In order to investigate the effects of different rubber replacement rates and steel fiber admixture on the long term stress performance of concrete,a load of 30％flexural strength was applied to concrete beams by bending creep experiments for 3 days,and the results showed that the addition of steel fiber could significantly improve the creep stiffness of rubber concrete.Meanwhile,based on the creep experi-mental data,a BP neural network model was established to predict the creep stiffness of steel fiber rubber concrete,and it was optimized by genetic ant colony algorithm(GA-ACO-BP),and the Mean absolute percentage error was reduced from 16.28％to 3.3％,which effectively improved the accuracy and stability of the model.
维普
万方数据