摘要
将多步预测自校正控制(Multi-step Predictive Self-Tuning Control)理论用于同步发电机励磁控制中来,选取多个状态量偏差的组合做为输出,用渐消记忆递推最小二乘法在线辨识同步发电机三阶非线性方程经线性化和离散化后的模型参数,并采用一种简单有效的方法离线确定b0(b0是前面提到的模型参数之一)。针对单机无穷大系统,设计出多步预测自校正励磁调节器(Multi-step Predictive Self-Tuning Excitation Regulator,MPSTER)。数字仿真结果表明,该励磁调节器控制效果好,能及时跟踪发电机运行工况的变化,在系统遭受的各种干扰下均表现出良好的动、静态性能。
The Multi-step Predictive Self-Tuning Control theory is applied in excitation control of synchronous generator. We select the combination of several deviations of state as output and estimate the model parameters on line, which generated after the linearization and discretization of third-order synchronous generator nonlinear equation, by Fading Memory Recursive Least Square Algorithm, At the same time, we adopt a simple but effective method for offline estimation of b1, which is one of the model parameters as mentioned above. For a single-machine infinite-bus power system, a Multi-step Predictive Self-Tuning Excitation R.egulator （MPSTER） is designed. The simulation result shows that the control effect of MPSTER is very good. Meanwhile, the regulator can track the variation of system＇s operation conditions timely, and has good dynamic and static performance under conditions of various disturbances. This work is supported by National Natural Science Foundation of China （No. 50747026）.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2011年第17期120-125,129,共7页
Power System Protection and Control
基金
国家自然科学基金资助项目(50747026)
广西科学基金资助项目(桂科自0728027)
南宁市市校科技合作专项项目(200801029D)
北海市市校科技合作专项项目(北科合200801027)~~
关键词
发电机励磁控制
多步预测自校正控制
递推最小二乘法
参数辨识
generator excitation control
multi-step predictive self-tuning control
recursive least square algorithm
parameter estimation
