• Login
    View Item 
    •   DSpace Home
    • Teses & Dissertações
    • 40001016030P0 Programa de Pós-Graduação em Métodos Numéricos em Engenharia
    • Dissertações
    • View Item
    •   DSpace Home
    • Teses & Dissertações
    • 40001016030P0 Programa de Pós-Graduação em Métodos Numéricos em Engenharia
    • Dissertações
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Aplicação de rede neural recorrente dirigida dinamicamente em previsão de formação de geada

    Thumbnail
    View/Open
    D - D - DANIEL PEREZ SANTANA.pdf (14.00Mb)
    Date
    2003
    Author
    Santana, Daniel Perez
    Metadata
    Show full item record
    Abstract
    Resumo: A geada é um fenômeno baseado no ponto de congelamento do orvalho, o qual pode ocorrer mesmo em temperaturas positivas porque a temperatura do solo pode variar cinco graus Celsius a menos do que a temperatura ambiente local. Algumas regiões sofrem grandes perdas em produção agrícola, principalmente no sul do Brasil e Uruguai, onde a ocorrência da "geada negra" pode atingir as células das plantas e destruir a sua membrana endoplasmática. O objetivo desta dissertação é desenvolver uma rede neural recorrente dirigida dinamicamente para gerar condições de prever a ocorrência deste fenômeno e identificar e classificar os padrões de ocorrência, garantindo condições para uma implementação futura do modelo em um sistema autômato para controlar automaticamente proteções por aspersão de água em macieiras. O trabalho foi todo desenvolvido aplicando-se o pacote de redes neurais do Matlab e depois verificado a sua eficiência.
     
    Abstract: The frost is a phenomenon based in dew freezing which occurs even in positive temperatures because the soil's temperature may acquire a temperature shift over 5 Celsius degrees in comparison with the surroundings. Some regions suffer great damage, mainly the southern part of Brazil and Uruguay, where the "black frost" reaches out the plants' tissues destroying their tissues' membranes. The aim of this work is to develop a recurrent neural network dynamically directed to forecast this phenomenon, identifies and classifies its patterns, guarantying future implementation in an automate system to control automatically protections to spill water onto apple trees to be protected. All the tasks were developed using Matlab's neural network toolbox and afterwards it was verified its certainty.
     
    URI
    https://hdl.handle.net/1884/81614
    Collections
    • Dissertações [170]

    DSpace software copyright © 2002-2022  LYRASIS
    Contact Us | Send Feedback
    Theme by 
    Atmire NV
     

     

    Browse

    All of DSpaceCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsxmlui.ArtifactBrowser.Navigation.browse_typeThis CollectionBy Issue DateAuthorsTitlesSubjectsxmlui.ArtifactBrowser.Navigation.browse_type

    My Account

    LoginRegister

    Statistics

    View Usage Statistics

    DSpace software copyright © 2002-2022  LYRASIS
    Contact Us | Send Feedback
    Theme by 
    Atmire NV