摘要
针对多机电力系统,提出了一种基于辨识的神经网络实时最优控制器(NNOEC),在所设计的控制器中,神经网络被用来根据系统状态量的变化实时调整最优控制的反馈增益矩阵,使控制器能够适应不同的运行点和干扰种类。并始终提供最优控制输出。针对多机系统中神经网络训练样本不易获得的问题,提出了一种等效的设计方法,并采用非线性最小二乘辨识法对系统参数进行辨识,在辨识的基础上通过线性最优控制理论计算出用于神经网络训练的样本。三机系统中的数字仿真结果表明,所训练出的NNOEC能够适应系统运行方式的大范围变化,在大小扰动下均表现出良好的控制性能。
A neural network optimal excitation controller based on system identification (NNOEC) for multi-machine power system is proposed in this paper. In this NNOEC, a neural network is used to adjust the optimal feedback gains according to the state variables of the generator in real-time. So the controller can be adapted to the changed operating conditions of the power system and the optimal control is always given. To simplify the design of the controller in multi-machine system, a non-linear least square identification method is proposed to identify the parameters of the system. Thus the training samples for the neural network can be calculated from the identification results and the disturbances in the system. The simulation results in a three-machine power system are show that the designed NNOEC can provide good control performance under various operating points and different disturbances.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2004年第7期80-84,共5页
Proceedings of the CSEE
关键词
多机电力系统
神经网络
最优励磁控制器
电力系统稳定器
Electric power engineering
Multi-machine power system
System identification
Neural network
Optimal control
Excitation control