Comparison and combination of two NDT methods with implications for compressive strength evaluation of selected masonry and building stones

The classic Schmidt hammer (SH) and the Equotip hardness tester (EHT) are the widely used nondestructive (NDT) instruments for measuring surface dynamic hardness (rebound hardness) of materials. In previous studies, both instruments have been used individually to establish empirical models for uniaxial compressive strength (UCS) evaluation of rock materials. However, no specific research seems to have been undertaken to investigate their relative UCS prediction effectiveness when performed on the same rock samples. In this study, we evaluated the relative UCS prediction performance of these two NDT instruments by performing laboratory tests on core specimens of some selected rock materials used as masonry and building stones. The correlations of Schmidt hammer rebound hardness (SHR) and Equotip hardness (HLB) to UCS were determined by using regression analyses and test statistics. Also, the variability of each test method was evaluated by employing variance analyses. The results indicated that the established UCS prediction models for both NDT methods were significant in the statistical sense (p value <0.05), with the EHT method showing somewhat better prediction accuracy (R-2 = 0.87; accuracy ratio = 0.60) compared to the SH method (R-2 = 0.84; accuracy ratio = 0.78). Data scatter graphs showed increasing standard deviation in UCS for increasing rebound indices. An average conversion factor (k) of 15.7 was derived in the zero-intercept regression model HLB = k . SHR, which may be used in practice as a guide to translate the measured hardness values from one instrument to another. For improved UCS prediction accuracy, an attempt was made to evaluate the potential use of a newly proposed combined NDT method involving the simultaneous use of SH and EHT test results. With this kind of approach, a relatively better prediction accuracy was achieved (R-2 = 0.89; accuracy ratio = 0.52) compared to the use of the SH and EHT methods individually.