Monitoring and Optimization in Coupled Natural Gas and Electric Power Networks

Virginia Polytechnic Institute And State University, Blacksburg VA

Investigators

Abstract

The pressing need for efficient, low-carbon, and fast-responding electric power generators and decreasing prices of natural gas are currently transforming natural gas networks. The intermittent operation of gas-fired power plants to compensate wind generation introduces spatiotemporal fluctuations in continuously increasing volumes of gas demand. Generation and network contingencies in power transmission systems can affect shortages in gas supplies; and vice versa, congested and suboptimally scheduled gas networks limit the integration of renewable energy. In this context, this project studies the interdependency between these two critical infrastructures through coordinated control and market operations to ensure stability, reliability, and efficiency. The research project will develop algorithmic tools for the analysis, optimization, and monitoring of natural gas networks and their interplay with electric power grids. The vision is threefold: develop a reliable computational toolbox for enhanced gas network operations leveraging advances in convex optimization; identify analogies, key differences, and opportunities by jointly dispatching the two energy systems; and engage data analytics for transforming both gas transmission and distribution networks into smarter cyber-physical infrastructures. The consequent gains in efficiency and awareness will be reflected on a wide gamut of industrial, commercial, and residential uses of natural gas; while enabling a smooth transition to the low-emission economy. Broader transformative impact will result from an educational dissemination plan, involvement of undergraduates in research, and outreach to the local community and high school students. At the heart of modeling, control, and monitoring of gas networks is a set of nonlinear equations relating nodal gas injections and pressures to flows over pipelines. Building on convex relaxations, the gas flow problem is formulated as a semidefinite program that outperforms existing alternatives in terms of region of convergence and which naturally leads to coupled gas and power flow formulations. The novel scheme is generalized for optimally dispatching natural gas networks, independently or in tandem with electric power systems. To cope with uncertainty on renewable generation, and adjusting to the slower gas transients, static and dynamic setups are considered via decentralized and stochastic implementations. Efficient modules are additionally devised for enhancing situational awareness via gas network state estimation. By developing a suite of gas pressure and injection data analytics, computational efforts will aim towards enhancing local gas utility distribution networks, thus resonating with efforts for smart cities and connected communities. The utility of the proposed research goes well beyond the envisioned application area to the broader fields of stochastic optimization, statistical signal processing, control of coupled networks, inference over graphs, and dynamic modeling of fluids.