Correlation of Damage With Seismic Intensity Measure for Ductile Concrete Bridge Columns in British Columbia

Abstract eng:
This study investigates the correlation of a number of ground motion intensity measures with the predicted seismic damage of ductile concrete bridges in British Columbia, Canada. The spectral acceleration at the fundamental period is recommended by bridge codes as the primary intensity measure for selecting and scaling of ground motions records. The suitability of this approach to estimate with confidence the expected damage in bridge columns is investigated in this paper. Three concrete bridge columns with periods of 0.5 s, 1.0 s, and 2.0 s were designed and detailed according to the 2014 Canadian Highway Bridge Design Code (CSA S6-14). Detailed models of the bridge columns with nonlinear displacementbased elements and fiber sections were developed. The columns were subjected to 30 ground motions records, including 10 records from each crustal, subcrustal, and subduction earthquakes, reflecting the complex seismicity of British Columbia. Following the code recommendations, the motions had been selected and linearly scaled individually for each column to match the target spectrum within a period range. Nonlinear time-history analysis was conducted for the 30 records scaled at six hazard levels of 50%, 10%, 5%, 2%, 1%, and 0.5% in 50 years probability of exceedance. To predict the damage to the columns, the maximum drift ratio was employed as the damage indicator. Six discrete damage states were considered for the columns, including minimal damage, yielding of the longitudinal reinforcement bars, spalling of the cover concrete, serviceability limit state of the longitudinal reinforcement bars, crushing of the core concrete, and fracture of the longitudinal reinforcement bars. The analysis results of the 30 records for maximum drift ratios were plotted against each of the candidate intensity measures for all the six hazard levels, and linear correlation of the response with the intensity measures was evaluated. It was observed that among the investigated intensity measures, PGV, and PGD had the strongest positive correlation with the response parameter. Based on this correlation, smaller suites of ground motion records with one third and half of the records in the original suite of 30 records were selected. Predictions for the mean maximum drift ratio and the corresponding damage in the columns were made based on the smaller suites of records, and the results were compared to those from the original suite of records. It was found that the predictions in most cases are reasonably accurate. In some cases the smaller suites may overestimate the response slightly. Employing a smaller number of records for timehistory analysis is advantageous for implementing a performance-based design, where a large number of trial and error is required to achieve multiple performance objectives at multiple hazard levels simultaneously. Therefore, the findings of this study could help addressing one of the major challenges ahead of implementing the CSA S6-14 performance-based design approach in practice.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket