Multi-parameter evaluation of concrete compressive strength using nondestructive techniques

Accurate assessment of in-situ concrete strength is critical for ensuring structural safety and durability. While nondestructive testing (NDT) methods such as the ultrasonic pulse velocity test and Schmidt’s rebound hammer are widely used, their reliability is limited when applied individually. This study introduces the StructureScan Mini velocity, a parameter obtained from the StructureScan Mini device, which operates on the principle of Ground Penetrating Radar (GPR), as an additional predictor to enhance strength estimation. Experimental tests were conducted on plain and reinforced concrete slabs with target strengths of 3,000 and 4,000 psi. To expand the dataset and improve robustness, a Generative Adversarial Network (GAN) was employed. Regression models were developed using single, two-variable, and three-variable combinations of ultrasonic pulse velocity, rebound number, and StructureScan Mini velocity. Results indicate that single-variable models performed poorly (R² ≤ 0.38), while two-variable models improved accuracy (R² up to 0.62). The three-variable model delivered the best performance, achieving R² values of 0.80–0.83 and mean absolute errors below 2% of strength ranges. Field validation further confirmed predictions within ± 8% of core test results. The inclusion of StructureScan Mini velocity proved essential for reducing scatter and improving reliability, offering engineers a practical, non-invasive tool for structural assessment and quality control.