节点文献

强地震形变过程的有限元数值模拟及其动力学环境的研究

FEM Numerical Simulation of Deformation of Strong Earthquake and Geodynamic Environment

【作者】 邵志刚

【导师】 傅容珊;

【作者基本信息】 中国科学技术大学 , 固体地球物理学, 2007, 博士

【摘要】 通过全球定位系统(GPS)和卫星孔径干涉雷达(InSAR)等大地测量观测强地震引起的同震和震后形变,结合地质构造条件分析,利用有限元数值模拟,使得我们可以对地震区域地球动力学环境有更加详细地了解,进而对地震孕育和地壳动力学演化等地质过程有更深入的认识和理解。本文利用固体力学理论,针对不同粘弹介质、孔隙介质以及粘弹—孔隙等介质的本构方程,建立和导出了其适用于FEPG(有限元程序生成系统)的虚功方程。以GPS观测所得地表形变做约束,用FEPG(有限元程序生成系统)实现了强地震引起的同震和震后形变有限元数值模拟,并掘此分析了以介质物性为中心的强地震区域的环境动力学特征。本文在前人震后形变分析和模拟基础之上,根据粘弹松弛和流体调整所引起的震后形变的物理含义,建立孔隙—粘弹本构方程以及适应于FEPG平台的有限元数值模拟所必需的虚功方程。同时针对震后形变所表现出的粘弹松弛特点,以Burgers体粘弹模型特点以及本构方程出发,建立相应虚功方程,实现了以Burgers体为粘弹介质模型的震后形变有限元数值模拟。以Burgers体为粘弹介质模型探讨了强地震引起的三个阶段的形变效应,同震的瞬时弹性瞬态响应、呈指数衰减的短期形变响应和线性增加的长期稳态形变响应。在理论模拟和分析的基础上,本文以实际地震为例详细讨论地震引起的同震和震后形变过程。本文以1999年Mw7.6级集集地震为例,讨论不同的位错模型和介质模型(均匀、层状均匀及横向不均匀等)对同震形变的影响。通过数值模拟分析本文认为,在条件允许的情况,在讨论地震形变时有必要考虑小断层位错模型和更加接近实际的非均匀介质模型,计算和探讨了集集地震引起的库仑应力变化及其与集集地震余震对应关系。本文以2001年Ms8.1级地震为例,以GPS观测形变为约束,通过有限元数值模拟,解释GPS所观测的震后形变表现出较为奇特的区域特征。用数值模拟结合网格搜索法确定了羌塘块体和柴达木盆地下地壳的粘滞系数分别约为5.0×1017Pa.s,9.0×1018pa.s相差十余倍。正是由于昆仑断裂带南北两侧下地壳粘滞系数的这十余倍的差异引起了断层两侧震后形变格局明显不同,这也可能是羌塘块体和柴达木盆地表现为不同现代地壳运动的直接原因。与此同时,本文还通过数值模拟定性地讨论了断层北侧地表形变在震后短期内转向的现象。我们发现,对于靠近断裂附近的测点上的形变转向也许是粘弹性松弛和孔隙流体调整共同作用的结果,而单独的粘弹性松弛或者孔隙流体调整不能引起这种转向。所以在分析短期震后形变时综合考虑粘弹—松弛介质物性和孔隙流体调整很有必要。本文探讨了震后不同时间尺度的形变机理问题。研究表明,简单的粘弹介质模型(Maxwell体)不能同时解释震后短期和长期的形变过程,该模型表现出对同一地震同一地区而言,使用震后不同时期内形变资料,将得到不同的介质粘滞性质。本文证明了。以Burgers体为粘弹介质模型可以较好地解决使用Maxwell体等简单的粘弹介质模型所不能解释的震后形变过程随时间变化的问题,即模拟预测表现出来的震后短期和长期形变不协调的现象。通过数值模拟分析了1960年Mw9.5级智利地震震后形变过程,较成功地解释了该强烈地震震后短期和长期形变观测数据。本文的研究表明,以观测地表形变为约束,结合实际地质条件,选取合理的介质物性模型,构建有限元模型的框架,利用有限元数值模拟的方法,研究强地震形变过程和规律,是分析了以介质物性为中心的强地震区域的环境动力学特征的有效方法。

【Abstract】 The coseismic and postseismic deformation observed by GPS and InSAR is used as the constraints in displacement of numerical simulation. We can learn more information about the regional geodynamic environment by the calculation and the geological condition and then research the some geological processes, such as, the gestation earthquake and dynamic evolution in crust. In this paper, we get the virtual work equations of FEPG (Finite Element Program Generator) on the base of viscoelastic, poroelastic and poro-viscoelastic constitutive relationship. We calculate the coseismic and postseismic deformation using FEM. We use the GPS data of crust deformation as the constraints and obtain different Fortran programs of different questions by FEPG, and study the role of physical characters in regional geodynamic behavior.In this paper, firstly, we get the constitutive relations and other equations used in FEM calculation by FEPG. These equations are based on the theories of viscoelastic relaxation and poroelastic rebound and other researches of postseismic displacement. As the characters of observed deformation of earthquake, we use Burgers body as the viscoelastic medium, get virtual work equations used in FEPG and simulate the postseismic deformation. The postseismic displacements based on the Burgers body include three kinds of deformations, which are the elastic coseismic deformation, the exponentially-decaying short-time deformation, the linearly-increasing steady-state long-time deformation.Based on these theoretical simulation and analyses, we study the coseismic and postseismic deformation of strong earthquake in detailed.In this study, we take 1999 Mw7.6 Chi-Chi earthquake for example to study the effect of sub-fault model and layered geological crust model to the coseismic deformation. From the simulation, it is essential to use the sub-fault model composed of many small faults and layered geological crust inhomogeneous model. And then calculate the Coulumb stress change caused by Chi-Chi earthquake and analyze the relation of the Coulumb stress change and the spatial distribution of the aftersocks. After the 2001 Ms 8.1 Kunlun earthquake, postseismic displacements were resolved with GPS. The observed data showed that there were many differences between the deformations on south and north side of the Kunlun fault. And after short-time adjustment, the deformation on the north side changed direction from westward to eastward. These GPS data are used as the constraints in FEM modeling of postseismic deformation. Using 3D poro-viscoelastic model, the modeling results show that the horizontal inhomogeneous viscosities in lower crust can play an important part in geodynamic behavior. The best-fitting viscosities of lower crust about are 5.0×1017Pa.s, 9.0×1018Pa.s, respectively in Qiangtang block and Qaidam basin. This difference of viscosity of two sides of Kunlun fault is the direct reason for the differences of postseismic deformation between these two blocks. The simulation suggests that the postseismic deformation after 2001 Kunlun Ms8.1 earthquake may be caused not only by the viscoelastic relaxation but also by poroelastic rebound.And we research the cause of the short-time and long-time postseismic deformation after one earthquake. Firstly we simulate the postseismic deformation by Maxwell model. If the viscosity is lower, the relaxation is going to balance, so there is no long-term postseismic deformation. On the other hand, if the viscosity is higher, the rate of strain may be a constant. Take 1960 Chile earthquake as example, the postseismic deformation modeled with Burgers model includes coseismic deformation, transient postseismic deformation and long-term postseismic deformation. As to the calculated postseismic deformation of 1960 Chile earthquake, there is no uncoordinated phenomenon caused by the Maxwell model, when we take the Burgers model as viscoelastic medium.From our research, based on the geological condition and physical characters of crust, we can study the seismic deformation of strong earthquake using FEM numerical simulation. Then we can analyze the role of physical characters in regional geodynamic behavior by this efficient the method.

  • 【分类号】P315.3
  • 【下载频次】550
节点文献中: 

本文链接的文献网络图示:

本文的引文网络