Ultrasonic Signal Characteristics for Nondestructive-Yield Detection in Steel Structures

Abstract

Nondestructive testing (NDT) methods for identifying stress levels in materials mostly rely on the theory of acoustoelasticity. However, the sensitivity and the accuracy of acoustoelasticity are affected by several factors such as the (1) type, (2) propagation, and (3) polarization directions of the used signals. This paper presents the results of an experimental investigation of longitudinal waves propagating perpendicular to the applied uniaxial tensile stresses in structural steel specimens. The changes in four ultrasonic signal characteristics were investigated with increasing stress levels ranging from below to above the yield stress of steel. The considered signal characteristics were the peak amplitudes and signal energy in the time domain, and the fast Fourier transform (FFT) and chirp-Z transform (CZT) in the frequency domain. Even though the acoustoelastic effect on the type ultrasonic signal used is very small, clear distinctions between prior to and postyielding are observed for all investigated parameters. The results are presented with a detailed statistical and receiver operating characteristics (ROCs) analyses. The results show that identifying damage to steel structures due to local yielding is possible using the simple ultrasonic signal classification.