基于多步回溯Q(λ)的PSS最优控制方法的研究  

Optimal control method of PSS based on multi-step backtrack Q(λ) learning

作　　者：余涛[1] 甄卫国[1] 

机构地区：[1]华南理工大学电力学院,广东广州510640

出　　处：《电力系统保护与控制》2011年第3期18-23,28,共7页Power System Protection and Control

基　　金：国家自然科学基金项目(50807016);广东省自然科学基金项目(9151064101000049);中央高校基本科研业务费专项资金资助~~

摘　　要：电力系统稳定器(PSS)是用来产生能抑制低频电力系统振荡的励磁系统辅助控制信号,具备自学习和参数在线整定能力是未来智能电网PSS控制器的一个发展趋势。提出一种基于多步回溯Q(λ)学习的新颖电力系统稳定器设计方法。利用多步回溯Q(λ)控制器代替整个传统PSS作为励磁附加控制,并与传统PSS和Q学习控制器进行比较。仿真研究显示,引入基于多步回溯Q(λ)学习的PSS控制后显著增强了整个系统的鲁棒性,有效提高了系统抑制低频电力系统振荡的能力,较好地解决了Q学习控制器收敛速度慢的问题。Power system stabilizers（PSS） are used to generatesupplementary control signals for the excitation system in order to damp the low frequency power system oscillations.With the development of smart grids,the multiply PSS controllers with the abilities of self-learning and self-tuning become the attractive trend. Anovel control method of power system stabilizer（PSS） based on multi-step backtrack Q（λ） learning is proposed in this paper.The multi-step backtrack Q（λ） controller is used to replace the conventional PSS to generate supplementary control signals for the excitation system,and is compared with the conventional PSS and Q-learning controller.Results show that the Q（λ） controllerstrengthens the robustness of the power system and enhances the ability of damping the low frequency power system oscillations. Besides,it can solve the problem of the slowconvergence rate of Q-learning controller.

关 键 词：电力系统稳定器(PSS) 马尔可夫策略(MDP) 强化学习 Q学习 多步回溯Q(λ)学习 

分 类 号：TM44[电气工程—电器]