Variability of dynamic characteristics of recycled aggregate concrete with Bayesian inference

Considering the limit research and data deficiency of dynamic characteristics of RAC, the variability of dynamic characteristics of RAC from Bayesian updating perspective is of much value. This paper presents an analysis of variability of dynamic characteristics of RAC based on Bayesian approach at the material level. Firstly, for both NAC and RAC, 27 specimens were tested for statistical distribution of dynamic properties by the forced resonance method. The statistical parameters and appropriate probability density functions of the dynamic characteristics were assessed with classical statistical inference. Then, based on Bayesian updating, probabilistic distribution models for dynamic characteristics were proposed with vague-informative priors. Finally, the estimation models of fundamental frequency and damping ratio of RAC based on the Bayesian linear regression were presented. It is found that the variability of both the fundamental frequency and damping ratio of RAC are 12% and 16.2% times higher than that of NAC from classical statistical analysis. The normal distribution and lognormal distribution are the most suitable to describe the dynamic characteristics for NAC and RAC. The variability of dynamic characteristics is also a random variable with normal distribution. The standardized fundamental frequency is positively correlated with generalized modulus of elasticity, while standardized damping ratio is negatively correlated with standardized fundamental frequency.