A Effective Solution Method Based on Bayesian Rule for Prediction of Long-term Tensile Strength of GFRP Bars in Concrete Beams

摘要

It is necessary to research on the long term structural behavior of GFRP bars in concrete by limited samples. The two models, Arrhenius Equation and Fick's Law were compared to form a long-term durability prediction thinking. Based on the work and ideas of the prior researchers, a detailed derivation of the iterative equation on the Bayesian prediction was conducted. Especially, a new likelihood function was set up to solve the long term behavior prediction of GFRP bars in concrete with less information. Then a new and effective solution method was developed. The residual strength behavior of GFRP bars in concrete beams after seven years' environmental exposure was analyzed. The results showed that the theoretical predictions and experimental data were very close and theoretical prediction model was reasonable.