LEAP-HI: Optimal Design and Life-Long Adaptation of Civil Infrastructure in a Changing and Uncertain Environment for a Sustainable Future

与天气有关的事件的频率和强度不断增加，对国家建筑环境的运作、完整性和寿命产生了负面影响，造成了重大的经济和其他损失。这部分是因为建筑物和民用基础设施一直是并将继续是根据历史数据和气候趋势设计的，而这些数据和趋势不再准确反映未来预期的相关不确定性。此外，虽然在设计、建造和维护国家建筑物和基础设施方面进行了大量投资，但这些活动都是独立管理的，导致了多余的成本和其他负面影响。最终，这些因素的融合导致美国人民和经济面临越来越大的风险。为了应对这些挑战，美国繁荣、健康和基础设施领先工程（LEAP-HI）项目将研究一个新的框架，改变国家建筑和基础设施的设计、维护和调整方式，以促进可持续未来的明智和负责任的决策。 该框架平衡了安全性和弹性与资源消耗和环境影响，同时考虑了当前和未来的不确定性。该项目将研究一个计算框架，用于基于概率优化建筑环境的设计、维护和终身适应生命周期指标，同时考虑不确定和非稳态的生命周期需求。该项目的目标是：根据最新的气候科学并考虑到各种不确定性，量化各种尺度的水文气象灾害;将这些不确定性传播到生命周期决策标准中（例如成本、环境影响、（与灾害有关的损失）通过在数学上考虑维护和自适应动作，通过使用深度强化解决的部分可观察马尔可夫决策过程问题，学习;设计一种方法和工具，通过调整平均风险和随机优势概念，针对多个生命周期决策标准，有效和严格地评估和比较一套设计和适应备选方案;了解各种不确定性来源如何影响区分设计备选方案的能力;确定在哪些情景下绿色基础设施组件可导致最佳设计和适应解决方案;并展示新的框架和关于实际设计情景的调查结果，例如受到洪水和老化影响的河流桥梁，受到气温上升和自然及运营压力威胁的城市中层建筑，以及受到海平面上升威胁的港口和基础设施资产系统。该项目将通过研讨会和教育活动，向决策者和最终用户传播新的知识和方法，从而提高应对非平稳气候的工程领导能力。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The increasing frequency and intensity of weather-related events is negatively affecting the operation, integrity, and life-span of the Nation’s built environment, resulting in significant economic and other losses. This is due, in part, to the fact that buildings and civil infrastructure have been, and continue to be, designed based on historical data and climate trends that no longer accurately reflect the relevant uncertainties about what can be expected in the future. Furthermore, although substantial investments are made in designing, constructing, and maintaining the Nation’s buildings and infrastructure, each of these activities is managed independently, resulting in redundant costs and other negative impacts. Ultimately, the convergence of these factors is leading to increasing risks for the American people and economy. To address these challenges, this Leading Engineering for America's Prosperity, Health, and Infrastructure (LEAP-HI) project will research a new framework that transforms the way in which the Nation’s buildings and infrastructure are designed, maintained, and adapted to facilitate informed and responsible decision-making for a sustainable future. That framework balances safety and resilience with resource consumption and environmental impacts, while accounting for current and future uncertainties.The project will research a computational framework for simultaneously optimizing the design, maintenance, and life-long adaptation of the built environment, based on probabilistic life-cycle metrics, while accounting for uncertain and non-stationary life-cycle demands. The objectives of the project are to: quantify hydrometeorological hazards at various scales based on the most current climate science and considering a variety of uncertainties; propagate those uncertainties to the life-cycle decision criteria (e.g. costs, environmental impacts, hazard related losses) by mathematically considering the maintenance and adaptation actions through a partially observable Markov decision process problem solved with Deep Reinforcement Learning; devise a methodology and tools for efficiently and rigorously evaluating and comparing a set of design and adaptation alternatives with respect to multiple life-cycle decision criteria by adapting concepts of mean-risk and stochastic dominance; understand how various sources of uncertainty affect the ability to discriminate among design alternatives; determine under which scenarios green infrastructure components can lead to optimal design and adaptation solutions; and demonstrate the new framework and findings on practical design scenarios, such as a riverine bridge threatened by flooding and age-related deterioration, an urban mid-rise building threatened by increasing temperatures and natural and operational stressors, and a port and systems of infrastructure assets threatened by sea level rise. The project will improve engineering leadership in response to a non-stationary climate by building research capacity and outreach through workshops and educational efforts that disseminate the new knowledge and methods to decision-makers and end-users.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.