The coupled horizontal -vertical response of elastomeric and lead-rubber seismic isolation bearings
Elastomeric and lead-rubber bearings are two types of seismic isolation hardware widely implemented in buildings, bridges and other infrastructure in the United States and around the world. These bearings consist of a number of elastomeric (rubber) layers bonded to intermediate steel (shim) plates. The total thickness of rubber controls the low horizontal stiffness and the close spacing of the intermediate shims provides a large vertical stiffness for a given bonded rubber area and elastomer shear modulus. Conceptually, a lead-rubber bearing differs from an elastomeric bearing only through the addition of a lead-core typically located in a central hole. During earthquake ground shaking, the low horizontal stiffness of elastomeric and lead-rubber bearings translates into large lateral displacements, typically on the order of 100--200% rubber shear strain, that might lead to significant reductions in the axial load carrying capacity and vertical stiffness of the individual bearings.
This dissertation presents an analytical and experimental investigation of the coupled horizontal-vertical response of elastomeric and lead-rubber bearings focusing on the influence of lateral displacement on the vertical stiffness. Component testing was performed with reduced scale low-damping rubber (LDR) and lead-rubber (LR) bearings to determine the vertical stiffness at various lateral offsets. The numerical studies included finite element (FE) analysis of the reduced scale LDR bearing. The results of the experimental and FE investigations were used to evaluate three analytical formulations to predict the vertical stiffness at a given lateral displacement. From component testing the vertical stiffness of the LDR and LR bearings was shown to decrease with increasing lateral displacement and at a lateral displacement equivalent to 150% rubber shear strain a 40--50% reduction in vertical stiffness was observed. One of the three analytical formulations, based on the Koh-Kelly two-spring model, was shown to predicted the measured reduction in vertical stiffness of the LDR and LR bearings at each lateral offset with reasonable accuracy. In addition, earthquake simulation testing was performed to investigate the coupled horizontal-vertical response of a bridge model isolated with either LDR or LR bearings. The results of simulations performed with three components of excitation were used to evaluate an equivalent linear static (ELS) procedure for the estimation of the vertical load due to the vertical ground shaking. The equivalent linear static procedure was shown to conservatively estimate measured maximum vertical loads due to the vertical component of excitation for most simulations.