ofContents - CAST) Div 2.pdfSt. Louis, MO 63167 (314) 694-8274 Director, 1989-1991 ArthurL. Parker...

•1

of Contents

Editorial NotesAbout This Issue, by Peter R. Rony and Joseph D. Wright 0 0 0 •••• '0 ••• 0 • • • • •• 1Chairman's Report, by Joseph D. Wright 0" 0 •• 0 • 0 ••••••• 0 0 •• 0 •• 0 •••••••••••••••• 0 • 0 ••• 0 00 0 o. 0 0 • 0 0 • 0 • 0 • 0 00 • 0 • 0 1CAST Executive Committee Meeting (March 20, 1990): Excerpts from Minutes, by Maria K. Burka 0 •••••• 0 • • • • • • •• 2Report of the CAST Division Programming Board (March 1990), by Jeffrey J. Siirola 00' • • • • • • • • • • • • • • • •• 4Report on Paper Retrieval Project, CAST Division, AIChE Meeting (November 1989) 0 0 • 0 0 • 0 •• 0 ••••••••••• 0 •••• 0 •• 0 0 5CAST Division Receives Outstanding Newsletter Award at AIChE Officers Meeting, by Ignacio E. Grossmann .0...... 5AIChE Recognition System . 0 •• 0 •••• 0 • ~ 0 •••• 0 0 ••• 0 0 • 0 •• 0 0 0 •• 0 0 ••••••••• 0 0 ••••• 0 • 0 0 0 •• 0 0 0 • 0 ••• 0 '00 ••••••••• 0 • 0 6

AwardsRoger W. H. Sargent is Honored With the 1990 Computing in Chemical Engineering Award 0 ••••• 0 ••••••••••••• '0 7George D. Byrne Wins the CAST Division Computing Practice Award 0 0 ••••••••••••• 0 ••••• 0 • • • • • • • • • • •• 8The Ted Petersen Student Paper Award Goes to R. (Krish) Krishnamurthy 0 •••••••• 0 ••• 0 •• 0 0 •• 0 • • • • • • • • • • • • • • •• 9

ArticlesChemical Process Engineering at Cray Research, by S. E. Zitney, R. Do La Roche, and R. A. Eades (Cray Research, Inc.)10The TI 500 Series Programmable Logic Controllers: Merging of TUTSIM Dynamic Simulation With APT

Control System Design, by K. E. Rony (Texas Instruments Incorporated) and P. R. Rony (Virginia Tech) 0 •••• 0.. 14

CommunicationsListening In, by Peter R. Rony 0 0 ••••••••••••• 0 ••••••• 0 • 0 0 0 •• 0 ••••••••••• 0 •••••••• 0 ••• 0 • 0 •••• 0 0 • • • • • • • • • • • •• 25Process Data eXchange Institute (PDXI), by John T. Baldwin 0 0 •• 0 0 0 • 0 • • • • • • • • • • • •• 26CAST Experiment - Chicago AIChE Meeting, by Alan Foss, University of California, Berkeley 0 0 0 0 •• 0 0 0 •••••••• 0 • •• 26

Meetings, Conferences, Short Courses, and WorkshopsProcess Integration Using Pinch Technology, Linnhoff-March, Inc., Chicago (October 2-5,1990),

Houston (October 30-November 3, 1990) .... 0 •••••••••••••••••••••••••••••••• 0 •••••• 0 • 0 • 0 • • • • • • • • • • • • • • • •• 27Statistical Process Quality Control (AIChE Course), Tennessee Eastman Company, Kingsport, Tennessee,

October 9-10, 1990 0 ••••••••••••••• 0 ••••••••••••••••••••• 0 ••• 0 • 0 0 •• 0 0 • • • • • • • • • • • •• 27Chicago AIChE Meeting, November 11-16, 1990 0 ••••• 0 •••• 0 • • • • • • • • • • • • • • •• 27Computer Process Control IV (CPC-IV), South Padre Island, TX, Feburary 17-22, 1991 0 • • • • • • • • • • • • • • • • • • •• 28Houston AIChE Meeting, April 7-11,1991 0 ••••••••••••••••••••••••••• 0 • • • • • • • • • • • • • • • • • • • • • • • • • •• 30Advanced Process Systems Engineering: Concepts and Practice, Carnegie Mellon University, June 3-7,1991 ..... 31Fourth World Congress of Chemical Engineering, Karlsruhe, West Germany, June 16-21, 1991 ... 0 • • • • • • • • • • • • •• 314th International Symposium on Process Systems Engineering (PSE '91), Montebello, Quebec, August 4-9,1991 .... 32Los Angeles AIChE Meeting, November 17-22,1991 0 0 •••••••••••••••• 0 •• 0 0 • o' ••••• 0 ••••••• 0 • • • • • • • • • •• 32New Orleans AIChE Meeting, March 29-April 2, 1992 0 ••• 0 •••••• 0 •••••••••••••••• 0 • • • • • • • • • • • • • • • • • •• 34Foundations ofComputer-AidedJ>Plant Operations (FOCAPO '92), Summer 1992 .... 0 ••• o' 0 ••••• 0 •• 0 • • • • • • • • • • • • •• 35Miami AIChE Meeting, November 1-6, 1992 0 ••••••••••••• 0 ••• 0 •••••• 0 0 0 0 0 • 0 ••• 0 0 •• 0 • 0 •••••••••••••• o. 35

Call for PapersHouston AIChE Meeting, April 7-11, 1991 0000 ••• 00 •• 0 • 0 0 • 0 • 00 •• 0 0 •• 00 •• 0 .0' ••••• 0' •• 0 ••• 0 ••••• o ••••••••••• o. 36Los Agneles AIChE Meeting, November 17-22,1991 0 0 ••• 0 0 • 0 •••• 0 ••• 0 •••• 0 •••• 0 0 0 o •••••••• 0 •••••••••••• 0 ••• o. 38New Orleans AIChE Meeting, March 29-ApriI2, 1992 000 •• 0 0 • 0 0 •• 0 •••••••••• 0 ••••••••• 0 ••• 0 ••• 0 ••••••••• 0 • 0 • •• 42

Advertisements1991 Award Nomination Form

..Join the CAsltDivision of AIChE

CAST Division of AIChEExecutive Committee

Elected Members

Past ChairmanBruce A. FinlaysonDepartment ofChemical EngineeringUniversity ofWashingtonSeattle, WA 98195(206) 543-2250Bitnet: FINLAYSON@MAX

ChairmanJoseph D. WrightXerox Research Centre ofCanada2660 Speakman DriveMississauga, Ontario, CanadaL5K2L1(416) 823-7091Fax: (416) 822-6984 (work)

(416) 332-1553 (home)Bitnet: [email protected]

1st Vice ChairmanG. V. (Rex) ReklaitisSchool ofChemical EngineeringPurdue UniversityWest Lafayette, IN 47907(317) 494-4075Bitnet: GVR@PURCHE

2nd Vice ChairmanIgnacio GrossmannChemical Engineering DepartmentCarnegie-Mellon UniversityPittsburgh, PA 15213(412) 268-2228Fax: (412)268-7139Bitnet: [email protected]

Secretary/TreasurerMaria K. BurkaChemical and Thermal System DivisionRoom 1115National Science FoundationWashington, DC 20550(202) 357-9606Fax: (202)357-7636Bitnet: [email protected]

Director, 1988-1990Herbert I. BrittAspen Technology, Inc.251 Vassar StreetCambridge, MA 02139(617) 497-9010

Director, 1988-1990ThomasJ. McAvoyDepartment ofChemical & Nuclear EngineeringUniversity of MarylandCollege Park, ~D 20742(301) 454-4593

Director, 1989-1991Henry H. ChienMonsanto800 N. Lindbergh BoulevardSt. Louis, MO 63167(314) 694-8274

Director, 1989-1991Arthur L. ParkerShell Oil CompanyP.O. Box 10Norco, LA 70079(504) 465-7627

Director, 1990-1992Lorenz T. BieglerChemical Engineering DepartmentCarnegie-Mellon UniversityPittsburgh, PA 15213(412) 268-2232Bitnet: [email protected]

Director, 1990-1992Mohinder K. SoodMobil Research and Development Corp.P.O. Box 1026Princeton, NJ 08540Office: (609) 737-4960Home: (215) 968-7862Fax: (609)737-5047

Ex-Officio LWembers

Programming Board ChairmanJeffrey J. SiirolaResearch Laboratories - B95Eastman Chemical CompanyKingsport, TN 37662(615) 229-3069Fax: (615)229-4558

A rea lOa ChairmanKris R. Kaushik (Krishna)Shell Oil CompanyP.O. Box 6249Carson, CA 90749(213) 816-2276

Area lOa Vice ChairmanChristodoulos A. FloudasDepartment of Chemical EngineeringPrinceton UniversityPrinceton, NJ 08544(609) 258-4595

Area lOb ChairmanDuncan A. MellichampDepartment ofChemical & Nuclear EngineeringUniversity ofCaliforniaSanta Barbara, CA 93106(805) 893-2821Fax: (805)893-4731Bitnet: [email protected]

Area lOb Vice ChairmanChristos GeorgakisChemical Process Modeling & ControlResearch CenterLehigh University111 Research DriveMountaintop CampusBethlehem, PA 18015-4781(215) 758-5432Bitnet: CGOO@LEHIGH

Area lOc ChairmanRajeev GautamUOPTarrytown Technical CenterSaw Mill River Road @ Route 100Tarrytown, NY 10591(914) 789-3206

Area lOc Vice ChairmanMark A. StadtherrDepartment of Chemical EngineeringUniversity of Illinois1209 W. California StreetUrbana,IL 61801(217) 333-0275Bitnet: MARKST@UIUCVMDInternet: [email protected]

Area 10d ChairmanDoraiswami RamkrishnaSchool ofChemical EngineeringPurdue UniversityWest Lafayette, IN 47907(317) 494-4066

Area 10d Vice ChairmanJeffrey C. KantorDepartment ofChemical EngineeringUniversity of Notre DameNotre Dame, IN 46556(219) 239-5797Fax: (219)239-8007

AIr-hE Council LiaisonThomas F. EdgarDepartment of Chemical EngineeringUniversity ofTexasAustin, TX 78712-1062(512) 471-3080

Other Members

Publications Board ChairmanPeter R. RonyDepartment ofChemical EngineeringVirginia Polytechnic Institute & State Univ.Blacksburg, VA 24061(703) 231-7658Bitnet: RONY@VTVMl

Associate Editor, CAST CommunicationsJoseph D. WrightXerox Research Centre of Canada2660 Speakman DriveMississauga, Ontario, CanadaL5K 2Ll(416) 823-7091Bitnet: [email protected]

Peter R. RonyandJosephD. Wright

CAST Communications regularly solicits articles-such as the one on SimuSolv (Dow) in the March 1989 issue orthe one on APT (Texas Instruments) inthis issue-that describe the characteristics of commercial software orientedtowards the specific needs of the CPI.We hope to publish these types of contributions in the Summer issue of thenewsletter. The Winter issue will usually contain feature articles from oneor more of the CAST award winners.Authors retain copyright ownership totheir contributions, and can distributeor publish them elsewhere withoutpermission from the AIChE or the CASTDivision.

Exciting things are happening at CrayResearch, and we thank Steve, Dick,and Bob for taking the time to reporton their program on supercomputingfor chemical engineering. All threehave provided their email and FAX addresses, and are very responsive toelectronic inquiries.

For the first time since your editor assumed responsibility for the newsletter, he is coauthor of a feature article.Karl, who clearly is related to the editor (one non-related Rony runs a commercial bakery north of Milan, Italy)received his B.S.ChE from VirginiaTech in 1987 and works three hoursfrom Blacksburg. He is interested inpromoting undergraduate ChE laboratory education, and reports in thisnewsletter on the adaptation of a simple, widely-used, simulation package,TUTSIM, for the creation of process simulation models that operate in realtime while communicating with a programmable logic controller. APT, forApplication Productivity Tool, is a13 Mb CASE tool based upon objectoriented programming (GOP), graphical user's interface (GUI), and an extensive database. Readers may be interested in studying this software just

to see how it is designed and struc. tured.

At the March 1990 Orlando AIChEmeeting, we were planning to publishin the Summer issue excerpts from thelead article, etVisualization: Expanding Scientific and Engineering Research Opportunities," by Thomas A.DeFanti, Maxine D. Brown, and BruceH. McCormick (see pages 12-25 in theAugust 1989 issue of_Computer magazine, an IEEE Computer Society publication). We finally decided not to reprint the article, but rather to directreaders to it.

CAST Communications received the"best division newsletter" award atthe spring 1990 AIChE Officer's Conference. Your newsletter editorial teamJoe Wright, Colette Totino, and PeterRony-thank the nominators and theAIChE for this kind recognition.

Chairman's Report

by JosephD. Wright, Chairman

In each issue of CAST Communications, the Chairman is invited topresent a status report of the Divisionor to present questions of interest tomembers for their consideration.

1

Much of the news in the Division hasbeen presented in other features ofthis issue. Therefore, it would seemmore relevant to raise matters of longterm importance to our members.Your Division Officers would verymuch like to hear your views on keyquestions facing our termtions.

One of A.I.Ch.E.'s most important objectives is to increase its membership.We in the CAST Division also subscribeto this objective since it is through theDivision fees and the generosity of ourAwards sponsors that we are able tooffer services to our members. Butwho are our members? Looking backto the founding era of the CAST Division, we observe t4at we have grownout of a group of chemical engineerswho were interested in machine computation. Then we expanded to include those interested in appliedmathematic's, control systems, andcomputer simulation. Today CAST iscomprised of four Areas: Systems andProcess Design, Systems and ProcessControl, Computers in Operations andInformation Processing, and AppliedMathematics and Numerical Analysis. To many of our colleagues inchemical engineering, this group ofeclectic engineers appears to be on theperiphery of where chemical engineersought to be. This might encompassthe design of chemical plants, thestudy and applications of the fundamental concepts of energy, mass andmomentum transfer, the application ofprinciples of chemical reaction engineering, and so on. But to play with,or even use Computing and SystemsTechnology? These are within the sacred domain of electrical engineers,applied mathematicians, and others not chemical engineers! However, mycolleagues in the CAST Division will bequick to point out that chemical engineers use these technologies alongwith a host of enablers to function better, lnore efficiently, and, even to solveproblems that otherwise could neverbe attempted.

.. "'_ .......... 0 (Joe Wright)

• Target: put together a set ofbylaws to present to AIChE Councilby August.

e Plan to have contractor starton guidelines by mid-1991.

by Maria Burka, Secretary/Treasurer

1. Process Data Exchange Institute(PDXI). Discussion on this subjectincluded:

• AIChE has given its blessing.

• Questions have arisen aboutthe scope of the data to be collectedand about potential conflict withDIPPER.

• Originally planned to raise$100,000 to get started. If thisdoes not materialize, will get going with whatever is in hand.

• Overhead is a problem. Howmuch will AIChE contribute? Probably 10% or less.

• Plan to charge a flat rate of$5,000 per company.

• Would take two years to comeup with standards (should becalled guidelines instead!).

• Will be in the mail by June1990.

and for new engineers, trained to usethe power to findcations not yet comprehended. Thiswill be the new mandate ofwe will be seen in the· long term as theDivision that encompasses somethingof what every A.I.Ch.E. member shouldknow. The outcome will be that virtually all members will a inthe CAST Division.

One of the key functions of the CASTDivision is to provide programmingfor our local and national A.I.Ch.E.meetings. As we look to the futurewith incredible new computer powerat our fingertips, a major contributionof CAST can be to provide tutorial andawareness sessions for the many engineers who will not have received thistraining. This will benefit the manymembers of A.I.Ch.E. seeking new skillsand may entice them to Division membership. We will look for leadershipfrom our many pioneering members

............ '.......... 'V' ........ _ ...... we may, as a Division, be focussed on needs more advancedusers of computing and systems technology, the impact of these new technologies continues to be astounding.

of us now have desktop computing power that far surpasses that of

or company computing centers in very recent past. In the near

we will have common access todesktop systems and engineeringworkstations that will be rated at 50to 100 MIPS or more. What are we go-

to do with all this power? The nextgeneration of computer tools will allow us as designers and practitionersof chemical engineering to use visualization techniques to ~(see" the effectsof variables in our designs, to investigate ~(what if' scenarios for extremelycomplex cases and to explore the sensitivity of the outputs by viewing the entire space of the design parameters.This power will influence the practiceof chemical engineering in ways notyet determined. Already we have begun to emerge from the world of unfriendly computer systems into therealm of graphical user interfaceswhich provide us with easy access tothe most powerful design tools wehave ever seen. They will be availableto novice engineers and will have captured the collective wisdom of research and development in the industrial and academic fields of chemicalengineering. The CAST Division mustrespond to these forces and provide itsmembers with the tools and theknowledge to update their skills.

This issue of CAST Communications includes an application form for our Division. Pass it on to a colleague who isnot a member of CAST. Help us grow.To help you respond to questions thatmay be posed, here are a few answers.The key benefits of CAST Divisionmembership include: two issues peryear of this newsletter; complete programming in A.I.Ch.E. meetings for sessions of interest to members; notification of ((Calls for Papers"; liaison withother Societies; short courses; andfrom time to time, special servicessuch as the Paper Retrieval Project described elsewhere in this issue. Not abad deal for $5 per year!

In today's technological world, one ishard pressed to find someone who doesnot use computing and systems technology daily. What engineer does notuse a calculator, a computer simulation, a filing or electronic mail system,a spread sheet, or at least a word processor? Even managers have a need touse critical path planners, personal information planners or company database systems. This then leads me tomy key questions. Who should belongto CAST and what should CAST provide?Given the wide acceptance of computing technology, should it not followthat all engineers (at least those whoare members of A.I.Ch.E.) be membersof CAST? Perhaps the view that CAST isreally an organization too narrowly focussed towards expert users of thetechnologies and not on common applications is correct. If this is true,should we broaden the scope of CAST toprovide services to a wider spectrum ofusers? As well, should we address theneeds of those who only require personal information managers, electronic mail, or word processors? Do weprovide sufficient help in simple evolutionary uses of our technologies?Probably we could offer more completeservice, but would this attract moremembers? We would appreciate receiving your suggestions on ~hese various points.

2

• Hope to have at least 20 com.panies involved.

• Are in the process of mailingflyers to prospective companies.

2. AIChE Council is presently goingthrough a strategic planning process. It is looking at the objectivesof the organization, detailedattributes, long-range goals, etc. Someof the ideas being considered include:

• Hopes to increase membershipof undergraduates by 30% between 1989 and 1992.

• During this time period, AIChEplans to increase overall membership by a couple of thousand newmembers. AIChE wants to developmeasurable criteria on what is agood chemical engineer.

• Hopes to have break-even finances.

• Increase the sales ofCEP.

• Headquarters will work with local sections more in the future.

• Study the overlap between divisions and how to resolve conflicts.

• Try to increase participation incontinuing education courses.

• Potential problems with thecreation of the new SeparationsDivision. Other divisions, such asthe Environmental Division, areworried that it will cut into theirmembership.

3. CAST needs to make some slidesshowing the mission, objectives,etc. of the CAST Division, to beshown at meetings, presentationsto potential members, etc.

4. AIChE wants multiple representatives at the June Officer's meetingfrom each division. The feasibilityof this is unclear.

Secretary/Treasurer's Report(Maria Burka)

We still have not received the moneyfor the advertisements that ran in thenewsletter. This needs to be straightened out with AIChE headquarters.

1st Vice Chairman's Report

Concerning long-range planning, JoeWright brought up one project that wemight begin thinking about: computer imaging (visualization). Area lOAwould like to have a session on this atsome future meeting. All officers wereasked to think about long-range planning for CAST and send any ideas theyhave to Rex Reklaitis.

The June Officer's meeting includeddiscussions about AIChE's long-rangeplan (three-year, long-range plan,1989-1992, as discussed by JoeWright). CAST needs to comment on it.

2nd Vice Chairman's Report(Ignacio Grossman)

As of March 1990, there are five nominees for the CAST Award. The procedure for the selection of award winners was reviewed. The committeethat makes the final decision consistsof the six CAST directors plus five others chosen by the 2nd Vice Chairmanwith appropriate representatives fromboth universities and industry. AIChEis planning to put out the book for thefall Chicago meeting by the summermeeting. This places a much earlierdeadline on when information onaward winners have to be submitted toAIChE headquarters.

Past Chairman's Report (BruceFinlayson)

Bruce sent a report on the paper retrieval project at the November 1989AIChE meeting. The information retrieval cost $2,500 and was very timeconsuming. Bruce spent a large portion of his time at the meeting justworking on software problems, andhad little chance to attend any sessions. CAST should not continue to dothis on its own; instead, the responsi-

3

bility should now be handed over toAIChE. Whether we should continuethis at the Chicago meeting on ourown was debated.

Report of the Directors

Herb Britt wrote an advertisement(published in the last newsletter) toencourage membership in the CAST Division. Jeff Siirola mentioned thatthere were no brochures about CAST,or the CAST slide at any of the sessionsat the Orlando meeting. A procedureneeds to be formalized about whose responsibility it is to make copies of theslides, to bring them to all AIChE meetings, and to train session chairmen touse them. Peter Rony suggested. thatwe display copies of the Newsletter atall sessions as a deliverable productthat comes with membership. Decisions will have to be made on these issues before the Chicago meeting.

Programming Board Chairman(Jeff Siirola)

CAST participation in the Chicagomeeting keeps growing. We are nowscheduled to have 26 sessions, withthree running in parallel at anyonetime. All of our requests were accepted.

PSE is finally getting going. It is behind schedule, but should be on tracksoon. FOCAPD 89 proceedings are in themail.

Jeff attended the AIChE workshop forprogramming board chairmen andbrought back the following: (1) AIChEhas hired Peter Knox as a permanentemployee to be in charge of programming. (2) The workshop concentratedon themes: specialty programming,block programming, etc. (3) Underrepresented groups (young engineers,etc.) who do not attend AIChE meetingswere identified and solutions soughton how to get them to the meetings(4) The number of attendees at meetings has been steady over the last fewyears. (5) Meeting Progranl Chairmen

(MPCs) are poorly trained; somethillgneeds to be done to remedy this. (6)Outcome of the workshop: plan tomeet every February to set themes fortwo years out. Each meeting will alsohave a theme. The theme for theHouston meeting is operations; CAST

fits in fine.

of the Area Chairmen

Area IOC wants to know how to get,and who pays for, visualizationequipment-movie projector or VCR-forthe Chicago meeting. Siirola will contact AIChE and request that they provide such equipment. Otherwise, CAST

will have to pay for it.

New Business

Larry Biegler was chosen to representthe CAST Division on the ResearchCommittee. Since the committee always meets during the day for a wholeafternoon during the middle of AIChE

meetings, CAST participation is onlytentative; it is questionable that anyone wants to miss out on a whole afternoon of presentations to attend theirmeetings. We will decide after a fewmeetings.

Report of the CAST DivisionProgramming Board

(March 1990)

by Jeffrey J. Siirola

The members of the 1990 Programming Board are listed on the insidefront cover of CAST Communications.CAST Programming activities can besummarized as follows:

1. 1990 Spring National AIChE Meeting, Orlando (March 18-22 1990):The CAST Division sponsored 13sessions (IDA, 4; lOB, 1; 10C, 7;and joint 10AIIOC, 1). This wasone less than originally planned.

2. 1990 Fall Annual AIChE Meeting,Chicago (November 11-16 1990):Two additional sessions were requested by Area B, bringing thetotal requested to 16.5 sessions(IDA, 6; lOB, 9; 10C, 6; lOD, 5; andjoint lOB/15C, 0.5). The MeetingProgram Chairman has devisedsessions of 2 hours, 3 hours, and3.5 hours length and allocatedsome of each length to each CAST

area. One of the lOB sessions willbe poster format. The MPC had seta March 12 deadline for his reviewof accepted Proposal to PresentForms, a date that was in conflictwith our published April 1 deadline for receipt of same. I assuredhim of the status of each session,but that we reserved the right tochange papers accepted until April1. Subsequently, he extended thedeadline to April 18.

3. Chemical Process Control IV

(South Padre Island, TX, February1991): Area lOB is continuingplans for the next specialty conference in the control series. HarmonRay is conference chairman.CACHE Corporation will facilitatethe meeting arrangements. Thetechnical program has been finalized.

4. Process Systems Engineering '91(Montebello, Quebec, August1991): The CAST division will cosponsor this continuation of the PSE

series of meetings with the Canadian Society for Chemical Engineering, the Institution of Chemical Engineers, and the Society forChemical Engineering, Japan.Gerry Sullivan is conferencechairman.

5. 1991 Spring National AIChE Meeting, Houston (April 7-11 1991):The CAST Division is requesting12.5 sessions (IDA, 4; lOB, 3; 10C,5; and joint 10C/5D, 0.5). Initialcalls for papers appeared in theWinter issue of CAST Communica-

4

tions; final calls will appear in theupcoming Summer issue.

6. 1991 Fall Annual AIChE Meeting,Los Angeles (November 17-221991): The CAST Division is requesting 27 sessions (IDA, 6; lOB,7; 10C, 6; 10D, 6;joint lOA/lOB, 1).Texts for the first call for paperswill appear in the Summer issue ofCAST Communications.

7. 1992 Spring National AIChE Meeting, New Orleans (March 29 toApril 2, 1992): Although the program is not finalized, it looks as ifCAST will request about 12sessions for this meeting. The preliminary program will also be published in the Summer issue of CAST

Communications.

8. Foundations of Computer-AidedPlant Operations (FOCAPO '92):Plans for the second FOCAPO conference, summer of 1992, are under way. Mark Stadtherr and JohnHale are conference cochairmen.The advisory committee met inOrlando. CAST area 10C is thetechnical sponsor and CACHE Corporation will again be approachedto facilitate the conference.

Programming Comment:

An AIChE programming retreat washeld just prior to the Orlando meetingby the planning team of the EBPC. Topics included block programming,meeting formats, joint sponsorships,regional meetings, outreach, and content and presentation quality. Resultswill be presented at a CAST ExecutiveCommittee Meeting.

Also, the evaluation team of the EBPC

is scheduled to meet in Orlando to devise a workable audit system to critique and improve our programmingeffort. The CAST Programming Chairman will represent the division atboth of these meetings and report results to the Executive Committee.

$ 438

Report on Paper RetrievalProject, CAST Division, AIChE

Meeting (November 1989)

At the AIChE meeting in San Francisco(November 1989), the CAST Divisionhad a paper retrieval project. Attendees at the meeting were invited toleave their address and a list of papersfor which they desired to receivecopies. This information was enteredinto a database program by sessionaides. After the meeting, the information was sorted by presented paper,and a list of mailing labels was printedand sent to each author. Each authorwas requested to send copies of his/herpaper; there was no mechanism eitherto ensure that the authors did sendtheir papers, or that all of the papersactually existed. This report summarizes information about the project.

Overall statistics are:

226 people used the service

170 papers were in 22 sessions

2,531 requests for papers were made

1,326 of these were for individual papers

1,205 of these were for papers, as partof all the papers in a session

154 requests were made for all papersin a session

For papers that were on the officialAIChE list of reprinted papers:

70 of 170 papers were reprinted (41%)

543 of 1,326 individual requests werereprinted (41%)

546 of 1,205 papers from session requests were reprinted (45%)

To ensure the success of this type ofproject:

The data must first be collected. Inthis experiment, it was entered into aportable computer (on loan fromXerox) by session aides using dBASE II

software that Bruce Finlayson hadavailable. The session aides liked theprocess, were pleased to use a computer, and were knowledgeable about database programs. Bruce wrote and ranthe computer programs after the meeting to print the mailing labels as afunction of paper or session requested.A secretary typed the mailing labelsfor the speakers using the abstractbooklet. Each author was then sent aletter requesting him/her to sendcopies of his/her paper to the namesenclosed.

Problems did arise. Somebody mustbe in attendance at the computer terminal during the meeting full time, orat least at the start of each day, beforeand after lunch, and at the end of theday. Bruce Finlayson performed thisservice, and missed most of the technical sessions. The address format usedin the data retrieval program was notwell suited to foreign addresses. Thenumber of papers and sessions thateach attendee could request was toolow in the computer program; someaddresses had to be entered twice. Thebiggest problem was that the projectwas so successful that it required significant time to prepare mailing labels. The initial program to do thiswas inefficient because there was alarge number of labels. The programwas run on an IBM-class personal computer.

The total cost of the project was $438plus cost itenlS absorbed by the University of Washington and by Xerox:

Session aides, two per day,4 days, budgeted at: $ 240

Postage, envelopes (Universityof Washington): 94

Secretarial time: 104

Rental of computer, 4 days(from Xerox):

Running dBASE, 3 days(University ofWashington):

Total, minimum cost:

5

If it is assumed that the only papersthat were actually sent were the onesthat were preprinted (1,089) by AIChE,the cost per paper to the CAST Divisioncan be computed to be $0.40, minimum. Not included 'were the costs ofphotocopying and mailing by authors.For an average of 20 pages per paper,the estimated cost for 1,089 copies was$2 per paper, or a total of $2,180 spentby authors.

CAST Division ReceivesOutstandingNewsletterAward at AIChE Officers

Meeting

by Ignacio E. Grossmann, 2nd ViceChairman

The 1990 AIChE Officers Conferencetook place in Buffalo on June 10-12,1990. The three major themes at thatmeeting were: (a) Volunteers and Effective Chapter Operation; (b) Resources for Successful Programming;(c) Communicating Effectively in the90's. A large percentage of the participants were from AIChE Local Sectionsaround the country, while there wererelatively few representatives fromAIChE Divisions.

A number of the sessions were conducted by facilitators who outlinedvarious strategies for recruiting members and for communicating effectively the activities of the AIChE. Information was also provided on the resources that are available from AIChEfor programming meetings, and forhelping local sections to identifyspeakers with current topics. Participants also had the opportunity to interact with members of AIChE Council.The meeting concluded with workshops that were offered in the areas oflocal section finances, public relations,government interaction committeesand the institute's long range planning.

Finally, in order to recognize the importance of disseminating useful andtimely information to AIChE membersthrough newsletters, awards were given for the first time to the best newsletters of the AIChE divisions, local sections and student chapters. Wearehappy to announce that the Computing and Systems Technology Divisionwas the 1990 recipient of the ~eOut

standing Newsletter Award" whichwas delivered by AIChE President Sheldon Isakoff. Congratulations are in order to our editors,Peter Rony and JoeWright, for this well deserved award.

AIChE Recognition System

Purpose:

The AIChE Recognition System is implemented beginning in 1989 to recognize and publicize the activities of agroup or individual in the organization whose performance showed significant accomplishment toward theInstitute's mission and objectives.Presentation of the e~AIChE RecognitionAward" is intended primarily forsingle-event performance, but can beused to recognize long-term performance in special cases.

Mechanics:

(1) No more than 10 awards will bepresented in one calendar year.

(2) Nominations for the award can bemade by individual member(s) orgroups in AIChE, but must be sponsored by a current council member. Nomination is accomplishedby completing form ~~1989 RS," acceptance by a Council member forsponsorship, with forwarding tothe current AIChE secretary forprocessing and review by the executive committee.

(3) Decision on awards presented willbe the responsibility of the executive committee, with the secretary

acting as coordinator of the system.

(4) Each award will be presented atan AIChE meeting, and will be subsequently publicized in AIChEXTRA. The award will consist of aplaque and an explanation letterfrom the current president of theInstitute.

(5) An official history of the system,including record of each recipientand sponsored nominee, will bekept in the ,member services department at the New York office.

(6) Nomination forms can be obtainedfrom the member services department at the New York office.

If you have questions about this newe~recognition" system, please contactGordie Ellis, the local section and division coordinator at AIChE headquarters; call (212) 705-7328.

Procedure:

(1) Nominator completes items 1, 2,and 3 of form 1989 RS.

(2) Nominator contacts potentialCouncil sponsor, forwards form tohim/her.

(3) If Council sponsor approves,he/she completes item 4 and forwards form to current secretary. IfCouncil member chooses not tosponsor, he/she contacts nomineewith explanation.

(4) Secretary submits all sponsorednominations~ to Executive Committee for review and decision.When the Executive Committeechooses a particular nominationfor award, the plaque and letterare prepared, and a Council member is designated to present theaward. The award will be keptconfidential until presentation tothe recipient(s). The form is further completed by the secretaryand forwarded to the member services department for recording. Ifthe Executive Committee declines

6

a nomination, the Council sponsoris notified.

(5) The plaque and letter are presented by a Council member at anAIChE meeting. AI,ChEXTRA publicity follows the presentation.

(6) The completed form is forwardedto the member services department for record keeping.

AIChE Recognition System Nomination Form:

(1) Group/individual Nominee:

(2) Summary of Accomplishment (Relating to AIChEMission and Objectives):

(3) Nominator(s), Date, and Telephone No(s):

(4) Executive Committee Approvaland Date:

(5) Date/Location of Plaque Presentation:

AIChEXTRA Publicity (Issue):

1989RS

----------------------------------

lectures atconferences will

his views andanyone work-

His method sea-his

mostprocess ~vc~tornc

eeWithout question, Roger is and remains the most distinguished chemical engineer active today in the process systems engineering field. Overthe 3D-plus years of his professionalcareer, he has contributed significantly and led in the number of importantresearch areas, including process dynamics and simulation, process control, nonlinear programming, flowshee~ing methodology, and solutionmethods for algebraic and differentialsystems. He has had a major impacton the field not only through his workbut also through that of his students,who have grown to form a very prolificacademic tree. He has led the development of the SPEEDUP system, which isbecoming an important commercialprocess simulatiOn/optimization software product. More recently, he was

If the process systems adventure wasinitiated by Roger's vision, then clearly it has been his ability to inspire thathas launched an international legionof enthusiastic voyageurs. His academic family, now six generationsstrong, comprises almost two hundredoutstanding scholars and industrialists who themselves are now makingprofound contributions in the processsystems field. This inspiration is alsoprojected to those far beyond the

processJ1.11\lJJl.. .Ii.""'';;'JI..,;:J. as well as of multi-

His achievements inae'VeJ.ODmf~ntof efficient numerical

methods solution and optimization of differential-algebraic systemshave been recognized by both thenumerical analysis and chemical engineering communities. This entirebreadth of work comes with both deeptechnical insight and an uncompromising standard of excellence.

ventedit was the eevision" Y"\1fO.'"),C'r'V"lI'il-·r.d"dl

that was the contribution.SPEEDUP is one of the TAl"on'1lAC~1t"

and simulation

Professor Sargent's interest in computing applications is evident from hisvery first publication (1958) entitled~~Applicationsof an Electronic DigitaiComputer in the Design of Low Tem-

The supporting statement for hisaward includes the following:

Proat Im-

Science, Technology,career has

over thirty years duringwhich time his name has become synonymous with Process Systems Engineering, is the recipient of the 1990Computing in Chemical Engineering

Sponsored by Simulation Sciences, Inc. and Dow Chemical, TheComputing in Chemical EngineeringAward is given '-in recognition of anoutstanding contribution in the application of computing and systems technology to chemical engineering. Thewinner receives a plaque, a check for$1500, and an invitation to be theafter-dinner speaker at the CAST

Awards Banquet at the Chicago National AIChE Meeting in November.

7

Other supporting letters stated:

Probably the most lasting contributions that George has made have dealtwith the development of new mathematical software tools. EPISODE andDISPL are probably the most widelyused and well know:n of the varioustools George has developed. Thesetools were spawned from collaborativeefforts with other top mathematicianssuch as Alan Hindmarsh, Gary Leaf,Mike Minkoff, and Peter Brown.George, Alan, and Peter ha~e recen~ly

released a new ordinary dIfferentIalequation solver called VODE.

The Institute of Mathematics and itsApplications recently el.ect~d ~eorge

as a Fellow, following hIS hIghlIghtedpresentation at their co~ference onnumerical ODEs at ImperIal Collegelast summer... The paper deals withcomputational experiments withVODPK a Krylov variant of vaDE.With these new methods, reductionsin storage requirements of 9: 1 and improvements 5:1 in execution speedhave been observed for heat and masstransfer and kinetics-diffusionconvection problems.

Dr. D. is one of thoserare scientists who has a

in twomathematics and

and this has'XTlIn,INIl:~n ·'·....... 1Id"'11I1I1Ul contributions to bothfi Id CAST should thoseIe s...

who have established atwo fields that are so vi

to the continuedenrichment of our profes-

The award nomination statement1;;-i-'ll"l>nirl.n,nc included

George received his B.S. de.gree. inMathematics at Creighton Universlty,and both M.S. (Mathematics) and Ph.D.

(Applied Mathematics) from IowaState University. At the University ofPittsburgh, from 1963 through 1980,he taught courses spanning mat~

ematics, computer science, and chemical engineering; he directed 6 Ph.D.

theses and served on over 30 Ph.D. committees. He is a.member of SIAM, ACM,

AIChE, IAMCS, IEEE, the AmericanMathematical Society, IMA (British),Sigma Xi, the National Speakers Association, and, of course, ToastmastersInternational. One supporting letter

George D. Byrne, a Research Associate at Exxon Research and Engineering Company, New Jersey, will be pr~

sented the 1990 Award at the CAST DIvision Award Dinner at the AIChE National Meeting in Chicago. He wascited e~for his contributions in numerical ~ethods, especially differentialequations, related software,. and ~hei~

applications in chemical engIneerIng.The Computing Practice Award, sponsored by Pergamon Press, honors ~n

outstanding effort that has resulted Ina specific embodiment, or poss.ibly. anindustrial or commercial applIcatIon,of computing and systems technology.The award consists of $1000 and aplaque.

..n......,.lI-lI"""ll'7f<C:.bt'lrt.l1 (static dynamic),synthesis of engi-

in the creation isU.II..II.JL"VJL.lI.VllA ....

the Director of theSystems

e~Roger's contributions to chemicalprocess engineering are legendary.These include not only fundamentaladvances in numeric methods, simulation, optimization, control, synthesis,and computer-aided design, but also,and especially, in education. He hasinspired many of his students with acontagious excitement for process design research. This has led to a dynasty of similarly-inspired researchers,which is represented in virtually every process design research organization in the world."

Rogermaintained very high standards of

t::'A..... CJ..lI.l(;;.ll..ll.'-'''-', leading to contributions oflasting impact.

academic scholarshipwith industrial relevance, he has been

to offer contributions of both archival value (e.g., his work on optimization non-linear systems' analysis)and ~ractical (e.g., SPEEDUP) significance."

8

Computer-aided design and simulation of multicomponent separationprocesses normally is carried out using the very well-known equilibriumstage model. Briefly, this model includes the assumption that the

Krish Krishnamurthy, the award winner, is a Lead Engineer at The HOC

Group, Inc. in Murray Hill, New Jer...sey. Krish, who performed his awardwinning research while a graduatestudent of Ross Taylor at ClarksonUniversity, was cited ~~for the development and application of a complete,mass-transfer rate-based model ofmulticomponent staged separationprocesses. "

The award nomination provided thefollowing identification and evaluation of the accomplishments of thenominee:

The Ted PetersonAward is given to recognize an outstanding published work, performedby a student, in the application of computing and systems technology tochemical engineering. This award,sponsored by ChemShare and IBM,

consists of $500 and a plaque. Theaward will be presented at the CAST

Division Award banquet at ChicagoAIChE meeting this fall.

~~In addition to his project contributions, George continues to be active, ashe has been for many years, in training others and disseminating themethodology of numerical computation in his field-through workshops,seminars, and presentations, eachyear. His dedication to his professionis very great. In personal affairs,George is a highly ethical, princely individual and one who is never too busyto help a colleague... "

~~At Exxon George has contributed tomany programs of chemical

analysis resulting in used in the development of new processes and products.One example concerns

Exxon DeNOx processsmoke stack gas which em-ploys ammonia. Optimizationwas used to adjust chemical kineticsrate coefficients so that ODE solutionsagree with laboratory and field data.The ability to use this code to simulatenew operating conditions has permitted novel applications of the technology in using unconventional fuels inwaste disposal incinerators, e.g., burning rubber tires, while obviating needfor expensive pilot plants. To date 60NH3-NOx plants have been commercially commissioned and more areplanned."

cial simulation tools. It is heavilychemical engineers because it

can chemical kinetics leadingto systems of stiff equations. A knowledgeable member of our staff remarksthat this integrator is so robust and ef-

that it has no competitor."

~~George's published work is highly regarded both by his colleagues in thenumerical mathematics field and bypeople in other disciplines. His list ofpublications includes at least 34 thatare in refereed journals and conference proceedings. Over the years, thecontent of George's papers has shiftedfrom abstract numerical analysis tostudies of methods in applied situations, but with the latter resting on asound analytical base. One of our mostrecent joint papers, an invited surveyof stiff ODE methods and software, onwhich George was the lead author, hasreceived considerable acclaim andaround 100 reprint requests to date,from allover the world. Another measure of the quality of his work is thepopularity of the software and he haswritten (or coauthored). One sequenceof ODE solvers that he and I wrote hasbeen requested by hundreds of people."

~~As a pioneer in the field he and A. C.Hindmarsh developed the well-knownEPISODE software package for the efficient solution of systems of ordinarydifferential equations. EPISODE wasdistributed to over 125 central computing sites world-wide by the National Energy Software Center, and isused in broadly distributed commer-

~~The application numerical methods and software to the solution of difficult problems in chemical engineering has always been a prominent partof George's work. Various kinds ofchemical kinetics models have provided the main motivation most of

work he has done on or withstiff ODE solvers, PDE solvers, and DAE

solvers... At George has con't1Y"lIhll'tOrt fundamentally to the solutionof numerous chemical engineering

a thermal DeNOx process involving kinetics simulation and parameter fitting; (b) an amine involving1-D transient reaction-diffusionconvection simulation; and fixed-

reactor simulations involving DAE

formulations and solution."

9

J

The chemical and materials based industries are using supercomputers ina similar manner to obtain molecularbased properties about their productsand processes. Thermodynamic data,such as heats of formation and heat capacities, are being calculated on supercomputers, providing informationwhich can often be too costly or timeconsuming to determine by experimental measurement. The molecularinteractions involved in catalytic processes are being studied on supercomputers, yielding knowledge aboutchemical processes that are typicallyunobtainable by experimental means.Supercomputers are also being used tostudy the impact of chemical productson the environment, such as in themodeling of the global warming potential from chlorofluorocarbons. Polymer flows are being modeled using

Supercomputers are R&D

activities in the chemical process industry (CPI). Companies in the pharmaceuticallbiotechnology industry areusing supercomputers to better focusand accelerate their R&D efforts in designing new drugs, vaccines, and herbicides. Large-scale molecular simulations of drug-enzyme complexes areused by researchers both to learn howa drug works on the molecular level,and to identify which of the many candidate compounds shows the mostpromise of being an effective drug.Further studies, requiring synthesis ofthe compound and experimental testing, can then be focused upon the mostpromising candidates.

J. D.and algo

Im'plem€~nt]Lng it could be thePV'-'Ip;;.,...Il.JLJI..A. .......... JL~ of a new era in tray-type

simulation."

supporting letter stated the following:

a plea in 1980 by the lateProfessor H. Sawistowski for a replacement of the equilibrium-stageapproach and its accompanying plateefficiencies for plate columns andHETPs for packed columns, R. Krishnamurthy and Ross Taylor began the development and experimental verification of a rigorous and robust algorithmfor a transport-rate model of continuous, steady-state, separation operations having countercurrent contacting of fluid phases. Incorporated intothis algorithm, the first of its kind,were reliable methods for computingmulticomponent transport coefficients; rigorous expressions for fugacities, enthalpies, and densities; androbust methods for converging thelarge set of highly nonlinear equations..."

causemost In

years after Krish'stion on subject, a J.J.VJl.A."",'-4 'Ul.J!.J1..JL"-',JL .JL~.,A.Jl.JL.A.

model has been Aspen one of the leading flowsheet simulation In a paper

Although models based on masstransfer principles are not new,Krish's model was designed to appealto computer simulation specialists.This approach must have worked, be-

Krish's first paper describing themodel was published in 1985. This paper was soon followed by other articlesshowing that the nonequilibriummodel was capable of true predictionsof both tray and packed column performance, in columns ranging from laboratory to comflltrcial scale.

streams a stage arein with other. How-ever, in actual operation, columns donot operate at ,-,y''I".lI..JLJI..JLp.JJI.

Beginning in 1980, R. (Krish) Krishnamurthy began the implementationof a model of multicomponent separation operations that considers them tobe the rate processes they really are.The model that Krish implementedavoids all a priori computations ofstage efficiencies (for tray columns)and HETP's (for packed ones). Materialand energy balance equations aresolved simultaneously with properlyformulated models of mass and energytransfer in multicomponent systems.

Equilibrium stage calculations alsoprovide the conventional route to designing packed columns. In place of a .stage efficiency, the ttHeight Equivalent to a Theoretical Plate" (HETP) isused. HETP's suffer from many of thesame shortcomings that plague stageefficiencies; they are not the same fordifferent components in a multicomponent mixture, for example.

conventional way of dealing withdepartures equilibrium in mul-tistage towers is through the use

and/or efficiencies. Thereare many different approaches to de

calculating efficiencies, andis no consensus on which is best.

For multicomponent systems, the efficiencies of different components maynot be the same.

10

computational fluid dynamics me.thods to better understand and designinjection molding processes.

More and more companies in the CPIare making the strategic decision toimplement supercomputing as an integrated part of their R&D activities. Table 1 shows Cray Research's customersin the CPI, along with chemistry related research institutions. Cray Research also has nearly 30 customers inthe petroleum industry, many ofwhich are running chemistry andchemical engineering codes along withtheir traditional seismic and reservoirsimulation applications. In addition,approximately 35 companies are usingCray Research systems for chemistry/chemical engineering at university and NSF-sponsored supercomputercenters.

Table 1. Cray Research's Customers in the CPI

Installation Customer Machine

1986 Dupont CRAY 1/ANCI CRAY X-MP/28

1987 Scripps CRAY X-MP/14seDupont (upgrade) CRAY X-MP/28

1988 Exxon R&E CRAY X-MP/14se

1989 Monsanto CRAY X-MP/116seSumitomo CRAY X-MP/116seScripps (upgrade) CRAY X-MP/116se

1990 Eli Lilly CRAY-2S/2-128DuPont (upgrade) CRAY Y-MP4/332

During the second half of the 1980s,CPI application programs for supercomputers have tmproved substantially in functionality and performance.Industry standard codes in the areas ofmolecular dynamics, quantum mechanics, computational fluid dynamics, and structural analysis are available and optimized for supercomputersystems. The increase in the numberof high-performance application codeshas generated a growing interest inusing supercomputer simulations forindustrial research.

Surprisingly, until recently, there hasbeen relatively little work done to develop or optimize application codes forlarge-scale chemical engineering design and optimization on supercom-

puters. However, initiatives at CrayResearch, along with leadingCPI companies, software vendors, and academic research groups, are signalingan end to this trend. We will discusssome of the chemical engineering design tools that are currently being tailored to run on Cray Research supercomputers and focus on the center ofthis application area, the process simulator.

Traditional sequential-modular simulators solve the material and energybalances of multi-component, multiunit processes at steady state. Themost widely used application codes inthis area are PROCESS and PROIII fromSimulation Sciences and ASPEN PLUS·

. from Aspen Technology.

These modular codes are reliable androbust design tools with a long historyof industry use and acceptance.

The CRAY version of PROCESS is currently running at about a dozen CrayResearch customer sites. Version 4.0is available from Simulation Sciencesfor running under both the cos andUNICOS operating systems.

DuPont recently conducted an evaluation of PROCESS running on a VAX !3800

as compared to running on a singleprocessor of their CRAYx- MP/28. In theOctober 1989 issue of Chemical Engineering Progress, Haley and Sarmareported that a single cpu eRAY X-MPincreased performance on an averageof 8 to 10 times over the VAX 8800. Theminimum ratio of CRAY to VAX performance was 1.9 and the maximum was67.9. A cost analysis showed that theCRAY X-MP supercomputer was themore cost-effective solution.

For DuPont useri, the average cost perrun on the CRAY X-MP computer isabout half as much as on the VAX. It isimportant to point out here that manyof the important PROCESS routinescould have run significantly faster ifthey were optimized or redesigned forvector and multiple cpu operation.

11

However, those involved in the Y"\f(O'... lln."I>1I"

from DuPont, Cray Research, andSimulation Sciences chose not to pursue this since the follow-up to PRO

CESS, namely PROIII, was due out in thenear future.

PROIII version 2.0.1 is already available for use on several personal computer platforms. In preparation forlater CRAY releases, version 2.0.1 wasported to the CRAY X-MP and CRAY Y-MPrunning UNICOS.

In June 1990, Cray Research obtainedPROIII 2.5.0 from Simulation Sciences.This is the pre-release of the officialversion 3.0.1, which will be releasedsimultaneously for mainframe computers and eRAY supercomputers dur..ing the fall of this year. Cray Research plans to pursue a highly optimized version of the PROIII simulator.

Cray Research has ported ASPEN PLUS8.2-6 to the CRAY X-MP, CRAY Y- MP,andCRAY-2 supercomputers. The code iscurrently running at three Cray Research customer sites for four customers. The ~tofficial" ASPEN PLUS shellscripts are available on Cray ResearchUNICOS systems. These scripts arenow in place for all variations of ASPEN PLUS runs (edit runs, user databanks, reusable load modules, etc.).

On the standard suite of 49 test problems from Aspen Technology, the unoptimized ASPEN PLUS 8.2-6 code wasan average of 14 times faster on oneprocessor of a CRAY Y-MP system thanon a VAX 8530 mainframe. The speedups typically ranged between 12 and17, but were as high as 25 and 32 fortwo of the test cases. It is important tonote 'here that these 49 problems testvarious features of the ASPEN PLUScode and are not necessarily very computationally intensive.

Optimization work at Cray Researchis well under way on the critical sections of ASPEN PLUS, which mainly include thermodynamic and equationsolving routines. To date, over 20 rou-

50-.----------------,

tines have been optimized for vectoroperation. Six large ASPEN PLUS problems with VAX solution times rangingfrom 12 minutes to 4.5 hours havebeen obtained from Aspen Technologyand a customer for use in Cray Research's optimization efforts. Thespeedup figures for the unoptimizedand optimized CRAY Y-MP versions overthe VAX 8530 version are given in figure 1 for the six large flowsheetingproblems.

ASPEN PLUS Opt:imi~~atic)n

Cray Research is also investigatingequation-based methods for solvingchemical process simulation problems.The equation-based methods differfrom traditional modular simulatorsin that they do not proceed sequentially from process unit to process unit inarriving at a final solution but insteadtreat the entire chemical process asone large equation set that is solved simultaneously. This approach allowsfor the efficient solution of problemsinvolving dynamic batch and cyclicprocessing, startup and shutdownstudies, control and operability studies, ,and operator training. Most importantly, the equation-based approach to process flowsheeting appears to be very amenable to the vec-

The General Algebraic Modeling System (GAMS) family of software is acomplete system for modeling andsolving optimization problems. AeRAY UNICOS version is supported anddistributed by the GAMS DevelopmentCorporation. The system consists of:GAMS modeling language and execution system, and solvers for linear,nonlinear, mixed-integer, nonlinearmixed-integer and nonlinear networkmathematical programming problems. GAMS can be used in such CPI application areas as: process scheduling,retrofitting, batch processing, heat exchanger network synthesis and production, and distribution planning.

SIMUSOLV, from the Dow ChemicalCompany, is an integrated, multifunctional software package for developingand using mathematical models ofphysical systems. This code can beused to model problems in such diverse fields as process engineering,toxicology, pharmacology, chemicalkinetics, environmental sciences, andagriculture. In early 1990, Cray Research obtained the SIMUSOLV code forporting and has begun to optimize itfor CRAY system architectures.

also working to provide highperformance solutions for thesetant design activities.

Aspen PlusCPU Time Comparison

6000

(J)

-g 5000

8~ 4000

Q.)

3000Ei=~

20000-0~ 10000f-

As noted above, chemical engineeringcomputing is centered around the process simulator, but other related application areas include the simulation ofphysical systems (e.g., reaction kinetics modeling), linear programming(e.g., process scheduling), and dynamic simulation (e.g., startup and shutdown evaluation). Cray Research is

RDIST1 RDIST2

Reactive Distillation Problems

Figure 2

Similarly, Simulation Sciences plansto release an X-windows version oftheir PROIII graphics package. Thiswill enable the use of graphics workstations for the visualization of theflowsheet design process, coupled withthe high-speed solution of the processsimulation equations on a CRA.Y Y-MP.

The result being a Network Supercomputing Environment which increasesthe abilities of chemical engineers tosolve critical chemical process problems.

Aspen Technology also producesModel Manager, a graphics front-endfor the ASPEN PLUS program. WithModel Manager, engineers can receiveexpert-system assistance to constructa rigorous flowsheet model. Later thisyear, a workstation version of ModelManager will be released.

frontal code with ASPEN PLUS will bepresented by S. E. Zitney at the AIChE

1990 Annual' Meeting in Chicago.This paper will be part of the Symposium on Parallel Computing sponsored by Area IOC of the CAST Division.

II Unoptimized CRAY Y-MP

ED Optimized CRAY Y-MP

RADFBIG RADFNON CRUDE OPTPROC IPAREAC DRSBIG

ASPEN PLUS Problem

40

Figure 1

Figure 2 demonstrates the impact ofredesigning solution algorithms to fully exploit advanced computer architectures. Two large reactive distillationproblems, RDISTl and RDIST2, requiredabout 90 minutes each to solve on theVAX 8530. A large portion of the totalcomputation time was spent in theHarwell MA28 subroutine, which is ageneral sparse linear equation solver.This traditional solver was replacedwith a new frontal solver. The motivation here is that the frontal routine exploits vector computer architecturesby treating parts of the sparse matrixas full submatrices, thereby allowingarithmetic operations to be performewwith full-matrix code (without indirectaddressing). Using the frontal code,the total time required to solve thesetwo problems in ASPEN PLUS was drastically reduced to only 1 minute forRDISTl and 30 seconds for RDIST2. As aresult, an engineer can now run thisdistillation model many times a day,instead of only once or twice. A moredetailed discussion on the use of the

sC")II)co)(

«>c:: 30:liE>->«a:: 20

£

12

tor and parallel processing capabilities of Cray Research computer systems.

There are many exciting challengesand opportunities for supercomputingsimulations in large-scale chemicalprocess synthesis/retrofitting, design!optimization, and operations/control.

For example, there can be an enormous set of alternatives for the synthesis of a grassroots process design.The search space for a retrofit designcan be even larger. But, with the development of more powerful mathematical programming techniques,combined with increasing supercomputer power, it will become feasible torigorously model, optimize, and automate the synthesis process. It is clearthat there is profit to be gained by optimizing processes to run at some optimal point, but optimization is rarelyperformed on a whole process in theCPI. We expect that the ability to runlarge supercomputer simulations maybring process optimization, even optimization of entire plants having several different processes, into the arena ofcommon industrial practice.

Connectivity between heterogeneoushardware is essential for developingan integrated process engineering environment. As network supercomputing makes it possible, on-line use ofsupercomputer simulation models willhelp to ensure more efficient plant operation, both during normal operatingconditions, and during changeoversand upsets. Cray Research will becontinuing its efforts in meeting thesechallenges in order to bring thesebenefits to the CPI. In the near term,we foresee that the continued improvements in chemical engineeringsoftware for supercomputers willmake process engineering on Cray Research systems a true partner to experimental efforts and pilot plantstudies as the CPI strives to developchemical processes that are more effi-

cient, less costly to operate, and environmentally sound.

We are always interested in hearingfrom others who are interested in supercomputing for chemical engineering. If you have any questions or comments, or would like further information on the work being done at CrayResearch, please contact one of the authors.

Biographical Sketch of Stephen E.Zitney, Richard D.. La Roche, andReA.. Eades

Stephen E. Zitney, Ph.D., joined the Industry, Science & Technology Departnlent of Cray Research in June 1989 as a Senior Chemical Engineer. He earned his B.S.Ch.E. degree at CarnegieMellon University. and his M.S.Ch.E. and Ph.D. degrees from the Univ. of Illinois at UrbanaChampaign. At Cray Research, he is involved inbringing the power of advanced architecturecomputing to the chemical process industrythrough the conversion, algorithmic analysis,optimization, and enhancement of large-scale,chemical engineering codes. His research interests include process synthesis, design, optimization, control, and scheduling. Phone: (612) 6833690, FAX: (612) 683-3099, Email:[email protected]

Richard D. La Roche, Ph.D., joined Cray Researchin May 1990 as a Senior Chemical Engineerwith the Industry, Science, and Technology Department. He received a B.S. degree in ChemicalEngineering from Montana State Universityand worked for two years with Dow Chemicalbefore returning to graduate school. He earnedhis M.S. and Ph.D. degrees in Chemical Engineering from the University of Illinois at UrbanaChampaign. Dr. La Roche joined the faculty atPenn State University in 1987 as an AssistantProfessor of Chemical Engineering. He main-

13

tains affiliation with Penn State as an AdjunctAssistant Professor and continues to conduct supercomputing and parallel computing researchfor chemical process engineering applications atCray Research. Phone: (612) 683-3696, FAX:

(612)683-3099, Email: [email protected]

Robert A. Eades, Ph.D., is the Manager of Chemistry Applications in the Industry, Science &Technology Department at Cray Research. Dr.Eades earned his Ph.D. in Physical Chemistryfrom the University of Minnesota at Minneapolis. He held positions at Argonne National Laboratory and Allied-Signal prior to his joiningCray Research in 1987. His interests and effortsat Cray Research are focused upon the development and application of computational methodsto solve industrial based R&D problenls in thefields of chemical engineering, chemistry, lifesciences, and materials science. Phone: (612)683-3669, FAX: (612) 683-3099, Email: ead[email protected]

TI 500 SeriesProgrammable Logic

Controllers: Merging ofTUTSIM Dynamic Simulation

APT Control SystemDesign

by Karl E. Rony (Texas InstrumentsIncorporated) and Peter R. Rony(Virginia Tech)

A basic question associated with thedesign of programmable logic controller (PLC) software is when to test thecontroller program logic (a) at startup,or (b) during the development of thecode? For alternative (b), appropriatesoftware or hardware tools must existboth to simulate the process and to ex- .ecute the simulation and controllercodes while they communicate witheach other.

The objective of this article in CASTCommunications is to introduce divisional members to both a new controlshardware platform-the Texas Instruments TI545 programmable logic controller (PLC)-and two software tools:

• High-level, sequential and continuous control design software: Texas Instruments Application Productivity Tool (APT).

• Dynamic simulation software external to the controller: Real-timeTUTSIM with a user-block driver for theTI 500 Series controllers.

In combination, the above hardwareand software provide, we believe,state-of-the-art capability in the implementation of inexpensive, midrange programmable logic controllers forthe control of chemical processes. APTshould be of particular interest tochemical engineering software designers because it is a complete, effectiveimplementation of object-oriented programming (OOP) in a chemical engineering process context.

In this article, we will briefly summarize the characteristics of the hardware and software-TI545 controller,APT, and TUTSIM-and then illustrate

their use in the creation of a preliminary control system design for a simple blending process [2].

In addition to their applicability forthe design of industrial processes, webelieve that the above hardware andsoftware have important implicationsfor chemical engineering controls education, specially in undergraduatecontrols and unit operations laboratories, where sufficient time may beavailable for a small-scale design project involving a process controlled by aPLC.

Limitations of space assigned to a feature article in an issue of CAST Communications means that neither thefigures (some of which are not largeenough) nor the accompanying textwill be able to convey-for two differentsoftware packages and one new hardware platform-the amount of information that was successfully conveyed fora single software package in a previous issue of the newsletter [1]. Torectify this deficiency, interested readers are invited to request supplementary information, including half-pagecopies of several of the figures, fromone of the coauthors (KER) at Texas Instruments Incorporated in JohnsonCity [3].

I. The Design Problem

The approach to unfamiliar applications software is usually facilitated bywell characterized problems that areused to test the characteristics of thesoftware. This technique was used byKuenker and Blau in their discussionofSimusolv [1]. Because of the varietyof software tools involved, we have selected the blender unit of a threecomponent blender/packaging process.As illustrated in figure 1, water (H20),an aqueous suspension of componentA, and a dilute aqueous polyelectrolyte suspension agent B are added to theblender tank, where they are mixedand heated. The solution is cooled andthen transferred to a packaging unit(not shown in the figure). Details of

14

the non-steady-state simulation theblender process will be inSection IV of this article.

II. Hardware and ~_."+"1I'IlI:l',",'IlI._Tools

The TI 505 System-a complete midrange control solution-consists of a backplane powered by a TI Model 545 Controller, which provides 0.78 ms/Krelay-ladder logic (RLL) execution,large memory (96K), and 64 preconfigured regulatory control loops. To illustrate the TI505 System I/O capability, a combination 8 input/4 outputanalog module uses surface-mounttechnology to combine on one printedcircuit board the following functionsthat normally require multiple packages: I5-bit analog input resolution;eight differential, isolated input channels; front-panel user calibration; selection of seven voltage ranges on eachinput channel, from 0 to ±50 mVallthe way up to 0 to ± 10 V; four individually isolated (1500 RMS) analog outputs; and 12-bit analog output resolution.

Priced at $2700 for the controller anda powered backplane, the TI545 is thenewest member of the TI 500 Series ofcontrollers, which also includes theTI530T, TI560, TI565, and TI565T.

The TI545 controller, as well as the entire 505 I/O system, is UL Listed, CSACertified, and Factory Mutual approved for Class I, Division 2 Hazardous Locations. The TI545 is manufactured in Johnson City, Tennessee,which received the 1989 ElectronicsFactory Automation Award for Manufacturing and the 1989 TaguchiAward by the American Supplier Institute Incorporated for the successfulpromotion of Taguchi Methods to improve quality and reduce the cost ofAmerican products.

John Arnold, one of the primary designers of the APT software, has developed, based upon his experience withAPT, a systematic approach to the design and development of a controlledprocess. This approach includes thefollowing elements: (a) partitioningthe process into functional areas; (b)identifying equipment areas (sharedequipment and processes); (c) telescopic focusing (changing focus readilyfrom system view to detailed view); (d)top-down design; (e) bottom-up implementation; (f) incremental development; and (g) testing as part of the development process. His bottom-up implementation methodology includesthree levels:

Level One Control: Manual operationand process safety.

Level Two Control: Set of objectivescontrol that imposes operational goalsonto the level-one control system, andprovides a primitive operator connection (interface primarily through thedevelopment system).

Level Three Control: System mergingand expansion, building of the complete control solution, exceptions handling, reports and operator transac-

A basic philosophy of APT is that thecontrol-design solution should reflectthe natural organization of the physical plant.' In arranging the physicallayout of the plant, design engineersusually subdivide the process into unitoperations (mixers, reactors, etc.).While we design each unit independently, we also consider the overall requirements of the integrated process.Process control engineers can use asimilar approach and divide the control solution along natural equipmentboundaries. For example, in figure 1the overall control requirements are to

APT is an object-oriented, program(OOP) design tool that permits a process control design problem to besolved by using a structured, modular,top-down approach. Modular designinvolves breaking a large control design problem into smaller, more manageable pieces. Each subproblem issolved independently, with the subsetsintegrated to provide the total controldesign solution.

The authors of this article subscribe tothis methodology and have applied thelower four layers to the blender process of figure 1.

Layer Ten: Structural displays.

Layer Eleven: Simplify operator interactions.

Layer Twelve: Operator messages andrecipes.

Layer Thirteen: High-level exceptionhandling.

These three levels are further subdivided into a set of thirteen layers:

Layer Zero: I/O definition and namingconvention determination.

tions; actual order and activities become fuzzier.

Layer One (optional): Dynamic simulation of the process.

Layer Two: Discrete device definition.

Layer Three: Equipment protectionvia interlock logic.

Layer Four: First-level continuouscontrol defined.

Layer Five: Translation of first-levelcontrol into operator station.

Layer Six: Procedural control.

Layer Seven: Structural control.

Layer Eight: Objective control.

Layer Nine: Merging with operatorinterface.

Systemrn.llO~SODn,v of

(;1;"1[ ] 110 syl'1'lbol ic nal'1'le table[ ] Deuice def in it ion tab Ie[ ] Dec larat ion tab Ie[ ] usage table

COttTROL~ OBJECT 10

DEVICE10 DECLAREMODULE RECIPEDEVICEDECLARE ---------------------CFCJ S---------------------RECIPE AUTO_RUtt [ ] Autol'1'latic run logic

TEMP_CTL [ ] Tel'1'lp control logicw:mIiIilI LEV_ALRM [ ] Tank leue I alarl'1'l

IttTERLJ< [ ] Deuice inter loel< protect ionSIMULATE [ ] Sil'1'lulation for the process

COMP ILE -----------------SAVED REPORTS-----------------SYMBOL 10 VAL [ ] 10 Va I idat i on Report

SIMULATE VAL [ ] CFC Va I idation Report

CONFIG

BLENDER DeVICES

Figure 1 Figure 2

15

move ingredients into a blender tank,mix them, and then transfer the mixture to a packaging system, which represents an entirely separate equipment area (called a Unit in APT). Eachequipment area (Dnit) has its own requirements that can be determined independently; integration requirements make it also necessary to consider the interconnectivity of the totalsolution. A successful modular designpartitions the problem according towhat is needed solely for each unit operation, and what is needed to interconnect the Units.

The APT hierarchy consists of threelevels: the Program Directory Level,the Program Content Level, and theU nit Content Level. The Program Directory Level contains a list of all programs. A program in APT is that portion of the process that can run on asingle controller. At this level, wework with entire programs (add, copy,delete, restore, archive, compile,download, debug, description, etc.).

The Program Content Level contains alist of all Units that are identified tobe a part of a single program. Thislevel of the hierarchy also includes ameans of identifying devices (for example, pumps, motors, and valves), 110

signals, and ingredients that need tobe accessed by more than one Unit, or

area of equipment; such ingredients,I/O signals, and devices are consideredto be global to the entire process. Atthis level, we work with tables, Units,and recipe templates for the program(add, edit, copy, delete, etc.).

The Unit Content Level contains thedefinition of the actual control processes, including sequential and continuous procedures as well as specialprocedures to handle emergency conditions. This level also includes ameans of identifying devices, I/O signals, and ingredients that are local toeach Unit. Sequential control of theprocess is defined in one or more Sequential Function Charts (SFCs); continuous control is defined in one ormore Continuous Function Charts(CFCs). Figure 2 illustrates the threelevel APT hierarchy.

The APT program, a major contribution to software for the CPI, requiresapproximately 13-14 Megabytes ofhard disk space on an IBM-class personal computer, preferably one thatcontains a 386 or 386SX microprocessor.

TUTSIM Block-Diagram SimulationLanguage [5]

TUTSIM, an acronym for Twente University of Technology SIMulation program, is a high-level, block-diagram

language that basically functions asan analog-computer-on-a-PC. Engineers from the late 1950s through early 1970s may recall that an analogcomputer was usually programmedfrom a block diagram sketch and a listof equations representing the systemto be modeled; the analog computersolved the problem by substitutingfunctional, analog-electronic hardware circuits (based upon operationalamplifiers) for real functions.

TUTSIM re-establishes the analogy between real-world systems and a computer modeling language. Ninetythree defined blocks from SUM, GAIn,

MULtiply, DIVide, and INTegrate throughx,Y FuNCtion, Laplace functions, logicblocks, Z-transform blocks, and thermodynamic property blocks combinethe convenience of the analog computer with the accuracy and speed of personal computers, specially those thatemploy an arithmetic coprocessor chip(8087, 80287, 80387). A very valuableblock is the SYC (synchronization)block, which synchronizes the TUTSIMsimulation time base to the real-timeclock within the personal computer.During a simulation, the output of SYCcan be monitored to determine howwell the synchronization is proceeding.

I1allle: ,~~:m~·. '0·· 'Y Type: .. Motor single driue / null feedback

feedback-----

Type:Descr ipt ion:

Energ ized state:Open/Close cOfl\lIland:

Open lilllit switch:Close lilllit switch:

Open/Close alarlll tillle:

Hallie :.:...~'# .:.': ::\:.....,!!;~ Type: :.:.Description:

Energ ized state:Open/Close cOlllllland:

Open/Close alarlll tillle:

Hallie: Bi'...·..'''-,· ....: ...·.aI Type::'·'Description:

Start/Stop cOfl\lIland:Start/Stop alam tillle:

BLENDER I/O

Figure 3 Figure 4

16

grams will be illustrated in this article. The TUTSIM/OrCAD interface includes a complete block library of allindividual TUTSIM blocks, an OrCAD-toTUTSIM conversion program, and aTUTSIM-to-OrCAD conversion programfor paralneter changes and plotblockscaling.

To create a new TUTSIM block diagramfunction, a user completes a C or fORTRAN source-code skeleton that is thencompiled into a file, USRBLK.EXE, that

Tutsim Products markets special software, the ~tUser-Defiried Block Option," that provides almost unlimitedextensibility to the block-diagramfunctions of TUTSIM. This ~tOption" allows a user to create a new type of TUTSIM block-usAlUSR-that may add functions such as memory access to I/O

hardware boards or, alternatively,may act as windows and"linkages tocomplex code sections that are notlimited to simple math functions.With USA/u8R blocks, different displays, recordin'g devices, or I/O devicesmay be addressed and incorporatedinto the TUTSIM model simulation.

The wide variety of TUTSIM blocks allow a user to describe virtually anymechanical control, process control, ordifferential equation, continuous ordiscrete. Of considerable importance,non-linear differential equations aregenerally no more difficult to. modelthan linear differential equations.

OrCAD Computer-Aided-nesignSoftware [5]

OrCAD/SDT III is an electronic designautomation tool that is widely used forthe capture of electrical schematic designs. In its common application, OrCAD/SDT III comes with 6200 unique library parts in over 40 integratedcircuit libraries. Structured designfeatures within OrCAD/SDT III provide ahierarchical approach to organizingdesigns that contain up to 4000 sheetsthrough the strategy of partitioningthe design into smaller, more manageable sections. Over 200 hierarchicallevels simplify complex tasks. Withover 30,000 packages sold, OrCAD isthe defacto industry standard for easyto-use schematic capture.

TUTSIM Products markets TUT8IM/OrCAD interface software that permits the creation of hierarchical TUT81M block diagrams and simulationmodels for direct input into TUTSIM using OrCAD/SDT III. Several such dia-

III.Controller

TUTSIM

TUTSIM for the TI545

11I"'111"11'11I"II'",_ Block " ...AI ....."""' ....

executes in parallel with the mainTUTSIM executable file TUTEXEC.EXEon an IBM-class personal computer,preferably one equipped with a mathcoprocessor. In late spring and earlysummer 1990, one of us (KER) developed a special version of USERBLK.EXEthat could· directly communicate withthe memory locations within a TI545programmable logic controller.

generate a special versionUSERBLK.EXE? We wanted to merge asuccessful simulation package, TUT

SIM, with the TI545 in order tonate the digital switches and analogdials currently used in a $5000 TexasInstruments hardware simulatorthe TI 500 Series of controllers. We assumed that all of the hardware I/O wasdisconnected and that USERBLK.EXEwas not executing on a running process. USERBLK.EXE has not entered TI's

quality control environment yet, so itis still a preliminary software product,for simulation and evaluation purposes only.

Once USERBLK.EXE has received proper TI testing, it could conceivably beused as part of advanced controlschemes on an operating process thatrequires significant computationswhile the process is running. TI offersa special I/O module called the ExpertSolutions Processor (ESP) module,

Module Table

Figure 5 ure 6

17

which is an IBM-compatible personalcomputer board located directly on theTI Series 500 I/O backplane. The ESP I/O

module with the optional C run-timeI .. h,,,,,.,,,,,..' .. r has excellent communicationspeeds for a PLC, namely, 480 words(max) or 2048 discrete bits (max) perscan of the PLC, where a scan can be asfast as 20 to 100 milliseconds.

USRBLK.EXE for the TI 500 Series

In this version of USRBLK.EXE, a special Texas Instruments TUTSIM blockwas defined: USAfUSR. The USA version had two parameters and three inputs, and the USR version, three parameters and three inputs. The firstparameter, PARAMl, was an integercode that corresponded to the type ofmemory location selected within the TI

500 Series controller. The second parameter, PARAM2, was the integer codethat corresponded either to the readfrom-PLC or write-to-PLC operations.The third parameter, PARAM3, was theinitial value for the USR block. The TI500 Series PLC memory address typesspecified by PARAMl, the read/writefunctions specified by PARAM2, and thethree types of USA/uSR block inputsare summarized in table 1.

dVdf=FH:!.o+FA+FB-F (I)

dVT/df = ( FH20 + FA + FB) Ti - FT - VA (T-Tf)/p Cp (2)

The ability to develop this Texas Instruments version of USRBLK.EXE hadits origin with Chris Manhard in theTexas Instruments APT developmentgroup. Chris had a proprietary C testprogram, with an associated Clanguage library, that communicatedwith the TI 500 Series controllers. Theprogram permitted him to test controller communications through the use ofcommand line prompts. We strippedthe user input section from this C testprogram, added it to the USRCODE.Cskeleton provided by Tutsim Products,compiled the result into an *.OBJ file,and finally, linked the object file to TIproprietary *. OBJ files to produceUSERBLK.EXE.

When TUTSIM and the USRBLK.EXE version for the TI 500 Series were executed, two independently clockedcomputers-the TI 500 Series controllerand IBM-class personal computerneeded to be synchronized. Several independent clock rates were involved:(1) the communication clock rate (ashigh as 19200 baud) between the personal computer and the TI 500 Seriescontroller, (2) the TUTSIM iterationclock rate (typically longer than 0.2sec) on the personal computer, and (3)one or more clock rates that triggeredthe usAfusR-block exchange of infor-

mation with the TI 500 Series controller. To accomplish thistion, withUSAfUSR block was triggeredrising edge of an clockwhich wassimulation

In our ofUSRBLK.EXEsents a in theuse of real-time TUTSIM within indus-try and academia. in con-....... ",.......~-.r-, ".... ,., with is the firstcustomized version TUTSIM that ex-changes in real time,with a programmable logic controller.In the specific context of the TI Series500 hardware and APT software, realtime TUTSIM with our USRBLK.EXEdriver provide, for the first time, anopportunity for a control system designer to segregate the design processinto: (a) a control design strategy, applied to (b), an independently developed process simulation.

Alternatively, both the process simulation and the control design strategycould be contained within the compiled APT program that is downloadedto the TI 500 Series controller; however,such a simulation cannot be segregat-

\"hert'

(3)

1200 .

B-V is the blender volume, nr~

II; is the jacket volume, m3

T is the blender temperature, C

T; is the jacket temperature, C

T;, is the inlet jacket temperature, C

T, is the inlet temperature of water, solution A, and solution B, C

FH20 is the inlet water flow rate, m3Is

FA is the inlet aqueous solution A flow rate, m3Is

FB is the inlet aqueous solution B flow rate, m3Is

F is the outlet flow rate, m31s

FJ is the jacket flow rate, m3I s

pCp is the product of the density and heat capacity of the blender inlet and outlet flow streams,

Jlm 3 C

p) CP1 is the product of the density and heat capacity of the jacket flow stream, Jlm3 C

product of the blender tank heat-transfer coefficient and heat exchanger surface area,

the product of the jacket tank heat-transfer coefficient and heat exchanger surface area,

Figure 7

18

Figure 8

BLENDER1.SHT

Jul\<20.1990Sh""t1

6

A sketch of the blender Unit is givenin figure 1. Under the assumptionsthat the jacket and the reactor areboth perfectly mixed, the volume ofthe jacket and all physical propertiesare constant, the densities and heat

BLENDERl.SIM OrCAD/TUTSIM Simulation Program

CFCs, Units,DEMO-OIB program, wereto check for information-entry errors .The validation capability of APT, specially its ability to check for errors insmall segments of the overall design,is outstanding. Validation reports canbe saved or ·printed to facilitate theprocess of error checking. Once validated, DEMO_OlE was compiled andthen downloaded for execution on a TI500 Series PLC, in our case, the TI545.

one oneCFB, one CFB,

and CFBs. andMath CFBs contained APT Math pro

remaining CFBs consistedI""lI.t"'hT'1II"'l>r~.nI forms. 6 illus-

CFB icons used in theTEMP_CTL control CFC: aPID loop CFB, a split-range CFB, and a

CFB. We can zoom an lI'1V"'r~',r·,.n!l'I"

as has been shown for the PID

Use of 32-discreteinput hardware simulation board

Automatic run logic

Temperature controllogic

Tank level alarm

Device interlock protection

..... VJll.V''-'I\. .......II.''-' A"lI .."" .... """......."r......... table.rilo'llntf7l>il.:lIC were made glo-

__ 'V ........... A~....... ,... sepa-feed tanks.

LEV_ALRM

INTERLK

SIMULATE

Five Continuous Function Chartswere created (figure 2):

A total of nine Continuous FunctionBlocks (CFBs) were used in these CFCs:

entered into the Module defini-tion table, each I/O was auto-

entered the APT programinto appropriate global or unit I/O

symbolic name table. Most of the design work occurred at Unit level.Figure 1 depicts the device names forthe BLENDER figure 3, the blender I/O commands the depicted devices;figure 4, an excerpt from the BLENDERUnit Device definition table; and figure 5, a portion of the Module definition table. The information given infigure 5 is significant, because it de-

. fines and summarizes the controllermemory types and addresses neededby the TUTSIM simulation program.

. AUTO_RUN

TEMP_CTL

mulation andSystem

1II.o4i.::!.1I01l1"ll0 OrCAD/TUTSIM and APT

ed unless a second TI 500 Series control-ler is OltY1I"Y'>II'),1IT.t"'hrtl

In control system development sequence we describe in this article,the first step of the sequence consistedof using ProductivityTool (APT) software to develop theblender regulatory-control strategy inits simplest implementation: a singlePI loop plus an analog alarm. Space isnot available to teach the APT designprocess for BLENDER, but perhaps figures that depict selected APTgenerated reports will give a flavor forthis object-oriented control designsoftware.

At the APT Program Directory level,the regulatory control strategy wasgiven a program name, DEMO_OlE,and a description. At the ProgramContent level, BLENDER was defined asa Unit, one of what ultimately wouldbe several independent Units-for example, BLENDER, PACKAGE, andSTORAGE-for the process. Also at theProgram Content level, four global devices were defined in the Device defini tion ta b le-BL_H 20 _FVL V,

BL_A_FVL V, BL_B_FVLV, ANDPK_BLN_FVLV-and both discrete andanalog I/O points were named and de-

Figure 9 Figure 10

19

capacities of all inlet and outletstreams are approximately the same,and heat losses are absent, the modelequations are given in figure 7. Forfurther details, contact KER [3].

A similar process example, a continuous stirred-tank reactor simulation, isdiscussed in Chapter 9 of Smith andCorripio [6]; this well-documented example was used by these authors todemonstrate different approaches tothe computer simulation of dynamicprocess models. Our article adds yetanother approach (TUTSIM) to controlsprocess simulation, one that we believe is both simpler and more intu~

itive. Further, this approach nowlends itself readily to the developmentof testable control design softwarewhen a Tl 500 Series controller and APT

software are used concurrently.

Equations (1) through (3) in figure 7were simulated by a TUTSIM model,called BLENDER1.SIM, that contained62 blocks. To save space, the TUTSIM

listing of BLENDER1.SIM is not published here; it is contained in the supplementary package of informationavailable from KER [3]. An undergraduate student who has taken a onesemester analog controls course wouldbe capable of developing and testingsuch a TUTSIM model.

We used CAD software to create anddocument the TUTSIM block-diagramsimulation model for the blender. Or-

CAD/sDT III (see Part II) was applied tothe task of simulating the system, including the exchange of values oftwelve variables contained in the global and BLENDER I/O symbolic name tables, BL_LEVEL.RAW, BL_TEMP.RAW,

BL_CLD_BV_OC; BL_HOT_BV_OC,

BL_CLD_AVLV, BL_HOT_AVLV,

BL_PRD_DV_oc, BL_PRD_DP_OC,

PK_BLN_FV_OC, BL_H20_FV_OC,BL_A_FV_OC, and BL_B_FV_OC between the BLENDERl.SIM program andthe corresponding APT program,DEMO_OlE, running in the TI 500 Seriescontroller.

Figures 8 through 12 illustrate thehierarchical, OrCAD approach to the

. generation of the BLENDER1.8IM simulation model for the system in figures1 and 7. Figure 8 represents the highest level of the hierarchy, which includes schematic blocks for the nonsteady-state total mass balance (topblock, Eq.1); the non- steady-state energy balance for the blender tank(third block, Eq. 2); the non-steadystate energy balance for the jacket(fourth block, Eq. 3); and the calculation of blender flow information (second block). The ten USR blocks in figure 12, each associated with a block oranalog valve, can be understood withthe help of figure 5 and table 1.

The unusual feature that distinguishes figures 8 through 12 from atraditional process simulation in ahigh-level language is the USR block,

specially those in figure 12. Not onlyhave we simulated the transient massand energy balances, we have also incorporated the behavior of the valvesand pump in our model. The flows ofH20, A, B, cold water, or hot water willnot occur unless the respective valvereceives an open command from theTI545 controller. This may seem to be asmall accomplishment, but it becomesimportant when the batch process control strategy incorporates automaticsequencing through the use of Sequential Flow Charts (SFCs), which determine the timing of the valve behavior .All of the 110 points can be simulatedin the TUTSIM model, and the testing ofthe controller strategy can be relatively complete. Nine-hundred ninetynine (999) TUTSIM blocks are sufficientto simulate an entire midrange process.

A reader who is familiar with the TUT

81M block diagram language would observe two recognizable features of thelanguage in the OrCAD figures 9through 12: (1) all values of the blockparameters are provided below thelower-left-hand corner of each block,(2) a comment can be provided aboveeach block.

The use of OrCAD increases the timeassociated with the creation of a TUT

81M model, but offers the academic andindustrial advantages of more effective process model documentation,communication, and maintenance. In

~TR

~OF

Figure 11 Figure 12

20

an academic environment, the cost ofOrCAD/SDT III is prohibitive for an individual student; unfortunately, an attractively priced,. restricted, studentversion tailored specifically to the TUT81M block diagram language does notexist.

v. Sample Results

In the control system development sequence that we describe in this article,the blender process (figure 7) was simulated in TUT8IM. In the TUTSIM simulation program, called BLENDERl.8IM,uSR-block tthooks" were provided tothe memory locations of the TI 500 Series PLC that executed the desired control strategy.

The next step in the development sequence consisted of using the Application Productivity Tool (APT) softwareto develop a regulatory control strategy in its simplest implementation: asingle PID loop (for the tank temperature), a single analog alarm (for thetank level), and a few interlocks. It isimportant to reiterate that the designof this simple regulatory control strategy was done at a high level, withoutthe use of relay ladder logic (RLL), atbest a difficult and inconvenient lowlevel language for chemical engineers.APT program DEMO_OlE was compiledinto TI545 object code that was immedi-

ately downloaded into the controllerwithout conversion into RLL.Relayladder-logic diagrams could have beenreverse generated (using Texas Instruments software, TISOFT) for use byplant technicians, but a discussion ofthe process for doing so is beyond thescope of this article.

For this article, two tests of the simultaneous execution of the simulation(TUTSIM) and controller (APT and TI545)codes were performed. The first was anisothermal test of the functioning ofthe blender inlet and outlet flowvalves. The second was a closed-looptest of the heating and cooling of theblender tank filled to a height of 1 meter.

At this point, at least three programsDEMO_OlE, APT, and TUTEXEC.EXE/BLENDERl.SIM-were ready for simultaneous execution. The first was locatedwithin the TI545 controller; the secondand third were located in an IBM PC/ATclass personal computer. As multitasking was not possible, first APT andthen TUTEXEC.EXEIBLENDERl.SIM wereexecuted. To start the execution ofDEMO_OlE within the TI545 controller,we entered the APT ttdebug" mode ofoperation, where we were, able tott run" the DEMO_OlE program. Afterinitializing variables (e.g., the setpoint, proportional gain, integral

time, and open/close or run/stop statesof the valves and pump), we left APT,entered TUTSIM, and monitored thetank level or temperature as a function of time. To perform a differentrun, we needed to exit TUTSIM, reenter APT" and repeat the process butwith different values for selected variables. It became a nuisance to constantly switch between APT and TUT81M for each different simulation/controls run. This nuisance waseliminated by using a second communications port on the controller.

Second Communications Port onthe TI545: Use of MultiplePersonalcolml~u'ter·s

The TI545 controller board (Model No.545-1101) contains two,asynchronous-serial communicationports, an RS-232C port and an RS-422port. A Metrabyte COM-422 board [7],purchased for $190, permitted two IBM

PC/AT-class personal computers to simultaneously communicate withcontroller memory. In the absence ofDOS multitasking, this is a preferredform of testing. One IBM PC/AT machine executed TUTSIM, while the other executed APT. APT ttdebug" operation permitted us to change controllerand valve variables, while TUTSIM execution permitted us to observe the effects of such changes on tank tempera-

BL_CLD_AVLVBL_CLD_BlJ_OCBL HOT AVLIJBL=Hor)v_oc

BL H20 FV OCBL=A_FV_OCBL_B_FIJ_OCPI( BU1 FV OCBL=PRD=DV=OCBL_PRD_DP_RC

BL_TEMPBL_LEVEL

BL TC .SPBL)C.)(CBL_TC. TIBL_TC. TD

BLENDERBLENDERBLENDERBLENDER

BUttDERBLENDER

BLENDERBLENDER

BLENDERBLENDERBLENDERBLENDER

¥4: Level18,8888

a 9,8888

TRUE a,888832aae 5TRUE

7,8888TRUE

FALSE 6,8880FALSE

4TRUE 5,8888TRUETRUE 4,8888

7a.394533,88883.6aaaa

375.aaaaa :2 ,88882.6aaaa1.aaaea 1,8888a.laaaa 2

8,88888.8888 'l'iUY&e

3 feed inputs" one output strearra

6 7

8

9

Figure 13

21

Figure 14

Figure 16 is a portion of the APT

thatprovides the specific addresses forsome of the dot-extension variablesshown in figure 15. BL_TC.SP is realmemory address Vl., a 32-bit floatingpoint quantity. PLC addresses VI., V5.,

and V7. were used as the basis for several additional TUTSIM blocks, shownin figure 17, that permitted both reading and writing the setpoint, proportional gain, and integral time directlyfrom our TUTSIM program, BLEND-

the.OUT, and .Tl dot extension variables were all variableswere ignored.

ParametersAPT or TUTSIM

mewas

A...., ... '..... J1. ... J.~ a steady10 me

and then (8)

Figure 15 is a screen capture that depicts the complete story of the behavior of PID control block, BL_TC,

which is an APT object. Associatedwith this object are both operationsand, as shown in the figure, a predefined set of discrete, integer, and realvariables. Each variable is represented by a ~~dot extension," for example,

wereand thewas closed to leave a static tank levelof 3.6 meters. This suc

valve seFunction

Six (Procedural

shown in 14.level was at 1 meter. The (2)

valve test was designed to quicklydemonstrate that simulationmodel, in terms of valves and flowrates, was correct. The manipulationof six input switc.hes produced the run

Use I/O

lator Boards

The test of the 101enaE~r

inlet and pump wasmost COllVE~nlently nal¥"1"A1¥"rlI"'lad"il with theaid of a 32DI simulatorYY'l\Jn..nIll1lrl1 , .... y ..... 'U" ...... "-I ... No. which

n""11111111"'l.V'I..f"'\,.rII with a set of seventeenswitches. Within the SIMULATE Continuous Function each of eight

was programmed in anactive MATH CFB to provide the open orrun commands for seven valves andone pump. Figure 13, an APT Monitor

summarizes the key variablesthat were monitored during a simulation run. To open the hot-water blockvalve, SWITCH34 was made TRUE,

which made BL HOT BV OC alsoTRUE.

1J66 INTEGERU69 UITEGER

V1. REALLTS1. REALC1a? BOOLEAN ARRAY OF 16lJ9. REAL1J7. REAL

063. REALVi. REALLUF1 INTEGERU21. REALU71. REAL

1J76. REALlJ78 BUEGERC769 RET BOOLEANC769 RET BOOLEA"C2e6 BOOLEAN

BL_IC •SMODEIlL_TC.SNUM

IL_IC.SPIL_IC.SIBL_IC .SIATUSBL_TC. IDBL_TC. TI

IL_IC.IMPIL_TC. TSPBL_IC "JFLAGBL_TC. YDABL_TC_SR .BR](P1

BL_IC_SR •BRXP2BL_IC_Sft. ICODEBL_IC_Sft .E"ABLBL_Te_Sft •mABL1BL_IC_Sft .EMBLD

SP PVHItl PUL

OUIlIN Forced

BIASERR

HHAHC HAII LAID LLA

AWS YDAS1 ODA

RCA

INYDAINODAUtRCAINBTA

OIJRUf1NRDY

H1AJ1 _

IAUIOICASCSMODE

PID Loop Block :BL_IC Blender tel'perature control

U1HHAU1HAIt4LA

nlLLA

16/Jul/1999 21 :58 :21 PAGE 16APT SYMBOL "AME TO PLC ADDRESS

PROGRAM "AME : DEMOJ)lC

Figure 15 Figure 16

22

with discussion onprocess

to understandcontrol."

Unit r>.-rt"I..",,_.n,'i!- ... r.~!"" controlstions

Batch sequence

Process data base manage-ment

topics which are consistent withuse TUTSIM, the TI545 in

iUl.'VI.'Vil.iUl.,J ............lIAA .... A .. 11'.... ' sequence.

Jim Doss (Tennessee Eastman), in thesame round-table discussion, stated[8]:

as PLCs

and their use in interlocks are

...... "',... I!t"\"''''''. concern is

sometimes even w<>""'''''''"'llr'''1l1l''ll1f'T

samethat isit. They are almost totally unawarethat the most control sys-tem in is not thecontrol system, but safety systern..... What I would like to see is arevision of the entire curriculum tomake it more model-based. Studentsshould connect basic ideanamic simulation to the solution of differential equations early on. Then, inthe various courses, the appropriatemodels would be developed as a pedagogical too1..... I have found in teaching process control that modeling iswhat the students are weakest in."

a rU~:11rne]ntclry

tor station using the Texasments Micro-TISTAR software.

The development and use of theUSRBLK.EXE driver for the TI500 Seriesof programmable logic controllers provides a bridge in academia betweenthe traditional undergraduate lecture/lab sequence on process controland industrial practice based upon increasingly powerful and cost-effectivePLCs. We believe that APT, TUTSIM andthe TI545 are, as a group, responsive tothe following comments made by IrvRinard (CUNY) in a recent round tablediscussion on process control education in the year 2000

¥4: TEMPa8,8888

Xc = 2.6, Ti =1, SP=15,35,55,75

72,S888

3S,8888

42,S888

27,S888

S7,S888

12,S808

Tit 1.~E~~~~:~ e~~t:h::::a:q:::tnurt ngBL_TC Controller PArameters

:. ze ocumen Nu ~r

A BL_CONT.SHT 1

Jul 201'3'30 h .... t 6

8.8888 Tillie 298.0888

Figure 17 ure 18

23

Figure 18. Closed-loop testing of theoperation of the blender cold- and hotwater streams. The proportionalwas 2.6, the integral time was 1.and the four setpoints were 1535°C, 55 °C, and 75 °e.

APT PID loop chart, obtained during ~~debug"operation.

Figure 16. A portion of the APTSymbol-Name-to-PLC-Address tableprovides information on the control~er

memory types and addresses. Such Information is needed for the TUTSIM USRblocks such as those in figure 17. Additio~al TUTSIM blocks provideread/write capability for the PLC variables BL_TC.SP, BL TC.KC, andBL_TC.TI.

Figure 17. The TUTSIM/?rCAD di.agram that provides read/wrIte capabIlity for BL_TC controller parameterssuch as the setpoint, proportionalgain, and integral time.

JU·"lIl"lI"... lII't\(.. .n. 14.. Testing of the operation

the blender tank inlet and outlet flowvalves and pump. This quick and simple test confirms that the valve-andpump simulation is correct.

13. APT Monitor ob-tained during ~~debug"operation.

Figure 12. The TUTSIM/OrCAD diagram that represents the blenderstreams.

Figure 11. The TUTSIM/OrCAD diagram that represents the oon- steadystate equation for the jacket l..I'Co.l!.,II..l!.IJ'Il... '"

ture.

state equation for the blender temperature.

................................. 5.. A portion of the Module

definition table for the DEMO_OlE APTprogram.

Figure 6. The three CFB icons associated with the TEMP_CTL ContinuousFunction Chart.

Figure 4. A portion of the Device ~efi

nition table for the BLENDER unIt offigure 1.

Figure 7. The three non-steady stateequations, along with variable and p~rameter definitions, for the blender Infigure 1.

......... ,...... ",.,.. '" 8. The highest level of the Or

CAD hierarchy for the design of theTUTSIM blender simulation: overallschematic for the BLENDERl.SIM blender simulation.

... ·lItrlflll1l1ll'&O 3. The BLENDER unit of Figure

1 with blender I/O signal names specified.

......... ",.................."'" and Table tJa,ptJlOl1lS

1. The BLENDER unit in thethree-component, blender/packagingprocess. Device names are specified.

lIWI'lItl"lf'lIl11'1i'&O 2. The three-level APT hierar

chy consisting of the Program Directory (highest level), Program Co~

tent Level (middle level), and UnItContent Level (lowest level).

8. Tom F. Edgar, ((Process Control Educationin the Year 2000: A Round Table Discussion," Chemical Engineering Education 7277 (Spring 1990).

7. KeithleylMetrabyte/AsysUDAC,440 MylesStandish Blvd., Taunton, MA 02780, (508)880-3000.

1. Kay E. Kuenker and Gary E. Blau, ((SimuSolv*: A Computer Program for BuildingMathematical Models," CAST Communications 12 (1), 10-21 (March 1989),

2. John A. Arnold, ((Process Control SoftwareDesign," presentation to Texas A & M, January 16, 1990. Contact John at Texas Instruments Incorporated, Industrial Systems Division, Erwin Highway, P.O. Drawer 1255, Johnson City, TN 37605-1255,(800) 284-9084, extension 2078. John received his PhD in Chemical Engineering Department at the University ofTennessee.

3. For additional information, contact KarlRony, APT Development Group, Texas Instruments Incorporated, Industrial Systems Division, Erwin Highway, P.O. Drawer 1255, Johnson City, TN 37605-1255,(800) 284-9084, extension 2078.

4. APT Programming Reference Manual, Texas Instruments Incorporated, IndustrialSystems Division, Johnson City, Tennessee, Apt-8102, Second Edition (March1990).

References

Peter R. Rony received his B.S. ChE in 1960 fromCaltech and his PhD from the University of California at Berkeley in 1965. During 1980

through 1985, he served as associate editor-inchief and then editor-in-chief of the IEEE Computer Society magazine, IEEE MICRO. He currently is editor of CAST Communications.

Karl E. Rony received his B.S.ChE in 1987 fromVirginia Polytechnic Institute and State University. In February 1988, he joined the TexasInstruments Industrial Systems Division, Johnson City, Tennessee, as a Control Systems Specialist. A member of the APT DevelopmentGroup, he is proficient with APT, T1STAR, and TUTSlM in the design, simulation, and testing of PLCbased control designs.

5. For further details, contact TUTSIM Products, 200 California Avenue, #212, PaloAlto, CA 94306, (415) 325-4800.

6. Carlos A. Smith and Armando B. Corripio,(tprinciples and Practice of Automatic Process Control," Wiley, 1985, pp. 471- 491.

Figure 9. The TUTSIM/OrCAD diagramthat represents the noo- steady-stateequation for the blender volume.

Figure 10. The TUTSIM/orCAD diagram that represents the 000- steady-

Table 1. Documentation forUSRBLK.EXE: TI 500 Series Controllermemory address types, read/writefunctions, and types of inputs for theUSA /USR blocks.

24

P ARAMl: Types of address

1 C read/write (Boolean variable)

2 X read/write (Boolean variable)

3 Y read (Boolean variable)

4 V read/write (16-bit integer)

5 V. read/write (32-bit floatingpoint)

6 WX read/write (16-bit integer)

7 WY read (16-bit integer)

8 C read/force write (Boolean variable)

9 X read/force write (Boolean variable)

10 WX read/force write (l6-bit integer)

P ARAM2: Read/write functions

1 Read value from TI 545 PLC locationIND_INPUT[ l] and stuff inUSR_OUTPUT

2 Write value IND_INPUT[2] to TI 545

PLC location IND_INPUT[ 1](USR_OUTPUT = 0)

P ARAM3: Initial value

The USR block can accept an initialvalue.

Inputs:

Must be in proper sequence, like theDIV block.

Inputl IND_INPUT[l] Integer valueof address type specified inPARAMI

Input2 IND_INPUT[2] Value of TUTSIM variable to be written toPLC

Input3 IND_INPUT[3] USRlUSA blocktrigger input; the USR blocktriggers on the positive edgeof a logical clock pulse; theUSA block triggers on a logic1 level.

Note: The difference between the USAand USR blocks is important. USR is aTUTSIM history block that requires aninitial condition, which is the first value that is output by the block. USA is aTUTSIM algebraic block, which is not ahistory block. History blocks executein parallel, while non-history blocksexecute in the order of the block diagram.

CODllIlunications

Listening In

by Peter R. Rony

John Hassler (University of Maine),who monitors USENET via BITNET, sentthe following information on July 11,1990, excerpts of which we are publishing as an example of the vitality ofelectronic communications in the fieldastronomy. Some of this informationwas subsequently published in theJuly 20, 1990 issue of Science magazine r~Hubble: It Could Have BeenWorse," page 242).

Like John, I admire the immediacy associated with the astronomy community. Do engineers tend to be more isolated in their activities and less likelyto participate in electronic conferencing and bulletin boards? I doubt it.But neither do I have a good, recentexample of this type of immediacy inchemical engineering.

25

HHST StatusNumber 60

091990

Current Status and Summary

~~This past weekend has been ratherquiet at HST. Since my last report wefinished Bootstrap Phase B part 5f andstarted part 5g. The biggest event ofthe past few days occurred yesterdayevening and today: a successful seriesof Foe exposures. Stars were seenwith good SIN and excellent overalldata quality. Very preliminary analysis suggests that significant coma wasNOT seen in the images, suggestingthat the primary mirror is the sourceof the spherical aberration (see below).Another piece of good news regardingthe characterization of the HST wavefront: modifications to the proceduresfor how we have been using wavefrontsensors (WFS) 2 and 3 have producedthe first good data from these sensorssince launch. The only piece of badnews was that some WFPC images werelost do to a series of operations eventsthat could have been avoided. Thisand the safing event of last Thursdayhas people taking a long hard look atreal time operations procedures."

/

pes

~~The only news on the PCS front regards the day/night terminator instability fix. A meeting has just concluded discussing current expectations forthe new control law. Optimism is stillhigh that they can meet pre-launchspecs, but I have noted that they havebacked off a bit on how much they canreduce the disturbance amplitudefrom an informal quote about 2 weeksago of a 10 to 100 reduction to a current guess of a 5 to 60 reduction factor.The current worst case disturbance isapproximately 110 milliarcsecondsRMS (2 milliarcsec RMS is the spec value). As of the moment the repair is on

The following items summarize information in the hands of the Editor byJune 30, 1990. The preferred deadlinesfor the two issues of CAST

Communications-called the Summerand Winter issues-will be approximately May 15 and December 15 sever.:.al weeks after the Spring and FallAIChE meetings, respectively. These revised deadlines will give CAST divisionmembers who are active in CAST programming activities sufficient timeafter AIChE meetings to send lastminute information to both thecations Board Chairman (Jeffand the Editor of the newsletterRony), We prefer that all communications with us be done in electronicform, either with MS DOS formatted diskettes or with messages sent electronically over BITNET. An up-to-date list-

Meetings and Conferences

facul ty membersPC instructional which

is of interest to other academics potential use in their own classes. To facilitate an information process, CAST will arrange for two or morePC-AT or 386 class machinesdisk drives and VGA graphics toused by academics who wish to demonstrate their software. Details of the location and times will be posted on theA.I.Ch.E. notice boards at the AnnualMeeting. In the meantime, will interested persons please contact ProfessorAlan Foss, Department of ChemicalEngineering, University of California,Berkeley, CA, 94720-9989, (415) 6424526. Fax: (415) 642-4778.

Alan Foss, UniversityBerkeley

Plans are being made for the first PDXI

meeting to take place at the ChicagoAIChE meeting in November. ContactPDXI at the address and phone numberon the brochure to pursue potential involvement.

Finally, there are similar activities inorganizations such as IChemE, GPA,

NIST, and ASTM. Although none of theother activities address the particularissue being addressed by PDXI, the PDXI

effort needs to proceed to ensure thatour interests are adequately catered.

PDXI focusses on a real and readily.A."-4 ............ v .......... \UI,p..JI .......... problem that is solvable!

needs to be wide parthe resulting solution

accepted. The memberfee is reasonable at $5,000 per

year per company, and the solutionbe forthcoming in two years if

enough join.

From a technical viewpoint, there isalso reason to be involved because thetechnology of data exchange does impact us but receives limited attentionby our profession. The organizing committee delved into the technologyenough to set funding requirementsand expected timeframe for the solution. During that experience, it wasfound that the technology of data exchange is progressing, that it continuesto have an increasing impact on ourdaily activities, and that there is achallenge to make good use of the technology.

By-of organization have been as-

sembled and accepted. With this issuethe Newsletter, a copy of a brochure

has been included summarizing then1l""r\nnl~.a.nl activity.

If'I.n.llll'll.11'U'''lIf'''\IT to GSFC in late

OCIUSJ..llif and .............. \,J'U2Hll,V

RonPolidan

Process Data eXchangeInstitute (PDXI)

by John T. Baldwin

The organizing of PDXI, which is supported by CAST, continues on schedule.The AIChE Council agreed to the formation of PDXI subject to sufficient mem-

HST

news isimages: 4 last night and 4

this The imagest'hl....n.l1INh a narrow (30 A) H-beta

a 12-13 magnitude star neareps has just begun, sothere are no quantitative results. Theimages have good SIN should yieldexcellent data. The images are quitesimilar to those taken with the pc: a

with a big (approximately 1.5-2arc lots of rings,rays, and the such. A quick look at thedata seems to show a stellar imagethat is similar gross detail) to thePC images with no obvious asymmetry(Le. coma). If this is confirmed withquantitative analysis, it would suggest that the primary mirror is responsible for the spherical aberration.There is also a rumor circulating thatsuggests a possible origin for thespherical aberration: it has been reported that if the null lens used to testthe primary were installed backwardsthat you would get a degree of spherical aberration similar to what is seenin HST. Obviously, more on this later.The only mirror movements have beento move the secondary to the ~~zero reference" position-the favored WFPC focus position. More mirror movementswill be occurring this week now thatwe have FOC images..... "

26

Quality Circles

Management ResponsibilityQuality Problems and Pitfalls

Course fee is $250 per person. Registration Deadline: September 10, 1990 (ifyour interest develops after this date,contact AIChE anyway... Ed.) For information, contact AIChE Symposium, POBox 511, Kingsport, Tennessee 37662,Attention Joe Parker.

Motivation for Quality

Tools of Statistical ....."" ....... ILlI..A.>-'IUW

(SQC)

Control Charts for Variables

Control Charts for Attributes

• Special Process Industry Applications

designed for engineers and managersin all aspects of process industry opera-

including technical and COAr'hllf""ilY"Ild'Il.n.

departments as well as productionoperations. Lectures and videoare utilized to present concepts and statistical methods. These presentationsare then followed by exercisesing small group discussionmerical calculations.Application of Pinch Technology in a

wide range industries has yieldedoutstanding results. Engineers reportenergy cost savings of 15% to 30%,capital cost flexibility and operability, and increasedplant capacity. To there havebeen hundreds of successful studies.

answers the question of how much aprocess can be improved. The techniqtle is Technology, which wasdeveloped by Professor B. Linnhoff.Pinch Technology provides a clear picture of energy flow in a process, andidentifies the most constrained of

process: the process

This is the standard introductorycourse in Pinch Technology thatLinnhoff-March offers most frequently.It features a combination of lecturesand working sessions, and covers energy targeting, heat exchange networkdesign, data extraction, process modifications, and other related topics. Theinstructors for the course are Mr. M. A.Rutkowski, Mr. J. D. Kumana, Mr. A.P. Rossiter, and Dr. Ravi Nath. For information, please call Linnhoff-March,Inc., 2 Cardinal Park Drive, Suite205A, Leesburg, VA 22075, (703) 777-.4145.

ing of proposed sessions and Calls forPapers will maintained electronically the Publications Board Chairman. CAST Division members can always request such information by1V'V"'J'''-4..1l.A.II.~ a BITNET message to RONY at

Authors are reminded that under current AIChE meeting policy, the meetingbooklet will contain only titles of thepapers presented. However, a book ofextended abstracts is distributed to attendees at the meeting. Moreover, authors may bring hard copies of theirpapers for distribution at their session, and hard copies or microfichemay be ordered at or after the meeting.

... Peter·R. Rony, Editor, CAST Communications

VTVM1.

Please send CAST Division session information, Calls for Papers, and meeting and short course announcementsto me by December 15, 1990 for inclusion in the ~tWinter1991" issue of CAST

Communications. For those membersof the CAST Division who are engagedin the presentation of seminars andshort courses, an advance (draft) copyof your announcement brochure wouldbe appreciated.

Process Integration UsingPinch Technology

Linnhoff-March, Inc.Chicago

October 2-5, 1990

HoustonOctober 30-November 3, 1990

Every engineer knows that there is often scope for improvement in processdesigns. But what about a particulardesign? Can it be improved? By howmuch? Until recently, there was nosatisfactory answer to these questions.But now, for the first time, there is anentirely fundamental approach that

Statistical Process QualityControl (AIChE Course),Tennessee Eastman Co.,

Kingsport, TennesseeOctober 9-10, 1990

Quality improvement in industry invol ves a management philosophygeared both toward excellence in ~very

detail of operation and a set of statistical tools necessary to achieve excellence through improved quality and increased productivity.

This two-day seminar, presented byDr. Victor Zaloom, Professor and Headof Industrial and Mechanical Engineering at Lamar University, has been

Chicago AIChE MeetingNovember 11-16, 1990

The Palmer House. Meeting ·ProgramChairman: Dr. Charles A. Wentz, Argonne National Laboratory, 9700South Cass Avenue, Argonne, IL60439, (708) 972-7693.

For further information details concerning CAST Division sessions andscheduling, contact Jeffrey J. Siirola(Area Programming Chairman), Research Laboratories - 895, EastmanChemical Company, P.O. Box 1972,Kingsport, TN 37662, (615) 229-3069.

27

4. To allow the industrial practitioner a better understanding of thenew tools available from the researchcommunity in order to stimulate widerimplementation where these are llse-

All conference sessions will be heldduring mornings and evenings, leaving the afternoons for informal discussions and recreational activities. Toprovide an atmosphere for active participation and substantial discussion,

Based on the experience from past CPC

meetings, we expect to have a group ofparticipants that is about 70% from industry and 300/0 from universities.

we should have the right mix toachieve the stated goals.

5. To stimulate the research community to adjust research directionsthrough (a) a better understanding ofperceived needs and challenges fromthe process industry, and (b) a tutorialintroduction to intellectual concepts inother areas of process control research.

1. To gain an appreciation thestate of process control practice ineach of the process industries, and toobtain from industry a list of futureneeds and challenges for the researchcommunity.

2. To present an overview thestate of the art in each of the important areas of process controland practice, presented in a way thatis understandable to the non-specialistand industrial practitioner.

3. To provide a forum for in-depthdiscussions between university researchers and industrial practitionerson the practical challenges in this fieldand the extent to which current research directions are satisfying perceived needs.

and used to define the practical needsand intellectual challenges of nextdecade. Thus, the theme of theence will be ~~Future Needs and Challenges in Process Control." The specific goals of CPC-IV are as follows:

IV(CPC-IV)

Island, TXFebruary17-221991

About the CPC-IV Conference:

The first Chemical Process ControlConference (CPC-I) was held in Asilomar, California in 1976, the second(CPC-II) in Sea Island, Georgia in1981, and the third (CPC-III) again atAsilomar. These conferences weresponsored by the AIChE and NSF, andthey brought together participantsfrom industry, government, and universities to discuss and critique thecurrent state of process control. Some15 foreign countries were representedby participants in the first three Conferences. These meetings have cometo be regarded as milestones in theprocess control field. They have not

served to highlight recent developments, but they have had a decisiveimpact on research efforts in the succeeding years. They have helped tonarrow the gap between process control theory and application.

As a continuation of the traditions ofthe first three conferences, CPC-IV willfocus on advances that have takenplace recently in the process controlfield. These will be put in perspective

1. Mathematical Analysis of ComplexSystems.

2. Applied Mathematics and N umerical Analysis.

3. Chaos in Deterministic Systemsand Applications in Chemical Engineering.

4. Novel Applications of Mathematicsin Chemical Engineering.

5. Recent Developments in Numerical Methods for ODE/OAE/PDE

Systems.

Systems and Process De-Area

Area lOB and Area 15CSession

1. Modeling and Control of Biochemical Processes.

Area tOe: ComputerstLAl.l.I.IIl."U1.ILll"~ and Information rr,oCles~,lnfl

1-2. Advances in Optimization I and II.

3-4. Parallel Computing I and II.

5. Visualization of Chemical Engineering Systems.

6-7. Application of Neural Networksin Process Engineering.

Area lOB: Systems and Process

Area IOD: Applied Mathematicsand Numerical Analysis

1-3. Recent Advances in Process Con- .trol I, II, and III.

4. Nonlinear Control.

5. Model Predictive Control.

6. Artificial Intelligence/Neural Networks in Process Control.

7. Process Control Education in the1990's.

8. Industrial Challenge Problems inProcess Control.

9. New Concepts in Dynamic Simulators.

1-2. Process Synthesis I and II.

3-5. Design and Analysis I, and III.

6. Modeling and Simulation for Process Design.

7. Batch Process Engineering.

8. Design for Process Innovation.

The CAST Division is planning the following sessions at the Chicago Annualmeeting.

28

the number of attendees will be limited, with selection made on the basis ofinvolvement in the field of process control.

Panel Discussion

Monday, February 18 (p.m..)

eeBeyond the Design of Predictive Controllers," by C. Garcia and B. Ramaker(Shell)

Panel Discussion

Panel Discussion

Thursday, February 21 (a.m.)

20 (a..m.)

Learning Systems, Adaptive and AI

Control (Session Chairman, T. J. McAvoy, Maryland)

eeModel Preditive Control - The Good,The Bad, and the Ugly," by M. Morari(Caltech)

e~The Industrial Importance of Nonlinear Control," by E. Longwell (DuPont) and R. MacFarlane (Amoco)

~eThe Differential Geometry Approachto Nonlinear Control," by C. Kravaris(Michigan), J. Kantor (Notre Dame),and Y. Arkun (Georgia Tech)

~eThe t:Direct' Approach to NonlinearControl," by P. Lee (Queensland) andC. Georgakis (Lehigh)

eeThe Optimal Control Approach toNonlinear Control," by J. Rawlings(Texas) and L. Biegler (CMU)

Wednesday, February 20 (p.m.)

Recent Approaches for the Control ofNonlinear Processes (Session Chairman, C. Georgakis, Lehigh)

~eAdaptive Generalized PredictiveControl," by D. W. Clarke (Oxford)

Wednesday,

Panel Discussion

e(The Robustness of Model PredictiveControllers," by E. Zafiriou (Maryland)

Model-Based Process Monitoring andControl II (Session Chairman, N. L.Ricker, Washington)

Panel Discussion

~(Dynamic Process Simulation-RecentProgress and Future 'Challenges," byW. Marquardt (Stuttgart)

uAn Industrial Perspective on Dynamic Flowsheet Simulation," by E. Vogel(Tennessee Eastman)

~(Qualitative Process Modelling," byK. Hangos (Hungarian Acad. Sci.)

uAdvances in Process Identification,"by H. W. Andersen, K. H. Rasmussen,and S. B. Jorgensen (T. U. Denmark)

Model-Based Process Monitoring andControl I (Session Chairman, D. G.Fisher, Alberta)

Panel Discussion

Tuesday, February 19 (p.m.)

~eModel Predictive Control - State ofthe Art," by N. L. Ricker (Washington)

eeExperience with DMC Inverse forModel Identification," by C. Cutler(DMC Inc.)

ee MPC Applied to an FCC Process," by J.M. Caldwell and J. G. Dearwater (Profimatics)

Tuesday, February 19 (a.m.)

e~Multivariate Statistics in ProcessControl," by J. MacGregor (McMaster)

e(Application of Neural Nets to Sensorsand Data Analysis," by M. Piovoso andA. Owen (DuPont)

On-Line Sensors and Data Analysis(Session Chairman, D. E. Seborg, Santa Barbara)

.Dynamic Process Simulation, Modelling, and Identification (SessionChairman, S. B. Jorgensen, T. U. Denmark)

Accommodations and Meals:

CPC-IV Program:

Monday, February 18 (a.m.)

Present Status and Future Needs The View From Industry (SessionChairman, W. D. Smith, Jr., DuPont)

~~The View from Japanese Industry,"by I. Hasimoto (Kyoto) and S. Yamamoto (Yokogawa Elec.)

eeThe View from European Industry,"by H. Shuler (BASF), F. Allgower, andE. D. Gilles (Stuttgart)

~eThe View from North American Industry," by J. Doss and J. Downs (Tennessee Eastman)

Location of the Conference:

Arrangements have been made forspecial room rates beginning at $60per night for a room with two beds.There is a conference meal package($282), which includes 3 ,meals a dayand all social functions.

The conference will be held at theSheraton South Padre Island BeachResort, (800) 222-4010, on South Padre Island, Texas. Situated on thesouth end of Padre Island, near theMexican border, the location offerssunshine, beach, swimming, fishing,sailing, golf, and tennis. The hotel isapproximately 40 minutes by car, limousine, or taxi from the Harlingen airport, which has frequent flights fromHouston and Dallas.

Keynote Talk - ~~The Best of the Bestin Advanced Process Control," by J. F.Swallow (A. T. Kearney, Inc.)

Sunday, February 17

29

Area lOB: Systems and ProcessControl

1. Intelligent Control. Ali Cinar(Chairman), Department of ChemicalEngineering, Illinois Institute of Technology, Chicago, IL 60616, (312) 5673042 and William D. McGraw, Aluminum Company of America, Alcoa Center, PA 15069, (412) 337-2755.

2. Industrial Applications ofProcess Control. James B. Riggs(Chairman), Department of ChemicalEngineering, Texas Technical University, Lubbock, TX 79409, (806) 7421765 and Gerardo Mijares (ViceChairman), M. W. Kellogg Company,

Joint Area lOA and Area 16B Session

1. Computer-Aided Engineering/ Computer-Aided Design (A Tutorial). C. K. Breaux (Chairman), ProsysTech, Inc., 7007 Gulf Freeway,Suite 133, Houston, TX 77087, (713)641-0940.

(713) 493-8866 and David R. Heltne(Vice Chairman), ShellCompany, PO Box 1380, .l.lUUi:)lJUIl ..

77251, (713) 493-7325.

3. Process andtion. Michael F. Malone (Chairman),Department of Chemical Engineering,University of Massachusetts, Amherst, MA 01003-0011, (413) 545-0838and Peter T. Cummings (Vice Chairman), Department of Chemical Engineering, University of Virginia, Charlottesville, VA 22903-2442, (804) 9784726.

4. Retrofit Designand Applications. Don R. Vredeveld(Chairman), Union Carbide Corporation, PO Box 8361, South Charleston,WV 25303, (304) 747-4829 and Michael F. Malone (Vice Chairman), Department of Chemical Engineering,University of Massachusetts, Amherst, MA 01003-0011, (413) 5450838.

Mailing Address:

Houston AIChE MeetingApril 7-11,. 1991

1. Applications of Artificial Intelligence in Process and ProductDesign. Bahu Joseph (Chairman),Department of Chemical Engineering,Washington University, St. Louis, MO63130, (314) 889-6076 and Krishna R.Kaushik (Vice Chairman), Shell OilCompany, PO Box 6249, Carson, CA90749, (213) 816-2276.

2. Industrial Applications of Optimization. Emilio J. Nunez (Chairman), Shell Development Company,PO Box 1380, I-Iouston, TX 77251,

Name:

Main Area of Research Interest:

Area lOA: Systems and ProcessDesign

Affiliation:

Conference Sessions of Main Interest:

The CAST Division is planning the following sessions at the Houston National meeting. The overall theme forthe meeting will be ~tTools for Chemical Engineers in Plant Operations andMaintenance." Deadlines and finalcall for papers for this meeting appearlater in this issue.

CACHE CorporationP. O. Box 7939Austin, TX 78713-7939

Telephone:

Meeting Program Chairman: Dr.John G. Ekerdt, Department ofChemical Engineering, University ofTexas, Austin, TX 78712, (512) 4714689, (512) 471-7060 (FAX).

Title:

Panel Discussion

Panel Discussion: Control Room of theFuture

Friday, February 22 (a.m.)

~tFuzzy Logic and Its Applications," byL. Zadeh (Berkeley)

~tReal-Time Estimation and ControlArchitectures - The Future," by J.Baras (Maryland)

t~Microelectronics Factory of the Future," by G. Larrabee (Texas Instruments)

Control Technology in the Year 2000(Session Chairman: T. F. Edgar, Texas)

f~Neural Networks - Overview, Programming, and Adaptation," by J.Hopfield (Caltech)

~~Towards Improved Process Supervision - Algorithms and KnowledgeBased Systems," by A. J. Morris, G. A.Mogague, and M. T. Tham (Newcastle)

~~Towards the Intelligent Controller Formal Integration of Pattern Recognition with Control Theory," by G.Stephanopoulos (MIT)

~~A Neural Network Methodology forProcess Fault Diagnosis," by V. Venkatasubramanian (Purdue)

In order to limit the conference size topreserve an atmosphere for creativediscussion, attendance at the CPC-IV

Conference will be limited and is byinvitation following receipt of this application form. You will be notifiedwhether you will be able to attend theconference. The conference fee of $450will cover registration and a copy ofthe proceedings. The meal package is$282 per person (registrant or guest).Please complete and send this application by September 30, 1990 to:

Application Form for CPC-IV Conference:

30

Three Greenway Plaza, Houston, TX77046-0395, (713) 960-2032.

Fourth World CongressChemical Engineering

Karlsruhe, West GermanyJune 16-21, 1991

This conference, which will directlyfollow the exposition ACHEMA '91, willbe held under the auspices of the European Federation of Chemical Engineering and is being organized by itsGerman Member Societies with participation by AIChE. The main subjectof the Congress will be ~~Strategies

2000" and will stress the interdependence of strategic societal goals, applications in the process industries, andfundamentals and methods of chemical engineering. Of particular interestto members of the CAST Division mightbe the sessions on Process Synthesisand Analysis, and Process Control,Simulation, and Optimization. Deadline for submission of proposed abstracts was June 15 1990, but paperselection will not be made until October. For additional information, contact Dr. Verle N. Schrodt, Bureau ofEngineering Research, University ofAlabama, Box 870201, Tuscaloosa, AL35487-0201, (205) 348-1591 or (205)348- 9455 (FAX).

The instructors of the course are Professors Larry Biegler, Ignacio Grossmann and Arthur Westerberg. For information please call (412)268-2207,or write to Post College ProfessionalEducation, Carnegie Mellon University, Mellon Institute, 4400 Fifth Avenue, Pittsburgh, PA 15213, Attn:Frank E. Nowak, Director.

terns, and planning and scheduling ofbatch processes. Course participantswill address these topics through lectures and hand-on workshops makingextensive use of powerful computersoftware. Also, a comprehensive set oflecture notes and handouts will be provided.

and Area 5D Ses-Area

No Sessions are planned.

Advanced Process SystemsEngineering: Concepts and

PracticeEngineering Design

Research CenterCarnegie Mellon University

Pittsburg, PAJune 3-7, 1991

Topics of this course include conceptsfor process synthesis (heat integration, separation systems, reactor networks), expert systems, modeling environments, flowsheet optimization,mixed-integer optimization models forprocess synthesis, strategies for retrofit design, differential/algebraic sys-

This 5-day course stresses the application of recently developed design concepts and optimization-based strategies to practical process problems.Geared to technical managers, industrial researchers and tool developers,this course provides practical information and exposure to powerful modeling tools for process synthesis, analysis, optimization and planning. In addition the course emphasizes systematic solution approaches and providesthe necessary background to understand the tools and apply them effectively and efficiently to process problems.

1. Applications of Robotics. Michael Tayyabkhan (Chairman),Tayyabkhan Consultants, Inc., 62Erdman Avenue, Princeton, NJ08540, (609) 924-9174 and John Jepsen (Vice Chairman), Clark MaterialsHandling Co., Route 2, Box 46, Highway 33, Versailles, KY 40383, (606)873-9973.

Computers in OperInl:-orlma.tlOln Processing

Areaations

I.David M. (Chairman),Department of Chemical Engineering,University of Texas, Austin, TX78712, (512) 471-7445 and VenkatVenkatasubramanian (Vice Chairman), School of Chemical Engineering, Purdue University, West Lafayette, IN 47907, (317) 494-0734.

2.. Plant-wide Management Systems. K. R. Kaushik (Chairman),Shell Oil Company, PO Box 6249, Carson, CA 90749, (213) 816-2276 and A.L. Parker (Vice Chairman), Shell OilCompany, PO Box 10, Norco, LA70079, (504) 465-7142.

3. Computer Integrated Manufacturing. C. E. Bodington (Chairman), Chesapeake Decision Sciences,PO Box 275, San Anselmo, CA 94960,(414) 453-4906 and Rufus A. Baxley(Vice Chairman), Digital EquipmentCorporation, 5555 Windward Parkway West, Alpharetta, GA 30201,(404) 772-2121.

4. Innovative Uses of Spreadsheets in Engineering Calculations. Randy A. Freeman (Chairman),Monsanto Company - F2WG, 800 N.Lindbergh Blvd., St. Louis, MO 63167,(314) 694-6068 and BruceM. Vrana(Vice Chairman), E. I. DuPont de Nemours & Co., PO Box 6090, Newark,DE 19714-6090, (302) 366-6211.

5. Applications of Expert Systems. James F. Davis (Chairman),Department of Chemical Engineering,Ohio State University, Columbus, OR43210-1180, (614) 292-0090 and Duncan A. Rowan (Vice Chairman), E.I.DuPont de Nemours & Company, POBox 6090, Newark, DE 19714-6090,(302) 366-6453.

31

Fourth InternationalSymposium on ProcessSystems Engineering

(PSE '91)Montebello, Quebec, Canada

August, 4-9 1991

• Statistical Process Control

• Plant Monitoring and Diagnosis

• Human Factors

Methods and Systems

Name:

Affiliation:

Address:

FAX:

Province/State:

Postal Code:

City:

Do

attachedto follow

For additional information, whetheryou are interested in presenting a paper or attending the conference, contact Gerry R. Sullivan (ConferenceChairman), Department of ChemicalEngineering, University of Waterloo,Waterloo, Ontario, CANADA, N2L 3Gl,

(519) 885-2196.

Los Angeles AIChE MeetingNovember 17-22, 1991

Meeting Program Chairman: Dr.Richard L. Bell, Department 01Chemical Engineering, University ojCalifornia, Davis, CA 95616, (916:752-8776.

Abstract:

I am considering attending and wouldlike to receive the Second Announce-ment: yes D no D

Country:

Telephone:

I would like to submit a paper:yes []no 0

New Frontiers

Timetable for Authors:

Data Bases

Integrated Process and Systems Engineerin,g

• Computing Environments

Integrating Marketing and Operations

• Expert Systems

Neural Networks

Education

December 31, 1990 - Full paper forrefereeing

April 30, 1991 - Final manuscript

International Programming Committee: T.F. Edgar, I. Hasimoto, P.L. Lee,M. Morari, E. O'Shima, J.D. Perkins,S. Pierucci, J.W. Ponton, R.G.H.Prince, G.V. Reklaitis, D.W.T. Rippin,R.W.H. Sargent, G. Stephanopoulous,T. Takamatsu, T. Umeda, C.F.H. vanRijn, A.W. Westerberg, and J.D.Wright.

Thinking of Attending:

If you might wish to attend PSE '91,

whether or not you plan to submit apaper, please send in the REGISTRA

TION OF INTEREST form now to ensurethat you receive a copy of the SecondAnnouncement, a form of application

,for registration, and the detailed program.

Following the tradition of the PSE series, the emphasis in 1991 will be onthe presentation of new informationon either technology or its application.Papers describing applications will beespecially welcomed, particularlywhere they contain detailed information relating to the value of a study.

Design

• Flowsheeting

• Steady-State, Batch and Dynamics

• Physical Properties and Data Bases

• Optimization

• Hazards and Reliability

Seven technical sessions are planned,each to be conducted by a ChairmanRapporteur, containing presentationsof five to six papers with full discussion. Following the successful postersessions in previous PSE conferences,similar sessions are planned this time.All of the papers to be presented willbe refereed and the Conference Proceedings will be published. Conferencetopics include:

Process Operations

This conference is being sponsored bythe Canadian Society for ChemicalEngineering (Systems and Control Division), the National Research Councilof Canada (NRC), and the American Institute of Chemical Engineers (CAST

Division). It is the fourth in a triennialseries entitled PSE, and follows highlysuccessful events in Kyoto in 1982,Cambridge in 1985, and Sydney in1988.

• Identification and Control

• Real-Time Optimization

The CAST Division is planning the following sessions at the Los Angeles Annual Meeting. Deadlines and first cal

32

for papers for this meeting appearlater in this issue.

lOA and lOB i:je~SSl~[)n: lOB and

1-2. and ofProcess I and II. Ifte-khar A. Karimi (Chairman), Department of Chemical Engineering, Northwestern University, Evanston, IL60208-3120, (708) 491-3558.

3. Personal Computers in PlantOperations. Michael T. Tayyabkhan(Chairman), Tayyabkhan Consultants, Inc., 62 Erdman Avenue, Princeton, NJ 08540, (609) 924-9174.

4. Computer Architectures.Gary D. Cera (Chairman), Mobil R&DCorporation, PO Box 1026, Princeton,NJ 08543-1026, (609) 737- 5299 andStephen E. Zitney (Vice Chairman),Cray Research Inc., 1333 NorthlandDr., Mendota Heights; MN 55120,(612) 681-3402.

5. Artificial Intelligence in Process Engineering. Venkat Venkatasubramanian (Chairman), School ofChemical Engineering, Purdue University, West Lafayette, IN 47907,(317) 494-0734 and Lyle Un\ger (ViceChairman), Department of ChemicalEngineering, University of Pennsylvania, Philadelphia, PA 19104, (215)898-7449.

6. Integration of Methodologiesfor Process Operations. Ignacio E.Grossmann (Chairman), Departmentof Chemical Engineering, CarnegieMellon University, Pittsburgh, PA15213, (415) 268-2228 and Mark A.

~t~ltl~;tlC~S and WUlalltv

Mohinder K. Sood (Cochairman), Mobil R&D Corporation, PO Box 1026,Princeton, NJ 08543-1026, (609) 7374960, S. H. Mah (Cochair-

Department Chemicalneering, Northwestern University,Evanston, IL 60208-3120, (708) 4915357, and Michael C. Wellonschairman), R&D Corporation,

Box Princeton, NJ 08543-1026, (609) 737-4454.

and ProcessArea lOB:

Recent Advances in ProcessJeffrey C. Kantor (Chair

man), Department of Chemical Engineering, University of Notre Dame,Notre Dame, IN 46556, (219) 239-5797and Paul Gusciora (Vice Chairman).

3. AI in ProcessControl. Melinda Golden (Chairman)and George Stephanopoulos (ViceChairman), Department of ChemicalEngineering, Massachusetts Instituteof Technology, Cambridge, MA 02139,(617) 253-3904.

4. Nonlinear Control. Dale Seborg (Chairman), Department ofChemical and Nuclear Engineering,University of California, Santa Barbara, CA 93106.

5. Robust Control. Evangelos Zafiriou (Chairman), Department ofChemical Engineering, University ofMaryland, College Park, MD 20742,(301) 454-5098 and Ahmet Palazoglu(Vice Chairman), Department ofChemical Engineering, University ofCalifornia, Davis, CA 95616, (916)752-8774.

6. Statistical Process Control.John MacGregor (Chairman) andChristos Georgakis (Vice Chairman),Chemical Engineering Department,Lehigh University, Bethlehem, PA18015, (215) 758- 4781.

7. Control of Discrete Event Processes. Erik Ydstie (Chairman), Department of Chemical Engineering,University of Massachusetts, Amherst, MA 01003-0011, (413) 545-2388and Ed Bristol (Vice Chairman).

1. Batch Process Synthesis andBill Luyben (Chairman) and

Michael F. Doherty (Vice Chairman),Department of Chemical Engineering,University of Massachusetts, Amherst, MA 01003-0011, 3) 545-.2359.

Area lOA: Systems and Process

6. Process Synthesis. James M.Douglas (Chairman), Department ofChemical Engineering, University ofMassachusetts, Amherst, MA 010030011, (413) 545-2252 and Robert L.Kirkwood, Polymers Products Department, E. I. DuPont de Nemours &Company, Wilmington, DE 198800262, (302) 695-3777.

WasteRakesh Govind

(Chairman), Department of Chemicaland Nuclear Engineering, Universityof Cincinnati, Cincinnati, OR 45221,(513) 475-5742 and Vasilios Manousiouthakis (Vice Chairman), Department of Chemical Engineering, University of California, Los Angeles,90024-1592, (213) 825-9385.

2. Information Man n tSystems for Process Design. MarkKramer (Chairman), Department ofChemical Engineering, MassachusettsInstitute of Technology, Cambridge,MA 02139, (617) 253-6508 and HeinzA. Preisig (Vice Chairman), Department of Chemical Engineering, UniversityofNew South Wales, PO Box 1,Kensington, N.S.W. 2033, AUSTRALIA.

3..4. Design and Analysis I and II.Ross E. Swaney (Cochairman), Department of Chemical Engineering,University of Wisconsin, Madison, WI53706, (608) 262-3641.

5. Batch Process Design. DavidW. T. Rippin (Chairman), ChemicalEngineering Department, Swiss Federal Institute of Technology, ETHZentrum, CH-8092 Zurich, Switzerland and Iftekhar A. Karimi (ViceChairman), Department of ChemicalEngineering, Northwestern University, Evanston, IL 60208-3120, (708)491- 3558.

33

Kramer (Vice Chairman), Departmentof Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, (617) 253-6508.

Area 10D: Applied Mathematicsand Numerical Analysis

1..2. Complex Chemical Engineer..ing Systems: Chaos, Fractals, andNeural Networks I and II. Julio M.Ottino (Cochairman), Department ofChemical Engineering, Stanford U niversity, Stanford, CA 94305, (415)723-9596 and Erik Ydstie (Cochairman), Department of Chemical Engineering, University of Massachusetts,Amherst, MA 01003-0011, (413) 5452388.

3. Instabilities and Bifurcationsin Chemical Engineering Applications. Ioannis G. Kevrekides(Chairman), Department of Chemical Engineering, Princeton University, Princeton, NJ 08544, (609) 258-2818 and H.Chia Chang (Vice Chairman), Department of Chemical Engineering, U niversity of Notre Dame, Notre Dame,IN 46556, (219) 239-5857.

4. PDE Simulations in ChemicalEngineering. Antony N. Beris(Chairman), Department of ChemicalEngineering, University of Delaware,Newark, DE 19716, (302) 451-8018and Lyle H. Ungar (Vice Chairman),Department of Chemical Engineering,University of Pennsylvania, Philadelphia, PA 19104-6393, (215) 898-7449.

5. Numerical Methods in Ordinary Differential Equations. S.Sundaresan (Chairman), Departmentof Chemical Engineering, PrincetonUniversity, Princeton, NJ 08544,(609) 258-4583 and Joseph F. Pekny(Vice Chairman), Department ofChemical Engineering, Purdue University, West Lafayette, IN 47907,(317) 494-7901.

6. Stochastic Models. KyriacosZygourakis (Chairman), Departmentof Chemical Engineering, Rice University, Houston, TX 77251-1892,(713) 527-8101 x3509 and Robert M.

Ziff (Vice Chairman), Department ofChemical Engineering, University ofMichigan, Ann Arbor, MI 48109-2136,(313) 764-5498.

For further information details concerning CAST Division sessions andscheduling, contact Jeffrey J. Siirola(Area Programming Chairman), Research Laboratories-B95, EastmanChemical Company, PO Box 1972,Kingsport, TN 37662, (615) 229-3069.

New OrleansAIChE Meeting

March 29 - April 2, 1992

The CAST Division is planning the following tentative program at the NewOrleans National Meeting:

Area lOA: Systems and ProcessDesign

1. Integration of Process Design,Optimization, and Control. HenryChien (Chairman), Monsanto Company, 800 N. Lindbergh Blvd., St.Louis, MO 63167, (314) 694-8274.

2. Training Simulators. Robert J.Farrell (Chairman), Department ofChemical Engineering, PolytechnicUniversity, Brooklyn, NY 11201,(718) 260-3628.

3-4. Modeling and Simulation Iand II. Babu Joseph (Cochairman),Department of Chemical Engineering,Washington University, St. Louis, MO63130, (314) 889-6076 and Rodolphe L.Motard (Cochairman), Department ofChemical Engineering, WashingtonUniversity, St. Louis, MO 63130, (314)889-6072.

Joint lOA and lOe Session:

1. Optimization of Batch UnitOperations. Lorenz T. Biegler(Chairman), Department of ChemicalEngineering, Carnegie Mellon University, Pittsburgh, PA 15213, (412)

34

268-2232 and Iftekhar A. Karimi(Vice chairman), Department ofChemical Engineering, NorthwesternUniversity, Evanston, IL 60208-3120,and S. Macchietto (Vice Chairman).

Area lOB: Systems and ProcessControl

Area tOC: Computers inations and Information Processing

1. Hazard and Operability Analysis. Randy A. Freeman (Chairman),Monsanto Company - F2WG, 800 N.Lindbergh Blvd., St. Louis, MO 63167,(314) 694-6068.

2. Process Data Managementand Reconciliation. Mohinder K.Sood (Chairman), Mobil R&D Corporation, PO Box 1026, Princeton, NJ08543-1026, (609) 737-4960 and Granville E. Paules (Vice Chairman), MobilR&D Corporation, PO Box 1026, Princeton, NJ 08543-1026, (609) 737-4593.

3. Promise of Integration (PanelDiscussion). Jerry Robertson (Chairman) and C. E. Bodington (Vice Chairman), Chesapeake Decision Sciences,PO Box 275, San Anselmo, CA 94960,(414) 453-4906.

4. New Environments for Engineering Computations. Gary D.Cera (Chairman), Mobil R&D Corporation, PO Box 1026, Princeton, NJ08543-1026, (609) 737-5299 and AlanB. Coon (Vice Chairman), Union Carbide Corp., PO Box 8361, SouthCharleston 1 WV 25303, (304) 7475470.

5. Technologies and Tools inSoftware Engineering. Stephen J.Zilora (Chairman), Creative SoftwareSolutions, PO Box 192, Flanders, NJ07836, (201) 927-8233.

6. Technical Desktop Publish ..ing. Peter R. Rony (Chairman), Department of Chemical Engineering,Virginia Polytechnic Institute andState University, Blacksburg, VA24061, (703) 231- 7658 and Richard S.H. Mah (Vice Chairman), Departmentof Chemical Engineering, Northwes-

Advances in "'1'""'U.&. ... JLJL ..... Ll·ll,A.OJ ... 'U' ... JL

_""""II .. ",,r,1I-"1I£1,""" of

Parallel Computing

Applications of Neural Networks

Automation

Area iOC:and I __ -i·- ·~nl ..'..n........ It'ro'ce~~sl.n2

The CAST Division is """"I .................... •........ rt' the following tentative program at Miami National.LvA. ......,v'ULJLJLj;;",.Mark A. Chair-

man), Department of ChemicalUniversity of Illinois, Ur-

IL 333- 0275 andJohn Hale (Conference Vice Chairman), E. I. DuPont de Nemours &Company, PO Box 6090, Newark, DE19714-6090, (302) 366-3041.

No ~e~,SH)nS

491-5357.tern

Re-search Laboratories - Eastman

........ v,L.ll.Jl..&.'..,(,,4J1. Company, PO Box 1lnO'C"'fn.f'\ll"'f" TN 37662, 229-3069.

----------------------------------

35

those applications that use particularlyproaches.

ap-

Vice Chairman

Peter T. CummingsDept of Chern EngUniversity of VirginiaCharlottesville, VA22903-2442(804) 978-4726(804) 924-6270 (FAX)

David R. HeltneShell

Box 1380nOllStOn .. TX 77251

493-7325

Vice Chairman

i'HD.""" ........ ,.,.D F. MaloneDept ofChem EngUn{versity of MassachusettsAmherst, MA 01003-0011(413) 545-0838(413) 545-1647 (FAX)

Chairman

....."V.II..Il.IIJ.Il..Il.U"-A.IIJ..L\.J.Il..A.O describing studies in ""J1.JI.'''''Jl.Jl.J~..Il.'''''LA.J1.

and simulation are requested.sign or their software Im·ple~m(;ntatlon.

plications of existing tools processare particularly welcome. will be given tothat are not covered in other sessions on retrofits, artificialintelligence, and optimization.

Chairman

Emilio J. NunezShell Development Co.PO Box 1380Houston, TX 77251

493-8866

to ses-intent

Area lOa: Systems and Process ....-.....,....,..................

5, 1990: an abstract and lI-I'fI",n-""I>""\.r,£"1r'!l

Present Form to the session chairman..

ucto[)er 1, 1990: Authors informed of selection and sessioncontent finalized.

1, 1991: Submit an extended abstract to be published for distribution at the meeting.

February 15, 1991: Final manuscript submitted to the session chairman.

1990:sion chairman.

names, addresses, and telephone numbers of the sessionchairmen are given on next several pages, as are briefstatements the to special emphasis in se-lecting manuscripts sessions. Prospective sessionparticipants should following .a. ,~

been established

Applications ofand Product Design.

............ .Jl"_.lL1Il....... Intelligence in Process4. Retrofit Design Techniques and AJ:)PIIC~ltllon,s.

2. Industrial Applications of Optimization.

Papers are invited in the areas of expert systems applied toprocess engineering, knowledge representation, architecture of knowledgebased systems for process engineering,product design using AI techniques, knowledge acquisition,and qualitative process modeling.

Papers are sought in all areas of industrial application of optimization techniques. Applications of linear, nonlinear, integer, and mixed-integer programming to the solution ofproblems in industry are welcome. Of particular interestare those applications that have resulted in significantbenefits over the previous method of operation as well as

Vice Chairman

Michael F. MaloneDept of ChemicalEngineeringUniv. of MassachusettsAmherst, MA 01003-0011(413) 545-0838(413) 545-1647 (FAX)

Chairman

Don VredeveldUnion Carbide CorporationPO Box 8361South Charleston, WV 25303(304) 747-4829(304) 747-5448 (FAX)

papers in this session should describe design techniquesand methodology for retrofitting plants or experiences inprocess retrofitting and how retrofit techniques were used toidentify process improvements. The process improvementsachieved might be in areas as improved raw material efficiency, reduced energy consumption, increased capacity, improved product quality, or improved process flexibility.gorithms and procedures and their embodiment in ,",V.lI..l.Jl.IIJUIIJ",;'J1.I

software can be described provided the focus is on the r~n.~nll-l

cal engineering technology or the commercial benefits ofing the software and not on the mathematical proceduresthe features of the software.

Vice Chairman

Krishna R. KaushikShell Oil CompanyPO Box 6249Carson, CA 90749(213) 816-2276(213) 816-2375 (FAX)

BabuJosephDept of Chern EngWashington UniversitySt. Louis, MO 63130(314) 889-6076

Chairman

36

~VI!i:t:tA1n~ and Process Vice

VenkatVenkatasubramanian

s;v~;telns_ neural netanl)rOa~J1~S in enhanc

sensor'It'.n.Bn''''~:ll'V'll'lt' "-,~"'J"U"" '-,.... ~ C'11111f'\d"l.""'llr1lCr"\1'I""l\T control and

aspects ofV't-¥\J .......... JL .... LdOU'"U ... "'''l.J1. may also

Vice

Alcoa PA 15069337-2755.337-2005 (FAX)

William D.Aluminum ~V.n.1V'!r'.lLA?IL""....",<Lln..,...,.a.~"Y ofeng

....................."" .... ...., Institute ofTechnologyChicago, IL 60616

567-3042567-8874 (FAX)

Papers are requested on topics relating to the industrial application of advanced process control methods. Desirableareas include nonlinear model based control, predictivemodel control, adaptive control, robustness, constraint handling, and nonlinear dead time compensation.

2. 1i1l'lll"""'fl1l1l1C''fi''1lt''lillnlll\.J)pJllc~atl.onlS of Process

R......... II.A\.,"-'CLIJ ............ ""

Shell Oil CompanyBox 6249

Carson, CA 90749(213) 816-2276(213) 816-2375 (FAX)

3.

A. L. ll-Io1volrJf"bW"

Shell Oil CompanyPO Box 10

LA 70079465-7142

1. On-line Fault Administration.

Area lOe: Computers in Operations and InformationProcessing

Both theoretical contributions and practical examples aredesired. Any theoretical presentation should have practicalpotential. Examples of fault detection and diagnosis shouldprovide sufficient details so that professionals might use theexamples in their own work. Papers will be reviewed bysenior researchers in' the field, and all authors will be informed of the decision about their paper at the end of the review process. A limited number of papers will be accepted,hence it is advisable to submit your paper as early as possible.

4. Innovative Uses of Spreadsheets in EngineeringCalculations.

Vice Chairman

Rufas A. BaxleyDigital Equipment Corp.5555 Windward ParkwayWestAlpharetta, GA 30201(404) 772-212

C. E. BodingtonChesapeake Decision SciencesPO Box 275San Anselmo, CA 94960(414) 453-4906

This session will cover the use of personal computer spreadsheet packages such as 'LOTUS 123 (DOS Machines) or Excel(Apple Macintosh) in the completion of engineering calculations. Possible topics include two-phase flow calculations,fluid bed design, reaction kinetics, network pressure drop

Chairman

This session will focus on the use of computers to integratethe planning, scheduling, and control of a sequence of eitherbatch or continuous processes or a combination of the two.For example, papers are invited that cover, but are notlimited to: scheduling of product mixing or blending, fromraw material to finished product~ planning and schedulingof multi-step processes for discrete part manufacture, e.g.,printed circuit boards; research results in scheduling or optimization of multi-step production; and fermentation andseparation sequences in pharmaceutical manufacture.

Vice Chairman

Gerardo MijaresM. W. Kellogg CompanyThree Greenway PlazaHouston, TX 77046-0395(713) 960-2032

James B. RiggsDept of Chern EngTexas Technical UniversityLubbock, TX 79409(806) 742-1765(806) 742-1900 (FAX)

Chairman

37

62 Erdman AvenueYfllrlCe1Lon. NJ 08540

924-9174

Vice

session

selection and session con-1, 1991: Authors lIn1r,("\ll"'nnnn

1, an abstract oftion to the session chairman.

A..lII._fiiii._IV/\t;I 1, 1991: an .n'V"f-nnil"'lln.1iI ............ ,.,1r_,""'''''''tr to belished for distribution at the ."...".,.",..n.1l"'1lnrT'

lJctot)er 1, 1991: Finallr"ll"'\OnllllClI'l>1V'l llnt- submitted to the sessionchairman.

Vice

Bruce M. VranaDuPont de Nemours

on which1LJI..Il.tJ'\Jq,.q,.ll..ll.'I,,;~ or are commissioned for installation. Ofinterest are papers on not the

JLllll."" \J.ll.J1.'J~'U'.I\.VS;'JI..'I,,;C.A.A a~eV€~10t)m4~nt of an but alsothe broader Impl€~m«~nt;at]lOnast)ec1tstions in process

are all of novel theoretical or meth-odological developments will also be considered for the ses-sion. abstract should indicate the status ofwork, including the degree of completion and whether theapproach has been applied successfully.

Chairman

A. FreemanMonsanto '-IVll.Jl.Jl. .....'uJ!J!

800 N. .L..IJl.J1.Jl.~...JLUll:lJl.

St.

Area lOe and Area 5D Session:

1. Applications of Robotics.

1. Process 1!LJJ''LJ .. :JI.JIL ....... Jll.lI\. for Waste Mj.ni]mi:l;a1~io]t1..

Area lOA: Systems and Process 1LJJ1'IIJ';:'1l...c::.AJl

Vice Chairman

Vasilios ManousiouthakisDept of Chern EngUniversity of CaliforniaLos Angeles, CA

90024-1592(213) 825-9385

Chairman

Rakesh GovindDepartment of Chemicaland Nuclear EngineeringUniversity of CincinnatiCincinnati, OR 45221(513) 475-5742

Vice Chairman

Duncan A. RowanE. I. DuPont de Nemours& CompanyPO Box 6090Newark, DE 19714-6090(302) 366-6453

Chairman

James F. DavisDept of Chern EngOhio State UniversityColumbus, OH 43210-1180(614) 292-0090(614) 292-9021 (FAX)

2. Information Management Systems for ProcessDesign

Papers are invited that deal with any and all aspects of applications of robotics which could include any or all of thefollowing: surveys, principles and theories, design, casestudies, etc. Presentations based on applications in operating plants, industrial laboratories, and uni versity/government laboratories are welcome.

Chairman

Mark A. KramerDept ofChem EngMassachusetts Instituteof TechnologyCambridge, MA 02139(617) 253-6508

Vice Chairman

Heinz A. PreisigDept ofChem EngUniv. of New South WalesPO Box 1Kensington, N.S.W. 2033AUSTRALIA

38

3-4. Design and Analysis I and II. Area lOB: Systems and Process Control

5. Batch Process Design.

Chairman

Ross E. SwaneyDept of Chern EngUniversity ofWisconsinMadison, WI 53706(608) 262-3641

This session will focus mainly on issues related to the designof noncontinuous or batch plants such as synthesis and sizing of multiproduct and multipurpose plants, design forflexibility, design under uncertainty, scheduling and planning in design, intelligent tools and/or knowledge base fordesign, task synthesis, retrofit design, flexibility and/or uncertainty analysis of designs, design of intermediate storage, etc.

Chairman

David W. T. RippinChemical Engineering DeptSwiss Federal InstituteofTechnologyETH-ZentrumCH-8092 ZurichSWITZERLAND

6. Process Synthesis.

Chairman

Vice Chairman

Iftekhar A. KarimiDepartment of ChemicalEngineeringNorthwestern UniversityEvanston, IL 60208-3120(708) 491-3558(708) 491-3728 (FAX)

Vice Chairman

1..2. Recent Advances in Process Control.

Chairman

Jeffrey C. KantorDepartment of Chemical EngineeringUniversity of Notre DameNotre Dame, IN 46556(219) 239-5797(219) 239-8007 (FAX)

3. AI Applications in Process Control.

Vice Chairman

George StephanopoulosDepartment of Chemical EngineeringMassachusetts Institute ofTechnologyCambridge, MA 02139(617) 253-3904

4. Nonlinear Control.

Chairman

Dale SeborgDepartment of Chemical and Nuclear EngineeringUniversity of CaliforniaSanta Barbara, CA 93106(805) 893-3411

5. Robust Control.

James M. DouglasDepartment of ChemicalEngineeringUniversity of MassachusettsAmherst, MA 01003-0011(413) 545-2252(413) 545-1647 (FAX)

Robert L. KirkwoodPolymer Products DeptE. I. DuPont de Nemours& CompanyWilmington, DE19880-0262(302) 695-3777

Chairman

Evangelos ZafiriouDept ofChem EngUniversity of MarylandCollege Park, MD 20742(301) 454-5098

Vice Chairman

Ahmet PalazogluDept ofChem EngUniversity ofCaliforniaDavis, CA 95616(916) 752-8774

Joint lOa and lOb Session:

1. Batch Process Synthesis and Control.

Vice Chairman

Michael F. DohertyDepartment of Chemical EngineeringUniversity of MassachusettsAmherst, MA 01003-0011(413) 545-2359(413) 545-1647 (FAX)

39

6. Statistical Process Control.

Vice Chairman

Christos GeorgakisChern Eng DeptLehigh UniversityBethlehem, PA 18015(215) 758-4781

7. Control of Discrete Event Processes.

Chairman

Erik YdstieDept of Chern EngUniversity of MassachusettsAmherst, MA 01003-0011(413) 545-2388(413) 545-1647 (FAX)

Joint lOB and lOe Session:

1. Statistics and Quality Control.

Increasingly tighter environmental regulations will requirethe process industries to significantly reduce the emission oftoxic and pollutant waste products. Statistical and qualitycontrol can playa big role in limiting waste product emissions while simultaneously improving product quality. Thissession will include papers describing both fundamentaltechniques and industrial applications of statistical andquality control. Papers describing new statistical tools andtechniques, applications to new areas, time series models,real-time applications, data reconciliation, multivariateanalysis, and interface with data acquisition and data basesare welcome.

certainty, scheduling and planning of multiproduct ,",V.A,.A."'.ll..U.U~

ous plants, scheduling in continuous plants, lJ.J.U•.ll..A.J.'J'VJ.Jl.aAA.\..ti

scheduling, planning/scheduling in refineries, artificialtelligence tools for scheduling and planning, etc.

Chairman

Iftekhar A. KarimiDept of Chernl EngNorthwestern UniversityEvanston, IL 60208-3120(708) 491-3558(708) 491-3728 (FAX)

3. Personal Computers in Plant Operations.

Chairman

MichaelT.TayyabkhanTayyabkhan Consultants, Inc.62 Erdman AvenuePrinceton, NJ 08540(609) '924-9174

4. Computer Architectures.

Chairman Vice Chairman

5. Artificial Intelligence in Process Engineering.

Cochairman

Mohinder K. SoodMobil R&D CorporationPO Box 1026Princeton, NJ 08543-1026(609) 737-4960

Cochairman

Michael C. WellonsMobil R&D CorporationPO Box 1026Princeton, NJ 08543-1026(609) 737-4454

Gary D. CeraMobil R&D CorporationPOBox 1026Princeton, NJ 08543-1026(609) 737-5299

Stephen E. ZitneyCray Research Inc.1333 Northland DriveMendota Heights, MN55120(612) 681-3402

Cochairman

Richard S. H. MahDept of Chern EngNorthwestern UniversityEvanston, IL 60208-3120(708) 491-5357(708) 491-3728 (FAX)

Papers studying the use of Artificial Intelligence in chemical engineering are solicited. Applications of AI to problem~

of process operations, including fault detection, diagnosisand control are sought, as are methodological contribution~

such as novel knowledge representation or reasoning techniques.

Area tOC: Computers in Operations and InformationProcessing

1-2. Scheduling and Planning of Process Operations Iand II.

These two sessions will focus on the operations of both continuous as well as batch processes. The main topics include,but are not restricted to, sequencing and scheduling in batchplants, resource-constrained scheduling, production planning in batch plants, scheduling/planning in the face of un-

40

Chairman

Venkat VenkatasubramanianSchool of Chern EngPurdue UniversityWest Lafayette, IN 47907(317) 494-0734

Vice Chairman

Lyle H. UngarDept ofChem EngUniv.ofPennsylvaniaPhiladelphia, PA19104-6393

(215) 898-7449

computing extends their applicability to more realisticmodels. We solicit papers on the application of these methods to models of chemical engineering processes, with emphasis on pattern formation and selection in distributed systems.

6. Integration of Methodologies for Process Operations.

The goal of this session is to examine the potential benefitsand the various approaches for combining different types ofmethodologies for process operations problems. In particular, this session will focus on the combination of AI techniques (e.g. expert systems, qualitative models, neural networks), with mathematical programming (e.g. linear, nonlinear, and integer programming) or with other search techniques (simulated annealing, genetic algorithms, tablesearch). Applications to scheduling and planning, and similar areas will be considered.

Chairman

Ioannis G. KevrekidesDept of Chern EngPrinceton UniversityPrinceton, NJ 08544(609) 258-2818

Vice Chairman

H.-Chia ChangDept ofChem EngUniversity of Notre DameNotre Dame, IN 46556(219) 239-5847

Chairman

Ignacio E. GrossmannDept of Chern EngCarnegie Mellon UniversityPittsburgh, PA 15213(415) 268-2228(412) 268-7139 (FAX)

Vice Chairman

Mark A. KramerDept of Chern EngMassachusetts Institute ofTechnologyCambridge, MA 02139(617) 253-6508(617) 253-9695 (FAX)

4. PDE Simulations in Chemical Engineering.

As chemical Engineering computing is coming of age, theuse of multidimensional simulations of systems of partialdifferential equations is becoming more widespread covering a large variety of applications ranging from free surfaceflows encountered in coating applications, to nonNewtonian flows and solidification problems encountered inmaterials processing.

Area IOD: Applied Mathematics and Numerical Analysis

1-2. Complex Chemical Engineering Systems: Chaos,Fractals, and Neural Networks I and II.

3. Instabilities and Bifurcations in Chemical Engineering Applications.

Analytical and numerical studies of stability and bifurcation of nonlinear systems allow us to decipher the underlying chemistry, physics, and geometric symmetries of reaction and transport processes. Successful implementations ofthese techniques are becoming available, and large scale

Joseph F. PeknySchool of Chern EngPurdue UniversityWest Lafayette, IN 47907(317) 494-7901(317) 494-0805 (FAX)

Vice Chairman

Lyle H. UngarDept of Chern EngUniv.ofPennsylvaniaPhiladelphia, PA19104-6393(215) 898-7449

Vice Chairman

Chairman

S. SundaresanDept of Chern EngPrinceton UniversityPrinceton, NJ 08544(609) 258-4583

5. Numerical Methods in Ordinary DifferentialEquations.

Papers are sought on novel mathematical concepts and algorithmic approaches in the solution of ordinary differentialequations; advanced numerical techniques for continuation, bifurcation canalysis, and tracking singularities suchas homoclinic and heteroclinic connections; and solution ofstiff differential equations. Emphasis will be more on novelnumerical methods as opposed to description of exampleswhere known numerical techniques are applied.

Chairman

Antony N. BerisDept of Chern EngUniversity of DelawareNewark, DE 19716(302) 451-8018

Vice Chairman

Erik YdstieDept of Chern EngUniv. of MassachusettsAmherst, MA 01003-0011(413) 545-2388

Chairman

Julio M. OttinoDept of Chern EngStanford UniversityStanford, CA 94305(415) 723-9596

A large number of recent developments in a variety of areasranging from mathematics to computer science to theoretical physics indicate that the behavior of many systems thaton first viewing appear quite complex can in fact be understood, modelled, and predicted in terms of simple fundamental rules. Papers are sought on applications of some of thesedevelopments, for example neural networks, lattice models,chaos, and fractals to systems of interest in chemical engineering.

41

6. Stochastic Models.

This session will focus on applications of probabilistic concepts to continuous and discrete models of chemical engineering systems. Topics of interest include (but are notlimited to) chemical reaction models, percolation processes,population balance models, turbulence, stochastic control,etc.

ic communication of manuscripts; and so forth. At least onepresentations describing personal experience with technicaldesktop publishing would be welcome.

Chairman

Kyriacos ZygourakisDept of Chern EngRice UniversityHouston, TX 77251-1892(713) 527-8101 x3509

Vice Chairman

Robert M. ZiffDept of Chern Eng

. University of MichiganAnn Arbor, MI 48109-2136(313) 764-5498(313) 763-0459 (FAX)

Chairman

Peter R. RonyDept of Chern EngVirginia Polytechnic Instituteand State UniversityBlacksburg, VA 24061(703) 231-7658

Vice Chairman

Richard S. H. MahDept of Chern EngNorthwestern UniversityEvanston, IL 60208(708) 491-5357

Preliminary Call for CAST SessionNew Orleans AIChE Meeting

March 29-April 2, 1992

Tentative deadlines, based upon the 1991 Houston AIChE

Meeting, are:

August 15, 1991: Submit a letter of intent and title to session chairman.

September 5, 1991: Submit an abstract and Proposal-toPresent Form to the session chairman.

October 1, 1991: Authors informed of selection and sessioncontent finalized.

January 1, 1992: Submit an extended abstract to be published for distribution at the meeting.

February 15, 1992: Final manuscript submitted to the session chairman.

Area lOC: Computers in Operations and InformationProcessing

6. Technical Desktop Publishing

Papers are solicited that provide tutorial information concerning current uses of fftechnical desktop publishing"namely, the desktop publishing of technical documents thatinclude equations, graphs, tables, and text-in industry, government, and academia. Of special interest is the issue ofthe standardization of equation capture; electronic communication of manuscripts to publishers and printers; custompublishing systems (e.g., McGraw-Hill) and related marketing issues; corporate practices; the role of professional societies and their journal policies directed toward the electron-

42

1.

2.

3. Present Position (exact title)

4. Education:

Institution

5. Positions Held:

Degree Received Year Received Field

Company or Institution Position or Title Dates

6. Academic and Professional Honours (include awards, memberships in honorary societies andfraternities, prizes) and date the honor was received.

7. Technical and Professional Society Memberships and Offices

8. Sponsor's Name and Address

_____________________________ Sponsor's Signature*A person may be nominated for only one award in a given year.

B. CITATION

1. A brief statement, not to exceed 250 words, of why the candidate should receive this award.(Use separate sheet of paper.)

2. Proposed citation (not more than 25 carefully edited words that reflect specificaccomplishments).

Each award has a different set of qualifications. These are described in the awards brochure. Afterreading them, please fill in the following information on the nominee where appropriate. Use aseparate sheet for each item if necessary.

1. Selected bibliography (include books, patents, and major papers published.)

2. Specific identification and evaluation of the accomplishments on which the nomination is based.

3. If the nominee has previously received any award from AIChE or one of its Divisions, an explicitstatement of new accomplishments or work over and above those cited for the earlier awards(s)must be included.

4. Other pertinent information.

D. SUPPORTING LETTERS AND DOCUMENTS

List of no more than five individuals whose letters are attached.

1.

2.

3.

4.

5.

Name Affiliation

Please send the completed form and supplemental sheets at any time to the CAST Division 2nd ViceChairman, Professor Ignacio Grosmann, Chemical Engineering Department, Carnegie-MellonUniversity, Pittsburgh, PA, 15213. Telephone: (412) 268-2228.BITNET: [email protected] Grossmann will forward your nomination packages to the next CAST Division 2nd ViceChairman who will be elected this Fall.

Join The CAST Division ofAIChE! Receive This Newsletter

Already a member? Please ask a friend to join.

The Computing and Systems Technology (CAST) Division of AIChE is responsible for the wide range ofactivities within AIChE that involve the application of computers and mathematics to chemicalengineering problems, including process design, process control, operations and applied mathematics.We arrange technical sessions at AIChE Meetings, organize special conferences, and publish thisnewsletter - CAST Communications - twice a year. These activities enable our members to keepabreast of the rapidly changing fields of computers and systems technology. Shouldn't you join theCAST Division now? The cost is only $5 per year, and includes a subscription to this newsletter.

Application For Membership

I wish to join the Computing and Systems Technology (CAST) Division of AIChE

Date:

Name:

Title:

Company/University:

Business Address:

City:

Home Address:

City:

Preferred mailing address: Home Office

I am a member of AIChE Yes N0

(If not, I understand that I must join AIChE within a one year period to continue as a CAST Divisionmember.)

___ My CAST dues of $5 are enclosed

___ I will pay my CAST dues with my annual AIChE dues

Please mail this application to:

American Institute of ChemicalComputing and Systems Technology Division345 East 47th StreetNew York, NY 10017

ofContents - CAST) Div 2.pdfSt. Louis, MO 63167 (314) 694-8274 Director, 1989-1991 ArthurL. Parker...

Documents

Transcript of ofContents - CAST) Div 2.pdfSt. Louis, MO 63167 (314) 694-8274 Director, 1989-1991 ArthurL. Parker...