TY - JOUR
T1 - An accelerating precursor to predict “time-to-failure” in creep and volcanic eruptions
AU - Hao, Shengwang
AU - Yang, Hang
AU - Elsworth, Derek
N1 - Funding Information:
This work is supported by National Natural Science Foundation of China (Grant no. 11672258), National Basic Research Program of China (Grant no. 2013CB834100) and Natural Science Foundation of Hebei Province (Grant D2015203398).
Publisher Copyright:
© 2017 Elsevier B.V.
PY - 2017/9/1
Y1 - 2017/9/1
N2 - Real-time prediction by monitoring of the evolution of response variables is a central goal in predicting rock failure. A linear relation Ω̇Ω¨−1=Ctf−t has been developed to describe the time to failure, where Ω represents a response quantity, C is a constant and tf represents the failure time. Observations from laboratory creep failure experiments and precursors to volcanic eruptions are used to test the validity of the approach. Both cumulative and simple moving window techniques are developed to perform predictions and to illustrate the effects of data selection on the results. Laboratory creep failure experiments on granites show that the linear relation works well during the final approach to failure. For blind prediction, the simple moving window technique is preferred because it always uses the most recent data and excludes effects of early data deviating significantly from the predicted trend. When the predicted results show only small fluctuations, failure is imminent.
AB - Real-time prediction by monitoring of the evolution of response variables is a central goal in predicting rock failure. A linear relation Ω̇Ω¨−1=Ctf−t has been developed to describe the time to failure, where Ω represents a response quantity, C is a constant and tf represents the failure time. Observations from laboratory creep failure experiments and precursors to volcanic eruptions are used to test the validity of the approach. Both cumulative and simple moving window techniques are developed to perform predictions and to illustrate the effects of data selection on the results. Laboratory creep failure experiments on granites show that the linear relation works well during the final approach to failure. For blind prediction, the simple moving window technique is preferred because it always uses the most recent data and excludes effects of early data deviating significantly from the predicted trend. When the predicted results show only small fluctuations, failure is imminent.
UR - http://www.scopus.com/inward/record.url?scp=85026299578&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85026299578&partnerID=8YFLogxK
U2 - 10.1016/j.jvolgeores.2017.07.009
DO - 10.1016/j.jvolgeores.2017.07.009
M3 - Article
AN - SCOPUS:85026299578
SN - 0377-0273
VL - 343
SP - 252
EP - 262
JO - Journal of Volcanology and Geothermal Research
JF - Journal of Volcanology and Geothermal Research
ER -