Surface reconstruction accuracy using ultrasonic arrays: Application to non-destructive testing

Abstract The accurate non-destructive inspection of engineering structures using ultrasonic immersion imaging requires a precise representation of the surface of the structure. Here we investigate the relationship between surface geometry, surface measurement error using ultrasonic arrays and the total focusing method (TFM) and how this impacts on the ability to image a feature within a component. Surfaces shaped as sinusoids covering combinations of surface wavelengths ( 0.8 to 32 λ w a t e r ) and amplitudes ( 0.6 to 9 λ w a t e r ) are studied. The surface reconstruction errors are shown to cause errors in imaging, such as reduced amplitude and blurring of the image of a side-drilled hole. These reconstruction errors are shown to increase rapidly with the maximum gradient of the sinusoid. Sinusoidal surfaces with maximum gradients 45 ° lead to average surface reconstruction errors λ w a t e r and amplitude imaging errors within 6 dB of the flat-surface case. It is also shown that very poor results are obtained if the surface gradient is excessively steep.