Newsletter Year 11-4

Year 8 n° 4 Winter 2011

Multi-objective Optimization with modeFRONTIER Applied to Systems Biology

EnginSoft CAE Conference 2011 Welcomes an Audience of 600 CAE users EnginSoft ha proposto una tavola rotonda sulla competitività d’impresa presso il nuovo centro di ricerca

Synergy between LS-DYNA and modeFRONTIER to Predict Low Velocity Impact Damage on a Composite Plate

Structural Optimization of a Car-body High Speed Train An Innovative Analysis and Design Methodology

Electromagnetic issues for a IEEE 1902.1 “RuBee” tag dipped in a fiber/composite laminate

FSO and Shuttle Tanker in Tandem Configuration Hydrodynamic Analysis

Newsletter EnginSoft Year 8 n°4 -

3

EnginSoft Flash CIRA, the Italian Aerospace Research Centre, For many of us, December is a time for illustrates the Synergy between LS-DYNA and reflection, for harvesting the fruit of our modeFRONTIER to predict low velocity work and our personal efforts of the year. impact damage on a composite plate. We Our Simulation and CAE environments hear from EnginSoft Nordic in Sweden on almost constantly see new developments, how multi-objective optimization is being upcoming software releases and changes. applied to systems biology. Here, we We are asked to be always ready for the encourage our readers to watch the movie “new”. While this is sometimes a challenge “Insulin Signaling (SignalPathways)” via the for most of us, every year also brings many link provided! new human encounters. In our fields of business, we can consider ourselves lucky We are pleased to introduce our customer to have the opportunity to meet people and ANSYS user the company Almacis, and from the CAE community, from around the AMD, our partner in the area of High world. While we learn about new and Ing. Stefano Odorizzi Performance Computing. different technologies, the human, the EnginSoft CEO and President Digimat is a powerful software for material engineer, its broad knowledge and modeling which is now distributed in Italy by experiences, always remain at the core of EnginSoft. More software news covers the LIONsolver by our attention.

Reactive Research, NVIDIA’s Tesla GPU, EnginSoft’s By sharing our knowledge, especially on occasions such as activities for composite materials with ESAComp and the EnginSoft International Conference, we help to shape ANSYS Composite Prep/Post as well as MAGMA’s release the future path of CAE and to support the next generation 5.2. The powerful Sculptor tool allows users to of CAE engineers. parameterize any mesh based on arbitrary cubic bezier In this Newsletter, we speak about the EnginSoft and control points. Sculptor was recently presented by ANSYS Italian Conferences 2011, the two annual events EnginSoft GmbH at the ANSYS Conference and 29th that offer one of the major knowledge platforms to CAE CADFEM Users’ Meeting in Stuttgart. users in Europe and beyond. ANSYS is the provider of the world’s leading software for engineering simulation and Furthermore, we hear about Gruppo Ferroli’s project with EnginSoft’s number 1 partner. EnginSoft and ANSYS were EnginSoft, the recent introduction of the BENIMPACT delighted to welcome 600 delegates to Verona on 20th project in China and about the Minimaster and the and 21st October, to a wealth of topics on today’s use of Training Programs of TCN and EnginSoft. simulation and design tools.

Our Japan Column tells us about the CAE University while In this issue, we also inform our readers about the Round some of the activities of JANCAE, The Japan Association Table Meeting of 100 Top Managers on the occasion of the for Nonlinear CAE, are explained to us in the article by opening of EnginSoft’s Research Center in the Scientific Hideo Takizawa.

Technology Park ”Kilometro Rosso”. The use of ANSYS Please mark your diary for the modeFRONTIER Users' Maxwell v.14 is shown in the article on electromagnetic Meeting 2012, which will be sponsored by ESTECO and issues for a IEEE 1902.1 “RuBee” tag dipped in a take place on 21st and 22nd of May 2012 in Trieste. fiber/composite laminate.

The capabilities of We hope that you enjoy reading the articles on the modeFRONTIER are described in AnsaldoBreda’s work for following pages of this last Newsletter of 2011. We always the structural optimization of a car-body high speed train. welcome your thoughts, your feedback as well as your Our readers also hear about the use of ANSYS AQWA and ideas for future publications! the ANSYS Workbench platform for the structural verification of the FSO Mooring System complemented by EnginSoft and the Editorial Team wish you and your EnginSoft’s broad experiences as a partner to the Oil&Gas families a very happy, healthy and a prosperous New Year industries.

2012!

The Università degli Studi di Ferrara presents their work with ANSYS CFX 13.0 while University of Debrecen Hungary Stefano Odorizzi informs us of how Grapheur can help its users with multiple criteria decision- making problems.

Editor in chief

4 - Newsletter EnginSoft Year 8 n°4

Sommario - Contents EVENTS

6 8 10

EnginSoft CAE Conference 2011: 600 partecipanti all’annuale appuntamento EnginSoft CAE Conference 2011 welcomes an audience of 600 CAE users EnginSoft ha proposto una tavola rotonda sulla competitività d’impresa presso il nuovo centro di ricerca

CASE STUDIES

12 15 18 19 20 23 26 29

Electromagnetic Issues for a IEEE 1902.1 “RuBee” Tag Dipped in a Fiber/Composite Laminate Structural Optimization of a Car-body High Speed Train - An Innovative Analysis and Design Methodology FSO and Shuttle Tanker in Tandem Configuration Hydrodynamic Analysis Finalized to the Structural Verification of the FSO Mooring System FEM analysis in Oil&Gas Industry Numerical Analysis of a Micro Gas Turbine Combustor Fed by Liquid Fuel Reconsidering the Multiple Criteria Decision Making Problems of Construction Workers Using Grapheur Synergy between LS-DYNA and modeFRONTIER to Predict Low Velocity Impact Damage on Composite Plate Multi-objective Optimization with modeFRONTIER Applied to Systems Biology

TESTIMONIAL

31

Eccellenza tecnologica e qualità: Almacis

SOFTWARE/HARDWARE NEWS

32 33 34 36 37

CAE Simulations and Innovations within the High Performance Computing HPC DIGIMAT per la modellazione avanzata dei materiali LIONsolver: Learning and Intelligent Optimization GPU Accelerated Engineering with ANSYS EnginSoft continua l’attività sui materiali compositi

EVENTS

38 39 39 40

EnginSoft presenterà la release 5.2 di MAGMA a METEF 2012 La simulazione di processo nella progettazione di radiatori modeFRONTIER Users’ Meeting 2012 EnginSoft GmbH Silver Sponsor at the ANSYS Conference & 29th CADFEM Users’ Meeting 2011

The EnginSoft Newsletter editions contain references to the following products which are trademarks or registered trademarks of their respective owners: ANSYS, ANSYS Workbench, AUTODYN, CFX, FLUENT and any and all ANSYS, Inc. brand, product, service and feature names, logos and slogans are registered trademarks or trademarks of ANSYS, Inc. or its subsidiaries in the United States or other countries. [ICEM CFD is a trademark used by ANSYS, Inc. under license]. (www.ansys.com) modeFRONTIER is a trademark of ESTECO srl (www.esteco.com) Flowmaster is a registered trademark of The Flowmaster Group BV in the USA and Korea. (www.flowmaster.com) MAGMASOFT is a trademark of MAGMA GmbH. (www.magmasoft.de)

ESAComp is a trademark of Componeering Inc. (www.componeering.com) Forge and Coldform are trademarks of Transvalor S.A. (www.transvalor.com) AdvantEdge is a trademark of Third Wave Systems (www.thirdwavesys.com)

.

LS-DYNA is a trademark of Livermore Software Technology Corporation. (www.lstc.com) SCULPTOR is a trademark of Optimal Solutions Software, LLC (www.optimalsolutions.us) Grapheur is a product of Reactive Search SrL, a partner of EnginSoft (www.grapheur.com) For more information, please contact the Editorial Team

Newsletter EnginSoft Year 8 n°4 -

RESEARCH AND TECHNOLOGY TRANSFER

42

BENIMPACT Suite has landed in China

TRAINING

43

Alta formazione: TCN punta ad una specializzazione sempre più avanzata

JAPAN CAE COLUMN

44 46

CAE Seminars in Japan “CAE UNIVERSITY” NPO Activity for Implementation of Anisotropic Elasto-plastic Models into Commercial FEM Codes

5

Newsletter EnginSoft Year 8 n°4 -Winter 2011 To receive a free copy of the next EnginSoft Newsletters, please contact our Marketing office at: newsletter@enginsoft.it All pictures are protected by copyright. Any reproduction of these pictures in any media and by any means is forbidden unless written authorization by EnginSoft has been obtained beforehand. ©Copyright EnginSoft Newsletter.

Advertisement For advertising opportunities, please contact our Marketing office at: newsletter@enginsoft.it

EnginSoft S.p.A. EnginSoft Event Calendar Corsi di addestramento software 2012

PAGE 8: ENGINSOFT CAE CONFERENCE 2011 WELCOMES AN AUDIENCE OF 600 CAE USERS

PAGE 12: ELECTROMAGNETIC ISSUES FOR A IEEE 1902.1 “RUBEE” TAG DIPPED IN A FIBER COMPOSITE LAMINATE

24126 BERGAMO c/o Parco Scientifico Tecnologico Kilometro Rosso - Edificio A1, Via Stezzano 87 Tel. +39 035 368711 • Fax +39 0461 979215 50127 FIRENZE Via Panciatichi, 40 Tel. +39 055 4376113 • Fax +39 0461 979216 35129 PADOVA Via Giambellino, 7 Tel. +39 49 7705311 • Fax 39 0461 979217 72023 MESAGNE (BRINDISI) Via A. Murri, 2 - Z.I. Tel. +39 0831 730194 • Fax +39 0461 979224 38123 TRENTO fraz. Mattarello - Via della Stazione, 27 Tel. +39 0461 915391 • Fax +39 0461 979201 www.enginsoft.it - www.enginsoft.com e-mail: info@enginsoft.it

COMPANY INTERESTS ESTECO srl 34016 TRIESTE Area Science Park • Padriciano 99 Tel. +39 040 3755548 • Fax +39 040 3755549 www.esteco.com CONSORZIO TCN 38123 TRENTO Via della Stazione, 27 - fraz. Mattarello Tel. +39 0461 915391 • Fax +39 0461 979201 www.consorziotcn.it • www.improve.it EnginSoft GmbH - Germany EnginSoft UK - United Kingdom EnginSoft France - France EnginSoft Nordic - Sweden Aperio Tecnologia en Ingenieria - Spain www.enginsoft.com

ASSOCIATION INTERESTS

PAGE 15: STRUCTURAL OPTIMIZATION OF A CAR-BODY HIGH SPEED TRAIN AN INNOVATIVE ANALYSIS AND DESIGN METHODOLOGY

NAFEMS International www.nafems.it www.nafems.org TechNet Alliance www.technet-alliance.com RESPONSIBLE DIRECTOR Stefano Odorizzi - newsletter@enginsoft.it PRINTING Grafiche Dal Piaz - Trento The EnginSoft NEWSLETTER is a quarterly magazine published by EnginSoft SpA

Autorizzazione del Tribunale di Trento n° 1353 RS di data 2/4/2008

50 52

6 - Newsletter EnginSoft Year 8 n°4

EnginSoft CAE Conference 2011: 600 partecipanti all’annuale appuntamento La Fiera di Verona ha ospitato l’edizione 2011 del maggiore appuntamento in Italia dedicato al calcolo scientifico: l’EnginSoft International Conference, CAE Technologies for Industry e l’ANSYS Italian Conference.

Oltre 600 i congressisti, esperti ed opinion leader in metodi e tecnologie CAE, che il 20 e 21 Ottobre scorso si sono incontrati, presso il Centro Conferenze del polo fieristico di Verona.

Molte le aziende presenti, tra cui: Ansaldo, Piaggio, Magneti Marelli, Avio, Tetra Pak, Ferrari, Iveco, ENI, a dimostrazione dell’utilizzo crescente del CAE in ambito industriale.

Tra gli obiettivi della Conference vi è stato quello di offrire ai partecipanti una visione d’insieme del comparto, attraverso il contributo di esponenti del mondo dell'industria, dell'università e della ricerca e dai numerosi sviluppatori di tecnologie intervenuti.

“La Conference – ha spiegato Stefano Odorizzi, CEO di EnginSoft – è nata nel 1984 quando le tecnologie in fatto di sperimentazione virtuale erano solo oggetto di ricerca da parte delle università. Convinti che queste tecnologie avrebbero avuto un’evoluzione importante, abbiamo deciso di abbracciare la sfida e oggi continuiamo a perseguire l’obiettivo di trasferire agli operatori del settore le informazioni e le conoscenze relative a questi ambienti di simulazione e supporto alla progettazione”.

Dopo la sessione plenaria di apertura che, oltre alla “Vision” da parte del Vice Presidente di ANSYS Inc., ha ospitato un mini simposio dedicato alla tematica del geo-modeling, l’evento è continuato su sessioni parallele, ognuna delle quali

Fig. 2 - Scorcio della sala conferenze di Verona nel corso di uno dei workshop.

Fig. 1 - Stefano Odorizzi - CEO di EnginSoft - in sessione plenaria.

focalizzata su una macroarea tecnologica o applicativa: meccanica, fluidodinamica, ottimizzazione, simulazione di processo, compositi, ecc.

Di grande appeal sui partecipanti e di interesse perchè d’attualità, l’esperienza presentata da Ansaldo Energia di Genova in tema High Performance Computing. Stefano Santucci, IT manager di Ansaldo, ha illustrato le ragioni della migrazione da una struttura formata da sole workstation ad un cluster in cui l’hardware distribuito e HPC non solo convivono felicemente ma si integrano in un tuttuno estremamente efficiente sia in termini di performance di calcolo che di ritorno dell’investimento per tutta l’azienda.

Newsletter EnginSoft Year 8 n°4 -

Nel corso dei lavori relativi alla sessione sulla simulazione meccanica sono stati presentati alcuni importanti progetti tra i quali lo sviluppo di un’innovativo sistema di contenimeto di argon liquido, commissionato dal CERN di Ginevra, che consentirà di approfondire la ricerca scientifica sui neutrini. EnginSoft ha inoltre illustrato il progetto di un veicolo filoguidabile, realizzato in collaborazione con WASS, finalizzato all’esplorazione subacquea sino a quattromila metri di profondità.

La sessione dedicata alla simulazione CFD (Computational Fluid Dynamics) ha, invece, reso evidente quello che è oggi, rispetto al passato, il ruolo centrale del progettista che, attraverso sofisticati strumenti di simulazione di cui può disporre, ha l’opportunità di focalizzarsi principalmente sull’aspetto ingegneristico del problema, delegando al software l’onere di governare gli aspetti matematici di base.

Progettare in CFD oggi si traduce nella necessità di avere: efficienti funzionalità di dialogo con i sistemi CAD, procedure automatiche di meshing e parametrizzazione del modello.

Tema centrale della sessione dedicata all’ottimizzazione è stata l’analisi dello stato dell’arte sulla simulazione multiobiettivo, tematica molto utilizzata in ambito automotive, dimostrato dalle testimonianze di Ferrari, Iveco e Continental.

Novità e successo di pubblico anche per il workshop dal titolo “La progettazione delle strutture in materiale composito” coordinato da Marco Perillo e dal suo team di ingegneri. Scopo del seminario è stato quello di condividere lo stato dell’arte dei metodi di progettazione e degli strumenti di analisi strutturale sia sul piano teorico/concettuale, sia sul piano applicativo.

A dimostrazione di molte tematiche verticali sostenute da EnginSoft, grazie anche all’esperienza nel progetto BENImpact, è stato inserito nel programma un workshop dedicato all’utilizzo del CAE in campo ECO-Building e progettazione sostenibile, il riscontro è stato notevolmente positivo e ha dimostrato l’ottima integrazione del CAE anche nelle tematiche “di frontiera”.

L’attività congressuale, inoltre, è stata affiancata da un’area espositiva, in cui quasi 30 tra le più importanti software house CAE, sviluppatori hardware e di applicazioni

7

complementari hanno condiviso con i partecipanti le novità relative ai loro prodotti.

Particolarmente emozionante la Cena di Gala organizzata presso il vicino Museo dell’Auto e della Tecnica Nicolis. Qui i visitatori, prima delle portate, hanno potuto osservare automobili, motociclette e oggetti unici da collezione di epoche differenti.

“Le tecnologie di simulazione rivoluzioneranno i processi progettuali attualmente adottati dalle aziende manifatturiere” ha concluso il CEO di EnginSoft. “Oggi si dice che queste tecnologie si integrano nel processo progettuale; in futuro oramai prossimo, queste tecnologie diventeranno il processo progettuale”.

Con questo messaggio diamo ai lettori appuntamento all’edizione 2012 della CAE Conference EnginSoft, sperando di accrescere ulteriormente la community di analisti e imprenditori che credono nell’innovazione attraverso l’utilizzo delle tecnologie di sperimentazione virtuale.

Per ulteriori informazioni: Luisa Cunico, EnginSoft info@enginsoft.it www.caeconference.com

ATTI DELLA CONFERENZA 2011 Sono disponibili in download gli atti della Conferenza EnginSoft 2011 all’indirizzo: www.enginsoft.com/proceedings2011

Fig. 3 - L’area espositiva in cui i congressisti hanno avuto l’opportunità di dialogare direttamente con i produttori di tecnologia presenti in sala.

8 - Newsletter EnginSoft Year 8 n°4

EnginSoft CAE Conference 2011 welcomes an audience of 600 CAE users The Exhibition Centre in Verona (Verona Fiere) hosted the 2011 edition of the major event in Italy on simulation based engineering and sciences, the EnginSoft International Conference, CAE Technologies for Industry, and the ANSYS Italian Conference.

EnginSoft and ANSYS had the great pleasure of welcoming over 600 attendees, among them many CAE experts and opinion leaders, to the Congress Centre in Verona on 20th and 21st October.

Representatives of large companies participated and contributed to the conference program as well: Ansaldo, Piaggio, Magneti Marelli, Avio, Tetra Pak, Ferrari, Iveco, and ENI, to name just a few. Their involvement underlined how CAE technologies are being used more and more in industry.

One of the goals of the Conference was to offer the participants an overall view of such technologies with presentations from industry, universities, research organizations, and technology developers.

“The Conference – explained Stefano Odorizzi, CEO of EnginSoft – was organized for the first time in 1984, when technologies in the field of virtual prototyping were just studied in universities. At the time, we saw great evolution, and this is what made us decide to invest in these technologies. Today, our goal is to transfer as much information and knowledge as possible about these simulation and design tools to the experts in this field”.

Fig. 2 - Welcome desk at EnginSoft area.

Fig. 1 - Swaminathan Subbiah - Vice President, Corporate Product and Market Strategy at ANSYS - during his speach talking about future developments.

The Plenary Session that opened the event, featured the “Vision” of the Assistant Director of ANSYS Inc. and a Mini-Symposium on geo-modeling. Later on in the afternoon, the program offered to the audience a number of parallel sessions focused on different technological fields: mechanics, fluid-dynamics, optimization, process simulation, composites, etc.

One of the particularly captivating presentations on current topics was the contribution by Ansaldo Energia of

Newsletter EnginSoft Year 8 n°4 -

9

Genova on High Performance Computing. Stefano Santucci, the IT manager of Ansaldo, explained the reasons why the company has left a structure with only workstations for a structure with a cluster, where the distributed hardware and the HPC were perfectly integrated thus generating an efficient computation performance and ROI for the company.

In the session about mechanical simulation, some important projects were presented, such as the development of an innovative storage system for liquid argon - committed by CERN (European Organization for Fig. 3 - Some beauties inside of the Nicolis Museum - Verona.

Nuclear Research) in Geneva – occasion, before the dinner started, our guests from that allows to perform in depth studies on neutrinos. On around the world enjoyed a guided tour of the large this occasion, EnginSoft explained the project of a wireexhibition rooms of the museum. guided vehicle, implemented with WASS, for underwater The CEO of EnginSoft closed the Conference saying that exploration activities of up to 4000 m under sea level.

“Simulation technologies will radically change the design The CFD session stressed the central role of the designer processes currently used in manufacturing companies. nowadays, compared to the past. Today, we can focus on Now we are saying that such technologies are integrated the engineering side of the problem, thanks to in the design process; but in the next years they will be sophisticated simulation tools, by entrusting the the design process itself”. management of the basic mathematical processes to the With this message in mind, we ask our attendees and software. Designing in CFD means: effective connections readers to keep an eye out for the 2012 edition of the with CAD systems, automatic mesh procedures and model EnginSoft International CAE Conference parameterization. The session about optimization www.CAEconference.com hoping that the Virtual emphasized the state-of-the-art of multi-objective Prototyping Community will grow further and further until simulation, a topic commonly discussed in the automotive we meet again! field – as Ferrari, Iveco and Continental assured us.

The workshop titled “The design of structures in composite materials”, managed by Marco Perillo and his For more information: team of engineers, also turned out to be a great success.

Luisa Cunico, EnginSoft The workshop’s aim was to share the state-of-the-art of info@enginsoft.it the diverse design methods and the structural analysis www.caeconference.com tools, both from a theoretical/conceptual and applicative level.

Another interesting workshop was connected to the CONFERENCE PROCEEDINGS 2011 BENImpact Project and the ECO-Building field. The results 2011 Conference Proceeding are now avaliable were incredibly positive and demonstrated how perfectly to download on: CAE is integrated in the “frontier” topics. www.enginsoft.com/proceedings2011 An important aspect of the annual event is the exhibition area. This year, nearly 30 of the most well-known CAE software houses showcased their hardware and software products. The conference attendees could hear about the latest developments and news in personal talks with some of the developers.

Finally, another highlight was the Conference Gala Dinner, held at the Nicolis’ Museum of Cars, Technology and Mechanics, which houses a private collection of vintage cars and motorbikes of Mr. Luciano Nicolis. On this

10 - Newsletter EnginSoft Year 8 n°4

EnginSoft ha proposto una tavola rotonda sulla competitività d’impresa presso il nuovo centro di ricerca Il 24 Novembre scorso si è tenuta a Bergamo, in occasione dell’inaugurazione del nuovo Centro di Ricerca EnginSoft presso il Parco Scientifico Tecnologico “Kilometro Rosso”, una Tavola Rotonda dal titolo “Lean Design e Competitività d’Impresa - Innovazione e moderni strumenti per il management strategico”. All’evento, al quale hanno partecipato oltre 100 Top Manager delle più importanti imprese manifatturiere italiane mentre al tavolo dei relatori si sono seduti: Roberto Formigoni (Presidente Regione Lombardia), Alberto Bombassei (Vice Presidente Confindustria), Antonello Briosi (Vice Presidente Confindustria Trento), Mirano Sancin (Direttore Generale e Consigliere Delegato del Parco Scientifico Tecnologico Kilometro Rosso), Massimo Egidi (Presidente della Fondazione Bruno Kessler), Giancarlo Michellone (già Presidente di Area Science Park di Trieste e ora Presidente GMC Consulting), Marie Christine Oghly (Presidente MEDEF, Parigi), Sergio Savaresi (professore al Politecnico di Milano) e Stefano Odorizzi (CEO EnginSoft).

Durante la tavola rotonda, condotta e moderata da Federico Pedrocchi - giornalista scientifico di ‘Radio 24-Il Sole 24 Ore’, gli opinion leader, provenienti dal mondo delle istituzioni, dell’impresa e della ricerca scientifica si sono confrontati sul tema dell’innovazione quale fattore chiave di successo e competitività d’impresa anche, ma soprattutto, in tempo di crisi di mercato. È Alberto Bombassei ad entrare in tema affermando che “… le strategie applicate dalla maggior parte delle aziende italiane - non solo PMI - fondate sull’innovazione incrementale e di processo, sostanzialmente finalizzate ad abbattere i costi di produzione e migliorare la qualità dei prodotti, non sono più sufficienti”. Aggiunge il presidente di Brembo Spa “in un mercato Globale, dove i paesi in via di sviluppo e con mano d’opera a basso costo la fanno da padrone, occorre sempre più innovare per essere competitivi e mantenere la leadership”. Gli fa eco Mirano Sancin, Direttore Generale di Kilometro Rosso, che aggiunge“… è l’innovazione radicale e di prodotto che contribuisce maggiormente a spostare le attività economiche, e produttive, da un’elevata concentrazione di manodopera (sempre più difficile da reperire) ad una elevata concentrazione di conoscenza (tipica dei sistemi più evoluti) e ad aumentare la competitività delle imprese a livello internazionale”.

Anche le istituzioni collaborano, con l’imprenditoria e la ricerca strutturata, alla causa comune della competitività dell’impresa-Italia attraverso veri e propri strumenti finanziari costituiti dai Bandi. “Chi non ricerca non cresce” è

Fig. 1 - Alberto Bombassei, Vice Presidente di Confindustria, che commenta il contesto di mercato entro cui le aziende italiane devono operare

lo slogan citato da Roberto Formigoni e promosso da Regione Lombardia che nel biennio 2009-2010 ha stanziato fondi per oltre 80 milioni di Euro destinati alla ricerca e all’innovazione industriale. “Nonostante le difficoltà, le aziende virtuose continuano ad innovare, innovare e ad investire nella crescita – accenna il Governatore di Regione Lombardia - in un momento di difficoltà generalizzata, le aziende investono in ricerca per cercare nuovi margini di profitto e aprirsi a quel contesto di conoscenza distribuita che caratterizza la società moderna. È questo il dato positivo - conclude Formigoni - che emerge dai primi risultati del Bando Regionale”.

“Le nuove tecnologie di simulazione e di analisi predittiva sono di fatto riconosciute da molte aziende un’effettiva rivoluzione dei processi progettuali” ha affermato Giancarlo Michellone.

In questo contesto di ricerca applicata ed incubatore tecnologico si inserisce a pieno titolo anche EnginSoft che da tempo collabora con l’R&D di Brembo per la simulazione di sistemi frenanti e con l’Istituto Mario Negri per applicazioni farmacologiche: realtà entrambe insediate nel Parco Scientifico. Con oltre 30 ricercatori ed ingegneri impiegati

Newsletter EnginSoft Year 8 n°4 -

Fig. 2 - Overview della platea di Imprenditori e Top Manager che hanno partecipato alla tavola rotonda organizzata da EnginSoft a Bergamo

nella sede di Bergamo, l’azienda investe sul proprio futuro e rilancia la presenza in Italia trasferendo una delle sedi all’interno di un incubatore tecnologico d’eccellenza qual è il Kilometro Rosso. “È dal 2007 che collaboriamo con il Consorzio Intellimech e con altri laboratori di ricerca inseriti nel Parco Scientifico Tecnologico - afferma Stefano Odorizzi, Presidente di EnginSoft – in questi anni abbiamo toccato con mano l’importanza di far parte di questa struttura che condivide la nostra stessa mission: sviluppo di tecnologia e innovazione”.

L’evento di oggi promosso da EnginSoft, in uno dei rari casi in cui istituzioni, ricerca universitaria e impresa si riuniscono a confronto su temi strategici e di vitale importanza per il sistema-Italia, è la riprova del consenso e dell’autorevolezza che l’azienda, negli anni, ha riscosso sul mercato.

Per ulteriori informazioni: Mosè Necchio - EnginSoft info@enginsoft.it

11

La gestione progetto in ottica Lean Design Sviluppare processi di progettazione e sviluppo-prodotto sempre più rapidi ed affidabili è oramai riconosciuta quale una necessità strategica imprescindibile. È quanto è emerso, in estrema sintesi, dal simposio di Bergamo. Per esplorare diverse alternative di soluzioni è necessario essere rapidi e tempestivi nell’apprendere i limiti e le potenzialità di ciò che stiamo ideando e progettando. La velocità e l’efficacia nell’esplorazione delle alternative, quindi, sono profondamente legate alla capacità di sperimentazione attraverso un numero significativo di prototipi ognuno funzionale alla verifica delle intenzioni di progetto e la loro corrispondenza alle necessità del cliente. Questo approccio, mediante l’impiego di prototipi fisici, potrebbe richiedere tempo e risorse in numero incompatibile con il budget disponibile. Anche nei processi di innovazione-prodotto esistono forme di “spreco” definibile in: qualsiasi attività che non crea Valore per il cliente. Il tema su cui riflettere è che tali sprechi non sono immediatamente visibili e non sono, quindi, facilmente aggredibili se non attraverso le giuste metodologie per individuarli. La riprogettazione dei processi di innovazione-prodotto, in chiave sperimentazione virtuale, può liberare enormi energie creative e di conoscenza che frequentemente sono già presenti negli uffici tecnici e di calcolo. EnginSoft, su questo tema, sta elaborando e sviluppando iniziative ad hoc finalizzate a diffondere le metodologie di Lean Design con la relativa valutazione del ROI soprattutto attraverso l’impiego della Simulazione e della Sperimentazione Virtuale.

12 - Newsletter EnginSoft Year 8 n°4

Electromagnetic Issues for a IEEE 1902.1 “RuBee” Tag Dipped in a Fiber/Composite Laminate The IEEE 1902.1 “RuBee” communication standard defines the air interface for radiating transceiver radio tags using long wavelength signals (up to 450 kHz). Conforming devices can have very low power consumption (a few microwatts on average), while operating over medium ranges (0.5 to 30 meters) and at low data transfer speeds (300-9600 bps).

In this article, the approach to model a loop tag operating at 131.072 kHz through ANSYS Maxwell v.14 is described when the sensor is dipped in a multilayer fiber/composite laminate. Some preliminary results are shown in terms of input inductance and magnetic fields.

Free standing antenna modeling Fig. 1 shows the prototype and the numerical ANSYS Maxwell model of a magnetic loop antenna for the short range “RuBee” protocol. The antenna (Fig.1a) is a 42mm radius multi-turn coil made of 33 loops of a copper wire with a section radius equal to 0.25mm. The numerical model is made of a solid single wire with a circular section

Fig. 2 (a) - Prototype of the multi-turn microstrip coil and (b) Maxwell 3D model. The PCB connector is visible in the bottom of Fig. 1a. The two side copper plates are helpful to tune the antenna input impedance.

CPW fed antenna is made of 16 properly distanced 0.6mm wide microstrip copper line turns.

The background scenario was modeled by imposing radiation boundaries to the problem region in order to simulate free emission into space. In the operational environment, the latter could be a lossy and/or conductive media like sea water and oil (see Table I for more details) and it should be consequently modeled with the correspondent electric characteristics.

Table I - Dielectric characteristics of some media compared with free space

Fig. 3 shows a sample of the electric current density along the loop and on the solid wire section. The imposed

Fig. 1 (a) - Prototype of the 33-turn copper wire coil and (b) geometrical details of the Maxwell 3D model. In the top of Fig. 1a the microstrip feeding line and the PCB connector are visible.

radius rls equal to 0,143cm. As indicated in the bottom of Fig. 1b, this value corresponds to the radius of a circumference with a surface equal to the sum of the 33 wire sections.

The second element is a multi-turn printed loop on a 0.8mm thick FR4 laminate and it is shown in Fig. 2. The

Fig. 3 - Sample of the current density distribution along the loop

Newsletter EnginSoft Year 8 n째4 -

stranded current, constant on the wire section, is visible in the bottom left detail and, as expected, the current is constant along the loop.

Fig. 4 shows a sample of the magnetic induction distribution in a plane containing the loop axis. This B field distribution is a well-known result, according to basic electromagnetic theory. Indeed, the loop length is much smaller than the free space wavelength at 131 kHz (around 2.3km), so resulting in an elementary loop design. For such elements the near field is mainly magnetic and completely decoupled by the electric field.

13

some other aspects could make the printed square loop preferable, like its mechanical stability and the more accurate repeatability of the prototyping.

Table II shows the simulated and measured values of the input inductance for the two configurations. For the solid loop case, the calculated value is obtained from a correspondent analytical

Fig. 6 - H field distribution along the loop axis for the wire solid ring and the printed microstrip square loop.

Table II Input inductance for the two antenna configurations Fig. 4 - Sample of the magnetic induction in a plane orthogonal to the sweep.

Even if the device is an antenna, this consideration justifies the use of ANSYS Maxwell 3D rather than ANSYS HFSS because the magnetic near field characterization provided by Maxwell 3D fully satisfies the design requirements.

model and this is in good agreement with the simulated one. The measured inductance is around 10% less than the previous cases. This disagreement results from the mismatch between the transverse section areas of the solid loop of the simulated and calculated cases and the 33-turn one of the prototype (see the bottom detail of Fig.1b). A 0.9 fill factor (Fig.1a) corresponding to the missing lighter areas of the prototype with respect to the numerical models should be considered to compensate it.

An excellent agreement between simulations and measurements is apparent for the printed element.

Electromagnetic modeling and analysis of the composite laminate The two prototypes would be dipped in a composite material as shown in the sample of Fig. 7.

A composite laminate can be schematized as a stack-up of several plies, each of them made of a sheet of fibers filled

Fig. 5 - Details of the mesh characteristics for the microstrip printed loop

Fig. 5 shows a sample of the mesh for the microstrip printed square loop. Around 161000 tetrahedra were used for the computational domain and around 24000 for the loop. For the solid wire loop 61000 tetrahedra were necessary for the computational domain and 14000 were used for the loop.

Fig. 6 shows the H field distribution along the loop axis, for both configurations. The H field is higher for the wire loop, suggesting the use of this antenna type. However,

Fig. 7 - Sample of rectangular loop dipped in a fiberglass composite laminate.

14 - Newsletter EnginSoft Year 8 n°4 isotropic, in the sense that only their intrinsic dielectric characteristics are known.

On the other hand, the structures in Fig. 8b and c are generally anisotropic, as a result of the applied methodology. The permeability and permittivity tensors need to be calculated according to the material properties and to the problem geometry, as:

Fig. 8 - Single composite ply: (a) schematic model, (b) equivalent model for the intermediate fiber/resin layer, (c) equivalent model for each single ply

where:

and g is a function of the ratio between the fiber and the resin volume in the intermediate layer of Fig. 8a.

Fig. 9 shows the Maxwell 3D model with 4 plies above and 4 plies below the wire antenna.

Fig. 9 - Example of a composite laminate made of 8 plies: 4 above and 4 below the wire loop antenna

with some dielectric resin, as shown in Fig. 8a. An df thick intermediate layer made of some fibers and resin lies between two dr thick single layers of resin. This structure could generally be dissipative, conductive and anisotropic, the latter depending on the characteristics and the distribution of the fibers.

Each ply has been modeled in Maxwell 3D, including all the material anisotropies and dielectric properties. The work on the analysis of the effect of a number of plies up to 64 is in progress. They have been fully parameterized in order to take into account a number of possible ply configurations and materials.

An effective approach to model this structure is to define an equivalent layer for each ply. Many models have been recently presented, resorting to different approaches but all of them afford a specific problem without deeply challenging a general approach. In the framework, the approach to model an equivalent layer for each ply is to apply the method described in for the intermediate layer of Fig. 8a, in order to get an equivalent anisotropic intermediate one, shown in Fig.8b.

Conclusions In this work, the approach to analyze the electromagnetic performance of a tag antenna for the IEEE 1902.1 “Rubee” protocol has been described through the use of ANSYS Maxwell. Preliminary results have been shown in terms of radiated magnetic field and input inductance for both numerical models and prototypes. Simulated and measured results are in excellent agreement, proving the tool reliability. The methodology to model a multi-ply composite fiber material has been defined and numerical analyses on the antennas’ performance in its presence will be the main topic of some future investigations.

Then, a circuital approach can be applied to the multilayer structure shown in Fig. 8b to result in a single layer equivalent anisotropic model. It is worth noticing that all the constitutive materials (fibers and resin) in Fig. 8a are

Per ulteriori informazioni: Andrea Serra, EnginSoft info@enginsoft.it Thanks to Federica Bolognesi, IDNOVA

Newsletter EnginSoft Year 8 n°4 -

15

Structural Optimization of a Car-body High Speed Train - An Innovative Analysis and Design Methodology In the past, the main challenge was to achieve a very high speed, but today the criteria such as energy efficiency, high transport capacity, comfort and low environmental impact are becoming more and more important. For this reason the philosophy of AnsaldoBreda is to combine a settled design process with innovative approaches to optimize the reliability, safety, low power consumption and an easy maintenance. In order to be competitive in the market, especially in this economically challenging period, it is necessary to push the envelope of the available technologies to ensure compliance with top level quality standards.

A new methodology approach has been developed by exploiting the new capabilities of the multi-objective design environment modeFRONTIER and it has been applied to the design of the carbody structure of a new generation of High Speed trains.

In this context, the aim of the activity was the design optimization of the aluminum carbody structure in terms of weight and dynamic behavior, respecting all project constraints according to the high standard structural and crash requirements of European EN 12663 - Category P-ll (Fixed units) and TSI Rolling Stock.

Starting from the CAD model of the original configuration, the FE comprehensive parametric model has been developed by ANSYS APDL procedure and integrated into the

modeFRONTIER optimization platform to achieve the requested goals. The FE parametric model has been divided into two different main parts: 1. The central parts of the carbody (named “fuselage”) – as shown in fig.1a; 2. The terminal tapered parts of the carbody – as shown in fig.1b.

The fuselage geometry (fig.2) is completely parametric in terms of:

Fig.2 - Section profile of carbody

a) number of the profile reinforcements; b) angle, position of reinforcements; c) thickness of reinforcements; d) thickness of external and internal skin of profiles.

The aims of the optimization process of a carbody in modeFRONTIER are: a. Minimizing weight b. Maximizing two first own frequencies Fig. 1a - Fuselage Parametric part of high speed train: it has been completely development in ANSYS APDL. Fig. 1b: No -parametric part of high speed train: terminal tapered parts are fixed geometry

with the following constraints: a. Max Von-Mises stress for static analysis

16 - Newsletter EnginSoft Year 8 n°4 b. Max Von-Mises stress for equivalent crash analysis c. Max Von-Mises stress for fatigue analysis d. Min buckling factor for linear instability analysis The original configuration, only referred to the parametric part of the carbody, weighs 5.927 Tons. The main goal is the weight reduction by min. 500 Kg, maintaining the first bending frequency of 11 Hz. The static structural analysis and fatigue analysis have been performed for both welded and unwelded region (fig.3), which have different material features:

Only the modal analysis has been performed to find out the best region for weight and frequency with no timeconsuming run (less than 1 hour on the cluster machine).

The results of this first optimization loop has been used as a starting DOE (Design of Experiment) for the second one, where objectives/constraints related to displacement under pressure loads and to the 5-6 strongest load cases (fig. 4) have been introduced. This step is more time-consuming than the first one (5 hours on the cluster machine).

After these optimization loops, some variables have been changed in agreement with AnsaldoBreda, and the final

Fig.3 - Section of a carbody structure

Due to the high number of time-consuming simulation and the high number of input variables, a progressive approach has been studied for the optimization analysis. Therefore, the optimization analysis has been carried out in three steps: • Step1: Screening, driving towards the best designs region; • Step2: Rough refinement, including the most important constraint conditions; • Step3: Final refinement, achieving the optimal solutions.

optimization run has been done to achieve the best solutions. Since this step was really time-consuming (15 hours on the cluster), the problem has become to monoobjective: only the weight has been considered, while the other objective has become constraints (fig. 5). The set of best designs belonging to the new Pareto frontier has been verified for each operative load condition and the best designs have been chosen using decision making tools.

The optimal designs selected on the basis of stress and weight values have a considerable variation of both external and internal skin thickness, which can cause manufacturing problems. In order to avoid such problems, another post processing analysis has been done to find out Pareto solutions with a homogeneous distribution of thickness

A total of 23 different working -load cases have been considered, with an additional specific comfort requirement about Static Pressure load (-8 KPa inside Tunnel) which constrained the side walls displacements (Uy < 3mm and Uz < 4.5 mm) The whole simulation took 3 weeks on cluster machine with 8 parallel simulation (32 core).

The first optimization step has been carried out taking into account the two most important objectives of the problem (increase of frequency and weight reduction) which lead the designs to the best region Fig.4a - History of weight convergence and allows to reduce the design (green points: 1st optimization loop; blue points: space of the input variables.

2nd optimization loop).

Fig.4b - displacements in y direction (mm)

Newsletter EnginSoft Year 8 n°4 -

along external and internal skin. New post processing using “parallel chart” applied on best design has been carried out in order to find a suitable solution matching the new requirements introduced a-posteriori (fig. 6).

Table II shows the comparison between the best design selected at the end of the optimization analysis (Design ID 378) and the best design after the last post processing considering a thickness uniformity (Design ID 339).

Thanks to the implemented methodology and the optimization routine, a considerable weight reduction has been reached. The chosen solution, Design ID 339, has a weight reduction of 546 Kg (- 9.2%) and it has a more uniform thickness variation which simplifies the carbody manufacturing.

This work aims to shows how to exploit new design methodologies and new technologies in order to manage industrial design processes that involve a large number of variables (more than 50), several constraints and objectives, finding the best solution according to industrial timing.

It is possible to summarize the most important steps of this activity, as follows: • The design optimization procedure developed has been completely automated: this allowed to make the most of all available hardware and software resources, completely exploiting the downtime (nights and holidays).

• The requested weight reduction has been achieved respecting every structural and comfort requirements: this has totally fullfilled the expectations of the modeFRONTIER industrial users.

• The additional requirement about manufacturing has been fulfilled without rerun any analysis thanks to the new methodology approach: this has been possible thanks to the really powerful capabilities of the post-processing tools of modeFRONTIER. • The optimization methodology can be completely re-used for other design processes: this activity was dedicated to a specific carbody but this approach can be easily adapted also to other railway vehicles.

For more information: Francesco Franchini, Enginsoft info@enginsoft.it

Fig.5 - The workflow of modeFRONTIER with all input and output variables, the final objective and constraints

Table I - The table above summarize the optimization strategy adopted.

The total number of design has been run in 20 days

Fig.6a - Parallel chart of the best designs

Fig.6b - The selected design (Design ID 339) with homogeneous thicknesses

Table II - comparison between the original solutions and the optimized solutions

Table III - Thickness comparison of the side walls of fuselage (profile ref. 5-6-7-8)

17

18 - Newsletter EnginSoft Year 8 n°4

FSO and Shuttle Tanker in Tandem Configuration Hydrodynamic Analysis Finalized to the Structural Verification of the FSO Mooring System Strength and Fatigue Verifications of an FSO mooring system have been performed basing the results on proper hydrodynamic analysis (developed inside ANSYS-AQWA) and structural analyses (developed inside ANSYS-Workbench) of the system and relevant components.

Hydrodynamic Analysis The FSO (109.000 DWT), operated by Edison, is moored on the Rospo Mare Offshore Oil Field. The FSO mooring is guarantees via 6 chains connected to a rotating turret, installed at the FSO bow. During the oil offloading operation, the Shuttle Tanker (45.000 DWT) is moored, via an hawser, at the FSO aft end.

The offloading operation takes place under proper sea conditions, with waves characterized by significant height (Hs) ad zero up-crossing period (Tz). To each sea state, consistent current and wind have been accounted for.

The hydrodynamic model (performed inside Ansys-AQWA suite), simulating the FSO and the Shuttle Tanker (this one moored, at its stern, to a Tug via a mooring cable), refers both to aligned and misaligned meteo conditions (current incoming at 50 degrees with respect to wave direction, wind incoming at 25 degrees with respect to wave direction).

On the model (FSO + Shuttle Tanker + mooring lines), time domain hydrodynamic analysis has been performed for each defined sea-state, obtaining, for each mooring chain and for the hawser connecting FSO and Shuttle Tanker, the axial tension as function of time.

In order to check the strength resistance of mooring components (such as Chain Stoppers and 'Ecubier') installed at the rotating turret, besides hydrodynamic analyses under offloading conditions, also hydrodynamic analyses of FSO in moored condition, for extreme storm case (100 years return period), have been performed.

Strength and Fatigue Verification of Chain-Stopper and “Ecubier” Based on results of hydrodynamic analysis performed for both extreme and offloading conditions, strength and fatigue verifications of Chain Stopper and ‘Ecubier’ have been performed.

Strength checks have been based on results obtained from contact non-linear analysis performed of Finite Element

Model of Ecubier + Chain Stopper under extreme load case (practically the chain minimum breaking load).

Fatigue checks have been developed according to spectral approach as required by DNV OS-E301 (Position Mooring), assuming proper S/N curve data as reported in DNV RP-C203 (Fatigue Design of Offshore Steel Structures).

The assumed hypothesis at the base of fatigue spectral approach is that the stress range, S, is a random variable characterized by a probability density equal to p(S) and that, for each sea-state, the number of cycles having stress variation in the range of S and S+dS is directly related to ni p(S), where ni is the total number of cycles of that sea-state.

Based on this and on the fact that, for offshore structures, the probability density of stress ranges, p(S), can adequately be represented by a Rayleigh distribution, the

Fig. 1 - Hydrodinamic Model of FSO, Mooring Lines, Shutter Tanker

Fig. 2 - Von Mises Stress distribution on Ecubier and Chain Stopper

Newsletter EnginSoft Year 8 n°4 -

19

damage, Di, for the i sea-state, is given by the following relation:

Fig. 3 - Finite Element Model of Ecubier and Chain Stopper

where a and m are factors of S/N curve (C curve has been considered for fatigue verification of Ecubier and Chain Stopper), while σs is the standard deviation of S distribution.

Finally, based on Miner-Palmgreen relation, the total damage, D, due to the summation of damages of each seastate, Di, is:

Enrico Miorin, Fabiano Maggio, Livio Furlan EnginSoft

Fig. 4 - S/N Curves in sea-water with cathodic protection

For more information: Livio Furlan, EnginSoft info@enginsoft.it

Design and FEM Analyses in Offshore and Oil&Gas Industry Besides competencies in Automotive, Aerospace and Industrial Engineering Simulations, EnginSoft has knowledge also in the Design and Analyses voted to the Oil&Gas and Offshore Industry. Many consultancy activities have been performed via collaborations with the most important Italian players in this sector: ENI, Saipem, Tecnomare, MIB Italiana, Petrolvaves, Cameron, FBM, Officine Resta, Nuovo Pignone, ATB, Foster Wheeler.

EnginSoft can supply a full range of services covering projects entire design route, from the earliest conceptual studies passing through FEED and basic design up to detailed design and installation engineering.

The following list reports some of the Oil&Gas Business Unit competences: • Conceptual and detailed design and structural analysis of fixed offshore platforms (jacket, top-sides, buoyancy tanks, stiffened structures) • Design and analysis of subsea foundation templates • Design and analysis of pressure vessels, valves, piping, rack, etc.

• Design and analysis of subsea manifold (even for installation, repairing and retrieval operations) • Detailed structural analysis of structural parts (Hulls, Deck, etc.) of Semi-Submersible Vessels • Detailed structural assessment of steel Gravity Based Structures (GBS) including stiffened plate code checks • Detailed design and structural analysis of risers and FPSO's mooring connectors • Revamping of fixed offshore platforms (assessment of structural reliability- re-certification and life extension), fracture and fatigue assessment of installed jacket structures (risk analysis) • Motion Analysis of Floating Vessels (even for Marine Pipeline Installations) The BU, which is located in EnginSoft Padova Office and is coordinated by Livio Furlan, has high skills also in the field of structural and mechanical applications in general (as an example the design and analysis of Roller Coaster structures and cars or the design of large valves for hydroelectric power plants).

For more information: Livio Furlan, EnginSoft - l.furlan@enginsoft.it

20 - Newsletter EnginSoft Year 8 n°4

Numerical Analysis of a Micro Gas Turbine Combustor Fed by Liquid Fuel This work presents a CFD analysis of the combustion chamber of a 50 kWel nominal power micro gas turbine. The purpose of the analysis is to investigate the combustion process and performance of the combustion chamber fed by liquid fuels, through 3D numerical simulations performed with ANSYS CFX 13.0. Firstly, a sensitivity analysis was carried out in order to determine the parameters for the correct modeling of the liquid injection. Then, a simulation campaign was conducted to investigate the case of Jet A feeding and the supply with different liquid fuels deriving from biomass.

Introduction Nowadays micro gas turbine (MGT) are one of the more flexible and effective system for the distributed and residential micro cogeneration, due to their compact size, the low operating and maintenance costs, their greater overall conversion efficiency and reduced environmental impact. The continuous flow operation of this system offers a greater flexibility with respect to the unsteady process of internal combustion engines that imposes constraints on fuel characteristics. In particular, MGTs can be supplied with fuel (both gaseous and liquid), characterized by a higher level of contamination thanks to their greater adaptability to different fuel supply. Among the renewable sources, an increasing interest has been shown in fuels derived from biomass since they are a predictable source, allowing the distributed grid-connected generation without causing discontinuities in the electric grid and frequency instabilities.

At the same time, vegetable oils have gained attention since they can be low-cost fuels and allow to implement

systems for the distributed energy production. MGTs are not well-established systems for straight vegetable oil feeding, yet, because the combustion of these oils had to be investigated due to the opposite physical and chemical characteristics, such as the chemical composition, the lower heating value (LHV), the molecular mass, the density and the viscosity, compared to diesel, biodiesel, dieselvegetable oils and their mixtures. In fact, the combustion performance depends on the atomization process and spray characteristics, which are directly related to the fuel composition and its physical properties, in particular the high viscosity of vegetable oils. The study presented below regards the preliminary analyses performed on a MGT combustion chamber fed by conventional fuel (Jet A), in order to find the correct settings for the simulation of biofuel feeding.

Computational domain and numerical models Geometry. The numerical analysis have been conducted on the combustion chamber of Solar T62-T32, a micro gas turbine of 50 kWel nominal power, fed by diesel fuel. The combustion chamber (Figure 1a) is a reverse-flow annular type combustor, with six fuel injectors, 24 dilution holes and a series of holes for the cooling of the liner wall. The air from the compressor enters the combustion chamber in counter-current with respect to the combustion gases, passing through the space between the external wall and the linerâ&#x20AC;&#x2122;s wall. The solid domain of the combustion chamber (Figure 1b) was obtained from the direct measurement of the real geometry (Fig. 1a). Thanks to the periodicity of the number of fuel nozzles, dilution holes and wall cooling holes, the fluid domain was reduced to a 60° annular sector of the combustor (Figure 1c).

Figure 1 - (a) real combustor geometry, (b) solid domain, (c) grid of the fluid domain.

Newsletter EnginSoft Year 8 n°4 -

21

parameters that better predict the behavior of this type of combustor, sensitivity analyses on the boundary conditions have been carried out.

Boundary conditions influence One of the ways to reduce the particle spray diameter and, therefore, to obtain a finer spray, is to increase the atomizing air mass flow, which also applies to high viscous fuels. A larger flow of atomizing air can be obtained by modifying the bypass Figure 2 - Comparison of the results of the air/fuel ratio variation: temperature distributions. from the main machine compressor or by adding external air from an auxiliary compressor. So the influence of the air mass flow coming from the compressor has been evaluated. The air mass flow to fuel mass flow ratio, AFR, was varied from the standard value of 70 to 50 (rich combustion) and 100 (lean combustion).

Figure 2 shows the comparison of the temperature contour plots in the nozzle mid plane: the flame increases in terms of extension and intensity as α increases, as Figure 3 - Comparison of the results of the particles’ diameter variation. expected. The quantitative results showed that the values of the turbine inlet Grid. Two unstructured tetrahedral grids with an overall temperature (TIT) and pollutant emissions, such as NOx and number of elements approximately equal to 1.5 and 2.5 CO, of the case of standard air/fuel ratio (ARF = 70) are in million respectively were generated using ANSYS ICEM CFD. good accordance with the measured pollutant Both grids are characterized by a uniform distribution of the concentrations and the calculation of the TIT by means of a elements inside the domain, with a more refined mesh gas turbine Cycle Deck. For these reasons, an air/fuel ratio inside the nozzle and combustion zone. of 70 was chosen for the subsequent simulations.

The sensitivity analysis of the grid showed that both grids achieved the numerical convergence and were robust with Spray parameters influence compared to the overall performances of the combustor. For The simulation of liquid fuel combustion has been carried these reasons the 1.5 million elements grid (Fig. 1c) was out defining a particle injection region placed nearby the used in the numerical analyses presented below. fuel inlet surface, which is closed to the exit of the fuel injection duct. Sensitivity analyses concerning the diameter Numerical models and boundary conditions. The numerical of the particles injected into the combustor and the angle models adopted are: the k-ε for turbulence, the Eddy of the injection cone have been performed: in particular, Dissipation (EDM) for combustion with a 2-steps reaction three diameter sizes (1, 10, 20 µm) and three injection scheme and a PDF model as the NOx formation method. A cone angles (10°, 20°, 30°) were investigated. particle injection region and the TAB (Taylor Analogy Breakup), as secondary breakup model, were set at the fuel In the case of variation of the particle diameter, the flow inlet surface in order to model the fuel spray, while the field and the temperature distributions in the nozzle mid primary breakup was not activated. An adiabatic boundary plane have not presented significant modifications. The condition was set for all the combustor walls. Fuel inlet values of TIT and pollutant emissions (NOx and CO) boundary condition was set according to the data provided calculated at the outlet surface of the combustor has by the manufacturer, while the air mass flow value was decreased as the particle diameter has increased, according obtained from literature. All the numerical simulations were to the liquid fuel combustion phenomena. The evaporation performed with ANSYS CFX 13.0. time of the particles has increased as they have increased in size, while the particle traveling distance has increased CFD Analysis of the combustion chamber in an irregular way, as shown in Figure 4. A great increase Case of conventional fuel feeding has occurred passing with diameter between 10 and 20 µm In these cases the simulation regards the supply with and a decrease has occurred with a diameter between 1 and conventional fuel, so the Jet A fuel of the CFX material 10 µm. This was probably due to the size of the grid library has been used. In order to determine the simulation elements. Nevertheless, numerical values were in

22 - Newsletter EnginSoft Year 8 n°4

Figure 4 - Comparison of the results between Jet A and mock biofuels: particle traveling distance and temperature distribution.

accordance with it. When the spray cone angle has varied, the vaporization time and the traveling distance of the particles increased as the cone angle has increased.

Temperature values into the primary combustion zone are lower in the case of a cone angle of 30°; the TIT value decrease accordingly. An anomalous behavior occurred when cone angle of 20°: there is a reduction of the particle traveling distance and the evaporation time; the TIT value is in accordance with the other simulated cases.

According to the results of the sensitivity analyses already performed, an air/fuel ratio of 70, a diameter of 10 µm for the particle injection and a spray cone angle of 30° have been chosen for all the simulation presented below.

Case of vegetable oil fuel feeding Subsequently, some simulations have been performed in order to investigate the behavior of the combustor in case of feeding with liquid fuel derived from biomass. As first attempt, two mock biofuel have been created starting from the Jet A characteristics and modifying only some of the parameters (density and viscosity values), in order to determine the influence of a single parameter each time.

The density value, equal to 914 kg/m3 at 20 °C and the dynamic viscosity value, equal to 40 cP at 20 °C, comes from a direct measurement of a sample of rapeseed oil derived from dedicated crops (experimental crops realized within a research project on short energy chain). As a reference, default Jet A density and viscosity values at 20 °C are 780 kg/m3 and 1.5 cP, respectively.

Figure 4 shows that the temperature distributions of the mock biofuels differ from the Jet A feeding in terms of intensity and flame morphology. The maximum temperature values in the mock biofuel cases are higher than the ones in Jet A case within the primary combustion zone, and the flame of Jet A case is more stable and there is less variation in temperature values. In terms of flame morphology, the base of the flame starts at the nozzle exit in Jet A case, while in mock biofuel cases it seems to even start inside the

nozzle. The highest values in the primary combustion zone are probably due to the lower flow velocity that produces an increase in the residence time, which come out from the analysis of the velocity field and the particle traveling distance pattern. The average values of TIT, NOx and CO calculated at the outlet surface of the combustor are not influenced by the density and viscosity variation.

Conclusions The aim of this work is to study the combustion phenomena related to the liquid fuel feeding of the annular combustion chamber of a micro gas turbine with an electric power of 50 kW. The main parameters of the fuel spray were investigated in the case of conventional fuel supply (Jet A) setting different values of particle diameter and cone injection angle. No significant modifications in terms of flow field and temperature distributions were noticed from the sensitivity analyses on spray parameters. The values of TIT and pollutant emissions (NOx and CO), calculated at the outlet surface of the combustor, decrease as the particle diameter increases, according to liquid fuel combustion phenomena. The evaporation time of particle and the particle traveling distance increase as dimension and cone angle increase, leading to slower combustion and, at the same time, a longer flame in the combustor. Particles with a diameter of 1 µm present an anomalous behavior in terms of the particle traveling distance and mean particle diameter, which is probably due to the size of the grid elements.

Subsequently, a numerical analysis was performed in case of biofuel supplying. A mock biofuel was used by setting the values of density and viscosity of a rapeseed vegetable oil obtain from mechanic extraction of dedicated crops. The setup of the model parameters was performed by starting from the sensitivity analyses carried out in case of Jet A feeding. The analysis of the particle track shows that there is an increase in the particle traveling distance and the particle time as the fuel viscosity increases and the consequent increase of the residence time. This leads to higher temperature values inside the primary combustion zone. The global performance of the combustor (TIT and pollutant emissions) are not influenced by changes in density and viscosity.

Michele Pinelli, Anna Vaccari Università degli Studi di Ferrara

Newsletter EnginSoft Year 8 n°4 -

23

Reconsidering the Multiple Criteria Decision Making Problems of Construction Workers Using Grapheur We are dealing with a series of multiple criteria decision making problems and analysis related to Canadian construction projects including waste management, productivity improvement, human and IT factors, emergy based lifecycle, and process optimization.

The urgent increase of using IT in construction projects has been considered as one way to improve the process of solving our problems. Construction project managers have to make tough decisions. They have been considering different IT tools and would like to invest on getting better data analysis tools for enhancing their decisions. However, making critical decisions for complicated and multiple criteria construction projects problems in which huge amount of data are involved is not a simple task to do. As the data-sets of our problems are often huge they can not easily be handled with the traditional means of data analysis. In order to better manage the data collected and make the most of our data-sets, we utilized the advanced interactive visualization tools provided by Grapheur and reconsidered our problems. Here the idea for solving the multiple criteria decision making problems is to visually model and clarify the whole dimension of problems. The effectiveness and performance of the interactive visualizations, made by Grapheur, are evaluated along with a number of our case study related to construction workers.

As the main result, the 7D plots and the option of sweeping through data have been found very useful for our applications. The achieved hidden information through Grapheur’s visualization tools would enhance our further decisions.

Introduction to Grapheur Grapheur is a data mining, modeling and interactive visualization package implementing the Reactive Business Intelligence approach, which connects the user to the software through automated and intelligent self-tuning methods on the basis of visualization. The principles of Grapheur were originated from researches on Reactive Search Optimization. The user friendly and innovative interface of provided visualization, via an interactive multiobjective optimization, facilitates the process of making tough decisions. Grapheur is a handy and simple tool which frees the mind from software complications and concentrates on mining the useful information data. It puts the user in an interactive loop, rapidly reacting to first results and visualizations to direct the subsequent efforts, in order to suit the needs and preferences of the decision maker. The Reactive Search is utilized within Grapheur to integrate some machine learning techniques into search

heuristics for visualization of complex optimization problems and interactive decision making accordingly. In Reactive Search for self-adaptation in an autonomic manner, we benefit from the past history of the search and the knowledge accumulated while moving in the configuration space.

Grapheur sample visualizations In one of the building construction projects a number of workers were surveyed with questionnaires and observations. Each row of our data-set is a construction worker with the corresponding columns, characterized by a series of parameters which are the ID and photo of each person, work time, looking for materials, looking for tools, specialization, moving, instruction, idleness and the other characteristics of the construction workers. The primary result of our survey clearly notes the urgent need for training programs to improve workers’ skill levels. However, the decision-making on how and with what rate the training programs should be arranged is not a simple task and it has to be considered from different perspectives and criteria. In order to learn how the training programs would affect team efficiencies, spirit, and perceptions of supervision, Grapheur, the flexible and powerful Business Intelligence and Interactive Visualization is utilized. With the aid of provided data mining and visualization some useful and hidden information are achieved which would enhance the process of solving the multiple criteria decision making problems of our case. After clarifying the dimension of the problem and finding out the relation between involved parameters and objectives, the effective decisions are easily made.

1. Supporting the decisions on workers’ skills Here the idea for solving the multiple criteria decision making problems is to visually and effectively model the problems and clarify the whole dimension of them. For instance we are trying to find out with which rate and how, the workers’ level of skills should grow in order to maintain their performance with regard to team perceptions of supervision. In order to study a part of this problem, we are considering the similarity map and the parallel filters for optimizing the idleness characteristic of the workers. The related multidimensional plot of the networks is created based on the collected data from the workers. The color code represents the specialization of the workers and the size of the bubbles is proportional to the idleness of workers. In our similarity map of the graphical visualization, the gray level of the edges and the generated

24 - Newsletter EnginSoft Year 8 n째4 displays the looking for tools characteristic of the workers.

In this figure (Fig. 2 a) we have found clustering tool very useful for a deep understanding of the different groups of workers. In this case, workers could be grouped according to the given characteristics. After grouping, one prototype case for each cluster is selected which is indeed a very effective way of compressing the information and concentrating on a relevant subset of possibilities.

Fig. 1a - and similary map

3. Sweeping though different characteristics of workers; tracking and examining the problem with the aid of animated graphs In the previous figure (Fig. 2 a), the relationship among work time, specialization, idleness status, looking for materials, and looking for tools characteristics of the construction workers were visualized. Moreover, sweeping though data and studying the generated animations on sweeping is an effective tool for further visualization along with advancing a particular objective. For instance, in our next visualization experience the previous graph is reconsidered by sweeping though looking material, Idleness and skill level as the time advances (illustrated in Fig. 2 b).

Fig. 1b - Parallel filter

clusters provide valuable information for the decision maker. In the following figure and with the provided video the capability of the similarity map for an effective clustering of the workers into meaningful clusters is illustrated (Fig. 1 a).

The parallel filters (Fig. 1 b) are other useful tools for optimization. The usefulness of parallel filters in reducing the complexity of the process of decision making is evaluated. We start from the matrix of work time in a multidimensional space while aiming at filtering particular workers and examining their performance within a particular group e.g. those who have had maximum idleness characteristic.

2. Displaying the precise condition of each construction worker For complete visualization of the condition of each construction worker over all parameters, the colored bubble chart is selected. In Fig. 2 a, the colored bubble chart shows work time versus specialization for each worker. The color code and the size of the bubbles represent looking for material characteristic and the idleness status of the workers respectively. Additionally, the shape of the bubbles

Fig. 2a - Bubble chart

Fig. 2b - Sweeping through data in the bubble chart

Newsletter EnginSoft Year 8 n°4 -

4. Analyzing a particular cluster of workers and their characteristics; sweeping through skill level and team perception of supervision In the new created bubble graph, Fig. 3 b, the idleness and specialization characteristic of a cluster of four workers is associated with the size and the color of the bubbles relatively. Here, by sweeping through team perception of supervision and the level of specialization of field workers in our building construction project, the achieved information from a limited cluster of workers can clarify the problem with more details in different scenarios. For instance, when the skill level of the workers and the team perception of supervision are monthly increased relatively by the rate of 10% and 5% within a year, the idleness characteristic is smoothly monitored. We can also play the resulted animation in smooth mode and track the past values (they appear in a lighter tone in the background of the plot), in order to focus on the changes which occur according to morning and afternoon working shifts.

25

Fig. 3a - 7D plot of data

5. Providing a reliable way to find the most productive workers With the aid of the 7D plot, the characteristics associated with the productivity can be presented within a single graph. In our case the size, the color and the shape of the bubbles relatively displays the specialization, the moving, and the following the instructions characteristic of the workers. Moreover the blinking feature displays the idleness characteristic of the workers who have been idle less than 100 hours (Fig. 3 a).

Conclusions In this short article, along with our case study, the aspects of data mining, modeling, and visualization of the data related to construction workers are considered and briefly presented by utilizing Grapheur. We made the most of IT applications via newly implemented data mining and visualization tools of Grapheur. Considering the ability of Grapheur, the interesting patterns are automatically extracted from the raw data-set via data mining tools. In addition, advanced visual analytical interfaces are involved to support the decision maker interactively. With additional features of Grapheur such as parallel filters and clustering tasks, construction managers can solve multi-objective optimization problems as it amends previous approaches.

Furthermore, the animations of sweeping through data and advanced visualizations including 7D plots stead managers and enable them to screen the data at their consulting room making decision interactively.

In one of our case studies, Grapheur provided a widespread view on how the throughput of the whole project would be affected by the increasing workersâ&#x20AC;&#x2122; specialization and supervision. We swept through different characteristics of workers in order to examine the whole dimensions of the problem. For instance, we assumed that the problem of having high level of idleness within the workers might be solved by increasing the supervision and team perception of

Fig. 3b - Sweeping through data

supervision. For this reason, workers are carefully clustered and analyzed with regard to their level of idleness and supervision. In this particular case, Grapheur has been a facile tool in modeling the problem with the aid of a 7D plot. Once a 7D plot is created the problem could be visually analyzed from seven different perspectives simultaneously.

In other words, a convenient way of concentrating on our objectives and further decision making is provided by simply observing the size, the color, the shape, and the blinking of the bubbles. Moreover, utilizing further visualization options such as similarity maps, parallel filters, and clustering would support making a confident decision. For our future studies aiming at making easier and faster decisions we will reconsider our problems with the aid of a developed issue of Grapheur called LIONsolver, Learning, and Intelligent OptimizatioN which is capable of learning from human feedback and previous attempts while benefiting from the Grapheur visualization tools.

M.Azodinia - University of Debrecen Faculty of IT, 4033 Debrecen, Hungary For more info on Grapheur: www.grapheur.com or email: battiti@reactive-search.com

26 - Newsletter EnginSoft Year 8 n°4

Synergy between LS-DYNA and modeFRONTIER to Predict Low Velocity Impact Damage on a Composite Plate Laminate composite structures do suffer from poor resistance to impact loading which results in internal damage that often cannot be detected by visual inspection. The damage can cause severe reductions in strength and can grow under load. Therefore the effect of foreign object impacts on composite structures must be taken into account during the design process. In order to simulate the impact event, an LSDYNA FE (Finite Element) model was developed and coupled with modeFRONTIER. The integrated procedure allowed to obtain a better understanding of the influence of some numerical parameters on simulation results (sensitivity analysis), moreover the configuration, which provided the best agreement with the experimental data, (optimization analysis) was computed.

LS-DYNA FE model As the plates’ length and width dimensions are large compared to the thickness, a 2D modelling approach was chosen. In particular layered shell elements with an element length of 3mm were used. The plate was associated to the linear elastic material model MAT54 which takes into account the progressive damage of the material. The elastic behaviour of the single ply is computed based on the lamina elastic material properties (Young modulus, shear modulus and Poisson’s ratio) which can be found in. Damage occurs as soon as one of the four failure indexes defined below becomes positive (Chang/Chang criteria):

Test Case Description The test case used to assess the capability of the procedure in investigating the impact event consists in a rectangular plate (whose dimensions are shown in Figure 1a) impacted at an energy level of 40J by a hemispherical steel impactor with a diameter of 25.4mm and a mass of 1.85kg.

The material of the plate was a laminate composite with a symmetric, quasi-isotropic lay-up of 24 plies [45°/0°/45°/90°]3s. In the unidirectional plies with the specification Cytec® 977-2-35-12K hts-134 the carbon fibres were impregnated with an epoxy matrix. The plies were then stacked and cured in an autoclave.

The resulting average cured plate thickness was 2.7mm. A length of 50mm of the specimen was clamped at each end reducing the free specimen length to 300mm (edge AD and BC fully constrained).

On the other hands, a simply supported condition was realized on the lateral sides (CD and AB).

After a failure is detected, elastic properties are degraded according to a specific degradation rule which depends on the kind of failure detected.

The impactor was modelled as a spherical rigid body with conventional shell elements and the material model MAT_RIGID. An initial velocity of 6.5m/s was imposed to the impactor by using the PART_INERTIA card. A very fine mesh was adopted in order to correctly compute the contact force between the impactor and the plate. The FE mesh used in the model is shown in Figure 1b. Finally, an automatic surfaceto-surface contact with the option SOFT=0 was defined between the composite plate and the rigid impactor.

Fig. 1 - a) Dimensions of impact test specimens (in mm); b) FE mesh

Newsletter EnginSoft Year 8 n°4 -

27

modeFRONTIER – LS-DYNA process integration The definition of some numerical parameters of the LS-DYNA model may be a considerable challenge because of several reasons (high uncertainty, no reliable data available in literature, etc.) and they usually are chosen on the basis of the analyst’s experience. In order to better understand the influence of such parameters on the simulation results, a sensitivity analysis was performed by coupling the LS-DYNA FE model with modeFRONTIER, a process integration and design optimization tool for exploring the design space (i.e. the free parameters dominions) and finding the configurations which fullfill several objective Fig. 3 - Scatter matrix chart functions. The integration of the LS-DYNA FE model described above into the modeFRONTIER environment of an efficient exploration of the design space. Looking at is roughly described by the workflow in Figure 2. From the the performances provided by these configurations, the top to the bottom the so-called “Data flow” can be seen.

“Scheduler” node starts to generate completely new designs based on various optimization algorithms with the aim of The blocks on the top define the input variables for which a achieving the defined goals. suitable range of variations was set. In particular the input variables object of this analysis were: the damping constant Sensitivity Analysis (variable “sf” in the DAMPING_PART_MASS card), the time In order to study the interaction between the input variables step size (variable “tssfac” in CONTROL_TIMESTEP card), the and the three chosen objectives a statistical analysis was penalty contact stiffness (variable “sfs” in the CONTACT card) performed by evaluating an initial population of 144 designs and the shear stress parameter (variable “alph” in MAT54 generated by using the Full-Factorial method with 3 levels for card). Each time a new combination of their values is the variables “alph” and “tssfac” and 4 levels for the proposed by the modeFRONTIER strategy the LS-DYNA input variables “sfs” and “sf”. file is updated and a new LS-DYNA analysis is performed in The scatter matrix chart, which is a very useful tool to batch mode. analyze the data of a statistical analysis, is shown in Figure 3. It is a matrix with 7 rows and 7 columns (4 input variables The output of each simulation is then post processed and the + 3 objectives) which contains in a matrix form three kinds objectives of the process are evaluated. The output of the of information: the Probability Density Function chart for analysis used in this study were the contact force time each variable (along the diagonal), all the pairwise scatter plot (above the diagonal) and the correlation values between the variables (below the diagonal). For example the first row and fifth column of the matrix represents the scatter plot of the variable “alph” vs. the objective “delta_energy”.

The correlation value is a normalized index spanning from -1 to +1: a value equal to +1 (-1) denotes a full direct (inverse) correlation, while a low absolute value means low correlation. The correlation value at the ith row and jth Fig. 2 - Sketch of the modeFRONTIER - LSDYNA workflow column can be also seen as the slope of history, the plate deflection time history and the absorbed the linear regression line shown in the scatter plot at the jth energy. These numerical results were compared to the row and ith column. experimental ones during the post-processing phase and the relative errors were computed. Such errors, which will be It was found that the variable “alph” is the least significant indicated respectively as “err_f_min”, “err_d_min” and input variable (low correlation with the 3 objectives) and “delta_energy”, were thus the objective functions to be thus it can be considered a constant in the next optimization minimized. The block DOE means “Design of Experiment”. The analysis reducing the number of input variables. On the other user can use this block to generate a suitable initial hands the damping constant and the penalty contact population (combinations of input variable values) in respect stiffness were found to affect significantly the results of the

28 - Newsletter EnginSoft Year 8 n°4

Fig. 4a - 3D Bubble Chart

Fig. 4b - Contact Force time history curve

analyses. Finally it was found that two pairs of objectives (“err_f_min”/“err_d_min” and “err_d_min”/“delta_energy”) are negatively correlated, that means that such objectives are conflicting and thus an optimization strategy should be used to find a good compromise.

The peak force and deflection, the impact duration and the energy absorbed by the plate are predicted by the model with a very good accuracy.

Optimization analysis The modeFRONTIER workflow was simplified according to the findings of the statistical analysis and a multi-objective optimization analysis with the algorithm MOGA-II was performed. The optimization strategy provided, in less then 3000 evaluations, several candidate optimal solutions.

They can be easily detected in the 3D bubble chart of Figure 4a where each solution is represented by a bubble in the 3D plane of the objectives.

A good configuration (the one which minimizes the three objectives) should stay bottom left in the chart and should be blue. Among the others the configuration 2940 highlighted in Figure 4a was considered a good compromise between the minimization of the three objective variables.

Conclusions An LS-DYNA – modeFRONTIER coupled procedure was proposed to simulate low velocity impact on composite plate. The procedure allowed to study the influence of some numerical parameters on the simulation results and to find the configuration that provide the best correlation between the numerical results and the experimental ones in terms of contact force, deflection and absorbed energy time history.

Moreover, the procedure allowed to take advantage from the modeFRONTIER automation: once the workflow was set in modeFRONTIER, the calculations run automatically and all the available time (night, weekend, etc.) is fully used.

Rosario Borrelli - CIRA - Italian Aerospace Research Centre, Capua, Italy www.cira.it

The correlation between the numerical results obtained with this configuration and the experimental ones, in terms of contact force, deflection and absorbed energy time histories are shown in Figures 4b, 5a and 5b, respectively.

Fig. 5a - Deflection time history curve

Fig. 5b - Absorbed Energy

Newsletter EnginSoft Year 8 n째4 -

29

Multi-objective Optimization with modeFRONTIER Applied to Systems Biology Systems biology, the art of simulating biological processes in a computerized environment, is of growing interest due to numerous applications for e.g. the pharmaceutical industry.

In this article modeFRONTIER was used to automate and optimize an analysis model written in MathModelica, a modelling and simulation software based on Modelica.

In the laboratory measurement process, human fat cells are exposed to insulin and the levels of certain indicator proteins are measured as the response.

Fig. 1 - Six insulin molecules assembled in a hexamer, the form in which the hormone is stored in the human body. Source: Wikimedia

Insulin signalling When the body detects glucose in the blood, e.g. after digesting a meal, the hormone insulin is released to signal various cells, such as fat cells, to absorb the glucose from the blood to prevent the blood sugar levels from becoming toxic.

In this study, a MathModelica model of this process, shown in figure 2, was run through modeFRONTIER for optimization and analysis.

Fig. 2 - The MathModelica model

The goal of the optimization process was not to identify a single solution to the model-fitting problem, but rather multiple solutions with acceptably small errors but at the same time with as widely varying parameters as possible. By identifying model properties shared among these different solutions, future experiments could be planned to further improve the model.

Optimization and clustering analysis A significant number (tens of thousands) of MathModelica simulations were run through modeFRONTIER (see figure 3 for the workflow setup) and several thousand solutions with an acceptably small error between measurement data and model predictions were identified. Since the goal was to identify different sets of solutions, a Partitive Clustering Analysis was carried out on the data.

In clustering, the goal is to identify groups (clusters) of similar solutions. A cluster is well-defined if the mathematical distances between its centroid (centre-point) and those of its neighbouring clusters are large compared to the distances between the points in the cluster and its centroid. The Davies-Bouldin index, best described as the ratio of intra-cluster to inter-cluster distances, is an indicator of the quality of the clustering. The lower the index, the better separated the clusters are from each other.

Figures 4 and 5 show the results of the clustering. In figure 4, all designs have been colour-coded according to which cluster they have been assigned. In figure 5, the centroids for each cluster are shown. The difference between parameters v1ak1 (first on the left) and v1rk1 (one step to

30 - Newsletter EnginSoft Year 8 n°4

Fig. 3 - The modeFRONTIER workflow

the right of the middle) is illustrated in this chart: for v1ak1, all the clusters have similar values, whereas for v1rk1, different clusters have different values. The conclusion to be drawn here is for any good fit of the model, v1ak1 will have the same value but we can find different values for v1rk1 which all generate good results. This matches the biological behaviour where different people have different body chemistries, yet still manage not to die from blood sugar poisoning.

Conclusions By using Partitive Clustering Analysis, one of the tools available in modeFRONTIER for Multivariate Analysis (MVA), information regarding complex system behaviour was identified that could not readily be understood using the normal tools available in the Design Space such as Scatter Charts and Parallel Charts.

Fig. 4 - The Partitive Clustering Analysis identified 8 separate clusters.

The data extracted from the analysis regarding the different solution clusters could then form a baseline for determining future experiments and measurements.

For more information: Adam Thorp, EnginSoft Nordic a.thorp@enginsoft.se Thanks to Elin Nyman at Linköping University for help with modelling and simulations.

For an animated explanation of insulin signalling, please watch the movie “Insulin Signaling (Signal Pathways)” at: http://www.youtube.com/watch?v=FkkK5lTmBYQ

Fig. 5 - A plot of the cluster centroids highlights the differing behaviours for the parameters

Newsletter EnginSoft Year 8 n°4 -

31

Eccellenza tecnologica e qualità: Almacis

L’Almacis è stata costituita nel 1987 nell’ambito del più ampio progetto di ristrutturazione dell’attività delle aziende del gruppo Marramiero, operante dal 1955 con l’impresa Marramiero. L’Almacis è in grado di progettare, realizzare chiavi in mano, garantire servizi di manutenzione, gestire, telegestire e curare tutte le pratiche relative ad impianti di

co/trigenerazione, centrali termiche ed elettriche, antincendio, idroelettrico e fotovoltaico in modo del tutto personalizzato e calzante con le differenti esigenze di ciascun cliente, tutto tramite personale proprio con costruzione e preassemblaggio nelle proprie officine.

Parallelamente al settore impiantistico, ugualmente leader nei propri ambiti, l’azienda dispone di un settore edile e di uno di realizzazione di reti gas ed acqua che, oltre ai propri mercati, garantiscono un’ulteriore completezza a progetti di cogenerazione ed impiantistica in tutte le loro fasi.

Fra le numerosissime referenze, l’Almacis può vantare collaborazioni con le aziende più importanti sul panorama mondiale, fra cui: Gucci, Procter&Gamble, Fater, Janssen Cilag, Angelini, Merck-Serono, Ibi Lorenzini, Marangoni, Merker, ecc…

Visitate il sito di Almacis all’indirizzo: www. almacis.it

L’utilizzo di ANSYS nella progettazione ANSYS verrà utilizzato per la verifica strutturale di scambiatori di calore, diverter, piping e strutture portanti di diverse apparecchiature.

Inoltre verrà utilizzato anche nell’area “ricerca” per tutti i prototipi da realizzare e testare nelle proprie officine Almacis.

Perché Enginsoft ed ANSYS in Almacis “A seguito di una attenta valutazione tecnica delle soluzioni software (ANSYS, Comsol…) abbiamo scelto ANSYS perché rappresenta a nostro avviso la miglior tecnologia attualmente presente sul mercato per le nostre esigenze” – ha dichiarato l’Ing. Emiliano Grande Responsabile Ufficio Analisi di Almacis. “La potenza del software, la sua versatilità e semplicità di utilizzo ci hanno convinto che ANSYS può rappresentare un vero e proprio fattore di vantaggio competitivo e di crescita tecnica per l’Almacis – ha continuato l’Ing. Grande - inoltre EnginSoft, a differenza di altre soluzioni che abbiamo preso in considerazione, ha dimostrato di essere un partner con forti competenze nel nostro settore industriale e ci affiancherà nella fase iniziale di start-up consentendoci di essere indipendenti nel più breve tempo possibile”.

32 - Newsletter EnginSoft Year 8 n°4

CAE Simulations and Innovations within the High Performance Computing HPC Finite Element Analysis, Computational Fluid Dynamics and all the CAE simulations and innovations within the High Performance Computing HPC are top-end market segments requiring the incredible processing power provided by today’s processors and system designs.

AMD has recently announced the availability of its AMD Opteron™ 6200 and 4200 Series processors (formerly code-named “Interlagos” and “Valencia”).

The new AMD Opteron processors are designed to provide: better performance for business, increased scalability for Virtualization and the best efficient economics for Cloud environments thanks to the reduced power consumption, up to less than 5W of power per core.

The new AMD Opteron™ 4200 Series processor is the world’s lowest powerper-core server processor. It has been built to deliver unparalleled efficiency different workloads. Designed for powerconscious cloud deployments and ideal for IT infrastructure and collaboration, the AMD Opteron 4200 Series processor was developed from the ground up to handle demanding server workloads at the lowest possible energy draw.

The new processors deliver unparalleled performance, scalability and efficiency with more cores than the previous Roberto Dognini, Commercial Account Executive AMD Italia generation for handling more threads per The new AMD Opteron™ 6200 Series processor is the world’s node, thanks to the new instruction set including AVX, FMA4 first and only 16-core x86 server processor, providing the and XOP, significant memory controller enhancements, the highest core density for incredible scalability to handle exclusive new Flex FP for 256-bit floating point processing demanding multi-threaded workloads such as cloud and additional features designed for HPC. computing, virtualization, high-performance computing (HPC) – this technology is already at the heart of several of The trends in enterprise computing are driving down two the fastest HPC systems in the TOP500 list and business distinct paths: toward greater performance and scalability or application datacenters. toward greater energy efficiency and value. AMD uniquely addresses both of these with its leadership technology.

With this in mind, AMD is proud to say that the new generation of AMD Opteron™ processors 6200 Series based on the innovative “Bulldozer” core architecture, have been chosen to power the most important AMD partners’ solutions, as HP ProLiant, Dell PowerEdge, Acer, E4 Computers, IBM, Cray, to provide a superior performance, efficiency and scalability with a greater CPU and memory density within the same or even less floor space and power envelope.

The new AMD Opteron processor

HPC is not a one-size-fits-all environment. It is a demanding one that requires new technologies to keep pace with customer’s needs. Whether its the “Interlagos” processor or AMD energy-efficient APU, AMD’s unique x86 and world-class graphics IP place AMD at the heart of some of the fastest systems as we push well beyond the PetaFlop moving towards to the next step: the ExaFlop.

Newsletter EnginSoft Year 8 n°4 -

33

DIGIMAT per la modellazione avanzata dei materiali EnginSoft ha recentemente firmato un accordo con la società Belga E-Xstream Engineering per la distribuzione in Italia del pacchetto software DIGIMAT per la modellazione dei materiali.

DIGIMAT, è un modellatore avanzato, non lineare, multi-scala di materiali e si pone come obiettivo quello di offrire una rappresentazione completa e rigorosa utile sia ai fornitori di materiali (“progettisti” di materiali), sia ai progettisti analisti CAE (end users) per i quali, il più delle volte, il materiale viene modellato in modo semplificato.

A seconda della complessità e della tipologia del problema in esame, DIGIMAT permette una trattazione analitica, semianalitica o numerica della microstruttura del materiale con la finalità, tramite algoritmi di omogeneizzazione, di fornire un modello rappresentativo in una scala dimensionale tipica delle strutture meccaniche e quindi utilizzabile dai maggiori solutori FEA commerciali.

L’utilizzo di DIGIMAT in combinazione con i solutori FEA permette lo studio dei complessi fenomeni di plasticità, di danno e di rottura di materiali quali tecnopolimeri, gomma, fibra di carbonio, metalli duri, nanocompositi, ecc. e può quindi rendere la simulazione numerica estremamente predittiva anche nell’indagine di fenomeni complessi per materiali non canonici.

Per ulteriori informazioni: Alfonso Ortalda, EnginSoft info@enginsoft.it

Per una previsione rapida e precisa del comportamento non lineare dei materiali multifase attraverso la tecnologia di omogeneizzazione Mean-Field.

Per una previsione precisa del comportamento non lineare a livello locale e globale di materiali multifase, attraverso un elemento di volume rappresentativo (RVE – Representative Volume Element).

Per la preparazione, l’archiviazione, il recupero e lo scambio in completa sicurezza dei modelli di materiali DIGIMAT tra i fornitori di materiali e gli utenti, favorendo contemporaneamente la protezione della proprietà intellettuale (IP – Intellectual Property).

Interfacce per lo stampaggio ad iniezione e codici strutturali FEA per una previsione precisa dei materiali compositi e delle prestazioni delle parti di plastica rinforzate, avvalendosi di una tecnologia di modellazione multiscala e non lineare.

Per una mappatura efficiente di grandezze scalari e vettoriali tra mesh di tipo shell e solido.

Per una progettazione facile e efficiente di pannelli sandwich a nido d’ape, avvalendosi delle più avanzate tecnologie di modellazione dei materiali.

34 - Newsletter EnginSoft Year 8 n°4

LIONsolver: Learning and Intelligent Optimization LIONsolver is a smart software to build models, optimize them and visualize them interactively.

The name LION stands for â&#x20AC;&#x153;Learning and Intelligent OptimizatioNâ&#x20AC;&#x153;, a combination of learning, modeling, problem-solving and optimization.

LIONsolver has been developed with the aim of offering a multi-purpose tool; a flexible framework that allows the users to reproduce their business or design process through modeling and data analysis.

Finance, logistics, bio-technology and, of course, engineering are some of the areas where a need for modeling, optimization and visualization frequently occurs; LIONsolver with its versatile modular structure has been designed to deliver accurate results in all possible different domains.

Whether your analysis requires a prediction of the right amount of material X in the design of component Y or to determine the optimal composition of your financial stock portfolio to minimize risk and maximize profit, you are dealing with problems that, even though very distant in the real world, can be both managed and solved with LIONsolver.

User-friendly interface One interesting feature of the software is the clear and intuitive interface.

The software works entirely with drag and drop.

Everything in LIONsolver is an item in a list that can be dragged and dropped to a workbench, where it can be connected to other items in a typical flow-chart process.

What it takes to start is: a valid set of data and the business process in your mind.

Then it is all about picking up the tools you need to recreate your business process by linking the icons (representing the tools you have selected) on the workbench and by connecting them to your data.

Fig. 1 - Interface

Build your model Do you want to perform predictive analysis? Do you need to shed some light on your forecast?

Connect the dots between your data inputs and outputs and find the rules that lie behind them.

LIONsolver comes with an ample set of integrated modeling tools: polynomial fitting, neural networks, local regression just to name a few.

Once you have determined a model that works well with your data, all you have to do is save it and then reuse it for your predictive analysis.

The process is just as simple as connecting two icons with an arrow on your workbench.

Fig. 2 - Modeling

Optimize your model Single and especially multi-objective optimization problems are often very tough to solve.

A crucial component of LIONsolver is Reactive Search Optimization (RSO), a robust and efficient method for solving difficult problems.

The word reactive hints at a ready response to events while alternative solutions are tested. Its strength lies in the introduction of high-level skills often associated to the human brain, such as learning from the past experience, learning on the job, rapid analysis of alternatives, ability to cope with incomplete information, quick adaptation to new situations and events.

RSO fast and accurate results in optimization problems are the trademark of LIONsolver.

Newsletter EnginSoft Year 8 n°4 -

35

Interactive visualizations LIONsolver offers a wide selection of data visualizations.

Spanning from the classic ones (pie charts, bubble charts, line plots) to the most advanced ones (7D plot, similarity maps). All visualizations are displayed in the dashboard area, where all panels can be rearranged according to preference; moreover all visualizations are refreshed in real-time, so that your filtering or focusing operations are displayed concurrently.

The RSO philosophy is fully embedded within LIONsolver's interactive visualizations: investigating locally optimal solutions is just a matter of a mouse click on a bubble chart.

For more information: www.lionsolver.com Roberto Battiti - Reactive Search battiti@reactive-search.com

Fig. 3 - RSO, pareto

Fig. 4 - Visualization

IL LIBRO CHE VI SUGGERIAMO

Strategia oceano blu.

Vincere senza competere Attraverso uno studio condotto in oltre trenta settori Kim e Mauborgne hanno elaborato un modello sistematico, replicabile da qualsiasi impresa, per raggiungere alti livelli di crescita. Dal "Modello T" della Ford allo "iPod" di Apple, essi hanno identificato i principi e gli strumenti per neutralizzare la concorrenza e creare uno spazio di mercato incontestato, dalle possibilità illimitate come quelle di un oceano blu. Strategia Oceano Blu porta un messaggio carico di ispirazione: il successo non dipende dalla concorrenza spietata né da costosi budget di marketing e R&S, ma da mosse strategiche brillanti, adatte a un uso sistematico da parte di tutte le imprese.

Dettagli del libro Titolo: Strategia oceano blu.

Vincere senza competere Autori: W. Chan Kim, Renée Mauborgne Editore: Etas Collana: Management Data di Pubblicazione: 2005 ISBN: 8845308480 ISBN-13: 9788845308482 Pagine: 288

36 - Newsletter EnginSoft Year 8 n°4

GPU ACCELERATED ENGINEERING with ANSYS also required for going parallel for greater than 2 CPU cores. For academic license users, the GPU capability is included with the base ANSYS Academic license that provides access to ANSYS Mechanical.

How much more could you accomplish if simulation times could be reduced from one day to just a few hours? As an engineer, you depend on ANSYS Mechanical to design high quality products efficiently. To get the most out of ANSYS Mechanical 13.0, simply upgrade your NVIDIA Quadro GPU or add a NVIDIA Tesla GPU to your workstation, or configure a server with NVIDIA Tesla GPUs, and instantly unlock the highest levels of ANSYS simulation performance.

Here is an example of the speed-up you can reach within ANSYS13.

With the upcoming ANSYS Mechanical 14.0 engineers will even more benefit from NVIDIA GPUs.

With ANSYS® Mechanical™ 13.0 and NVIDIA® Professional GPUs, you can improve your product quality with 2x more design simulations or you can develop high fidelity models with practical solution times. This accelerates your timeto-market by reducing engineering cycles.

The amount of acceleration achievable when using the GPU will vary greatly depending mostly on the model of the simulation, but also on the hardware configuration being used. To get the best speed-up the simulation should spend most of its time in the matrix solver operations rather than other tasks, such as matrix assembly. Also the problem size should be between 500K to 8,000K DOFs for the sparse direct solver and 500K to 5,000K DOFs for PCG/JCG iterative solvers.

To unlock the GPU feature in ANSYS Mechanical 13.0, you must have an ANSYS HPC Pack license, the same scheme

NVIDIA and ANSYS have collaborated to bring you the power of GPU computing for ANSYS. With the latest release of ANSYS R13, NVIDIA GPU acceleration enables faster results for more efficient computation and job turnaround times, delivering more license utilization for the same investment.

This will continue with even more features and optimizations in the upcoming release of ANSYS.

Newsletter EnginSoft Year 8 n°4 -

37

EnginSoft continua l’attività sui materiali compositi La progettazione di un componente in materiale composito rappresenta una sfida complessa ad elevato contenuto tecnologico che coinvolge attualmente settori industriali profondamente diversi, dall’aerospace alla nautica, sino all’automotive ed applicazioni sportive più spinte. Il CAE svolge un ruolo sempre più importante in questo senso, rappresentando lo strumento in grado di riprodurre in maniera fedele ed accurata un prototipo virtuale dei componenti realizzati in materiale composito. ESAComp ed ANSYS Composite Prep/Post rappresentano ad oggi lo stato dell’arte dei software per la simulazione dei compositi; il loro avvento, sin dalle prime release, ha permesso di superare definitivamente i limiti intrinseci del classico approccio seguito per la progettazione delle strutture in composito, consentendo una caratterizzazione dettagliata dei materiali di base (fibre, matrici, core in schiuma o honeycomb, ecc.), una accurata gestione della laminazione attraverso la simulazione delle fasi tecnologiche di stesura dei tessuti (Draping & Flat Wrap) e una dettagliata verifica degli stati tensionali avvalendosi di Failure Criteria polinomiali (Tsai-Hill, TsaiWu, ecc.) e basati sulla natura fisica dei compositi (Hashin 2D/3D, Puck 2D/3D, ecc.).

EnginSoft, attraverso l’organizzazione di seminari a tema dal taglio fortemente tecnico, è costantemente impegnata in attività di formazione avanzata; l’obiettivo principale è quello di sensibilizzare le società leader nel settore ed i principali istituti di ricerca nella valutazione dell’efficienza dei nuovi software numerici al fine di affrontare in maniera efficace anche le problematiche più ostiche e profonde.

Il seminario “Progettazione delle strutture in materiale composito”, svolto il 21 ottobre a Verona nel contesto dell’”EnginSoft International Conference 2011”, è stata un’ottima occasione di ritrovo per tutti coloro che quotidianamente si ritrovano a dover affrontare tematiche complesse relative al mondo dei compositi; i consensi raccolti dimostrano che l’evoluzione del CAE, attraverso l’avvento di strumenti di prototipazione virtuale come ESAComp ed ANSYS Composite Prep/Post, ha generato un nuovo modo di concepire la fase di progettazione delle strutture, attraverso una sensibilità completamente rinnovata focalizzata all’efficienza computazionale, alla

drastica riduzione del time-to-market ed all’accuratezza dei risultati raggiunti. Il seminario è stato replicato il 4 novembre a Marina di Ravenna, nell’ambito dei “Seminari Nautilus” organizzati dalla Facoltà di Ingegneria dell’Università di Bologna.

L’evento anche in questa occasione è stato seguito con particolare interesse da operatori del settore industriale, in particolare nautico, e della ricerca scientifica. La formazione avanzata sui nuovi software di simulazione per le strutture in composito rappresenta senz’altro un punto cardine per EnginSoft, che continuerà ad investire in eventi e seminari mettendo a disposizione competenze e strumenti CAE d’avanguardia.

Per ulteriori informazioni: Fabio Rossetti, EnginSoft info@enginsoft.it

38 - Newsletter EnginSoft Year 8 n°4

EnginSoft presenterà la release 5.2 di MAGMA a METEF 2012 Il Metef, la fiera di riferimento per l’industria metallurgica, si terrà presso la Fiera di Verona dal 18 al 21 Aprile 2012.

EnginSoft, come consuetudine, sarà presente con uno spazio espositivo in cui verranno presentate le nuove release dei sotware sostenuti relativi alla simulazione di processo, in particolare ci sarà la preview di MAGMA 5.2.

A novembre 2009 è uscita la prima versione di MAGMA 5, la 5.0, che permetteva, in un ambiente completamente nuovo,

di affrontare virtualmente tutti i processi di fonderia basati su sabbia (ferrosi e non ferrosi).

Con la versione 5.1, attualmente disponibile, sono stati integrati tutti i moduli, consentendo agli utenti di affrontare lo studio di tutti i processi di fonderia, dalla conchiglia in gravità alla bassa pressione, alla pressocolata in camera calda e in camera fredda ecc.

Per i primissimi mesi del 2012 è prevista l’uscita della versione MAGMA 5.2. In questa versione sarà disponibile MAGMA C+M, un nuovo modulo, che permetterà di simulare la produzione delle anime con diversi tipi di leganti e di sabbie. Questo modulo consentirà di simulare la fase di riempimento delle casse d’anima e la fase di indurimento delle anime, permettendo di valutare le problematiche del processo produttivo e porvi rimedio con soluzioni correttive.

MAGMA 5.2 consentirà inoltre, nell’ambiente di visualizzazione dei risultati (postprocessore), di confrontare direttamente simulazioni di differenti versioni permettendo di analizzare i risultati sia come singola immagine che come filmato in stato di avanzamento.

Sarà possibile sincronizzare i filmati delle versioni a confronto per garantire una più semplice ed

efficace comparazione dei risultati selezionati. Grazie al tool “User Results”, presente nell’ambiente di visualizzazione dei risultati, sarà possibile elaborare nuovi criteri di valutazione, combinando i risultati forniti dal software. Tale procedura sarà resa possibile da un fornito compilatore matematico.

Sarà infine possibile sfruttare la visualizzazione dei risultati sfruttando sistemi 3D, che permetteranno una visualizzazione in profondità dell’oggetto analizzato.

MAGMA 5.2, come l’attuale versione 5.1, sfrutta la tecnologia Java, che permette un diretto interfacciamento con gli attuali sistemi operativi Linux e Windows a 64 bit, a garanzia delle più elevate performance di calcolo.

METEF-FOUNDEQ, giunto alla nona edizione, rappresenta l'evento di riferimento per le tecnologie per l'alluminio e la fonderia. Grazie alle tante iniziative messe in campo, anche nel 2012 METEF-FOUNDEQ, attrarrà buyer da tutto il mondo interessati ad acquistare impianti, macchine, attrezzature per la produzione e la trasformazione dei metalli; componenti estrusi, colati e laminati; prodotti e materiali per il trattamento e la finitura.

Per ulteriori informazioni: Piero Parona, EnginSoft info@enginsoft.it Sito web dell’evento: www.metef.com

Newsletter EnginSoft Year 8 n°4 -

La simulazione di processo nella progettazione di radiatori Estetica e integrità di prodotto sono due fattori fondamentali per la produzione di radiatori, ma altrettanto importante è saper rispondere alle esigenze del mercato in tempi rapidi con costi competitivi. Il processo produttivo utilizzato per questo genere di produzione corrisponde alla colata in alta pressione. Tale processo, per le caratteristiche e i ridottissimi tempi di produzione dovuti all’iniezione forzata della lega negli stampi, richiede la massima precisione ed il controllo assoluto dei parametri imposti alle macchine da pressocolata.

Per assolvere a queste richieste è fondamentale ridurre al massimo gli sprechi di produzione, comprimendo il più possibile i tempi di progettazione/realizzazione del prodotto.

In questo contesto la progettazione prodotto/processo assume un ruolo di considerevole importanza: è infatti in questa fase molto delicata dove vengono valutate le soluzioni più efficaci per la realizzazione delle attrezzature ed i più adeguati parametri di processo. Sviluppare ed ottimizzare un processo produttivo significa identificare le variabili che maggiormente influiscono sulle caratteristiche del prodotto, valutandone gli effetti. Questo può essere perseguito attraverso un approccio al lavoro di progettazione che includa la simulazione di processo.

Il caso che verrà proposto all’High Tech Die Casting 2012, che si terrà a Vicenza il 9 e 10 Febbraio 2012, riguarda la produzione di una specifica linea di radiatori progettati e prodotti dal Gruppo Ferroli. EnginSoft è stata coinvolta nell’attività di riprogettazione delle attrezzature al fine di ridurre al massimo gli scarti presenti nella linea produttiva, incrementando al massimo la qualità estetica e di tenuta del prodotto. Lo studio svolto ha avuto come obiettivo principale la ricerca del miglior sistema di colata per ottenere la massima qualità del componente e ridurre al minimo il rischio di inglobamenti d’aria, principale causa di scarto nel processo di pressocolata in produzione. La riprogettazione delle attrezzature ha inoltre permesso di incrementare la produttività e ridurre i costi. La collaborazione fra il Gruppo Ferroli ed EnginSoft ha determinato il successo del progetto, permettendo un rapido avvio della produzione.

Per ulteriori informazioni: Giampietro Scarpa, EnginSoft info@enginsoft.it

39

modeFRONTIER Users’ Meeting 2012

This year the International modeFRONTIER Users' Meeting 2012 (UM12), sponsored by ESTECO, will take place on 21st and 22 May 2012 at the Savoy Palace in Riva del Mandracchio Excelsior in Trieste.

UM12 provides a unique forum to discover how engineering and academic experts apply the latest methods and techniques to optimize simulation design processes. The meeting of global significance has traditionally brought together experts from leading companies such as FIAT, Honda, Jaguar, Bombardier and many others. The issues relate to the operating logic of modeFRONTIER and its applications in different companies with high technological interest.

ESTECO's biannual event is coming to its 5th edition, marking 10 years of exchange of best practices and ideas among modeFRONTIER enthusiasts.

The leit-motiv of 2012 is Collaboration: nowadays sharing knowledge and team working are imperatives for any successful company. Technology helps breaking the barriers between disciplines, teams and field, encouraging knowledge sharing and enhancing working in team. It is not by chance that this concept is the main theme of the upcoming event, as modeFRONTIER provides a unique multidisciplinary software platform utilized in a wide range of fields all over the world.

UM12 is not just a meeting of modeFRONTIER users, but it’s open also to the academic world: students and researchers are welcome to attend the event, and have the chance to look closely at industrial applications while getting the possibility to present in front of a knowledgeable audience.

Guest of honor of the 2012 edition is David Edward Goldberg, the leading expert of genetic algorithms, although his expertise spans multiple disciplines. He has been Director of Illinois Genetic Algorithms Laboratory (IlliGAL) and Professor at the Department of Industrial and Enterprise Systems Engineering (IESE) of the University of Illinois. He will present two talks: one about collaborative engineering as part of the official UM12 agenda, and another one, open to the general public, concerning the relationship between higher technical education and society.

For more information: http://um12.esteco.com/um12/

40 - Newsletter EnginSoft Year 8 n°4

EnginSoft GmbH Silver Sponsor at the ANSYS Conference & 29th CADFEM Users’ Meeting 2011 EnginSoft GmbH sponsored the ANSYS Conference and Users’ meeting, held this year at the Stuttgart International Congress Center. The ANSYS Conference focused on Electric Mobility Technologies, Machine Tools, Wind Energy Systems, Electronic Products Design, Building and Environmental design and Bio-Engineering Simulation.

that are difficult to combine when running aerodynamic shape optimization, although both the concepts are required to advance the transportation industry. High Speed Trains have to withstand the increasing efficiency requirements and emissions restrictions, hence major efforts are ongoing to innovate their aerodynamic design. In particular, train shape

The aim of the conference was to inform about the most recent methodologies for virtual prototyping and simulations.

From 19th to 21st October, the Stuttgart International Congress Center hosted engineers and researchers from industry, research and education institutions, who shared best practices and recent outcomes from their simulation projects.

The 29th ANSYS Users' Meeting started with a welcome speech by Jim Chashman (ANSYS CEO). The conference this year hosted over 1000 attendees, 200 technical presentations from Industrial Companies and Universities and 27 Technical Seminars.

Thanks to the wide exhibition area available, the conference also gave the opportunity to engineers and ANSYS partners of a fruitful exchange of ideas.

In addition to the more established engineering applications -like structural-mechanics, fluid-dynamics, electrical mechanics, a number of lectures and seminars focused on Engineering Systems Simulation and Optimization have been performed. Today, Engineering Systems like Car Engines, can be holistically simulated, accounting the physical and behavioral interactions between the subsystem parts.

In the spirit of the conference, EnginSoft GmbH presented and time-lined an aerodynamic shape optimization process, presenting the paper “High Speed Train Aerodynamic Shape Optimization Methodology and Framework Comparison” [T. Newill - G. Buccilli, EnginSoft GmbH].

Train speed and aerodynamics efficiency are two concepts

contributes substantially to the overall aerodynamic performances. Typically, a 3D train design should guarantee an improved ratio between aerodynamic lift force and drag force with respect to reference designs.

To pursue the High Speed Train aerodynamic optimization, EnginSoft GmbH proposed a methodology which used a baseline mesh model of the Train and a set of mesh-morphing control points. Then, instead of re-CADing and re-MESHing, the model was morphed using Arbitrary Shape Deformation algorithms. Finally, Latin Hypercube methods have been used to generate the Design Of Experiments and to identify the optimal Train Shape.

To mesh-morph the Train model, EnginSoft GmbH used Sculptor™ software. Sculptor™ directly modifies any geometry or any mesh model, without using CAD or meshing tools. The software enables CFD analyses of different geometries in short time, without re-generating CAD geometries and meshes. This means that more design variations can be calculated in the same amount of time.

Sculptor™ proved to be useful to find optimal High Speed

Newsletter EnginSoft Year 8 n°4 -

Train design easier and quicker.

Coupled with ANSYS-FLUENT, it allowed finding an improved train shape in just a few days, while with the traditional re-CAD and re-MESH approach, it would have taken several weeks. With subtle shape modifications, a sound 2% increase of overall lift/drag ratio and over 80% simulation time reduction was achieved – without affecting the overall geometry constraints.

Sculptor™ avoided time consuming operations on the CAD model and on the computational grid, since the morphing took place over the ANSYS-FLUENT model directly.

Besides the Train Aerodynamics optimization paper, at the 29th ANSYS Conference EnginSoft held a Seminar on “Product Design Chain Innovation thorugh Manufacturing Process Simulation” [N. Gramegna - Enginsoft Italy].

Today the whole product development chain can be simulated, from manufacturing process to thermalmechanical fatigue behavior and several CAE software are available for that purpose. More in particularly, the design of the manufacturing process (like casting, forging and machining) is gaining importance in product development, as all those processes directly impact mechanical properties and component behavior.

41

During the seminar, EnginSoft presented innovative methodologies for Manufacturing Process simulation, all aimed at reducing product development time and resources needed. Nicola Gramegna gave the attendees an overview of the most relevant manufacturing processes, like Casting Process and Heat Treatment (simulated with the software MAGMASOFT), Forging (simulated with FORGE software) and Machining (simulated by the means of Advantedge software tools). Nicola showed how residual stresses-strains and local mechanical properties can be calculated through computer simulation. Finally, he showed how the non-uniform stressstrain and mechanical properties previously calculated, can be integrated into the FEM model (like ANSYS) to simulate the macro component behavior.

For more information on Sculptor™: Giorgio Buccilli, EnginSoft GmbH info@enginsoft.it

SCULPTOR Sculptor is a powerful tool that allows a user to parameterize any mesh based on arbitrary cubic bezier control points. It can be linked to your existing fluid-flow (CFD) and/or structural (FEA) analysis tools and then deform these meshes and maintain quality in real time.

Enabling the user to optimize a product without the need to remesh, saving you days, weeks, even months.

42 - Newsletter EnginSoft Year 8 n°4

BENIMPACT Suite has landed in China Dal 19 al 26 ottobre si è tenuto a Dalian, in Cina, il primo summit planetario sul basso impatto ambientale - Low Carbon Earth Summit (LCES 2011), che ha riunito esponenti della politica, della ricerca, delle tecnologie e pubblico con l’obiettivo di creare un tavolo intorno al quale scambiarsi le conoscenze oggi disponibili per riuscire a controllare l’impatto sul clima: “Spronare la green Economy per tornare in armonia con la natura”. Al “Forum 8: Clean Sciences and Technology for Low Carbon Environment - Today’s R & D, Tomorrow Industrial Revolutionization” di questo importante incontro è stato presentato anche il progetto CASA ZERO ENERGY, un edificio progettato e realizzato con un approccio “filosofico” che mira ad una visione integrata della sostenibilità.

L’edificio, progettato dall’Università di Trento, è stato realizzato dal Gruppo Polo Le Ville Plus con il supporto della Regione Friuli Venezia Giulia. Numerose analisi di simulazione e ottimizzazione delle prestazioni energetiche ed ambientali sono state eseguite da EnginSoft con l’ausilio di BENIMPACT Suite.

Portavoce dell’attività è stato il prof. Antonio Frattari, Responsabile Laboratorio Progettazione Edilizia (LPE) Direttore del CUnEd dell’Università di Trento.

BENIMPACT Suite è il risultato di un progetto di ricerca cofinanziato dalla Provincia Autonoma di Trento - Legge Provinciale n° 6/99 Programma Operativo FESR 2007-2013 Obiettivo 2.

Last month, from 19th to 26th the first Low Carbon Earth Summit (LCES 2011) was held in Dalian, China.

It brought together important politicians, researchers, and a large audience. The aim of this meeting was to create a round table where people could share their knowledge about controlling the environmental impact: “Leading the Green Economy, Returning to Harmony with Nature”.

Prof. Antonio Frattari, the chief of Building Design Lab of the University of Trento, presented the project ZERO ENERGY HOUSE at “Forum 8: Clean Sciences and Technology for Low Carbon Environment - Today’s R & D, Tomorrow Industrial Revolutionization”. A philosophical approach towards integrate sustainability characterizes this building.

The University of Trento has designed this house, Gruppo Polo Le Ville Plus has built it, and the local administration, Regione Friuli Venezia Giulia, has given its support.

For the simulation of the building behavior BENIMPACT Suite has been used.

BENIMPACT Suite is the outcome of a research project cofounded by the Autonomous Province of Trento (Italy) – Provincial Law n° 6/99 Operative Program FESR 2007-2013 Objective 2.

Newsletter EnginSoft Year 8 n°4 -

43

CasaZeroEnergy can be called in this way because it has a very low energy consumption, it does not use any fossil fuels, and its energy demand is produced using renewable energetic sources. It anticipates the EU directive 31/2010/CE that requires the realization of near zero energy buildings, starting from 2020.

The main features of CasaZeroEnergy are: • a strong bioclimatic characterization; • the use of natural, renewable and recycled materials for the construction of the building; • the development of a new and innovative timber frame system; • the set up of an intelligent system (home automation) to manage the energy consumption; • the integration with energy systems that use alternative and clean sources: photovoltaic plant of 14.6 kWh, solar thermal panels for DHW, horizontal geothermal plant with water – air heat pump integrated with a radiant floor heating and it can also work as a cooling system in summer.

The building behavior has been simulated using BENIMPACT Suite and then compared with the real house behavior, which is monitored.

Comparing the simulation with the monitoring results it is possible to observe some interesting things: • the performed simulation (with only two thermal zones) has been validated with the monitoring; • the building behavior is very good and it meets in a perfect way the expected predictions for summer. except from some temperature picks in the hottest days (June 29th and July 4th) the comfort in the house has been always achieved in the monitored period.

Alta Formazione: TCN punta ad una specializzazione sempre più avanzata Anche per l’anno 2012 il Consorzio TCN erogherà corsi di formazione specialistici e corsi a calendario. Continuerà l’attività di organizzare corsi personalizzati a seguito di specifiche richieste da parte dell’industria.

A questi si aggiungeranno una serie di Minimaster con programmi formativi più intensi ed approfonditi rispetto a quelli dei corsi base ed avanzati.

Per questo nuovo anno c’è anche l’intenzione di inserire corsi che trattano argomenti inediti di attuale interesse.

Tutto sarà coordinato tra il Consorzio TCN ed i responsabili della formazione delle varie realtà lavorative.

Per informazioni vi consigliamo di visitare il sito: www.consorziotcn.it oppure contattare la segreteria organizzativa: Mirella Prestini info@consorziotcn.it

For more information: Angelo Messina, EnginSoft info@enginsoft.it

44 - Newsletter EnginSoft Year 8 n°4

CAE Seminars in Japan “CAE UNIVERSITY” Cybernet Systems Co.,Ltd. (with its headquarter located in Tokyo) has offered a wide range of leading-edge CAE solutions and services for many years since its establishment in 1985. Today, Cybernet sells more than 80 CAE products for diverse applications in mechanical, electrical and control engineering, optics, civil engineering and construction, optimization, bioengineering, nanotechnology and other sectors. To complement its software portfolio for its clients across Japan, Cybernet provides different types of services, such as technical support, training, and consultancy, to companies in manufacturing, as well as to universities and research institutes…and far more than this: The engineers of Cybernet are passionate about supporting MONOZUKURI in Japan, as one of the country’s leading CAE providers. The company’s corporate message states: “Energy for your Innovation”.

To foster the interest in CAE and to support the next generation of CAE engineers, Cybernet developed and introduced an educational system called “CAE UNIVERSITY”.

CAE University’s primary objective is to grow the use of CAE techniques among engineers.

In Japan, the use of CAE technologies in manufacturing has expanded significantly in recent years. While we witness a growing interest in CAE, we also hear that many companies ask for additional support and know-how for improving their application skills and for making the use of CAE more efficient to solve their real problems. CAE UNIVERSITY is a new type of educational system which enables students to learn CAE systematically and continuously. It provides students with the necessary skills to use CAE technologies flexibly and efficiently for the actual requirements in their product design and development activities.

Lectures and practical examples CAE UNIVERSITY offers both, 1-day or 2-day courses, in short periods, on each single topic in different fields.

In lectures and hands-on sessions, participants study intensively theories of mathematics, physics and engineering, which are used in CAE today. By combining different courses, they are able to acquire theoretical knowledge in each field systematically. For example, by attending the 1.5 day lecture on

“Design and CAE Mechanics through Numerical Experiments“, the students learn many applications, from the basic numerical experiment and its theoretical consideration using beam and frame structure, to the modeling of solid structures, thermal stress and anisotropic materials.

FEM Laboratory Nowadays, performing simulation by using CAE has become quite common in design and development departments.

However, engineers sometimes are facing problems when simulation results differ from testing results. By performing testing and by comparing test results with simulation results in the FEM Laboratory, students can study and discuss the factors which sometimes lead to such errors. This helps them to understand the background and how the different steps and techniques are linked; they can now evaluate and verify simulation results correctly and make efficient use of them in their real design and development work.

I had a pleasure to conduct the following interview with Mr. Takashi Sakurai, Manager of CAE UNIVERSITY. Please can you tell us about the positioning and the features of CAE UNIVERSITY? The main features of CAE UNIVERSITY are to offer the curriculum, which meets certain standards based on the University’s educational system and to invite active teachers from universities. The courses are linked with each other and the learning content has been examined carefully to avoid overlapping and insufficiency. Teachers who are in charge of the computational mechanics courses get together for

Newsletter EnginSoft Year 8 n°4 -

curriculum meetings periodically, to check and modify the interaction of each lecture and practice. CAE UNIVERSITY can be regarded as a new CAE education system with a universitylike philosophy that offers high CAE knowledge levels.

Concerning the content, some may think that mostly advanced CAE theory is being taught. The aim of CAE UNIVERSITY though clearly is not limited to the teaching of theory, it also provides the knowledge of advanced techniques of how to use CAE in the right way. Hence the theory is just an element of the teaching content. We believe that a combination of both: learning how to apply CAE and studying theory, will enable us to use CAE effectively for the actual job. Many companies have learned in the meantime that CAE is not a magical tool that will help just by introducing it. Introducing CAE also requires learning methods for its correct use. CAE UNIVERSITY is a reasonable system to learn techniques for CAE usage because it is systematic and continuous.

45

about them. This means touching and trying shapes their imagination and deepens their understanding. Therefore, the “FEM Laboratory”, which offers a combination of lectures and testing is very effective. FEM Laboratory maintains a high reputation and gathers many registrations always. The size of testing is not so large, because it needs to be done on the desk. However, the testing is very well conducted and sufficient to understand the essence. Presently, we offer 2 FEM Laboratory courses, we are planning to add more in the future. Attending these courses, in which the students perform testing in groups, also provides a good platform for them to share information and to get to know each other and each other’s work easily. Often, students from different companies keep in touch afterwards and discuss their own problems with other engineers in similar environments.

What type of students do you have mostly?

Many different types of employees participate. We welcome designers, R&D engineers, analysis specialists and trainers.

We have students from universities too. They usually have different goals, for example reviewing things they have learned many years ago and solving specific problems on the job. We ask everybody who registers to complete a questionnaire before the course; we ask to tell us about requests, expectations and backgrounds. Teachers prepare and try to arrange the course as much as possible following the students’ satisfaction questionnaires. There is another questionnaire that is submitted after each course for future improvement of the courses. With these efforts, we are able to maintain good quality and to constantly gain reputation.

The number of students who register for subsequent courses is rather high.

Recently, we received several requests to hold on-site CAE UNIVERSITYs from customers who greatly appreciate the philosophy and the intent. Our clients more often book now on-site Courses and arrange for their engineers to participate in the entire range of lectures and practical sessions over months, to provide thorough CAE employee training.

Please tell us about your future vision for the “CAE UNIVERSITY” Currently, students need to come to our seminar room to attend CAE UNIVERSITY. This is difficult for someone who has to travel a long way or for those whose schedules are tight.

To improve this, we are planning an alternative way of CAE education. Actually, we already have experiences in delivering the customized CAE UNIVERSITY on the customer’s intranet so that all their engineers can learn while being at work. By using cloud computing, the system can be applied and extended to a wider area and audience. We can indeed offer CAE UNIVERSITY to many more people. Also, if we develop other language versions, it will be possible to share this education system globally. We want to achieve new CAE oriented design innovation by collaborating with as many engineers as possible who have studied and overcome engineering challenges. To reach this goal, we want to build a CAE community, to foster comprehensive CAE development in Japan and in other countries around the world. This is our ultimate vision and goal.

What is the students’ general reaction to the “FEM Laboratory”?

CAE is a tool for design. Designers have got into the habit of doing, looking at and touching real things and thinking

This article has been written in collaboration with CYBERNET SYSTEMS Co.,LTD.: http://www.cybernet.co.jp/english/ Akiko Kondoh, Consultant for EnginSoft in Japan

46 - Newsletter EnginSoft Year 8 n°4

NPO Activity for Implementation of Anisotropic Elasto-plastic Models into Commercial FEM Codes The nonprofit organization JANCAE, The Japan Association for Nonlinear CAE (chairperson: Kenjiro Terada, Tohoku University), offers several activities to companies, universities and software vendors [1] to gain a deeper understanding of nonlinear CAE including its main work, the nonlinear CAE training course held twice a year. This article introduces JANCAE’s efforts to implement anisotropic elasto-plastic models into commercial FEM codes as one of the initiatives of the “Material Modeling Committee”.

2. Background and outline 2.1 The efforts of the Material Modeling Committee: When we think of comprehensive advancements in the accuracy of a simulation, we are aware that all capabilities of the material modeling, the boundary conditions as well as the definition of the geometric modeling have to be improved at the same level. The capabilities of geometric modeling for FEM simulations have drastically improved along with the growth of the 3D CAD market, the advancements in auto-meshing capabilities and the progresses made in hardware speeds and capacities, over the past 10 years. Yet material modeling capabilities have not progressed as significantly as the advances achieved in geometric modeling. Users need to be involved in the definition process of material modeling, which means that they have to choose the appropriate material model from huge amounts of available material models offered by each FEM code. As a next step, the parameters of the material properties have to be determined by performing material tests.

These processes are still necessary, even now, at a time when many sophisticated commercial FEM codes are available.

In this situation and independently from its CAE training course, which mainly consists of classroom lectures, JANCAE organizes “The Material Modeling Committee” as a practical approach to the study of nonlinear materials. The Committee was originally established in 2005 to study mainly hyperelasticity and viscoelasticity. Then, its research activities have diversified into all material nonlinearity including metal plasticity. In the frame of the Committee, members learn about typical nonlinear material modeling by studying the basic theory of the constitutive equations, material testing methods, and how to handle test data and parameter identification techniques.

2.2 User subroutines for constitutive law in FEM Codes There are many constitutive equations of materials, as we can see from the many researchers’ names which appear in the titles of the equations. Although such variety of material models contributes to the improvement of simulation accuracy, not all material models, especially new models, can be applied

to various commercial FEM codes. With regard to yield functions, which are a core concept for metal plasticity, it has been pointed out that the yield surfaces of the actual metal materials cannot be represented well enough by the classical anisotropic yield functions [2]. However now, many different types of new yield functions are proposed especially in sheet metal forming; they are able to represent real plastic deformation much better than before [3].

LS-DYNA provides specific capabilities for sheet metal forming simulation, it also offers a considerable number of new anisotropic yield functions [4]. On the other hand, when we think about other commercial general purpose FEM codes, they usually have only limited kinds of yield functions, such as the classical Hill quadratic anisotropic function.

These commercial codes offer user subroutine capabilities to extend material models. By using these capabilities and defining material models following the programming rules that each code provides, users can implement the required constitutive laws. However in reality, it is difficult for ordinary users who are not familiar with the framework of continuum mechanics, numerical simulation and the theory of plasticity, to perform such processes only from released text books or available information, as the manual definition in FEM codes requires professional skills.

2.3 The development activity in the Material Modeling Committee The Material Modeling Committee started its unique R&D activity in 2009. For this activity, engineers with various backgrounds and skills engaged in the CAE field got together to jointly work on making subroutines for the constitutive laws. The members are from industrial companies and CAE software vendors.

As mentioned above, it is impossible to create such subroutines without understanding the basic concept of elastoplastic models for FEM. In the first year, in 2009, we studied the basics of plastic constitutive equations and the framework

Fig. 1 - Framework of the subroutine “UMMDp”

Newsletter EnginSoft Year 8 n°4 -

of constitutive law subroutines by referring to some text books [5], to obtain a better understanding of their principles. In parallel, we summarized the characteristics of each code’s user subroutine and proposed a framework for the user subroutine development. [Fig.1] In this framework, stress integration and calculation of consistent tangent modulus, which represent basic capabilities of constitutive law subroutines, have been determined as “Unified Material Model Driver for Plasticity (UMMDp)” and separated from each code’s specified rule in order to be able to be used commonly. Additionally, yield functions were isolated as a modularized subroutine so that we can implement different types of yield functions easily. In the second year, in 2010, the members worked on the programming based on this framework and by sharing each role.

3. Development and verification of the user subroutine 3.1 Basic equations of elasto-plastic constitutive laws Here we show the basic part of the subroutine for elasto-plastic constitutive laws, which is crucial for this programming.

Tensor is represented by the Voigt notation arraying components as vector. The stress to be calculated is , and the strain increment given to the subroutine is .

Following are basic equations for elasto-plastic constitutive laws.

(1) (2) (3) (4) (5)

Equation (1) shows the yield condition and represents that stress point on the yield surface. The shape of the yield surface is determined by the yield function , the magnitude is given by the hardening curve showing isotropic hardening and the center of the yield surface is provided by the back stress showing kinematic hardening respectively. Equation (2) shows that the elastic and the plastic strain increments are given by additive decomposition, and the elastic strain increment gives the stress increment

by Hooke’s law shown as Equation (3). Equation (4) gives the plastic strain increment and here the associated flow rule is used, in which the outward normal of the yield surface and the plastic strain increment have the same direction. Equation (5) is the evolution equation of the back stress. p shows the equivalent plastic strain which has a conjugate relation with the equivalent stress in the plastic work. UMMDp uses backward Euler’s method for the stress integration algorithm.

In this method, nonlinear simultaneous equation is solved, assuming the stress and internal variable (back stress and equivalent plastic strain pn+1) after the completion of ”n+1” increment satisfy the basic equations (1) – (5). We now define residual functions as follows.

(6) (7) (8)

47

Now is the trial stress (initial estimate of stress integration) assuming all strain increments are elastic components and given by

Equation (6), Equation (7) and Equation (8) correspond to the yield condition of Equation (1), Equation (2) – (4) and the back stress evolution equation of Equation (5) respectively, and the stress after integration and the internal variable ( and ) are obtained by converging , and to 0 using Newton-Raphson method. In UMMDp, it is predicted that the convergence calculation will be difficult because of implementation of higher order anisotropic yield functions. So we relaxed the condition of (6) by using Multistage Return Mapping [6] which leads to gradual convergence.

3.2 The idea of UMMDp The variables used for convergence calculation of the stress integration are the yield function , the isotropic hardening curve , the back stress evolution equation and their first and second order differentials. In the static implicit method, tangent matrix (Material Jacobian) consistent with stress integration algorithm, is also required to obtain the quadratic convergence in equilibrium calculation. In this calculation, as with the stress integration, yield function, isotropic hardening curve, back stress evolution equation value and its differential value are going to be needed. The frameworks of calculation for stress integration and consistent tangent modules are in common regardless of forms of yield function, hardening curve and back stress evolution equation.

Therefore, if we could make a unified interface for those various sets of functions, the function group of a variety of constitutive equations described above would be able to be modularized and have higher expandability. When we think about the variable names and the stored formats in subroutines of commercial codes, of course they vary from code to code.

However, the role of the constitutive law in FEM codes is to provide “local stress-strain relation at integration point” and there is no difference on this point. By using proper variable conversion for code-independent user subroutines, they can be linked to UMMDp correctly. From this standpoint, the great variety of constitutive equations, such as yield function and hardening law can be externalized. Additionally, if we develop the interface for different commercial codes using their specific user subroutines, which we call “Plug”, it will kick-off an open effort and a discussion which will not be limited to a specific code.

3.3 Yield function subroutine The yield function subroutine is developed mainly by CAE users from industrial companies. Following are the yield functions for the implementation. (von Mises is used for verification of the implementation.) von Mises[7] Hill(1948[8], 1990[9])

48 - Newsletter EnginSoft Year 8 n°4 Gotoh's bi-quadratic yield function [10] Barlat yield function (Yld89[11], Yld2000[12], Yld2004[13]) Banabic yield function (BBC2005[14], BBC2008[15]) Cazacu 2006[16] Karafills & Boyce[17] Vegter[18] The yield function subroutine receives the stress component as the argument, and then returns the corresponding equivalent stress , its first order differential and its second order differential .

To demonstrate objectively that the developed subroutine works correctly, also numerical verification is being performed.

For this verification, we also provide a main routine so that only the capability of the yield function’s subroutine itself can be checked separately without mixing up its bug with other bugs in UMMDp (the parent routine of the yield function’s subroutine), and “Plug” for commercial FEM codes. By doing so, the members can work independently. The verification was performed by the comparison between the yield surface in the original paper, which proposed anisotropic yield functions, and the output from our developed subroutine as shown in Fig.2, as well as by the comparison between analytical and numerical differential values to secure correctness.

3.4 Development of the interface “Plug” for commercial codes The “Plug” subroutine, which becomes an interface to commercial FEM codes, is developed mainly by engineers from CAE software vendors. This subroutine links to UMMDp correctly through each different manner depending on commercial codes. The name of the ‘Plug’ is based on the functional analogy of plug-adopter for AC power socket which differs by nation.

The Plug needs to offer overall capabilities for communication with commercial codes, such as storing and updating internal variables, and variable output adjustment to result data. On this point, it was very helpful to gain the cooperation of engineers from software vendors, who are familiar with each commercial code. We appreciated their cross-border cooperation.

The verification of the developed Plug was also performed. For this verification, we used the basic benchmark test provided by the NAFEMS guidebook [19] for “Code to Code Verification”. We

(a) Yield locus in original paper [12]

(b) Output from ummdp_checkyf

Fig. 2 - Verification of yield function subroutine (eg:Yld2000).

compared the result using default elasto-plastic models prepared in each commercial code (von Mises type isotropic yield functions) and the result using the von Mises type yield function through UMMDp, and we confirmed that these stress histories are matching as shown in Fig.3.

Fig. 3 - Comparison with result of commercial code (von Mises model)

3.5 Implementation of combined hardening law We finalized the development and the verification of the program for the standard isotropic hardening models in 2009.

It is difficult to simulate deformation behavior accurately when the direction of stress is reversed. So we are promoting the development of the combined hardening model including kinematic hardening shown in the basic equations. Kinematic hardening behavior is modeled by back stress evolution equation. For this evolution equation, various types of models are proposed, and we need to accept this diversity as with yield functions. At this point in time, we are developing a framework to modularize the function shown in Equation (5) as a subroutine.

3.6 Total verification For total verification of the developed program, we analyzed problems which come to the surface by the influence of plastic anisotropy, and we compared them to the reliable result. We simulated a hole-expansion test of a steel sheet [20] and a hydraulic bulge test of aluminum alloy [21] in cooperation with Prof. Kuwabara, Tokyo University of Agriculture and Technology. Fig.4 shows the simulation result of the holeexpansion test. We can see that the thickness decrease around the center hole varies with angle from the rolling Fig. 4 - Simulation example of holeexpansion test direction. Afterwards, we verified that the developed subroutine group worked rightly, by comparing the UMMDp simulation result and the reliable simulation result. The aim of the verification at this stage is not the comparison with experimental results, instead it is absolutely for Code to Code Verification. We think that using the middle scale problem, which is positioned between small scale problems like material testing and large scale problems in realistic sheet metal forming, is more important for the material model validation rather than jumping to a complicated large scale problem.

4. Closing In this article, we introduced an activity of the NPO “JANCAE” working group. As the volume of tasks becomes larger, the development is still in progress. In 2011, the development of a common subroutine for resin and rubber has been planned as a subsequent activity of the working group. The effort this

Newsletter EnginSoft Year 8 n°4 -

time is the implementation of the yield functions which were already proposed in previous papers, hence there is no academic novelty. Meanwhile, it is not just about a limited activity for a specific commercial code only. This is why the topic is not really suitable to be presented in academic societies or at specific users’ conferences by CAE vendors. We introduced this work as an example of the activities featuring NPO’s neutrality. Following NPO’s guidelines, it is planned that the subroutine group will be opened to the public a year after activity completion. Yet more than 30 engineers from different organizations have already joined the working group. Their backgrounds are different, some have already obtained permissions from their managers, some join to support their own personal development. In any case, their motivation is the most important driving force for the activity.

C.A Coulomb, when he was a building engineer in the military corps of engineers, expressed the reason to write a paper by making an analogy to an artisan when he submitted the paper to the French Académie des sciences in 1773, as follows.[22] “Besides, the Sciences are monuments consecrated to the public good. Each citizen ought to contribute to them according to his talents…. While great men will be carried to the top of the edifice where they can mark out and construct the upper stories, ordinary artisans who are scattered through the lower stories or hidden in the obscurity of the foundations should seek only to perfect that which cleverer hands have created.” We think the reason why so many engineers were eager to be involved in the work is because of their motivation to understand in a deeper way and to express their sympathy for the activity based on Coulomb’s words. We, ordinary artisans, have great responsibility in the present apprehensions regarding the gap between computational mechanics and CAE [23].

5. References [1] http://www.jancae.org/ [2] The Japan Society for Technology of Plasticity ed.: StaticImplicit FEM – Sheet metal forming (process simulation series), Corona Publishing, pp.198, 2004. (in Japanese) [3] ibid. pp.172.

[4] LSTC, JSOL: LS-DYNA Version 970 User’s Manual Vol.2, 2003.

[5] F.Dunne, et al.: Introduction to Computational Plasticity, Oxford Univ. Pr., 2005.

[6] J.W.Yoon, et al.: Elasto-plastic finite element method based on incremental deformation theory and continuum based shell elements for planar anisotropic sheet materials, Comp. Meth. Appl. Mech. Engrg., vol.174, pp.23-56, 1999.

[7] R.von Mises: Mechanik der festen Körper in plastischendeformablem Zustand, Göttinger Nachrichten math.-phys.

Klasse, pp.582, 1913.

[8] R.Hill: A theory of the yielding and plastic flow of anisotropic metals, Proc. Roy. Soc. A: vol.193, pp.281, 1948.

[9] R.Hill: Constitutive modeling of orthotropic plasticity in sheet metals, J. Mech. Phys. Solids, vol.38, no.3, pp405417, 1990.

49

[10] M.Gotoh: Improvement of orthotropic theory by implementation of forth order yield function (plane stress) I, JSTP journal, vol.19, no.205, pp.377-385, 1978.

[11] F.Barlat, et al.: Plastic behavior and stretchability of sheet metals. Part-I, Int. J. Plasticity, vol.5, pp.51-66, 1989.

[12] F.Barlat, et al.: Plane stress yield function for aluminum alloy sheet: part 1:theory, Int. J. Plasticity, vol.19, pp.1297-1319, 2003.

[13] F.Barlat, et al.: Linear transformation-based anisotropic yield functions, Int. J. Plasticity, vol.21, pp.1009-1039, 2003.

[14] D.Banabic, et al.: Influence of constitutive equations on the accuracy of prediction in sheet metal forming simulation, Proc. of NUMISHEET2008, 2008.

[15] D.Banabic, et al.: Plane-stress yield criterion for highlyanisotropic sheet metals, Proc. of NUMISHEET2008, 2008.

[16] O.Cazacu, et al.: Orthotropic yield criterion for hexagonal closed packed metals, Int. J. Plasticity, vol.22, pp.11711194, 2006.

[17] A.P.Karafillis, M.C.Boyce: A general anisotropic yield criterion using bound and a transformation weighting tensor, J. Mech. Phys. Solids, vol.41, no.12, pp.18591889, 1993.

[18] H.Vegter, et al.: A plane stress yield function for anisotropic sheet material by interpolation of biaxial stress states, Int. J. Plasticity, vol.22, pp.557-580, 2006.

[19] A.A.Becker: Understanding Non-linear Finite Element Analysis Through Illustrative Benchmarks, NAFEMS, pp.20, 2001.

[20] Kuwabara, T., Hashimoto, K. Iizuka, E. and Yoon J.W., Effect of anisotropic yield functions on the accuracy of hole expansion simulations, J. Mater. Processing Technol., 211 (2011), 475-481.

[21] Daisaku Yanaga, Toshihiko Kuwabara, Naoyuki Uema and Mineo Asano: Material Modeling of 6000 Series Aluminum Alloy Sheets with Different Density Cube Textures and Effect on the Accuracy of Finite Element Simulation, Proc.

NUMISHEET 2011, Seoul, Korea, 21-26 August, 2011, pp.800-806. (AIP Conference Proceedings, Volume 1383) [22] Timoshenko, S.P.: History of Strength of Materials, Dover publications, pp.47, 1983.

[23] N.Kikuchi: Computational Solid Mechanics –Trend and Future, JSCES Journal, vol.11, no.1, pp.1290-1295, 2006. (in Japanese) Hideo Takizawa (Mitsubishi Materials Co, Japan) Vice-chairman of JANCAE Material Modeling Committee For more information about this article, please e-mail: takizawa@mmc.co.jp

By courtesy of Mechanical Design & Analysis Corporation, an original version of this article was presented at the 4th Mech D&A Users’ Conference, 1 July 2011 (Tokyo, Japan) and published in the Conference Proceedings.

50 - Newsletter EnginSoft Year 8 n°4

EnginSoft Event Calendar ITALY For more information on the next EnginSoft Seminars and Webinars, please contact: eventi@enginsoft.it Stay tuned to: www.enginsoft.com (Events) Download the 2011 Conference Proceedings now on: www.enginsoft.com/proceedings2011 and stay tuned for the dates/venue of the 2012 International Conference: www.caeconference.com Every year, the conference program features applications of CAE in: mechanics, industrial applications, structural engineering, optimization, manufacturing process simulation, computational fluid dynamics, emerging technologies, durability and fatigue, rapid and impact dynamics, CAD/CAE integration, …

9-10 February - High Tech Die Casting, Vicenza EnginSoft will present a case history of the process simulation applied to Ferrioli radiators. www.metallurgia-italiana.net 18-21 April - METEF 2012, Fiera Verona EnginSoft will present the MAGMA 5.2 release. www.metef.com 15th European Conference on Composite Materials. 24-28 Giugno; Venezia. www.eccm15.org.

3rd Dolomites Workshop on Constructive Approximation and Applications; 9-14 Settembre; Canazei events.math.unipd.it/dwcaa2012/?q=node/1 GERMANY 15-16 November - NAFEMS European Conference: Simulation Process and Data Management (SDM). Munich If you would like to hear more about EnginSoft Germany’s presentation on: Methodology and Validation for Bidirectional, Homogeneous Simulation Data Flow Management in a Fluid-Structure Interaction Problem Utilizing Workflow Management and Shape Deformation Tools, please contact our team at: a.hauschopp@enginsoft.com EnginSoft Germany. Regular Webinars and On-site Presentations 2011 & 2012: EnginSoft Germany hosts regular Webinars to present the company’s products and services, as well as specific Webinars to discuss our customers’ current applications and needs.

To hear more and to fix an appointment for your company,

please contact: a.hauschopp@enginsoft.com.

Please stay tuned to: http://www.enginsoft-de.com/ FRANCE Flowmaster Roadshow 2012 Pour accompagner le lancement de Flowmaster V7.9 et présenter ses principales nouveautés, Enginsoft France organise des conférences dans plusieurs villes de France.

Vous y découvrirez notamment l’analyse diphasique, le temps réel, et le couplage avec modeFRONTIER. Inscrivez-vous vite!

Book your place now, for the Conferences that EnginSoft France will host in 2012 – Hear about Flowmaster V7.9 and the coupling with modeFRONTIER! Voici les lieux et dates – Dates & venues: • 2 février 2012 après midi à Nantes • 7 février 2012 après midi à Lyon • 9 février 2012 après midi à Toulouse • 14 février 2012 après midi à Aix en Provence • 16 février 2012 après midi à Paris Pour vous inscrire, appelez vite le +33 (0)1.41.22.99.30 ou visitez http://www.enginsoft-fr.com/ EnginSoft France 2011 & 2012 Journées porte ouverte dans nos locaux à Paris et dans d’autres villes de France, en collaboration avec nos partenaires.

Pour plus d'information visitez: www.enginsoft-fr.com, contactez: info.fr@enginsoft.com UK The workshops are designed to give delegates a good appreciation of the functionality, application and benefits of modeFRONTIER. The workshops include an informal blend of presentation plus ‘hands-on’ examples with the objective of enabling delegates to be confident to evaluate modeFRONTIER for their applications using a trial license at no cost. modeFRONTIER Workshops Warwick Digital Laboratory, Warwick University • Thursday 10th March • Tuesday 12th April • Tuesday 21st June • Wednesday 17th August • Tuesday 1st November • Wednesday 14th December modeFRONTIER Workshops at Warwick Digital Laboratory, Warwick University

Newsletter EnginSoft Year 8 n°4 -

• • • • • •

Thursday 10th March Tuesday 12th April Tuesday 21st June Wednesday 17th August Tuesday 1st November Wednesday 14th December

modeFRONTIER Workshops at Cranfield University • Thursday 27th May modeFRONTIER Workshops for InfoWorks CS at Warwick Digital Lab • Tuesday, 8th February • Thursday 26th May • Wednesday 20th July • Thursday 13th October • Tuesday 22nd November To register, please visit: www.enginsoft-uk.com 7th December - CIWEM Innovations Showcase, Coventry EnginSoft has been selected to present 'Exploring the full range of possible solutions to DG5 schemes by combining modeFRONTIER's smart algorithms with InfoWorks CS maximising customer choice between performance and cost' http://bit.ly/InnSC SWEDEN 2011 Training Courses on modeFRONTIER – Drive your designs from good to GREAT EnginSoft Nordic office in Lund, Sweden The Training Courses are focused on optimization, both multi- and single-objective, process automation and interpretation of results. Participants will learn different optimization strategies in order to complete a project within a specified time and simulation budget.

Other topics, such as design of experiments, meta modeling and robust design are introduced as well. The two day training consists of a mix of theoretical sessions and workshops.

The following dates are scheduled for 2012.

All courses are held at the EnginSoft Nordic office in Lund, Sweden.

• 1st-2nd December • 25-26th January • 8th-9th February • 6th-7th March • 2nd-3rd April • 3rd-4th May • 5th-6th June • 4th-5th September • 3rd-4th October • 6th-7th November • 6th-7th December To discuss your needs, for more information and to register, please contact EnginSoft Nordic, info@enginsoft.se

51

SPAIN EnginSoft Iberia. Programa de cursos de modeFRONTIER and other local events. To enquire about the next events in Spain and for more information, please contact: tel: +34 938.945.092. email: info@enginsoft.com Stay tuned to: http://iberia.enginsoft.com/empresa El 14 de diciembre de 2011 a las 09:30 Webcast: Add-On para LabVIEW de modeFRONTIER para la Optimización de parámetros y Prototipado Rápido de Control El 20 de diciembre de 2011 a las 10:00 (45 minutos) Webcast: Metodologías que aumentan su valor añadido a sus clientes For more information on the 2 Webcasts, please visit: http://www.aperiotec.es/agenda.php USA TMS 2012 Annual Meeting & Exhibition; 11-15 March; Orlando. www.tms.org/meetings/annual-12/AM12home.aspx Courses and Webinars on Design Optimization with modeFRONTIER Sunnyvale, CA. For more information, please contact: info@ozeninc.com www.ozeninc.com ISRAEL AUVSI; 20-22 Marzo; Tel Aviv event.pwizard.com/auvsi2012/index.py?p=376 EUROPE, VARIOUS LOCATIONS modeFRONTIER Academic Training Please note: These Courses are for Academic users only. The Courses provide Academic Specialists with the fastest route to being fully proficient and productive in the use of modeFRONTIER for their research activities. The courses combine modeFRONTIER Fundamentals and Advanced Optimization Techniques.

For more information please contact: modeFRONTIER University Program, info@enginsoft.com To meet with EnginSoft at any of the above events, please contact us: info@enginsoft.com

Corsi di addestramento software 2012 L'attività di formazione rappresenta da sempre uno dei tre maggiori obiettivi di EnginSoft accanto alla distribuzione ed assistenza del software ed ai servizi di consulenza e progettazione.

Per ciascuno dei possibili livelli cui la richiesta di formazione può porsi (quella del progettista, dello specialista o del responsabile di progettazione), EnginSoft mette a disposizione la propria esperienza per accelerare i tempi del completo apprendimento degli strumenti necessari con una gamma completa di corsi differenziati sia per livello (di base o specialistico), che per profilo professionale dei destinatari (progettisti, neofiti od analisti esperti).

La finalità è sempre di tipo pratico: condurre rapidamente all'utilizzo corretto del codice, sviluppando nell'utente la capacità di gestire analisi complesse attraverso l'uso consapevole del codice di calcolo. Per questo motivo ogni corso è diviso in sessioni dedicate alla presentazione degli argomenti teorici alternate a sessioni 'hands on', in cui i partecipanti sono invitati ad utilizzare attivamente il codice di calcolo eseguendo applicazioni guidate od abbozzando, con i suggerimenti del trainer, soluzioni per i problemi di proprio interesse e discutendone impostazioni e risultati.

Anche per il 2012 EnginSoft propone una serie completa di corsi che coprono le necessità di formazione all'uso dei diversi software sostenuti. Le novità proposte, confermano l’idea che EnginSoft ha della formazione: non è una realtà statica che si ripropone uguale a se stessa di anno in anno, ma è un divenire, guidato dall'esperienza accumulata negli anni, dall'evoluzione del software e dalle esigenze delle società che si affidano a noi per la formazione del proprio personale. In tale contesto EnginSoft organizza e sviluppa anche attività didattiche attraverso un programma formativo personalizzato, soluzioni di progettati in relazione alle necessità e alle specifiche esigenze aziendali del committente.

L’offerta dei corsi ANSYS viene ridefinita ogni anno per adeguarsi, sia all’evoluzione del software ed alle caratteristiche dell’ultima versione disponibile, che all’introduzione di nuovi moduli e solutori. In tale senso si segnala in campo fluidodinamico l'introduzione, accanto ai corsi tradizionalmente erogati, del corso ANSYS FLUENT: Corso Avanzato sulla Combustione.

Sono stati inoltre rivisti ed aggiornati i corsi relativi a tutti gli altri software sostenuti da EnginSoft per adeguarli allo stato attuale delle relative distribuzioni.

Si segnala infine l'introduzione del nuovo corso DIGIMAT, modellatore avanzato, non lineare, multi-scala di materiali che si pone come obiettivo quello di offrire una rappresentazione completa e rigorosa utile sia ai fornitori di materiali (“progettisti” di materiali), sia ai progettisti analisti CAE (end users) per i quali, il più delle volte, il materiale viene modellato in modo semplificato.

Dal punto di vista organizzativo nel 2012 tutte le sei sedi EnginSoft saranno impegnate nella formazione, dando la possibilità agli utenti di scegliere la location a loro più conveniente in termini di vicinanza geografica alla propria società.

Tutto questo a riprova dell'impegno nella formazione che, per EnginSoft, è e rimane un punto fondamentale della politica aziendale, un impegno costante verso l'eccellenza, un servizio per fare crescere i suoi clienti e, se lo desiderano, crescere con loro.

Per maggiori informazioni: www.enginsoft.it/corsi Per richiedere una copia del libretto: corsi@enginsoft.it