CAREER: Assessment of Infrastructure Risk Under Natural Disasters in a Multiscale Probabilistic Framework

这项教师早期职业发展（Career）奖的研究目标是创建一种新的地震期间区域地面震动的概率描述，用一组具有代表性的场景取代一个地区几乎无限数量的可能的灾难场景。受自然灾害影响的基础设施系统的分析受到限制，因为需要捕捉复杂的网络效应，而且难以描述灾害对大区域影响的复杂变化。当前的基础设施分析通常只对单一灾难情景下的破坏进行估计。很少考虑到这种情况发生的可能性，从而妨碍了基于成本效益分析的风险管理。这些限制是由于在可能发生的大量灾难场景下，不可能对复杂网络进行计算中断。然后将开发用于地面震动和运输网络中断的多尺度模型，以便尽可能地促进粗尺度建模，并根据需要进行更精细的建模。这种多尺度处理导致了量化灾害和基础设施网络特征空间尺度的新方法。然后，这些分析技术将应用于其他生命线网络和危害，以更全面地了解特征空间尺度与基础设施破坏模式的关系。在过去的十年中，单个结构的基于风险（基于性能）的建模一直是结构工程的主要推动力，而多尺度建模已经成为计算力学中的一种革命性方法。将这些概念扩展到基础设施系统的分析，将为基础设施网络的精细尺度特征如何影响区域尺度上的性能提供新的见解。自然灾害中关键的生命线中断会增加痛苦并阻碍灾后恢复，但升级这些系统的资源有限，因此必须谨慎地确定减少风险的目标。这项研究的见解将指导有关基础设施开发和维护的决策制定。随着决策者利用这些知识，他们最终将生产出更安全、更有弹性和更经济的基础设施。该项目的扩展目标将促进其他研究人员的教育和培训（通过生产开源软件和网络内容），本科生（通过为未来的工程专业学生开发基于项目的研讨会）和高中生（通过继续参与斯坦福大学的高中生暑期研究实习计划）。对这些拓展项目的评估将通过利用现有的大学项目来跟踪参与者的教育/职业道路来实现。

The research objective of this Faculty Early Career Development (CAREER) award is to create a new probabilistic description of regional ground shaking during earthquakes, which replaces the near-infinite number of possible disaster scenarios in a region by a small set of representative scenarios. Analysis of infrastructure systems subject to natural disasters is limited by the need to capture complex network effects and difficulties in describing the complex variation of the disaster's impact over a large region. Current infrastructure analyses often estimate disruption given a single scenario disaster. The probability of occurrence of that scenario is rarely incorporated, hindering risk management based on cost-benefit analysis. These limitations are caused by the impossibility of computing disruptions to complex networks under the huge number of possible disaster scenarios that might occur. Multiscale models for ground shaking and transportation network disruption will then be developed to facilitate coarse scale modeling wherever possible, coupled with finer-scale modeling as needed. This multiscale treatment leads to new methods for quantifying the characteristic spatial scale of disasters and infrastructure networks. These analysis techniques will then be applied to other lifeline networks and hazards to more generally understand how characteristic spatial scales relate to infrastructure disruption patterns. Risk-based (performance-based) modeling of individual structures has been a major thrust of structural engineering in the past decade, and multiscale modeling has been a revolutionary approach in computational mechanics. Extension of these concepts to analysis of infrastructure systems will provide new insights into how fine-scale features of infrastructure networks affect performance at the regional scale.Critical lifeline disruption in natural disasters increases suffering and impedes post-disaster recovery, but resources to upgrade these systems are limited and so risk reduction efforts must be carefully targeted. Insights from this research will guide decision making regarding infrastructure development and maintenance. As decision-makers utilize this knowledge, they will ultimately produce safer, more resilient and more economical infrastructure. The project's outreach objectives will facilitate education and training of other researchers (through the production of open-source software and web content), undergraduates (through development of a project-based seminar for prospective engineering majors) and high school students (through continued participation in Stanford's Summer Research Internship Program for High School Students). Assessment of these outreach programs will be achieved by leveraging existing university programs to track the educational/career paths of participants.