Acoustic emission from low-cycle high-stress-intensity fatigue

This paper presents the findings of an Advanced Research Projects Agency (ARPA) study, the overall objective of which was to develop a nondestructive testing technique to determine flaw criticality based on acoustic emission. The research included an evaluation of sensors and instrumentation systems, using several materials and material conditions loaded in low-cycle, high-stress-intensity fatigue.The materials used for the study were D6aC tempered at 600 and 1100°F, annealed and solution-treatedand-aged 6A1-4V titanium and 7075-T6 aluminum. The test specimen was the precracked, single-edgenotch tension specimen; macrocracking was detected by crack-opening-displacement (COD) gage and micro-cracking by acoustic emission. The acoustic-emission system utilized 400 and 1000 kHz band-pass filtering at 100db gain. The output signals of the totalizer and the COD gage were recorded on a single strip chart using a dual-pen recorder. The specimens were subjected to low-cycle, high-stress-intensity fatigue at 6 c/min. In some tests, cycling was begun in air and finished in water.Acoustic emission was demonstrated to be highly effective as a non-destructive test method for following crack growth in low-cycle high-stress-intensity fatigue; acoustic emission confirmed the existence of periods of dormancy punctuated by periods of active fatigue crack growth. Using a dual-pen, strip-chart recorder displaying both crack-opening-displacement and stress-wave count on the same chart, it was a simple matter not only to observe if there was crack growth in each individual cycle but also where in the cycle it occurred. Moreover, the process of stress-corrosion cracking during low-cycle, high-stress-intensity fatigue was readily detected by a marked increase in the stress-wave count rate.The utility of acoustic emission as a precursor of imminant failure was demonstrated for low-cycle, high-stress-intensity fatigue as well as for the case of environmentally assisted fatigue. Plots of cumulative stress-wave count vs cycle number consistently showed a marked increase in count rate several (10–20 or more) cycles before fracture.