Syfte: Att utreda hur digitala 3D-modeller bör anpassas inför utskrift i en 3D-skrivare samt undersöka hur en sådan modell kan användas i byggprocessens olika skeden. Metod: En fallstudie genomförs med en befintlig digital 3D-modell som utgångspunkt där intervjuer och action research används som datainsamlingsmetoder. Empirin jämförs och analyseras med det teoretiska ramverket som tagits fram genom litteraturstudier. Resultat: En fysisk 3D-modell skulle kunna användas i flera skeden i byggprocessen, huvudsakligen i idéskedet, produktionsskedet och genomgående processen som ett kommunikationsverktyg och vid reklam/försäljning/presentation för ökad förstående. Inför utskrift bör alla byggnadsdelar vara solida, detaljer bör raderas beroende på skala och komponenter bör bestå av samma material. Konsekvenser: Då intervjuerna utförs med personer med varierande kunskap och erfarenhet är det viktigt att beakta det faktum att förslag på användningsområden eventuellt inte är genomförbara i praktiken då dessa är önskemål. Trots detta kan användning av fysiska 3D-modeller rekommenderas i flera av byggprocessens skeden för ökad förståelse och bättre kommunikation, vilket även styrks av det teoretiska ramverket. Gällande anpassningar av en digital modell krävs en digital 3D-model som utgångspunkt och viss vana av 3D-projektering. Begränsningar: Då denna studie är en fallstudie utförd på ett specifikt fall, kan kunskap och rekommendationer inte generaliseras statistiskt på andra typer av byggnader. Dock kan resultatet i denna studie implementeras på liknande projekt om små justeringar tillämpas. På grund av det faktum att studien är kvalitativ med ett begränsat antal respondenter finns möjlighet till ett annat resultat om utförandet skett med andra förutsättningar. Nyckelord: BIM-modell, fysisk byggnadsmodell, 3D-modell, 3D-skrivare, 3D- utskrift / Purpose: To investigate how digital 3D models should be adapted to enable 3D printing for use in the construction process in its various stages. Method: A case study is conducted with an existing digital 3D-model where interviews and action research is used as a data collection method. The empirical data are compared and analyzed with the theoretical framework developed through literature studies. Findings: A physical 3D model can be used at several stages in the construction process, mainly in idea development stages, the production stage and throughout the process as a communication tool and for advertising/sales/presentation for increased understanding. Prior to printing, all parts of the building should be solid, details should be erased depending on the scale used and components should consist of the same material. Implications: Based on interviews with people with varying knowledge and experience within the subject, it is important to take into consideration the fact that the proposals on the fields of use may not be enforceable in reality since they are requests. Despite this, the use of physical 3D models can be recommended in several construction phases of the process for greater understanding and better communication, which is corroborated by the theoretical framework. Adaptions of a digital model require a digital 3D model as a prerequisite and a certain experience of 3D design. Limitations: Since this study is a case study conducted in a specific case, knowledge and recommendations cannot be generalized statistically to other types of buildings. However, with small adjustments, this study can be implemented in similar projects. Due to the fact that the study is qualitative with a limited number of interviewees, there is a possibility of a different result if the execution occurred with other conditions. Keywords: BIM model, physical building model, 3D model, 3D printer, 3D printing

Durch die Verwendung von 3D-Druck-Verfahren wird die Gestaltung der Komponenten revolutioniert, weil die Form nicht mehr abhängig vom Fertigungsverfahren ist. Dabei werden auch optimale Gitterstrukturen innerhalb der Komponenten immer wichtiger. Diese Stützstrukturen können in Creo Parametric 4.0 mit dem neuen «Lattice-Feature» modelliert und Creo Simulate analysiert werden. Parallel dazu kann man mit ProTopCI (Hersteller CAESS, PTC Partner Advantage, Silver) eine Topologieoptimierung mit Stützstrukturen durchführen. Der Vortrag beleuchtet die Unterschiede dieser 2 Methoden.

[pt] O objetivo deste trabalho foi estudar a influencia das condições da impressão 3D nas propriedades finais do polietileno de alta densidade usando a modelagem por fusão e deposição, FDM. Foram impressos protótipos com formato de corpos de prova para teste de tração tipo V segundo norma ASTM D638, a três temperaturas de processamento: 220, 240 e 260 Graus Celsius. Para a impressão das amostras foram mantidos constantes os parâmetros de controle, entre eles a espessura da camada de impressão. As amostras impressas foram caracterizadas por difração de raios X, espectroscopia infravermelha, calorimetria diferencial de varredura, análise termogravimétrica, ensaio de tração, índice de fluidez e teste de contração. Os resultados das caracterizações das amostras impressas foram comparados com os resultados do material sem processar, cujas propriedades foram obtidas usando os mesmo métodos de caracterização. Estes resultados demostraram que as condições de impressão por FDM empregadas neste trabalho causaram apenas uma leve mudança nas características estruturais das amostras processadas do PEAD em relação ao material original sem processamento. Houve um leve aumento da cristalinidade no PEAD impresso (em torno de 1,3 a 3 porcento). Além disso, foi comprovado que por causa do resfriamento desigual na superfície e no interior da amostra impressa, o grau de cristalinidade foi levemente maior no interior que na superfície do corpo de prova impresso. A leve mudança no grau de cristalinidade não foi suficiente para causar mudança no módulo de elasticidade e no limite de escoamento em relação ao PEAD original. Outros resultados demostraram que não houve mudança significativa envolvendo formação de ligações duplas, quebra de cadeias e degradação térmica por efeito da condição do processamento utilizada durante a impressão. / [en] The aim of this work was to study the influence of process conditions for 3D printing on the final properties of prototypes of high density polyethylene (HDPE) using the method of the fused deposition modeling. Prototypes for type-V tensile testing according to ASTM D 638 were printed; They were made to three processing temperatures: 220, 240 and 260 Celsius degree. Control parameters for printing were kept constant in all the samples. The printed samples were characterized by X – ray diffraction, infrared spectroscopy, thermogravimetric analysis, differential scanning calorimetry, tensile test, melt flow index test, and, shrinkage test. The results of the characterization of the printed samples and of the original material were compared. These results demonstrated that the printing conditions employed in this study caused a slight change in the structural characteristics of the printed samples compared to the unprocessed original material, there being a slight increase in crystallinity (about 1,3 to 3 percent) for HDPE which was printed. In addition, it has been proven that the degree of crystallinity was slightly greater on the inside that on the surface of the printed samples, because of uneven cooling on the surface and inside of these samples. The slight change in the degree of crystallinity was not enough to cause change in the elastic modulus and yield strength compared to the original HDPE. Other results showed that there was not significant change involving bond formation, break chains, and, thermal degradation by the effect of the processing conditions used during printing.

L’accroissement de l’espérance de vie s’accompagne d’une détérioration de l’état de santé général des seniors et d’une recrudescence des maladies chroniques. Parmi les manifestations de la sénescence, l’atteinte de l’appareil locomoteur est particulièrement invalidante et accélère considérablement l’entrée en dépendance. C’est également le cas chez les plus jeunes lors d’atteintes traumatiques ou pathologiques. Ainsi, au niveau mondial 2,2 millions de greffes osseuses sont pratiquées chaque année, mais le taux de complications post-opératoires demeure élevé et est estimé à 15 % des interventions. Ces faits dessinent les contours d’un enjeu sociétal majeur ; les matériaux d’origine animale posent des problèmes d’histocompatibilité, de pathogénicité et donc de rejet. C’est pourquoi les efforts de recherche ciblent prioritairement le développement de biomatériaux synthétiques aptes à promouvoir la régénération osseuse. Actuellement les principaux substituts osseux sur le marché sont les « céramiques » bioactives (phosphates de calcium, verres bioactifs) qui présentent comme caractéristiques d’être biocompatibles, de se lier spontanément aux tissus osseux, de promouvoir l’adhésion des cellules osseuses et enfin d’être biorésorbables. Cependant, malgré cet ensemble de caractéristiques très satisfaisantes, la fragilité de ces matériaux en limite les applications. Pour pallier ce défaut, une solution ingénieuse est de s’inspirer de la structure particulière du tissu osseux. Celle-ci mêle intimement une phase inorganique, le minéral osseux constitué de cristaux d’apatite (phosphate de calcium résorbable), à une phase organique qui est majoritairement du collagène. De manière remarquable, une telle structure associe la rigidité de la partie inorganique à la ténacité des fibres de collagène. Pour obtenir des implants aux propriétés mécaniques proches du tissu osseux, la stratégie consiste donc à combiner céramiques bioactives et matière organique. À cette fin, l’équipe Biomatériaux du Laboratoire de Physique de Clermont (LPC) a récemment mis au point un procédé innovant qui permet la synthèse de matrices tridimensionnelles d’hybrides organique-inorganique à base de verre bioactif et de polymère biocompatible aux caractéristiques variées. Dans la continuité des travaux, il était alors question d’exploiter ce procédé afin de développer un substitut osseux hybride aux propriétés optimisées. Il s’agissait tout d’abord de sélectionner le polymère le plus adéquat pour la régénération osseuse, qui s’est avéré être le polycaprolactone, puis d’optimiser la synthèse (notamment la source de calcium), la structure macroporeuse et la proportion organique-inorganique. Le matériau hybride résultant a ensuite été dopé en éléments thérapeutiques à faible dose (< 5 % de la masse totale) avec des ions strontium ou des nutriments tels que la fisétine et l’hydroxytyrosol qui possèdent un effet ostéogénique. Les mousses hybrides ainsi développées ont finalement été caractérisées in vitro afin de déterminer leurs propriétés physico-chimiques et biologiques, et in vivo afin d’évaluer leur performance. Après 3 mois d’implantation dans un défaut critique de la calvaria de souris, les résultats démontrent le potentiel de ce substitut osseux: comparé au matériau commercial de référence (os bovin traité) qui conduit à une reconstruction osseuse de 16% (± 5%), l’hybride permet une reconstruction allant de 32% (± 3%) lorsqu’il n’est pas dopé, jusqu’à 55% (± 7%) voire 58% (± 7%) lorsqu’il est dopé respectivement en fisétine ou en strontium. Ces travaux de thèse laissent entrevoir des perspectives prometteuses telles que l’association des dopants et l’impression 3D des mousses hybrides polycaprolactone-verre bioactif. / The increase in life expectancy results in the decline of seniors’ health conditions and the resurgence of chronic diseases. Among the expressions of senescence, disorders of the musculoskeletal system are particularly disabling and considerably accelerate the state of dependency. This is also the case for young people who suffer from traumatic injuries or pathologic conditions. Thus, about 2.2 million bone grafts are performed worldwide every year. Yet, the level of postoperative complications remains high and is estimated at 15% of surgical operations. These facts outline a major societal concern: animal-based materials present a risk of histocompatibility issues and pathogenicity that may lead to implant failure. This is the reason why research efforts focus on the development of synthetic biomaterials capable of promoting bone regeneration. Currently, commercialised bone substitutes are mainly made of bioactive “ceramics” (calcium phosphates, bioactive glass) that are known to be biocompatible, to spontaneously bond to bone tissues, to promote bone cell adhesion and finally to be bioresorbable. However, despite these remarkable properties, the brittleness of these materials limits their applications. An ingenious solution to this brittleness can be learned from the particular structure of bone tissue. Bone tissue intimately blends an inorganic phase, the bone mineral, which is made of apatite crystals (resorbable calcium phosphates), with an organic phase that is mainly collagen. Such a structure associates the stiffness of the inorganic part with the toughness of collagen fibres. Therefore, in order to obtain implants with mechanical properties similar to that of bone, the strategy consists in combining bioactive ceramics with organic matter. To this end, the Biomatériaux team from the Laboratoire de Physique de Clermont (LPC) recently developed an innovative process that allows the synthesis of tridimensional organic-inorganic hybrids comprised of bioactive glass and biocompatible polymer. The objective of the thesis was to exploit this process in order to develop a hybrid bone substitute with optimal properties. First of all, polycaprolactone was selected as the polymer, especially because of its adequate degradation rate for long-term applications such as bone regeneration. Then, the synthesis process was improved (in particular, the calcium source was changed), the macroporous structure was optimised and the organic-inorganic ratio was chosen. Afterwards, elements that are known to induce an osteogenic effect were incorporated in the hybrid at low doses (< 5% of total weight): an inorganic doping was performed using strontium ions and an organic doping was performed using nutrients such as fisetin or hydroxytyrosol. The resultant hybrid scaffolds were eventually characterised in vitro in order to determine their physicochemical and biological properties and in vivo in order to evaluate their performance. After 3 months of implantation in a mouse calvarial critical defect, results demonstrate the potential of this bone substitute: compared to the reference commercial material (treated bovine bone) that leads to a bone reconstruction of 16% (± 5%), the hybrid allows a reconstruction going from 32% (± 3%) when it is not doped, to 55% (± 7%) and even 58% (± 7%) when it is doped respectively with fisetin or strontium. This thesis paves the way to promising perspectives like the association of doping agents and the 3D printing of polycaprolactone-bioactive glass hybrid scaffolds.

Estudos recentes em medicina regenerativa utilizam estruturas de crescimento celular conhecidas como scaffolds: um scaffolds é uma estrutura porosa feita de material biodegradável. Essas estruturas ajudam na formação e reconstituição de novos tecidos, servindo como suporte para o crescimento e proliferação celular. Esses podem ser fabricados utilizando processos de manufatura aditiva (MA). O processo utilizado na pesquisa foi o FFF (Fused Filament Fabrication), que se baseia em fundir polímero e extrudá-lo em forma de filamento para produção de peças tridimensionais para confecção dos scaffolds. O polímero utilizado na pesquisa foi o ácido poliláctico (PLA), por ser biocompatível. Uma possível forma de aumentar a rugosidade de superfície é gerar uma geometria diferente na seção transversal do filamento extrudado que forma o scaffold. Assim, o scaffold adquiriria a micro-rugosidade ou nanorrugosidade (rugosidade superficial do material utilizado e inerente do processo), e a macrorrugosidade (geometria gerada ao longo de seu comprimento): a micro-rugosidade, para fins deste trabalho, será considerada como a rugosidade (acabamento de uma superfície); a macrorrugosidade (ou macrogeometria) será considerada, para fins deste trabalho, como a textura que se apresenta na forma de uma seção transversal de um filamento extrudado. Para gerar essa macrotextura, é necessário gerar na ferramenta bico extrusor uma geometria diferente na saída do bico. Geralmente, na saída do bico extrusor é feito um furo de geometria circular. Foi proposto neste estudo utilizar uma geometria diferente de um círculo na saída do bico, geometria essa a ser transferida ao filamento durante o momento da extrusão. A alteração desta geometria requer a utilização de técnicas não convencionais de usinagem. Neste projeto foram produzidos filamentos modificados a fim de melhorar a citocompatibilidade no scaffold. A morfologia foi analisada in vitro e o novo bico gerou filamentos impressos com diferenças na citocompatibilidade e com mudanças nos aspectos morfológicos das células quando comparadas àquelas aderidas aos filamentos convencionais. / Recent studies in regenerative medicine use cell growth structures known as scaffolds: a scaffold is a porous structure made of biodegradable material. These structures aid in the formation and reconstitution of new tissues, serving as a support for cell growth and proliferation. They are manufactured using additive manufacturing (AM) processes. The technology executed in this research was the FFF (Fused Filament Fabrication), which is based on melting polymer in the form of a filament and extruding it for the production of three-dimensional parts for the manufacture of scaffolds. The polymer used in this research was the polylactic acid (PLA), which was used because it is biocompatible. One possible way to increase surface roughness is to generate a geometry in the cross section of the extruded filament forming the scaffold. Thus, the scaffold would acquire the micro-roughness or nano roughness (surface roughness of the material used and inherent in the process), and the macro-roughness (texture generated along its length): the micro-roughness, for purposes of this work, will be considered a roughness (surface finish); The macro-roughness (or macro-geometry) will be considered, for purposes of this work, as the texture which is in the form of a cross-section of an extruded filament. To generate this texture (macro-roughness) it is necessary to generate a macro-geometry in the nozzle extruder tool. Usually at the exit of the extruder nozzle is made a hole of circular geometry. It was proposed in this study to use a geometry different from a circle at the exit of the nozzle to be transferred to the filament during the moment of the extrusion. Altering this geometry requires the use of non-conventional machining techniques. In this project modified filaments were produced in order to improve cytocompatibility and for cell differentiation in the scaffold. The morphology was tested in vitro and the new nozzle was able to generate printed filaments with differences in cytocompatibility and with changes in the morphological aspects of the cells when compared to those adhered to the conventional filaments.

3D-skrivare är inte längre något som endast stora företag har råd att använda sig av, utan någonting som har tagit sig ända till konsumentmarknaden. Detta har givit både hobbyanvändare och småföretag tillgång till ett kraftfullt verktyg för iterativ design. Det görs idag även stora framsteg inom robotik som ger upphov till möjligheter för tillämpningar av både industri- och konsumentrobotar i framtiden. Komplexiteten i robotar gör 3D-skrivare till användbara verktyg när det kommer till att ta fram robotprototyper. Målet med detta projekt var därför att analysera fördelar och nackdelar med 3D-skrivning i samband med prototypframtagning av en robotarm med sex axlar som kunde utföra enkla pick-and-place-rörelser. Roboten designades i Fusion360 och skrevs ut med 3D-skrivare i PETG. En Arduino MEGA användes till att styra stegmotorer i robotens axlar och styrningen implementerade en analytisk lösning till robotens kinematisk modell. Resultatet var en fungerande prototyp som kunde programmeras till att flytta och rotera föremål. Ett antal förbättringar av prototypen går att göra, till exempel går det att öka robotens rörlighet genom att designa om delar för att tillåta att axlar rotera längre och att implementera ytterligare funktionalitet i programmet som accelerationsrampning av stegmotorer och rörelse längs förvald bana. Utifrån analys av 3D-skrivningens inverkan på projektet framgick det att 3D-skrivning verkar användbart i prototypframtagning om dess möjlighet till iterativ design är önskvärd, men att det passar ännu bättre till mindre projekt då robotens storlek hindrade möjligheten till att utnyttja iterativ design till fullo och på så sätt hindrades även effektiviteten i 3D-skrivningsprocessen.

Próteses auriculares tem o objetivo de proporcionar uma maior confiança a pacientes com perdas parciais ou totais da orelha. A aparência realista de uma prótese é fator fundamental para a recuperação psicossocial dos pacientes, pois auxilia na dissimulação da perda. O procedimento convencional de elaboração de próteses auriculares envolve o trabalho de um especialista que utiliza processos manuais para recriar a forma singular de uma orelha. Este é um processo lento, complexo e naturalmente pode ocasionar imperfeições. Este estudo apresenta um procedimento que, a partir de exames de imagem, utiliza técnicas computacionais de reconstrução, espelhamento e inversão de modelos tridimensionais para manufatura aditiva (impressão 3D) de moldes. Os resultados mostram que a prótese elaborada utilizando o processo exibe alto grau de realismo, apresentando 97,8% de similaridade dimensional comparada com a orelha sã do paciente. / Auricular prostheses are intended to provide confidence to patients with partial or total loss of the ear. The realistic appearance of a prosthesis is a key factor for the psychosocial recovery helping in dissimulation of loss. The standard procedure of auricular prosthesis design involves the work of an expert using manual processes to recreate the complex shape of the ear. This slow and complex process can easily lead to imperfections. This study presents a procedure that starting from an image exam uses computational techniques of reconstruction, mirroring and inversion of three-dimensional models for additive manufacturing (3D printing) of molds. The results show that the designed prosthesis using the process has a high level of realism, with 97.8% of similarity compared to the healthy ear of the patient.

<div>This dissertation addresses two significant issues in the effective application of machine learning algorithms and models for the physical and engineering sciences. The first is the broad challenge of automated modeling of data across different processes in a physical system. The second is the dilemma of obtaining insightful interpretations on the relationships between the inputs and outcome of a system as inferred from complex, black box machine learning models.</div><div><br></div><div><b>Automated Geometric Shape Deviation Modeling for Additive Manufacturing Systems</b></div><div><b><br></b></div><div>Additive manufacturing systems possess an intrinsic capability for one-of-a-kind manufacturing of a vast variety of shapes across a wide spectrum of processes. One major issue in AM systems is geometric accuracy control for the inevitable shape deviations that arise in AM processes. Current effective approaches for shape deviation control in AM involve the specification of statistical or machine learning deviation models for additively manufactured products. However, this task is challenging due to the constraints on the number of test shapes that can be manufactured in practice, and limitations on user efforts that can be devoted for learning deviation models across different shape classes and processes in an AM system. We develop an automated, Bayesian neural network methodology for comprehensive shape deviation modeling in an AM system. A fundamental innovation in this machine learning method is our new and connectable neural network structures that facilitate the transfer of prior knowledge and models on deviations across different shape classes and AM processes. Several case studies on in-plane and out-of-plane deviations, regular and free-form shapes, and different settings of lurking variables serve to validate the power and broad scope of our methodology, and its potential to advance high-quality manufacturing in an AM system.</div><div><br></div><div><b>Interpretable Machine Learning</b></div><div><b><br></b></div><div>Machine learning algorithms and models constitute the dominant set of predictive methods for a wide range of complex, real-world processes. However, interpreting what such methods effectively infer from data is difficult in general. This is because their typical black box natures possess a limited ability to directly yield insights on the underlying relationships between inputs and the outcome for a process. We develop methodologies based on new predictive comparison estimands that effectively enable one to ``mine’’ machine learning models, in the sense of (a) interpreting their inferred associations between inputs and/or functional forms of inputs with the outcome, (b) identifying the inputs that they effectively consider relevant, and (c) interpreting the inferred conditional and two-way associations of the inputs with the outcome. We establish Fisher consistent estimators, and their corresponding standard errors, for our new estimands under a condition on the inputs' distributions. The significance of our predictive comparison methodology is demonstrated with a wide range of simulation and case studies that involve Bayesian additive regression trees, neural networks, and support vector machines. Our extended study of interpretable machine learning for AM systems demonstrates how our method can contribute to smarter advanced manufacturing systems, especially as current machine learning methods for AM are lacking in their ability to yield meaningful engineering knowledge on AM processes. <br></div>

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2018-04-11T16:07:01Z No. of bitstreams: 1 Pedro Nascimento de Lima_.pdf: 14461751 bytes, checksum: 473cf6676a373e7639a581816a08b7a5 (MD5) / Made available in DSpace on 2018-04-11T16:07:01Z (GMT). No. of bitstreams: 1 Pedro Nascimento de Lima_.pdf: 14461751 bytes, checksum: 473cf6676a373e7639a581816a08b7a5 (MD5) Previous issue date: 2018-01-26 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / A avaliação de decisões estratégicas em condições de profunda incerteza é um desafio significativo para as organizações. Em condições onde informação disponível permite que stakeholders cheguem a um consenso sobre o futuro que será mais provável, ferramentas de planejamento baseadas em predição podem suportar o processo decisório de modo confiável. No entanto, em situações de instabilidade, onde o futuro é altamente incerto, a avaliação de decisões estratégicas utilizando predições pode levar a decisões equivocadas. Tais condições de incerteza frequentemente ocorrem em mercados nascentes, onde há alta incerteza relacionada ao processo de difusão de um novo produto. Na Indústria da Manufatura Aditiva, enquanto alguns especialistas estimam que a indústria pode chegar a faturar 21 bilhões de dólares em 2020, outros estimam que este mercado pode valer até 550 bilhões até 2025. Esta pesquisa emprega a simulação computacional de dinâmica de sistemas utilizando o método Robust Decision Making (RDM) para avaliar decisões estratégicas de fabricantes de sistemas de impressão 3D profissional. Para tanto, este trabalho amplia modelos de dinâmica competitiva e difusão de novos produtos para permitir a simulação no contexto da manufatura aditiva. Em seguida, são desenvolvidos algoritmos necessários para a análise RDM. Para avaliar decisões estratégicas em um amplo conjunto de futuros plausíveis, 10.800 simulações são realizadas. Em seguida, a robustez das estratégias avaliadas é testada, e as vulnerabilidades da estratégia mais robusta localizada são examinadas utilizando técnicas estatísticas. Finalmente, o trabalho identifica estratégias alternativas à estratégia mais robusta. Os resultados da simulação sugerem que fabricantes de sistemas de impressão 3D profissional deveriam perseguir uma estratégia de dominação do mercado agressiva, com um modelo de Pesquisa e Desenvolvimento e proteção intelectual fechado. Finalmente, o trabalho discute implicações gerenciais e teóricas relacionadas à avaliação de decisões estratégicas em condições de incerteza profunda. / Strategic Decision Making under deep uncertainty is a relevant challenge to organizations. When the available information allows sound decision making based on predictions, traditional decision making tools based on maximum expected value can lead to the right decision. Under conditions of deep uncertainty, however, decision making based on predict-then-act approaches might mislead and build overconfidence. In the 3D printing industry, uncertainty is highly relevant. While some experts forecast that this industry will worth 21 billion dollars by 2020, other estimates point that this market can have an economic impact of 550 billion by 2025. This dissertation leverages system dynamics simulation, using the Robust Decision Making (RDM) approach as the analytical framework to evaluate 3D printing Systems Manufacturers’ strategic decisions. I extend an existing competitive dynamics model allowing it to take into account expiring patents dynamics, an important aspect of the 3D printing industry. Then, I test 54 different strategies under 200 different scenarios, highlighting the most robust strategies. Afterwards, I examine the vulnerabilities of a candidate strategy using machine learning algorithms. The experiments showed that aggressive strategies dominate their conservative counterparts, using robustness as a criteria. Also, the results do not lend support to open source Research and Development strategies. Finally, I discuss managerial implications to the 3D printing industry, and theoretical contributions to the Strategic Decision-Making literature.

O bambu é considerado um dos materiais naturais com a melhor relação resistência por peso. Entre as características responsáveis encontram-se os feixes fibrosos do esclerênquima, que protegem os elementos de condução de água e solutos na planta, além de serem preenchidos pelo parênquima. Os feixes estão distribuídos gradual-mente, da parte interna à externa do bambu, aumentando a rigidez do colmo. Contudo, a morfologia da seção dos feixes fibrosos e a presença da matriz parenquimática no caule ainda não foram totalmente estudados, quanto à sua importância estrutural, nem aplicados em estruturas biônicas. Esse estudo trata do design de estruturas biônicas baseada na caracterização do bambu, por meio de técnicas não-invasivas, de proto-tipagem e de análise por elementos finitos. Um protocolo de amolecimento e secagem para secionamento foi seguido em uma amostra de bambu (Bambusa tuldoides Munro). A superfície da amostra foi analisada por microtomografia computadorizada de raios X de alta resolução. As imagens resultantes permitiram a segmentação dos tecidos constituintes e caracterização em nível celular. Os modelos 3D do parênquima e do esclerênquima foram discretizados para uma análise por elementos finitos não-linear. Os resultados mostraram que o parênquima é configurado como uma matriz celular de baixa densidade e que distribui as tensões em todos os elementos de re-forço na planta, sendo considerado um tecido com importante função estrutural. De mesmo modo, a geometria da seção dos feixes fibrosos apresentou um desempenho mecânico superior às seções referenciais. Ambas características foram aplicadas no desenvolvimento de duas estruturas biônicas, sendo analisadas como mais eficientes em comparação a um modelo da literatura, e impressas em 3D para análise visual. / Bamboo is considered one of the natural materials with the best strength-to-weight ratio. Among the features responsible for its properties are the sclerenchyma’s fiber bundles that protect the conducting elements of water and solutes in the plant, and are filled by parenchyma. Bamboo fibers are gradually distributed from the inner to the outer side of the plant, thus increasing the culm stiffness. However, the morphology of the section of the fiber bundles and the presence of the parenchymatic matrix in the stem have not been fully studied regarding its structural importance, neither applied in bionic structures. This study addresses the design of bionic structures based on the characterization of bamboo, by means of non-invasive techniques, rapid prototyping and finite element analysis. A bamboo sample (Bambusa tuldoides Munro) was col-lected and softening and drying protocols were followed for proper sectioning. The surface of the sample was analyzed by high-resolution X-ray microcomputed tomog-raphy. Resulting images allowed the segmentation of the sample’s constituent tissues and the characterization at a cellular level. The 3D models of the parenchyma and sclerenchyma were discretized for a non-linear finite element analysis. The results showed that the parenchyma is set as a low-density cellular matrix by distributing the stresses among all reinforcement elements in the plant, being considered a tissue with great structural importance. Likewise, the shape of the fiber bundles’ section showed superior mechanical performance compared to reference sections. Both characteris-tics were applied in the development of two bionic structures that were analyzed as more efficient than a literature model, and 3D printed for a visual analysis.