Scaling errors of a seismic isolation system with a shear key

Abstract A shear key is installed on a seismic isolation bearing to resist long-term service and small earthquake loads. In the event of a large earthquake, the shear key is cut off to allow for the sliding of a bearing that isolates and dissipates the earthquake's energy. A scale in the bearing will satisfy the shaking table test's demands, however, the bearing will undergo disturbance when cutting off the shear key, including friction variability, gravity distortion, spring and viscous damping. In order to solve the problem prior to a real shaking table test, a simplified single-degree-of-freedom isolation system (SDOFS) can be numerically scaled. In this case, seismic responses are generated from scaled models and then compared with those of the prototype model. Response errors found during comparison demonstrate that friction variability and gravity distortion lead to significant inaccuracies in the scaled models. However, acceleration and displacement errors can be reduced by increasing the strength of the shear key. Moreover, such acceleration and relative displacement errors can also be reduced by increasing peak ground acceleration (PGA). These conclusions have been validated through the adoption of a real shaking table test.