Safe Deep Reinforcement Learning for Hydrogen-Blended Integrated Electricity-Gas Systems with Wind Power and Stochastic EV Load
水素混入統合電力・ガスシステムにおける安全な深層強化学習：風力発電と確率的EV負荷を考慮して (AI 翻訳)
Lixin Tan, Xian Zhang, Yidian Zhu, A. Sayed, Mahmoud Mohamed Sayed Mohamed Hemdan
🤖 gxceed AI 要約
日本語
本研究は、水素混入統合電力ガスシステムのオンライン運用決定のための安全な深層強化学習手法を開発した。制約付きマルコフ決定過程として定式化し、制約付きソフトアクター・クリティックアルゴリズムを採用、物理ガイド環境下で電力潮流、ガス動態、水素混入率、貯蔵制約を考慮する。数値実験では、提案手法が最適化ベースのベンチマークと比較して高い実現可能性と計算時間の大幅短縮を達成。20%の水素混入率が運用コストとCO2排出の最良のトレードオフを示した。
English
This study develops a safe deep reinforcement learning (DRL) scheme for online operational decision-making in hydrogen-blended integrated electricity-gas systems with wind power and stochastic EV loads. The problem is formulated as a constrained Markov decision process and solved with a constrained soft actor-critic algorithm in a physics-guided environment. Numerical results show high feasibility and significantly reduced computation time compared to optimization-based benchmarks. A 20% hydrogen blending ratio provides the best trade-off between operating cost and carbon emissions.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本では水素基本戦略に基づき水素混焼が注目される。本手法はリアルタイム運用の効率性を示し、SSBJやTCFD関連の情報開示にも応用可能性がある（運用データに基づく排出量管理）。
In the global GX context
This work is relevant to global efforts in decarbonizing gas grids and integrating variable renewables and EVs. The safe DRL approach offers a scalable solution for real-time operation, with implications for emissions reductions and system flexibility.
👥 読者別の含意
🔬研究者:A benchmark for safe DRL in multi-energy systems; demonstrates the trade-off between cost and carbon with hydrogen blending.
🏢実務担当者:Operators can adopt the DRL-based controller to reduce computational burden and maintain feasibility in hydrogen-blended systems.
🏛政策担当者:Provides quantitative evidence that moderate blending (20%) yields optimal environmental and economic outcomes, supporting policy for hydrogen injection limits.
📄 Abstract（原文）
This study develops a safety-aware deep reinforcement learning (DRL) scheme for online operational decision-making in hydrogen-blended integrated electricity-gas systems (HB-IEGS) with wind power and stochastic electric vehicle (EV) loads. The problem is converted to a constrained Markov decision process (CMDP), and a constrained soft actor-critic (C-SAC) algorithm is adopted within a physics-guided environment that captures power flow, gas dynamics, hydrogen blending, and storage constraints. Numerical results on a 14-bus/8-node system show that the proposed method achieves high feasibility and significantly reduces online computation time compared with optimization-based benchmarks. Moreover, a moderate hydrogen blending ratio is found to provide the best trade-off between operating cost and carbon emissions, with 20% blending yielding the most favorable performance.
🔗 Provenance — このレコードを発見したソース
- semanticscholar https://doi.org/10.1109/epsic70071.2026.11590187first seen 2026-07-13 06:39:18
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。