Attention Network Forecasts Time-to-Failure in Laboratory Shear Experiments

被引:14
|
作者
Jasperson, Hope [1 ]
Bolton, David C. [2 ]
Johnson, Paul [3 ]
Guyer, Robert [3 ,4 ]
Marone, Chris [5 ,6 ]
De Hoop, Maarten, V [7 ]
机构
[1] Rice Univ, EEPS, Houston, TX 77005 USA
[2] Univ Texas Austin, Inst Geophys, Austin, TX 78712 USA
[3] Geophys Grp, Los Alamos Natl Lab, Los Alamos, NM USA
[4] Univ Nevada, Dept Phys, Reno, NV 89557 USA
[5] Penn State Univ, Dept Geosci, State Coll, PA USA
[6] Sapienza Univ Roma, Dipartimento Sci Terra, Rome, Italy
[7] Rice Univ, CAAM, Houston, TX USA
基金
欧洲研究理事会;
关键词
ACOUSTIC-EMISSION; NEURAL-NETWORKS; SEISMIC CYCLE; EARTHQUAKES; FRICTION; SIGNALS; DAMAGE; ROCK;
D O I
10.1029/2021JB022195
中图分类号
P3 [地球物理学]; P59 [地球化学];
学科分类号
0708 ; 070902 ;
摘要
Rocks under stress deform by creep mechanisms that include formation and slip on small-scale internal cracks. Intragranular cracks and slip along grain contacts release energy as elastic waves termed acoustic emissions (AE). AEs are thought to contain predictive information that can be used for fault failure forecasting. Here, we present a method using unsupervised classification and an attention network to forecast labquakes using AE waveform features. Our data were generated in a laboratory setting using a biaxial shearing device with granular fault gouge intended to mimic the conditions of tectonic faults. Here, we analyzed the temporal evolution of AEs generated throughout several hundred laboratory earthquake cycles. We used a Conscience Self-Organizing Map (CSOM) to perform topologically ordered vector quantization based on waveform properties. The resulting map was used to interactively cluster AEs. We examined the clusters over time to identify those with predictive ability. Finally, we used a variety of LSTM and attention-based networks to test the predictive power of the AE clusters. By tracking cumulative waveform features over the seismic cycle, the network is able to forecast the time-to-failure (TTF) of lab earthquakes. Our results show that analyzing the data to isolate predictive signals and using a more sophisticated network architecture are key to robustly forecasting labquakes. In the future, this method could be applied on tectonic faults to monitor earthquakes and augment early warning systems. Plain Language Summary Earthquake forecasting is a daunting task, but advances in machine learning can help us toward this goal. In this study, we combine machine learning with expert knowledge about earthquake formation to forecast synthetic earthquakes made in a laboratory. Prior to an earthquake, rocks emit small bursts of energy as they are stressed, which are called acoustic emissions (AE). We identified all the AEs in the raw seismic data from our lab earthquakes and then used a machine-learning algorithm called a Conscience Self- Organizing Map to divide them into groups based on their characteristics. We examined the groups to determine which ones would be useful for earthquake forecasting. Finally, we used the data from the chosen groups along with another machine learning approach (LSTM and attention networks) to forecast the time of future labquakes. Our results show that earthquake forecasting is improved by processing seismic data to find predictive signals and by using machine-learning networks that are specifically made for time series data.
引用
收藏
页数:20
相关论文
共 50 条
  • [21] SPREAD OF TIME-TO-FAILURE MEASUREMENTS IN THIN METALLIC FILMS
    BOBBIO, A
    SARACCO, O
    THIN SOLID FILMS, 1973, 17 (01) : S13 - S16
  • [22] A study of bimodal distributions of time-to-failure of copper via electromigration
    Lai, JB
    Yang, JL
    Wang, YP
    Chang, SH
    Hwang, RL
    Huang, YS
    Hou, CS
    2001 INTERNATIONAL SYMPOSIUM ON VLSI TECHNOLOGY, SYSTEMS, AND APPLICATIONS, PROCEEDINGS OF TECHNICAL PAPERS, 2001, : 271 - 274
  • [23] ELECTROMIGRATION MEDIAN TIME-TO-FAILURE BASED ON A STOCHASTIC CURRENT WAVEFORM
    NAJM, F
    HAJJ, I
    YANG, P
    PROCEEDINGS - IEEE INTERNATIONAL CONFERENCE ON COMPUTER DESIGN : VLSI IN COMPUTERS & PROCESSORS, 1989, : 447 - 450
  • [24] INFLUENCE OF TIME-TO-FAILURE ON STRENGTH OF CCA-TREATED LUMBER
    WINANDY, JE
    FOREST PRODUCTS JOURNAL, 1995, 45 (02) : 82 - 85
  • [25] Predicting time-to-failure in rock extrapolated from secondary creep
    Hao, Sheng-Wang
    Zhang, Bao-Ju
    Tian, Ji-Feng
    Elsworth, Derek
    JOURNAL OF GEOPHYSICAL RESEARCH-SOLID EARTH, 2014, 119 (03) : 1942 - 1953
  • [26] Consistent Time-to-Failure Tests and Analyses of Adhesive Anchor Systems
    Nincevic, Kresimir
    Boumakis, Ioannis
    Meissl, Stefan
    Wan-Wendner, Roman
    APPLIED SCIENCES-BASEL, 2020, 10 (04):
  • [27] Bearing time-to-failure estimation using spectral analysis features
    Reuben, Lim Chi Keong
    Mba, David
    STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL, 2014, 13 (02): : 219 - 230
  • [28] A parametric study on the time-to-failure of wood framed walls in fire
    Clancy, P
    FIRE TECHNOLOGY, 2002, 38 (03) : 243 - 269
  • [29] Electron Device Subjected to Temperature Cycling: Predicted Time-to-Failure
    Ephraim Suhir
    Reza Ghaffarian
    Journal of Electronic Materials, 2019, 48 : 778 - 779
  • [30] Optimal Incomplete Maintenance for Systems with Discrete Time-to-Failure Distribution
    Kahle, Waltraud
    MATHEMATICAL AND STATISTICAL MODELS AND METHODS IN RELIABILITY: APPLICATIONS TO MEDICINE, FINANCE, AND QUALITY CONTROL, 2010, : 123 - 132