Beam splices are typically located at moment contraflexure points (where M=0). Most design specifications require these splices to develop a strength either to meet design forces or a minimum value set by specifications. The design forces are typically determined through elastic analysis, which does not include flexibility of splice connections. In this research, two types of splice connections, an extended end plate splice connection and a flange and web plate bolted splice connection, were tested and analyzed to investigate the effect of the partial strength splice connections on structural response. The splices were designed to resist 40% and 34% of connecting section capacities using current steel design codes, respectively. It has been observed from the experiments and FEM analysis results that splice connections designed under capacities of connecting steel members can result in changes in design moment diagrams obtained from analyses without splice connection simulation and can also significantly decrease the rigidity of the structure endangering serviceability. The differences in design moment diagrams can go up to 50 % of elastic analysis without connection flexibility. The vertical displacements can increase to 155% of values obtained from elastic analysis with no splice connection simulation. Therefore, connection flexibility becomes very important to define in analysis.

Este trabalho apresenta um estudo do comportamento estrutural das ligações metálicas parafusadas viga-pilar com chapa de topo estendida. Enfatiza-se a fundamentação teórica, abordando os principais conceitos, os métodos de dimensionamento apresentados em normas internacionais e os principais sistemas de classificação de ligações existentes. O objetivo deste trabalho é apresentar uma metodologia capaz de aprimorar a análise numérica de ligações, com relação aos estudos anteriores, obtendo resultados com melhor representatividade do comportamento real das ligações. Desenvolve-se análise numérica, via Método dos Elementos Finitos (MEF), com a elaboração de diversos modelos numéricos tridimensionais de ligações parafusadas com chapa de topo estendida. Utilizou-se na modelagem numérica o código de pré-processamento TrueGrid® e o código de análise numérica ANSYS®. Os modelos foram desenvolvidos e validados a partir de resultados experimentais. Na análise dos resultados foram comparados os resultados numéricos obtidos com resultados experimentais previamente conhecidos. Faz-se uma abordagem das propostas de dimensionamento apresentadas por normas internacionais em comparação com os resultados numéricos. Através da variação de parâmetros, como espessura da chapa e diâmetro dos parafusos, observam-se os modos de falha das ligações, e a plastificação da chapa de topo e dos parafusos. Por fim, apresenta-se a conclusão do trabalho e sugestões para pesquisas futuras. / This work presents a study about the structural behavior of extended end-plate bolted beam-to-column steel connections. Emphasizes the theoretical foundation, addressing the main concepts, the design methods presented in international standards and the main classification systems for connections existing. The objective of this work is to present a methodology capable of improving the numerical analysis of connections, in comparison with previous studies, getting results with better representation of the actual behavior of the connections. It develops numerical analysis via Finite Element Method (FEM) with the development of various numerical models of three-dimensional bolted end plate connections. It was used in numerical modeling the preprocessing code TrueGrid® and the numerical analysis code ANSYS®. The models were developed and validated from experimental results. In the analysis of the results were compared the numerical results obtained with experimental results previously known. It approaches the design proposals presented by international standards in comparison with the numerical results. By varying the parameters such as end-plate thickness and bolts diameter, it was observed failure modes of the connections, as well as yielding mechanism at the end-plate and bolts. Finally, it presents the conclusion of the work and suggestions for future researches.

<p>У овом раду спроведене су опсежна експериментална и нумеричка<br />испитивања понашања веза са чеоном плочом и четири завртња у реду,<br />као и карактеристичне компоненте ових веза - Т-елемента.<br />Експериментална испитивања укључују тестирања 36 узорака<br />Т-елемената у природној величини, као и четири везе греда-стуб са<br />чеоном плочом и четири завртња у реду, такође у природној величини.<br />Нумеричке анализе, спроведене у софтверском пакету ABAQUS, су<br />обухватале све типове нелинеарности, али и отказ материјала, те се као<br />такве сврставају међу ретке којима су слични проблеми нумерички<br />испитивани. Параметарска студија обухвата 33 различите конфигурације<br />Т-елемената са четири завртња и представља највећу такву студију<br />изведену за ову компоненту. Као основни још увек нерешен проблем,<br />круцијалан за понашање Т-елемената, препознати су положај сила услед<br />ефеката полуге, али и савијање завтњева. Приказан је и предложен нови<br />аналитички модел прорачуна носивости Т-елемената са четири завртња,<br />и квантификована његова предност над до сада постојећим моделима.</p> / <p>U ovom radu sprovedene su opsežna eksperimentalna i numerička<br />ispitivanja ponašanja veza sa čeonom pločom i četiri zavrtnja u redu,<br />kao i karakteristične komponente ovih veza - T-elementa.<br />Eksperimentalna ispitivanja uključuju testiranja 36 uzoraka<br />T-elemenata u prirodnoj veličini, kao i četiri veze greda-stub sa<br />čeonom pločom i četiri zavrtnja u redu, takođe u prirodnoj veličini.<br />Numeričke analize, sprovedene u softverskom paketu ABAQUS, su<br />obuhvatale sve tipove nelinearnosti, ali i otkaz materijala, te se kao<br />takve svrstavaju među retke kojima su slični problemi numerički<br />ispitivani. Parametarska studija obuhvata 33 različite konfiguracije<br />T-elemenata sa četiri zavrtnja i predstavlja najveću takvu studiju<br />izvedenu za ovu komponentu. Kao osnovni još uvek nerešen problem,<br />krucijalan za ponašanje T-elemenata, prepoznati su položaj sila usled<br />efekata poluge, ali i savijanje zavtnjeva. Prikazan je i predložen novi<br />analitički model proračuna nosivosti T-elemenata sa četiri zavrtnja,<br />i kvantifikovana njegova prednost nad do sada postojećim modelima.</p> / <p>In this research, extensive experimental and numerical investigations of<br />behaviour of end-plate moment connection with four bolts per row are<br />performed, as well as specific component of the connection &ndash; T-stub element.<br />Experimental research comprises tests on 36 unscaled specimens of T-stub<br />elements and 4 specimens of beam-to-column joints with end-plate connection<br />and four bolts per row, also in full-size. Numerical analyses, performed with<br />software ABAQUS, included all types of nonlinearities, and material<br />deterioration, hence they can be considered as one of the few most<br />sophisticated analyses used for investigation of the considered problem. The<br />parametric study comprises 33 different configurations of the T-stub element<br />with four bolts per row and it presents the biggest study of this kind for this type<br />of component. Bending of the bolts and position of the prying forces are<br />recognized as the largest, still unsolved problem, crucial for the behavior of the<br />T-stub elements. A new theoretical model for the resistance of the T-stub<br />element with four bolts per row is proposed and its advantage over the existing<br />models is presented and quantified.</p>