Analysis of Acoustic Emissions from a Steel Bridge Hanger

Abstract

A steel bridge hanger with three fatigue cracks was monitored for acoustic emission (AE) using combined source location, strain gauge monitoring, and waveform analysis. AE activities from all three cracks were clearly identified and classified as crack growth or noise signals using location, strain magnitude, position on strain cycle, and uniqueness of waveforms as the primary criteria. A vast majority of AE from the cracks was found to be due to crack face rubbing and the crushing of corrosion products between the crack faces while limited crack growth emissions were detected. Results from laboratory tests on A588 compact tension specinlens under variable-amplitude tension-tension fatigue loading were used to aid in interpreting AE data from the hanger. Crack growth AEs from these tests were detected only on overload cycles mostly above 92% of the maximum load while AE due to crack face rubbing occurred throughout the load cycle.

Damage evolution and acoustic emission of wood at tension perpendicular to fiber

“…Sison et al 1998). In wood science the ®rst application of the acoustic emission technique to detect fracture development was performed by Adams 1969.…”

Optimisation of acoustic emission wavestreaming for structural health monitoring

“…Referring to previous discussion, this is the peak frequency band which is thought to owe itself to crack face rubbing. 41,42,44 With regard to the frequency centroid data, both the 100-150 and 150-200 kHz bands bear a good similarity to the crack growth activity as determined by DIC, although the 150-200 kHz band is the most comparable as it most closely matches the changes in the gradient. Again, this is the frequency centroid band which was predicted to have a connection to crack face rubbing.…”

“…As well as primary signals from crack growth itself, secondary signals, such as those from crack-face rubbing, can be valuable indicators of crack activity [45]. It is understood that crack face rubbing signals are typically lower amplitude than crack growth signals and can occur at a range of amplitudes during a single test [42,45,46]. It is therefore challenging to attempt to isolate crack face rubbing signals in this experiment based on amplitude and location, however there are discussions of crackface rubbing in the literature from which to draw upon.…”

“…It is possible that this is due to the fact that the sensors used are resonant over this range, thus they are likely to record a certain amount of energy at this frequency for all AE arriving at the sensor, no matter what the source. Thus, even if the 250-300 kHz frequency centroid band was related to the crack growth mechanism, since the occurrence of crack growth signals during fatigue is typically relatively low [45], they might not be able to drastically change the appearance of that frequency band above and beyond the trends dictated by the resonant-noise.…”

“…44 It is understood that crack face rubbing signals are typically lower amplitude than crack growth signals and can occur at a range of amplitudes during a single test. 41,44,45 It is therefore challenging to attempt to isolate crack face rubbing signals in this experiment based on amplitude and location; however, there are discussions of crack face rubbing in the literature from which to draw upon. Noting that the frequency of a signal recorded by an acquisition system is dependent on the transfer function of the signal from the source, including material and geometric effects, and sensor and preamplifier characteristics, it can be difficult to compare the results found in various literature on characterisation of mechanisms using frequency with high precision.…”