IEEE Access (Jan 2024)

Decentralized Privacy-Preserving Distributionally Robust Restoration of Electricity/Natural-Gas Systems Considering Coordination of Pump Storage Hydropower and Wind Farms

  • N. Nasiri,
  • S. Zeynali,
  • S. Najafi Ravadanegh,
  • S. Kubler,
  • Y. Le Traon

DOI
https://doi.org/10.1109/ACCESS.2024.3354891
Journal volume & issue
Vol. 12
pp. 13747 – 13762

Abstract

Read online

A swift power system restoration in a post-blackout event is one the most important challenges faced by the transmission system (TS) operators (TSO), which is particularly essential in the presence of wind farms, as their potential can be great in a fast restoration. In this study, we propose a bi-level decentralized approach to examine the influence of natural gas network (NGN) constraints on the bulk power system restoration process, taking into account the concurrent effects of pump storage hydropowers (PSHs) and wind farms. At the upper level of the problem, the transmission system (TS) operator (TSO) submits the amount of natural gas fuel consumed by the gas-fired units (GFUs) to the NGN by observing the restoration, operational, and topological constraints. The objective of the TSO is to maximize load servicing in the power grid restoration process. The equilibrium of the proposed bi-level problem is calculated by the analytical target cascading (ACT) algorithm, preserving the privacy of both electricity and NGNs. In the proposed study, an investigation has been conducted into the impact of the gas storage system (GSS) and linepack technology on enhancing the restoration process. Moreover, a moment-based distributionally robust optimization (DRO) approach has been deployed to model the uncertain behavior of wind farms in the restoration process. The proposed approach comprehensively examines the effects of the decentralized interconnection between electricity and NGNs in the restoration process. This facet holds great significance for the advancement of future sustainable energy systems. The results show that ignoring the NGN model leads to 11.92% higher level of unserviced loads.

Keywords