Characterizing acoustic signals and searching for precursors during the laboratory seismic cycle using unsupervised machine learning

David C. Bolton, Chris J. Marone, Parisa Shokouhi, Jacques Riviere, Bertrand Rouet-Leduc, Claudia Hulbert, Paul A. Johnson

Research output: Contribution to journalReview article

1 Citation (Scopus)

Abstract

Recent work shows that machine learning (ML) can predict failure time and other aspects of laboratory earthquakes using the acoustic signal emanating from the fault zone. These approaches use supervised ML to construct a mapping between features of the acoustic signal and fault properties such as the instantaneous frictional state and time to failure. We build on this work by investigating the potential for unsupervised ML to identify patterns in the acoustic signal during the laboratory seismic cycle and precursors to labquakes. We use data from friction experiments showing repetitive stick-slip failure (the lab equivalent of earthquakes) conducted at constant normal stress (2.0 MPa) and constant shearing velocity (10 μm=s). Acoustic emission signals are recorded continuously throughout the experiment at 4 MHz using broadband piezoceramic sensors. Statistical features of the acoustic signal are used with unsupervised ML clustering algorithms to identify patterns (clusters) within the data. We find consistent trends and systematic transitions in the ML clusters throughout the seismic cycle, including some evidence for precursors to labquakes. Further work is needed to connect the ML clustering patterns to physical mechanisms of failure and estimates of the time to failure.

Original languageEnglish (US)
Pages (from-to)1088-1098
Number of pages11
JournalSeismological Research Letters
Volume90
Issue number3
DOIs
StatePublished - Jan 1 2019

Fingerprint

machine learning
acoustics
cycles
earthquakes
stick-slip
earthquake
acoustic emission
shearing
laboratory
fault zone
slip
friction
experiment
sensor
broadband
trends
sensors
estimates

All Science Journal Classification (ASJC) codes

  • Geophysics

Cite this

@article{b78c8d40f8fc4b2d8ddc601844f32659,
title = "Characterizing acoustic signals and searching for precursors during the laboratory seismic cycle using unsupervised machine learning",
abstract = "Recent work shows that machine learning (ML) can predict failure time and other aspects of laboratory earthquakes using the acoustic signal emanating from the fault zone. These approaches use supervised ML to construct a mapping between features of the acoustic signal and fault properties such as the instantaneous frictional state and time to failure. We build on this work by investigating the potential for unsupervised ML to identify patterns in the acoustic signal during the laboratory seismic cycle and precursors to labquakes. We use data from friction experiments showing repetitive stick-slip failure (the lab equivalent of earthquakes) conducted at constant normal stress (2.0 MPa) and constant shearing velocity (10 μm=s). Acoustic emission signals are recorded continuously throughout the experiment at 4 MHz using broadband piezoceramic sensors. Statistical features of the acoustic signal are used with unsupervised ML clustering algorithms to identify patterns (clusters) within the data. We find consistent trends and systematic transitions in the ML clusters throughout the seismic cycle, including some evidence for precursors to labquakes. Further work is needed to connect the ML clustering patterns to physical mechanisms of failure and estimates of the time to failure.",
author = "Bolton, {David C.} and Marone, {Chris J.} and Parisa Shokouhi and Jacques Riviere and Bertrand Rouet-Leduc and Claudia Hulbert and Johnson, {Paul A.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1785/0220180367",
language = "English (US)",
volume = "90",
pages = "1088--1098",
journal = "Earthquake Notes - Seismological Society of America, Eastern Section,",
issn = "0895-0695",
publisher = "Seismological Society of America",
number = "3",

}

Characterizing acoustic signals and searching for precursors during the laboratory seismic cycle using unsupervised machine learning. / Bolton, David C.; Marone, Chris J.; Shokouhi, Parisa; Riviere, Jacques; Rouet-Leduc, Bertrand; Hulbert, Claudia; Johnson, Paul A.

In: Seismological Research Letters, Vol. 90, No. 3, 01.01.2019, p. 1088-1098.

Research output: Contribution to journalReview article

TY - JOUR

T1 - Characterizing acoustic signals and searching for precursors during the laboratory seismic cycle using unsupervised machine learning

AU - Bolton, David C.

AU - Marone, Chris J.

AU - Shokouhi, Parisa

AU - Riviere, Jacques

AU - Rouet-Leduc, Bertrand

AU - Hulbert, Claudia

AU - Johnson, Paul A.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Recent work shows that machine learning (ML) can predict failure time and other aspects of laboratory earthquakes using the acoustic signal emanating from the fault zone. These approaches use supervised ML to construct a mapping between features of the acoustic signal and fault properties such as the instantaneous frictional state and time to failure. We build on this work by investigating the potential for unsupervised ML to identify patterns in the acoustic signal during the laboratory seismic cycle and precursors to labquakes. We use data from friction experiments showing repetitive stick-slip failure (the lab equivalent of earthquakes) conducted at constant normal stress (2.0 MPa) and constant shearing velocity (10 μm=s). Acoustic emission signals are recorded continuously throughout the experiment at 4 MHz using broadband piezoceramic sensors. Statistical features of the acoustic signal are used with unsupervised ML clustering algorithms to identify patterns (clusters) within the data. We find consistent trends and systematic transitions in the ML clusters throughout the seismic cycle, including some evidence for precursors to labquakes. Further work is needed to connect the ML clustering patterns to physical mechanisms of failure and estimates of the time to failure.

AB - Recent work shows that machine learning (ML) can predict failure time and other aspects of laboratory earthquakes using the acoustic signal emanating from the fault zone. These approaches use supervised ML to construct a mapping between features of the acoustic signal and fault properties such as the instantaneous frictional state and time to failure. We build on this work by investigating the potential for unsupervised ML to identify patterns in the acoustic signal during the laboratory seismic cycle and precursors to labquakes. We use data from friction experiments showing repetitive stick-slip failure (the lab equivalent of earthquakes) conducted at constant normal stress (2.0 MPa) and constant shearing velocity (10 μm=s). Acoustic emission signals are recorded continuously throughout the experiment at 4 MHz using broadband piezoceramic sensors. Statistical features of the acoustic signal are used with unsupervised ML clustering algorithms to identify patterns (clusters) within the data. We find consistent trends and systematic transitions in the ML clusters throughout the seismic cycle, including some evidence for precursors to labquakes. Further work is needed to connect the ML clustering patterns to physical mechanisms of failure and estimates of the time to failure.

UR - http://www.scopus.com/inward/record.url?scp=85070757766&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070757766&partnerID=8YFLogxK

U2 - 10.1785/0220180367

DO - 10.1785/0220180367

M3 - Review article

AN - SCOPUS:85070757766

VL - 90

SP - 1088

EP - 1098

JO - Earthquake Notes - Seismological Society of America, Eastern Section,

JF - Earthquake Notes - Seismological Society of America, Eastern Section,

SN - 0895-0695

IS - 3

ER -