Search

[Display omitted] • Experimental and theoretical research on the fire resistance of reinforced concrete beams is summarized. • The important parameters are selected by parameter correlation analysis. • The fire resistance prediction equation is proposed by Machine Learning and regression analysis. Reinforced concrete (RC) beams are widely applied in building structures, and fire resistance performance becomes a focus of fire research with the high frequency of building fires. This paper aims to summarize the fire resistance performance of 216 specimens in 80 references and propose a fire resistance prediction equation. Firstly, the factors of fire resistance are considered in these specimens such as load ratio, concrete cover thickness and longitudinal reinforcement ratio. The influence rules of these factors are summarized. And the theoretical analysis, finite element simulation and fire resistance prediction equation of RC beams are also summarized. It is found that the parameters such as load ratio and concrete cover thickness have a great impact on fire resistance, while the parameters such as stirrup spacing have a minor impact on fire resistance. Secondly, the Pearson correlation analysis of 10 parameters is conducted in 216 specimens and 6 parameters with a correlation degree greater than 0.1 are identified as important parameters. The database is established by 4 Machine Learning (ML) algorithms. It is found that the database established by the Random Forest (RF) algorithm has high performance. Finally, the fire resistance prediction equation of RC beam is obtained by the database and nonlinear regression analysis. The prediction equation is compared with existing equations, and it is found that the proposed equation with R 2 of 0.935 has high precision in predicting fire resistance. The proposed equation can be adopted to guide the fire resistance design of RC beams. [ABSTRACT FROM AUTHOR]