STUDY ON PROBABILISTIC SHEAR STRENGTH MODEL FOR DEEP FLEXURAL MEMBER

Bayesian theory is widely used in various fields, for its advantages that not only the model information and data information, but also transcendental information is used adequately. A shear model for reinforced concrete deep flexural members based on influence parameters is developed by using Bayesian multivariate statistical theory and 271 test results collected. The model is simplified through a removal process of Bayesian parameters. The performance of the developed probabilistic model is confirmed by comparison with the shear strengths predicted by the shear models in codes of GB 50010-2010, ACI318-08, CSA and EC2. Good agreements between the shear strengths obtained by the simplified model is achieved, based on Bayesian and test results. And the results updating Bayesian dynamics are closer to the tests than the codes. The simplified model can be used in calculating shear strength for deep flexural members.