the most current resources to define and attack the problem
assignment. The central strategies of the program, however,
have remained consistent throughout the last fifty-four years.
These strategies seek to develop in students, through coop-
erative learning, skills for life-long learning, critical think-
ing, effective communication, self-evaluation, problem solv-
ing and decision making, and leadership and team building.
Motivation for the year-long project is that the best way
for students to learn the above skills is to practice them.
Through this project, students learn:
How a large team works. They learn that each team member
can and must have different responsibilities, but that
everyone's contribution is important. They also teach each
other.
To identify what they need to know in order to solve a
problem, to develop skills for learning technical subject matter
not explicitly taught in the standard curriculum, and to develop
strategies to teach each other.
The need for, and develop the skills necessary for, extensive
library research work, including patent searches.
To communicate with professionals from industry and
government agencies-possibly their first contact with
professionals who work outside of the university setting.
About management, whether or not they are group leaders or
the chief engineer. All are involved in negotiating, communi-
Copyright ChE Division ofASEE 1996
Chemical Engineering Education

eating, planning, and evaluating. In addition, the chief engineer
and the group leaders gain leadership and management skills
and experiences not normally available in an undergraduate
experience.
In general, all students get the "big picture" view of a
chemical process design, including responsibility within an
organization, not usually encountered at the undergraduate
level. Results of our outcomes assessment plan provide some
support that students are acquiring the skills listed above.7'

DESCRIPTION AND ORGANIZATION
The year-long design project is a component of the two
senior-year, four-hour design classes. In the fall semester
there is a typical classroom component consisting of design,
economics, safety, and professional ethics. There are also
two individual design projects (called "Majors").'4"
Faculty play roles in the project (a similar concept has
been used by others'', with one faculty member assuming
the role of the "client." (At times we are fortunate enough to
have someone from outside the University assist in this
role.) The client's role may be one of a venture capitalist
looking for a profitable investment, or the client may own a
company that has an excess of raw material. The client
"hires" the student company by communicating initially with
their "vice-president," another faculty member. The group
"assigned" to the client is the senior class, under the direc-
tion of the student chief engineer.
The goal of the fall semester is to do a feasibility study and
to present alternatives to the client at the end of the semester.
(This first-semester group project counts for 20% of the
course grade.) Before the beginning of the spring semester,
the client makes a decision based on the alternatives pre-
sented in the feasibility study. In the spring semester there
are no formal classes. There is the third Major and the
second semester of the group project, with the latter being
75% of the course grade. The final product is a detailed
preliminary design, as per the client's wishes, presented
in a public forum.
The essence of the senior design is that the students, not
the faculty, are responsible for the project. The student chief
engineer is free to organize the class in any manner; the
usual result is one layer of management, with the class
divided into groups of 4-6 students, with each group under
the direction of a group leader. The chief engineer coordi-
nates and distributes tasks among the groups, and the
group leaders then assign the group members component
parts of the task. Typical group tasks include researching
a patent containing kinetic data, designing a reactor or
separation sequence, performing a HAZOP study, or ana-
lyzing process economics.
The group leaders are also responsible for ensuring that
the task is completed, ensuring that all group members con-
tribute as equally as possible, and reporting the results to the
Winter 1996

The vertical integration of design through the
curriculum, the extensive oral and written
communication assignments accompanying these
experiences, and the communication assignments
required in the senior laboratory, create the
backbone of what we call "The Holistic Curriculum."

chief engineer and/or the client. Communication is impor-
tant, between students and the client (described below) as
well as among students. Internal memoranda allow all stu-
dents to keep track of what has been done and to avoid
duplication of work.
There is usually a clear chain of command. If the chief
engineer is to be away from campus for any period of time,
someone (usually a group leader) is appointed as interim
chief engineer. The interim chief engineer should be some-
one already familiar enough with the big picture to step
in and act as chief engineer. The same situation is true
for group leaders.
If the students need help, they go to the vice-president-
not to the client. The client maintains a professional distance
in the context of this project. In fact, part of the role-playing
for the client can be to act deliberately ornery, or even
unreasonably, forcing students to learn the art of negotiation.
Since students participate in the initial definition of the project,
any change in its scope involves negotiation between the
students and the client.
It should be noted that 30-35 students appears to be a
critical number of students. When class sizes exceed this
number, we have found that the group is too large for one
chief engineer to manage. When this happens, we have two
different projects, two groups, two chief engineers, and two
clients and vice-presidents (still with two faculty members
total, each assuming a role in each project). Our enrollments
are such that we have never needed more than two groups.
When there are two groups, we make the projects very
different. One rationale for this is that the groups will not
perceive themselves to be competing with each other. They
feel they can help each other and, in fact, we have seen that
they do just that. Another important rationale is that the
groups learn something about another project in addition to
their own. Both rationales are good examples of students
teaching students.
The initial problem statement is deliberately vague. Fig-
ures 1 and 2 are examples of initial memos from the client to
the vice-president. It is up to the students to define more
specifically the goals of the project and to obtain written
approval from the client. The students must determine what
tasks are required to complete the project. They learn to
brainstorm and to develop a Gantt chart (milestone chart)"'9
such as the one shown in Figure 3. Once the Gantt chart is
approved by the client, weekly meetings are held for progress
reports. Students are expected to meet the deadlines indi-
71

cated on the Gantt chart, to justify the situation if they are
falling behind, or to justify why the schedule should be
renegotiated. Minutes are taken at the weekly meetings
and become official once approved by the client and the
chief engineer. The minutes serve as a permanent record
of agreements and commitments made by the students
and by the client.
The primary deliverables for this project include a final
oral presentation each semester for which all faculty, college
administrators, and underclassmen are invited. The final pre-
sentation in the second semester is also open to the public;
all of those individuals within and outside the University
who provided information during the year are invited, and
those who are nearby, in addition to members of the College
administration, often attend. An extensive
written report accompanies these oral pre-
sentations, and, in recent years, a poster
presentation has been available prior to OMICROI
GEORGE B. BERRY
the formal oral presentation. The students CHAIRM
also give midterm presentations each se- CHIEF ZXECUTEIVE OF
master in order to convince the client that
the class is on schedule and to provide an Dr. R. C. Bail
opportunity for each student to make a Vice President
Technocats, In
presentation. Written reports often accom- % College of E
pany the midterm presentations, although West Virginia
the number, scope, and timing of mid- P.O. Box 6102
term reports are determined by the client Morgantown, WV
for the specific project. Dear Dr. Baili
Among the students' unique experiences
are extensive library work and communi- SUBJECT: E
cation with professionals in industry and As we have
government agencies. Since the project is investment opp
ill-defined initially, students must learn omicron is
to search the various data bases for enough C-1 chemicals
information to define the project better, assumption the
world sources
Much of the information necessary to pro- would result i
duce alternatives for the client in the first Omicron pr
semester and to gather equilibrium and economic feasJ
kinetic data for the final design is in the chemicals which
patent literature. Often, information is detailed proce
available only in obscure resources not in Your propose
our library, so students must learn to use provide a gene
and describe
other libraries. Increasingly, students are second phase
finding the Internet to be an additional products ident
valuable resource. Since information on Omicron wi]
chemical processes or on regulations is Manager, who
always necessary, students must contact with any quest
companies, vendors, and regulatory agen-
cies for specific information. Here they
learn that when the information they want
is not in the public domain, most compa-
nies are very cooperative in providing the cc: Dr. Eugene
information. We are fortunate that there Figure 1. Initia
is a network of those who have previ- Berry is an alur
72

ously gone through the year-long design course-a first
contact to one of our alumni or to a company close to West
Virginia University requires no explanation of why the in-
formation is needed and usually produces the necessary data.
Student evaluation is also unique and fits the concept of
student responsibility:
The chief engineer is responsible for developing, early in the
project and with the approval of the faculty, detailed grading
criteria and for presenting them to the students. Typical
criteria are such that every student can earn an A or every
student can earn an F. Students are not competing against
each other but against an absolute scale. Past grade
distributions are comparable to those for other capstone
project courses.

I CAPITAL CORPORATION
Y

yFICER
August 15, 1991
ie
-Engineering Design
C.
engineering
University

26506-6102

e:

IGINEERING STUDY OF C-i CHEMICALS FROM METHANOL

discussed, Omicron Capital Corporation is constantly looking for
ortunities as a part of our corporate long range plan.
currently considering investment in a facility for the manufacture
from methanol. Our rational for this program is based on the
t methanol will experience a substantial price reduction as various
return to full capacity and technical advances come to bear. This
n favorable economics for various C-1 chemicals.
oposes to engage your firm to provide preliminary engineering and
ability studies leading to a suitably prioritized listing of C-1
:h would offer the greatest economic incentive and a proposal for
ss design.
al to perform this first stage should describe the work to be done,
ral statement of design criteria, scope and limitation to the work,
costs, rates and payment terms. The engagement could result in a
which would provide process development of one or more of the
ified in phase one.
li be represented in Morgantown by Dr. Richard Turton, our Project
will be fully authorized to act in our behalf. If this leaves you
ions, please do not hesitate to contact Dr. Turton or me.

Sincerely

George B. Berry

V. Cilento, Dr. Richard Turton, Dr. Wallace B. Whiting

'l memorandum for class of 1992 design project. Note: George
nnus and Omicron Capital Corporation is a real company.
Chemical Engineering Education

* At the end of each semester, the chief engineer meets with
each group member individually for a performance evalua-
tion, discusses the grade, and provides written justification
for assigning a grade. Students have the right to include
comments or a rebuttal in the written material provided to the
faculty. The evaluation process can be a negotiation, and the
chief engineer may assign a different grade than originally
intended based on a convincing argument made by the
student.
* The faculty have the right to alter any recommended grade,

up or down, by one letter.
The faculty assign a grade to the chief engineer.

Students are evaluated based on their assignments and
(especially) on their initiative to identify, develop, and ex-
ecute additional tasks that enhance the team effort. Everyone
is judged on the quality of the product produced. Therefore,
peer pressure is one mechanism ensuring each student's

participation,

Davy Jones, Inc.
P.O. Box 6102
Morgantown, WV 26506-6102
(304) 293-2111
August 24, 1993
Dr. J.A. Shaeiwitz, Vice President for Engineering
Technocats, Inc.
417 Engineering Sciences
Morgantown, WV 26506-6102

Dear Dr. Shaeiwitz:
As we discussed, Davy Jones, Inc., is prepared to engage Technocats to perform a preliminary engineering
evaluation of opportunities in the chlorine-replacement arena. Increasingly, attention has been drawn to the adverse
health and environmental effects of chlorine and chlorine-based products. While there is still considerable debate about
the extent of these effects and appropriate regulatory actions, we feel that there should be a niche for us to invest in a
plant to produce substitute chemicals that can gain considerable market penetration, regardless of the outcomes of the
debates. While others debate, we shall move forward. Thus, time is of the essence in this project; our timetable is
ambitious.
Specifically, we need for you to accomplish the following:
1. Report on affected chlorine and chlorine-based chemicals. Research the potentially affected products, paying close
attention to health, environmental, and regulatory concerns. On September 30, 1993, please submit a written report and
present your findings orally.
2. Preliminary comparison between process options. Based on the above report, we will choose a relatively small number
of options to investigate further. You will analyze these options in enough detail that we can choose one niche for
detailed evaluation. On December 2, 1993, please submit a written report and present your findings orally.
3. Design of proposed plant. You will design a plant to produce whichever chemical or chemicals that we choose based
on your December 2 report. This design should include costs accurate to approximately +40%. For the kind of high-risk,
high-return projects that we like to invest in, this level of accuracy is usually sufficient for financial commitment. On
April 21, 1994, please submit a written report and present your findings orally.
Although these tasks are broadly stated, I will be available for clarifications and modifications, as the project
progresses. Our president, Dr. R.C. Bailie, will be making two trips to Morgantown to be briefed on your progress this
fall. His first visit is on Monday, August 30, and we have scheduled the briefing for 1:00 P.M. Please bring your draft
Gantt Chart and a statement of your design team's capabilities.
Please let me know who at Technocats will be assigned to this project. I expect at least weekly briefings, preferably
on Thursday afternoons.
I look forward to meeting your design team.
Sincerely,
W.B. Whiting
Vice President for Strategic Planning
Davy Jones, Inc., is a closely held corporation that seeks to identify environmentally conscious opportunities in
the chemical process industries for investment. With capitalization of approximately $200 million, Davy Jones is
primarily a venture capital organization. After concept definition and preliminary analysis, auxiliary investment
partners are sought through preferred-stock or bond offerings.
The office of the president is in New Bern, North Carolina. However, operations are concentrated at the
Morgantown, West Virginia, office. Several of the faculty at West Virginia University serve on the Board of
Directors.
Figure 2. Initial memorandum for class of 1994 design project. Note: Davy
Jones, Inc., is a "fake" chemical company, "owned" by Emeritus Professor
Richard C. Bailie.
Winter 1996

since one student's grade is partially deter-
mined by the result of the group. The group
leaders are also evaluated based on how
effectively their group worked. The chief
engineer is evaluated based on the success
of the project as a whole and how realistic
the evaluations of others were. For ex-
ample, a chief engineer who devoted sig-
nificant amounts of time to routine calcu-
lations, distracting from coordinating the
group, would probably not receive a good
grade. Similarly, a group leader whose
group contributed less than expected would
receive a poor grade. Evaluation of the
chief engineer is usually facilitated by an
anonymous questionnaire given to all stu-
dents and seen by the faculty and the chief
engineer.
The chief engineer is also responsible
for recommending next year's chief engi-
neer. During the spring semester, the chief
engineer visits the junior class, informs
them of the role of the chief engineer, and
solicits applications. Then the class is asked
to respond to a questionnaire about each
applicant; in effect, the juniors are taking
partial responsibility for selecting their se-
nior-year chief engineer. The chief engi-
neer interviews the candidates. Part of this
interview includes the candidates' re-
sponses to management problem scenarios.
The chief engineer then ranks the candi-
dates, and the faculty choose the top-ranked
candidates) unless there is a compelling
reason not to do so. No current faculty
member can remember a case when the
top candidates) was not chosen. It is im-
portant to note that chief engineers have
come from all GPA levels.

DISCUSSION
Every student benefits from the com-
plex group interaction. Arguably, those
who benefit most are those with the lower
GPAs and students with high GPAs who
might be described as "book smart." The
73

project requirements are too extensive for a small subset of
the class to do most of the work; everyone must share the
work load equally. All students learn that their contribution
is important and develop confidence in their ability to con-
tribute to the success of the project. They learn that all
colleagues, regardless of their successes on timed tests, can
make quality contributions to the group effort. For example,
in the class of 1992, the most tenacious library researchers
did not have high GPAs. In the same year, another student
who did not have a high GPA found a fundamental flaw in
the project, missed by management, that saved the project.
There are always students who become so involved in this
project that the amount of effort they put forth could only be
described as awesome.
In addition to learning responsibility in a team structure,
the students also benefit from other singular experiences.
There are always components of the project that require
knowledge of processes not taught in the standard curricu-
lum, such as the design of a furnace, a rotary vacuum filter,
or a pressure swing adsorption bed. Students develop life-
long learning skills by using library resources (such as pat-
ents), corporate contacts, and government contacts that un-
dergraduates do not normally use. They learn valuable non-
verbal and informal communication skills (short internal
memos, e-mail) in addition to the formal oral and written
skills that pervade the curriculum and are reinforced during
the year-long project. Through role playing, students learn to
communicate with a "client." In general, students get a per-
spective of chemical engineering far beyond what can be
obtained from textbooks or more narrowly defined design
projects. It is noted that the class of 1992 visited the same
type of plant that they had designed. The equipment sizes,
separation scheme, and operating conditions were very simi-
lar to their preliminary design.
This type of project gives the instructor tremendous
flexibility. The project is extremely open-ended, as can
be seen from the memos in Figures 1 and 2. Students
learn that faculty do not have all of the answers, espe-
cially during the first semester's feasibility study. The
instructors usually have a general direction in mind c
when the project begins, but the project often goes in s
unanticipated, student-initiated directions (we encour-
age this). As one example, the class of 1990 found a
more profitable direction than the one originally antici- P
pated. Their job was to find a use for the acrolein side- P
product made during the production of acrylic acid. It
was expected that a three-carbon (like acrolein) chemi- ,
cal would be the final product, but students determined
that production of d, 1-methionine, a racemic amino acid
mixture used as a nutritional additive for farm animals,
would be far more profitable. The instructors also have
the flexibility to include as components of the final
product rapidly emerging areas of chemical engineer- Fi
ing. For example, safety (e.g., HAZOP) and environ- p
74

mental considerations were a part of all of these projects
well before formal courses were developed on these topics
and before ABET showed interest in them.
The chief engineer and group leaders also acquire special
skills. The chief engineer must learn to delegate responsibil-
ity, to lead a large group, and to manage people. He or she,
usually an organized student with a strong technical founda-
tion, is faced with a disorganized group (at least at the
beginning) and a situation in which there is no time to do
calculations. Indeed, the toughest part of the job for the chief
engineer (and for many group leaders) has been described as
"learning to let go." Learning to trust others is something top
students do not ordinarily get a chance to do. The group
leaders learn a similar lesson. They learn to divide a task into
smaller parts, to delegate responsibility, and to synthesize
the results into a coherent product. Both the chief engineer
and the group leaders gain experience in managing people,
all of whom are different and must be treated differently
without compromising the group effort.
With a group of 20-30 students, many nontechnical prob-
lems arise. In dealing with these problems, together with the
delegation of responsibility and the evaluation of personnel,
the chief engineer and the group leaders can discover if they
like and are suited for management. It is not uncommon for
group leaders to be changed (without penalty) during the
year, and once a chief engineer resigned when he found that
he did not like management.
As with any unique educational program, there are draw-
backs. The most common is dealing with the inevitable
conflicts that arise between students. In a class divided into
groups of three or four students, students can choose their
groups (or groups can be assigned) so as to avoid conflicts,
but this is not so when the whole class works as a unit.
Perceived favoritism, past disagreements, and former per-

Milestone Chart Chlorine Replacement Study
Aug22 Aug29 SqLp5 Sptl12 Sept l9 Sept126

A

timiumarnfnnun Sarh
itial Chant M going
FC Reach
rn in CFC Information
*mt Recarch
iiun MSolvnti information
pmp/Blach Research
rn in PapdoBleach lnfo
VC Re-tanr
m PVC Information
.tcid R-eseah
Sm Pcti.cid Infomation
aintc Ro dich
m in Caustic Information
5hoinm Reach
/inta T tmmnt Research
IF in Wmtle Innintinoa
trgai All Informnaion
Wr Reporn
initial Dra of Rpot
Vork on Presenation
Client Mctiiui
Orgamnational Meetings

figure 3. Example of a milestone (Gantt) chart for the initial
portion of the class of 1994 design project.
Chemical Engineering Education

A A

MM_

A
A
A

A

A
A
A A
A A

sonal relationships are among the most common conflicts
that arise. If the chief engineer cannot resolve the problem,
the faculty intercede (as faculty, not as client or vice-presi-
dent). On occasion, this has been time consuming. Since the
scope and direction of the project are so open-ended, it is
difficult to predict what problems might arise and to prepare
fully for them. As mentioned earlier, some students have
been known to put an inordinate amount of time into this
project. The negative is that, on occasion, it has been known
to affect their performance in other classes.

IMPLEMENTATION
Since every chemical engineering department has a cap-
stone design course, the obvious question is why implement
a project like the one described here? We believe that the
extremely open-ended nature of this type of project gives
students insights they do not normally get in the traditional
class or in a typical capstone design course. More impor-
tant, each student learns to work responsibly in a group
structure, and the chief engineer and group leaders gain
leadership experience.
It is not necessary to have a two-semester sequence to
implement a portion of what has been described here. For
example, the whole class section (which is defined here as
less than 30-35 students) could work on a more well-defined
project for one semester. The faculty would have to narrow
the scope of the project such that the starting point is the
beginning of our second semester. It would also be possible
to divide the class into groups of 8-12, appoint or elect chief
engineers or group leaders, and work for one semester on a
more open-ended project than those normally done by groups
of 3-4 students in a capstone design class.
From our extensive experience with this type of project,
we can also suggest some key problems to be wary of upon
implementation. Regular meetings with the client are neces-
sary. We suggest weekly meetings because they allow suffi-
cient time between meetings for significant accomplishments.
It is important that students appreciate the need to make
continuous progress, even in the face of other courses and
assignments, and regular, weekly meetings achieve this. It is
also important that all students be involved. It might appear
easy for a few students in a large group to contribute signifi-
cantly less than other group members, but this problem can
be controlled by a firm chief engineer and by peer pres-
sure-if someone does not work, other students complain.
Faculty and the chief engineer use interim evaluations to
identify these students and attempt to correct the problem.
Faculty also adjust requirements to ensure that everyone's
contribution is needed to complete the project successfully.
Continuity is also important, both within a given assign-
ment and from year to year, as is the chain-of-command
idea. Since continuous progress must be made, when a chief
engineer or group leader is absent for extended periods of
time (plant trips, visiting graduate or professional schools),
Winter 1996

someone must be responsible for making key decisions,
suggesting new directions, and maintaining project continu-
ity. From year to year, it is not necessary for new chief
engineers to "rediscover the wheel." Grading criteria, evalu-
ation forms, etc., from previous years are made available to
all chief engineers. Although each new chief develops fresh
materials, each can use departmental archives as a starting
point. Most important, since underclassmen informally ob-
serve the seniors, a standard is passed from "generation to
generation." This "culture" that develops over time is a
tremendous advantage to the continuity of the program.
A guide for chief engineers was written by one of the
authors of this paper (DV). It is available to anyone
interested in implementing a design project such as the
one described.

CONCLUSIONS
For over fifty years, a unique, two-semester, senior design
course has been the focal point of our curriculum, providing
students with experience similar to what they will later en-
counter in industry. They learn responsibility through work-
ing as part of a large team; they work on an open-ended
project, learning how to gather the information from a wide
variety of available resources. The student chief engineer
and group leaders gain management and leadership experi-
ence and can discover if they are suited for management. All
of this learning occurs in a low-risk, university environment.
From a faculty perspective, in addition to providing a unique
educational experience for students, this type of project per-
mits new material to be added to students' learning experi-
ence and it can be done more rapidly than when it is intro-
duced in a traditional classroom setting.

REFERENCES
1. Wales, C.E., A.H. Nardi, and R.A. Stager, Professional Deci-
sion Making, West Virginia University, College of Engi-
neering (1986)
2. Shaeiwitz, J.A., and R.C. Bailie, "Incorporating Design into
the Sophomore and Junior Years," Proc. of 1992 ASEE
Conf., p. 1266
3. Bailie, R.C., J.A. Shaeiwitz, and W.B. Whiting, "An Inte-
grated Design Sequence: Sophomore and Junior Years,"
Chem. Eng. Ed., 28, 52 (1994)
4. Turton, R., and R.C. Bailie, "Chemical Engineering Design:
Problem-Solving Strategy," Chem. Eng. Ed., 26, 44 (1992)
5. Shaeiwitz, J.A., W.B. Whiting, R. Turton, and R.C. Bailie,
"The Holistic Curriculum," J. Eng. Ed., 83,343 (1994)
6. Gardner, A.A., P.H. Whiting, and A.F. Galli, "From Raw
Materials to Profit: Career Role-Playing in a Senior Design
Project," paper #74c, Annual AIChE Meeting, Los Angeles,
CA (1982)
7. Shaeiwitz, J.A., "Outcomes Assessment in Higher Educa-
tion," Proc. of 1995 ASEE Conf., p. 1387
8. Woods, D.R., D.W. Lawson, C.A. Goodrow, and R.A. Romeo,
"Career Planning and Motivation Through an Imaginary
Company Format," Chem. Eng. Ed., 16, 44 (1982)
9. Dewar, J.D., "If You Don't Know Where You're Going, How
Will You Know When You Get There?" Chemtech, 19, 214
(1989) 0

classroom

FRESHMAN DESIGN COURSE

FOR CHEMICAL ENGINEERS

CAROL MCCONICA
Colorado State University Fort Collins, CO 80523
Freshman design courses are problematic because stu-
dents do not yet have the fundamental engineering
background necessary to solve real problems. Yet, for
students to be "caught" by the excitement of an engineering
career, they must experience the thrill of understanding a
problem, of using a rational approach to create a solution,
and finally, of watching the public enthusiastically receive
that solution. In chemical engineering, this becomes difficult
because we do not traditionally make "widgets." We make
processes, and frankly, flowsheets are just plain boring.
Chemical engineering in academia is quite abstract, and
yet many problem solutions benefit from the capabilities of
practical minds. One departmental goal is to intrigue and
retain students who learn and work with practical styles.
Studies have shown that women enter engineering because
of high-level mathematical skills, but leave with very low
self-esteem, in part because they have little hands-on confi-
dence."' Therefore, it is also critical to help the abstract
thinkers gain self-efficacy through hands-on experiences.
At Colorado State University (CSU), each department be-
gins its core course sequence in the freshman year. The goals
include developing a sense of belonging among the students,
familiarization with the campus facilities, and building close
personal relationships between engineering faculty and new
students. The freshman core in each department consists of a
one-semester programming course and a one-semester de-
sign course. The chemical, agricultural, and environmental
engineering students are grouped into one class of 60-85
students. There are no pre- or co-requisites, and the course is
meant to give very basic skills as a foundation for the sopho-
more fall course on mass and energy balances. Students are
allowed to substitute design course credits from another
department should they choose to change majors.

COURSE DESCRIPTION
Chemical and Bioresource Engineering 102 is a three-
credit freshman design course. The students use classical
design steps to build a lab-scale pilot plant that solves an
open-ended process design problem. The pilot plant must be

Carol McConica, a full professor at Colorado State University, earned
her MS and PhD degrees from Stanford University. Prior to joining
CSU, she spent three years developing new integrated circuit (IC)
processes for Hewlett Packard. Her research areas include waste mini-
mization during IC processing, multimedia education, and power/gen-
der issues in the workplace. She co-advises students in psychology and
counseling. On the weekends she can be found racing her Austin
Healey bug-eye Sprite with her husband, mountain biking with her son,
or rock climbing with her daughter . or kayaking, snowboarding,
windsurfing, etc.

safe to operate, require minimal lab space and machine-shop
time, cost less than $10 per student, be used by the public at
Engineering Days (E-Days), and demonstrate chemical en-
gineering principles. It is the professor's challenge to iden-
tify a problem difficult enough to be solved in at least fifteen
unique ways.
The course maximizes structure within the design process
and creativity within the laboratory. It is taught in a way that
integrates the traditional and modem approaches, as defined
by Dym.[2] Students are required to take each step in the
design process, from material characterization through pro-
totype testing. At the same time, they are required to develop
their own laboratories associated with each of these steps.
The text, Design in Agricultural Engineering (by Christianson
and Rohrback) is used as a reference when discussing the
design process. While the text content is directed more to-
ward objects than processes, the design concepts apply di-
rectly to any type of a problem. One goal of the course is to
have students learn that design results from a rational pro-
gression of thought and action.
The skills that the students must learn in this introductory
course are: methods of measurement, remedial statistics,
computer graphing packages, computer drawing packages,
word processing, computer spreadsheets, lab notebook man-
agement, engineering drawing, time management, team dy-
namics, product design, product testing, failure analysis,
project costing, mole balances, mass balances, report writ-
ing, and finally, creating a display and giving a market-
ing talk. These skills are taught in the context of the
design problem.
There is one combined two-hour lecture each week, fol-
Copyright ChE Division ofASEE 1996
Chemical Engineering Education

lowed by two official hours of bers of a team, they are

lab with sections of fifteen stu-
dents. Teaching assistants
(TAs) grade lab books and su-
pervise labs late in the semes-
ter when the pilot plants are
well defined. Laboratories are
held open on nights and week-
ends for four weeks prior to
E-Days so that students can
have a place to build their pro-
totype pilot plants. Because the
students work so many hours
prior to E-Days, held in late
April, they are not given ex-
ams and are released two
weeks early in the semester.
The course schedule in shown
in Table 1.
For the past three years, the
design problems have been
separation problems that can-
not be solved trivially. They
were designed to be like a
Disneyland ride: molecules
blown up to the scale where
they can be seen and measured
(counted, if possible). This
helps the students to visualize
mass and mole balances for
their sophomore year. In the
first year, the students were
given candy, gum, and foam
spheres. In the second year,
they were given iron, zinc,
sawdust, salt, and glass beads.
In 1994, they were given plas-
tic, glass, and polymer beads,
in addition to metal shot.
Throughout this article, these
various materials will be re-
ferred to as beads. In every
case the particle size distribu-
tion of the beads were nearly
identical. Filtering is not al-
lowed to be the sole design solut

th
pi

... for students
the excitement of a
ey must experience thd
-oblem, of using a rati4
solution, and finally,
enthusiastically re

TA]
Lecture/Lab T

Jan 17-22 Course purpose
Course outline-syllabus
Grading
Rules for notebooks/Gradir
Personality styles; Social in
Form lab groups
Assign the semester pro
Planning laboratory
Keeping a time log on p
Jan 24-28 Taking measurements: mul
Measuring physical proj
(teach micrometers, soil
Feb 1-5 Science vs engineering; dej
Continuation of physical
Feb 8-12 Using Quattro Pro/plotting
Lab in computer facility
Feb 15-19 Separation technology/the
Begin building first prot
Feb 22-26 Time management
Continue on prototype b
Mar 1-5 Engineering drawing/classr
Finalize first prototype
Mar 8-12 Product testing and evaluate
Test prototype I
Mar 22-26 Project costing/rebuild for
Plan and build prototype
Mar 29- Project planning/create crit
Build prototype II
April 5- Mole and mass balances/vi
Testing mole and mass
April 12- Marketing the product: visu
Creating visuals for E-D
Create talk for E-Days
Apr 19- Write report
Apr 22-23 Engineering Days

ion.

Initially, the class is introduced to the social type profiles
used by several companies.'31 Through the use of self-admin-
istered worksheets, students identify themselves and each
other as amiable, driver, analytical, or expressive. The assets
of each social style are taught and then teams of three are
made by combining students with different social styles.
Because studies have shown that women and minorities
routinely lose self-esteem when they are the minority mem-
Winter 1996

to be "caught" by grouped to be the ma-
jority of their team.[4"61
n engineering career, For example, a team
e thrill of understanding a might be two women
onal approach to create a and one man, or three
of watching the public women. The teams
that solution. draw on the strengths
receive that solution.
of each social style as
they progress through
their design experience.
BLE 1 Three makes a good
opics: Spring 1994 team size because no
one sits idle.
The students keep a
corporate quality team
traction styles and strengths: Teamwork lab book with carbon
copies that are turned
ject in at the end of each
lab session. Every ac-
roject tion on the team is
tiple samples, std, error analysis documented in the form
perties of materials
sieves, balances, etc.) of a lab write-up, with
grees/selecting a process a purpose, equipment
I property measurements description, procedure,
packages to present data results, discussion, and
conclusion. Lab books
design process (text) are written in ink, and
otypes all team signatures, as

build well as the date, are re-
oom examples quired on every page
and draw views on computer in order to protect the
ion/failure paretos and mean time to failure team's patent rights.
The lab books are
cost reduction graded on neatness,
e II format, effort, and
ical path method plot through end of project thought.
sa distiation When the students
sual distillation
streams, distribution coefficients meet in the first lab,
ual, verbal communication they are given pure
)ays samples of the beads
they will have to sepa-
rate and are asked to
develop labs to quan-
tify the physical prop-
erties. They are not told what or how to quantify. They think
up as many tests as possible and consult with the instructor
or TA about how to use the tools in the lab. The teams meet
in a soil's laboratory stocked with balances, micrometers,
graduated cylinders, sinks, outlets, tools, yardsticks, sieve
sets, drying ovens, and other standard equipment. The in-
structor supplements with tape, staples, cardboard, cans, fab-
ric, velcro, magnets, and extraneous items that might be
useful to a creative mind.

lowed by two official hours of

bers of a team, they are

For two weeks the students measure, record, and run sta-
tistical studies on the physical properties of the beads. They
use computers to plot particle size, weight, density, bounce,
roll, and any other distributions they may have measured.
Each student calculates a mean value and the standard devia-
tion for each property for each bead type. This is the time
when the class is taught to use the correct number of signifi-
cant digits and to estimate errors.
The students are encouraged to become unbounded in
their creativity. For example, some students measure the rate
at which each of the beads rolls through velcro when held at
different angles. Others measure the distance each bead de-
viates from its path of travel when a hair dryer is used to
blow perpendicular to the roll path. There are an endless
number of properties to be measured. The students then
make a pareto* of physical properties that shows the greatest
variation from bead to bead. They realize that this pareto is
unique to pairs of beads and that different properties can be
used in series to separate one bead type from the mixture.
After identifying the physical properties that can form the
basis of a separation, the teams are asked to brainstorm a
series of unit operations that can be linked together to make
a separation pilot plant. The rules are that the design must
accommodate between 0.5 and 1.0 liter of batch feed. There
can be no human judgement involved in the separation.
Human power is allowed if it is blindfolded. Pilot plants are
rewarded for simplicity of design, structural integrity, speed
of separation, manufacturability, quality of separation, mini-
mized cost (floor space, labor, utilities, materials, etc.), inde-
pendence of human involvement, continuous flow capabil-
ity, and ease of use.
The students spend several more weeks in the lab design-
ing and drawing views of their proposed plant. The lectures
at this time are focused on engineering drawing. The teams
are required to submit top, side, and front views of their
proposed pilot plant. They gather materials from dumpsters,
dorm rooms, home, and other free sources to build their
separation pilot plant. Because we do not have the time or
money to teach the students to work with metal and wood in
a machine shop, they are left with tape, cans, staples, and
other less durable materials. Durability becomes a relative
concept-the pilot plant only has to hold together through
Engineering Days. Students with previous shop skills are al-
lowed to use the shop at CSU's research center, but pilot plants
welded from metal do not receive higher scores than those
made from cans and tape, if both meet all other criteria equally.
The students have two weeks to build their first prototype
separation pilot plants. This is a time of maximum frustra-
tion as they realize that going from plans on paper to product
in bins is not so easy. It does not take long for reality to set
in. Peer relationships change and new respect is found for
* The pareto is a bar graph of% difference versus physical
property, from greatest to smallest.171
78

those creative hands-on students who may not excel on ex-
ams-honor students sometimes struggle to control cardboard
in a 3-D world. Competition between teams causes the stu-
dents to work hard, in and out of class, to have the first
working prototype pilot plant.
When the prototype pilot plant meets the mass balance
goals of 95% total recovery and 50% enrichment in each unit
operation, the students are asked to develop a simple statisti-
cally designed experiment to test the separation capability
over a range of compositions."8' They typically test a grid of
inlet stream compositions and count beads to determine the
composition in each "out" bin. The unit operations within the
pilot plant are drawn as a series of black boxes. The students
write in the exit stream compositions on this diagram and then
calculate the mass and mole balances on each black box and
on the pilot plant as a whole. They are required to develop a
pareto of failure modes (from most to least frequent). Ex-
amples of failure modes include: candy in the gum bin, jam-
ming in inlet funnel, or beads flying out of the pilot plant. The
team outlines a strategy for understanding and solving each of
the failure modes, in order of importance. Here they learn to
spend effort where there will be the greatest payoff.
One lab requires the teams to disassemble their pilot plants,
down to each piece of tape, in order to complete a costing of
the prototype. Sometimes this means weighing gobs of tape
and converting this to feet, knowing the mass per length. They
go to the library or to stores to find a cost value for each
element of the pilot plant and use spreadsheets to cost out the
project. The spreadsheet lists the elements in the pilot plant,
designer time, and any time that could have been charged to a
technician, such as routine testing. Each item is given a unit
value in one column, the amount used in another column, and
the associated cost in a third column. The expenses are summed,
overhead is included, and the students discover that over 99% of
the project cost is in engineering time. It is a real "light bulb"
experience to learn that their $30 pilot plant actually costs over
$5,000 when engineering time is considered. As one student
stated, "Now I know why my sunglasses cost so much."
Credit is given in proportion to the success of the team in
implementing a cost reduction, a rebuild of the first prototype
pilot plant. When they realize that the most leverage in cost
reduction comes from saving engineering time, they are ready
to learn about time management and the critical path method
of project management."9' Each team is required to develop a
pert chart for their project for the remainder of the semester,
through E-Days. They are asked to find ways to reduce the
schedule and to work in parallel.
After the second prototype pilot plant is complete, more
mass balances are performed for a range of inlet composi-
tions. The students are asked to cost out this second pilot
plant. They are then given dollar values for the pure elements
and asked to calculate a profit for each inlet composition,
based on a measured processing rate. For simplicity's sake,
Chemical Engineering Education

Three very individual and unique solutions to the same separation problem.

the feed is assumed to have no cost. A more appropriate
model would be to give the students product values based on
the purity of the streams.
The students are asked to think of their pilot plants as a
series of unit operations, each tailored to enrich one species.
They imagine their mixture as a liquid and the "out" bin for
that operation as a vapor. Equilibrium is assumed in each unit
and distribution coefficients are calculated as a function of
composition.
For E-Days, the teams are required to have a typed profes-
sional report with an abstract, an introduction, the procedure,
the results, and a conclusion. All drawings and spreadsheets,
developed over the course of the semester, are included in the
appendix. Reports are 15-25 typed pages in length. A display is
made with poster board, and a marketing talk is written. All lab
books are displayed with the final version of the separation pilot
plant. In many cases this is the second prototype, cleaned up and
painted. Judges are brought to CSU from Hewlett Packard,
Kodak, NCR, Woodward Governor, and other local companies
to judge and score all projects in the college. For the last three
years, this class has won the freshman E-Days award, finally
beating the mechanical engineers.

GRADING
Grading team projects is always difficult. For the first two
years, each student was required to keep his/her own lab book,
and exams were given. While this made grading easy, with
70% of the grade being earned individually, it created far too
high a workload for the students and graders. Now, each team
keeps one lab book, with a different person accepting respon-
sibility each week. Students earn individual grades for each
week they are the scribe. In-class computer projects, exer-
cises, and quizzes are also graded individually. The TAs take
Winter 1996

attendance and record observations on individual effort dur-
ing each lab. The students fill out evaluations of each other's
performance at the end of the course. Under this system,
40% of the grade is earned individually through homework,
lab write-ups, quizzes, attendance, and peer evaluation. The
project, display, written report, and marketing talk con-
tribute to 60% of each grade. Generally, team members
receive grades within a letter of each other. When this is
not the case, it is because one member clearly ignores his
or her responsibilities and simply does not show up in
class much of the time.

DISCUSSION
It is interesting that even with the three sphere problem
and forty students, no two pilot plants were at all alike, as
can be seen in the photographs on the preceding page. E-
Days is a perfect demonstration of the value of diversity, as
all of the different populations incorporate vastly different
tools into their solutions. These range from cross-stitch fab-
ric and blow dryers to power drills. The open-ended problem
also allows cross-cultural education. Male students, who
have never been in fabric stores, are out looking for lace of
different mesh sizes, and women students snoop around the
machine shop. One woman probed her kitchen and showed
her male colleagues how Karo syrup and corn oil separate
candy and gum spheres.
The class of 60-85 students is either team taught with two
faculty members or split into two sections. With one com-
mon two-hour lecture and four to six two-hour labs a week,
each faculty member is responsible for 5-7 contact hours per
week. During the month before E-Days, the lab is often held
open an extra fifteen hours a week. Fortunately, by this time
in the project, TAs are more than capable of sharing the
79

TABLE 2
Results of Course-Evaluation Questionnaire
Very Mostly Rarely Never Not
True True True True Appropriate

1. I am more active in class time in this class than in other classes. 12 28 11 4 0
2: 1 feet er about my partcipatifle iit cldaslsl in ot erclasse 11 32 8 4 0
I ifeel beer about m\ iacomplishmeni in itl% !lja than in the progranimming hi 2, 1 -,
4.-- .-a2A---..mh-2-. 27 .3 1 0
6. I consider the design problem open-ended with no known answer. 14 22 14 5 0
'7.'-en-c0lpa c.ipk.0f-ofabJ nfSfC --- :- 40 14 -. 0
8. This cilaa requires me to be creatie. 411 14 I .
9. I fel is go textbook ollonto tis prohle 34 13 8. 0 0 0
II, I fee! mr icn em remher are pulling Iheir ,eighi on th., probiei 149
11. I would rather eam through active ico thy than lifOn to lectures, 37 18 00 0 -
12. This represents real-world problems in the difficulty and unknown involved. 21 23 9 2 0
15. I would come to extra tutoring sessions to leam Quattro Pro. 9 19 19 7 I
17. I know Word Perfect from high school. 27 7 8 12 1
18. 1 know iore about social styles than used to; 5 30 16 3 1
19. 1 can recognize different social styles in people noA. 12 Q3b 0
20. 1 sit around and do less inthis class-that inA tectidaus. 6 13 24 12 0
21. I teel \orse about m) performance in thdi cl.i-j ithn in the programming cld,, 5 13 11 I 6
22. My grup members do not conribute t-my e&firiinjsolv thaproblem 1 12 13 29 0
23. The design problem is too eascy 1 2b 11
24. The design problem should have an answer that teprofesior posts. 0 3 14 38 0
25. 1 expect tihe professor to tell me e er) atung I rned knno, aind do lor ithi, cIj I 1 0
26. I feel totally defeated by the diffculty-of the design problem. I 17 26 I1 0
27. 1 enjoy conung to the lecture part of class. 1 24 19 II I.
28. I enjoy coming to the lab. 17 32 4 2 0
29. I would learn more by having guest speakers than lectures from the book. 20 26 8 1 0
30. know lcan use the Professor as a consultant on ay design project. 20 30 5 0 0

supervision of the teams and ensuring safety in the lab. For
the design class to be taught in this time-consuming man-
ner, there has to be a strong commitment by the department
head to the value of personal and open-ended education.
Because it has been shown that women and minorities are
positively influenced by increased interpersonal interac-
tion with faculty,'r" 12] this type of course is a superb tool for
increasing the retention of these students.
The class is very successful in meeting its goals. A
summarized course evaluation is shown in Table 2. While
it is time-consuming for both the instructor and the stu-
dents, all come away with feelings of pride and accom-
plishment. The students certainly develop a strong sense of
camaraderie and close ties to the faculty members. It is not
unusual for "D" students to become motivated by doing
something "real" and to earn an "A" grade in the process.
All students learn teamwork and an appreciation for per-
sonality types and work styles other than their own. They
learn that there are many solutions to any given problem and
the goal of engineering is more than the manipulation of
equations. Best of all, they learn that solving problems and
building "real things" can have a contagious excitement.
80

REFERENCES
1. McIlwee, Judith, and J. Gregg Robinson, Women in Engineering:
Gender, Power, and Workplace Culture, State University of New
York Press (1992)
2. Dym, Clive L., "Teaching Design to Freshmen: Style and Content,"
J. Eng. Ed., 83, 303 (1994)
3. Managing Interpersonal Relationships, Wilson Learning Corpora-
tion (1989)
4. Tannen, Deborah, You Just Don't Understand, Ballantine Books,
New York, NY (1990)
5. Kanter, R.M., A Tale of "0": On Being Different in an Organization,
Harper & Row, Publishers, New York, NY (1980)
6. Felder, R. M., G.N. Felder, M. Mauney, C.E. Hamrin, and E.J. Dietz,
"Women in Engineering: Falling Into the Gender Gap," American
Society for Engineering Education, Annual Conf. Proceedings (1994)
7. Christianson, L.L., and R.P. Rohrbach, Design in Agricultural Engi-
neering, American Society of Agricultural Engineers (1986)
8. Box, George E.O., Statistics for Experimenters: An Introduction to
Design, Data Analysis, and Model Building, Wiley, New York, NY
(1978)
9. Monks, Joseph G., Operations Management: Theory and Problems,
McGraw-Hill, New York, NY (1987)
10. Anderson, James A., "Cognitive Styles and Multicultural Popula-
tions," J. of Teacher Ed., 39(1), 2 (1988)
11. Rayman, P., and B. Brett, Pathways for Women in the Sciences,
Pathways Project, Center for Research on Women, Wellesley, MA
(1993) a
Chemical Engineering Education

tNITED STATES Statement of Ownership, Management, and Circulation
SPOSTAL SERVICE. (Requald by 39 U.S.C. 3685)

13 PubWicatcon Name 14 I.sue O.le lo Cculaton Daota Belo
CHEMICAL ENGINEERING EDUCATION Summer 1995
I Ea mad Nhare o Ck culatlon Ava No. Copa Each Ila Actual No. Caopes el i o lse I
During Ptlcedkl 12 Months Pubnlhd NHIsaalt I Flin Deta
a. ToalNo. CoasIfNetPrsnRun 2100 2100

b. PeId aWcna Rltatd cCaldUon
(1) Sitm Da and Cerni. S tat Vndo. and Cou s -0- -0-

(2) Pa da Oaeqmrtad aMpla"
ltad. oAmrc a fCamm..,g cauts, 1952.50 1990

r Tl sPaldadw aReetseda iCi an 1952.50 1990
DfSn ora tn)td t5 3'(2))
d oactia Mi 38 37
ISeml.a Carnplmnily a ll.tar 3F e

Fr lrbuln O mutad el Ma Ul (Cam rs Onoer Mean -0- -0-

I TotuFrM~O~Itriuetonumortln tas) 38 37

o ^talbiat(S ofIt "1990.50 2027

h opkl Nl t Dtlibuad
(1) OICe Urs Laovns.Spaaed 109.5 73

(2) Rarcan tman nai Aga -0- -0-

I Trct I(sco a r, ag s, and t52)) 2100 2100

I1/Ips ia 98.09% 98.17%
(t& /tbyo 8 5o 98.09
to Th~n muSmnt oms w cn p Ra primed in ih Winter 199 ,uofteubcan 0 Chkbaoiaomt ilo
1. Sigt.aa UInd Tia ol Editaor. PubttEtr. Bucus Managr. n Owner oato

t Catalt cicthsanl m1omnctton tbhadan tIh c l tti Ot and ciatemplet I tad antd eat InyOna w-o tutnsoc fale cstolaaduin tIotmataton ihn Iotnc or
)MtigcnA an tmgta d amlsan cepWMalti,

Instructions to Publishers
1. Complete and tile one copy of his lorn wth you postmaster on or before October 1. annually. Keep a copy of the completed Iorm for
yoar cotda
2. Include n iems 10 and 11, in cases where e sockholder or security holder is a trustee. h name of the peron or corporaton for whom
Oth i"nm l actng. Also include the nants and addresses of Idividuals who are slockholders who own or hold I percent or moe of Ihe
total amount o bond., monigages. o other securlis of the publishing cooratin. In Item 1, if nome, check box. Use blank ahatts f
more space sla requld.
3. Be sure to furnish als Iormation called lao In item 15. regarding circulation. Free circulation must be show n in hems 15d, a. and I.
4 If mhe publcatlon had second-clacs authorization as a general or requester publication. tnm Slalement of Ownership. Management, and
Cireuteion must be published : i must be printed in any issue m October or the first printed issue after October, if the publication is not
pubaised ditg Oclober.
5. In hem 16. indctal date oa the issue in which Ihn Statement ol Ownership will be pnnted.
6. Item 17 must be signed.
Faure to ise or publish a statement of ownership may lead to suspension of second.class authorization.

PS Font 3526, Octab IMcm (Reaasl

1. Putumcaulln T P. ubco N. a 3. Ftg Dam
CHEMICAL ENGINEERING EDUCATION I11|0 1 l- b00 0 9/18/95
4. ISm eFnquMicy 5. No. d asu Pubatashd 6. AnnualSubcripdonPri
Quarterly Ataty 4 See Attached
7c_ Correll. P-- qhpp _____

LAW so ism Cat, Coown Se uc.a

University of Florida, Gainesville, Alachua, Florida 32611-6005

i. camplrt Maling Addtrm d Headq uaMem Osral Buisineu Olrc o PulMr (f NOt Pin)
Chemical Engineering Division, American Society for Engineering Education,
11 DuPont Circle, Washington, DC 20030

o. Fulmanmesar Comta Miaing Add s a Piublhea. Edhie ad Macu Ettar (Dao n tainank mE)
rPubiler fNameiandiComtplateMactgA)ddu

ASEE-Chem. Eng. Division, 11 DuPont Circle, Washington, DC 20030

Edaor Nam h and CoryAls, MaIg Adrnesa)
Ray W.Fahien, Chem.Eng.Dept., Room 319, U of Fla., Gainesville, FL 32611-6005

Masnaga Editar (Namet and Comeitnemi end Addarea)
Carole C. Yocum, Chen.Eng.Dept., Room 317, U. of Fla., Gainesville, FL 32611-6005

omdn ay a atndartf a.cmt-otn uwnnpnadc fnt LUa u an mttdmddtn a aaa ntlt o l act dnddc .t taa cc ac
oy noanoiltardatto. ttanB aa incntnusdt> ) ( BroNolnLealiBank.)
Fu ll ame Complsle MalinOi Addrdas
Official publication of Publisher Any mail addressed to owner should go

as listed above Editor or Managing Editor listed above

11. Kon Bonaholdeta Monla geas,i cml ieOdts y Holers Owning or Hoaldg 1 IPercan or Mora ol Totl Amount of Boands, Mortgages, aS Othe
Samcsaa. it nona, cailk are, na
Ful Naama Complale aling Addmsa

12. Fm comnisn by notrvmt oi anzavtann aumodzcad Io tal 1 i raies hit purpOw lutlon, ad anoenptll status of a aocrgilzafo and the sxatt
slarus lar deal ncom ct epurpont: (Chla ) p Ha ,iNot Changed during PCrcedinm 12 Momn
0 Hal Changad Dina Praoldcig 12 Mnict e
()tcatnltIogtdpni-etmut submitelmntatenionocWang mE casaoutmmn)
PS FIlm 3526, Ocalb5 1a94a (Sact asticont on Rttvat)

POSAL BuL'rEm

57812 9-1-94 PAGE 7

PAGE 18, 9-1-94, 21875

POSTAL BULLET

MEET ONE OF THE

WORLD'S LEADING SCIENTISTS.

The way he studies a ladybug. Plays with his belly button. Stacks A-B-C blocks.
Thoughtful, skeptical, inquisitive. A born scientist.
The question is, will we let him outgrow his natural inclinations?
Only if we engage our kids with science in a hands-on, knees-on, minds-on way,
will we prompt them to become better thinkers, and eventually, lifelong learners.
As a research-based company with major businesses in health care, chemicals
and imaging technologies, Bayer supports hands-on programs that encourage
kids to question and explore. Such as raising butterflies at school in Pittsburgh. And
predicting thunderstorms in Elkhart, Indiana.
Our goal is to keep the scientist in each of us alive. And, of course, to make the
world see that some of our greatest minds are, indeed, rather small.

Bayer

	Front Cover
	Table of Contents
	University of South Carolina
	Positions available
	Deran Hanesian of the New Jersey...
	Application of pseudo-steady-state...
	Book reviews
	Computational results: How reliable...
	Applications of some modern management...
	Application of quality management...
	The warm winds of change
	Changing vapor-liquid traffic in...
	Teaching transport phenomena with...
	New books
	EPIC: The engineering program for...
	Low-cost experiments in mass transfer:...
	On selecting appropriate control...
	On using a boundary perturbation...
	Design of separation units using...
	Book reviews
	A large-group senior design experience:...
	Freshman design course for chemical...
	Back Cover

MILLISECOND	CLASS.METHOD	MESSAGE
0	sobekcm_page_globals.constructor
0	sobekcm_page_globals.constructor	Application State validated or built
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.constructor	Navigation Object created from URI query string
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.display_item	Retrieving item or group information
0	sobekcm_page_globals.get_entire_collection_hierarchy	Retrieving hierarchy information
0	sobekcm_assistant.get_entire_collection_hierarchy
0	cached_data_manager.retrieve_item_aggregation
0	cached_data_manager.retrieve_item_aggregation	Found item aggregation on local cache
0	item_aggregation_builder.get_item_aggregation	Found 'all' item aggregation in cache
0	system.web.ui.page.page_load (ufdc.page_load)
0	sobekcm_page_globals.constructor.on_page_load
0	html_echo_mainwriter.add_style_references	Adding style references to HTML
0	html_echo_mainwriter.add_text_to_page	Reading the text from the file and echoing back to the output stream
5	html_echo_mainwriter.add_text_to_page	Finished reading and writing the file


UFDC Home	<%MYSOBEK%> \| Help \| RSS