同步发电机励磁快速模型预测控制研究

Research on Fast Model Predictive Control of Synchronous Generator Excitation

介绍了一种新的控制器设计方法——在线优化快速预测控制方法,并将这种方法应用到同步发电机励磁控制中,以有功功率、角速度、机端电压和励磁电压的偏差量组合作为全局变量,以预测步数内系统状态轨迹和控制量的偏差最小为优化控制指标,并考虑以系统微分方程、状态和输入为约束条件。算法中对滚动优化模型进行简化,运用区块排除法、温启动等有效地降低了模型预测控制在线计算复杂度,可加快在线运算速度。对单机无穷大系统的仿真结果表明,该励磁控制方法优于一般励磁控制方法,控制效果好。此外,控制器的控制效果可由预测步数控制,灵活方便。

The paper introduces a new controller design method, online optimization fast predictive control method and applies this method in synchronous generator excitation control. Combined amount of dispersion of active power, angular velocity, generator terminal voltage and excitation voltage are taken for global variable, minimized error of system state traj- ectory and controlled variable in predicted steps is taken for optimized control index, and consideration of differential equation, state and input are taken as constraint condition. In the algorithm the scrolling optimization model is simplified and complexity of online calculation on model prediction control is decreased by block elimination method and warm start for the purpose of speeding online calculation up. Simulation on single machine infinite system shows that the excitation control method is better than any ordinary methods in controlling; furthermore, the controller can be determined by prediction steps, flexible and convenient.