基于改进近似动态规划的安全约束机组组合算法  

A Security-Constrained Unit Commitment Algorithm Based on Improved Approximate Dynamic Programming

作　　者：曾恺 朱建全[1] ZENG Kai;ZHU Jian-quan(School of Electric Power Engireering,South China University of Technology,Guangzhou 510640,China)

机构地区：[1]华南理工大学电力学院,广东广州510640

出　　处：《电工电气》2024年第2期1-7,共7页Electrotechnics Electric

基　　金：国家自然科学基金项目(51977081);广东省自然科学基金项目(2022A1515011193);直流输电技术国家重点实验室开放基金项目(SKLHVDC-2021-KF-09)。

摘　　要：针对电力系统的安全约束机组组合问题,提出了一种基于改进近似动态规划的求解算法。考虑到安全约束机组组合是一个多时段混合整数非线性规划问题,难以直接求解,所提算法将其建模成马尔可夫决策过程,并利用近似动态规划实现解耦求解。为了处理近似动态规划决策空间过大的问题,所提算法通过决策空间缩减技术提前获得大部分机组的开停机状态,有效提高求解效率。在IEEE 39节点系统和一个实际系统上进行了算例分析,验证了所提算法的有效性。In this paper,an improved approximate dynamic programming algorithm is proposed for the security-constrained unit commitment in power systems.First,considering security-constrained unit commitment is a multi-period mixed integer non-linear programming problem,which is difficult to solve directly,the proposed algorithm models it into a Markov decision process,and uses approximate dynamic programming to achieve decoupling.Then,in order to deal with the problem of excessive decision-making space for approximate dynamic programming,the proposed algorithm obtains the start-up and shut-down states of most units in advance through the decision space reduction technique,which effectively improves the solution efficiency.Finally,case studies are conducted on the IEEE 39-bus system and a practical system to validate the effectiveness of the proposed algorithm.

关 键 词：安全约束机组组合 混合整数非线性规划 近似动态规划 决策空间缩减技术 

分 类 号：TM715[电气工程—电力系统及自动化] TM76