(373ar) Dynamic Optimization with Intertemporal Uncertainty in Natural Gas Pipeline Operations Using Stochastic Finite Volume

Roughly 25% of worldwide electricity generation is fueled by natural gas, which is viewed as playing a major role in the transition to net-zero energy systems [1]. Natural gas-fired generators emit far less carbon dioxide and other pollutants than coal power plants and are much easier to site, permit, and build than nuclear generation stations [2]. In addition to combined-cycle plants that provide base load, gas-fired peaking plants can quickly ramp up output to compensate for shortfalls in production by variable and intermittent renewable sources. This has resulted in a reliance of the power grid on gas-fired generation that is fueled by just-in-time delivery of natural gas through pipeline systems.

A standard approach to uncertainty management for this problem is to use stochastic optimization, which results in well-known issues of poor computational scaling arising due to sampling of parameters in the uncertainty region. In this study, we utilize a recently developed stochastic finite volume representation for optimization of the nonlinear gas flow equations with uncertain boundary conditions to manage intertemporal uncertainties for a pipeline system [3]. Intertemporal uncertainty in this context is used to denote a temporary increase in load starting at a time that is randomly distributed (e.g. uniformly on an interval). This represents the activation of a peaking plant at an a priori unknown time in the day ahead planning interval. In conjunction with the proven method of chance-constraints, this time dependent uncertainty management scheme provides robustness guarantees with a minimum of extra reserve requirements.

References:

[1] IEA, “Gas,” https://www.iea.org/fuels-and-technologies/gas, 2023, [Online; accessed 27-June-2023]

[2] J. Bistline, “Natural gas, uncertainty, and climate policy in the US electric power sector,” Energy Policy, vol. 74, pp. 433–442, November 2014.

[3] Tokareva, S., Zlotnik, A., & Gyrya, V. (2024). Stochastic finite volume method for uncertainty quantification of transient flow in gas pipeline networks. Applied Mathematical Modelling, 125, 66-84.