多目标励磁控制的神经网络逆系统方法

根据逆系统理论分别分析了以机端电压和功角为被控量的励磁系统的可逆性，并采用不同的神经网络逆系统结构实现了对单目标量的励磁控制，针对单目标励磁控制不能够同时改善功角稳定性和机端电压稳定性的不足，在以机端电压为被控量的神经网络逆系统结构中引入功角偏差反馈信号，设计了多目标神经网络励磁控制器，研究结果表明，应用该系统能够实现多目标励磁控制，综合性能优于目标的神经网络逆系统控制方法和常规PSS控制方法。

MULTI-OBJECTIVE EXCITATION CONTROL BASED ON ANN INVERSE SYSTEM METHOD

This paper analyses the invertibility of excitation system considering two different objective variables--terminal voltage and power angle. By adopting two corresponding ANN inverse system structures, single objective excitation control is realized. In order to overcome the drawback of single objective excitation control (i.e. the stability of power angle and terminal voltage can not be improved simultaneously), the signal of power angle error is introduced into the ANN inverse structure (with terminal voltage as objective variable). Simulation results prove that the proposed control strategy can realize multi objective excitation control, and its comprehensive performance is better than that of single objective ANN inverse system excitation control method and standard PSS method.