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

点击翻译按钮获取中文摘要

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.