Abstract
In this paper, the split spectrum processing (SSP) protocol, a data processing method is offered to make the use of ultrasound of MHz frequency possible for improved nondestructive defect detection and layer thickness determination of concrete.
Reflected ultrasonic signals can be obtained from defects within and from boundaries of concrete, but when the frequency is high, they are typically corrupted by echoes from randomly distributed scatterers in the form of (coherent) noise. This results in a low signal-noise ratio; that is, it is difficult, if not impossible, to distinguish defect signals from scatterer noise that cannot be reduced by time averaging. It is shown in this paper that defect visibility in concrete call be enhanced by combining the high-frequency pulse-echo test with SSP. The basis of SSP is that the scatterer noise is highly sensitive to shifts in the operating frequency, whereas the echoes from targets of interest are not. Thus, if an ultrasonic signal with broad frequency content is captured after passing through concrete, and its Fourier spectrum is led through a Gaussian filter bank that splits the signal into different frequency bands, the computerized reassembly of these bands enhances the signal-noise ratio because the scatterer noise contents obtained at different frequencies cancel out.
SSP has been successful with metals and polymers. The test results presented demonstrate the SSP's noise reduction potential also far improved nondestructive testing of concrete with ultrasound of frequencies up to 1 MHz. It is also shown, however that further research is needed for field application.