Can mechanics-based simple masonry constitutive models help simulate infilled RC frames' response?

A commonly used building typology today is the unreinforced masonry-infilled reinforced concrete (RC) frames. Masonry is used primarily as a non-structural component (e.g., partition). The frames are responsible for transmitting the vertical loads safely to the foundation. Given the complex behaviour of masonry panels and their interaction with the surrounding frame (when used as an infill), design standards often ignore the contribution of infill. The masonry panel is generally represented using one or more struts spanning along or around its diagonal. In most cases, the constitutive behavior for these struts is defined through a mechanics-based functional form calibrated using experimental data. A recent study has also suggested functional forms based on extensive statistical analyses. This paper considers seven and one functional forms, respectively, belonging to each of the two categories. These constitutive models were integrated with one-strut, two-strut and three-strut models with suitable assumptions on the geometrical and material parameters. Combinations of strut and constitutive models were used to develop the numerical models for 106 experimentally studied one-bay-one-story specimens. A pushover analysis was performed for each of these 2,544 models. It was found that the mechanics-based definitions for the parameters of the constitutive model can most effectively capture the global lateral force-displacement behaviour of the infilled RC frames, if effect of the masonry prism strength was suitably accounted for. For example, for infill strength 0-10 MPa, the peak strength of the masonry strut defined as the product of the masonry prism strength and the area of the strut led to simulated peak strength of infilled RC frames closest to the experimental observations on an average, and in terms of uncertainty in the results. Similarly, the initial stiffness of masonry panel was simulated in line with experimental results using the product of masonry elastic modulus and the strut area divided by the diagonal length of the strut. Flexure failure in frames was captured well regardless of the choice of strut or constitutive model. Shear failure in frames was found sensitive to the choice of both strut and constitutive models. Single strut model was found inadequate in capturing the shear failure in frames. Further, performance of different models vis-à-vis the properties of experimental specimens (e.g., masonry prism strength) is discussed in the paper.