Towards energy-based ultrasonic imaging for the detection of microscopic defects in strongly scattering materials

Detecting microscopic defects-particularly sub-millimeter defects much smaller than the ultrasonic wavelength-in strongly scattering materials remains a significant challenge due to intense grain-induced noise. To address this issue, an energy-based ultrasonic imaging method is proposed, which is capable of revealing such subtle defects under heavy noise interference. Grain scattering noise and defect-reflected signals were simulated using the finite element method, and the spectral characteristics of both were analyzed. The results show that grain noise and defect-reflected signals differ significantly in both peak frequency distribution and effective bandwidth. Signal energy is then computed in the frequency domain using Parseval's theorem, and a frequency-dependent weighting function is introduced to enhance defect-reflected signals for imaging. Experimental validation was conducted on 304 stainless steel and two-phase Ti-6Al-4 V alloy specimens containing sub-millimeter defects at various depths. It is shown that, while conventional amplitude-based imaging fails to visualize these defects, the proposed energy-based approach successfully detects them with high sensitivity. Notably, the technique is implemented using only standard ultrasonic equipment, involves no complex algorithms, and achieves high resolution-offering a practical and effective strategy for non-destructive evaluation of microscopic defects in strongly scattering materials.