Documents disponibles écrits par cet auteur

Fillet weld groups loaded with out-of-plane eccentricity / A. M. Kanvinde in Journal of structural engineering, Vol. 139 N° 3 (Mars 2013) 

Public ISBD [article]  in Journal of structural engineering > Vol. 139 N° 3 (Mars 2013) . - pp. 305–319 Titre : Fillet weld groups loaded with out-of-plane eccentricity : simulations and new model for strength characterization Type de document : texte imprimé Auteurs : A. M. Kanvinde, Auteur ; J. Liu, Auteur ; X. Fu, Auteur Année de publication : 2013 Article en page(s) : pp. 305–319 Note générale : structural engineering Langues : Anglais (eng) Mots-clés : welding;  welded  connections;  finite  element  method;  simulation Résumé : The strength of fillet weld groups, loaded with an out-of-plane eccentricity, is controlled by complex interactions of weld yielding, as well as bearing between the connected parts. Current models that characterize connection strength, including those used in North American design specifications, are highly conservative, leading to oversized welds. These models are phenomenological, because the internal stress distribution within the welds is difficult to characterize experimentally. A new model is proposed for characterizing the strength of these connections. The model is based on insights developed from sophisticated finite-element (FE) simulations that feature accurate measurements of weld profiles, multiaxial plasticity, and simulation of contact and gapping phenomena that strongly influence connection response. The FE simulations reveal that current models do not reflect key aspects of force transfer within the connection, especially on the compression side. The proposed model incorporates these insights by using stress profiles and mechanisms consistent with those implied by the FE simulations. The model is evaluated against 79 experiments from three test programs. It is determined that the new model greatly reduces the conservatism of the existing models, resulting in an average test-to-predicted ratio of 1.01. This is in contrast to previous models, for which the average test-to-predicted ratios are in the range of 1.33–1.77. The efficacy of the proposed model is analyzed with respect to various parameters, and its limitations are outlined. En ligne : http://ascelibrary.org/doi/abs/10.1061/%28ASCE%29ST.1943-541X.0000641 [article] Fillet weld groups loaded with out-of-plane eccentricity : simulations and new model for strength characterization [texte imprimé] / A. M. Kanvinde, Auteur ; J. Liu, Auteur ; X. Fu, Auteur . - 2013 . - pp. 305–319. structural engineering Langues : Anglais (eng) in Journal of structural engineering > Vol. 139 N° 3 (Mars 2013) . - pp. 305–319 Mots-clés : welding;  welded  connections;  finite  element  method;  simulation Résumé : The strength of fillet weld groups, loaded with an out-of-plane eccentricity, is controlled by complex interactions of weld yielding, as well as bearing between the connected parts. Current models that characterize connection strength, including those used in North American design specifications, are highly conservative, leading to oversized welds. These models are phenomenological, because the internal stress distribution within the welds is difficult to characterize experimentally. A new model is proposed for characterizing the strength of these connections. The model is based on insights developed from sophisticated finite-element (FE) simulations that feature accurate measurements of weld profiles, multiaxial plasticity, and simulation of contact and gapping phenomena that strongly influence connection response. The FE simulations reveal that current models do not reflect key aspects of force transfer within the connection, especially on the compression side. The proposed model incorporates these insights by using stress profiles and mechanisms consistent with those implied by the FE simulations. The model is evaluated against 79 experiments from three test programs. It is determined that the new model greatly reduces the conservatism of the existing models, resulting in an average test-to-predicted ratio of 1.01. This is in contrast to previous models, for which the average test-to-predicted ratios are in the range of 1.33–1.77. The efficacy of the proposed model is analyzed with respect to various parameters, and its limitations are outlined. En ligne : http://ascelibrary.org/doi/abs/10.1061/%28ASCE%29ST.1943-541X.0000641