Simulation of shake‐table test for a two‐story low‐damage concrete wall building

Abstract

Summary To support the development of low‐damage concrete structures, a system‐level shake‐table test of a two‐story concrete wall building implementing state‐of‐the‐art design concepts was conducted using the multi‐functional shake‐table array at Tongji University as part of an international collaborative project. The test building was designed with a perimeter frame and exterior post‐tensioned concrete walls in both directions. Different floor systems and wall‐to‐floor connections were incorporated in the test building to compare a number of design concepts and construction details. A range of energy dissipation devices were installed at the wall base and/or slotted‐beam joints of the test building. To simulate the test building response during the shake‐table tests, numerical models of the test building in both the longitudinal and transverse directions were established in OpenSees. Because the test building utilized flexible and isolated wall‐to‐floor connections that specifically reduced potential wall‐to‐floor interaction, planar frame numerical models were selected to represent the test building in each direction of loading. In the models, fiber hinge elements were used to simulate the unbonded post‐tensioned walls and a modified multi‐truss spring method was adopted to simulate the slotted‐beam joints with and without dampers. Inelastic time‐history analyses of the numerical models were conducted considering three earthquake intensities and comparisons of the test and simulation responses of the building are presented. The simulation results showed that the analytical model established in this study could reasonably predict both the global and local responses of the test building under different shaking intensities.

Planar frame models considering system‐level interactions of a two‐story low‐damage concrete wall building

“…The dynamic analysis setting was the same as that of the original planar frame models. 26 The total seismic weight of the test building, including the self-weight and additional mass, equaled 522 and 357 kN for L1 and L2, respectively. In the planar models that represented half of the building lateralresistance, half of the total seismic mass was accumulated at the wall node of each level.…”

“…Test building overview and floor systems in each level. 26 was designed primarily as a gravity frame in Design 2 (D2) and Design 3 (D3) configurations, while in the Design 1 (D1) configuration the perimeter frame also contributed to lateral resistance. The key difference of those design configurations is the presence of energy dissipation devices at the beam-column and beam-wall connections to provide moment reaction in D1, as opposed to acting more like a pinned connection in D2 and D3.…”

“…The slottedbeam joint modelling and the isolated connection modelling were validated by investigating the local response results. 26 In the test building, the gravity columns were connected to the foundation by unbonded PT bars that had zero initial prestress. 10 As a result, only minimal moment resistance was possible, so the boundary condition of column bases in the models was assumed to be pinned.…”

“…1,24 A fiber hinge model was applied in a numerical study of a 3D finite element model in SAP2000 for the E-defense test building with UPT walls. 25 Gu et al 26 had established planar frame models utilizing the fiber hinge approach to model the UPT walls for the lowdamage concrete wall building tested in 2019. In the planar frame models, the critical height of the nonlinear beam-column element at the wall base was 1.6 times the wall thickness t w , which is within the recommended range suggested by prior researchers (smaller than 2.5t w and twice the neutral axis depth), 4,16,25 and consistent with the neutral axis depth observed at the wall bases during testing.…”

“…The model results tended to underpredict the peak overturning moments for all loading cases, even when the simulated peak roof drift exceeded the test peak roof drift under the maximum considered earthquake (MCE) intensity. 26 In addition, peak roof drifts were underpredicted for loading cases under the design-based earthquake (DBE) intensity. The disparity between the simulation and test results indicated that the interactions between different structural components and the response of connections detailing during the system-level tests of building cannot be ignored in these planar frame models.…”