Correlation between non-destructive testing indicators and local transverse compressive strength and density of wood

Objective The differences in correlation between different types and quantities of non-destructive testing (NDT) indicators and the local transverse compressive strength and density of wood were studied, and the optimal combination of NDT indicators were determined, in order to provide reference for the evaluation of transverse local compressive strength and density of wood based on NDT methods.

Method The beam components from a typical Tibetan ancient building wooden structure were taken as the research object, and the local transverse compressive samples were sawn off. By conducting physical and mechanical tests and NDT tests, the density (ρ), local transverse compressive strength (f_lc), longitudinal and transverse stress wave velocity (v_l and v), and micro drilling resistance (F) were measured. The linear correlation coefficients between different NDT indicators or indicator combinations and ρ, f_lc were calculated, and then comparatively discussed.

Result (1) The correlation coefficient between v_l and ρ was the highest when using single NDT index, at 0.662, after supplementing the F_m (the mean value of F), it can be increased by 15.56%, reaching 0.765. The correlation coefficient between F_m and f_lc was the highest when using single NDT indicator, at 0.526, after supplementing the v_m (mean value of v), it can be increased by 26.80%, reaching 0.667; after further supplementing the v_l, it can be further increased by 7.20%, reaching 0.715. (2) When the basic indicators in two sets of NDT indicator combinations are consistent, even if there are differences in the form of basic indicators, the correlation coefficients between the two sets and the predicted parameters are approximate. (3) When two NDT indicators with good correlation with the predicted parameters are combined, the correlation coefficient of the combined model may further increase; when NDT indicators with good correlation are combined with indicators with poor correlation, the correlation coefficient of the combination model is close to the former; when NDT indicators with poor correlation are combined with unrelated indicators, the correlation coefficient may increase significantly or remain poor; when two NDT indicators that are not significantly correlated are combined, the combination of NDT indicators still does not have a significant correlation with the predicted parameters.

Conclusion Determining reasonable NDT indicators has a greater impact on improving the correlation than simply increasing the number of indicators. Taking into account the improvement effect of correlation coefficients and the potential increase in testing workload, it is recommended to use the combination of “v_l, F_m” and “F_m, v_m” for predicting the density and local transverse compressive strength of wood, respectively.