DESC0023601

C55-01b-270620Infrastructure systems such as electricity distribution and transmission grids, transportation networks, and water or oil/gas distribution pipelines are critical from an economic and national security standpoint. Expedited risk quantification of infrastructure network systems is of utmost importance for resiliency assessment against natural or man-made hazards. Currently available techniques for network probabilistic risk assessment suffer from large-scale computational challenges. A practical means for expedient and accurate quantification of risk associated with infrastructure networks is needed to safeguard critical infrastructure systems against potential high-risk threats. Protection Engineering Consultants, LLC (PEC) proposes to develop novel technology for expedited risk analysis of large electric grid systems. The proposed technology will utilize a Random Matrix Theory (RMT)- based probabilistic modeling technique to exploit the structure of large, sparse matrices representing electric grids. The framework will leverage various ASCR-funded numerical packages thus offering improved efficiency relative to the existing state-of-the-practice commercial platforms. The proposed framework will be encoded into a suite of numerical tools – Network Risk Assessment Toolkit (NetRAT) for commercial deployment. During Phase I, PEC, with support from consultant Dr. Michael Shields, will develop Network-level Risk Assessment Module (NetLevRAM) that will utilize RMT-based technique and leverage various ASCR-funded libraries to exploit the structure of large, sparse matrices representing energy grids. A part of the technical approach will focus on careful manipulation of matrix algebra computations to estimate eigenvalue bounds for network risk assessment. Another focus will be on utilization of various ASCR-funded numerical packages, e.g., STRUMPACK, Trilinos, Dakota, and OpenMPI, to gain computational efficiency, and their integration with opensource package UQpy to develop an extensible commercial module ECMUQpy. The performance of the toolkit will be assessed across several computational platforms using testbed data. The feasibility of our proposed technology will be demonstrated during Phase I, and a Phase II effort will focus on advancing our network risk analysis algorithms, expanding our validation dataset, and building out the front-end graphical user interface (GUI) for the field-deployable platform. Our finished toolkit will be at market-ready status at the end of Phase II; interested users will be broad-ranging, from field engineers to energy grid stakeholders and local emergency response/planning personnel. We also envision a potentially broader user base comprising military and civilian critical infrastructure owners/users (e.g., NNSA sites, commercial NPPs, DOE, DOT, EPA).