Dam Alteration Identification by Extended Full-Waveform Inversion and Multiphase XFEM

大坝是水电、供水和防洪的重要设施。然而，今天在运行的大坝中，有很大一部分的年龄超过50年，因此其结构完整性的真实状况尚不完全清楚。大坝的恶化可能是由于短期影响（如地震、冲击载荷等）和长期影响（如老化、物理和化学风化、疲劳等）造成的。虽然这些影响可能不会导致大坝立即倒塌/破坏，但为了确保这些结构的最佳性能，有必要快速识别可能危及大坝安全的区域。不幸的是，传统方法繁琐、耗时、昂贵且效率低下。本项目提出的方法从全波形反演（FWI）在地下勘探和岩土工程无损检测（NDT）中的成功应用中获得灵感。在这里，我们不仅打算利用FWI在提供精确和可靠的结果方面的优势，而且还打算对多相多场效应占主导地位的结构的损伤识别方法进行适应和扩展。为了提高所提出方法的鲁棒性和可靠性，将考虑结构运行周围的所有必要现象。此外，开发和应用一种新的两阶段方法将增加识别结果的稳定性（鲁棒性）和灵活性，该方法由基于扩展有限元（XFEM）的反演和随后的FWI相结合。该方法是通用的，甚至可以应用于其他反地震层析成像问题。该技术还利用了最新的计算技术，使用XFEM、FWI、FEM-BEM混合方法和试验优化设计（DoE）进行损伤识别。目标是以非破坏性的方式高效、可靠地检测大坝异常，并将全波形反演（FWI）和传统结构健康监测技术在大坝中的应用所带来的复杂性和成本降至最低。

Dams are important structures employed in the generation of hydro-electricity, provision ofwater supply and for flood defense. However, a large percentage of dams in operation today are more than 50 years old and as such the true condition of its structural integrity is not fully known.The deterioration of dams may result both from short term effects (e.g earthquakes, impact loads, etc) and from long term effects (e.g. aging, physical and chemical weathering, fatigue etc). Though these effects may not lead to an immediate collapse/failure of the dam, it is necessary, in order to ensure optimum performance of these structures, to identify quickly regions which may jeopardize the safety of the dam. Unfortunately the conventional methods are tedious, time consuming, expensive and inefficient.The method proposed in this project draws inspiration from the recorded success of full-waveform inversion (FWI) application in subsurface exploration and for non-destructive testing (NDT) in geotechnical engineering. Here not only do we intend exploiting the advantages of FWI in providing precise and reliable results but also the adaptation and extension of the method for damage identification in structures with dominant multiphase-multifield effects. In order to increasethe robustness and reliability of the proposed method, all the necessary phenomena surrounding the operation of the structure will be incorporated. Additionally, the stability (robustness) and the flexibility of the identification results will be increased be developing and applying a novel two-stage approach which consists of a combination of an extended Finite Element (XFEM)-based inversion and a subsequent FWI. The approach is general and may be applied even to other inverse seismic tomography problems. The proposed technique also takes advantage of recent advancements in computing technology, works in damage identification using XFEM, FWI, FEM-BEM hybridmethods and the optimal design of experiments (DoE). The goal is to efficiently and reliably detect anomalies in dams in a non-destructive manner and to minimize the sophistication and cost incurred by the application of full-waveform inversion (FWI) and conventional structural health monitoring techniques in dams.