An assessment of the feasibility of ultrasound as a defect detector in lumber

This study investigated the limitations of certain attributes for use in detecting internal defects using ultrasonic methods in spruce dimension lumber. Wave propagation in the radial direction and the effect of induced defects on ultrasonic wave characteristics were examined. Ultrasonic velocity, a parameter often applied, was found to be ineffective due to changes in wavelength and frequency when large defects were present. Amplitude, peak frequency magnitude and the area under the frequency spectra showed statistically significant differences between clear and defect-bearing samples, but among-sample variation was found to be too high because of localized density differences and microstructure, so that these methods were not consistently reliable. When sound intensity level was calculated, the variation among samples decreased and distinctions could be made between clear and defect-bearing wood, both within and among the samples tested. Certain attributes of the frequency spectra also showed clear changes with the introduction of defects.