魏大钧, 孙树敏, 程艳, 管荑, 郝然, 艾芊. 基于库存理论考虑局部不确定性的互联微电网分布式鲁棒调度[J]. 现代电力, 2020, 37(1): 19-26. DOI: 10.19725/j.cnki.1007-2322.2019.0121
引用本文: 魏大钧, 孙树敏, 程艳, 管荑, 郝然, 艾芊. 基于库存理论考虑局部不确定性的互联微电网分布式鲁棒调度[J]. 现代电力, 2020, 37(1): 19-26. DOI: 10.19725/j.cnki.1007-2322.2019.0121
WEI Dajun, SUN Shumin, CHENG Yan, GUAN Ti, HAO Ran, AI Qian. Distributed Robust Dispatch for Interconnected Microgrids Considering Local Uncertainty based on Inventory Theory[J]. Modern Electric Power, 2020, 37(1): 19-26. DOI: 10.19725/j.cnki.1007-2322.2019.0121
Citation: WEI Dajun, SUN Shumin, CHENG Yan, GUAN Ti, HAO Ran, AI Qian. Distributed Robust Dispatch for Interconnected Microgrids Considering Local Uncertainty based on Inventory Theory[J]. Modern Electric Power, 2020, 37(1): 19-26. DOI: 10.19725/j.cnki.1007-2322.2019.0121

基于库存理论考虑局部不确定性的互联微电网分布式鲁棒调度

Distributed Robust Dispatch for Interconnected Microgrids Considering Local Uncertainty based on Inventory Theory

  • 摘要: 为实现广域分布式能源系统的高度自治性和鲁棒性,提出了一种用于互联微电网(interconnected microgrids, IMG)的分布式鲁棒经济调度(distributed robust economic dispatch, DRED)。首先,应用多智能体理论搭建IMG的分层分布式调度框架;其次,为了解决分布式鲁棒优化问题,将原问题分解为多层问题,在上两层进行可行性测试和联络线功率优化。在此研究结果的基础上,运用库存理论对可再生能源和负荷的不确定性进行了研究;最后,以一个实际的IMG系统为例,通过对分散和集中式鲁棒方案的比较,验证了该策略的收敛性和可行性。该模型在计算复杂度上也优于传统的DRED模型,利用概率信息实现了某一置信水平下的最优调度。

     

    Abstract: A distributed robust economic dispatch (DRED) for interconnected microgrids (IMG) is proposed to realize high autonomy and robustness of wide-area distributed energy system. Firstly, the framework for the distributed scheduling of IMGs is built based on multi-agent theory. Secondly, in order to solve robust optimization (RO) in a distributed manner, the original problem is decomposed into a hierarchical problem, where feasible test and tie-line power optimization are carried out in upper two layers. Based on the research results, the uncertainties of renewable energy and load are considered in local agents via inventory theory. Finally, taking an actual IMG system as an example, numerical simulation results verify the good astringency and feasibility of the proposed strategy by comparing decentralized and centralized robust schemes. The proposed model outperforms the conventional DRED in computation complexity by exploiting probability to achieve optimal dispatch under a certain confidence level.

     

/

返回文章
返回