• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Doctoral Thesis
DOI
https://doi.org/10.11606/T.3.1999.tde-11032004-160112
Document
Author
Full name
Ademar de Azevedo Cardoso
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Paulo, 1999
Supervisor
Committee
Augusto, Oscar Brito (President)
Costa Neto, Alvaro
Dias, Carlos Alberto Nunes
Ragazzo, Clodoaldo Grotta
Schmidt, Hernan Prieto
Title in Portuguese
Redes neurais artificiais na avaliação de concentração de tensões em juntas tubulares soldadas.
Keywords in Portuguese
juntas tubulares soldadas
método dos elementos finitos
redes neurais artificiais
Abstract in Portuguese
Neste trabalho está apresentada uma alternativa para o cálculo do fator de concentração de tensões (FCT) em juntas tubulares soldadas do tipo Y. Redes Neurais Artificiais (RNA) foram utilizadas para representar a distribuição de tensões ao longo da junta tubular para os casos de carregamento força axial no plano e momento fletor no plano. As RNA podem aprender a partir de um conjunto de dados sem a necessidade de uma expressão matemática entre as variáveis dependentes e independentes; representa uma vantagem sobre o procedimento normalmente utilizado, ou seja, as equações paramétricas. O modelo proposto representa um avanço no projeto de juntas tubulares, uma vez que evita a necessidade de se conhecer uma expressão matemática para representar a distribuição de tensões na junta e fornece um método mais preciso para avaliar a distribuição de tensões ao longo da junta soldada. O conjunto de dados utilizado foi formado a partir de simulações numéricas das juntas soldadas através do MEF, nas quais foi considerada a geometria do cordão de solda.
Title in English
Artificial neural networks to calculate stress concentration factors in welded tubular joints.
Keywords in English
ANN
artificial neural networks
SCF
stress concentration factor
tubular joints
Abstract in English
An alternative approach to calculate stress concentration factors (SCF) in Y-type welded tubular joints is presented. Artificial Neural Networks (ANN) were used to represent the stress distribution along the tubular joints in both in-plane axial force and in-plane bending moment load cases. ANN can learn from a database without establishing a mathematical expression between dependent and independent variables, which is an advantage over the usual parametric equations approach. The proposed model represents an improvement in the tubular joints design, since it avoids the previous knowing of a mathematical expression to represent the stress distribution in the joint and provides an accurate method to evaluate the stress distribution along the welded fillet joint. The database herein used was completed with FE simulations of tubular joints which consider the geometry of the weld fillet.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
tese_ademar.pdf (1.25 Mbytes)
Publishing Date
2004-03-29
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.