Realistic behavior of infilled steel frames in seismic events: experimental and analytical study

Abstract

An experimental and analytical study is carried out to investigate the effects of lateral loading type on the behavior of masonry infilled steel frames. During earthquake the lateral load is applied as distributed loading to the top beams and columns through rigid floors; however, in most available experimental studies the lateral loading is applied as concentrated loading. In this study, two identical specimens are tested and their behavior is compared under distributed and concentrated lateral loadings. Finite element models of the specimens are also developed and validated against the experimental results. To have a better view, the influence of loading type is studied on another experimental specimen having concentrated loading. A parametric study is also conducted on the influence of loading type in multi-span frames and infills with different aspect ratios. The obtained experimental results show that the distributed loading results in 18.5 and 29% increase in the strength and stiffness of Infilled frames, compared to the case with concentrated loading. Less strength and stiffness in specimens subjected to concentrated loading is a result of stress concentration at the infill corner near the loading point which leads to premature corner crushing. Therefore, it is believed that the codes’ formulas, mostly based on specimens with concentrated loading, are conservative, underestimating the real ultimate strength and stiffness of masonry infilled steel frames.

Appendix

See Fig. 25.

Fig. 25

Simplified micro modeling procedure. (1) Based on a recommendation by Choudhury et al. (2015) the values of 10° is chosen for the dilation angle. The value of 0.15 is used for Poisson’s ratio (Lourenco 1996). (2) In accordance with a study by Bolhassani et al. (2015) the behavior of prism specimens is determined with material test for using as compressive behavior in FE software. Then the average of parameters including stiffness, maximum strength and its corresponding strain for different specimens is obtained. Using the recommendation by Agnihotri et al. (2013) the Kashik et al. (2007) model was employed to reach a general behavior of the masonry specimens and this behavior was defined in FE software. (3) Using the results of a study by Lumantarna et al. (2012), tensile and shear  strength of interface are determined as 0.031 and 0.055 of compressive strength of mortar. The friction coefficient was considered equal to 0.75 based on the study by Angelillo et al. (2014). (4) The interface damage starts when the quadratic function equals one (Nazir and Dhanasekar 2013). Using the recommended values by Angelillo et al. (2014) $G_{IC}$ and the ductility index for shear fracture energy ($\frac{{G_{IIC} }}{{t_{s}^{0} }}$) are considered equal to 0.012 N/mm and 0.093 mm. (5) The bilinear curve shows the cohesive behavior and its softening in different conditions (Bolhassani et al. 2015). Power model by Tao (2013) is employed to model decohesion under mixed mode loading. (6) In accordance with results of study by Kaushik et al. (2007) $E_{b}$ and $E_{m}$ are equal to 300 times compressive strength of brick ($f_{b}$) and 200 times compressive strength of mortar ($f_{m}$), respectively. (7) To define the cohesive behavior of the interface normal stiffness ($K_{N}$) and shear stiffness ($K_{s}$) should be determined using the formulas by Lourenco et al. (1996). (8) The mortar layer is not modeled and half of the mortar joint thickness is added to the adjacent brick layers. Instead, the cohesive behavior of mortar in three conditions of pure tensile, pure shear and mix mode is assigned to the bricks interface as shown in the chart. As mentioned the friction coefficient is assumed to be 0.75 (Angelillo et al. 2014)