Proposta de modelos para estimativa do fator de redução da resistência à compressão de pilares castelados carregados axialmente

Abstract

Currently, castellated columns have been increasingly used in buildings due to their flexibility, easy compatibility and potential savings given their lower steel consumption. However, there is a lack of research and design methods for these elements. The influence of the openings on the load capacity of castellated columns has not yet been clearly defined in the literature. Studies on castellated beams can be easily found, however, studies on such profiles acting as pillars are scarce. In this context, numerical models based on finite elements developed with the aid of ABAQUS® software are developed. This parametric study with 1328 numerical models aims to evaluate the influence of the three main types of castellated openings (Peiner, Litzka and Anglo-Saxon), three expansion factors (k - 1.33, 1.50 and 1.67) and eight slenderness indices (λ - 25, 50, 75, 100, 125, 150, 175 and 200) in the behavior of axially loaded columns. In order to validate the numerical model, the load capacity obtained through a nonlinear analysis was compared with the valuesfound in three experimental studies and in ABNT NBR 8800 (2008 After validation, the resistant capacity of each model was determined through a nonlinear physical and geometric analysis around the largest and least inertia axis. The loads obtained by the castellated and full-web column were compared. The resistance reduction factors associated with global buckling (χ) for castellated columns were also compared to the values proposed by ABNT NBR 8800 (2008). Castellated columns had a lower resistance capacity than the full-web elements, especially the short columns. Castellated columns with larger openings showed a greater reduction in resistance capacity. Furthermore, the curve proposed by ABNT NBR 8800 (2008) may overestimate the resistant capacity of castellated columns for reduced slenderness indices (λ0) lower than 0.75. Finally, equations based on regression models were developed (with R2 values above 94.00) to estimate the resistance reduction factor associated with global buckling (χ) of the columns evaluated within the scope of this study.