- 1Department of Earth Sciences, University of Turin, Turin, Italy
- 2Central South University, Changsha, China
Acoustic Emissions (AE), the laboratory analogs for seismic activity, offer a controlled environment to study deformation and failure mechanisms. By leveraging high-precision three-dimensional localizaton techniques, researchers can analyze ongoing these mechanisms during rock deformation experiments. The spatial resolution is thus crucial for increasing our capability to understand and predict failure modes in natural and engineered systems.
Despite recent advances in localization techniques, automated AE localization faces significant challenges. Conventional AE processing systems generally extract a limited set of parameters, such as arrival time defined as the first overcoming a given amplitude threshold. While these parameters provide the bulk information, they often overlook critical signal aspects, underestimating AE phenomena complexities and compromising the localization accuracy. In fact, amplitude thresholds may not capture accurately the signal onset, particularly in noisy or complex waveforms.
This study proposes a new methodology to improve source location accuracy and AE event classification by developing an automatic picking system tailored to seismic signal characteristics and set on Signal-to-Noise Ratio (SNR). The novel algorithm introduced here overcomes conventional amplitude-based thresholding by including broader waveform characteristics, source-receiver distance and wave propagation path. The analysis operates on multiple signal windows and provides uncertainty estimates, enabling more accurate AE location.
The AE source localization process was carried out using the Time Difference Of Arrival (TDOA) method, widely applied in rock deformation laboratory experiments. This approach considers signal arrival time differences at multiple transducers and, with a velocity model, estimates the three-dimensional coordinates of AE sources. The localization quality was assessed via four key parameters: 1) RMS (Root Mean Square) between observed and calculated arrival times, 2) localization errors along the three principal coordinates, 3) MAPE (Mean Absolute Percentage Error quantifying arrival time differences), and 4) the average azimuthal gap across three principal planes. These parameters quantify discrepancies in location accuracy and are employed to assess the final localization.
We applied our methodology to waveform data of AE recorded by an array of twelve 1 MHz piezo-electric transducers during conventional triaxial deformation of a 40 × 100 mm Darley Dale sandstone cylindrical sample (King et al., 2021) at 20 MPa confining pressure. We show that our approach reduces the localization errors and improves the AE detection and localization accuracy. By addressing conventional AE location limitations, this work advances AE-based monitoring toward more accurate and reliable results, providing higher resolution and improved information to describe the damage micro-mechanisms driving failure in stressed rocks.
How to cite: Adinolfi, G. M., Chen, Y., and Vinciguerra, S. C.: Acoustic Emission Source Localization Techniques: A Methodology for Improved Localization Accuracy in Laboratory Deformation Experiments, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12108, https://doi.org/10.5194/egusphere-egu25-12108, 2025.