Prediction of creep of recycled aggregate concrete using back‐propagation neural network and support vector machine
Xian Rong, Yinbo Liu, Pang Chen, Xueyuan Lv, Chen Shen, Boqiang Yao
Abstract
Abstract The promotion of recycled aggregate concrete (RAC) can effectively reduce carbon emissions from the construction industry and increase the recycling rate of waste concrete. Owing to the significantly higher creep of RAC than that of natural aggregate concrete (NAC), the accurate prediction of long‐term RAC creep deformation is vital. Therefore, a predictive model for RAC is investigated in this study. A creep database of RAC with 106 groups of 1309 experimental data points considering 15 influencing parameters is established. Furthermore, back‐propagation neural network (BPNN) and support vector machine (SVM) models are adopted to process and predict data. The accuracies of the prediction of the existing RAC creep, BPNN, and SVM models are comparatively analyzed. The results reveal that for the prediction of RAC creep, the BPNN and SVM models are more accurate than the existing RAC creep model. Finally, the extended parameters of the model are analyzed based on the BPNN model to further clarify the effects of the recycled coarse aggregate (RCA) replacement ratio, RCA residual mortar content, and RAC strength on the creep properties of RAC.