Combined Research-Curriculum Development Program
1997 Grantees Conference Proceedings


Experiences in Team-Teaching a Process Design Course
Covering Steady-state Synthesis, Optimization and Control

Warren D. Seider1
University of Pennsylvania

Christodoulos A. Floudas2
Princeton University

William L. Luyben3
Lehigh University


Abstract: In 1996-97, two courses are being taught cooperatively at Penn, Princeton, and Lehigh with the instructors presenting six two-hour common lectures at each of the three schools. Since the courses are in progress, this paper provides a status report with emphasis on the format of the courses, the courseware being used and developed, the successes and problems, the methods of evaluation and dissemination, and plans for the future.

Vision, Goals, and Objectives: Our three-year CRCD Project, which began in September, 1995, has been designed to enable the three investigators to translate the results of their ongoing research in the design and control of chemical processes into a modern two-course sequence in process design. The three investigators are involved in various aspects of research to improve the designs of chemical processes by taking advantage of the latest optimization and control strategies. Their work has been helping to shift the emphasis of conventional design strategies from analysis in the steady state to a more balanced approach using optimization and dynamic analysis.

Since the inception of the project, the three investigators have been working to create a two-course sequence, and the associated courseware (text and multi-media modules), that provides the first comprehensive treatment of process design, optimization, and control for the chemical engineering profession. It is blending heuristic approaches with their more analytical counterparts that involve computer simulation and optimization. It is the first two-course sequence to include instruction on: (1) the strategies for using the latest process simulators (both steady state and dynamic) and optimization packages, (2) methods for performing optimization in the synthesis of process flowsheets, and (3) techniques for achieving better process designs through the consideration of plant-wide control and process dynamics.

Accomplishments: During the first year of the project, the three investigators met for six afternoon sessions in which they designed their courses cooperatively. A sequence of 18 common lectures, to be delivered at the three schools, six by each of the investigators, was designed; note that Table 1 shows the schedule for the lectures at Penn. Obstacles to be overcome included the somewhat different academic year calendars at the three schools and, of more significance, the fact that, unlike Penn and Lehigh, Princeton does not require a second course in process design. Furthermore, the Princeton undergraduates do not take a required course in process control. Since most undergraduate curricula do not include a second course in process design, it was decided that our cooperative efforts should confront this obstacle in a direct manner. Prof. Floudas scheduled a course with emphasis on process control, to be offered as an elective, which the three instructors hoped would be a popular elective as a consequence of their combined efforts in the first course. To overcome the lack of background in process control, Prof. Floudas planned to cover the basics of process control in five-weeks at the beginning of the elective course. In future years, he planned to teach the sophomores a course in which dynamics are introduced as well as some of the principles of process control.

Also, during the first year, the three investigators discussed their ongoing work to develop text and multi-media courseware, so as to achieve some cooperation in these efforts. Profs. Floudas and Luyben agreed to class test some of the courseware being developed by Prof. Seider. In addition, Profs. Seider and Luyben agreed to adopt the book recently completed by Prof. Floudas entitled Nonlinear and Mixed-integer Optimization:Fundamentals and Applications[1]. Furthermore, Prof. Luyben decided to continue writing a textbook entitled Plant-wide Process Control with Michael Luyben and Bjorn Tyreus. He would use portions of this courseware in his common lectures, as well Chapters 5 and 6 of his recently completed text, Essentials of Process Control [2].

The three investigators also discussed the possibility of purchasing computers for their courses cooperatively. One meeting was held with a major computer vendor and an attempt was made to negotiate a sizable discount through a combined purchase. The computers marketed by this vendor, like the others, could not run all of the commercial packages (e.g., ASPEN PLUS and HYSYS) as well as the multi-media software using Macromedia's DIRECTOR package. Since each of the investigators had somewhat different priorities with regard to the speed of computation and software, separate purchases were made by Profs. Floudas and Seider. Prof. Luyben is completing the arrangements for the purchase of his computers.

Now, about one and one-half years into the project, we are completing the first offering of our two-course sequence. It has been very successful in many respects, but has also encountered problems, some of which have not yet been overcome. At this point, the students at Penn and Lehigh are completing their design projects and it is too early to fully assess the success of this first offering. Still, in the next section, an attempt is made to make a preliminary assessment.

Successes and Problems: By far the greatest success, thus far, has been in the interactions of the three investigators in planning their courses, built around the common lectures. At all three schools, one two-hour common lecture was presented each week. In addition, at Penn and Lehigh, two lectures were presented by the local instructor, whereas at Princeton three lectures were presented by Prof. Floudas. Each local instructor supplemented the common lectures with the materials, homeworks, and exams that he selected. For example, at Penn, additional lectures and homeworks were assigned to supplement the materials presented by Prof. Floudas, and consequently, the students gained a good comprehension of his materials. Each instructor attended all of the common lectures and experienced, first hand, the approaches of his colleagues in teaching the materials in ways that he would not have attempted on his own. This was a major advantage of our interactions.

Throughout the two course sequence, Prof. Seider continued to write his text and multi-media courseware entitled Computer-aided Chemical Process Design [3]. This courseware is being coauthored by J.D. Seader (University of Utah) and D.R. Lewin (Technion). An outline is provided in Figure 1 and it is noteworthy that 10 of the 14 chapters were sent to John Wiley for review on March 17, with the remaining chapters due on September 15. The interactions with Profs. Floudas and Luyben and their students have had a major impact on this courseware. On the basis of the responses of the students and faculty at all three schools, major sections were rewritten and clearer examples were added. Chapter VII on Heat and Power Integration has not yet been written, but will be structured on the basis of the lectures presented by Prof. Floudas. The approach used by D.R. Lewin in teaching the basics of plant-wide process control differs somewhat from that used by Prof. Luyben in his lectures. When Prof. Lewin visits Penn this summer, some of Prof. Luyben's ideas will be incorporated. It is important to note that Prof. Luyben is writing a complete text entitled Plant-wide Process Control [2] and the treatment by Prof. Lewin is designed to provide a brief, self-contained introduction to the subject that can be included in a single text on process design. Clearly, the cross-talk between the investigators on this project will continue to enable these approaches to develop more effectively.

Another success related to the cooperation in teaching these courses has been the usage of the multi-media CD-ROM entitled Steady-state Simulation of Process Flowsheets [4,5], prepared by Prof. Seider and his students, at Penn and Princeton. This enabled the students at these two schools to learn to use the ASPEN PLUS simulator more easily. It wasn't used at Lehigh because Prof. Luyben preferred to use the HYSYS simulator, which also has the ability to simulate process flowsheets dynamically when studying the controllability of a design.

Probably the biggest problem this year was that the students at Princeton did not select the elective course in process control. Only two of the 21 students registered for the course, and consequently, it could not be offered. In this regard, although five elective courses were offered in chemical engineering, the vast majority of the students elected courses offered by other departments. One speculative explanation which is difficult to evaluate is that the Fall course, with three one-hour lectures, a two-hour common lecture, and the requirement to complete homeworks and a comprehensive design project, was too demanding. In effect, the Princeton students received about the equivalent of two-thirds of the two-course sequence being taught at Penn and Lehigh. They rated the course as excellent; in fact, once again, it was rated amongst the best in chemical engineering at Princeton. The students may have felt that, although they were lacking the process dynamics and control aspects, their exposure had been sufficiently comprehensive. Had Prof. Floudas covered less ground in one semester, perhaps the students would have responded positively in electing the second course. Perhaps this will be an option for next year. Note that the three faculty will begin to plan for their collaboration in May after the Spring semester is completed and evaluations of the courses have been completed by the students, faculty, and industrial consultants. The latter provide the design projects at Penn and Lehigh and interact with the students and faculty as the projects are being completed.

Yet another problem was that it was difficult to present the common lectures so that they meshed properly with the remaining lecture materials. Undergraduate students, for the most-part, are accustomed to attending lectures and solving homework problems that are closely related to the lectures. In these common lectures, the lecturer could only recommend that certain problems be worked, but could not assign and grade homeworks. As a consequence, in some areas, the students did not solve problems needed to reinforce their understanding of the materials. Clearly, their comprehension was not as complete as intended by the common lecturer who might have set his objectives at a lower level had he taken this into consideration.

A related problem was that, when the local instructor intended to provide supplemental lectures and problem assignments to accompany the common lectures, this was much more difficult to accomplish than anticipated. While attending a common lecture, and seeing the material being presented for the first time, the local instructor would have to decide how to supplement the lectures and which homeworks to assign just two days later. It was a challenge to select the best homework problems having not taught using this approach in the past. Consequently, many of the assignments were not coordinated with the lecture material as effectively as usual. The net result is that at Penn many of the students found faults not encountered previously (over 15 years of teaching process design). Several indicated in a Questionnaire that they were severely overloaded and that the materials were not sufficiently well planned and organized. Yet, many rated the common lectures by Prof. Floudas to be either very good or excellent. Prof. Luyben's common lectures at Penn in the Spring have not yet been evaluated, although it is evident that some of the design groups are using HYSYS to carry out plant-wide control studies. This is the first time this level of analysis is being utilized in the senior design projects at Penn. Furthermore, to better prepare the Junior class for next year, Prof. Seider is teaching the Process Control course this semester using the Essentials of Process Control by Luyben and Luyben.

Lessons Shared with the Engineering Education Community: It seems clear that the results of our interactions will benefit the community of chemical engineering educators mostly through the courseware that is being written and class-tested as our collaboration proceeds. As mentioned in the previous section, Prof. Seider benefitted significantly by the feedback he received when preparing and delivering his common lectures. His courseware is on target to be completed and in production by John Wiley on September 15, 1997. It is anticipated that Wiley will have the courseware prepared for distribution to the chemical engineering community in time for the Fall Semester in 1998. A textbook will be prepared with a CD packaged on the inside of the back cover, and a Web page will be maintained by Wiley to contain information about the latest updates of the commercial software. Fortunately, the ASEE Summer School for Chemical Engineering Faculty (held every five years) is scheduled to take place in August and Prof. Seider will be presenting his courseware to the faculty assembled, representing a large fraction of the departments in the United States and Canada. In addition, these interactions have provided Prof. Luyben good experiences as he assembles his text entitled Plant-wide Process Control. And, finally, Prof. Floudas received good feedback concerning the ability to use his book effectively in the undergraduate design courses at other schools. These experiences are building closer bonds between the three faculty which in the long run should have a very beneficial effect on the courseware that they prepare separately and together (possibly in the future), and consequently, upon the community of chemical engineering educators.

Evaluation: The two-course sequence is being evaluated by the students through their participation and informal feedback as well as questionnaires designed to solicit their opinions about the strengths and weaknesses. As mentioned above, some of the questionnaires have been received and analyzed. Others await the end of the Spring Semester after the Seniors have completed their design projects. In addition, at Penn and Lehigh, we plan to request an evaluation by our industrial consultants, many of whom have provided us with excellent design projects. Finally, at Penn, several of our industrial consultants are reviewing the chapters of the courseware we are writing. Throughout the last year, we have been receiving critiques from colleagues at ARCO and Aspen Tech. The March 17th Edition has been sent to many more colleagues and reviewers selected by John Wiley. We anticipate receiving detailed critiques from several of these reviewers, especially those who will receive an honorarium from Wiley in return for their reviews.

Dissemination: As mentioned above, plans have been completed for John Wiley to disseminate the courseware being developed by Prof. Seider. Prof. Luyben is negotiating a contract with another publisher. Presentations will also be made at the ASEE Summer School (see above) and the Annual Meeting of the AIChE in Los Angeles in November. To the extent desirable, we are prepared to share our experiences with the community of engineering educators beyond those in chemical engineering. As sessions in education-oriented conferences are planned, we will accept invitations to participate.

Plans for the Third Year: As mentioned above, these plans will be initiated at the end of the Spring Semester, at which time we will have the results of the evaluations. Each faculty member is reserving time for another round of common lectures. We also plan to explore the possibility of using television communications in a distance learning format. If possible, this should save considerable time in travelling amongst the three schools.

Building Upon our Foundation: With the rapid development of software packages and multi-media communications, there is an increasing need to develop and share courseware in the design and control areas. We can envision the development of a more comprehensive project over the next few years involving schools like the University of Utah, Carnegie-Mellon University, and the Technion, in addition to Penn, Princeton, and Lehigh. The key in this regard is to identify faculty members that have much to contribute to a coalition, rather than trying to find suitable participants from amongst the existing coalitions (e.g., GATEWAY).

Acknowledgments: Much of the progress described in this paper was made possible through the generous support of our NSF Project No. EEC-9527441 administered by the Combined Research-Curriculum Development Program. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

References:

  1. Floudas, C.A., Nonlinear and Mixed-integer Optimization: Fundamentals and Applications, Oxford Univ. Press, 1996.
  2. Luyben, W.L., and M.L. Luyben, Essentials of Process Control, McGraw-Hill, 1997.
  3. Seider, W.D., J.D. Seader, and D.R. Lewin, Computer-aided Chemical Process Design, submitted to John Wiley, 1997.
  4. Seider, W.D., S. Winters, M. Ali, and D.M. Miller, Steady-state Simulation of Process Flowsheets - A Multi-media CD-ROM, Univ. of Pennsylvania, 1997.
  5. Winters, S., M. Ali, and W.D. Seider, "An Interactive Approach to Teaching Steady-state Simulation of Chemical Processes," Comp. Applic. Eng. Educ., 4, 4, 1996, 261-268.

Mailing Addresses:

 1.



Professor Warren D. Seider
Department of Chemical Engineering
University of Pennsylvania
Philadelphia, PA 19104/6393
 2.



Professor Christodoulos A. Floudas
Department of Chemical Engineering
Princeton University
Princeton, NJ 08544/5263
 3.



Professor William L. Luyben
Department of Chemical Engineering
Lehigh University
Bethlehem, PA 18015/4791


COMPUTER-AIDED CHEMICAL PROCESS DESIGN

TABLE OF CONTENTS

ChapterTitle
1The Design Process
2Process Synthesis
3Simulation in Process Synthesis
4Heuristics for Process Synthesis
5Synthesis of Separation Trains
6Second-law Analysis
7Heat and Power Integration
8Heat Exchanger Design
9Capital Cost Estimation
10Profitability Analysis
11Optimization of Process Flowsheets
12The Interaction of Process Design and Process Control
13Flowsheet Controllability Analysis
14Dynamic Simulation of Process Flowsheets

Figure 1 Table of Contents for Computer-aided Chemical Process Design [3]


Table 1 Common Lectures Presented at the University of Pennsylvania
DateLecturerTopic
Sept. 4,6*SeiderProcess Synthesis - Overview, Steps, Example
Sept. 9SeiderFlowsheet Simulation - Architecture, Recycle, Degrees-of-freedom
Sept. 16LuybenDynamics and Plant-wide Control
Sept. 25,27*SeiderHeuristic-based Process Synthesis
Sept. 30SeiderDistillation Trains for Nearly-ideal Mixtures
Oct. 7SeiderDistillation Trains for Non-ideal Mixtures
Oct. 16,18*SeiderThermodynamic Efficiency of Chemical Processes
Oct. 21FloudasConvex Analysis for Optimization
Oct. 28FloudasLinear and Mixed-integer Linear Optimization
Nov. 4FloudasHeat-exchanger Networks - Minimum Utilities Costs
Nov. 18FloudasHeat-exchanger Networks - Minimum Number of Matches
Nov. 25FloudasNonlinear Optimization
Dec. 2FloudasHeat-exchanger Networks - Minimum Investment Cost
Jan. 13LuybenReview of the Basics of Process Control
Jan. 20LuybenControl of Unit Operations
Jan. 27LuybenControl of Distillation Columns
Feb. 3LuybenPlant-wide Control
Feb. 10LuybenControl of Multivariable Processes
Feb. 17LuybenDesign/Control Tradeoffs

*Two one-hour lectures. All others are two-hour lectures.