Overview: This presentation uses hetero-functional graph theory to understand the asset-level resilience and national level sustainability of the American Multi-Modal Energy System (AMES)
Context: The challenge of global climate change necessitates a fundamental and holistic re-design of the AMES. Traditionally, the electric grid, natural gas, oil, and coal systems have been studied individually. More recent efforts have tried to study pairs of these systems but for a number of theoretical and practical reasons such efforts have fallen short of combining these four systems into an integrated model.
Purpose: The "American Multi-Modal Energy System Synthetic and Simulated Data (AMES-3D)" project funded by the (American) National Science Foundation seeks to fill this gap and produce an open-source, physically-informed, machine-learning, structural and behavioral model of the AMES for potential use and expansion by the broader scientific community. This presentation seeks to introduce the audience to the AMES' data, model and development in several steps.
Approach: Hetero-functional graph theory is briefly introduced as the scientific basis for producing structural and behavioral models where an arbitrary number of infrastructure networks of arbitrary topology are connected arbitrarily.
Insights: We briefly compare the network statistics and resilience measures of the (traditional) formal and hetero-functional graphs of the American electric power system. Such a comparison motivates the use of hetero-functional graphs in systems with multiple energy carriers and facilities with fundamentally heterogeneous function. In particular, we show how such systems require 3D degree distributions that differentiate the processes and energy carriers in a hetero-functional graph. Finally, we demonstrate how such 3D degree distributions in the states of New York, California, Texas, and the entire United States differ as a result of their fundamentally different geographies and policies. The presentation concludes with several directions for future work using the AMES model and the hetero-functional graph theory toolbox.