UBC Theses and Dissertations
Probabilistic drift capacity models for reinforced concrete columns Zhu, Ling
The need to ascertain the drift capacity of reinforced concrete columns frequently arises in the displacement-based seismic design and assessment method. Drift capacity at column lateral strength degradation, as well as the column capacity to sustain axial load after the loss of lateral strength, is sought for each column. Such predictions can only be made in a probabilistic sense, due to lack of experimental data, model imperfections and variability in the intervening parameters. The probabilistic models developed in this study incorporate these sources of uncertainty and are capable of estimating the probability of column failure for a given drift demand. The probabilistic models not only provide engineers the prediction of column drift capacities but also help engineers to determine the critical columns and the priority of column retrofit. The prediction of column response is also one of the objectives in this study. Two approaches are proposed to classify columns. First, a probabilistic failure mode index model is developed to identify column failure modes: flexure failure, flexure-shear failure and shear failure. Second, a two-zone column classification method is proposed to approximately distinguish the shear response dominated columns (Zone S) and the flexural response dominated columns (Zone F) based on three column parameters. Two probabilistic drift capacity models at 20% reduction in lateral strength are developed; one for columns classified in Zone S, and another for columns in Zone F. The key parameters affecting the drift capacity at 80% lateral strength for reinforced concrete columns are identified through the assessment of probabilistic drift capacity models. Based on the shear-friction mechanism, a probabilistic drift capacity model at axial failure is developed for shear-damaged columns. Several examples are presented to demonstrate the applications of the probabilistic models. The probabilistic failure mode index model can be employed to assess the probability of each failure mode for a given column. The probabilistic drift capacity models can be used to construct the fragility estimate of reinforced concrete columns with respect to the drift demand. The variability in the fragility estimate due to epistemic uncertainties is captured by confidence bounds at specified probability levels. The assessment of failure mode probability, as well as the fragility estimate, for a column damaged during the Northridge Earthquake is presented. Other results derived from the probabilistic capacity models include the assessment of numerical acceptance criteria for column drift in FEMA 356, the fragility estimates for a range of column properties, and the application of normalized fragility curves. An additional advantage of the probabilistic drift capacity models is the potential incorporation in structural reliability analysis to assess the probability of global structural system collapse.
Item Citations and Data