CRISP Type 1: Data-driven Real-time Simulation for Adaptive Control of Interdependent Infrastructure Systems

水、电、气、运输和电信等相互依存的关键基础设施的运作高度依赖于由计算机硬件和软件系统实现的传感器、数据网络和控制服务，而没有电力和足够的冷却能力，这些基础设施又无法运作。这些网络物理系统的相互依赖和相互联系的性质增加了一个基础设施的轻微干扰可能会引发多个基础设施系统的区域中断的可能性。无论是在供应方面还是在需求方面，人类对这种中断的反应都是至关重要的，而且主要受到对新出现的风险的感知和做出理性决定的能力的影响。该项目正在开发一个框架，用于建模由人类认知能力和偏好驱动的协作自适应能力，以尽量减少跨基础设施系统级联故障的风险。本项目研究的网络-物理-心理相互作用将为基础设施管理者、应急响应团队和政策制定者带来广泛的好处，使他们能够更有效地应对新出现的危机。除了研究生辅导外，该项目还为本科生和代表性不足的学生提供跨学科研究机会。研究目标是通过将仿真模型架构与人类自适应偏好相结合，提高相互依赖的关键基础设施级联故障的实时预测能力，从而在面临前所未有的风险时做出理性决策。为了实现这一目标，将承担三个相互关联的任务：(1)将使用认知任务分析技术对基础设施控制室操作员（及其所代表的组织）的认知能力和适应偏好进行建模；(2)采用系统在环框架，通过对单个动态仿真模型进行时间同步，建立电-气-水网络综合实时仿真模型；(3)将评估计算智能技术（如细胞计算网络）在预测近未来系统状态方面的能力。将特别关注基础设施运营商通过开发的模型架构可视化新威胁的能力，并将其自适应偏好嵌入预测建模框架，以使响应决策合理化。该项目将通过使用来自SCADA系统的实时监测数据的模拟模型的持续学习，推进对级联故障的空间和时间范围的理解。随着几个方面的进展，研究成果将有助于实现关键相互依存基础设施的自主自适应控制。

The functioning of interdependent critical infrastructures such as water, electricity, gas, transportation and telecommunications is highly reliant on sensors, data networks, and control services that are enabled by computer hardware and software systems, which in turn cannot function without electric power and sufficient cooling capacity. The interdependency and interconnected nature of these cyber-physical systems has increased the possibility that a minor disturbance in one infrastructure can cascade into a regional outage across several infrastructure systems. The human response to such outages, both on the supply and demand sides, is crucial and mainly influenced by the perception of emerging risk and the ability to take rational decisions. This project is developing a framework for modeling collaborative adaptive capabilities that are driven by human cognitive abilities and preferences in order to minimize the risk of cascading failures across infrastructure systems. The cyber-physical-psychological interplay investigated in this project will have widespread benefits to infrastructure managers, emergency response teams and policy makers enabling them to more effectively deal with emerging crises. This project also offers inter-disciplinary research opportunities for undergraduates and underrepresented students in addition to graduate student mentoring.The research objective is to advance real-time predictive capabilities of cascading failures across interdependent critical infrastructures by aligning the simulation model architecture with human adaptive preferences to enable rational decision making in the face of emerging unprecedented risks. Three interconnected tasks will be undertaken to achieve this objective: (1) the cognitive abilities and adaptation preferences of infrastructure control room operators (and organizations they represent) will be modeled using cognitive task analysis techniques; (2) an integrated real-time simulation model for electricity-gas-water networks will be developed through time-synchronization of individual dynamic simulation models using a system-in-the-loop framework; and (3) the capabilities of computational intelligence techniques such as cellular computational networks in predicting near-future system states will be evaluated. Particular attention will be paid to the ability of infrastructure operators to visualize an emerging threat through the developed model architecture and embedding their adaptive preferences in the predictive modeling framework to rationalize response decision making. This project will advance understanding of both spatial and temporal extents of cascading failures through continuous learning of the simulation model using real-time monitoring data from SCADA systems. With advancements on several fronts, the research outcomes will contribute to realizing autonomous adaptive control of critical interdependent infrastructures.