Reliability Assessment of the Serviceability Performance of Geotechnical Structures Using Neural Networks

在本研究项目中，探讨了在基于可靠性的框架中集成有限元方法和神经网络技术来评估土工结构（如支撑开挖系统）的使用可靠性的可行性。城市地区的深基坑开挖是一项复杂的施工作业，经常会造成不可接受的地面移动和对邻近建筑物的破坏。对潜在的地面位移和相邻建筑物的破坏进行可靠性评估，有助于此类岩土结构的安全设计和施工。根据可靠性分析的结果，建立了进行可靠性评估的框架和程序。还制定了使用现场监测/测量数据更新可靠性的规定。从智力上讲，这项研究可能会导致一种显著改进的方法，设计和施工的深支撑挖掘，既健全又安全的环境。开发的方法将有限元法、神经网络和观测方法集成到基于可靠性的框架中，被认为是创新的。该方法如果成功开发，将是复杂岩土结构可靠性评估的重大突破。拟议研究的更广泛影响也很重要。首先，本研究有可能改进深基坑支护的设计和施工方法，使其既健康又对环境安全。研究结果可能会被工程界采用，从而为保护公众及其财产做出重大贡献。第二，研究结果将通过多种方式广泛传播。本项目强调的技术转让是综合研究和教育的总目标中的一个重要因素。第三，拟议的研究将通过其强大的国内和国际合作，加强克莱姆森大学的研究基础设施。第四，拟议的研究将以一种显著的方式扩大代表性不足群体的参与，通过分包合同让一名来自本科院校的女性工程教师和克莱姆森大学（Clemson University）的一名女性博士后研究助理参与，并从代表性不足的群体中招募和指导研究生和本科生助理。第五，也是最后一点，但同样重要的是，拟议的研究将通过纳入有助于研究和教育一体化总体目标的各种活动，在促进教学、培训和学习的同时，推进知识状态。

In this research project the feasibility of integrating finite element method and neural network technologies in a reliability-based framework for assessing serviceability reliability of geotechnical structures such as braced excavation systems is explored. A deep braced excavation in an urban area is a complex construction operation that often causes unacceptable ground movement and damage to adjacent buildings. Reliability assessment of potential ground displacement and damage to adjacent buildings aids in the safe design and construction of such geotechnical structures. The framework and procedures for conducting reliability assessment are established based on the results of reliability analyses. Provisions for reliability updating using field monitoring/measurement data are also established. Intellectually, this research may lead to a significantly improved method for design and construction of deep braced excavations that is both sound and safe for the environment. The developed methodology, which integrates the finite element method, neural networks, and the observational method into a reliability-based framework, is considered innovative. This methodology, if successfully developed, represents a significant breakthrough in assessing reliability of complex geotechnical structures. The broader impacts of the proposed research are also significant. First, this research is likely to result in an improved method for design and construction of deep braced excavation that is both sound and safe for the environment. The results of the research may be adopted by the engineering community, and thus, contribute in a significant way to protecting the public and their properties. Second, the results of the research will be disseminated broadly through a number of ways. Technology transfer, which is emphasized in this project, is an important element in the overall objective of integrating research and education. Third, the proposed research will enhance the infrastructure for research at Clemson University through its strong national and international collaboration. Fourth, the proposed research will broaden participation of underrepresented groups in a significant way, by participation of a female engineering faculty from an undergraduate institution through a subcontract and a female postdoctoral research assistant at Clemson University, and by recruiting and mentoring graduate and undergraduate assistants from underrepresented groups. Fifth and lastly, but equally important, the proposed research will advance the state of knowledge while promoting teaching, training, and learning by incorporating various activities that contribute to the overall objective of research and education integration.