一种新颖的智能发电控制策略

A Novel Policy for Smart Generation Control

本文提出了一种新颖的智能发电控制策略以解决新能源接入所带来强随机电网环境下的互联电网自动发电控制。各区域电网包含电网实时状态输入智能体、分布式协调算法智能体和控制输出智能体。分布式相关均衡Q(λ)学习被集成到控制智能体中作为主要的协调控制方法。此外,本文还详细讨论了JADE技术智能体的通信、行为和动作的作用机理。该平台已通过南方电网四省实时频率偏差模型证实能实现AGC区域的协调控制,尤其是当某区域电网控制误差ACE不能完全消除时。多智能体控制模块可以嵌入不同的控制算法并能评估其控制性能,这表明该平台具有很好的可扩展性、灵活性和高效性。

A novel policy is proposed in this paper to deal with smart generation control under strong random grid environment considering emerging renewable energy sources. Every control area contains input grid status agents, synergetic algorithm implementation agents, and output controller agents. Decentralized correlated equilibrium Q(λ) learning was integrated in control agent as the main synergetic control method. Besides, the mechanism of agent communication, behavior and action of this platform are discussed in details. It has been proved that this platform applied to China Southern Power Grid could achieve coordination control between automatic generation control areas, especially when one of the areas could not drive its area control error to zero. Different control algorithms can be embedded into control agent and its control performance can be evaluated, which demonstrates the effectiveness, scalability and flexibility of this platform.