Repeatability and Parasitic Out‐of‐Plane Motion in Controlled Rocking Structures With Cylindrical Columns: Results From Triaxial Shake Table Tests

ABSTRACT Self‐centering rocking structures outperform conventional ones since they show zero post‐earthquake damage and excellent recentering; however, their applications remain limited due to concerns that rocking is extremely sensitive and nonrepeatable. This paper investigates the repeatability and parasitic out‐of‐plane motion of controlled rocking structures with cylindrical columns through large‐scale triaxial shake table tests. The specimen comprised four slender reinforced concrete columns restrained with unbonded steel tendons that provided positive post‐uplift stiffness, representing a bridge‐like structure at 1:5 scale. A total of 16 shake table tests were performed: five pairs of nominally identical triaxial tests to assess repeatability and six tests with varying directional excitation to quantify out‐of‐plane motion. The ground motions induced large displacements. Results show high repeatability of the shake table response. Displacements showed correlation coefficients of 0.85 for entire time histories. Maximum responses, which govern design, showed near‐perfect repeatability with correlation coefficients of 0.90 and errors below 10%. No systematic relationship between response amplitude and repeatability was observed. Under planar (one‐directional) seismic excitation, parasitic out‐of‐plane motion was low compared to uncontrolled rocking systems, with average out‐of‐plane to in‐plane displacement ratios below 25%. Independent planar analyses, when geometrically combined, can estimate the bidirectional response with reasonable accuracy. The results contradict the perception that rocking behavior is nonrepeatable and show that controlled rocking structures with inherent frictional and impact damping can exhibit repeatable seismic response even without supplemental energy dissipation.