Title:
Probabilistic Prediction Model for Flexure-Shear Capacity of Reinforced Concrete Girders
Author(s):
Xinmin Zhang, Chaoyuan Wu, Zengwei Guo, Fanxiang Xia, and Xianhu Ruan
Publication:
Structural Journal
Volume:
122
Issue:
4
Appears on pages(s):
113-124
Keywords:
Bayesian estimate; Markov chain Monte Carlo (MCMC); probability model; reinforced concrete (RC) member; shear capacity
DOI:
10.14359/51745491
Date:
7/1/2025
Abstract:
It is well known that the estimates of most shear capacity prediction
models for reinforced concrete (RC) components are of high dispersion due to their elaborate failure mechanisms and elusive material variability. A probability prediction model is more appropriate for estimating the shear capacity of RC members than a deterministic prediction model. Therefore, this study proposed a probabilistic model to evaluate the shear capacity of RC T-beams and employed a Bayesian-Markov chain Monte Carlo (MCMC) approach to determine the posterior parameter in the shear strength prediction
model by Bayesian updating. The analysis results indicate that the
probabilistic model achieves minimal variance, offering the most
accurate predictions that closely match test data compared with
other prediction models. The shear capacity of a T-beam increases
with changes in flange width and flange height ratio, but remains
constant once beyond a certain level. The shear capacity varies
rapidly when the shear-span ratio (λ) is less than 2.5 or larger
than 4.0 due to a notable shift in the failure mechanism. Besides,
the shear capacity raises linearly by increasing the characteristic
value of stirrups (ρvfyv).