A dedicated ultrasound transmission method is used for in-situ detection and monitoring of fatigue cracks. For this purpose, piezoelectric transducers are embedded at the ends of hourglass-shaped specimens of HSLA steel. One transducer emits constant-amplitude time-limited pulses which travel as longitudinal and surface waves. The waves are received by the other transducer and their times-of-flight and amplitudes are readily monitored by a digitizing oscilloscope. During constant amplitude fatigue cycling at constant R, a decrease in surface wave amplitude indicates crack initiation while crack propagation is sensed by the change in the amplitude of both waves. During initial fatigue cycles, the transmitted wave amplitudes may decrease due to cyclic plastic strain and the accumulation of dislocations. In subsequent cycles, the lattice defects reach a saturation level giving rise to a steady state level of the transmitted wave amplitudes. Wave amplitudes monitored during saturation stage serve as a reference for subsequent measurements of amplitude decrease that indicates the growth of a fatigue crack. Changing the applied load in the early stage of fatigue cycling has a little effect on the saturation level. Such an effect diminishes with fatigue cycling and does not interfere with the ability of detecting crack initiation. The method is capable of monitoring the entire history of fatigue damage evolution from initial strain hardening, through strain saturation, crack nucleation, crack growth to failure.

No abstract

The method proposed is a further development of an earlier work described by Hazony et al. [‘‘Ultrasonic monitoring of tensile, fatigue, or creep specimens in situ,’’ 2279(A) (1993)]. Two transducers are imbedded at the front and back faces of the specimens along the principal axis. These transmit and receive arbitrarily sharp stress pulses that probe both the volume and the surface of the specimen. The respective times of flight are sufficient for the monitoring of longitudinal and transversal strains. Moreover, intensity tracking of the various received acoustic signals also serves well for monitoring onset and growth of cracks. Sensitivity and calibration issues will be discussed and experimental data on high-strength low-alloy steel and other alloy specimens will be presented. The method lends itself to high precision measurements of material deformation and damage and may be used at high temperatures and in aggressive environments. [Work supported by ONR and TecSonics, Inc. of Twinsburg, Ohio.]

Of concern is noninvasive gage-less monitoring fatigue experiments as well as detection and tracking of cracks, surface, and bulk. Accordingly, a digital signal decomposition method is developed for the analysis of sharp, time-limited ultrasonic pulses traveling through a tensile specimen on-site. The method extracts from the propagating pulses, respective elongation, and diameter reduction due to stress, associated with signal delays and changes in amplitude and shape. The analytical process is based on a nonlinear multivariate regression analysis. When applied to the current experiments, the method extracts delay variations accurate to the order of ±1 ps and quantified signal-deformation data. These results relate to a tensile-stress experiment under a five-cycle strain-cycle experiment, extending some of the cycles well into the plastic regime of a stainless steel specimen. The three-section cylindrical specimen consists of two gripper heads and a slender cylindrical section 6 mm in diameter and 20 mm long. Transducers are implanted on both specimen faces along the main axis [Mostafa et al., Int. J. Fract. 85, 99–109 (1997)]. [Work supported by NASA.]

Fatigue tests are often encumbered by lack of the specimen’s exact length due to mechanical grip effects and by environmental noise factors. Such issues may be mitigated when differential measurements are undertaken between consecutive data states. Our principle-monitoring tool is an ultrasonic pulse-echo process where the primary and secondary echoes along the principle specimen’s axis provide both length and gage length, respectively [D. Hazony, Circuit Systems Signal Process. 14(4), 525–538 (1995)]. The process is also useful for crack detection [I. Mostafa et al., Int. J. Fracture 85, 99–109 (1997)]. Compared with absolute measurements, differential stress-strain monitoring is shown to be highly reproducible and more sensitive providing changes in Young’s modulus and Poisson’s ratio as well as a focus for the detection of an emerging crack. [Work supported by NASA.]