• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • Tagged with
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 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.
1

Avalia??o da aplicabilidade de correla??es matem?ticas e redes neurais na predi??o de par?metros de especifica??o do diesel

Oliveira, Fernanda Maria de 31 July 2014 (has links)
Submitted by Automa??o e Estat?stica (sst@bczm.ufrn.br) on 2016-06-23T19:59:24Z No. of bitstreams: 1 FernandaMariaDeOliveira_DISSERT.pdf: 5023454 bytes, checksum: ad313de2391d707658adc05fbcb2a8e0 (MD5) / Approved for entry into archive by Arlan Eloi Leite Silva (eloihistoriador@yahoo.com.br) on 2016-06-27T19:42:35Z (GMT) No. of bitstreams: 1 FernandaMariaDeOliveira_DISSERT.pdf: 5023454 bytes, checksum: ad313de2391d707658adc05fbcb2a8e0 (MD5) / Made available in DSpace on 2016-06-27T19:42:35Z (GMT). No. of bitstreams: 1 FernandaMariaDeOliveira_DISSERT.pdf: 5023454 bytes, checksum: ad313de2391d707658adc05fbcb2a8e0 (MD5) Previous issue date: 2014-07-31 / No Brasil, o controle da qualidade do ?leo Diesel comercializado ? realizado por monitoramento de propriedades f?sico-qu?micas, caracter?sticas do combust?vel atrav?s das Resolu??es ANP n? 65 de 09 de dezembro de 2011 e n? 45 de 20 de dezembro de 2012, que determinam os limites de especifica??o para cada par?metro e as metodologias de an?lise que devem ser adotados. No entanto, esses m?todos, apesar de bastante consolidados, possuem alguns inconvenientes t?cnicos, que levaram ao estudo de m?todos alternativos mais r?pidos e de menor custo. Este trabalho realizou uma avalia??o da aplicabilidade de equa??es matem?ticas dispon?veis na literatura e de Redes Neurais Artificiais (RNAs), na determina??o de par?metros de especifica??o do ?leo diesel. Foi realizado um levantamento bibliogr?fico das principais correla??es adequadas para a determina??o de propriedades do diesel, as quais foram aplicadas para obten??o do ponto de fulgor e do ?ndice de cetano, al?m da aplica??o, de forma mais resumida, para predizer propriedades do petr?leo. Para este estudo, foram utilizadas 162 amostras de diesel, com teores m?ximo de enxofre, 50 ppm, 500 ppm e 1800 ppm, que foram analisadas, em laborat?rio especializado, por meio de metodologias ASTM normatizadas pela ANP, com um total de 810 ensaios. Resultados experimentais das amostras de diesel, destila??o atmosf?rica, ASTM D86, e massa espec?fica, ASTM D4052, foram utilizados como vari?veis b?sicas de entrada para as equa??es avaliadas. As RNAs foram avaliadas para a predi??o do ponto de fulgor, ?ndice de cetano e teores de enxofre (S50, S500, S1800) e, nesta parte do trabalho foram testados dois tipos de arquiteturas de rede, feed-forward backpropagation e generalized regression, variando os par?metros da matriz de entrada de forma a determinar o grupo de vari?veis e melhor tipo de rede para predi??o das vari?veis de interesse. Os resultados obtidos pelas equa??es e pelas RNAs foram comparados com resultados experimentais obtidos por metodologias padr?o, utilizando o teste n?o param?trico de Mann-Whitney e test t de student, ao n?vel de signific?ncia de 5%, assim como pelo coeficiente de determina??o e erro percentual. Os resultados obtidos pela equa??o aplicada para o ponto de fulgor apresentou um erro de 27, 6 %, contudo observou-se uma tend?ncia um tanto similar aos resultados padr?o. O ?ndice de cetano foi obtido por tr?s equa??es, e ambas apresentaram bons coeficientes de correla??o, com destaque para equa??o baseada no ponto de anilina, que apresentou o menor erro de 0,8 %. As equa??es aplicadas ao petr?leo foram trabalhadas de forma a identificar perspectivas de trabalhos futuros, que se mostram bastante promissores. As RNAs para predi??o do ponto de fulgor e do ?ndice de cetano mostraram resultados bastante superiores ao observados com as equa??es matem?ticas, com erros respectivamente de 2,5% e 0,2%. Para os teores de enxofre S50, S500, e S1800, as RNAs constru?das utilizando como dados de entrada a destila??o D86 (T10, T50, T85 e T90), Massa Espec?fica, ?ndice de Cetano e Ponto de Fulgor apresentaram os melhores resultados. Dentre os teores de enxofre as RNAs conseguiram melhor predizer o S1800, e o erro obtido a partir dos resultados das RNAs aumentou com a diminui??o do teor de enxofre. De um modo geral, as redes do tipo feed-forward mostraram-se superiores as generalized regression. / Diesel fuel is one of leading petroleum products marketed in Brazil, and has its quality monitored by specialized laboratories linked to the National Agency of Petroleum, Natural Gas and Biofuels - ANP. The main trial evaluating physicochemical properties of diesel are listed in the resolutions ANP N? 65 of December 9th, 2011 and N? 45 of December 20th, 2012 that determine the specification limits for each parameter and methodologies of analysis that should be adopted. However the methods used although quite consolidated, require dedicated equipment with high cost of acquisition and maintenance, as well as technical expertise for completion of these trials. Studies for development of more rapid alternative methods and lower cost have been the focus of many researchers. In this same perspective, this work conducted an assessment of the applicability of existing specialized literature on mathematical equations and artificial neural networks (ANN) for the determination of parameters of specification diesel fuel. 162 samples of diesel with a maximum sulfur content of 50, 500 and 1800 ppm, which were analyzed in a specialized laboratory using ASTM methods recommended by the ANP, with a total of 810 trials were used for this study. Experimental results atmospheric distillation (ASTM D86), and density (ASTM D4052) of diesel samples were used as basic input variables to the equations evaluated. The RNAs were applied to predict the flash point, cetane number and sulfur content (S50, S500, S1800), in which were tested network architectures feed-forward backpropagation and generalized regression varying the parameters of the matrix input in order to determine the set of variables and the best type of network for the prediction of variables of interest. The results obtained by the equations and RNAs were compared with experimental results using the nonparametric Wilcoxon test and Student's t test, at a significance level of 5%, as well as the coefficient of determination and percentage error, an error which was obtained 27, 61% for the flash point using a specific equation. The cetane number was obtained by three equations, and both showed good correlation coefficients, especially equation based on aniline point, with the lowest error of 0,816%. ANNs for predicting the flash point and the index cetane showed quite superior results to those observed with the mathematical equations, respectively, with errors of 2,55% and 0,23%. Among the samples with different sulfur contents, the RNAs were better able to predict the S1800 with error of 1,557%. Generally, networks of the type feedforward proved superior to generalized regression.

Page generated in 0.2468 seconds