• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 71
  • 57
  • 8
  • 4
  • 3
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • Tagged with
  • 163
  • 163
  • 163
  • 51
  • 48
  • 45
  • 38
  • 29
  • 29
  • 28
  • 28
  • 28
  • 25
  • 25
  • 24
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
101

Utilização de redes neurais auto-organizativas para identificação de regimes de escoamento bifásico horizontal ar-água / Self-organizing neural networks for the identification of air-water horizontal two-phase flow regimes

Kelen Cristina Oliveira Crivelaro 31 August 2004 (has links)
Um dos principais problemas relacionados ao transporte e manipulação de fluidos multifásicos refere-se a existência de regimes de escoamento e sua forte influência sobre importantes parâmetros de operação. Um bom exemplo disto ocorre em reatores químicos gás-líquido no qual máximos coeficientes de reação podem ser alcançados mantendo-se um escoamento a bolhas disperso, maximizando a área interfacial total. Assim, a habilidade de identificar automaticamente regimes de escoamento é de importância crucial, especialmente para operação adequada de sistemas multifásicos. Este trabalho se constituirá no desenvolvimento e implementação de uma rede neural auto-organizativa especializada ao problema de identificação de regimes de escoamento bifásico ar-água em tubo horizontal. Os regimes reconhecidos em tubo horizontal são seis: estratificado liso, estratificado ondulado, estratificado rugoso, intermitente, bolhas e anular. Para tanto, pretende-se tomar como ponto de partida as medidas capacitivas, variação de pressão da tubulação e medida de pressão \"flutuante\" como padrões associativos a serem armazenados na rede neural. Posteriormente, a fase de treinamento da rede neural consistirá em identificar os coeficientes sinápticos apropriados, a partir de um conjunto representativo de ensaios. Nesse trabalho optou-se por uma arquitetura com 10 neurônios na camada de entrada, portanto uma quantidade maior do que o número de regimes que se deseja identificar. O objetivo é ver se a rede neural consegue encontrar de forma autônoma os seis regimes já conhecidos, mesmo tendo 10 neurônios na grade. Isso demonstra a habilidade da rede neural auto-organizativa em identificar regimes de escoamento mesmo em situações onde não há conhecimento prévio dos mesmos. Serão feitas simulações para verificar o desempenho da rede neural a partir de dados experimentais coletados no oleoduto piloto do Núcleo de Engenharia Térmica e Fluidos da Escola de Engenharia de São Carlos. / One of the main problems related to the transport and manipulation of multiphase flow is the existence of flow regimes and its strong influence on important parameters of operation. As an example of it occurs in gas-liquid chemical reactor in which maximum coefficients of reaction can be reached by keeping a maximal flow bubbly dispersed into a total interfacial area. Thus, the ability to identify flow regimes automatically is very important, specially in order to obtain a properly operation of multiphase systems. This work consists in the development and implementation of a self-organizing neural networks specially for the problem of identification of regimes of two-phase flow air-water in horizontal pipe. The regimes recognized in horizontal pipe are six: estratified smooth, estratified wavy, estratified rugged, intermittent, bubbly and annular. The capacitive measures, variation of pressure of the piping and measure of flutuante pressure were considered as a starting point to be stored as associative standards in the neural networks. After, the phase of training of the neural networks consisted of an appropriate identification for the sinaptic coefficients from a representative set of tests. An architectural input-layer of 10 neurons was opted. Therefore a bigger amount of regimes than the ones usually required. The objective is to see if the neural network is able to find independently the six regimes already known, even having 10 neurons in the grid. This demonstrates the ability of the self-organizing maps neural network in identifying flow regimes in situations where one does not have previous knowledge of them. Simulations will be made to verify the performance of the neural network from experimental data collected in the pilot pipe-line of the Nucleus of Thermal Engineering and Fluids of the School of Engineering of São Carlos.
102

Análise de agrupamentos baseada na topologia dos dados e em mapas auto-organizáveis. / Data clustering based on data topology and self organizing-maps.

Clodis Boscarioli 16 May 2008 (has links)
Cada vez mais, na conjuntura das grandes tomadas de decisões, a análise de dados massivamente armazenados se torna uma necessidade das mais variadas áreas de conhecimento. A análise de dados envolve a realização de diferentes tarefas, que podem ser realizadas por diferentes técnicas e estratégias como análise de agrupamento de dados. Esta pesquisa enfatiza a realização da tarefa de análise de agrupamento de dados (Data Clustering) usando SOM (Self-Organizing Maps) como principal artefato. SOM é uma rede neural artificial baseada em aprendizado competitivo e não-supervisionado, o que significa que o treinamento é inteiramente guiado pelos dados e que os neurônios do mapa competem entre si. Essa rede neural possui a habilidade de formar mapeamentos que quantizam os dados, preservando a sua topologia. Este trabalho introduz uma nova metodologia de análise de agrupamentos a partir de SOM, que considera o mapa topológico gerado por ele e a topologia dos dados no processo de agrupamento. Uma análise experimental e comparativa é apresentada, evidenciando a potencialidade da proposta, destacando, por fim, as principais contribuições do trabalho. / More than ever, in environment of large decision making, the analysis of data stored massively becomes a real need in almost all knowledge areas. The data analyzing process covers the performing of different tasks that can be executed for different techniques and strategies as the data clustering analysis. This research is focused on the analysis task of data groups, called Data Clustering using Self Organizing Maps (SOM) as principal artifact. SOM is an artificial neural network based on competitive and unsupervised learning, what means that its training is entirely driven by the data, such the neurons of the map compete themselves for doing it. This neural network has the ability to build the mapping task that quantifies the source data, but preserving the topology. This work introduces a new clustering analysis methodology based on SOM, considering the topological map produced by it and also the topology of the data obtained in the clustering process. The experimental and comparative analysis are also presented to demonstrate the potential of the proposal, highlighting at the end the mainly contributions of the work.
103

Metodologia para previsão de carga de curtíssimo prazo considerando variáveis climáticas e auxiliando na programação de despacho de pequenas centrais hidrelétricas

Bordignon, Sérgio 29 June 2012 (has links)
Submitted by Sandro Camargo (sandro.camargo@unipampa.edu.br) on 2015-05-09T18:11:33Z No. of bitstreams: 1 107110004.pdf: 2957226 bytes, checksum: b15ec66f6abfaa78dc10c29127881a4b (MD5) / Made available in DSpace on 2015-05-09T18:11:33Z (GMT). No. of bitstreams: 1 107110004.pdf: 2957226 bytes, checksum: b15ec66f6abfaa78dc10c29127881a4b (MD5) Previous issue date: 2012-06-29 / A previsão de carga é uma atividade de grande importância no Setor Elétrico, tendo em vista que a maioria dos estudos de planejamento e operação dos sistemas elétricos necessita de uma boa estimativa da carga a ser atendida. Na literatura encontram-se diversas metodologias para projeção de carga elétrica nos distintos horizontes de planejamento, porém limitadas a sistemas elétricos de médio e grande porte e poucas são as propostas de projeção de demanda no horizonte de curtíssimo prazo, principalmente para pequenas empresas do Setor Elétrico. O objetivo deste trabalho é apresentar uma metodologia inovadora de previsão de carga, a curtíssimo prazo, que considere as influências das condições climáticas e que possa auxiliar na programação do regime de operação de uma Pequena Central Hidrelétrica (PCH), principalmente em épocas de estiagem, quando a disponibilidade de água é restrita. A metodologia proposta envolve a criação de um modelo probabilístico discreto (cadeia de Markov) a partir da classificação dos dados históricos em um Mapa Auto-Organizável (SOM). Assim, é possível se estimar a probabilidade de um determinado nível de demanda acontecer dada uma condição climática atual, bem como o número de intervalos de tempo (horas) até que isso aconteça. Com estas informações é possível elaborar a melhor agenda de funcionamento da PCH de forma que a mesma esteja em funcionamento nos momentos em que a demanda atingir os valores máximos. O método proposto apresenta como diferencial em relação aos demais métodos existentes o fato de considerar a influência das variáveis climáticas (temperatura, umidade relativa do ar e velocidade do vento) para a previsão de demanda de energia elétrica no curtíssimo prazo, além de que os valores de entrada de demanda de energia e das variáveis climáticas (temperatura e umidade relativa do ar) são obtidos em tempo real, através de um sistema SCADA. Esta metodologia foi aplicada utilizando-se os dados reais de uma pequena concessionária de distribuição de energia elétrica do Rio Grande do Sul, mostrando resultados satisfatórios, suficientes para permitir a sua aplicação prática. / The electrical charge forecast is an activity of great importance in the Electricity Sector, considering that most studies of electrical systems planning and operation require a good estimative of the charge to be fulfilled. In books, there are various methodologies to have the electrical charge projection in different planning horizons, but limited to medium and large electrical systems. Furthermore, there are only a few demand projection proposals in the very short-term horizon, especially for small Electricity Sector companies. The aim of this paper is to present an innovative methodology in order to have the charge forecast, in a very short-term, which considers the climatic conditions influence and is able to assist the operation system programming of a Small Hydroelectric Power Plant, particularly in times of drought when water availability is restricted. The proposed methodology involves creating a discrete probabilistic pattern (Markov chain) from the historical data classification in a Self-Organizing Map (SOM). It is therefore possible to estimate the probability of reaching a certain demand level, taking the current climatic condition, as well as the periods of time (hours) until it happens. With this information it is possible to develop the best plant operation schedule so that it operates when the demand reaches its maximum numbers. The proposed method presents as differentials upon the other existing methods, the fact of considering the climatic variables influence (temperature, air humidity and wind speed) to forecast electricity demand in the very short-term, as well as the energy demand input values and climate variables obtainment (temperature and air humidity) in real time via a SCADA system. This methodology was applied using real data from a small electricity distribution plant in Rio Grande do Sul, showing satisfactory results, enough to allow their practical application.
104

Emprego de mapas auto-organizáveis para localização de faltas em redes de distribuição

Cavalheiro, Franciele Cristina 19 October 2012 (has links)
Submitted by Sandro Camargo (sandro.camargo@unipampa.edu.br) on 2015-05-09T19:03:09Z No. of bitstreams: 1 107110001.pdf: 3453553 bytes, checksum: fb266cf11dce80833ef41b2adb20cd21 (MD5) / Made available in DSpace on 2015-05-09T19:03:09Z (GMT). No. of bitstreams: 1 107110001.pdf: 3453553 bytes, checksum: fb266cf11dce80833ef41b2adb20cd21 (MD5) Previous issue date: 2012-10-19 / As distribuidoras de energia elétrica devem estar preparadas para restabelecer o fornecimento de forma ágil, sempre respeitando os critérios de segurança e os índices de continuidade e qualidade compatíveis com o estabelecido na legislação vigente. A possibilidade de localizar defeitos remotamente acelera o processo de restabelecimento. Apesar dos próprios relés e religadores microprocessados fornecerem uma indicação com base na impedância de curto-circuito, ela não é precisa para sistemas de distribuição, uma vez que as redes apresentam várias bifurcações (ramais) protegidas por elementos fusíveis e diferentes bitolas de condutores. Entretanto, atualmente há vários métodos como revelam as pesquisas, que tentam estabelecer maior precisão dos dados e resultados ali envolvidos, buscando criar uma inovação e satisfação às empresas do setor elétrico. Assim, a proposta deste projeto é desenvolver uma metodologia diferencial para localização de faltas em redes de distribuição a partir de estudos da rede combinados com a monitoração remota de dispositivos de proteção (relés e religadores microprocessados). Estas variáveis de entrada: corrente de carga (corrente pré-falta), corrente de curto-circuito, corrente pós-falta; serão obtidas em tempo real pelo SCADA no momento da ocorrência de uma falta na rede. As medidas obtidas serão comparadas e classificadas por Mapas Auto-Organizáveis ou SOM (Self-Organizing Map) de acordo com os padrões de dados simulados pelos estudos da rede. A partir do método proposto será possível estimar o local do defeito ocorrido na rede elétrica por meio do mapa que melhor se identifica com os dados medidos. Essa pesquisa foi aplicada no estudo de caso de uma concessionária da região central gaúcha, a qual alcançou resultados bastante satisfatórios, obtidos a partir de dados reais. / The power utilities must be prepared to restore the supply in an agile way constantly respecting the safety criteria and indexes of continuity and quality consistent with the current legislation. The ability to remotely locate defects accelerates this healing process. Despite own microprocessor relays and reclosers provide an indication based on the short circuit impedance, it is not applicable to distribution systems, since the networks present various bifurcations (branch lines) protected by fuse elements and different wire gauges. However, today there are several methods such as surveys show, trying to establish more precisely the data and results here involved, seeking to create an innovation and satisfaction to the electric companies. Thus, the aim of this project is to develop a methodology for distinct fault location in distribution feeders. It will combine network studies with remote monitoring of protective devices (microprocessor relays and reclosers). These input variables: load current (pre-fault current), short circuit current, post-fault current; will be obtained in real time by SCADA when a fault occurs in the network. These measures will be compared and ranked by Self- Organizing Maps (SOM) according to the patterns obtained by simulation studies of the network. With this method, it will be possible to assess the location of the defect occurring in the power grid, identifying the best map that resembles with the measured data. This research was applied in a power utility in the South of Brazil as case study, which achieved satisfactory results obtained from real data.
105

Utilização de redes neurais auto-organizativas para identificação de regimes de escoamento bifásico horizontal ar-água / Self-organizing neural networks for the identification of air-water horizontal two-phase flow regimes

Crivelaro, Kelen Cristina Oliveira 31 August 2004 (has links)
Um dos principais problemas relacionados ao transporte e manipulação de fluidos multifásicos refere-se a existência de regimes de escoamento e sua forte influência sobre importantes parâmetros de operação. Um bom exemplo disto ocorre em reatores químicos gás-líquido no qual máximos coeficientes de reação podem ser alcançados mantendo-se um escoamento a bolhas disperso, maximizando a área interfacial total. Assim, a habilidade de identificar automaticamente regimes de escoamento é de importância crucial, especialmente para operação adequada de sistemas multifásicos. Este trabalho se constituirá no desenvolvimento e implementação de uma rede neural auto-organizativa especializada ao problema de identificação de regimes de escoamento bifásico ar-água em tubo horizontal. Os regimes reconhecidos em tubo horizontal são seis: estratificado liso, estratificado ondulado, estratificado rugoso, intermitente, bolhas e anular. Para tanto, pretende-se tomar como ponto de partida as medidas capacitivas, variação de pressão da tubulação e medida de pressão \"flutuante\" como padrões associativos a serem armazenados na rede neural. Posteriormente, a fase de treinamento da rede neural consistirá em identificar os coeficientes sinápticos apropriados, a partir de um conjunto representativo de ensaios. Nesse trabalho optou-se por uma arquitetura com 10 neurônios na camada de entrada, portanto uma quantidade maior do que o número de regimes que se deseja identificar. O objetivo é ver se a rede neural consegue encontrar de forma autônoma os seis regimes já conhecidos, mesmo tendo 10 neurônios na grade. Isso demonstra a habilidade da rede neural auto-organizativa em identificar regimes de escoamento mesmo em situações onde não há conhecimento prévio dos mesmos. Serão feitas simulações para verificar o desempenho da rede neural a partir de dados experimentais coletados no oleoduto piloto do Núcleo de Engenharia Térmica e Fluidos da Escola de Engenharia de São Carlos. / One of the main problems related to the transport and manipulation of multiphase flow is the existence of flow regimes and its strong influence on important parameters of operation. As an example of it occurs in gas-liquid chemical reactor in which maximum coefficients of reaction can be reached by keeping a maximal flow bubbly dispersed into a total interfacial area. Thus, the ability to identify flow regimes automatically is very important, specially in order to obtain a properly operation of multiphase systems. This work consists in the development and implementation of a self-organizing neural networks specially for the problem of identification of regimes of two-phase flow air-water in horizontal pipe. The regimes recognized in horizontal pipe are six: estratified smooth, estratified wavy, estratified rugged, intermittent, bubbly and annular. The capacitive measures, variation of pressure of the piping and measure of flutuante pressure were considered as a starting point to be stored as associative standards in the neural networks. After, the phase of training of the neural networks consisted of an appropriate identification for the sinaptic coefficients from a representative set of tests. An architectural input-layer of 10 neurons was opted. Therefore a bigger amount of regimes than the ones usually required. The objective is to see if the neural network is able to find independently the six regimes already known, even having 10 neurons in the grid. This demonstrates the ability of the self-organizing maps neural network in identifying flow regimes in situations where one does not have previous knowledge of them. Simulations will be made to verify the performance of the neural network from experimental data collected in the pilot pipe-line of the Nucleus of Thermal Engineering and Fluids of the School of Engineering of São Carlos.
106

Cartes auto-organisatrices pour la classification de données symboliques mixtes, de données de type intervalle et de données discrétisées. / Self-Organizing Maps for the clustering of mixed feature-type symbolic data, of interval-valued data and of binned data

Hajjar, Chantal 10 February 2014 (has links)
Cette thèse s'inscrit dans le cadre de la classification automatique de données symboliques par des méthodes géométriques bio-inspirées, plus spécifiquement par les cartes auto-organisatrices. Nous mettons en place plusieurs algorithmes d'apprentissage des cartes auto-organisatrices pour classifier des données symboliques mixtes ainsi que des données de type intervalle et des données discrétisées. Plusieurs jeux de données symboliques simulées et réelles, dont deux construits dans le cadre de cette thèse, sont utilisés pour tester les méthodes proposées. En plus, nous proposons une carte auto-organisatrice pour les données discrétisées (binned data) dans le but d'accélérer l'apprentissage des cartes classiques et nous appliquons la méthode proposée à la segmentation d'images. / This thesis concerns the clustering of symbolic data with bio-inspired geometric methods, more specifically with Self-Organizing Maps. We set up several learning algorithms for the self-organizing maps in order to cluster mixed-feature symbolic data as well as interval-valued data and binned data. Several simulated and real symbolic data sets, including two sets built as part of this thesis, are used to test the proposed methods. In addition, we propose a self-organizing map for binned data in order to accelerate the learning of standard maps, and we use the proposed method for image segmentation.
107

Self-organizing maps for virtual sensors, fault detection and fault isolation in diesel engines

Bergkvist, Conny, Wikner, Stefan January 2005 (has links)
<p>This master thesis report discusses the use of self-organizing maps in a diesel engine management system. Self-organizing maps are one type of artificial neural networks that are good at visualizing data and solving classification problems. The system studied is the Vindax(R) development system from Axeon Ltd. By rewriting the problem formulation also function estimation and conditioning problems can be solved apart from classification problems. </p><p>In this report a feasibility study of the Vindax(R) development system is performed and for implementation the inlet air system is diagnosed and the engine torque is estimated. The results indicate that self-organizing maps can be used in future diagnosis functions as well as virtual sensors when physical models are hard to accomplish.</p>
108

Self-organizing maps for virtual sensors, fault detection and fault isolation in diesel engines

Bergkvist, Conny, Wikner, Stefan January 2005 (has links)
This master thesis report discusses the use of self-organizing maps in a diesel engine management system. Self-organizing maps are one type of artificial neural networks that are good at visualizing data and solving classification problems. The system studied is the Vindax(R) development system from Axeon Ltd. By rewriting the problem formulation also function estimation and conditioning problems can be solved apart from classification problems. In this report a feasibility study of the Vindax(R) development system is performed and for implementation the inlet air system is diagnosed and the engine torque is estimated. The results indicate that self-organizing maps can be used in future diagnosis functions as well as virtual sensors when physical models are hard to accomplish.
109

Predicting The Effect Of Hydrophobicity Surface On Binding Affinity Of Pcp-like Compounds Using Machine Learning Methods

Yoldas, Mine 01 April 2011 (has links) (PDF)
This study aims to predict the binding affinity of the PCP-like compounds by means of molecular hydrophobicity. Molecular hydrophobicity is an important property which affects the binding affinity of molecules. The values of molecular hydrophobicity of molecules are obtained on three-dimensional coordinate system. Our aim is to reduce the number of points on the hydrophobicity surface of the molecules. This is modeled by using self organizing maps (SOM) and k-means clustering. The feature sets obtained from SOM and k-means clustering are used in order to predict binding affinity of molecules individually. Support vector regression and partial least squares regression are used for prediction.
110

Patterns and dynamics of ocean circulation variability on the West Florida shelf

Liu, Yonggang 01 June 2006 (has links)
Patterns of variability and the dynamics of the ocean circulation on the West Florida Shelf (WFS) are investigated using multi-year, shelf-wide oceanographic observations from moored Acoustic Doppler Current Profiler (ADCP) arrays,hydrographic cruises, High-Frequency (HF) radars, satellites, and coastal tide gauges.Novel neural network techniques, Self-Organizing Map (SOM) and Growing Hierarchical Self-Organizing Maps (GHSOM), are introduced as feature extraction methods in physical oceanography. The SOM is demystified and demonstrated to be a useful feature extraction method in a series of performance evaluations using artificial data sets comprising known patterns. It is then applied to velocity time series from moored ADCP arrays and to a joint HF-radar and ADCP data set, respectively, to extract patterns of ocean current variability, and it is shown to be a useful technique for extracting dynamically consistent ocean current patterns. The extracted characteristic patte rns of upwelling/downwelling variability are coherent with the local winds on the synoptic weather time scale, and coherent with both the local winds and thecomplementary Sea Surface Temperature (SST) patterns on the seasonal time scale. Thecurrents are predominantly southeastward during fall-winter and northwestward during summer. The GHSOM is used to describe the SST seasonal variation. As feature extraction methods, both the SOM and the GHSOM have advantages over the conventional Empirical Orthogonal Function method.The circulation dynamics are examined, first through depth-averaged momentum balances at selected locations and then via sea surface height (SSH) estimates across the inner shelf. Dominant dynamics of the shelf circulation are diagnosed and a method is discussed for estimating along-shelf currents from coastal sea level and wind data. Nontidal coastal sea level fluctuations are related to both the offshore SSH and the dynamical responses of the inner shelf to wind and bu oyancy forcing. The across-shelf distribution of the SSH is estimated from the velocity, hydrography, wind, and coastal sea level data.Subtracting the variability that may be accounted for by inner shelf dynamical responses yields a residual at the 50 m isobath that compares well with satellite altimetry data. This suggests the possibility of calibrating satellite SSH data on the continental shelf.

Page generated in 0.055 seconds