岩石和地壳非均匀应变场演化规律的揭示和灾变破坏前兆分析

Geological media and civil engineering materials, such as rocks, are always heterogeneous. However the induced heterogeneous strain-field evolution is rarely examined in classical mechanics theory. Therefore, some practically significant phenomenoa that may be controlled by the evolution are poorly understood, for example, localization and catastrophic rupture of heterogeneous brittle media. This may be a key why we still cannot effectively predict some serious disaster, such as earthquake. In the project, we are planning to focus on the reconstruction of the heterogeneous strain -field, based on the measured displacements of rock samples and available GPS data. The measured displacement data of rock samples in Lab tests will be used to form and check the methods for processing and analyzing the evolution of heterogeneous strain-field, whose characteristics will then be investigated carefully; while the GPS data will be used to check the precursors to massive catastrophic earthquake. For this sake, firstly, we are trying to form an algorithm for the reconstruction of the displacement-field from the observed displacement with inhomogeneously spatial and temporal distributions (particularly the GPS data). The algorithm should satisfy the requirement for the study of the evolution of heterogeneous strain-field. Then based on the reconstructed results, the characteristics of heterogeneous strain-field evolution and the universal features of localization will be investigated in order to find out whether precursors to catastrophic rupture observed in Lab experiments (e.g. power-law singularity) also present before earthquake.

岩石等土木工程和地质介质大多是非均匀的，经典的力学理论很少涉及这类材料非均匀性导致的非均匀应变场的演化，及其所控制的尚未被充分理解、并且有重要实际价值的现象，例如：地壳变形和损伤局部化和突发灾变破坏，这很可能是许多重大灾害，如地震，尚不能有效预测的一个关键。 本申请项目拟以室内岩石位移场测量和已有的大地震震前GPS测量数据为对象，前者主要用于建立和校核非均匀应变场演化的数据处理和分析方法，揭示非均匀应变场演化的特征；后者主要用于做大规模突发灾变破坏预测的后验检验。 要实现这个目标，首先要建立基于非均匀时间和空间间隔的位移测量（特别是GPS测量）的应变场演化的重构方法，该重构场要能满足揭示非均匀的应变场的演化规律的要求。基于此，要进一步分析应变场非均匀性演化的特征和局部化的共性特征，比较地震前是否会和室内实验类似出现幂律奇异性等前兆。

灾变破坏是岩石类非均匀脆性材料破坏的一种典型形式，大地震的发生往往也属于这种破坏类型。在灾变点附近，许多物理量（如应变、位移、声发射事件数、能量释放率等）呈现出加速演化的行为，并且其响应量在破坏点处按幂律规律发散到无穷大。确定幂指数的特征，对于准确地建立灾变预测方法至关重要。.本项目通过大理岩和花岗岩试样的准静态单轴压缩实验，发现灾变岩样的整体响应函数R = du/dU在灾变破坏前呈现出幂律奇异性，R的幂指数在−1和−1/2之间分布。在理论上阐明了响应函数R的幂律奇异性是灾变破坏能量准则的必然结果，而幂指数的分布范围可由幂函数近似来分析。.基于幂指数的分布范围，提出了一个实时的灾变预测方法，得到了破坏点的一个上界及相应的预测区间。.通过数字散斑相关方法测量了试样的表面变形场，观察到了灾变破坏前试样表面变形场的变形局部化演化。灾变破坏前，局部化区域内的应变演化呈现出幂律的奇异性，幂律指数接近−1。基于变形场的局部信息建立起了灾变点预测方法。.利用汶川地震（2012年5月12日）震前4期GPS流动站数据和同震的位移数据重构了龙门山破裂带附近的应变场，观察到了在汶川地震发生前，垂直破裂带方向的压应变分量呈现局部化现象。利用GPS连续站构成的计算网格，观察到了汶川地震的幂律前兆。基于此，将建立的实时预测方法应用于对汶川地震发震时间的后验性预测，实时地给出了发震时间的上界。本文的研究表明，利用局部区域的GPS数据进行大地震的预测是可行的，并且不需进行变形场的重构，这给应用带来了极大的灵活性。