PSI - Issue 64

Austin Martins-Robalino et al. / Procedia Structural Integrity 64 (2024) 418–425 Martins-Robalino and Palermo / Structural Integrity Procedia 00 (2019) 000 – 000

422

5

Longitudinal reinforcement bars were modeled as discrete truss elements and defined to consider buckling, based on the formulations proposed by Dhakal & Maekawa (2002). Headed mechanical couplers were also defined as discrete truss elements with cross-sectional area of 962 mm 2 , d b of 35 mm, F y of 690 MPa, F u of 795 MPa, E of 200 GPa, ε sh of 40 mε, and ε u of 350 mε. To capture the lower bond inherent with the smooth surface of SE-SMA bars, implementation of link elements and bond models capable of accounting for slip were appropriate. For consistency in the models, for both Wall SWS and SWN all bar truss elements starting at a depth of 350 mm in the foundation to a height of 900 mm in the wall were defined as imperfect bond-link elements; defined as either having an “embedded deformed rebar” model for the steel reinforcement or a n “embedded smooth rebar” model for SE -SMA bars. Mesh discretization maintained that all concrete elements have a height of 50 mm while width of elements was kept at 50 mm where possible to allow for an aspect ratio of 1. There was variation in the element width between 40 mm and 60 mm along the length of the wall to accommodate for the concrete cover and reinforcing bar spacing in the boundary regions. Concrete elements in the foundation and cap beam sections beyond the length of the wall had widths of 100 mm to reduce the number of non-critical elements. The preliminary models (V1) for Wall SWS and SWN are illustrated in Fig. 3.

a) b) Fig. 3. Preliminary model of (a) Wall SWS; (b) Wall SWN.

The numerical lateral load-displacement responses for each wall are shown superimposed on experimental results in Fig. 4. The preliminary models predicted average peak loads similar to those from experimental testing. Wall SWS achieved experimental peak loads of 125 kN and -111 kN while the preliminary VecTor2 model reaches peak loads of 128 kN and -122 kN. A similar level of accuracy was captured in Wall SWN preliminary models which predicted peak loads of 126 kN and -123 kN compared to experimental peak loads of 116 kN and -121 kN. Examining the ultimate displacement, the preliminary models significantly overestimate the capacities of Wall SWS and Wall SWN. The numerical model of Wall SWS predicted failure during completion of the 108 mm displacement cycle while experimentally the wall failed during the 72 mm displacement cycles. Likewise, the numerical model of Wall SWN predicted failure during the completion of the 168 mm displacement cycle while during testing the wall failed during the 132 mm displacement cycle in the positive direction and during the 108 mm displacement cycle in the negative direction. Note that the asymmetrical failure of Wall SWN was in part due to issues of sliding during experimental testing. Experimentally, both walls failed due to rupture of the steel reinforcement along the base of the wall; boundary region steel reinforcement bars bucked and ruptured in Wall SWS while the steel reinforcement in the web of Wall SWN ruptured and buckled with no notable damage to the SE-SMA bars. Both preliminary models failed to capture