|UFDC Home||myUFDC Home | Help ||
ALL VOLUMES CITATION PDF VIEWER
STANDARD VIEW MARC VIEW
This item is only available as the following downloads:
Vol. 44, No. 2, Spring 2010 89 DEPARTMENT 90 Chemical Engineering at NC State University Lisa G. Bullard, Peter S. Fedkiw, and Francis P. ODell EDUCATOR 97 Robert C. Armstrong of MIT Kenneth A. Smith and Sarah H. Wright CLASSROOM 111 The Soccer Ball Model: A Useful Visualization Protocol for Scaling Con cepts in Continua Pedro E. Arce, Jennifer Pascal, and Cynthia Torres 134 A Simple Refraction Experiment for Probing Diffusion in Ternary Mixtures Cecil A. Coutinho, Bijith D. Mankidy, and Vinay K. Gupta 147 MetstoichTeaching Quantitative Metabolism and Energetics in Bio chemical Engineering Kelvin W.W. Wong and John P. Barford RANDOM THOUGHTS 109 The Link Between Research and Teaching 1. Does It Exist? Richard Felder LABORATOR Y 105 A Laboratory Experiment on How to Create Dimensionless Correlations Robert V. Edwards 127 Testing a Constrained MPC Controller in a Process Control Laboratory Luis Ricardez-Sandoval, Wesley Blankespoor, and Hector Budman CURRICULUM 119 Making a Chemical Process Control Course an Inductive and Deductive Learning Experience David L. Silverstein and Gifty Osei-Prempeh 140 Closing the Gap Between Process Control Theory and Practice Carlos Velzquez, Nelson Cardona-Martnez, and Edwin Velzquez 157 The Microbial Fuel Cell as an Education Tool Alim Dewan, Bernard Van Wie, Haluk Beyenal, and Zbigniew Lewandowski 166 Industrial Scale Synthesis of Carbon Nanotubes Via Fluidized Bed Chemical Vapor Deposition: A Senior Design Project York R. Smith, Alan Fuchs, and M. Meyyappan TEACHING TIP 172 Skits, Stockings, and Senioritis Ale: Creative Chemical Engineers Lisa G. Bullard Chemical Engineering Education Volume 44 Number 2 Spring 2010 CHEMICAL ENGINEERING EDUCATION (ISSN 0009-2479) is published quarterly by the Chemical Engineering Division, American S ociety for Engineering Education, and is edited at the University of Florida. Co r respondence regarding editorial matter, circulation, and changes of address should be sent to CEE, Chemical Engineering Department, University of Florida, Gainesville, FL 32611-6005. Copyright 2010 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 120 days of pu b lication. Write for information on subscription costs and for back copy costs and availability. PO S TMA S TER: Send address changes to Chemical Engineering Education, Chemical Engineering Department., University of Florida, PUBLICATIONS BOARD EDITORIAL AND BUSINESS ADDRESS: Chemical Engineering Education Department of Chemical Engineering University of Florida Gainesville, FL 32611PHONE and FAX : 352-392-0861 EDITOR Tim Anderson ASSOCIATE EDITOR Phillip C. Wankat MANAGING EDITOR L ynn Heasley PROBLEM EDITOR Daina Briedis, Michigan State LEARNING IN INDUSTRY EDITOR William J. Koros, Georgia Institute of Technology CHAIRMAN John P. OConnell University of Virginia VICE CHAIRMAN C. Stewart Slater Rowan University MEMBERS Lisa Bullard North Carolina State Jennifer Curtis University of Florida Rob Davis University of Colorado Pablo Debenedetti Princeton University Dianne Dorland Rowan Stephanie Farrell Rowan University Jim Henry University of Tennessee, Chattanooga Jason Keith Michigan Technological University Suzanne Kresta University of Alberta Steve LeBlanc University of Toledo Ron Miller Colorado School of Mines Lorenzo Saliceti University of Puerto Rico Stan Sandler University of Delaware Margot Vigeant Bucknell University
Chemical Engineering Education 90 ChE department Much has changed in the 30 years since CEE last published an article, in 1979, on the chemical engineering program at North Carolina State Uni versity. That is an understatementalmost everything has changed: our name, our location, our faculty, the advent of online graduate distance learning, and departmental diversity, to name a few. the broader impact of life sciences on our discipline. The same year we moved from Riddick Engineering Laboratories, a building that we occupied on the NC State main campus for 55 years, to a showcase facility on the universitys new and rapidly growing Centennial Campus. We have expanded our Ph.D. graduate research program and have developed an internationally subscribed online M.S. degree program that is a model for such activity. And compared to the old days, we have become more diverse in faculty and students. What has remained unchanged over these 30 years is the quality of the education that our faculty provides undergraduate and graduate students. THE UNIVERSITY, THE COLLEGE, AND THE DEP ARTMENTTo many folks, the mention of North Carolina brings to mind the four seasons as experienced from the scenic Outer Banks in the east to the Appalachian Mountains in the west; basketball traditions among NC State, Duke, and UNCChapel Hill. The state is also increasingly well known for its technology focus, however, as evidenced by the renown of the Research Triangle Park; the states investment in edu cation; its position as a growing hub of biotechnology and biomanufacturing (currently more than 500 companies); and the international recognition that the NC State Centennial Campus has gained as a model for the promotion of university and industrial interactions and cooperation. The states busi ness climate has been ranked the best in the nation by Site Selection the last nine years. Chemical Engineering at... NC State UniversityLISA G. BULLARD, PETER S. FEDKIW, AND FRANCIS P. ODELL Copyright ChE Division of ASEE 2010
Vol. 44, No. 2, Spring 2010 91 North Carolina State University is located in Raleigh, NC, the state capital. NC State is the largest university of the 16 campuses in the University of North Carolina system, and it receives more applications than any of the others. The university has a current enrollment of almost 25,300 under graduate and 8,600 graduate students. The academic campus has more than doubled in size since 1979 with the opening of our Centennial Campus in 1985which at the time consisted of 805 wooded acres within walking distance of the main campus. Centennial Campus, named to celebrate NC States founding in 1887, is a unique environment for a university and an extraordinary success storya research park and partners close proximity to world-class research facilities and a highly educated workforce, all in an amenity-rich environ ment (
Chemical Engineering Education 92 F ACULTY Our faculty members and the universities of their doctorate degrees are listed under the group photo, facing page. The interdisciplinary aspects of our research and teaching bers from other NC State colleges (Flickinger, of Agriculture and Life Sciences, and Pourdeyhimi, of Textiles) with associ ate faculty member status, as well as the appointment of the status at two institutions (DeSimone, joint with Chemistry UNC Chapel Hill). In addition, we are pleased to include a number of adjunct and research faculty who are listed at our department Web site along with the staff who underpin our enterprise (
Vol. 44, No. 2, Spring 2010 93 Kenneth Beatty Michigan Emeritus Chase Beisel Cal Tech (starting 2011) Lisa Bullard Carnegie Mellon Director of Undergraduate Studies Ruben Carbonell Princeton Joseph DeSimone VA Tech Michael Dickey UT Austin Peter Fedkiw Cal Berkeley Department Head Richard Felder Princeton Emeritus Michael Flickinger Wisconsin Jan Genzer Pennsylvania Associate Department Head Christine Grant Georgia Tech Associate Dean of Faculty Development and Special Initiatives, College of Engineering Keith Gubbins Kings College, University of London Carol Hall Stony Brook Jason Haugh MIT Wesley Henderson Minnesota Harold Hopfenberg MIT Emeritus Robert Kelly NC State Saad Khan MIT Director of Graduate Studies Henry Lamb Delaware P.K. Lim Illinois David Ollis Stanford Gregory Parsons NC State Steven Peretti Cal Tech Behnam Pourdeyhimi Leeds Balaji Rao MIT Gregory Reeves Princeton George Roberts MIT Emeritus Richard Spontak Cal Berkeley Orlin Velev Phillip Westmoreland MIT Hubert Winston NC State Emeritus Faculty members of CBE at 2009 retreat (not listed in order of appearance).
Chemical Engineering Education 94 dental school, business school, or pharmacy school. Selected achievements of our graduates include: presidents and CEOs of Fortune 500 corporations; company owners and entrepre neurs; membership in the National Academy of Engineering; military leaders, including a four-star general; leadership positions in academia; renowned M.D.s and surgeons; and court judges, among others.STUDENT ORGANIZA TIONSOur AIChE student chapter was the 15th chapter established place continuously since 1930. It has won a national student chapter of the year award 14 times in the past 20 years. Our students have been instrumental in establishing student chap ters of the International Society of Pharmaceutical Engineers in 1996 and Omega Chi Epsilon in 1997, and many are ac tive in the Society of Women Engineers, National Society of Black Engineers, Society of Hispanic Professional Engineers, and Engineers Without Borders. Our graduate students have had a Graduate Student Association (GSA) chapter since the Francis ODell now the departments director of development). The GSA plays an active role in the social fabric of the department and graduate students.GRADUA TE PROGRAM We offer two thesis degrees: the Doctor of Philosophy (Ph.D.) and the Master of Science (M.S.). The emphasis, however, is on the Ph.D. program, with 95% of our gradu ate students being part of it. A unique aspect of our doctoral program is the replacement of the traditional written qualify ing examination with a research propositiona change that we implemented 18 years ago and was described by Ollis elsewhere ( CEE 222-29, Fall 1995). The research proposi tion has evolved into a pair of research-related courses taken tion to Research during the Fall semester and the Research Proposition course in the Spring. We offer these two novel courses to better prepare students to function at high levels of productivity in both their Ph.D. studies and careers, and to teach the mechanics of effectively communicating to a tech nical audience through proposals (hypotheses) and papers (results). In the Introduction to Research course, the student independently develops an original chemical engineering research paper and defends it orally to the course instructor. The class also covers issues related to research ethics. In the follow-on Research Proposition course, the student creates an NSF-format-like proposal, typically in the area of his/her anticipated Ph.D. study, under the guidance of the course instructor and thesis advisor, and defends it before a faculty committee. Successful completion of these two courses and continue for the Ph.D. Creativity, collaboration, and innovation are characteristics emphasized and nurtured in our research programs, which encompass a variety of aspects of chemical and biomolecular engineering. Fundamental studies include investigating na noscale phenomena, reaction pathways in living cells, protein applied topics include fabricating functional nanostructures developing alternative sources of clean energy, to name a few. Following are the present areas of research strength in the department and the faculty members involved with them: Biofuels and Renewable Energy Technology: Dickey, Fedkiw, Henderson, Khan, Lamb, Parsons, Peretti, and Westmoreland Biomolecular Engineering and Biotechnology: Beisel, DeSimone, Carbonell, Hall, Haugh, Kelly, Rao, and Reeves Catalysis, Combustion, Kinetics and Electrochemical Engineering: Fedkiw, Lamb, and Westmoreland Computational Nanoscience and Biology: Gubbins, Hall, and Westmoreland Environmental Studies: Grant, Ollis, and Peretti Nanoscience and Nano-Engineering: DeSimone, Dickey, Genzer, Khan, Parsons, and Velev Polymers: DeSimone, Dickey, Genzer, Hall, Khan, and Spontak Figure 3 shows the initial placement of our Ph.D. graduates over the last 10 years. The majority of them continued their studies in post-doctoral research positions at other institu Figure 2. The initial placement of our undergraduates over the last 10 years.
Vol. 44, No. 2, Spring 2010 95 tions. Over the 17-year period from 1953-1980, 17 of our Ph.D. students went into academia, while 31 have done so from 1981 through 2009. The department was a pioneer in establishing a distanceeducation-based Master of Science in Chemical Engineering program, which originally began with videotapes that were distributed through the mail and has evolved to delivery of content through online streaming video. Students in the pro gram come from all over the United States and abroad. The distance program provides an opportunity for individuals in the work force to complete their studies while maintaining full employment, and also for non-chemical engineers to train themselves to be chemical engineers. This non-thesis M.S. in chemical engineering is a 30-credit-hour program that offers Web access to all core graduate classes and most CBE graduate electives. To earn the degree students must take at least 10 three-hour courses with at least seven being in chemical engineering. BIOMANUF ACTURING TRAINING AND EDUCA TION CENTERThe Golden LEAF Biomanufacturing Training and Educa tion Center (BTEC) opened on Centennial Campus in 2007. BTEC simulates a biomanufacturing facility capable of pro ducing sterile bulk biopharmaceutical compounds, and the building includes classrooms, laboratories, and high-purity building and process utilities. The facility delivers a handson educational experience for all levels of post-secondary students and does so using large-scale, state-of-the-art equip ment and process systems in a cGMP environment. This only a few in the world. BTEC was founded under the leader ship of our former department head, Peter Kilpatrick (now dean of engineering at Notre Dame), and its present director is also a former CBE department head, Ruben Carbonell. Our undergraduate students in the biomanufacturing con centration take BTEC classes and graduate students perform research under the guidance of CBE faculty who participate in BTECs mission.OTHER F ACULTY INTERESTSWhile our faculty members maintain very high standards in both their teaching and their research, they also enjoy diverse personal interests and all contribute immensely to creating a collegial environment. Since retiring as the ace pitcher on the graduate student softball team, Peter Fedkiw has become an avid fan of NHL hockey and is a season ticket holder of Raleighs Carolina Hurricanes (as well as the Broadway South stage series). Sporting his Lord of the Rings jacket, Rich Spontak is our entered in the U.S. Congressional Record) and likes to coach teenagers in Odyssey of the Mind and Math Counts, as well as playing chess and squash. Carol Hall is a ballroom-dance enthusiast and participates in regional competitions. In the summer, she can often be seen paddling her yellow kayak on local lakes. Christine Grant treasures the time that she spends working to support womens ministries at her church. One of her dreams is to work on the set of a movie as a specialeffects engineer. Orlin Velev relaxes by traveling with his family and often schedules intense programs of sightseeing and cultural events on such trips. At home Orlin enjoys watching movies; hes keen on documentaries on history, especially those featuring events of the Cold War. Several faculty members are musi cians: Jason Haugh is a capable guitarist/multi-instrumentalist and songwriter; Chase Beisel is a drummer whos played in diverse musical groups ranging from concert bands to a col lege drumline; Rich Felder plays classical guitar; Jan Genzer played piano for about a decade as a youngster (about a tion in teasing the keyboard; Steve Peretti is a Gilbert and Savoyards; and Lisa Bullard sings in the praise team at her church. Hubert Winston practices tai chi and qi gongenergy techniques based in Chinese medicineand has also served as a mediator in local small claims court. Jan Genzer, Jason Haugh, and Chase Beisel are unabashed beer enthusiasts (and quite picky about what they drink). When not trying to keep up with his four teenagers and their activities, Henry Lamb still plays basketball a few times a week, goes horseback riding, and tends to his various critters (horses, dogs, and cats). He enjoys camping and hikingespecially in the NC mountains and the western United States. P.K. Lim has Figure 3. The initial placement of our Ph.D. graduates over the last 10 years.
Chemical Engineering Education 96also made recent visits to the western United States and is unafraid to set off on cross-country road trips with his three small children in tow. Jan Genzer is a talented mimic who can wittily impersonate graduate students and faculty members. Bala Rao reads books on economics in his spare time, while Dave Ollis is the man to see if you ever want to learn about the history of porcelain or re-discover ancient technology of China. Dave is also often accompanied by Teddy, a lovable pound hound who whiles away most days sleeping on the (and getting much attention from just about any visitor who with his son, Game 4 of the World Series in Philadelphia against the NY Yankees. Joe enjoys playing tennis and riding bikes on Holden Beach with his family. Greg Parsons coaches his daughters soccer team, and Bob Kelly coached his sons baseball team. Bob is also a big Mets and Giants fan. Saad Khan whips up a mean curry (besides picking up occasional takeout from Burger King). Wesley Henderson regales the students with tales from his days as an Airborne Ranger parachuting out of planes in was an extra in the movie Outbreak. Michael Dickey is a rabid Wolfpack sports fan who has been known to chase down football or basketball coaches to get a photo taken with them. This phenomenon has been witnessed by several necessary. Greg Reeves likes to swim and play water polo, and is a big fan of Atlanta Braves baseball as well as col lege football. Greg also likes to ponder the big questions in life and is an active participant in his church ministries. Hal Hopfenberg, in addition to being a famous local gourmet cook and our former department head, has the distinction of having been athletic director of NC State. When he was appointed, the newspapers winked that his previous athletic experience had been on the basketball team at MIT, but in the Michael Flickinger is an amateur architect and a lifelong house carpenter, hand woodworker, and furniture designer who donates his carpentry time providing safe, warm, dry, and affordable housing locally and in Appalachia. When hes not gardening with his wife, or working in his shop, you can but his heritage chicken breeds) which provides his family, students and colleagues with fresh eggs every week. Keith Gubbins, besides being known as a snappy dresser, enjoys bird watching, swimming laps, and boating. Keith owns an extensive art collection. Ruben Carbonell recently obtained his Captains license from the U.S. Coast Guard (100-ton vessels, 100 miles offshore) and has charter cruises planned to the West Coast of Florida (Sanibel, Useppa, Marco Islands), the San Juan Islands in the state of Washington, and the Turks and Caicos in the Caribbean. Ruben likes to dance to Cuban music and cook Cuban food, and he is the only faculty member who has had three children graduate from NC State, one from the departmentluckily for both, Ruben never and museums. For proof, check out Problem 7-7 in Chemical Reactions and Chemical Reactors his recent book published by Wiley. George is also the local faculty expert on American movie classics and is infamous for telling Pat and Mike jokes at departmental events with a rollicking Irish accent. Rich Felder enjoys listening to music (mostly classical and opera, some jazz and folk and bluegrass and s and s rock); tacking vacations onto the international trips he takes to give teaching workshops with his wife Rebecca, doing the Sunday New York Times crossword, and above all, playing with his seven grandkids. Newcomer Phil Westmoreland has found the Triangle to be an amazingly hip part of the coun trya blend of grassroots and the very sophisticatedand enjoys cuisine from barbecue to vegan, music from Piedmont Blues to the Cats Cradle, crafts and art, and hiking trails in the Eno River and Umstead State Parks. * We hope this description gives an idea of who we are and what we do. We also hope that it conveys the great collegiality and friendship we enjoy in our department. Not only do we have fun, we also work together well, as evidenced by nearly 40% of our graduate students being co-advised by more than current faculty who received their ChE degrees at NC State eventually returned to the department to teach and carry out our students and graduates.
Vol. 44, No. 2, Spring 2010 97 ChE educator Yesterdays good answer is todays great question. Thats what Bob Armstrong of MIT tells his ChemE mechanics or a global problem in carbon emissions, answers are the best place to start asking, What were their methods? How can we account for the unexpected? Where can we go from here? The Chevron professor of chemical engineering at MIT phenomena for more than 30 years. In 2009 he was elected a member of the Board of Directors of the American Institute of Chemical Engineers (AIChE), and he has held a leadership role in addressing the worlds energy problems, serving as deputy director of the MIT Energy Initiative (MITEI) since 2006. A native of Baton Rouge, Louisiana, Bob grew up immersed in the new technologies that were responding to Americas needs for materials and energy. Innovations in the chemical, petroleum, shipping, and other industries were burgeoning. With its port on the Mississippi and its relative safety from hurricanes and floods, the states capital city embodied Americas post-war boom. Robert C. Armstrong of MIT Robert C. Armstrong at MITs Chemical Engineering Department. Copyright ChE Division of ASEE 2010 BY KENNETH A. SMITH, Massachusetts Institute of TechnologyAND SARAH H. WRIGHTphoto/Webb Chappell
Chemical Engineering Education 98 Esso (now Exxon) and Shell had great jobs for chemical engi neersBobs father, Wallace, worked for Ethyl Corporation, a major chemical company, throughout his career. Cars, malls, and homes hummed with air conditioning. In 1960, when Bob was 12, crude oil cost $3.00 a barrel. Gasoline was 31 cents a gallon. I took energy for granted for most of my life, Bob says. The energy crisis Bob now works to resolve took root in 1960, too. In a move that attracted little attention at the time, Iran, Iraq, Kuwait, Saudi Arabia, and Venezuela formed the Organization of Petroleum Exporting Countries (OPEC). American energy consump tion and dependence on foreign oil would affect global industries, economies, politics, and energy supplies in years to come. Like most Americans, Bobs personal life was more affected by events in the United States than by OPEC or the Middle East. The Vietnam War coincided with his high school and college yearsin the 1969 draft lottery, Bobs number was 308. His family, active in church and civic life, were known in Baton Rouge as inclusive and pro-integration. His parents, Wallace and Eileen, received wide local recognition for their community service. neers and growing industries. My family was very service-oriented. I look on what I do today as service, as giving back to society, Bob says. A gifted and serious student, Bob focused closely on school. In summertime, he played tennis and golf and waterand shoe-skied on Louisianas Amite River and on Lake Maurepas, near Baton Rouge. You had to hold those short skis just so, he says. I got pretty good at it. When he headed off to Georgia Tech in 1966, he already knew he would major in chemical engineering. It was a great way to combine the subjects I liked bestmath, physics, and chemistry, he says. But dont let Bobs modesty and studious demeanor mislead you. He has a lively wit and love of adventure, warns Debbie Armstrong, Bobs wife of 40 years and also a Baton Rouge native. Bob found out he loved math when he got sent to detention and met the math teacher. He and a friend once blew up a shack in a bad chemistry experiment, she recalls. When we were in high school, he was a great lab partner. At the same time, he was a drummer in a rock groupThe Capersthat featured him playing solo drums GEORGIA TECHWhen he got to Atlanta, Bob was pleasantly surprised to discover a familial aspect of the chemical engineering community: Tech ChemE professor Jesse Mason had taught Bobs father at the Uni versity of Florida. Thats one of many connections I had with my father, Bob says. And I had a similar experience at MIT: I taught the daughter of two of my former ChemE students. Characteristically, Bob devoted his college years to academics. He kept up his musical life, too, playing snare drum for the Geor gia Tech Marching Bands half-time shows at football games. That was big-time sportsuniforms, full sta diums, TV coverage. The Yellow Jackets even went to the Orange Bowl, he says. With typical understatement, Bob says he did alright at Georgia Techmeaning, he earned As in all his courses; passed the grueling Tech swimming test with its bizarre requirement that students swim hog-tied for one hour; and drummed for the band. He did get one F, in track. I had no endurance for distances, he reports. He spent his college summers working in the energy industry, working for the Gulf South Research Insti tute; Kaiser Aluminum; Ethyl Corporation, now the Albemarle Group; and at Esso Production Research (now Exxon) in Houston, where he studied bottom-hole samples to discern how much oil was left in the ground after initial production. His other experiences with oil production were more for red snapper off the Gulf Coast, boating from Grand Isle or Empire out to offshore drilling platforms. It was great. Fish just love those platforms, he says. THE ICY NORTHBob and Debbie married in August 1969. A year later, they packed their blue 1970 Chevrolet Monte Carlo, a graduation gift from Bobs parents, loaded their U-Haul trailer, and drove from Houston to the University of Wisconsin, Madison. Their destination, renowned as an epicenter of the student anti-war and environmental movements, would have felt far more foreign had not a catastrophe quieted everything down, Bob recalls. The Capers rockin drummer.
Vol. 44, No. 2, Spring 2010 99 Protestors had bombed the Army Math Research Cen ter, killing one and injuring others, two days before we arrived. Marches and demonstrations had stopped. It wasnt so different from Baton Rouge. to love Madisons frozen lakesonce they learned how to stay warm. What do Louisianans know about 27 degrees outside. The paint had literally fallen off our car, Bob says. According to Professor Emeritus Robert B. Bird of Wisconsins Chemical and Biological Engineering De partment, Bobs graduate career got off to a slow start. In Birds course on applied mathematics in chemical engineering, the new student looked rather sleepy. But I soon learned that I could depend on him to come up with correct results. In 1968, Bird had shifted the aims of his research program from a purely continuum approach to poly Armstrong and Ole Hassager, now ChemE professor at the Technical University of Denmark, were among Bob completed his Ph.D. thesis, Obtaining Constitutive Equations from Molecular Models in 1973. Debbie typed it on a Smith-Co rona typewriter while Bob hand-drew the equations. Just before that was done, he and Hassager proposed an astonishing project, Bird recalls. They told me, Wed like to write a book with you. I had a set of very rough notes for my classes, and they wanted to volunteer to turn the notes into a book. It would involve an immense amount of work, but they did not appear to be dismayed, he says. Bird knew whereof he spoke. He is co-author of the 780-page Transport Phenomena (1960) and the 1,200-page Molecular Theory of Gases and Liquids (1954), known around ChemE as the Green Monster. Undaunted, the three roughed out an outline for Dynamics of Poly meric Liquids which was published in 1977. DPLs second edition appeared in 1987 and was named a Citation Classic in 1988 in recognition of how frequently it is cited in papers and books. It remains in print Book-writing is a very severe test of personal relations. There were heated discussions and technical disagreements. Bob was invariably pleasant and easy to work with, says his advisor-turnedcolleague. Bob and Debbies wedding, August 1969. Fishing with sons David and Eric.
Chemical Engineering Education 100 ON TO MITIn 1973, Bob accepted an assistant professorship in chemi year chemical engineering program had been founded. He and Debbie, now a clinical social worker, moved to the land of the Red Sox and the New England Patriots, settling in Lexington, Mass., a community known for its green space, good schools, and youth sports programs. Their son David was born in 1976; son Eric, in 1979. Over the next 10 years, Bobs growing leadership skills, his love of sports, and his familys example of community service would take him out to the ball game as a Little League coach and, later, president of the local league. varsity team in high school and in Division One baseball at Lafayette College in Pennsylvania. His parents were in the stands for many of his AAU and college games. Now a U.S. Marine Corps pilot, Eric was deployed in Iraq in summer 2008, and he headed out with the 24th Marine Expeditionary Unit at the beginning of 2010. Eric loves adventure and the outdoors. His patriotism led him to enlist after 9/11, Debbie says. According to Debbie, teaching and research in chemical engineering sometimes inspired Bobs parenting in funny, creative ways. When the boys were little, Bob would talk to them about how gunky and gooey materials worked. Hed illustrate by sculpting with peanut butter and squirting toothpaste out of the tube, she says. called Stretch Armstrong. His superpower was super-stretch In conversation, Bob readily cites ordinary household productsshampoo, ketchup, mayonnaiseas examples of a glass of water, and a little polymerto show how a nonSometimes I connect what the students are seeing to his tory, Bob says of his ChemE lecture style. I tell them, this isnt new: The Roman historian Tacitus described the same mechanics. When they harvested bitumen, a tarry substance, they found it would siphon itself into a boatvery different than water. A GIFT FOR TEACHINGKnowledgeable and enthusiastic, Bob shone as a new Outstanding Faculty Award in 1976 and the DuPont Young Faculty Award for 1974-75 and 1975-76. Former student Zubair Anwar (MIT Ph.D. 2008) is an associate at McKinsey & Co. In his graduate classes on Bob often stressed that most analysis was a departure from known analytic solutions. He encouraged us to propose an answer and test its validity with fundamental equations, Anwar recalls. Antony Beris (MIT Ph.D. 1985), professor of chemical engineering at the University of Delaware, says Bob inspired us to give our best. More than inspiring students to do their best, Bob has sometimes inspired students to do things theyd never con sideredlike include a trip to a balloon festival on a research consultation visit. Beris had gone to meet with Bob, who was working at Exxon Corporate Research in New Jersey. He included me with his family at the festival. He was very cordial and open an academic career on my own. Promoted to full professor of chemical engineering in 1988, Bob continued to publish research in the journals of Rheology and of Non-Newtonian Fluid Mechanics (among others), Bob and Debbie, 2009.
Vol. 44, No. 2, Spring 2010 101 teach, serve as a respected and demanding thesis advisor and show up for Little League games. Micah Green (MIT Ph.D. 2007) guesses its Bobs sense of humor that helps him balance it all. He has a corny joke chambered for any occasion. Once he told a huge audience of professors and alumni a groan-worthy joke about a duck buying Chapstick at a store and telling the cashier to put it on his bill, says Green, now assistant ChemE professor at Texas Tech. Susan Muller (MIT Ph.D. 1986) is professor of chemical engineering and associate dean, Graduate Division, at the University of California, Berkeley. Bob, one of her thesis advisors, was a great classroom instructor. As a research advisor, he was incredibly patient and generous with his time. Im not sure I am as patient with my students, but I think of that experience as a touchstone, she says. In 1992, Bob won the prestigious Professional Progress Award for Outstanding Progress in Chemical Engineering, of chemical engineering through discovery, invention, or Outside of ChemE, Bobs spirit of discovery is replenished drilling platforms, by time and activity outdoors, preferably wilderness areas. I dream of hiking, he says. Debbie and I have climbed half of Mount Desert Islands 26 mountains. Beautiful scenery, vistas of ocean and other mountains, followed by popovers and teathats hardly roughing it. Its my favorite place. One of Bobs memorable outdoor adventures took place in 1993, when he and his son David joined Bob Bird and John Wiest (currently Associate Dean of Engineering at the University of Alabama) for a two-week canoe trip in Quetico Provincial Park in western Ontario, Canada. We had a great trip, but the weather didnt co operate, Bird says. Once, it poured portages. Bob held up quite well as an outdoorsman. David held up best of all, his father and Bird recall. Our worst daya mile-long portage in the rainwas his favorite, Bob says. He liked that strenuous challenge. Later, Bob tried to return his advisors favor by taking him on a climbing trip to New Hampshires Mount Monadnock. Sleet and snow above the tree line defeated them before they reached the summit; both men describe the cold, thwarted hike as interesting, like an experiment that goes awry yet opens vistas for new work. You might say Bob hikes as he teaches. Hes well prepared, enthusiastic, a good map in handand open to surprises. FROM TEACHER TO LEADERIn 1996, Bob was named head of MITs department of chemical engineering; he served in that role until 2007. As department head, Bob focused on, well, everythingteaching, research, and leadership. Yet he maintained his naturalists ability to take in a wide landscapethe big picturewhile managing challenges and change. In the late 1990s, exciting developments in materials sci an academic leader. How could the department engage in these new disciplines yet retain its focus? More broadly, its professional identity? These new questions called for both patience and innova tion. Bobs leadership of the department included hiring 11 MIT ChemE faculty and introducing a new undergraduate pro gram, Course 10-B, Chemical-Biological Engineering. This has proven to be a magnet for undergraduates: just over half of the departments undergraduates are now enrolled in it. Bobs vision and commitment to ChemE education ex when he became department head. Little League Coach Bob, with clipboard.
Chemical Engineering Education 102 Active in discussions about the professions identity and how it should evolve, he led one study on the subject that was funded by the NSF, and he was prominent in AIChE delibera tions on the topic. In 2006, these culminated in his publishing, as sole author, A Vision of the Chemical Engineering Cur riculum of the Future in Chemical Engineering Education and a chapter by the same name in Chemical Engineering: Visions of the World which was published in both English and Chinese. took him to universities in Britain and Canada and across the United States; to industries including General Electric, Eastman Kodak, and Dupont; and to national conferences of the AIChE and the Society of Rheology. OLD SCHOOL TIES, NEW GLOBAL T ASKSBack in Georgia, the Rambling Wrecks kept track as Bobs career advanced, naming him to the Georgia Tech Academy of Distinguished Engineering Alumni in 1996. It had been 30 years since he arrived at the Atlanta cam pus, eager to get down to work. He enjoyed success as a professor, researcher, administrator, husband, and father. College-sports-wise, he had switched loyalties: The Wrecks former Marching Band drummer now roots for LSU base ball and football teams, along with his Baton Rouge family and friends. Plus, hed erased that pesky F in track. Right after I de fended my thesis, I took up jogging and long-distance running as a hobby. Five or six miles became easy, he says. week, doing his own landscaping and gardening at home, and walking with Josie, the familys black Lab. (They also have a cat, MidKnight.) In 1999, the drum set came out once more, for a rock and roll faculty skit at a department Christmas party, according to Barry Johnston, senior ChemE lecturer at MIT. Bob recalls the performance as a good chance for students to laugh. Outside academic departments and beyond national borders, the global energy crisis cast a pall over many Y2K galas. By 2000, world leaders and scientists were hotly debating the gravity of climate change and energy needs. In 1997, the Kyoto Protocol set binding targets for reducing greenhouse gas emissions, to be implemented between 2008 and 2012. In 2005, as the Protocol went into effect, ChemE professor Bob Brown, then-MIT Provost and now Boston University President, announced the formation of MITs Energy Research Council (ERC). Bob and Ernest Moniz, former undersecretary for the U.S. Department of Energy and an MIT professor of physics, co-directed the ERC. They were charged with crafting MITs overall energy initiative and coming up with a structure to sustain it. Bob took up that challenge as he has so many others me thodically and thoroughlywhile pursuing his teaching, ad vising, and research. Whether approaching a research problem or roughing it in the rain, he sticks with the hallmarks of a careful theoretician. Establish the basic principles, clearly explain your methods, then make systematic advances towards your goal, says Bobs former student, Gareth McKinley (MIT Ph.D. 2007). Old school, some would call it. Thesis advisor and fellow outdoorsman Bob Bird of the University of Wisconsin, Madison, left, with Bob.
Vol. 44, No. 2, Spring 2010 103 Ever enthusiastic about research and innovation in ChemE and energy studies, Bob acknowledges he does rely on one old-school technologythe blackboard. Once I worked all night to prepare PowerPoint slides, only to forget to bring them to class. The students were so relieved I used the ing, he says. THE BINGHAM MEDALISTIn 2006, Bobs colleagues and peers presented him with the prestigious Bingham Medal. Sponsored by the Society of Rheology, the Bingham Medal recognizes outstanding Bobs published research provides a record of those contri butions. He has co-authored more than 70 papers with Brown in a particularly productive collaboration. Bob and Brown also co-advised numerous MIT graduate students. They formed a unique team, says McKinley. Working for Bob and Bob was a statement that carried a lot of weight and commitment. Bobs Bingham lecture, Rheology and Energy: What bringing his academic research and energy research more closely together. world. Understanding how they move is crucial. Rheologists work to understand how sea ice breaks up due to climate change and how the stringy slurry from bio-fuel production ment clear at the prospect of new challenges. Certainly, Bobs research has contributed to the next genera tion of leaders in ChemE and related research. McKinley, now head of MITs Non-Newtonian Fluid Dy namics Research Group, describes his labs experimental work Beris of the University of Delaware currently studies mod according to Green, Bobs theoretical work helps provide context and direction for his own research on novel nanoma terials like carbon nanotubes and graphene. Bobs enthusiasm for passing on his commitment to ChemE education and research is contagious. His eyes light up when he describes his vision for linking students skills and altruism to grand problems in energy, health, water, and security. Id love to have student groups work on projects like cleaning up Superfund sites or designing concentrated solar power plants, he says, Id like to say to them, This is one of the worlds big problems, and we need your help. Id like them to leave MIT with a belief they can solve anything and a sense of responsibility to do so. Between 1998 and 2006, Bob had to believe he could solve anything: He balanced his roles as teacher, advisor, researcher, and department head while traveling almost constantly. In 2004 alone, he traveled to China, Germany, and South Korea to speak on chemical engineering research or education. Meanwhile, as Bob traveled around the world, oil prices were traveling out of this world. In 2007, the price of crude oil surged over $100 a barrel. Gasoline hovered at $3.00 a gallon market events unthinkable just a few years before. Carbon footprints, solar panels, bio-fuels, algae-based smog-eaters a new lexicon of concern about fossil-fuel consumption sprang into popular use. A different man might have tooted his own horn to say hed told them so. But thats not Bob. Back in 1999, he and Debbie got rid of one of two family cars to conserve another scarce resourcetime. Then, as now, Bobs busy schedule left too few moments to indulge his love for books. Since then, Bob has taken public transportationa bus, then a trainso he can read and work during his commute. that feature science or scientists; he prefers authors Clive Cussler, Dan Brown, John Grisham, and Michael Connelly. Diamonds Guns, Germs, and Steel Bobs wide reading and boundless curiosity add spice to his lectures, keeping students interested, he says. My advice to a new teacher is: Connect the material youre covering to current global problems. Invite your students to get involved in thinking about the worlds future. Another strategy is to bring in history. Students enjoy my Romans siphoning bitumen. They like learning that the Debo rah Number got its name from the Biblical Book of Judges, Bob says. It opens their eyes.ENERGY STUDIES, C hem E A W ARDSThe MIT Energy Initiative (MITEI) was established in 2006. Bob and Moniz serve as deputy director and director, respectively; they coordinate MITs existing energy activi ties and guide the development of relationships with other institutions, industry, and government agencies. Ever enthusiastic about research and innovation in ChemE and energy studies, Bob acknowledges he does rely on one old-school technologythe blackboard.
Chemical Engineering Education 104In 2008, Bob was elected to the National Academy of Engineering, a measure of his scholarship and leadership in academia, according to Bird. The NAE praised Bob for his co-authoring landmark textbooks, and providing leadership in chemical engineering education. That year, he presented the Barnett F. Dodge Distinguished Lecture at Yale, on The Energy/Environment Challenge and MITEI, and the L.T. Fan Distinguished Lecture at Kansas State, on The Global Energy Challenge: Opportunities for Chemical Engineering Research and Education. Bobs current projects often combine ChemE and energy research. He and McKinley are working with Chevron on de position of waxy crude oils on surfaces of pipes in deepwater drilling operations, for example. Since assuming the deputy directorship of MITEI, Bob has lectured widely on energy and chemical engineering educa tion. He devoted a sabbatical year to establishing MITEI, traveling to Saudi Arabia, Italy, Spain, China, and Singapore on the Initiatives behalf. In spring 2010, Bob returned to the MIT classroom to coteach a graduate class, Macromolecular Hyrodynamics, with McKinley and to develop a new ChemE senior design module on concentrated solar power. Looking ahead, hes quick to cite great examples of how innovative approaches to ChemE education, supported by MITs Energy Initiative, can prepare students for their futures The 2007 MIT Coal Study is one such innovation, he says. The Coal Study compared different strategies for carbon capture and sequestration (CCS). ChemE senior design stu dents did all the economic evaluations of capture technologies. They had to see energy issues from new perspectives, and the result was a rigorous and consistent economic analysis of one of the worlds biggest problems and a sense of the relevance of their education. MIT and MITEI are bringing together faculty and students from every discipline to focus on energy. Were getting out of disciplinary silos and traditional textbooks. There are no solve the problems in chapter 10 in the real world. Were facing complex and urgent global problems, he says. The MIT Energy Studies Minor is a great example of this new interdisciplinary approach. Ever the ChemE educator, Bob believes ChemE students need a balance of academic and international experience to take on those global problems. Some MIT ChemE students, for example, work at Novartis in Switzerland, at BP in Britain, pressing world problems. U.S. solutions dont always apply in other cultures, in cluding the developing world. Even if graduates work in a big international energy company, they will need to adapt to local cultures, he says. Whether in a classroom, a lab, or an unfamiliar culture, he wants chemical engineers to be ready to ask themselves What kinds of new solutions can we synthesize to solve societys big challenges? Those who know him have no doubt that Bobs impact on the energy world will be at least as great as his impacts have
Chemical Engineering Education 105 Dimensionless correlations are a prominent feature in however, to create a student experiment that will experiment can be set up where pressure drop is measured with the published dimensionless correlation, but the results relevant parameters have been varied. This paper discusses a laboratory experiment in our junioryear laboratory that shows how dimensionless correlations should be constructed. Balls of various densities and diameters are dropped from various heights into a pool of water, and the maximum depth reached by the ball is recorded for each drop. The variables are the liquid density, the ball density, the the greatest depth of penetration. The experimental apparatus is shown in Figure 1. For many years, this experiment was a great frustration to the students. They kept futilely attempting to use the Bucking In fact, the Buckingham Pi Theorem is an existence theorem. It tells us that given m quantities describing a physical situa tion and n fundamental units (mass, length, etc.), a dimension less description of the situation can be written as a function of m-n dimensionless groups. The proof is a construction proof wherein an algorithm is constructed to compute example m-n dimensionless groups. Students are sometimes aware that the version of the theorem found in chemical engineering texts also states that the product of any dimensionless group to any power times any other dimensionless group to any power is also a legitimate dimensionless group. What they dont always problem. The appropriate groups can be obtained only by a low-probability accident. The latter method will fail in situ ations like time-dependent heat transfer problems where the theorem predicts too many dimensionless parameters. There are more elaborate versions of the theorem in the literature that claim to guide the user in which variables and parameters can be combined, but they contain no physics and are thereby suspect as a guide to constructing a physical theory. A LABORA TORY EXPERIMENT ON HOW TO CREA TE DIMENSIONLESS CORRELA TIONSROBERT V. EDWARDS 1 Case Western Reserve University Cleveland, OH 44106-7217 1 Prof. Robert Edwards passed away after submitting this paper. Final revisions to the paper were made by Prof. Daniel Lacks, Department of Chemical Engineering, Case Western Reserve, daniel.lacks@case. edu. Correspondence regarding this paper can be directed to Prof. Lacks. Copyright ChE Division of ASEE 2010 ChE laboratory
Vol. 44, No. 2, Spring 2010 106 THEORY The appropriate method is the one suggested decades ago by Bird, Stewart, and Lightfoot,  and reinforced by Sides.  Basically it says that the physics that control the process less groups. The method consists of writing down the gov erning differential equation and then making that equation dimensionless using the boundary and initial conditions. The functional forms of the governing dimensionless groups will reveal themselves after some manipulation. Importantly, this procedure is valid for approximate models where only the dominant variables are treated. The core of this problem consists of determining the veloc ity of the ball in the water as a function of the balls physical parameters, the parameters of the water, and time. When the velocity goes to zero, the ball is as deep as it is going to get It is easy to estimate the initial ball velocity in the water, v o, when dropped from a height L, by assuming that the initial potential energy is completely converted to kinetic energy. vg L o 2 1 () where g is the gravitational acceleration. This assumes that that velocity is well under the terminal velocity and the ef fect of air friction is negligible. The ball will lose some speed when it penetrates the water surface, but it will be assumed that loss is negligible; recall that the purpose of this analysis dure, rather than to solve the full equations rigorously, and as pointed out above, reasonable approximations do not damage our ability to obtain a reasonable correlation. Once in the water, the differential equation that describes the acceleration of the ball is given by Newtons second law, b b f Dd v dt FF 3 6 2 () where v, D, and b are the velocity, diameter, and density of the ball. The buoyant force, F b, according to Archimedes, is simply F D g b b w 3 6 3 () where w is the density of water. The friction force is estimated in typical chemical engineering fashion using another dimen sionless correlation (see Bird, Stewart, and Lightfoot), Fv D f f w 1 24 4 2 2 Re () where the dimensionless friction factor f is a function of the Reynolds number Re= w vD and The Reynolds number is a function of the velocity of the ball, The equation of motion can now be written b b w w Dd v dt D gD vf 3 3 22 66 1 8 5 Re () In principle, this equation could be solved numerically, but it is not necessary to do so to get the appropriate dimension less groups. By dividing Eq. (5), by the parameters on the vv gL 2 the result is 2 3 2 6 2 0 L g dv dt L D vf v bw b w b *R e* () where Re 0 2 w gL D is the initial Reynolds number hand side are already dimensionless, the left-hand side must tt g L 2 the equation obtained is dv dt L D vf w b w b * Re 1 3 2 2 ( () 7 Eq. (7) implies that the solution for the dimensionless velocity will be of the form vv t L D w b ** *; ,R e ( ) 0 8 which demonstrates the dimensionless groups that affect the trajectory of the ball. The maximum depth the ball reaches, h, is obtained as hv td t t 0 9 ma x () where t max is the time at which the ball reaches its maximum and time, this leads to hg L L g vt dt t 2 2 10 0 ** *( ) ma x Figure 1. Ball-Dropping Apparatus.
Chemical Engineering Education 107 h L vt dt t 2 1 1 0 ** ( ) ma x Eq. (11) implies that h L h L t L D b w *; ,, Re () ma x 0 12 Now, the value of t max is determined by the condition v*(t* max)=0, which in combination with Eq. (8) implies tt L D b w ** ,, Re () ma xm ax 0 13 By combining the above two equations, the result is found that h L h L L D b w ,, Re () 0 14 which shows how the dimensionless depth to which the ball falls will depend on a set of dimensionless physical parameters. At this point, the solution is not known, but common engi neering practice is to attempt a correlation of the form h L A L D b w a b c Re 0 ( () 15 where the constants A,a,b,c are the parameters that minimize a least squares sum. There is not enough space here to cover logarithm of the expression above and doing a least squares simple using an optimizer like Solver in Excel ciple, if the experiment is repeated, different parameters will result. The important question is How large a variation is expected from experiment to experiment? The answer to this maximum likelihood estimate  of the parameter variance for the case where the errors in the depth estimate are Gaussian. Under these circumstances, the variance of the parameters from experiment to experiment is proportional to the average measurement error and inversely proportional to the sharpness of the least squares minimum, taken as the expected second derivative at the minimum. It can be shown that the parameter estimate error can be estimated by the following procedure : First, the Fisher information matrix, F s yy pq n p n q n 1 16 2 () p derivatives are evaluated at the set of parameters (Re 0, D, and B ) used for the n th experimental trial. The parameter s 2 is the estimated measurement variance, which is the average square of the deviation of the theory from the measurement estimated variance for each of the parameters is given by pp p F 21 17 () where F pq 1 is the inverse matrix of F pq. This methodology is described in detail elsewhere.  While these calculations look cumbersome, they are easily done using a spreadsheet program like Excel RESULTS AND DISCUSSION A randomly selected students set of laboratory data is ana lyzed as described above. This data included the maximum depths reached by 14 different balls (with various D and B ) dropped into water from 3 separate heights above the water; the parameter ranges were 2.5 cm < D < 6.5 cm, 0.55 g/cm 3 < B < 0.90 g/cm 3, and 34 cm < L < 141 cm, and the maximum depths that the balls reached were in the range 11 cm < h < 47 of the measured value of the depth vs. the theoretical depth Fisher information matrix and its inverse are calculated, F pq 704 93 331 06 2973 0 13571 331 06 198 01 1 .. .. 4 437 0 6388 8 2973 0 1437 0 1311 0 57099 13571 .. .. 6 6388 8 57099 262630 F pq 1 0 3399 0 00542 0 01138 0 01522 0 0054 .. .. .2 20 02554 0 00194 0 00048 0 01138 0 001937 00 0 . .. 1 194 0 00021 0 01522 0 00048 0 00021 0 00076 .. . tion to the experimental data, A= 1.60 0.51 a= 1.32 0.14 b= -0.61 0.04 c= 0.062 0.024 p, which describe the standard deviations for the uncertainties in the p are associated with p are associated with 95%
Vol. 44, No. 2, Spring 2010 108 Students often try other functional forms, however, which hc mD L b 0 3089 18 07 91 06 02 6 ( ) .. 2 calculated 2 is a measure of, much less how large a change is meaningful. That should be pointed out to students, but even more impor them nothing about what might happen if the experiment were correlation should do a good job of predicting the change in behavior since it is dimensionless and contains the essential physics of the problem, including relative densities and the effect of viscosity. The method demonstrated to derive the appropriate dimen sionless groups above is quite robust, but it can give different functional forms. For instance, if the velocity were made dimensionless by embedding it in the Reynolds number, a different form would have been obtained. Following the same methodology as above, the form for the correlation would be w b w a w hgD A gD 22 2 2 23 b w gL D 2 c () 19 With no loss of generality, this expression can be multiplied by the second dimensionless group to give h D A gD b w a w 2 23 b w c gL D 2 20 () This functional form is legitimate, but not as desirable as Eq. (15) because it stresses the viscosity, which was not varied in the experiment. In contrast, in Eq. (15), the viscosity only ap peared in the friction factor correlation where the dependence on the Reynolds number is well tested. All three versions of the correlation have the same depen dence, within experimental error, on the variables used: the ball density, ball diameter, and height of the drop. The dimen sionless versions are preferred if only because they give the engineer some guidance as to what experiments need to be liquids other than water are used. More experiments should be done where the surface tension and viscosity are varied. The example shown here gives an indication what informa tion can be extracted from experimental data. Another instruc tor has used this same experiment as a vehicle to demonstrate randomness in experiments and examines the sample size dependence for the uncertainty in the estimate of the mean, and whether or not the probable distribution is Gaussian. NOMENCLA TURE A proportionality constant in correlation a, b, c exponents of dimensionless terms in correlation D diameter of ball f friction factor for a sphere F b buoyancy force F f frictional force F ij ij th element of the Fisher information matrix g gravitational acceleration h maximum depth reached by ball L height ball is released above water Re Reynolds number Re 0 s 2 Estimated measurement variance t time t* dimensionless time tg L 2 v velocity of the ball v* dimensionless velocity of ball vg L 2 k k th b density of ball w REFERENCES 1. Bird, R.B., W. Stewart, and E.N. Lightfoot, Transport Phenomena John Wiley and Sons, New York (2002) 2. Sides, P.J., Scaling of Differential Equations, Analysis of the Fourth Kind, Chem. Eng. Educ. 36 232 (2002) 3. Edwards, R.V., Processing Random Data: Statistics for Engineers and Scientists, Figure 2. Relationship between the measured maximum depth of the dropped balls and the tted (theoretical) val ues. The line denotes points of equality of the measured and theoretical values.
Vol. 44, No. 2, Spring 2010 109 For the past half-century, research performance has been the mainand sometimes the onlycriterion for tenuring and promoting engineering faculty at research universities, and its becoming increasingly important at insti tutions whose primary mission has traditionally been teaching. This trend has had unfortunate consequences. Intense pres sures to bring in grants and publish papers force professors to spend most of their time on their research and the minimum they can get away with on their teaching, relationships, and healthand the quality of the latter three often shows it. Faculty members with strong research records and belowaverage teaching routinely get to be full professors, while outstanding teachers with below-average (and sometimes average) research productivity dont get tenure. Depress ingly many research papers are published that have little or no impact on technology or society and are never cited by anyone other than their authors, and core engineering courses stagnate, even though globalization has dramatically changed the skills engineers will need in the coming decades. If university administrators were being honest, they would state that they need massive amounts of external research funding to function, and while teaching also matters, the main determinant of a faculty members value to them is scholarly achievement. No administrator would dare say that publicly, though, since to many stakeholdersparents, potential and current students, alumni, donors, and legislatureseducation is more important than research. The chancellor of a university that proclaimed teaching to be of secondary importance would have to face some hard and unwelcome questions. So what happens instead is rationalization. Chancellors, provosts, and deans routinely declare that teaching is their institutions most important function, and to justify the heavy dominance of research in the criteria for faculty hiring, ten -THE LINK BETWEEN RESEARCH AND TEACHING RICHARD M. FELDER North Carolina State University Copyright ChE Division of ASEE 2010 Random Thoughts . .ure, and promotion, they claim that research and teaching are inextricably linkedso much so that only productive researchers can be good teachers. They offer that proposition as a self-evident truth with (ironically, considering the subject) no supporting evidence whatever. There is no logical reason to expect productivity in research and effectiveness in teaching to be closely related, since research and teaching have different goals and require differ ent skills and personal attributes. The goal of research is to advance knowledge, while that of teaching is to develop and enhance abilities. Excellent researchers must be observant, objective, skilled at drawing inferences, and tolerant of am biguity; excellent teachers must be skilled at communication, familiar with the conditions that promote learning and expert at establishing them, approachable, and empathetic. Having both sets of traits is clearly desirable but not at all necessary 40 hours a week, so that time spent on one activity is inevita bly time taken from the other. It should therefore come as no research productivity and teaching effectiveness.
Chemical Engineering Education 110As it happens, many studies have been performed and thats exactly what they reveal. Most arguments for requiring all fac ulty members to be active researchers relate to how research can enhance teaching, but a recent review of the literature  demonstrates that the potential enhancements are not gener ally found in practice. The next few paragraphs list the most common arguments and summarize what the studies show about them. For details and citations, see Reference 1. * Argument: Research productivity correlates positively with teaching effectiveness. Fact: Wrong. Correlations between numbers of papers and grants and measures of teaching quality such as student evalu ations, peer evaluations, and learning outcomes are mostly negligible and sometimes negative. Argument: Research-intensive universities provide the best undergraduate education. Fact: been found between a universitys research orientation and numerous student learning and satisfaction outcomes. Argument: in science and engineering to be viable teachers. Fact: Never demonstrated, and almost certainly wrong for all but advanced graduate courses on the instructors research specialties. In recent decades applications of most core under graduate and graduate courses have expanded and impressive resources for teaching those courses have become available, but basic course content has not changed by all that much and little research is now done on that content. Pedagogical experts are much more likely than disciplinary researchers to know how to modernize most core courses appropriately. Argument: Faculty with active research programs bring their research into the classroom and use it to inform and enliven their teaching. Fact: Usually wrong, especially in undergraduate classes, and when research is integrated into teaching its not always a good thing. Most current research is well beyond the scope of all but advanced graduate courses, and rigid curricula make it challenging to bring in new material. Some instructors do discuss their research in class and some of their students ap preciate their enthusiasm, but other students complain about excessive digressions from basic course content and/or the in structors apparent lack of interest in teaching that content. Argument: Research experiences enhance undergraduate education. Fact: True for some students. Participation in undergradu of African-American students (but not of other groups), a number of self-reported growth measures and research skills (but not externally measured cognitive skills), and pursuit of graduate study. Even when the argument is supportable, however, it does not justify requiring all faculty members to be active researchers. For one thing, it presumes that active researchers are likely to be better than their more teaching-ori ented colleagues at designing and supervising undergraduate research. No supporting evidence exists for this presumption; in fact, much undergraduate research directed by research fac ulty has students functioning more as unpaid lab technicians than as true researchers. Moreover, undergraduate research is resource-intensive, and at most universities relatively few undergraduates engage in it. Incorporating inductive methods such as inquiry-based, problem-based, and project-based learning into core class instruction could produce many of at a lower cost. * * In short, the unwritten rule that all university faculty should be active researchers places unreasonable and unhealthy demands on faculty members (especially untenured ones); weakens departmental teaching programs; keeps potentially outstanding teachers from devoting enough time and energy to teaching to realize their potential; deprives students of some inspirational and possibly life-changing instructors, mentors, and role models; and is unsupportable by either logic or research. Which leaves us with two questions. (1) If most of the po tential synergies between research and teaching are not being achieved in practice, what can be done to better achieve them? (2) How can schools and departments recognize, reward, tenure, and promote outstanding teachers with little interest in traditional research without compromising their institutions the next column. REFERENCE 1. Prince, M.J., R.M. Felder, and R. Brent, Does Faculty Research Im prove Undergraduate Teaching? An Analysis of Existing and Potential Synergies, J. of Eng. Educ. 96 (4), 283 (2007),
Vol. 44, No. 2, Spring 2010 111 The major characteristic that sets an engineer apart from every other profession in the world is his/her ability to apply the concepts of scaling/up-scaling to a variety of situations. What do we mean by scaling? Well, take for instance a chemist working in the laboratory designing a new take that laboratory synthesis and convert it to a process that produces thousands of tons of that drug per year? Probably not; however, a chemical engineer would be an excellent can didate. Similarly, if building an airplane, scientists (physicists, come to mind, in spite of the obvious useful roles of their professions. An aeronautical engineer would most likely be the selection that makes everybody comfortable. The same can be said for building structures (bridges, buildings, etc.) where civil engineers are the masters, and for the scaling of industry where industrial/managerial engineers are very skillful. The list is long, but these few examples illustrate the basic concept: Engineers are masters of scaling/up-scaling. Therefore, it is imperative when training engineering students, that they fully grasp the concept of scaling/up-scaling to be able to implement it for practical applications, such as the ones mentioned above. One important class of up-scaling in engineering educa tion is the different scales involved in describing quantities related to the physics of transport (mass, momentum, energy). THE SOCCER BALL MODEL: PEDRO E. ARCE, JENNIFER PASCAL, AND CYNTHIA TORRES Tennessee Technological University Cookeville, Tenn. 38505 Copyright ChE Division of ASEE 2010 In many high school or college-level courses, students are introduced to velocity, density, energy, etc., from a discrete scale point of view.  In many engineering applications, however, when studying the physics of transport, it is neces sary to develop conservation equations for system properties, ChE classroom
Chemical Engineering Education 112 such as total mass, energy, and momentum for a continuum, or microscopic scale.  To accomplish this, the concept of a continuum scale must be introduced to students. Since most students have only been exposed to the physical and chemical concepts related to total mass, energy, and momentum, from a discrete scale point of view, the concept of a continuum scale can be very challenging. In making the transition from a discrete scale to a continuum scale, one very important pedagogical aspect to keep in mind is that students already have substantial knowledge related to calculating the total mass, velocity, and momentum of a single particle (discrete domain). So from the students learn ing point of view, how does the instructor use their previous experience and knowledge with the discrete domain to scale it up to the continuum domain? Most textbooks do not address this issue. In fact, many of them have suppressed or hidden the process associated with the up-scaling b, on the assumption that all steps and concepts are familiar to the learner, when in fact they are not. This can be frustrating to students and does not enable them to fully understand the importance of the idea of a continuum. More over, some textbooks  have approached the problem from discrete case: mV pp p () 1 where m p is mass of the particle, p is density of the particle, and V p is volume of the particle. They have then simply c(t) as follows: md V s Vt c () 2 where m s is the total mass of the system under study. Based on Eq. (1) and Eq. (2), it seems that as suggested in Figure 1, two domains exist: A non-applicable (for the system descrip tion), or old domain (discrete domain) and a new domain (continuum domain). As Figure 1 shows, the discrete, or old, domain is valid for very small scale systems (order of molecules), whereas the continuum, or new, domain adequately describes the mass of the system for domains of a larger or continuum (microscopic) scale. It is interesting to note that the so-called old domain in Figure 1 is at the mo lecular level and the concepts learned by students during, for example, high school or college physics are not necessarily at this scale. The molecular scale is a discrete domain, however, and this characteristic offers a bridge for student learning that is effectively used in the Soccer Ball Model (SBM) protocol described in this paper. The pedagogical challenge described in Figure 1 is that the old domain is the domain in which the students are most comfortable and more knowledgeable with the con cepts. Students, in general, are unfamiliar with the new domain indicated in Figure 1. Many teaching approaches (in the literature) focus on the new domain and mostly forget the level of knowledge that students already have on the old domain. This situation is prob ably very familiar to most students, unfortunately, as oftentimes when learning new concepts they are told to forget everything they already know; this type of learning approach that the students have already ac quired. Another option that instructors sometimes use is to force students to imagine a new system where the boundary (or boundaries) are no lon students to apply old concepts to the new (suddenly introduced) system. These two options illustrate the many disadvantages for the students when they are not engaged in the process b The word up-scaling here is used to in dicate the change of the description of a property from one scale to another, such as, for example, from the microscopic to the macroscopic scales. Density Figure 1. Sketch of the material density as a function of the size of the system indicating the two scales or domains of interest.
Vol. 44, No. 2, Spring 2010 113 of transforming and adapting what they already know. This suggests the need for adopting a procedure in which students are fully engaged in the process of learning (up-scaling), then coaching them on how to move from one scaling level to the next. Moreover, such a process allows students to build on what they already know about the discrete point of view, and to integrate this knowledge with the new view of matter, i.e., the microscopic or continuum scale. concepts from the discrete point of view, students have an adequate background in many complementary subjects including calculus, integral concepts, and algebra. It appears that the mathematical background to help students in catalyzing the transformation from one scale to the other one in an effective way from the students learning point of view. In other words, instead of hiding the details about the scaling-up process, by several activities in which students are exposed to and can  In this contribution, we propose a visual c process to help with the transformation of scales (domains), i.e. from discrete to continuum, by using soccer balls in conjunction with geometrical domains, mathematical principles, and physical properties. The student is exposed to a very powerful set of pedagogical activities to construct a learning environment that is both practical and effective. An introduction to this environment is given in the next section. From the learning environment point of view, the SBM protocol is an effective Principal Object of Knowledge, or POK, a tool introduced in the Colloquial Approach [6,7] and later adapted to include other learning environments. [8,9] POKs are tools that allow the facilitator to focus students learning on a collection of topics or variables conducive to visualizing the process of understanding the different aspects. In this sense, the SBM presents scaling, packing, geometrical, mechanical, the process of students learning. DESCRIPTION OF THE SYSTEM(S): THE SOCCER BALL MODEL ELEMENTS Soccer d is the worlds most popular sport. It is played on the beaches of Brazil, on the grassless surfaces of Argentina, Europe and North America. Therefore, soccer balls are geometrical objects that are popular among college students in a large number of countries in the world. The International Federation of Association Football, FIFA,  (international government of game at the professional level. The smaller sizes (numbers four, three, two) are used in games depending on player ages, and the smallest ones (number one) are mostly given as souvenirs. One can observe from Figure 2a that one of the attractive features of the set of soccer balls (decreasing size from the largest one to the smallest one) is the fact that they are all like objects of the same geometry. Although the balls are made of a shell with air at a given pressure, in the soccer ball model it is assumed that they are all made of the same material as the shell (see Figure 3). This assumption usually promotes c The research on brain-based learning suggests that vision is the most powerful tool for the brain to add new knowledge (Medina, 2008). d We have used the name mostly used in the United States, however in the rest of the world this foot-based sport is simply called football. Figure 2. Elements of the soccer ball model (SBM). (a) Set of the ve sizes of soccer balls approved by the FIFA. (b) Container of a given volume, V s Figure 3 Visualization of the soccer ball material.
Chemical Engineering Education 114 strong discussions of why this can be proposed and allows the instructor to bring previous vs. new student knowledge to the discussion. In addition to N balls of a given size, k (k=1, 2, 3, 4, 5), the soccer ball model uses a container of a given volume, V s (t) (see Figure 2b). This container could be either rigid ferent geometries, i.e. rectangular, cylindrical, or spherical. For simplicity, a rigid, cylindrical vessel is assumed for the analysis (see Figure 4). The idea of control volume is dis cussed in connection with the vessel of cylindrical shape. In general, students are introduced to this idea and also to the concept of dimensions associated with the control domains. In fact, they are made aware that these domains could be of one dimension (line), two dimensions (surface), and three dimensions (volume) e. Figure 4 shows a typical situation that is helpful for the system (vessel + soccer balls) can be viewed as a composite, or a two-phase system with one phase made completely of the void space between the soccer balls. In many classes students are presented with a transparent vessel containing soccer balls to show the different phases and spaces. If the experiment is with air; however, a discussion is conducted for several the mass associated with it by m F. The mass associated with the N soccer balls inside the vessel is denoted by m SB. Since in engineering many different types of practical systems exist f, students are introduced to a variety of systems that may have these types of characteristics where the soccer balls can easily One interesting characteristic from the didactic point of view is that these particles are discrete objects that the students can see and touch. Once the system in Figure 4 is understood, then a procedure for the mathematical formulation for the total mass of such a system can be developed. The process should start with a system such as the one shown in Figure 4, then steps are made where the students can be coached until the formulation (or scale-up) to a continuum scale can be reached. This is the focus of the next section. LEARNING PROCESS: TRANSFORMA TION OF SCALES To compute the total mass of the system depicted in Figure 4, lets start by stating some assumptions. We will assume that the volume of the vessel of cylindrical shape is given by V s. Also, we will assume that a set of N soccer balls of the same size, SB N k has a volume, V pj (j=1, 2,...,N), density, j (j=1, 2,...,N), and mass, m j (j=1, 2,...,N). Since they are discrete objects, one can easily compute the mass of ball j as in the classical college physics textbook , i.e. : mV jN jj pj ,, ,, () 12 3 Because of the assumptions stated above, it is immediately recognized that the total mass of the soccer balls can be computed as: mm V SB j j N j j N pj 1 1 4 () and, therefore, the total mass of the system, m s, with control volume v s g can be computed (for the volume of the container) as follows: mm m SS BF () 5 where m F is the mass of the F material. Then in view of Eq. (4), one can express Eq. (5) as: mV m S j j N pj F 1 6 () Now, one question arises: How can we reduce the mass as sociated with F (m F) and simultaneously increase the mass pletely. The idea of fractions is possible or, alternatively, the choice of V S is discussed. f In general, colloidal and non-colloidal suspensions are very good also be discussed. e The idea of domain is connected to the domain concept of a mathemati cal function, which students are familiar with from calculus courses. Figure 4. Sketch of the different components associated with the container identied in Figure 2.
Vol. 44, No. 2, Spring 2010 115 of the soccer balls (m SB) while maintaining the volume of the whole system as constant ( i.e., V s =constant ) ? To answer this question, one should recognize that within the container there are spaces ( i.e. void spaces that do not include soccer ball (for example, air, see above) that is located between the dif ferent balls (see Figure 4). The rest of the spaces within the container are occupied by the soccer balls. Coaching Point 1: The instructor may want to discuss with the students several examples of particle packing systems: marbles of different sizes and sand are excellent examples. The discussion should be focused on the role played by the size of the particles and the void spaces in a given container to help connect the previous knowledge with the analysis of the situation. The instructor should strongly refuse to give answers, and instead act as a facilitator being ready to offer counter examples to the situations brought up by the students. The discussion should lead to the conclusion that by reducing the particle size, the void spaces are also reduced Now as a corollary: What would the effect of this reducing process be on the number of soccer balls? Should N increase or de of coaching point 1 may be tested by using the soccer ball model. Here, for example, the number 5 soccer balls should into the container. Both m SB and m F should be determined or estimated. This is a very useful exercise h to acquire a solid idea of the systems characteristics. The instructor could as sign vessels of the same volume but of different geometries and ask students if N is the same, or what would change. Coaching Point 2: The instructor may want to coach the students in calculating the mass of particles in a given vol ume. The idea of voids and porosity of a packed bed can be easily connected to the problem. Experiments to measure the properties should be discussed. This exercise will produce intense discussions among students regarding very relevant aspects of the different geometries (see coaching point 1, 1 has been understood, students should be able to check it by using the soccer ball model. By using the idea of the size of soccer balls, the process sketched in Figure 5, one should change the number 5 soccer balls to number 4, again measure the mass of soccer ball material and the mass related to the void space, m F. Once the process or experimental protocol is at what iteration should it be stopped? The idea of an ap proximation in engineering becomes useful to address this question. Recall what is intended; to minimize the mass of the void spaces (m F) up to a point where: mm SS B () 7 Coaching point 3: The instructor may want to discuss with students at this point the implications or approximations if the protocol were to be implemented in the laboratory. Some of the relevant aspects may include: 1. How do we stop the iteration process to produce the desired approximation in Eq. (7)? Hint: The idea of the sequence and the comparison of the mass of the system in iteration k with the k-1 would be helpful: mm s k s k 1 2. The step in the sequence ( i.e., k) may be determined by the accuracy of the instrument being used in the measurements. 3. How valid is the approximation in Eq. (7) for the pur poses of reducing m F and increasing m SB? By stressing the various geometrical and experimental as pects of the protocol, students gain a very useful hands-on and concrete view of the transformation proposed in the process shown in Figure 5, where the unloading and reloading of the vessel with the different-size soccer balls is sketched. Students soon realize that the set of soccer balls is incomplete for the purposes of perhaps reaching a valid approximation in order for Eq. (7) to hold. This is another great advantage so they can develop possibilities for other systems that will help them to achieve the results. In this sense, the soccer ball model is just h A very powerful visualization of this protocol can be achieved by using actual soccer balls and containers of different geometries. Figure 5. Sketch of the Protocol of reducing the mass of the F phase and increasing the mass of soccer ball material.
Chemical Engineering Education 116 a pedagogical promoter, or initiator of a process that allows students to visualize the transformation in the scales. At the end of the process when the approximation given in Eq. (7) is reached, the mass of the system is given by mV S j pj j N 1 8 () since m F is very small it can be neglected compared to m SB for all practical purposes. Now the next question is to check how Eq. (8) can be improved. One excellent possible solution is to continue using small objects (smaller than the smallest soccer ball) as most likely students have proposed, and going to sizes such as, for example, grains of sand and even molecular sizes. Mathematically, this implies m V S N V jp j j N pj li m ( ) 0 1 9 geometrical situation in the vessel to a mathematical-based domain with incremental volume V j (see Figure 6). It is useful to discuss with the students the dimensions of the volume of this tiny domain (with respect to the volume of the vessel)  with the mathematical concept of incremental volume. From this, now Eq. (9) becomes: m V S N V jj j N j li m ( ) 0 1 10 Eq. (10) is nothing but a representation of the Riemann sum,  that in the limit produces the Riemann integral, i.e. li m ( ) N V jj j N Vt j c Vd V 0 1 11 From Eq. (10) and Eq. (11) now we can write: md V s Vt c () 12 It is very straightforward to conclude that Eq. (12) allows us to compute the total mass of the system from a continuum point of view whose control volume V c (t)=V s This equation (valid for a continuum) is derived directly from the discrete objects ( i.e. particles=soccer balls) and therefore every physical concept that was valid for a discrete domain is also valid for the continuum domain. By using the visualization protocol as described in this section, we have introduced a different scale in the computation of a physical property, i.e. for this case the mass (total) of the system. The total mass of the system (for a single component system, the total mass coincides with the mass of the component) is the primary variable or property that allows us to compute others that are proportional to it (see the section below). Therefore, the transformation from a discrete scale point of view to a continuum scale point of view is relatively straightforward. Students never have to deny that what they learned in the discrete scale is valid for the continuum scale. It is, at the end, a different mathematical description of the same property since the scale has changed. EXAMPLES AND APPLICATIONS: OTHER VARIABLES OF INTEREST The learning protocol described in the previous section may be applied to other variables that are relevant for the formulation of conservation principles, such as linear and angular momentum and energy.  The steps are identical as for the case of total mass. First, one should start with the Figure 6. Mapping of the container volume to a geometrical domain of size V s (a): Container lled up with N soccer balls. (b) Side view of the space occupied by the N soccer balls. (c) Geometrical domain showing N incremental volumes of size V j and with a density j 1 2 3 4 5 6 N j
Vol. 44, No. 2, Spring 2010 117 section to reach the proper mathematical equation for the new property. For example, the linear momentum, p for a discrete particle is: pm v pp () 13 for each particle of mass, m p, and velocity, v p From a continuum point of view, (by using the protocol previously described) we can conclude, p vdV Vt c () 14 There is a shortcut approach by realizing the mass of the sys tem is given by Eq. (12) and then, by replacing the velocity of the particle by the one of the medium, one arrives to Eq. (14) from the suggested form given by Eq. (13). Didactically, this is consistent with the fact that students have a protocol in mind of how the transformation works and is similar to the mathematical tricks used frequently in analysis courses to obtain results in a quicker manner. Similarly, energy, E, for the discrete point of view is given by: Ep v p () 15 into:  Ev vd V Vt c () 16 Note: Students may want to use the relation, pv dV Vt c () 17 vv x could, however, be a function of the position inside the control volume, V c(t), and therefore Eq. (17) will not capture this situation. Eq. (14) and also Eq. i.e. Eq. (16) is the most general equation for describing the energy of the system for a continuum scale. More complicated functions or properties can be expressed from a continuum point of view. For example, the moment around a point, i.e. torque, M is given by: Mr F p () 18 where r p is the position vector of the force, F It is known from mechanics that dp dt d dt mv R p () 19 From a continuum point of view Eq. (19) can be expressed as, R d dt vdV Vt c () 20 If R is the net force applied to the system, i.e. the one en closed within the control volume, V c(t), then from Eq. (18) and Eq. (20): M d dt r vdV R p Vt c () 21 in the case that rr t pp Caution must be kept in mind re garding the interpretation of the meaning of the derivative, d dt in Eq. (19); also, the formulation of Eq. (21) and similar ones requires a careful analysis and discussion that are not part of the scope of this contribution.[12, 14] The protocol of the soccer ball model is actually a helpful tool, from a didactic as well as from the conceptual point of view, since, in prac tice, all key variables for the description of the conservation principles in a continuum scale can be systematically derived by using such a protocol; or, alternatively, shortcuts based on the protocol are possible. IMP ACT ON STUDENT LEARNING The SBM protocol was introduced some years ago  and it has been systematically implemented in various core courses both at the undergraduate and graduate level. The comments by students in course exit interviews have indicated the healthy action of the protocol in helping students build an excellent level of knowledge based on the previous level they bring to the classroom as well as avoid misconceptions. In addition, the protocol has been extremely effective for introducing macroscopic or integral balances from a continuum point of Furthermore, the connection between mathematical concepts learned in calculus and engineering applications, such as the change of scales, is effectively integrated by using elements of the SBM.  This, in turn, assists the students in understand ing the relevancy of the mathematical tools in engineering applications and enhances the appreciation of their power in, for example, simulating engineering processes. We believe the protocol of the SBM is an effective tool in removing the students frustration in understanding a very different description (from the students point of view) of matter, momentum, energy, and related concepts from a new and more sophisticated scale, i.e. the continuum scale. SUMMARY AND CONCLUDING REMARKS This contribution summarizes some of the typical ap proaches used to introduce students to scaling/up-scaling for variables and properties related to conservation principles in continua. The key aspect is the introduction of a new learn ing protocol, the soccer ball model, that engages students in every step of the process of transforming scales from a discrete level to build a continuum. The soccer ball model
Chemical Engineering Education 118approach allows students to use what knowledge they have already acquired in previous courses from the discrete point of view, to apply it in a systematic manner, and to obtain the description of properties such as mass, energy, and momen tum; these properties are used in conservation equations for learning environment of the soccer ball model is powerful since students never lose track of the discrete nature of the objects when engaging in building a continuum. They reach this level at the end of the protocol and simultaneously they have been able to develop an excellent idea of the continuum with an equation to compute the given property or proper ties of the system.ACKNOWLEDGMENTSThe authors are grateful to students of the transport phe nomena sequence (mass, momentum, energy) at the College of Engineering (Florida State and Florida A&M universi ties), Tennessee Technological University, and Universidad Nacional Mayor de San Marcos (Lima, Peru) for their helpful comments. A presentation with preliminary results was made at the AIChE Annual Meeting (1999) and another one, with additional aspects, as a student poster at the AIChE Annual Meeting, Centennial (2008). We are grateful for the input received. Student (JA) is supported by a diversity fellowsip froom TN Tech. Student (CT) is supported by a fellowship from Chile. REFERENCES 1. Halliday, D., R. Resnick, and J. Walter, Fundamentals of Physics 7th Ed., New York, J. Wiley (2004) 2. Bird, R.B., W. Stewart, and E.N. Lightfoot, Transport Phenomena New York, John Wiley (2001) 3. Arce, P., M. Oyanader, and S. Whitaker, The Catalytic Pellet: A Rich Learning Environment for Up-Scaling, Chem. Eng. Ed. 41 (3) 187 (2007) 4. McCabe, W., J. Smith, and P. Harriot, Unit Operations of Chemical Engineering New York, McGraw-Hill (2001) 5. Arce, P., and L. Schreiber, High Performance Learning Environments, Chem. Eng. Ed. 38 (4) 286 (2004) 6. Arce, P., Principal Objects of Knowledge (POKs) in Colloquial Ap proach, Annual Conference Proceedings of the American Society for Engineering Education, Section 3413 (2000) 7. Arce, P., The Colloquial Approach: An Active Learning Technique, J. of Science Education and Technology 3 (3): 145-160 (1994) 8. Arce, P., How Do We Introduce Continuum Mechanics Concepts to Engineering Undergrads? (Session: Undergraduate Forum, Paper 175e, suppl. to Chem. Eng. Progress 82 (1999) 9. Arce, P., Role of Teams in High Performance Learning Environments, Hi-PeLE. Purdue University, School of Engineering Education, West Lafayette, IN, Department Teaching Seminar (2007) 11. Stewart, J., Calculus Company (1999) 12. Whitaker, S., Introduction to Fluid Mechanics Malabar, FL, Krieger Publishing Co. (1981) 13. Bird, R.B., The Equations of Change and Macroscopic Mass, Momen tum, and Energy Balances, Chem. Eng. Science 6 (3) 123 (1957) 14. Whitaker, S., Newton, Euler, and the Speed of Light, Chem. Eng. Ed. 43 (2), 96 (2009)
Vol. 44, No. 2, Spring 2010 119 Process control has often stood out in the chemical engineering curriculum as a necessary topic that is oddly disconnected from the rest of the curriculum. While control modeling still relies on conservation laws and other fundamentals of chemical engineering, its mathemati cal focus on process descriptions in the Laplace domain has made it appear to students as a course distinct from regular chemical engineering. In reality, process control is key to in dustrial practice and will draw upon an engineers theoretical knowledge and practical experience to be effective. One goal of the addition of inductive and deductive labo ratory exercises to this course has been to improve the stu dents understanding of the linkage between the course and engineering practice. Additionally, the changes are driven by more effective methods of instructioninductive learning and experiential learning. This approach assumes that physi cal laboratories are preferred to virtual labs or simulations whenever possible, although virtual labs and simulations can effectively be used to meet objectives for this course. This paper expands upon the initial experience with in corporating inductive laboratory experiences previously reported.  EXPERIENTIAL AND INDUCTIVE LEARNING Experiential learning is one approach to engaging students actively in the learning process. Farrell and Hesketh  suggest that students typically recall only a low percentage of what they hear, while if they hear and see something done, they may recall closer to half of the experience. If they actually do something, such as conduct an experiment, they are likely to recognized as contributing to common student learning styles in engineering.  There are numerous examples of incorporation of experi ential learning in process control courses. [4-9] Most involve development of experiments, typically required as a part of a distinct one-hour laboratory section extending the course length from three to four semester hours. Clough  incorpo rated experiments directly into the lecture course prior to the addition of the one-hour laboratory section.  Others have attempted to add this active-learning component through use of Web-accessible experiments.  More recent efforts to include experimentation in process control courses include development of kits using LEGO RCX brick and quickdisconnect piping to build desktop process control equipment for in-class use.  Inductive learning refers to the organizational approach to more general conclusions. This is effectively the inverse approach of deductive learning, where general principles Most teaching is performed in the deductive mode, but most inductively. This suggests that induction is a more natural CHEMICAL PROCESS CONTROL COURSE ChE DAVID L. SILVERSTEIN AND GIFTY OSEI-PREMPEH University of Kentucky Paducah, KY 42002-7380 Copyright ChE Division of ASEE 2010
Chemical Engineering Education 120 learning style and more effective for many student learners.  Studies on the implementation of different inductive teaching methods in science and engineering have shown that students conceptual understanding and attitudes toward learning sig or comparable to traditional teaching methods.  Moor and Piergiovanni  describe their application of classroom kits for inductive experiments in a process control course. An inductively structured course in Heat and Mass Transfer is described by Farrell and Hesketh.  Hesketh, Far rell, and Slater  describe the role of experiential learning when using an inductive style of teaching. COURSE DESCRIPTION The course described here is a three-hour lecture course offered during the Spring of the senior year. There is no for mal prerequisite other than consent of instructor, although it draws heavily upon a course in modeling offered during the Spring of the junior year. The expected outcomes for the course are that students should be able to: Apply knowledge of mathematics and science to process dynamics and control Analyze and interpret different control systems transient and frequency response data Design simple control systems for distillation columns and chemical reactors Identify, formulate, and solve linear control problems Use engineering tools to analyze control systems When preparing to modify the course to add experiments and increase inductive content, the following topics were selected for emphasis: Instrumentation eters to responses of real systems Empirical modeling Signal conditioning and interpretation PID controllers and tuning MIMO interaction To modify the course and still conform to reasonable student expectations of time formally committed to the course, lecture time was reduced 10 minutes for every 30 minutes of expected laboratory time. Labs were scheduled at least one to two weeks for the lab work was given as part of the student homework Figure 1. (right) Pressure regulation apparatus. Figure 2. (below) Con trol panel for pressure regulation apparatus.
Vol. 44, No. 2, Spring 2010 121 grade, which was increased to account for 25% of the total grade for the course. The lab reports were kept simple (mostly of textbook-type problems assigned was reduced. Three class sections at the University of Kentucky have consisted of 10 students, the second class two students, and the third three students. The quantitative assessment described THE EQUIPMENT The commercially available equipment described herein is typical of many devices offered by a number of vendors, including Creative Engineering,   and Feedback Instruments Limited.  Two devices were used over the course of the semester. The of a pneumatic control valve, various pressure gauges, an a storage tank. The apparatus can be connected to a control panel (Figure 2) that incorporates an ammeter, a voltmeter, signal conditioner ports, and an industrial-type digital PID controller. The second device is a Process Plant Trainer (Figure 3), which combines three plate heat exchangers, two feed tanks, Figure 3. (left) Process Plant Trainer. Figure 4. (below) Control panel and interface board for the plant trainer. a dead-time segment of tubing, various solenoid valves, level device can be connected to a control panel with an interface board (Figure 4). For some experiments, the apparatus is con
Chemical Engineering Education 122 nected to a PC with MS-DOS-based acquisition and control software. A PLC is also available for use with the system (Figure 5). THE EXPERIMENTS Students were presented a syllabus and a homework assign ment (including the laboratory assignment), and told to leave their books in the classroom (which was then locked) and come down to the controls lab. Students were briefed on safety rules for the lab, had the exercise explained to them, and then for this 30-minute assignment were to: Induce a conceptual understanding of process time con stant and gain Demonstrate intuitive use of proportional control Sketch process behavior Introduce elements of instrumentation Not all objectives were immediately met by the laboratory exercise, but the student experience was used in subsequent lectures to form a foundation for discussion. For example, students were asked to maintain a particular pressure in a (leaking) tank by adjusting the current signal sent to the I/P transducer connected to the pneumatic control valve. We discussed immediately after the lab what everyone did to set the pressure in the tankstarting with big changes when the tank pressure was far from the desired pressure and making smaller adjustments when the error was smaller. Proportional control was introduced a month later referring to this initial experience. Students were able to describe the meaning of a time constant, noting how quickly the pressure apparatus responded (small time constant) compared to a level control response in the process plant trainer (larger time constant). We later modeled both processes and determined the relative tions. Figures 6a and 6b are the assignment sheet provided to students. Note that the deliverables were kept simple so that students could focus on observation and not on record ing data. This also helped maintain class morale, as students were leery of the added workload of labs in a traditionally lecture-based course. This exercise was clearly inductive, since students had no background in control prior to the lab. After completing the experiments, we returned to the classroom and discussed what following class meeting, since students understood why they were learning control. Student discussion quality in that class Figure 5. PLC con nected to the plant trainer.
Vol. 44, No. 2, Spring 2010 123 t Figure 6a. Page one of lab exercise conducted during rst class meeting.
Chemical Engineering Education 124 2.For the level apparatus, sketch desired level, actual level, and relative pump rotation on the same plot. Include a legend to distinguish between traces. t 3.Draw process schematics of both pieces of equipment. For multiple lines going to the same place, draw a single line. Do not include equipment not in use. A process schematic uses pictures and lines to represent equipment and connectivity. See Figures 1.1 and 1.2 in your text. 4.Draw a control schematic of each system. A control schematic contains equipment representations along with indication of what is being measured and what is being manipulated. See Figures 1.4 and 1.5 in your text. 5.What was the key difference (in context of control) between the pressure and level systems? Revised 1/2/2007 Figure 6b. Page two of the initial exercise.
Vol. 44, No. 2, Spring 2010 125 was far better than in previous offerings, since there was a common experience to form a basis for that discussion. The into more general principles of process control. The second laboratory exercise was designed to reinforce the students understanding of dynamic modeling. Students collected data for a two-tank gravity drainage system, with both tanks connected at their bases. Students then prepared an analytical model of the system, and compared their results match experimental data in their chemical engineering labo ratory experience. Since students had prior experience in writing balances and designing experiments, this exercise was designed deductively, but still engaged students actively in the process of convincing them of the validity of theory discussed in class. deductively as well, since students had previously performed regressions of experimental data. Students collected step response data of a multi-step heat exchange process con trolled by changes in three process variables. After the lab, they selected an appropriate model and determined model the highlight of the exercise was the opportunity to wire a controller to a feed valve to eliminate the need to manually maintain an adequate level in a feed tank. Students were provided wiring and a manual, and were told the controller or so required to allow the system to come to steady state, all three groups managed to deduce the appropriate wiring connections to set up automatic feedback control. This experi as deadband and direct action. tems. Since closed-loop behavior, PID controllers, controller tuning, and stability had not been discussed in class, this was an inductively designed exercise with detailed instructions on what to do and what to observe. The pressure apparatus was varied the controller gain to make the system marginally stable and unstable. They then observed the difference in response with P and PI control. They rewired the system to bypass the observed the difference in system behavior, and were asked to show why it was required. observe which variables interacted, and then to suggest why from general models. They also observed the effect of detun ing to account for interaction. The last exercise required students to run through an exer cise in PLC usage involving a ladder program simulating a exercise was not completed due to communications issues between the PC and PLC. Time required for each of these exercises varied widely. class period. The second and fourth, which also involved preparation of the students ( did you read the instructions? ). The remaining exercises took about three hours, due to the thermal transfer dependence of the experiments. ASSESSMENT of this course structure were asked to submit responses to a free-answer survey assessing their perceptions of the labs. Of the 10 students in the sample, nine responded. 1. Were the laboratory exercises a valuable part of the course? 2. Did they help you better understand the course material? All nine students indicated that the exercises were valuable and helped in understanding the course material. 3. Were they more valuable when they served as an in troduction to course concepts, or when they reinforced lectures and reading? This question was intended to determine whether students preferred the inductively designed labs or deductively de signed labs. One-third preferred the inductively designed labs, while the remaining two-thirds preferred the deductive labs. The learning styles of these students were not assessed, so the only conclusion by this instructor is that the students were not considering the learning value of the experience, but were focused on their comfort level during the lab. to subsequent sections of the course. 4. Which labs would you recommend be kept? Should any be removed from the course? Students expressed a distinct preference for the faster exper iments, since much of the time spent on the slower (thermal) labs was spent idly waiting for the system to reach steady state. Later offerings had the apparatus preheated, but lab times were still longer than students would have preferred. 5. What changes to the lab/lecture balance would you recommend? Students were concerned about the time spent on the thermal labs, and preferred having a regularly scheduled lab section. Due to the number of hours in the current cur riculum, this is not an option at this time. This is, however,
Chemical Engineering Education 126the historical evolution of such improvements to a process control course.[9,10] Other requests included make the equipment work, refer ring to problems with a peristaltic pump and with entrapped air in a pressure measurement line. The labs seemed to improve student performance on exams, but more noticeably students seemed more tunedin during lectures in which the lab results were used as examples. Students were more eager to ask questions in class after labs had been run. These observations were consistent during all course offerings from both authors. Interviews with students indicated that the experiments truly did improve their understanding and provided a framework from which they were able to better analyze process control problems. The key improvements made for the second offer for inductively designed labs, and efforts to reduce waiting times during thermal labs. No direct assessment of any improvement in student learn ing was possible. For one instructor, the course was only taught one time. For the other instructor, the addition of these laboratories was concurrent with a change in textbook and course structure. ADDRESSING LARGER CLASSES Clearly, it is not feasible in all institutions to require all students to participate in a half-dozen lab exercises in small groups over the course of a semester. For larger classes, some options include in-class kits such as those developed by Moor and Piergiovanni,  or remote, Internet-based labs available at all hours, such as those developed by Henry.  One additional option takes advantage of newer equipment that is controlled via computer. By adding a remote video camera and using a remote access technology (such as Remote Desktop in Micro soft Windows), students can get familiar with the equipment in one hands-on lab exercise, and then use the remote access technologies to control the equipment in later labs. This para digm emulates the experience of an industrial operator, where most of the control is performed by wire with only occasional visits to the equipment being controlled. Computer software could also be used to accomplish many of the objectives of laboratory experiences. Control-oriented simulation software is commercially available as a stand-alone software package ( e.g. Loop-Pro Trainer ) or as an add-on to MATLAB.  Either approach should be suitable as a substitute for some laboratory exercises. CONCLUSIONS Restructuring a chemical process engineering course to to improve student learning. Inductively designing some of those exercises seems noticeably more effective at introduc ing new topics in process control than traditional lectures alone, as evidenced by the quality of classroom discussion as perceived by the instructors. Six laboratory exercises were developed within the context and schedule of a traditional lecture course, and were integrated into the course to improve student learning. Student feedback indicates students value the lab experiences, provided they do not perceive they are wasting time waiting for systems to reach steady-state. REFERENCES 1. Silverstein, D.L., An Experiential and Inductively Structured Process Control Course in Chemical Engineering, Proceedings of the 2005 ASEE Annual Conference & Exposition, ASEE, (2005) Also appeared as Silverstein, David L., An Experiential and Inductively Structured Process Control Course in Chemical Engineering, CACHE News 62 Summer 2006. 2. Farrell, S., and R.P. Hesketh, An Inductive Approach to Teaching Heat and Mass Transfer, Proceedings of the 2000 ASEE Annual Conference & Exposition, ASEE (2000) 3. Felder, R.M., and L.K. Silverman, Learning and T eaching Styles in Engineering Education, Eng. Ed. 78 (7), 674 (1988) 4. Lennox, B., and M. Brisk, Network Process Control Laboratory, Chem. Eng. Ed. 32 (4), 314 (1998) 5. Skiliar, M., J.W. Price, and C.A. Tyler, Experimental Projects in Teaching Process Control, Chem. Eng. Ed. 32 (4), 254 (1998) 6. Joseph, B., C. Ying, and D. Srinivasagupta, A Laboratory to Supple ment Courses in Process Control, Chem. Eng. Ed. 36 (1), 20 (2002) 7. Ang, S., and R.D. Braatz, Experimental Projects for the Process Control Laboratory, Chem. Eng. Ed. 36 (3), 182 (2002) 8. Muske, K.R., Simulation and Experiment in an Introductory Process Control Laboratory Experience, Chem. Eng. Ed. 37 (4), 306 (2003) 9. Toghiani, H., R.K. Toghiani, D.O. Hill, and C. Wierenga, Enhancement of Instrumentation and Process Control Studies at the Undergraduate Level, Proceedings of the 2000 ASEE Annual Conference & Exposi tion, ASEE (2000) 10. Clough, D.E., The Integration of Laboratory Experience with a Senior Course in Chemical Process Instrumentation and Control, Proceed ings of the Frontiers in Education Conference, Champaign-Urbana (1977) 11. Clough, D.E., Bringing Active Learning into the Traditional Class room: Teaching Process Control The Right Way, Proceedings of the 1997 ASEE Annual Conference & Exposition, ASEE (1997) 12. Henry, J., Web-based Controls Laboratory Hardware and Software,
Vol. 44, No. 2, Spring 2010 127 Model predictive control (MPC) is a widely used control methodology in the chemical industry for the control of multivariable processes under con straints. Although the implementation of predictive control lers in chemical plants has been traditionally subcontracted to specialized companies, there is a need for the process control engineer to understand the algorithm for the purposes of maintenance and tuning and there is an increasing trend for in-house implementation of these controllers by process control engineers with limited control experience. Thus, there is a great incentive to familiarize chemical engineering graduates with this control methodology. Predictive control theoretical concepts are presented in different undergradu ate control textbooks. [1-3] Also, although there are published experiments of multivariable controllers in undergraduate control laboratories, [4-6] undergraduate-level experiments of constrained predictive controllers based on linear models are less common. The experimental system discussed in this work is used in a process control laboratory course offered as a fourth-year elective in the chemical engineering undergraduate program at the University of Waterloo. The experiment involves the ap plication of a constrained model predictive control algorithm for the control of a double pipe heat exchanger (DPHE). The elective laboratory course is composed of two main experi ments: multivariable control of the DPHE described in the current paper and single variable temperature control in a stirred tank heater. Each of these experiments is conducted in two three-hour sessions. In addition, the course includes a weekly one-hour tutorial lecture in which theoretical concepts LUIS A. RICARDEZ-SANDOVAL, WESLEY BLANKESPOOR, AND HECTOR M. BUDMAN University of Waterloo Waterloo, ON, N2L 3G1, CanadaTESTING A CONSTRAINED MPC CONTROLLER ChE laboratory Copyright ChE Division of ASEE 2010 related to the experiments are covered, e.g., topics on system predictive control. The presentation of MPC theory is done during four one-hour tutorial sessions and it includes a discus
Chemical Engineering Education 128 sion of the output prediction equations, the analytical solution of the unconstrained case using least squares, the effect of tuning parameters such as weights, control, and prediction horizons, and a brief discussion about constrained optimiza tion. Since it is challenging to deliver all the necessary theory of MPC in the tutorial hours, the students are provided with a clear and concise 10-page manual containing both theory and the experimental procedure  and they are referred to the material in the tutorial of the MPC toolbox in MATLAB  and to lecturer notes. The experiment presented in this paper was tailored to illustrate several particular features of constrained MPC vis--vis other control methodologies such as decentralized control. First, the experiment is used to illustrate system systems, and how this interaction varies with changes in operating condition due to the systems nonlinearity. Then, the experiment is used to demonstrate how the constrained optimization of the predictive controller can be used to ef fectively regulate the system at operating conditions where that a nonlinear simulation of the system could be used to illustrate the capabilities of MPC, the implementation of the controller to a real system offers the unique opportunity to observe challenges that may be encountered in industrial practice. For example, the fact that different process models are obtained on different sessions suggests the occurrence of time-varying conditions, nonlinearity, valve stiction, and measurement noise. Furthermore, the fact that the students the theoretical concepts. The constrained MPC optimization problem is given as follows: [9-10] at each samp ling inte rv al k: min u yr i ki kk i ik kj k j j m t p 2 1 1 2 1 u () s ubject to p pr oce sss cons tr aint s Where y is the outputs prediction vector at interval k, r is u is the vector of the manipulated variables at interval k. The matrices i and j represent the output and input weighting matrices for the MPC algorithm and p and m are the prediction and control horizon, respectively. Another important feature of the experiment is that the controller has been implemented by interfacing a stan dard MATLAB MPC function through a LabView-based future to easily change the constrained predictive control ler based on linear models by other controllers that can be programmed in MATLAB, e.g. nonlinear or adaptive predictive controllers. An additional consideration favoring the use of a LabView-based interface is that previous experi ences with the MATLAB real time interface toolbox exposed limitations in controlling the sampling interval  whereas in LabView the sampling interval could be accurately controlled. Furthermore, the University of Waterloo has a license agree ment with National Instruments that naturally motivated the use of this software/hardware combination. The paper is organized as follows: Section 2 describes the chemical process and the hardware used in the experiment. Section 3 presents an overview of the LabView/MATLAB hybrid program created by the authors to interface with the process. Section 4 presents the experimental procedure used to perform the on-line testing of the constrained MPC controller. Section 5 presents the experimental results and the experience gained by the students following the experiment. Concluding remarks are presented in Section 6. 2. EXPERIMENT AL EQUIPMENT The double pipe heat exchanger (DPHE) consists of six sections of concentric tubing set out as shown in Figure 1. Each section of the pipe is made of steel of approximately  three sections, saturated steam is supplied to the outer tubes (3 a 0.3 m 3 storage/surge tank from which it is recycled to the Four type-T thermocouples are located in the process unit to measure the inlet and outlet temperature of the oil and the tap water, respectively. Similarly, stainless steel vane-type Figure 1. Double pipe heat exchanger apparatus.
Vol. 44, No. 2, Spring 2010 129 a gate valve. The pressure of saturated steam is controlled using a manually operated globe valve and is measured by a local pressure gauge. All sections of the heat exchanger are insulated to prevent heat loss and for safety. Figure 2 tion used to interface the process with a PC was based on National Instruments interface devices. 3. LABVIEW AND MA TLAB A HYBRID SOFTW ARE INTERF ACE The role of the software interface in the process labora tory is to bridge process signals, human-machine interfacing (HMI), and the management of external data. LabViews wire and block programming  facilitated the use of visually informative dynamic plots, dials, and data entry for use as an HMI tool. The implementation of the MPC calculations in the LabView environment was expected to be laborious, however. Thus, it was decided that MATLAB (TM) along with its MPC add-on toolbox  is a more suitable computational environment for performing the MPC calculations. The new key feature in LabView software is the use of a MATLAB Script Node block that allows for wiring of variables to and from the LabView programming environment into an m-script text-based command line format of MATLAB. Details regarding the state-space based constrained MPC al gorithm implemented in this work can be found in Bemporad, et al.,  and Maciejowski.  Figure 3 shows the graphical HMI developed for this experiment. As shown, it displays four x-y plots, mode selection switches, signal parameters, MPC parameters, and an MIMO Laplace transfer matrix pro cess-model format. A copy of the LabView code developed by the authors is avail able upon request from
Chemical Engineering Education 130 4. EXPERIMENT AL PROCEDURE Due to the length of the tests, the MPC experiment is involves closed-loop control tests. For safety, the students and the lab instructor are required to wear protective glasses and insulated gloves during the lab sessions. The electrical equip any contact between them. the oil and water temperatures at the outlet. Thus, four trans fer function models that describe the dynamics between the water and oil valve and the oil and water temperature must be the students design a series of tests based on step changes on the input variables following a two-factorial design. The procedure to perform the step tests is available online.  The graphical display in Figure 3 shows a snapshot of the evolu tions of oil and water temperature following a step change in the water valve from 90% to 40% of opening. From these graphs, the students are able to observe the effect of distur bances that may affect the process during the step tests, e.g. changes in the inlet water temperature. Likewise, they can notice the interactions occurring in the process. Based on the exponential nature of the step responses, the students conclude that the process can be approximated by a set of First Order plus Time delay (FODPT) models. These models are used by the constrained MPC algorithm [9-10] in the second lab session to calculate the moves in the oil and water valves (inputs) that will drive the oil and the water temperatures toward a To commence the MPC closed-loop control testing, the DPHE is brought to an initial steady state. Students con struct their own set point tracking and disturbance rejection intuition into the effect of different tuning parameters values Based on the results obtained from this session and discussion with the lab instructor regarding the closed-loop performance this experiment. Details on the experimental procedure to perform this test are available online. 5. TYPICAL RESULTS AND DISCUSSION: EXPERIENCES GAINED BY THE STUDENTS Figure 4 shows one of the set point tracking tests performed in the second ses sion of the lab. As shown, both the water and the oil temperature smoothly track the set point sig nals. The water and the oil valves are manipulated by the MPC al gorithm to reach the reference sig nals. The water valve, however, reached an input limit (60% of Figure 4. Closed-loop performance in the DPHE: Con strained MPC test for a set point change in both the oil and water tempera ture.
Vol. 44, No. 2, Spring 2010 131 valve opening) for a period of time and then returned back This result shows that the implemented constrained MPC algorithm works properly but it also demonstrates that the performance is limited by the presence of process constraints. Similarly, the controller performance-to-disturbance rejection is tested by closing for 350 seconds the steam valve that sup plies steam into the process. Figure 5 shows the case where this disturbance test is performed in open loop and closed loop. The performance is judged by comparing the sum of square errors for the water and oil showing a 35% and 65% improvement for water and oil, respectively, obtained with the closed loop system as compared to open loop operation. In the event that one or both manipulated variables reach and remain at constraints, offsets will occur in the controlled variables. The instructor explains to students that the offsets in the two controlled variables can be altered, one vs. the other, by properly selecting the output weight matrix to be different from the identity matrix. Upon completion of this laboratory, the students gain a number of practical and insightful experiences that can be categorized as experiential, analytical, and design oriented. Experiential: Students gain experience in multivariable control of a typical chemical process, i.e. a heat exchanger. This experiment evolved from a previous implementation that ran for several years and used an unconstrained MPC algorithm coded in a rudimentary DOS-based BASIC envi ronment. To ensure that the new imple mentation will enhance the students learning experience, an undergraduate student that is also one of the coauthors of this publication participated in the development of this project. Clearly, the key advantage of the new implementa tion is the ability to enforce constraints. In addition, in the earlier implementa tion, the graphical interface and the recording capabilities were very limited and the students could only manipulate tion test and the temperature set points in the closed-loop. With the new Lab View/MATLAB implementation, the students use a graphical interface that shows the complete process response to a particular change, they can manipu late the MPC tuning parameters during the control testing session, and they can record up to 3.5 hours of process data. Many of these software and HMI im provements were introduced following the suggestions of the coauthor that had experienced the lab as a student. In particular, the DPHE experi ment allows the student to learn about nonlinear processes, the techniques and the performance of model-based control algorithms that are typically implemented in the industry for multi Figure 5. DPHE response to a distur bance in the steam: a) oil response and b) water response.
Chemical Engineering Education 132 variable control. The real-time operation of a physical system also allows for the development of intuition into background noise and system disturbances inherent in a real system. The challenges posed by noise and disturbances become evident different valve positions to a linear FOPDT MIMO plant model. Moreover, Session 2 also shows that processes like the DPHE are challenging to control because they are highly nonlinear and show a high degree of variable interaction. This systems nonlinearity and interactions are evident in Table 1, FOPDT around different operating conditions. As shown in from 0.48 to 1.3 depending on the selected operating point, whereas for an ideally linear system the RGA should remain constant for all operating points. In view of these systems characteristics, one of the key challenges is to obtain process models that mimic the true process behavior and to select con troller tuning parameters based on the identified dynamic behavior. Another experiencebased observation is that due to the open-loop time constants being relatively large (10-15 minutes), the students found that the were somewhat lengthy. The waiting time was used for discussions about issues. Moreover, the stu dents that performed the lab at the beginning of the term and that were not previously exposed to the MPC theory found it somewhat challenging to study this theory on their own. This was partially addressed by mak ing available, at the early stages of the lab, the theoretical MPC principles in the manual pages and in the tutorial of the toolbox. Future planned improvements in this course include the development of questionnaires that must be answered during the lab experiment. Also, to further address the lack of theoretical background, the lecturer of the course is plan ning to deliver the fundamentals of the MPC theory at the beginning of the term. Analytical: Students gain experience about data set analysis. nonlinear optimization where they minimize the sum of the the model predictions. Addressing noise in the process data, solving the nonlinear optimization problem and obtaining meaningful process model parameters for each test are just a few of the challenges confronted by the student following During the second session of the lab, the students test the controller performance using different MPC tuning param eters. For example, they observe that increasing the weights on the movements of the manipulated variables produces a sluggish response with a large settling time. To verify the importance of interaction, the students are instructed to implement a decentralized control strategy using the MPC algorithm by setting the off-diagonal process model Figure 6. Closed-loop performance in the DPHE: Unconstrained MPC test for a set point change in both the oil and water temperature. T ABLE 1 RGA Analysis for the DPHE at Different Operating PointsOpen-loop Tests 11 Step tests 1 and 5 0.4827 Step tests 2 and 8 1.3006 Step tests 3 and 7 1.2603 Step tests 4 and 6 0.4724
Vol. 44, No. 2, Spring 2010 133 parameter gains to zero and testing a set point change in both outputs. Since this test generally results in closed-loop insta bility the students conclude that accounting for interaction is of utmost importance corroborating the need for a centralized multivariable MPC control strategy. Design-oriented: From the results obtained in Session 1, the students realize that the process is highly nonlinear (see Table nominal operating condition for the process in the second part of the lab. That is, if they select an operating region for which the process model parameters do not accurately represent the process and the proposed set point changes to be tested are close to the process operating limits, then MPC will perform poorly resulting in input saturation and outputs far away from their corresponding set-point values. Thus, operability considerations must be addressed when estimating the pro cess model parameters and the MPC tuning parameters. On the other hand, the students can also analyze the trade-offs when either considering or ignoring constraints in the input variables. Figures 4 and 6 shows the process response to a set point change when the MPC algorithm takes into account input constraints (Figure 4) and when it does not (Figure 6). As time to reach the reference signal when the inputs constraints in the MPC algorithm are active (Figure 4). Therefore, the students learn from this test that process limitations can dras tically affect the systems closed-loop performance and that process design considerations, such as process constraints, have a direct impact on the controllability of the process. The learning outcomes presented in this section were assessed based on both the discussions between the laboratory assistant 6. CONCLUDING REMARKS This paper presented an implementation of a linear con strained MPC in a process control laboratory at the Univer sity of Waterloo. Upon completion of this experiment, the fourth-year chemical engineering students are expected to appreciate the capabilities of MPC over conventional feed back controllers. From the experiment, the students conclude that a decentralized strategy for highly interactive processes such as the DPHE cannot provide a satisfactory performance as it was demonstrated in the lab and that an MPC control ler is more suitable for this task. Likewise, they conclude that the selection of the MPC parameters plays a key role in the closed-loop performance. The students also learn that formance and that care must be given to the selection of the operating point in a process. They also learn to appreciate that data analysis and relatively accurate models are essential in the development of a model-based control strategy. In sum mary, the DPHE experiment represents an educational and practical tool that shows the challenges usually involved in the industrial deployment of MPC strategies to multivariable processes with high degree of interaction and in the presence of constraints. The University of Waterloo operates a large co-operative program where students spend at least one term per year in industry. Many of the students that have been ex posed to MPC applications in their co-operative terms have expressed that the current experiment have offered them a unique opportunity to experience and understand the design and implementation of this advanced controller. ACKNOWLEDGMENTS The authors would like to acknowledge the Waterloo En gineering Endowment Fund (WEEF) and the Department of Chemical Engineering at the University of Waterloo for their REFERENCES 1. Seborg, D.E., T.F. Edgar, and D.A. Mellichamp, Process Dynamics and Control 2nd Ed., John Wiley & Sons, USA (2004) 2. Marlin, T.E., Process Control: Designing Processes and Control Sys tems for Dynamic Performance 2nd Ed., McGraw Hill, USA (2000) 3. Bequette, W., Process Control: Modeling, Design and Simulation, Prentice Hall New Jersey, USA (2003) 4. Joseph, B., C. Ying, and D. Srinivasagupta, A Laboratory to Supple ment Courses in Process Control, Chem. Eng. Ed. 36 (1) 20 (2002) 5. Long, C.E., C.E. Holland, and E.P. Gatzke, Experimental Air-Pressure Tank Systems, Chem. Eng. Ed. 40 (1) 24 (2006) 6. Rusli, E., S. Ang, and R. Braatz, A Quadruple Tank Process Control Experiment, Chem. Eng. Ed. 38 (3) 171 (2004) 7. Gatzke, E.P., E.S. Meadows, and F.J. Doyle III, Model Based Con trol of a Four-Tank System, Computers and Chem. Eng. 24 1503 (2000) 9. Bemporad, A., M. Morari, and N.L. Ricker, Model Predictive Control Toolbox (TM) 3, The Mathworks, Natick, USA (2009) 10. Maciejowski, J.M., Predictive Control with Constraints Prentice Hall, Great Britain (2002) 11. Ricker, N.L., Using MATLAB/Simulink for Data Adquisition and Control, Chem. Eng. Ed. 35 286 (2001) 12.
Chemical Engineering Education 134 Diffusion in mixtures is one of the most important molecular transport processes in the understanding of mass transfer operations in chemical engineer ing,[1, 2] and it is the fundamental phenomenon underlying chemical processes ranging from mass transport in living cells and separations to corrosion. Consequently, reports in the literature on chemical education have frequently fo cused on simple experimental demonstrations for diffusion in liquids. [3-11] Here, we describe an experiment based on Wieners method  for measuring one-dimensional diffusion in ternary aqueous solutions and incorporate several new features that provide an opportunity for students to practice computer calculations. The method relies on using refraction of laser light from the concentration gradient near a solvent-solution interface. Past reports[7, 10, 11] have used this optical technique for undergradu ate-level demonstrations of diffusion in a physical chemistry laboratory. In these past reports, diffusion of single solutes such as KCl or CsCl in liquids such as water or glycerol was examined experimentally to demonstrate one-dimensional diffusion in binary mixtures as well as the impact of solvent technique described in literature can be also used to analyze diffusion in ternary mixtures of simple species, namely KCl and sucrose in water. In the experiment described here, we chose mixtures of KCl and sucrose but the method can be extended to other solutes. In other work, we have explored the application of this method to solutions of linear and globular polymers.  KCl and sucrose, in addition to being inexpensive, have vari ous other advantages. For example, well-established literature A SIMPLE REFRACTION EXPERIMENT CECIL A. COUTINHO, BIJITH D. MANKIDY, AND VINAY K. GUPTA University of South Florida Tampa, FL 33620 Copyright ChE Division of ASEE 2010 ChE classroomexists on diffusion studies of these chemicals in water. [14-17] them conducive for a laboratory session, and the materials are nontoxic for students. Furthermore, transparent solutions with a stable solution-water interface can be easily prepared for both KCl and sucrose. From an educational perspective, the laboratory experiment described in this paper has many useful features. It demon strates to students how a molecular scale process manifests
Vol. 44, No. 2, Spring 2010 135 itself into macroscopic observables and provides them a visual (Figure 1), which is governed by the diffusion equation in one dimension. c t D c z A AB A 2 2 1 () New elements are incorporated in the implementation of the experiment to give students practice in skills such as digital capture and digitization of experimental data, plotting and baseline correction, familiarity with simple statistical con Excel TM. The experiment can also be extended to reinforce a variety of concepts ranging from simple statistical and error analysis to more advanced concepts of mass transfer such as the importance of cross-diffusional contributions in mixtures. Thus, the experiment is suitable for incorporation into the curriculum at various stages depending on the em phasis placed on the different aspects of the experiment. For example, it can be used as an introductory hands-on learning opportunity at a freshman or sophomore level in chemical biochemical engineering, physics, and physical chemistry to emphasize the mathematical and computational aspects of the experiment with lesser emphasis on diffusion theory. Alternatively, it can be incorporated in a chemical engineer ing laboratory at a junior or senior level in coordination with the typical course on mass transfer theory that covers topics of diffusion in binary and ternary systems. LABORA TORY DESCRIPTION A schematic of the experimental setup is shown in Figure 2. A 12 18 aluminum breadboard (MB1218 from Thorlabs, breadboard using the threaded holes, which is convenient as it makes the whole assembly portable. An inexpensive alter A 5 mW laser diode module (31-0508 from Coherent Inc., stand and clamp. It was fanned through a commonly available laboratory glass rod (OD=3mm) to create a thin diagonal band of laser light at 45 degrees to the horizontal to scan several depths in the cuvette simultaneously. We found that a laser leveler readily available from home improvement stores can be a reasonable inexpensive substitute to the laser diode. As a safety measure, students were explicitly warned to avoid direct eye contact with the laser beam. The distances A and B shown in Figure 2 were adjustable and could be used to on the screen. Solution (KCl and/or Sucrose) Solvent (water) Interface 0.00.51.0 Normalized Concentration Position Figure 1. (right) Schematic representation of initial step concentra tion gradient (solid line) and concentration prole as diffu sion of solute occurs across the interface (dashed line). Figure 2. (below) Schematic representation of experimental arrangement and the optical deection curve.
Chemical Engineering Education 136 In our setup, the distances A and B were approximately 13 cm and 50 cm, respectively. A rectangular refraction cuvette 60 30 water used in experiments was obtained from an EasyPure removed particulate matter. The ternary solutions studied contained a total of 10wt% solute in the following ratios: (A) 25% KCl, 75% sucrose; (B) 50% KCl, 50% sucrose; (C) 75% KCL, 25% sucrose. In addition, experiments were also performed with binary mixtures containing 10wt% of either KCl or sucrose. Initially, the cuvette was raised or lowered using a laboratory jack until the air-solution inter face (marked by a discontinuity) could be located on the screen. The experiment was carried out by carefully pouring the solvent (water) upon the aqueous mix ture (KCl and/or sucrose) in the cuvette to create a sharp bound ary (solid line in Figure 1). The diffusion of the solute molecules from concentrated solution into the water phase with time leads to a decrease in the sharpness of the interface (dashed line in Figure 1). As a consequence, the refractive index gradient at the boundary changed with time and a skewed Gaussian curve developed on the screen as the beam traversed the mix ture-water interface. Over time, diffusion of the solute caused the optical curve to become broader and smaller in peak height. The optical trace was projected on a screen (a box covered with blank paper or graph paper) and was photographed using a digital Web camera (Rosewill RCM32301 from Newegg, Whittier, tervals of time. A free software (Automouseclicker18, Version 2.10) allowed automatic image capture with the Web camera at pre-determined intervals by clicking the capture button for the camera. THEORY At any particular time during the diffusion process, the mixtures of a single solute species in a solvent this gradient is best described by the Gaussian function : n x n Dt t xx Dt t o 4 4 0 0 2 0 ex p () 2 o represents the difference in the refractive index contrast between the aqueous solution of the solute (KCl or 1cm y (cm) x (cm) t=9min t=17min t=28min t=42min t=67min 8 6 4 2 0 y (cm) 8 6 4 2 0 x (cm) y y curve baseline x curve (A) (B) (C) Figure 3. (A) Image of the optical deection curve illustrating the digitization procedure. (B) Digitized curves at different times. (C) Plot illustrating the baseline correction procedure.
Vol. 44, No. 2, Spring 2010 137 sucrose) and water at the initial time (t 0 o depends on the initial weight fraction of the solute and is typically available in literature  or can be measured using sion of Eq. (2) for the overall gradient in refractive index. n x nz Dt t xx Dt t o 4 4 0 0 2 0 ex p nz Dt t xx o 4 0 0 ex p 2 2 0 4 3 Dt t () ) for each species are scaled by the relative composition in solution using fractions (z z ). It should be noted that the implicit assumption in Eq. (3) is that the individual species in the mixture diffuse independently of each other. The non-zero contribution of cross-diffusional terms in multi-component mixtures is well known in research literature.[15-17, 20-22] Neglecting ate laboratory experiment, however. This simplifying assumption keeps the data analysis simple and introduces relatively small error for the case of 10wt% total solute concentration. DA T A ANALYSIS AND RESULTS be digitized with respect to an arbitrary origin. Use of digitizing software provides an and plots, which is a skill that can also be useful in other contexts when students need to perform data analysis using published literature. Digitizing software widely available from the World Wide Web can be used ( e.g. shareware license products such as Data thief TM TM, Graph Digitizer Scout TM or open source products such as Engauge Digitizer). The use of this software typically requires importing the location of each mouse-click and provides a table of coordinates that can be imported imposed on a graph paper and recorded into a spreadsheet. This approach rapidly becomes laborious, however, when several curves have to be analyzed and also does not promote the use of computational tools. Figure 3B shows a collection of the digitized curves at various times during the decreasing peak height. The skewed nature of the curve is corrected by subtracting the linear baseline from each point (Figure 3C) and plotting y(=|y curve y baseline|) vs. x (=x curve). Since the y-coordinate is directly proportional to the refractive index gradi 1 and K 2 describes the dependence of y on position and time as yx tK n x Kn Dt t Kx x D o e xp 1 1 0 20 2 4 4 t tt 0 4 () K 1 and K 2 refraction of the laser light by the acrylate cuvette wall. Prior to characterizing the ternary systems, the scaling factors K 1 and K 2 were determined experimentally using only KCl in water (in an under graduate laboratory, this step may be completed beforehand by a teaching assistant). Data was collected using 10%, 15%, and 23% solutions of KCl. For each solution, the value of o was taken from literature and as a function of both variables for a common set of parameters K 1, K 2, x 0, t 0, and DKCl using least squares TM using the Solver add-in. The average K 1 and K 2 values of these runs were used as constants in all further experimental studies. As a check, the binary diffusivity of sucrose in 10% solution in water was also measured. Table 1 shows that there is a good agreement between the measured binary diffusivity val ues for KCl and sucrose with those found in literature. In the case of ternary mixtures of KCl-sucrose-water, the total solute concentration was held constant at 10wt% and relative ratio of KCl and sucrose (z :z ) in the solution was TABLE 1 Substance D(cm 2 /s) D(cm 2 /s) (from literature  ) KCl a 1.81 10 -5 1.87 10 -5 Sucrose b 4.45 10 -6 4.77 10 -6 a n o(KCl) = 1.215 10 -4 c o(g/L)+1.3334 b n o(sucrose) = 1.601 10 -4 c o(g/L)+1.3318 using literature data  Use of digitizing soft ware provides an op portunity for students extracting data from plots, which is a skill that can also be useful in other contexts.
Chemical Engineering Education 138 (y ,y yx tK n x Kn z Dt t Kx x o e xp 1 1 0 20 4 2 0 1 4 4 Dt t Kn z Dt o t Kx x Dt t 0 20 2 0 4 ex p () 5 o,D) were held constant for both species while the parameters (z z x 0, and t 0) were changed. The optical curve at several different times was regressed TM using the Solver add-in (illustrative worksheets are available at
Vol. 44, No. 2, Spring 2010 139 10. Rashidnia, N., R. Balasubramaniam, J. Kuang, P. Petitjeans, and T. Fluids Using Both Interferometry and Wieners Method, Int. J. Ther mophysics 22 (2), 547 (2001) 11. Sattar, S., and F.P. Rinehart, Diffusion of CsCl in Aqueous Glycerol Measured by Laser Refraction: A Physical Chemistry Experiment, J. Chem. Ed. 75 (9), 1136 (1998) 12. Wiener, O., Darstellung gekrummter Lichtstrahlen und Verwerthung derselben zur Untersuchung von Diffusion und Warmeleitung, Annals of Phys. Chem. 49 105 (1893) 13. Mankidy, B.D., C.A. Coutinho, and V.K. Gupta, Probing the Interplay of Size, Shape, and Solution Environment in Macromolecular Diffusion Using a Simple Refraction Experiment, J. Chem. Ed. (in press) 14. Jamshidi-Ghaleh, K., M.T. Tavassoly, and N. Mansour, Diffusion J. Physics D: Applied Physics 37 (14), 1993 (2004) 15. Kim, H., and G. Reinfelds, Isothermal Diffusion Studies of the WaterSucrose-Potassium Chloride System at 25.deg, J. Solution Chemistry 2 (5), 477 (1973) 16. Kim, H., G. Reinfelds, and L.J. Gosting, Isothermal Diffusion Studies of Water-Potassium Chloride-Hydrogen Chloride and Water-Sodium J. Physical Chemistry 77 (7), 934 (1973) 17. Reinfelds, G., and L.J. Gosting, Measurements of Isothermal Diffusion at 25 Deg with the Gouy Diffusiometer on the System Water-SucrosePotassium Chloride, J. Physical Chemistry 68 (9), 2464 (1964) 18. automouseclicker
Chemical Engineering Education 140 There is increasing pressure on the manufacturing indus tries around the globe to meet new tougher demands and regulations.  Higher product quality, expensive raw materials, larger production volume, environmental and safety regulations, global economy, and other factors have forced industries to rethink the way manufacturing is executed. Process control or automation is a tool that can be employed by companies to deal with these challenges. Therefore, the de mand for people well-educated in process control, especially in chemical-based processes, is increasing. Therefore, univer sities must also rethink the way process control is taught. The Experiential Learning  theory establishes that learn ing is a cycle that begins with experience, continues with conceptualization, and active experimentation steps, in that order. This means that the learning process is enhanced with hands-on activities  in which teams of students  act on the fundamentals. Based on the teaching needs for process control, the Chemi cal Engineering Department of the University of Puerto Rico at Mayagez (UPRM) is tackling the challenge of modifying the material taught in the classroom and including hands-on CLOSING THE GAP BETWEEN PROCESS CONTROL THEORY AND PRACTICECARLOS VELZQUEZ, 1 NELSON CARDONA-MARTNEZ, 1 AND EDWIN VELZQUEZ 21 University of Puerto Rico at Mayagez Mayagez, PR 00681 2 Automation Technologies Isabela, PR Copyright ChE Division of ASEE 2010 ChE
Vol. 44, No. 2, Spring 2010 141 experience with real industrial control systems and industry practices. The new approach at UPRM integrates process con trol theory, where all the basic and indispensable concepts and rationale are discussed, with a unique real practice of chemical process control. This paper describes the changes in the course material, the infrastructure to support the innovation, and the hands-on project. The main feature of the project is the use of real industrial technologies and practices to ensure a rookie engineer has a solid basis in process control.MODIFICA TION OF COURSE MA TERIALA survey of process control course syllabi demonstrates that the typical teaching method used consists of lectures on introduction to process control; principles-based modeling of processes, sensors, and actuators; stability analysis using sev eral techniques; control loop tuning; cascade; feed forward; and maybe an additional strategy if time allows. Most of the process control textbooks are written with a large focus on these topics [5,6] including, in some cases, material related to control practice and standards. At UPRM, the course has been aligned with a recent trend of several textbooks [7-9] that orient the course toward more practice experience. Table 1 presents the syllabus established to accomplish this alignment; it can be seen that the course starts with control practice topics including laboratory work, real-life example, and seminars offered by an industry expert. This mate rial is followed by a reduced portion of the use of mathematical concepts to support the real application of control. In parallel, the students work in the innovative hands-on experience with an industrial control system (more details ahead). During the semester the students dedicate 45 hours to classroom time, plus the corresponding time for the exams, plus approximately 35 hours to complete their corresponding tasks for the special project. Therefore, the students dedicate on average eight hours per week to this course, which is taken course is taught both Spring and Fall semesters to typically between 45 and 60 students. grade. The remaining 25% corresponds to the special proj ect, which most of the time is a group grade. The professor reserves the right to compensate or penalize the student for however. The objective with the change in material content is to familiarize students with basic experience in most of the T ABLE 1 Organization of C ourse T opics and T asks Hours Course Topics Project T asks (executed out of the class period) 2 Introduction to Process Control 1 Discussion of project, formation of teams and workgroups 1 Project Management 1 Basic Components of Control Systems Sensors Seminar 1: Validation (3 hrs) 2 Basic Components of Control Systems Actuators Seminar 2: Distributed Control Systems (3 hrs) 2 Basic Components of Control Systems -Controllers Preparation and hand-in of Gantt chart (3 hrs) 1 Discrete Control, Boolean Logic Training with assigned operation and presentation to show how to use the equipment (5 hrs) 1 Control Design (P&ID and SAMA) 5 Example First Partial Exam 3 Modeling of Dynamic Systems Balances Dynamics Simulation Progress report by each of the workgroups: process modeling, interface, control algorithm, and validation (5 hrs) 2 Process Parameter Estimation Hints for experiments Execution by each workgroup of corresponding tasks. (Industry-expert support is provided) (10 hrs) 3 Design of Single-Loop Feedback Control Systems Meeting between workgroups and progress evaluators (1 hr) 1 Tuning of Feedback Controllers Completion of modeling task (5 hrs) Second Partial Exam 2 Cascade Control Completion of interface (5 hrs) 1 Ratio control Completion of control algorithm and experiments (5 hrs) 2 Feed forward Completion of validation process and demonstration of performance of the controlled operation (5 hrs) Third Partial Exam
Chemical Engineering Education 142 issues concerning automating real manufacturing operations but maintaining the basic concepts. Students are exposed to issues such as communication protocols between accessories, integration of software, tuning of real controllers, industry standards, and validation of processes and systems. In the end, the student should be able to connect the control practice with control fundamentals. The deliverable is a cluster of students aware of the issues of hardware implementation, control strat egy selection, and process understanding. Therefore, they are able to contribute more to their employers from Day One of being hiredwhich contrasts sharply with the current situa tion, in author Velazquezs experience, in which newly hired engineering graduates need from six to 12 months to acquire enough experience to start contributing to companies.DET AILS ON COURSE MA TERIALThe course begins with students visiting the laboratory to see the industrial control system and the sensors and actua view of the control of one of the operations of the industrial control system. The course continues with lectures describing characteristics of sensors, actuators, typical communication loops ( e.g. PID, discrete, dead band) for a chemical process. These control loops are designed and represented through Apparatus Marketing Association drawings) as well as the well-known Process and Instrumentation Diagram (P&ID).  lectures are enhanced with experiences and practical details and aspects of implementation of a process control project from process control engineers. The main idea is to provide students with as much knowledge as possible of real-life applications, such as control logic, for safety of humans and material covered is administered at this point. The second part of the semester is focused on the fundamen tals of control. The topics include 1) modeling of processes (low-order transfer functions), actuators, and sensors using empirical data; 2) closed-loop transfer function and stability; and 3) tuning. For the modeling of the process including the sensors and actuators, the students perform experiments using transfer functions. For this task, they use the graphical method but they could also use Matlab or Excel. The material offered in the classroom comes directly from the textbook and is enhanced with control practice details especially for the tuning part. This is followed by another individual exam. The third part then focuses on cascade, feed forward, and ratio control, if time permits. The main idea here is to guide students to learn when and how to implement these strate gies to improve the strategies learned before. This objective is basically the same as used in textbooks. A third individual exam is administered after these last topics are covered. In summary, the students should have learned practical aspects for a process control project, the basic feedback control strategy and its practical aspects, and three additional strategies designed to enhance the basic feedback strategy, all along with the hands-on project.DESCRIPTION OF THE HANDS-ON PROJECTThe project starts early in the semester by dividing the subdivided into four working groups. Each working group is then assigned one of four tasks: 1) modeling of the assigned operation; 2) control loop design, implementation, and tuning; 3) control interface; and 3) hardware and software validation. The last two tasks come from the control practice in industry. Each team is provided with a scope-of-work document that describes the project assigned and the objectives, the hardware and software available, and the requirements for grading. ment techniques to prepare a Gantt chart of the remaining tasks to achieve the desired scope, including the overall deadline for the entire project. For this, either the instructor or an expert from the industry (preferred) lectures on projectmanagement basics. Before students start working with the system, additional seminars and workshops are offered in process systems validation. Typically, two or three industry experts help us with the seminars and the direct support to students. Another requirement is that students must demonstrate to the instructor that they know how to run the particular unit operation and the control system. For this, each team visits the unit operation laboratory and the control room to famil iarize themselves with the different accessories, and gather information on how to run their operation. The deliverable is a cluster of students aware of the issues of hardware imple mentation, control strategy selection, and process understanding. Therefore, they are able to contribute more to their employers from Day One of being hired.
Vol. 44, No. 2, Spring 2010 143 After that, the group in charge of modeling prepares the procedure to generate and collect the adequate dynamics data for the low-order transfer function. At the same time, the group in charge of control loops designs the different loops through the SAMA drawings and prepares the P&ID drawings. The group in charge of the interface must collect information from the other groups to design the interface. At this point the four working groups must hand in a progress report, which should include the dynamics data and the model of the working groups of each team should communicate with each other to ensure each working group has any information required from the other working groups so that the work can This interaction is captured in Figure 1. The project continues with the implementation of the inter face and the control loops. During this period, several control experts from local system integration companies coach the students. This approach is similar to the mentoring approach used by Kavanagh and Crosthwaite.  Once everything is programmed, the entire group must run experiments in the automated system. The experiments must include at least a step change in set point and one disturbance. The student must characterize the performance of the control system using the standard criteria taught in class such as overshoot and decay ratio. In parallel, the validation group, which at this point should have prepared the validation document, executes it, collect ing data from the other working groups. After this, the four handed in by the deadline. The project implementation follows an identical project implementation life cycle to projects currently implemented in industry, to make this experience as valuable as possible. As can be deduced from the above description, the students must employ project-management techniques, prepare prog ress reports, have project status meeting between the students from each working group and the professors, and in some cases work in interdisciplinary environments. The interdisciplinary environment is simulated by includ ing in the teams students from electrical engineering who are pursuing a specialization in process control. This experience is typically done only during the Spring semester. At the project completion, an open house is coordinated sometimes with industry leaders to give the students the opportunity to present and discuss their projects with future employers and professional partners. This exposition to indus trial representatives also gives the university an opportunity to get industry feedback in terms of the latest trends and future industrial requirements, in order to continuously focus the CONTROL TECHNOLOGY A T THE LABORA TORYThe infrastructure to support this innovation consists of a control room in the unit operations laboratory, which houses two industrial process control systems identical to the ones currently used in the bulk chemical processing industries. One control system (DeltaV from Emerson Process Management) consists of the controller, a 24V power supply, three analog input cards, one analog output card, one discrete input card, istrative computer, and two workstation computers. The other control system (PCS7 from Siemens) consists of the controllers, the power supply, two analog cards (input, output), and two discrete cards (input, output). This system and the communication cards. Five unit operations are connected to the systems: 1) a cool ing tower, 2) a chemical reactor, 3) a distillation column, 4) a heated tank and level control, and 5) a heat exchanger. The cooling tower has three industrial pneumatic control valves and a variable frequency driver as actuators. In addition, it has three industrial RTDs (resistance temperature device) to measure the air inlet temperature, the water inlet temperature and the air outlet temperature. With these devices, there are four control loops: 1) air inlet temperature, 2) water inlet temperature, 3) water outlet temperature, and 4) air outlet temperature. The heated tank and level control apparatus has two con Figure 1 Interaction of working groups under the supervision of a team leader.
Chemical Engineering Education 144 apparatus has an RTD, and for the level it has an industrial pressure cell. Figure 2 depicts an example of the interface the students de veloped for the heated tank. They used the symbols provided by the controller software, which are the same they would use or see if they were working in a company. The heat exchanger also has two industrial pneumatic rate. An RTD is installed at the exit of the heat exchanger for meter at the entrance is used for the control loop of the water The chemical reactor has two control valves; one is elec an analog pH meter. The pH meter is to control the outlet (electronic) for the feed rate, and one (pneumatic) for the relay connected to the heating device of the boiler to manipu late the heat supplied to the solution. The output variables (temperature at the top and bottom plates) are measured with RTDs. One additional component of the control infrastructure is the software called PI from OSIsoft, which is designed to collect data from industrial control systems. This software allows students to transfer their raw data from the historian of the control systems to spreadsheets like Excel. Once the data is transferred to the spreadsheet, the student can use all of the features of the spreadsheet to compute many different values and prepare plots. The software is installed at a server connected to the university network so that the students can access the data from any computer in the university. To facilitate the availability of experienced process engi neers for class lectures or support, a virtual classroom with videoconference capabilities has been implemented. This permits colleagues from industry to interact with students directly from their respective industrial sites without aban doning their working areas. Students receive the lectures or suggestions in real time and they are able to see, hear, and interactquestioning and clarifying doubts with their virtual professor at their regular class time.IMP ACT ON THE STUDENTS 2004 and since then it has transformed approximately 150 chemical engineering and 30 electrical engineering students. Many of these students have used experience from the project in their jobs. In some cases, the students have been the leaders in automation projects of several manufacturing operations Table 2 presents the average, the maximum, and the mini As can be seen, the semester average after 2004 (75 pts) is 11% higher than the average before 2004. The maximum has increased substantially and the passing percentage has been higher, too. After 2004, withdrawals have been zero, which suggests that the students are more motivated to try until the end even if the grades are not too encouraging. T ABLE 2 Grade Distribution Before and After the InnovationSpring 1998 Spring 1999 Spring 2002 Spring 2003 Spring 2004 Spring 2006 I Spring 2006 II Spring 2007 Spring 2008 Avg 69.0 52.8 72.6 73.3 65.6 77.8 82.0 71.2 77.0 Max 81.1 89.4 91.9 92.3 81.6 92.0 96.8 95.9 95.0 Min 42.7 17.2 52.1 25.0 21.4 54.5 64.2 52.0 48.0 Pass % 95.0 77.5 87.5 93.1 72.2 88.9 96.9 95.0 90.9 W 4 2 Figure 2. Example of interface.
Vol. 44, No. 2, Spring 2010 145 Looking at the individual exams under the new scheme, standing of the fundamentals), it can be seen that the students are still earning similar grades as before. The reasons stem from the fact that the change described herein was aimed at modifying the material offered in the classroom and providing the students with hands-on experience with a real industrial control system. Independent of the reason, this provides another opportunity to improve the course once more, by in corporating a strategy to strengthen learning of fundamentals using the very same special project. During the Spring semester, as mentioned above, students from both departments, electrical and chemical engineering, work together in the project. The dynamics between the students is similar to the one that develops in industry. This interaction helps the students with their interdisciplinary skills, which is one of the outcomes required by ABET (Ac creditation Board for Engineering and Technology). Table 3 describes the outcomes that this innovation in process control teaching and learning impacts. Most of the students do enjoy working with the project and see the value of the approach. Even the teaching of the course is more interesting, from the instructor perspective, especially since it allows the professor to get involved in the project with the students as they develop it during the semes ter. Many students comment after completing the project that it was a great experience and that they would have preferred to spend more time in the project to go deeper and gain more value out of the experience. Comments from industry profes sionals are very encouraging and supportive, too. Comments like Finally, a project that teaches hands-on experience to the students are heard from them. The feedback from students has been valuable to keep course. This helps students reduce their initial stress caused by a topic quite different from the core courses of chemical engineering. The students are more motivated to take the course when compared with those that received or are receiv ing the classical teaching approach.CONCLUSIONSThe teaching of the course since 2004, although more de manding on the professor, has been more interesting than in previous years. Most of the students enjoy and appreciate the project and most of them improve their opinion about process control as the semester progresses. The industry has been very supportive and consistently has considered the approach very innovative and of positive impact for them. This approach not only provides practical experience in the visualization of the practical applications of all the theory learned. It also stimulates the students to continue with their careers, as there is a direct association of the theory learned and the future use of this knowledge in their professional careers. The teaching of standards of industrial process control implementation and the experience acquired by implementing the projects complement the theoretical knowledge and help the students visualize and value all the theory learned. Also, the project implementation experience helps students to de velop other important skills for their future professional lives, such as: project management, time management, presentation skills, leadership, work under pressure, and documentation and coordination between multidisciplinary groups. This project strongly supports the ABET outcome list. One issue with the current approach is the time the students spent on the class and the number of credits received. The ground on which to start their careers, outweigh the issue of no proper credit recognition, which can be addressed admin presented here could mean a greater contribution for industry in general and, even more important, for the career of young, new engineers.ACKNOWLEDGMENTThe authors are grateful of the support of Abbott Labo OSIsoft, INDUNIV (local industrial group), Air Products, T ABLE 3 ABET Outcomes That the New Innovation Impacts1. An ability to apply knowledge of mathematics, science, and engineering. 2. An ability to design and conduct experiments, as well as to analyze and interpret data. 3. An ability to design a system, component, or process to meet desired needs. 4. An ability to function in multidisciplinary teams. 5. An ability to identify, formulate, and solve engineering problems. 6. An ability to communicate effectively. 7. Recognition of the need for, and an ability to engage in, lifelong learning. 8. An ability to use the techniques, skills, and modern engineering tools necessary for engineering practice. 9. Recognition of basic leadership skills.
Chemical Engineering Education 146Automation Technologies, and Invision Engineering in the development of the required infrastructure.REFERENCES 1. Clements, C., the Life Sciences Industry, August 2008,
Vol. 44, No. 2, Spring 2010 147METSTOICH KELVIN W.W. WONG AND JOHN P. BARFORD The Hong Kong University of Science and Technology Clear Water Bay, Kowloon, HONG KONG Biochemistry is one of the important foundation courses in a biochemical engineering curriculum. It provides a basic introduction of cellular metabolism to engineering students to teach them how raw materials can be converted into valuable metabolic products by microor ganisms in various bioprocesses. Teaching metabolism in biochemical engineering courses normally adopts the tradi tional biochemistry approach. Students are presented with a number of reaction pathways that make major cell components ( e.g. protein, RNA, DNA, lipids, cell walls) as well as major catabolic products using a qualitative description. Traditional chemical engineering courses, however, focus on product yield, selectivity, reaction rate, and reactor/process design. It is similar for biochemical engineers that product yield, biomass yield, and ATP yield are important parameters for bioreactor design. All these goals, if applied to a biochemical system, require a quantitative knowledge of metabolism. Therefore, a quantitative description in metabolism can complement the major skill base of engineering students and is more consis tent with the overall philosophy or learning outcomes of an engineering degree. As a subset of system biology, metabolic engineering fo cuses on the metabolism of one organism. It is the practice DNA technology along with mathematical analysis to opti mize genetic and regulatory processes within the cell. This desirable objective. [1-4] Metabolic Flux Analysis (MFA, also etc.), is a practical tool for understanding and analyzing metabolic pathways, pathway interaction, and control. Varma and Palsson for MFA, namely: 1) to quantify metabolic physiology, 2) to simulate and interpret experimental data, 3) to analyze meta bolic pathways for metabolic engineering, 4) to optimize cell culture medium, and 5) to design and optimize bioprocesses. MFA is an analytical tool developed based on stoichiometric network models,  and it is assumed that those metabolic other processes. Unlike simulations based on mechanistic models that require detailed enzyme kinetic data, MFA is used reaction pathway details and stoichiometry, growth metabo  Even if several restrictions are enforced, for complete me needs to be analyzed to accurately represent the interactions between the various metabolic pathways. Practically, such that simulate the metabolic networks. The analysis provided by MFA is also good for demonstra tion of the quantitative aspects of metabolism to students. Most analytical software packages, however, are developed for research purpose and mainly focus on pathway control ( i.e. metabolic control analysis, MCA). Metstoich was ini tially developed to focus on teaching metabolism and to link practical biochemical engineering parameters with metabolic ChE classroom Copyright ChE Division of ASEE 2010
Chemical Engineering Education 148 EXISTING SOFTW ARE PACKAGESTo explore the physiological properties of biological sys tems, a system of equations must be solved. Such a task can be easily done with the aid of modern personal computers and metabolic engineering software packages. Some important and/or widely used software packages are: GEPASI 3 [7-8] is a widely used free biochemical reactions simulation software package. GEPASI simulates the kinet ics of biochemical reaction systems and provides functions such as metabolic control analysis (MCA), elementary mode version of GEPASI released was 3.30 in September 2002. COPASI  was developed based on GEPASI with different simulation techniques, optimization routine, etc. Jarnac (a.k.a. Scamp II)  simulates the steady state and transient behavior E-Cell  is an object-oriented, whole-cell simulation software package. MIST  performs dy namic simulations, stoichiometric calculations, and MCA. JWS On line  is an Internet-based meta bolic simulator with collections of several metabolic models, and it can provide MCA to analyze the simulation results. KINSIM [14,15] is a rate equation-based numerical simulator and it was used for the simulation of enzymatic reaction system kinetics. FluxAnalyzer  is a MATLAB package with GUI for stoichiometric analysis of metabolic networks. It can pro features detection, and pathway analysis. In one of the more extensive examples, Klamt, et al.  carried out a metabolic lyzer. This model involved 30 of the most important catabolic branchpoint-metabolites (intermediate metabolites to which at least three reactions are linked) and 41 catabolic reactions1 for growth rate, 25 for central metabolic pathways, and 7 for photosynthesis, cyclic electron transport during photosynthe sis, respiration, ATP synthesis, and maintenance. The model also involved 46 anabolic reactions using the stoichiometries presented in Neidhardt, et al. Except for FluxAnalyzer, all above simulation packages focus on the dynamic behavior of metabolic pathways. They require reaction kinetics as input and some of them can even perform metabolic pathway analysis such as MCA. Other than the above-listed software packages, there are many packages/projects developed or under development. Figure 1 summarizes part of the metabolic engineering soft ware packages/projects found. Most of them are ODE solvers, some of which can perform sensitivity analysis (MCA). Some, however, were developed for various purposes, such as: CellDesigner  is for gene-regulatory and biochemi cal networks. Cellerator  is a Mathematica package designed for modeling with automated equation generation. It was designed with the intent of simulating signal transduction. InNetics  was developed for genomic-based drug discovery. The JigCell project  explores the cell physiology from the scope of molecular regulatory networks.There are two trends for metabolic software development. T ABLE 1 Parameters as Input and Output for Various Calculation Modes in Metstoich Parameters Problem Types (a) Theoretical Y XS(b) Experimental Y XS(c) Y X/ATP(d) Experimental Y XS X/ATP(1) Cell composi tions Input Input Input Input (2) Glucose usage for energy gen eration process Input Input Input Input (3) P/O ratio Input Input Output Output Input Output Input Output (5) Y XS Output Input Output Input (6) Y X/ATP Output Output Input Input Figure 1 (left). Some existing metabolic engineering software packages/projects.
Vol. 44, No. 2, Spring 2010 149 pathway models. The second is the development of Web-based ap plications. Nowadays, the Internet is already part of daily life and Web-based applications are a good choice, especially for database projects to collect and share data. The common advantage of these packages is that you can input any model to the package for analysis. Their practical use for en gineering purposes, however, is limited and is not their primary purpose. They do not address issues of energetics and ATP usage, the production of biomass yield, etc.METSTOICHMetstoich was initially developed for teaching purposes [23-24] and S. cerevisiae.  Met stoich includes the following major pathways: 1) central metabolic pathways, such as glycolysis, tricarboxylic acid (TCA) cycle and pentose-phosphate pathway (PPP); and 2) biosynthetic pathways. The central metabolic pathways serve to provide precursors for bio synthetic pathways, and for generating energy (ATP) to support cell growth and maintenance. tribution among pathways with practical engineering parameters encountered in a standard biochemical engineering course, such as biomass yield (Y XS), product yield (Y PS), ATP yield (Y X/ATP), etc. branch points. There are several important inputs necessary for Metstoich to 1) Cell macromolecular composition; 2) Glucose distribution (usage) in central metabolic pathways for energy generation process; 3) P/O ratio; consumed in biosynthetic reactions; 5) Biomass yield, Y XS ; 6) ATP yield, Y X/ATP .There are four problem types that can be solved by Metstoich with above inputs:a) Calculation based on theoretical yield; or b) Calculation based on experimental biomass yield, Y XS ; or X/ATP ; or d) Calculation based on experimental biomass yield, Y XS and X/ATP .Table 1 summarizes a matrix of problem types, inputs, and outputs, and Figure 2 shows part of the input interface. Users can specify: the cell composition (Figure 2); carbon source and electron donor if CO 2 is the carbon source (Figure 3); electron Figure 2. Part of Metstoich input interfacebasic information and cell compositions. Figure 3. Part of Metstoich input interfacecarbon source.
Chemical Engineering Education 150 acceptor (Figure 4); and energetic issues of the microorganism of difference in percentage. The results generated by Metstoich are organized into several levels of detailed worksheets with biochemical detail and illustrative reaction pathways included to make it more understandable. Levels of organized results are: Cell Y ield and Energetics (Figure 7) This worksheet is the executive summary of the overall performance of the cell with the inputted common engineering parameters; Fate of Glucose (Figure 8) This worksheet summarizes Figure 4. (top left) Part of Metstoich input interface electron acceptors. Figure 5. (bottom left) Part of Metstoich input interface energetic and other parameters. Figure 6. (above) User can specify highlight values that changed larger than the given percentage.
Vol. 44, No. 2, Spring 2010 151 Figure 7. Cell Yield and Ener getics, the cell yield (either estimated or given), Y X/ATP amount of ATP generated directly from reac tions or oxidative phosphor ylation are summarized. Figure 8. Fate of Glucose, glucose directly linked with bio synthesis or energy genera tion process is analyzed.
Chemical Engineering Education 152 Figure 9. All detailed biomass compositions, such as amino acid, etc., are summarized in composition summary. Figure 10. All Detailed Reactions shows all biochemical reactions.
Vol. 44, No. 2, Spring 2010 153 Composition Summary (Figure 9) This worksheet sum marizes cell compositions and their detail; and All Detailed Reactions (Figure 10) This worksheet shows metabolic pathways used for either biosynthesis purpose or energy generation purposes.Metstoich already contains amino acid production pathways and it is capable of analyzing amino acid production. Since Metstoich already contains information on major catabolic and anabolic pathways, it is easy to further include more production formation pathways such as antibodies, biofuel, etc. and therefore enzyme concentrations, kinetic expressions, intermediate concentrations, and thermodynamics have not been incorporated. An extension of Metstoich that incorpo rates thermodynamics and reaction kinetics, etc., has been developed and reported. [26-28]The core calculation module of Metstoich is written using Microsoft Excel 2002 with VBA Macro. This core Excel module map. The front-end graphical user interface was written in Visual Basic. Metstoich runs on Microsoft Windows 98, 2000, XP, and This is an example problem that students undertake as an exercise. It is taken from a number of problems included in the Metstoich package:The biomass composition (weight %) of a given yeast is as follows: Protein = 39%, DNA = 1%, RNA = 11%, Lipids = 3%, Phospholipids = 5%, Cell Wall = 38%, and Ash = 3% For energy generation, 10% glucose is used by pentose phosphate pathway, 60% glucose is used by the TCA cycle and 30% glucose is used by the fermentation pathway. The reported biomass yield is 0.4 g-biomass / g-glucose and let P/O ratio be 2.2 mol-ATP / mol-NADH. What is the corre sponding Y X/ATP between P/O ratio and Y X/ATP ?Since Y XS with P/O ratio are given, the Experimental Y XS calculation mode should be used. With given input values, Metstoich returns Y X/ATP = 7.85 g-biomass / mol-ATP and ATP X/ATP and P/O ratio is shown in Figure 11 at various P/O ratios: XS and cell compositions, glucose directly g-biomass. The total glucose consumed is 2.5 g-glucose / gbiomass for the given Y XS = 0.4. Therefore glucose consumed to generate energy is always 1.07 g-glucose / g-biomass, and it always generates 17.3 mmol-ATP and 50.1 mmol-NADH per 1.07 g-glucose consumed in assigned pathways. Therefore, total ATP generated in energy generation process = ( 17.3 + Figure 11. Relationship between Y X/ATP and P/O ratio.
Chemical Engineering Education 154 P/O x 50.1 mmol-ATP ) / 1.07 g-glucose. And Y X/ATP = 1 gbiomass / total ATP generated in energy generation process. It is suggested that normal Y X/ATP is around 10.5 g-biomass / mol-ATP. Using the Experimental Y XS and Fixed Y X/ATP calculation mode, it is found that the P/O ratio = 1.56 molFigure 14 to understand the quantitative use of glucose by the cell and how much energy had been generated. By combin ing Figure 13 and Figure 14, students can generate an overall COMMENTS ON METSTOICHProfessional evaluation was undertaken by Learnet of Hong Kong University. Metstoich had been reviewed by four lead ing academics in biochemical engineering from the U.K., the United States, Australia, and Singapore: Prof. D. Bogle from University College London, Prof. L. Nielsen from University of Queensland, Prof. D. Trau from National University of Singapore, and Prof. P. Fu from the University of Hawaii at Manoa. It was considered an excellent tool for learning of major biochemical engineering concepts such as Y X/ATP, yield, also highly regarded (Figure 6). In general, Metstoich has different aspects such as interface design, quality of content, and learning potential.METSTOICH AS TEACHING T OOLMetstoich has been applied in biochemical engineering and biochemistry classes at HKUST and it has been rated as easy to use by students. Students have been interviewed by the Center of Enhanced Learning & Teaching (CELT) of the Hong Kong University of Science and Technology (HKUST). It is agreed that Metstoich is easy to use, since the help functions and labels and buttons of the software are clear. The advantage of Metstoich is it can compare two sets of calculated results by highlighting the difference. Students felt that Metstoich contained too much information, however, since it covers from networks of reactions to energetics and cell yield, etc.CONCLUSIONEngineering students are accustomed to quantitative con cepts from their foundation courses. Biochemistry can also be taught quantitatively and when this is done, engineering students can appreciate the importance of metabolism in understanding and optimizing bioprocesses. Metstoich, a metabolic calculator for teaching purposes, was developed to introduce metabolism to students using quantitative prin ciples. As such, it is useful to both engineering students and biochemistry/life sciences students, who normally do not have strong backgrounds or training in quantitative methods. Metstoich has many novel features:1. Linking practical engineering parameters with cell growth, product yield, energetics, etc.; 3. Calculating how many nutrients are required for cell growth.Such analysis can provide useful information about how product yield is related with biomass yield, cell energet ics, etc. Students can explore different metabolic options and are challenged to further explore their relationship to bioreactor/medium design. The package has been well received by both academic ex perts in biochemical engineering and undergraduate chemical engineering and biochemistry students at HKUST.ACKNOWLEDGMENT port (Project : HKUST 00409E) of the Center For Enhanced Learning & Teaching (CELT) as well as their technical par ticipation in the project.REFERENCES 1. Nielsen, J., Metabolic Engineering: Techniques for Analysis of Targets for Genetic Manipulations, Biotech. Bioeng. 58 125-132 (1998) 2. Olsson, L., and J. Nielsen, The Role of Metabolic Engineering in the Improvement of Saccharomyces cerevisiae : Utilization of Industrial Media, Enzyme Microb. Technol. 26 785-792 (2002) 3. Stephanopoulos, G.N., A.A. Aristidou, and J. Nielsen, Metabolic Figure 12. Glucose used for biosynthesis and energy gen eration purposes, drawn based on Metstoich results.
Vol. 44, No. 2, Spring 2010 155 Figure 14. (below) Fluxes among central metabolic pathways for energy generation purpose, drawn based on Metstoich results. Figure 13. (left) Fluxes among central meta bolic pathways for biosynthesis, drawn based on Metstoich results. Engineering, Principles, and Methodologies Academic Press (1998) 4. Lee, S.Y., and E.T. Papoutsakis. Ed., Metabolic Engineering, Marcel Dekker, Inc. (1999) and Practical Use, Bio/technology 12 994-998 (1994) 6. Weichert, W., Modeling and Simulation: Tools for Metabolic Engineering, J. Biotech. 94 37-63 (2002) 7. Mendes, P., Biochemistry by Numbers: Simulation of Biochemical Pathways with Gepasi 3, Trends Biochem. Sci. 22 361-363 (1997) 8. Mendes, P.,
Chemical Engineering Education 156 13. Olivier, B., and J. Snoep,
Vol. 44, No. 2, Spring 2010 157 According to Accreditation Board for Engineering and Technology (ABET) outcome 3d, graduates from engineering programs must have the ability to func tion on multidisciplinary teams.  Unfortunately, evaluation shows that engineering students are not well positioned to understand new concepts from a variety of disciplines and integrate them into what they learn in their own disciplines.  This is especially true for concepts in emerging areas, such as life sciences. Obviously the emergence of technological breakthroughs in new arenas is stimulating faculty members to include related multidisciplinary concepts in their course designs so that students can be prepared to meet the industrial challenges presented in applying new technologies within industrial settings. Motivated by the lack of appropriate tools that can be used to teach chemical engineering undergraduate students, especially to teach them how to integrate life sciences and THE MICROBIAL FUEL CELL AS AN EDUCA TION TOOL ChE ALIM DEWAN, BERNARD VAN WIE, AND HALUK BEYENAL Washington State University Pullman, Washington 99164-2710ZBIGNIEW LEWANDOWSKI Montana State University Bozeman, Montana 59717-3980 Copyright ChE Division of ASEE 2010 engineering concepts, we have developed the microbial fuel cell education module (MFCEM), a hands-on learning mod ule that can be used for learning multidisciplinary concepts in an active group-learning modality. This module uses the principles of mass and energy conversions applied in a mi crobial fuel cell (MFC) to integrate various concepts taught in biology, chemistry, electrochemistry, and engineering. In this paper, our goal is to show how the MFCEM can be used as an aid in teaching a senior-level course in chemi cal engineeringIntroduction to Bioprocess Engineering (ChE 475). Figure 1 shows the components of the module and how we implemented it in the classroom. Figure 1. Methodology for implementing the microbial fuel cell education module.
Chemical Engineering Education 158 MICROBI A L FUEL CELL E D UC A TION MO D ULEClass lectures were used to introduce the theory of various processes important to understanding microbial respiration, the thermodynamic and kinetic principles of the processes involved in energy conversion in MFCs, and the basic calculations used in electrical engineering ( e.g., current and power). The hands-on work consisted of the operation of an MFC in the laboratory. For the laboratory exercises, we prepared a manual to instruct the students about safety issues; equipment needed to run the T ABLE 1 The Concepts, Related Topics, and Mathematical Expressions Introduced Using the MFCEM Concepts Mathematical expression* ReferenceCellular respiration Microbial growth kinetics Metabolic pathway Electron transport chain Redox reactions in respiration Microbe-solid interactions Monod kinetics dX dt X S KS X Y S XS dS dt S KS S ma x ma x / [6,9] Electrode potential (Electrochemical equilibrium) Nernst equation EE M M EE AM o ox re d C o 0 059 2 0 059 4 0 2 lo g lo g g 1 2 4 pH o [6,7,10] Overpotential (Electrode kinetics) Butler-Volmer equation Electrode polarization ii F RT F RT 0 05 05 ex p ex p [6,10] Current Faraday constant Calculation of current from material balance and growth kinetics in MFC Faraday constant = electrical charge of an electron Avogadro constant I V R ex t  Power Differences between current & power and energy & power PV I V R IR ex t ex t 2 2  Charge conservation c t Id t Fn S M 0 [7,11] Energy conservation Material and energy balance for MFC E t ci n ex t t ci n VI dt Hm IR dt Hm 0 2 0 .  Sustainability Sustainability of power generation in MFCs [12,13] Details on the development of these equations, and example calculations using experimental data, were included in the MFCEM handout given to the students.MFC, with photographs; step-by-step procedures, also with photographs; sample experimental results; sample calculations; groups, ran the MFC, computed the energy conversion, and presented the results to their classmates. Debates, moderated by the instructor, on the results obtained by the various groups, were aimed at reinforcing the concepts discussed in the lectures. Last, we assigned a set of problems to test understanding of the concepts and to evaluate the role of the hands-on active experience in the classroom.
Vol. 44, No. 2, Spring 2010 159 The incorporation of the MFCEM into the bioprocess en gineering course, ChE 475, gave us an opportunity to teach the concepts through an active-learning process. Compared to standard lecture-based, passive learning, the hands-on activelearning process helps students to visualize and more fully think through what they learn and helps them to make connections between concepts that they learned before.  As claimed in the well-known learning retention pyramid,  students remember concepts best when they see a demonstration (50%), engage in a debate or discussion (70%), or have a chance to do some thing real and apply their knowledge immediately (90%). With the MFCEM we particularly emphasized practice by running experiments that were an immediate application of the in-class lecture and having the students prepare reports, perform home work assignments, and hold in-class debates with the active involvement of the other students in the audience, who asked questions or expressed opinions on one side or the other of the debate. The remainder of the material in ChE 475 was taught in a passive manner, with the professor lecturing and the students taking notes and completing homework assignments based on their notes and reading. We expected that the introduction of the retention of the topics in ChE 475 and of the multidisciplinary concepts introduced by the MFCEM.IN-CLASS LECTURE: THEORYIn the ChE 475 course, we used Bioprocess Engineering written by Shuler and Kargi (2002).  After completion of mentals needed to understand MFCs and we then introduced multidisciplinary concepts using the MFCEM. We do not discuss all the concepts in this paper because of space limitations; however, they were discussed in considerable detail in the classroom and in the MFCEM manual given to the students. While some of the concepts had been taught in previous courses, we reintroduced them so that students could connect the new concepts with previously learned concepts. The concepts, related topics, and mathematical expressions [6-8] that were introduced using MFCEM are summarized in Table 1. The topics were presented by asking a series of questions paragraphs present selected questions we asked and give a brief discussion of how they were implemented. microbes? Our theory section started with this question and it was actively discussed in the classroom in a group format to give the main idea behind the MFC. Then we introduced the concept of how a single cell grows, duplicates, and gains energy. We showed a single cell (Figure 2A) and described the main idea behind MFCs, which is separating the oxidation and reduction reactions in the respiration system. After show ing Figure 2A, we asked the students to build an MFC. The students worked to separate the two environments and make the MFC depicted in Figure 2B. This process helped them to understand the basic principle of the MFC, that of separating the oxidation and reduction reactions using a proton-exchange membrane (PEM) and connecting the two reaction environ ments through an external circuit. Later we discussed in detail and explained why we need to use a proton-exchange mem brane (Figure 2B). This helped the students better understand what a cathode and an anode are. They learned that oxidation happens at the anode and reduction at the cathode. How are the electrons transferred from bacteria to the solid electrode? The interaction of microbes with solid materials is a fascinating new topic, not only in MFC research but also in microbiology and environmental science. Electron transfer mechanisms were introduced and the students were taught why electrons cannot jump directly from microbes to solid surfaces, i.e. that electrons must be transferred by a redox reaction via: 1) a mediator, a chemical that accepts electrons resulting from the microbial respiration process and transfers them to the solid electrode, or perhaps 2) linkage of the microorganisms with the electrode surface by nanowires or cytochromes. The students were excited about these topics, which constitute cutting-edge research questions in microbiology. What are the source of and the sink for the electrons? This question was answered by revisiting the major metabolic path way concepts, taught earlier in the course. We showed how electrons are derived from the microbial respiration system and transported to an electron acceptor (in this case a solid electrode) through the electron transport chain. This used to be a mundane subject for the students, but now there was a Figure 2. The original idea of a microbial fuel cell de scribed using a single cell. (A) The redox reactionsoxi dation and reductionin the microbial energy generation process. (B) The separation of the oxidation and reduc tion reactions, using a proton exchange membrane, to build a microbial fuel cell.
Chemical Engineering Education 160 real-life application for the metabolic pathways and therefore they appeared to be engaged in the subject as evidenced by the energy demonstrated by students when they expressed their ideas and the content of their group discussions. To further facilitate discussions, the idea of using two different types of bacteria was introducedone using lactate and the other using glucose as the electron donor. What do we mean by the electrode potential, current, energy, and power of a fuel cell? The electrode potentials are thermodynamic properties and are calculated using the Nernst equation (Table 1). When current is passed through the electrode, however, the thermodynamic equilibrium does not exist anymore; rather the current needs to be calculated using the Butler-Volmer equation describing electron transfer kinetics. As a result of this discussion, the students improved their understanding of the differences between equilibrium and non-equilibrium processes. The concept of overpotential is vital to understanding batteries, fuel cells, corrosion processes, and electrochemical sensors. Using the MFCEM, students were introduced to the concepts of polarization curves (cell potential vs. current) and overpotential. In addition, they learned basic electrical engineering concepts such as electrical potential, current, power, and energy, and how to perform measurements for and calculations of their values in MFCs. How much electrical charge or how many electrons can we derive from microbial reactions? This question was asked to in how to calculate the maximum possible number of electrons transferred as a result of the differences in oxidation states of chemicals and the concentrations of chemicals in the growth medium. Students had to do material balances to calculate helped to introduce the concept of charge conservation in elec trochemical systems and the Faraday constant (Table 1). How much power can be harvested from an MFC for a given amount of substrate? The calculation of power from current and potential helped students to understand simple electrical engineering concepts and taught them the differ ences between potential and current, and between power and energy. These concepts often confuse students in the electrical circuit course required during their undergraduate studies. Thus, the MFCEM allowed them to integrate elec trical engineering concepts with chemical and biochemical engineering concepts. How sustainable is power generation in an MFC? Students are often exposed to general terms, such as the sustainable development of the economy of a country, and they need to extend the application of the term sustainable to power generation in MFCs. Sustainability of power generation in rate over long periods, and it is evaluated by monitoring the energy production and consumption using different loads on the MFC. There is no general criterion by which one cell produces power in a sustainable manner and another does not: it all depends on the ratio of the power generated to the power consumed. When the rate of energy consumption is higher than the rate of energy generation by the microorganisms, the MFC does not produce power in a sustainable manner. The opposite is true as well, however, the MFC produces power in a sustainable manner when the rate of energy consump tion is lower or equal to the rate of energy generation by the microorganism. HANDS-ON WORK: EXPERIMENTS We had nine students run the experiment (selection of students was on a volunteer basis), and the group of experi menters was divided among two teams, each of which was as signed one of two options. One team ran the MFC experiments using Shewanella oneidensis while the other used Klebsiella pneumoniae The goal of using two different microorganisms was to observe the difference in power generation and later to arrange a debate on the role of the different respiration systems in accounting for that difference. During the experiment each team tested the effects of the selected variable (microorgan isms, see Table 2). To implement the pedagogy in crowded classes, multiple groups could be used and each group could be subdivided into smaller teams: each team would investigate back to their other group members. This paper describes our classroom experience, in which only one group consisting of two teams worked on understanding the roles of different microorganisms in MFCs. The remaining 14 students received information via lecture, from listen ing to debate presentations by the two groups, and from participation during and after the debate through asking questions and expressing their own opinions. At the end of the experiments, the teams prepared reports and gave class presentations. The presenta tions had the format of a debate. The two teams, which had used different microorganisms, compared their T ABLE 2 Experimental Conditions, Variables Tested, and Topics for the Debate Variable (for groups) Conditions in the MFCs (for teams) Topics for debates between teams Microorganisms 1. Shewanella oneidensis 2. Klebsiella pneumoniae Cellular respiration Electrode material 1. Graphite 2. Stainless steel Microbe-solids interaction Substrate 1. Glucose 2. Natural biomass Renewable energy source Load 1. Low resistor 2. High resistor Sustainability
Vol. 44, No. 2, Spring 2010 161 power generation. Debates centered around why one of the microorganisms produced more power than the other, and a teams position had to be substantiated using the calcula involvement of the students, interesting questions, and many recommendations on how to increase the power generation in such devices. Construction of the MFC. To run the experiments, students used a two-compartment MFC. The components of the MFC and the steps required to construct it are shown in Figure 3. The compartments were made of polycarbonate (8 cm 8 cm 3.7 cm) and were furnished with openings at the top to make electrical connections with the electrodes. Cation exchange membrane (ESC-7000, Electrolytica Corporation) was used to separate the compartments. The cover plates were made of polycarbonate and had three openings for inlet and outlet tubing connections. To prevent leakage, rubber gaskets were used between compartments and between the compartments and cover plates. Screws with wing nuts were used to hold the reactor together. Silicone rubber tubes were used to deliver liquids and gases and to remove them from the respective compartments. The electrodes were made of graphite plates (GraphiteStore.com, Inc.) with surface areas of 23 cm 2 for the anode and 63.4 cm 2 for the cathode. These were placed against the cation exchange membrane, parallel to each other. To construct the MFC, students followed the steps shown in Figure 3. The steps are: 1) inserting the electrodes into the compartments, 2) placing a gasket on the inner side of each compartment, 3) placing the cation exchange membrane over of the membrane, 5) putting the compartments together, 6) placing gaskets outside of the compartments, 7) placing the cover plates, 8) holding the compartments and cover plates together using wing nuts and bolts, and 9) connecting the tubing and then autoclaving the MFC. After autoclaving, the reactor was cooled down to room temperature and the and catholyte, respectively. The anode was inoculated with the selected bacteria for each group. Then in step 10 students placed the reference electrode in the cathodic compartment and connected the electrical wire, resistor, etc. These stepby-step procedures are described in a written manual, and it is available upon request from the authors. ASSIGNMENTSThe assignments were designed to evaluate understanding of the multidisciplinary concepts taught using the MFCEM. In constructing assignment questions we considered the dif ferent levels in Blooms taxonomy  and the levels of learner knowledge described by Apple and Krumsieg(2001)  that are expected to be evident in responses generated by college graduates. We summarize the assignment problems in Table 3 (next page), and match the assignments with levels in Blooms taxonomy and with Apple and Krumsiegs levels of learner knowledge. The questions are discussed in the following sec tions; however, because of space limitations we give only selected answers here. The full range of answers is available upon request. Figure 3. Major com ponents used and steps followed to construct the microbial fuel cell used by the students.
Chemical Engineering Education 162 students understood the principles of the process and could perform basic calculations quantifying processes in MFCs, electrochemistry, and electrical circuitry. Solving this problem required use of Blooms levels of knowledge, comprehension, application, and evaluation, and Apple and Krumsiegs cor responding levels of learner knowledge, as shown in Table 3. In Problem 1a, the students described the basic idea of MFCs using their knowledge of microbial respiration and the electron transport mechanisms. In part 1b, they used the concepts of material balance and charge balance to calculate current, power, current density, and power density. In part 1c, the students used their conceptual understanding of electri cal circuits and material balances to calculate the amount light bulb for one hour. In part 1d, they were asked to think creatively and apply learned concepts to MFCs. This problem also tested the students ability to integrate concepts from electrical engineering with those from chemical engineering and apply them to scaling up a device. In part 1e the students were asked to discuss the future of MFCs and evaluate their potential for providing an alternative energy source. The second problem was designed to relate concepts surrounding cell-growth kinetics in bioreactors with those relevant to MFCs, both of which are taught in this course. This problem was designed to include the analysis and synthesis, and evaluation levels of Blooms taxonomy, and the corresponding working expertise and research levels of Apple and Krumsiegs taxonomy. In part 2a, students were asked to develop a mathematical model to quantify variations in substrate concentration and power generation over time and to construct the plots shown in Figure 4A and 4B. This part was designed to determine whether students could inte grate the idea of the Faraday constant with mass and energy conservations laws, which are taught in physical chemistry/ electrochemistry and basic chemical engineering courses. Constructing the plots in Figure 4 required the solving of taught in our sophomore-level numerical methods course. Part 2b required the calculation of power generation using different values for microbial growth kinetic parameters and columbic time to reach the maximum power depend on the Monod kinetic con stant (Ks). The students needed to draw plots similar to that in Figure 4C to evaluate the effects of the maximum growth rate and of the problem was designed to evaluate the students abilities to interpret the physical meaning of the growth kinetic parameters in the context of the MFC. In part T ABLE 3 Levels in Blooms Taxonomy and Apple and Krumsiegs Levels of Learner Knowledge Used in Constructing Questions to Evaluate Learner Performance in the MFCEM Levels Blooms level  Apple and Krumsiegs level of learner knowledge  Problems as signed I Knowledge Information 1b, 2a II Comprehension Conceptual Understanding 1a, 1c III Application Application 1d IV, V Analysis and Synthesis Working expertise 2a, 2b, 2c, 2d VI Evaluation Research 1e, 2d change if an external electron acceptor (oxygen) were present in the anodic compartment. This question was asked to assess their understanding of how electron transport is involved in the microbial respiration system. Part 2d was open-ended and matches with Blooms levels V and VI, synthesis and evaluation, respectively, and Apple and Krumsiegs Level V, research. In this part students were able to assess the variation in power generation as a function of actual process variables including temperature, pH, and conductivity. ASSESSMENT OF THE MFCEMThe assignments (Box 1, pg. 164) could be completed if the students could successfully integrate multidisciplinary concepts. For example, to construct a model equation (Prob lem 2a) for the substrate concentration, current and power generation in a batch MFC, students needed to integrate the concepts of microbial growth kinetics, mass and charge The effectiveness of the MFCEM was assessed by: 1) comparison of the results of the assignments completed by the students who had run hands-on experiments with those of students who had not; 2) our observations during the experi mental activity and the debate; and 3) the students comments. Since the students running the experiment had volunteered we expected them to be more curious, take more initiative and be self-motivated, and therefore to be better prepared to learn the MFC concepts. We indeed found this to be true, as they earned 42% more points on average than the students who did not run the experiments. The result was shown to = 0.05) and the null hypothesis that the two averages were different. In a two-tailed t-test with 18 degrees of freedom the observations made during the debate between the two teams doing the experiments: 1) The students discussed aspects of microbial respiration and the electron transport processes, including concepts 2) Current and power calculations were vividly discussed, and one of the students even commented that potential,
Vol. 44, No. 2, Spring 2010 163 sense to him; and 3) Many students who previously were silent in class were effectively involved in the discussions.We are aware that our study is limited and a more detailed assessment of the effectiveness of this tool is needed. We expect to collect more data in the upcoming semesters. First, it will be important to come up with other performance mea sures besides homework assignments; we will likely include a critical-thinking rubric being developed in other companion work,  concept inventories that our group will develop based on similar strategies taken by Streveler, et al. (2008).  Also, it will be important to eliminate the possibility that only the more motivated students elect to participate in the active experimental aspect of the course and that such students are already inclined toward a more independent learning compo nent. To safeguard against this we will organize student groups by random selection or based on a fairly equal distribution of GPAs between students within the active experimental groups and those exposed to passive lectures. We could also have a second experimental activity of equivalent rigor in which the activity and formed debate squads would only have the passive lecture for the second experiment, and vice versa. Finally, a detailed survey on student perspectives could be used in which the students themselves compare the learning environments, their growth in understanding, and their ability to work with other group members independent of instructor input. Figure 4. (A) Microbial cell concentration and substrate concentration vs. time plotted using the model equation derived in Problem 2a. (B) Current vs. time calculated for Problem 2a. (C) Power generation vs. time calculated for various Monod constants (Ks). Similar gures were plotted by the students to show the effects of the maximum growth rates and columbic efciencies.
Chemical Engineering Education 164 BOX 1: Problems assigned after introducing the MFCEM Problem 1. In a continuous microbial fuel cell (MFC) the cells are grown in the planktonic phase, anaerobically, using glucose as the electron donor, and the electrons are transferred to the solid electrodes without any kinetic limitation. Suppose the cell growth rate can be described by Monod kinetics as r SX S x 09 07 . At a steady state, the sterile growth medium is fed at a rate of 1 L/h. The working volume of the anodic compartment x/s) is 0.5. Electrons are derived from the substrate according to the following reaction. CH OH OC OH e 61 26 2 2 66 24 24 a) Can you explain why and how the electrons are transferred from cells to the solid electrode? solid electrode). If the cell potential is maintained at 0.5 V, what will the power produced by the MFC be? What will the current density and power density be if the anode surface area is 5 m 2 potential of 0.5 V? c) How much glucose would be needed to power a 60-W light bulb for an hour, if the power generation were the same as that in part a? d) MFCs are an energy source characterized by low power generation. If the energy generation of the MFC remains the same as hour? Note that an Apple Mac laptop operates at 97 watts. e) How feasible is it to design a MFC that will power a laptop directly? Can you predict the future for MFCs as an alternative energy source? Problem 2. A laboratory-scale batch MFC is started with a cell concentration of 0.1 g/L in an anaerobic anode chamber with a volume of 0.25 L. The growth kinetic parameters are ma x =0.9 hr -1 and K s=0.4 g/L. The initial substrate (glucose) concentration is 1 g/L. The cell growth can be described using Monod kinetics, and the substrate consumption rate is dS dt X Y xs / where x/s is assumed to be equal to 0.7. a) Construct a mathematical model to quantify the variation of substrate concentration, current, and power over time. How long will it take to consume 90% of the substrate? What will the concentration of cells be at that time? When will the MFC produce a maximum current? What would the maximum theoretical power be if the cell potential could be maintained at 0.8 V? b) What would happen to the power generation if the Monod half rate constant (K s ), the maximum growth rate, and the Colum d) Qualitatively predict the variation of power generation with the variation of reactor temperature, pH, and conductivity.
Vol. 44, No. 2, Spring 2010 165 CONCLUDING REMARKSWe successfully developed and implemented the MFCEM to teach the concepts of microbial respiration, electrochemical equilibrium and kinetics in a fuel cell, charge conservation, energy, current, and power. The senior-level bioprocess engi neering course was appropriate for incorporating our MFCEM. Initial assessments based on student assignments give strong supportive evidence that the MFCEM is an effective tool for teaching multidisciplinary concepts and that active experimen tation surrounding its implementation is superior to learning through a passive lecture. Expanded activities and a more rigorous learning assessment are planned for the future. ACKNOWLEDGMENTSThe authors would like to thank the Gene and Linda Voil and School of Chemical Engineering and Bioengineering for startup funds in support of the research (Dr. Beyenal) and the Assurances at Washington State University for approving the research study and processing the exemption determination application. The authors gratefully acknowledge the U.S. 0217) (activities through this grant laid the groundwork for 0618872 (activities through this grant laid the groundwork for the hands-on approach and its assessment), and 3M on the development of MFCEM.NOMENCLA TURE E A Anode potential (Volts) E C Cathode potential (Volts) E M 0 Standard reduction potential of mediator (Volts) E O 2 0 Standard oxygen reduction potential (Volts) F Faraday constant (coulombs/mole of electrons) i o Exchange current (A) I Current through a resistor (A) K s Growth constant (g/L) M ox Mediator concentration at oxidation state (mole/L) M red Mediator concentration at reduction state (mole/L) m in n Number of moles of electrons produced per mole of fuel P Power (Watt) PEM Proton Exchange Membrane P O 2 Partial pressure of oxygen (atm) R ext External resistor (ohms) S Substrate concentration (g/L) S o Initial substrate concentration (g/L) t Time (sec) V Potential drop across the resistor (Volts) X Cell concentration (g/L) X o Initial cell concentration (g/L) Y x/s c Heat of combustion of fuel (J/mole) c E ma x REFERENCES 1. Felder, R.M., and R. Brent, Designing and Teaching Courses to Satisfy the ABET Engineering Criteria, J. Eng. Ed. 92 7 (2003) 2. Lattucha, L.R., P.T. Terenzini, and J.F. Volkwein, Engineering Change: A Study of the Impact of EC2000, ABET Inc.
Chemical Engineering Education 166 The University of Nevada, Reno, Chemical Engineer ing, Capstone Senior Design class is a yearlong design piping and instrumentation diagrams, process simulation, engineering economics, heuristics, and systems engineer ingmultiple unit operations, environmental health and safety, engineering ethics, and teaming. During the second that continue for the following one-and-a-half semesters. One of these projects is a collaborative effort with the Center for Nanotechnology at the NASA Ames Research Center in Moffett Field, CA. The objectives for this design project were prepared in close collaboration with Dr. M. Meyyap pan at NASA Ames. This particular project and the project approach have been used for two consecutive years with the design effort was primarily aimed at the detailed design and economics of a 10,000 tonne per year single-wall carbon nanotube (SWNT) plant. In the second year this project was offered, the design class, besides the students own design effort, included building a pilot-scale reactor that was used to synthesize carbon nanotubes as well as to measure selected design parameters. The group sizes for this particular Senior Design Project ranged from three to four students. The students worked collectively to manage their own groups to meet weekly pro jected goals with each student acting as the group leader on a rotating basis. The weekly assignments consisted of either a brief written report or a presentation to the class. Weekly reports included updates in the following areas: literature review, experimental instrumentation, experimental results, and full-scale design calculations. At the end of the semester, A SENIOR DESIGN PROJECT ChE YORK R. SMITH, 1 ALAN FUCHS, 1 AND M. MEYYAPPAN 21 University of Nevada Reno, NV 89557 2 Center for Nanotechnology, NASA Ames Research Center Moffett Field, CA 94035 Copyright ChE Division of ASEE 2010
Vol. 44, No. 2, Spring 2010 167 and poster. Students presented their results at a preliminary project was evaluated by the students ability to stay on task and accomplish their overall objectives of the project. CENTER FOR NANOTECHNOLOGY AT NASA AMES RESEARCH CENTER Started in 1996, the Center for Nanotechnology has several scientists and engineers working on various aspects of nanotechnology. The early focus was on carbon nanotubes (CNTs) because of their potential in a broad array of applications. Right from the beginning the Center has had a strong educational component through an active undergraduate student research program (USRP) and a high school student research program (HSRP) both of which are paid internship programs during the summer. Each program has had more than a hundred students in the last 10 years and a high percentage of the students have been coauthors in publications. Several of them have returned for more than one summer. NASA tracking indicates nearly 100% of the undergraduates going on to graduate school and the high school students moving on to elite universities across the United States for science and engineering. During the semester, the students are typically from the local universities. In addition, the NASA team created an Introduction to Nanotechnology course for engineering majors at the senior undergraduate and  This senior ChE design course participation is an extension of this long-standing education focus at NASA Ames. DESIGN PROJECT Carbon nanotubes have attracted much attention due to their extraordinary mechanical properties and unique electronic properties, and been considered for applications in electronpanel displays, nanoelectromechanical systems, composites, and several others.  Major challenges to commercialization currently include control of diameter and hence the properties and the cost. The latter is an issue because production of even 1 kg/day of SWNTs is not common. At present SWNTs cost $500/g. Unless the production volume goes up, thus bringing the price down to a few hundred dollars per kg, their great potential will not be realized. Production of SWNTs by chemical vapor deposition covers all the bread and butter subjects taught in chemical engineering education: catalysis, reaction others. Future commercial production plants will be designed, built, and operated by chemical engineers. The objective of the senior design project in collaboration with NASA was to design a process that would have the capability to synthesize 10,000 tonnes of SWNTs per year. Further, the students were also asked to design and conduct experiments to verify predictions made in their proposed design. For this design, the students chose to synthesize nique (FBCVD). This technique was chosen due to its ease of scalability  along with its practicality and costs for their experiments as opposed to other CNT synthesis techniques such as arc discharge [4-6] and laser ablation.  In the case of arc discharge, the scale up procedure is not evident because of the elaborate design of the system and also because the SWNT product purity is not attractive. Laser ablation is simply not scalable and practical. EXPERIMENTAL WORK The work of Mauron, et al.  along with See and Harris  bed was designed and constructed consisting of a vertically mounted quartz tube (OD 40mm, ID 36mm) within a tube from one side, and a second frit of the same porosity located the reactor can be seen in Schemes 1 and 2. Scheme 1. Dimensions of quartz tube used as a uidized bed reactor for CNT synthesis. Scheme 2. Depiction of experimental setup. Quartz tube reactor within vertically mounted tube furnace.
Chemical Engineering Education 168 All chemicals, magnesium oxide (MgO, Fischer Chemical, powder), ferric nitrate (Fe(NO 3 ) 3 .9H 2 O, Acros Organics, 99+%), ethyl alcohol (C 2 H 5 OH, Acros Organics, 95%), and hydrochloric acid (HCl, Fischer Chemical, technical grade) were received and used without any further treatment. CNT Synthesis An iron catalyst (5 wt%) supported by magnesium oxide was homogenized in ethanol through ultrasonication. The solution was dried overnight in an air oven and ground lyst-precursor was used to promote SWNT growth and the system was then allowed to reach the synthesis temperature introduced as the carbon source and mixed with the nitrogen stream at a rate well under the explosive limit of the exit gas stream. [9, 10] Separation of the nanotubes from the catalyst-precursor was performed in 0.1M HCl bath while stirred for 15 hours at  The cleaved nanotubes were decanted from solution and allowed to dry in air. The products were then analyzed through energy dispersion spectroscopy (EDS) and scanning electron microscopy (SEM) techniques. Health and Safety A major pedagogical issue for the students to understand was that health and safety issues are paramount. Because it Figure 1. SEM images of bundles of web-like CNT structures formed on iron particles through uidized bed chemical vapor deposition at 900 C for 30 minutes. Figure 2. SEM and EDS of CNT formed on iron particles through uidized bed chemical vapor deposition at 900 C. (a) SEM image of CNT bundle ~10 nm in diameter, (b) EDS spectrum zone, (c) and (d) EDS results.
Vol. 44, No. 2, Spring 2010 169 was necessary to deal with explosive gases at high temperatures, many safety precautions were taken. The students were required to have their experimental design approved by the Environmental Health & Safety Department (EH&S). In addition, a hazard and operability (HAZOP) analysis was done by the students and approved by EH&S before any experiments were conducted. These operating conditions and procedures were also incorporated into their full-scale design. EXPERIMENTAL RESULTS Fluidized Bed Reactor Experiments Figures 1 and 2 show from SEM images that carbon nanotubes formed web-like structures across the surface of the iron catalyst. The diameters of these nanotubes have a large variation, which indicates that MWNTs along with bundles of SWNTs could have been formed. Energy dispersion spectroscopy (EDS) indicates that the web-like formations are indeed carbon deposits. SWNTs have diameters within the range of a few nanometers, which is beyond the resolution of the SEM. Further characterization techniques such as Raman spectroscopy and/or high-resolution tunneling morphology. DESIGN METHODOLOGY scale reactor were determined following the texts of Kunii and Levenspiel and Geldart. [12, 13] Reaction Kinetics and Bed Sizing Methane (CH 4 ) as carbon source and an iron catalyst supported by magnesium oxide (MgO) were chosen in this project. For a low carbon-to-hydrogen ratio, methane is ideal for SWNT production. A kinetic model was adapted from the work of Pinilla, et al.,  which provided a reaction rate as a function of partial pressure of methane for carbon This study also provided a pseudo steady-state activation energy of 230 kJ/mol. design equation can be implied. This type of model is a good estimate since it can account for distribution of particles as well as the residence time. The design equation for a CSTR is given by:  V dC dt FC Cr V CH AA A 4 0 1 () r A is the reaction rate and C A0 and C A are the initial and exit concentrations of the methane, respectively. To solve for the reactor volume, Eq. (1) was solved for steady-state operation and the reaction rate was solved by Eq. (3), V FC X r A A 0 2 () rk ep A E RT A 0 3 () where X is the conversion factor, E is the activation energy, R is the gas constant, p A is the partial pressure of methane, T is temperature, and k 0 is the reaction rate constant that was back-calculated by linearization of the Arrhenius equation  and use of the pseudo steady-state Arrhenius plot for carbon formation over carbonaceous catalyst with methane gas  on the basis of 1.5 g of catalyst/0.5 g of CNT.  As prescribed by Fogler,  the design equation for the volume of a CSTR is determined from the following parameters: reaction rate equation, stoichiometric constants, gas concentration, and pressure. Eq. (2) was used to solve for the reactor volume as a function of conversion for various partial pressure, which is given in Figure 3. The minimum 3 reactor volume. These values were determined through iterating a system of nonlinear equations derived from the Ergun equation to determine the bed characteristics (discussed later). Convergence of the system of equations was achieved via a Levenberg-Marquardt algorithm implemented in a program written in Mathcad 13. Although a CSTR model is a fair approximation, this model is normally used for a homogeneous system. This is not the catalyzed, gas-phase reaction, a model presented by Kunii and Levenspiel  expresses the rate per unit volume of catalyst, Figure 3. Volume of uidized bed reactor (m 3 ) as a function of reaction conversion (X) for various volumetric ow rates of gas.
Chemical Engineering Education 170 1 4 3 3 V dN dt KC K mg as ms olid s s A rA r () where K r is the reaction rate constant and V s considers the solid as nonporous. Given a feed rate v of reactant gas C A0 to a catalyst bed containing solids of volume V s, integration gives Eq. (5) in terms of the outlet concentration C Ai or the outlet fractional conversion X, and the reactor ability is given in Eq. (6). 1 1 1 5 0 X C CK A Ai r () volume_of _catalys t volumet ri c_ fl ow _of _ga s V v s () 6 This leads to the dimensionless reaction rate group of, KK L u rr ii o 1 7 () o is the From the dimensionless rate group of Eq. (7), the height of the bed can be determined and from choosing a cross-sectional area, the volume can be approximated. This method has been shown to correlate very well with experiment.  Furthermore, this correlation can be adjusted to incorporate deactivation of the catalyst. Through both these models an appropriate volume can be determined for optimal operation and cost. Bed Characteristics that is a mixture of methane and nitrogen, and estimating Geldart chart to fall within the regime of a Geldart Group B powder. that contains most solids in the mean size and density range p p = 4 g/cm 3 which has been particles), interparticle forces are negligible and bubbles ity.  It is also assumed that this material has a sphericity in calculations. nonlinear equations derived from the Ergun equation must be tion constants is given in Table 1 for a reactor of diameter of 3 density of the methane-nitrogen mixture was calculated using the Peng-Robinson equation of state for a binary system, along with other thermodynamic values such as enthalpy, entropy, compressibility, and fugacity. is to ensure that particle entrainment occurs. This determina tion is used to design the freeboard space required along with by examining single-particle calculations. A summary of these is given in Table 2. Once the gas velocity exceeds the terminal velocity of a single particle, entrainment is likely to occur. Since one cannot physically account for every particle, determination of the suspension bed expansion T ABLE 1 Summary of Results for Minimum Fluidization Properties of Industrial-Size Fluidized Bed Reactor Ar Archenemies number [dimensionless] 295 Re mf Reynolds Number of minimum 0.331 u mf Gas velocity of minimum 0.025 u o 0.125 BE Bed expansion [m] 1.06 L mf Bed height of minimum 10.5 mf Void fraction of minimum 0.45 Pressure drop [psi] 30 T ABLE 2 Summary of Single Particle Calculations for Our Catalyst and Feed Gas u t Terminal particle velocity [m/s] 0.62 C D 5.82 Re p Particle Reynolds Number [dimensionless] 8.22 T ABLE 3 Summary of Results for Determining Bed Suspension maxMaximum void fraction [dimensionless] 0.785 u f 0.07 u ps 0.055 Q f 3 /min] 16.23
Vol. 44, No. 2, Spring 2010 171 is required. The single particle calculations can be used to sion concentration. If u f > u ps entrainment is likely to occur and the bed will operate in a stable condition. This relation is shown in Table 3. When designing the gas distributor, minimal pressure drop zation. Through design criteria based on reactor dimensions, the data obtained indicate that the process would require very high natural gas demand. Further characterization and design would include more detailed mass and energy balances on the reactor and eventu be used to determine the overall feasibility of this process. An economic analysis was conducted using the aid of the program CAPCOST  for equipment and operating costs. By implementing a rudimentary balance sheet of income vs. expenses it is estimated that the break-even cost for SWNT selling price was estimated around $350-$400/kg. Further economic and market analysis is required, however, to deter mine how readily the current market would be captured, thus diluting the current market price. This analysis would yield to the results presented, the designs of the catalyst synthesis and separation processes were included, but are not presented in this paper.CONCLUSIONSSenior-year chemical engineering students designed a process that would produce 10,000 tonnes of SWNTs per year and also conducted bench-top experiments to synthesize SWNTs. This was an excellent pedagogical experience for the students because it related to real-world design issues. The success of the students project was evaluated on the basis of completion of weekly assignments and project milestones. The experiments resulted in carbon nanotubes, which were characterized using SEM. Detailed design of the reactor was presented, and the break-even cost of the nanotubes is esti mated to be approximately $400/kg.ACKNOWLEDGMENTThe authors acknowledge support from the Nevada NASA Space Grant Consortium for this project and diligent work by the 2007 senior design class students: Oliver Daniel, Kevin Farley, Cherish Hoffman, and Jeff Nicholson and the 2008 senior design class students: Jignesh Patel, Kylen Smith, and York Smith.REFERENCES 1. Meyyappan, M., Nanotechnology Education and Training, J. Mat. Educ. 26 311 (2004); also see, Meyyappan, M., A Course on Intro duction to Nanotechnology for Undergraduate Science and Engineering Majors, in Nanoscale Science and Engineering Education Sweeney, A.E., and S. Seal, (Eds.) American Science Publishing (2008) 2. Meyyappan, M., (Ed.), Carbon Nanotubes: Science and Applications CRC Press, Boca Raton, FL (2004) 3. Mauron, P., C. Emmenegger, P. Sudan, P. Wenger, S. Rentsch, and A. Zttel, Fluidized-bed CVD Synthesis of Carbon Nanotubes on Fe 2 O 3/MgO, Diam. Relat. Mater. 12 780 (2003) 4. Iijima, S., Single-Shell Carbon Nanotubes of 1-nm Diameter, Nature 354 56 (1991) 5. Bethune, D.S., C.H. Kiang, M.S. de Vries, G. Gorman, R. Savoy, J. Vasquez, and R. Beyers, Cobalt-Catalysed Growth of Carbon Nano tubes With Single-Atomic-Layer Walls, Nature 363 605 (1993) 6. Journet, C., and P. Bernier, Poduction of Carbon Nanotubes, Appl. Phys. A 67 (1998) 7. Guo, T., P. Nikolaev, A. Thess, D.T. Colbert, and R.E. Smalley, Catalytic Growth of Single-Walled Nanotubes by Laser Vaporization, Chem. Phys. Lett. 243 49 (1995) 8. See, C.H., O.M. Dunens, K.J. MacKenzie, and A.T. Harris, Process Parameter Interaction Effects During Carbon Nanotube Synthesis in Fluidized Beds, Ind. Eng. Chem. Res. 46 7686 (2008) 9. Air Products, MSDS Report,
Chemical Engineering Education 172 SKITS, STOCKINGS, AND SENIORITIS ALE: CREA TIVE CHEMICAL ENGINEERSLISA G. BULLARD North Carolina State University Copyright ChE Division of ASEE 2010 ChE the column should be approximately 500 words. If graphics are included, the length needs to be Wankat
The laboratory experience in chemical engineering education has long been an integral part of our curricula. CEE encourages the submission of manuscripts describing innovations in the laboratory ranging from large-scale unit operations experiments to demonstrations appropriate for the classroom. The following guidelines are offered to assist authors in the preparation of manuscripts that are informative to our readership. These are only suggestions, based on the comments of previous reviewers; authors should use their own judgment in presenting their experiences. A set of general guidelines and advice to the author can be found at our Web site: Manuscripts should describe the results of original and laboratory-tested ideas. allow and motivate others to adapt the ideas to their own curricula. It is noted that the readership of CEE is largely faculty and instructors. Manuscripts must contain an abstract and often include an Introduction, Laboratory Description, Data Analysis, Summary of Experiences, Conclusions, and References. Introduction should establish the context of the laboratory experi ence (e.g., relation to curriculum, review of literature), state the learning objectives, and describe the rationale and approach. Laboratory Description section should describe the experiment in to implement a similar experiment on his or her campus. Schematic dia grams or photos, cost information, and references to previous publica should be addressed as well as any special operating procedures. Data Analysis section should be included that concisely describes the method of data analysis. Recognizing that the audience is primarily faculty, the description of the underlying theory should be referenced or brief. The purpose of this section is to communicate to the e.g., treatment of reac tion-rate data in a temperature range that includes two mechanisms). Summary of Experiences section is to convey the results of laboratory or classroom testing. The section can enumerate, for example, best practices, pitfalls, student survey results, or anecdotal material. Conclusions (as opposed to a summary) of your experiences should be the last section of the paper prior to listing References.Author Guidelines for the LABORATORY Feature