EXPERIMENTAL STUDY OF LOW DAMAGE SHAPE MEMORY ALLOY BRIDGE COLUMNS IN LONG-DURATION EARTHQUAKES

Amos Jesse Black

doi:10.7273/000007324

Modern bridge design relies on energy dissipation through the formation of a plastic hinge at the base of columns. Low damage bridge columns that remain functional after earthquakes have been previously studied, with some studies incorporating advanced materials such as superelastic shape memory alloys (SMA) that provide column recentering. Previous research on columns with SMA’s has often used alternative materials, such as engineered cementitious composite, in combination with SMA rather than conventional concrete. In this study, the concept of low damage seismic response in concrete columns was examined for the combination of nitinol longitudinal reinforcement and conventional concrete with steel or FRP jackets. The nitinol is a shape memory allow that provide re-centering, and the jackets provided confinement of concrete to prevent crushing and buckling of longitudinal reinforcement. The combination of nitinol and conventional concrete with jackets has not been explored in previous experimental studies. An experimental study was conducted for three columns with SMA longitudinal reinforcement and jackets at the base. Characteristics of bridge columns built after 1976 in Washington State were compiled based on several design properties, including but not limited to ratio of longitudinal reinforcement, compressive strength of concrete, and longitudinal bar size. The characteristics were used to inform the parameters selected for the tests columns. The cantilever test columns were nominally identical, with 24” diameter, 96” height to lateral loading, and a longitudinal reinforcement ratio of 1.35% The longitudinal reinforcement consisted of conventional reinforcement mechanically spliced to 11.25” lengths of nitinol centered at the base of the column. The primary test variables were the loading protocol and the jacket thickness and type. Two of the three half-scale test columns were tested with a fully reversed cyclic protocol with increasing amplitudes, while the third was tested with a modification of this protocol. The modified protocol was characterized by a large excursion to 5.0% drift after the completion of force-controlled cycles, followed by decreasing cycle amplitude before the resumption of progressively increasing cycle levels starting beyond 5.0% drift. The fully-reversed cyclic loading protocols used in the tests had more cycles than those used in previous tests on columns with SMA reinforcement, and this was intended to better simulate the effects of a long duration earthquake on a system using superelastic SMA. The confining jackets were either 3/16” thick steel or 0.32” thick FRP. The columns were tested quasi-statically as cantilevers, with a roller and pin used at each end of an axial load assembly to properly simulate the P-delta effect. The tested columns exhibited the desired low-damage behavior, characterized by no observed crushing or spalling of concrete and recentering to within 1.0% drift of center until first bar fracture. First bar fractures typically occurred between 5.0% (4.0 [Delta]) and 6.25% (5.0 [Delta]) drift and were the primary cause of strength degradation. The variation in loading protocols was found to have a minor effect on the overall load deformation response. Peak lateral load resistance and recentering characteristics had minimal sensitivity to the variation in jacket type and thickness. Based on the overall favorable performance of the test columns, it was concluded that the use of nitinol and conventional concrete with confining jackets is a viable alternative to previous approaches used to achieve low-damage, recentering seismic response, such as those that have used nitinol and engineered cementitious composite (ECC). For the test columns, approximately 70% of lateral drift at cycle peaks came from rotation at the base. A hinge-rotation model was formulated to model the columns. This model consisted of an elastic element with a spring at the base. Calculated moment-curvature was used for the relationship between moment and curvature, with the hinge rotation obtained as the product of curvature and the plastic hinge length, which was taken as the 11.25” length of smooth nitinol centered at the base of the column. The model provided a suitable fit for the test column with the near-monotonic protocol to 5.0% drift. For the other columns tested, with progressively increasing increments to 5.0% drift, the model predicted an upper bound capacity for the seismic responses.

EXPERIMENTAL STUDY OF LOW DAMAGE SHAPE MEMORY ALLOY BRIDGE COLUMNS IN LONG-DURATION EARTHQUAKES

Files and links (1)

Abstract

Metrics

Details