Response of seismically isolated buildings with buffers subjected to near-source ground motions and possible alternative isolation systems

Abstract

The response of a seismically isolated building with lead rubber bearings (LRB) to near source ground motions from large earthquakes was investigated. The building was assumed to have a buffer to limit the maximum bearing displacement in a rare event of large magnitude and the buffer gap was assumed to be only 150mm (the level of maximum isolator displacement used in the 1980s). The structure was assumed to be designed for 1.5 times the NS component of the 1940 El Centro record. The 15% damped (an amount of damping which is close to the equivalent damping ratio for an seismically isolated building at its isolator design displacement) displacement spectrum of the design motion is only 40% that of the Sylmar County Hospital Parking Lot record from the 1994 Northridge earthquake (M_w= 6.7) in the period range around the first modal periods of both isolated and un-isolated structure used in the present study. Among the near-source records that are available, the near-source Sylmar record from the 1994 Northridge earthquake was found to have a very large displacement demand in a period range of 2.0 - 3.0s and this record is thought to be a better representation of the expected near-source motions than the 1.5 times the 1940 El Centro record.

Structure-buffer impact was found to impose very large inter-storey drifts and produce very large storey accelerations, when the building was subjected to the excitation of the Sylmar record. The structure-buffer impact was found to be detrimental to the structural response if the structure was not designed to provide inelastic deformation capacity, and the structural response did not improve when the gap was increased to 200-250 mm, the expected maximum displacement capacity of the LRBs used in the building. An alternative isolation system of LRBs and hysteretic dampers was investigated and found to be adequate for resisting near-source motions. A large initial damper stiffness and relatively small buffer stiffness (compared with the total initial stiffness of LRBs) were found to be effective in reducing inter-storey drifts and storey accelerations at floors except for the base and roof of the structure. A disadvantage of such a system is the relatively large base and roof accelerations. The system has relatively large inter-storey drifts and storey accelerations compared with an isolated structure using LRBs only when the structure was subjected to either the 1940 El Centro type ground motions or the Joshua Tree type ground motions with backward directivity effect. Such an isolation system would still enable the structure to respond essentially elastically under the excitation of the Sylmar record even though the isolated structure was designed for a much lower level of ground shaking.

As the upper structure of a seismically isolated building is usually designed to respond essentially elastically, the detailing used in the design of a reinforced concrete structure to provide inelastic deformation capacity was generally uncommon and was not fully accounted for in the present study.