Strength prediction of alkali-activated cemented gangue backfill materials based on ultrasonic parameters

Utilizing mining waste rock for the production of cemented gangue backfill material (CGBM) represents a crucial pathway for the resource utilization of solid waste. However, the practical engineering application and promotion of this material have been limited due to the challenges in real-time monitoring of its strength development at the construction site. Ultrasonic non-destructive testing was employed to evaluate the strength properties of CGBM. This study systematically investigated the effects of key parameters, including activator dosage (D), activator ratio (N), and curing age (t), on the ultrasonic pulse velocity (UPV), with the mechanism by which the activator affects UPV elucidated from a microstructural perspective. A compressive strength prediction model (GSA-UPV) was developed based on multi-parameter coupling relationships, utilizing the Gravitational Search Algorithm (GSA) to optimize decision variables. The results indicated that the UPV of CGBM initially increased and then decreased with higher D and N, a trend which was effectively fitted by a quadratic polynomial. The optimal dosage ranges were identified as 12%–18% for D and 1:2–1:1 for N. The GSA-UPV model demonstrated significantly improved prediction accuracy, with a mean absolute percentage error (MAPE) of 5.5405% and a coefficient of determination (R²) of 0.9667.