Statistical Analyses of Strength of Slender RC Columns

Abstract

Probabilistic analyses of the modeling errors of several selected strength prediction models for slender reinforced concrete (RC) columns are carried out. The selected strength models include the moment magnifier methods that are recommended in the ACI and CSA design codes and the stability-based theoretical model. A relatively large amount of test data on slender RC columns is collected from the literature and the test results are compared with the ones obtained from different strength prediction models. Both normal- and high-strength concrete columns are included in this study. Probabilistic analyses of the modeling error include the use of pseudolikelihood estimation method. Analysis results suggest that the coefficient of variation of the modeling error for slender RC columns can be as high as 20%, which is considerably larger than those suggested and employed for reliability analysis in the literature. The results also suggest that the modeling error for slender RC columns depends on concrete compressive strength, the load eccentricity, and the slenderness ratio. However, the effect of the slenderness ratio on the modeling error is negligible. Sets of probabilistic models of the modeling errors by considering different strength models for slender RC columns are suggested.