We present a novel method for example-based simulation of fluid flow. We reconstruct fluid animation from physically based fluid simulation examples. Our framework shows how to decompose a given series of fluid motion example data into small units and then recompose them. We capture the properties of local fluid behavior by dicing the fluid motion example data into sequences of fragments, which have smaller volume and shorter length. We build a database out of these fragments, and propose a matching strategy to generate new fluid animation. To achieve highly efficient database query, we project our fragments onto lower dimensional subspace using Principal Component Analysis (PCA) approach, and construct our data structure as a kd-tree by treating each fragment as a point in this subspace. Our method has been implemented in synthesizing both two-dimensional (2D) and three-dimensional (3D) fluid’s velocity fields.

We present a novel method for example-based simulation of fluid flow. We reconstruct fluid animation from physically based fluid simulation examples. Our framework shows how to decompose a given series of fluid motion example data into small units and then recompose them. We capture the properties of local fluid behavior by dicing the fluid motion example data into sequences of fragments, which have smaller volume and shorter length. We build a database out of these fragments, and propose a matching strategy to generate new fluid animation. To achieve highly efficient database query, we project our fragments onto lower dimensional subspace using Principal Component Analysis (PCA) approach, and construct our data structure as a kd-tree by treating each fragment as a point in this subspace. Our method has been implemented in synthesizing both two-dimensional (2D) and three-dimensional (3D) fluid’s velocity fields.

Inom avloppsvattenrening finns ett intresse att mäta olika parametrar, främst i inkommande och utgående vatten, för att få kunskap om hur väl reningen fungerar samt hur reningen eventuellt ska styras beroende på vattnets egenskaper. De vanligaste metoderna för provtagning och analys av avloppsvatten är antingen med fysiska sensorer eller genom analyser på lab. Ytterliggare ett alternativ för att ta reda på egenskaper hos vattnet är att konstruera softsensorer, en sensor som bygger på modellerade värden av den parameter som ska mätas utifrån andra parametrar som mäts med fysiska sensorer. Syftet med denna studie var att undersöka möjligheten att skapa softsensorer för COD, tot-N, tot-P, NH4-N, NO3-N och PO4-P i olika steg av vattenreningen. Skapandet av softsensorer utfördes genom datainsamling, modellering och validering. Datainsamlingen har utförts på en småskalig reningsanläggning på Hammarby Sjöstadsverk i Stockholm. Värden från de befintliga fysiska sensorerna samt analyserade värden från laborativt arbete har tillsammans förbehandlats och använts i mjukvaran SIMCA för att skapa modeller för givna parametrar. Modelleringen bygger på multivariat dataanalys där syftet är att utvinna värdefull information ur ett stort set med data. Vid ett senare tillfälle insamlades oberoende data som inte ingått i modelleringsarbetet och som användes för att validera modellerna. Resultatet av själva modelleringen gav goda modeller för tot-P och PO4-P i inkommande vatten. Ingående data beskrivs för dessa två parametrar väl av respektive modell och de ger båda en god prediktionsförmåga. För övriga parametrar erhölls inte tillräckligt bra modeller. Valideringen visade dålig överenstämmelse mellan predikterade värden från modellerna för tot-P respektive PO4-P i inkommande vatten och observerade värden. Detta resultat beror med stor sannolikhet på ett byte av inlopp för inkommande vatten till reningsanläggningen. Något som gav stora skillnader mellan de förhållanden som varade vid datainsamlingen för skapandet av modellerna och de förhållanden som varade vid datainsamling för valideringen. Förändringen av inkommande vatten var antagligen för stor för modellerna att klara av. Trots en dålig validering antas det utifrån själva modellerna och tidigare studier finnas goda möjligheter att skapa softsensorer som kan implementeras på Hammarby Sjöstadsverk.

In this thesis, a novel method for multi-pulse with equal chirp characteristics and more efficient THz generation through two photon absorption (TPA) are investigated and demonstrated. These are all the first time, for our best knowledge, odd multi-pulses generation and innovative approach for efficient THz through TPA are proposed and studied. By modulating the amplitude and phase of the spatial light modulator with the pulse shaper, the number of multi-pulses can be adjustable without the limitation of even number only. Meanwhile, the chirp properties of the generated pulses are with the same characteristics and tunable also. For the case with bandwidth of 10nm, the generated multi-pulses with equal chirps varying from -20000fs^2 to 20000fs^2 are demonstrated and the results have a good agreement with the theoretical estimation. We also discuss the number of the multi-pulses and inequalities of the amplitude of the pulses are limited by the spectral resolution of SLM within the pulse shaper. Regarding to efficient THz radiation, it can be generated more efficiently from a low-temperature-grown GaAs (LT-GaAs) photoconductive (PC) antenna by taking into account the TPA induced photo-carrier in the photoconductor. A rate-equation-based approach using the Drude-Lorentz model taking into account the band-diagram of LT-GaAs is used for the theoretical analysis. The super-linear power dependent photocurrent clarifies the role of TPA. Previously unnoticed THz pulse features and anomalously increasing THz radiation power rather than saturation were observed. These are in good agreement with the theoretical predictions.

The purpose of this study is to understand the composition of the secondary forest of Nanjenshan area. A 5-ha plot was established at altitudes 200-264m on the north ridge of Nanjenshan. All free-standing woody plantsin the plot with DBH¡Ù1cm were identified, measured, tagged and mapped. A total of 108 tree species belonging to 40 families and 77 genera was recorded. The dominant species are Acacia confusa, Lagerstroemia subcostata, Psychotria rubra, Ficus septica¤ÎDendrocnide meyeniana. The dominant families are Fabaceae, Moraceae, Euphorbiaceae, Rubiaceae¤ÎLythraceae. TWINSPAN analysis reveales four groups. Group 1 is dominanted by Acacia confusa, Glochidion philippicum and Psychotria rubra. Group 2 is dominanted by Acacia confusa, Psychotria rubra and Lagerstroemia subcostata. Group 3 is dominanted by Acacia confusa, Psychotria rubra and Glochidion rubrum. Group 4 is dominanted by Acacia confusa, Lagerstroemia subcostata and Dendrocnide meyeniana. There are four kind of distribution pattern in this plot. Machilus obovatifolia and Osmanthus marginatus occurred on the hilltop; Dendrocnide meyeniana and Ficus virgata occurred on the middle-woody site; Zanthoxylum ailanthoides and Wendlandia uvariifolia only occurred along the steamside; Acacia confusa and Psychotria rubra were commonly distributed over the plot.

Modern IT development infrastructure has a large number of components that must be monitored, for instance servers and network components. Various system-metrics (build time, CPU utilization, queries time etc.) are gathered to monitor system performance. In practice, it is extremely difficult for a system administrator to observe a correlation between several systemmetrics and predict a target system-metric based on highly correlated system-metrics without machine learning support. The experiments were performed on development logs at Ericsson. There were many system-metrics available in the system. Our goal is use machine learning techniques to find correlation between buildtime and other system-metrics and predict its trends in the future.

We present a novel method for example-based simulation of fluid flow. We reconstruct fluid animation from physically based fluid simulation examples. Our framework shows how to decompose a given series of fluid motion example data into small units and then recompose them. We capture the properties of local fluid behavior by dicing the fluid motion example data into sequences of fragments, which have smaller volume and shorter length. We build a database out of these fragments, and propose a matching strategy to generate new fluid animation. To achieve highly efficient database query, we project our fragments onto lower dimensional subspace using Principal Component Analysis (PCA) approach, and construct our data structure as a kd-tree by treating each fragment as a point in this subspace. Our method has been implemented in synthesizing both two-dimensional (2D) and three-dimensional (3D) fluid’s velocity fields.

Orientador: Lino Lemonica / Resumo: Com o rápido avanço que foi observado nos últimos anos nas técnicas cirúrgicas e anestésicas, os procedimentos se tornaram cada vez mais invasivos. Como houve progressivo envelhecimento da população, o período mais delicado de recuperação, ou seja, o pós-operatório, passou a receber maior atenção. O surgimento de novas técnicas de analgesia, como a Analgesia Controlada pelo Paciente (PCA), vem preencher as necessidades da comunidade médica, cada vez mais atenta à qualidade dos serviços prestados. O Hospital de Clínicas da Faculdade de Medicina da UNESP, Botucatu, atento a essa nova realidade constituiu o Serviço de Dor Aguda (SEDA) para que esta lacuna fosse também preenchida em nosso meio. No intuito de identificar a atuação do SEDA, realizou-se levantamento, de fevereiro de 1995 a dezembro de 1997, com a pesquisa das evoluções de 679 pacientes seguidos pelo SEDA e que fizeram uso do método PCA de analgesia. Observou-se que os resultados obtidos pelo Serviço estavam acima da média relatada pela literatura internacional, com excelentes níveis de analgesia atingidos, baixa ocorrência de efeitos colaterais, e nenhuma complicação fatal no período do estudo. / Abstract: In the past years a fast developing has been observed in the surgery and anesthetic technique, with more invasive procedures being performed. As the general population has becoming older, the critical recovery period, it means the post surgery period, became focus of attention. The developing of new analgesia techniques, such as Patient Controlled Analgesia (PCA), has the intention of fulfill the needs of the medical community, day by day more aware about the quality of the services. The Hospital of the Sao Paulo State Medical School, Botucatu, aware of this new reality had decided to form the Acute Pain Management Service (SEDA). With the goal of identify the way the SEDA acts this research was realized in period between February, 1995, to December, 1997. Data of 679 patients who used the PCA device were evaluated. The results in this study were as good as the international literature shows, with high quality analgesia, low side effects and no fatal complications on the period observed. / Mestre

Made available in DSpace on 2014-06-12T18:07:13Z (GMT). No. of bitstreams: 2 arquivo7925_1.pdf: 1184307 bytes, checksum: 2503378983f5e07ea3147d8bde83ec7f (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2005 / O álcool combustível, por ser uma fonte renovável de energia e menos poluente que os combustíveis fósseis, está se tornando uma das principais alternativas para substituição destes. Novas tecnologias e principalmente o Protocolo de Kyoto vêm incentivando o aumento na produção e consumo do álcool. O Brasil, por ter grande capacidade de produção e suficiência nesta área, pode vir a ser um dos maiores exportadores de álcool combustível. Para isto, além de aumentar sua produtividade, o álcool produzido deve também ter uma qualidade que atenda às exigências do mercado exterior. Sendo assim, um processo de produção que seja controlado de forma eficiente, utilizando metodologias analíticas confiáveis, rápidas e que possam determinar simultaneamente vários parâmetros de qualidade é de fundamental importância. A espectroscopia na região do infravermelho próximo (NIR) e intermediário (MID), por ser rápida, não destrutiva, de relativo baixo custo e possibilitar a determinação de vários constituintes simultaneamente em um curto espaço de tempo, em conjunto com métodos quimiométricos para tratamento dos dados, vem sendo amplamente empregada no controle de processos e produtos, inclusive em usinas de produção de açúcar. Com o objetivo de caracterizar, classificar e prever algumas propriedades do álcool combustível produzido na região foram coletadas amostras de álcool etílico anidro e hidratado em três usinas de Pernambuco e uma de Sergipe e determinados 15 parâmetros de qualidade. Foram também obtidos os seus espectros na região do infravermelho próximo e intermediário. A técnica estatística multivariada de análise de componentes principais (ACP) foi empregada na análise exploratória dos dados obtidos, com o objetivo de avaliar existência de possíveis agrupamentos referentes a alguma característica específica. Os melhores resultados foram as ACP dos espectros de MID que mostraram uma melhor separação dos agrupamentos por usinas. Foi investigada a possibilidade de utilizar a espectroscopia MID como alternativa para a determinação de algumas propriedades e contaminantes do álcool combustível, através de modelos de regressão por mínimos quadrados parciais (PLS - Partial Least Squares). Os resultados para o álcool anidro não foram satisfatórios, apenas a densidade foi bem prevista, com erro de previsão comparável ao valor estimado para a precisão intermediária do método de referência. Para o álcool hidratado, parâmetros como metanol, densidade, acetato de etila, álcool n-propílico e pH, apresentaram correlações satisfatórias, sendo que para a densidade, pH e álcool n-propílico forma obtidos erros de previsão equivalentes aos valores de precisão intermediária dos métodos de referência. Além disso, foi investigada a possibilidade de utilização da espectroscopia na região MID, aliada a técnica de classificação multivariada SIMCA (Soft Independent Modeling of Class Analogies), para a identificação da fraude em que o álcool anidro é rehidratado (álcool molhado) e vendido como álcool hidratado, sem pagamento de impostos. Os resultados mostraram-se promissores para a identificação deste tipo de fraude. Através do modelo construído, nenhuma das amostras de teste foi classificada incorretamente. Entretanto, 15% foram classificadas como pertencentes às duas classes (molhado e hidratado)

We present a novel method for example-based simulation of fluid flow. We reconstruct fluid animation from physically based fluid simulation examples. Our framework shows how to decompose a given series of fluid motion example data into small units and then recompose them. We capture the properties of local fluid behavior by dicing the fluid motion example data into sequences of fragments, which have smaller volume and shorter length. We build a database out of these fragments, and propose a matching strategy to generate new fluid animation. To achieve highly efficient database query, we project our fragments onto lower dimensional subspace using Principal Component Analysis (PCA) approach, and construct our data structure as a kd-tree by treating each fragment as a point in this subspace. Our method has been implemented in synthesizing both two-dimensional (2D) and three-dimensional (3D) fluid’s velocity fields.