Scholarlink

<p>As global energy demand rises and carbon reduction targets intensify, natural gas is gaining prominence as a clean energy source. To balance natural gas supply and demand, reduce load volatility, and enhance system stability, this paper proposes a bi-level model combining price-based and incentive-based demand response (DR) strategies. The upper-level model uses dynamic gas pricing to guide users in adjusting consumption, thereby reducing peaks, filling valleys, and optimizing resource allocation. The lower-level model considers factors like weather and heating, creating incentives to boost user participation and flexibility. This model is solved using multi-population ensemble particle swarm optimization (MPEPSO) and Deep Q-Network (DQN) algorithms. Additionally, a spectral clustering algorithm is applied to classify load peak and valley times. For engineering applications, the model is validated using load data from a natural gas station in Xi’an, providing tailored DR strategies for various user types across heating and non-heating periods. The results demonstrate that the proposed strategy effectively smooths gas load fluctuations, alleviates supply-demand imbalances, secures supplier revenue, and maximizes user economic benefits, thereby enhancing the overall flexibility and applicability of DR.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Natural gas bi-level demand response strategies considering incentives and complexities under dynamic pricing

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月24日

查看全文
  • Huibin Zeng,
  • Jie Zhou,
  • Hongbin Dai

摘要

As global energy demand rises and carbon reduction targets intensify, natural gas is gaining prominence as a clean energy source. To balance natural gas supply and demand, reduce load volatility, and enhance system stability, this paper proposes a bi-level model combining price-based and incentive-based demand response (DR) strategies. The upper-level model uses dynamic gas pricing to guide users in adjusting consumption, thereby reducing peaks, filling valleys, and optimizing resource allocation. The lower-level model considers factors like weather and heating, creating incentives to boost user participation and flexibility. This model is solved using multi-population ensemble particle swarm optimization (MPEPSO) and Deep Q-Network (DQN) algorithms. Additionally, a spectral clustering algorithm is applied to classify load peak and valley times. For engineering applications, the model is validated using load data from a natural gas station in Xi’an, providing tailored DR strategies for various user types across heating and non-heating periods. The results demonstrate that the proposed strategy effectively smooths gas load fluctuations, alleviates supply-demand imbalances, secures supplier revenue, and maximizes user economic benefits, thereby enhancing the overall flexibility and applicability of DR.