44
School of Engineering ICP Institute of Computational Physics Zürcher Fachhochschule www.zhaw.ch Research Report 2008

ZHAW ICP Research Report 2008

Embed Size (px)

DESCRIPTION

Reserach Report 2008 vom Insitute of Computational Physics (ICP) der ZHAW School of Engineering.

Citation preview

Page 1: ZHAW ICP Research Report 2008

School of Engineering

ICP Institute ofComputational Physics

Zürcher Fachhochschule www.zhaw.ch

Research Report 2008

Page 2: ZHAW ICP Research Report 2008

Die Titelgrafiken zeigen einen Ausschnitt aus einergemischt ionisch-elektronisch leitenden Elektrode einerHochtemperatur-Brennstoffzelle. In der Elektrode findenLadungstransport und Ladungstransfer (d.h. die Elek-trochemie) statt. Die Bilder zeigen die ionisch-elektronischen Phasen (grune und blaue Partikel) sowiedie berechneten Verlaufe elektrischer Potentiale fur ver-schieden gut leitende Elektroden. So lassen sich inSESES Modellen reale Alterungsprozesse abbilden, diezu Leistungseinbussen fuhren.

The illustrations on the frontpage show sections of amixed ionic/electronic conducting electrode of a high tem-perature fuel cell. Charge transport and electrochemicalcharge transfer occur in this electrode composed of ionicand electronic phases, see green and blue particles, re-spectively. The electrical potential distribution is calcu-lated with our software SESES for electrodes of varyingmorphology and conductivity. This way one can study realaging processes that lead to performance reduction.

Page 3: ZHAW ICP Research Report 2008

Contents

1 Einleitung 5

2 Introduction 7

3 Sensors and Actuators 93.1 Finite element software for the industry . . . . . . . . . . . . . . . . . . . . . . . . . 103.2 Numerical modeling of laminates anisotropic plastic deep-drawing . . . . . . . . . . 113.3 Beruhrungsloses Signalanalyseverfahren fur Schichtsystem . . . . . . . . . . . . . . 123.4 Velocimetry in railcars by a run-time measurement of thermal markers . . . . . . . . 133.5 Diagnostischer Atemanalysator mit elektronischer Mikronase . . . . . . . . . . . . . 14

4 Fuel Cells 154.1 Modeling PEM fuel cell performance with a focus on porous layer properties . . . . . 164.2 Simulation of polymer-electrolyte-membrane fuel cells and stacks . . . . . . . . . . . 184.3 Enhancing the lifetime of SOFC stacks for combined heat and power applications . . 194.4 Numerical simulation for the development of the Hexis SOFC-system . . . . . . . . . 204.5 Feasibility of a portable SOFC for Hilti’s mobile machine tools . . . . . . . . . . . . . 22

5 Organic Electronics and Photovoltaics 235.1 Advanced experimentally validated integrated OLED model . . . . . . . . . . . . . . 245.2 Efficient areal organic solar cells via printing . . . . . . . . . . . . . . . . . . . . . . . 255.3 Cost efficient thin film photovoltaics for future electricity generation . . . . . . . . . . 265.4 Organic electroluminescent pictograms for push-button applications . . . . . . . . . 275.5 Modeling, simulation and loss analysis of dye-sensitized solar cells . . . . . . . . . . 295.6 Photovoltaics inspector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305.7 Optoelectronic research laboratory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

6 Student Projects 336.1 Wirtschaftlichkeit und Okobilanz eines Blockheizkraftwerks . . . . . . . . . . . . . . 346.2 Analyse der U–I–Kennlinie von SOFC Hochtemperatur-Brennstoffzellen . . . . . . . 356.3 Optimierung der Holzvergaseranlage in Wila . . . . . . . . . . . . . . . . . . . . . . 366.4 Prufstand fur thermo-mechanische Belastungstests von Brennstoffzellen . . . . . . 376.5 High Performance Grating Spectrometer . . . . . . . . . . . . . . . . . . . . . . . . . 38

7 Appendix 397.1 Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407.2 Conferences and Workshops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407.3 Prizes and Awards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427.4 Teaching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437.5 ICP Team Members . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447.6 Location and Contact Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

3

Page 4: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

ZHAW 4

Page 5: ZHAW ICP Research Report 2008

Chapter 1

Einleitung

Das Institute of Computational Physics hat im Jahr 2008 schulintern wie auch in der nationalenund internationalen Hochschulszene seinem Namen Ehre erwiesen. Es wurden verschiedeneForschungsprojekte im Bereich der numerischen Modellierung von Multiphysik-Systemen abge-schlossen und mindestens so viele neu lanciert. Wir durften in unserer Forschungstatigkeit vielespannende Herausforderungen zusammen mit unseren Industrie- und Hochschulpartnern mei-stern. An der Schnittstelle zwischen akademischer und angewandter Forschung sehen wir unstaglich neuen Herausforderungen gegenuber. Wie konnen wir mit unseren Modellen und Sim-ulationen fur unsere Forschungspartner einen Mehrwert schaffen? Die Antwort ist vielschichtig.Zum einen mussen wir das Vertrauen unseres Partners gewinnen und zeigen, dass wir mit un-seren Berechnungen Erkenntnisse erzielen, welche allein durch experimentelle Untersuchungennur schwierig zu erreichen sind. Andererseits mussen wir auch uberzeugend darlegen, dass wirnumerische Simulationen nicht nur unserem Handwerk zuliebe, quasi ”l’art pour l’art”, durchfuhren,sondern im Sinne der Projektziele optimal einsetzen. Eine Zusammenarbeit beginnt denn typis-cherweise auch mit einer umfangreichen Analyse des vorhandenen Problems und der Bedurfnissedes Partners. Oft ist es dann wahrend der Zusammenarbeit zentral aufzuzeigen, wie unsere Simu-lationsresultate mit der Realitat verknupft sind. Dies erfolgt, indem wir z.B. mit Hilfe des Vergleichszwischen Experiment und Simulation, Parameter bestimmen sowie ein System beschreiben odermit Simulationen Wege zur Verbesserung der Systemeigenschaften aufzeigen. Auch wenn dieVerfugbarkeit von schnellen Rechnern und benutzerfreundlicher Simulationssoftware weiter zu-nimmt, stellen wir fest, dass dies alleine z.B. einem Industrieingenieur nicht genugt um eigenhandigmit Multiphysiksimulationen Fortschritte zu erzielen. In der Regel ist einschlagige Erfahrung in derBeschreibung physikalischer Phanomene auf verschiedenen Langenskalen unabdingbar. Es gibtzu viele Beispiele, bei denen in der Industrie oder an anderen Hochschulen Simulationssoftwarebeschafft wird, welche dann aufgrund von Schwierigkeiten in der Einarbeitung kaum weiter genutztwird. Es gibt also viele Anzeichen dafur, dass wir auch in Zukunft als Partner attraktiv bleiben. DieOpportunitatskosten fur den Fall, dass man der Simulation als Entwicklungsmethode fern bleibt,steigen weiter an.Um diesen Leistungsauftrag zu erfullen, durfen wir auf finanzielle Unterstutzung von verschiede-nen Forderagenturen und Forschungsstiftungen zahlen. Letztlich sind fur unsere Arbeit aberauch die Mitarbeiterinnen und Mitarbeiter zentral. Im Jahr 2008 durften wir mehrere neue Mi-tarbeiter am ICP begrussen, welche unsere Kompetenzen verstarken. Mit Roger Hausermannund Evelyne Huber sind zwei ETH-Diplomanden zu uns gestossen, welche an unserem Instituteine Doktorarbeit absolvieren und von Professoren der ETH Zurich begleitet werden. Unser Um-feld mit internationalen Forschungspartnern und -Themen und der Status als wissenschaftlicheAssistenten macht eine Dissertation an der ZHAW attraktiv. Mit Dr. Yasser Safa und Dr. MatthiasSchmid sind zudem zwei Wissenschafter von der ETH Lausanne zu uns gestossen. Die Attrak-tivitat einer Tatigkeit an unserer Schule bzw. unserem Institut wurde auch erhoht mit dem imHerbst 2008 lancierten Masterprogramm MSE (Master of Science in Engineering), das es denMaster-Studenten ermoglicht, neben dem Unterrichtspensum in aktuellen Forschungsprojektenmitzuarbeiten. Dies ist denn auch ein Unterscheidungsmerkmal zu den Masterprogrammen anuniversitaren Hochschulen. Die ersten Erfahrungen mit dem MSE Masterprogramm zeigen, dass

5

Page 6: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

dies die Studenten wie auch die Dozenten gleichwohl fordert. Der Regelabschluss bleibt jedochweiterhin der Bachelor und nur die bestqualifizierten Studenten werden im MSE Masterprogrammaufgenommen. Im Bachelor Unterricht waren Dozenten unseres Instituts in Modulen zur Physik,Mathematik, Signalverarbeitung sowie Thermodynamik tatig.Unsere Sichtbarkeit als Institut in der Schweiz haben wir im Jahr 2008 auch mit verschiedenenTeilnahmen an Messen und Tagungen gepflegt, so zum Beispiel am Winterthurer Technologietagim Fruhling sowie der Tagung der Electrosuisse zum Thema MEMS im Herbst. Am lokalen Anlasszum nationalen Tag der Technik und dem Thema Energieeffizienz referierten zwei ICP Dozen-ten, Thomas Hocker und Nils Reinke, zur Brennstoffzellen- bzw. Beleuchtungstechnologie. An-fang 2008 hat Jurgen Schumacher das ”5th Symposium on Fuel Cell Modelling and Experimen-tal Validation” an der ZHAW organisiert. Dies ist ein Beweis fur die internationale Anerkennungunserer Forschung auf dem Gebiet der Brennstoffzellenmodellierung und hat dem ICP weitereImpulse verliehen. Im Jahr 2008 gelang uns die Lancierung zweier europaischer Projekte ”AE-VIOM” und ”Apollo” im Bereich der organischen Elektronik, eines davon finanziert durch das 7.Forschungsrahmenprogramm der EU.Thematisch haben wir im vergangenen Jahr die Energieforschung deutlich verstarkt. So sindzwei grossere Photovoltaik-Forschungsprojekte angelaufen und auch das OLED EU-Projekt istauf Energieeffizienz ausgerichtet. Auch ein Forschungsprojekt zum Thema Holzvergasung fur diedezentrale Kraft-Warmekopplung wurde am Ende des Jahres bewilligt. Somit leisten wir mit diesenProjekten einen Beitrag zu einem gesellschaftlich hochst relevanten Themengebiet.Erwahnenswert ist auch der Aufbau des Mess- und Validierungslabors am ICP (kurz ”O-LAB”).Dank diesem Labor konnen wir kunftig vermehrt eigene Messungen zur Validierung von Simula-tionsresultaten ausfuhren oder sogar einen Beitrag in der Entwicklung von Messtechnik leisten.Letztlich ist dieses Labor auch bestens geeignet um Studentenarbeiten auf Stufe Bachelor undMaster im Bereich Mechatronik, Elektrotechnik und Informatik auszufuhren.Dass unsere Forschung anwendungsorientiert ist, zeigt nicht zuletzt auch der bisherige Erfolg derSpin-off Firma Fluxim AG, welche im vergangenen Jahr mit der Anstellung von Mitarbeitern und derEinmietung von Buroplatzen am ICP erste Schritte in die Unabhangigkeit unternommen hat. Mitder Kommerzialisierung der Software SETFOS macht diese Firma einen Teil der Forschungsresul-tate des ICPs einer internationalen Gruppe von Anwendern zuganglich. Die Anerkennung dieserSpin-off Firma auf nationaler und internationaler Ebene wurde bezeugt durch die Einladung zumTransferkolleg 2008 ”Industrial Photonics” der SATW (Schweizerische Akademie der TechnischenWissenschaften) und KTI (Kommission fur Technologie und Innovation) sowie durch die Teilnahmeam EU Projekt AEVIOM unter Leitung von Philips Research in Eindhoven. Die Zusammenarbeitunseres Instituts mit Spin-off Firmen hat sich auch mit Numerical Modelling GmbH erneut bewahrt,welche im vergangenen Jahr gleich zwei neue KTI Projekte mit dem ICP lancierte. Numerical Mod-elling GmbH ist weiterhin ein strategisch wichtiger Partner.Bei einem Blick ins kommende Jahr mussen wir davon ausgehen, dass sich die Finanzkrise auchin der Realwirtschaft auswirkt. Ob sich das auch in unserer Forschungstatigkeit bemerkbar macht,ist jedoch unklar. Es besteht sogar die Aussicht auf vermehrte Industrieforschungsprojekte, daFirmen in schwieriger wirtschaftlicher Situation gut beraten sind, vorhandenes Know-how und Per-sonalresourcen von Hochschulen einzubinden. Der Beschluss des Bundesrates anfang des neuenJahres zur Erhohung des Budgets 2009 der KTI ist diesbezuglich ein deutliches Signal.An dieser Stelle mochte ich im Namen des ICP Teams der School of Engineering der ZHAW fur dieUnterstutzung und insbesondere unseren Forderagenturen fur die erfolgreiche Zusammenarbeitdanken. Ich freue mich, die genannten Herausforderungen gemeinsam im ICP Team anzupackenund mochte mich bei jedem einzelnen ICP Mitglied fur den personlichen Einsatz im vergangenenJahr bedanken.

Beat RuhstallerInstitutsleiter

ZHAW 6

Page 7: ZHAW ICP Research Report 2008

Chapter 2

Introduction

The Institute of Computational Physics has lived up to its name in 2008 with respect to its activitieson national and international level. Several research projects in the field of numerical model-ing of multiphysics systems were concluded and new ones launched. We enjoyed tackling manychallenges in our R&D activities together with academic and industrial partners. At the bound-ary between academic and applied research we are facing many challenges on a daily basis.How can we generate value to our research partners with our numerical models and simulations?The answer is not straightforward. On the one hand we must earn the trust of our partners anddemonstrate that our calculations indeed help gaining insights which could hardly be obtained byexperimental means. On the other hand we must explain convincingly, that we do not performnumerical simulations just for the fun of it but rather because they are the most efficient way ofachieving the objectives. Thus a collaboration typically starts with an analysis of the problemsat hand and the needs of the partners. Generally it is crucial to demonstrate how the simulationresults are connected to reality. This happens with the help of comparisons between experimentand simulation, parameter extraction or by formulating possible ways of improvement of the sys-tem properties. Although the availability of fast computers and user-friendly simulation software isfurther advancing, we can state that this alone won’t allow an engineer to get up to speed with mul-tiphysics simulations. In general, appropriate experience is indispensable. There exist too manyexamples of simulation software purchases in industry and at universities, in which the software ishardly used any longer after initial training difficulties. Thus there are numerous signs that we willremain attractive partners also in the future. The opportunity cost for the case in which simulationsoftware is not taken advantage of is large.In order to fulfill our mission we count on financial support of various research funding agencies.But after all, our staff members are crucial. In 2008 we welcomed new staff members whichstrengthen our competences. Roger Hausermann and Evelyne Huber are two ETH graduatesthat joined us for writing a Ph.D. thesis at our institute and in collaboration with professors ofETH Zurich. Our work environment with international research partners and topics and the statusas scientific assistant makes a Ph.D. thesis at the ZHAW attractive. In addition, two scientistsfrom ETH Lausanne joined us, Dr. Yasser Safa and Dr. Matthias Schmid. The attractivity of theworking environment at our school and institute was increased also with the launch of the Masterof Science in Engineering (MSE) program in Fall 2008. It allows the students to get practicaltraining in research projects besides attending classes. This is a differentiation from the otheracademic master programs. The first experiences with the MSE show that it is demanding for boththe students and involved lecturers. However, the bachelor degree will remain the standard degreeat our school and only the best bachelor graduates are accepted as master students.Lecturers of the ICP were teaching courses in Physics, Mathematics, Signal Processing and Ther-modynamics this year. The visibility of our institute was maintained with the attendance of variousconferences and symposia. For instance, the Winterthur technology day in Spring and the Elec-trosuisse MEMS meeting in fall. At the local event of the national day of technology with its focuson energy efficiency two ICP lecturers, Thomas Hocker and Nils Reinke, gave presentations inthe field of fuel cell and lighting research. Moreover, Jurgen Schumacher organized the 5th Sym-posium on Fuel Cell Modelling and Experimental Validation (ModVal) at the beginning of 2008 in

7

Page 8: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

Winterthur. This is one proof of the international recognition of our fuel cell research. The sameyear we succeeded in launching two European resarch projects (AEVIOM and Apollo) in the field oforganic electronics, one of which financed by the seventh framework programme of the EuropeanUnion.In view of the research topics, we have strengthened our research activities in energy. Two largerresearch projects in photovoltaics as well as the OLED EU project are targeting energy efficiency.Moreover, a project in wood power was approved at the end of the year. Therefore we contributein part to this societally highly relevant topic.We are also proud to have set up a laboratory for measurement and validation in the past year.Thanks to this laboratory we will be able to carry out some of the measurements for validation ofthe simulation results on our own. We may even contribute to the development of suitable mea-surement techniques. This laboratory is also ideally suited for student thesis projects at bachelorand master level in the fields of mechatronics, electrical engineering and computational science.The application-oriented nature of our research is proven by the success of the spin-off companyFluxim Inc., which has undertaken several steps towards an independent business by hiring acouple of staff members and renting office space at the ICP. With the commercialization of thesoftware SETFOS, this company gives an international user community access to some of theresearch results of the ICP. The recognition of this spin-off company was proven by the invitationto the Transferkolleg ”Industrial Photonics” by the SATW and the CTI in November 2008 as wellas the participation in the EU project AEVIOM under leadership of Philips Research in Eindhoven.The ICP collaboration with spin-off companies has proven valueable also in case of NumericalModeling GmbH, which launched two CTI projects with the ICP. Numerical Modeling GmbH thuscontinues to be a strategic partner of the ICP.Looking ahead into the coming year, we must assume that the financial crisis will also affect realeconomy. It is unclear, however, whether this is noticeable also in our research activities. There iseven a chance for more research projects with industry, since companies are advised to considerengaging more know-how and human resources of the universities. The decision of the federalcouncil to increase the budget for the commission for technology and innovation (CTI) for 2009 isa clear signal.On behalf of the ICP team I would like to express our gratitute to the School of Engineering forsupporting us and the research funding agencies for the successful collaboration. I am lookingforward to tackle the above-mentioned challenges together with our team and would like to thankeach ICP team member for her/his commitment in 2008.

Beat RuhstallerHead of the institute

ZHAW 8

Page 9: ZHAW ICP Research Report 2008

Chapter 3

Sensors and Actuators

9

Page 10: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

3.1 Finite element software for the industry

Contributors: Guido Sartoris, Hansueli Schwarzenbach, Markus Roos

Partners: ZHAW, Numerical Modelling GmbHFunding: Commission for Technology and Innovation (CTI)Duration: 2007 – 2009

The goal of this project is to provide the indus-trial partner NM Numerical modeling with upto date software for the modeling of Maxwell’sequations in the low frequency regime wheredisplacement currents may be neglected, i.e.the eddy current model. Of particular interestfor SMEs is the modeling of magnetic proximitysensors with common PC hardware and com-putational time limitations of a few hours. AtICP we are developing our in-house softwareSESES, a general purpose multi-physics nu-merical simulation tool for the analysis of micro–macro, sensor–actuator devices by the finite el-ement method in 2D and 3D domains. Its ma-jor strengh is the possibility to solve and cou-ple together almost any governing equation ofclassical physics. In order to greatly improvethe flexibility of the software, we have developeda functional input language allowing the user todefine within the input generic material laws andcoupling effects between the different physicalfields. Although SESES may already be usedto compute some eddy current problems, theircomputation in 3D is far from being optimal andefficient and one aim of this project is to improvethis situation with up to date numerical algo-rithms. In particular, the multi-physics conceptbehind SESES allows to quickly couple togetherthese new models with the available ones andfor example to perform optimized computationsby mixing up vector and scalar potential formu-lations which are coupled together along theircommon boundary. On the other side, sinceeven optimized 3D eddy current computationsare still computationally expensive, from the in-dustrial partner there is a need to know if re-duced and less demanding models may be usedto replace fully fledged 3D computations. Sucha comparison is shown in Fig. 1. The new nu-merical models will be used to validate such ap-proximations and this part of the work will bedone by the industrial partner.From the software point of view which is the con-tribution from ZHAW to the project, the majorwork is the introduction of edge finite elementswithin SESES. These elements are the natural

framework when working with the ∇ × ∇ oper-ator since they do not enforce the normal com-ponent of the vector field to be continuous anddifferently from Lagrangian or nodal elementsare well suited to model sharp corners andsharp discontinuous material laws. However,the eddy current problem cannot be gauged withedge elements and the linear system is singu-lar. If the equations are assembled in a con-sistent manner, this is not a drawback for it-erative solvers and here edge elements showin general a better convergence behavior thannodal elements. However, the good conver-gence is lost at middle-high frequencies or canbe destroyed by using optimizations to reducethe number of unknowns, e.g. by using a mixedformulation with the magnetic potential togetherwith boundary interface conditions, a cotree dof-reduction, a symmetrization or removal of theelectric potential in the conductors. A two matri-ces approach may be useful to overcome theformer difficulty at high frequencies, whereasthe other optimizations must be combined withmatrix preconditioning and are just used if theydo not globally impair the computational expen-diture. The greatest unknown is the conver-gence speed of the iterative solver and numer-ical investigations on the preconditioning’s per-formance are required here. These numericalalgorithms are pivotal for the overall success.

-1.20e-04

-1.00e-04

-8.00e-05

-6.00e-05

-4.00e-05

-2.00e-05

0.00e+00

2.00e-05

0 2 4 6 8 100.00e+00

2.00e-05

4.00e-05

6.00e-05

8.00e-05

1.00e-04

1.20e-04

Res

ista

nce

R01

(Ohm

) Reactance X01 (O

hm)

Frequency (kHz)

R01 R01-approx

X01 X01-approx

Fig. 1: Reactance matrix coefficients for a system oftwo coils computed by the eddy current model and asecond order low frequency approximation.

ZHAW 10

Page 11: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

3.2 Numerical modeling of laminates anisotropic plastic deep-drawing

Contributors: Guido Sartoris

Partners: Alcan Technology and Management AGFunding: Commission for Technology and Innovation (CTI)Duration: 2008 – 2009

Modern laminates for the production of pharma-ceutical blister packages have a complex innerstructure. Several layers of different plastic andalumina sheets of thickness between 5−100µmare packed, stacked and glued together with theintent to provide the best environmental protec-tion from external agents. Each layer plays adifferent but important role, as for example, alu-minum protects against light, oxygen and wa-ter vapor. When forming the blisters by deep-drawing, the overall functional role of the layersin the original laminate should be preserved andno holes, cracks or layer delamination are al-lowed to evolve during deformation. This is oneof the major concern of Alcan, which providesits customers with laminates for the produc-tion of pharmaceutical packaging blisters andwhere the mechanical properties of the laminatemust be guaranteed. Illustrations of blisters areshown in Fig. 1. One may try to get phys-ical insight on the disruptive processes of thesandwiched layers by running complex, time-consuming experiments, but they do not neces-sarily yield ultimate and correct answers on thecauses of failure. As well the material proper-ties of the single layers have great influence onthe overall results and there are large variationsbetween production cycles which complicatesthe analysis further. Therefore a numerical ap-proach to laminate deep-drawing is almost a ne-cessity. Numerical modeling is expected to pro-vide a better and more detailed insight of thedeformation processes with respect to a solelyexperimental approach. At the present time, thenumerical models used at Alcan are not able tocorrectly reproduce experimental data and moreadvanced and precise tools are required. This isthe major goal of the project. The task is com-plex and includes both the writing and fitting ofnew material laws and optimizing the numericalsoftware.From the material point of view, the major con-cern is the correct description of the anisotropicelasto-plastic behavior at large strains, since atthe present time, a theoretical definitive treat-ment is not available and several formulationshave emerged.

Our intent is to use the more physically soundedmultiplicative approach of plasticity instead ofthe additive one. The intent is to write genericanisotropic elasto-plastic laws for orthotropicmaterials at finite strains to be fitted with mea-surements performed at Alcan. This task iscompletely independent from the underlying nu-merical tools and for the in-house developedfinite element tool SESES, can be performedcompletely in user space, i.e. without the needto compile any software code. This is also truefor the curve fitting process based on simple ex-periments at constant strain values not requiringnumerical simulations.Because of the anisotropic nature of the de-formation process, a 3D numerical non-linearanalysis at large strains must be generally per-formed. These computations are still very ex-pensive and the second goal of the project isto provide various optimizations at the codinglevel to speed-up the numerical analysis. Heregeneric non-linear volume elements are too ex-pensive and non-linear shell elements support-ing the same 3D material laws should be pre-ferred. Forseen is also a mixing of both shelland volume elements by providing special inter-connecting conditions. It is therefore possibleto model the deep-drawing process on regionsof small radii with the more precise volume ele-ments and otherwise to use the less expensiveshell elements.

Fig. 1: Blisters formed by deep-drawing of a laminate.

ZHAW 11

Page 12: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

3.3 Beruhrungsloses Signalanalyseverfahren fur Schichtsys-tem

Contributors: Nils A. Reinke, Andreas Tiefenauer, Beat Ruhstaller, Andor Bariska (IDP)

Partners: Flo-IR und weitere 5 IndustriepartnerFunding: Kommission fur Technologie und Innovation (KTI)Duration: 2007 – 2009

Wird nach einem optischen Energiepulsdie IR-Abstrahlung von einer beschichtetenOberflache zeitlich, ortlich und spektralgerastert gemessen, lassen sich Ruckschlusseauf die Schichtdicken oder auf thermo-physikalische Stoffdaten ziehen.Auf der Grundlage dieser Technik wird einAuswertealgorithmus fur die Messdatenanalysesowie ein robuster Messprototyp entwickelt. Diefolgenden beiden ZHAW Institute ubernehmendabei den Kernteil dieser Entwicklung in folgen-der Arbeitsteilung:

• IDP: Entwicklung eines Auswerteal-gorithmus: Aus Messdaten einesMehrschichtensystems, sollen mittelseines physikalisch analytischen ModellsParameter wie Schichtdicke, und thermis-che Materialeigenschaften extrahiert wer-den konnen.

• ICP: FEM-Simulationen: Simulation vonAnwendungssituationen; Entwicklung desMessprototypen: Es ist ein Messaufbau(Messhardware und Software) zu entwick-eln, der den hohen Anforderungen an dasSystem genugt (siehe unten).

In der Anfangsphase des Projekts wurdenam ICP FEM-Analysen von Anwendungsfallengemacht. Dazu wurde ein instationaresSchnittmodell der Schichtstruktur erstellt. Un-tersucht wurde beispielsweise der Einflussvon Semitranzparenz, Delamination oder auchKontaktwiderstanden. In der Hauptprojekt-phase wurde nach einer geeigneten Losungder Datenerfassung gesucht. Die Heraus-forderung lag darin, ein System zusammen-zustellen, welches die hohen Anforderungenan die Messdatenrate erfullt (> 10 kHz), einehohe Empfindlichkeit im IR-Bereich besitzt, unduber geeignete Moglichkeiten zur Ansteuerunguber einen PC besitzt. Ferner sollten dieDatenaufnahme der Kamera, sowie auch dasAuslosen des Blitzes uber eine Software koor-diniert angesteuert werden konnen. Der Proto-typenaufbau entspricht weitgehend dem in Fig.

1 abgebildeten Schema.

Fig. 1: Messaufbau am ICP bestehend ausJustierungseinheit mit Probe, der datenerfassendenEinheit (Kamera, U/IKonverter, Photodiode) und derdatengerenierenden Einheit (Kamera, Blitz, Pul-skonverter, Kondensator) sowie der Steuereinheit(Datanaqusisationsgerat (DAQ), Messsoftware).

Fig. 2: LabView basierte Software zur Ansteuerungdes Blitzes und Datenaufnahme inkl einfachemDatenmanagement.

Mit diesem Messaufbau wurden im ICP-eigenenMesslabor verschiedene Messreihen an denvon Industriepartnern bereitgestellten Probendurchgefuhrt, welche der Entwicklung desAuswertealgorithmus am IDP dienen.

ZHAW 12

Page 13: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

3.4 Velocimetry in railcars by a run-time measurement of ther-mal markers

Contributors: Nils A. Reinke, Andreas Tiefenauer, Beat Ruhstaller, Andor Bariska

Partners: Siemens TSFunding: Commission for Technology and Innovation (CTI)Duration: 2008 – 2009

State-of-the-art velocimetry using the Dopplermethod, dynamic image analysis and global po-sitioning systems suffers from unreliabilities to agreater or lesser extent. This disaffection arisesfrom a liability to interference towards environ-mental influences. Leaves, snowfall, rainfall andartificial lighting (light pollution) may lead to se-vere distortions in the measured velocity.This feasibility project verifies a patent pend-ing velocimetry technique which is based ona run-time measurement of thermal markers.The thermal-marker is applied from the railcarby optical excitation and subsequently photo-thermally detected by an infrared sensor. Thevelocity of the railcar can be determined by mea-suring the run-time from excitation to detectionand the spatial distance between excitation anddetection setup.

Fig. 1: Corroded steel disc spinning and a germaniumzoom lens.

In recent years significant progress has beenmade in developing powerful, compact and in-expensive solid-state lasers. Nowadays, solid-state lasers replace traditional gas-lasers due totheir higher reliability and low maintenance cost.

At the same time, fast and inexpensive sensorsfor the mid-infrared regime become more andmore available. Improvements in laser tech-nique and MIR sensors enable to realize this in-novative velocimetry technique.This project subdivides into three sequentialsteps:

• Computational modeling for estimating theamount of light necessary to deposit thethermal marker with different optical exci-tation sources (Laser, LED, flashlamps).The optical deposition of the thermalmarker leads to a transient temperaturedistribution inside the rail. In order to de-tect the deposited thermal marker by theinfra-red sensor, the amplitude of this tem-perature distribution has to be sufficientlyhigh.

• An experimental setup has to be devel-oped from scratch. This experimentalsetup has to reflect conditions close to re-ality, e.g. the optical and thermal proper-ties of the rail. Computational modelingwill help to specify the individual compo-nents used in this experiment. The exper-imental setup will be controlled and dataacquired by LabView.

• In a final step of the project test measure-ments will be performed and analyzed toverify results from computational simula-tions and to proof the concept of the pend-ing patent.

A prototype experimental setup comprising a fix-ture for a spinning steel disc and an infra-reddetector is shown in fig. 1. Results of this feasi-bility project will be further investigated in a fieldtest, if the third step turns out to be successful.

ZHAW 13

Page 14: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

3.5 Diagnostischer Atemanalysator mit elektronischer Mikro-nase

Contributors: Hansueli Schwarzenbach, Roman Gmur

Partners: Interstaatliche Hochschule fur Technik Buchs, Hochschule Furtwangen,Universitatsklinik Innsbruck, Forschungszentrum Karlsruhe, SYSCA AG, GDS

Funding: Internationale BodenseehochschuleDuration: 2007 – 2009

Das auf zwei Jahre angelegte Verbundpro-jekt namens DAMINA beinhaltet die Mach-barkeit eines kostengunstigen, handlichenAtemgasanalysesystems fur diagnostischeZwecke1. Es basiert auf einem Gassen-sormikrochip, der nach dem Prinzip der elek-trischen Leitfahigkeitsmessung mit Metalloxid-Gassensoren die Atemgaszusammensetzungmisst. Es unterscheidet sich von teuren,konventionellen, Grossgerateanalysesystemen(Gaschromatographen, Massenspektrometer)durch seine Handlichkeit und ist in Arztpraxenzum kostengunstigen Einsatz ohne Spezialaus-bildung gedacht.

Fig. 1: Behalter Gasaufbereitung: A: Ventil, B: Ver-tikales Zuflussrohr Atemsack zum Messzylinder, C:Zuflussrohr Referenzgas zum Atemsack (Reinigung),D: Behalterheizung (vgl. Fig. 2)

Fig. 2: Behalterheizung (an Ruckwand am Behalterzur Gasaufbereitung montiert). A: Lufter mit Heizung,B: Thermostat, C: Netzteil.

Das Projekt umfasste die Herstellung von Funk-tionsmustern zur Diagnose von Lungenkarzi-nomen. Bei Feldtests stellte sich heraus,dass die Nachweisgrenze des heutigen Sys-tems fur Formaldehyd bei 1 ppm in Atem-gas liegt. Damit ist die Zielsetzung um einenFaktor 100 nicht erreicht. Zwar konnte dernachteilige Temperaturdrift der Sensorsignaledurch ein verandertes Auswerteverfahren un-terdruckt werden. Die Langzeitstabilitat des Mi-crochip ist heute dadurch wesentlich besser alsvor 2 Jahren. Aus heutiger Sicht mussten alter-native Messverfahren bezuglich der verlangtenNachweisgrenze fur Formaldehyd und fur an-dere Gaskomponenten evaluiert werden. Diesubersteigt die Moglichkeiten von SYSCA hin-sichtlich einer Produkteentwicklung bei weitem.Auf der wissenschaftlichen Seite gibt es seitdem Hinschied von Kollege Goschnick amFZK im vergangenen Jahr keine weiterenMoglichkeiten fur experimentelle Untersuchun-gen. Aus diesen Grunden wurde das Projektvorzeitig gestoppt.

1ICP Research Report 2007

ZHAW 14

Page 15: ZHAW ICP Research Report 2008

Chapter 4

Fuel Cells

15

Page 16: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

4.1 Modeling PEM fuel cell performance with a focus on porouslayer properties

Contributors: Yasser Safa, Jurgen O. Schumacher

Partners: Paul Scherrer InstituteFunding: Swiss Federal Office of EnergyDuration: 2007 – 2010

Proton exchange membrane (PEM) fuel cellsare promising clean energy converting deviceswith high conversion efficiency, high power den-sity and low to zero emissions.Before reaching the full potential of this technol-ogy, some challenging tasks need to be over-come. In particular, the optimal control of thewater flow in the fuel cell appears as a key issuein the design of fuel cell devices.Under condensing conditions (the generalcase), liquid water will occupy part of the voidin the porous structures of the fuel cell, and thuschange the gas transport properties.Actually, the water saturation inside the PEMfuel cell should be low enough in order to notblock the porous Gas Diffusion Layer (GDL). Inother words, the flow rate of water should behigh enough to ensure the release of the pro-duced water from the fuel cell. On the otherhand, water in the cell is essential to main-tain a high protonic conductivity of the pro-ton exchange membrane. Consequently, for asuccessful water management in the PEM fuelcell, an accurate insight in the water distribu-tion at different operating conditions is inevitablyneeded.

Fig. 1: Operation of a PEM fuel cell, the Gas Diffusionlayer (GDL) is a thin material (fibers) located between

the Flow Plate and the Catalyst Layer 1

Fig. 2: The water distribution in a sample part of aGas Diffusion Layer (GDL) presented in side view.The red color corresponds to the fibers and the bluecolor corresponds to the water. 2

Fig. 3: The water distribution (blue color) in a samplepart of a GDL presented in top view.

A combination of the measurement techniquesthat are applied at the Paul Scherrer InstitutePSI and the numerical model that is being de-veloped at the ICP is a practicable way to inves-tigate the two-phase phenomena in a PEM fuelcell.In this project, we are developing an advancednumerical two-phase flow model to describe thewater transport in PEM fuel cells. Roughlyspeaking, we are applying the so called Multi-phase Mixture (M2) formalism to describe thewater transport. This formulation allows to in-clude the effects of vapor bubbles in the liquid

1www1.eere.energy.gov/hydrogenandfuelcells/animation/mod1.html2Tomography performed at PSI, Switzerland

ZHAW 16

Page 17: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

water flow as well as the effects of liquid dropletsin the water vapor flow.Compared to the traditional multi-phase flowmodel, different phases are considered as a sin-gle mixture fluid. The M2 formalism is based onmixture variables for velocity, pressure and othervalues as viscosity, mobility and permeability.Mathematically, we solve a time-dependentmodel that is composed of a convection-diffusion equation for the water saturation andof the Darcy-law with incompressibility conditionfor the fluid flow.Due to a degeneracy of the equations, the trans-port coefficients are perturbed to obtain a non-degenerate problem with a smooth solution.The regularized solution should converge to theoriginal as the perturbation parameter goes tozero with a specific convergence order.To deal with space discretization, the finite ele-ment approximation is used for solving the regu-larized problem. It is combined with a stabilizedmixed finite element method for solving simulta-neously for velocity, pressure and water satura-tion.Also, due to its convection-dominated feature,the stabilization tool SUPG (Streamline UpwindPetrov Galerkin) is applied when solving the sat-uration equation.The time discretization of the global system isapplied via an implicit scheme. Since the timescale of the total flow is longer than that of thewater saturation evolution, we are choosing tointroduce a multi-time step discretization in ourapproach.In Figures (4-5) we present samples of numer-ical results on the liquid water distribution andthe total pressure along the thickness directionof the GDL for an academic test case. The nu-merical simulations are achieved by developinga Mathematica code.

50 100 150 200 250 300z µm

30

40

50

60

70 Sw

Liquid Saturation

Fig. 4: Numerical results on the liquid water distribu-tion in the thickness direction of the GDL.

50 100 150 200 250 300z µm

99 200

99 400

99 600

99 800

100 000P Pa

Total Pressure

Fig. 5: Numerical results of the total pressure drop inthe thickness direction of the GDL.

20 40 60 80 100Sw

40 000

30 000

20 000

10 000

Pc Pa

Capillary Pressure

Fig. 6: The capillary pressure as a function of theliquid water saturation according to the Brooks-Coreymodel.

A change of the slope of the water saturationprofile is observed in Figure 4 around the valueof Sw = 50 %. The capillary effects (Figure 6)appear as a new contribution on the pressuredriven flow phenomena. This has an impact onthe distribution of the water saturation.

Finally, the introduction of the measurement re-sults in the computational work is the future taskdevoted for an advanced technical exploitationof the code. The local water saturation in theGas Diffusion Layers is determined experimen-tally at the Paul Scherrer Institute using x-raymicro-tomography. After phase separation ofthe image data, the water distribution inside theporous structure could be resolved on the scaleof the pores. This enabled the measurement ofthe liquid saturation as function of the through-plane coordinate of the GDL and the capillarypressure. First experiments with complete cellsare also performed. A comparison of the spa-tially distributed data of water saturation to thesimulation results is an upcoming crucial task.

ZHAW 17

Page 18: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

4.2 Simulation of polymer-electrolyte-membrane fuel cells andstacks

Contributors: Jurgen O. Schumacher, Jens Eller, Guido Sartoris

Partners: Paul Scherrer InstituteFunding: Gebert Ruef StiftungDuration: 2005 – 2008

We developed a computationally efficient modelof a proton exchange membrane (PEM) fuelcell. In the framework of this ”2+1D”-approachthe anodic and cathodic fields are discretizedin two dimensions with our finite element soft-ware SESES. The coupling between the an-odic and cathodic side is established by a one-dimensional model representing the gas diffu-sion layer (GDL) and the membrane electrodeassembly (MEA).This approach is suitable for taking the high as-pect ratio between the in-plane and the through-plane dimensions of fuel cells into account. Thenumber of degrees-of-freedom (DOF) variablesis significantly reduced in comparison to a full3D model approach. Therefore, parameter vari-ations for design studies of PEM fuel cell canbe performed on standard PCs. Further, thismodeling approach can also be applied to otherfields, e.g. for the description of coupled elec-trochemical processes.The gas flow channels and bipolar plates of theanode and cathode are discretized in two di-mensions. Coupling between the anodic andcathodic side is established by a 1D modelrepresenting the membrane electrode assembly

(MEA). Coupling between the 1D model of theMEA and the 2D models is achieved by usingthe values of the DOF variables of the 2D modelas boundary conditions of the 1D model.The 2+1D FEM model was applied to simulatea simplified micro polymer electrolyte fuel cellwithout gas diffusion layers (see Fig. 1). Thegas flow-field of this fuel cell is made of micro-structured glassy carbon plates. The electriccurrent density distribution in the through-planedirection is shown in the figure. This currentdensity is predicted by the 1D model, that is,in the direction perpendicular to the cell area.The plot corresponds to an average current den-sity of 250 mA/cm2. The current density is low-est near the gas inlet (top right), it increasesto a maximum value in the cell center, and de-creases towards the gas outlet (top left). Theminimum value near the gas inlet is due to alow value of the membrane water content, i.e.the protonic conductivity in this region is lowesthere. An increase of the membrane water con-tent was found in the cell center.The model approach is suitable to analyze dif-ferent operation scenarios of proton exchangemembrane fuel cells.

Fig. 1: Electric current density of a micro PEM fuel cell in the through-plane direction (left: anode, right: cath-ode).

ZHAW 18

Page 19: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

4.3 Enhancing the lifetime of SOFC stacks for combined heatand power applications

Contributors: Thomas Hocker, Rolf Weiss (ITFE)

Partners: Hexis AG (Winterthur), EPFL Laboratoire d’Energetique IndustrielleFunding: Swiss Federal Office of Energy, SwissElectric ResearchDuration: 2007 – 2010

Within 2008, our focus has been on modelingthe thermo-mechanical and thermo-fluidic be-havior of the overall system as well as on elec-trochemical micro-models for predicting the per-formance and the degradation behavior of singleSOFCs.

Fig. 1: Typical temperature-distribution within theHexis SOFC-system as obtained from ANSYS-CFXthermal-flow modeling.

Together with our project partner ITFE, i. e.ZHAW’s Institute of Thermo-Fluidic Engineer-ing, a 3D thermal-flow model of the Hexis sys-tem has been implemented with the commercialCFD-software ANYS-CFX. The model consistsof the stack, the heat exchanger, the reformer,and the thermal insulation, see Figure 1. Heattransfer by conduction, convection and ther-mal radiation (surface-to-surface and gas radia-tion with the post-combustion zones) have beenconsidered. Chemical and electrochemical re-actions are not explicitly modeled, but the cor-responding heat-sources and -sinks (in case ofmethane steam reforming) have been taken intoaccount. A typical simulation-run takes about24 hours on a 12-CPU Linux cluster with 32 GBof RAM. Results show that the predicted stacktemperatures are in good quantitative agree-ment with experimental results. The model de-velopment has been almost finished. A more re-alistic implementation of gas radiation (for CO2and H2O molecules as well as soot particles) to

accurately predict the heat transfer between thepost-combustion zones and the stack remainsto be included.

0.0 0.2 0.4 0.6 0.8 1.0

x

a

0.005

0.010

0.015

0.020

0.025

0.030

0.035

Η

Ueq

0.2 0.4 0.6 0.8 1.0

x

a

500

1000

1500

2000

2500

3000

i (A/ m2)

iio

iel

a=10 Μm, ASR(tot , act , ohm )=(112, 81, 31) m� cm2

ASRtot = 112 m! cm2

ASRact = 72%

ASRohm = 28%

Reaction-limitation in 10 !m electrode

electrode-location

curr

en

t-d

en

siti

es

electrode-location

ov

er-

po

ten

tia

l

0.0 0.2 0.4 0.6 0.8 1.0

x

a

0.005

0.010

0.015

0.020

0.025

0.030

0.035

Η

Ueq

0.2 0.4 0.6 0.8 1.0

x

a

500

1000

1500

2000

2500

3000

i (A/ m2)

iio

iel

a=50 Μm, ASR(tot , act , ohm )=(91, 16, 75) m� cm2

ASRtot = 91 m! cm2

ASRact = 18%

ASRohm = 82%

Transport-limitation in 50 !m electrode

curr

en

t-d

en

siti

es

electrode-locationelectrode-location

ov

er-

po

ten

tia

l

Fig. 2: Over-potentials predicted from the Costam-agna electrode model for two limiting cases: elec-trode performance dominated by (a) reaction limita-tion, and (b) transport limitation.

The model by P. Costamagna et al.3 for mixedion-/electron conducting electrodes has beenimplemented into Mathematica, see Figure 2.Comparisons of the model predictions with ex-perimental data have been made. They showgood qualitative agreement and provide insightinto the influence of particle coarsening on theelectrode performance.

3P. Costamagna et al., Electrochemical Acta, vol.43, pp. 375–394, 1998.

ZHAW 19

Page 20: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

4.4 Numerical simulation for the development of the Hexis SOFC-system

Contributors: Yasser Safa, Thomas Hocker

Partners: Hexis AG (Winterthur), Industrial Energy Systems Laboratory, EPFL (Lausanne)Funding: Swiss Federal Office of Energy, SwissElectric ResearchDuration: 2007 – 2010

The Solid Oxide Fuel Cell (SOFC) is a highlyefficient producer of heat and power. As anelectro-chemical device it consists of two porouselectrodes, separated by a solid electrolyte.The oxygen molecule accepts an electron fromthe cathode and then an oxygen ion migratesthrough the electrolyte to the anode where it re-acts with fuel. The electrons are then releasedand conducted through an external circuit to thecathode and thereby creating the electric cur-rent (Fig. 1-2).The concept of SOFC heater “Galileo System”,established by our partner Hexis AG, is de-voted to cover the essential power and heatrequirement for a representative Central Euro-pean single-family home.A challenge task is to ensure a long lifetimeservice of Galileo with satisfactory performance.Therefore, an investigation of the long termdegradation mechanism is highly needed.Due to its novelty, little is known about thedegradation behavior of the Hexis SOFC systemand an experimental investigation may prove tooexpensive. In this respect, numerical simula-tion appears as an inexpensive tool to providea computational prediction on the failure modesrelated to either electro-chemical or thermo-mechanical phenomena.In this project we develop an advanced compu-tational reactive flow modelling of the coupledtransport phenomena in SOFC.

Fig. 1: Electro-chemical basic principle of SOFC 4

Fig. 2: SOFC Process (Courtesy HEXIS AG)

The physical model, that we study, is gov-erned by a system of nonlinear partial differen-tial equations describing the conservation lawfor mass, charges, energy and species. Thefluid flow in the porous media is described byDarcy’s Law.We apply the finite element approximation forthe space discretization of the continuity equa-tions. The convection dominant cases are sta-bilized by SUPG (Streamline Upwind PetrovGalerkin). To ensure the conservation proper-ties in the results when solving for Darcy’s prob-lem, we implement a stabilized mixed finite ele-ment method to solve simultaneously for veloc-ity and pressure.The mathematical system is linearized bymeans of a pseudo time step discretization viasemi-implicit scheme. This stable iterative pro-cedure is applied to a transient problem and itconvergences unconditionally to the steady so-lution.In order to overcome the geometrical complex-ities we start with a restriction in a representa-tive sub-domain located in the so-called Cell Re-peated Unit (Fig. 3).

4science.nasa.gov/headlines/y2003/18mar fuelcell.htm

ZHAW 20

Page 21: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

Fig. 3: A Cell Repeated Unit in a planar SOFC5.

For an advanced insight in the coupled fieldswe solve, firstly, the system as one-dimensional(1D) problem through the thickness of the CellRepeated Unit.Nevertheless, in Figures (4-7) we present sam-ples of numerical results on the gas-phasespecies distribution along the thickness direc-tion in the anode for an academic test case. Thenumerical simulations are achieved by develop-ing a Mathematica code.

5 10 15 20z µm

50.0

50.5

51.0

51.5

52.0

52.5

H2O

Fig. 4: Water vapor distribution along the thicknessdirection in the anode: the H2O molecules are pro-duced at the anode-electrolyte interface.

5 10 15 20z µm

24.0

24.5

25.0

25.5

26.0

H2

Fig. 5: Hydrogen distribution: the H2 molecules areconsumed by reaction with Oxygen ions.

5 10 15 20z µm

1.0

1.5

2.0

2.5

CO

Fig. 6: Carbon Monoxide distribution: the COmolecules are consumed by Water Shift Reaction(WSR): CO + H2O −→ H2 + CO2.

5 10 15 20z µm

6

7

8

9

CO2

Fig. 7: Carbon Dioxide distribution: the CO2

molecules are produced by water shift reaction.

Clearly, it is inevitable to account for an inter-action between the 1D computed fields and thetwo-dimensional (2D) in-plane flow in the gaschannel. For this purpose, we are choosingto implement a 2+1D numerical model for solv-ing the full problem with reduced computationalcost. Finally, for a realistic simulation moremimicking the in-situ observations, we intendto adapt our inputs to new experimental values(Micro Structural Data).

5Craig Fisher: http://people.bath.ac.uk/cf233/sofc.html

ZHAW 21

Page 22: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

4.5 Feasibility of a portable SOFC for Hilti’s mobile machinetools

Contributors: Roman Gmur, Thomas Hocker

Partners: ETHZ Institute of Nonmetallic Inorganic Materials, Hilti Research CenterFunding: HiltiDuration: 2007 – 2010

In contrast to conventional SOFCs, µ-SOFCsare characterized by membrane thicknesses ofonly a view micrometers and operation temper-atures as low as 500 ◦C. Feasibility studies ofsuch systems suggest electrical power outputsas low as 1 W. µ-SOFCs could be employedin combination with or as replacement of batter-ies to power small to medium electronic devices,such as laptop computers, camcorders, or ma-chine tools. Compared to energy storage de-vices such as Li-ion and NiMH batteries, higherenergy densities (per volume or weight) are pos-sible.This project investigates the feasibility of aportable SOFC for Hilti’s mobile machine tools.Having studied the current profiles of Hilti’s mo-bile machine tools and specifications for mobilepower sources, it is anticipated that a SOFCalone will not be sufficient for Hilti’s high currentpeak applications. Therefore, the SOFC has tobe coupled with an energy storage device suchas a battery. The SOFC will charge the storagebattery continuously, while the battery itself willdeliver peak power during operation of the elec-trical tool. The SOFC can be designed as anexternal charger or as an internal unit.

4

4

OneBat Specification Sheet

! Accurate and self-contained description of all factors relevant for system design

by a simple model based on global balances

! Checked against FE- and thermodynamic equilibrium calculations

! Fast and flexible design tool to analyse dimensions, mass-, energy flows, overall

efficiencies and applications

Specification InputSpecification Input

Specification Calculation

Spec Summary

1 + 2

Standard Input 1, 2.5, 5, 20 Wel Solid Material

SpecificationEPFL, LTNT, NMW, NTB

Insulation Mat.Specification

EPFL, LTNT, NMW, NTB

FluidSpecifications

General SystemSpecification

EPFL, LTNT, NMW, NTB

GPUSpecification

LTNT

Manual input

Click button

Auto

matic in

put

Au

tom

atic input

Spec Comparison

1 + 2

Fig. 1: Specification sheet for dimensioning theOneBat SOFC-system.

Figure 1 gives an overview of the newly de-veloped specification sheet that has been de-veloped for dimensioning the OneBat SOFC-system. It represents a accurate and self-contained description of all factors relevant forsystem design by a simple model based onglobal balances. This specification sheet hasbeen checked against FE- as well as thermo-dynamic equilibrium calculations. It is a fastand flexible design tool to analyse dimensions,mass- and energy flows, as well as overall effi-ciencies and possible applications.

11

11

0

500

1000

1500

2000

2500

3000

0 10 20 30 40 50 60 70 80

Fuel capacity [h]

Vo

l. !

E [

Wh

/l]

.

SOFC

Li-Ion Battery

Fuel Capacity Influence on SOFC Volume

Volumetric Energy Density

! SOFC shows big volume advantage at higher fuel capacities in comparison with

Li-Ion batteries

! Potential for HILTI applications?

SOFC 2.5 W at !el 40%

= ⋅ ⋅v E LHVel

ρ ρ η

Butane: 2935 Wh/l

Issue / Problem:The influence of energy capacity on the system volume is analysed. SOFC and Li-Ion battery are compared

Discussion:The volumetric energy density is an important parameter to compare different power systems, especially in the portable application sector.Battery: The volumetric energy density of a Li-Ion battery is independent of the battery energy capacity (figure in presentation). So an energy capacity enhancement by a factor of 2 causes a volume enhancement by a factor of 2 (figure bottom).SOFC: The volumetric energy density of an SOFC strongly depends on the energy capacity. For high energy capacities the SOFC volumetric energy density can theoretically reach the same value as the energy density of the fuel (butane).

To get a higher SOFC energy capacity only the fuel cartridge has to be enhanced. In case of the battery, the hole cell has to be enhanced.

Attention: In case of an SOFC it only makes sense to talk of volumetric energy density if the energy capacity is known.

Results:The SOFC shows big volume advantages for high energy capacity applications in comparison with Li-Ion batteries

SOFC Assumptions:" 2.5 W constant power

" !el 40%

" Active cooling: Tsurf 80°C

" Optimistic cell and insulation characteristics

0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80

Fuel capacity [h]

Vo

lum

e [

cm

3]

.

SOFC

Li-Ion Battery

System Volume

Fig. 2: Volumetric energy density of a SOFC as afunction of its fuel capacity.

Figure 2 shows how the volumetric energy den-sity of a SOFC increases with an increase in thefuel capacity. (The calculations were peformedfor an SOFC with an electrical power of 2.5 Wand an electrical efficiency of 40 %.) In the limitof a very high fuel capacity, the SOFC reachesthe electrical capacity of butane which has beenchoosen as fuel. These results demonstrate thebig volume advantage of an SOFC at higher fuelcapacities in comparison with Li-Ion batteries.

ZHAW 22

Page 23: ZHAW ICP Research Report 2008

Chapter 5

Organic Electronics andPhotovoltaics

23

Page 24: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

5.1 Advanced experimentally validated integrated OLED model

Contributors: Evelyne Huber, Beat Ruhstaller

Partners: Philips Research Aachen and Eindhoven, Technical University Dresden, FLUXiM,University of Cambridge, University of Groningen, Eindhoven University of Tech-nology, Sim4tec

Funding: European Union FP7Duration: 2008 – 2011

White Organic Light-Emitting Diodes (OLEDs)are highly efficient light sources that can po-tentially be used for general lighting. They canalso be printed on flexible surfaces resulting inlight-emitting flexible foils. Moreover, OLEDshave entered the small- and medium-size flatpanel display market. The increasing complex-ity of OLEDs (layer composition and structures)demands an experimentally validated compre-hensive OLED device model to design efficientand rational OLEDs. It has recently becomeclear that today’s first generation models, basedon conventional understanding of transport andphotophysical processes, are insufficient for re-alistic OLED materials. The aim of the AE-VIOM project1 is to develop a numerical simu-lation tool for achieving breakthroughs in whiteOLED technology. Within the project, a second-generation OLED device model will be devel-oped, within which the full three-dimensionalcharacter of the charge transport, recombina-tion, and excitonic processes is taken into ac-count. Thereby, the disordered nature of theorganic materials is treated more properly thanin today’s first-generation OLED device mod-els. The resulting integrated model will de-scribe the complete interplay between OLEDlayer structure, charge carrier mobilities, exci-tonic processes and light-outcoupling. The taskof the ZHAW within this project is the develop-ment of an effective 1-dimensional model thatintegrates the entire chain of physical processesthat takes place in OLED operation. The mainchallenge is to incorporate novel physical mod-els in the course of the project, which lead tostrongly coupled partial differential equations,into an efficient solver. In this first year a solverhas been implemented that solves the system ofpartial differential equations in a coupled man-ner and includes novel physical models like theExtended Gaussian Disorder Model (EGDM),the generalized Einstein diffusion and injectioninto Gaussian density of states. With the aid of

the simulation we can calculate a) what is hap-pening in the OLED (carrier profiles, see Fig. 1)as well as b) data that can be measured in anexperiment (IV-curves, see Fig. 2).

0 20 40 60 80 10010

−5

10−4

10−3

10−2

10−1

100

device [nm]

rela

tive

ca

rrie

r d

en

sity,

n(x

)/n

(0)

8 V

2 V

0.5 V

0 V

Fig. 1: Carrier concentration profile of a hole-only de-vice. The dotted lines show the solution for constantmobility and diffusion while the solid lines representthe solution for the EGDM at different voltages.

10−2

10−1

100

101

10−4

10−2

100

102

104

106

voltage [V]

curr

ent density [A

/m2]

EGDM

const

Mott−Gurney

Fig. 2: IV-curve for constant mobility and diffusionand the EGDM. The Mott-Gurney law is plotted forconstant mobility and neglecting diffusion.

1www.aeviom.eu

ZHAW 24

Page 25: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

5.2 Efficient areal organic solar cells via printing

Contributors: Roger Hausermann, Nils A. Reinke, Beat Ruhstaller

Partners: CSEM Basel, Ciba Inc., TU Eindhoven (NL), Universitat Jaume I (E)Funding: Swiss Federal Office of EnergyDuration: 2008 – 2011

This project with acronym Apollo aims to com-bine plastic electronics expertise in Europe forrealizing organic solar cells for empoweringprinted electronics applications. Existing solarcell technologies cannot provide the attributessuch as printability in ambient conditions, flexi-bility and low cost.So far, the development of organic solar cellswas to a large extent a semi-empirical trial-and-error process, in which organic semiconductingmaterials were selected on the basis of theirknown or partially known separate properties.It has recently become clear that this providesan insufficient sound basis for a further develop-ment. Therefore an interdisciplinary approachwith new materials, device concepts, modelsand characterization methods is critical.The focus of this project is on single cells andtandem cells with record efficiency that featureease of production and proof-of-principle for theinterdisciplinary research approach. The de-tailed understanding of device operation allowsfor steady improvements in efficiency.

Fig. 1: Maximum achievable current in the two sub-cells of a tandem solar cell. The white line indicatescurrent matching of the subcells, which is a precondi-tion to reach high efficiencies.

A comprehensive device model for the studyof operation mechanisms and the interpreta-tion of measured data will be developed. Itwill cover the whole process chain from lightabsorption, exciton dissociation, charge carriertransport and collection by electrodes. The ICP

has expertise in numerical modelling of opto-electronic processes in transient and steadystate in OLEDs which is used to develop a simu-lation software for organic solar cells. The max-imum achievable short circuit-current will be ad-dressed with optical simulations that provide thespatial exciton generation rate density. It is in-tended to distinguish the detrimental effects ofcharge trapping, recombination and collectionlosses by the use of drift-diffusion simulations.

-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0

-30

-20

-10

0

10

Normal AM 1.5 Intensity

Two Times AM 1.5 Intensity

Three Times AM 1.5 Intensity

Curr

ent [m

A/c

m2]

Bias [V]

Fig. 2: Simulated current-voltage curves of an organicsolar cell dependent on the illumination intensity.

In this early stage of the project an optical modelhas been developed to simulate multilayer or-ganic solar cells. It is used to calculate the max-imum achievable current in the solar cell. In Fig.1 such a simulation is shown, where the gener-ated current in a tandem solar cell is calculatedfor each subcell dependent on the thickness ofthese layers. This optical model has been cou-pled to a drift-diffusion solver to also simulatethe electrical behavior of such a solar cell. Firstresults can be found in Fig. 2 where calcu-lated current-voltage curves are shown. Thesecurves are simulated dependent on the illumina-tion intensity which is varied from one sun inten-sity to three times the sun intensity. In ongoingwork we investigate the exciton dissociation intofree carriers and validate the results with exper-iments.

ZHAW 25

Page 26: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

5.3 Cost efficient thin film photovoltaics for future electricitygeneration

Contributors: Roger Hausermann, Nils A. Reinke, Jurgen Schumacher, Beat Ruhstaller

Partners: EMPA, ETHZ, Univ. of Neuchatel, EPFL, PSI OC Oerlikon Solar, Solaronix, FlisomFunding: SwissElectric ResearchDuration: 2007 – 2010

The design of emerging thin film photovoltaiccells has several aspects in common: 1) Thecollection of ambient light in multilayers can beenhanced with the help of structured surfaces,2) it shall be absorbed inside the thin film stack,3) generate excited states that relax at nearbydissociation sites into electrons and holes, 4)which finally shall be transported to the elec-trodes with minimal loss.In the course of the second ThinPV2 project yearthe optical model has been finalized. For coher-ent optics an arbitrary layer setup can be sim-ulated. Not only can the achieved short-circuitcurrent Isc be estimated, but also the absorptionin each layer. This is shown in Figure 1. In thetop part the optical field distribution is calculatedinside a multilayer thin film solar cell stack. Fromthis calculation the amount of absorbed photonscan be calculated at every position inside thedevice, which can be seen in the lower part ofFigure 1.Subsequently, one can calculate the maximumachievable short-circuit current. By varying thelayer thickness and materials one can optimizea given thin film solar cell. The extension ofthe multilayer optical model for mixed incoher-ent/coherent layer stacks is underway. This willprovide the basis for designing several of the ex-isting thin film photovoltaics technologies underinvestigation in this project.

0 100 200 300 4000.0

2.0x1019

4.0x1019

6.0x1019

8.0x1019

1.0x1020

Num

ber o

f abs

orbe

d ph

oton

s [n

m-1 s

-1]

Position inside the stack [nm]

Spectral EM Field

5 0 100 150 200 250 300 350 400 450Position

400

425

450

475

500

525

550

575

600

625

650

675

700

725

750

775

Wav

elen

gth

(nm

)

Glass ITO PEDOT P3HT:PCBM

ZnO Silver

Highly absorbingspectral region of P3HT:PCBM

Fig. 1: Optical field distribution within a multilayer thinfilm cell (top) and profile of absorbed photons persecond (bottom) under AM1.5 illumination.

2http://thinpv.empa.ch

ZHAW 26

Page 27: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

5.4 Organic electroluminescent pictograms for push-button ap-plications

Contributors: Nils A. Reinke, Roger Hausermann, Beat Ruhstaller

Partners: CSEM, Ciba, EAOFunding: Commission for Technology and Innovation (CTI)Duration: 2007 – 2009

The novelty of the envisioned product in thisproject is the possibility to produce very thin,custom-designed self-luminous pictograms forapplications in push-buttons. They combine ex-isting pushbuttons (EAO) with patterned emis-sive pictograms (CSEM) made from electrolu-minescent semiconductor materials (Ciba SC).The characterization and optimization of mate-rial properties, device and system architectureis supported by computational engineering andsimulation software3 development at ZHAW asillsutrated by prototypes in Fig. 1.

Fig. 1: Transparent PLED (left) and backlighted PLEDprototype (right) developed in this project. The com-bination of blue backlight and yellowish emission fromthe PLED yields a white color emission.

The numerical simulation activities subdivideinto electronic and optical device modeling aswell as advanced analysis algorithms. Theelectronic device model contains charge drift-diffusion and exciton rate equations for mod-eling organic semiconductor devices includingmultiple excitons, several charge mobility and in-jection models and charge doping and trapping.The numerical algorithm delivers charge distri-bution profiles, transient electroluminescence,as well as current-voltage curves. The opticalsolver evaluates the angular emission charac-teristics as well as the substrate, thin-film andplasmon modes. The modes can be visualizedand integrated to extract relative mode contri-butions and position-dependent lifetimes. Keyperformance figures such as luminous efficacy,CIE, CRI, etc. are calculated. Advanced analy-sis features allow to extract device parameters,

for instance the shape of the emission zone,from experimental data. Our model fits experi-mental current-voltage curves of PLEDs consid-ering a field dependent charge carrier mobility(cf. Fig. 2).

Fig. 2: Experimental and simulated current-voltagecurves considering a field dependent mobility.

In large-area OLEDs non-uniformities in the lu-minance arise from potential drops across alow-conductive electrode. FEM-simulations al-low to study such potential drops (cf. Fig. 3).

0 V

0 V

10 V

Fig. 3: SESES simulation of the potential drop.

3SETFOS Semiconducting Emissive Thin Film Optics Simulator, Fluxim AG, Switzerland, www.fluxim.com.

ZHAW 27

Page 28: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

5.5 Modeling, simulation and loss analysis of dye-sensitizedsolar cells

Contributors: Matthias Schmid, Jurgen O. Schumacher

Partners: Laboratoire de Photonique et Interfaces (LPI), EPFLFunding: Gebert Ruf StiftungDuration: 2008 – 2010

Dye-sensitized solar cells (DSCs) are an inno-vative technology for the production of electric-ity from solar energy. In contrast to conven-tional silicon semiconductor solar cells, light isabsorbed in the DSC by dye molecules boundedto the surface of a highly porous structure ofnanoparticles of transparent TiO2. The pro-duction process of DSCs is based on relativelysimple and inexpensive techniques, like e.g.screen printing. In addition no cost-intensive,exhaustible raw materials are needed. There-fore, the contribution of DSCs to an economicproduction of solar energy could be essential inthe future.

Fig. 1: SEM image of the F:SnO2 glass used as frontelectrode in the DSC (image courtesy of EPFL). TheSEM images are used to estimate the thickness andsurface roughness of the F:SnO2 layer.

In September 2008 we started a researchproject for the modeling of dye-sensitized solarcells. The project is carried out under the lead ofthe ICP in a close collaboration with the Labora-tory of Photonics and Interfaces (LPI) at EPFL.The objective of this project is to develop val-idated mathematical models for the DSC. Themodels aim at describing the coupled optical,electrical and electrochemical processes takingplace within the solar cell. This allows to ana-lyze quantitatively the different loss channels ofthe energy conversion process. Research anddevelopment of DSCs could be accelerated to a

large extent, if these new device models wereimplemented into accurate and efficient numeri-cal algorithms.Currently the ICP is developing an optical modelof the DSC in order to simulate absorption, re-flection loss and the spatially resolved sensi-tizer excited state generation rate. The opti-cal model is validated by optical reflection andtransmission measurements of the layers thatconstitute a DSC (F:SnO2-glass, nanoporoussemiconductor film, monolayer of dye, and elec-trolyte). The measurements are carried out byEPFL, see e.g. Fig. 1.Simultaneously, a complete 1D through-planemodel of the DSC is formulated, which uses thesimulated generation rate profile from the opti-cal model as an input. The through-plane modelis based on a coupled nonlinear 1D system ofpartial differential equations (PDEs) to describethe electrochemical reactions and the transportprocesses. This 1D model is validated on lab-oratory DSCs by different measurements tech-niques such as photovoltage and photocurrentresponse and spectrally resolved quantum effi-ciency, see Fig. 2.

Fig. 2: Comparison of the simulated (solid line) andmeasured (dashed line) quantum efficiency QE(λ) ofthe DSC as a function of wavelength of the incidentlight. The dotted line is the simulated fraction of ab-sorbed light fabs(λ) and the dash-dotted line showsthe simulated reflection loss R(λ) of the front elec-trode.

ZHAW 28

Page 29: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

5.6 Photovoltaics inspector

Contributors: Beat Ruhstaller, Nils A. Reinke

Partners: Fluxim AGFunding: Transferkolleg ”Industrial Photonics” funded by CTI and SATWDuration: 2008 – 2009

The production of thin film solar cells demandshigh accuracy, quality and control on the pro-cess parameters in order to ensure highest pos-sible power conversion efficiency, lifetime aswell as yield. Quality-monitoring during and atthe end of a production line is thus mandatory.Traditionally both optical and electrical methodsare used to assess the quality, compare withFig. 1. It is, however, not yet possible to auto-matically extract crucial material and device pa-rameters with the help of an accurate numericaldevice model. Such a method would allow anearly identification of process parameter devia-tion in production. In addition, a reliable numer-ical model is useful in the cell design phase inR&D. The software SETFOS is commercializedby Fluxim AG for organic semiconductor devicesimulation in R&D. So far, SETFOS does notyet provide physical models for charge trans-port and light scattering in inorganic semicon-ductor devices. Both is relevant to inorganic so-lar cells that dominate the photovoltaics markettoday. The extended numerical device modelshall combine a) an accurate description of therelevant physical processes, b) high computa-tion speed and c) algorithms for parameter ex-

traction.In this feasibility study, the project idea is beingevaluated and discussed with potential stake-holders and project partners. It includes a mar-ket study and analysis of industry needs and ex-isting inspection methods.

Fig. 1: Illustration from Xenon Corp. of xenon flashlamps for photovoltaic module inspection.

ZHAW 29

Page 30: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

5.7 Optoelectronic research laboratory

Contributors: Nils A. Reinke, Andreas Tiefenauer, Beat Ruhstaller

Partners: Various project partnersFunding: ICP, School of EngineeringDuration: ongoing

True-to-life modeling always is based on accu-rate input parameters from the real world. Thepurpose of the optoelectronic research labora-tory (O-LAB) in this context is both providingsuch input parameters and experimentally val-idating simulation results. Combining advancedcomputational methods and state-of-the-art ex-perimental methods allows us to develop inno-vations in measurement engineering.

Fig. 1: Photo-thermal characterization of an infrared-sensor.

The O-LAB is therefore a consequent expan-sion of our existing computational toolbox, fa-cilitating the collaboration with experimentally-oriented partners and industry. Our current in-vestigations comprise the following fields of in-terest:

• Ultra-fast photo-thermalheating and detection

• Thin film spectroscopy

• Transient optoelectronicdevice characterization

An exemplary setup developed at the OLAB isshown in Fig. 1. This setup allows to measure

photo-thermal response times of infrared sen-sors down to 100 µs. An experimental setupmeasuring transient currents and electrolumi-nescence from organic light emitting devices isdepicted in Fig. 2. Our experimental toolboxcomprises fast oscilloscopes, signal generators,IR-sources and detectors, flashers and spec-trometers.The O-LAB supports industrial needs in thecharacterization of sensors and actuators andseveral ICP projects related to thermal labeling,thin-film spectroscopy, organic light emitting de-vices and photovoltaics. In addition, our labora-tory is ideally suited for educational purposes inteaching B.Sc. and M.Sc. students in mecha-tronics, electronics and informatics. The O-LABgives young scientists and engineers the possi-bility of getting in contact with R&D and workingon exciting issues in ongoing projects.

Fig. 2: Transient measurement of currents and elec-troluminescence in organic light emitting devices.

ZHAW 30

Page 31: ZHAW ICP Research Report 2008

Chapter 6

Student Projects

31

Page 32: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

6.1 Wirtschaftlichkeit und Okobilanz eines Blockheizkraftwerksmit regenerativen Energietragern und 5 bis 20 kW elek-trischer Leistung

Students: Martin Hauser, Marcel Ott

Category: Projektarbeit MaschinentechnikMentoring: Christoph Meier, Thomas HockerPeriod: Februar 2008 – Mai 2008

Ein Blockheizkraftwerk (BHKW) ist eine modu-lar aufgebaute Anlage zur Erzeugung von elek-trischem Strom und Warme, die vorzugsweiseam Ort des Warmeverbrauchs betrieben wird.Sie setzt dazu das Prinzip der Kraft-Warme-Kopplung ein. Der hohere Gesamtnutzungsgradgegenuber der herkommlichen Kombination vonlokaler Heizung und zentralem Großkraftwerk(fur die Stromproduktion) resultiert aus derNutzung der Abwarme der Stromerzeugung di-rekt am Ort der Entstehung. Als Antrieb furden Stromerzeuger konnen Verbrennungsmo-toren, aber auch Gasturbinen verwendet wer-den. Gasmotoren werden meist mit Erdgas be-trieben. Jedoch auch ein Betrieb mit regen-erativen Energietragern wie z. B. Holzgas undRapsol ist denkbar.

Thomas Hocker, ZHAW 28. März 2008

Energiebilanz um Wärmespeicher

Da die Masse innerhalb des Speichers

konstant ist, kann man die Energiebilanz

für ein geschlossenes System anwenden.

Sie lautet in ihrer differentiellen Form:

WärmespeicherEnergiebilanz

QzuIsolationsdicke

Hzu Hab

U = ρVs cTs

Oberfläche As

dU = Qdt − pdV

Umformungen

dU = ρVs cdTs

pdV = 0

Qdt = (Qzu + Hzu − Hab)dt

Einsetzen

Weitere Umformungen

Wärmeverluste

TwQu

dTs

dt=

Qzu − Qu + Hzu − Hab

ρVs c

Qu = As α(Tw −Tu)

Wärmeverbraucher Qv = Hzu − Hab

Wärmeleitung

Endergebnis

d

Qu = Aκ(Ts −Tw)/d

Wandtemperatur Tw = d αTu +Ts κ/(d α+κ)

dTs

dt=

Qzu −As α(Tw −Tu)+ Qv

ρVs c

Thomas Hocker, ZHAW 28. März 2008

Lösung der Energiebilanz des Wärmespeichers

Die obige Differentialgleichung für Ts(t) ist linear in Ts und vom Typ

Falls C2 konstant ist, d.h. falls die Wärmezufuhr durch das BHKW und die Wärmeabfuhr durch den Verbraucher konstant sind, folgt als Lösung

dTs

dt= C1 Ts +C2(t), Ts|t=0 = Ts,0

Ts(t) =(

C2

C1+Ts,0

)exp(C1 t)− C2

C1

Für C2 konstant verändert sich Ts also exponentiell mit der Zeit. Da C1 < 0, gilt für grosse Zeiten Ts|t!! = –C2/C1. Die Lösung für beliebige C2(t) hat die Form

C1 und C2(t) sind wie folgt definiert:

C1 ≡−As α

ρVs c( ακ/d +1)

< 0

dTs

dt=

Qzu(t)−As α(

Ts+ ακ/d Tu

( ακ/d +1) −Tu

)+ Qv(t)

ρVs c

C2(t) ≡Qzu(t)+ As αTu

( ακ/d +1) + Qv(t)

ρVs c

Ts = exp(C1 t)[

Ts,0 +Z t

0exp(−C1 ψ)C2(ψ)dψ

]

Fig. 1: Einfaches Modell eines Systems aus Warme-verbraucher, -quelle und -speicher.

Obwohl die BHKW-Technologie wegen derdamit verbundenen CO2-Reduktion (s. a. Kyoto-Protokoll) in vielen europaischen Landernstaatlich gefordert wird, blieb deren Verbreitungbisher unter den Erwartungen zuruck. De-shalb ist das Ziel des Projektes, die BHKW-Technologie in Kombination mit regenera-tiven Energietragern auf ihre Wirtschaftlichkeitund ihren okologische Nutzen hin zu unter-

suchen. Fur den warmegefuhrten Betriebwurde in Matlab/Simulink ein einfaches Mod-ell aus Warmeverbraucher, -quelle (BHKW) und-speicher implementiert, siehe Abbildung 1.Es konnte gezeigt werden, dass Mikro-BHKW-Anlagen unter der Voraussetzung eines mini-malen Warmebedarfs zu ahnlichen Kosten be-trieben werden konnen wie konventionelle Pel-letheizkessel, siehe Abbildung 2.

!"#$%&%'()*+,-.(/+0-% 12(34+(%5)3463478+%92(%:";+<."=-+%>066+"634.9-+"%

10 Wirtschaftlichkeit %

?@%=0+%A5B>&%:"8.;+"%".34% 04(+(%>0(-634.9-8034,+0-%/+7(-+08+"%C7%,D""+"E%

<7(=+% *+<+086% +0"% ,)"F+"-0)"+88+(% '+88+-&% 5+0C,+66+8% .86% G+9+(+"C."8.;+%

4+(.";+C);+"H%I./+0%<7(=+%=0+%:@)(-06.-0)"6=.7+(%92(%=0+%,)"F+"-0)"+88+%

5+0C."8.;+% 7"=% 92(% =+"% JK0-C+",+66+8% .79% LM% N.4(+% .";+6+-C-E% <O4(+"=%

A8)3,4+0C,(.9-<+(,% 7"=% 5)8CF+(;.6+(% /+=0";-% =7(34% 04(+"% ;(D66+(+"%

#+(6348+066%/+(+0-6%".34%PHM%N.4(+"%./;+634(0+/+"%60"=H%:766+(=+@%<7(=+%

92(% =0+% A(+""6-)99,)6-+"% +0"% '+88+-K(+06% F)"% Q% GKHR,>4% +0";+6+-C-H% I0+% C7%

+(<.(-+"=+% !0"6K+06+F+(;2-7";% <7(=+% 0"% :"8+4"7";% ."% =0+%

J-()@F+(6)(;7";6F+()(="7";% PSTHPL% F)@% LH% :K(08% UVVQ% .79% SV% GKHR,>4%

9+6-;+8+;-H%

I0+%;+6.@-+%B)6-+"(+34"7";%7"=%<+0-+(+% W"9)(@.-0)"+"%=.C7% 60"=% 0"%=+(%

$./+88+%X#+(;8+034%JC+".(0+"%A5B>Y6Z%0@%:"4.";%+(6034-8034H%

%

Kostenvergleich

0.00%

50.00%

100.00%

150.00%

200.00%

250.00%

300.00%

EFH

_alt

EFH

_neu

MFH

_alt

MFH

_neu

MFH

_neu

_198

2

MFH

_neu

_199

6

MFH

_neu

_200

0

Sch

ulhau

s_ne

u

Rest

aura

nt

Jäh

rlic

he G

esam

tko

ste

n

BHKW und SK

konv. Heizkessel

Abbildung 14: Kostenvergleich BHKW mit SK vs. konventioneller Heizkessel

W@% )/0;+"% I0.;(.@@% 8O66-% 6034% 6+4(% ;7-% +(,+""+"E% =.66% =0+%

A8)3,4+0C,(.9-<+(,+% ./% =+@% JC+".(0)% [\5% .8-% "7(% 7"<+6+"-8034% ;(D66+(+%

*O4(8034+% ]+6.@-,)6-+"% F+(7(6.34+"% .86% =0+% *+<+080;+"% ,)"F+"-0)"+88+"%

G+9+(+"C."8.;+"E%<)/+0% =0+% A5B>&% :"8.;+% 0@% J34784.76% 6);.(% ;2"6-0;+(%

<0(=H%W@%]+;+"6.-C%=.C7%6-+88-%6034%.734%=+7-8034%4+(.76E%=.66%=0+%A5B>Y6%0"%

=+"% !0"9.@080+"4O76+("% =+7-8034% -+7(+(% <+(=+"% .86% =0+% ,)"F+"-0)"+88+"%

5+0C."8.;+"H% I0+6% 8O66-% 6034% @0-% =+@% ;+(0";+"% >O(@+/+=.(9% 7"=% =+(%

=.=7(34% 6+4(% ,8+0"+"% ^+06-7";% =+6% A5B>Y6% % +(,8O(+"H% I7(34% =0+% ;+(0";+%

](D66+% <+(=+"% =0+% 6K+C090634+"% :"8.;+",)6-+"% 6+4(% ;()66% 7"=% ,D""+"%

<0+=+(7@% "034-% =7(34% =+"% ,8+0"+"% !(-(.;% .76% =+(% !0"6K+06+F+(;2-7";%

;+=+3,-%<+(=+"H%

$()-C%=+(%F0+8+"%?"6034+(4+0-+"%0"%=+(%A+(+34"7";%8O66-%6034%.86)%/+4.7K-+"E%

=.66% =0+% A5B>&% $+34")8);0+% ./% +0"+@% ;+<066+"% >O(@+/+=.(9%

<0(-634.9-8034%,)",7((+"C9O40;%<0(=%_60+4+%J34784.76%"+7`H%

%

A8)3,4+0C,(.9-<+(,% J+0-+%UL%F)"%UQ%J+0-+"%

a--%R%5.76+(RVLHVTHUVVQ%

Fig. 2: Kosten BHKW mit Standardkessel fur ver-schiedene Applikationen im Vergleich zu einem kon-ventionellen Heizkessel.

Fur Einfamilienhauser und andere kleineBauten ist der Warmebedarf allerdings zugering. Die Moglichkeit von Warmeverbun-danlagen, wo mehrere EFH durch dasselbeBHKW beheizt werden, stellt hier sicherlicheine interessante Alternative dar. Vorausset-zung fur einen storungsfreien Betrieb der An-lage sind technisch ausgereifte Komponenten.Dieser Entwicklungsstand ist bei gasbetriebe-nen Blockheizkraftwerken erreicht worden, beiden kleinen Holzvergasern muss allerdingsnoch weiterer Entwicklungsaufwand betriebenwerden. Ein weiterer ungeklarter Punkt ist dieAuslegestrategie verschiedener Hersteller: ent-gegen unserer Auslegung werden keine, odernur sehr kleine Pufferspeicher verwendet, wasden Deckungsgrad und somit den Anteil gekop-pelt erzeugter Energie deutlich verkleinert.

ZHAW 32

Page 33: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

6.2 Modell-basierte Analyse der U–I–Kennlinie von SOFCHochtemperatur-Brennstoffzellen

Students: Markus Linder

Category: Vertiefungsarbeit Masterstudium (MSE)Partner: Hexis AGMentoring: Thomas HockerPeriod: Oktober 2008 – Februar 2009

Die Hexis AG, Winterthur, gehort zu denfuhrenden Unternehmen in der Entwicklungstationarer Hochtemperatur Brennstoffzellen-systeme. Hexis entwickelt Brennstoffzellen-Heizgerate im kleinen Leistungsbereich, die furdie dezentrale Warme- und Stromversorgung inEinfamilienhausern eingesetzt werden.Um die Brennstoffzellen in ihren elektro-chemischen sowie thermo-mechanischenEigenschaften gezielt weiter zu entwickeln,mussen diese moglichst exakt charakterisiertwerden. Dazu wird im Betrieb die elek-trische Spannung in Abhangigkeit des elek-trischen Stromflusses gemessen und als so-genannte Strom-Spannungskennlinie ((U, i)-Kennlinie) aufgezeichnet. Diese Kennlinie istvon zentraler Bedeutung, da damit die Leis-tungskurve dargestellt und Verluste typisiertwerden konnen. Das Ziel dieser Vertiefungsar-beit war die Entwicklung eines Modells, mitdem die umfangreichen Messdaten von Zellen-Prufstanden effizient analysiert werden konnen.Dazu wurden die unterschiedlichen Effekte undEinflusse auf den Gesamtverlust quantifiziert.Fur den Betrieb der Brennstoffzelle wird Erdgas,das primar aus Methan (CH4) besteht, ver-wendet. Auf der Zelle selbst wird Wasser-stoff (H2) umgesetzt. Dazu wird das Erdgas,in einem Reformer in Wasserstoff und Kohlen-monoxid CO umgewandelt. Uber die Mengedes verbrauchten Wasserstoffes an der Zelle,wird dann die ideale (U, i)-Kennlinie berech-net. Die (U, i)-Kennlinie aus den Messun-gen, unterscheidet sich von der berechneten(U, i)-Kennlinie. Diese gut erkennbare Differenzergibt sich aufgrund der Verluste, die im realenSystem auftreten. Mit dem entwickelten Mod-ell konnen, nebst dem Erzeugen der idealen(U, i)-Kennlinie, die Leckage- und Repeat Unit-Verluste (eine Repat Unit eines Zellenstapelsbesteht aus der Zelle und den Stromsammler-

Platten) quantifiziert werden, vgl. mit Abbil-dung 1. Dazu wird die ideale (U, i)-Kennliniein mehreren Schritten an die gemessene (U, i)-Kennlinie angepasst.Das in Mathematica implementierte Modell er-laubt eine weitgehend automatisierte Auswer-tung der Messdaten. Einzig die Zusammenset-zung des Erdgases nach der Reformierung undder Verbrauch an der Zelle mussen vorgangigermittelt werden. Eine Zusammenfassung derResultate und Inputdaten der Modellrechnungwird als Excel-Sheet ausgegeben. Fur dieUberprufung auf mogliche Fehler wurden dieResultate des Modells mit denen einer altenAuswertung verglichen. Da diese Verifizierungerfolgreich war, darf das Modell fur Auswer-tungen und Analysen eingesetzt werden. Inweiteren Arbeiten ist geplant, die Benutzerfre-undlichkeit des Modells weiter zu verbessern.Es soll kunftig bei Hexis fur die routinemassigeAnalyse von (U, i)-Kennlinien eingesetzt wer-den. Vorgesehen ist zudem eine Auswertungund statistische Aufbereitung einer grosserenAnzahl von (U, i)-Kennlinien.

Diskussion

Markus Linder 33

!Abbildung 18 Auswertung mit der Mai Messdaten mit dem neuen Modell

4.4 Einfluss von variablen Grössen

"#$%! &'(($)!$#)!*++%!,-&.#%/-)0$)!+-1!2+%#+3($!4%5&&$)!-)6$%&-786!-)9!9$%$)!:#)1(-&&!+-10$;

<$#06!.$%9$)=!

>?%! $#)$! 0$0$3$)$! :%90+&! @-&+AA$)&$6<-)0! /+))! 9#$! B'(A+&&$! 3$&6#AA6! .$%9$)!

C20(=!D+*=!E=FG=! >?%! #9$+($&! :%90+&H! +(&'! %$#)$&! B$68+)! CI"JGH! #&6! 9#$&$! KL!0MA'(=! N$+($&! :%90+&!

8+6H! +38O)0#0! 2')! 9$%! @-&+AA$)&$6<-)0H! $#)$! -A! %-)9! K!0MA'(! 0%5&&$%$! B'(A+&&$=! P#$! #A!

2'%0O)0#0$)! D+*#6$(! C20(=!D+*=!J=QG! 3$%$#6&! +)0$9$-6$6H! .#%9! 3$#! $#)$%! 0%5&&$%$)! B'(A+&&$!9$%!

B'($)&6%'A!3$<'0$)!+-1!9$)!:%90+&!R)*-6H!9$%!+(&!B+&&$)&6%'A!2'%0$0$3$)!.#%9H!2$%/($#)$%6=!

S#$&!1?8%6!9+<-H!9+&&!9$%!T6%'A1(-&&!+3)#AA6=!S#$&$%!@-&+AA$)8+)0!#&6! #)!,33#(9-)0!KU!0%+;

1#&78!9+%0$&6$((6=!S+!9$%!:#)1(-&&!9$%!B'(A+&&$!%$(+6#2!0%'&&!#&6H!&'((6$!1?%!$#)$!'*6#A+($!,-&.$%;

6-)0!9$%!B$&&9+6$)!9#$!:%90+&<-&+AA$)&$6<-)0!<-A!@$#6*-)/6!9$%!B$&&-)0!3$/+))6!&$#)=!

:#)$!V)9$%-)0!9$&!B+&&$)&6%'A&!9$&!:%90+&#)*-6&!8+6!+-1!9#$!I8+%+/6$%#&6#/!9$%!#9$+($)!WX;

D$))(#)#$!9$)!0($#78$)!:#)1(-&&! .#$! 9#$!V)9$%-)0! 9$%!B'(A+&&$!9$&!:%90+&$&=! Y$9'78! -A0$;

/$8%6!*%'*'%6#')+(=!S+!9$%!:%90+&!R)*-6A+&&$)&6%'A!#A!@O8($%!1?%!9#$!Z$&6#AA-)0!9$&!B'($);

&6%'A!&6$86! C20(=!D+*=!E=[!4($#78-)0!CE=KQG=!S$&8+(3! #&6!$&!2')!Z$9$-6-)0!9$)!B+&&$)&6%'A!9$&!

:%90+&#)*-6&!A50(#78&6!$\+/6!<-!$%1+&&$)=!

S$%!:#)1(-&&!9$%!T6+7/6$A*$%+6-%!.-%9$!3$%$#6&!#)!9$)!D+*#6$()!Q=J!-)9!J=K!$#)0$8$)9!9#&/-;

6#$%6=! P#$! 3$%$#6&! $%(O-6$%6! .-%9$H! #&6! 9#$! :%1+&&-)0! 9$%! $11$/6#2! +)! 9$%! @$(($! 2'%8$%%&78$)9$)!

]$A*$%+6-%! $)6&78$#9$)9=! :#)$! ,3.$#78-)0! 9$%! ]$A*$%+6-%! 3$.#%/6! $#)$! 2$%6#/+($! ^$%&78#$;

3-)0!9$%!#9$+($)!WX;D$))(#)#$=!S#$&$%!:11$/6!.#%9!3$#!9$%!R)6$%*'(+6#')!<.#&78$)!9$)!]$A*$%+;

6-%$)!+-&0$)-6<6=!

! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !!!

!

" " " " " " " " " " " " " " " " " " ""

# # # # ## # # # # # ## # # # # # ###

#

$$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $

$$

$

!"" #"" $""% !

&'

(&#"

")#

")*

")+

"),

!)"

- #.$

/011234!.564782 /0192:

$ Ideal #AsTher $

# Leakage " 25.7#

" Stack Data

! ASR RU " 420 m$ cm2

;<=>4 " *)"?? &.

Fig. 1: Modellbasierte Analyse der Spannungs-Strom-Kennlinie einer Hochtemperatur-Brennstoffzelle.

ZHAW 33

Page 34: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

6.3 Modell-basierte Optimierung der Holzvergaseranlage in Wila

Students: Roger Bachtiger, Fabian Kekeis

Category: Diplomarbeit MaschinentechnikPartner: Woodpower AG, Wila, Globutech Services, Freiburg im BreisgauMentoring: Thomas HockerPeriod: August 2007 – Oktober 2007

Unter Holzvergasung versteht man den ver-fahrenstechnischen Prozess der Teilverbren-nung unter Luftmangel um aus Holz dasbrennbare Holzgas zu gewinnen. Holzgaswird beispielsweise als Brennstoff fur stationareGasmotoren eingesetzt, um uber einen Gen-erator elektrischen Strom zu produzieren. DieWoodpower AG und die Elektrizitatswerke desKantons Zurich (EKZ) betreiben seit Anfang2007 die erste kommerzielle Anlage fur Holzver-gasung in der Schweiz. In der Anlage werdennach der oben beschriebenen Methode ca. 450kW Strom und zusatzlich ca. 1000 kW Warmeaus Altholz erzeugt. Projektziel war die modell-basierte Optimierung der Holzvergaseranlage inWila uber Massen- und Energiebilanzen sowiethermodynamische Gleichgewichtsmodelle. Diehierfur benotigten Grossen wie Temperaturen,Konzentrationen und Massenflusse wurden ausMessungen vor Ort bestimmt. Aus den Mod-ellen lassen sich wichtige Kenngrossen wie dieHohen der verschiedenen Reaktionszonen imReaktor ermitteln, siehe Abbildung 1.

Diplomarbeit: EnVT Holzvergaser

Roger Bächtiger • Fabian Kekeis52

Zürcher Hochschule für Angewandte Wissenschaften

Auswertung Partikelform8.3.3.1

Die Partikelform hat bei gleichem Volumen einen grossen Einfluss auf die aktive Zonenhöhe (Abbildung 22). Die Pyrolyse geht mit wachsender Partikeloberflä-che (kleinere Sphärizität) schneller voran und verkürzt dadurch auch die aktive Zone.Bei dieser Auswertung wurde immer mit gleichem Volumenstrom und glei-cher Schüttdichte gerechnet. Die Partikelform würde jedoch in Wirklichkeit die Schüttungsdichte beeinflussen.

Partikelgrösse8.3.4

Die Partikelgrösse (Volumen) ist ein weiterer Bestandteil der HUFF-Gleichung und wir haben untersucht wie stark sich die Zonen mit dem Volumen verän-dern. Die Rechnung wurde mit dem durchschnittlichen Schnitzel gemacht.

Auswertung der Partikelgrösse8.3.4.1

Die aktive Zone wird mit grösserem Volumen länger und kleinerem kürzer (Ab-bildung 23). Die Schüttungsdichte wurde wieder als konstant angenommen, allerdings müsste diese bei den grösseren Schnitzel kleiner werden.

Flammenpyrolyse

Kohlenausgasung

1.15 m

1.75 m

0.9 m

1.5 m

0.7 m

1.15 m

0.25 m

0.55 m

0.8

m

1.3

5 m

2.4

m

2.9

m

Volumen: 22 cm3

Holzfeuchte: 18 %Holzdichte: 550 kg/cm3

Volumenstrom: 1.67 m3/hPyrolysetemperatur: 600 °CBettgeschwindigkeit: 1.47 m/h

Flammenpyrolyse

Kohlenausgasung

1.15 m

1.75 m

0.9 m

1.5 m

0.7 m

1.15 m

0.25 m

0.55 m

0.8

m

1.3

5 m

2.4

m

2.9

m

Volumen: 22 cm3

Holzfeuchte: 18 %Holzdichte: 550 kg/cm3

Volumenstrom: 1.67 m3/hPyrolysetemperatur: 600 °CBettgeschwindigkeit: 1.47 m/h

Flammenpyrolyse

Kohlenausgasung

1.15 m

1.75 m

0.9 m

1.5 m

0.7 m

1.15 m

0.25 m

0.55 m

0.8

m

1.3

5 m

2.4

m

2.9

m

Volumen: 22 cm3

Holzfeuchte: 18 %Holzdichte: 550 kg/cm3

Volumenstrom: 1.67 m3/hPyrolysetemperatur: 600 °CBettgeschwindigkeit: 1.47 m/h

Zonenhöhen im Wilareaktor bei verschieden Körpern mit selben VolumenAbbildung 22:

Flammenpyrolyse

Kohlenausgasung

V= 22 cm3

V = 11 cm3

V = 44 cm3

Holzfeuchte: 18 %Holzdichte: 550 kg/cm3

Volumenstrom: 1.67 m3/hPyrolysetemperatur: 600 °CBettgeschwindigkeit: 1.47 m/h

0.5 m

0.75 m

0.7 m

1.15 m

1.05 m

1.65 m

1.2

5 m

1.8

5 m

2.7

m

Zonenhöhen eines durchschnittlichen Holzschnitzels mit verschiedenen VoluminaAbbildung 23:

Fig. 1: Berechnete Hohen der Flammenpyrolyse- undReduktionszonen in Abhangigkeit von der Form derverwendeten Holzschnitzel.

Hierfur muss man wissen, wie lange die Py-rolyse (d. h. der Ausgasungsprozess) einzelnerPartikel dauert. Die in Abbildung 2 gezeigten

Messungen von Pyrolysezeiten stimmen gut mitden von E. R. Huff1 publizierten Werten uberein.

Diplomarbeit: EnVT Holzvergaser

Roger Bächtiger • Fabian Kekeis48

Zürcher Hochschule für Angewandte Wissenschaften

Pyrolyseversuch in Bildern

VersuchsschnitzelAbbildung 14:

TemperaturmessungAbbildung 16:

ReduktionAbbildung 18:

VersuchsaufbauAbbildung 15:

PyrolyseAbbildung 17:

Fig. 2: Messung der Pyrolyse- und Reduktionszeit aneinzelnen Holzpartikeln.

Ausserdem wurde berechnet, wie sich derunerwunschte Austrag an Holzkohle mit demFeuchtegehalt der verwendeten Holzschnitzelsowie den Warmeverlusten uber die Reaktor-wand verandert, siehe Abbildung 3. Bei einemWassergehalt von 15 % und einem Warmever-lust von 20 kW ergibt sich ein Holzkohleaustragvon knapp 5 Gew.-%. Dieser Wert stimmt gutmit den Erfahrungswerten der Woodpower AGuberein.

Diplomarbeit: EnVT Holzvergaser

Roger Bächtiger • Fabian Kekeis 45

Zürcher Hochschule für Angewandte Wissenschaften

Holzkohleaustrag - Analyse mit Gleichgewichtsmodell7.2

Im Rahmen der Projektarbeit wurde der Holzkohleaustrag bereits untersucht und dabei bemerkt, dass der Wasseranteil im Holz ein wichtiger Bestandteil ist. Da der Wasseranteil in den Schnitzeln von Wila zwischen 15 und 20 Prozent schwankt, ist es weiterhin von Interesse dessen Einfluss zu untersuchen.Mit Hilfe des Gleichgewichtsmodells wurden weitere Messdaten der Holzver-gaseranlage in Wila analysiert und die Wärmeverluste am Reaktorgehäuse in Funktion des Holzkohleaustrages aufgezeichnet (Abbildung 12).Da wir das Modell nur an die CO- und H

2-Werte fitten sind noch zu viele wichti-

ge Faktoren wie zum Beispiel CO2 unbekannt. Die Diagramme zeigen nur einen

Trend auf, welcher Einfluss der Wasseranteil auf den Holzkohleaustrag hat. Die Wärmeverluste könnten höher oder tiefer liegen, da sich Einflüsse wie Brenn-wert des Holzes und die Luftzahl im Reaktor, welche wir beide nicht im Reaktor messen können, ebenfalls sehr stark auswirken.

Die Diagramme in Abbildung 12 zeigt auf wie gross die Wärmeverluste bei ei-nem bestimmten Kohleaustrag im Reaktor sind. Die dicken Linen entsprechen jeweils dem Mittelwert, die Schattierungen der Standartabweichung der be-rechneten Daten.Interessant ist, dass die Wärmeverluste des Reaktors bei 15 Prozent Wasser-anteil in der Biomasse am stärksten fallen. Bei 20 Prozent ist der Wärmeverlust praktisch stabil, allerdings wäre der Betrieb nach unserem Modell nicht möglich, da man zusätzlich Wärme zuführen müsste um ein Gleichgewicht einzustellen.

Auswertung der Wärmeverluste über den HolzkohleaustragAbbildung 12:

5 10 15 20C Austrag

150

100

50

50

100

Q kW

Pel 350kW Datum : 27.07.2008

___ !"## $%&&'(%)*'+,

___ -!"##$%&&'(%)*'+,

___ ./"##$%&&'(%)*'+,

5 10 15 20C Austrag

150

100

50

50

100

Q kW

Pel 350kW Datum : 26.07.2008

___ !"## $%&&'(%)*'+,

___ -!"##$%&&'(%)*'+,

___ ./"##$%&&'(%)*'+,

5 10 15 20C Austrag

150

100

50

50

100

Q kW

Pel 350kW Datum : 28.07.2008

___ !"## $%&&'(%)*'+,

___ -!"##$%&&'(%)*'+,

___ ./"##$%&&'(%)*'+,

5 10 15 20C Austrag

150

100

50

50

100

Q kW

Pel 350kW Datum : 13.08.2008

___ !"## $%&&'(%)*'+,

___ -!"##$%&&'(%)*'+,

___ ./"##$%&&'(%)*'+,

Fig. 3: Einfluss des Feuchtegehalts der Holzschnitzelund der Warmeverluste uber die Reaktorwand aufden Austrag an Holzkohle.

1E. R. Huff, ”Effect of Size, Shape, Density, Moisture and Furnace Wall Temperature on Burning Times of Wood Pieces,“in Fundamentals of Thermochemical Biomass Conversion, Ed. R. P. Overend, 1985.

ZHAW 34

ICP Institute of Computational Physics Research Report 2008

6.4 Prufstand fur thermo-mechanische Belastungstests vonBrennstoffzellen

Students: Matthias Kocher, Florian Baumberger

Category: Diplomarbeit MaschinentechnikPartner: Hexis AGMentoring: Thomas HockerPeriod: August 2008 – Oktober 2008

Zur Erhohung der Zuverlassigkeit und Lan-glebigkeit des Hexis Brennstoffzellensys-tems mussen die Materialeigenschaftender Brennstoffzellen weiter optimiert wer-den. Insbesondere mussen sie den thermo-mechanischen Beanspruchungen im Betriebbesser standhalten. Zur Uberprufung der Sta-bilitat der Zellen wurde deshalb ein Prufstand furthermomechanische Belastungstests entwick-elt. Die Idee ist, die Zellen bereits vor Einbauins System einem realistischen Belastungstestzu unterziehen.

Allgemeiner Maschinenbau

Energie- und Verfahrenstechnik EnVT

Matthias Kocher - 30 / 93- 03.10.08 Florian Baumberger

Mit

glie

d d

er

rch

er

Fa

chh

och

sch

ule

3.3.4 Instationäre Rechnung

In der instationären Berechnung wird ermittelt, wie lange es dauert, bis die Zelle die gefor-

derte Temperatur erreicht hat. Durch diese Simulation kann nur bedingt abgeschätzt werden, wie träge das System ist, da im realen Modell nicht mit einer konstanten Leistung

geheizt wird. Abbildung 18 zeigt die Veränderung der Temperaturen im Prüfstandaufbau nach 36s, 636s und 6885s.

Abbildung 18: Instationäre Temperaturverteilung

Veränderung der Zelltemperaturen

0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70

Radius r in mm

Te

mp

era

tur

in °

C

6

76

156

316

636

1242

2270

4036

6885

10800

Diagramm 10: Veränderung der Temperaturverteilung

Fig. 1: Temperaturplot aus thermischem FE-Modell zur Analyse des Aufheizens des thermo-mechanischen Zellenprufstands.

Im Prufstand mussen sich deshalb ahnlicheTemperaturverteilungen wie im Zellenstapel desrealen Systems einstellen lassen. Hierfur wurdezunachst ein ”virtueller Prufstand“ mit Hilfeder Finite Element Software SESES entwick-elt, siehe Abbildung 1. Das Modell beruck-sichtigt neben der Warmeleitung in den Fest-stoffen den Warmetransport uber thermischeStrahlung in Kavitaten sowie die Warmeabfuhruber die Oberflachen via Naturkonvektion.Durch Vergleich einer grossen Anzahl ver-schiedener Designs wurde via FE-Analyse ein

Konzept erarbeitet und anschliessend umge-setzt. Abbildung 2 zeigt den fertigen Prufstand.Daran ist besonders, dass sich hohe Temper-aturgradienten in den Zellen auch ohne eineerzwungene, externe Kuhlung einstellen lassen.

Allgemeiner Maschinenbau

Energie- und Verfahrenstechnik EnVT

Matthias Kocher - 52 / 93- 03.10.08 Florian Baumberger

Mit

glie

d d

er

rch

er

Fa

chh

och

sch

ule

6.5 Umsetzung Prüfstand

6.5.1 Aufbau des Prüfstands

Alle Teile des Prüfstands wurden bis am 02.10.2008 geliefert und konnten anschlies-

send zusammengebaut werden. Die innere, und die äussere Heizung wurden

von Hand gewickelt und in den darin vorge-sehenen Platz eingesetzt. Bei der inneren Heizung wurden die nach oben laufenden

Heizdrähte, für die thermische und elektrische Isolation, mit Keramikperlen ummantelt.

Die Gesamtlänge des eingebauten Drahtes der inneren Heizung ist 3900 mm. Bei der äusseren Heizung wurden 5500 mm

Draht verwendet. Die unterschiedlichen Drahtlängen der

Heizungen führen zu verschiedenen elektrischen Widerständen. Aus diesem Grund können die beiden Heizungen nur

umständlich über eine Steuerung geregelt werden.

Bild 1: 1 Zellen Prüfstand

Bild 3: innere Heizung Bild 2: gesamte Anlage

Fig. 2: Prufstand fur thermomechanische Belastung-stests von Hochtemperatur-Brennstoffzellen.

Um die Temperaturverteilung innerhalb derZelle in einer hohen Auflosung zu messen,wurden 20 Thermoelemente verbaut. Hier-bei handelt es sich um eine Spezialanfer-tigung der Winterthurer Firma Sawi. DieThermoelemente halten sehr hohen Tempera-turen stand und lassen sich uber Feingewindesehr genau positionieren. Aus den gemesse-nen Temperaturverteilungen auf den Zellenwurden in einem nachfolgenden Schritt dieinduzierten mechanischen Spannungen viathermo-mechanischer FE-Analyse berechnet.

ZHAW 35

Page 35: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

6.4 Prufstand fur thermo-mechanische Belastungstests vonBrennstoffzellen

Students: Matthias Kocher, Florian Baumberger

Category: Diplomarbeit MaschinentechnikPartner: Hexis AGMentoring: Thomas HockerPeriod: August 2008 – Oktober 2008

Zur Erhohung der Zuverlassigkeit und Lan-glebigkeit des Hexis Brennstoffzellensys-tems mussen die Materialeigenschaftender Brennstoffzellen weiter optimiert wer-den. Insbesondere mussen sie den thermo-mechanischen Beanspruchungen im Betriebbesser standhalten. Zur Uberprufung der Sta-bilitat der Zellen wurde deshalb ein Prufstand furthermomechanische Belastungstests entwick-elt. Die Idee ist, die Zellen bereits vor Einbauins System einem realistischen Belastungstestzu unterziehen.

Allgemeiner Maschinenbau

Energie- und Verfahrenstechnik EnVT

Matthias Kocher - 30 / 93- 03.10.08 Florian Baumberger

Mit

glie

d d

er

rch

er

Fa

chh

och

sch

ule

3.3.4 Instationäre Rechnung

In der instationären Berechnung wird ermittelt, wie lange es dauert, bis die Zelle die gefor-

derte Temperatur erreicht hat. Durch diese Simulation kann nur bedingt abgeschätzt werden, wie träge das System ist, da im realen Modell nicht mit einer konstanten Leistung

geheizt wird. Abbildung 18 zeigt die Veränderung der Temperaturen im Prüfstandaufbau nach 36s, 636s und 6885s.

Abbildung 18: Instationäre Temperaturverteilung

Veränderung der Zelltemperaturen

0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70

Radius r in mm

Te

mp

era

tur

in °

C

6

76

156

316

636

1242

2270

4036

6885

10800

Diagramm 10: Veränderung der Temperaturverteilung

Fig. 1: Temperaturplot aus thermischem FE-Modell zur Analyse des Aufheizens des thermo-mechanischen Zellenprufstands.

Im Prufstand mussen sich deshalb ahnlicheTemperaturverteilungen wie im Zellenstapel desrealen Systems einstellen lassen. Hierfur wurdezunachst ein ”virtueller Prufstand“ mit Hilfeder Finite Element Software SESES entwick-elt, siehe Abbildung 1. Das Modell beruck-sichtigt neben der Warmeleitung in den Fest-stoffen den Warmetransport uber thermischeStrahlung in Kavitaten sowie die Warmeabfuhruber die Oberflachen via Naturkonvektion.Durch Vergleich einer grossen Anzahl ver-schiedener Designs wurde via FE-Analyse ein

Konzept erarbeitet und anschliessend umge-setzt. Abbildung 2 zeigt den fertigen Prufstand.Daran ist besonders, dass sich hohe Temper-aturgradienten in den Zellen auch ohne eineerzwungene, externe Kuhlung einstellen lassen.

Allgemeiner Maschinenbau

Energie- und Verfahrenstechnik EnVT

Matthias Kocher - 52 / 93- 03.10.08 Florian Baumberger

Mit

glie

d d

er

rch

er

Fa

chh

och

sch

ule

6.5 Umsetzung Prüfstand

6.5.1 Aufbau des Prüfstands

Alle Teile des Prüfstands wurden bis am 02.10.2008 geliefert und konnten anschlies-

send zusammengebaut werden. Die innere, und die äussere Heizung wurden

von Hand gewickelt und in den darin vorge-sehenen Platz eingesetzt. Bei der inneren Heizung wurden die nach oben laufenden

Heizdrähte, für die thermische und elektrische Isolation, mit Keramikperlen ummantelt.

Die Gesamtlänge des eingebauten Drahtes der inneren Heizung ist 3900 mm. Bei der äusseren Heizung wurden 5500 mm

Draht verwendet. Die unterschiedlichen Drahtlängen der

Heizungen führen zu verschiedenen elektrischen Widerständen. Aus diesem Grund können die beiden Heizungen nur

umständlich über eine Steuerung geregelt werden.

Bild 1: 1 Zellen Prüfstand

Bild 3: innere Heizung Bild 2: gesamte Anlage

Fig. 2: Prufstand fur thermomechanische Belastung-stests von Hochtemperatur-Brennstoffzellen.

Um die Temperaturverteilung innerhalb derZelle in einer hohen Auflosung zu messen,wurden 20 Thermoelemente verbaut. Hier-bei handelt es sich um eine Spezialanfer-tigung der Winterthurer Firma Sawi. DieThermoelemente halten sehr hohen Tempera-turen stand und lassen sich uber Feingewindesehr genau positionieren. Aus den gemesse-nen Temperaturverteilungen auf den Zellenwurden in einem nachfolgenden Schritt dieinduzierten mechanischen Spannungen viathermo-mechanischer FE-Analyse berechnet.

ZHAW 35

Page 36: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

6.5 High Performance Grating Spectrometer

Students: Martin Neukom, Simon Hubscher

Category: Project Work in MechatronicsMentoring: Nils A. Reinke, Christoph StammPeriod: Oktober 2008 – Januar 2009

In this work a grating spectrometer with a fo-cal length of 30cm is constructed from scratch.The exploded assembly drawing in Fig. 1 showsthe external and internal compartments of thedevice. The spectrometer offers both fixed en-trance slits ranging from 20µm to 1mm and anadjustable entrance slit. An additional mount al-lows to connect optical fibers to the device. Aholographic grating with 1200 lines/mm is usedas a dispersive element. The concave cur-vature of the grating defines the focal lengthof the device making additional optical compo-nents needless. The curved grating thus allowsa very compact design of the device and re-duces optical aberrations and losses. The spec-trum is detected with a highly sensitive CCD linesensor comprising 3000 pixels.

Fig. 1: Exploded assembly drawing of the spectrom-eter comprising the internal and external body, themounted holographic grating and the CCD line sen-sor.

The inner shielding of the device encapsulates

the CCD sensor from optical noise. The de-tectable spectrum ranges from 450nm to 750nmat a maximum spectral resolution of 0.1 nm de-pending on the width of the entrance slit. Thesensor is both powered and controlled by anUSB 2.0 interface. The developed LabView soft-ware allows calibrating both spectral positionsand absolute intensities by a build-in polyno-mial fitting routine and reading calibration datafrom a file, respectively. A screenshot of the de-veloped LabView software is displayed in Fig.2. The software offers the basic functionalityof background subtraction and setting the inte-gration time. During long acquisition processesthe user is informed about remaining recordingtime. Transient processes can be recorded at aspeed of 300 scans per second which can beaccelerated up to 3000 scans per second usingan alternative line sensor.

Fig. 2: Spectrometer software developed in LabView.

This high performance spectrometer offers highresolution, high sensitivity and high speed mak-ing it competitive with upper-class commercialproducts. This spectrometer will thus be a valu-able tool for prospective projects in the field ofthin-film optical spectroscopy and bio-sensingapplications.

ZHAW 36

Page 37: ZHAW ICP Research Report 2008

Chapter 7

Appendix

37

Page 38: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

7.1 Publications

• Light extraction and optical loss mechanisms in organic light-emitting diodes: influence ofthe emitter quantum efficiency: S. Nowy, B. Krummacher, J. Frischeisen, N. A. Reinke, W.Brutting, J. Appl. Phys. 104, 123109 (2008).

• Surface plasmon resonance sensor utilizing an integrated organic light emitting diode: J.Frischeisen, C. Mayr, N. A. Reinke, S. Nowy, W. Brutting, Opt. Express 16, 18426-18436(2008).

• B. Ruhstaller, T. Flatz, D. Rezzonico, M. Moos, N. A. Reinke, E. Huber, R. Hausermann, B.Perucco: Comprehensive simulation of light-emitting and light-harvesting organic devices,Proceedings of the SPIE Conference, Vol. 7051, ISBN 9780819472717, San Diego, USA(2008).

• R. Hausermann, N. A. Reinke, E. Huber, B. Ruhstaller: Towards Comprehensive Simulationof Organic Solar Cells, Proc. EOS Annual Meeting (2008).

• N. A. Reinke, B. Perucco, M. Moos, W. Brutting, B. Ruhstaller: Emission Zone Extraction andAnalysis of Loss Channels in Organic Light-emitting Devices, Proc. EOS Annual Meeting(2008).

• S. Nowy, N. A. Reinke, J. Frischeisen, and W. Brutting: Light extraction and optical lossmechanisms in organic light-emitting diodes, Proc. SPIE (2008).

• J. Frischeisen, N. Reinke, C. Ostermayr, J. Neumann, S. Nowy, W. Brutting: Surface plasmonresonance sensor based on a planar polychromatic OLED light source, Proc. SPIE (2008).

• K. Steinkamp, J.O. Schumacher, F. Goldsmith, M. Ohlberger, C. Ziegler: A non-isothermalPEM fuel cell model including two water transport mechanisms in the membrane, Specialissue of the Journal of Fuel Cell Science and Technology, Volume 5, Issue 1, 011007 (2008).

• A. Bieberle-Hutter, D. Beckel, A. Infortuna, U. Muecke, J. Rupp, L. Gauckler, S. Rey-Mermet,P. Muralt, N. Hotz, M. Stutz, N. Bieri, D. Poulikakos, P. Muller, A. Bernard, R Gmur, T. Hocker:A micro-solid oxide fuel cell system for battery replacement, Journal of Power Sources, 177,123ff (2008).

• H. Schwarzenbach: Modellbildung in der Mikrosystemtechnik - Kreativitat ist gefragt, SWISSENGINEERING STZ Nummer 7/8 (2008).

• H. Schwarzenbach: Multiphysikalische Werkzeuge im Kleinstformat, Technica 7 (2008).

• R. Alther, R. Weiss, E. Lang, T. Hocker: Self-opening closure with air inlet channel for com-posite packaging or for container necks to be sealed with foil material, International PatentApplication PCT/CH2008/000339 (2008).

7.2 Conferences and Workshops

• E. Huber, R. Hausermann, N. A. Reinke, B. Ruhstaller, T. Flatz, M. Moos: Mobility Models forOLED Device Simulation, International Krutyn Summer School, Krutyn, Poland, May (2008).

• R. Hausermann, E. Huber, N. A. Reinke, B. Ruhstaller, T. Flatz, M. Moos: Design Rulesfor Organic Photovoltaic Cells: Reflective Electrode Material, Olla Summer School, Krutyn,Poland, May/June (2008).

• N. A. Reinke: OLED Light Outcoupling, 1st International User Workshop, Winterthur, Switzer-land, July (2008).

• R. Hausermann: Organic Photovoltaic Cell Simulation, 1st International User Workshop,Winterthur, Switzerland, July (2008).

ZHAW 38

Page 39: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

• B. Ruhstaller, T. Flatz, D. Rezzonico, M. Moos, N. A. Reinke, E. Huber, R. Hausermann, B.Perucco: Comprehensive simulation of light-emitting and light-harvesting organic devices,SPIE Optics & Photonics Conference, San Diego, USA, August 2008.

• B. Ruhstaller, T. Flatz, D. Rezzonico, M. Moos, N. A. Reinke, E. Huber, R. Hausermann, B.Perucco: Comprehensive simulation of light-emitting and light-harvesting organic devices,IBM Almaden Research Center, San Jose, USA, August 2008.

• R. Hausermann, N. A. Reinke, E. Huber, B. Ruhstaller: Towards a Comprehensive Simulationof Organic Solar Cells , EOS Annual Meeting, Paris, France, September/October (2008).

• N. A. Reinke, R. Hausermann, D. Rezzonico, B. Perucco, E. Huber, B. Ruhstaller: Emis-sion Zone Extraction and Analysis of Loss Channels in Organic Light-emitting Devices, EOSAnnual Meeting, Paris, France, September/October (2008).

• D. Rezzonico, B. Ruhstaller, T. Flatz, M. Moos, N. A. Reinke, E. Huber, R.Hausermann, B.Perucco: Comprehensive simulation of light-emitting and light-harvesting organic devices,Organic Semiconductor Conference, Frankfurt, Germany, October 2008.

• R. Hausermann, N. A. Reinke, E. Huber, B. Ruhstaller: Opto-Electronic Simulation of OrganicSolar Cells , ThinPV Workshop: “A look inside solar cells”, Ascona, Switzerland, November(2008).

• R. Gmur, T. Hocker: Thermal-Flow Modeling of the OneBat Micro-SOFC System, 5th FuelCell Research Symposium on Modeling and Experimental Validation, ZHAW, Winterthur,2008.

• T. Hocker, B. Iwanschitz, J. Sfeir, A. Mai, R. Denzler: Model-Based Analysis of Degrada-tion Phenomena and Performance Losses in Hexis SOFC-Stacks, 5th Fuel Cell ResearchSymposium on Modeling and Experimental Validation, ZHAW, Winterthur, 2008.

• J. Eller, J.O. Schumacher, G. Sartoris, M. Roos: Computationally Efficient Simulation of PEMFuel Cells and Stacks, 5th Fuel Cell Research Symposium on Modeling and ExperimentalValidation, ZHAW, Winterthur, 2008.

• A. Bieberle-Hutter, D. Beckel, A. Infortuna, U. Muecke, J. Rupp, L. Gauckler, S. Rey-Mermet,P. Muralt, N. Hotz, M. Stutz, N. Bieri, D. Poulikakos, P. Muller, A. Bernard, R Gmur, T. Hocker:Micro-Solid Oxide Fuel Cells: From Thin Film Membranes to a Micro-Fuel Cell System, 8thEuropean Fuel Cell Forum, Lucerne, Switzerland, 2008.

• AJ. Sfeir, A. Mai, B. Iwanschitz, U. Weissen, R. Denzler, D. Haberstock, T. Hocker, M. Roos:Status of SOFC Stack and Material Development at Hexis, 8th European Fuel Cell Forum,Lucerne, Switzerland, 2008.

• H. Schwarzenbach: Uberblicksvortrag Numerische Modellierung, electrosuisee Fachtagungder ITG Fachgruppe Hardeware-Technologie: Elektronik der Zukunft, Trilogie MEMS - NANO- ORGANIC, September 2008.

• Th. Hocker: Mit Hochtemperatur-Brennstoffzellen effizient und dezentral Strom und Warmegewinnen, Nationaler Tag der Technik zum Thema Energieeffizienz, ZHAW, Winterthur, 5.November 2008.

• N. A. Reinke: Effiziente Lichtquellen der Zukunft, Nationaler Tag der Technik zum ThemaEnergieeffizienz, ZHAW, Winterthur, 5. November 2008.

ZHAW 39

Page 40: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

7.3 Prizes and Awards

• Thomas Hocker received the title ”Professor ZFH” from the board of the Zurich University ofApplied Sciences.

• Roger Hausermann received the Springer Presentation Award for Students for his talk ”To-wards a Comprehensive Simulation of Organic Solar Cells” at the European Optical Societymeeting in Paris at the end of September.

• Nils A. Reinke received the ”Augsburger Universitatspreis” for his Ph.D. thesis entitled ”Pho-tophysical Processes and Light Extraction in Organic LEDs” at the University of Augsburg.

ZHAW 40

Page 41: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

7.4 Teaching

• ”Physik 1 & 2” fur Unternehmensinformatiker (PhUI 1 & 2) - N. A. Reinke

• ”Physik 2” fur Unternehmensinformatiker (PhUI 2) - B. Ruhstaller

• ”Applied Photonics” (Master of Science in Engineering) - B. Ruhstaller

• ”Fluid-Thermodynamik“ fur Maschinentechniker (FTh1) - Th. Hocker

• ”Mensch, Technik, Umwelt” fur Maschinentechniker (MTU) - Th. Hocker

• ”Diskrete und Numerische Mathematik” fur Mechatroniker (DNM2) - H. Schwarzenbach

• ”Signale und Systeme 1 & 2” (SiSy 1 & 2) - J.O. Schumacher

• ”Physik 1 & 2” fur Wirtschaftsingenieure (PhWI 1& 2) - M. Roos

• ”Material Properties of Crystals / Tensors for Engineers” (Master of Science in Engineering)- M. Roos

ZHAW 41

Page 42: ZHAW ICP Research Report 2008

ICP Institute of Computational Physics Research Report 2008

7.5 ICP Team Members

The ICP team members as of December 2008 are listed below.

Name Title FunctionEller, Jens Dipl. Rech. Wiss. ETH Research AssistantGmur, Roman Dipl. Ing. FH Research AssociateHausermann, Roger Dipl. Phys. ETH Research Assistant, PhD studentHocker, Thomas Prof. Dr., Dipl. Verfahrensingenieur LecturerHuber, Evelyne Dipl. Rech. Wiss. ETH Research Assistant, PhD studentLindner, Markus Dipl. Ing. FH MSE StudentPerucco, Benjamin Dipl. Ing. FH Research Assistant, MSE StudentReinke, Nils A. Dr. rer. nat., Dipl.-Phys. LecturerRoos, Markus Prof. Dr., Dipl. Phys. ETH LecturerRuhstaller, Beat Prof. Dr., Dipl. Phys. ETH, e-MBA Lecturer, Head of the ICPSafa, Yasser Dr. es sc., MSc. Research AssociateSartoris, Guido Dr., Dipl. Phys. ETH Research AssociateSchmid, Matthias Dr. es sc., Dipl. Phys. ETH Research AssociateSchwarzenbach, Hansueli Prof. Dr., Dipl. Math. ETH LecturerSchumacher, Jurgen Dr. rer. nat., Dipl.-Phys. LecturerTiefenauer, Andreas Dipl. Ing. FH Research Assistant, MSE StudentToniolo, Lilian Administrative Assistant

The following visiting scientists were staying at the ICP for some time during 2008:

• Jose Montero, Ph.D. candidate, Universita Jaume I, Castellon, Spain: February - March 2008

• Kemal Celebi, M.Sc., MIT, Boston, USA: June - September 2008

7.6 Location and Contact Info

Institute of Computational PhysicsZurich University of Applied SciencesWildbachstrasse 21 (TK)8401 WinterthurSwitzerland

Prof. Dr. Beat RuhstallerPhone +41 58 934 78 36Fax +41 58 934 77 97

www.icp.zhaw.chE-Mail: [email protected]

Administration:Lilian TonioloPhone +41 58 934 73 06E-Mail: [email protected]

ZHAW 42

Page 43: ZHAW ICP Research Report 2008
Page 44: ZHAW ICP Research Report 2008

Zürcher Hochschule für Angewandte Wissenschaften

School of Engineering

ICP Institute ofComputational Physics

Technikumstrasse 9PostfachCH-8401 Winterthur

Telefon +41 58 934 71 71Fax +41 58 935 71 71E-Mail: [email protected] www.icp.zhaw.ch