NEESR-SG: High-fidelity site characterization by experimentation, field observation, and inversion-based modeling

摘要0619078该奖项是NSF 06-504计划征求“乔治E。小布朗地震工程模拟研究网络（NEESR）”竞赛。该项目小组包括卡内基梅隆大学、格鲁吉亚理工学院、加州大学圣巴巴拉分校、德克萨斯大学埃尔帕索分校和德克萨斯大学奥斯汀分校。 本研究的主要目标是开发估算单个站点和完整盆地的地质结构和力学性质的能力，并在加纳谷井下阵列（GVDA）和整个加纳谷的nees@UCSB站点上展示这种能力。具体而言，这种高保真估计将基于整合：（a）使用德克萨斯大学奥斯汀分校的NEES设备的现场动态激励;（B）来自新强震和宽带传感器网络的地震记录;（c）基于偏微分方程约束优化的新反演方法。该项目代表了一个无与伦比的机会，将最先进的现场实验与最先进的计算工具相结合，以便以直到最近还无法实现的分辨率和长度尺度对地下进行成像。选择这两个NEES地点的原因如下：（1）NEES@UTexas地点的移动的振动台可以在广泛的频率和载荷水平下施加载荷。收集到的数据加上地震记录，就可以估计一次波和剪切波的速度。此外，土壤阻尼将包括在反演模型中，与两个速度同时估计;（2）GVDA（nees@UCSB）是位于加州南部高地震区狭窄山谷中的试验场地，非常适合监测地面运动。在过去的15年里，在该地点和整个山谷的其他自由表面位置记录的数百次小地震使其成为区域深部结构反演的宝贵数据来源，并通过现场测试对现场反演进行独立验证。为了提高反演模型的保真度，该数据集将增加来自EarthScope的USAray组件的仪器的数据，这些仪器将在与主动测试期间重叠的时间段内部署。现场测试将为GVDA现场和整个山谷的上层特征提供数据。这些试验的观测结果，包括井下和自由表面的观测结果，将与地震观测结果一起沿着用于获得二维和三维高保真剖面。此外，这些观测结果将与表面波频谱分析（SASW）方法一起使用，以获得测试位置的剪切波速度剖面。该速度剖面将用作基于全局优化的逆局部和区域模型的盲测验证套件的一部分。 虽然综合方法将适用于南部加州的一个特定地区，但其结果将适用于美国和国外的许多其他地区。 新的综合方法将同样适用于波用于探测的类似问题，例如石油和天然气勘探社区中出现的问题，医学成像应用，其他基础设施状况评估问题，例如结构缺陷识别问题，评估候选核废料场的适用性，甚至是远程控制的行星探索任务，其中场地特征是至关重要的。由于土壤在基础设施设计中发挥着关键作用，可靠的土壤特性鉴定方法的加速可用性将对公共安全和福利产生直接影响。

Abstract 0619078 This award is an outcome of the NSF 06-504 program solicitation "George E. Brown, Jr. Network for Earthquake Engineering Simulation Research (NEESR)" competition. The project team includes Carnegie Mellon University; Georgia Institute of Technology; the University of California at Santa Barbara; the University of Texas, El Paso; and The University of Texas at Austin. The main objective of this research is to develop the capability for estimating the geological structure and mechanical properties both of individual sites and of complete basins, and to demonstrate this capability on the nees@UCSB site at the Garner Valley Downhole Array (GVDA) and the entire Garner Valley. Specifically, this high-fidelity estimation will be based on integrating: (a) in-situ dynamic excitation using the NEES equipment at the University of Texas at Austin (nees@UTexas); (b) earthquake records from new strong-motion and broadband sensor networks; and (c) new inversion methods based on partial-differential-equation (PDE)-constrained optimization. This project represents an unparalleled opportunity to couple state-of-the-art field experimentation with state-of-the-art computational tools for the purpose of imaging the subsurface at resolutions and length scales until recently unattainable. These two NEES sites were selected for the following reasons: (1) The mobile shakers at the nees@UTexas site can apply loads at a wide range of frequencies and loading levels. Coupled with earthquake records, the data collected will permit the estimation of the primary- and shear-wave velocities. In addition, soil damping will be included in the inversion models, to be estimated simultaneously with the two velocities; and (2) The GVDA (nees@UCSB) is a test site located in a narrow valley in a highly seismic region in southern California, which is ideally suited for monitoring ground motion. The hundreds of small earthquakes that have been recorded at the site and at other free surface locations throughout this valley in the last 15 years make it an invaluable source of data for the regional deep structure inversions, and for an independent verification of site inversion through field tests. To increase the fidelity of the inverted models, this dataset will be augmented with data from instruments of the USArray component of EarthScope that will be deployed over periods of time that will overlap the periods of active testing. In-situ tests will provide data for characterizing the upper layers at the GVDA site and throughout the valley. Observations from these tests, both downhole and on the free surface, will be used, along with earthquake observations, for obtaining two-dimensional and three-dimensional, high-fidelity profiles. In addition, these observations will be used with the spectral-analysis-of-surface-waves (SASW) method to obtain the shear-wave velocity profile at the test locations. This velocity profile will be used as part of the blind-test validation suite of the global optimization-based inverse local and regional models. While the integrated methodology will be applied to a specific region in southern California, the results will be applicable to many other regions in the United States and abroad. The new integrated methodology will be equally applicable to similar problems where waves are used for probing, such as those arising in the oil and gas exploration communities, medical imaging applications, other infrastructure condition-assessment problems such as in structural flaw identification problems, assessing the suitability of candidate nuclear waste sites, and even remotely-controlled planetary exploration missions where site characterization is of paramount importance. Because of the critical role that soils play in infrastructure design, the accelerated availability of reliable soil characterization methods will have a direct impact on public safety and welfare.