基于多智能体的分布式发电系统协调优化

本文采用评价体系的动态合同协议, 解决分布式发电系统的运行优化问题. 根据分布式发电系统的结构特点, 建立了一种分布协调的Multi-Agent体系. 将各基本可再生发电单元作为电源Agent, 能量管理系统看作是管理协调Agent. 同一地域内的电源Agent组成一个Agent域, 由管理协调Agent对域内各电源Agent电源登记管理. 各Agent域聚集成一个分布式的网络结构. 在此网络结构中, 采用评价体系的动态合同网协议进行协调发电以简化投标过程, 以减少通信量, 提高系统的优化效率. 管理协调Agent作为招标方, 各电源机组作为投标方并按照偷懒原则进行投标. 最后对一个小型的分布式发电系统进行仿真实验, 验证了所提出方案的可行性, 并保证了分布式发电系统的电能质量和可靠供电.

Coordinated optimization of distributed hybrid generation system based on multi-agent system

A novel dynamic contract-net-protocol algorithm based on the evaluation system is proposed for the coordinated optimization of hybrid generation system. A distributed coordinated multi-agent system (MAS) is built based on the characteristics of the distributed hybrid generation system. Each basic renewable generation unit is viewed as an electrical source agent (ESA). The energy management system is viewed as the management coordinated agent (MCA). All ESA in an area constitute an agent domain registered and controlled by the MCA. All agent domains are gathered together to form a distributed network structure. In this structure, a novel dynamic contract-net protocol algorithm based on the evaluation system optimally coordinates the power generation, which simplifies the bidding procedures, reduces the amount of correspondences, and improves the optimization efficiency. In the bidding process, the MCA is the tenderee, the ESA is the bidder. Bidding is carried out based on the idleness principle according the local environment. The simulation on a small-scale hybrid renewable power system shows that the proposed optimization method realizes the stable operation of the whole generation system.