电力系统神经网络逆控制中的闭环控制器设计

由于实际电力系统输入限幅、工作区域受限等因素的影响,采用神经网络逆系统方法对被控电力系统进行反馈线性化解耦后的复合系统并非理想的解耦线性系统.在对此复合系统设计闭环控制器时需要确定合理的期望响应特性.本文通过对被控电力系统的输入输出特性分析,得到线性化解耦后各子系统的主导物理特性,从而为确定合理的闭环期望响应特性提供依据.结合神经网络逆系统方法在励磁与汽门系统线性化解耦控制中的应用,本文给出了详细的设计方法与计算机仿真结果.

Closed-loop controller design method in the ANN-inversion-based decoupling and linearizing control of power systems

As the system's controlling inputs and operating region are constrained in real power systems, the decoupled and linearized system derived by ANN inversion method is not an ideal one. So the anticipant system's responding characteristic should be appropriately determined when designing a closed loop controller for the decoupled system. By analyzing the input output responses of the controlled power system and the dominant physical characteristic of each decoupled subsystem, this paper presents a systematic method of how to determine reasonable closed loop anticipant responding characteristics and how to design the closed loop controller for each subsystem. As an example, the detailed design procedure of closed loop controller and simulation results for the excitation and valve control based on ANN inversion decoupling and linearizing method are given.