Scholarlink

Microgrid loads, represented by distributed renewable energy sources, are naturally characterized by intermittency, fluctuation, randomness. Their large-scale integration on the demand side poses significant challenges to the safe, reliable, and economical operation of distribution grids. Meanwhile, the existing centralized optimization in the collaborative operation of integrated energy microgrid clusters faces the issues of privacy protection and the inability to share parameters. In response, this paper presents a two-stage power distribution system (PDS) optimization based on the encapsulation of microgrid demand response characteristics using deep learning. The encapsulation enables the direct mapping between distribution grid prices, meteorological data and the trading power at the points of common coupling (PCCs). Then, a two-stage optimization model for PDS is constructed. The first stage focuses on economic dispatch, while the second stage addresses reactive power optimization to improve voltage quality. The model solution is achieved through second-order cone programming. Simulations on a modified IEEE 33-node distribution system, which includes microgrids, validate the proposed encapsulation and optimization models. And the results confirm the effectiveness of these models.

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

Two-Stage Operational Decision Making for Active Distribution Network Based on Microgrid Demand Response Encapsulation

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jiabei Ge,
  • Zehao Song,
  • Dongqi Wu,
  • Junyi Li,
  • Yong Yan,
  • Zhiyi Li

摘要

Microgrid loads, represented by distributed renewable energy sources, are naturally characterized by intermittency, fluctuation, randomness. Their large-scale integration on the demand side poses significant challenges to the safe, reliable, and economical operation of distribution grids. Meanwhile, the existing centralized optimization in the collaborative operation of integrated energy microgrid clusters faces the issues of privacy protection and the inability to share parameters. In response, this paper presents a two-stage power distribution system (PDS) optimization based on the encapsulation of microgrid demand response characteristics using deep learning. The encapsulation enables the direct mapping between distribution grid prices, meteorological data and the trading power at the points of common coupling (PCCs). Then, a two-stage optimization model for PDS is constructed. The first stage focuses on economic dispatch, while the second stage addresses reactive power optimization to improve voltage quality. The model solution is achieved through second-order cone programming. Simulations on a modified IEEE 33-node distribution system, which includes microgrids, validate the proposed encapsulation and optimization models. And the results confirm the effectiveness of these models.