Three-Dimensional Distinct Element Simulations of Dynamic Process of Fault Gouge Evolution

脆性断裂带在滑动过程中经历断层表面的连续磨损，导致断层表面粗糙度的连续变化和断层泥带厚度的逐渐增加。断层泥带特征的动态变化可能影响断层的摩擦强度、稳定性和地震特征。尽管以往的野外、实验室和数值研究在探讨断层泥带的几何和物理性质及其对断层带摩擦和力学行为的影响方面取得了重大进展，但这些结果仍然对断层泥带在各种地质环境下发育的完整过程以及这些物质如何影响地震周期提供了有限的理解。最近对离散元法（DEM）的修改，包括增加颗粒间结合和破碎，使我们能够模拟黏结岩的断裂过程，并使直接观察断层泥的聚集和修改成为可能，模拟自然断裂带变形。断层泥演化的二维DEM模拟成功地再现了现场和实验室的许多观测结果，产生了以前断层泥带研究中没有报道过的有趣发现。这包括泥断层带发育的阶段及其对剪切位移、正应力和岩石强度的依赖。这些令人鼓舞的结果表明，改进数值模型并将目前的研究扩展到未来的可行性、必要性和重要性。该方案描述了一个程序来模拟断层带演化的动态过程，重点是约束断层泥厚度、粒度分布、剪切带强度和表面粗糙度之间的关系。我们将使用更大的集合，这将允许更大范围的表面和颗粒尺寸和几何形状，并在新的Rice Cray XD1超级计算机上并行进行这些模拟。这些数值实验的结果将有助于解释更复杂的自然断层的观测结果，从而提高对断裂带过程的理解。这项工作将支持博士后助理和本科生的培训，并为社区提供断层泥演化的增强3D代码。了解断层带是如何变形的可以帮助我们更好地了解断层是如何滑动的，从而了解地震的危险性。

Brittle fault zones undergo continual wear of fault surfaces during slip, whichleads to a continuous variation in fault surface roughness and a progressive increase inthe thickness of the fault gouge zone with net slip. The dynamic changes in gouge zonefeatures may affect fault frictional strength, stability, and earthquake characteristics.Although previous field, laboratory, and numerical studies have made significantprogress in investigating the geometrical and physical properties of fault gouge zones andtheir effects on fault zone frictional and mechanical behavior, these results still providelimited understanding of the complete process of gouge zone development under a widerange of geological settings, and how these materials might affect the earthquake cycle.Recent modifications to the Distinct Element Method (DEM), including additionof interparticle bonding and breakage, allows us to model the fracture process of cohesiverocks, and makes it possible to directly observe gouge accumulation and modification,mimicking natural fault zone deformation. 2-D DEM simulations of gouge evolutionhave successfully reproduced many observations in the field and laboratory, yieldinginteresting findings that have not been reported in previous gouge zone studies. Theseinclude stages of gouge zone development and their dependence on shear displacement,normal stress, and rock strength. These encouraging results suggest the feasibility,necessity, and importance of improving numerical models and extending the currentstudies into the future.This proposal describes a program to simulate the dynamic processes of gougezone evolution, focused on constraining the relationships between gouge thickness, grainsize distribution, shear zone strength, and surface roughness. We will use much largerassemblages that will allow a wider range of surface and particles sizes and geometries,and conduct these simulations in parallel on the new Rice Cray XD1 supercomputer.Results of these numerical experiments will help in interpreting observations of morecomplex natural faults, leading to an improved understanding of fault zone processes.The work will support training of a postdoctural associate, undergraduates, and provide enhanced 3D codes of fault gouge evolution to the community. Understanding how fault zonesdeform can help improve understanding of how faults slip and thus earthquake hazards.