26 European Symposium on Computer Aided Process Engineering PDF
Preview 26 European Symposium on Computer Aided Process Engineering
Elsevier Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, USA Copyright © 2016 Elsevier B.V. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress ISBN (Part A): 978-0-444-63873-1 ISBN (Set): 978-0-444-63428-3 ISSN: 1570-7946 For information on all Elsevier publications visit our website at https://www.elsevier.com/ Publisher: John Fedor Acquisition Editor: Kostas Marinakis Editorial Project Manager: Sarah J. Watson Production Project Manager: Mohanapriyan Rajendran Designer: Greg Harris Typeset by SPi Global, India Preface These proceedings contain contributions presented at the 26th European Symposium on Computer Aided Process Engineering (ESCAPE 26) held in Portorož, Slovenia from the 12th to 15th of June 2016. ESCAPE 26 is the 26th edition of ESCAPE series. Previous editions were held in Denmark (1992, 2001, 2015), France (1992, 2008), Austria (1993), Ireland (1994), Slovenia (1995), Greece (1996, 2011), Norway (1997), Belgium (1998), Hungary (1999, 2014), Italy (2000, 2010), The Netherlands (2002, Finland (2003, 2013), Portugal (2004), Spain (2005), Germany (2006), Romania (2007), Poland (2009) and United Kingdom (2012). The ESCAPE series is a prominent European yearly event in the field of Computer Aided Process Engineering (CAPE) bringing together people from around the world dedicated to Process Systems Engineering. Scientists, industrial experts, academics and students exchange and share the most recent knowledge about CAPE concepts, methods, tools and applications, either regarding continuous developments of existing technologies or innovative developments of new ones based on new discoveries and inventions. The ESCAPE 26 focused on themes: Process-product Synthesis, Design and Integration; Modelling, Numerical Analysis, Simulation and Optimization; Process Operations and Control; Green Bioprocess Engineering and Advances in Biomedical Engineering; CAPE/PSE in Environmental Engineering; CAPE/PSE in Sustainable Energy Applications; CAPE Applications; and Education in CAPE/PSE. Out of 406 contributions, prepared by authors from 47 countries (Europe, the Americas, Africa, Asia and Australia), the ESCAPE 26 International Scientific Committee selected 176 oral presentations, 7 being plenary lectures, 18 keynote lectures, and 230 poster presentations. We beleive that the contributions contained in these proceedings will serve as a reference and motivate new ideas, developments and collaborations in the field of computer aided process engineering and beyond. Yours sincerely, Zdravko Kravanja Miloš Bogataj Conference Chair Conference Secretary International Scientific Committee Conference Chairman Zdravko Kravanja, University of Maribor, Slovenia Themes Coordinators Process-product Synthesis, Design and Integration Mariano Martin, University of Salamanca, Spain Modelling, Numerical analysis, Simulation and Optimization Stratos Pistikopoulos, Texas A&M University, USA Process Operations and Control Sebastian Engell, TU Dortmund University, Germany Green Bioprocess Engineering and Advances in Biomedical Engineering David Bogle, University College London, UK CAPE/PSE in Environmental Engineering André Bardow, RWTH Aachen University, Germany CAPE/PSE in Sustainable Energy Applications Petar Varbanov, Pázmány Péter Catholic University, Hungary CAPE Applications Flavio Manenti, Politecnico di Milano, Italy Education in CAPE/PSE Antonio Espuña, Universitat Politecnica de Catalunya, Spain Members Mariano Martin, University of Salamanca, Spain Thomas Adams, McMaster University, Canada Mario Eden, Auburn University, USA Fengqi You, Northwestern University, USA Michael Fairweather, University of Leeds, UK Antonis Kokossis, National Technical University of Athens, Greece Rafiqul Gani, Technical University of Denmark, Denmark Andrzej Kraslawski, Lappeenranta University of Technology, Finland Jose Caballero, University of Alicante, Spain Gonzalo Guillén-Gosalbez, The University of Manchester, UK Stratos Pistikopoulos, Texas A&M University, USA Marianthi Ierapetritou, Rutgers University, USA Andreja Nemet, University of Maribor, Slovenia xxviii International Scientific Committee Ruth Misener, Imperial College London, UK Alexander Mitsos, RWTH Aachen University, Germany Michael Georgiadis, Aristotle University of Thessaloniki, Greece Lazaros Papageorgiou, University College London, UK Vivek Dua, University College London, UK Chrysanthos Gounaris, Carnegie Mellon University, USA Pei Liu, Tsinghua University, China Benoit Chachuat, Imperial College London, UK Carl Laird, Purdue University, USA Nikolaos Sahinidis, Carnegie Mellon University, USA Hermann Feise, BASF, Germany Davide Manca, Politecnico di Milano, Italy Moisès Graells, Universitat Politècnica de Catalunya, Spain Sebastian Engell, TU Dortmund University, Germany Ana Barbosa-Póvoa, Técnico Lisboa, Portugal Luis Puigjaner, Universitat Politecnica de Catalunya, Spain Gintaras Reklaitis, Purdue University, USA Sigurd Skogestad, Norwegian University of Science and technology, Norway David Bogle, University College London, UK Paul Agachi, Kazakh British Technical University, Kazakhstan Andreas Linninger, University of Illinois at Chicago, USA Ioannis Androulakis, Rutgers University, USA Costas Kiparissides, Aristotle University of Thessaloniki, Greece Francois Marechal, École polytechnique fédérale de Lausanne, France Tony Kiss, AkzoNobel, The Netherlands Filip Logist, University of Leuven, Belgium Michael Narodoslawsky, Graz University of Technology, Austria André Bardow, RWTH Aachen University, Germany Jiří Klemeš, Pázmány Péter Catholic University, Hungary Peter Mizsey, Budapest University of Technology and Economics, Hungary Fabrizio Bezzo, University of Padova, Italy Ferenc Friedler, University of Pannonia, Hungary Niall Mac Dowell, Imperial College London, UK Petar Varbanov, Pázmány Péter Catholic University, Hungary Hon Loong Lam, The University of Nottingham Malaysia Campus, Malaysia Zainuddin A. Manan, University of Technology, Malaysia Sharifah Rafidah Wan Alwi, University of Technology, Malaysia Raymond Girard R. Tan, De la Salle University, Philippines Petro Kapustenko, Kharkiv Polytechnic Institute, Ukraine Igor Bulatov, The University of Manchester, UK Panos Seferlis, Aristotle University of Thessaloniki, Greece Sakis Papadopoulos, Centre for Research and Technology Hellas, Greece Tomislav Novosel, University of Zagreb, Croatia Franjo Cecelja, University of Surrey, UK Hella Tokos, University of Surrey, UK International Scientific Committee xxix Niyi Isafiade, University of Cape Town, South Africa Michael Walmsley, The University of Waikato, New Zealand Flavio Manenti, Politecnico di Milano, Italy Tilman Barz, Austrian Institute of Technology, Austria Mattia Vallerio, University of Leuven, Belgium Guido Buzzi-Ferraris, Politecnico di Milano, Italy Massimiliano Barolo, The University of Padova, Italy Soledad Diaz, Planta Piloto de Ingeneria Quimica, Argentina Christos Maravelias, University of Wisconsin, USA Alessio Frassoldati, Politecnico di Milano, Italy Chi Wai Hui, The Hong Kong University of Science and Technology, China Rubens Maciel Filho, University of Campinas, Brazil Antonio Espuña, Universitat Politecnica de Catalunya, Spain Il Moon, Yonsei University, South Korea Henrique Matos, Técnico Lisboa, Portugal Zorka Novak Pintarič, University of Maribor, Slovenia Emilia Kondili, Technological Education Institute of Piraeus, Greece Valentin Plesu, Polytechnic University of Bucharest, Romania Iqbal Mujtaba, University of Bradford, UK Elvis Ahmetović, University of Tuzla, Bosnia and Herzegovina Dimitrios Gerogiorgis, The University of Edinburgh, UK Lidija Čuček, University of Maribor, Slovenia Alexandra Elena Bonet-Ruiz, Polytechnic University of Bucharest, Romania Jordi Bonet-Ruiz, University of Barcelona, Spain Ignacio Grossmann, Carnegie Mellon University, USA Iiro Harjunkoski, ABB Corporate Research, Germany Lorenz Biegler, Carnegie Mellon University, USA Pedro Castro, University of Lisbon, Portugal Carlos Mendez, National University of Litoral, Argentina Cesar de Prada, University of Valladolid, Spain Ton Backx, Eindhoven University of Technology, Netherlands Claudio Scali, University of Pisa, Italy Rajagopalan Srinivasan, Indian Institute of Technology Gandhinagar, India Nilay Shah, Imperial College London, UK Peter Singstad, Cybernetica AS, Norway Stefan Kraemer, TU Dortmund University, Germany Igor Plazl, University of Ljubljana, Slovenia Xavier Joulia, ENSIACET, France Jean-Marc Le Lann, ENSIACET, France Jan Thullie, Silesian University of Technology, Poland Luis Cisternas, University of Antofagasta, Chile Sauro Pierucci, Politecnico di Milano, Italy Vladimir Mahalec, McMaster University, Canada Local Organising Committee Chairman Zdravko Kravanja, University of Maribor, Slovenia Conference Secretary Miloš Bogataj, University of Maribor, Slovenia Members Igor Plazl, University of Ljubljana, Slovenia Neven Duić, University of Zagreb, Croatia Zorka Novak Pintarič, University of Maribor, Slovenia Mojca Slemnik, University of Maribor, Slovenia Andreja Nemet, University of Maribor, Slovenia Lidija Čuček, University of Maribor, Slovenia Zdravko Kravanja, Miloš Bogataj (Editors), Proceedings of the 26th European Symposium on Computer Aided Process Engineering – ESCAPE 26 June 12th -15th, 2016, Portorož, Slovenia © 2016 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/B978-0-444-63428-3.50005-9 Modelling and iterative Real-time Optimization of a homogeneously catalyzed hydroformylation process ReinaldoHerna´ndezandSebastianEngell ChairofProcessDynamicsandOperations,TechnischeUniversita¨tDortmund, Emil-Figge-Strasse70,Dortmund44227,Germany {reinaldo.hernandez,sebastian.engell}@bci.tu-dortmund.de Abstract Inthiscontribution,thereal-timeoptimization(RTO)ofanovelmultiphasehomogeneouslycat- alyzedprocessisinvestigated. Specifically,theconversionoflongchainolefinsisstudiedforthe exampleofthehydroformylationof1-dodecene. Afirstprinciplemodelwasbuiltandvalidated withexperimentaldataavailablefromaminiplant. Aniterativeoptimizationschemeisproposed inordertoensureconvergence totheplantoptimuminpresenceofmodeluncertainties. Local quadraticapproximationcombinedwithderivative-freeoptimizationisusedfortheestimationof theplantgradientsunderpresenceofnoisymeasurements.Thesimulationresultsshowsignificant improvementsintheeconomicperformanceoftheprocessdespitemodeluncertaintiesconcerning thegassolubilityandtheamountofactivecatalystinsolution. Keywords:real-timeoptimization,homogeneouscatalysis,modifieradaptation. 1.Introduction Thehighselectivityandactivityatmildconditionsthatcanbeachievedbymeansofhomogeneous catalysishavemotivatedsignificantresearch,aswellasthedevelopmentofimportantindustrial applications,wherethehydroformylationofpropeneintheRuhrchemie/Rhone-Poulencprocess (RCH-RP)isthebestknownexample(Frey,2014). Theprocessingoflongchainedolefinispar- ticularlychallengingduetotheirlowsolubilityinthepolarcatalystphase.Novelprocessconcepts havebeendevelopedduringthelastyearsinordertoovercometheselimitations,includingtheuse ofcomplexsolventsystemswhichareabletoperformthereactionhomogeneouslyathightem- peratures,whileatalowtemperaturethemixtureseparatesintotwophasesandefficientcatalyst recoveryispossibleinsuchaThermomorphicMulticomponentSolvent(TMS)System(Brunsch andBehr,2013). Fortheeconomicsuccessofthistechnology,minimizingcatalystleachingwhileachievinghigh conversionandselectivityoftheprocessarecrucial. Processoptimizationcanbeappliedbased on kinetic models and thermodynamic descriptions of the phase equilibrium (Hentschel et al., 2015;Schaferetal.,2012). Inthiscontribution,optimaleconomicaloperationofahydroformy- lationprocesshasbeenaddressedbymeansofReal-timeOptimization(RTO).Inthisapproach the problem is decomposed hierarchically: an upper layer uses a rigorous static model to de- termine the optimal set-points that are tracked by a lower layer operating at a higher sampling frequency(Darbyetal.,2015). However,duetomodelinaccuracies,thiswillnotleadtoanoptimaloperationoftherealplant.We proposetheuseofaniterativeReal-timeOptimizationscheme(GaoandEngell,2005)tooptimize thedosingofthecatalystandthereactiontemperatureinpresenceofmodel-plantmismatch.The 2 R.Herna´ndez andS.Engell Figure1:Left:ThermomorphicMulticomponentSolvent(TMS)system.Right:Reactionnetwork ofthehydroformylationof1-dodecene. algorithmemployedheremakesuseofalocalquadraticapproximationfortheestimationofthe plant gradients which leads to reduced sensitivity to measurement noise (Gao et al., 2015). In addition,atrustregionframeworkisappliedtoavoidoscillations. This contribution is structured as follows: section two describes the integrated process for the hydroformylationof1-dodeceneinaThermomorphicMulticomponentSolvent(TMS)systemand insection3themodelispresented. SectionfourgivesageneraloverviewoftheproposedRTO scheme.Finally,inthesectionsfiveandsixtheresultsandthemainconclusionsarepresented. Acknowledgment: ThisworkwassupportedaspartoftheCollaborativeResearchCenter: ”In- tegratedChemicalProcessesinLiquidMultiphaseSystems”(SFB/Transregio63InPROMPT)by theDeutscheForschungsgemeinschaft(DFG). 2.ProcessDescription InFigure1,thediagramofaTMSsystemisshown. Aswasstatedintheintroduction,theidea is to use a predefined mixture of a nonpolar solvent (solvent 1) and a polar solvent (solvent 2) suchthatathightemperaturesthesystemishomogeneous; thereforethelimitationsinthemass transferofthereactanttothecatalystphaseareavoided. Atlowtemperatures,phaseseparation takes place, and the expensive catalyst in the polar phase can be recovered and recirculated to the reactor while the product in the nonpolar phase goes to further purification steps (Brunsch andBehr,2013).TheTMSconcepthasbeenalreadyappliedtodifferenthomogeneouscatalyzed reactions, includingthehydroformylationoflongchainalkenes. Theprocessconcepthasbeen proveninacontinuouslyoperatedminiplantatTUDortmund(Zagajewskietal.,2015). Besides themainreactioncorrespondingtotheproductionofthelinearaldehyde(tridecanal),isomeriza- tiontoiso-dodecene,hydrogenationtododecaneandformationofbranchedaldehydetakeplace; asisshowninthereactionnetwork(Figure1). Highselectivityatmildconditions(90◦Canda syngas pressure of 20 bar) has been achieved by using a catalyst-ligand complex consisting of theprecatalystRh(acac)(CO) andtheligandbiphephos. Assolventsystem,amixtureofthepo- 2 lar solvent dimethylformamide (DMF) and the nonpolar solvent decane was used. The cost of rhodiumandligandmakesthecatalystrecoverycrucialforthesuccessoftheprocessconcept. 3.ModelDescription Afirstprinciplesmodelfortheintegratedoperationofthereactoranddecanterispresentedinthis section. 3.1.Reactormodel ThereactorismodelledasanidealCSTR,andtwo-filmtheoryisusedfordescriptionofthemass transferbetweenthegasandtheliquidphases. Accordingtothematerialbalance,theconcentra- ModellinganditerativeReal-timeOptimizationofahomogeneouslycatalyzed hydroformylationprocess 3 tionofthedifferentliquidcomponentsC (i=1-dodecene,tridecanal,dodecane,etc)isgivenby i equation(1). M isthemassofactivecatalystinthereactorandV isthereactorvolume. The cat R inflowandoutflowvolumetricflowratesaregivenbyV˙in andVo˙ut,whileνi,l arethecoefficients ofthestoichiometricmatrixandr isthereactionrateforthereactionl. Theinterestedreaderis l referredtotheliteratureforthekineticmodel(Hentscheletal.,2015). VRddCti =V˙in·Ci,in−Vo˙ut·Ci,out+McatN∑reactνi,l·rl (1) l=1 Forthe jgascomponents(j=COandH ),theconcentrationintheliquidphasedependsonthe 2 masstransfercoefficient(k )andtheequilibriumconcentrationattheinterface(Ceq): eff ddCtj =−keff·(Cj−Ceqj)+V˙in·Cj,in−Vo˙ut·Cj,out+McatN∑reactνj,l·rl. (2) l=1 TheGLequilibriumismodelledbymeansofHenry’slaw(3),whichdescribestherelationbetween theliquidconcentrationstothepartialpressureP.ThedependenceoftheHenrycoefficientH on j j thereactiontemperatureT ismodelledbyanArrheniusexpression(4). P =Ceq ·H (3) j j j (cid:2) (cid:3) −E Hj=Hj,0·exp RTj (4) Aswasstatedinequation(1)thereactionrateisproportionaltothemassofactivecatalystand thereforeproportionaltoitsconcentrationC .Athighcarbonmonoxideconcentrations,catalyst cat deactivationtakesplacebyformationofinactiveRh-dimers. Thisphenomenonhasbeenapprox- imatelyquantifiedaccordingtoequation(5),wheretheactivecatalystconcentrationisexpressed asafunctionoftheconcentrationofcatalystprecursorCRh,precursorandtheequilibriumconstants Kcat,1andKcat,2: C = CRh,precursor . (5) cat 1+Kcat,1·CCO+Kcat,2·CCO/CH2 3.2.Decantermodel LLEisassumedbetweenthephasesinthedecanter. Simpleexpressionsoftheequilibriumcon- stantK asafunctionofthedecantertemperatureT areused: i decanter (cid:4) (cid:5) Ki=exp Ai,0+TAi,1 +Ai,2Tdecanter , (6) decanter whereAi,0,Ai,1 andAi,2 wereobtainedbyregressionofexperimentaldata(Schaferetal.,2012). Thesplitfactorζiandthemolarflowofthecomponentsintheproductstream(ni,product)andthe catalystrecycle(ni,catalyst)asafunctionoftheinletflowtothedecanter(ni,decanter)canbedefined accordingto(7)-(9): K ζ= i , (7) i 1+K i ni,product=ζi·ni,decanter, (8) ni,catalyst=(1−ζi)·ni,decanter. (9) Based on experimental data, expressions similar to (7)-(9) are used for the description of the catalystleaching. 4 R.Herna´ndez andS.Engell 4.IterativeReal-timeOptimization 4.1.Problemformulation InRTO,thesolutiontothestaticoptimizationproblem(10)isaddressed: min φ(u) (10a) u s.t. G(u)≤0, (10b) whereustandsforthesetofinputs/degreesoffreedomwhichminimizethecostfunctionφ(u). G(u) is the vector of constraint functions which includes the model equations and the process constraintsthatmustbesatisfied.Duetomodellinginaccuraciesanddisturbances,thevalueofthe costfunctionandconstraintsfortheactualprocess(φ(u)andG respectively)willbedifferent p p fromthemodel.Asaconsequence,thesolutionoftheproblem(10)canleadtoasuboptimalorto anunfeasibleoperatingpoint. Iterativeoptimizationaddressesthemodel-plantmismatchbyamodificationoftheproblem(10) withcorrectiontermsthatrepresenttheactualvaluesoftheobjectivefunctionandconstraintsas well as their gradients. In each iteration k, the problem (11) is solved and a new input uk+1 is computed(GaoandEngell,2005). (cid:2) (cid:3) (cid:2) (cid:3) T min φ(uk+1)+φp(uk)−φ(uk)+ ∇φp(uk)−∇φ(uk) uk+1−uk (11a) uk+1 (cid:2) (cid:3) (cid:2) (cid:3) T s.t. G(uk+1)+GP(uk)−G(uk)+ ∇Gp(uk)−∇G(uk) uk+1−uk ≤0. (11b) Thealgorithmensuresconvergencetotheoptimalpointundermodeladequacyconditions(Chachuat etal.,2009). 4.2.Estimationofthegradientsandquadraticapproximation Themostimportantissueintheapplicationofiterativeoptimizationistheestimationofthegradi- entsundernoisyconditions.Differentapproachescanbefoundintheliterature,includingamong others, nested modifier-adaptation (Navia et al.,2013) and dual-modifier adaptation (Marchetti, 2015). Inthiswork,localquadraticapproximationoftheplant,asproposedbyGaoetal.(2015) incombinationwithatrustregionframework(Biegleretal.,2015)isapplied. Thequadraticap- proximationcanbeconsideredasareducedmodel(RM)ofthetrueplantmodel,whiletheoriginal detailedmodel(ODM)istheplantitself. Giventhenumberofinputsn ,theideaofquadraticapproximationistoconstructalocalquadratic u modelofthecostfunctionandtheconstraintsoftheplantbasedonatleastn =[(n +1)(n + r u u 2)/2−1] regression pointsU ={u(cid:6)k,uk−1,...,uk−nr(cid:7)+1}. The problem is reduced to finding the valuesofthesetofparametersθ:= ai,j,...,bi,...,c whichminimizethesquareoftheestimation errors,accordingtoproblem(12): (cid:2) (cid:3) min ∑nr φp(uk−i+1)−φq(uk−i+1,θ) 2 (12a) θ i=1 s.t. φq(uri,θ)=∑nu ∑nu ai,juiuj+∑nu biui+c, (12b) i=1j=1 i=1 Aftertheproblem(12)hasbeensolved,thegradientsareestimatedusingthequadraticapproxima- tionoftheobjectivefunction(φ ≈φ)andtheoptimizationproblem(11)issolved.Anadditional q p constraintisintroducedtotheproblem(11)whichrepresentsboundsontheinputssuchthatthey staywithinthetrustregion. Thetrustregionisupdatedineachstepandthealgorithmisrepeated untilconvergence.
The list of books you might like
