Evaluation of controlled low strength materials containing red brick construction waste based on mechanical experiments and BP neural network

Abstract

In this study, two kinds of controlled low-strength materials (common and rapid-hardening) were designed and prepared by using red brick construction waste fines. Fourteen mixture specimens were prepared with the control parameters of binder to construction waste fine aggregate ratio (α), accelerator to binder ratio (β), and water to solid ratio (γ). The flowability of fresh mixtures, compressive strength of hardened materials were studied. Test results show that it is applicable to use the red brick construction waste fine aggregate to prepare controlled low-strength materials. Increasing α and γ can increase the flowability, however with different the influence mechanisms, where γ has a more significant effect. The 2 hour-strength of rapid-hardening CLSM (controlled low strength material) based on sprayed concrete accelerator can be up to 0.41∼0.79MPa, which meet the needs of emergency backfill engineering. Using on the BP neural network, the long-term strength prediction model of CLSM is established, which can help the civil engineers in the material mix-design.