中文摘要：

提出一种基于智能多代理技术的能量协调控制方法，用于直流微网的能量管理与电压控制。对直流微网的架构进行设计．实现对光伏电池、燃料电池及蓄电池的数学建模。设计了一个2层的智能多代理系统，对直流微网的并网、解列．以及光伏电池、燃料电池、负荷及蓄电池的能量管理进行协调控制。从微源和负荷各种状态中提取了8个特征量及13种不同动作作为神经网络的输入和输出参数，训练并实现了控制中心Agent的决策器。通过MATLAB／Simulink仿真对光照变化、并网时负荷增加后断网及孤岛时负荷增加后并网这3种算例进行分析．验证了该智能多代理能量协调控制策略的可行性．

英文摘要：

An energy coordination control based on intelligent MAb（Mulh Agent bystem） IS proposed tor the energy management and voltage control of DC microgrid. The framework of DC microgrld is designed and the mathematic models are established for PV,fuel cell and battery. A two-layer MAS is designed for the coordination control：the grid-connection and islanding of DC microgrid and the energy management of PV cells,fuel cells,loads and battery. 8 characteristic variables and 13 operations are extracted from various conditions of micro-sources and loads as ,respectively,the input and output parameters of a neural network, which is then trained to obtain the decision maker of central Agent. The proposed strategy of energy coordination control is verified by simulation with MATLAB/Simulink for three operating conditions：variable light intensity,islanding after load increase and grid-connection after load increase.