Effet de la classification des sols sur la demande sismique des piles de ponts sous l'effet des séismes bi-directiononnels

Abstract

A study has been conducted to examine the effect of the soil class on the ability of the 30% combination rule to adequately predict the seismic demand and its impact on the required longitudinal reinforcement in bridge piers subjected to bidirectional earthquake ground motions. The study consists in a series of linear spectral analyses and multiple time-history analyses on a regular bridge model, typical of those encountered the most in eastern canadian highways. The study was conducted for a moderate seismic hazard (Montreal, QC) using a set of five (5) pairs of recorded historical ground motion earthquakes. In the first part of the study, the exact critical seismic demand, in terms of interacting moments (Mx, My) at the base of the bridge columns, was determined for each soil class by performing a series of multiple time-history analyses under pairs of orthogonal seismic ground motion time-histories. Response spectrum analyses were also carried out, using the acceleration spectra of the orthogonal ground motion components, to determine the maximum bi-axial response in each direction and the critical bi-axial response was estimated using the 30%-Rule. In the second part of the study, the column longitudinal steel ratio required from response spectrum analysis with the 30% combination rule was computed for each soil class and compared to the ratio determined from the results of multiple linear dynamic time history analysis of the bridge structures under bidirectional ground motions. Comparison of the bi-axial moments (Mx, My), calculated at the base of the bridge columns, and the corresponding required longitudinal reinforcement ratios show that 30% combination rule as specified by the seismic codes do not adequately estimate the bidirectional seismic demand and the required reinforcement ratio. The error limits obtained were deemed too high. The results also show that the combination rule is highly dependent on the soil class. More specifical.y, the error in estimating the seismic demand and the required longitudinal reinforcement ratio increase significantly from a soil of class A (hard roc) to a soil of class E (very soft soil). These results indicate that the 30% combination rule can be much more improved by taking into account the classification of the soil.

Date	20 Sept 2013
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Amar Khaled (Supervisor)