A new method to estimate location and slip of simulated rock failure events
A new method to estimate location and slip of simulated rock failure events
复制标题
DOI：
10.1016/j.tecto.2015.03.009
复制
发表时间：
2015-05-31
期刊：
影响因子：
2.9
通讯作者：
Miller, Stephen Andrew
中科院分区：
文献类型：
--
作者：
Heinze, Thomas;Galvan, Boris;Miller, Stephen Andrew
在实验室尺度上，识别和定位声发射(AEs)是岩土材料短期失效预测的常用方法。高于平均值的声发射通常先于故障过程，并且很容易测量。在更大的尺度上，微震活动的增加有时先于大地震(如东北，海洋变换)，并可用于评估地震风险。这项工作的目标是开发一种方法和数值算法来提取可测量的量，类似于岩石变形方程的解所产生的声发射。由于没有可从控制方程中导出的物理性质来量化声发射，因此需要找到合适的岩石力学模拟。在这项工作中，我们确定了岩石破坏前声发射产生过程的一般行为。这种行为包括在破坏前阶段任意定位低震级事件，然后增加数目和幅度，最后在宏观破坏时定位在早期破坏平面附近。我们提出偏差应变率作为模拟这一行为的数值模拟，并开发了两种不同的算法，旨在利用移动平均来检测偏差应变的快速增加。数值模型求解完全多孔-弹塑性连续介质模型，并耦合到两相流动模型。我们通过将模拟结果与排水压缩和注液实验数据进行比较来验证我们的模型。我们发现，对于这两种情况，我们的声发射模拟的发生和幅度都模拟了观测到的声发射产生过程的一般行为。我们的技术可以扩展到现场尺度上的建模，可能为从偶尔发生大地震前的地震活动评估地震危险性提供一个力学基础。(C)2015爱思唯尔B.V.保留所有权利。
At the laboratory scale, identifying and locating acoustic emissions (AEs) is a common method for short term prediction of failure in geomaterials. Above average AE typically precedes the failure process and is easily measured. At larger scales, increase in micro-seismic activity sometimes precedes large earthquakes (e.g. Tohoku, oceanic transforms), and can be used to assess seismic risk.The goal of this work is to develop a methodology and numerical algorithms for extracting a measurable quantity analogous to AE arising from the solution of equations governing rock deformation. Since there is no physical property to quantify AE derivable from the governing equations, an appropriate rock-mechanical analog needs to be found. In this work, we identify a general behavior of the AE generation process preceding rock failure. This behavior includes arbitrary localization of low magnitude events during pre-failure stage, followed by increase in number and amplitude, and finally localization around the incipient failure plane during macroscopic failure. We propose deviatoric strain rate as the numerical analog that mimics this behavior, and develop two different algorithms designed to detect rapid increases in deviatoric strain using moving averages.The numerical model solves a fully poro-elasto-plastic continuum model and is coupled to a two-phase flow model. We test our model by comparing simulation results with experimental data of drained compression and of fluid injection experiments. We find for both cases that occurrence and amplitude of our AE analog mimic the observed general behavior of the AE generation process.Our technique can be extended to modeling at the field scale, possibly providing a mechanistic basis for seismic hazard assessment from seismicity that occasionally precedes large earthquakes. (C) 2015 Elsevier B.V. All rights reserved.