基于虚拟狼群策略的分层分布式自动发电控制

Hierarchical and Distributed Control Method for Automatic Generation Control Based on Virtual Wolf Pack Strategy

为解决传统集中式自动发电控制(AGC)无法解决的大规模新能源接入电网所带来的强随机扰动问题,从AGC角度提出一种面向分布式能源、基于虚拟狼群策略的分层分布式控制方法,以实现区域的最优协同控制。该方法简称虚拟狼群策略,它结合新赢输评判标准、爬山算法和资格迹,通过智能体精确计算赢输评判标准并快速收敛到纳什均衡;在总功率指令分配部分采用协同一致性算法。通过对IEEE标准两区域电力系统改进模型以及湖北电网模型进行仿真,表明所提方法能实现区域协同控制、改善闭环系统性能、减少碳排放,同时具有更快的收敛速度。

To solve stochastic disturbance problems caused by massive integration of new energy to the power grids that cannot be solved by traditional automatic generation control（AGC）,a virtual-wolf-pack-strategy based hierarchical and distributed control（VWPS-HDC）method is proposed to obtain the optimal collaborative control of the region,which orients to distributed energy.This method is called virtual wolf pack strategy for short.In this method,the fast convergence to Nash equilibrium and the accurate calculation for agents are obtained based on the new win-lose judgment criterion,policy hill-climbing,and eligibility trace.Moreover,the active power is distributed by a collaborative consensus algorithm.The simulation conducted on the modified load frequency control model of IEEE two-area power system and that of Hubei power grid model in China demonstrates that the proposed method can obtain the regional optimal collaborative control,improve the closed-loop system performances,reduce the carbon emission,as well as achieve a faster convergence speed.