Filtering of Acoustic Emission Signals for the Accurate Identification of Fracture Mechanisms in Bending Tests

Abstract

In this manuscript, acoustic emission (AE) analysis is used to identify the fracture mechanisms that take place during a three-point bending test of steel specimens. However, an important source of AE was detected and related to the deformation of the supports and the surface of the steel specimen in contact. This source creates AE signals that are very high in number and amplitude, and prevent determining accurately the onset stress for fracture mechanisms in specimens. A signal filtering is proposed based on the properties of the initial part of the recorded AE waveform, combined with a linear localization. The filtering successfully allows the AE signals to be classified according to their source as background noise, damage due to contact of the specimen surface to the supports and fracture mechanisms occurring in the specimen microstructure as a result of the bending test. The aforementioned filter has been successfully applied in case of a cold work tool steel DIN 1.2379 to determine accurately the stress level at which the first damaging mechanisms start to occur in the microstructure in situ during the three-point bending test. Filtered AE signal results indicating damage in the microstructure have been corroborated by inspection of the specimen's surface in a Confocal Microscope. In default of using the proposed filter, unfiltered signals have been estimated to lead to an overestimation of critical stresses of about 20%, what is undesirable for most applications.

Stick-Slip Phenomena and Acoustic Emission in the Hertzian Linear Contact

“…These modifications are usually composed of discrete events so that the emission of elastic waves is in the form of bursts or pulses of distinctive characteristics. Therefore, the analysis of signals captured in the AE test is usually performed considering this discrete nature, using a rather complex range of parameters related to the wave pulses, such as the number of counts, the rise and disappearance time, or frequency content-related variables [29][30][31][32][33]. This analytical strategy, combined with the frequency content, typically in the ultrasound range (>20 kHz), leads to the need for specific equipment, both hardware and software, to carry out these very specific tests.…”

Deep Learning-Based Acoustic Emission Signal Filtration Model in Reinforced Concrete

“…Signals with extremely short durations can be discarded using duration filters, whereas electric noise with significantly high signal strengths can be filtered using a signal strength filter [18]. Some researchers also define frequency filters to eliminate noise based on very low-frequency signals [19,20]. Moreover, [21] proposed a filtering standard based on the root mean square of AE signals to separate mechanical friction from concrete cracking.…”

Acoustic emission processes occurring during high‐pressure sand compaction

“…The study of AE waveforms in both the time and frequency domains has been developed by many researchers in different study areas (Theobald, Zequiri and Avison 2008;Nunez, Nishino and Yoshida 2008;Shrivastava and Prakash 2009;Martínez-González et al 2013). Most of these authors conclude that it is generally possible to associate a given phenomenon with some specific frequency content of the AE signal.…”