Resilient and Sustainable Infrastructure Systems Design and Analysis

近年来，气候变化的影响变得更加频繁、明显和极端，科学界预测的未来影响越来越不利和不可预测。因此，与经济、环境和社会考虑相关的基础设施系统的工程设计相关的挑战和复杂性增加了。这些复杂性自然来自于这些系统在其整个生命周期中所需的操作参数的变化，这些参数会影响几个重要指标的性能。考虑到相关的风险和未来趋势数据的缺乏，有必要修改复杂基础设施系统的设计方法。此外，建筑环境已经被确定为工业部门，它对我们的全球可持续性施加了最大的压力，因为这些结构的建造和运营消耗了相关的能源、材料和资源。从设计过程的早期阶段开始进行灵活性设计，可以帮助显著提高这些系统的可持续性、弹性和适应性潜力，这些系统具有与联合国可持续发展目标（UN SDGs）相关的重要正外部性。然而，传统的建模和设计方法往往需要比可获得的准确性更多的数据，并且存在与计算强度和建模过程的复杂性相关的主要缺点。因此，它们往往只针对一个目标，在设计决策过程的早期阶段是不可行的。考虑到多学科相互作用在最终确定寿命性能方面的重要性，需要一种更广泛适用的工具。为了确保投资带来弹性和可持续系统的部署，成功的灵活性分析还必须在考虑定性因素的同时，通过定量措施为决策者提供可信度。因此，有必要开发一种新的数据驱动方法，该方法可以适用于跨基础设施系统，用于实物期权分析和灵活性设计的决策规则方法。这一体系在提高全球适应气候变化能力方面的长期可持续性和有效性与联合国可持续发展目标的整合密切相关，而不是简单的传统经济价值优化方法。因此，机器学习（ML）为这些问题提供了一个巨大的潜在解决方案，因为它可以允许快速预测和参数变化，从而更有效地探索替代设计，并最终在各种场景中优化寿命性能。使用机器学习可以帮助降低设计过程中的风险和不确定性，并且比传统方法的计算成本更低，从而产生更强、适应性更强的基础设施系统，在面对气候变化时更具弹性和经济可行性。项目目标本项目提出的核心问题是：如何将机器学习应用于基础设施系统的设计和分析，通过利用灵活性来增强弹性、可持续性和经济绩效？项目具体的研究问题可能是：1)不同决策层面的利益相关者的角色如何随时间和不确定性而变化？什么时候在设计过程中考虑不同的灵活性是最优的？2)工程师如何识别更适合灵活性和简单性的替代设计或配置？在不确定的情况下，要完成这样的任务需要什么样的输入？3)如何将这种数据驱动的方法应用于发展中国家，在缺乏当地结构的情况下，通过优化分布式发电能力和布局来增强能源获取和政策？

The impact of climate change has become more frequent, visible and extreme in recent years with increasingly adverse and unpredictable future effects predicted by the scientific community. As a result, there are increased challenges and complexities associated with the engineering design of infrastructure systems related to economic, environmental and social considerations. These complexities naturally arise from the changing operational parameters required by these systems throughout their lifetime which impact performance across several important metrics. Given the associated risk and lack of future trend data, there is a need to revise the design approach for complex infrastructure systems. Furthermore, the built environment has already been identified as the industrial sector which puts the most pressure on our global sustainability resulting from the associated energy, material and resource consumption of the construction and operation of these structures. Designing for flexibility from the early stages of the design process can help significantly increase the sustainability, resilience, and potential for adaptability of these systems with important positive externalities related to the United Nations Sustainable Development Goals (UN SDGs). Traditional modelling and design approaches, however, tend to require more data than is available for accuracy and present major drawbacks related to computational intensity and complexity of the modelling process. Consequently, they tend to target only one objective and are not feasible to use in the early stages of the design decision making process. Given the importance of multidisciplinary interactions in ultimately determining lifetime performance, there is a demand for a more broadly applicable tool. To ensure that investments lead to the deployment of resilient and sustainable systems, a successful flexibility analysis must also provide credibility for decision makers through quantitative measures while accounting for qualitative factors. It is therefore essential to develop a new data driven methodology which can be applicable across infrastructure systems for a decision rule approach to real option analysis and flexibility design. The long-term sustainability and effectiveness of such a system in increasing our ability for climate change adaptation globally is closely linked to integration with the UN SDGs rather than simple traditional economic value optimization methods. Machine learning (ML) thus offers a great potential solution to these issues as it can allow fast prediction and parameter variation for more efficient exploration of alternative designs and ultimately optimized lifetime performance across a wide range of scenarios. Using ML can help to mitigate risks and uncertainty in the design process at lower computational costs than traditional approaches, yielding enhanced and adaptable infrastructure systems which are more resilient and economically viable in the face of climate change.Project ObjectivesThe central question this project asks is: How can ML be applied to the design and analysis of infrastructure systems by leveraging flexibility to enable enhanced resilience, sustainability and economic performance? Project specific research questions may be:1)How does the role of stakeholders at different levels of the decision-making change over time and with uncertainty? When is it most optimal to include different flexibility considerations in the design process?2)How can engineers identify alternative designs or configurations which are better suited for flexibility with simplicity and early on? What sort of input would be necessary to achieve such a task under uncertainty?3)How can this data driven methodology be applied to enable enhanced energy access and policy in developing countries through optimized distributed generation capacity and placement in the absence of local structures?