Adaptive array processing for ultrasonic non-destructive evaluation

In non-destructive evaluation (NDE) applications, most materials like concrete, stainless steel and carbon-reinforced composites used extensively in industries and civil engineering exhibit heterogeneous internal structure. When inspected using ultrasound with the conventional delay-and-sum imaging, the signals from defects are significantly corrupted by the echoes form randomly distributed scatterers, even defects that are much larger than these random reflectors are difficult to detect. We propose to apply adaptive beamforming to the received data samples to reduce the interference and clutter noise. Beamforming is to manipulate the array beam pattern by appropriately weighting the per-element delayed data samples prior to summing them. The adaptive weights are computed from the statistical analysis of the data samples. This delay-weight-and-sum process can be explained as applying a lateral spatial filter to the signals across the probe aperture. Simulations show that the background noise level is reduced by 30dB when adaptive beamforming is applied. In experiments inspecting a steel block with side-drilled holes, good quantitative agreement with the simulation results is demonstrated.