基于ADMM算法的多主体综合能源系统分布式协同优化研究

针对多园区综合能源系统的分布式调度问题,本文构造了多园区综合能源系统的多主体架构,提出了一种分布式经济调度的基本框架,建立了多决策主体的集中式优化调度模型；利用交替方向乘子法(Alternating Direction Method of Multipliers,ADMM),引入功率协调变量解耦园区间互济功率,将集中式优化调度转换为具有一致性的各园区分散自治子优化问题,构建了基于ADMM分布式调度模型与算法,实现了与集中式优化调度几乎一致的优化结果,同时克服了集中式方法信息隐私安全性低、通信成本高、建模复杂、求解难度大等缺陷。最后,分别以简单和复杂的多园区综合能源系统为算例,验证了ADMM算法的计算性能及其相较于集中式方法的特点和优势。

Research on distributed Cooperative Optimization of multi-agent integrated energy System based on ADMM Algorithm

Aiming at the distributed scheduling problem of multi-park integrated energy system, this paper constructs the multi-agent architecture of multi-park integrated energy system, proposes a basic framework of distributed economic scheduling, and establishes a centralized optimal scheduling model of multi-decision agents. Alternating Direction multiplier Method is used to introduce the power coordination variable to decouple the mutual aid power between parks, The distributed scheduling model and algorithm based on ADMM were constructed to achieve optimization results almost consistent with that of the centralized optimization scheduling. Meanwhile, the defects of the centralized method, such as low information privacy security, high communication cost, complex modeling, and great difficulty in solving, were overcome.. Finally, taking simple and complex multi-park integrated energy system as examples, the computational performance of ADMM algorithm and its characteristics and advantages over centralized method are verified.