Study on the effects of steam-curing regimes on the fracture characteristics of RAC based on acoustic emission and machine learning

The application of recycled aggregate concrete (RAC) is an effective way to address ecological and environmental problems. To investigate the fracture characteristics of steam-cured RAC, specimens under standard curing, 60 °C–12 h steam curing, and 80 °C–9 h steam curing were prepared. Acoustic emission (AE) technology was used to monitor damage evolution during static and cyclic three-point bending tests. An information-dimension method based on AE parameters was developed to divide the damage stages of RAC, and a semi-parametric clustering-support vector machine (SPC-SVM) method was established to identify damage modes. The results show that, under monotonic loading, AE activity increased with load and was concentrated near the peak load. Under cyclic loading, AE signals exhibited obvious hysteresis. The information-dimension method accurately captured staged damage evolution, showing a three-stage pattern under monotonic loading and a two-stage pattern with pronounced cyclic response under cyclic loading. The SPC-SVM method identified the damage modes effectively. Combined with damage stage division, it revealed the fracture characteristics at each stage. The results were consistent with the fracture evolution of RAC under different curing regimes.