Directed Graph Signal Processing Approach for Analyzing Interdependence of Electric Power and Natural Gas Systems

Encouraged by the success of Graph Signal Processing (GSP) in power systems, researchers have begun exploring its application to other infrastructure networks, such as water distribution networks, which also exhibit graph-like structures but pose domain-specific challenges, such as nonlinear hydraulic behavior requiring customized GSP models. Similarly, the growing interdependence between power and gas systems, driven by gas-fired electricity generation and power-to-gas technologies, highlights the need for integrated modeling to ensure reliability and resilience. Motivated by these developments, this work extends GSP techniques to natural gas pipeline systems using circuit-theoretic analogies. Specifically, Weymouth's approximation of Euler's equations is used to represent quadratic pressure drops and volumetric flows, analogous to voltage and current in power networks. This forms a physically grounded foundation for building GSP frameworks that capture the dynamics of gas networks within the context of interconnected energy systems. Our experimental results show a low-pass nature in the gas pipelines thereby proposing a graph signal processing framework in analyzing interdependence of electric power and natural gas systems.