Made available in DSpace on 2014-06-11T19:28:21Z (GMT). No. of bitstreams: 0 Previous issue date: 2012-07-06Bitstream added on 2014-06-13T19:36:43Z : No. of bitstreams: 1 moia_dfg_me_bauru.pdf: 2028523 bytes, checksum: a4957573090c9f2211022b6a4fceb9b6 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / O processo de retificação confere à peça o acabamento final, minimizando as irregularidades superficiais através de interações entre os grãos abrasivos de uma ferramenta (rebolo) e peça retificada. O desgaste do rebolo devido ao atrito entre o rebolo e a peça retificada torna a ferramenta inadequada para nova utilização, sendo necessária a realização do processo de dressagem do rebolo para remoção e ou avivamento dos grãos gastos de sua superfície de corte, de forma e deixá-lo em condições para novo uso. O presente trabalho tem como objetivo classificar a condição do rebolo durante a operação de dressagem utilizando o sinal de emissão acústica (EA) e estatísticas derivadas desse sinal, por meio de redes neurais artificiais. Nos experimentos realizados usou-se um rebolo de óxido de alumínio instalado em uma retificadora plana, um sistema de aquisição de sinais e um dressador de ponta única de diamante. O processamento digital de sinais foi obtido através do software MATLAB. Os ensaios foram realizados com diferentes graus de recobrimento e profundidade de dressagem. A partir dos dados obtidos de EA puro, calculou-se o valor médio quadrático (RMS), bem como mais duas estatísticas, as quais já foram empregadas com sucesso em trabalhos de detecção de queima, no processo de retificação. Essas estatísticas também se mostraram bons indicadores para o monitoramento da operação de dressagem. Uma rede neural perceptron multicamadas (MLP) foi utilizada com o algoritmo de aprendizado Levenberg-Marquardt, cujas entradas foram as duas estatísticas mencionadas e o valor RMS de EA. Os resultados mostram que o método empregado foi capaz de classificar as condições do rebolo no processo de dressagem, identificando o rebolo como afiado (com capacidade de corte) e rebolo se afiação (com perda de capacidade de corte), viabilizando a redução do tempo e custo dessa operação e minimizando a remoção excessiva / The grinding process gives the piece a final finish by minimizing surface irregularities through interactions between the abrasive grains of a tool (wheel) and the part to be ground. The wear of the grinding wheel due to excessive friction between the grinding wheel and ground workpiece makes the tool unsuitable for further use; it is imperative the accomplishment of the process of dressing the grinding wheel to remove or resharpen the worn grains of its surface in order to make if suitable for use again. The present study aims to classify the condition of the grinding wheel during operation using acoustic emission (AE) signal and statistics derived from this sinal through artificial neural networks. In the experiments an aluminum oxide grinding wheel installed to a surface grinding machine was used along with a data acquisition system and a single point diamond dresser. The digital processing of these data was obtained using the MATLAB software. Tests were performed with different overlap ratio and depth of cut. The root mean square value of the AE signal as well as two other statistics were obtained from the raw acoustic emission signal, which have been successfully used in grinding burn detection. These statistics were also good indicators for monitoring the dressing operation. A multilayer perceptron neural network (MLP) was used with the learning algorithm Levenberg-Marquardt, whose inputs were the statistics previously mentioned and dressing conditions. The results show that the method used was able to classify the conditions of the grinding wheel in the process of dressing, identifying the wheel as sharp (cutting capacity) and dull (with loss of cutting capacity), enabling the reduction of time and cost of operation and minimizing the excessive removal of the wheel abrasive material

Orientador: Paulo Roberto de Aguiar / Banca: Carlos Elias da Silva Junior / Banca:m Eduardo Carlos Bianchi / Resumo: O processo de retificação confere à peça o acabamento final, minimizando as irregularidades superficiais através de interações entre os grãos abrasivos de uma ferramenta (rebolo) e peça retificada. O desgaste do rebolo devido ao atrito entre o rebolo e a peça retificada torna a ferramenta inadequada para nova utilização, sendo necessária a realização do processo de dressagem do rebolo para remoção e ou avivamento dos grãos gastos de sua superfície de corte, de forma e deixá-lo em condições para novo uso. O presente trabalho tem como objetivo classificar a condição do rebolo durante a operação de dressagem utilizando o sinal de emissão acústica (EA) e estatísticas derivadas desse sinal, por meio de redes neurais artificiais. Nos experimentos realizados usou-se um rebolo de óxido de alumínio instalado em uma retificadora plana, um sistema de aquisição de sinais e um dressador de ponta única de diamante. O processamento digital de sinais foi obtido através do software MATLAB. Os ensaios foram realizados com diferentes graus de recobrimento e profundidade de dressagem. A partir dos dados obtidos de EA puro, calculou-se o valor médio quadrático (RMS), bem como mais duas estatísticas, as quais já foram empregadas com sucesso em trabalhos de detecção de queima, no processo de retificação. Essas estatísticas também se mostraram bons indicadores para o monitoramento da operação de dressagem. Uma rede neural perceptron multicamadas (MLP) foi utilizada com o algoritmo de aprendizado Levenberg-Marquardt, cujas entradas foram as duas estatísticas mencionadas e o valor RMS de EA. Os resultados mostram que o método empregado foi capaz de classificar as condições do rebolo no processo de dressagem, identificando o rebolo como "afiado" (com capacidade de corte) e rebolo "se afiação" (com perda de capacidade de corte), viabilizando a redução do tempo e custo dessa operação e minimizando a remoção excessiva / Abstract: The grinding process gives the piece a final finish by minimizing surface irregularities through interactions between the abrasive grains of a tool (wheel) and the part to be ground. The wear of the grinding wheel due to excessive friction between the grinding wheel and ground workpiece makes the tool unsuitable for further use; it is imperative the accomplishment of the process of dressing the grinding wheel to remove or resharpen the worn grains of its surface in order to make if suitable for use again. The present study aims to classify the condition of the grinding wheel during operation using acoustic emission (AE) signal and statistics derived from this sinal through artificial neural networks. In the experiments an aluminum oxide grinding wheel installed to a surface grinding machine was used along with a data acquisition system and a single point diamond dresser. The digital processing of these data was obtained using the MATLAB software. Tests were performed with different overlap ratio and depth of cut. The root mean square value of the AE signal as well as two other statistics were obtained from the raw acoustic emission signal, which have been successfully used in grinding burn detection. These statistics were also good indicators for monitoring the dressing operation. A multilayer perceptron neural network (MLP) was used with the learning algorithm Levenberg-Marquardt, whose inputs were the statistics previously mentioned and dressing conditions. The results show that the method used was able to classify the conditions of the grinding wheel in the process of dressing, identifying the wheel as sharp (cutting capacity) and dull (with loss of cutting capacity), enabling the reduction of time and cost of operation and minimizing the excessive removal of the wheel abrasive material / Mestre

Orientador: Carlos Kenich Suzuki / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecanica / Made available in DSpace on 2018-07-22T19:32:43Z (GMT). No. of bitstreams: 1 Ogata_DanielaYuri_M.pdf: 14373801 bytes, checksum: 37390613963c28ba54f71a665d3cefb7 (MD5) Previous issue date: 1997 / Resumo: Neste trabalho procurou-se investigar a influência do tratamento da superficie (tensão/rugosidade provocado pelo processo de desbaste com pó abrasivo) na devitrificação e na dissolução da sílica vítrea. Para tanto, utilizou-se de métodos de caracterização de materiais, tais como difratometria de raios-X, espectroscopia de infravermelho, microscopia óptica e eletrônica e espalhamento de raios-X a baixo ângulo. Para o estudo da influência do tratamento superficial na devitrificação, os resultados de difração de raios-X e microscopia, mostraram que a devitrificação possui forte correlação com o tratamento superficial da sílica vítrea, onde amostras com maior rugosidade, apresentaram também uma maior devitrificação em comparação à outra com menor rugosidade. A dissolução da sílica vítrea por ácido fluorídrico ("etching") também foi estudado, e os resultados mostraram que a taxa de dissolução é dependente da rugosidade superficial, concentração do ácido e tempo de "etching". Resultados adicionais mostraram a formação de cavidades arredondadas na forma de "larvas" (padrão superficial) após "etching". Após um tempo de "etching" entre 33 e 88 horas, as amostras apresentaramum padrão superficial semelhante a "grãos" e no interior dos "grãos", outros "sub-grãos" da ordem de 4-10 'mu¿m, que denotam a existência de microtensões na estrutura amorfa. Neste trabalho, este resultado inédito foi denominado "grão ou sub-grão amorfo", que podem estar relacionados com heterogeneidades em nível atômico da estrutura da sílica vítrea / Abstract: The effect of surface stress and roughness, produced by polishing process with abrasive powder, was investigated on the devitrification and dissolution of silica glass. The characterization studies of the silica glass treated by grinding, polishing and etching and their effect on the devitrification were conducted by X-ray diffraction, optical and electronic microscopy, inftared spectroscopy and small-angle X-ray scattering. The samples were heat treated at 1250°C for a period of 15 min and 150 mino X-ray diffraction and optical microscopy observation showed more intense devitrification effect into a-cristobalite for higher degree of surface roughness, and the samples treated with HF etching presented the lowest devitrification effect. The silica glass surfaces exposed to the furnace atmosphere (air) presented a much bigger crystallization effect, inversely to the bottom surface in contact with a silica base. The dissolution of silica glass by chemical etching (HF solutions) was studied, and the results showed that dissolution rate is dependent on the surface roughness, HF concentration and etching time. For example, the dissolution rate of samples treated with SiC #320 and 'Ce IND. 2¿¿O IND. 3¿, were estimated as 0.16 'mu¿m/min and 0.14 'mu¿m/min, respectively. The results showthat sampleswith longtime etching(~1000 min) revealed grain-like etch pattern of 4 to 10 11min diameter, which suggests reveals the existence of micro-stress in the amorphous structure. This "amorphous grains or sub-grains" might be correlated with atomic into inhomogeneities of amorphous silica structure / Mestrado / Materiais e Processos de Fabricação / Mestre em Engenharia Mecânica

Im Bauwesen kommen verschiedene Kantenausführungsarten von Glas zum Einsatz. In Abhängigkeit ihrer Art erfüllen sie Anforderungen an den Schnittschutz, die Maßhaltigkeit und die Ästhetik. Nach DIN 1249-11 erfolgt die Einteilung entsprechend des äußeren Erscheinungsbildes in geschnittene (KG), gesäumte (KGS), maßgeschliffene (KMG), geschliffene (KGN) und polierte (KPO) Kanten. Die mechanische Festigkeit der Glaskante ist jedoch gesondert zu betrachten. Die charakteristische Biegezugfestigkeit von Glas ist maßgeblich von dessen Oberflächenzustand abhängig. Die Herstellung und Bearbeitung der Kante erfolgt durch einen Materialeingriff mit harten Schneid- und Schleifwerkzeugen. Dabei werden die Oberflächenbeschaffenheit verändert und das Bruchverhalten beeinflusst. Bisher regelt die europäische Normung Kantenfestigkeiten in der Bemessung in Form von Beiwerten, welche die charakteristische Biegezugfestigkeit pauschal oder in Abhängigkeit der Kantenausführungsart abmindern. Bestehende Untersuchungen zeigen jedoch wesentliche Unterschiede der Kantenfestigkeit in Abhängigkeit von Kantenausführungsart und Herstellungsprozess. Die Bemessungswerte der Kantenfestigkeit gelten als untere Grenze der auf dem Markt verfügbaren Qualitäten. Wissenschaftlich belegte Beurteilungen der visuellen Kantenqualität mit Einschätzung ihrer mechanischen Festigkeit liegen bisher nur für die geschnittene Kantenausführung vor. Für den industriellen Schneidprozess wurden auf Basis systematischer Untersuchungen Parameter detektiert, die sich positiv auf die Kantenfestigkeit auswirken. Es ist unbekannt, wie sich der Oberflächeneingriff durch Schleif- und Polierprozesse auf die mechanische Festigkeit der dabei hergestellten Kantenausführungsarten auswirkt und welche Ursachen dafür zu benennen sind. Zudem fehlen geregelte, einheitliche Methoden, um die Kantenqualitäten optisch und mechanisch zu erfassen. Im Rahmen dieser Arbeit wird eine Versuchsmethodik entwickelt, welche über mikroskopische Analysen und Bruchversuche die Erfassung vergleichbarer optischer und mechanischer Kennwerte ermöglicht. In einer Zusammenführung der Methoden erfolgt erstmalig die Charakterisierung bruchverursachender Fehlstellen und deren Rückführung auf den Entstehungsort im Herstellungsprozess. Anhand einer systematischen Untersuchung verschiedener Kantenausführungsarten eines Herstellers findet die Erprobung und Bewertung der entwickelten Versuchsmethodik statt. Anschließend werden Prozessanpassungen zur Fehlstellenreduzierung vorgenommen und in Bezug auf eine Steigerung der Kantenfestigkeit untersucht. Weitere Analysen des Herstellungsprozesses eines zweiten Herstellers erfolgen für verschieden polierte Kanten. Entsprechend der ermittelten bruchverursachenden Fehlstellen werden Hypothesen für zukünftige Untersuchungen abgeleitet. Die Erkenntnisse dieser Arbeit stellen die wissenschaftliche Grundlage für verfahrenstechnische Handlungsempfehlungen zur Herstellung von Glaskanten dar. Der aktuelle Stand bedeutender europäischer Bemessungsregeln wird für abschließende normative Empfehlungen einbezogen. Die gesäumte Kante ergibt sich beim ersten Hersteller als Kantenausführungsart mit der höchsten Kantenfestigkeit. Für die maßgeschliffene, geschliffene und polierte Kante ergeben sich im Vergleich zur geschnittenen und gesäumten Kante geringere Festigkeiten. Die Untersuchung der polierten Kantenausführung des zweiten Herstellers ergibt, verglichen mit der polierten Kante des ersten Herstellers, eine höhere Kantenfestigkeit. Daraus leitet sich der Einfluss der Maschinenkonfiguration als einflussreicher Prozessparameter ab. Die Analyse der bruchverursachenden Fehlstellen zeigt, dass höherfeste Kanten mit einer Reduzierung von mikroskopisch erfassbaren Fehlstellen für die geschnittene, gesäumte und polierte Kante korrelieren. Darüber hinaus gewährleisten die in dieser Arbeit entwickelten Fehlstellenanalysen eine Detektion festigkeitsmindernder Fehlstellen, die auf eine Einbringung nach der Herstellung hindeuten und somit die Notwendigkeit von Kantenschutzmaßnahmen nach sich ziehen. Definitionen allgemeingültiger Prozessparameter in Schleif- und Polierprozessen, die eine positive Auswirkung auf die Kantenfestigkeit bearbeiteter Kanten haben, sind bisher in der Literatur nicht vorhanden. Die in dieser Arbeit beschriebenen Ergebnisse dienen als Grundlage zur Optimierung der Prozessparameter für hohe Kantenfestigkeiten. Experimentelle Nachweise der mechanischen Festigkeit sind dabei unabdingbar. Die Kantenqualität ist in optische und mechanische Eigenschaften zu unterscheiden. Die optisch als am hochwertigsten geltende polierte Kante geht nicht zwangsläufig mit einer hohen Kantenfestigkeit einher, was eine in der Praxis weitverbreitete Annahme widerlegt. In diesem Kontext leistet die vorliegende Arbeit einen Beitrag zum Verständnis der Herstellungsprozesse und der Bemessung von Glasbauteilen, um den steigenden Anforderungen im konstruktiven Glasbau gerecht zu werden.:1 Einleitung 2 Grundlagen zu Glaskanten 3 Wissensstand zur Kantenfestigkeit 4 Entwicklung einer Versuchsmethodik zur Erfassung der Kantenqualität 5 Systematische Analyse der Kantenqualität eines Herstellers 6 Einfluss von Prozessparametern auf die Kantenfestigkeit nach Kantenausführungsart 7 Diskussion der Ergebnisse 8 Bemessungsansätze für Glaskanten 9 Handlungsempfehlungen 10 Zusammenfassung und Ausblick 11 Literatur / In the building industry, different types of glass edges are used. Depending on their type, they fulfil requirements for cutting protection, dimensional accuracy, and aesthetics. According to DIN 1249-11, the classification is based on the visual appearance and includes cut (KG), arrissed (KGS), ground (KMG), smooth ground (KGN), and polished (KPO) edge finishing type. However, the mechanical strength of the glass edge requires additional consideration. The characteristic bending tensile strength of glass depends mostly on its surface condition. During production and processing the edge comes in contact with hard tools, which modifies the optical appearance and influences the strength. Presently, the European standardization regulates the edge strength in the design by general coefficients, which reduce the characteristic bending tensile strength depending on the edge finishing type. Existing studies show a large range of values in edge strength depending on the edge finishing and the manufacturer. The design edge strength considers the lower limits of the available glass edge finishing types. Scientifically based evaluation of the optical edge quality with assessment of the mechanical strength is available only for the cut edge. For the industrial cutting process, parameters have been determined on the basis of systematic investigations that show a positive influence on edge strength. It is still unknown how surface interferences by grinding and polishing processes affect the edge strength of processed glass edges and what are the underlying mechanisms. In addition, there is a lack of unified methods for assessing the mechanical edge quality. Within the scope of this work, a test methodology is developed that enables the determination of comparable optical and mechanical characteristics by means of microscopic analysis and destructive tests. A combination of different methods is used to characterize fracture-causing defects and to identify the location of their creation in the manufacturing process. The testing methods are approved and evaluated on the basis of a systematic examination of different edge types of one manufacturer. Subsequently, process adjustments for defect reduction are conducted and investigated with regard to an increase in edge strength. Moreover, the manufacturing process of various polished edge types of a second manufacturer are examined. Thereby, fracture-causing flaws are identified and hypotheses for future investigations are derived that form the basis for processing recommendations for the manufacturing of glass edges. The current status of relevant European design rules is finally discussed for normative recommendations. The arrised edge finishing type revealed the highest edge strength. For the ground, smooth ground, and polished edges, the edge strength is lower compared to the cut and arrised edges. The examination of the polished edge from the second manufacturer resulted in a higher edge strength compared to the polished edge from the first manufacturer. Therefore, the influence of the machine configuration is derived as an impacting process parameter. The analysis of fracturecausing defects has confirmed that a higher edge strength correlates with a reduction in microscopically detectable defects for the cut, arrised, and polished edge types. The defect analysis also enables the detection of strength-reducing defects that are introduced after production and require edge protection measures. Definitions of generally valid process parameters in grinding and polishing that positively affect the edge strength of machined edges are unknown. However, the results of this work enable manufacturers to optimize processes specific to their (manufacturing) process in order to ensure high values of edge strength. Nonetheless, experimental verifications are indispensable in this respect. Optical and mechanical properties of glass edges need to be considered separately. For example, the polished edge is considered to offer the highest optical quality. However, contrary to a widespread assumption, it does not necessarily display high edge strength. A better understanding of the manufacturing processes and the verification of edge strength are necessary for the design to meet the increasing demands in structural glass applications.:1 Einleitung 2 Grundlagen zu Glaskanten 3 Wissensstand zur Kantenfestigkeit 4 Entwicklung einer Versuchsmethodik zur Erfassung der Kantenqualität 5 Systematische Analyse der Kantenqualität eines Herstellers 6 Einfluss von Prozessparametern auf die Kantenfestigkeit nach Kantenausführungsart 7 Diskussion der Ergebnisse 8 Bemessungsansätze für Glaskanten 9 Handlungsempfehlungen 10 Zusammenfassung und Ausblick 11 Literatur

A utilização mais eficiente dos recursos naturais tem levado a indústria a aumentar a eficiência dos produtos manufaturados. Veículos de transporte Diesel estão incluídos neste cenário. Nos motores Diesel, um componente importante do sistema de injeção é o bico injetor. Uma legislação recente no Brasil introduziu limites de emissões EURO5. Estes novos parâmetros tiveram impacto no projeto do bico injetor com tolerâncias sendo reduzidas, em especial na área da sede do corpo do bico injetor, alterando o parâmetro funcional de pressão de abertura do produto. Neste cenário, o impacto desta superfície cônica retificada sobre a pressão de abertura é grande, bem como, sobre o desempenho do motor Diesel. Este estudo tem como objetivo analisar o defeito de topografia superficial típico e recorrente no processo de retificação interna cônica da sede do corpo dos bicos injetores. Um experimento planejado foi realizado de acordo com método Taguchi e determinou-se a relação sinal-ruído para dois parâmetros de topografia 2D. A superfície da sede também foi analisada utilizando análise topográfica 3D. Os resultados deste estudo incluem a indicação de possíveis causas do defeito recorrente, caracterização dos principais elementos do processo de retificação, caracterização da superfície retificada e otimização dos parâmetros do processo de retificação. / Industries of manufactured products have increased their efficiency optimizing the natural resources usage and Diesel commercial vehicles are included in this scenario. For Diesel engines, one of the most important components of the injection system is the nozzle injector. In Brazil, EURO5 legislation was recently introduced, bringing new emission limits for Diesel engines. Because of this, the nozzle injector design has changed and some manufacturing tolerances were reduced, in special the body seat geometry. This also changed the nozzle opening pressure. In this new process, the body seat grinded conical surface impacts on this functional parameter and consequently the Diesel engine performance. This study has as target to analyze a recurrent defect in the internal conic grinding process of the nozzle body seat. A trial was performed in this process according to Taguchi method and signal / noise ratio for 2D topographic parameters were defined. The body seat surface was also analyzed using 3D topographic analysis. The results of this study include the possible cause of the recurrent failure, characterization of the ground surface, process main elements integrity assessment and optimization of the grinding process parameters.