Adaptive Load Shedding Strategies for Interdependent Energy Networks under Temperature Rise

Abstract

Temperature rise and associated heatwaves, driven by climate change, create severe stresses on interdependent power and natural gas networks by simultaneously inflating electricity demand for cooling and constraining supply through thermal derating of overhead transmission lines, reduced efficiency of gas-fired generation, and limitations on gas pipeline operations. Gas fired plants, often critical for peaking during heat events, face cooling limitations and potential fuel delivery constraints, while electric compressors in gas networks may lose power support, creating bidirectional cascading risks. When preventive measures demand response, energy storage discharge, interregional imports, and redispatch are exhausted, adaptive load shedding becomes essential to maintain frequency and voltage stability while minimizing total unserved energy, economic losses, and societal harm. Traditional under-frequency or under-voltage load shedding (UFLS/UVLS) schemes are often static or semi-adaptive and fail to account for temperature-dependent constraints, network interdependencies, or equity considerations, leading to disproportionate impacts on vulnerable communities in urban heat islands. This research paper develops adaptive load shedding strategies for coupled power-gas networks under rising temperatures, formulating multi-stage, optimization-based, and intelligence-driven approaches that incorporate real-time temperature and weather data, dynamic line ratings (DLR), vulnerability-weighted priorities using social vulnerability indices (SVI), and grid Gini coefficients for fairness. Models leverage mixed-integer nonlinear programming (MINLP), robust optimization against forecast uncertainty, and hybrid methods combining particle swarm optimization, reinforcement learning, or adaptive algorithms that dynamically adjust shedding amounts and locations based on measured imbalances, DER ramping capabilities, and inter network couplings.