Experimental simulation of earthquake rupture processes

当地壳中的断层突然破裂时，就会发生地震。束缚断层的岩石块在小地震中会滑动一小段距离，在大地震中会滑动几米。在这次滑动中，断层减弱，这种减弱为地震继续破裂提供了能量。显然，这是一个复杂的过程，在这个过程中，断层的机械特性在几秒钟内就会发生剧烈的变化。探索这些过程的经典实验方法是以恒定速度或非常慢的速度进行的；这两种情况都与自然地震的破裂无关。我们提出了一种新的实验方法，有望揭示与地震有关的关键参数。这些实验将利用我们在俄克拉何马大学的地震实验室在能源供应和电力控制方面的独特能力。这些能力的影响在初步工作中得到了证明，成功地模拟了M=8级地震（类似于1906年旧金山地震）期间的断层行为。为什么这样的实验很重要？首先，总的来说，更好地了解地震过程对减少地震危害至关重要。其次，本文的研究对于建筑结构设计中使用的地震地震动计算具有重要意义。这些计算是基于地震中断层的力学行为，更准确和相关的计算需要对这种行为有更好的了解。理论模型表明，地震破裂是一个复杂的过程，具有运动学和动力学特性的非平凡演化（例如，Tinti et al. 2005）。最近模拟实际地震速度历史的实验（例如，Sone和Shimamoto, 2009； Fukuyama和Mizoguchi, 2010; Liao， Chang和Reches，在EPSL的修订中）表明，在低速，短距离摩擦实验中确定的本构关系并不一定适用于自然地震的可变，高滑动速度。目前的目的是模拟地震破裂过程，通过加载实验断层类似于地震加载，并得出在这些条件下的本构关系。实验将使用高速旋转剪切装置进行，在滑移速度高达1 m/s，法向应力高达30 MPa，滑移距离大的情况下，沿岩石块进行摩擦滑动。该装置具有两个独特的功能：(1)通过存储在飞轮中的有限能量（高达10^7 J/m^2）加载实验故障（Chang et al., 2012）；(2)控制施加在断层上的功率密度的实时反馈系统（功率密度=滑移速度*剪切应力）。这些方法允许应用任何加载历史来揭示瞬态地震作用下的断层响应。两种方法的初步结果在断层强度（减弱和加强）、滑移速度、能量耗散和上升时间的演化方面与断层行为的理论模型具有很强的相似性。我们将观察到的本构关系与断层弱化的机制联系起来，为此，我们将用超微观方法（AFM， XRD， SEM和TEM）表征断层的性质。

An earthquake occurs when a fault in the earth crust ruptures abruptly. The rock blocks that bound the fault, slip for tiny distance in small earthquakes, and up to a few meters in large events. During this slip, the fault weakens and the weakening provides energy to continue the earthquake rupture. Obviously, this is a complex process in which the mechanical properties of the fault drastically change within seconds. Classical experimental approaches to explore these processes are conducted at constant velocity or at very slow velocity; neither of these conditions is relevant to the rupture of natural earthquakes. We propose a new experimental approach that is anticipated to reveal key parameters that are relevant to earthquakes. These experiments will utilize the unique capabilities of energy supply and power control of our earthquake laboratory in the University of Oklahoma. The impact of these capacities was demonstrated in the preliminary work that succeeded in simulations of fault behavior during earthquakes up to magnitude M=8 (similar to 1906 San Francisco earthquake). Why are such experiments important? First, in general, better understanding of earthquake processes is essential for the reduction of seismic hazard. Second, this study is particularly important for the calculations of seismic ground shaking, which are used for building structural design. These calculations are based on the mechanical behavior of faults during earthquakes, and more accurate and relevant calculations require better knowledge of this behavior. Theoretical models suggest that earthquake rupture is a complex process with non-trivial evolution of kinematic and dynamic properties (e.g., Tinti et al. 2005). Recent experiments that simulated the velocity history of actual earthquakes (e.g., Sone and Shimamoto, 2009; Fukuyama and Mizoguchi, 2010; Liao, Chang and Reches, in revision for EPSL) show that constitutive relations which were determined in low-velocity, short-distance friction experiments do not necessarily hold for variable, high slip-velocity of natural earthquakes. The present objective is to simulate the earthquake rupture process by loading experimental faults similarly to earthquake loading, and to derive the constitutive relation at these conditions. The experiments will be conducted with a high-speed rotary shear apparatus that applies frictional sliding along rock blocks under slip-velocity up to 1 m/s, normal stress up to 30 MPa, and large slip distance. This apparatus has two unique capabilities: (1) Loading the experimental fault by a finite amount of energy (up to 10^7 J/m^2) that is stored in a flywheel (Chang et al., 2012); and (2) A real-time feedback system that controls the power-density applied on the fault (power-density = slip velocity * shear-stress). These methods allow the application of any loading history to reveal the fault response under transient earthquake loading. The preliminary results of both methods show strong similarities to theoretical models of fault behavior in terms of the evolution of fault strength (weakening and strengthening), slip velocity, energy dissipation, and rise-time. We will link the observed constitutive relations to the mechanisms of fault weakening, and for this purpose, we will characterize the fault properties with ultramicroscopic methods (AFM, XRD, SEM, and TEM).